The Extractive Metallurgy of Gold in South Afri (B-Ok - Xyz)

THE EXTRACTIVE METALLURGY
OF GOLD IN SOUTH AFRICA
VOLUME 2
SAIMM PUBLICATIONS
THE MONOGRAPH SERIES
Ml Lognormal-De Wijsian Geostatistics for Ore Evaluation

(2nd ed 1981)
D.G. Krige
M2 An Introduction to Geostatistical Methods of Mineral
Evaluation (2nd ed 1981)
J.-M.M. Rendu
M3 Principles of Flotation (1982)
Edited by R.P. King
M4 Increased Underground Extraction of Coal (1982)
Edited by C.J. Fauconnier and R.W.O. Kersten
MS Rock Mechanics in Mining Practice (1983) (3rd imp. 1986)
Edited by S. Budavari
M6 Assay and Analytical Practice in the South African Mining
Industry (1986)
W.C. Lenahan and R. de L. Murray-Smith
M7 The Extractive Metallurgy of Gold in South Africa,
2 volumes (1987)
Edited by G. G Stanley
THE SPECIAL PUBUCATIONS SERIES
SPl Proceedings, Underground Transport Symposium (1986)

Edited by R.C.R. Edgar
SAIMM PUBLICATIONS
THE SYMPOSIUM SERIES
S1 Mathematical Statistics and Computer Applications in Ore

Valuation (1966)
S2 Planning Open Pit Mines (1970) (4th imp.)
Edited by P.W.J. van Rensburg
S3 Application of Computer Methods in the Mineral Industry
(APCOM 1973)
Edited by M.D.G. Salamon
S4 Infacon 1974
Edited by H. W. Glen
S5 Proceedings of the 12th CMMI Congress, 2 volumes (1982)
Edited by H. W. Glen
S6 Rockbursts and Seismicity in Mines (1984)
Edited by N.C. Gay and E.H. Wainwright
S7 The Planning and Operation of Open Pit and Strip Mines
(1986)
Edited by J.P. Deetlefs
SS GOLD 100: Proceedings of the International Conference on
Gold (1986)
Volume 1: Gold Mining Technology
Edited by H. Wagner and R.P. King
Volume 2: Extractive Metallurgy of Gold
Edited by C.E. Fivaz and R.P. King
Volume 3: Industrial Uses of Gold
Edited by G. Gafner and R.P. King
S9 APCOM 87: Proceedings of the Twentieth International
Symposium on the Application of Computers and
Mathematics in the Mineral Industries (1987)
Volume 1: Mining
Edited by L. Wade, R.W.O. Kersten and J.R. Cutland
Volume 2: Metallurgy
Edited by R.P. King and I.J. Barker
Volume 3: Geostatistics
Edited by I.C. Lemmer, H. Schaum and F.A.G.M. Cami-
sani-Calzolari
The South African Institute of Mining and Metallurgy
Monograph Series M7
The Extractive
f 1 in S
Edited by
G. G. Stanley
VOLUME 2
This book was commissioned by the Sub-Committee of Consulting

Metallurgists of the Chamber of Mines of South Africa
Johannesburg
1987
Published by The South African Institute
of Mining and Metallurgy
Kelvin House, 2 Hollard Street, Johannesburg,
2001
©The Chamber of Mines of South Africa, 1987

P.O. Box 809, Johannesburg, 2000
ISBN 0 620 10944 0 Volume 2

0 620 10109 1 Volume 1
0 620 10945 9 The Set
Typeset by Scientific and Technical Typesetting

Kelvin House, 2 Hollard Street, Johannesburg 2001
Printed in the Republic of South Africa

by National Book Printers, Goodwood
Outline of Contents
VOLUME 1
page
Preface .......................................................................... . XXXl
Acknowledgements ........................................................... . xxxiii
Foreword by President, Chamber of Mines of South Africa ..... . xxxvii
Chapter 1 Geology and Mineralogy of the Principal Goldfields

in South Africa ............................................. . 1
Chapter 2 Sorting and Crushing ...................................... . 69
Chapter 3 Milling and Classification ................................ . 121
Chapter 4 Gravity Concentration .................................... . 219
Chapter 5 Flotation ...................................................... . 235
Chapter 6 Cyanidation ................................................... . 277
Chapter 7 Recovery of Gold from Solution by Cementation .. 331
Chapter 8 Pyrometallurgy of Gold .................................. . 345
Chapter 9 Application of Activated Carbon to Gold Recovery 379
VOLUME 2
Chapter 10 Refining of Gold at the Rand Refinery .............. . 615
Chapter 11 Disposal of Residues ...................................... . 655
Chapter 12 Retreatment of Residues and Waste Rock ......... .. 707
Chapter 13 Process Evaluation ......................................... . 745
Chapter 14 Process Control ............................................. . 793
Chapter 15 The Chemistry of the Extraction of Gold ........... . 831
Chapter 16 Plant Design and Commissioning ...................... . 907
Chapter 17 Loss Control and Safety Management ............... . 971
Chapter 18 Laboratories and Pilot Plants .......................... .. 1013
Tables and Formulae . .. .. .. .. . .. .. .. . .. .. . . . . .. .. .. .. .. . . .. .. . . . .. .. . . . .. .. .. 1053
Index.............................................................................. 1073
Vll
Contents
VOLUME 2
page
Chapter JO Refining of Gold at the Rand Refinery, by
K.G. Fisher . . . ........ ..... .. . ... . . .. . . . . . . .. . .. . . .. . .. . 615
10.l Introduction .......................................................... 615
10.2 Historical Development .. ..... ... .. .. . .. .... . . .. . .. . . .. .. .. .. .. .. . . 615
10.3 Function ............................................................... 616
10.4 Metallurgical Operations .. .. . . . . ......... .. .. . .. . .. . .. .. . . .. . . . .. . 618
10.4.1 Summary description .................................. 618
10.4.2 Gold Refining Branch .. . . ...... ... . . . . . . . .. . . .. . . . .. . . 618
10.4.2.1 Flow sheet 618; 10.4.2.2 Melting and
sampling 619; 10.4.2.3 Refining 621; 10.4.2.4
De-golding 626; 10.4.2.5 Electrolytic gold re-
fining 626; 10.4.2.6 Fume recovery and sweep
treatment 628
10.4.3 Silver Refining Branch ... . . ... . . ... . .. . . . ... ... . . . .. .. 630
10.4.3 .1 Flow sheet 630; 10.4.3 .2 Leaching and
reduction630; 10.4.3.3 Electrolytic silver refining
633
10.4.4 Coin blank production .. ........ .. . . . . .. .. . .. . .. . . . .. . 637
10.4.4.1 Historical 637; 10.4.4.2 Process de-
scription 638
10.4.5 Smelter ...... ... . .. . ... . . . . . . .. .. . .. . . . . . . . . . . . .. . . ... . .. . . 641
10.4.5.l Flow sheet 641; 10.4.5.2 Sampling 641;
10.4.5.3 Blending and sintering 643; 10.4.5.4
Blast furnace smelting 644; 10.4.5 .5 Cupellation
648; 10.4.5.6 Panfurnacesmelting649; 10.4.5.7
Fume collection 651
10.4.6 Assaying and chemical analysis . .. . . . . . . .. . .. .. . .. . . 652
10.5 Future Developments . .... . .. . . .. . .. . . .. .. . . . . .. . . . .. . . . .. . . . . . . .. . . 652
10.6 References ............................................................. 653
Chapter II Disposal of Residues, by G.I. MacPhail and

J.C. Wagner ............................................ 655
11.1 Aim and Scope .. .. .. .. .. .. .. .. .. .. .. .. .. . .. . . . . .. .. .. .. .. . .. .. .. .. .. 655
11.2 Pianning . .. . . . . .... . .. .. . . .. . .. . .. . . .. . . . . . . . .. . . . .. . . . .. . . . .. . . .. . . .. . . 656
ix
CONTENTS
Components of a residue disposal system ...... .

11.2.1 656
Sizing ..................................................... .
11.2.2 657
11.2.2.1 Depositional area requirements 657;
11.2.2.2 Allowances for services 659; 11.2.2.3
Return water system660; 11.2.2.4 Total area660
11.2.3 Site selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 660
11.2 .4 Site exploration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663
11.2.5 Pollution control considerations . . . . . . . . . . . . . . . . . . . . 665
11.3 Statutory Requirements .. . . .. . . .. . ... . .. . .. . ... ... . . ... . . . . . . . .. . . . 665
11.3 .1 Legal considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 665
11.3 .1.1 The Water Act 665; 11.3 .1.2 The Mines
and Works Act 666
11.3 .2 Licensing ................................................. . 666
11.3.3 Environmental impact assessment ................. . 666
11.3.3.1 Methods of assessment 667; 11.3.3.2
Remedial measures 667; 11.3 .3 .3 Monitoring 667
11.4 Design, Specification and Construction . . . . . . . . . . . . . . . . . . . . . . . 668
11.4 .1 Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668
11.4.1.1 General 668; 11.4.1.2 Method of for-
mation of tailings dam 669
11.4.2 Tailings delivery ....................................... . 675
11.4.2.1 Pulp concentration 675; 11.4.2.2
Pumping and piping systems 676; 11.4.2.3 Pump
and pipeline design 678; 11.4.2.4 Tailings distri-
bution 678
11.4.3 Return water management .......................... . 678
11.4.4 Stormwater management ............................ . 683
11.4.4.1 General 683; 11.4.4.2 Top surface of
dam 684; 11.4.4.3 Side slopes 684
11.4.5 Slope stability .......................................... . 685
11.4.5 .1 Influence of foundation soils 685;
11.4.5 .2 Influence of tailings product 685;
11.4.5.3 Influence of rate of rise 687; 11.4.5.4
Influence of phreatic surface 687; 11.4.5.5 In-
fluence of slope geometry 689; 11.4.5 .6 Stability
analysis 690
11.4.6 Repair and reinstatement of operational and aban-
doned impoundments .................................. 691
11.4.6.1 Common problems with old and aban-
doned dams 691; Remedial measures 693
11.5 Observation and Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 695
11.6 Closure Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 698
x
CONTENTS
11.6.1 General ................................................... . 698

11.6.2 Boundary fence ........................................ . 698
11.6.3 Access roadways ....................................... . 700
11.6.4 The underdrainage system ........................... . 700
11.6.5 The solution or effluent trench .................... . 701
11.6.6 The slimes delivery pipeline ........................ .. 701
11.6.7 Surface drainage ....................................... . 701
11.6.8 Ground level catchment paddocks ................ . 701
11.6.9 Berms or step-backs ................................. .. 701
11.6.10 The dam surface ....................................... . 703
11.6.11 Return water dam ...................................... . 706
11. 7 References . .. . . .. . . . . .. . . . . .. . . .. . . . . .. . . .. .. . .. . . .. . . .. . . . .. . . .. . . . .. . 706
Chapter 12 Retreatment of Residues and Waste Rock, by

D.W. Bosch ............................................ 707
12.1 Introduction .......................................................... 707
12.2 Origin of Residues .. . .. .. .. .. .. .. .. .. .. . .. .. .. . .. .. .. .. .. .. .. .. .. .. . 707
12.2.1 Sand dumps and slimes dams ....................... 707
12.2.2 Rock dumps .... . .... .. .... ..... .. .... . .... .... .. .. .. .... 708
12.3 Sampling, Surveying and Valuation ............................ 709
12.3.1 Sampling .................................................. 709
12.3.1.1 Grab sampling 709; 12.3.1.2 Groove or
channel sampling 709; 12.3.1.3 Drilling 710
12.3.2 Surveying ................................................. 711
12.3.2.1 Tacheometric surveys 711; 12.3.2.2
Electronic distance measuring 712; 12.3.2.3
Photogrammetry 712
12.4 Sand/Slime Retreatment Operations ........................... 715
12.4.1 Slime reclamation at Blyvooruitzicht Gold Mine 716
12.4.1.1 The bucket-wheel excavator 716; 12.4.1.2
Repulper feed conveyor 719; 12.4.1.3 Repulper
719
12.4.2 Slime reclamation at ERGO 719
12.4.2.1 Location of pump stations 720; 12.4.2.2
Reclamation operations 720; 12.4.2.3 Processing
at ERGO 723
12.4.3 Sand and slime reclamation at Rand Mines and
Milling . .. . . . .. . .. . . . .. . . . .. . . .. . . .. . .. . .. .. .. . .. . . .. . . . .. . . 724
12 .4. 3 .1 Process development for sand treatment
724; 12.4.3.2 Reclamation operations 726;
12.4.3.3 Plant operations 727
Xl
CONTENTS
The Anglo American Corporation Joint Operating

12.4.4
Metallurgical Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 731
12.4.4.1 Historical overview 732; 12.4.4.2 Feed
sources 733; 12.4.4.3 Flotation plants 733;
12.4.4.4 Treatment of flotation tailings 734;
12.4.4.5 Treatment of pyrite concentrates 735;
12.4.4.6 Treatment of calcine 735
12.5 Waste Rock Recovery and Treatment ......................... 736
12.5.1 Development waste ..................................... 736
12.5.2 Waste rejected from ore setting .................... 737
12.5 .3 Radiometric sorting of waste rock . . . . . . . . . . . .. .. . . 738
12.5.4 Waste rock reclamation at the Blyvooruitzicht
Gold Mine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 738
12.6 Heap Leaching of Dumps ........ ................................ 740
12.7 Bacterial Leaching of Dumps and Slimes Dams ............ 741
12. 7 .1 Theoretical aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 741
12.7.2 Dump leaching by bacteria .......................... 742
12.7.3 Slimes dam leaching by bacteria ................... 743
12.8 References . . . .. . . .. .. . . . . . . .. . . .. . . . . . . . . .. . .. . . . . . . . . . . . .. . . . . .. . ... .. 743
Chapter 13 Process Evaluation, by H.E. Bartlett and

D.M. Hawkins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 745
13 .1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 745
13.2Process Evaluation ............................................ .... .. 745
13.3Statistical Concepts ... .................... .......................... 746
13 .3 .1 Population, sample, increment and composite
samples .. . . .. . . .. . . .. .. . . .. . . . . . . .. . . . . . . . . . . . .. . . . .. . .. . . . 746
13.3.2 Accuracy and precision ............................... 746
13 .3 .3 Standard deviation, variance, coefficient of varia-
tion . .. . . . . .. . . . . . . .. . . . . . . . . . . .. . .. . . . . .. . . . .. . . . . . . .. . . . . . . 746
13.3.4 Properties of variances and standard deviations 747
13.3.5 Normal and log-normal distribution .............. 748
13.4 Sampling Theory .................................................... 751
13.4.l Calculation of the precision of sampling......... 751
13.4.2 Theoretical calculation of sampling precision using
Gy's formulae ...................................... ..... 751
13.4.3 Gy's '50-piece' experiment ........................... 753
13 .4.4 Precision of sampling using geostatistical methods 756
13.4.4.1 The semi-variogram 756; 13.4.4.2 Nug-
get effect 757; 13.4.4.3 Estimates of variance
from semi-variograms 757
13.5 Sampling Practice ................................................... 759
13.5.l Run-of-mine ore feed - stop-belt sampling .... 761
CONTENTS
13 .5 .1.1 Calculation of precision of manual stop-

belt sampling 762; 13 .5 .1.2 The effect of com-
positing increments into shift and daily samples
763; 13.5.1.3 Accuracy of the sampling proce-
dures 765; 13.5.1.4 Crushing of the sample be-
fore sub-division 765
13.5.2 Automatic sampling of run-of-mine ore ........ . 766
13 .5 .2.1 Testing for bias in automatic sampling
766; 13.5.2.2 Examples oft-tests 768
13.5.3 Crusher product sampling ........................... . 770
13.5.4 Sampling of pulps .................................... .. 771
13.5.4.1 Cross-stream cutters 771; 13.5.4.2
Rotary samplers 773; 13.5.4.3 Precision of
sampling with cross-stream cutters 774
13.5.5 Other sampling equipment of pulps .............. . 775
13.5.5.1 Pipe sampler 776; 13.5.5.2 Projector
sampler 777
13.5.6 Sampling of loaded carbon, slag, flue dust and
other secondary materials at Rand Refinery ... . 777
13.5.7 Sampling of bullion .................................. .. 778
13.5.8 Sampling of liquids .................................. .. 778
13.6 Sampling Preparation .............................................. 778
13 .6.1 Sample preparation equipment and practice .. .. 778
13.6.1.1 Pulps 778; 13.6.1.2 High grade mate-
rials 778; 13.6.1.3 Run-of-mine ore 780
13.6.2 Statistical methods to quantify bias in sub-
sampling . . . . .. . . . .. . . . . . .. . . .. . . .. . . . . .. .. . . . . . . . . .. . . . .. . 780
13.7 Mass Measurement ................................................. 780
13. 7 .1 Truck and train weigh bridges .. .. .. .. .. .. .. .. .. .. . .. 781
13.7.2 Electromechanical belt massmeters ................ 782
13.7.3 Nuclear belt massmeters .............................. 785
13. 7 .4 Moisture determination for mass measurement 785
13. 7 .5 Batch mass determination of pulps .. .. . .. .. .. .. .. . 786
13.7.6 Batch tanks on load cells............................. 786
13. 7. 7 Mass of pulps using flow and density meters .. . 786
13 .8 Chemical Analysis .. .. . .. .. .. .. . .. .. .. .. .. .. .. .. .. .. . .. . .. .. .. .. .. .. 787
13 .9 Metallurgical Accounting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 787
13.9.l Call system ............................................... 788
13 .9 .2 Check-in/ check-out system .. .. .. .. .. .. .. .. .. .. .. .. .. 789
13 .10 Further Developments .. .. .. . .. .. .. .. .. .. .. .. .. .. .. .. . .. .. .. .. .. .. . 790
13 .11 References . .. . . . . . . . . . . . . . . .. . . . . . . .. . . .. .. . . . . . . . . . . .. . . . . . . . . .. . .. . . . 791
Xlll
CONTENTS
Chapter 14 Process Control by I.M. MacLeod and H.E.

Bartlett . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 793
14.1 Introduction .......................................................... 793
14.2 Historical Background ............................................. 794
14.3 Principles of Continuous Process Control ................... 795
14.3.1 Control objectives ...................................... 795
14.3 .2 Classification of process variables .... .. ... .. .... .. 796
14.3.3 Control system structuring ........................... 797
14.3.4 The standard three-term industrial controller ... 800
14.4 Process Measurements .. .... .. .... .. .. . ...... .. .. .. .. .. .. . .. .. .. .. . 801
14.4.1 Introduction .............................................. 801
14.4.2 Terminology of instrumentation systems ......... 802
14.4.3 The primary sensing element ........................ 803
14.4.4 Accuracy and calibration ............................. 803
14.4.5 Basic physical measurements ........................ 804
14.4.5.1 Flow rate of solids 804; 14.4.5.2 Flow
rate of liquids 804; 14.4.5.3 Liquid level in a
vessel 806; 14.4.5.4 The water content of a pulp
807; 14.4.5.5 One-line particle size measurement
807; 14.4.5 .6 Basic chemical measurements 810;
14.4.5.7 Carbon and resin concentration meter
815; 14.4.5.8 Carbon activity meter 815
14. 5 Actuators and Final Control Elements .. .. .. .. . .. .. .. .. .. .. .. . 817
14.5.1 Introduction ............................................. . 817
14.5.2 Control valves .......................................... . 817
14.5.3 Material feeders ........................................ . 819
14.5.4 Variable-speed drives ................................. . 819
14.6 Process Control System Design Philosophies and Architec-
tures .................................................................... 819
14.6.1 Introduction .. .. . . .. . . . . . . . . . . . . . . .. . .. . . . . . .. . . . .. . . . .. . . 819
14.6.2 Manual control versus fully automatic control . 820
14.6.3 Hard-wired analogue control systems ............. 821
14.6.4 Centralized computer control systems ............ 822
14.6.5 Distributed computer control systems .... .... . .. .. 823
14.6.6 Examples of current practice ........................ 824
14.7 References ............................................................. 829
Chapter 15 The Chemistry of the Extraction of Gold, by

M.J. Nicol, C.A. Fleming and R.L. Paul ...... 831
15 .1 General Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831
15 .1.1 The chemistry of gold compounds .. . .. .. .. .. .. .. .. 831
15 .1.2 An equilibrium approach to the hydrometallurgy
of gold .................................................... 836
xi·v
CONTENTS
15 .1. 3Electrochemical aspects . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 843

15.1.3.1 Anodic reactions of gold 843; 15.1.3.2
Anodic reactions of zinc 848; 15.1.3.3 Cathodic
processes 849
15 .2 The Dissolution of Gold .. . . .. .. . .. .. . . .. . . . . .. . . . .. .. .. .. .. .. .. .. 854
15.2.1 The mechanism of cyanidation ..................... 854
15 .2 .2 The effect of other constituents of the pulp . .. . 857
15.2.2.1 Iron sulphide minerals 858; 15.2.2.2
Copper sulphide minerals 859; 15.2.2.3 Arsenic
sulphide minerals 860; 15.2.2.4 Antimony sul-
phide minerals 860; 15.2.2.5 Anions containing
sulphur 861; 15.2.2.6 Cyanide complexes of
various base metals 861; 15.2.2.7 Calcium ions
862; 15.2.2.8 Carbonaceous material 862;
15 .2.2.9 Flotation reagents and other surface ac-
tive reagents 863
15.2.3 Other leaching processes .............................. 863
15.2.3.1 Thiosulphate 864; 15.2.3.2 Thiourea
864; 15.2.3.3 Thiocyanate 865; 15.2.3.4 Chlo-
ride 866
15.3 The Recovery of Gold from Solution ........... .............. 866
15. 3 .1 Cementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 866
15 .3 .2 Extraction by activated carbon, ion-exchange
resins or organic solvents .. . . .. .. .. . . . .. .. . . . .. . . . . . . . 869
15.3.2.1 The activated-carbon process 870;
15.3.2.2 The ion-exchange process 877; 15.3.2.3
The solvent-extraction process 885
15.3.3 Chemical precipitation ............................... . 889
15 .3 .4 Electro winning ......................................... . 890
15.3.4.1 Celldesign890; 15.3.4.2 Electrodereac-
tions 893
15.4 Clean-up, Calcination and Smelting ........................... 894
15.4.1 Acid treatment .......................................... 895
15 .4.2 Calcination .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . .. . . . .. . . . . . . .. . . 896
15.4.3 Smelting ................................................... 896
15. 5 Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 899
15 .6 References . . . . . .. . . .. . . . . . . .. . . . .. . . . .. . . .. . . . . . . . . . . . . . . .. . . . . .. . . . . . . 899
Chapter 16 Plant Design and Commissioning, by G.G.

Stanley ... . . ... . . . . .. ..... ....... ....... ..... .......... .. . 907
16.1 The Importance of Good Plant Design and Punctual Com-
missioning . . . . . .. . . .. . . . .. . . . . . .. . . . .. . . . . . . . . . . . . . . . .. . . . .. .. . .. . . . .. . 907
16.2 The Design and Construction Organization .................. 907
xv
CONTENTS
16.3
Status, Responsibilities and Qualities of the Design Metal-
lurgist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 908
16.4 General Procedure for Plant Design ......................... .. 908
16.4.1 The procedural plan ................................... 908
16.4.2 Ore testing ................................................ 909
16.4.2.1 Obtaining the sample 909; 16.4.2.2
Mineralogical examination 911; 16.4.2.3 Sample
preparation 912; 16.4.2.4 Laboratory testing 913;
16.4.2.5 Pilot plant testing 917
16 .4. 3 Process design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 917
16.4.3.1 Process design criteria 917; 16.4.3.2
Flowsheet design 917; 16.4.3 .3 Quantified flow-
sheet 918; 16.4.3.4 Equipment sizing and selec-
tion 924; 16.4.3 .5 Flowsheet evaluation and
selection 926; 16.4.3 .6 The conclusion of the pro-
cess design phase 933
16.4.4 General arrangement .................................. 933
16.4.4.1 Site selection 934; 16.4.4.2 Elements of
good layout 935; 16.4.4.3 Minimizing transpor-
tation requirements 935; 16.4.4.4 Maximizing
ease of operation 937; 16.4.4.5 Maximizing ease
of maintenance 938; 16.4.4.6 Maximizing ease of
supervision 938; 16.4.4.7 Maximizing security
939; 16.4.4.8 Maximizing safety 940; 16.4.4.9
Provision for expansion 940; 16.4.4.10 The use
of CAD and models 941
16.4.5 Detailed design .......................................... 941
16.4.5 .1 Plant layout and design details 941;
16.4.5.2 Construction details 953; 16.4.5.3
Piping and instrumentation diagrams 957; 16.4.5.4
The definitive and revised cost estimates 957
16.5 Plant Construction and Commissioning .............. ........ 957
16.5.1 Metallurgical involvement in the construction
phase ....................................................... 957
16.5.2 Preparation for commissioning ..................... 958
16.5 .3 Commissioning . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 958
16.5.3.1 Personnel 958; 16.5.3.2 Cold commis-
sioning 958; 16.5.3.3 Hot commissioning 959;
16.5 .3 .4 Some practical commissioning tips 960;
16.5 .3 .5 Acceptance runs 961
16.6 Conclusion ............................................................ 961
16.7 References ............................................................. 961
Appendix 16.1 Pipeline Design .. .. . . ... . . . . . .. . . .. .. . . .. . . . .. . . . . . . .. .. .. . 963
Appendix 16.2 Pump Selection ... . . . . . . .. . . .. . .. .. . . . . . . . . .. . . . . . . .. . . .. . . 967
xvi
CONTENTS
Chapter 17 Loss Control and Safety Management, by

F.K. Vice and D.J. Nel .. .... .. .. .. .. .. ...... .. .. .. .. 971
17 .1Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 971
17 .2Loss Control Department .. .. .. .. . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 971
17 .3Physical Condition Rating Programme .. .. .. .. .. .. .. .. .. .. .. . 973
17.4 The Manager's Policy.............................................. 974
17 .5Training . . . . . . . . . . .. . . . . . . . . . . . . . . .. .. . . .. . . .. . . . . . . . . .. . . .. . . . . . .. . . . . . 974
17 .6Planned Inspections .. .. .. .. .. .. .. .. .... . .. .. .... . .. .. .. .. . .. .. .. .. . 974
17.6.1 The general inspection ................................ 974
17.6.2 Critical parts inspections ............................. 975
17 .6 .2 .1 Pre-use inspection of equipment 97 5
17 .7 Accident/Incident Investigation and Analysis ............... 975
17. 7 .1 The investigation .. .. .. .. .. .. .. . .. .. . .. .. .. . .. .. .. . .. .. . 977
17. 7 .2 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 977
17.7.2.1 Statistical data 977
17. 7 .3 Benefits derived from accident investigations and
analysis . . . . . . . . . . . . . . .. . . . . . . . . . . . . .. . . .. . . . . . . . . . . . . . . . . . . 982
17. 8 Communications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 982
17.8.1 Group meetings ......................................... 982
17 .8.2 Personal communication .. .... .... . .. .. .. .. .. .. .. .. .. 983
17 .9 Emergency Preparedness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 983
17 .9 .1 Care of the injured and ill .. .. . .. .. .. .. .. .. .. .. .. .. .. 985
17 .9 .2 Hazardous chemicals .. .. . .. . .. .. . .. .. . .. .. . .. .. .. . .. .. 985
17.9.3 Fires........................................................ 985
17.9 .4 Miscellaneous . . . . . . . . . . . .. . . . . . . . . . .. . . . .. . . . . . . . . . . . . . . . 986
17 .10 Physical Examination of Employees .. .. .. . .. . .. .. .. . .. .. .. .. .. 987
17 .10.1 Medical examinations . .. .. .. .. . .. .. .. .. . .. .. .. .. . .. .. . 987
17.10.2 Eyesight examinations ................................. 987
17 .10.3 Audiometric testing ... ............................ ..... 989
17 .11 Personal Protective Equipment .. .. .. .. . .. . .. .. .. .. . .. .. . .. .. .. .. 991
17 .12 Task Analysis and Procedures .. .. .. .. .. . .. .. .. . .. . .. .. .. .. . .. .. . 993
17 .12.1 Identification of critical tasks .. .. . .. . .. .. .. . .. .. .. .. 993
17.12.1.1 Severity994; 17.12.1.2 Repetitiveness
994; 17.12.1.3 Probability995; 17.12.1.4 Iden-
tification of tasks requiring procedures, standards
or rules 995; 17.12.1.5 Task analysis 996;
17.12.1.6 Critical task inventory summary 996
17 .13 Planned Task Observations .. . .. .. . .. .. .. . .. .. .. .. .. .. . .. .. . .. .. . 999
17 .14 Engineering and Purchasing Controls . .. . .. . .. . ... .. .. .. .. .. .. 1001
17 .15 References . . . . . . . . . .. . . . . .. . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1001
Appendix 17 .1 Checklist for General Planned Inspection and Physi-
cal Condition Ratings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1002
XVll
CONTENTS
Chapter 18 Laboratories and Pilot Plants, by C. W .A. Muir 1013

18.1Introduction .......................................................... 1013
18.2Characteristics of Metallurgical Laboratories of the Gold
Mining Industry .. . . . .. . .. . . .. . . . . ... . . . . .. . . . . . . . . . . . .. . . . .. .. . . ... . 1015
18.2.1 On-mine laboratories ................................... 1015
18.2.1.1 Layout 1015; 18.2.1.2 Equipment 1017;
18.2.1.3 Functions 1019; 18.2.1.4 Staffing 1019;
18.2.1.5 Comment 1020
18 .2 .2 Group laboratories .. .. .. .. .. .. .. .. .. .. .. .. .. .. . .. .. .. .. 1020
18.2.2.1 Layout 1022; 18.2.2.2 Equipment 1022;
18.2.2.3 Functions 1023; 18.2.2.4 Staffing 1025
18.3 Mintek 1026
18.3.1 Introduction.............................................. 1026
18 .3 .2 Categories of sponsorship and conditions pertain-
ing to them . . . .. . . .. . . . .. . . . . .. . . . .. .. . . . . .. . . .. . . . .. . . .. . 1027
18.3.3 The facilities at Mintek ............................... 1028
18.4 Pilot Plants ........................................................... 1029
18.4.1 Reasons for piloting ................................... 1029
18.4.2 Types of pilot plants .................................. 1030
18.4.3 Scale-up factors ......................................... 1030
18.4.4 Design of pilot plants ................................. 1031
18.4.5 Operation of pilot plants ............................. 1031
18.4.6 Pilot plant location .................................... 1031
18 .4. 7 Type of investigation carried out in pilot plants 1032
18. 5 References .. . .. . . .. .. . . .. . . .. . .. . . . .. . . . . .. . . . .. . . . . . . .. .. . .. . . . . .. .. . . 103 3
Appendix 18.1 Determination of Relative Grindability using a Laber
ratory Batch Rod Mill .. .. .. .. .. .. .. .. .. .. .. .. .. . .. .. . .. .. . .. .. .. .. 1034
Appendix 18.2 Standard Procedure for Routine Size Analysis of
Finish-ground Products ............................................ 1036
Appendix 18.3 Determination of Thickener Area .................... 1041
Appendix 18.4 Filtration Rate Determination ...... .. . .... .. ..... . .. .. 1044
Appendix 18.5 Cyanidation Tests using Rolling Bottle Method .. 1047
Appendix 18.6 Muffle Roasting of Sulphide Concentrates ........ 1049
Appendix 18.7 Determination of Activated Carbon Parameters . 1050
Tables and Formulae, by W.R. Flook ........................... 1053

Note on conversion tables ... .. .. .... .. .... .. ... .. .. .. .. .. .. .. .... 1054
Table 1 Conversion table: length .... ... .. ...... .... .. .. .. .. .. 1055
Table 2 Conversion table: area .. .. . .. .. .. .. .. .. .. .. .. . .. .. .. .. 1056
Table 3 Conversion table: volume .. .. .. .. .. .. .. .. .. .. .. .. .. .. 1057
Table 4 Conversion table: velocity .... .. . .... .. .. ........ .... 1058
Table 5 Conversion table: rate of flow . .. .. .. .. .. .. .. .. .. .. . 1058
Table 6 Conversion table: mass . .. . .. .. .. . .. .. .. . .. . .. . .. .. . .. 1059
x·v111
CONTENTS
Table 7 Conversion table: pressure . . .. .. . .. .. . . . .. . . . . .. . .. . . 1060

Table 8 Conversion table: energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1060
Table 9 Conversion table: power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1061
Table 10 Conversion table: density . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1061
Table 11 Some fundamental constants . . . . . . . . . . . . . . . . . . . . . . . . 1062
Table 12 Moisture content, density, etc. of pulps . . . . . . . . . 1062
Table 13 Formulae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1064
Table 14 Base units of the International System . . . . . . . . . . . 1067
Table 15 Derived units of the International System ... . . .. 1067
Table 16 SI prefixes ... . . ...... .. . ... . ....... .. .. ....... .. ... . . .. . . 1068
Table 17 Critical speed of mills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1068
Table 18 Te~ting-sieve series . .. . .. . . .. .. . . . .... .. . . . ... . . . ... .. .. 1069
Table 19 The periodic table .. .... ...... ... . . . . . . . .. . . . ....... .. . . 1071
Table 20 The Greek alphabet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1072
Index ...................................................................... 1073
VOLUME 1
Note: These contents have been abbreviated by omission of the third level
of subheading.
Chapter 1 Geology and Mineralogy of the Principal

Goldfields in South Africa, by C.R. Anhaeusser,
C.E. Feather, W.R. Liebenberg, G. Smits and
J.A. Snegg ............................................. .
1.1 Historical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Gold in the Sediments of the Witwatersrand Triad . . . . . . . . 5
1.2.1 Geological aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.2 Macroscopic description of the placers ... .. . .. ... 13
1.2.3 Mineral composition of the placers ..... ...... ... . . 15
1.2.4 Nature and mode of occurrence of the gold ... . 16
1.2.5 Distribution of gold and heavy minerals . . . . . . . . . 26
1.2.6 Chemical composition of the gold . . . . . . . . . . . . . . . . . 28
1.2.7 Gold in residues ......................................... 30
1.2.8 Other minerals of economic interest . . . . . . . . . . . . . . 33
1.3 Gold in the Barberton Mountain Land . . . . . . . . . . . . . . . . . . . . . . . 35
1.3 .1 Introduction .. . . . . . .. . . .. . . . . . . . . .. . . . . . .. . . . . .. . . .. . .. . . . 35
1.3.2 General geology ......................................... 37
1.3 .3 Gold distribution in the Barberton Mountain Land 41
xix
CONTENTS
1.3.4 Mineralogical aspects of the Barberton gold ores 49

1.3.5 Refractory ores and their treatment .............. . 58
1.3.6 Origin of the Barberton gold deposits ........... . 60
1.4 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Chapter 2 Sm1ing and Crushing, by G. G. Stanley . . . . . . . . . 69

2.1 Introduction .......................................................... 69
2.1.1 Mining methods ......................................... 69
2.1.2 Underground waste sorting .......................... 70
2.1.3 The nature of the run-of-mine ore ................. 70
2.1.4 The operations carried out in the sorting and
crushing plant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
2.2 Ore Transport to the Treatment Plant .. . .. . . .. . . . . . . .. . . . .. . . 75
2.2.1 Skip transport ........................................... 75
2.2.2 Endless rope haulages .. . .. . . .. . . .. . . .. . . . . .. .. . . . .. . . . 75
2.2.3 Rail transport . . . . . . .. .. . . . . .. . ... . . .. . . .. . . . . .. .. .. . .. . .. 75
2.2.4 Road transport . . . . .. . . ...... ... . . ... ... . . .. . .. .. .. . .. . . . 76
2.2.5 Overhead ropeways . ... ... .. . . . .. . .. . . . .. . .. . . . . . . . . . . . 76
2.2.6 Belt conveyors . .. . .. . . . . . . . . .. . . .. . . .. . . .. . . .. . . .. . . . .. . . 76
2.3 Ore Storage, Feeding and In-Plant Conveying ... . . .. . . ... . . 76
2.3.1 Ore storage ............................................... 76
2.3.2 Feeders .................................................... 80
2.3.3 In-plant conveying ..................................... 85
2.4 Washing ............................................................... 91
2.4.1 Washing in chutes ...................................... 92
2.4.2 Washing on conveyor belts .......................... 92
2.4.3 Washing on grizzlies ................................... 92
2.4.4 Washing in trammels .................................. 92
2.4.5 Washing on vibrating screens ....................... 93
2.4.6 Dealing with washings . .. . .. . . .. . . .. . . .. . . .. . . .. . . . ... . 94
2.5 Sizing ........ .. ..... .. .. . . . ... . . . . . . . .. . . . .. . . . .. . .. . . .. . .. . .. .. .. . ... . . 94
2. 5 .1 Grizzlies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
2.5.2 Vibrating screens .............................. ......... 95
2.6 Crushing .... ... .. . . .. .. . . . .. . . . . . . ....... .. .. . . .. .. .. . . .. . . . . . . .. . .. . . . 98
2.6.1 Types of crusher ........................................ 98
2. 7 Waste Sorting and Reef Picking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
2.7.1 Definition of waste sorting .......................... 107
2. 7 .2 Manual sorting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
2.7.3 Mechanized sorting .................................... 108
2. 7 .4 Reef picking . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . .. . . . . . . . . 111
2.7.5 Disposal of reject rock ................................ 111
2.7.6 Economics of waste sorting.......................... 113
xx
CONTENTS
2.8 Selection
of Grinding Media for Pebble Milling . .. . . . . . . . . . 114
2.8.1
Manual pebble selection .............................. 114
2.8.2
Pebble selection by grizzley or vibrating screen 114
2.8.3
Pebble selection by deflection of the entire ore
stream . ..................... ............... ......... ....... 115
2.8.4 Pebble storage and transport . . . . . . . . . . . . . .. . . .. . . . . . 115
2.9 Mass Measurement, Sampling and Control .... .............. 115
2.9.1 Mass measurement ..................................... 115
2.9.2 Sampling .................................................. 116
2.9.3 Control .................................................... 116
2 .10 Safety and Health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
2.10.1 Regulations ............................................... 117
2.10.2 Dust suppression .. . . . . . . .. . . . . . . . . . . .. . . . . . . . . .. . . . . . . . . 117
2.10.3 Noise suppression . .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . 119
2.11 References . . .. . . . .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . .. . . . . . . . . . 119
Chapter 3 Milling and Classification, by G. G. Stanley . . . 121

3 .1 Evolution of the Milling Techniques Used in the Treatment
of South African Gold Ores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
3 .2 Theory of Milling .. . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . .. . . . . . . 125
3 .2.1 Terminology and definition . . . . . . . . . . . . . . . . . . . . . . . . . . 125
3 .2.2 The reasons for grinding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
3.2.3 The boundary between crushing and grinding .. 127
3 .2.4 The process of brittle fracture resulting from corn-
pression ............................. , .................... . 128
3.2.5 The particle size distribution resulting from brittle
fracture ................................................... . 129
3.2.6 The determination and reporting of size distribu-
tion ........................................................ . 129
3.2.7 Mathematical descriptions of size distributions 131
3.2.8 Mechanisms of size reduction and transport within
tumbling mills .......................................... . 135
3.2.9 Energy requirements of comminution ............ . 139
3.2.10 Stage grinding .......................................... . 143
3.2.11 Closed circuiting ....................................... . 147
3.2.12 Summary of basic milling theory ................. . 152
3.3 Types of Tumbling Mill and their Characteristics.......... 152
3.3.l General description .................................... 152
3.3.2 Rod mills ................................................. 157
3 .3 .3 Ball mills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
3.3.4 Autogenous mills ....................................... 160
3.3.5 Milling combinations .................................. 162
3 .4 Mill Feeders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
3.4.1 Scoop feeders ............................................ 163
xxi
CONTENTS
3.4.2 Drum feeders ........................................... . 164

3.4.3 Spout feeders .......................................... .. 164
3.4.4 Hopper feeders ......................................... . 165
3 .5 Mill Liners ... . .. .. .... . . .. . . .. .. . . .. . . .. .... . . . .. . . . . . . .. . . . . . . . . . .... 166
3 .5 .1 Trunnion liners ........................................ .. 166
3.5.2 Inlet end liners ......................................... . 167
3.5.3 Outlet end liners ...................................... .. 167
3.5.4 Shell liners .............................................. .. 170
3.5.5 Summary of South African liner practice ...... .. 175
3.5.6 Relining methods ..................................... .. 175
3 .6 Grinding Media .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 178
3.6.1 Rods ....................................................... 178
3 .6.2 Balls .. .. . . .. . .. .. . . .. . ... .. . . .. . ... . . . . .. .... . .. . . .... .... .. 179
3 .6.3 Pebbles .. . . .. . . . . .. .. . . .. . . .. .. . . .. . .. . . . . . . .. . . . . . . . . . .. . . 179
3 .6.4 Grinding medium size for given ore feed size .. 179
3 .6.5 Nature of grinding medium wear .. .. .. .. .. .. .. .. .. 181
3 .6.6 The ball-wear theories of Davis and Prentice .. . 182
3.6.7 Ball rationing............................................ 183
3.7 Ancillary Equipment ............................................... 184
3 .7 .1 Feeders . .. . . .. .. . .. . . . . .. . . .. . .. . .. .. . . . . . .. .. . . . . . . . . . .. .. 184
3. 7 .2 Massmeters .. .. . . .. .. . .. . . .. . .. . .. .. .. . . . . .. .. . . . . . . . . .. .. 187
3.7.3 Pumps ..................................................... 187
3.7.4 Pipes ....................................................... 188
3.7.5 Valves...................................................... 188
3.7.6 Distributors .............................................. 188
3. 7. 7 Water reticulation .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 189
3 .8 Principles of Mill Operation .... .............. .... .. .. .. .. .. .... . 189
3.8.1 Basic objectives ......................................... 189
3 .8.2 Maximizing energy availability .... .. .. .... .. .. .. .. .. 189
3.8.3 The effect of feed size distribution in ROM and
SAG mills . . .. . . . .. . .. . . . . . .. . .. . .. . . . . . . . .. . . .. . .. . . . .. . . 195
3 .8.4 Maximizing utilization of the available power .. 196
3.8.5 Automatic control ...................................... 199
3 .9 References . . .. .. . . . . . .. . . . .. . . .. . . . .. . . .. .. . . .. . . . . . . . . . . . .. . . . . . . . .. . . 201
Appendix 3 .1 Hydrocyclone classifier theory, modelling and design 204
Appendix 3 .2 Classification formulae .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. . 209
Appendix 3 .3 Pebble competence testing and autogenous grindabili-
ty determination .. ...... .. .. .. .. .... .. .... .. .. .... .. .... .... .. .. .... . 210
Appendix 3.4 Mill design methods ....................................... 212
Appendix 3.5 Determining grinding energy requirements for
materials that are already finer than the feed size (3350 µm)
specified for the Bond Grindability Test .. ...... .. .. .. .. .... . 216
Appendix 3 .6 Derivation of Equation 3 .22 relating ball mass in load,
daily ball feed rate, new ball diameter and daily ball wear
rate ...................................................................... 217
XXll
CONTENTS
Chapter 4 Gravity Concentration, by D.W. Penman ...... 219

4 .1 Introduction .. . .. . . .. . . . . .. . . .. . .. .. . .. . . .. . . .. . . . .. . . .. . . .. . . . .. . .. .. 219
4.1.1 Historical ................................................. 219
4.1.2 The rationale of concentration ...................... 219
4.1.3 Historical background ................................. 220
4.1.4 Present philosophy and practice .................... 221
4.2 Advantages and Disadvantages of Gravity Concentration 221
4.3 Equipment ............................................................ 223
4.3.1 Corduroy blankets ..................................... 223
4.3.2 Johnson drum concentrator ..................... .... 224
4.3.3 Plane table ............................................... 225
4.3.4 Jigs ......................................................... 226
4.3.5 Endless riffle belts ...................................... 226
4.3.6 Shaking tables ........................................... 227
4.3.7 Mill liners ............................................. .... 228
4.4 Concentrate Treatment ............................................ 228
4.4.1 Re-dressing and smelting ............................. 228
4.4.2 Roasting and cyanidation ............................ 228
4.4.3 Amalgamation ........................................... 229
4.4.4 Intensive cyanidation ...... .... .... .... ... ...... .... ... 230
4.4.5 Iron removal . .. . .. .. . .. . .. . .. .. . .. . .. . . . .. .. . . . . . .. . . .. . . 230
4.5 Gravity Concentration Circuits .................................. 231
4.5.1 President Steyn gravity circuit ...................... 231
4.5.2 Gravity concentration at Sheba ..................... 232
4.6 Development .......................................................... 233
4.7 References ............................................................. 234
Chapter 5 Flotation, by B.R. Broekman, L.A.C. Carter and

R. C. Dunne . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
5 .1 Introduction .. .. . .. . .. . . . .. .. .. . .. . . .. . . .. . .. . . .. . .. .. . . . . .. .. . . . . . .. . 23 5
5.2 Flotation Processing ................................................ 236
5 .2.1 Mineralogical aspects .. .. .. .. .. .. . .. .. .. . .. .. .. .. .. .. .. 236
5 .2.2 Process aspects ...... .... .... .......... ... .... .. .... .... . 238
5.2.3 Flotation plant practice ............................... 251
5.3 Current Flotation Trends ......................................... 260
5.3.1 Equipment ............................................... . 260
5.3.2 Design ................................................... .. 261
5.3.3 Control ................................................... . 265
5.3.4 Trouble-shooting methods .......................... . 266
5.3.5 New flotation applications .......................... . 272
5.4 References
xxiii
CONTENTS
Chapter 6 Cymmidation, by G.J.C. Young .................... 277

6.1 Introduction .......................................................... 277
6.2 Thickening ............................................................ 277
6.2.1 Batch thickeners or collectors ....................... 277
6.2.2 Continuous thickeners ................................. 278
6.2.3 High-rate thickeners .... .. . . . ... ...... .......... ....... 290
6.3 Leaching ............................................................... 291
6.3.1 Leaching reagents ...................................... 291
6.3.2 Batch and continuous leaching ..................... 295
6.3.3 Leaching vessels ......................................... 301
6.3.4 Leaching conditions .................................... 303
6.3.5 Gold losses ............................................... 304
6 .4 Liquid-solid Separation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
6.4.1 Filtration .................................................. 305
6.4.2 Filter practice ... .... .. ..... ....... .... ... ....... ... .... .. 312
6.4.3 Maintenance ............................................. 316
6 .4 .4 Capacities .. . . .. . . .. . . . .. . . . .. . . . .. . . .. . . .. . . . .. . .. .. . .. . .. 317
6.5 Clarification .......................................................... 317
6. 5 .1 Clarifier types . . . .. .. .. .. .. .. .. .. .. .. .. . .. . . .. . . .. .. .. . . . 318
6.6 References .. .. . ... ... . .. .. . .. ... . . . ... . . . .... . ..... ... . .... .. . .. . . ... . . 323
Appendix 6.1 Thickener design calculations ........................... 324
Appendix 6.2 Leaching design calculations ....... ..... ....... .. .. ... . . 326
Appendix 6.3 A note on the development of an on-stream analyser
for cyanide concentration, pH and alkalinity . . . . . . . . . . . . . . . 328
Appendix 6.4 The determination of nett dissolved gold loss (DG
Nett) .................................................................... 328
Chapter 7 Recovery of Gold from Solution by Cementation,

by D. Bosley .. .. .. .. ... .. ... .. ..... ...... .. ... . ... ..... 331
7 .1 Introduction . . . . . .. . . .. . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . .. . . . . .. . . .. . . 331
7 .1.1 Characteristics of gold bearing solutions . . .. .. . . . 331
7 .1.2 Recovery methods . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . . . .. . . . 331
7 .2 Zinc Dust Precipitation .. .. .. .. .... .. .... .... .... .. ...... .... .... . . 333
7 .2.1 Solution preparation .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. . 333
7 .2.2 Precipitation practice . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. . 337
7 .3 References . ... ... .. . . .. .... .. . ... ........................ ... . . . .. . ... . . 343
Chapter 8 Pyrometallurgy of Gold, by G.I. Gossman .... 345

8 .1 Introduction .. . . .. .. . .. . . . .. . . . .. . . . .. . . . . . . . . . . .. . . . . . . . . .. . . . . . . . . . . 345
8 .2 Roasting of Concentrates .. . . . . .. . . . . . . . . . . . . . . . . .. . . . .. . . . . . . .. . . 345
8 .2.1 Introduction .. .. . . . .. . . . .. .. . .. . .. .. .. .. . .. .. .. . .. .. .. .. .. 345
8.2.2 Basic chemistry of the roasting process .......... 346
XXlV
CONTENTS
8 .2.3 Roasting furnaces .... .. .. .. .... .. .. .. .. .. .. .. .. .. .. .. .. 348

8.2.4 Roaster gas treatment ......................... ...... .. 350
8 .2.5 Status of roasting in South African gold metallurgy 351
8.3 Retorting of Gold Amalgam ........ ............................. 352
8.3.1 Introduction.............................................. 352
8.3.2 The retort furnace .............................. .... .... 352
8.3.3 Mercury hazard ......................................... 353
8.3.3.1 Precautions when handling mercury 354
8.4 Calcining of Zinc-Gold Precipitate ............................. 354
8.4.1 Introduction.............................................. 354
8 .4.2 Clean-up of zinc-gold precipitate .. .. .. .. .. .. .. .. .. . 355
8.4.3 Acid treatment of zinc-gold precipitate ........... 356
8.4.4 Dewatering of gold slime.............................. 357
8.4.5 Calcining of gold slime ............................... 361
8.5 Smelting gold bearing Material ............. ................. .... 363
8.5.1 Introduction.............................................. 363
8.5.2 Fusion furnaces ......................................... 363
8 .5 .3 Direct smelting of gold concentrates .. .. .. . .. .. .. . 369
8 .5 .4 Melting of sponge gold .. .. .. .. .. . .. .. .. .. .. .. .. .. .. .. 369
8.5.5 Smelting of calcined slime ........................... 370
8.5.6 Smelting of steelwool cathodes ..................... 372
8.5.7 Cleaning and sampling gold bars .................. 373
8.5.8 Slag handling ............................................ 375
8.5.9 Safety hazards in the smelthouse ................... 375
8.6 Dust and Fume Extraction ....................................... 376
8. 7 Precautions against Gold Theft .. .. .. .. .. .. .. .. .. .. .. .. .. .. . .. .. 376
8.8 Record Keeping ................................ .................. .... 377
Chapter 9 Application of Activated Carbon to Gold

Recovery, by P.R. Bailey ....................... .... 379
9 .1 Introduction . .. . . . .. . .. . . . . . . . . . . .. . . .. . . . . . . . . . . .. . . . . . . . . . . . .. . . . . .. 379
9 .1.1 Historical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
9 .1.2 Gold recovery .. . .. .. .. . .. . .. .. .. . .. .. .. .. . .. .. .. .. .. .. .. 379
9 .1.3 Process route comparisons .. .. . .. .. .. . .. .. .. .. .. .. .. . 382
9.2 Activated Carbon .............................................. ..... 383
9.2.1 Physical and chemical properties .................. . 384
9.2.2 Gold absorbing properties .......................... .. 387
9.3 Absorption ............................................................ 393
9.3.1 Modelling ................................................. 393
9.3.2 Stage, loading and barren values ................... 399
9.3.3 Equilibrium effects ..................................... 417
9. 3.4 Carbon activity . .. .. . .. .. .. .. .. .. .. .. .. .. .. . .. .. .. .. .. .. . 429
9.3.5 Physical effects .................... ...................... 439
xxv
CONTENTS
9.3.6 Operating considerations ............................ . 443

9.3.7 Carbon-in-leach operations ........................ .. 449
9.3.8 Other feed stocks ..................................... .. 455
9.3.9 Equipment ............................................... . 457
9.4 Elution 481
9.4.1 Introduction ............................................. . 481
9.4.2 Physical and chemical effects ...................... . 483
9.4.3 Kinetics of elution ..................................... . 489
9.4.4 Anglo American Research Laboratory (AARL)
elution .................................................... . 492
9.4.5 Zadra elution ........................................... . 503
9.4.6 Equipment ............................................... . 511
9.5 Regeneration.......................................................... 530
9.5.1 Introduction.............................................. 530
9.5.2 Thermal regeneration .................................. 530
9.5.3 Rotary kilns .............................................. 541
9.5.4 Rintoul kiln regeneration ............................. 547
9.6 Electrowinning ....................................................... 550
9.6.1 Electrochemical reactions ............................. 551
9.6.2 Cell design ................................................ 552
9.6.3 Electrowinning cells .................................... 554
9.6.4 Mintek cell - operational data .................... 561
9. 7 Acid Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 570
9. 7 .1 Acid treatment conditions . .. .. .. .. . .. .. . .. .. .. .. .. .. . 571
9.8 Carbon Breakage .................................................... 575
9.8.1 Breakage by pumps .................................... 575
9.8.2 Breakage by eductors .................................. 581
9.8.3 Breakage by mixing .................................... 583
9.8.4 Breakage during regeneration ....................... 583
9.8.5 Plant carbon loss ....................................... 584
9.9 Carbon in Solution ................................................. 585
9.9.1 Adsorption systems .................................... 585
9.9.2 Design considerations ................................. 599
9.10 References ............................................................. 608
xxvi
Chapter 10
Refining of Gold at the Rand Refinery

K.G. Fisher
10.1 Introduction
Since 1921, all gold produced in South Africa has been refined at the Rand
Refinery (Adamson, 1972). To give the reader a complete picture of gold
refining in South Africa, the historical development and all operations of
the refinery are described in detail in this chapter, although some activities
cannot strictly be defined as gold refining.
10.2 Historical Development

In the early days of South African gold mining, the crude bullion produced
had to be shipped to London for refining. The concept of a local refinery
had been considered for several years before the first positive step was taken
in 1909, when the Transvaal Chamber of Mines founded the Witwatersrand
Co-operative Smelting Works, to treat the gold and silver bearing by-products
from reduction works. These by-products required special treatment and were
obviously best handled by a central, joint concern under the sponsorship of
the Chamber.
Later, in 1917, the South African government proposed the establish-
ment of a local mint and consequently the Pretoria branch of the London
Royal Mint was opened in 1922. During the same period the Transvaal
Chamber of Mines investigated the possibility of establishing a local gold
refinery. The technical aspects were discussed with Sir T.K. Rose, a leading
authority on gold metallurgy, and the chlorine refining process was chosen
specifically with a view to the rapid and effective treatment of large amounts
of bullion. This process had been patented by F.B. Miller, Assayer of the
Sydney Mint, in 1867 and used in Australia since 1872 in contrast to the Lon-
don and continental refineries which were still using the sulphuric acid (par-
ting) process.
On 27 November 1920, Rand Refinery Limited was registered as a private
company with initial capital of £50 OOO, raised from shares limited to gold
mining companies that were members of the Transvaal Chamber of Mines.
A further £165 OOO was borrowed and secured by registered debentures.
Following a comprehensive survey, the site at Germiston, 16 kilometres east
of Johannesburg, was selected because services and rail facilities were
available and its position was the geographical centre of the Witwatersrand
gold mining area.
By December 1921, the refinery was completed with a capacity of 12
million troy ounces (373 250 kg) of fine gold per annum. Designed capacity
615
THE EXTRACTIVE METALLURGY OF GOLD
Table 10.1. Throughput of the Rand Refinery

(Year ending 30th September).
Bullion Fine gold Fine silver

received produced produced
Year (kg) (kg) (kg)
1922 133 230 110 362 6 710

1930 381 732 337 230 31 985
1940 518 982 457 949 39 273
1950 436 855 380 196 42 191
1960 745 761 660 675 67 189
1970 1 149 945 1 007 603 86 100
1980 806 505 695 189 77 293
1983 802 529 690 006 80 275
1984 800 583 678 199 73 459
was reached within ten years and since then the throughput has increased
considerably, as Table 10.1 illustrates.
Modifications and improvements were made to cater for this increasing
throughput, but by 1965 it had become necessary to undertake a major four-
year expansion programme at a cost of R4 million. Electrolytic gold and silver
refining facilities were installed and many new concepts were introduced,
including large-scale refining (up to 500 kg per furnace), induction furnaces,
mechanical handling and modern analytical techniques.
At the same time, the old Witwatersrand Co-operative Smelting Works
(by then known as By-Products Limited) located in Johannesburg was also
in need of large-scale renovations but, because of site limitations, it was con-
sidered more economical to build a new plant elsewhere. A suitable site was
available at Rand Refinery, so it was decided to amalgamate the two com-
panies and in 1966 the new by-products treatment plant (smelter) was erected
alongside the gold refinery. The interchange of products between refinery
and smelter was greatly facilitated and support functions (administration,
laboratory, assay office, workshops, etc.) were combined with consequent
economies in operation.
In 1969, embryonic facilities were installed for the production of alloy
and blanks for the Krugerrand family of coins. Owing to the subsequent suc-
cess of the coin marketing programme, a purpose-built coin blank complex
was erected in 1984 at a cost of R4 million. Thus Rand Refinery Limited
has developed into a fully integrated complex capable of refining all South
African gold bullion and gold-containing by-products and is the largest
establishment of its kind in the world, being responsible for approximately
62% of the free world's production of fine gold.
10.3 Function
Rand Refinery Limited is a subsidiary of the Chamber of Mines and operates
on a non-profit making basis for the benefit of member companies, but also
services the requirements of non-members. The refinery's principal functions
616
THE REFINING OF GOLD
are to refine bullion and recover precious metals from by-product materials
received from member gold mines and to sell the refined gold to the Reserve
Bank, which credits the mines accordingly. A subsidiary function is now the
production of blanks for the Krugerrand and its sub-denominations.
Depositors are credited with fixed percentages of the gold and silver con-
tents of their deposits, less appropriate treatment charges. These percentages
are currently (mid-1986):
Bullion: 99,85% of the gold

94,00% of the silver
By-products: 97 ,50% of the gold
90,00% of the silver
Periodically a pro rata distribution of the value of any metallurgical

surplus is made to members. A vital feature of the whole operation is that
of metallurgical control, whereby each of twelve metallurgical centres is re-
quired to compile a monthly balance for gold and silver and to explain any
loss or gain.
Crude bullion J
By-products
GRB 86%1Au ~~~~~~ ~;___ ,

:- - - - - - -- - - - - - -S~~pl~~ -: Sampling /
: 1 : ,--- _§!ll3___ -; Sitr :

: Electro- 99,65% Chlorine :chlorides: Silver leach :
: refining Au refining : :
l
Blast furnace
'
~'. /
,/ <1
99:99~;r,,\;--:,7' - -- - - - S/Qt'~t t'.j'
_I
't'r
:
Electro-
:
•30% Au:
I l
~~in-~,:~\ refining :10010 Ag: Cupe!
production: 99 •6 % Au 1 i ---s~-elt~;---
--1~~~k~
'
, Bar casting !
I
'
99,9~~;.g_r ____ _
. Coins Ad . . . .
S.A. Mmt_____... mm1strat10n ----~
i
Sales
To Reserve Bank (fine gold bars)
To banks and dealers (coins & silver bars)
(Credited to depositors)
Not shown for simplicity are:

(a) Transfer of by-products to smelter from GRB and SRB
(b) Transfer and returns of samples to assay and chemical
laboratories from all departments
Figure 10.1. Simplified flow sheet of Rand Refinery.
617
10.4 Metallurgical Operations
10.4.1 Summary description

The metallurgical operations of Rand Refinery are summarised in Figure 10.1,
in which it can be seen that the operations can be considered under the follow-
ing five headings: Gold Refining Branch (G.R.B.), Silver Refining Branch
(S.R.B.), Coin blank production, Smelter, and Assaying and chemical
analysis.
10.4.2 Gold Refining Branch

10.4.2.1 Flow sheet
As shown in Figure 10.2, incoming bullion is weighed into the melthouse
where it is melted down for sampling purposes and recast into bars suitable
for the refining furnaces. When the assay and mass have been agreed with
the depositing mine, the bullion is delivered to the refinery, where it is refin-
ed to 996,0 parts per thousand purity (fineness 9960) by the chlorine injec-
tion technique. Most of the refined bullion is cast into 12,6 kilogram monetary
bars and is delivered to the Administration department, for sale to the Reserve
Bank. Some crude bullion is contaminated with the platinum group metals
Bullion receipts
l
Av. 86% Au Smelter
Weigh twice
Assay Dip Melting Fume Electro-

& ,_____ _ _ ____., static
Office samples sampling precip.
Casting
Weighl
Storage
C1z & flux
Chlorides
Electro- Chlorine >----------i>i De-
refining 99,6% anodes refining
Au/ Ag button
99,99%jcathodes Chlorides
9960 Fine gold bars t
Casting Quench
Weigh
---
Coin blank -
production
---
Pack & despatch
Figure 10.2. Flow sheet of the Gold Refining Branch (G.R.B.).
G
618
Figure 10.3. Weighing mine bullion on arrival at Rand Refinery.
osmium and iridium (osmiridium), which are not removed by chlorination

and such deposits must be further refined by electrolysis. The product of
the electrolytic gold section (fineness 9999) is delivered to the Coin Bank Com-
plex for the manufacture of Krugerrand blanks.
Molten chlorides from the refinery are first de-golded, in the de-golding
section, by partial reduction with sodium carbonate, then quenched and
transferred to the Silver Refining Branch for subsequent silver recovery.
10.4.2.2 Melting and sampling

The average assay of the crude bullion received from the mine is gold 84%,
silver 11 OJo and base metals 5%, although any gold assay over 50% is nor-
mally acceptable. A standard shape of tapered bar has been adopted by all
mines, especially designed to fit face to face in a melting crucible, giving a
maximum bar mass of 31 kg. Bars from a single mine are combined in groups
of up to four bars each to make up nominal 125 kg" deposits". Each deposit
has its mass determined on a beam balance and checked on a second balance
(Figure 10.3), the mass being determined to the nearest gram.
It is considered that the only reliable method of sampling heterogeneous
alloys such as crude gold bullion is by taking dip samples from the molten
mass. For sampling, therefore, each deposit is melted down in an induction
furnace crucible and, after a few minutes at peak temperature, four dip
samples for fire assay are taken simultaneously by means of a special carbon-
619
Figure 10.4. Crude gold sampling spoon. The right-hand vertical

dark bar is the shadow of the spoon shaft.
coated mild steel spoon (Figure 10.4), together with a disc sample for X-ray
analysis. The mixing induced in the melt by the induction heating method
guarantees the homogeneity necessary for accurate sampling. The samples
are despatched via a pneumatic conveyor system to the assay office for gold
and silver determination. Assays are done in duplicate in two separate assay
offices, each office taking two of the spo9n samples.
The bullion is cast out manually into 12,5 kg bars on a 1,6 m diameter
casting wheel and the solidified ingots are automatically tipped on to a slat-
ted conveyor which conveys them via a cooling trough to a marking table
(Figure 10.5). Each ingot is stamped with its original deposit number and
the deposit is reweighed to determine its "mass after melting". Strict con-
trol is exercised over any spillage, and crucibles are cleaned after each melt.
The deposit is stored in a strong-room, normally overnight, to await its assay
results.
The refinery assay is compared with the depositor's advised assay and
any discrepancy greater than 15 parts per 1 OOO results in possible re-sampling
and re-assaying. Upon agreement of the fine gold content, the deposit loses
its identity and is subsequently transferred to the refining section. It is seldom
that a genuine discrepancy arises in the assays, but such disagreements are
generally due to the presence of platinum group metals which interfere with
the assay or of excessive amounts of iron which form a separate phase at
normal sampling temperatures.
620
Figure 10.5. Casting crude bullion after melting and sampling.
The six furnaces used in this department are stationary 2 OOO Hz induc-
tion furnaces, employing clay/graphite crucibles which are lowered into and
raised out of the furnace on an hydraulically operated platform. A normal
deposit is heated for 20 - 25 minutes using 70 kW at 1200 V to reach a
temperature about 100°C above its melting point of approximately 1050°C.
Power generation and the cooling water system are shared with tbe refining
room, described in the following section.
10.4.2.3 Refining
Refining of the gold bullion is carried out in seven 500 kg capacity induction
furnaces, using the Miller chlorine injection process, whereby all the base
metal impurities are converted to their chlorides and hence separated from
the gold either as a fume/gas or liquid slag, according to Table 10.2.
In practice, the iron is removed first as gaseous FeC1 3 followed by lead
and zinc chlorides, also in gaseous form. The reaction is therefore quite
vigorous and the chlorine injection rate must be limited to control the tur-
bulence. Only when these three elements have been almost completely remov-
ed do the copper and silver start to form their chlorides, which are liquids
at the reaction temperature of 1150° C, enabling the injection rate to be in-
creased. As the end point is approached, gold chloride fumes start to form
and considerable turbulence again occurs. Figure 10.6 shows how a typical
reaction proceeds.
621
Table 10.2. Chlorination reaction data.
Free energy Melting point Boiling point

LlG at 1!50QC of chloride of chloride
Reaction (kCal) (QC) {QC)
2Au + 3Cl2 - 2 AuC13 Positive Decomp. 254

2Ag + Cl2 - 2 AgCl -33 455 1 550
2Cu + Cl2 - 2 CuCI -30 430 I 490
Fe + Cl2 - FeC12 -48 672 1 026
Pb + Cl2 - PbC12 -48 501 950
Zn + Cl, - ZnC12 -69 383 732
2Fe + 3Cl2 - 2FeCl3 NIA 306 315
Each 450 kg furnace charge is made up in the melthouse by selecting

from the deposits whose assays have been agreed. The ingots, normally
around thirty-six in number, are stacked on a pallet, delivered to the refin-
ing room and loaded manually into a preheated furnace. Approximately 3 kg
of a borax/silica/salt flux is added to absorb extraneous impurities, protect
the crucible and inhibit volatilisation.
In normal operation the charge melts in 45 minutes. Thereafter, two
tapered graphite tubes, internal diameter 13 mm, are inserted into the molten
bullion and linked to the chlorine distribution system by rubber hoses. As
shown in Figure 10.6 and described above, chlorination takes approximate-
ly 35 minutes. As the molten chlorides fill the crucible they are bailed off
into a preheated 6 litre crucible and delivered to the de-golding section. Fur-
ther flux is added after bailing and three or four bailing/ fluxing operations
i::
0
·~
~u
§ lOOr....,,,,.--~-r------i---Typical initial concentrations
u
~ ~~~
.S Cu 4,0o/o
.~
Ci Pb 0, 7%
Fe 0,2%
0
v Zn 0,1 o/o
Oll
~ 50+--t----H+----'t---'<-1-----+------+----+
~
0.
Time of chlorination (min)
Figure 10.6. Chlorination graph.
622
6 mm 105 mm dia.
A No. 5 fireclay crucible

B Slag
a C Chloride
a D Bullion
0
"' E Clean bullion sample
F Graphite stopper
Figure 10. 7. Section through sampling crucible.
are required, depending on the original bullion quality.

The end point of the refining operation is indicated by the smell of free
chlorine in the fume and by the colour of the fume, which begins to turn
reddish-brown as gold chloride forms. At this point a disc sample is taken
from the metal, through the slag, using a special crucible as shown in Figure
10.7.
The sample is conveyed to the laboratory via a pneumatic conveyor,
where it is immediately analysed for silver using an EDX (energy dispersive
X-ray) spectrometer. If the silver content is less than 0,35%, the bullion is
considered sufficiently refined, then chlorine injection is stopped, the pipes
removed, and the residual chlorides bailed off and finally skimmed off after
thickening with bone ash.
After heating the refined bullion to l 100°C, the furnace is tilted and
the bullion poured into a pre-heated transfer ladle which is conveyed by fork
truck to a 1 OOO kg capacity holding furnace. It is normal to transfer two
refined charges to the holding furnace before casting out into bars.
Approximately 14 kg of gold is poured from the holding furnace into
a pre-heated crucible from which the metal is poured manually into a mould
placed on a platform scale, until an increase of 12,6 kg is indicated (Figure
10.8). A soft reducing gas flame is then directed on to the surface of the
bar until the surface is solidified. In this way the bar cools from the bottom
upwards and shrinkage cavities are avoided; the flame also polishes the bar
surface and drives small particles of refractory dust to the edges.
Five such arrangements of scales and polishing flames are placed in front
of the holding furnace and as soon as a bar is solidified, it is tipped out of
its mould on to a copper carrier and immersed in a quench tank. The mould
is replaced ready for another bar. All moulds are pre-heated and given a
mould dressing of diesoline smoke. For every twelve bars poured, a pair of
dip samples is taken and sent for fire assay. The cooled bars are removed
from the quench tank and stamped with a series of code numbers, the
trademark and the name "Rand Refinery Limited". They are then weighed
623
Figure 10.8. Fine gold casting. Polishing flame burner in raised position.
twice on separate balances (Figure 10.9) and stacked on pallets to await the
assay results.
When the assay results are received, the fineness is stamped on the bars,
the bars inspected for surface defects and the finally approved bars are
delivered to the Administration department, which handles sales and deliveries
to the Reserve Bank.
The acceptable mass range for monetary bars is 12,441 4 to 12,752 4
624
Figure 10.9. Second weighing of fine gold bars.
kg corresponding to a range of 400 to 410 troy ounces and the fineness is

a minimum of 995 ,0 parts per thousand, average 996,5.
Seven identical furnaces are used for refining, these being tilting, 2 OOO
Hz induction furnaces. Normally, power up to 100 kW at 1 200 Vis employed
for melting but is reduced for refining. The 500 kg capacity clay-lined
clay/graphite crucibles are fixed into position with a castable refractory.
Crucible life is normally 21 - 27 heats, depending on the bullion refined; base
metals, particularly lead, have a deleterious effect on the lining.
The two holding furnaces are of similar design but are fitted with 1 OOO
kg unlined crucibles, are rated at 220 kW and employ two coil voltages, i.e.
800 V for holding molten metal and 1 200 V for melting ingots or bars.
The refining, holding and melthouse melting furnaces are all fed with
2 OOO I!.,z power from a central station, consisting of five 350 kW frequency
converters, up to four of which operate at any one time. The converters deliver
power at 800 V which is transformed to 1 200 V and 600 V. These three
voltages are supplied to the furnaces as required via a busbar system. Each
furnace is equipped with overload protection, coil voltage control, automatic
power correction and power metering. The overall load is balanced between
the departments to maintain a steady maximum demand. Constant
temperature cooling water is distributed to the furnaces from an ice bank/ am-
monia compressor system, with provision for emergency flushing with mains
water in the event of power failure.
625
10.4.2.4 De-go/ding
The de-golding room is an extension of the refining room in which the molten
chlorides from the refining operation are freed of contained gold before be-
ing delivered to the Silver Refining Branch.
The process comprises the addition of sufficient sodium carbonate to
the molten chlorides to precipitate some 4 % of the contained silver. The reac-
tions are:
2 AgCl + N~C0 3 - Ag2 C03 + 2 NaCl (10.1)

Ag2C03 - AglO + co2 (10.2)
2Agp - 4Ag + 0 2 (10.3)
Gold is present either as mechanically entrained particles or as a result

of the dissociation of gold chloride, which is unstable. In either case it is
more soluble in the metallic phase and is collected by the myriad fine silver
droplets produced by reaction (10.3), and the resulting gold-silver alloy set-
tles to the bottom of the melt.
In practice, the molten chlorides are transferred in a 15 kg crucible from
the refining furnaces and poured into the crucible of a gas-fired tilting fur-
nace. A furnace charge of approximately 120 kg is heated to 1 100°C, at
which temperature the borax silica flux from the refining process floats to
the surface of the melt and is used to inhibit the volatilisation of silver
chloride. Three additions each of 1,5 kg of sodium carbonate are required;
the reaction is quite vigorous and must be allowed to subside between addi-
tions. The reaction is complete after about 10 minutes but the furnace is allow-
ed to stand for a further 10 minutes, after which the viscous slag layer is
removed and most of the chlorides are decanted into a pre-heated 250 kg
capacity transfer ladle. The remaining chloride plus the gold-silver alloy are
poured into a 6 litre crucible and cooled to 800° C, at which temperature the
chloride is still fluid and is poured off and returned to the gas furnace. The
solidified gold-silver button remaining assays about 25% gold, 72% silver
and 3 % base metal and is tipped out of the crucible and recycled to the refin-
ing furnaces.
The ladle containing the "de-golded" (less than 100 ppm) chlorides
is hoisted above a rubber-lined mild steel quenching tank containing 350 litres
of water and the chlorides are slowly poured via a funnel into the tank. This
granulation process produces a finely divided product ideally suited to the
leaching process that follows. The tank and contents are delivered to the silver
leaching department (see section 10.4.3 .2).
10.4.2.5 Electrolytic gold refining

The electrolytic gold refining section fulfils two purposes in the refinery: firstly
it provides the means of producing high purity gold, i.e. 9999 fine as oppos-
ed to the 9960 fineness of monetary bars, and secondly it provides the means
ofremoving platinum group metals (P.G.M.s) which cannot be removed by
chlorination. In practice these two objectives are combined by selecting
deposits high in P.G.M.'s for electrolysis (after chlorination) to satisfy the
626
demand for high purity gold.

The Miller chlorination process is capable of producing high purity gold,
but the inevitable volatilisation of gold chloride at low impurity concentra-
tions necessitates elaborate fume collection arrangements to prevent excessive
gold losses. Therefore the W ohlwill electrolysis process is preferred, despite
its inherent lock-up of gold (and hence operating capital) in electrodes and
electrolyte.
The electrolyte used is acidic gold chloride, which ionises thus:
HAuC14 - H+ + Auel; (10.4)
At the cathode:
Auel; - Au3 + + 4c1- (10.5)
Au1+ + 3e- - Au (10.6)
At the anode:
Au - 3e- - Au1+ (10.7)
Au1+ + 4c1- - Auel; (10.8)
Also at the anode, some gold dissolves in the monovalent state:
Au - e- - Au+ (10.9)
Au+ + 2C1- - Auel; (10.10)
Consequently, some gold is precipitated close to the anode as a fine sludge:
AuCI; + H+ - HAuCl2 (10.11)
and 3HAuC~ - 2Aul + 2HC1 + HAuC14 (10.12)
The formation of monovalent gold is known to occur, since the fine
gold precipitated near the anode (equation 10.12) has been found to be of
higher purity than the anode; also cathode efficiencies based on trivalent gold
(equation 10.6) are often greater than 100% (the electrochemical equivalents
of Au+ and Au1+ are 2,45 and 7 ,35 grams per ampere-hour, respectively).
This formation of monovalent gold is undesirable because the
''disproportionation'' reaction (equation 10 .12) results in excessive cell clean-
ing requirements and also because, at the high current densities employed,
the deposition of monovalent gold at the cathode produces a rough dendritic
deposit. Therefore, considerable investigation work has been carried out on
this aspect, resulting in various changes to the operating data previously
reported. The present operating data are given in Table 10.3.
Anodes are produced in the refining room, by pouring chlorine-refined
gold into vertical moulds. The cathode starters are rolled from specially cast
billets of recycled cathode. Fresh electrolyte is made up by dissolution of
refined gold in aqua regia.
Copper, zinc, platinum and palladium are the major impurities which
are soluble in the electrolyte, and because they do not deposit at the cathode
they accumulate until the electrolyte must be discarded. The insoluble im-
purities, mainly iridium, osmium, ruthenium and silver chloride (the elec-
trolyte is saturated with silver), fall to the bottom of the cell and are remov-
ed as a sludge. The sludge is largely metallic gold and is returned to the refin-
ing room and refined back to anodes. All anode scrap is recycled, so in this
627
Table 10.3. Operating data - Electrolytic gold refining.
Electrolyte Au content, g/l 80 - 100

Free HCl content, g/l 80 100
Temperature 60°C
Circulation Natural convection
Heating Immersion element
Electrical Anode current density, A/m2 800
Current per cell, A 200
Voltage per cell, V 1,0 - 1,5
Electrode spacing, mm 100
Anodes Composition 9960 fine
Mass, kg 16,0
Dimensions, mm (L x Bx Th) 230 x 280 x 12
No. per cell 2
Life, hr 22
Cathodes Composition of starters 9999 rolled gold
Dimensions, mm (L x Bx Th) 320 x 74 x 0,5
No. per cell 12 (3 rows of 4)
Final cathode weight, kg 1,0
Final cathode purity 999 fine
Cells Material Glazed porcelain
Dimensions, mm (L x Bx H) 465 x 405 x 250
4 banks of 10
way the insoluble P.G.M.s accumulate in the circuit until it becomes necessary
to treat the sludge for their removal.
All the anodes and cathodes are renewed daily. The cathode deposit ex-
hibits a very open porous structure, so in order to remove entrained elec-
trolyte completely, a thorough washing procedure is necessary, comprising
two hot water washes, a sodium thiosulphate wash (to remove silver chloride)
and two further hot wat~r washes. The final product is delivered to the refin-
ing room where it is melted and cast into bars for delivery to the Coin Blank
department or, occasionally, as required, cast into special fine gold bars to
be sold for industrial use.
10.4.2.6 Fume recovery and sweep treatment

The nature of the various processes in the Gold Refining Branch and also
the Silver Refining Branch (see the following section) dictates that comprehen-
sive ventilation measures be applied to the working areas, particularly the
furnaces. Furthermore, because the fumes are very rich in gold and silver,
the fume and dust collection equipment must be very efficient. Unfortunately,
because of the chlorides and moisture in the gas stream, a fabric filter can-
not be used, so in this case electrostatic precipitators were selected.
Fumes from all areas are drawn into a common flue delivering to three
separate banks of precipitators. Basically the three units are of similar design,
being tube and wire types, with mild steel wires suspended in 3 m long,
628
Figure 10.10. Electrolytic gold refining plant.
250 mm diameter tubes. Wire current is 1 mA at the maximum potential of

50 OOO V and the gas velocity through the tubes is 0,8 ml sec. The total volume
extracted and treated is 31,3 m3I sec, made up as follows:
Melthouse 2,8 m3/sec

Refining 14,3 m 3/sec
De-golding 7,2 m3/sec
Silver bar casting 2,3 m3/sec
Silver anode casting 4,7 m 3/sec
629
All fume and dust collected from the precipitators is delivered to the
smelter plant, and in this way some 350 kg gold and 1 OOO kg silver per an-
num are recovered.
Spent crucibles, furnace linings, chlorine pipes, borax slags etc. are col-
lected in the "sweep" room where they are crushed, rod milled with the ad-
dition of zinc dust to reduce chlorides and passed over a spiral concentrator.
The metallics are returned to the refinery, while the fines are sent to the
smelter plant as "refinery sweep".
10.4.3 Silver Refining Branch

10.4.3.1 Flow sheet
As shown in Figure 10.11, the quenched chloride is delivered to the silver
leaching department, where the silver is separated from the base metals by
a leach/reduction process. The resultant silver sponge is cast into anodes and
sent to the electrolytic silver plant for final refining.
Dore bullion received from smelter plant and silver bullion purchased
from depositors are also refined electrolytically.
The electrolytic silver crystals are finally melted and cast into silver bars
and delivered to the Administration department for sale to commercial dealers
and banks. Occasionally, unmelted silver crystals are also sold.
Dore bullion
Silver ex smelter Quenched chlorides
bullion ex ORB
receipts 30% Au
70% Ag
Electro- Anodes Lea~ing Aqueous

refining 99% Ag reduction effluent
Sludge to Ag crystals
ORB
Cementate
(base metals)
Admin. Dept. I Sm!lter I

For sale
Figure 10.11. Flow sheet of the Silver Refining Branch (S.R.B.).
10.4.3.2 Leaching and reduction

Depending on the bullion refined, crude silver chloride received from the
G.R.B. assays within the ranges:
630
AgCl 48 70%
Cu Cl 10 30%
PbC~ 1 7%
NaCl 10 20%
The sodium chloride is formed during the de-golding operation and, be-
ing highly soluble, is easily removed during leaching, so it is necessary here
to describe only the leaching of copper and lead, and the problems of silver
chloride solubility.
A detailed flow diagram for the leaching and reduction is given in Figure
10.12, which shows that the main leaching step actually takes place in the
250 kg
molten chlorides
,___s_a_m~p_le_ _,.. Lab
350 I H 2 0 + 25 I HCI
Quench tank
+ 3 kg NaC10 3
Scrap iron
100 1 H 20 at 60°C Wash

+ 1 1 HCl (x!O) Pan filter
liquor
3inc slurry Conical tank

Smelter
150 I Hp at 60°C 1) Reduction
+ 1 1 HCI (X 3l 2) Wash
Anodes
Electro-refining
Figure 10.12. Flow sheet for the silver leaching and reduction process.
631
de-golding section during the quenching of the molten chlorides. To the 350
litres of water in the quench tank are added 25 litres of 30% hydrochloric
acid and 3 kg sodium chlorate. This causes the insoluble cuprous chloride
to be oxidised to the highly soluble cupric chloride, according to the follow-
ing reaction, which is very rapid:
6CuCl + NaC103 + 6HC1 - 6CuCl2 + NaCl + 3Hz0 (10.13)
At the same time the sodium chloride in the chlorides is dissolved and
this in turn increases the solubility of any unreacted cuprous chloride. Most
of the copper is thus leached during the quenching operation. The remain-
ing traces are removed during the subsequent washing operations.
The contents of the quench tank after quenching some 250 kg of molten
chlorides are deposited on to a pan filter and the filtrate pumped to waste,
via scrap iron cementation tanks.
The solids on the filter (i. e. the leached silver chloride) are given a wash
with approximately 100 litres of water at 60°C acidified with hydrochloric
acid. The solubility of lead chloride is relatively high under these conditions
of low chloride ion concentration and high temperature, but six to ten such
washes are required to give a filtrate free of lead and copper (as tested with
NH4 0H).
The solubility of silver chloride increases at high NaCl and HCl con-
centrations and also with temperature, so although most of the silver dissolved
in the effluent is recovered from the cementation tanks, it is advisable to
control these conditions carefully.
Traces of gold exist in the crude silver chloride and some is dissolved
during the leach process by chlorine which is formed by reaction between
the hydrochloric acid and sodium chlorate thus:
6HC1 + NaC103 - 3C~ + NaCl + 3H2 0 (10.14)
Any dissolved gold is ultimately precipitated in the cementation tanks.
The next step is the reduction of the silver chloride to metallic silver
by zinc dust. The six reaction vessels used are 550 litre rubber-lined conical
tanks with a more acutely angled second conical bottom section (Figure
10.13). A tank is first part filled with approximately 100 litres of water, then
a charge comprising 250 kg of leached, washed chlorides is added. The charge
is agitated by rubber-covered mechanical stirrers and air which is admitted
through 25 mm diameter pipes from a manifold below the tank. The
stoichiometric amount of zinc required is calculated from the reaction equa-
tion, which is:
2AgCl + Zn - 2Ag + ZnC~ (10.15)

The zinc dust is added in slurry form over a period of 30 minutes. Care
632
Figure 10.13. Silver leaching and reduction plant.
is needed, since the reaction is very rapid and exothermic. At the end of the
reduction period, hydrochloric acid is added to pH 2 to dissolve any surplus
zinc and the agitation is stopped. The supernatant liquor is decanted through
a valve on the side of the tank, leaving the fine silver behind. Three washes
with hot acidifed water are given, before discharging the silver sponge on
to a pan filter via the bottom discharge valve of the tank. The silver sponge
from the pan filter is dried in a centrifuge before being charged to a 300 kg
capacity gas-fired tilting furnace, where it is melted and cast into anodes for
the electrolytic plant.
10.4.3.3 Electrolytic silver refining

Approximately 5 500 kg of anodes from the silver leaching section plus over
1 OOO kg from the smelter are refined per month in the electrolytic silver plant.
The relatively pure leach anodes are refined in vertical (Moebius) cells
(Figure 10.14), whilst the smelter anodes with their high gold content are
refined in horizontal (Thum) cells (Figure 10.15). In both types of cell, the
basic electrochemistry is the same and the operating data are summarised
in Table 10.4. The significant difference is the gold content of the anodes
and hence the amount of sludge fall-out (gold is insoluble in nitrate solutions).
633
Figure 10.14. Removing silver crystals from vertical (Moebius) cells.
The theory of silver electro-refining is basically simple by virtue of silver's

position in the Electromotive Force Series, silver being more noble (elec-
tropositive) than the electrolyte impurities copper, lead, iron, zinc etc. and
so is plated preferentially. However, it is necessary to ensure that the elec-
trolyte cannot be depleted of silver ions at the cathode, since copper would
then plate out. In the horizontal cells, the convectional flow suffices, but
in the vertical cells the electrolyte must be constantly agitated by a circula-
tion pump and the cathode scrapers. Daily analysis ensures that there is suf-
ficient free acid present to dissolve chemically some anode silver and so main-
tain the silver tenor. Left to itself a slow depletion would result, due to the
634
Table 10.4. Operating data - Electrolytic silver refining.
Moebius Thum
Type of cell (vertical) (horizontal)
Electrolyte Ag content, g/l 30 - 40 30,- 100

Cu content, g/l 4 - 70 4 - 80
Pb content, g/l 5 5
Free HN0 3 content, g/l 10 20
Temperature 30°C 30°C
Circulation Scrapers Convection
& pump
Electrical Anode current, A/m2 390 290
Current per cell, V 2,0 3,0
Electrode spacing, mm 55 110
Anodes Composition 990Jo Ag 700Jo Ag- 300Jo Au
Mass, kg 11 8
Dimensions, mm (L x Bx Th) 290 x 220 x 20 400 x 220 x 8
No. per cell 8 (4 rows of 2) 1 (consists of six
8 kg slabs)
Life, hours 48 24 - 36
Cathodes Material 316 S/S 316 S/S
Dimensions, mm (L x B) 290 x 480 800 x 780
No. per cell 5 1
Production, kg/ day/ cell 19 11
Purity of crystals 99,93 - 99,99% 99,90 - 99,99%
Cells Material Fibre glass Fibre glass
coated PVC coated PVC
Capacity, litres 300 150
fact that silver plus impurities dissolve at the anode, whilst only silver deposits
at the cathode. Thus it is necessary to maintain a higher free acid level in
the horizontal cells where the anodes are less pure.
Apart from copper nitrate which is deliberately added to fresh electrolyte
to increase its conductivity, the only electrolyte addition reagent used is tar-
taric acid, which is added to the horizontal cells to produce harder, denser
and larger crystals, although the exact mechanism of this effect is not known.
In the vertical cells, the anodes are suspended in pairs inside an anode
bag made of polypropylene, mounted on a PVC frame. The sludge collects
in the bottom of the bag and is removed monthly. Anode scrap amounting
to less than 15 % of the original anode mass is transferred into special cells
equipped with anode baskets made of perforated titanium plate, the titanium
carrying the current to the residual anode pieces, enabling them to be com-
pletely dissolved. The deposit on the cathodes is crystalline and dendritic and
is removed by mechanical scrapers. The crystals fall to the bottom of the
cell and are removed daily by scraping them up the inclined side of the cell
into a stainless steel trolley.
635
Figure 10.15. Horizontal (Thum) electrolytic silver cells.
In the horizontal cells, a stainless steel sheet on the bottom of the cell
serves as the cathode; suspended above it is a PVC-framed basket, covered
with a polypropylene cloth. The anodes are laid flat on top of the cloth and
the electrical connection is made by a 100 mm diameter carbon rod. In this
case, the anodes dissolve completely and there is no recycle, but the high
sludge fall-out dictates that the cloth must be cleaned off every three to five
days. The cathode crystals are scraped off the bottom of the cell daily.
Sludge from both sets of cells is mainly gold and is washed off the
bags/cloths with spent electrolyte and filtered in a Buchner filter. Thereafter
it is dried, melted and cast into bars which are returned to the gold refinery
for chlorine refining.
Fresh electrolyte is made up by dissolving silver crystals in nitric acid
and remains in use until the lead level has built up to around 5 grams per
litre, at which point it must be discarded. In this case the silver is precipitated
with sodium chloride and the copper-bearing liquor is sent to the cementa-
tion tanks; the precipitated silver chloride is sent to the leach plant.
Crystals from both types of cell are washed in a centrifuge, first with
1 OJo HN03 solution and then with further water washes. The crystals are
then melted into standard 34 kg bars, which assay:
636
Figure 10.16. Silver bar casting.
Silver 99,9% minimum Iron 10-15 ppm

Oxygen 50-70 ppm Lead 1-2 ppm
Copper 5-15 ppm Zinc 1 ppm
The melting furnace is a 2 OOO Hz, tilting, induction furnace. Crucible

capacity is 342 kg and 1,25 hours are required to melt the charge, using 100
kW at 1 200 V (Figure 10.16).
10.4.4 Coin blank production

10.4.4.1 Historical
Historically, coins have all had a face value given in the currency of the issu-
ing country. However, in 1966 provision was made to issue a South African
coin with a face value of one ounce of fine gold. Krugerrands were produc-
ed in limited quantities in 1968 and 1969. Towards the end of 1969 the
Chamber of Mines of South Africa decided to increase the production of
these coins and to sell them internationally. Rand Refinery Limited under-
took to produce the gold blanks, and started production in February 1970
at the rate of 2 OOO blanks per day. These initial blanks were produced on
equipment borrowed from the South African Mint and an old refurbished
rolling mill.
637
During the next 10 years demand for the coins rose and, in order to keep
pace, production was increased in stages to 40 OOO blanks a day. This was
done largely by scaling up the existing processes, but additional equipment
such as a locally manufactured rolling mill was purchased.
In 1980 production of three smaller coin denominations containing
respectively a half, a quarter and a tenth of an ounce of gold was started.
The small size of these coins, particularly the tenth, demanded very close
dimensional tolerances in the rolled strip. This meant that for the one-tenth
Krugerrand only about 8% of the gold cast into strip finally appeared as
coin. To overcome this high scrap recycling problem and to permit produc-
tion of large quantities of small coins, it was subsequently decided to install
modern high precision rolling equipment.
A prerequisite for this type of equipment is that the strip should be as
long as possible, and certainly much longer than the starting length of 500
mm that was being produced at the refinery. This meant that a continuous
casting machine would also be needed. Additional advantages expected from
a continuous casting machine were much higher quality strip and the ability
to produce this in a wide range of widths and thiclrnesses. To accommodate
this equipment, a new building was planned for the production of blanks
from gold-copper alloy and, in addition, provision was made for the min-
ting of coins at the refinery, should this be required. Construction started
in September 1982 and the first production runs took place in October 1983.
The installation of the new plant has considerably improved produc-
tion efficiencies, particularly for the smaller sized coins. Compared to the
old plant, this has enabled more than four times as many tenth Krugerrands
to be produced for every kilogram of alloy cast. The casting, rolling and
punching operations all produce far less scrap, resulting in improved securi-
ty. The overall result is that production rates of the smaller sized coins have
been increased and the number of staff reduced.
10.4.4.2 Process description
Refined gold
Copper
Rejects
Weigh
S.A. Mint - - - - l &
inspect
Figure 10.17. Flow sheet for the production of coin blanks.
638
The new coin blank production process is summarised in Figure 10.17. To

produce the required 22 carat alloy containing 91,670Jo to 91, 720Jo gold, refin-
ed gold from the Gold Refining Branch is weighed and the calculated amount
of copper added. The charge is melted in a 3 OOO Hz induction crucible fur-
nace and, when molten and well stirred to ensure uniform composition, is
sampled and poured directly into a holding furnace. The holding furnace
is also a 3 OOO Hz induction furnace with a cylindrical crucible. The bottom
of the crucible comprises a set of graphite dies between water-cooled copper
blocks. These dies, which are the heart of the continuous casting machine,
have shaped channels to give the required cross section of alloy strips. Molten
metal flows by gravity to fill the channels and the cooling system causes the
metal in the channels to set. The solid metal is then withdrawn by a few
millimetres, allowing more liquid metal to flow into the die, which in turn
sets and is withdrawn. In this way a continuous strip of solid metal is drawn
in small steps from the bottom of the holding furnace. The strip is coiled
and cut when the mass approaches 600 kg. Continuously cast strip is ideal
for coin production as it has a very uniform cross section and virtually no
defects.
The coils are weighed and passed to the rolling mill. The mill's prin-
cipal features are two coilers which allow the strip to be rolled in one piece,
from one coiler to the other, and thickness gauges which automatically ad-
just the roll gap to maintain consistent thickness of the product (Figure 10.19).
Strip thickness is reduced by 200Jo at each pass through the rolls, so after
six to eight passes the strip is reduced from its original 8,0 mm to 2,34 mm
for production of Krugerrands, or 0,94 mm for tenth Krugerrands. By keeping
the allowable variation in thickness to a minimum, the number of blanks
rejected as under- or over-weight is reduced.
The strip, now at the required thickness, is weighed again and taken
to the punch section. Here the coils are straightened and fed into a punch
which produces four blanks at each stroke. The blanks are tumbled to remove
sharp edges and oil, then dried and counted into bags.
Bags of blanks, each containing 500 one-ounce pieces, are delivered to
the weighing and final inspection department. Here the blanks are fed one
by one into automatic sorting machines. The blanks are weighed to within
a milligram on an electronic balance and the mass of each blank is fed into
a microcomputer programmed to determine the mass category of the blank.
The microcomputer then operates the appropriate gate through which the
coin passes into one of seven buckets according to category, and controls
the feeding of the next blank into the machine.
The blanks are thus sorted by mass. The first bucket contains blanks
which are of an acceptable mass; for Krugerrands this is between 5 and 50
milligrams over the nominal one ounce of gold. The next bucket contains
blanks which are either below the minimum mass, or are too heavy for grind-
ing to the correct mass, and are therefore rejected. The remaining buckets
contain blanks which vary from 50 to 250 milligrams overweight. These blanks
have their edges ground by a set amount to bring the mass down into the
"correct" range. The edges are ground and not the faces, in order to avoid
639
Figure 10.18. Continuous casting of Krugerrand alloy strip.
scratches on the faces which might show up on the final coin. A similar pro-
cedure is followed for the other denominations.
The ground blanks, together with the coins of acceptable mass from the
first bucket, are given a final mass check and a visual inspection of the sur-
face and edges to ensure that they are free of any faults which could mar
the finish of the final coin. The blanks are then weighed and counted into
bags for despatch to the South African Mint where they are minted.
640
Figure 10.19. The precision strip rolling mill and coiler.
10.4.5 Smelter
10.4.5.1 Flow sheet
Gold mines and other processors of gold, including Rand Refinery itself,
produce various gold- and silver-bearing by-products from which the gold
and silver are most efficiently recovered by pyrometallurgical processes. These
materials are thus delivered to the smelter department of Rand Refinery where
the treatment process is generally as shown in Figure 10.20, the end product
being Dore bullion which is delivered to the refinery for electrolysis.
There are six basic operations involved: sampling, blending and sinter-
ing, blast furnace smelting, cupellation, pan furnace smelting and fume
collection.
10.4.5.2 Sampling
Over 300 tons of material are received for treatment every month, contain-
ing some 450 kg gold and 1 200 kg silver. By far the largest component is
borax slag from mine smelting operations, followed by miscellaneous con-
centrates, ashes, bricks, pots and liners, activated carbon, jewellers'
''sweeps'', base bullion bars, computer scrap and various sweeps from the
refinery's gold and silver refining branches.
Considerable effort has been put into developing reliable and accurate
sampling methods for these materials. The first prerequisite is that the material
641
By-products
from mines etc.
Assay & lab.
Sampling
Bag
Fluxes
Blending filter
------i 7"'"' II
I
Sinter plant --,..--:..7 - - - -->- - - - 1- - - --- - - - - - Fufe
,.."'!' A I
,..,.. I I
--- ,...._--~~~,..~
Blast furnace
Lead
I
Cupe!
I
Pan
Coal
I furnace
Litharge
furnace
Matte Slag bullion
1
Granulate
Dore
anodes
Barren
lead
t
~
t
Sale Sale Sale
Figure 10.20. Flow sheet of the smelter department.
be fine enough to pass through a 2 mm screen; therefore all oversize materials

are crushed in a Symons crusher prior to sampling. To sample accurately
material with a bulk density of 2 - 3 tl ml containing gold bullion particles
of density up to 18 t/ml is not easy, and a further prerequisite is that the
gold particles be distributed uniformly throughout the bulk. To achieve this
condition each lot of material is charged via a pneumatic conveying system
to either a 2 m3 or a 6 ml conical mixer and allowed to mix for approx-
imately fifteen minutes. The mixer is then discharged into 0,2 ml drums and
two "rod samples" are taken from each drum. The rod samples are ac-
cumulated to give a 20 - 30 kg sample representative of the whole. In theory,
much larger samples are necessary to achieve the desired accuracy of± 2,5%
but experience has justified the technique described. Experience has also pro-
ven the necessity of achieving thorough mixing; providing this is done, the
rod sampling technique is adequate and there is no need for elaborate sample-
splitting operations. After mixing and sampling, the complete lot is weighed
to obtain its wet mass. Two small moisture samples are withdrawn from the
main sample, weighed, dried in an electric oven and reweighed to enable the
true dry mass of the material to be calculated. The remaining sample material
is coned and quartered and split into 'A' and 'B' fractions which are then
reduced to suitable form for assay, as shown in Figure 10.21. Screening of
the metallics is essential to leave the bulk of the sample more uniform in
642
'A' SAMPLE 'B' SAMPLE

9 kg 9 kg
!
Dry at 105°C
+ 30 metallics
t
~------..---30 Tyler mesh
I
i
Cone & quarter
i
Cone & quarter
to 500 g to 500 g
t t
Crush in Crush in
disc mill disc mill
+ 120
t +
l
metallics 120 Tyler mesh 120 Tyler m e s h l
- 120 fines -1201 +120t
To assay & lab. (retained)
N.B. A 1 & B1 sent for assay

A 2 & B2 retained for queries
Figure 10.21. Flow sheet for the preparation of smelter samples.
value. The metallics and fines are assayed separately and the assays combin-
ed according to the masses of the different fractions.
The above sampling method applies to the majority of materials receiv-
ed, but metallics and bars require special techniques normally involving
melting and fluxing with galena, followed by sampling of the lead metal and
matte produced.
After sampling, weighing and assaying, all materials are purchased from
the depositor and stored pending treatment.
10.4.5.3 Blending and sintering

Having reduced all materials to fines to ensure accurate sampling, the next
step is to agglomerate these fines into a form suitable as feed for the blast
furnace. The sinter process was chosen in preference to pelletising or bri-
quetting because of the superior physical characteristics of sinter with regard
to size, strength and porosity.
All purchased materials are analysed not only for gold and silver but
also for the chemical components which affect blast furnace and sinter plant
performance. These are silica, iron, lime, alumina, zinc, lead, copper, sulphur
and arsenic. From the stock of materials available, a sinter charge of roughly
250 tons is calculated, chemical composition of the total charge being ad-
643
justed by the addition of iron mill scale, pyrite and coke breeze. The charge
calculation is governed by the requirements of good quality sinter, minimum
barren flux usage and good blast furnace slag. Generally speaking, the rules
are:
i) Borax slags must comprise less than 35%. Above this figure the sinter
bed slags over, restricting the sintering action.
ii) The approximate composition must be:
Si02 , 12-20%; Fe, 18-24%; S, 6-8%; Pb, 8-10%.
iii) Alumina must be less than 80Jo to avoid forming a viscous blast fur-
nace slag.
iv) Excessive zinc must be avoided, also to prevent the formation of viscous
slag.
The total calculated charge is divided into identical 15 ton batches. Each batch
must be thoroughly mixed to give a uniform feed to the sinter plant and this
is done by heaping all materials on the floor and turning over the heap several
times with a front-end loader, dust being suppressed by means of water sprays.
This method of blending the charge has proved to be the simplest and most
economic way of achieving the required throughput of some 40 tons per day
of blended feed.
The blended mix is loaded into drums which are in turn discharged by
a fork-lift truck into a hopper from which feed material is elevated to the
sinter plant.
The object of the sintering process, as previously stated, is to agglomerate
the feed prior to smelting in the blast furnace. In practice, this agglomera-
tion is achieved by igniting the fuel constituents in the blended mix laid on
the sinter strand. The heat generated by the combustion reaction is suffi-
cient to soften many ingredients and to produce localised melting and slag-
ging of the lower melting point materials such as litharge, borax slag and
matte. The resulting product, after the hot zone has passed through the bed,
is a hard porous frit, ideal for the blast furnace. The two fuels used are pyrite
(FeSJ and carbon in the form of coke breeze, although recently breeze has
been largely replaced by activated carbon, which is now being received from
the mines in increasing amounts.
The sinter plant itself is a small conventional down-draught machine
as depicted in Figures 10.22 and 10.23. One of the main features is the use
of - 25 + 6 mm recycle material as a hearth layer to protect the grate bars.
The important parameters are as follows:
Feed moisture= 9% - determined by carbide bomb apparatus.
Bed depth= 175 mm + 50 mm hearth layer
Strand width= 1,0 m
Strand speed=0,18 m/min (but variable)
Down draught= 1,9 m3/sec each windbox, at 250-450 mm w.g.
Rated output= 35 t/ day product + 10 t/ day recycle.
10.4.5.4 Blast furnace smelting

The general theory of blast furnaces can be found in most metallurgical text-
644
filter
-25 mm +6 mm
-6 mm
Skip
Blended---_,_.---
feed material
Figure 10.22. Schematic layout of the sinter plant.
books, so it is more important to describe here the metallurgical principles

governing Rand Refinery's specific operation.
The basic process is the reduction of the lead oxides in the charge to
metallic lead, which percolates through the charge column and collects the
precious metals. At blast furnace temperatures the basic oxides combine with
silica and other gangue materials to form slag which, having been freed of
its precious metal content by the lead, is low enough in value to be discard-
ed. Sulphides present in the charge cause the formation of a matte phase
which serves to recover the copper content of the charge and also collects
silver.
The important reactions are:
2C + 02 - 2CO (10.16)
PbO + co - Pb+ co2 (10.17)
Reaction (10.16) generates most of the heat within the furnace, so there
must always be sufficient coke included in the charge, not only to ensure
a stable operation, but also to prevent the formation of carbon dioxide which
645
Figure 10.23. The sinter strand.
occurs in preference to carbon monoxide at lower temperatures, i.e.
(10.18)
The proportion of carbon converted to CO rather than C02 is most im-

portant since it affects the degree of reducing conditions and hence the ex-
tent of reaction (10.17). This explains the success of the oxygen enrichment
technique in which increased oxygen and reduced nitrogen contents of the
air blast cause more efficient combustion, hence hotter conditions and a
higher CO: C02 ratio, thereby increasing the lead formation by reaction
(10.17). A further benefit of oxygen enrichment, indeed the main reason for
its use, is that the more efficient combustion permits the use of less coke
per unit charge and allows a higher feed rate.
The composition of the slag leaving the furnace must be carefully con-
trolled. The slag formation reaction can be written:
x Si02 + y FeO + z CaO - x Si02 .y FeO.z CaO (10.19)
In practice the ratio of (y + z):x, i.e. the amount of iron and calcium
oxides combining with the silica, must be sufficient to displace lead oxide
from the slag. Unless this is so, the silver and gold in the slag will be unac-
ceptably high. The ratio of x: y: z is also very important to ensure a fluid,
low melting point slag. Figure 10.24 illustrates this point.
646
Si02
coo
rv257QO
Figure 10.24. Ca0.Fe0-Si02 phase diagram (after Levin et al., 1969).
It can been seen that the lowest melting point slags are found where
Si02 : FeO : CaO is 35 : 46 : 18, corresponding to an olivine-type slag.
Because of the other constituents present such as zinc oxide, alumina and
magnesia, which raise the melting point, and borax, which lowers the melting
point, slag analyses obtained in practice are as follows:
Si02 20 240Jo MgO 20Jo

FeO 35 390Jo MnO 1OJo
Cao 11 140Jo Cu 0,30Jo
Alp 3 60Jo s 1OJo
B/)i 50Jo Nap 2,50Jo
Pb 1OJo
Figure 10.24 clearly shows that deviating too far from these values moves
the slag into areas of rapidly increasing melting point and hence difficult
furnace operations.
The physical condition of the charge is also very important. All com-
ponents should be of approximately the same size to avoid segregation of
647
the materials as they are fed into the shaft and to ensure an even distribution
of the gases as they rise through the column. If "channelling" of the blast
is allowed to occur, the hot zone of the furnace rises from its normal level
(just above the tuyeres) and furnace operation becomes difficult. This is why
the sinter quality is so important; the sinter must be strong enough to with-
stand the weight of the burden above without becoming finely crushed.
Rand Refinery's furnace is a modified lead blast furnace with a rec-
tangular hearth of area 1,22 m2 and four tuyeres on either side. The hearth
slopes towards the lead taphole, which is in the centre of one side, whilst
the slag overflows a slag notch at the front end of the furnace. If necessary
the slag can be run out of a notch at the opposite end. The furnace is water-
j acketed to a height of 1,5 m then lined for the remaining 4,5 m with high-
alumina firebrick. A fan delivets up to 90 m3lmin of air to the tuyeres via
a bustle pipe into which oxygen is injected to an enrichment level of 2,50Jo
(i. e. the air blast is 23,5% oxygen). At the top of the shaft, a cast iron collar
is fitted through which the charge descends. The off-gases are drawn from
around the collar into a side offtake and thence to the flue system.
The furnace charge, generally comprising coke, sinter, litharge, return
matte and the requisite fluxes, is carefully calculated and adjusted daily to
maintain the slag, lead and matte compositions within the correct limits.
Metallic iron (turnings and chopped discarded drums) is added to reduce lead
from the matte while limestone and iron mill scale are added to adjust the
slag composition. The various ingredients are each weighed into a scale car
and transferred into a feed skip which tips directly into the top of the fur-
nace. Slag and matte overflow the hearth continuously via the slag notch
into three settling pots in series. Slag overflowing the third pot is granulated
in a water launder, conveyed to a storage bin and sold for use as a grit blasting
medium or as a colour additive to glass. The assay of the slag is normally
around 1 git Au and 50 git Ag.
Matte, being denser than slag, collects in the bottom of the settling pots
and is tapped periodically into pig moulds and recycled to the furnace, both
directly and via the sinter, until its copper content exceeds 30%. At this point
the lead tapping becomes difficult due to copper dross freezing in the taphole.
The matte is therefore withdrawn from the circuit and accumulated for sale
to a suitably equipped refinery which can extract the copper, lead, gold and
silver. Matte for sale typically analyses 30% Cu, 15% Pb, 20% Fe, 20% S,
0,5% As, 400 git Au and 8 OOO git Ag. The lead bullion is tapped every
four hours into pig moulds and the solidified ingots, each weighing 25 kg
and assaying typically 1OJo Au and 3,50Jo Ag, are delivered to the cupel
furnace.
Blast furnace throughput is around 40 tons of total charge per day, in-
cluding 10 tons of new material. The furnace is operated for campaigns of
three months after which the flues must be cleaned and the accretion in the
shaft removed.
10.4.5.5 Cupellation
The cupel is an oil-fired, brick-lined furnace as shown in Figure 10.26, with
648
Figure 10.25. Blast furnace slag and matte stream.
a tiltable, removable hearth (referred to as a "test") made of rammed

magnesite. The lead bullion is loaded into the feed chute where it melts and
trickles down into the hearth. A low pressure jet of air across the hearth causes
the lead to oxidise to litharge and directs this litharge into a shallow channel
and thence into a collecting pot. Any other base metals are also oxidised and
removed either as fume or with the litharge. The silver and gold are not ox-
idised and, being denser than lead and litharge, settle in the bottom of the
hearth and accumulate until the hearth is full. At this point, lead addition
is stopped, the temperature raised and fluxes sprinkled over the melt to remove
the last traces of lead as a slag. The cleaned Dore bullion is ladled out into
flat anode moulds. More than one ton of these anodes, assaying roughly 70%
Ag, 30% Au and up to 0,2% Pb and 0,6% Cu, are delivered each month
to the electrolytic silver section.
10.4.5.6 Pan furnace smelting

The pan furnace is a small, oil-fired reverberatory with a hearth area of
6 m2 • In the past this was used for smelting purchased materials using a lead
collection technique similar to that of the blast furnace. However, the residue
slag was never low enough in value to be discarded and so had to be retreated
in the blast furnace. Because of the adverse economics, this process has been
discontinued and the furnace is now used only to reduce surplus litharge from
the cupel furnace back to lead metal. The process is quite straightforward,
involving the charging of batches of 5 t litharge plus 0,6 t coal, firing for
four hours and tapping into pig moulds.
The lead is sold to a lead works, where the residual gold and silver are
649
Bullion bath
Air
Litharge
ELEVATION
Oil burner
Furnace lining Furnace lining
Furnace floor
Inspection door
E
-.
0 -----
Air
..-,..~----
. ...........:-:-.,..,
To flue
1,60 m
PLAN
Figure 10.26. Diagram of cupel furnace.
650
Figure 10.27. The cupel furnace, showing litharge running into collecting pot.
extracted by drossing with sawdust and zinc in a Parkes kettle, the drosses
being returned to the Refinery's blast furnace circuit.
10.4.S. 7 Fume collection

Because of the lead hazard, all operations in the smelter must be vented to
efficient gas-cleaning equipment before the gases are released to atmosphere.
Their precious metal content makes it economically imperative to recover
the dust and fume. A further contaminant in the gas stream is sulphur diox-
ide; its concentration is too low to permit economic extraction and it is
therefore dispersed via a 61 m stack. Fumes from all the furnaces and the
sinter plant are extracted via an underground flue system leading to four iden-
tical electrostatic precipitators of the tube and wire type. The electrodes are
1,6 mm-diameter titanium wires suspended in 280 mm-diameter tubes. Elec-
trode current is 1 mA at a maximum potential of 65 OOO V. The linear gas
velocity is 0,76 mlsec.
Since the dust consists mainly of heavy metal oxides which are difficult
to ionise, the precipitators are only about 60% efficient, so the gas stream
from the precipitators is passed to a bag filter plant. Precipitated dust assays
typically 25% Pb, 11 OJo Zn, 400 git Au, 800 git Ag, and is recycled to the
sinter charge. The bag filter has a designed capacity of 34 m3I sec and con-
sists of 48 compartments, total cloth area 2880 m 2 , two compartments at a
651
time being cleaned "off line" by the action of a pulsating reverse air flow.
The low air: cloth ratio of 0,7 m/min results in a very high filtration effi-
ciency of around 99,9% and minimum failures of the "Dralon T" filter
sleeves. The bag dust recovered, assaying typically 40% Pb, 15% Zn, 60 git
Au, and 800 git Ag, is pelletised and recycled to the sinter charge.
10.4.6 Assaying and chemical analysis

It is essential that the assay and analysis of all samples are of the highest
integrity because of the value of the materials being handled. Errors in the
assay of purchased or sold materials could cause either the Refinery or the
depositor to suffer financially, while errors in the process control sample
analyses could lead to expensive metallurgical losses and process failures.
The assay and analytical techniques used at the Rand Refinery are fully
described in the book Assay and Analytical Practice in the South African
Mining Industry (Lenahan and Murray-Smith, 1986).
One of the main features of the assay procedure, designed to ensure the
required accuracy, is that routine gold assays are done on a twin-stream
duplicate system to reveal any errors or bias which would otherwise go
undetected. The assay department is therefore divided into two separate assay
offices, each with its own staff, equipment and rooms. Despite the duplica-
tion of labour and equipment, the cost is no greater than running a single
stream with the comparable degree of control.
10.5 Future Developments

In any business it is essential to seek ways and means of improving process
efficiency and the working environment by developing and incorporating new
technology. Precious metal refining is no exception to this rule.
At the time of writing this chapter (1985), Rand Refinery is embarking
on a project to refurbish the smelter plant. The plans include the adoption
of electric furnace smelting to replace the sinter plant/blast furnace route.
This will lead to increased capacity, improvements in air quality index, reduc-
tion in operating costs and a reduction in the labour requirement. For similar
reasons the cupel furnaces will be replaced by a top-blown rotary converter
- a high intensity, energy efficient reactor. The same converter furnace will
also be used for the litharge reduction, simply by changing to reducing con-
ditions. The latest mechanical handling techniques will also be introduced
to save labour and eliminate dust emission. Further, there are medium term
plans to rebuild the refinery, incorporating sophisticated electronic/ electrical
equipment applied to new induction furnaces, improved mechanical hand-
ling and ventilation, and process changes designed to eliminate the aqueous
effluent from the silver refinery. Computers have already found applications
in the analytical field, are currently being applied to the metallurgical ac-
counting system, and in the future will undoubtedly be used for on-line pro-
cess control of the smelting and refining operations.
652
10.6 References
Adamson R.J. (1972). Gold Metallurgy in South Africa. Chamber of Mines of South Africa,
Johannesburg.
Lenahan, W.C. and Murray-Smith R. (1986). Assay and Analytical Practice in the South African
Mining Industry. Chamber of Mines of South Africa, Johannesburg.
Levin, Robbins and McMurdie (1969). Phase Diagrams for Ceramicists, 2nd edn. p.204. The
American Ceramic Society.
653
Chapter 11
Disposal of Residues
G.I. McPhail and J.C. Wagner
11.1 Aim and Scope

No metallurgical textbook would be complete without a chapter dealing with
residue disposal. In this chapter an overview is given of the aspects which
require consideration when planning, operating or closing a tailings dam;
it should therefore provide background knowledge for the effective manage-
ment of tailings dams. Emphasis has been placed on those aspects which re-
quire investigation by specially qualified personnel. In this regard the extent
and nature of the requisite investigations have been discussed and the
minimum qualifications of those undertaking the work have been given.
The bulk of this chapter concentrates on surface disposal, since at the
time of writing underground stope backfilling had only been carried out on
a pilot scale on South African gold mines. It is however recognised that the
volume of tailings sent underground will probably increase and that the
remaining fraction will become finer and more difficult to manage.
It is of particular interest to note that the responsibility for safe and
effective disposal of the tailings underground no longer falls to the discipline
of metallurgy. The overflow or surplus will, however, always be disposed
of on surface, and this is likely to remain within the metallurgical scope.
Particular aspects of surface disposal of residue covered in this chapter
include:
• sizing, siting, designing and operating gold tailings dams;
• legal and environmental regulations governing the establishment, opera-
tion and closure of a tailings dam;
• some operational "do's and don'ts";
• observation and control of the tailings disposal operation;
e present state of the art in closure and rehabilitation of gold tailings
dams.
The chapter has been confined to hydraulically placed gold tailings impound-
ments. Dumps have not been covered, since, in terms of potential for
generating problems, tailing dams have a much more significant history of
operational, environmental and safety problems, they are generally poorly
understood, and require considerably more attention.
655
Plant
Toe starter wall
PLAN
Daywall
Delivery Pool
main
Fence Road Solution
trench
SECTION
"' /
Under drainage
Figure 11.1. Typical layout of a tailings dam.
11.2 Planning
11.2.1 Components of a residue disposal system
By way of introduction to this section on planning a surface tailings disposal
scheme, it is useful to list the various components of a tailings disposal com-
plex and briefly describe their purpose. Consideration of each of the corn-
656
DISPOSAL OF RESIDUES
ponents during the siting, sizing and investigation of potential impoundment

sites is important in view of their influence on cost, area provision and future
operation.
e Containment wall or toewall - a wall, usually of compacted earth, to
contain the tailings either as a total confining structure or during the early
development of the impoundment.
• Underdrainage system - a system of filter drains to control seepage within
the impoundment.
e Decant system - a pipe system to facilitate the removal of clarified water
and stormwater accumulated on the impoundment.
• Delivery system - pipework, valves and discharge points to convey the
pulp to, around and into the impoundment.
e Return water system - a system of dams, sumps, pumps and pipelines
for holding and conveying excess effluent and stormwater from the tail-
ings dam to the plant.
e Stormwater diversion system - a system of trenches and bunds con-
structed around an impoundment to control and divert external storm-
water around the impoundment.
• Stormwater catchment paddocks - a system of paddocks constructed
around the toe of the impoundment to capture, control and store storm-
water and sediment eroded from the side slopes of the impoundment.
• Ancillary system - comprising access roads, power supply system, etc.
These components are illustrated in Figure 11.1, which depicts a plan layout
of a typical gold tailings dam.
11.2.2 Sizing
11.2.2.1 Depositional area requirements
The first step in planning and providing for tailings disposal is the definition
of the required area of ground which will meet the constraints of total
volumetric capacity required for a manageable height of dam and rate of
rise or operating cycle time. At this stage of planning, a maximum height
of 35 - 40 m would be assumed and a rate of rise selected on the basis of
pulp density and anticipated foundation soils. Figure 11.2, as proposed by
Wates (1983), serves as a guide.
Typically, gold tailings have a settled density in a tailings dam of be-
tween 1250 and 1 650 kg/m3 depending on the depth of tailings. A
reasonable figure to assume for sizing calculations is 1450 kg/m 3• The re-
quired tailings dam area can then be calculated very simply from the equation:
Area, m2 = (t.p.m. x 12) I (1,45 x ROR) (11.1)
where t.p.m. = dry tons deposited per month, and ROR = allowable rate
of rise (m/yr).
This area takes no account of ground topography, or of reduction of area
as the dam grows. Figure 11.3 gives an indication of factors which could
be applied to the depositional area of approximately square dams on sites
with a significant ground slope.
657
THE EXTRACTIVE METALLURGY OF GOLQ
b!l
.5
lo-!
o a
~~
~~
3,0 § .g
a
·a >~
I:'
ro ·-::;; "u
0
...."'
>.
Recommended.max. rate of rise /
"'0. :<.eS /
/
"'
~
"'.... e; /
OJ /
s ·~~/
'Q<,O
"'
"'
·;::: 'O<>'
"'
2,0 Q.'1><$> ':>
""'0 ~I .,o"'
~"'.... o1:'P' ~ ~"'°'o/:
"'
::0 { .c'"
ro ~ ·~"'
::: ~o' o<S',
..s / oey:
<'. "-..O'
1,0 v<S',
~<$',
~'I>
r.d. 1,0 1, 1 1,2 1,3 1,4 1,5 1,6

I I I I I I
w/s oo 6,0 3,0 2,0 1,5 1,0
Density of tailings pulp

Figure 11.2. Rate of rise versus tailings pulp density.
The actual rate of rise on a sloping site will tend initially to be above
the calculated figure and thereafter will decrease to just above this figure
when all ground is covered. From this stage the rate of rise will begin to in-
crease again due to the reducing area as the side slopes of the dam are formed.
In the case of paddock dams, at rates of rise above the limits implied
by Figure 11.2, cycle times between lifts become too short to allow suffi-
cient drying and desiccation of each lift. Access becomes difficult and, while
rates of rise in excess of 2,5 m/yr can be tolerated for limited periods of
time, above a height of 2,5 to 5 m, stability problems will eventually result.
658
Area required = Area x multiplying factor
j
1,3
--;;;- I
s., I
'O
.,.... ,I
<!)
;:I
~ ,
><
....0 1,2 I
~
0.
0. I
.... I
~
<lS /
bJl
/
f
.,s
<!)
1,1
v
/
.I
/
....
<t:
/
1,0 /
,,.,,
"
0,5 1,0 1,5 2,0 2,5 3,0 3,5
Average ground slope(%)
Figure 11.3. Correction to dam area for various ground slopes.
The anticipated period of usage of the site may be calculated from:

Time (years) (hmax X 1,45) X (area) I t. p. m. X 12 (11.2)
hmax maximum height, m
area average area, m 2
t.p.m. dry tons deposited per month.
At this level of planning, a maximum height of 35 m would be a reasonable
assumption. Detailed design and the provision of suitably positioned drainage
media could facilitate an increase in this height to 50 m and more, provided
rate of rise constraints are not exceeded. If the period of usage of the site
from the above calculations proves inadequate it is advisable to reverse the
calculation to give the depositional area required to attain a height of 35 m.
This will also result in a lower average rate of rise.
11.2.2.2 Allowances for services

To the above areas should be added a perimeter of sufficient width to allow
for the formation of:
659
e a stormwater cut-off trench (4- 7 m);

• an access road (5 m);
• a delivery/return water pipe servitude (2 m);
• the solution trench (4 - 7 m);
e catchment paddocks (20 m).
Allowing for earthwall widths, a perimeter width of 50 m is common. The
area implications of this perimeter depend on the plan shape of the dam,
which will define the perimeter length. For a square dam of area A the ser-
vice area (SA) would be reasonably approximated by:
SA = 200 (50 + A 0 •5 )m2 for A in m2 (11.3)
11.2.2.3 Return water system

Allowance should be made at this stage for a return water system. In some
cases, however, no return water can be accepted and it is necessary either
to treat the water to meet metallurgical or stream standards, or to provide
for evaporation of the excess water. In the case of evaporation, the required
areas can equal or exceed tailings dam areas and will require specific plan-
ning outside the scope of this chapter.
In most cases a water dam, pumps and pipeline are provided with suffi-
cient capacity to capture and control rainfall under specified conditions. The
area occupied will be dependent on topography and soil cover, but in this
initial planning stage it is reasonable to assume that an area equivalent to
25% of the tailings dam area should be allowed.
11.2.2.4 Total area

A provisional estimate of the area of dam required will be the sum of:
e depositional area whether dictated by rate of rise or by capacity con-
straints;
• the area of services;
• the area of the return water system.
11.2.3 Site selection

The area and volume calculations in Section 11.2. l provide a first indication
of the size of the site required. Using this information and a regional
topocadastral map, it is possible to identify a region of interest within which
candidate sites can be located. This exercise is referred to as "regional screen-
ing" and facilitates the elimination of areas obviously unsuitable owing to
the following factors:
• too far from metallurgical plant;
e topography too steep;
• access too difficult;
• catchment area too large;
• unsuitable geology or mineralisation;
• ground water discharge area;
e important land use zones;
e sensitive ecological areas.
660
Within the areas remaining after the above elimination process, a number
of possible impoundment sites may be identified. For each of these the follow-
ing characteristics should be listed:
o capacity;
• toewall height (based on a maximum rate of rise of 2,5 m/yr);
o distance from the plant;
• land use constraints;
• other salient characteristics that might affect tailings disposal.
A ''fatal flaw'' analysis will eliminate sites flawed by characteristics suffi-
ciently unfavourable or severe to preclude use of the site. Table 11.1 pro-
vides a list of possible flaws.
Table 11.1. Criteria for fatal flaw screening.
Visual
1. Unacceptable visual impact
Land use/ecological
1. Mineralisation (economic)
2. Man-made features, e.g. pipelines, power lines, major roads
3. Actual or potential urban areas
4. Land with high agricultural potential
5. Historic and archaeological sites
6. Important recreation areas
7. Sensitive or unique ecosystems, wild life or fish habitat, endangered species
Airborne release
1 . Dust/ erosion - high wind exposure
2. Proximity to human habitation
Seepage release
1.• Foundation
2. Discharge to ground water area
3. Discharge to streams
4. Flood plain
Stability
I. Topography (too steep)
2. Faults (active)
3. Foundation conditions (poor)
Operational
I. Capacity (too small)
2. Access (too difficult)
3. Technical feasibility (not implementable)
Cost
1. Development cost (uneconomic project)
661
Sites not eliminated in the fatal flaw analysis should be visited and data
gathered about factors such as: visibility, land use, meteorology/ climate,
hydrology, geology, soils, vegetation, and ground water.
A conceptual design for each site should be formulated from which an
estimate of costs for land, construction, operation and closure should be made
and tabulated.
A qualitative evaluation and ranking of the sites subjectively into very
good, good, moderate, poor and very poor, on the basis of site selection
elements such as those given in Table 11.1, can prove useful at this stage
in deciding on those sites to be put forward for detailed investigation. A
minimum of two and preferably three sites should be selected for detailed
analysis. Detailed investigation may involve drilling if the geology and ground
water conditions are of particular concern, as would be the case in dolomites.
Further costing and environmental studies and impact analysis would be car-
ried out to define more accurately the implications of using each of the sites.
Semi-quantitative evaluation and ranking procedures are available; they
are set out in detail in the paper by Caldwell and Robertson (1983). In the
course of these procedures, numbers are assigned to the subjective ratings,
categorised in terms of importance and then weighted. These procedures may
be called for in cases where site selection is particularly sensitive, and various
parties such as government and local authorities, boards of directors, hostile
property owners and environmental action groups require a referable, less
subjective system for justifying the selection of a particular site. In most cases
cost will predicate a site and the subjective analysis will be adequate.
11.2.4 Site exploration

The potential sites from the selection exercise in Section 11.2.3 above should
be explored in detail by registered and experienced geotechnical engineers
and/ or engineering geologists who specialise in investigations of this nature.
A soil and drainage report should be produced, and only on the basis of the
findings in this report should a final decision on the suitability of the site
be taken.
The principal aims of a soil and drainage investigation are to:
e identify and characterise areas of potential foundation problems;
• determine ground water seepage models;
• locate possible construction materials.
The soil and drainage survey should establish the different types of soil
horizons covering the site, and their boundaries. The following data should
be gathered:
G typical soil profiles, depths to bed rock, the position of the ground water
table, preliminary estimates of the permeability of the soils and bedrock,
in situ shear strengths, and consolidation data;
o indications of the presence of dykes, faults, sills and other geological
features;
e the presence of collapsing sands, cavernous dolomites, soft, active or
dispersive clays, etc.
662
The soil survey will often include deep drilling into the regional water table
and the installation of sampling piezometers to establish regional and near
surface ground water quality. These data would form background water quali-
ty data against which changes can be measured once the tailings dam is put
into operation.
Trial holes in the surface soils are excavated using either a large diameter
auger or a backhoe, depending on soil depths, and, from these holes, disturb-
ed and undisturbed samples of the soils are recovered. Since trial holes give
data on soil conditions at specific points only, in certain cases it may be
necessary to use geophysical methods such as gravimetric surveys,
microseismic surveys and resistivity surveys to locate sinkhole areas, depths
to bed rock and water table depths respectively. The actual siting, distribu-
tion and number of holes and the extent of geophysical testing will be decided
by the geotechnical engineer or engineering geologist on the basis of the an-
ticipated characteristics of a particular site. Generally, since the slopes of
walls of residue deposits are critical in the design, exploration holes will tend
to be around the perimeter of the tailings dam.
During the site investigation an appraisal of materials suitable for the
construction of earthwalls and filter drains should also be undertaken. This
involves the determination of compaction characteristics and the gradings
of available sands and gravels.
11.2.5 Pollution control considerations

By virtue of their size, poor ability to resist wind and water erosion and the
hydraulic fill method of construction, tailings dams can, if adequate provi-
sion is not made to control these factors, pose serious long term pollution
problems. •
Around many of the gold mines on the Witwatersrand and elsewhere,
sizeable communities have evolved to serve the peripheral industries required
by these mines. Often housing development encroaches on the mine proper-
ty boundary. Consequently the extent and nature of the potential pollution
hazards have increasingly become factors which can influence the choice of
a particular site.
In the course of the site evaluation, if the factors given in Section 11.2.3
are taken into consideration, the predominant wind directions and meteoro-
logical characteristics of the mine will have been determined. Also, the soil
exploration exercise in Section 11.2.4 will give data from which the poten-
tial for ground water contamination through uncontrolled seepage under the
dam can be assessed. These characteristics of the site fall within the orbit
of certain legislation (see Section 11.3). In planning a site for a new tailings
dam, therefore, particular attention must be given to pollution control con-
siderations. Remedy of pollution problems, particularly ground water pollu-
tion, once these have begun, can be costly, with only limited prospects of
complete success.
663
Table 11.2. Legislation affecting tailings dams.
Act Aspects covered Relevant sections Controlling body
Government Gazette All actions directed Section24 Dept. Environment

of29May1987, at or liable to have Affairs
Notice353. Depart- an influence on the
ment of Environ- environment
ment Affairs Draft
Bill on Environment
Conservation
Water Act No. 54 Safety of dams, Sections 9, 11, 12, Dept. Water Affairs
of 1956, as amend- treatment and dis- 12A, 20, 21, 22, 23,
ed posal of effluents, 24, 30 and Regula-
water use and en- tions No. 287 of 20
titlements, surface February 197 6,
water management, R991 of 18 May
water pollution con- 1984 & R2834 of 27
trol, water care December 1985 as
works, disposal of amended in toto
underground water,
effluent water qua-
lity
Health Act No. 63 Air pollution con- In toto Dept. Health and
of 1977 trol; water pollution Welfare
control
Atmospheric Pollu- Dust control Sections 27 and 28 Dept. Health and
tion Preventio.n Act Welfare
No. 45 of 1965
Soil Conservation Soil erosion control 2,3.(k) and 3 (m) Dept. Agriculture
Act No. 76 of 1969
Conservation of Maintenance of In toto Dept. Agriculture
Agricultural Re- production potential
sources Act No. 43 of land, combating
of 1983 and prevention of
erosion and
weakening or de-
struction of water
sources, protection
of vegetation, corn-
bating of weeds and
invader plants
Physical Planning Zoning and use of Sections 2, 4, 6B Dept. Environment
Act No. 88 of 1967 land for industrial and 8 Affairs
purposes, quarry Dept. Mineral and
development and Energy Affairs
rehabilitation
Weeds Act No. 42 Control of weeds In toto Dept. Agriculture
of 1937 and pest plants
Agricultural Pests Prevention and In toto Dept. Agriculture
Act No. 36 of 1983 combating of agri-
cultural pests
664
Act Aspects covered Relevant sections Controlling body
Nuclear Energy Act Control over radio- Section 5 .1 Dept. Mineral and
No. 92 of 1982 active wastes Energy Affairs
Hazardous Sub- Control of trans- Sections 2, 3 and 29 Dept. Health and
stances Act No. 15 port, dumping, Welfare
of 1973 storage or disposal
of any hazardous
waste, including
radioactive material
Mines and Works Regulations con- Section 12 and re- Dept. Mineral and
Act No. 27 of 1956 cerning protection gulations 2.10, 5.9, Energy Affairs
of public safety, en- 5.10, 5.12, 5.13 and
vironmental protec- 5.14 Dept. Water Affairs
tion, rehabilitation
plans, removal of
topsoil, stream and
river diversion,
waste disposal, re-
vegetation, storm
water and effluent
control and disposal
11.3 Statutory Requirements

11.3.1 Legal considerations
Environmental control legislation in South Africa is widely fragmented among
numerous Acts of Parliament which bedevils the systematic integration of
environmental planning into the design of mines and associated activities.
The environmental aspects which are, to some extent, provided for in ex-
isting legislation are: water use and disposal, water pollution, air pollution,
soil conservation, conservation of vegetation, rehabilitation, and dam stability
and safety.
Existing legislation which provides for environmental control is listed
in Table 11.2.
The main Acts controlling the design, construction and operation of a
gold residue disposal facility are the Mines and Works Act (Act 27 of 1956)
and the Water Act (Act 54 of 1956) as amended from time to time. Relevant
aspects of these Acts are discussed below.
11.3.1.1 The Water Act

Although many sections of the Water Act are applicable to gold residue
disposal, the most pertinent provisions can be summarised as:
e safety of dams (Section 9B and 9C);
• use of public water (Sections 11 and 12);
• alteration of water course (Section 20);
e disposal of effluents (Section 21);
665
• prevention of water pollution (Section 22);

e pollution of water to be an offence (Section 23);
e right of access of inspectors (Section 24);
o regulations relating to the prevention of the pollution of water (Section
26);
• penalties and fines for non-compliance (Section 170);
e award of damages (Section 171).
Regulations in terms of the Water Act of 1956 which are currently in effect
and of relevance to tailings disposal systems are:
e R287 of 20 February 1976, which deals with the prevention of water
pollution;
e R991 of 18 May 1984, which sets out quality criteria for waste water,
effluent and seepage in terms of the General Standard and Special
Standard.
11.3.1.2 The Mines and Works Act

Environmental regulations have been made in terms of Section 12 of the Mines
and Works Act. These regulations were promulgated under Government
Notice R992 of26June1970 and R537 of21March1980. Provisions which
are pertinent to gold residue disposal are listed below:
e rehabilitation of dumps (Sections 5.10 and 5.13.2);
e waste disposal (Sections 5.13.1 and 5.13.4);
e location of dumps adjacent to watercourses (Section 5.14.3);
• poisonous water (Sections 5.9.1 and 5.9.2).
11.3.2 Licensing
At the present time no licence or permit is required to operate a gold residue
disposal facility as such. Generally, however, all mines will need to make
applications for the following permits, which involve matters concerning
slimes dams:
Department of Minerals and Energy Affairs:
Application for a permit in terms of Section 8(1)(a) read with Section 6B/4(2)
of the Physical Planning Act (Act 88 of 1967).
Department of Water Affairs:
Application for permits/exemptions/registrations in terms of Sections 12,
21 and 30(5) of the Water Act (Act 54 of 1956 as amended). These applica-
tions are made on what is known as the M2 Questionnaire and pertain to
the use, treatment and disposal of water at a mine.
11.3.3 Environmental impact assessment

By placing some emphasis on environmental planning, a mine will realise
a number of benefits:
e a good public image;
e compliance with legislation;
666
• the preservation of the environment, if not actual environmental improve-

ment;
• proof that industrial development can go hand-in-hand with environmen-
tal protection.
Environmental planning begins with the production of an environmental im-
pact assessment (EIA). An environmental impact assessment is a formal ap-
proach to evaluating the environmental effects of a project and to minimising
environmental degradation n the most cost-effective manner. As part of the
feasibility study of a project, it can detect problems at an early stage or
minimise expenditure on mitigatory measures.
Procedures adopted involve:
e an assessment of existing environmental conditions;
• a review of the impact on the environment of proposed actions;
e an evaluation of the quality of this impact;
e a definition of appropriate mitigatory and remedial measures;
• monitoring to ensure the effectiveness of control measures.
11.3.3.1 Methods of assessment

An environmental impact assessment is approached in the following way:
• existing environmental conditions (locations, topography, climate, hydro-
logy, flora) are defined;
• those aspects of the proposed project which may have an impact on the
environment are defined, e.g. movement of men and machines, struc-
tures, waste disposal, aesthetics, changes to storm runoff, changes to the
groundwater regime;
• the significance of the impact of each aspect is evaluated.
Good judgement is required to quantify the key environmental parameters.
In this regard a multidisciplinary and broad-minded approach is required
to define the scope of environment impacts for decision-makers.
11.3.3.2 Remedial measures

Once the aspects most likely to have a significant impact on the environ-
ment have been identified, the appropriate remedial or mitigatory measures
can be designed. This will normally involve controls built into the design-of
the project to counter negative effects.
11.3.3.3 Monitoring
Finally, a monitoring system is incorporated. Baseline data on the site and
its environs are established before the project starts. Monitoring during the
operational phase ensures that the effectiveness of the control measures can
be evaluated and, if necessary, improved.
A thorough environmental impact assessment, followed by an on-going
monitoring programme, helps to ensure
• that the prevailing legislative requirements regarding environmental pro-
667
Particle size analysis
100
90
80
~
~
"'
E 70
""'
.D
~
60
"'<:::
c:
50
"'
bl)
"' 40
""'"'
i: 30
""' 20
10
0
0,001 0,01 0,10 1,0
Particle size (mm)
M.l.T. Classification
Clay Fine Medium Coarse Fine Medium Coarse Fine Medium Coarse
fraction Silt fraction Sand fraction Gravel fraction
Typical grading curves for platinum mine tailings

a) Fine
b) Coarse
-·-
.......,.......
Typical grading curve for copper mine tailings (total)
Typical grading curves for gold mine tailings

c) Coarse
d) Fine
-·- Typical grading curve for diamond mine tailings
Figure 11.4. Typical grading curves for various tailings materials.
tection and pollution control are satisfied;

e that the environment can be protected from air and water pollution in
both the short and long term.
11.4 Design, Specification and Construction

11.4.1 Operation
11.4.1.1 General
The method of disposal of hydraulically placed tailings to ensure optimum
distribution of the particle size fractions varie.s considerably from one tailings
product to the next, principally as a result of the following factors: gradings
(particle size distributions), solids content, and solids density.
All these will, to varying degrees, affect the tendency of the material
668
to segregate, but the most important factor is the grading of the product.
A well-graded product with a particle size range from 2 mm to clay fraction,
such as is common in copper or platinum operations, will segregate rapidly
so that within 30 to 50 m after discharge, the solids remaining in suspension
are predominantly fine silt particles. On the other hand gold tailings, which
comprise predominantly silt particles and have a reasonably uniform particle
distribution, have a much reduced tendency to segregation. Typical grading
curves are given in Figure 11.4.
The nett result of segregation differences is that certain methods of
disposal which rely on there being minimal gravitational segregation can be
used only for tailings similar in grading uniformity to gold tailings. On the
other hand methods commonly used in the disposal of copper and platinum
tailings, which are designed to optimise the effects of gravitational segrega-
tion, are less effective on gold tailings. Gravitational segregation significantly
affects the 'beaching' profile attained by the tailings. This can, in the case
of gold tailings, have significant freeboard implications.
The single most important determinant of beaching slopes for gold
tailings is the density of the pulp discharged on to the dam. In this regard
typical variations in grading result in much smaller variations in beach slope
than those due to changes in pulp density:
Pulp water:solids, by mass Slope
2,15 1:400
1,03 1:200
It therefore follows that the first aspect to be considered when embarking
on a design for a site is the method of development and operation of the
proposed tailings dam. The method of operation will dictate: toewall re-
quirements, drainage requirements, penstock locations, and tailings delivery
details.
Factors influencing the selection of a particular method of operation
and brief descriptions of these methods are set out below.
11.4.1.2 Method of formation of tailings dam

A number of possible methods for hydraulically placing gold tailings exist,
namely, the paddock system, the cyclone system, the spigot system, and open-
end discharge qehind a mechanically formed containment wall.
The choice of disposal methods for a particular project will be deter-
mined by a number of factors:
• cost, both capital and operating;
e previous mine experience with one or more of the methods and hence
mine preferences;
• site topography - a valley site favours cyclone deposition due to the
higher rates of rise while a flat or hillside site favours the paddock system;
• climatic conditions as these affect drying characteristics and freeboard
requirements;
• pulp density.
669
Paddocks filled - - - - - - - - - .
sequentially
Delivery main
Pool
(a) Perspective view
Night wall or floor area
Earth starter
wall
Phreatic line
(b) Typical section
c. 20m
Hand-packed or
plough-formed
dyke
(c) Typical daywall section
Figure 11.5. Paddock system of tailings dam construction.
670
The systems for placing the tailings are described briefly below:
Paddock system
The paddock system for dam operation has been developed empirically over
the past 100 years and seems particularly suited to the semi-arid and temperate
climatic conditions in which most of the gold mines in South Africa are
located. The method is illustrated in Figures I l.5a to I l.5c.
Initial construction is carried out to form earth starter walls and under-
drainage. The starter walls are sized such that when the tailings reach their
maximum height the rate of rise is not greater than 2,5 m/yr. At that stage
the tailings themselves are used for wall building. An outer wall is raised
by ploughing up a dyke on either side of the wall areas as shown in Figure
I I .Sc. The wall area is further subdivided into paddocks by cross dykes.
During the day shift, pulp is led from the delivery points along furrows
into each of the paddocks and deposited to a depth of about 150 mm. Ex-
cess or supernatant water from pulp is decanted into the inner part of the
dam. The slurry is left to dry and crack for up to 3 weeks before the next
deposition. Deposition in this wall area is carried out only during daylight
owing to the high degree of control required on pulp depths. Uncontrolled
deposition could easily result in over-topping.
During the night, the tailings are discharged directly into the interior
of the dam behind the walls formed during the day.
In view of the emphasis on the two shifts, the outer wall area is referred
to as the·' day wall'', while the inner area is referred to as the 'night area',
'night pan' or 'floor'. Supernatant water is drawn off the dam by means
of penstock decants or by barge and pump.
On the gold fields of South Africa evaporation generally exceeds rain-
fall. Provided that rates of rise are low enough, therefore (see Figure 11.2),
the surface, with the exception of the pool area, becomes desiccated and large
shrinkage cracks develop. These cracks are filled and re-filled by successive
lifts of tailings. This desiccation is a fundamental requirement of paddocked
dam construction. Drying results in densification, which gives the gold tail-
ings the required strength. In addition the cracks tend to become filled with
coarser material, which improves vertical drainage. Rate of rise must be con-
trolled in order to ensure that desiccation does occur. The allowable maximum
rate of rise is dependent on tailing moisture content. Figure I I .2 provides
a guideline for the maximum rate of rise which can be imposed on the dam.
Cyclone system
Increased rates of rise can be tolerated by the gold tailings (up to 7 m/yr
and more) by making using of a hydrocyclone to split the incoming pulp
into two components:
• cyclone underflow which contains the coarser particles and significantly
reduced water content;
• the cyclone overflow which contains the finer particles and most of the
water.
Figure l l .6a to I 1.6e illustrate the principle of distribution of materials using
671
Undulating crest profile
Delivery pipe located on crest
Coarse
underflow
Under-
drainage
(b) Centreline method (c) Downstream method (d) Upstream method
0 Successive delivery
positions
-v--v- Successive cyclone
Upstream positions
(e) Upstream/downstream method
Figure 11.6. Hydrocydone system of tailings dam construction.
672
a hydrocyclone.
The cyclone underflow generally has improved shear strength proper-
ties due to the lower water content, is relatively more free-draining than pad-
docked tailings, and will form a cone on discharge. The cyclone overflow
material, on the other hand, is wet and of lower permeability due to the in-
creased proportion of fines (International Commission on Large Dams, 1980).
The objective in using the hydrocyclone is to place the underflow pro-
duct in such a way as to create an impoundment for the containment of the
overflow. Owing to the lower moisture content and improved drainage
characteristics, this material tends to perform better under seismic loading
conditions, although in the case of gold tailings where the largely uniform
grading of the material results in only a slight difference in coarseness be-
tween underflow and overflow, seismic performance is not improved unless
the underflow material is well drained and of low moisture content. The most
important factor in a hydrocyclone operation is to ensure that the growth
of the underflow embankment is at all times significantly ahead of the growth
of the fines beach infill behind the embankment. Sufficient underflow
material must therefore be available to build the embankment.
The tailings are deposited in a certain sequence using one or other of
the downstream, centreline, upstream or upstream/ downstream placement
techniques. These are illustrated in Figures 1l .6b to l 1.6e. The choice of
separation technique depends on topography. The shape of the site defines
the required volume of the underflow embankment to contain the overflow
product. Geotechnical considerations or drainage constraints may also in-
fluence the width of the underflow embankment.
The number and size of the cyclones will also be determined by operator
preferences and ease of moving and handling the cyclones, the quantity of
tailings, and the need to distribute overflow around the dam perimeter for
pool control. Systems in which multiple cyclones are required generally in-
volve more pipework and valves than the paddock system, and depending
on the need for cycloning 24 hours per day, can result in increased labour
to control operations, particularly during the early development stages of
the tailings dam.
These systems tend, therefore, to be used in special cases where:
• high rates of rise of the dam can occur due to the tonnage and topography;
o topography is steep and favours a valley fill tailings dam;
• large freeboard is required for stormwater control purposes in the early
stages of formation of the dam;
• high rainfall and low evaporation will reduce natural drying of the tail-
ings and preclude the adoption of a paddock system.
Cyclone systems become less attractive as the perimeter increases. Cycloning
is therefore most attractive for valley or hill side dams. The relatively poor
split for gold tailings also makes cyclone wall construction difficult.
Spigot system
The spigot deposition system is based on the need to ensure adequate drying
673
Pool
Phreatic line
____ Steep beach
Under-drainage
(b) Typical section
~--- Containment wall

constructed of tailings
Spigoted pipe
( c) Typical pipe support and wall construction system
Figure 11. 7. Spigot system of tailings dam construction.
674
and drainage of the tailings in the outer wall area by maximising the effects
of natural evaporation and drainage. The system involves the use of a pipeline
with multiple outlets referred to as a spigoted pipe. Regulated delivery in
limited (200 mm maximum) layer thicknesses using a spigoted pipe is carried
out. The spigoting encourages runoff of supernatant water directly to the
pool concurrent with deposition.
By depositing in thin layers, with a drying period between successive
layers, the drainage of each newly deposited layer and evaporation effects
are enhanced. The spigot arrangement is illustrated in Figure 11. 7 a to 11. 7 c.
The tailings deposition front is spread over a length of the ring main, the
actual length being a function of the number and size of spigots in relation
to the tonnage.
Operations involved in sustaining this system are:
• the construction by hand or machine packing of an outer crest wall for
containment of the newly placed material;
e periodic raising of the pipe and extension of the feeds from a ring main.
This usually requires dismantling of the pipe, raising of the supports and
re-assembling.
Mechanically formed containment systems
The three systems described above have one major common feature - they
all facilitate containment dyke construction using hydraulically placed tail-
ings. On large impoundments with long perimeters this feature results in a
very considerable cost saving over a system which would involve mechanical
formation of the containment dykes using imported materials.
Generally, the South African gold fields are characterised by gently roll-
ing slopes, the exception being the Barberton gold fields. Most tailings dam
sites therefore necessitate containment wall construction around the entire
perimeter of the impoundment, making long-term operation using mechani-
cally formed (as opposed to hyd_raulically formed) systems prohibitively ex-
pensive. However, mechanical systems do find particular application in the
following cases:
• short valley impoundments with a favourable storage to height ratio;
• sites in which seismic activity is of particular concern since earthfill dykes,
when correctly designed and constructed, perform better under seismic
conditions;
e small operations of limited life.
In view of these factors there are relatively few examples of this method
of gold residue disposal in South Africa.
11.4.2 Tailings delivery

11.4.2.1 Pulp concentration
Within limits it is usually preferred to pump solids at as high a concentra-
tion as possible, since in many cases this reduces the size of pumps, pipelines,
motors, and power costs. Means for increasing the concentration of solids
in the pulp include one or more of the following: spiral classifiers, cyclones,
675
screens and pressure filters, and thickeners.

All of these have significant capital and operating cost implications which
should be offset against: the value and re-use potential of the water recovered,
i.e. reduced loss of fluid by avoiding delivering it to the tailings dam; re-
duced costs of return water recovery; and higher allowable rates of rise.
Maximum concentration may be limited by the thickeners or the metal-
lurgical plant process. In other instances, the maximum concentration may
be limited by the adverse effects of high viscosity on pump performance and
pipeline losses. Typically, the tailings water: solids ratio varies between 2,8
and 1,0 for uranium and gold and gold-only extraction, respectively. Tail-
ings density is set essentially by process requirements, and modification is
not generally economically feasible.
11.4.2.2 Pumping and piping systems

A tailings delivery system comprises pumps, pipelines and valves.
On the vast majority of mines in South Africa, these components are
made up as follows:
Pumps - usually centrifugal in series, with one variable speed pump to cater
for surges. Pumps are usually rubber-lined to reduce impeller wear, and are
made of cast iron.
Pipelines - in certain cases rubber-lining of the pipe is used to reduce cor-
rosion under acid conditions and infrequently to reduce wear. Delivery
columns to the dams are supported on concrete plinths with .vertical guide
rods to control expansion and contraction. HDPE (high density polyethylene)
and asbestos cement pipelines with lower friction and better wear charac-
teristics are finding application on more and more mines. The cost of HDPE
pipes is high and problems of fire damage and expansion often necessitate
burying the pipe.
Valves - types depend on pressure head and the capital cost/ operating cost
ratio. Certain makes of diaphragm valves which operate at low pressures can
be adequate, provided a stock of repaired valves is maintained to facilitate
rapid replacement and a routine valve repair programme is maintained. For
higher pressures, pinch valves are likely to be the most reliable.
The following guidelines are recommended for the design and opera-
tion of tailings pumping and piping systems:
e A void sharp bends and steep troughs in the pipeline route. Support the
pipe on plinths to regulate vertical alignment. A void fixing the pipes to
the plinths; opt rather for guides.
c; Provide measures to counter expansion and contraction of the pipelines.
Position these as close to valves and tees as possible to minimise flange
stressing. Expansion control measures can take the form of loops, sec-
tions of corrugated armoured rubber piping, free corners or specially
manufactured expansion pieces.
• Provide spigots along the delivery line to facilitate flushing in the event
of a blockage. Drainage points may be provided at convenient points to
676
facilitate emptying the pipe for maintenance purposes.

• Implement pipe rotation programmes in areas of anticipated high wear.
• Ensure that valves are placed directly adjacent to tees to avoid blockages
between the tee and the valve.
e Allow for surging when deciding on the ratings of valves and flanges.
Surge pressures are commonly double design pressures.
9 Although external coatings are only required where the pipes may come
into contact with the tailings, steel pipes may be coated with a reflecting
paint such as high metal content aluminium to minimise temperature dif-
ferentials. Pipe supports should be high enough to prevent conta~t of the
pipe with the ground or deposited tailings.
11.4.2.3 Pump and pipeline design

Pipeline design and pump selection are dealt with in an Appendix to Chapter
16. Detailed design of the pump and pipeline systems should be carried out
by people experienced in the field of pulp transportation.
11.4.2.4 Tailings distribution

In general, the more delivery points there are, the better for operation of
the dam. In paddocked dams the spacing of delivery stations should preferably
be less than 400 metres. As the relative density of the tailings decreases, con-
trol becomes more difficult and it may be necessary to reduce the spacing.
For low relative densities particular attention must be paid to locating the
delivery stations at corners to prevent the formation of low points. A ringfeed
serves the delivery stations which commonly take the in-wall configuration
shown in Figure 11.8a. In paddock dams it is becoming more common today
to opt for this type of delivery.
In the case of a spigoted delivery, delivery points in the form of spigots
are spaced at intervals of between 1 m and 5 m, and are operated in groups
of 20 to 40 spigots. The spigots are sequentially opened and closed so that
a delivery front progresses around the perimeter of the dam. A ring main
around the toe of the dam is provided to feed into the spigot pipe (Figure
11.8b). The number and positions of deliveries into the spigot line depend
on th~ pipe-lifting cycle.
Cyclone systems usually involve connection of the cyclone units into a
'manifold' pipe which is maintained along the crest of the dam. One or more
cyclones tee off the manifold pipe at, commonly, 30 m centres. The length
of the manifold pipe is a function of operating cycles and sequences, and
can be 100 to 400 m. The positions at which the main delivery column con-
nect into the manifold depend on the length of the manifold. An example
is illustrated in Figure 11.8b.
11.4.3 Return water management

Storage policies
The Bafokeng disaster of 1974 (Jennings, 1979) demonstrated the importance
of sound return water management policies whose cornerstone should be the
stipulation that water should not be stored on the top surface of the tailings
677
Night (floor) delivery
Daywall delivery
Section illustrating
in-wall delivery
Feed from plant
(a) Paddock dam
_ _ Cyclones (at least one

either side of delivery)
,,----"<c:---Distribution pipe with branch

connections for cyclones
,__-~-Ring main
~-~~--r-~~---t+==--1----~Delivery from ring main
Feed from plant
(c) Spigot dam
Figure 11.8. Pulp delivery systems for the various types of dam
construction.
dam. Ideally water should be decanted off the dam concurrently with deposi-
tion so as to maintain a minimum pool area and a depth just sufficient to
ensure adequate water clarity. Inherent in this operating philosophy is the
need for a return water system which either re-uses or treats the water at
a rate at or above the rate at which the water can be decanted, or provides
adequate storage in a return water storage dam.
In the first case it may be adequate to decant into a stilling/ settling pond
whose capacity is limited to that required to ensure adequate water clarity.
In the second, the storage facility may act as an evaporation system to pre-
678
vent discharge of the return water into public streams. In this case the facili-
ty may assist in overcoming plant make-up problems during the dry season,
and the required storage capacity m,ay be between 30 OOO m 3 and
I OOO OOO m3 depending on circumstances.
Two separate objectives must be recognised in the design and operation
of an effective water management programme. These are: i) Minimisation
of water loss and hence the minimisation of the cost of water. ii) Elimina-
tion of pollution caused by discharge of effluents which do not comply with
statutory requirements.
The two objectives are not always compatible, since the cost of water
has not yet reached the level where it is a significant proportion of the cost
of gold production. For this reason the second objective above must be ac-
cepted as a responsibility which cannot always be justified economically.
In order to implement and maintain an effective water management pro-
gramme, a clear understanding of the plant and mine reticulation network
must be obtained. In this regard networks and balances must be prepared
and kept up to date. Thereafter, an understanding of the effluent genera-
tion and consumption problems on an hourly, daily, monthly and seasonal
basis must be sought. Given this information, the reticulation and storage
facilities required to eliminate uncontrolled discharge, except in extreme
weather conditions, can be designed. The tailings dams produce the most
variable and unpredictable quantity of effluent on the mine. In order to
understand the effluent generation problems hydrological models can be used.
Elements of such a model are:
Inflows: water with the tailings
precipitation
any extraneous disposals such as sewage or concentrated effluents.
Outflows: return water re-use
evaporation
seepage losses
interstitial water (water retained in the pores of the tailings).
Figure 11.9 illustrates typical components of a tailings dam water
balance. The difference between inflows and outflows gives the change in
storage for the period in question and, in the case of water excess, would
be storage capacity required to prevent spillage. In the case of water short-
ages the change in storage would be the water shortfall for the period. Govern-
ment regulations (R287) stipulate water storage capacity requirements on the
basis of a balance which incorporates average monthly rainfall, and rainfall
of recurrence interval I: 100 years of 24 hours duration. Above the storage
level obtained from the calculation a freeboard of 0,5 m has been stipulated.
This stipulation seems severe but is actually quite realistic. Water balance
simulations calculated using actual monthly rainfalls over an extended record
period of 50 to 70 years, and using actual area-storage characteristics for
the various structures to calculate evaporative losses and rainfall, show that
often the maximum volume to be stored calculated over the record is slight-
ly larger than the volume calculated in accordance with the regulation.
679
Key: External
Tw - Tailings water make-up supply
I - Interstitial moisture
Swb - Seepage from wet beach
Dr -
Sp - Seepage from pool HBalancing
Ewb - Evaporation from wet beach ~ tank
Ep - Evaporation from pool
P - Precipitation Process plant
Rb - Runoff from beach
Rp Rainfall input to pool
Rxt - Runoff from external pool
D - Decant water from slimes dam
V - Volume stored in Rwd
Dr - Return draft
R wd - Return water dam
p Penstock intak;\ A EAb
Beach~~ ~~Pool lf lfwet beach ,;-;w

/ · == ,,,'.) n ,:~ l, ,~A)
~Sp
???
Horizontal penstock outlets

Swb
External catchment
(diverted around dam)
Return
Figure 11.9. Tailings system water network.
Decant systems
Systems for dec~ting and removing supernatant water from the dam com-
monly comprise one of the following:
e A penstock decant comprising a pipe outlet of steel or concrete, and a
series of inlet boxes with vertical risers.
e Floating barge systems. These systems comprise a primed pump mounted
on a raft and located centrally within the pool area. A decant pipe is
680
'floated' across the pool to the barge and supported above the tailings
surface. The pipe is laid over the wall crest.
Commonly, precast concrete penstock rings with effective height of 0, 1 m
and internal diameter of 0,51 mare used. Alternative materials such as steel
and HDPE can also be used. After storms, rings may be temporarily removed
to accelerate decanting and then, once stormwater has been removed, the
rings may be replaced as appropriate. A decant system which was common-
ly used until 10 years ago comprised a vertical steel or concrete box with one
side open to receive wood stoplogs or weir planks. Some of these systems
are still in use today but owing to their poor long-term reliability have been
largely superseded by the concrete ring system. The steel boxes and the wood
stoplogs degrade, becoming structurally hazardous, and have resulted in many
decant failures over the past 30 years.
Steel columns are also used on a number of tailings dams. The columns
are extended in sections as the dam rises and are joined by bolts. The column
diameter and width of the inlet weir depend on the size of the dam and the
rate at which water must be decanted. The weir level is regulated with steel
slats which are added or removed to control water level.
Guidelines and cautions on the operation of the concrete ring decant
systems based on past experience are given below.
Penstocks
This system is most commonly used on South African gold mines. Despite
this fact, a simple, reliable system for locking adjacent concrete rings to yield
a riser capable of withstanding tension, such as could be required in seismic
events, has yet to be developed. The end faces of the rings are profiled to
provide interlocking between adjacent rings against transverse forces.
Dimensional control on the rings has been criticised since out-of-true
rings have been found to be responsible for considerable deviation from the
vertical (lateral deflections of700 mm have been measured). A draft specifica-
tion which will form an appendix to the South African Bureau of Standards
Specification 677 for spun reinforced concrete pipes has been prepared (1986)
and is currently under review. This draft specification provides for con-
siderable tightening of dimensional control during manufacture.
Penstock risers which have gone off vertical should be brought back
into plumb by using a mastic filler between rings.
A number of fatalities have occurred during the addition and removal
of rings. Causes have been mainly associated with under-estimation of the
force of the water flowing into the penstock. In one case the operator was
drawn far enough into the penstock to drown, despite being attached to the
platform via a safety harness. Under storm conditions, therefore, manage-
ment of the rings can prove dangerous and it is advisable to provide suitable
platforms and equipment around the inlets to reduce lifting and replacement
difficulties with the rings.
Penstock decant structures buried within the tailings dam require struc-
tural design where the depth of tailings above or surrounding the structure
exceeds 10 m. Elements requiring detailed design are:
681
• The decant pipeline outfall, which should be able to withstand at least

the full total overburden stress resulting from the tailings and differen-
tial settlements. The bedding and jointing system should allow for
movements between pipe sections. The outfall usually comprises a pipe
on a bedding. In these cases it is preferable to lay the pipe in a trench
to reduce load effects. Both steel and concrete pipes are often used.
• The riser or decant, which, as noted above, usually comprises a stack
of spun reinforced concrete rings. This riser will have to withstand the
lateral pressure of the tailings and frictional down-drag caused by con-
solidation and settlement of the tailings.
• The penstock decant base: this forms the junction between riser and out-
fall and should be able to support the riser without transferring excess
load to the outfall pipe.
Guidelines on the design of the riser and the outfall are given in the
Chamber of Mines Handbook of Guidelines (1983). It is, however, advisable
to seek professional assistance on decant structures in tailings dams of heights
greater than 10 m. By far the majority of failures within tailings dams over
the past 10 years have involved penstock structures. None of the failures has
been singly responsible for abandonment of the tailings dams and, in most
cases, a replacement 'floating' penstock has been installed.
Loss of a penstock decant usually necessitates temporary decommission-
ing of the tailings dam to allow drying for access. The drying and construc-
tion period may take 3 to 4 months. Owing to susceptibility to penstock
malfunction or failure it is advisable to divide the tailings dam into two com-
partments with separate penstocks to provide sliming area in the event of
failure of one penstock.
Barge systems
Barge systems are seen as being structurally more reliable than the penstock
ring system above; however, they do require considerably more maintenance
to ensure dependable operation of the pump and to maintain the pipeline
above the tailings. In most cases, a barge system is only practical where a
large year-round pool is likely to persist. On dams on which the pools are
allowed to dry up completely between deposition cycles, the barge tends to
'bog down' and is difficult to re-float. To overcome this, some large valley
dams have made use of barges on a rail track to raise and lower the barge.
Access to penstock inlets and barges commonly involves one of two alter-
natives described below:
• A catwalk which usually comprises gumpoles (i.e. eucalyptus poles) for
uprights between which pine planks are spanned as a walkway. Uprights
are periodically extended and a new walkway formed as the tailings level
rises. After a number of extensions and a period of exposure most of
these catwalks become rickety and dangerous. In addition, over the total
period of operation of the dam, catwalk costs mount up and can become
excessive.
• A pool wall and limited length of catwalk. This involves the formation
682
~-~-------Delivery required
at pool wall
Permanent
penstock ~--~~--Pool wall formed
inlet similarly to
typical daywall
)!Q[?":-~----+-to-+--~c--- Short permanent

Temporary catwalk
penstock
inlets
Figure 11.10. Typical pool wall arrangement.

of a day-wall, in the case of a paddock system, a cycloned embankment
in the case of a cyclone operation, or a spigoted wall, stretching from
the closest external wall to the penstock. Only a limited length of cat-
walk connecting pool wall to penstock is required and is easy to main-
tain. This is illustrated in Figure 11.10.
The pool wall can also be useful in maintaining pool control on dams
where the penstock is located close to the perimeter. The wall can be used
to increase the effective length of beach by forcing the water to flow a
longer route to the penstock, thereby improving freeboard along the
critical perimeter.
11.4.4 Stormwater management

11.4.4.1 General
A general rule for stormwater management is that all stormwater should be
removed from the top of the tailings dam and berms (i.e. paddocks) within
one to three days of falling. Penstock inlet or barge pump capacity and any
berm decants should be sized on the basis of these average flow rates. As
far as is practical, all water accumulating from catchments up-contour of
the dam should be routed around the complex and should remain uncon-
taminated. The regulations stipulate that stormwater which falls within the
confines of the tailings dams has to be retained on that dam. This is usually
interpreted as meaning that all water which has come into contact with the
tailings or blended with any tailings liquor should be regarded as contaminated
and cannot be discharged from the complex without a permit from the
Department of Water Affairs.
In addition, in view of the need to ensure that water is not stored on
top of the tailings dam for any protracted period, the regulations are usually
interpreted to mean that the water should be stored for re-use, evaporation
or treatment in a separate, specifically designed water-retaining structure.
which will constitute part of the tailings dam complex.
683
For the top surface of the dam where adequate surge capacity should
exist, the intensity of the 1 in 100 year storm of 24 hours' duration may be
taken as the total depth of rain divided by 24 hours to give an average inten-
sity. For certain regions of the dam such as the berms or toe catchment pad-
docks, unlike the top surface, surge capacity is limited and variation in
intensity of rainfall within the 24-hour period can prove important. Specific
methods to calculate design inflows and outflows have been developed by,
among others, the U.S. Soil Conservation Service. These involve the deriva-
tion of a simulated rainfall hydrograph with intensities equivalent to
10-minute, 2-hour, 6-hour, etc. storms built into the 24-hour storm. These
methods have been computerised and form the basis for most urban and
agricultural drainage design.
11.4.4.2 Top surface of dam

The decant system for return water control is usually sized to serve storm-
water control requirements as well. The number and sizes of inlets to
penstocks and barge pumps are based on an outflow hydrograph over one
to three days. Over this period some surge capacity must be provided, and
this is usually taken up on the beach slope of the dam. It is considered good
practice to ensure that the edge of the pool remains more than 50 m from
the edge of the dam at all times. Freeboard of 0,5 m must be maintained
over and above the maximum pool level at which this minimum distance is
retained.
11.4.4.3 Side slopes

The present approach to stormwater and erosion control of the side slopes
of the dam is to maintain a stepped profile of the dimensions shown in Figure
11.11. The rationale for adopting this profile is as follows:
e Limiting the length of the slope face reduces erosion at the toe of the
slope. Further reduction in slope lengths would continue to reduce ero-
sion but would prove unwieldly operationally, as each step-back requires
the establishment of a new starter wall.
e The berm widths are based on the minimum width to accommodate a
crest wall and facilitate machine access for the dressing and maintenance
of berms.
It is not recommended that stormwater be retained on the berms and
allowed to evaporate as this requires diligent maintenance of the crest and
cross walls to ensure that freeboard remains adequate at all times. Uncon-
trolled spillage over the crest walls can result in severe erosion damage on
the side slopes particularly if a cascade develops. In certain cases retention
of the water on berms for a protracted period can induce localised stability
problems.
It is therefore recommended that berm penstocks be provided as sliown
in Figure 11.11 at the low points between deliveries (in the case of a paddock
dam) or at mechanically formed low points. More details on berm penstocks
are given in Section 11.6. Ring inlets will facilitate raising of the inlet invert
684
Daywall
Nominal
crest walls
Geometry
9 m rise in dam height
7-10 m step-in
Figure 11.11. Typical wall profile.
as eroded tailings accumulate from the side slopes on the berms. It is ad-
visable to provide flexible pipes to service the berm penstocks to accommodate
anticipated settlements. Further it is recommended that coated pipes or
preferably plastic (HDPE) pipes be used to provide long-term corrosion
resistance. When constructing the pipe outlets, it is advisable to puddle the
trench backfill into place using fresh slurry in order to minimise erosion pro-
blems around the pipe at a later stage.
11.4.5 Slope stability

11.4.5.1 Influence of foundation soils
In the design of a tailings dam, and in particular the evaluatkm of slope stabili-
ty, considerable emphasis should be placed on obtaining an understanding
of the following aspects of the foundation soils:
• grading, and Atterberg limits for correlation of soil types;
• permeability at varying stress levels;
• consolidation characteristics in cases where rates of rise are to be par-
ticularly high or where soils of low permeability occur;
• shear strength.
Where feasible the above parameters should be obtained from tests in
which actual tailings liquor is used in order to verify that no obvious
hydrogeochemical effects occur which could change the behaviour of the
foundation material. Figure 11.12 illustrates the influence that foundation
drainage can have on the position of the water table or, in geotechnical terms,
the phreatic surface, for four drainage conditions. Slope stability is particular-
ly sensitive to the position of the phreatic surface, in relation to whether it
intersects the slope or passes into the foundation before reaching the slope.
Low permeability soils with poor consolidation characteristics will, particular-
ly under high rate of rise conditions, generally have low shear strength and
could impose limitations on the height and slope angle of the tailings dam.
11.4.5.2 Influence of tailings product

Geotechnically the gold tailings product may be described as a rock flour
comprising 0-15% fine sand and 0 lOOJo clay fraction with some 80% of
685
kr = 0
Dam permeability kd
Foundation permeability kr
///// /T////777/T//7/T/// Impervious /T///7/7 //7/T////T//T///T//

stratum
Figure 11.12. Effect of differing foundation (kr) and dam (kd)
permeabilities on position of phreatic line (after Blight and
Steffen, 1979).
the material falling within the silt classification. The clay fraction arises from
the grinding of fresh rock and therefore possesses few of the properties usually
associated with clay. All particles tend to have a granular as opposed to plate-
like shape as is the case with clay. Gold tailings exhibit no plasticity and no
cohesive characteristics. Although in pyrite-rich products some crusting of
the dried outer surface may occur, this crusting is purely superficial and is
not evidence of cohesion within the body of the dam. The effective angle
of friction of fine gold tailings such as would be found around the penstock
is the same as for coarse tailings at the delivery on to the dam. The impor-
tant difference between the coarse and fine products lies in their permeabili-
ty and consolidation characteristics. The fines will take longer to consolidate
and, if loaded too rapidly, the available shear strength can be reduced by
excess pore pressures.
Differences in grading, permeability and shear strength between fine and
coarse products are, in the case of gold tailings, not as pronounced as with
coarser products such as platinum or copper tailings. The maximum difference
in permeability can be one order of magnitude (factor of 10) but is usually
less. Some layering does occur due to variations in directions of deposition
and paddock filling. Variations in the milling consistency, however, may give
rise to greater variations than segregation during deposition. However, desic-
cation cracks and the preferential filling of these cracks with coarser material
on the next deposition cycle, tend to reduce the effects of layering so that
vertical permeability is only some 1,5 to 3 times lower than horizontal
permeability. Distribution of finer and coarser fractions along a beach will
be similar for the paddock and the spigot systems, and provided delivery
points are positioned in accordance with the recommendation in Section
11.4.2.4, this will be the case along the day/ crest wall as well. Finer material
will therefore be concentrated towards the pool area with coarser material
around the perimeter.
686
Cyclone deposition is usually specifically designed to give a materials

distribution where adequate coarser fraction is placed as a confining embank-
ment to retain the finer overflow product. In this case differentiation of
material properties from a slope stability view point may well prove more
important particularly as regards permeability and the control of the phreatic
surface.
11.4.5.3 Influence of rate of rise

Rate of rise is the term globally applied to describe qualitatively the con-
trolling influence of drainage and consolidation. High rates of rise on poorly
draining soils can severely affect
• the position of the phreatic surface - with severe stability implications;
• the degree of consolidation of the foundation soil and hence the shear
strength of the tailings;
• operation of the dam in cases where cycle times prove inadequate to
facilitate drainage drying and desiccation of the daywall tailings. In the
case of paddock dams this could result in piping through the pack-out
walls (i.e. the dykes) when new deposition takes place. Inadequate dry-
ing will also affect access for packing-out tailings (i.e. dyke building).
The influence of rate of rise on slope stability is illustrated in Figures
ll.13a and ll.13b. The figures were derived for a factor of safety= 1,5.
These do not necessarily give the controlling condition since operational con-
straints such as access may be more stringent. Reference in these instances
should be made to Figure 11.2.
11.4.5.4 Influence of phreatic surface

The position of the phreatic surface is the parameter to which slope stability
is the most sensitive (where stability of the tailings rather than the founda-
tion soil is critical). This has been demonstrated both in the theoretical analysis
and in field observations of failures. It therefore follows that in cases where
slope stability has to be improved, any measures which beneficially influence
the phreatic surface will prove the most effective. Up to the 1960's it was
common practice to initiate and operate a gold tailings dam by dozing up
a toe starter wall and commencing deposition. Fortunately in most cases the
rates of rise and foundation drainage conditions were such that no severe
stability problems were experienced, at least not usually until a dam height
of 15 to 20 m was reached. Seepage erosion at the toe of the dam has in the
past often been remedied with expensive rockfill buttressing. The efficiency
of these buttresses was, however, reduced as the interstices of the rock became
clogged with slime. This clogging causes the phreatic surface to rise further;
eventually large scale instability can occur. Correct design of buttresses is
therefore a prerequisite for long term performance.
In 1959 the first underdrain was placed at the toe of a new section of
a gold tailings dam in the Orange Free State (National Building Research
Institute, 1959). This drain was constructed of specially selected materials
to ensure a gradation from fine tailings to coarse rock so that only water
687
Dam has completely pervious base
35
100
0) 30 0)
onc on
@
""0.
0) 0)
0.
.g 25 80 .g
~ ~
ill ill
20
""
~
0)
t; 60 ""
s
0)
E 15
""
"O
"""'
0 40
..c:Oil 10
·o:;
:i:::
20
..c 5
R = Rate
of rise
0 ~~~~~~~~~~~~~~~~~~~~O
30° 32° 34° 36° 3go 40°
Safe slope angle (3°
(a) Pervious foundations
Figure 11.13. Height of dam versus wall slope angle for various
rates of rise (after Blight and Steffen, 1979).
and no tailings would enter the drain. The drain still operates to this day.
The objective in deciding on placement of drains is to provide a control point
well inside the toe of the dam to which the phreatic surface can be drawn,
thus maintaining a dry slope face. Today drains in slimes dams are the norm
and this coupled with other modern design concepts has increased feasible
dam heights from a maximum of 30 m to 60 m and more. Figure 11.14 illu-
strates the effect of an underdrain.
Even with the inclusion of an underdrain, poor deposition practice, high
rates of rise, variable material distribution, and poor pool control can negate
many of the positive influences of the underdrainage. It therefore follows
688
Dam has impervious base
35
"'
30
.0 JOO
v ~
bii Factor of v
c
(rj safety 8 bii
'- c
v '<
....
p.
.g 25
= 1,5 "'v
p.
80 .g
v
""""
ill v
""""
ill
'oil 20
'"'
~
t) 60 ~"'
E """"
E
E 15 "'
"O
"'
"O """"
0
""""0 40 .:<::
co
.:<:: Ti
co 10 ::r::
'i)
::r::
..c::
..c:: 20
5
0
0300 32° 34° 36° 38° 40°
Safe slope angle (3°
(b) Impervious foundations
Figure 11.13 (continued)
that the provision of underdrainage goes hand in hand with good operation
in improving slope stability and increasing maximum final height.
11.4.5.5 Influence of slope geometry

Figures ll.13a and b above illustrate the influence of slope angle on slope
stability. It is of interest to note from the graphs that a relatively small reduc-
tion (2 to 6 degrees) in slope angle can result in a very large increase in height
of dams and a proportionally large increase in dam capacity. Present prac-
tice is to adopt even flatter effective slope angles than those considered in
drawing up the graphs in Figures 11.13a and b. The reason lies in erosion
689
Phreatic surface - no drains

Pool
I
Phreatic surface
with drains
LToe drain Blanket drain

Figure 11.14. The influence of underwall drains on the posi-
tion of the phreatic surface.
and stormwater control requirements. In Section 11.4.4 a recommended slope

profile was described which involved the creation of stepbacks and the pro-
vision of berm penstocks. This profile results in an effective slope angle of
25 degrees. It is the authors' opinion that the slope volume sacrificed by adopt-
ing a flatter angle is offset by the benefits of improved slope erosion control
and the potential for even greater allowable height.
11.4.5.6 Stability analysis

Several methods for analysing slope stability exist, ranging from those in-
volving charts developed from specific geometric conditions to sophisticated
computerised calculations able to cater for almost any slope geometry. By
far the majority result in the calculation of a factor of safety which would
represent a ratio of resistance to disturbance as a quotient of forces, moments
or shearing stress. Other analytical approaches exist, the most important of
which is probably that of finite elements. In this analysis the slope is divided
into a mesh of elements with connecting nodes.
There is a trend in slope stability analysis at present to opt for prob-
abilistic type analyses; most of these involve the calculation of a distribution
of factors of safety and from them in turn a deduction of the probability
that the factor of safety is less than a given amount. The principal rationale
behind probabilistic analysis is that unlike pure factor of safety evaluations,
probabilistic analyses are linear. By this is meant that halving the probabili-
ty of failure makes a slope twice as safe, whereas doubling the factor of safety
makes the slope safer by an unquantifiable amount. Factor of safety is
therefore really a qualitative measure of safety, while probability of failure
is a quantitative measure. Probabilistic methods are still very much in the
early stages of development and have a short track record compared with
tried and proven factor of safety methods.
690
11.4.6 Repair and reinstatement of operational and abandoned

impoundments
11.4.6.1 Common problems with old and abandoned dams
It has only been over the past 10 to 15 years that a trend towards systematic
design coupled with detailed geotechnical evaluations of new tailings sites
has developed. Consequently there is a large number of tailings dams which
are still operational but are exhibiting problems. In addition, as a result of
increased land prices and environmental issues, there is an increasing tendency
for mines to look to re-commissioning old, previously abandoned, impound-
ments. Repairing existing troublesome impoundments for which infrastruc-
ture such as pipe networks and access are in good working order usually in-
volves expenditure on civil works similar to that for a new dam. However,
cost savings considerably larger than these expenditures result from
eliminating the need to provide new pumping and piping systems and the
avoidance of land purchase costs. In the case of re-commissioned dams, land
purchase costs are avoided and, since these dams are usually located closer
to the plant than the newer sites, the pump and pipework costs are usually
lower.
The commonest problems exhibited by troublesome dams are: seepage
erosion at the toe resulting in slope instability; a high phreatic surface resulting
in general slope instability; inadequate deposition area resulting in excessive
rates of rise, poor consolidation and drainage, in turn inducing the previous
two problems. Inadequate area more often than not results in a commitment
to obtaining additional area for a supplementary new development or a dam
extension. Even in this instance savings can be effected by purchasing only
sufficient land to supplement the repaired existing dam as opposed to com-
pletely replacing the dam.
Almost invariably the source of these problems can be traced back to
phreatic surface control. Even in cases where the dam is on weak founda-
tions, significant lowering of the phreatic surface c~n improve stability
drastically. Often the old dams suffer from the following geometric problems:
• The penstock is located close to the edge of the dam, causing the pool
to be off centre. Usually, owing to the original ground topography, the
penstock is established in the topographically low regions which corres-
pond with the tallest side of the dam. Close proximity of the pool at the
edge serves to accentuate seepage erosion and phreatic.surface problems
identified above.
II Too few deliveries are provided, resulting in long runs along the daywalls,
poor pool control and low freeboard furthest from the deliveries.
Ii Slope faces are maintained too steep with few stepbacks. This has im-
plications concerning both slope stability and stormwater erosion.
• The penstock pipes and inlets are rickety and unsafe through deteriora-
tion of pipes and inlet structures. Often these 'hole through' and draw
tailings into the penstock with the resultant formation of a sinkhole
around the inlet or along the penstock line.
691
Pool
Drainage blanket
Rip rap rock for~
erosion protection
(a) Mechanically placed buttress
Hydraulically _ _~ Pool
placed tailings
Toe starter
wall
(b) Hydraulically placed buttress
Pool
Rip rap rock for

erosion protection
(c) Cut-to-fill buttress
Cyclone -v~Overflow
Buttress of Under~ --------~=~P~o~o=l~

cyclone underflow
Blanket drain---~
( d) Cycloned buttress
Figure 11 .15. Various types of wall buttress.
692
11.4.6.2 Remedial measures

Remedial measures usually involve one or more of the following:
Adjustment of slope geomet1y
This may involve:
e Buttressing using mechanically placed materials such as earth and ro\:k
or tailings (see Figure l 1.15a). Note the placement of a drainage medium
between the original tailings and the buttress for phreatic surface con-
trol. Mechanical buttressing is usually undertaken where remedial action
is urgently required.
e Buttressing using hydraulically placed tailings by forming 'mini' dams
around the toe. Each dam has its own penstock, and is underdrained (see
Figure 1 l.15b). This buttressing method can be cheaper to form than a
mechanically placed buttress, depending on buttress height.
Since buttresses can normally only be built on the dayshift, freeboard
is sometimes lost on the main dam if insufficient slime is available to build
daywalls. This becomes a serious limitation where rates of rise are high
or where slime density is low. However, formation of hydraulically placed
buttresses takes time and needs to be done with care to avoid artificially
lifting the phreatic surface in the main dam and inducing slope failure.
e Pushdown operations where the effective slope of the dam wall is reduced
(see Figure ll.15c).
e Effective reduction of the slope angle using cyclone underflow (see Figure
1l .15d). Disposal of the cyclone overflow can become problematic if the
dam already has a stepback. In this case a slurry tank and pump may
be required to lift the overflow into the floor area of the dam.
Generally, hydraulically placed options are low on capital cost but high
on operating costs. Mechanical buttressing therefore often proves to be more
effective when all costs are considered.
Adjustment of pool location and top surface geometry
Seepage and phreatic conditions within the slope can be considerably im-
proved by relocation of the pool so that the edge of the pool is maintained
a distance of 100 to 200 m from the crest of the slope. This usually involves
the installation of a 'floating' or elevated penstock. A pool wall can assist
in reducing penstock pipe lengths and reducing catwalk costs as illustrated
in Figure 11.16.
Improved delivery system

The addition of delivery points to meet criteria proposed in Section 11.4.2.4
can increase cycle times and drying characteristics of the dam. Pool control
and freeboard can also be improved.
Installation of drainage
A number of methods for installing slope drainage in an existing slope exist,
two of which are:
• Elevated drainage - this involves the construction of drainage blankets
693
Main inlet close

to daywall
at highest point
(a) Original penstock position
Pool
High phreatic surface .~
Weeping
Section A-A
New 'floating'
Old penstock or elevated penstock
abandoned
New inlet Pool wall

Lower phreatic surface~ ~-i
Reduced weeping
~ ,,,,,,, ...
?J/)~//F'''j))
Section B-B
(b) Revised penstock position
Figure 11.16. Pool relocation.
694
Future phreatic surface - no d r a i n - - - - - -

Pool
Future phreatic surface -
Elevated drain
Drain outlet----
Pool
Phreatic surface
at time of
installing
drain
(a) Elevated drainage
Phreatic surface Pool

without drains
Phreatic surface
with drains
Push-in
drain
(b) Push-in drains
Figure 11.17. Incorporation of drainage into established walls.
on the top surface of the dam with the aim of controlling the phreatic
surface during future operations (see Figure 11.17 a). Water accumulating
in the graded filter blankets is drawn off the dam through flexible outlets.
• 'Push-in' drains - these comprise slotted pipes wrapped with geofabric
(i.e. synthetic filter membrane) and are drilled or wash bored into the
toe of the dam as shown in Figure 11. l 7b. The pipes form wick drains
and, if placed closely enough, can be effective in controlling seepage at
the toe. Access problems and wet and often dangerous working condi-
tions at the toe often preclude this method as a viable option.
Construction difficulties and sensitivity to design assumptions make
these methods difficult to implement correctly.
11.5 Observation and Control

The adage, a stitch in time saves nine, is nowhere more appropriate than
when used in reference to tailings dams. It is the authors' experience that
695
the introduction by mine management of formal observation and control

procedures has, almost without exception, led to an improvement in mine
and operator personnel awareness. On the majority of mines, however, slimes
dam observation and control are still usually related only to pumping prob-
lems and problems associated with plant downtime. Routine aspects of
matters such as the following are either ignored or glossed over: pool con-
trol, rate of rise, slope geometry, physical condition of key structures such
as penstocks and access ways, erosion control on side slopes and along step-
backs, catchment paddock and solution trench maintenance, deposition se-
quencing, phreatic surfac.e control, and drainage operation. In many cases
one or more of these aspects is responsible for serious problems with the
tailings dams. Problems could either have been avoided or remedied while
access was easier and costs lower had monitoring been effective. It is recom-
mended that formal procedures for reviewing and discussing these features
be entrenched in all tailings disposal operating procedures on a quarterly or
a six-monthly basis. What is most important is that the review be done at
senior management level to ensure that the consequences of developments
on the tailings dam are appreciated at first hand by those legally responsible
for the safety of the dam, and that discussions on remedial action and ap-
proval of monies to pay for this action are taken with minimum lead time.
Depending on the specific state of the dam and the consequences of a
failure, it may be prudent to retain the services of a professional engineer
specialised in the fields of geotechnical and tailings dam engineering to review
the dam operation.
A good observation and control system will necessitate the services of
a surface surveyor to:
e survey cross sections regularly and establish the heights and positions of
instrumentation, calculate rates of rise, and check on slope geometry;
e monitor instrumentation such as piezometers which are installed to locate
the phreatic surface;
e make observational notes on the location of the pool, the status of the
penstock inlets and access ways, erosion problems, maintenance aspects
and drainage operation.
Instrumentation which is commonly installed on tailings dams includes:
e Piezometers, which are used to measure pore water pressure in the tail-
ings or foundation soils, and provide an indication of the position of the
phreatic surface. Types of piezometers are:
i) Standpipe piezometers which comprise a vertical length of pipe posi-
tioned in the side slope or on the top surface of the dam with a porous
tip inserted below the phreatic surface. Water enters the tip and rises
in the tube to a level indicative of the piezometric head or pore pressure
at the tip. The water level is then measured by using a remote sensor
lowered down the tube to the water surface. The depth of water is
usually plotted on a cross section for interpretation purposes.
ii) Pneumatic piezometers which are installed at specific points in the
696
Piezometer P2 Piezometer P4
I Pool
--l---'--H-=-- ~-- Gumpole markers

to indicate
distance L
Figure 11.18. Typical piezometer locations.
tailings and foundation soils. These units comprise a porous tip and
a diaphragm. Tubes filled with nitrogen are led from the unit to a
point outside the dam where they are connected to an instrumenta-
tion board. A remote read-out unit which pressurises the nitrogen
until the external water pressure on the diaphragm is balanced is used
to measure the pore pressures.
iii) Vibrating wire piezometers which rely on water entering a porous tip
as with pneumatic piezometers, and causing a diaphragm to distort.
The diaphragm is connected to a thin length of specially designed wire
surrounded by a coil. Contact leads are led to a terminal board out-
side the dam from which a current is passed through the coil. The
current induces vibration of the wire at the diaphragm and, since the
frequency of vibration of the wire is a function of the tensile stress
in the wire, through a series of conversion and calibration factors,
the amount of movement undergone by the diaphragm as a result of
the water pressure may be deduced from a remote read-out.
e Inclinometers, which are used to measure progressive slope movement.
These instruments comprise specially shaped, vertically grooved tubes
which are drilled vertically into the section until the bottom of the tube
is in material which is unlikely to move. The tube is then fixed at the
base using concrete. A probe is lowered into the tube along the grooves.
Within the probe is an inclinometer which, through a read-out, permits
measurement of the angle of inclination of the probe. By taking a series
of measurements up the tube starting at the bottom it is possible to com-
pare consecutive measurements and plot a profile of the inclinomet~r
readings showing cumulative changes in slope. Trends from one set of
measurements to the next can be used to assess the rate of movement
and the horizon along which movement is predominantly occurring.
e Other simpler forms of instruments which may comprise survey pegs or
extensometers which would be used to detect local movements of the
surface.
Figure 11.18 shows a typical section and locations of standpipe piezo-
meters.
The various depths which are measured and recorded are indicated in
a cross section on the record sheet. It is worth noting also an estimate of
697
the distance from the daywall to the pool. This is useful when interpreting
piezometer levels, since a close pool will often explain a sudden rise in phreatic
surface.
Concluding this section on observation and control, it is worth stres-
sing the importance of documenting all discussions and review comments
in minute form. If an action column is included on the minutes there will
be increased urgency for the person responsible to carry out the decisions
taken during the review. It should also then be recorded when the action
is taken so that a chronological record is developed of all problems and
remedies. Attached to the minutes should be the piezometer measurement
records, plots of cross sections and any other reports worth putting on record.
At regular intervals the measurements, together with the recorded comments,
and the present dam status sl}.ould be reviewed by a professional geotechnical
engineer who should then undertake to produce a report documenting all
comments and formalise into report format instrumentation readings and
notes made during the period of review.
11.6 Closure Considerations

11.6.1 General
Closure procedures for a gold tailings dam must comply with the requirements
of the Air Pollution Prevention Act, the Mines and Works Act, the Health
Act and the Water Act, in order to prevent pollution of surface and
underground water resources and of the atmosphere and to treat the dam
so that it and its appurtenant works require as little maintenance as possible.
The closure of a tailings dam involves considerable expenditure by the
owner. If the desired result is achieved, this expenditure is justified because
an unstable generator of nuisance and a source of pollution is converted into
a stable, permanent topographical feature with little adverse environmental
impact and a low maintenance cost.
In addition to these basic objectives, attention should 'also be given, at
closure, to the other possible benefits of properly closed dams. For example,
whether it will be possible to use the tailings dam as a clear water catchment,
transform the tailings dam into a scenic feature, or to use the tailings dam
or any of its appurtenances for recreational purposes.
11.6.2 Boundary fence

Uncontrolled access of people, animals or other traffic to a closed tailings
dam must be prevented by installing a properly constructed and maintained
security fence. Current costs for installing a 2 m security fence are three to
four times those for a 1,2 m high fence, as shown in Figure 11.19. As the
cost difference between the two systems is substantial, the requirements of
the Chamber of Mines Guidelines should be interpreted with particular
reference to the cost effectiveness of the chosen system. The location of the
tailings dam should dictate the type of fence that is necessary. For example,
a security fence should be considered in cases where the dam is adjacent to
a built-up or residential area. However, lower cost fencing should be suffi-
698
101 mm f1 posts Steel droppers spaced at
Top cap spaced at 50 m cl c 2 m cl c between standards
0
§ \1600 ". - ~ ...... --;~ f----->.:
I I ™N.G.L.
6 mm th base plate I300J

sq
15 MPa !3001
sq
concrete
H-section standards
2,5 mm 10 WA double at 10 m clc
150 sq x 6 m base plate stranded barbed wire
Figure 11.19. Typical fencing details (after Smith, 1983).
I I
CL CL Overfill to allow tl
.......
! I for subsidence due C/J
Existing effluent trench Existing effluent trench to consolidation '"O
0
C/J
Overfill to allow --~
for subsidence due
I Selected earth backfill
(nominal compaction)
I Clean dump rock
or filter stone >
0
t""'
to consolidation 0
~
<"l
§
C/J
.......
Geofabric lining tl
Cl'I etr:I
"~
~ C/J
Alternative 1 Alternative 2
Figure 11.20. Closure methods for effluent trenches (after Smith. 1983).
cient for dams in sparsely populated rural areas. The effectiveness of the
fence system is dependent on the level of maintenance. The only feasible ap-
proach appears to be to form a body of rangers whose duties are regularly
to inspect and repair all fences in their region.
11.6.3 Access roadways

The access road requires periodic maintenance to ensure continued access
to all parts of the tailings dam for inspection. The maintenance envisaged
is limited to low-cost operations. Typically, this involves repair of erosion
gullies and large potholes and clearing of encroaching vegetation.
11.6.4 The underdrainage system

To ensure continued stability of the slopes the phreatic surface should be
kept permanently low. This requires that the existing underdrainage system
function continuously after closure. The drainage outfall pipes should be
marked clearly or staked and kept free of silt and vegetation to allow
unhindered flow at the outfalls.
11.6.5 The solution or effluent trench

Problems associated with effluent trenches, both during operation of the dam
and particularly after closure, can be summarised as follows:
• sloughing of the side walls, which leads to siltation of the trench, is a
common phenomenon (the extent of sloughing depends on the soil type
and the depth of the excavation);
• water velocities which depend on the grade of the trench can result in
either scouring or siltation. Uncontrolled vegetation in the trenches can
adversely affect the flow characteristics and performance of the trench;
e although the dam may be fenced, open trenches up to 2 m deep in places
are a danger to trespassers such as stray livestock, domestic pets and par-
ticularly children.
To eliminate or reduce these problems, recommended closure procedures
for effluent and solution trenches are:
• where no underdrains discharge into the solution trench and where the
benefit of the solution trench acting as a seepage cut-off trench is not
evident, the closure procedure as described in Figure 11.20 (Alternative
1) is recommended, e.g. backfill the trench with selected, nominally com-
pacted earth;
• where underdrains discharge into the effluent trench and/ or where the
trench acts as a seepage cut-off facility, construct a maintenance-free
'closed' drain system as described in Figure 11.20 (Alternative 2), i.e. line
the sides and bottom of the trench with a geofabric prior to backfilling
with a clean, free drainage material (clean dump rock or aggregate).
Complete the 'drainage sock' by extending the side wall geofabric over
the top of the rock and then backfill the top 300 to 500 mm section of the
trench with selected earth backfill as in Alternative 1. The high capital costs
700
required to implement these procedures are justified when they are compared
with the cost of maintaining open trenches for the continued life of the
deposit.
11.6.6 The slimes delivery pipeline

The slimes delivery column from the plant to the dam and around the dam
perimeter and all valves and slimes delivery pipes should be dismantled and
removed from site. If the slimes delivery pipes are 'in wall', the closure pro-
cedures illustrated in Figure 11.21 are recommended. The pipe ends should
be blanked off after the pipe has been filled with mass concrete. To prevent
piping erosion where the pipe enters the toe of the dam, a filter arrangement
should be installed. This can be achieved by cutting back into the toe and
installing a geofabric and rock or sand and rock filter.
11.6. 7 Surface drainage

Drainage of the surface and sides of the tailings dam is influenced by the
geometry of the dam. The recommended water control features for each of
the tailings impoundment components as illustrated in general in Figure 11.22
are outlined below.
11.6.8 Ground level catchment paddocks

After closure of a dam, a detailed survey of the catchment paddock area
is undertaken to optimise the location, number and dimensions of compacted
crosswalls. Crosswalls should have interleading, hydrologically sized spillways
to ensure an equal distribution of stormwater around the dam perimeter.
These measures maximise the evaporation potential of the available paddock
area. Interleading spillways and an emergency spillway from the topo-
graphically lowest paddock to the evaporation or storage dam area minimise
the risk of overtopping or breaching or both of catchment paddock walls
during excessive rainfall.
Before the construction of spillways and crosswalls, the catchment pad-
dock wall should be examined and checked for structural competence and
size (i.e. available freeboard and capacity). Consideration should be given
to covering the outer paddock walls, division walls, and the lower reaches
of the impoundment wall with dump rock to protect them against erosion
when water is stored against them.
11.6.9 Berms or step-backs

There are principally two ways to control stormwater that falls on berms
and the side slopes above the berms. The first involves discharging water
from the terraces by means of spillways or penstocks into the ground level
catchment paddocks. Spillways and penstocks are located at each of the low
level points along the berm about midway between each of the original slimes
delivery stations. The spillways are formed by combination of rock mattresses
and gabions (i.e. rock-filled wire baskets) over a prepared, compacted
geofabric-lined surface. At the base of the spillway, within the catchment
paddock, a reinforced concrete apron is provided. To ensure that water flows
701
Existing daywall slimes Existing night slimes
delivery outlet removed delivery outlet removed
-..] _ _ Existin_g slimes delivery pipeline --3

0
l'-' removed on closure of dam ::r::
trJ
Existing slimes delivery blanked off trJ
Mass concrete to fill -------.. ><
--3
existing slimes delivery pipes :;:i:::I
;:.:.
(")
N.G.L. profile
-<
--3
trJ
Figure 11.21. Closure procedures for 'in wall' delivery pipelines (after Smith, 1983). ~
trJ
~
r
r
Blanket and toe drains c::
:;:i:::I
Daywall (stabilised with cement)
0
Daywall cross wall ~
Catchment paddock cross wall with spillway
Terrace or berm
Catchment paddock wall
~
Solution trench 0
·Pool wall 0
r
0
Penstock
Figure 11.22. Plan of typical paddock wall (after Smith, 1983).

freely along the berm to the spillway or penstock, and to pr.event spillage
and erosion of the lower side slopes, terrace preparation is important. One
method includes shaping of the berm with a 5% fall inwards towards the
slope of the tailings dam and provision of stone pitching along the flow path
to prevent scouring or erosion. Another alternative consists of a lined or pro-
tected trench or bund on the outer extremity of the berm.
The second method of controlling storm water on berms and side slopes
entails the storage of all water on the terraces. Design criteria are the same
as those for ground level catchment paddocks. The lowest level berm pad-
dock is provided with an overflow penstock or spillway to route surplus water
to ground level paddocks. For both methods, the establishment of vegeta-
tion cover of the surface of the terrace is recommended to minimise con-
tamination of rainwater by transportation of solids, and wind erosion during
dry windy periods.
Of the two methods, the first is preferred as it ensures that all water
and silt, eroded off the slopes above the berm in question, are transferred
to the catchment paddocks at ground level. It is easier to deal with water
and eroded material at ground level than part way up the dam. The effec-
tiveness of the second method is less assured since the ability of walls formed
with tailings to impound water is questionable.
11.6.10 The dam surface

The surface of the tailings dam is subdivided into two areas, the daywall
and the basin. The existing surface topography should be used to control
rainwater falling on the area. The most economic solution is to provide rock
mattress spillways to discharge water from the daywall into the basin area.
Spillways should be provided at low level points, midway between each of
the original delivery stations. Water discharged into the ,basin flows to the
central penstock or spillway. All surface runoff is directed to and controlled
at the central penstock or spillway. Froin there it is discharged to a storage
or evaporation dam. As shown in Figure 11.23, there are two ways to decant
water from the surface of the dam.
The first method (Alternative I) is applicable to dams with adequate
penstocks. In this case penstocks may be used as spillways to discharge water
to the evaporation or storage dams. Details include no-fines concrete penstock
rings above the slimes level, surrounded by a sand filter layer to allow drainage
of water. Penstock rings can be added to a level higher than that of the basin
in order to ensure that only clear water is discharged, even during heavy
precipitation.
The 1;econd method is applicable to retired dams not equipped with ade-
quate penstock system or systems of doubtful long term integrity. In this
case penstocks should be sealed as shown in Figure 11.24, and a surface
spillway or trench arrangement provided as detailed in Figure 11.23. This
system consists essentially of a rock mattress-lined trench from the pool and
towards and through the daywall. Flow down the sides is controlled by a
gabion spillway similar to that described for the berm spillways. A weir at
the edge of the pool area should be provided for discharge control. As in
703
'No fines' concrete-

penstock rings Cement-stabilised
basin
Excavate profile prior Existing tailings deposit

to cement stabilisation
Filter sand layer
Flow to existing
effluent trench Existing penstock
ring intake tower
Existing penstock outlet
Typical section of penstock decant arrangement (Alternative I)
30 OOO Daywall Central basin
B Desilting weir
Existing fall
0,25% Existing fall 0,2507°
Fall 1:250
B
N.G.L.
Concrete slab with Penstock decant trench lined

dissipation blocks with geofabric and Reno mattress
Typical longitudinal section of decant trench
Stabilised
central basin
Existing tailings
deposit
Rock mattress
Figure 11.23. Details of water decant arrangements (after Smith,

1983).
704
~------ Air vent pipe

~--- Reinforced
~~---11---...,.,......,,., concrete plug
Penstock closure (top)
Penstock
Residue backfill
Concrete plug
Temporary plug support
(a) Penstock closure (bottom)
Graded rock
Graded filter or
filter fabric
Residue dam Perforated pipe
Penstock pipe
(b) Penstock closure at outlet using a graded buttress
Figure 11.24. Penstock closure methods (after Chamber of

Mines of South Africa, 1983).
the case of the outer berms, the top surface of the dam perimeter and
crosswalls, daywall and basin should be compacted and stabilised using
vegetation. This minimises and reduces wind and water erosion from the top
of the dam.
705
11.6.11 Return water dam

Only clear water from the top of the dam should be discharged to the return
water dam, so that the return water dam may be used as a clear water reser-
voir. If necessary, at the time of closure, it must be put in a condition such
that it will operate safely and with minimum maintenance for the foreseeable
future. The water in the return water dam could be used for irrigation,
discharged into the natural drainage system, used to recharge underground
water, or used for fish-farming, recreational fishing or water sports. To en-
sure high quality water, the owner should be prepared to consider stabilisa-
tion and hardening of the top surface of the slimes using lime or cement to
minimise contamination by silt and dissolved salts.
11. 7 References
Blight, G.E. and Steffen, O.K.H. (1979). Geotechnics of Gold Mine Waste Disposal. Geot.
Div. ASCE. pp. 152.
Caldwell, J .A. and Robertson, A.R.M. (1983). Selection of tailings impoundment sites. Civil
Eng. in S. Afr. Vol. 25, 73-95.
Chamber of Mines of South Africa (1983). Handbook of Guidelines for Environmental Pro-
tection: The Design, Operation and Closure of Metalliferous and Coal Residue Deposits. Vo ls
1 and 2. Chamber of Mines of South Africa, Johannesburg.
International Commission on Large Dams (1980). Manual on Tailings Dams and Dumps. Bulletin
No. 45.
Jennings, J.E. (1979). The failure of a slimes dam at Bafokeng. Mechanisms of failure and
associated design considerations. Civ. Eng. in S. Afr. Vol. 21, 135-141.
National Building Research Institute (1959). An investigation into the stabilty of slimes dams
with particular reference to the nature of the material of their construction and the nature
of their foundation. CSIR Report, Contract 5033/8002. Council for Scientific and Industrial
Research, Pretoria.
Smith, M. (1983). Avoiding costly closures of retired ring dyke gold deposits. S. Afr. Min. World.
Vol. 2, No. 10, 119 130.
Wates, J.A. (1983). Tailings disposal, the real cost of excess water in residues. J. S. Afr. Inst.
Min. Metall. Vol. 25, 257-262.
706
Chapter 12
Retreatment of Residues and Waste Rock

D.W. Bosch
12.1 Introduction
From the commencement of gold mining on the Witwatersrand in 1887 up
to 1984, a total of approximately 4,2 billion tons of gold ore has been milled
in South Africa. The deposition of gold mine residues has left the country,
particularly Johannesburg and its neighbouring cities in the central Witwater-
srand, with a legacy which has not only been an eyesore but also a source
of irritation and dust pollution in the dry, windy season, and contaminated
run-off water in the rainy season. Large amounts of money have been spent
combating this pollution and as a result of successful vegetating of their sur-
faces, the residue dumps in the Johannesburg area have now become well
known as landmarks identifying the Golden City.
Extraction of the low grade gold content of the dumps has long been
researched by metallurgists and recently a number of developments have com-
bined to make the retreatment of gold residues a profitable proposition, not
the least of which have been:
e the rise in the rand price of gold;
• modern metallurgical technology leading to more efficient extraction pro-
cesses with lower costs; ·
• the high value of the property which becomes available for further
development after removal of the residues.
12.2 Origin of Residues

12.2.1 Sand dumps and slimes dams
Residues contained in sand dumps and in slimes dams consist essentially of
three different products, namely separate sand and slime and slime from the
all-sliming process.
In the early days gold was recovered from stamp-milled material by gravi-
ty concentration only. Later, in 1890, the product of stamp milling was split
into sand and slime fractions. The sand was treated by percolation cyanida-
tion, whereas the slime was accumulated for lack of a suitable treatment
process.
In 1894 the decantation process for slime treatment was introduced. Both
accumulated and current slime were leached. This resulted in two discard
products; sand which contained about 11 OJo minus 75 µm, and slime which
contained about 95% minus 75 µm fines. Gold values of the material depend-
707
ed on a number of factors but ranged from 0,3 to 1,5 git for sand and 0,05
to 0,50 git for slime.
Tube mills were introduced in 1904 to grind the stamp mill product fur-
ther, and this reduced the ratio of sand to slime. This development was ac-
celerated by the introduction of the all-sliming process in 1918. However,
sand treatment plants were slow to be phased out and the method persisted
for many years. By 1936, 40% of the ore milled was still leached in sand
treatment plants. Almost 20 million tons of sand per year was still deposited
onto dumps. By 1946 this figure was considerably reduced and in that year
only 13% o-f the ore, amounting to some 7,3 million tons, was treated in
sand plants.
Today tailings from the all-sliming process constitute virtually all the
residues from gold plants, The size range of this material is between 65 and
80% minus 75 µm and typical gold values range from 0,1 to 0,5 git. In a
few exceptional cases accumulated tailings have values between 0,5 and 1,0
git. Gold plant tailings at present amount to a little over 100 million tons
per year and this figure is expected to increase in the future. The tendency
is to mine lower grades, which will result in lower gold values in residues.
A new development in the industry is the classification of the residue
by cycloning and the use of the cyclone underflow for backfilling
underground. During cycloning a small degree of upgrading of the gold value
occurs in the cyclone underflow and this gold will obviously not be available
for recovery at a future date unless treated underground. The cyclone
overflow which will be deposited on the slimes dams will be considerably
finer than the original residue (>95% minus 75. µm) and of a lower grade.
The backfilling programme is still in the early stages of development
and the amount of residue being placed underground is insignificant.
However, it is predicted that most South African gold mines will be using
residues for backfilling within 10 years. The amount of residue which can
be used for backfilling is up to 45 % of the tonnage mined. f\_ significant reduc-
tion in residue tonnage for deposition on slimes dams, as well as a reduction
in the gold content of this material, can therefore be expected.
12.2.2 Rock dumps

Waste rock arises from two main sources, namely underground development
and sorting. In both cases the waste is mineralised to a greater or lesser ex-
tent. The values are very heterogeneously distributed amongst the rock pieces
and vary so widely in magnitude that the grade of a bulk sample is only mean-
ingful if no upgrading is contemplated. Waste rock from off-reef develop-
ment can become contaminated during transport to surface by mineralised
rock from unpay and marginal areas, from mistakes in tramming and tip-
ping and by gold bearing fines with which it may have come into contact.
Rock pieces rejected by sorting should only be those mineralised to a
lesser extent than the cut-off grade. However, as currently available sorting
methods, whether manual or machine, rely on some secondary characteristic
of the ore which seldom correlates precisely with gold content, and as machine
sorting is carried out on the basis of interrogation of individual particles at
708
RETREATMENT OF RESIDUES AND WASTE ROCK
high speed, statistics dictate that there will be sorting errors. Whether the
waste rock comes from development underground or from the reject frac-
tions of sorting, it is conveyed to large dumps where it is stockpiled until
some use can be justified for it, for instance aggregate preparation for civil
engineering works, ballast for railway tracks, rockfill or treatment for gold
recovery. By far the largest quantity of material from the waste rock dumps
goes into civil engineering works without prior treatment for gold extrac-
tion. This was understandable when gold was $35 per ounce, but disposal
in this way needs regular review in the modern climate of fluctuating gold
price and currency exchange rates, continuously rising costs and the advent
of new technologies.
12.3 Sampling, Surveying and Valuation

When the reclamation of residues for retreatment is considered, informa-
tion pertaining to the dumps is required in order to carry out feasibility studies
that will determine whether a proposed retreatment programme can be pro-
fitable. Part of this information is acquired from the mine surveyors, who
would normally be responsible for supplying plans of dumps, dams and sur-
rounding areas as well as computing volumes and values of the residues to
be treated. The computation of dump or dam tonnage from volume obviously
requires the determination of the in situ bulk density of the material.
The plans, compiled by aerial photogrammetric means, provide
topographic and cadastral information essential for the planning and layout
of the plant and other buildings, pipe lines, conveyor belts, and for the ap-
plication for and registration of legal rights such as servitudes, way-leaves,
surface rights and dump permits (Section 161 of the Mining Rights Act and
Regulations (Act No. 20of1967)). A brief description of the sampling, survey-
ing and valuation techniques used to obtain the volumes and values of residues
is given below.
12.3.1 Sampling
A number of sampling methods are employed for the determination of residue
values. Time, cost and relevant requirements will dictate the method to be
used for any one set of circumstances.
12.3.1.1 Grab sampling

Samples are taken, preferably on an even grid system covering the entire dump
or dam surface. Easy access makes this a cost-effective method for the ac-
quisition of either small samples for assay purposes or bulk samples for
metallurgical testing, especially in a preliminary investigation. There is,
however, the danger of the samples not being representative in their response
to treatment processes due to surface oxidation.
12.3.1.2 Groove or channel sampling

This method is used quite effectively to sample a slimes dam that has been
covered by sand or, as is often the case; by building rubble or other refuse,
and only the sides of the dam are exposed for sampling. The samples are
709
cut from the slime in even, equidistant grooves or channels, on the line of
true dip over the entire length of the face slope. Care must be exercised to
reduce the measured length of channel sampled to the true vertical depth
of the dam. As for grab sampling, the method is useful in providing limited
information with reasonable ease.
12.3.1.3 Drilling
Auger drilling
Little use has been made of auger drilling for sand or slime sampling except
with lightweight hand-held augers, where penetration depth is severely limited.
Figure 12.1. Solid auger flight arrangement.
The Sandrill
A power-driven rig known as the 'Sandrill' is an hydraulically operated auger
drill which is extremely mobile and can reach virtually any site. The drill
is constructed with an outer casing having removable hardened cutters, within
which counter-rotating flights draw the sample up into the sample barrel.
Extensions of aluminium tubing are fitted with an inner drive tube on ball
bearings. The complete unit weighs approximately 73 kg (Figure 12.2). The
rig is capable of penetrating to depths in excess of 90 metres at a drilling
rate of between 50 and 300 metres per day, depending on the local condi-
tions and the depth from which the sample is being taken (MacDonald, 1983).
The Sandrill is the most effective of the sampling methods mentioned.
The other techniques, including the hand auger, are limited largely to pro-
viding surface information. To obtain relevant data it is necessary to penetrate
all parts of a dump or dam and the Sandrill accomplishes this task with
reasonable ease at a relatively low cost. However, this method of drilling
is not without limitations as the lightweight drill cannot negotiate hard ob-
jects such as rocks, wood and scrap iron frequently encountered in dumps
and dams.
710
Figure 12.2. Core drilling on a Witwatersrand sand dump using a 'Sandrill'.
12.3.2 Surveying
Surveys of gold residue dams and dumps and surrounding areas are required
for:
• the compilation of accurate working plans;
e the determination of volumes and tonnages for planning purposes, and
• the determination of volumes and tonnages for grade predictions and
metallurgical comparisons.
The surveys necessary to obtain the above plans, volumes and tonnages can
be carried out using a variety of instruments and methods. The method of
survey chosen will depend on the degree of accuracy required, the urgency
of the survey, and financial limitations.
12.3.2.1 Tacheometric surveys

The term tacheometry is derived from the Greek 'takhos' (speed) and 'metron'
(measure) and therefore signifies the art of rapid measurement. However,
the advent of electronic distance measuring instruments has to a large extent
superseded the 'tachy' survey. Nevertheless, when surveying small dumps
and especially stockpiles, the method is used to good effect. A tacheometric
survey requires a graduated tacheometric staff and a telescope for taking
readings from the staff at distance. The telescope forms an integral part of
the theodolite or tacheometer, which is an instrument used for the measure-
ment of angles in both horizontal and vertical planes.
The tacheometric survey of a dump or stockpile is carried out by setting
up the theodolite at a suitable survey station and observing positions of the
staff at a sufficient number of points on the dump surface, so that when
the co-ordinates and elevations of these points have been calculated and plot-
ted, a contour plan of the dump may be compiled. Volumes of dumps or
711
stockpiles are deduced by the construction of sections through the dump at

regular intervals. Sections may be transverse, longitudinal or horizontal, and
the area of each section is accurately determined by planimeter and substituted
in one of the following formulae to obtain the volume:
i) Trapezoidal Rule
V = D/2 (A + 2S + B), (12.1)
whereD is the regular interval between sections, A andB are end areas,
and S is the sum of intermediate areas.
ii) Simpson's Rule
V = D/3 [end areas + 2(odd areas) + 4(even areas)] (12.2)
The total number of areas must be odd and they must be at regular intervals
of D units apart. It is advisable, when applying the above rules, to use both
formulae and, if in reasonable agreement, accept the mean of the volumes.
The above rules have been derived from similar formulae used to ob-
tain areas of irregular figures but have proved very effective in the above
application (Dennis, 1974). The accuracy of plans and volumes derived by
the methods described above will depend on the number of points observed,
the number of sections drawn and the shape of the dump.
The use of electronic field books for capturing initial data, computer
programs for the calculation of co-ordinates, elevations and volumes, and
computer aided draughting for the compilation of plans, have greatly reduced
the amount of time spent in the office on· calculating and draughting.
12.3.2.2 Electronic distance measuring

The infrared distancer or distomat has simplified the measurement of
distances in the field and eases considerably the task of the surveyor in the
mapping of sand dumps and slimes dams. A number of different electronic
distance measuring (E.D.M.) instruments are available, which can either be
used independently, mounted on a theodolite, or as 'total stations' where
the theodolite and distomat are constructed in a single unit with automatic
recording of field data. As with tacheometry, detail is surveyed by observ-
ing distances and angles between a known survey station (beacon) and re-
quired points.
12.3.2.3 Photogrammet1y
Aerial photogrammetry
Photogrammetric methods have a wide range of use in survey, mapping and
the determination of volumes of slime and sand. Mapping from aerial
photography has many advantages over conventional ground surveys, the
most important being the speed and ease of carrying out the survey, and
secondly the high degree of accuracy which can be attained. Photographs
are taken from an aircraft flying in straight line parallel strips. Ground con-
trol surveys of strategic points, which are easily identifiable on the
photographs, are carried out and form the link between the ground survey
system and the stereoscopic model in the plotter. The stereoscopic plotter
712
Figure 12.3. (Upper photographs) Terrestrial photogrammetry:

pair of stereoscopic photographs showing working dump face
and 'ringed' survey points.
(Lower plan) Contour plan of the working dump face compil-
ed from the pair of photographs.
Note the unmapped area due to the loss of stereoscopy where
a minor 'sand-slide' occurred during the time lapse between
taking the two photographs.
SCALE 1:2 500
/
713
is used to compile an accurate contour plan of the dam or dump and, in

addition, can produce a digital terrain model (D. T.M.). The value of accurate
contour plans to the metallurgist to ensure optimum planning in conveyor,
pipe and drainage layout requires no fu_rther elaboration, but the many uses
of the D. T.M. are fairly new concepts and will be discussed more fully later.
A convenient scale for the mapping of residues is 1: 1 OOO, with a one
metre contour interval where, for standard photogrammetric procedures, the
scale of photography will be between 1:6 OOO and 1:8 500. Should a two-
metre contour interval suffice for any particular mapping requirement,
photography to a scale of 1: 10 OOO can be used for 1: 1 OOO mapping. It should
be noted that contour lines on a plan are generally accurate to within half
of the contour interval.
Useful visual information can be obtained from photographic mosaics,
which are aerial photographs pieced together, re-photographed and enlarg-
ed. As distortion will be present, especially towards the edge of an aerial
photograph, the scale of these mosaics can at best only be considered
approximate.
Terrestrial photogrammetry
For this survey the camera is on the ground and not in an aircraft. The re-
quirements of the photographs are similar to those for aerial photography
in that ground control survey points are required once again to establish the
relationship between ground detail and the photographic model in the
stereoplotter. The co-ordinates and elevations of a number ( Ciependent on
the size of dump or dam) of well-selected targets, easily identifiable on the
photographs, are determined by theodolite and E.D.M. (distomat) methods.
The points over which the camera is positioned are similarly ascertained. This
information, together with the photographs, is used to compile a plan view
of the working dump/dam face.
Volumes and values of residues

Successive photographs are taken at intervals determined by demand, and
as the latest analytical stereoplotters can accommodate a number of
stereomodels at any one time, it is comparatively simple to compile a plan
of two or more face positions including the intervening ground contours.
A D. T.M. of this mapped area is generated and can be integrated with the
D. T.M. of the original aerial survey. In addition, the borehole data and values
are added to this database and a statistical value distribution of the entire
dump/ dam is completed. From this information the volume and value of
sand or slime removed between successive surveys, as well as the recovery
grade, are calculated and compared with actual results as obtained by the
metallurgist during routine sampling and volumetric surveys.
Future planning in the recycling of residues

The computer-aided system described above is now used in the forecasting
of tonnages and values, either monthly for short term, or annually for long
term planning. Volumes and values are computed for any given tonnage to
714
be processed in the plant. Where the facility to blend grades from different
sources exists, various simulations are carried out on the computer, giving
the metallurgist the opportunity to choose the sequence best suited to his
requirements. Reports on tJ:ie simulated removal of slime and sand, giving
the tonnages and values expected, are produced, as well as plans depicting
the rate of advance of all working faces.
The exploitation of residues is a relatively new industry and undoubtedly
the means and methods described in this chapter will be continually improv-
ed and implemented.
12.4 Sand/Slime Retreatment Operations

The increase in the price of uranium in the early 1970's led to the construc-
tion of plants at a number of existing gold mines to retreat accumulated
residues for the production of uranium. The 1970's also saw the first of a
number of independent retreatment operations come on stream for the
recovery of gold, uranium and pyrite. One of the first of these new plants
was installed at the Blyvooruitzicht Gold Mine for the recovery of uranium
from accumulated gold plant tailings. This plant, which was commissioned
in July 1977 and closed down in 1984, incorporated several novel features,
including dry reclamation of the tailings by means of a bucket-wheel ex-
cavator, the first CCD (counter-current decantation) circuit to be used in
place of filters in South Africa, and the first NIMCIX continuous iop ex-
change circuit in a uranium plant.
This plant was followed by the Chemwes uranium plant, which was built
to recover uranium from old slimes dams at Stilfontein and Buffelsfontein
and later by a similar plant at Merriespruit to treat both current and reclaimed
slimes. Liquid/solid separation was accomplished on belt filters at both of
these plants; other unit processes were similar to those employed at
Blyvooruitzicht.
South Africa's largest reclamation operations are at ERGO (East Rand
Gold and Uranium Company Limited), which was commissioned in 1977
on the far East Rand and in W elkom, in the Orange Free State, where the
Joint Metallurgical Scheme has been operating since 1977. ERGO presently
reclaims about 1,6 million tons a month of old slimes dam material for pro-
cessing to recover gold, uranium and pyrite. In 1985, ERGO commissioned
a 1,5 million tons per month carbon-in-leach (CIL) plant to recover gold from
the current flotation plant tailings. ERGO is to commission an additional
one million tom per month CIL plant in 1986 at the old East Daggafontein
mine site. This plant will process those dams that are more amenable to direct
CIL treatment than to treatment in the flotation-CIL process of the original
ERGO plant. Tailings reserves for this operation include the flotation plant
tailings generated in the original ERGO plant prior to the erection of the
CIL circuit.
The Rand Mines Milling and Mining Company (RMMM) established
a treatment plant at a site a few kilometres south west of central Johannesburg
to reprocess 50 million tons of sand and 20 million tons of slime tailings deriv-
ed mainly from the old Crown Mines. At the time of its commissioning in
715
1982, the plant, which was designed to treat 370 OOO tons per month for the
recovery of gold and pyrite, included the 'largest CIP circuit yet built (Laxen
and Brown, 1984). A similar plant is to be commissioned by RMMM early
in 1987 at the old City Deep mine. Approximately 42 million tons of tail-
ings, largely sand, will be reprocessed.
The first of the large reclamation operations is the Joint Metallurgical
Scheme (JMS), which is an arrangement amongst those gold mines in the
Orange Free State administered by the Anglo American Corporation of South
Africa (AAC). By means of this arrangement, gold plant residues are retreated
in a number of flotation plants for the production of gold, uranium and
sulphuric acid from pyrite concentrates, and directly for uranium production.
There are a number of smaller retreatment operations on both the East
and West Rand. Egoli, a company which is less well known than the larger
South African mining houses, owns in excess of 70 million tons of surface
material on the East Rand. At Modderfontein 74, some 55 OOO tons per month
of sand and slime are being reprocessed in a plant which incorporates one
of South Africa's earliest carbon-in-pulp circuits. This company also operates
a plant on the West Rand, retreating sand from several old Randfontein
Estates dumps (Anonymous, 1982). Village Main, one of the best known
mines on the central Witwatersrand, would have been forced to close down
had it not embarked on retreatment of accumulated surface residues. In the
eastern Transvaal, the Fairview mine reclaims and treats some 25 OOO tons
per month of accumulated flotation tailings.
Descriptions of reclamation procedures and outlines of subsequent treat-
ment of reclaimed residues at some of these operations now .follow.
12.4.1 Slime reclamation at Blyvooruitzicht Gold Mine

Hydraulic redamation by means of high pressure water is the most commonly
used method for slime reclamation. However, in cases where this method
is undesirable or not permitted, for example in dolomitic areas such as
Blyvooruitzicht where sink-hole formation is a danger, mechanical means
must be used.
After examining several alternative mechanical reclamation systems, a
bucket-wheel excavator was selected for slime recovery at Blyvooruitzicht.
The system comprised a mobile excavator with its two integral transfer con-
veyors discharging onto a movable conveyor belt placed parallel to the face
being worked. This conveyor fed a movable repulper. After repulping to the
required relative density, the slurry was screened to remove unwanted over-
size material including rocks, vegetation and timber contained in the dam,
and then pumped to mechanically agitated surge tanks from where it was
transferred to the plant.
12.4.1.1 The Bucket-wheel excavator

The excavator used was an Orenstein and Koppel Model SH 250 bucket-wheel
excavator and consisted of a five-metre-diameter wheel fitted with 10 tooth-
ed buckets, each with a nominal capacity of 250 litres (Figure 12.4). The
bucket-wheel deposited material onto a short conveyor belt which discharg-
716
I '
I ' ... ',
(=-----~
ELEVATION
PLAN
/
Figure 12.4. 0 & K SH 250 bucket-wheel excavator.
ed via a chute. The superstructure housing the bucket-wheel and its discharge
conveyor could be slewed through 360 degrees. The chute discharged onto
a second transfer conveyor which could be slewed through 180 degrees relative
to the superstructure. The bucket-wheel and the rear transfer conveyor could
move in both horizontal and vertical planes. The maximum digging height
was 7 ,8 metres and the reach of the rear transfer belt was 15 metres. The
excavator was mounted on metal crawler tracks and had a maximum travel
speed of 20 metres per minute. The rear transfer conveyor was fitted with
a nuclear belt weigher to enable the operator to control the digging rate.
The excavator was originally powered by two 130 kW, 8 cylinder diesel
engines, which drove pumps to provide hydraulic power to all of the machine
drives. However, excessive engine maintenance was required due to the dus-
ty environment in which the machine operated, and the engines were subse-
quently replaced with a 261 kW electric motor.
717
Slimes dam
Bucket wheel Repulper

excavator
ELEVATION
100 m cross-feed conveyor

unitised with
200 m working face a movable repulper
PLAN
Figure 12.5. Mode of operation of slimes dam reclamation at

the Blyvooruitzicht Gold Mine.
718
12.4.1.2 Repulper feed conveyor

The repulper feed conveyor structure consisted of 10 metre sections joined
by plates connected with swivel pins. Each section had a pair of motor-driven
wheels at either end, running in heavy duty steel channel sections. The entire
structure could be advanced toward the working face by single push-button
operation of the wheel drives. The structure was aligned by advancing or
reversing individual wheel drives. Advancing and repositioning of the repulper
conveyor was an operation that took 2 to 4 hours. In practice, conveyors
were installed on two working faces so that reclamation continued on one
while the conveyor was advanced on the other.
12.4.1.3 Repulper
The repulper consisted of a 3,8 m x 2 m x 1,6 m rubber-lined trough. It was
fitted with a horizontal stainless steel shaft carrying 72 blades. The shaft was
driven by a 75 kW motor through a gearbox, and revolved at 20 rev/min.
Semi-flexible repulper blades were found to be most effective in coping with
oversize scrap, including stones, wire rope, timber, etc., contained in the
dump. The blades consisted of 575 mm lengths of 30 mm diameter steel rope
encased in rubber hose and sealed at both ends with an epoxy compound
to protect the rope from the acidic corrosive effect of the repulping solu-
tion. The blades were clamped to the shaft in a spiral. Blade life was about
three months. The repulper was mounted on crawler tracks and was towed
into position by means of a tractor or front-end loader.
Repulped slime, at a controlled liquid: solid ratio, was pumped via a 200
mm diameter flexible pipe over stationary screens into surge/ storage tanks
adjacent to the reclamation site, prior to being pumped to the plant (Figure
12.5). A detailed description of the rest of the Blyvooruitzicht uranium plant
is beyond the scope of this book. The unit processes employed are outlined
in 12.4 above. The development of the process and a description of the unit
operations involved have been described (Boydell, Laxen et al., 1977).
After uranium extraction, the tailings were subjected to flotation, where
a carboniferous/ sulphidic concentrate, containing about 9,5 g Ault was pro-
duced. This concentrate was treated for gold recovery in the existing gold
circuit. The tailings treatment plant closed down in November 1984 when
.the accumulated tailings reserve was depleted.
12.4.2 Slime retreatment at ERGO

The ERGO division of the East Rand Gold and Uranium Company Limited
will reclaim over forty slimes dams during its proposed life by monitoring
the dams using high pressure water. Many of the slimes dams have been
disposal sites for municipal refuse, abattoir effluents and sewage plant ef-
fluents, and over the years, to reduce air pollution, have been vegetated or
clad with broken rock. The dams were built on ground of varying topography,
ranging from flat sites to steep valleys. All of these factors influence reclama-
tion and subsequent processing, and must be taken into account.
719
12.4.2.1 Location of pump stations

The correct location and elevation of pump stations is vital to subsequent
smooth operations. Contours under the dam to be reclaimed are estimated
using borehole data, 1 in 1 OOO mapping with 1 m contours and larger scale
mapping of the surrounding area. Once these contours have been drawn,
the pump station is located adjacent to the wall with the lowest contour.
Testwork at ERGO has shown that the pulped material will gravitate down
an earth furrow with a slope of 1 in 150 (0,4°). This slope is a function of
the size and shape of the particles and the density of the pulp. Based on this
slope, a number of suitable sites for the pump station can be selected. It is
necessary to check the elevations of possible sites for the pump station, since
the cost of the pump station is approximately related to the cube of its depth.
Cost exercises must be carried out to establish the optimum arrangement.
If only one pump station is required at the site, it is normally equipped
with vibrating trash removal screens, the screened pulp then being pumped
to either a booster station or direct to the treatment plant. The screens are
required to remove oversize material which will cause pumping or pipeline
problems. These are fitted with polyurethane decks with 3mmx12 mm slotted
openings. Where more than one pump station is required, the main pump
station (i. e. the pump station to which the greatest tonnage will gravitate)
is usually a transfer station and the other pump stations are satellite stations.
In a satellite station, the pulp gravitates directly into a small sump feeding
an all-metal pump. Finger screens in the launder feeding the sump protect
the pump from large rocks or other debris .. The pulp is then pumped via a
rubber-lined pipe to vibrating screens situated in the transfer station.
At one dam the transfer station was built above ground with two satellite
stations feeding onto the screens at the transfer station. This reduced the
capital cost of the civil construction because of the depth that would have
been necessary for a conventional below-ground transfer station. The velocity
of pulp in the pipelines is maintained in the range 1,9 to 2,1 m/s. Details
of pumping tests and the results of practical experience have been published
(Sabbagha, 1982).
A catchment area is required downstream of the transfer sump to pre-
vent pollution in the event of an overflow. In a number of cases, a main
catchment area has been installed upstream of the pump station to which
the pulp is diverted in the event of any breakdown or operational problem.
A small downstream catchment area is then provided for any overflow from
the transfer sump. Oxidation of pyrite in the dams causes a lowering of the
in situ pH of the slime. In most cases, the reduction of pH is not sufficient
to warrant measures to prevent pipeline corrosion, but a facility for lime ad-
dition is installed where considered necessary.
12.4.2.2 Reclamation operations

Prior to reclamation of a dam, borehole samples are drilled on a 100 metre
grid. These samples are assayed and the distribution of values deduced from
the results. In addition, an aerial survey of the' dam is carried out and a
volumetric estimate of the dam is made on a 50 metre grid. Each quarter,
720
RETREA TMENT OF RESIDUES AND WASTE ROCK
the aerial survey is repeated and this has been found to be an accurate, quick
and cost effective method of calculating the volume of the dam remnants.
Various tests have been carried out to establish an in situ dry bulk density
of the slimes, and a figure of 1,473 t/m3 is currently used.
The monitor guns used are similar to those used by the English China
Clay Company. They are water nozzles constructed with a curved inlet and
a ,swivel bearing in such a way that the reaction thrust of the water jet is
balanced. The design of the curved inlet is critical to ensure the stability of
the gun irrespective of the direction of the water jet. The water pressure at
the gun is maintained at about 2 OOO kPa. The water velocity leaving the
gun is of the order of 45 m/s. The liquid: solid ratio of the pulp obtained
from the monitoring operation varies between 0,89 and 1,21 and a procedure
has been established which ensures that a fairly consistent density is produced.
At ERGO, the full height of the dam is monitored, whereas at some
other operations a benching method is used, where approximately four metre
layers of the dam are monitored, with the slurry gravitating between ben-
ches. The advantage of benching is that the pump stations do not have to
be deep in the ground to reclaim the majority of the dam. The gun can be
set up very close to the face and the full force of the water can be used, thereby
giving higher pulp densities. However, the final clean-up of the dam to natural
ground level will require either the transfer station to be sited at the correct
elevation or portable satellite stations to be used.
In order to minimise final ground clean-up costs at ERGO, the slimes
dams are monitored downslope to ensure that ground once cleared is not
re-covered with slimes in the event of a breakdown or stoppage at the transfer
station. The downslope method of reclamation also offers a benefit with
respect to rain water control, since rain water on the previously cleared sec-
tions of the dam can be routed away from the pulp launders. Final clean-up
can be carried out by a number of means as the face of the dam advances.
It has been found that routine mechanical clean-up is more cost effective
than a final mechanical clean-up once the dam has been removed.
Experiments on the automation of the operation of the monitor guns
are being undertaken and a sweep action for its operation during absence
of personnel has been installed. Further work using television cameras at each
gun with one operator at a central control room, and programmable guns
that will operate to a set programme for a period of time, are being
considered.
Three slimes dams are worked at any one time. Each of these dams is
monitored at a reclamation rate of about 550 OOO tons per month, providing
a total plant feed of about 1,65 million tons per month. A fourth supplemen-
tary dam is available when necessary to provide a higher sulphur grade in
the feed, to ensure that sufficient pyrite of a suitable grade is available to
the acid plant. After monitoring, and screening to remove trash, the pulp
is pumped in three separate streams in 450 mm diameter pipes to a main
booster station for pumping to the flotation section of the central plant
complex.
721
-..J >-3
N
N :::r::
i4NDWATER"
~>-----------------------------71----------, tI1
~
--------.. I
GROOTVLEI
PUHPING I
~ I >-3
~
SLURRY SERVICES
HONITOR RECLAMATION PROCESS WATER - RAND WATER
SlHHERGO WATER __ WATER BOARD
- ()
'T:I PYRITE BY DAH DAM DAM DAM RESERVOIR RESERVOIR
o;;· ROAD 2 3 I, ---- ---- >-3
......
.:
.....
~
! 7'/. CCNUM:D hOISTURt
(!) LEA(HED PYRITE
......
N
ll : I~ ~
°'0 I II AP THICKENER OVERFLOW I II
1
1
1
1 I II tI1
~
:~~;~ATE
(!)
::I t""'
~ t""'
e:.. FLOTATION
PfAITT
SULPHURIC c::
tT:I ACID PLANT ~
:;;;
[
-- 1-~M~----~ a
0 >-<
0 >----'--~ TAILINGS!-,~~=-~
$.A.LANDS 0
(J
0
.§
~ I ~ -APCALCJNE
a
>Tj
II 0
~
:::i
0
:;:
SIMMERGO
TAILINGS
c. 1. L. PLANT TAILINGS LIME SLAKING LIME I

I ! I I
GOLD PLANT I s
"'::r Q6!:! ~
ly..,..rr-,.-,-J------'=='-l~~___J
JI
(!)
~ I
I
~n I BULUON
~~~.ELUATE I
0
~~A~~o~ I
BY ROAD
12.4.2.3 Processing at ERGO

A general plant flowsheet is presented in Figure 12.6. Slurry from the three
slimes dams is treated in three separate circuits in the flotation plant to ac-
commodate any differences in flotation characteristics that may exist among
the slimes dams. The slurry is first diluted to the required liquid: solid ratio
and the flotation reagents added. A pyrite concentrate, containing a portion
of the gold and uranium, is floated. The tailings from the flotation cells is
thickened in four 138 metre diameter thickeners.
The overflow from the thickeners is recirculated to reclamation. The
thickened underflow is pumped to the carbon-in-leach (CIL) plant, where
it is first screened to remove coarse sand to prevent plugging of the inter-
stage screens downstream. Screen oversize is milled in a ball mill before be-
ing returned to circuit ahead of the desanding screens. The slurry is pumped
to a series of mechanically agitated contactors, where calcium cyanide is add-
ed. The contactors contain activated carbon granules which are kept agitated
in the slurry and which adsorb the gold dissolved by the cyanide. The car-
bon is moved counter-current to the slurry flow, ensuring low levels of dissol-
ved gold in the tailings. Carbon loaded with gold is separated from pulp,
which is pumped periodically from the first contactor; the carbon is strip-
ped of its gold content in the elution section using the AARL elution pro-
cedure. Eluted carbon is thermally regenerated and returned to circuit while
the gold-rich eluate is pumped to the gold plant for gold recovery. Slurry
tailings from the CIL plant are pumped to the tailings dam, situated 11 km
to the south of the processing plant, where the coarser fraction is separated
by cyclones for construction of the dam walls.
The pyrite concentrate from the flotation plant is thickened and transfer-
red to air-agitated pachucas where uranium is extracted using sulphuric acid.
The leached slurry, now containing uranium in solution, is filtered on drum
filters. The filter cake is re-slurried to 70% solids and pumped to the acid
plant. The uranium-bearing pregnant solution is clarified and then proces-
sed in a solvent extraction plant, where the uranium is purified and concen-
trated. Uranium is precipitated from the resulting uranium-rich liquor as am-
monium diuranate (ADU) by addition of ammonia. The ADU slurry is sent
to the Nuclear Fuels Corporation of South Africa (Nufcor) plant for filtra-
tion, drying and calcining to uranium oxide concentrate prior to export.
The leached pyrite from the uranium plant is fed into three fluid bed
roasters. The sulphur dioxide off-gas from the roasters is fed into two Lurgi
double contact acid plants. The total production capacity of the two plants
is 1 500 tons of acid per day. Part of the smaller plant's production is in
the form of oleum (fuming sulphuric acid) for sale to the explosives industry.
The calcine from the acid plant fluid bed roasters is quenched with water,
thickened and transferred to a cyanidation plant similar to those used on
all major South African gold mines. After being milled in one of two ball
mills, the calcine is pumped to air-agitated Browns tanks where the gold is
leached from the calcine using a sodium cyanide solution. The calcine from
the cyanide leach tanks is filtered and the filter residue pumped to a tailings
dam.
723
The gold in the filtrate is precipitated from the solution by the addition
of powdered zinc. This precipitate is filtered and the zinc-gold filter cake
calcined and smelted in an electric arc furnace. The resultant gold bullion
bars are sent to the Rand Refinery for final refining.
12.4.3 Sand and slime retreatment at Rand Mines Milling and Mining
Tailings were first reclaimed at Crown Mines during the early 1970's to sup-
plement the feed to the old plant when underground mining operations were
being run down. However, the lower grade of this material and the limited
rate of throughput imposed by plant capacity made total substitution an
uneconomic proposition using this process route.
Following the introduction and operation of a free market for gold in
the early 1970's, the potential of the accumulated tailings at Crown Mines
was re-evaluated. The result of these investigations was the decision to erect
the Rand Mines Milling and Mining (RM3) dump retreatment plant on a site
on the Crown Mines property central to the dumps intended for retreatment.
The primary function of the plant is to recover gold from the accumulated
residues, with pyrite as a by-product, and with the added benefit that land
will be released for future development.
Removal of the pyrite will remove a significant source of acid pollution
of the surface water on the Central Rand. The plant was designed to process
370 OOO tons per month of accumulated residue made up from 324 OOO tons
per month of sand and 46 OOO tons per month of slime. Construction began
in June 1980 and commissioning on slime in February 1982. The commis-
sioning of the milling circuit on sand commenced in April 1982. By early
1984 tonnage throughput had increased to a monthly average of450 OOO tons.
12.4.3.1 Process development for sand treatment

The flowsheet shown in Figure 12.7 was evolved from laboratory and pilot
plant testwork. The minus 75 µm fraction in the sand contains a significant
proportion of the total gold content and reacts favourably to direct cyanida-
tion. This indicated that a classification stage would immediately yield a pro-
duct which could be processed for gold recovery. As fine screening at high
tonnage feed rates was not a well established process, the application of
hydraulic classification was investigated and proved to be successful in
separating a proportion of this fine gold bearing material.
The pulp generated by mixing sand and water has a low pH. It was found
that the settlement of the fines produced by hydraulic classification was
superior at low pH to that obtained in tests conducted at the more conven-
tional gold plant pH of about 11. It was concluded that thickening could
be used as a means of recovering solids from the dilute suspension resulting
from hydraulic classification. The low pH and the presence of a pyritic frac-
tion in the sands suggested the possible application of flotation for separa-
tion of the sulphides. That this was feasible was confirmed in laboratory and
pilot plant tests, and sulphur recoveries in excess of 850Jo were achieved.
As gold was also recovered in the sulphide concentrate, the gold con-
tent of the tailings from flotation was lower than the original head value.
724
~From
sand dump
T WLJ[J;lllJLJLrLJ[f' --
Cyanidation
1---t.,
~
CIP
Lime
-------<---- ----------------,
Lime t To waste
• Loaded carbon
L.------~1
'
L-------~---: Slime
+6~'-sl-im_e_s_d_am~\ : ''
'
Reactivated :
carbon ~,
··.. I
Solution
to
I slimes
monitoring
'
~
I
~-~ :
ULJLJ ' \:'.::.:>--~Carbon
!
' Zinc dust
uuu vstorage
i
To dewatering
paddocks ULJLJ 1 dO:''
A cih Elution
·--_Storage Reacttvatmn
ULJLJ I
I
I
was
1
I
I
I I
I ~I r>C-----,
I/~
I I
Cyanidation :~LoadectL>I L-~
carbon
1
L__ ,...J Li
! A
1
LJuuuu ~- -------- --~

L1l Calcining
r
CIP
inecarbon
recovery
i
Smelting
To slimes dam
i
Bullion
Figure 12.7. RMMM Plant flowsheet.
Leaching tests carried out on the tailing over a range of grinds established
that it was unnecessary to grind the float tailing any finer than 45% passing
75 µm. It was observed, however, that dissolution from the fraction coarser
than 150 µm was poor. As a fine grind was not required it was considered
that closed circuit grinding would be unnecessary, and tests conducted on
a 1,68 metre diameter ball mill indicated it was feasible to satisfy these criteria
725
in an open milling circuit. Having achieved the size reduction and establish-
ed the leaching parameters for these materials, it was suspected that conven-
tional air agitated vessels could prove unsuitable for continuous process ap-
plication, as the coarse particles might settle out. Tests were conducted in
mechanically agitated tanks and their metallurgical efficiency and the suspen-
sion of the coarse solids were found to be satisfactory.
Whilst all testwork on leaching included filtration, the cost of a filtra-
tion installation for the tonnage throughput rate would have been high and
mechanical problems could be anticipated in suspending the coarse material
in the feed hoppers. The potential of the carbon-in-pulp (CIP) process was
evaluated and the decision was taken to follow this route.
12.4.3.2 Reclamation operations

As noted above, two forms of feed materials are recovered for treatment,
the one being described as sand and the other as slime. The mineralogical
composition of these materials is similar, the difference arising mainly from
their different particle size distributions.
Slime
This is reclaimed by conventional hydraulic methods with high pressure
monitors using water from various sources, including return dam solutions,
and operating at a discharge pressure of 1 600 to 2 OOO kPa. The pulp flows
by gravity to a flat deck vibrating screen fitted with 0,5 mm x 12 mm slotted
polyurethane panels for the removal of tramp materials such as oversize par-
ticles and vegetation. The underflow of the screen is pumped to the sieve-
bend screens at the processing plant and the oversize is rejected to waste.
The design pulp liquid: solid ratio is 1,00, but in the early stages of opera-
tion this varied widely. Current operations are much more stable despite the
throughput rate having increased to a level in excess of 60 OOO tons per month
from the original design capacity of 46 OOO tons per month.
Sand
The material is reclaimed with front-end loaders. The concept of this method
of reclamation is simple, but practical difficulties have arisen due to a varie-
ty of factors which had not been anticipated. The dumps have been used
in the past for disposal of all sorts of refuse. Metallic scrap in the form of
drill steel, rails and metal sleepers has been a major cause of belt damage,
and coarse rock and vegetation has caused blocked chutes and damage to
plant equipment when introduced to the system, resulting in operating pro-
blems, delays and additional cost.
Reclamation is further complicated by the nature of the deposit, where
alternating layers of sand and slime do not provide the stable footing re-
quired for the equipment. In addition, the angle of repose of the working
face has proved to be variable and is relatively unstable, resulting in hazar-
dous conditions when reclaiming only from the base of the dump (Figure
12.8 a and b). Testwork is in progress to assess the effectiveness of profiling
the working faces with water sprays, which should serve the dual purpose
726
of dust abatement and stabilisation of the face. Where possible the in-
termediate layers of slime are reclaimed by conventional hydraulic methods.
The material reclaimed by the prime movers is fed to movable hoppers
fitted with grizzly bars and vibrating feeders which feed onto conveyor belt
systems, which transport the sand to the stockpile at the plant. The stockpile
is circular and material is reclaimed from it by means of a scraping chain
conveyor feeding a draw-down chute onto the feed belt to the screens ahead
of the process.
12.4.3.3 Plant operations

Sand preparation
Sand reclaimed from the stockpile is fed by conveyor to a coarse scalping
screen, fitted with 3 mm x 19 mm slotted polyurethane panels and washing
sprays. The liquid: solid ratio of the minus 3 mm material is adjusted to 1, 7
and the resulting pulp is pumped to a four-way distributor which feeds the
primary cyclones. The overflow from the cyclones is fed by gravity to two
15 m diameter acid duty thickeners. The underflow is fed to mechanically
agitated primary conditioners prior to flotation. The pyrite concentrate from
flotation is collected in a 10 m diameter acid duty thickener. The thickener
underflow is fed to the discharge sump of a 1,68 m diameter ball mill in closed
circuit with a hydrocyclone. The overflow of this hydrocyclone reports to
a 25 m diameter thickener.
The flotation tailings are classified in hydrocyclones. The sand underflow
is dewatered to about 18% moisture on dewatering screens and is then
distributed to four 3,66 m diameter by 6,71 m ball mills operating in open
circuit. The hydrocyclone overflows and the underflow from the dewatering
screens are distributed to the primary float conditioner and the flotation tail-
ings sump for dilution control, and to the two 15 m acid duty fines thickeners.
A feed of alkaline return dam solution and lime is added to the dewatering
screen overflow at the mill inlet for density control and protective alkalini-
ty. Thickened fines from the 15 m thickenersjoin the mill discharge, which
is fed onto 0,5 mm x 11 mm slotted screens at the mill outlet. A transfer
pump then conveys the pulp to a leach circuit.
Leach and CIP

Leaching of the milled sand/ slime pulp is carried out in parallel in four rows
of agitators fitted with draft tubes. The tanks are 8 m in diameter by 16 m
high. The first stage of each row is currently used as a preconditioner. Calcium
cyanide is added to the second stage. At this point, partially loaded carbon
is also introduced by transfer of pulp and carbon from the adsorption sec-
tion by means of vertical spindle pumps. This is a recent innovation, effec-
tively converting the leaching section into a co-current CIL system. Air is
introduced at a point close to the bottom of all the leach tanks. The leached
pulp flows by gravity to the feed box ahead of the CIP adsorption tanks
where the four leach streams combine. Air-swept static screens retain the load-
ed carbon in the last stage leach vessels.
The six 8 m diameter by 16 m adsorption contactors of the CIP plant
727
'A' Dump
i Slime furrow
I (a)
I
cJPump sump
----Potentially
unstable area
Sand
Slime
~~~~~~sa-nd~~~-
Slime
Slime
(b)
0Pump sump
Figure 12.8 a, Plan of reclamation b, Elevation of reclamation.
728
are also fitted with draft tube agitators and have a variety of interstage, air-
swept screens to retain the carbon whilst the pulp flows through the system
by gravity. The carbon is moved counter-current to the flow of the pulp by
means of submerged vertical spindle pumps. The pulp leaving the final ad-
sorption stage is screened to recover fine carbon and then pumped to residue
dams which are approximately 7 km from the plant.
There are other feeds to the plant. Material from the monitoring of slimes
dams, after trash removal by means of sievebend screens situated above the
preconditioning tanks, may either be fed directly to the preconditioning stage
of the leaching circuit, or to the mill discharge screens. Slime found underlying
the sand is also monitored and may be introduced either directly to -the
leaching circuit or to the screen house.
Concentrate
The milled flotation concentrate is pumped from the 25 m thickener via a
mechanically agitated surge tank to a five stage leach circuit where lime and
cyanide are added. The leached concentrate is screened on a high frequency
inclined vibrating screen of 0,75 mm aperture, the overflow of which is
discarded and the underflow fed to the eight stage concentrate adsorption
section. As in the main plant, the adsorption tanks are fitted with vertical
air swept screens and the pulp flows by gravity whilst the carbon is moved
counter-currently. Discharge from the CIP plant is screened to collect escaping
carbon and is then pumped to settlement paddocks for drying prior to "ship-
ment to acid plant operations. The supernatant water is decanted and returned
to circuit to recover dissolved gold losses.
Gold recovery
Loaded carbon is removed from the last CIL stage (main circuit), or from
the No. 1 adsorption contactor in the pyrite circuit onto external circular
vibrating screens. The underflow from the screens is returned to the ap-
propriate adsorption/leach stage. The loaded carbon is transferred
hydraulically to columns where it is subjected to a hot acid wash with dilute
hydrochloric acid, rinsed and then transferred hydraulically to the elution
columns. It is then contacted with a caustic/ cyanide solution at temperatures
up to 120° C and rinsed with water. The resulting eluate is collected, de-aerated
by vacuum and the gold is recovered conventionally by zinc dust precipita-
tion, filtration, calcining and smelting to bullion. The post-precipitation tail-
ings solution is returned to the adsorption section.
The eluted carbon is recycled to the final stage contactor of the main
adsorption circuit via any one of the following three routes: direct hydraulic
transport, via rotary kiln regeneration, or via Rintoul kiln regeneration.
New activated carbon is added to the adsorption circuit manually to make
up for losses. Only fresh carbon is used in the concentrate adsorption circuit.
729
Solution balance
Three sources of water are available to the plant:
a) Rand Water Board - for make-up, to feed the CIP processing section
and for gland service.
b) Return dam solution - for mill dilution and the slimes reclamation
monitors.
c) River water - used as make up on the acid side of the circuit and for
applications such as hosing and flushing.
Internal water recycling
The overflow from the 15 m acid thickeners is pumped to a40 m water treat-
ment thickener where the acid solution is contacted with lime to precipitate
metal salts as their hydrated oxides. The overflow recirculates, via a storage
tank, to the acid side of the circuit to be used, for example, as dilution at
the screen house. The underflow from this thickener is pumped to the residue
surge tank and can, if assay values of the solids in the underflow demand
it, be pumped to the main leach circuit via the mill transfer pumps.
Problems encountered in plant operations

Metal scrap and rocks in the sand feeds can block the screen-house feed chute
and damage the screen decking. Also, if the fines content of these feeds in-
creases, they can form a cake on the screen, which causes overloading and
damage to the screen support and drive mechanism.
The performance of the flotation section has not matched the predic-
tions for sulphur recovery. A number of possible explanations exist to ac-
count for this:
• lack of pH control in the 40 m thickener circuit, which causes the pH
of the water to rise;
• mechanical conditions within the flotation circuit, i. e. inefficient froth
removal;
• the addition of surplus return dam solution to the acid water circuit,
which thus may contain cyanide, a known depressant for pyrite.
The establishment of stability in the secondary cyclone (dewatering) circuit
proved to be a major obstacle in tonnage throughput; however, this has been
solved by the following actions:
• false bottoms were installed in the pump sumps to minimise the conse-
quences of sand build-up and slumping;
• the vortex finders in the hydrocyclones were reduced in diameter.
The 3 ,66 m diameter by 6, 71 m mills were originally installed with grate and
scoop discharge ends and metal shell liners. High steel consumptions were
encountered as well as massive deformation of the shell linings. A rubber
lining was tested in one mill but did not prove to be cost effective. One mill
was converted to an overflow configuration and fitted with a different design
of rubber lining, and both steel consumption and liner life improved. All
730
mills have now been converted to an overflow configuration and the reduc-
tion in steel consumption has been maintained. The deformation of steel lining
systems also appears to have been eliminated.
The conversion of the first stage leach tank in each bank to a precondi-
tioner is a consequence of the observation that feeds other than milled flota-
tion tails were high lime, cyanide and oxygen consumers. The move has been
partially successful in that cyanide constimption has been reduced, but lime
consumption continues to fluctuate and saturation oxygen levels have only
been achieved in the leach by the use of liquid oxygen. Whilst CIP is dealt
with in detail elsewhere in this volume, some of the problems encountered
in the application of this technique at RM3 deserve mention here. The capacity
of interstage screens and their response to fluctuations in flow rate continues
to be a problem. The basic cause of this is believed to be contaminants such
as fibre and near mesh size siliceous particles which blind the screens. The
introduction of these contaminants to the carbon processing section also
causes problems with the screens inside the elution column and can possibly
cause interference with the carbon regeneration, particularly in the resistive
heating furnace. An additional contaminant to the circuit is tramp car-
bonaceous material which is similar in appearance to activated carbon. This,
it is believed, distorts the carbon concentration profile in the adsorption train
and the loaded carbon values.
In order to remove as much as possible of these contaminants, a sieve-
bend screening system was installed above the leach tanks of the main leaching
section.
Plant performance
Present plant throughput is about 440 OOO tons per month and extraction
efficiency is running between 65 and 73 OJo. Steel consumption is approximately
1 kg/ton of sand feed and power consumption by the mills approximately
8 kWh/ton sand feed. As mentioned previously, lime consumption fluctuates,
but is of the order of 7 kg/ton treated, whilst cyanide consumption is 400
g/ton treated. Water recovery from the slimes dams is currently estimated
to be between 30 and 35% of water fed to the dams.
12.4.4 The Anglo American Corporation Joint Metallurgical Scheme

The Joint Metallurgical Scheme (JMS) is an arrangement amongst the gold
mines in the Orange Free State (OFS), administered by the Anglo American
Corporation of South Africa Limited (AAC), to retreat gold plant residues
for the production of gold, uranium oxide and sulphuric acid. In summary,
gold plant residues, either current or reclaimed from slimes dams, are treated
in three flotation plants at Free State Geduld Mines Limited, President Brand
Gold Mining Company Limited and President Steyn Gold Mining Company
Limited (Figure 12.9). The pyrite concentrates are transported to a central
metallurgical complex at President Brand, where they are treated for the pro-
duction of uranium oxide and roasted for the production of sulphuric acid
with the calcines being treated for gold. The tailings from the President Brand
flotation plant are further treated in the uranium plant at President Brand.
731
From Pres. Brand Free State Geduld

Pres. Brand Flotation Plant From Western Flotation Plant
Holdings
r-·-- ---, ,----- ---, r - - ----1 , - - - - - - -,
t Gold • ' Pyrite ; j Gold j
,--- --,
j Pyrite • i Pyrite j
I concentrate
'------ ___ J
! I concentrate
L__ _ _ _ _ _ _ .J
f
c~L
L _ ___ j L_c~n~~t~a~_i • CIL
L_~
'
Residue
Residue
From Western URANIUM ACID Calcine CALCINE Residue

Holdings PLANT PLANT PLANT
,----- :-·1 Vibro energy

! Sulphunc i milling and
Pyrite
! acid j leach
stockpile
! production j
L,. _ _ _ ---...J
'f
;-·------ --·---·--·1
r·---------, ! Gold '
i i
"''~ J
Gold
L.!'.'Cr':'.~_J l____
Figure 12.9. The JMS Metallurgical Complex.
12.4.4.1 Historical overview

In the 1950' s, shortly after the start of mining operations in the Orange Free
State, two uranium plants were brought into operation. Production was
discontinued in the sixties as a result of weak markets. However, in anticipa-
tion of an upturn in the market, a 180 OOO tons per month uranium plant
was constructed at President Brand in the late sixties. In the event it became
necessary to mothball the plant until 197.6, when the uranium market even-
tually firmed.
At that time it was decided not only to commission the uranium plant,
but also to produce sulphuric acid and gold from residues. In order to take
full advantage of potential feedstocks owned by various AAC mines, the JMS
came into being on 1 January 1977. Essentially, the JMS is a financial ar-
rangement between participants in a joint venture whereby mineral recoveries,
operating costs and service charges (paid by plant users to plant owners for
amortisation purposes) are allocated among the participants. Initially em-
phasis was placed on uranium production and a major plant expansion took
place in 1980, when slimes treatment capacity was increased to 500 OOO tons
per month with the construction of a counter-current decantation thickener
circuit and a larger (1 700 m3/h) solvent extraction plant.
However, with yet another downturn in the uranium market in the ear-
ly 1980's, the emphasis was changed to maximise gold production whilst
reducing uranium production.
732
12.4.4.2 Feed sources

Prior to 1982, the uranium plant and the three flotation plants had indepen-
dent feed sources, with the uranium plant residues being stored for future
flotation campaigns. However, it was decided in 1982 that the uranium plant
should be fed with flotation tailings, which, besides being simpler, satisfied
the objectives of maximising gold recovery and minimising the cost of
uranium production.
Originally, the uranium plant residue had been a flotation plant feed
source, but entrained solvent in the slurry significantly reduced gold recovery.
Uranium plant residues were subsequently stored on a slimes dam, and it
is not yet known whether the depressing effect of the solvent will persist after
ageing. However, the effect of solvent has now been completely negated by
the present practice of flotation before treatment for uranium.
When the JMS was started some of the flotation feed was current gold
plant residue. However, residual cyanide in this material severely depressed
pyrite flotation, and since much of the residual gold is associated with the
pyrite, gold recovery was below expectations. While it is possible to over-
come the depressing effect of the cyanide by long conditioning with acid,
it was decided to concentrate efforts on reclaimed slimes. Also, the mineral
content of the reclaimed slime is higher than that of current material. The
slime is reclaimed from slimes dams constructed by the conventional ring
dyke method. Grade and size distribution variations between walls and ponds
are noticeable but these do not affect the flotation process significantly.
Reclamation is by means of hydraulic sluicing using Tourell monitoring guns
operating at a water pressure up to 3 500 kPa. A successful reclamation cam-
paign requires careful planning. The most important factors are the posi-
tion and elevation of the pulp pump station(s). These should be located so
that the pulp can gravitate from the operating faces. A practical minimum
design slope to obtain adequate pulp flow is 1: 150. In the case of large dams
located on flat terrain, extensive pumping from movable satellite stations
becomes necessary. Most of the OFS dams are free from vegetation and
dumped rubbish and consequently only basic screening is required. The sur-
face areas of slimes dams can be up to 200 hectares and as these form natural
catchments for rainfall, facilities must be provided for the diversion of
stormwater.
Efficient monitoring practices have been developed by the mines together
with their slimes dams contractors. Benches about four metres high are
developed which allow the monitoring guns to operate close to the face. High
impact velocities are thus maintained which maximise pulp density. It has
been found preferable to use small lightweight monitoring guns mounted on
wheeled supports which can be manhandled into position. This enables small-
bore water piping (100- 200 mm) with quick-fit couplings to be used, which
facilitates handling whilst keeping line pressure losses to an acceptable level.
12.4.4.3 Flotation plants

The three flotation plants at Free State Geduld, President Brand and Presi"
dent Steyn have a combined capacity of 1,53 million tons per month. Reclaim-
733
ed slime is diluted to a liquid: solid ratio of 1,73 and then conditioned for
at least four hours at the pH for flotation of 3,8 in air-agitated pachucas.
In the case of President Brand, hot acidic effluent water from the sulphuric
acid plant is available and has been found to be particularly beneficial, but
on the other plants pH adjustment is made solely by means of concentrated
sulphuric acid. Not only does the conditioning circuit provide surge capaci-
ty between the reclamation and flotation sections, thus permitting control
of the feed rate to the rougher banks, it also allows for removal from the
slime of acid soluble constituents such as lime added originally in the gold
plants. Mechanically agitated conditioners were originally installed for in-
tense conditioning with reagents but the high maintenance costs and the lack
of measurable benefit have led to their elimination.
Reagents are continually evaluated both on the mines and at the Anglo
American Research Laboratories (AARL) to evaluate new reagents on the
market and to maintain process efficiency with changing feedstocks. Original-
ly, yellow dextrin was used as a depressant for pyrophyllite, a fast floating
phyllosilicate mineral. A change to a modified guar gum led to an immediate
improvement in recoveries whilst maintaining the grade of pyrite concen-
trate to the roasters. The main collector is sodium mercaptobenzothiazole
(SMBT). On one plant this is enhanced by the use of a methylisobutylcar-
binol (MIBC)/paraffin mix which recovers kerogen, a light organic mineral
containing finely disseminated gold and uranium. The most widely used
frother is a polyglycol ether. These reagents are added simultaneously im-
mediately ahead of the pipe columns which feed the rougher banks, which
allows for thorough mixing. Copper sulphate solution is added as the pulp
enters the rougher banks.
The flotation cells are arranged in a conventional rougher-cleaner cir-
cuit. Steady state conditions are maintained by automatic control of pH of
the feed (3,8), liquid: solid ratio (1,73), reagent addition and feed rate to the
rougher cells.
Laboratory tests have shown repeatedly that pyrite floats relatively quick-
ly, but that other gold-bearing minerals respond more slowly. It has become
standard practice, therefore, to create froth conditions that allow for a high
mass pull in the first set of cells of the rougher banks in order to remove
as much pyrite of final product grade as possible. To achieve this, considerable
attention has been given to ensuring that froth removal launders and pipes
are of adequate size. It has also been shown that gold recovery varies in-
versely with sulphur grade in the final product. However, a minimum grade
of 300Jo has been found necessary for the sulphuric acid plant.
12.4.4.4 Treatment of flotation tailings

Cyanidation of flotation tailings can yield gold dissolutions up to 400Jo.
However, current gold grades in the flotation tailings are relatively low and
recovery of gold from this source is not necessarily economic. Previously-
deposited flotation tailings are of higher grade, but reclamation and redeposi-
tion, besides being costly, will require large amounts of water, which is a
scarce resource on the OFS goldfields.
734
At President Brand it has been possible to take advantage of the

amenability of flotation tailings to cyanidation with only minor capital ex-
penditure as the flotation tailings are presently treated for uranium recovery.
Cyanide is added in surge pachucas ahead of the acid uranium leach. Depen-
ding on the feedstock, up to 40% of the gold is dissolved, of which about
half is subsequently reprecipitated in the uranium leach. Unprecipitated gold
follows the solubilised uranium through the circuit and ultimately reports
in concentrated form to the solvent regeneration aqueous solution, which
is a discard stream of the solvent extraction plant. Gold is recovered by elec-
tro winning in open bath cells.
12.4.4.5 Treatment of pyrite concentrates

The pyrite concentrates are delivered in slurry form to the central metallurgical
complex at President Brand. Road tankers are used for transport from the
two distant flotation plants. After weighing and sampling, the concentrates
are leached for uranium by addition of sulphuric acid and injection of steam.
Liquid-solid separation is carried out on rotary filters with the residue being
repulped to 70% solids for direct feed to the sulphuric acid plant roasters.
Direct cyanidation of pyrite concentrates yields a gold dissolution of
between 40 and 60% depending upon the grade and origin of the pyrite.
However, after roasting dissolutions of up to 90% are possible. Hence the
1 150 tons per day sulphuric acid plant has an important role in the overall
flowsheet.
Roasting parameters have a significant effect upon subsequent cyanida-
tion efficiency. Highest gold dissolution and maximum sulphuric acid pro-
duction are obtained when there is a very low residual sulphur content
( < 0,05%) in the calcine. However, 'over-roasting' in the presence of an ex-
cess of oxygen results in the formation of some sulphur trioxide and conse-
quently in high costs for neutralisation of effluent water, corrosion of gas
cleaning equipment, damage to the vanadium pentoxide catalyst and the pro-
duction of a visible emission from the stack. Optimal control of the
pyrite/ oxygen ratio is therefore essential.
Pyrite arisings are in excess of sulphuric acid plant capacity. The market
for the excess pyrite has been found to be variable, and an investigation into
direct gold recovery from pyrite has been carried out. The most attractive
option is fine grinding. At the time of writing, a CIP plant incorporating
fine grinding using Vibro-Energy mills (V.E.M.) is under construction. In
these mills, a pyrite slurry flows through a bed of vibrating ceramic cylinders
where coarse particles are selectively ground. This technique has been wide-
ly used in other industries.
12.4.4.6 Treatment of calcine

Originally, the gold recovery plant included belt filters, clarification and zinc
precipitation. Persistently high soluble losses motivated the construction of
the first full-scale CIP plant in South Africa for treatment of calcine. After
cyanidation the leached calcine slurry is screened and then gravitates through
a train of seven flat-bottomed mechanically agitated contactors. Interstage
735
screening is by means of peripheral woven mesh panels. Owing to the low

linear throughput it has not been necessary to consider later screening
developments such as 'EPAC' (Equalised Pressure, Air Cleaned).
Selection of carbon for gold adsorption from calcine pulp has proved
to be critical due to the competition from base metal ions. Tests on synthetic
gold cyanide solutions have given misleading results, so all activity tests are
carried out on plant solutions. Carbon testing is also carried out by suspen-
ding carbon in baskets in the contactors. Gold loading on carbon, consis-
tent with an acceptable dissolved gold profile down the contactor bank, has
been below initial expectations. Carbon activity decreases with time in the
contactors, owing to progressive poisoning by lime, silica and base metals.
Pulp dilution from recycle streams, gland seal water and apron spillage has
a noticeable effect on solution gold concentration. Other adverse factors are
the relatively high slurry temperature and free cyanide strength.
Elution is carried out using the process developed by the AARL. Bat-
ches of carbon are transferred to a column where they are treated with hot,
dilute hydrochloric acid then soaked in a mixture of sodium cyanide and
sodium hydroxide at elevated temperature and pressure and finally eluted
with softened water. All the carbon is then regenerated in a' rotary kiln at
700°C. Electrowinning cells developed by AARL, which use an ion-exchange
membrane and pure sodium hydroxide solution as anolyte, are used to treat
the eluate. This type of cell permits the use of a high mass ratio of gold to
steel wool cathode which assists the subsequent calcining and smelting
operations.
12.5 Waste Rock Recovery and Treatment

The origins of waste rock are outlined in Section 12.2.2 above. The major
sources are underground development and sorting.
12.S.1 Development waste

The average grade of development waste is likely to be less than 1,0 g Au/ton,
although there are times when the value sent to the waste rock dump exceeds
this figure. At this value and assuming about 700Jo gold extraction, the revenue
from its treatment allows room for transport and treatment costs, provided
a fully amortised plant is available nearby. This contribution in itself has
often been justification enough to top up the plant when the supply of run-
of-mine ore is insufficient to match the available treatment capacity.
But consideration of average grade alone masks other possibilities. It
is important to determine the gold grade of the whole spectrum of size ranges
and the mass distribution of material in each, and from this to decide whether
one or more of these size fractions is worth setting aside for immediate or
future treatment. It is also important to determine the gold content of a
representative number of individual rock pieces within each size range above
about 25 mm, to be able to determine the feasibility of sorting the larger
rock pieces.
Because of the relatively high capital cost of a screening and washing
plant to classify run-of-mine waste into various size fractions, it is unlikely
736
that any development waste stream will warrant this treatment, particularly
if coupled to automatic electronic sorting. But if part of the infrastructure
already exists, for instance an aggregate preparation plant on site or a redun-
dant crushing, screening and washing plant, then these possibilities become
more realistic. An aggregate preparation plant can be used for washing waste.
The fines, which are a nuisance to the operator and likely to carry the highest
concentration of gold, can be screened out and classified. The slimes frac-
tion can be pumped to the mill thickeners and the sands conveyed to the mill
feed bin or submitted to heap leaching.
If a sorting plant is available, washed and sized waste can be conveyed
to the plant, sorted, the accept fraction sent forward for gold extraction and
the reject returned for aggregate preparation. If a sorting plant is not
available, there is still merit in examining the various products from the ag-
gregate plant. Representative samples of each size fraction, especially the
popular minus 6 mm crusher sand, should be subjected to simulated heap
leaching tests in large diameter columns to determine the period and chemical
conditions required to obtain optimum gold dissolution and its recovery on
activated carbon. On this basis it will be possible to calculate the economic
viability of a heap leach operation on any size fraction as obtained from the
aggregate plant, or as reduced by secondary crushing to a smaller size range.
Typically it has been found that some 75% of the gold in minus 6 mm crusher
sand can be heap leached in two weeks using a caustic solution containing
0,3 kg NaCN per ton of ore. The pregnant solution is collected, passed
through columns of activated charcoal for adsorption of the gold while the
barren is made up to strength and recycled through a spray distribution system
onto the heaps.
The ideal is for this operation to be carried out in close co-operation
and association with the aggregate producers drawing their raw material under
contract from the owner's waste rock dump. They carry out the crushing,
washing and sizing operations and return to the owner the high grade fines
for treatment in the plant. Having the necessary transport and handling equip-
ment, they can also build the heaps on the leach pad. After heap leaching
and washing is complete the heap leached material can be loaded into trucks
for sale as aggregate. Such a plant is under construction at the Kinross mine.
12.5.2 Waste rejected from ore sorting

Because each rock piece is interrogated individually, a well run ore sorting
plant should reject waste with an average grade significantly below that set
on the sorter, and the average grade of the reject fraction should not exceed
a value that equates in cents per ton to the downstream marginal cost of treat-
ment between the sorting plant and the slimes dam.
If the above criteria are met it is unlikely that extraction of gold from
the reject fraction from ore sorting could bear the costs of further crushing,
screening and heap leaching. Indeed it appears from tests that the percen-
tage dissolution of gold from fresh reject material crushed to minus 6 mm
is considerably less than the 75% obtained from this fraction when derived
from run-of-mine waste. Under these conditions rejects from sorting can be
737
sent as raw material to the aggregate preparation plant.
12.5.3 Radiometric sorting of waste rock

At the Hartebeestfontein mine, reef is picked manually from development
waste in the size range from 50 mm to 150 mm. About 35% of the sortable
rock is accepted as reef and the remaining waste is dumped at a grade of
about 0,5 g Ault. The total rock stockpiled on the dumps at present amounts
to about 30 million tons containing gold worth about R350 million at cur-
rent prices. Samples of this rock were tested to evaluate its sortability with
a radiometric sorter. Most of the individual rocks assayed at about 0,1 git
with very few assaying at above 0,5 git. A production scale radiometric test
facility was subsequently erected to evaluate the feasibility of sorting the rock
by radiometric means. Rock samples tested were typically 1 OOO ton lots.
Great care was taken to ensure that sub-samples for assay purposes were
accurate and extensive facilities were provided to sample the various pro-
ducts. The tests showed that the sorter accept fraction assayed at 1,5 git and
the reject fraction assayed at only 0,15 git. About 25% of the feed to the
sorter was accepted as reef. The rock sorted was in the size range 25 to 65
mm and 65 to 115 mm. Good washing and screening facilities were considered
to be essential for the success of this type of operation. High screening effi-
ciencies ( > 90%) were recommended. Jaw crushers were chosen for this pro-
gramme for primary crushing prior to screening because they tended to pro-
duce less fines than cone crushers. It was noted that waste material tended
to break as slabs in the jaw crusher.
The economic viability of radiometric sorting of the run-of-mine waste
proved to be marginal in this particular case. It was established that gold
in run-of-mine waste could be more economically recovered by means of
screening and washing of the waste, and processing the fines. However, the
replacement of the original reef-picking operation with radiometric sorting
appears to be promising, but two-stage sorting will have to be done.
12.5.4 Waste rock reclamation at the Blyvooruitzicht Gold Mine

Development waste generated in the underground mining operations at
Blyvooruitzicht Gold Mine has been hoisted at a single shaft over a number
of years, and deposited on a dump presently containing about 4,37 million
tons. This dump has an average value of 0, 7 - 0,8g Ault while the current
waste value is about 0,3g Ault. The -10 mm fraction, however, contains
about 2g Ault. This fraction comprises about 15% of the current waste, but
it can go as high as 40% in the dump. A considerable portion of the gold
content can therefore be recovered in a relatively small fraction of the total
tonnage in a simple screening and washing plant.
The waste rock reclamation plant at Blyvooruitzicht is designed to treat
both current and reclaimed waste at rates of about 25 OOO and 40 OOO tons
per month respectively. A flowsheet of the circuit is shown in Figure 12.10.
During the periods when waste is being hoisted, plant feed is drawn from
the shaft waste bin; at other times, rock is reclaimed from the waste dump
by means of a front end loader via a ground bin. Rock is transferred from
738
Shaft bin
Front end loader
Magnet
~ Cl
c
Mass meter
________ - 50 mm screen
~-~~---------10 mm screen
i
~
- -
- - -
New dump Mill
bin
Secondary Primary
mill mill
~--~g--------- To main plant

Mass meter
Figure 12.10. Waste rock reclamation.
the ground bin or the shaft bin to the washing screen on separate conveyors.
The washing screen is a 4,3 m long by 2, 7 m wide double-deck vibrating
screen with a manganese steel top deck of 50 mm opening and a polyurethane
bottom deck of 10 mm opening. Washed oversize material ( + 10 mm) is
deposited via a series of conveyors and a spreader onto a new 'treated' dump;
- 10 mm material is fed to a 1,2 m Akins Simplex classifier. The Akins
classifier underflow is fed by conveyors via a mill bin to a two-stage milling
circuit. The circuit comprises two 1,6 m diameter by 6, 7 m milis. The fines
are fed to the first (primary) mill which discharges into a common sump.
From the sump the material is pumped to a 610 mm primary hydrocyclone.
The overflow of the primary cyclone is fed to a 610 mm secondary cyclone.
The overflow of the secondary cyclone, containing about 75% minus 75 µm
material, is fed to two 15,2 m diameter thickeners. The underflow of the
739
primary cyclone is the feed to the secondary mill and the secondary cyclone
underflow is returned to the sump. The overflow of the Akins classifier is
also fed to the mill sump. The grinding media in the mills are 70 mm diameter
balls.
The thickener underflow is pumped to a surge Browns tank from where
it is transferred to the main gold plant, some distance away, for treatment
with current low grade ore. Gold extraction is about 88 % . Waste rock feed
rate to the plant is monitored by means of nuclear belt weighers, while a
mass- flow meter is used to measure the tonnage transferred to the gold plant.
12.6 Heap Leaching of Dumps

Heap leaching is a process which is quite widely used, particularly in the
western United States, for the leaching of low grade gold ore. The process
has not yet been applied to South African run-of-mine gold ores, although
in situ leaching in underground stopes is being tested in a number of mines.
Heap leaching has also recently been applied to leach reclaimed sand and
waste rock. One operation on the East Rand treats about 25 OOO tons per
month of accumulated sand on three leaching pads. The pad loading-leaching-
unloading cycle is three weeks. Gold is recovered from the leach solution
in a cascade of upflow columns on activated carbon. Considerable difficul-
ty has been experienced in controlling the pH in the heap because of the
variable acidity in the old dump. Pre-liming of the sand has been found to
be essential to prevent the solubilisation and subsequent precipitation of base
metals, which result in blinding of the heap, and to contain cyanide
consumption.
The two major variables which must be considered in the selection of
a heap leaching process are the permeability of the ore and the percolation
of the leach liquor through the heap (Thorndycraft, 1984). The former refers
to the ability of the leach solution to penetrate ore particles. This penetra-
tion takes place through micro-cracks in the ore particles or along grain boun-
daries. The rate of penetration may be improved, in the case of very compe-
tent rock, by fine crushing to expose the gold. It is clear, however, that
penetration of the ore by the leach solution is a prerequisite for successful
heap leaching.
The second requirement for successful heap leaching is the free percola-
tion of the leach liquor through the heap to ensure adequate wetting of the
ore. Poor percolation may result in channelling and the leaching of only a
small percentage of the heap, the wetting and leaching of only an outer layer
of the heap, or flooding, which can cause severe erosion of the heap. This
problem usually relates to the particle size distribution, and particularly to
the amount of fines in the heap. Fines migrate downwards with the leach
solution and cause blinding. Fines generated in the heap during leaching due
to ore degradation can similarly result in heap blinding. Poor percolation
can sometimes be overcome by crushing and agglomerating the ore, but this
may not be economic.
The procedures used in heap construction, particularly when leaching
740
run-of-mine ore, are extremely important. These should be designed to

minimise size segregation in the heap, which is a cause of channelling, and
to minimise compaction which hinders percolation.
12. 7 Bacterial Leaching of Dumps and Slimes Dams

12.7.1 Theoretical aspects
It has long been known that valuable metals remaining in dumps of sulphidic
waste material become solubilised with time and can be recovered in the rain-
water run-off. This was frequently believed to be due to 'weathering' and
the result of chemical reaction. Acidic run-off also occurs from South African
gold mine waste dumps and generally becomes apparent following alkalini-
ty depletion in the dump. In this case the run-off contains iron salts and is
not worth recovering. However, it has been found that gold previously en-
capsulated in sulphide particles, becomes liberated.
It was not until about 30 years ago that it was found that solubilisation
of metals in dumps and mines was due mainly to bacterial action. The bacteria
responsible are generally referred to as chemolithotrophic or 'rock eating'
micro-organisms, as they obtain their energy from the oxidation of inorganic
material such as sulphides, sulphur and ferrous iron. Oxygen and carbon
dioxide from the air are used for this purpose. The carbon dioxide provides
the 'building blocks' for the organic body composition of the micro-
organisms. Trace amounts of salts of magnesium, calcium, phosphorus and
potassium are also necessary. These are obtained from minerals and associated
mineral water, but supplementary sources are sometimes necessary. Nitrogen
is essential, but as these micro-organisms do not fix nitrogen from the air,
their only source is whatever ammonium salts are dissolved in local
waters.
There are a large number of chemolithotrophic bacteria, but those most
commonly known to be involved in the leaching of ores are listed below with
the inorganic species they oxidise.
Micro-organism Oxidised lnorganics

Thiobacillus ferrooxidans s0 , s2 - , s2 0;-, Fe2 +
Thiobacillus thiooxidans s s2 - , s2 0;-
0
,
Leptospirillum ferrooxidans Fe2+
Like all other micro-organisms, there are a great number of strains of

each type which have a symbiotic relationship with others. Natural selection
therefore develops a mixed strain of various bacteria that are best adapted
to decompose the minerals present in the particular environment.
Autotrophic bacteria decompose most sulphide minerals. The reactions,
as illustrated below, result in the formation of various compounds, many
of which react further by normal chemical routes. Of particular interest is
the formation of jarosite, which occurs in substantial amounts under the con-
ditions in which the autotrophic bacteria proliferate. It will be noted that
the formation of jarosite releases acid, which helps maintain the environ-
741
ment at the pH most suited to the micro-organisms. It is of interest to note

that the colour change from grey to yellow on the surface of slimes dams
is the result of the formation of basic iron salts such as jarosite.
Bacteria
Bacteria
4FeAsS + 1302 + 6Hz0
Bacteria
CuFeS 2 + 402 CuS04 + FeS04
Bacteria
4FeS04 + 0 2 + 2H2 S04 2Fez(S04) 3 + 2H2 0
Chemical
FeS2 + Fe2(S04) 3 3FeS04 + 2S 0
Chemical
2H3 As04 + 16FeS04 + 5H2S04
+ 2S 0
Chemical
Fez(S04) 3 + 6Hz0 2Fe(OH) 3 + 3H2 S04
Chemical
Fe(OH) 3 + H 2 S04 Fe(OH)(S04) + 2Hp
Chemical
3Fez(S04 ) 3 + M2 S04 + 12H2 0 2MFe3(S04)i(OH) 6
+ 6H2 S04
where M = Na, K, NH4 , H.
Ideal conditions for bacterial leaching are as follows:

• the ore particles should be as small as possible;
e there should be an adequate nutrient supply;
• bacteria usually occur in heaps by natural inoculation. However, an
inoculum may have to be supplied in relatively new heaps;
• air and water must be able to permeate easily through the mass;
• the heap must be situated on a suitably impervious base.
12. 7 .2 Dump leaching by bacteria
Dumps usually contain fairly coarse particles and so could be amenable to
the technique of bacterial leaching. Application of the nutrient solution would
generally be done by spraying for a specific period followed by a rest period
while the solution drains dcwn through the dump. In the case of base metal
leaching, the solution draining from the dump is treated by means of a suitable
recovery process and the barren liquor recycled back to the dump with the
addition of nutrients and pH adjustment. The recycled solution should not
contain chemicals, for example solvents, which may prove harmful to the
bacteria. Sulphide decomposition and thus metal recovery depends on the
extent to which the leach solution can diffuse into the particles. Sulphide
decomposition in a dump is generally only about 40 - 60%.
In the case of dumps that contain gold encapsulated in sulphide minerals,
bacterial leaching will liberate the gold; however, as bacterial leaching results
in the dump becoming acid, it would then have to be rendered alkaline prior
742
to cyanidation. Bacterial leaching of dumps is still in the developmental phase

in South Africa.
12.7.3 Slimes dam leaching by bacteria

The material in a slimes dam is generally very fine and therefore highly com-
pacted, with the result that the interior of the dam is almost impenetrable
to water and air. Thus the slime is largely preserved as when originally laid
down. As air and water penetrate the surface and sides of the dam, with
natural bacterial inoculation, oxidation of the surface sulphide mineral takes
place fairly rapidly. However, within the dam oxidation is found to take place
to a depth of only half to one metre, depending on the age of the dam. Gold
is liberated in the oxidised zone and recovery may be achieved by removing
the oxidised material and cyaniding. The oxidised slime, however, is only
a small portion of the dam. Recovery of gold from the remainder of the dam
could be achieved by ploughing and inoculating successive layers of slime.
Provided the moisture content of the ploughed mass is sufficient to ensure
all particles are wetted, decomposition of slime containing 4% sulphide
minerals will be completed in about 35 days.
The above procedure can be carried out on the dam or, for closer
moisture control, the material could be transported to a suitable concrete
pad for bacterial oxidation. Another method, which is receiving increasing
attention, and which may have application for treating ores such as the
arsenopyritic gold deposits found in the Eastern Transvaal, is the biological
oxidation of ore pulps or concentrate in agitated vessels prior to cyanidation.
12.8 References
Anon. (1982).There's still good profit the second time around. Coal, Gold Base Miner. 30, 59- 63.
Boydell, D.W., Laxen, P.A., Bosch, D.W. and Craig, W.M. (1977). The new uranium recovery
circuit at Blyvooruitzicht. CIM Bulletin, 72(805), 127 -134.
Dennis, E.F. (1974). An exercise on the calculation of the volume of a small irregular-shaped
chrome dump. J. Inst. Mine Surv., S.A., 17, 179-222.
Laxen, P.A. and Brown, T.D. (1984). The carbon-in-pulp plant at Rand Mines Milling and
Mining: problems encountered and developments introduced. In Proceedings of the Interna-
tional Conference on Recent Advances in Mineral Science and Technology (MINTEK 50),
edit. L.F. Haughton. Council for Mineral Technology, Randburg.
MacDonald, E.H. (1983). Alluvial Mining - The Geology, Technology and Economics of
Placers. Chapman and Hall, London and New York, 203-209.
Sabbagha, C.M. (1982). Practical experiences in pumping slurries at ERGO. In Proceedings
of the 8th International Conference on the Hydraulic Transport of Solids in Pipes. BHRA
Fluid Engineering and the South African Institute of Civil Engine'!!rs, Johannesburg.
Thorndycraft, R.B. (1984). Heap leaching methods in the western United States and applicability
to Canada. In Proceedings of the J6th Annual Meeting of the Canadian Mineral Processors.
Ottawa.
743
Chapter 13
Process Evaluation
H.E. Bartlett and D.M. Hawkins
13.1 Intrnduction
Two kinds of information proce§sing are important in the control of a metal-
lurgical plant:
1) Systems for gathering and analysing data of a long-term nature for statis-
tical and accounting purposes. For these purposes, composite informa-
tion over a minimum time span of a shift is required. For some purposes,
shift data are summed for daily, weekly and monthly information. This
is the long-term strategic control aspect - termed process evaluation in
this chapter.
2) The control of a plant as it is operating in order to attain specific objec-
tives, be they even flow, maximum throughput or maximum recovery.
For this control, information on the performance of the plant is required.
This information is analysed immediately and appropriate action taken.
This is the real time dynamic aspect and is termed process control: it is
dealt with in Chapter 14.
The requirements for these are different. For process evaluation, the
primary requirement is accurate information on the historical performance
for metallurgical accounting, cost control, etc. In this case, rapid feedback
of information is not required.
For process control, however, it is vital that information is available
rapidly enough so that the behaviour of the plant can be influenced or con-
trolled. For this, it is not essential that the information is absolutely accurate.
The requirement is for reproducibility or repeatability to indicate whether
a particular control parameter has changed so that the appropriate action
can be taken.
Because the requirements are so different, these two aspects of control
are covered in separate chapters. This chapter deals with process evaluation.
13.2 Process Evaluation

The components of process evaluation are:
1) sampling of process streams to obtain representative samples;
2) preparation of these samples, usually by drying, crushing, pulverization
and sub-division to obtain a sample of the correct size for analysis;
3) measurement of the mass flow-rate of the process streams;
4) analysis of the samples; and
745
5) analysis of the data to calculate metallurgical balances.

Before dealing with the individual topics, it is necessary to define some of
the terms that are used. Extensive use is made of statistical techniques and
notation. It is not appropriate, however, to delve too deeply into statistical
concepts in this book. The approach, rather, is to introduce those aspects
which are of direct application and to refer the reader to statistical texts for
more detail (Spiegel, 1961; Rendu, 1978; Box, Hunter and Hunter, 1978).
13.3 Statistical Concepts

13.3.1 Population, sample, increment and composite samples
A fundamental statistical concept is that of a population which, in the con-
text of process evaluation, is every possible outcome to assays of a particular
batch of materials, mass measurements or any other observation that may
be made. However, instead of obtaining a whole population, a small part
of the group called a sample is usually examined.
This definition of samples from a population is distinct from a sample
of a batch of ore, residue etc. These physical samples are usually composites
comprising a number of increments which, in turn, are a group of particles
extracted from a lot of ore in a single operation of a sampling device.
Although the same word is used for a sample from a population and a sample
of an ore etc., the distinction between them is usually clear from the context.
13.3.2 Accuracy and precision

It is vital that a clear distinction is made between the concepts of accuracy
and precision, terms which are very commonly confused. Accuracy implies
an absence of bias. The meaning of bias is best illustrated by way of an
example. Assume that the content of gold in a sample is 6,00 git. This value
is not known but let us suppose that by repeated analytical determinations
an average result of 6, 1 git ( ± 0,2 git) is obtained. The figures in the bracket
are those within which the results could be expected to lie in 95 out of 100
cases. These are the upper and lower confidence limits at the 95% level.
In this particular example the confidence limits are ± 0,2 git. However,
by analysing more and more samples, making the number of determinations
very large, let us suppose that 95% confidence limits fall to ± 0,01 git and
the mean becomes 6,11 git. As the true value is 6,00 git, there is a bias of
0, 11 git and the precision of the measurement is ± 0,01 git.
Repeated measurements improve the precision to give a sample mean
value within closer limits of its long-term mean but do not eliminate bias.
Accuracy has to do with closeness to truth, precision only with closeness
together.
13.3.3 Standard deviation, variance, coefficient of variation

The standard deviation of a sample of N observations x 1, x 2 , x 3 , xn in a
population is denoted by and is defined by
746
PROCESS EVALUATION
.x)2 + (xn - x)2

(13 .1)
n - 1
n
2:,; (X; - x)2
i = I
(13.2)
n - 1
where x is the mean of all the observations.
A mathematic11lly equivalent formula which is easier to manipulate on
a calculator is
(13 .3)
Variance v is the square of the standard deviation.

v = i2. (13 .4)
The coefficient of variation, which is usually presented as a percentage, is
given by
CV = (six). 100% (13.5)
where x is the mean of a series of observations.
13.3.4 Properties of variances and standard deviations

There are a number of properties of standard deviations or variances that
are of importance. These are
a) If a number of independent operations is involved in a particular deter-
mination, the total variance for the whole operation is the sum of the
variances of each individual operation. As an example, the total variance
vt in the determination of the mass of gold in a process stream is com-
posed of the variance of sampling, v,, the variance of assaying, v a• plus
the variance associated with mass measurement vm.
(13.6)
A consequence of this is that there is little point in one component in the
measurement process being much more precise than the others as the
overall precision is largely determined by the most imprecise measurement.
b) The standard error in the mean of a sample taken from a population is
given by the expression
Standard error in mean = s/ ~n (13 .7)
where s = standard deviation and n number of measurements.
c) Upper and lower confidence intervals.
The mean value of a sample taken from a population will be within ap-
proximately twice the standard error of the mean in 95 out of a 100 cases.
747
0,4
0,3
s
~ 0,2
0,1
µ-2<J µ-<J µ µ+ (J µ+2<J
Figure 13.1. The normal distribution.
Therefore the 95 OJo confidence interval for the mean is

± 2s/ ~n.
This figure is often quoted in brackets after a measurement to indicate its
confidence interval.
d) For a 'normal' distribution as defined below, for sample sizes greater than
30, about 68% of the results will be within one standard deviation of the
mean, about 95% will be within two standard deviations and about
99,73% within three standard deviations.
13.3.5 Normal and log-normal distribution

In the sampling of ores and other solid materials containing gold, we are
looking at particles of gold distributed within a barren host. If there are a
large number of particles and these are randomly located within the host,
repeated measurements will be normally distributed. A normal distribution
is. characterized by the symmetrical bell-shaped distribution curve of the type
given in Figure 13 .1.
The important features of this distribution are that it is symmetrical
about the mean, that is, there are as many high as low values, and that the
number of observations within a particular range are known from the mathe-
matical definition of the normal distribution. For example, 68% of the obser-
vations will be between x - s and x + s, and the number of observations
above or below three standard deviations from the mean will be 0,14%. In
contrast to the above, if the particles of gold are in clusters or are large and
therefore relatively few in number, a normal distribution need not be follow-
ed. Typically, in sampling of gold, distributions are skewed to the right or,
in other words, have a greater number of high values than the normal distribu-
tion would suggest.
An example of a type of distribution that is encountered is given as Figure
13 .2, which is the distribution of the gold values in a set of increments taken
from a feed belt to a crusher plant. The top 5% of the ore had a particle
size greater than 150 mm.
748
PROCESS EVALDA TION
300
"' 200
Q)
0.
Ei
ill
4-<
0
....
Q)
.D
Ei
z"' 100
8 12 16 20 24 28 32
Assay value git
Figure 13.2. Typical histogram for run-of-mine ore samples.
Figure 13.2 shows that in a set of 741 observations, 10 or 1,3% of the

values were greater than 4 standard deviations from the mean and about 17
or 2,3 % were greater than 3 standard deviations from the mean. This com-
pares to the normal distribution, where it would be expected that only 0, 14%
of the observations would be greater than 3 standard deviations from the
mean and the number above 4 would be vanishingly small.
The extent to which a distribution deviates from a normal distribution
can be quantified by two parameters, namely the skewness, which is zero
for a normal distribution, and the kurtosis, which is three for a normal
distribution.
Skewness is defined by
m3
Skewness = - (13.8)
s3
where m 3 is the third moment about the mean and s is the standard
deviation.
The rth moment about a mean is defined by
1 n
mr = - I; (x; - x)'. (13.9)
ni =I
Kurtosis is defined by the equation

m4
Kurtosis = (13.10)
7·
749
It is a measure of the tendency of observations to give occasional data many

standard deviations from the mean. For a distribution with a kurtosis greater
than three, there is a more 'peaked' distribution than the normal distribu-
tion. Conversely, with a kurtosis of less than three, a 'flatter' distribution
is obtained.
The example on the distribution of gold values on a feed belt demon-
strates that it could be misleading to apply normal statistics without first
testing to see if the distribution is in fact normal. On the basis of the statistics
of normal distribution, the assumption could have been made that all results
more than three standard deviations from the mean should be regarded as
arising from analytical blunders and eliminated from the data set. This would
result in an under-evaluation in this particular case of about 4% in the head
grade. It must be emphasized, however, that the rejection of high (or low)
values should only be done if there is definite proof of analytical or other
blunders.
The distribution of gold in run-of-mine ore, or crusher feed, from a
specific ore body is probably the most unfavourable to be encountered in
plant sampling from the point of view of conformity to a normal distribu-
tion. As the particle size of the ore is reduced by crushing, milling etc., and
as a result of mixing in bins, mills, thickeners and other processes, the ·distribu-
tions become more nearly normal.
In underground evaluation skewed distributions with occasional very
high gold values are commonly encountered. These data can be transformed
to give a symmetrical distribution curve by taking natural logarithms of the
values and plotting these on a histogram. If a normal distribution is obtained
after this transformation, the distribution is said to be log-norm111. A further
refinement is to introduce a constant which is added to the value before taking
logs and establishing whether this produces a normal distribution when a
histogram is constructed. This is called a 3-term log-normal distribution.
On the basis of the foregoing discussion, it is wise, as a first step in
defining a sampling procedure, to establish the likely distribution of the
results. This can be done from previous history or from a preliminary
sampling campaign. The results should be plotted on a histogram, possibly
using an appropriate computer package, and examined visually to see if there
are any very high or very low results and to see if a smooth bell-shaped curve
is obtained.
The values of the skewness and kurtosis quantify the extent to which
the distribution is not normal but it is difficult to decide at which level of
these two parameters it is necessary to consider transforming data to logs
or other functions and re-examining the distributions. Fortunately, however,
in most situations where plant products and feeds are examined, normal
distributions apply because of the mixing that occurs before the ore reaches
the surface. Where this does not apply, the reader is referred to Rendu (1978)
for the detail on the use of log-normal statistics.
750
PROCESS EVALVA TION
13.4 Sampling Theory

13.4.1 Calculation of the precision of sampling
If a set of assays on samples from a process stream is obtained, it is a simple
enough procedure to process the results and calculate the mean, standard
deviation and variance. The variance that is obtained is the sum of the
variance of sampling and assaying and of the true variability of the process
stream. The variance associated with assay can be separately determined by
appropriate experiments but one is left with the variance associated with
sampling per se and the variance associated with the true variability of the
process stream.
One of the central problems in sampling is to obtain a measure of the
precision (or variance) of sampling as distinct from the true variability of
the process stream, which is in fact the quantity one wishes to determine.
The other major problem is that of the bias in the mean from the unknown
true value. This problem is addressed later under the question of accuracy.
There are several methods available to determine the precision of
sampling and these can be split broadly into theoretical methods based upon
probability theory, and statistical methods which make no presuppositions
regarding the underlying frequency distributions of mineral particles in the
ore or process stream.
13.4.2 Theoretical calculation of sampling precision using Gy's formulae

The best known of the theoretical methods for calculating the precision of
sampling are those propounded by Gy (1979). Gy sub-divides the errors in-
volved in sampling into seven different classes without distinguishing among
accuracy, precision of measurement, or the natural variability of the material
being sampled. These error classes, as defined by Gy but sub-divided into
precision, natural variability and accuracy are given in Table 13 .1.
Table 13.1. Classification of sampling errors by Gy.
Type Gy notation Origin
Precision Fundamental error, FE Particulate nature of ore

Group and segregation Inhomogeneous mixing of ore
error, GE
Weighting error, WE Uneven flow of ore
Natural variability Long-range quality Natural variability
fluctuation error, QE2
Periodic quality The quantities to be
fluctuation error, QE3 measured
Accuracy Increment delimitation Incorrect cutter design
error, DE
Increment extraction Incorrect cutter speed
error, EE
751
Wih regard to precision, Gy derived a formula to calculate the probability

that particles of mineral, randomly distributed in a host matrix with a
specified top particle size, would be collected in a sample of specific mass.
This probability can be used then to determine the fundamental error as given
in Table 13 .1, for certain ore types. This formula is
M = c.l.f.g.d3 (13.11)
s 2
8
where M, = mass of representative sample for a selected precision, c =

mineralogical composition factor, l = liberation factor, f = shape factor,
g = size range factor, d = nominal or maximum particle diameter in cen-
timetres, and s = standard deviation which is selected.
The mineralogical composition factor is a measure of the mass of gold
in unit volume of host rock and is approximated to by the ratio
c = f../a (13.12)
where f..a is the relative density of gold and a is the gold content expressed
in grams of gold per gram of ore.
The shape factor f is defined by the relationship
v = f.d 3 (13.13)
where Vis the volume of a particle of ore and d is its mean diameter. For
a sphere, f = 0,524, and for a cube, f = 1. Gy recommends usingf = 0,5
in most cases.
g is defined by the equation
(13.14)
where dis the mean particle diameter in centimetres for a lot of mass Min
grams. M; and d; are the mass and mean particle diameters of the different
size fractions within the lot. Gy states that g usually deviates little from an
average value of 0,25.
The liberation factor, !, is defined by the relationship
(13.15)
d 1 is the liberation diameter, which is the maximum particle diameter ensur-

ing complete liberation of the critical component.
As an example of the use of this formula to calculate the minimum sample
size, typical values for the various factors are substituted in the formula.
The liberation size for gold is taken as being 74 µm. Assuming a top particle
size of 120 mm, using the factors given above and choosing a coefficient of
variation of 8%, formula (13.11) gives a sample mass of about 900 tons for
an ore containing 6 git Au. This is a very large mass in relation to typical
production figures of between 1 OOO and 10 OOO tons per shift. This implies
752
PROCESS EVALUATION
that if the underlying assumptions of the Gy theoretical formula are correct,

it is not practically feasible to sample ore with a top size of 120 mm. However,
statistical methods applied to actual results and an alternative method pro-
posed by Gy using the 50-piece experiment, both indicate that far smaller
sample masses are in fact required. These methods are given in later sections.
Although the Gy theoretical formula is not directly applicable to gold,
it is not unreasonable to postulate that sampling variance is indirectly pro-
portional to the cube of the top particle size,
. Cd 3
i.e. s2 = - (13.16)
M,
where C is a constant, not necessarily as calculated from the factors in the
full Gy formula, and sis the precision of sampling. This abbreviated formula
is useful in determining the size to which a sample must be crushed before
sub-division.
Gy also proposes grouping and segregation errors which arise because
of the tendency for the minerals in the ore either to group or to segregate.
There is no way of determining these theoretically.
13.4.3 Gy's 'SO-piece' experiment

Instead of using the theoretical formula for calculation of the sampling preci-
sion, Gy suggests two methods which involve experimental determinations
from which the fundamental error could be calculated. The first of these
methods involves grading a sample into different size fractions and then
dividing the size fractions into different density classes which are analysed.
This method, besides being fairly complex and requiring extensive laboratory
work, is probably not applicable to gold ores.
The second method, the so-called '50-piece' experiment, involves taking
50 or more of the larger lumps of ore selected at random from a belt con-
veying run-of-mine ore. The mass of each specimen is measured, its density
determined and it is analysed. From these measurements, the fundamental
variance, fv, is calculated using the formula
f
v
= _l [
M,
g v ~
a2. Mi = I
(ai- ii)2
Vi
.M2]
l
(13.17)
where g is the size range factor as defined previously.

ai is the assay of the ith particle, n the number of pieces (50 or more) and
Mi is the mass of the ith particle and vi its volume.
a is the weighted mean of the assay results and v is the mean volume of the
particles. M is the total mass of the n particles and M, is the minimum
sample mass required to obtain the specified fundamental variance.
Equation (13 .17) relies upon the assumption that the specimens of the
larger ore particles that are used are related in composition to the whole ore
stream. This is an assumption that may not be valid because richer portions
of ore tend to be more friable and therefore the larger lumps will tend to
be lower in grade. Another disadvantage is that the size range factor g is
753
required in the estimate.

As an example of the use of the 50-piece formula, 50 specimens of the
larger particles of a gold ore were selected by picking from a belt feeding
a run-of-mine mill and these were analysed for gold, their individual masses
were measured and their specific gravities determined. The results are
presented in Table 13.2.
Table 13.2. Results of 50-piece experiment on large lumps of gold ore.
Gold Mass Volume Variance*

Number (git) (g) (ml) function
268,00 3042,9 788,72 762463712

2 2,48 1932,2 702,11 605382
3 1,93 2651,0 964,00 917758
4 0,40 1208,l 438,03 541648
5 4,81 2961,0 1041,14 585734
6 6,57 1886,4 671,56 229423
7 0,63 3277,4 1175,54 1432290
8 3,69 2574,l 935,70 633721
9 4,49 3266,l 1109,79 720868
10 0,40 2789,9 1001,04 1263997
11 0,40 1995,5 730,15 886567
12 0,40 2172,4 781,72 981406
13 3,46 1783,0 642,75 464414
14 1,70 2641,4 972,89 940189
15 1,21 1896,4 667,28 768356
16 0,40 2500,8 901,51 1127732
17 14,70 1964,2 640,85 14464
18 0,40 2691,4 967,09 1217620
19 84,00 3929,l 1154,94 67097618
20 58,70 2024,6 674,87 12601935
21 0,55 2443,9 904,14 1048747
22 0,67 2944,l 1052,22 1283008
23 12,50 3049,7 1084,92 3622
24 0,40 2858,6 1057,57 1256087
25 0,40 1809,7 645,86 824319
26 4,43 2472,3 909,94 510768
27 0,52 2646,8 931,97 1199074
28 9,78 2424,8 83$,45 79641
29 0,40 2636,1 950,63 1188316
30 0,40 1612,0 579,44 729025
31 5,03 1715,2 596, 18 325362
32 0,40 2049,1 745,13 916044
33 17,40 1867 ,5 634,34 99306
34 0,41 1995,3 744,51 867924
35 1,29 4769,6 1726,86 1852998
36 2,91 2504,1 917 ,59 716564
37 0,40 3160,4 1146,32 1416443
38 4,68 3270,8 1155,35 664295
754
PROCESS EVALVATION
Gold Mass Volume Variance*

Number (git) (g) (ml) function
39 0,40 3916,3 1420,49 1755226

40 0,40 3219,9 842,07 1038990
41 0,55 2112,5 770,70 919278
42 2,85 4526,6 1622,44 1339833
43 1,76 2295,5 835,03 818652
44 0,63 2738,3 995,75 1180381
45 4,97 1925,2 633,08 391741
46 0,77 3832,9 1354,38 1662475
47 31,90 2215,4 774,34 2228294
48 0,40 2463,9 899,23 1097473
49 16,50 1413,3 506,92 44220
50 1,54 1039,7 381,26 382171
Totals 126217 883335113

Mean 13,15 892,38
(Gold - mean gold/ x mass2

*Variance function =
volume
From these, substituting in Equation (13 .17),

1 0,25 x 892,88 x 883335113
fv =
M, 13,15 x 13,15 x 126217
9034,19
M,
If it is now required to measure the grade of ore with a coefficient of varia-
tion of 8%, which represents a value of (0,08) 2 for fv, the minimum sam-
ple mass required for a representative sample is given by
M = 9034,19
5
fv
9034,19
(0,08)2 = 1411592 gl= 1,41 tons.
This mass of sample relates to a particular top particle size. The mean volume
v is related to the particle size thus:
v = fd 3
where f is the shape factor and d the top particle size in centimetres. For
this example, taking f = 0,5, i.e. for approximately spherical particles,
892,37 = 0,5 .d 3
d = 12,1 cm.
An estimate can be made of the size of sample required for differing top
755
particle sizes by using Equation (13.16).

As an example, if d were 6 cm instead of 12, 1 cm, the minimum sample
mass m would be
m ~ ( ;,l )'. 1,41

1
~ 0,171 tons.
Examination of Table 13 .2 shows that the estimates of minimum sample mass

are sensitive to high observations. If the first observation, where a gold value
of 268 git was measured, were to be deleted, the minimum sample mass for
12,1 cm particle size becomes 0,665 ton.
Occasional high values are a feature of individual particles of ore because
of the distribution of the gold within the ore body. However, if an incre-
ment consisting of a large number of smaller particles is taken, as is the case
when an increment is taken from a belt, such high values would not be
common due to the averaging effect of the large number of particles.
Despite its limitations, the Gy 50-piece experiment is useful in that it
gives a sample size of the same order of magnitude as is measured from
statistical analysis for samples taken from belts. Gy's 50-piece experiment
indicates the minimum mass of composite sample that is likely to be represen-
tative. The actual precision that is attained can, however, be calculated directly
using the geostatistical methods described in the next section.
13.4.4 Precision of sampling using geostatistical methods

In classical statistics of independent samples there is no reason to suppose
that the difference in the results from two samples or increments taken adja-
cent to one another would be less than the difference between two widely
spaced samples or increments. It is, however, reasonable to suppose that in
spatial data the difference in results from two closely spaced samples would
be less than that for samples which are far apart.
The correlation between adjacent samples has been observed and ap-
plied to underground evaluation by using what are termed 'semi-variograms'
(Rendu, 1978). In underground evaluation the situation is more complex in
that one is concerned with three dimensions. For plant evaluation there is
on:ly one dimension, namely, the horizontal distance between sample in-
crements which, on a belt, translates to the time between increments.
13.4.4.1 The semi-variogram

The semi-variogram 'YII is defined as one half the variance of the difference
between sample grades at different points. To understand this, consider the
distribution in grade of ore on a conveyor belt from which a sampling incre-
ment is taken at fixed time intervals. The grade at these intervals is given
by Y1, Y2, Y3, Y4 · · · Yn·
The semi-variogram for a spacing of 1 is
'Y1 =
1 n--=----1
2. 1 [(yl - yi) 2 + CY2 - Y3) 2 ... (yn-1 - Yn) 2] (13.18)
for a spacing of 2 it is
756
PROCESS EVALUATION
Sill
C+N ---------------------'.-"'--~--------___,
a
~
0
-~ 'Y(h)
>
·§
<l)
(/)
N
I. Sample,\ spacing (time)

Range (a)
Figure 13.3. Typical semi-variogram.
'Y2 = -
1 1 [
· - - (y1 - Y3)
2
+ 2
(y2 - Y4) · · · (yn-2 - YJ
2]
2 n - 2
etc. for spacings of 3, 4, 5, etc. The semi-variogram is calculated for all
possible pairings of the grades and the results plotted on a graph of Y1i
versus spacing h. A typical plot of a variogram is given in Figure 13.3.
13.4.4.2 Nugget effect

Referring to Figure 13.3, C + N is the sill, a the range and N the nugget
effect, which is the intercept on the y-axis of '}'(h) at very close intervals. The
nugget effect is the variance between samples and the immediately adjacent
true grade. This variance is the sum of the variance introduced by assaying
errors, by errors involved in the sub-sampling procedures and in the variance
that comes about because of the inhomogeneity of the distribution of mineral
particles within the _host rock.
13.4.4.3 Estimates of variance from semi-variograms

If a number of increments (n) are contributed to a sample taken at close in-
tervals, the variance of the mean value of the increments has a minimum
value of Nin. On the other hand, if the interval between the increments is
greater than the range a, the variance for the mean will be given by (C + N)ln,
that is
(13.19)
where af,, is the variance of the mean of n increments and is called the
estimation variance. It is sometimes also called the extension variance.
For many purposes, knowing the estimation variance between these limits
is sufficient but, if it is not, a mathematical model needs to be fitted to the
757
semi-variogram and the estimation variance calculated for different sample

spacings.
The techniques for calculation of the estimation variance are given by
Rendu (1978). The first step is to fit a mathematical model to the semi-
variogram established experimentally. There are numerous mathematical
models available but for plant sampling, where one is normally considering
relatively short time scales of a shift at a time, the two types of semi-
variograms that are of interest are the pure nugget effect where there is no
spatial correlation and the spherical semi-variogram as shown in Figure 13 .3.
a) Pure nugget effect
'Yh = N (13.20)
where N is a constant.
To calculate the estimation variance when there is a pure nugget effect, divide
the nugget effect by the number of increments n.
b) Spherical model
This is characterized by three parameters C, Nanda, where C + N is
the sill, a the range, and N the nugget effect.
3
'Yh = C -3h - 3h ] + N for h ~ a (13.2i)
[ 2a a
'Yh = C + N for h > a.
The nugget effect N is an additive constant and is the intercept on the y-axis
at the origin.
To calculate the estimation variance for a spherical semi-variogram, one
can use the following formula, which gives an approximation for a;,.
(J~ = _!_f N
+ C + N. !_ + 3(C + N) .. !____] (13.22)
n( 4 a 160 a3
where t is the time between increments and a is the range in units of time.
More exact methods for calculation of ai
are given in Rendu (1978).
There are other models that could be applied. In particular the
logarithmic, or de Wijsian, model has been found to be useful for
underground valuations where log-normal population distributions are found.
However, in plant sampling such a model is not called for because of the
mixing that occurs before ore reaches the surface. Examples are given of the
calculation of estimation variances when practical sampling problems are
discussed (see Section 13. 5).
13.4.5 Accuracy of sampling

Irf contrast to precision, which as shown above can be measured by various
statistical techniques, accuracy can be assessed only by
a) Comparing the result of measurement to a known value. Unfortunately,
in sampling it is not usually possible to obtain a known value for a lot
of ore or residue.
758
PROCESS EVALUATION
b) Comparing the results obtained from one sampling device to those from
another which is known to be unbiased.
c) Assessing the results of a sampling programme in terms of the mass balance
across a plant or process. From mass balance considerations, it is known
that ore fed to a plant will report ultimately as either product or residue.
Therefore, if the gold content of ore fed to a plant matches the gold con-
tent of the products and residue, there is a high probability that all the
sampling, mass measurement and assaying procedures are accurate. If,
however, there is an imbalance between input and output, this need not
necessarily be because of inaccuracies in the measurements but could be
due to unknown losses, theft, or long time lags between feeding material
into a plant and residue and products being produced.
d) Examining the sampling scheme for potential sources of bias. If there are
no sources of bias, then the sample will be presumed accurate.
The fundamental requirement for a bias-free sample is that every particle
within the process stream must have an equal probability of being selected
for the final sample for analysis. This requirement precludes the use of any
type of sampler which takes an increment only from part of the process
stream. An unbiased sampler must take an increment from the whole pro-
cess stream.
For there to be no bias in the sample submitted for analysis, there must
be no bias in the taking of the primary increments which are used to make
up the final sample, and there must be no preferential loss of fine particles
as dust, or upgrading of heavy mineral particles in the sample preparation
stages.
In most plant situations, ore is usually transported at some stage on a
conveyor belt or as a slurry in a pipeline. Under these circumstances, all the
ore passes a single point at which an increment could be taken'. If this is done,
each particle within the ore has the same probability of being selected for
the final sample provided no bias is introduced in taking the primary incre-
ment. This does not apply to the extraction of increments from a hopper
and therefore, to ensure accuracy, sampling from a conveyor either as a stop-
belt sample or as a falling stream is recommended in preference to hopper
sampling.
After crushing and milling, ore and residues are in the form of pulps
and suitable sampling devices to obtain increments across the whole of the
process streams are commonly used. In sampling of process streams, bias
can arise from taking proportionately more from one part of the stream than
from another, the delimitation error as defined by Gy, or by taking more
from one size fraction than from another, the increment extraction error.
The listing in Table 13 .3 identifies these sources of bias and gives the measures
that need to be taken to eliminate them.
13.5 Sampling Practice

In the practical application of sampling for process evaluation, two funda-
mental principles should be observed:
759
Table 13.3. Sources of bias in sampling (after Burton et al., 1982).
Sonrce of bias: A
Non-uniformity of sampling ratio of different parts of the stream cross section.
Canse and solution: A
Cause
Inadequate cutter design.
Solution
a) Straight path samplers must be designed so that the speed of cutting is the same for all points
of the cutter. Thus the width of the cutter should be constant and parallel.
b) Arc-path samplers follow a circular path, therefore the speed at any point is proportional
to the cutter radius. Thus to maintain constant sampling ratio, the width of the cutter must
be proportional to the radius. This is achieved by .having the edges radial. Rectangular openings
introduce bias.
c) Hand sampling. It is never possible to be sure that the sampling ratio is uniform for all parts
of the stream for hand sampling a flowing stream. It is, however, possible to sample a stop-
ped belt using a former to delineate a sample increment. The former that is used must have
parallel sides and must be shaped to fit the belt to prevent material outside the former leak-
ing in.
Cause
Non-uniformity of cutter speed.
Solution
a) Hydraulic and pneumatic drives can never ensure a constant speed and thus should be avoided.
b) Electric drives should be powerful enough to keep the speed up under load, thus they should
be oversized.
c) Hand drive cannot ensure a constant speed and thus is not recommended.
Cause
Incomplete cutting of stream during cut.
Solution
a) Cutter path covering only part of stream section. The length of the cutter should be at least
twice that of the stream thickness.
b) Dust from batch escaping sampling can be avoided by atomizing water at the end of the
belt, by avoiding air currents around the end of the belt and by enclosing the end of the
belt in an air-tight box.
c) Moist ore adhering to the belt will escape the parabolic stream cut by the cutter. To avoid
this a scraper must be positioned such that the scraped material falls inside the surface covered
by the cutter travel.
d) Defective design of collecting chute. Some cutters deliver the increments into a chute; this
chute must collect the entire increment.
Sonrce of bias: B
Collection of alien material between cuts.
Cause and solntion: B
a) Cutter remaining in stream. The stopped position of the cutter should be far outside the
stream. A good rule is to design a cutter_ path twice the width of the stream.
b) Cutter receiving alien dust at standstill can be avoided by using a protective cap.
Source of bias: C
Non-uniformity of sampling ratio of different classes of particles.
Canse and solution: C
a) Distance between cutter edges. It has been recognized that a minimum ratio should be pro-
vided between the smallest distance between the cutter edges and the diameter, d, of the largest
particle to be collected. The table below is in good agreement with experience.
760
PROCESS EVALUATION
Sampled material Minimum width
d greater than 3 mm 3d
d less than 3 mm 10 mm
b) Inadequate cutter speed. When the cutter speed is too high (greater than 600 mm/s), the
effect is the same as when the cutter is too narrow.
c) Partial dedusting of increment in cutter. Scoop-type cutters can be crossed by an upward
air draught that has a dedusting effect on dry ore. Air draughts should be controlled.
Source of bias: D
a) Partial loss of increment due to bouncing, splashing out, or overflowing.
Cause and solution: D
a) Open-scoop cutters.
Inadequate depth. The depth depends on the size of the largest particles and on the ton-
nage sampled. A minimum depth of 3d is usually safe, but for tonnages of 1 OOO t/h or more
5d is safer.
Inadequate slope: Dry ores 45° slope minimum. Wet ores, 60° slope minimum.
Bad surfacing is caused by using ordinary steel and there being rust, rivets or welded joints,
etc. on the surface. Stainless steel is preferred.
Sharp angles should be avoided; the sides and bottom of the chute should be in one piece
and the junction between them radiused, not sharp.
Sticky ores such as clay or filter cake should have a cutter designed so that the width is
5d, the depth is 5d, the slope 65°, the surface mac;le of stainless steel and curved in one piece.
The cutter could also be heated to reduce the surface tension of the water and so prevent
sticking.
b) Bucket type cutters.
Inadequate capacity. The volume of the bucket cutter should be twice that of the incre-
ment. Sticky ore should not be collected using a bucket cutter owing to the difficulty in
discharging the increments.
c) Pulp cutters.
Pulp cutters are always of the scoop type. The minimum depth at the high end should
be 75 - 100 mm, and the minimum slope should be 30°.
a) The accuracy of the sampling must be guaranteed by proper design of

the sampling equipment so that there are no sources of bias.
b) Once accuracy has been guaranteed, precision should be measured using
the procedures already given.
To illustrate the application of these two fundamental principles,
examples are given for the sampling of the various feeds, residues and pro-
ducts that are encountered in the gold mining industry.
13.5.1 Run-of-mine ore feed - stop-belt sampling

It has been stated that, in general, manual sampling is not acceptable because
one cannot guarantee its accuracy. The exception to this is the use of stop-
belt sampling where a feed belt is stopped, a former placed on it and all the
material within the former carefully removed into a sample bucket before
restarting the belt. This method is recognized by certain national and inter-
national standards as the reference sampling method when checking for bias
in automatic samplers (Gy, 1979).
761
Because of the capital costs and mechanical difficulties involved in

automatic sampling of run-of-mine ore, manual stop-belt sampling is used
on a routine basis on some Witwatersrand gold mines. An example of the
practice on a particular mine, where run-of-mine ore with about 5% by mass
greater than 150 mm is sampled, is given below.
A belt conveying the ore to a crushing plant is stopped every 30 or 60
minutes and a former 500 mm long with parallel ends is placed on the belt
at a fixed position. All the material within the former is placed in a con-
tainer, with care being taken to brush all the fines adhering to the belt into
the container. The individual increments are separately dried and then passed
through a 300 x 200 mm jaw crusher after first breaking any very large par-
ticles. The primary increment typically has a mass between 20 and 30 kg.
After being sub-divided twice with a Jones divider, the sample is passed
through a cone crusher to give a product with a particle size less than 6 mm.
This is further divided using a spinning riffler (Section 13.6.1.2) to give eight
sub"samples of 1 to 2 kg in mass for the analytical laboratory.
13.5.1.1 Calculation of precision of manual stop-belt sampling

The precision of sampling using stop-belt sampling was calculated using tlie
semi-variogram procedure described earlier. For purposes of illustration, the
methods used to calculate the precision of stop-belt sampling are dealt with
in some detail but the same techniques of obtaining a semi-variogram and
calculating the estimation variance for a particular sample spacing apply to
all sampling problems.
To obtain the initial semi-variogram, increments were taken by stop-
ping the feed belt at regular intervals, placing a 500 mm long former on the
belt and then sweeping all the material within the former into a bucket. The
increments were dried, crushed, sub-sampled and assayed for gold. The results
of the assays were plotted on a histogram presented as Figure 13 .2 and the
semi-variogram is presented as Figure 13.4.
In order to quantify the nugget effect, 30 increments were taken im-
mediately adjacent to one another and a separate semi-variogram plotted
for these closely spaced increments. This semi-variogram for closely spaced
samples showed no spatial correlation and the points for this have been plotted
on the y-axis in Figure 13.4.
A spherical model was fitted with the following parameters:
'Y11 = 6,25 + 7 -3h - -1 . 3h

[ 2a 2 a
3
J
for h < a and 'Y1i = 13,25 for h > a.
The nugget effect N is 6,25 (glt)2. The sill, C + N, is 13,25 (g/t) 2 • The
range a is taken as 5 hours.
Using the formula
a~ = nl [N + C+4 N. !__a + 3(C160+ N) !___]

a 3
762
PROCESS EVALUATION
13
.. . .. . ..
12
:2
11
..
?=:' 10
§Oil
9
0
-~ 8
.2\
~
[/)
7
4~--~---~---~--~---~
0 8 16 24 32 40
Sample spacing (hours)
Figure 13.4. Semi-variogram for samples at half-hourly inter-
vals on a crusher feed belt.
where n is the number of samples, t is the time between samples and a is

the range,
a2
E
= _1 6 25
16 ( '
r + 13,25 . Q2_ + 3.13,25 . 0,5
4 5 160 3
5
3
]
= 0,411 (glt) 2
for 16 samples taken at half-hourly intervals during an 8-hour shift.
The mean gold value for all the increments was 7 ,54 git. For the pur-
poses of illustrating the effect of different sample spacings, it is assumed
that this grade applies to a particular shift. On this basis the coefficient of
variation for a half-hourly increment is 8,5%.
In Table 13 .4, the coefficients of variation and the upper and lower 95%
confidence limits for different sampling intervals are presented. This table
demonstrates that the precision for sampling on a shift basis is poor. However,
on a monthly basis, the precision is improved because of the number of shifts
involved. As an example, if stop-belt sampling every half-hour is employed
over a 90-shift month, the coefficient of variation (CV) would be
CV (per month) = 8,51-./90 %
= 0,89%.
Accordingly, the head grade for the month is 7 ,54 ( ± 0,13) git, the numbers
in the brackets giving the 95% upper and lower confidence limits.
763
from the primary increment.
13.5.2 Automatic sampling of run-of-mine ore

The sampling and analysis of separate increments taken by stopping the belt
at regular intervals is a cumbersome procedure and requires stringent super-
vision. In some situations it is not practicable to stop the belt because of
the disruptive ~ffect this would have on plant operation. Therefore, it is often
advisable to automate the procedures by taking a primary increment from
the discharge of a head pulley, and passing this through one or more stages
of crushing and sub-division to give a sample for analysis. There are, however,
considerable mechanical problems associated with automatic sampling of run-
of-mine ores and the technology for this has not yet been fully developed.
Figure 13 .5 is a diagrammatic representation of a run-of-mine sampler
installed as a prototype. A heavy duty sampler with an opening of 450 mm
takes an increment of the run-of-mine ore with a top particle size of 200 mm.
An increment is taken every 5 minutes and falls through a 500 x 275 jaw
crusher where the sample is reduced to 100% minus 38 mm.
The product is fed to a secondary Vezin sampler to give a sub-sample
2,5% by mass of primary increment. This sub-sample is passed through a
200 x 125 jaw crusher to reduce its particle size to 100% passing 12 mm.
It is then further subdivided using the tertiary Vezin to give a final sample
for the laboratory and reject which is returned to the plant downstream of
the sampling point.
13.5.2.1 Testing for bias in automatic sampling
The absence of bias in the automatic sampling can be established by con-
ducting a campaign whereby the results from the automatic sampler are com-
pared to the results for stop-belt .samples taken from the belt feeding the
automatic sampler. This test can take two forms, namely:
a) A paired t-test between shift results for the automatic and manual samples.
This involves a comparison, preferably over at least 30 shifts, of the mean
result from the automatic sampler with the mean over every shift of
manual stop-belt samples taken at fixed time intervals. For this comparison
the differences between the two means for every shift are analysed
statistically to determine whether there is a significant difference between
the two means.
For this significance test, a so-called null hypothesis is proposed that
there is no difference between the two means and the probability that this
hypothesis can be rejected is determined from calculated t statistics and
tables that give probabilities for Student's t distribution (Box et al., 1978).
-b) A paired t-test using individual increments rather than averaged shift
results. For thiflest, the belt feeding the sampler is stopped and an incre-
ment taken for analysis. The belt is restarted and an increment taken with
the automatic sampler. This increment is also analysed and the difference
between the results for the manual and automatic samplers obtained for
about 50 pairs of increments. For this test, the size of the former used
to take the manual stop-belt sample should be of a size calculated to give
766
PROCESS EVALUATION
Sample
Figure 13.5. Run-of-mine automatic sampling plant.
the same mass of sample as is taken by the primary cutter of the automatic
sampler.
The t- test can also be used to test whether there has been a significant
change in a mean value before and after a process change.
Detailed descriptions oft-tests and null hypotheses are available in stan-
dard statistical texts but, as an illustration of their use for testing samples,
the average shift values of samples cut with a manual belt taken at half-hourly
intervals are compared in Table 13 .5 to a composite shift sample taken with
an automatic cross-stream cutter.
Visual examination of the automatic sampler had indicated that there
were two potential problems, namely, fine material from the belt falling con-
tinuously into the cutter and, secondly, large particles of ore missing the cut-
ter. Both of these would tend to bias the automatic sampler high, relative
to the manual stop-belt sample, which is taken as the reference, because fines
tend to be higher in grade then coarse material. A dribble tray was fitted
to the automatic sampler to prevent the extraneous fine material falling into
the sample after shift 45.
It was required to establish whether:
767
Table 13.5 Exampled of paired and 2-sample t-test for comparison of an automatic and manual
sampler.
Manual (x 1) Automatic (x2)

git git Difference
Shift gold gold (x 1 - X 2)
Before trays installed
5,59 5,32 0,22

2 3,30 5,45 -2,15
3 3,68 4,66 -0,98
44 4,51 4,53 -0,Q2
Mean 3,81 4,72 -0,91

Std dev. 0,88 0,95 1,19
After trays installed
45 2,86 3,89 -1,03

46 4,06 4,83 -0,77
47 3,41 3,08 0,33
90 3,68 3,20 0,48
Mean 3,67 4,01 -0,34

Std dev. 1,00 1,14 1,16
(i) there was a significant different between the automatic and manual
samplers,
(ii) the difference changed after the dribble trays were installed, and
(iii) did the cutter need to be changed to catch the larger particles?
To answer these questions, paired-t tests were applied.
13.5.2.2 Examples oft-tests

The null hypothesis H 0 , that there was no difference between the paired
results x 1 and x2 for the two samples, was proposed, so that
t = X1 - X2 • .Jn (13.23)
s
where n = number of pairs and s = standard deviation of the difference
x 1 - x2 • From Table 13.5 for the results for the first 44 pairs,
t = -0, 9 l. v'44 = -5,07.
1,19
768
PROCESS EVALUATION
From tables of t values for (n - 1) = 43 degrees of freedom, a value of t

greater than about 3,55, ignoring the minus sign, means that there is a less
than 0,01 OJo probability that the difference between the two values is zero.
Therefore, the null hypothesis can be rejected at the 99,9% confidence level
and there was a statistically significant difference between the manual and
automatic samplers before the dribble trays were installed.
The numerical value of the bias is - 0,91 with a standard deviation of
1,19 git. For 43 degrees of freedom (44 observations) the value oft, from
tables, is approximately 2,02 and therefore the confidence interval for the
bias, at the 95 OJo level, is given by
t = +2 02 l,l 9 = 0 36
- ' . ,J44 ± ' .
There is therefore a 95% probability that the bias between the two samplers
lies between - 0,55 git and - 1,25 git for all the observations.
This bias changes, however, when the dribble trays are installed. The
value of the bias can be established after the installation of the trays by
applying a t-test similar to the example already given.
In order, however, to establish whether the change in bias is significant
statistically, an alternative form of the t-test is applied. For this, a null
hypothesis is postulated that there is no difference between the differences
between the two samplers before and after the installation of the trays. A
t-test using the formula below is used:
f = (i\ - X2)Before - (.Xl - X2) After
(13.24)
(~ + ~ )'.s, .
For this t-test a pooled estimate is made of the standard deviation
s~ (n - 1) + ~ (m - 1)
s21 = ~~~~~----''-=---~~~
n+m-2
where s8 = standard deviation before, sA = standard deviation after, n =
number of observations before, and m = number of observations after. x1
and x2 are the mean values for the separate increments.
Therefore
2
s2 = (1,19) x 43 + 1,16 x 45
1,29
I 88
.". S 1 = 1,13,
and
t = (-0,91) - (-0,34)
1
44 + 46 . 1,13
-2,39.
769
()•
. r--,::_, gooq;
r. -.. .------ 0
>-t[J
I
l_j __ "o~
Bucket sampler
,
00
ocf) Roll ccmhcc
~ !;
Sample
Figure 13.6. Automatic sampling plant for crusher product.
From tables, for 88 degrees of freedom, this value for t shows that the null
hypothesis can be rejected with more than 95% confidence.
13.5.3 Crusher product sampling

Figure 13 .6 is a diagrammatic representation of a sampler for a crusher prod-
uct with a nominal top size of 13 mm. For this sampler, the primary in-
crements are collected in a bucket with a gap of 40 mm travelling at a speed
of 0,703 m/s. The cutter speed is slightly higher than the 0,600 m/s recom-
mended by Gy. The cutter is set to take an increment every 8 minutes and
this is discharged into a set of continuously operating rolls set to give a prod-
uct 100% minus 12 mm.
A spinning disc sampler operating at 75 rev/min and with a radius of
140 mm at the sampling point takes a secondary sample consisting of about
14% of the primary increment. The final sample falls into a bucket and the
reject joins the process stream. This particular sampler operates on a
gold- antimony ore with a very irregular gold distribution. Therefore, despite
the smaller particle size of crusher product as compared to the run-of-mine
ore in the previous example, the precision of sampling is only ± 30% per shift.
For a month of 90 shifts, the precision of the monthly mean will be 30/ v90
= 3,16%. This shift precision was established by tests in which the shift assays
for a period of several months were analysed using the variogram as defined
earlier.
770
PROCESS EVALUATION
Splash
curtain
Stilling
box Flexible hose
/
Drain
To process
Double-headed
rotary sampler
Reject to process
Figure 13 .7. Automatic cross-stream sampler for pulps.
13.5.4 Sampling of pulps

After milling, the ore is in the form of a pulp as a cyclone overflow, thickener
underflow, feed to leach or a residue. For the sampling of these materials,
manual sampling, although commonly employed, is not satisfactory because:
a) there is no guarantee that the samples will be unbiased, and
b) the precision of the measurements is generally much lower than can be
obtained with properly designed mechanical systems.
The mechanicai systems are usually of two types, namely, those with cross-
stream cutters and the arc-type rotary cutters.
13.5.4.1 Cross-stream cutters

Figure 13. 7 is a diagram of a typical cross-stream installation. It consists of
a stilling box into which the pulp is discharged. The purpose of this is to
771
~
- ~e~h~~'ism
Cutter blades
(hardened steel
!Omm
minimum gap)
780 mm
!Omm
~
''j''
11
!I 11
11
11 11 - Sample
11 1!
Figure 13.8. Sample cutter for pulps.
smooth out the flow as the pulp rises. The dimensions of the box are
calculated to give rising velocities high enough to prevent the' settling of
mineral particles but low enough to give a non-turbulent flow. Typical rising
velocities would be of the order of between 2 and 10 metres per second.
The dimension D between the bottom of the stilling box and the bot-
tom of the discharge pipe is important. If it is too small, the end of the pipe
becomes blocked when the plant is shut down and the solids settle to the
bottom of the box. From the stilling box, the pulp overflows down the launder
and into the discharge box. The height of the crest as the pulp flows into
the launder can be calculated from the modified Francis formula (Perry and
Chilton, 1973). This height will decrease as the pulp accelerates down the
launder and can be calculated. The sample cutter is positioned to intercept
the pulp stream at right angles to the flow. It is vital that the cutter is large
enough to take the full flow without pulp being ejected back out of it. Figure
13.8 is a sketch of a pulp cutter that has been found to be of adequate
proportions.
The cutter is parked well out of the way of the process stream and once
every five minutes, or at some other interval as determined by experiments
on precision, is activateel and moves across, taking an increment. The incre-
ment flows down the flexible hose into the secondary sampler which, in this
particular example takes duplicate sm;nples for separate analysis. In order
to ensure that the cutter is kept clean, water-sprays, activated by the same
timing mechanism as the sample cutter, are situated to clean the cutters in
the parked position.
There are numerous variations of this basic design. For instance, in some
772
PROCESS EVALUATION
Pulp
i Drive
Cutter
Reject
\Sampl1
Figure 13.9. Rotary sampler.
cases a sample collection launder is used instead of the flexible hose or two
single cut rotary samplers one above the other are used to obtain duplicate
s.amples. The mass of each single increment, I, in kg, is given by the formula:
I= D.FIS (13.25)
where D = cutter gap in mm, F = pulp flow rate in kg/ s, and S = cutter
speed in mm/ s.
The primary increment is sub-divided as it passes through the secondary rotary
sampler. Figure 13 .9 is an expanded view of a rotary sampler. These are usual-
ly manufactured to give a 2,5%, 5% or 10% cut by having l, 2 or 4 arms
on the rotating hollow shaft for collecting the sample.
13.5.4.2 Rotary samplers

An alternative to a cross-stream cutter, particularly for smaller process flows,
is a rotary sample cutter similar to those used for sub-sampling (Figure 13 .9).
Where the sampler is used for primary increments, there would only be one
sample cutter mounted on the central shaft. The sampler is activated to make
a single revolution at specified time intervals to give a primary increment
which could be sub-sampled with the same type of equipment operating
continuously.
Similar rules to those for cross-stream cutters apply also to rotary cutters,
i.e. the minimum cutter gap should be 10 mm and the tangential velocity at
the point at which the sample is being cut should not exceed 600 mm per
second.
The composite sample for a shift is usually collected in a bucket. General-
ly, for ease of handling, the volume of the shift composite sample is less than
10 litres. This pulp is filtered using a pressure filter, the filter cake is dried
in an oven or under infrared heaters, and the dried cake is sent to the
773
laboratory for analysis.

At one mine, the pressure filtration step has been automated. Here, the
increment from the secondary sampler feeds into the pressure filter via a pipe
fitted with a solenoid valve. Once the primary cutter has reached its parking
position and has been washed with automatic water sprays, this solenoid valve
closes and another opens to introduce compressed air into the pressure vessel.
At the end of the filtration cycle, the air pressure is released, and the valve
to the sampler opened ready for the next cycle. At the end of the shift the
operator removes the filter cake and replaces the filter paper for the next shift.
13.5.4.3 Precision of sampling with cross-stream samplers

The accuracy of sampling is guaranteed by adherence to the criteria given
above but the precision depends upon the number and size of increments
taken per shift, the precision of the analytical measurements and upon the
underlying variability of the process.
The size of increments is governed by the speed of the cutters and the
cutter gaps while the precision of the analytical procedures and underlying
variability of the process are usually fixed. This leaves the major controls
for precision to the number of increments per shift and to the number of
replicate analyses for samples that are produced.
The overall precision per shift, incorporating the variance due to assay
and that due to sampling, can be calculated using the semi-variographic
technique already mentioned. Alternatively, particularly where it is required
to obtain separate estimates of the assaying and sampling variance, an analysis
of variance technique as described by Box et al. (1978) and Merks (1985)
can be used.
It is outside the scope of this chapter to deal with analysis of variance.
However, in order to illustrate the type of information that such a technique
can produce, an t:;xample is given of analysis of variance of samples from
the tailings from a carbon-in-pulp plant treating Witwatersrand gold ore.
For this, 29 separate increments were taken from a tailings stream us'ing a
cross-stream sampler and each analysed four times. The assays were fed to
a computer program, PROC OLM, in the SAS system statistics (SAS Users
Guide, 1985), which performs an analysis of variance after making allowance
for any missing assay data. In this particular example only 106 observations
could be used in the analysis as there were 10 missing values. A summary
of the analysis of variance is presented in Table 13.6.
For this table the variances attributable to assay and sampling have been
separately determined and expressed as coefficients of variation (CV).
In this particular example, the CV for sampling is 34,8% and that for
assay was 11,5%. From these values, it is possible to calculate the precision
that is attainable with different sampling strategies. For instance, the variance,
vt' that could be expected from a composite sample consisting of 16 half-
hourly increments taken during a shift and analysed 4 times by fire assay,
would be given by
v, = v, + v.
774
PROCESS EVALUATION
Table 13 .6. Summary table for analysis of variance on CIP tailings values using the SAS system
OLM procedure.
Dependant variable: Assay grade of tailings, mean .value = 0,232 git Au.
Source of Degrees of Sum of Mean Standard

variation freedom squares square Variance* deviation CV(%)
Model 31 0,18935031 0,00610807

Error 74 0,07395074 0,00099933
Corrected
total 105 0,26330105
Sampling 28 0,18720475 0,0066858 0,00650701 0,0807 34,8
Assay 3 0,00214556 0,00071517 0,0007151-7 0,0267 11,5
Note:
*Variance for assay = mean square for assay.
Variance for sampling = mean square for sampling - (mean square assay/4).
where v, = sampling variance, v. = analytical variance, and

(34,8) 2 (11,5) 2
vt _1_6_ + --4-
108,75.
Therefore the variance in this situation is 108,75, which gives a coefficient
of variation of 10,4%.
Table 13. 7 gives a summary of the calculated CV' s for different numbers
of increments and assays. Tables of this type would be used to determine
optimum sampling strategies in terms of the costs of analyses, required preci-
sions and practical constraints such as the size of sample that can be handled
conveniently.
Table 13.7. Coefficient of variation (CV) for combined sampling and assay precisions for CIP
tailings.
Increments per shift CV(%) CV(%) CV(%)

(number) Single assay 2 Assays 4 Assays
4 20,86 19,21 18,32

8 16,84 14,75 13,58
16 14,42 11,91 10,43
32 13,04 10,20 8,42
96 12,04 8,87 6,76
13.5.5 Other sampling equipment for pulps

It is sometimes expensive to install cross-stream cutters in a pulp flow, par-
ticularly in older plants, as the flow has to be broken to include the cross-
stream or rotary arc cutter. Also, another stage of pumping may be necessary
to convey the pulp after the sample has been taken. For these reasons other
775
Process
flow
Rotary sampler
Reject
[ Y a m pie
Figure 13.10. Pipe sampler for pulps.
sampling devices that do not necessarily conform to the requirement of prob-

abilistic sampling may have to be used. Probabilistic sampling means that
every particle within the process stream has an equal chance of ending up
in the sample. For non-probabilistic sampling, not every particle has the same
chance of being in the sample but samples are taken from well-mixed streams
in order that the samples are as representative as possible.
If the mixing is good enough, the process stream will be homogeneous
and therefore it will not matter if there is a greater tendency to collect a sample
from one part or another of the stream. Two examples of such non-
probabilistic sampling are given below.
13.5,5.1 Pipe sampler

Figure 13.10 is a diagram of a pipe sampler which consists, essentially, of
50 mm diameter pipe introduced into a main process pulp stream at a bend
where there would be considerable turbulence to assist in obtaining
homogeneity. This is further assisted by increasing the diameter of the line
in this case from 300 mm to 350 mm. This change in diameter changes the
area, and hence the flow velocity, by 36%, which would give considerable
turbulence. The turbulence could be further increased by introducing rods
at right angles to the flow, but these tend to erode away leaving only the
holding nuts on the outside and deluding one into thinking that the mixing
rods are still there.
The diameter of 50 mm for the sample pipe is chosen as anything smaller
776
PROCESS EVALUATION
Ram
Column carrying
process stream
"" Seals
/~ ,\:I--\---'
Sampler
Sampler retracted
extended
Sample
Figure 13.11. Injector sampler.
tends to block. A smaller sample for the assay laboratory is obtained with
a Vezin secondary sampler as indicated on the diagram.
13.5.5.2 Injector sampler

Figure 13.11 is a diagram of an injector sampler. When an increment is re-
quired, the pneumatic, or hydraulic, ram is actuated, pushing the sampling
tube into the process stream. The sample enters the sampler tube through
the orifice shown, flows down the tube and passes via the exit branch and
flexible hose into the sample bucket. After the sample has been taken the
tube is returned to its housing. The flexible disc at tbe end of the sample
tube seals against the inside of the process column preventing pulp from
escaping when the sampler is inactive. The injector sampler is usually installed
on a rising main.
Sampling with the injector sampler is non-probabilistic because a) the
hole does not traverse the complete process stream, and b) the sample mass
is not proportional to the process flow rate.
For these reasons the injector sampler cannot be guaranteed to be
unbiased.
13.5.6 Sampling of loaded carbon, slag, flue dust and other secondary
materials at Rand Refinery
This type of material is sampled at the Rand Refinery, whose analyses are
used as the basis for payment to clients.
At the refinery, great care is taken to homogenize the materials before
taking "rod" samples using augers to take samples from each drum of
material. This sampling is non-probabilistic but the careful mixing ensures
that samples are representative. The reader is referred to Chapter 10 for the
details.
777
13.5.7 Sampling of bullion

Gold bars are sampled at the various mines by drilling on a fixed pattern
on each bar. The procedures followed by Rand Refinery for gold bullion
sampling are given in Chapter 10.
13.5.8 Sampling of liquids

The sampling of clear solutions, for example, of pregnant gold feed solu-
tions to zinc precipitation or of barren solutions after the Stellar filters, is
often done with drip samplers. These consist of a small valve teed into the
main process flow line. The valve is opened to allow a steady drip of solu-
tion into a bottle.
At some mines, instead of a steadily dripping valve, a solenoid valve
is opened on a fixed time sequence to obtain the sample. In both cases, the
representativeness of the sample depends upon adequate mixing of the solu-
tions and constant flow rate.
13.6 Sample Preparation

For a single assay of gold in a sample, a maximum of about 150 grams of
finely divided, well-mixed, dried sample is required from the original samples
that are taken on the plants. In many cases, particularly with run-of-mine
sampling, the samples that are taken on the plant are very much larger than
required and consist of a range of particle sizes, including very large par-
ticles. It is necessary, therefore, to reduce the mass and particle size of many
samples withollt introducing bias.
There are two aspects of sample preparation that are dealt with, namely
1) The equipment and operating practice employed in sample preparation,
and
2) statistical methods used to quantify the variance and bias introduced by
sub-sampling.
13.6.1 Sample preparation equipment and practice

13.6.1.1 Pulps
The preparation of pulp samples is a critical area in which there are numerous
pitfalls that must be avoided. For example, the use of aluminium or stainless
steel trays in which to dry samples should be avoided as these metals react
with gold in solution so causing bias in the results. A detailed account of
recommended procedures is given by Lenahan and Murray-Smith (1986).
13.6.1.2 High grade materials

For higher grade materials, for example slag or loaded carbon, samples are
sub-divided using either a spinning riffler (Figure 13 .12) or a turning tube
divider (Figure 13.13). The spinning riffler has a carousel on which are
mounted 8 to 12 cups and is used mainly to obtain a number of replicate
samples of the same material for assay at different laboratories. The dried
sample is fed by the vibrating feeder to the rotating cups, so giving the
replicate samples.
778
PROCESS EVALUATION
Sample
hopper
Vibrating
feeder
Turntable
Figure 13.12. Spinning riffler.
Where only one or two sub-samples are required from a mass of 10 to

100 kg of primary sample, a turning tube divider is used. With this, the
material is fed into the rotating tube which discharges the sample through
one or two slots in the side of the inverted cone of the machine. The sub-
sample size can be regulated by changing the width of the slots, subject to
the minimum opening being greater than 10 mm or three times the maximum
particle size.
-tt---t--Vibrating
feeder
Turning----1-<+-
tube
Slots---\~
10 mm (min)
Sample
Reject
Figure 13 .13. Turning tube divider.
779
13.6.1.3 Run-of-mine ore

Increments taken by stop-belt sampling with 500 mm formers have a mass
of 20 kg or more and the top particle size is about 200 mm. There are a variety
of methods that can be used to prepare these samples but they can be
separated broadly into manual methods, which require little capital equip-
ment, and automatic procedures. One of each of th,ese is described below.
Manual method
The first stage in the sample preparation procedure is to crush the whole
sample to minus 38 mm using a 330 x 203 mm jaw crusher. This is sub-
divided by an alternate shovelling technique (Gy, 1979). In this technique,
alternate shovels of the crushed sample are placed in one of two wheelbarrows
to reduce the mass by a half. One of these two wheelbarrows is selected at
random and its contents sub-divided to give one quarter of the mass of the
original sample.
Further splitting into buckets is then performed using a square-sided
scoop instead of the shovel until about 1 kg of sample is obtained. This is
then dried, crushed to minus 10 mm and passed repeatedly through a Jones
divider until about 150 grams of sample are obtained. A preferred method
of sample sub-division, used in some mining companies, is to pass the crushed
sample through a turning tube divider. The sub-sample is then pulverized
using a vertical spindle pulverizer to give a finely divided sample for assay.
Mechanized method
The primary increment is first dried and then passed through two stages of
jaw crushing to reduce the particle size to minus 10 mm. The crushed sample
is fed, via a small bin and vibrating feeder, to a turning tube divider (Figure
13 .13), the sample gap for which is adjusted to at least 30 mm. A sub-sample
of between 500 and 1500 grams is obtained. This is pulverized in a vertical
spindle pulverizer and the fine powder fed to a spinning riffler to give 8
separate samples for the assay laboratory.
13.6.2 Statistical methods to quantify bias in sub-sampling

A very useful statistical technique to quantify the bias in sub-sampling and
analysis is to perform a nested analysis of variance, ANOVA experiment
Box et al., 1978). For this an increment is taken and successively divided
to give 8, 16 or more separate sub-samples, all of which are then analysed
in replicate. The data from these analyses are then treated using a nested
ANOVA design. A computer package such as the SAS system can con-
veniently be used for this purpose (SAS Users Guide, 1985).
13, 7 Mass Measurement

Mass measurement· is often a subject of heated debate on a mine with
underground personnel claiming a certain tonnage broken, trammed and
hoisted and the metallurgical department receiving a somewhat lower ton-
nage. The underground mass measurement involves survey measurement,
calculation of hopper and skip factors and sometimes a direct measurement
780
PROCESS EVALUATION
Comp~;ote<
Figure 13.14. In-motion train scale.
of the mass of hoisted tons using underground flask measurement.

Underground mass measurement is outside the scope of this work and
therefore the reader is referred to texts such as Storrar (1981).
13.7.1 Truck and train weighbridges

In situations where ore is transported by lorry or train, it is sometimes con-
venient to weigh these before and after loading to obtain the mass of ore.
There are two distinct types of weighbridge:
a) The static system in which a lorry is driven onto the weighbridge and its
mass recorded. For a train, the trucks are uncoupled before weighing
b) The in-motion system where the train does not stop as it goes over the
weighbridge.
There are a number of manufacturers for static systems. These can be assized
by the Department of Trade and Metrology, who can give a certificate veri-
fying the accuracy of a particular installation to 0,1 % . These weigh bridges
consist of a platform which, in the more modern types, is mounted on load
cells which give a mass readout. The reader is referred to the various manufac-
turers' catalogues for detailed designs· and specifications.
The static weighbridge, being both accurate and precise, is used as
reference method for calibrating other devices. For the routine measurement
of large tonnages, however, they are inconvenient because of the necessity
to stop each truck on the weighbridge.
A better method is to use an in-motion train scale. For these, the train
with its hoppers is driven at a speed of less than 10 km per hour over the
weighbridge, which consists of rails suspended on load cells. Each axle is
weighed as it passes over the bridge and the mass of the trucks computed
and printed out on a tonnage return. Figure 13 .14 is a diagram on an in-
motion train scale. Calibration is performed by measuring the mass of a truck
781
Total
Integrate
Velocity from tachometer

Multiply - ~ - - - - - - - - - I
I
I
I
Figure 13 .15. Electromechanical belt massmeters.
on a static weighbridge and then passing it over the in-motion.system. The

precision for in-motion train weighing is better than 0,3% for a good in-
stallation. If required, an in-motion train scale can be assized by the Depart-
ment of Trade and Metrology.
There are more than 40 in-motion train scales in operation in South
Africa, most of them on the coal mines, but there is one installation on a
gold mine and one on a platinum mine. Of these 40, about 12 are assized
units. It is a requirement for the in-motion train scale that the speed does
not exceed 10 km/h.
13. 7 .2 Electromechanical belt massmeters

The most commonly used method for mass measurement of ore on gold mines
is by belt massmeter, of either the electromechanical or nuclear type. Detail-
ed discussion on belt weighers is contained in Colijn (1983). Figure 13.15
is a diagram of a typical electromechanical belt weightometer.
Load cells can be used to measure the force, although in one make of
high precision assizable weightometer, the force is measured by the change
in frequency of vibrations in a wire as the force changes.
For the massmeter shown in Figure 13 .15 the mass on the belt is measured
over a single idler. For greater precision, there are other models where 2,
3 or 4 idlers are included on a single weighframe. Additionally, more than
one load cell could be used. In general, the more idlers on the weighframe
782
PROCESS EVALUATION
the less the effects of belt tension and alignment and the longer the instru-
ment will remain in calibration.
Electromechanical belt weightometers are sensitive to the alignment and
condition of the belt and idlers. For high precision it is necessary to have
specially balanced weigh-class idlers before and after the weighframe and
on the weighframe itself. In addition, a gravity take-up is required on the
belt. It is also a requirement that the conveyor frame is rigid and does not
deflect under load. The belt speed is measured with a tachometer and the
product of belt speed and belt loading gives the tonnage flow rate.
Calibration of an electromechanical belt scale can be performed in four
ways, namely:
a) By suspending known mass pieces on the weighframe and recording the
integrated mass for a minimum of three revolutions of the belt, with the
belt empty.
With well-balanced idlers and well-aligned belt in good condition,
this method is, in the absence for facilities for conducting a reliable bulk
test, the best available. A useful check on the effect of belt alignment
is to conduct the calibration several times with different masses on the
gravity take-up for the belt. Varying tensions should have no effect on
the masses recorded.
b) By placing a roller chain of known mass per unit length on the belt and
recording the totalized mass for at least three revolutions of the belt. The
chain should be long enough to span beyond the second fixed idler on
either side of the weighframe. The chain is anchored while the belt is
rotating and therefore drags over the belt. The positioning of the chain
is critical and errors can arise from a variable number of links of the chain
resting on the weighframe. The effective weighlength is from halfway
between last fixed idler and first weighlength idler to halfway through
the corresponding gap at the far end of the weighlength.
c) Modern microprocessor-based belt weightometers are provided with a
facility to check the load cells electronically. This is a useful check between
the calibration with static weights, chains or bulk tests. In addition, the
microprocessor-based weightometers can be fitted with automatic zero
corrections.
d) By bulk load tests. In a bulk load test a known mass of ore is passed over
the weightometer and the mass recorded compared to the known mass.
The bulk load test is the preferred method and is the one that is obligatory
if an assize certificate from the Department of Trade and Metrology is
required. For a bulk load test at least three revolutions of the conveyor
and at least 10 minutes' running time with the belt running at between
25 and 100% of its rated capacity are required.
The requirements for bulk load tests usually translate into a large tonnage
of ore that has to be weighed in hoppers or on an assized weighbridge before
being passed over the weightometers. In addition, it is not uncommon for
ore to be sticky and for it to hang up to varying extents in the receiving bins
783
~Source
Detector
0000000000 0
I
: Tachometer
Signal '
Velocity Total
Figure 13 .16. Nuclear massmeter.
ahead of the feed belts. Therefore, although the mass of ore placed in the
bins may be known, the mass of ore passing the weightometer will be af-
fected by changes in the quantities hung-up in the bins.
For these reasons, it is sometimes difficult to perform adequate bulk
loading tests and reliance has therefore to be placed on the other calibration
methods and other cross-checks.
An alternative procedure to weighing the ore before passing it over the
weightometer is to pass the ore over the weightometer and then divert the
ore into a flask on load cells or into a truck to be weighed later.
Several manufacturers make electromechanical belt weighers of assizable
quality where an accuracy of 0,5% is required in a bulk load test. This type
of equipment has been used in gold mining applications, not always in ideal
situations. At one mine, for example, 4-idler electromechanical belt
weightometers have been installed on the belts from the shaft head gears to
a load-out point. In this position there is a very wide fluctuation in belt loading
that is difficult to.control. The mass recorded by these is compared routine-
784
PROCESS EVALUATION
ly with the corresponding masses recorded with the same type of equipment
at the treatment plants, where steadier feeds are obtained and agreement
within about 1,5% is obtained.
13.7.3 Nuclear belt massmeters

In the nuclear belt weigher (Figure 13.16), a source of gamma rays is placed
either below or above a belt opposite a detector. The attenuation of the
gamma rays gives a measure of the mass of material on the belt. This
multiplied by the speed of the belt, as measured by the tachometer, gives
the mass flow. The mass signal is sensitive to the profile of the material on
the belt and to the moisture content of the ore, but is insensitive to changes
in the alignment, position and tension of the belt.
Ideally, calibration is done by a bulk load test in which the belt is fed
for a minimum period of 20 minutes at 50% of the maximum capacity and
the read-out adjusted to the known mass of material fed during this period.
The same problems as were mentioned for bulk calibration of an electro-
mechanical belt weigher apply to the calibration of a nuclear belt weigher
and it is therefore not always possible to perform a bulk load test as indicated
above. In this situation, for the initial calibration, static samples are located
on the stationary conveyor for several different belt loadings and the
massmeter electronics adjusted for linearity and amplification to give the cor-
rect mass reading. As an alternative, different masses of ore are spread evenly
on a thin sheet that is slid along the belt at a constant speed under the nuclear
source while the electronics are being adjusted. The calibration can be checked
by taking belt cuts and relating these back to the readings of the massmeter.
For routine calibration, to correct for the natural decay of the source,
a calibration plate is often used. This is a flat sheet of metal which, from
the initial calibration exercise, is known to have the same absorption
characteristics as a known mass of ore. Because a nuclear weightometer is
sensitive to ore profile and moisture content, manufacturers generally claim
a poorer precision for a nuclear scale than for an electromechanical scale.
One manufacturer, for example, claims a precision of 1to3% for a nuclear
massmeter as compared to 0 ,25 to 0 ,5 % for a 4-idler electromechanical device.
In practice, however, this difference in precision may not be significant.
Nuclear scales, because of their insensitivity to the condition of belts
and idlers, and their lack of moving parts, are easier to maintain than electro-
mechanical units. A major consideration, however, is that the accuracy of
all measurements made with a nuclear massmeter depends on the initial cali-
bration exercise or bulk load tests.
13.7.4 Moisture determination for mass measurement

All the methods for mass measurement given above measure the total mass
and a correction has to be made for moisture content to obtain the dry mass.
Several methods have been proposed for on-line measurement of moisture
content - microwave measurement, neutron absorption or electrical con-
ductivity - but none of these is in routine use on mines. A common prac-
tice is, however, to take samples from feed belts and determine the moisture
785
content by mass measuring, drying and measuring the mass again. For some
mines, a constant moisture factor, determined after sampling campaigns, is
used whereas at others samples are taken from the belts daily and a running
average of the moisture content is used.
13.7.5 Batch mass determination of pulps

For those mines where ore pulp is leached in batches, a convenient method
for mass determination is to fill the leach tank to a known volume, take a
sample of the pulp as it is mixing and determine its water: solids ratio. Tables
are then used to relate the water: solids ratios of the pulp and the depth of
filling of the tank, with the air agitation stopped to the mass within the tank.
This method is often satisfactory but relies upon the dip sample being
representative of the density of the total volume.
13. 7.6 Batch tanks on load cells

An alternative to the measurement of the volume and the water: solids ratio
of the pulp is to mount a tank of known volume on load cells. At one par-
ticular mine, readings from flow and density meters are integrated using a
computer control system, and at predetermined times, typically once a week,
the whole flow is diverted to a tank large enough to accommodate the total
flow for an hour, which is a volume in excess of 100 m3 • The level in the
tank is determined by ultrasonic means before it is filled and is measured
again when the tank is full.
Similarly, the mass of the tank is measured with the load cells at the
start and end of the filling process. Hence, from the measured volumes and
mass, the flow and density meters are automatically re-calibrated.
13. 7. 7 Mass of pulps using flow and density meters

The dry mass flow of solids in a pulp is commonly determined from
measurements of the volume flow using magnetic flowmeters and the
water: solids ratio of the pulp using nuclear density meters. The design and
operation of flow and density meters is described in Chapter 14. However,
a problem with these measurements is in the initial calibration of the in-
struments. Flowmeters are usually calibrated before installation and the
calibration can be checked by measuring the rise, or fall, of the level in a
tank as it is fed or emptied by the flow.
Calibration of nuclear density meters is normally performed by first fill-
ing the pipe on which the instrument is fitted with water to set the reading
at one. The other densities are then set by recirculating liquids or pulps of
known water:solids ratios or by inserting metal plates obtained from the
manufacturers between the source and detector that will give attenuations
corresponding to pulps of particular densities in the pipes.
In operation, the readings given by the density meter can be compared
to the results of water: solids determinations on samples taken from the pulp
stream. These samples are often taken from a T-piece into the main pulp
stream in the vicinity of the density meter. There is no guarantee, however,
that these samples are truly representative, particularly if there is non-
786
PROCESS EVALUATION
Table 13.8. Precision of fire assay at different gold grades (single assays).
Gold ore grade Standard deviation Coefficient of

(git) (git) variation (%)
0,5 - I 0,12 12
1 - 3 0,2 10
3 - 5 0,4 10
5 - 7 0,5 8
7 - 10 0,5 6
10 - 15 0,68 5
15 20 0,68 4
60 - JOO 2,5 3
turbulent flow through the measuring section.

A more satisfactory method of checking the calibration of flow and den-
sity measurement is to divert the flow to calibration tanks on load cells. This
is a practice employed at several gold mines to obtain accurate measurements
of tonnage.
13.8 Chemical Analysis

In the gold mining industry the most commonly used method of analysis for
ores and plant products is the fire assay technique. This, and other methods
for gold analysis, are contained in other texts (Lenahan and Murray-Smith,
1986). However, from the point of view of process evaluation it is impor-
tant to have a measure of the precision of fire assay.
Table 13.8 gives typical precisions for single fire assays on samples of
gold ore in different concentration ranges, using 100 g portions for samples
of less than about 1 git and 50 g for samples with more than 1 g/t. These
precisions, although typical, will vary from laboratory to laboratory, with
the ore type and with the mass of sample used. It is advisable to use an analysis
of variance technique as described before to measure the precision directly
in those cases where it is important.
13.9 Metallurgical Accounting

All the various components that have already been dealt with - sampling
precision and accuracy, mass measurement, sampler design - come together
in metallurgical accounting systems for a plant or process. The purpose of
these systems is to provide accurate information on the operation of a plant
so that there are reliable statistics on recoveries, grades and tonnage for
strategic decisions and to provide a measure of security against theft or other
malpractices.
A number of methods have evolved for meta)lurgical accounting, and
with the advent of the computer, changes are taking place to these systems.
787
13.9.1 Call system

The most common method used for metallurgical accounting is the 'call'
system. For this, the mass of gold that should be produced, as calculated
from the mass and assays of the inputs and outputs to a plant, is compared
to the actual production or some other independent measure of the mass of
gold.
As an illustration, consider a plant treating gold ore by cyanidation,
filtration, zinc precipitation and smelting of the zinc sludge to produce gold
bars and a slag. The mass of ore treated is taken from the readings of the
mill weightometers, after adjusting for moisture content, and the grade of
ore is obtained from samples of thickener underflow taken with an automatic
cross-stream cutter of the type already described. From this grade and ton-
nage a daily input of gold is calculated, together with progressive totals as
the month proceeds.
The mass of gold in the tailings is calculated from the grade of tailings,
as measured from samples taken with automatic cross-stream cutters, and
from tonnage of ore treated, as recorded by the mill weightometers, after
correcting for the mass of any flotation concentrate that may be produced.
The mass of gold produced is small relative to the original mass of ore and
can be neglected in calculations of the tonnage of tailings. The difference
between the gold input in ore and gold output in tailings is the gold call from
solids.
A gold call from solution is calculated from the difference in analyses
of gold-bearing cyanide solution fed to zinc precipitation and barren solu-
tion from precipitation multiplied by the mass of these solutions as measured
with flowmeters. Finally, there is the gold produced which is calculated from
the mass and assay of bullion and slag with appropriate corrections for dif-
ferences between masses and assays on mine as compared to those from the
Rand Refinery.
A comparison is made on a monthly and progressive basis after each
smelt, of the solid and solution calls and gold produced.
Two recovery figures are also used, namely:
a) Theoretical recovery defined as
. Gold in input - gold in tailings x 100010
Th eore t ica1 recovery = .,,
Gold in input
and
b) Actual recovery defined as
Gold produced
Actual recovery = . . x lOOOJo
God1 m mput
The use of the milied tonnage using mill feed massmeters as a primary measure
of both ore and tailing mass stems from the fact that in several mining groups,
there is greater confidence in the accuracy and precision of this measure-
ment as compared to tonnage measurement with flow and density meters.
This is not, however, universal. On some mines, the tonnage treated is taken
as being that determined by batch mass measurements and on others more
788
PROCESS EVALUATION
reliance is placed on the readings from flow and density meters.

Similar arguments on relative accuracy and precision apply to the choice
of thickener underflow samples as a primary measurement of input grade.
The thickener underflow is finely divided and the flow from the thickener
underflow to a sampler is even. With an appropriate sampler design, one
can therefore take a reliable sample at this point.
A consequence, however, of taking a sample at the thickener underflow
is that one does not have a measure of the input of the mass of gold into
the mill on a daily basis because of the time required for the gold to go from
the feed belts, through the mills and thickeners to the thickener underflow
sampler. Purely on the basis of the volumetric capacity of a mill and thickener,
the residence time for ore should be of the order of one to three days.
It is known, however, that gold tends to be entrapped in mill liners and
the time taken for equilibrium to be attained between the gold input to a
mill and the output can be of the order of months (Mokken et al., 1975).
Also, if gold is recovered by concentration from mill pulp streams, the
thickener underflow assay cannot be representative of the mill feed assay.
Therefore, although the thickener underflow sample may be more precise
than a belt sample of material fed to a mill, it is not necessarily a good measure
of gold input to a plant even over a period as long as a month.
Furthermore, for the same reasons, the solid gold call as often calculated
using the gold value in thickener underflow could be unrealistically close to
gold production because any gold entrapped in the mills and thickeners is
excluded from the calculations.
Time lags are also evident when comparisons are made between solu-
tion and solid calls. These comparisons only become meaningful after several
days of each accounting cycle have passed and the progressive figures ap-
proach one another.
13.9.2 Check-in/check-out system

As an alternative to the call system of accounting, some mines have adopted
a check-in/ check-out system whereby a plant is divided into a number of
units. Overall accounting, or primary accounting, is then performed by com-
paring the input into the plant as run-of-mine ore to the output from the
plant as gold or tailings.
The main purpose of primary accounting is to obtain an overall balance
across the whole plant. However, as with the call system, because of time
lags and lock-ups, meaningful comparisons are only obtained over a period.
In order to identify where lock-ups are occurring, the time lags involved and
whether imbalances are arising at any particular stage, secondary account-
ing is performed across sub-units of the process. For this secondary account-
ing a running total is kept of the difference between input and output. This
difference is due to
a) changes in the inventory in the process unit, or
b) inaccuracies in the mass measurement, sampling or assay procedures, or
c) theft or undetected loss of valuable material, or
d) a combination of the above factors.
789
200
'Cl)
c,
0.
"'u 100
""'"
2
"O
0
0
0
0 100 200 300 400 500 600

Days from start
Figure 13 .17. Cumulative difference between the gold input and
output of a run-of-mine mill.
As an example of secondary accounting, using a check-in/check-out system,

Figure 13 .17 gives the cumulative difference between the gold input to a run-
of-mine mill and the output from it from the day the mill was commission-
ed. The input is calculated from the mass of ore, as measured with an electro-
mechanical belt massmeter, and the grade of ore, as determined by stop-belt
sampling of the run-of-mine ore. For the output, the mass is taken as being
the same as the input mass and the grade is obtained from samples of cyclone
overflow taken with a well-designed cross-stream sampler.
Figure 13 .17 shows that there was a rapid accumulation of gold in the
mill, up to about 150 kg in the first 100 days of operation. Thereafter there
was a steady increase in gold lock-up with some releases as indfcated by in-
flections in the curve until there was a maximum lock-up of 300 kg after about
330 days. Since then gold has been released from the mills, possibly because
of falling feed grade.
This simple cumulative sum technique (cusum) is useful also to quan-
tify accumulations in other units of a plant. More sophisticated cusum techni-
ques designed to determine whether changes such as those shown qualitatively
on Figure 13 .17 are statistically significant are available but are not discussed
here.
13.10 Further Deveiopments

The preceding sections of this chapter give the present state of the art for
process evaluation in the gold mining industry. Increasing use, however, is
being made of computers. At present, in some groups, personal computers
are being used to automate the metallurgical accounting functions, and in
others some aspects of metallurgical accounting are being handled on pro-
790
PROCESS EVALUATION
cess control computers as described in Chapter 14. The use of computers,

either as an addition to a process control computer or as a separate entity,
will become more common as more sampling and mass measurement points
are introduced to obtain closer control.
Historically, it was considered adequate to take the minimum number
of samples needed to establish a balance. Indeed, taking any more than the
minimum set of measurements raised the problem of how to reconcile the
resulting redundancy. A set of measurements containing redundancy will
almost certainly not balance as they stand, but will need 'adjustment' to meet
the balance assumption.
Typical cases that can lead to redundancy are flowsheets with recycle
loops, or flowsheets where more than one constituent is assayed, or with
several mass flow measurements available.
A large number of papers deal with the problem of how to adjust
measurements so that they satisfy the balance assumption. These papers are
mostly to be found in the mining, metallurgical, and chemical engineering
literature. A good review paper is by Reid et al. (1982), who also cover com-
puter packages designed to solve the materials balance problem. Most of the
computer techniques involve defining some goodness of fit criterion for the
measurements which is then minimized with nonlinear optimization methods.
Some papers also give some measure of the confidence in the obtained
balance; for example Smith and Ichiyen (1973), or Laguitton and Wilson
(1979), who use Monte Carlo simulation to generate information experiment-
ally about variances and covariances.
A statistical approach is taken by Hunt and Hinde (1985), who set out
to provide a framework within which to address broader questions such as:
How much is sampling information worth? What are the objectives and how
can they be met? This framework has the potential to tackle other statistical
problems, such as devising tests of the bias at a particular sample point.
Work is in progress to incorporate balancing and statistical techniques
into process evaluation procedures making extensive use of computer
facilities. It is to be expected that over the next few years progress will be
made in these areas.
13.11 References
Box, G.E.P., Hunter, W.G. and Hunter, J.S. (1978). Statistics for Experimenters. Wiley, New
York.
Burton, P.F., Douglas, W.D. and Tumilty, J.A. (1982). A limited review of sampling and pro-
cedures. Anglo American Research Laboratories Project D/82174, Report 2, Johannesburg.
Colijn, H. (1983) Weighing and Proportioning of Bulk Solids. In Series on Bulk Materials
Handling. Trans Tech Publications, Clausthal-Zellerfeld.
Gy, P. (1979). Sampling of Particulate Materials. Theory and Practice. Elsevier, Amsterdam.
Hunt, C.A. and Hinde, A.L. (1985). Statistical aspects of materials balancing in the minerals
industry. J. S. Afr. Inst. Min. Metal!., 85, 131-135.
Laguitton, D. and Wilson, J.M. (1979). MATBALII, A Fortran program for balancing mineral
processing circuits, 18th Annual Conference of Metallurgists, CIM Metallurgical Society,
Sudbury.
791
Lenahan, W.L. and Murray-Smith, R. (1986). ASS"ay and Analytical Practice in the South African
Mining Industry. Chamber of Mines of South Africa. Johannesburg.
Merks, J.W. (1985). In Series on Bulk Materials Handling. Trans Tech Publications, Clausthal-
Zellerfeld.
Mokken, A.H., Blendulf, G.K.l. and Young, G.J .C. (1975). A study of the arrangements for
pulp discharge on pebble mills, and their influence on mill performance. J. S. Afr. Inst. Min.
Metal/., 75, 249- 289.
Perr~, R.H. and Chilton, C.H. (1973). Chemical Engineers Handbook, 5th edn., Section C-17.
McGraw-Hill, Kogabusha, Tokyo.
Reid, K.J., Smith, K.A., Voller, V.R. and Cross, M. (1982). A survey of material balance com-
puter packages in the mineral processing industry. Proceedings of the 17th APCOM Sym-
posium. Society of Mining Engineers of AIME, New York.
Rendu, J.M. (1978). An Introduction to Geostatistical Methods of Mineral Evaluation. S. Afr.
Inst. Min. Metall., Johannesburg.
SAS Users Guide (1985). Statistics, 5th edn., pp. 433- 506. SAS Institute, Cary, North Carolina.
Smith, H.W. and Ichiyen, N. (1973). Computer adjustment of metallurgical balances. Bull.
CIMM, 66, 97-160.
Spiegel, M.R. (1961). Statistics, Schaum 's Outline of Theory and Problems of Statistics. McGraw-
Hill, New York.
Storrar, C.D. (1981). South African Mine Valuation. Chamber of Mines of South Africa,
Johannesburg.
792
Chapter 14
Process Control
I.M. MacLeod and H.E. Bartlett
14.1 Introduction
In general terms, control is concerned with the manipulation of the inputs
to a system (a machine, process or plant) so that the outputs meet certain
specifications. Control is thus a broad concept which encompasses long-term
process operating strategy based on process evaluation, as discussed in
Chapter 13, as well as manual control and the various forms of real-time
automatic control such as logic or sequencing control, single-variable or multi-
variable continuous control and supervisory control. These various forms
of real-time control are collectively known as process control and form the
subject of this chapter.
In order to give a satisfactory treatment of the subject, both the science
and the technology of process control must be dealt with. The science of
process control includes the theories of dynamic modelling, feedback, stabili-
ty, disturbance rejection, interaction and controller design. These principles
are therefore initially reviewed. The technology of process control includes
the vast technology of process measurement and monitoring as well as aspects
such as the interfaces between the process, the operator and the control system
and centralized and distributed control system architectures. These aspects
are discussed further in subsequent sections.
An important characteristic of ore treatment plants is the difficulty in
obtaining on-line measurements of key process variables and the large
maintenance effort required to keep the process sensors operating reliably.
This is primarily due to the aggressive properties of process streams such
as pulps, the harsh environment encountered in extraction plants and the
difficulty of designing reliable and representative sampling systems. These
are perhaps the most important distinguishing characteristics of process con-
trol in ore treatment. Process measurement technology for gold extraction
is discussed later in this chapter.
It is by no means easy to answer the questions that arise when consider-
ing the introduction of process control in ore treatment plants. Some of these
questions are:
a) What are the key process variables that should be controlled?
b) Is there an economic justification for control?
c) What can be controlled?
d) What control philosophy should be used?
e) Will the reliability of the proposed control system be high enough?
Questions (a) and (b) are closely interrelated and concern the determina-
793
tion of the need for process control. In the past, control system design was
usually done only after the process design had been completed. This still hap-
pens in many cases. However because of advances in on-line sensors and the
availability of flexible and sophisticated computer-based control systems, the
interaction between process flowsheet design, equipment selection and pro-
cess control system design can now be exploited. Much can be gained by link-
ing process design and control system design for new plants. It remains dif-
ficult, however, to quantify the economic benefits of a proposed process con-
trol system. Some financial mo9els have been proposed for this purpose (Pur-
vis and Erickson, 1982; Chang and Bruno, 1983).
Question (c) reduces to the following three sub-questions:
What can be measured?
Are suitable actuators available?
What configuration of control loops should connect the sensors and
actuators, and will it be possible to obtain satisfactory dynamic
performance?
This last sub-question will be discussed under the principles of automatic
process control. Question (d) relates to operational aspects such as whether
to use manual or fully-automatic control and whether to use analogue or
computer control. Question (e) concerns detailed system design, equipment
selection and maintenance issues. This chapter aims to provide essential
background material on process control theory and technology for ore treat-
ment applications in the gold industry.
14.2 Historical Background

Although standard commercial instrumentation and process control equip-
ment had been widely and successfully applied in the chemical and petro-
chemical industries since the 1930's, as Fletcher (1972) pointed out, instrumen-
tation first found its way into South African gold extraction plants only in
the 1950's. Even through the 1960's, instrumentation and process control
were not regarded as being as vital in the gold industry as they were in the
advanced chemical industries. Fletcher stated that by the early 1970's two
schools of thought had emerged. One encouraged instrumention and
automatic process control in spite of the poor performance of the equipment
then available and of the general shortage of qualified instrument maintenance
personnel and engineers who were experienced in the special requirements
of ore treatment. The other believed that efficient process operation could
be achieved more readily via manual control by better-skilled operators. At
the time of writing (1986), the latter opinion is rapidly disappearing and the
benefits of process controi and optimized operation are becoming more widely
accepted. Fairly advanced automatic control systems using mimic panels for
the operator interface have been extensively used during the last 10 years
and computer-aided monitoring and control and the use of sophisticated video
displays for the operator interface are now fairly common in modern ore
treatment plants.
794
PROCESS CONTROL
14.3 Principles of Continuous Process Control

14.3.1 Control objectives
It is necessary to approach the subject of process control system design from
a wide perspective. The control of each individual unit in an ore treatment
plant is complex and control of the overall operation of an entire plant is
an extremely complex undertaking. Furthermore, it is well known that it is
very difficult to reconcile the large-scale real problems which are encountered
in process control practice with the simple well-defined problems treated in
standard control theory textbooks (Dorf, 1986; Stephanopoulos, 1984; Smith
and Corripio, 1985).
The first and usually also the most difficult step in control system design
is to understand and formulate clear control objectives. A good understand-
ing of long-term process operating strategy is essential for this. The com-
plexity of the problem can be appreciated by considering the following fac-
tors underlying operating strategy in gold extraction plants.
The most important factor is that ore that is brought to the surface must
be treated in the gold plant as quickly as possible in order to minimize the
ore inventory held on surface. The reason for this is that the mining opera-
tion represents approximately 90% of both capital investment and operating
cost, and any untreated ore inventory is thus an extremely costly investment.
Also, especially in the case of older plants with a large number of small in-
dividual process units and little or no instrumentation, both plant design and
plant operating procedures must be very conservative so that the fluctuating
throughputs dictated by mining can be accommodated with only a small ef-
fect on gold recovery. Another important factor is that such vital informa-
tion as the amount of gold in the incoming ore feed to the plant and its
physical properties is not available to process operators and metallurgists
until the following day. The plant must therefore be run without knowledge
of the material currently passing through it. This also dictates conservative
operation.
Operating strategy based on the above factors has as its prime require-
ment the need to maximize mill throughput which is therefore a typical key
control objective. Other possible control objectives are improved recovery
of the valuable minerals, improved head grade to cyanidation or reduced
operating costs - for example, by savings in reagents or more productive
use of plant personnel.
It is normal practice to focus on certain specific high-level control ob-
jectives. These would vary from plant to plant and would also change from
time to time within one plant. When the demand for gold is high, for exam-
ple, the objective might be to maximize throughput. When demand is low,
improved recovery and/ or lower operating costs could possibly become the
key objectives. It may be very important to design sufficient flexibility into
a process control system to allow for such changes in objective.
The control objectives for each individual process within an ore treat-
ment plant must also be clearly stated in relation to the overall plant objec-
tives. For example, in a grinding circuit, possible objectives would be to main-
tain the finest possible product size at constant throughput, to maximize
795
Disturbance
inputs
Control Plant or Measured

input process output
Unmeasured
outputs
Figure 14.1. Classification of process variables.
throughput and keep product size within a limited range or to maximize

downstream plant performance (Lynch, 1977). Such different objectives can
require significantly different control schemes.
The high-level objectives discussed above all imply some form of per-
formance optimization. Before any such optimization can be attempted, it
is essential to be able to steer the process towards the desired steady-state
operating point and to keep it there. The primary control objective in any
overall process control scheme is therefore that certain key physical variables
(e. g. flows, concentrations, densities, levels, temperatures, pressures and
speeds) are kept as close as possible to their target values, called set-points,
for as much of the time as possible. It is necessary for the control system
to manipulate suitable process inputs to force this to happen in spite of distur-
bances. The presence of disturbances and uncertainty as to the exact behaviour
of the plant are in fact the fundamental reasons for using control.
As mentioned above, the practice of conservative plant operation is
natural in older plants with complex flowsheets and little or no instrument-
ation. Developments in gold process plant design during the 1970's and 1980's,
such as the trend towards simpler flowsheets using high-capacity run-of-mine
mills, a small number of high-capacity thickeners and continuous downstream
recovery processes, together with improvements and greater flexibility in in-
strumentation and control system technology, have resulted in a growing abili-
ty to control plant operation and have therefore also created new options
in process operating strategy. A clear specification of control objectives is
the key to exploiting these developments for the optimization of plant
economic performance.
14.3.2 Classification of process variables

In dealing with process control concepts, it is useful to classify the physical
process variables into two categories:
796
PROCESS CONTROL
Outputs: These are the key process variables to be kept as close as pos-
sible to their set-points, that is, controlled. The outputs can be further
sub-classified as measured outputs and unmeasured outputs.
Inputs: These are the variables that, when changed, cause one or more
outputs to change. The inputs can be further sub-classified as control
inputs and disturbance inputs.
Sometimes the control objectives relate to properties that cannot be
measured, that is, unmeasured outputs. In such cases, other variables which
can be measured easily and reliably must be used to infer the unmeasured
output.
The control inputs are also called manipulated variables. These are the
variables that are changed by the controller to drive the outputs towards their
set-points. All other process variables that affect the outputs in any way are
called disturbance inputs or process load variables. Variations in these distur-
bance inputs cause unwanted changes in the process outputs. These defini-
tions are summarized in Figure 14 .1.
14.3.3 Control system structuring

As indicated in Section 14.3.1, the real problems that must be tackled in con-
trol system design are large and complex. On the other hand, control theory
deals only with certain small, well-defined problems. Unfortunately there
is, as yet, no systematic approach for bridging the gap between these two
extremes. The design of a plant-wide control system must be approached on
the basis of successive refinement by trial and error; it is therefore highly
dependent on the experience and ingenuity of the designer. Two main ap-
proaches are used to guide the structuring of an overall process control system.
These are known as the bottom-up approach and the top-down approach.
The bottom-up approach is used most often in practice. It begins with
the choice of individual output variables to be measured and controlled and
the choice of control inputs. Simple, standard control configuratiop_s are then
used as building blocks. These are introduced in various combinations until
a control system that meets the control objectives is obtained. A detailed
exposition of this approach is given by Stephanopoulos (1984). The individual
local controllers rely on the principles of single-variable feedback and
feedforward.
The control configuration shown in Figure 14.2 is referred to as feed-
back control. The principle of this approach is to reduce the effect of a distur-
bance input by first measuring its effect on a process output and then
calculating the necessary correcting input. It is necessary for the disturbances
to enter the local system in a well-defined way for this to work. It is also
necessary to be able to adjust a control input that enters the system in the
neighbourhood of the disturbance input.
Control based on the feedback principle has the major advantage that
it incorporates an inherent self-correcting action and that it is not necessary
to have detailed information about the characteristics of the process, pro-
vided that a high gain can be used in the loop. For the same reason it can
797
Disturbance
Inputs
Actuator
PLANT
Process
sensor
Cont-
roller
Figure 14.2. The principle of feedback control for reduction

of the effect of a local disturbance input.
deal with disturbance inputs even though these may not be measurable.
The basic feedback control configuration is by far the most important
and widely used building block in process control systems des'igned accor-
ding to the bottom-up approach. Many common industrial processes behave
as combinations of capacity and pure delay and can tolerate a high loop gain
(Dorf, 1986; Stephanopoulos, 1984). A simple standard method of calculating
the correcting control input can therefore be used in most cases. This has
led to the standard three-term controller which is discussed further in Section
14.3 .4.
Situations commonly occur in the bottom-up approach where standard
multiple control loop configurations can give improved control. These in-
clude the cascade control configuration, multiple actuator (i.e. control in-
put) control systems and the selective control configuration. These are treated
thoroughly by Shinskey (1979). An example illustrating the use of two cascade
control configurations is the ball mill grinding circuit control scheme shown
in Figure 14.3. This scheme is based on maintaining constant cyclone feed
water: solids ratio.
If the disturbance inputs can be directly measured, the alternative con-
trol configuration shown in Figure 14.4 can be used. This is referred to as
feedforward control. A control input that attempts to counteract the effect
of the disturbance is generated from the measured disturbance input. This
presupposes that a reasonably accurate model of the characteristics of the
process is available for calculating the necessary control input. For this reason,
feedforward control is not as robust as feedback control and is therefore
usually used in conjunction with a feedback control loop in order to improve
its robustness to unmodelled effects.
The top-down approach to control system structuring begins with a
798
PROCESS CONTROL
Legend
A Level control by water

B Density control by
fresh feed rate
DC= Density controller
DT Density transmitter
FC Flow transmitter
r~~;-
FY Current-pressure converter
LC Level controller
LT Level transmitter
SC Speed controller
SP Set point
Water jI
1-----1
Figure 14.3. Ball mill grinding circuit control scheme based on

maintaining constant cyclone feed density.
careful analysis of the control objectives. This leads to the selection of the
control principles to be implemented. In a grinding circuit, for example,
typical control principles could be cyclone inlet solids flow control, cyclone
underflow density control or mill power maximum-seeking control. Based
on each selected control principle, the variables that should be controlled
are identified. These variables may not be measurable and alternative
measured outputs that are closely related to the variables given by the con-
trol principle must be chosen. Suitable control inputs must also be chosen.
The next step is to group related outputs and the control inputs that
affect them into small sub-systems such that there are only weak couplings
between the sub-systems. Unfortunately there are no general rules to guide
this grouping process and it must be done by trial and error, possibly guided
by a detailed analysis or simulation of process behaviour. A suitable con-
troller must then be designed for each sub-system to' maintain its outputs
at their set-points. Often the same standard single-variable feedback and feed-
forward control configurations discussed above are adequate for this pur-
pose, but in other cases, due to strong interactions between variables within
a sub-system, it is necessary to resort to the use of a more sophisticated multi-
variable controller. Good examples of the top-down approach applied to
grinding circuit control which do lead to the need for multi-variable con-
trollers are available in the reports by Hulbert and Braae (1981) and Hulbert
(1983). A number of techniques such as the Inverse Nyquist Array method,
799
Controller
Measured
disturbance input
PLANT
Actuator
Figure 14.4. The principle of feedforward control.
the Dyadic Expansion method and the Characteristic Locus method are
available for carrying out the detailed design of multi-variable controllers.
Owens (1978) provides a good treatment of these methods. All these design
techniques lead to an essentially identical final controller implementation.
Another excellent example of control system design according to the top-
down approach is provided by Pauw and co-workers (1985). In this work,
a clear identification of the control objective leads to a control system struc-
ture based on a combination of single-variable feedback and feedforward
controllers, and logic control.
14.3.4 The standard three-term industrial controller

As stated in Section 14.3.3, a standard generalized form of automatic con-
troller meets the majority of industrial requirements. The application of such
a controller is illustrated in Figure 14.5. These controllers are generally design-
ed as modules suitable for panel mounting and provide displays of the
measured process output variable and of the controller output to the actuator
as well as a convenient way of supplying the set-point. They are available
from a large number of suppliers. The same standard generalized control
module is also commonly implemented as a program in process control com-
puter systems. Here, flexible colour video displays of measured process out-
put and controller output as well as other controller parameters are usually
provided.
Internally in the block shown as Ge these controllers carry out signal
processing according to the following equation:
m(t) = Kc e(t) + _l_ \ e(t) dt + T de(t) (14.1)

T; J d dt
.where
m(t) controller output,
e(t) error between set-point and measured process output,
Kc proportional gain (reciprocal 100/Kc is known as proportional
band),
integral constant (alternatively known as reset time),
derivative constant (alternatively known as derivative time).
800
PROCESS CONTROL
r------------,
Set-point I
generator I m
I Ge PLANT
I Output
I + _ _ _ _ _JI
L ____ _
Controller
Figure 14.5. The standard industrial controller.
Because of the form of Equation (14.1), this type of control is known

as proportional, integral, derivative or PID control, or alternatively as three-
term control. The constants Kc, T. and Tct can be adjusted easily. The opera-
tion of determining suitable values for these constants is known as controller
tuning. Often only the first two terms are used in practice, the controller
then being called a PI controller.
Full discussions of the standard three-term controller, its dynamic per-
formance when used to control various common types of process and the
significance of each of the terms are available in standard texts
(Stephanopoulos, 1984; Shinskey, 1979).
Methods for tuning this kind of controller based on experimental tests
are of particular importance because they can be effectively applied on site.
Ziegler and Nichols (1942) originally introduced the idea of optimal settings
for the adjustable parameters of conventional PI and PID controllers. These
settings could easily be calculated using given formulae once an approximate
model of process behaviour had been found by means of simple on-line tests.
The Ziegler and Nichols approach has subsequently been refined and im-
proved and has been widely used in practice. Modern versions (Fertik, 1975)
provide detailed practical guidelines on how to tune a standard controller.
In effect, the Ziegler and Nichols approach provides a direct way of incor-
porating a simple model of the dynamic behaviour of the process into a PID
controller.
14.4 Process Measurements

14.4.1 Introduction
Industrial instrumentation is a vast subject, and emphasis in this section will
therefore be placed only on those techniques that are most important in gold
ore processing. Adequate treatment of general industrial instrumentation is
available in standard texts (Considine, 1985; Doebelin, 1983; Bentley, 1983).
Before discussing specific measurement techniques, it is worth re-
emphasizing the special practical problems that are encountered in ore pro-
cessing plants due to the aggressive nature of the process streams and the
harsh plant environment. In the gold industry, it is necessary to measure pro-
perties such as the flow rate of solids on conveyor belts, the flow rate and
density of dirty liquids and pulps in pipes and launders, the particle size and
size distribution of mill products in pulps, the flow rate and concentration
801
Readout
Signal Signal
Trans- Recording
condit- process-
ioning ing mission
Controller
Computer
Figure 14.6. General structure of a measurement system.
of reagents in pulps and other chemical and physical properties of components

that occur only in very small concentrations. Many of these properties can
only be measured indirectly by employing appropriate primary and secon-
dary sensing elements in series. Moreover, many standard industrial in-
struments developed for use in the chemical, petrochemical and similar in-
dustries are suitable only for relatively clean environments and cannot with-
stand abrasive pulps.
It is also vital to appreciate that successful and reliable operation of an
industrial instrument depends upon all of the components associated with
the measurement system. These include the process stream sampling devices,
the sensing element, the signal transmission cable, the data presentation
device, the auxilliary electricity, air and water supplies and the trained per-
sonnel and equipment needed to maintain and calibrate the measurement
system. When selecting an instrument it is therefore very important to con-
sider the total measurement system life-cycle costs, especially maintenance
and spares-holding costs, in addition to initial capital costs.
The growing importapce of process control and the increasing use of
process control computers (which are less forgiving of measurement errors
than a human operator) place even greater demands on the reliability and
repeatability of instrumentation. In certain cases, very accurate data will also
become an increasingly important requirement. Successful industrial measure-
ment requires a thorough understanding of the basic principles of measure-
ment and careful attention to all the components of the measurement system.
14.4.2 Terminology of instrumentation systems

Figure 14.6 shows the functional elements found in typical industrial measure-
ment systems. The physical property to be measured is known as the
measurand. The primary sensing element interacts with the process medium
and produces an output that varies as some known function of the measurand.
This output variable is often a physical property that is not in a form suitable
for further processing or transmission. It is the function of the signal condi-
tioning element or transducer to convert the output of the primary sensing
element into a more suitable form such as a voltage signal, while preserving
the information contained in the original output of the primary sensing ele-
ment. The resulting signal may then require subsequent signal processing,
for example amplification, filtering and analogue to digital conversion, in
order to make it suitable for transmission.
The equipment to carry out the above three functions is usually housed
802
PROCESS CONTROL
in one or more enclosures located close to the point of measurement in the

plant; it is commonly known as the instrument, field transmitter or sensor.
The remainder of the measurement system consists of the transmission
medium (usually copper wires) to carry the signal to a convenient point such
as a central control room and possible further signal processing in the con-
trol room. Thereafter the signal is available for uses such as driving read-
out equipment, recording, providing the feedback signal for a controller or
input to a process control computer.
14.4.3 The primary sensing element

The primary sensing element is the most critical part of a measurement system.
The chosen physical or chemical principle of operation is the dominant fac-
tor in determining the performance characteristics of the whole measurement
system. The materials of construction which can be used for the primary sen-
sing element determine the ability of the instrument to withstand the process
medium and the environment. In practice, the primary sensing element is
always unintentionally sensitive to physical properties other than the desired
measurand. These are called interfering inputs and they can, of course, vary
as a function of time and thereby cause the output of the primary sensing
element to vary erroneously. The user of the output of the measurement
system is usually unaware of the error. Temperature is often the most
troublesome interfering input.
All the other functional elements of a measurement system are also af-
fected by interfering inputs. Adherence to good instrument engineering prac-
tice can, however, usually reduce the effect of these other interfering inputs
to negligible proportions.
14.4.4 Accuracy and calibration

In process instrumentation, the accuracy of a measurement is the closeness
of the displayed or output value to the true value of the measurand. It is
quantified in terms of the measurement error or inaccuracy, that is, the dif-
ference between the measured value and the true value. The true value of
a variable is defined as the measured value given by a standard instrument
of ultimate accuracy.
In spite of this seemingly clear definition, the question of accuracy in
process instrurp.entation is ambiguous and confused and the user should be
very careful when interpreting accuracy specifications. Part of the problem
is that the so-called true value is generally unknown except in certain specializ-
ed laboratories. The problem is also compounded by the influence of the
environment and all the measurement system components (primary sensing
element, circuitry and read-out equipment) on total measurement system ac-
curacy. The accuracy of an instrument depends on the environment in which
it must operate.
A proper accuracy specification first states the allowable ranges for all
the significant interfering inputs and then gives an estimate for the maximum
total error or inaccuracy. This is usually quoted as a percentage figure based
on the full-scale reading of the instrument. Another method sometimes used
803
gives the error as a percentage of the particular reading with a qualifying

statement which applies to the low end of the scale. Note that an accuracy
specification does not give the actual error for a particular measurement,
but rather gives worst-case bounds for its error.
Repeatability is the ability of the measurement system to give the same
output for the same input, repeatedly applied to it. The non-repeatability
error is only one of the components contributing towards the total error in-
cluded in the accuracy specification. Good repeatability is sometimes all that
is necessary when the measurement is to be used for process control pur-
poses. Note that repeatability is closely related to the statistical concept of
precision discussed in Chapter 13.
The static characteristics of a measurement system element can be found
experimentally by holding all the inputs except one at known constant values.
The one input under study is then varied through a range of known constant
values and the corresponding element outputs are noted. The input/ output
relationship obtained in this way comprises a static calibration valid under
the stated constant conditions of all the other inputs. In order to describe
the overall static b.ehaviour of the instrument, this calibration procedure is
carried out for the measurand or desired input and for all significant interfer-
ing inputs as well. The calibration procedure may include the operation of
adjusting the instrument in order to reduce certain error components.
The static calibration procedure described above clearly refers to an ideal
situation which can only be approached, but never reached, in practice. It
is particularly difficult to ensure that all interfering inputs are held constant
and to achieve the required known constant input values. Usually it is only
the instrument manufacturer who performs such comprehensive tests. The
user of the instrument should always be aware of errors introduced by dif-
ferences between the ideal situation and the real plant environment. It is also
important to note that calibration is a static test and that in practice the
measurand varies dynamically. However, ore treatment plant applications
generally involve the measurement of quantities that vary only quite slowly
and the dynamic response of instrumentation systems is therefore seldom
a serious limitation.
14.4.5 Basic physical measurements

14.4.5.1 Flow rate of solids
This important measurement has been dealt with under mass measurement
in Chapter 13 (Section 13.7).
14.4.5.2 Flow rate of liquids

Measurement of the flow rate of a liquid in a pipe is straightforward in many
industries; simple primary sensing elements such as the orifice plate are widely
used (Considine, 1985; ISO 5167, 1980). In ore processing, however, the most
important flowing media to be measured are pulps and water containing
significant amounts of solids in suspension. These materials cause major pro-
blems with many of the standard industrial liquid flow measuring devices
804
PROCESS CONTROL
E = BDV
Magnet coil
Figure 14. 7. Principle of the electro-magnetic flow meter.
due to erosion and blocking. For this reason, the electro-magnetic flow meter,
which causes essentially no obstruction to flow, is the most successful and
widely used flow measuring device. This has a typical accuracy of ± 1% of
indicated flow, and measurement is independent of changes in viscosity,
pressure, density and temperature.
The principle of the electro-magnetic flow meter is Faraday's Law of
induced e.m.f. as shown in Figure 14.7. A magnetic field is set up across
a pipe and a fluid flowing at average velocity, V, cuts the magnetic flux of
flux density, B, at right angles. The induced e.m.f. is normal to both Vand
B and is detected by a pair of electrodes on the inner surface of the pipe
as shown. The magnitude of the e.m.f. is given by the equation
E=BDV
where D is the diameter of the pipe.
The electro-magnetic flow meter is therefore actually a velocity measuring
device. Assuming a full pipe and constant pipe diameter, the flow velocity
can however easily be converted to a volume flow rate. It is, of course,
necessary for the liquid to be conductive. However, with high-impedance
electronic amplifiers, very low conductivities are acceptable and this is seldom
a practical limitation.
A. C. electro-magnetic flow meters use a 50 Hz alternating magnetic field
derived from the standard a.c. line current. This eliminates polarization ef-
fects (formation of an insulating layer of gas) at the electrodes and also per-
mits the use of high-gain a.c. amplifiers which have far less drift than com-
parable d.c. amplifiers. A.C. electro-magnetic flow meters have the advan-
tage of a relatively low cost and are very widely used. A disadvantage with
a.c. systems, however, is that the strong alternating magnetic field induces
spurious e.m.fs. into the sensitive measuring circuitry and the amount of
interference is variable and changes with time. The so-called d. c. electro-
magnetic flow meter has been introduced to overcome this problem for ap-
plications requiring high accuracy. These actually use a magnetic field that
is switched on and off at a low frequency (usually around 10 Hz). When
the field is switched off, any residual instrument output is considered to be
an error and is stored and then subsequently subtracted from the total out-
805
put obtained when the field is next switched on again. This provides a con-
tinuous, automatic zeroing effect that largely overcomes the problem of
spurious induced e.m.fs. An additional advantage of d.c. systems is that they
consume far less power than equivalent a.c. systems. Problems have however
occurred when using d.c. systems with certain mining pulps. In these cases
a severe spurious noise signal is induced across the electrodes. Expert advice
should be sought in doubtful cases.
Certain basic guidelines apply to all liquid flow meter installations. The
flow meter should ideally be installed in a vertical pipe section with an up-
wards flow direction to ensure that the pipe remains full. There should be
no bends or other restrictions immediately upstream of the flow meter. Final-
ly, for electro-magnetic flow meters it is important to have a good earth con-
nection between the instrument casing and the measured liquid.
14.4.5.3 Liquid level in a vessel

It is often necessary to measure the height or level of the surface of the li-
quid in some vessel such as a tank or sump. A number of standard industrial
level measurement techniques operate successfully in extraction plants. These
include air bubblers in which the back-pressure of air in a dip tube is pro-
portional to the head of liquid (assuming constant liquid density), various
arrangements using floats or bobs suspended from wires or cables which are
attached to a displacement sensing unit mounted above the vessel, electrical
capacitance or conductivity probes and measurement of the differential
pressure between the bottom of the vessel and the gas above the liquid surface.
Bubbler tube and differential pressure methods are simple and reasonably
reliable, but have the disadvantage of being sensitive to the liquid density
and subject to erosion, while floats and capacitance probes tend to be sub-
ject to fouling and need regular maintenance.
In addition to the above methods, level measurement instruments using
ultrasound have recently become popular in extraction plants. These operate
on the principle of electronically measuring the time taken for the echo of
a pulse of radiated ultrasonic energy to return from the liquid surface to a
transceiver. No moving parts are involved and it is not necessary to have
any probes or sensing elements in contact with the liquid (a non-invasive
technique).
One source of error in ultrasonic level measurement is the dependence
of the speed of sound on the temperature of the medium through which it
travels (air temperature). Also, the surface being measured must reflect suf-
ficient sound energy. Froth or bubbles on the liquid surface can cause prob-
lems due to sound absorption. It is also important to ensure that the ultrasonic
transceiver is installed on a vibration-free mounting.
Finally, it can be mentioned that this technique has also been successfully
used to measure the height of the liquid/ solid interface in thickeners where
the solids form a mud layer below the water surface. Here the sound pulses
are radiated downwards through the water to the mud line and are then
reflected back to the transceiver.
806
PROCESS CONTROL
14.4.5.4 The water content of a pulp

The water:solid ratio (w/s) of a suspension of ground solids in a flowing
water stream is a very important variable for process control in a gold ex-
traction plant. This is measured by sensing the average density of the pulp
stream as it flows through a pipe. The nuclear density meter is the standard
device for performing this measurement.
The operation of the nuclear density meter relies on the fact that the
gamma rays produced by a radioactive source arrive at a detector with an
intensity that is inversely proportional to the density and thickness of the
material through which the radiation passes. The radioactive source, .usual-
ly caesium-137, is located on one side of a pipe and a radiation detector is
placed on the opposite side. With suitable amplification and correction for
radiation attenuation by the walls of the pipe and assuming a constant inter-
nal pipe diameter, an electrical signal proportional to the instantaneous
average density of the flowing pulp can be produced from the radiation detec-
tor output. This measurement technique is simple, non-invasive and has been
found to be reliable in practice. The ionization chamber is the most satisfac-
tory kind of radiation detector in these applications.
Important sources of error in nuclear density measurement include the
presence of entrained air bubbles in the flowing pulp, the build-up of material
on the inside walls of the pipe and the natural exponential decay of the
radioactive source. The density of air is very much lower than that of pulp,
and air bubbles therefore can cause a significant measurement error. This
problem is particularly severe when the nuclear density meter is installed close
to the delivery of a pump that is sucking in air due to air entrainment in
its feed or due to poor seals or is experiencing cavitation. These points should
be carefully checked. Problems due to material build-up can be reduced by
installing the density meter in a vertical pipe section with an upward flow
direction. Finally, compensation for errors due to radioactive source decay
can be included in the density meter electronics. This is particularly impor-
tant if small density changes are to be measured. With dilute pulps (w/ s ratio
> 10) the sensitivity to the above errors becomes worse due to the non-linear
relationship between density and wIs ratio; the precision of density measure-
ment therefore becomes poor. The calibration of nuclear density meters is
discussed in Chapter 13 (Section 13.7.3).
14.4.5.5 On-line particle size measurement
A key parameter in assessing the performance of a milling circuit is the par-
ticle size distribution of a cyclone overflow. This is measured by a screen
analysis of a representative sample for record purposes. For dynamic con-
trol, two automated measurement methods are available, namely, the Arm-
co Autometrics Particle Size Monitor (PSM) and the Leeds and Northrup
Microtrac analyser.
The Armco Autometrics PSM
The principle of the Armco Autometrics PSM is the measurement of the at-
tenuation of ultrasonic waves as they pass through a sample of the pulp.
From this attenuation, the pulp water: solids ratio and the percentage of
807
1?]~315:~ To vacuum source

(water aspirator)
Impeller drive
\motor
'\
)
\
Discharge from
impeller below
fluid level in t a /
\/
Typical fluid level
in impeller
section
/'
To electronics
Figure 14.8. Armco Autometrics PSM - air eliminator.

section
material passing a particular size fraction are calculated. The instrument com-
prises three sections: the air eliminator, the sensor section and the electronics
section.
The air eliminator draws a sample from the process pulp stream and
removes entrained air bubbles. The de-aerated pulp then passes between the
sensors where particle size and w/ s ratio are sensed by ultrasonic means and
the pulp returns to the main stream below the sampling point.
The air eliminator is shown in Figure 14.8. This shows the impeller sec-
tion where air is removed and the sample cell where the two pairs of ultrasonic
·transducers are situated. One pair is set at an ultrasonic frequency which
is not sensitive to particle size and the second set at a frequency that is af-
fected by the particle size range selected for the particular instrument.
The instrument requires between 55 and 75 litres per minute of pulp ex-
tracted from a position within the process stream .where there is sufficient
turbulence to ensure a representative sample. Figure 14.9 shows a diagram
808
PROCESS CONTROL
Figure 14.9. Armco Autometrics PSM - sampling system.
of a typical sampling box. It is possible to use the same instrument for more
than one stream by multiplexing the sample extraction step. However, the
time taken for the air eliminator and electronics to process and stabilize a
measurement is lengthy. To sample one stream, switch to a second and then
return to the first takes about 18 minutes.
The instrument is initially pre-calibrated by the manufac;turer and then
calibrated on site by taking samples from the discharge pipe of the PSM and
relating the readings of the instrument at the time of taking the samples to
a screen analysis and pulp w/ s ratio measured later in the laboratory. The
manufacturers claim an accuracy of± 1,5% to 2% of actual cumulative per
cent passing a given mesh size and ± 2 % on the w/ s ratio.
There are 27 operating Armco Autometrics PSMs in South Africa in
the gold and platinum mines. They are particularly important in mill con-
trol strategies.
The Leeds and Northrup Microtrac Analyser
The Leeds and Northrup Microtrac measures the scatter from a laser beam
projected through a liquid stream (Figure 14.10). Scatter angle is a function
of particle size. Small particles scatter light at larger angles than large par-
ticles. The light scattered by the particles is passed through an optical filter,
and its intensity and angle is detected and transmitted to a microprocessor
which calculates the size and the quantity distribution of particles in the li-
quid. In operation, a sample of pulp is introduced into the instrument where
it is diluted with water and then recirculated through the measurement cell
where it is illuminated by the laser beam.
809
Laser beam
Figure 14.10. L & N Microtrac optical measurement system.
The calibration of this instrument is fixed by the optical filter design

and the microprocessor software. There is automatic background compen-
sation provided by a set zero control. In operation, continuous sub-samples
of about 150 litres per minute are taken from the cyclone overflow using
fixed launder or pipe samplers. These samples are screened to remove wood
fibre and the pulp is passed through a 40 mm pipe from which an Isolok
valve takes as many 3 ml samples as required. The Microtrac is able to give
a complete particle size distribution every two minutes.
The Microtrac has been used in South Africa mainly as an off-line in-
strument but it is planned to use it in mill control strategies. By suitable design
of the sampling system, it is possible to use a single Microtrac to measure
the particle size distribution from several mills. The penalty would be a longer
delay time between measurements on an individual mill. Whether this can
be tolerated depends on the particular control strategy being employed.
14.4.5.6 Basic chemical measurements

A number of fundamental chemical measurements are required for the con-
trol of an ore treatment plant. Examples include the measurement of pH,
oxidation-reduction potential, the degree of alkalinity of a pulp, the oxygen
content and the cyanide concentration. The usual laboratory methods
employed for determining these quantities involve removing a sample of the
pulp, filtering it to separate the solids and then applying normal analytical
chemical methods for the particular determination. For example, cyanide
concentration in solution is determined by adding potassium iodide to a
known volume of clear solution and then titrating with silver nitrate. These
methods can be obtained from Lenahan and Murray-Smith (1986).
810
PROCESS CONTROL
Standard industrial instruments are available for some measurements

such as pH and oxidation-reduction potential (Considine, 1985). There have
also been recent developments in automating certain of the traditional pro-
cedures to obtain rapid on-line measurements. Some of these are dealt with
below:
The On-Line Gold Analyser (OLGA)
The first on-line gold analyser, known as the Tell-tale, was developed by Cor-
ner House Laboratories during the period 1965 - 1969. This design is still
successfully used in a number of plants.
The AAC OLGA Mark III Gold Analyser incorporates many features
of the original Tell-tale and is designed specifically to determine gold con-
centration in the barren solutions resulting from the zinc precipitation and
filtration step in the production of gold by the Crowe-Merrill process (Brandt
and co-workers, 1980).
The analyser system is situated in a small room centrally sited within
the filter section of the plant. A maximum of 9 process streams in addition
to the composite are continuously sampled from the filter plant at a rate of
approximately 1 litre per minute. Each process stream flows into a constant-
head sampling vessel that provides an analyser feed at a rate of 50 ml per
minute. The analyser feed-lines from the sampling vessel are furnished with
control valves that are actuated in a programmed sequence by a
microcomputer.
The basis of the automatic analysis technique is preconcentration of the
trace levels of gold in the barren solution by continuous solvent extraction,
followed by flame atomic absorption spectrophotometric (AAS) determina-
tion of the gold concentration in the organic extract. The solvent extractant
employed is a solution of 1% Aliquat 336 (a proprietary extractant) in di-
isobutylketone (DIBK). The continuously extracted gold, preconcentrated
40 times, is collected and automatically analysed by the AAS spectrophoto-
meter.
The instrument can measure gold in the concentration range 0,005 to
0,500 mg/l. In operation, the gold concentration in the composite solution
is measured every three minutes. If the concentration exceeds a preset level,
an alarm is sounded and the instrument then measures the effluent from each
of the filters until the individual filter that is malfunctioning is identified.
Figure 14.11 shows the OLGA schematically.
Automatic on-line analysis using AAS with electrothermal atomization
As an alternative to continuous solvent extraction, the use of AAS coupled
with electrothermal atomization has been developed for measuring gold in
solution. The basis for this method is the graphite tube atomizer, on to which
between 5 and 50 microlitres of solution are injected automatically. The fur-
nace goes through a heating cycle, controlled by a microprocessor, ending
up with a temperature in excess of 2 000°C to atomize the gold for deter-
mination with the AAS.
For automatic on-line analysis, solutions from the process are delivered
continuously to the room where the AAS is sited. Sub-samples are extracted
811
-
00
N
~l~ 1%
standard EDTA
Ace-
tone
DIBK
Di st.
water >-3
::r:
tI1
tI1
>::
>-3
~
n
>-3
......
~
Printer ~
Sample streams
from plant
>
t"""
t"""
c:
~ ~ ~
?;::!
Alarm
a
>-<:
l Constant
head AA spectro- 0
sampling photometer 'T1
Microprocessor
vessels a
0
~ control unit t"""
tl
Gas control
unit
Q)
i::
Q)
>.
';:)
u
<t:
Compressor
Figure 14.11. Schematic diagram of AAC Mark III on-line gold analyser.
PROCESS CONTROL
arm
Wash cup
Figure 14.12. Sampling dispenser.
from the main sample lines and fed to a position where the sample dispenser
can extract an appropriate aliquot. The sample dispenser feeds samples of
the solution to be analysed together with modifiers required for the AAS
analysis to the graphite tube atomizer. The dispenser is also programmed
to take standard solutions for calibration of the equipment. Figure 14.12 1s
a diagram of a typical sample dispensing arrangement as used by Robert and
Ormrod (1985).
Mintek have developed this technique for measuring gold in solution
in CIP plants. Solutions are extracted from the absorption tanks using a filter
tube immersed in the tank. Solution permeates through the filter tube under
the hydrostatic head between the surface of the slurry and the tube which
is immersed to a depth of about 1 metre (Figure 14.13). On a timed sequence,
the probe is pressurized by nitrogen and a sample is transported to the analysis
room for analysis on the AAS as described above.
Alternatively, sampling probes have been used in which the solution is
drawn into the filter probe by vacuum and then transported to the AAS room
by compressed air. This technique has been incorporated in an electrother-
mal atomization AAS system, developed by the J.C. I. Minerals Processing
Research Laboratory, for the on-line monitoring of soluble gold loss from
a conventional filter plant. A microcomputer controls the filtration opera-
tion and the sequencing of streams being monitored. It also performs the
regressions, prints out results and triggers alarms when selected concentra-
tion levels are exceeded. This AAS system has been further developed by
J.C.I. for the CIP process to cover the range 0,002 ppm to 70 ppm gold in
solution. This is achieved by exploiting the features offered by the furnace
system used. Eight furnace programmes and three calibration curves are us-
ed to obtain the dynamic range, all of which are selected by the microcom-
puter. The most suitable curve is automatically selected, depending upon the
absorbence obtained for the solution being measured. This system monitors
each absorption tank and, by using the higher calibration range, informa-
tion can also be obtained on parts of the elution cycle.
813
r.i+--- Sample-transpprt
tube
,.___ _ Porous filter
--111-----Filtrate
Figure 14.13. Sampling probe.
Automatic spectrophotometric analysis

There are on the market instruments which allow for continuous colorimetric
determinations. With these instruments, a peristaltic pump takes up a
measured volume of a solution for analysis which can then be mixed with
known volumes of reagents taken with the same peristaltic metering pump
but with different tubes. In this way coloured solutions are produced and
these are passed through an optical measurement cell through which light
passes. The absorption of the light gives a measure of the concentration of
the particular element in the original solution. This technique has found wide
application in many industries and has lately been applied to the metallurgical
industry.
One particular application is the Gold Fields cyanide monitor. In this
instrument a sample of solution is extracted from a leach tank using a filter
tube sample extractor. This consists of a permeable filter tube about 500 mm
long immersed in the pulp. A vacuum is applied inside the filter tube to draw
in solution. After about 5 minutes, the vacuum is removed and compressed
air fed into the tube to transport the sample to the analyser room and to
blow the cake off the filter tube. The solution samples are piped to the out-
side of the analysis room and sub-samples led through the wall into overflow
sampling pots.
The auto-analyser is programmed to take an aliquot from the sub-
sampling pots, and then to take measured volumes of copper sulphate and
EDTA to mix with the cyanide solution. Cyanide reacts with copper sulphate
to produce a colourless solution. However, copper sulphate is always added
814
PROCESS CONTROL
in excess and therefore cyanide concentration is inversely related to the cop-

per sulphate concentration which is determined as the solution flows through
the optical cell.
On-line analysis of cyanide in solution
Automatic cyanide monitors have been developed for use in the control of
cyanide addition to leach circuits. These determine the free cyanide content
of leach pulps by automatic titrimetric analysis. A solution sample is extracted
from the leach tank by means of a vacuum filter tube. In one instrument
(the AAC Cycad monitor), the sample is then blown by compressed air to
a cell fitted with a combination gold-thallium electrode and titrated with 0, 1
N silver nitrate until an endpoint is reached at 700- 750 mV. An output in
the range 4- 20 mA is then produced. The filtration/titration cycle takes
approximately 3 minutes. A rinse cycle is performed periodically. This uses
5 OJo sulphamic acid to remove calcium carbonate and sulphate scaling.
Another instrument which gives an adequate measurement for control
purposes transfers the sample by means of a metering pump. This sample
is mixed with a metered equivalent of standard silver nitrate and the mixture
passes through a cell fitted with a gold-thallium electrode so that a continuous-
ly varying e.m.f. is produced corresponding to the sensitive section of the
titration curve. Suitable signal processing then gives an output in the range
4-20 mA.
Lime and pH measurement
A logical extension to the above measurement technique is the sequential
determination of both free cyanide and protective alkalinity on the same
measured volume of pulp filtrate. Work on this extension is in progress. After
the alkalinity due to cyanide has been neutralized by the addition of one
equivalent of silver nitrate, the remaining, so-called protective alkalinity is
determined by titration with 0, 1 to 0,5 N sulphuric acid, using pH electrodes
to achieve an endpoint at pH 8,3.
14.4.5.7 Carbon and resin concentration meter

This device being developed at Mintek is based on the phenomenon that
ultrasonic waves are attenuated by particulate matter as described for the
Armco Autometrics particle size monitor. However, this device uses a robustly
constructed probe which can be lowered into the pulp. By transmitting and
receiving ultrasound at two frequencies, tht; concentration of carbon or resin
can be measured against the constant concentration of the pulp. Normally
the attenuation of the lower wavelength is constant and due to the solid par-
ticles of the pulp, whereas attenuation of the high frequencies is due to the
resin or the carbon and varies with their concentration. The probes use
piezoelectric ceramic transducers encapsulated in epoxy resin. Figure 14.14
shows the plan of a typical probe.
14.4.5.8 Carbon activity meter

Different batches of activated carbon, whether new or reactivated, can have
different gold-loading capacities. The activity can be determined simply by
815
92 kHz Peak
burst VT!
/ Transmitter 1 detector
generator
Receiver 2
burst
generator
)))mill)))) I Amp
Peak
detector
---~
VR2
/~----1 Peak
detector
YR!
Receiver 1
Peak ~-------------. VT2

detector
Figure 14.14. The Mintek carbon and resin concentration meter.
the contacting of a known amount of carbon with a known concentration

of gold in the presence of sodium hydroxide and sodium cyanide (simulated
plant conditions) and the periodic removal of an aliquot portion of the solu-
tion for the determination of its gold content by atomic absorption spec-
trometry. The rate of the gold absorption and the final gold capacity can
be obtained from a plot of the gold concentration remaining in solution ver-
sus time. This approach yields excellent results and is straightforward, but
does require laboratory facilities such as an atomic absorption spec-
trophotometer and, more important still, a person to perform the analysis
over a period of 3 to 4 hours. However, this approach is not favoured on
most plants because it means that the sample must be removed from the reac-
tivation kiln and transported to a laboratory for assay, all of which can lead
to long delays before the results of the analysis are obtained.
A small and relatively simple instrument has been developed at Mintek.
This instrument, which is microprocessor-based and extremely easily
operated, is designed for use on the plant in close proximity to the reactiva-
tion kilns and is based on the electro-chemical technique known as anodic-
stripping voltammetry.
A 500 ml portion of a stock solution containing sodium hydroxide (0,01
mol/ dm3) sodium cyanide (0,02 mol/ dm3), gold (100 µgl g) and a small
amount of either lead or thallium nitrate (1 to 4 µg/ g) is added to a beaker.
The electrode probe of the instrument, which consists essentially of a platinum
electrode and a gold reference electrode, is positioned, and the stirrer is set
in motion.
816
PROCESS CONTROL
When the readings on the analyser have stabilized, 1 g of carbon is add-

ed and the instrument is allowed to run automatically. The gold concentra-
tion is measured every 3 minutes and plotted on a chart recorder. Operation
of the instrument requires no special expertise; the operator needs to carry
out only a few mechanical operations, e.g. measuring out a volume of stock
gold solution with a measuring cylinder, weighing 1 g of carbon sample and
pressing a button on the instrument to initiate the measuring process.
Besides its main functions of measuring the activity of reactivated car-
bon, the instrument can also be used to measure the activity of carbon samples
removed from various absorption stages in the plant, which is of assistance
in studies of possible poisoning effects during absorption.
14.5 Actuators and Final Control Elements

14.5.1 Introduction
Successful process control requires that all key process variables can be chang-
ed conveniently. The term controllability is often used in this context. To
obtain a process that is controllable, it is first necessary to have a sufficient
number of actuators or final control elements. Also, actuators should be
selected so that the process variables can be changed over a sufficient range
with good resolution. The actuator characteristic should be such that the gain
does not change very much over the whole operating range. A common prob-
lem in flow systems, for example, is that the control valve is oversized, giving
a very non-linear relationship between valve opening and flow. The actuators
used in ore treatment plants fall into three main categories: control valves,
material feeders and variable speed drives.
14.5.2 Control valves

The main types encountered in ore treatment plants are globe valves, ball
valves, diaphragm valves and pinch valves as shown in Figure 14.15. See Con-
sidine (1985) for details of other types of control valves. Many valves in-
stalled in plants are operated manually, but for process control we are con-
cerned with those versions which can be actuated by pneumatic, hydraulic
or electrical means.
Globe and ball valves are suitable only for use on clean or slightly dirty
fluids. Small ball or needle valves are often used for reagent flow control
duties. Diaphragm valves are the traditional means of manipulating the flow
rate of pulp in a pipe line. They have the advantage of simplicity and
robustness, but the disadvantage of having a large mass and a high actuator
power requirement. Pinch valves consist of a length of rubber pipe which
is pinched between bars or, preferably, by a pneumatic or hydraulic sphinc-
ter. The rubber pipe has significant hysteresis, and a valve that provides a
pull action as well as a pinch action may be required to overcome this. Where
the flbw rate of an exceptionally abrasive or viscous substance must be
manipulated, a proportioning diverter valve operated in on/ off mode by a
variable mark-to-space ratio pulse signal is often the only solution.
Various linear and non-linear flow versus valve stem position
817
Diaphragm Needle Pinch
Split-body globe 3-way globe Angle-body globe
Figure 14.15. Typical control valves.
characteristics are available (Hutchinson and Merwick, 1976), and manufac-

turers' catalogues should be consulted for details. It should be borne in mind
that such characteristics apply for a constant pressure drop across the valve.
The characteristic for the valve installed in a line is quite different. For
automatic control applications, the valve characteristic should be chosen so
that the overall system flow versus valve stem position characteristic is ap-
proximately linear.
Actuators for control valves are of three main types, namely, electrical,
pneumatic, and electro-hydraulic. Pneumatic actuators have remained the
most popular type over many years because they offer the combined advan-
tages of a high power to mass ratio, simplicity, low cost and a well-defined
fail-safe action. The electrical output signal from the control centre is initially
processed by a current-to-air-pressure transducer which can be either separate
or integral with the actuator. Clean compressed air is supplied to this
transducer at pressures of up to 700 kPa, and the actuator action on air supply
failure can be either to open or to shut the valve, depending on design.
Actuators are available to operate valves in both infinitely variable and fully
open/fully shut modes. Valve positioners which operate as local position feed-
back controllers can be attached to pneumatic actuators. These are used when
the valve is the final control element in a critical control loop and a strictly
linear relationship between electrical current and valve position is required.
Electrical valve actuators are available but are usually powered from
the main electricity supply and therefore suffer from the problem that in the
event of supply failure, the valve stays put. Electro-hydraulic actuators can
overcome this problem by provision of a reserve accumulator to permit a
definite failure action. These have the advantage of not requiring an air supply
but are relatively expensive due to the high-precision construction required.
818
PROCESS CONTROL
14.5.3 Material feeders

The main types used for automatic process control are vibrating feeders and
belt feeders. The vibrating feeder is a flexible device and can be used for
feeding material ranging from fine powder to run-of-mine rock. It has the
advantages of high throughput capability, low cost and low maintenance re-
quirements. Its chief disadvantage, however, is that the actual mass feed rate
varies erratically when the material changes through the range from wet fines
to dry rock. Attempts to overcome this problem by measuring the actual mass
feed rate with a belt weigher and using this as a feedback signal to control
the vibrating feeder have not been successful, probably due to the transport
delay caused by the distance between the feeder and the belt weigher.
Belt feeders are more expensive and require more maintenance than
vibrating feeders, but have the advantage that a belt weighing action can be
directly incorporated to give a so-called weighfeeder. These give a precise
measurement and control of material mass feed rates.
14.5.4 Variable-speed drives

The two main types encountered use hydraulic control and solid-state elec-
tronic control. Hydraulic variable-speed drives are based on a variable fluid
coupling which is placed between a conventional a.c. induction motor and
its load. These are reliable, robust and are available in a range of power
ratings. Versions specially designed for use with pumps are also available.
Electrical variable-speed drives can be subdivided into a.c., d.c. and eddy
current types. A.C. drives consist of a solid-state rectifier/inverter which con-
trols the frequency and voltage supplied to a specially rated a.c. induction
motor. These are very reliable, but can be more expensive than other types.
D.C. drives have simpler electronics than a.c. drives, but require a d.c.
motor. The brushes in such motors have specialized maintenance
requirements.
Eddy·current drives comprise a variable magnetic coupling which is plac-
ed between a conventional a.c. induction motor and its load. These have very
good reliability and are easy to maintain.
14.6 Process Control System Design Philosophies and Architectures

14.6.1 Introduction
Preceding sections of this chapter have discussed certain important com-
ponents of process monitoring and control systems. In this section, we ex-
amine the issue of integrating these components into overall working systems.
One of the most basic tasks of all process control systems is to gather
data from the various process instrumentation subsystems and to present this
data in a suitable form to the process operating personnel o·r, in other words,
to provide the interface between the process and the operator. Various alarm
monitoring and reporting tasks together with the task of implementing
automatic safety shut-down logic are also usual basic requirements. Closely
connected with basic data presentation are the tasks of storing and subse-
quently processing the data and presenting information for process study,
819
evaluation and problem investigation purposes. There is a trend towards pro-

viding sophisticated process study, development and management tools in
modern plants. Another important task of process monitoring and control
systems is to implement the various kinds of real-time automatic control refer-
red to earlier in this chapter such as single-variable or multi-variable con-
tinuous control, supervisory control, process optimization and logic or se-
quencing control.
Although the above tasks are relatively standard, there are a number
of different ways in which they can be implemented, based on different pro-
cess operating philosophies and to a certain extent, process design. There
are also a number of different technologies available for implementing pro-
cess control functions.
Some of the most important issues relating to process control system
architectures and design philosophies will be discussed under the subsequent
headings. Considine (1985) provides additional information.
14.6.2 Manual control versus fully automatic control

Manual control is one major process operating philosophy. This is not, as
is sometimes mistakenly supposed, inherent in process control systems im-
plemented with the older technologies. Some of the very latest systems bas-
ed on digital computer technology with sophisticated colour video displays
are designed to support essentially a manual control philosophy.
The basic idea in this approach is to provide the process operator with
a good dynamic display or other means of visualizing the state of the pro-
cess, but then to leave the manipulation of many of the most important pro-
cess control inputs to the operator. The process control system would also
be designed to provide him with convenient means of making the necessary
adjustments. In other words, the human operator is retained as an essential
element of most feedback and feedforward loops. A recently proposed varia-
tion on this idea is that the process monitoring system should calculate and
display some meaningful measure of economic performance for the entire
process. This, in theory, then enables the human operator to use his ingenuity
to find the most profitable operating condition for the plant.
Perhaps the major advantage of manual control is that the design of
the process control system is relatively simple. It can also be argued that there
are certain tasks where it will always be virtually impossible to better manual
performance. On the other hand, a major disadvantage is that the use of
manual control where a simple automatic device could do the job is a gross
misuse of the human resource. Human performance is especially bad when
simple tasks requiring concentration over extended time periods have to be
performed. Human concentration is a limited resource and should only be
called on in short bursts.
At the opposite end of the spectrum one finds the fully automatic pro-
cess control systems. Here the philosophy is that whenever an adjustment
based on a measurement needs to be made to a control input, this should
be done automatically by the process control system without requiring human
intervention. This particular philosophy can also extend upwards to the super-
820
PROCESS CONTROL
Process unit 1 Process unit 2

==::i
Key: P Protection logic system

M Monitoring system
C Controller
Figure 14.16. Process control using hard-wired analogue
equipment.
visory control of overall process unit operation, the coordination of the opera-
tion of the various units and overall plant economic optimization. Implemen-
tation of fully automatic process control has only really become feasible due
to the flexibility of the newer programmable computer-based process con-
trol technologies, although it is certainly possible to implement a large number
of the individual tasks required using hard-wired analogue process control
technologies.
Most present-day mineral processing plant control systems fall
somewhere between the above two extremes, with a steady trend towards
the use of fully automatic control.
14.6.3 Hard-wired analogue control systems

Process control equipment that was available commercially during the period
from the early 1960's to the mid 1970's utilized analogue electronic or occas-
ionally pneumatic technology. The form of control system architecture that
can be readily implemented using this kind of equipment is shown in Figure
14.16. Signals from field transmitters are individually wired to a central con-
trol room where various items of equipment such as measurement read-out
displays, recorders, controllers, alarm monitoring systems and safety shut-
down logic systems are located. These are often mounted on or behind a
large control panel to give a degree of centralization of information display
in order to aid the operator in visualizing the state of the process.
An important property of such systems is that although the various func-
tions are physically centralized in the control room, there is a logical separa-
tion or distribution of the various process control tasks among individual
standard modules such as process controllers and shut-down logic modules.
This gives these systems their major advantage of simplicity and a degree
of fault tolerance. If one controller fails, for example, it does not necessari-
ly affect the operation of the rest of the process control system seriously.
The disadvantages of hard-wired analogue systems are, firstly, that it
is difficult to achieve a properly coordinated and centralized display of the
state of the process. The various measurement read-outs are usually spread
over a fairly large section of the control panel and the operator has to move
821
Process
computer
Process unit 1
Figure 14.17. Process control using centralized process control

computer.
from one read-out to another in order to build up a mental picture of the
operation of the plant. Secondly, with these systems, flexibility to implement
advanced control functions for process study, development and management
information tools and to modify and improve existing functions as the needs
of the process evolve is severely limited. A third disadvantage is that the cost
of such equipment is relatively high.
In spite of these disadvantages, hard-wired analogue control systems have
a good reputation for reliability and remain the most appropriate choice for
many present-day applications.
14.6.4 Centralized computer control systems

The development of relatively cheap and powerful minicomputers during the
early 1970s meant that by about 1975 it had become feasible and attractive
to incorporate such minicomputers into process control systems. The benefits
of doing this were that the vast data manipulation and storage capabilities
of the digital computer could be used to collate and present data and deriv-
ed information to the process operator in a far more meaningful and power-
ful way than had previously been possible. Historical process operating data
could also be stored, retrieved on demand and processed into management
reports and information concerning process dynamic behaviour. In addition
to this, many of the previously hard-wired control tasks could be program-
med, thereby giving far greater flexibility and the opportunity to implement
more integrated and advanced forms of control. Details of how the controllers
discussed in Section 14.3 can be implemented in digital computer form are
available in the texts by Astrom and Wittenmark (1984) and Bennett and
Linkens (1984).
The way that these early computer process control systems are arrang-
ed is illustrated in Figure 14.17. Because of limitations of the early computing
hardware and cost considerations, individual signals from all the field
transmitters are wired into one central computer. This computer carries out
all of the process control tasks including information display, continuous
control, alarm monitoring and safety shut-down logic functions. These
systems certainly allow the implementation of more integrated and co-
ordinated information display and control strategies, but suffer from the
major problems of very high complexity and poor reliability, particularly
822
PROCESS CONTROL
r------,
Back-up
1
process 11
L_ _:~m!~e:_j
Process unit 1 Process unit 2
Figure 14.18. Process control using centralized process control

computer with distributed PLCs.
in connection with the programs or software. There has consequently been
a trend in more recent computer control systems to return to a greater distribu-
tion of the process control tasks among simpler separate hardware modules.
In modern systems these separate modules themselves would typically
be computer-based. Figure 14.18 shows the structure of an example of
this kind of system which uses a number of programmable logic controllers
(PLCs) connected to a minicomputer in a star arrangement. The PLC modules
would perform scanning and pre-processing of the sensor inputs and also
possibly logic control and PID control tasks. The minicomputer would
typically provide information display, historical data logging, report genera-
tion and process study tools.
14.6.5 Distributed computer control systems

The problems experienced with earlier centralized computer control systems
based on a single minicomputer and the increasing availability of cheap digital
computing power in the form of the microprocessor have spurred on the
development of a new generation of digital computer-based process control
systems since the late 1970's. Known as distributed computer control systems,
these attempt to combine the advantages of hard-wired analogue process con-
trol systems and centralized computer control systems without incurring the
disadvantages of either. The basic idea is that a distributed computer con-
trol system consists of a collection of fairly autonomous digital processing
modules which communicate with one another over a shared high-speed
digital communication channel or data highway.
There is no master module that directs overall system operation in a truly
distributed system. System co-ordination relies instead on co-operation among
the autonomous processing modules. This eliminates the possibility of total
system failure due to the failure of any one processing module.
The various tasks of process control are distributed among these separate
modules which may also be physically distributed along the data highway
as shown in Figure 14.19. This retains the advantages of simplicity and fault
tolerance through the use of simple standard modules and also gives a poten-
tial reduction in wiring costs. On the other hand, the data collation and pro-
823
Shared
(data highway
Figure 14.19. Distributed computer control system using five

autonomous processing modules. Note that processing modules
3 and 4 operate in active redundancy to give fault tolerance.
cessing capabilities of the digital computer are retained, allowing powerful.

,centralized information display and operator interface capabilities and the
implementation of advanced ·a:nd flexible forms of control.
14.6.6 Examples of current practice

This section discusses three examples of recently installed process control
systems. Figure 14.20 shows the overall structure of the control system for
Operators' terminals & printers
Signal marshalling racks
Process
Figure 14.20. Structure of control system for East Driefontein

gold plant.
824
PROCESS CONTROL
Figure 14.21. East Driefontein gold plant control room.
the East Driefontein gold plant. The system includes two process computers
and two PLCs. Although similar to the basic centralized computer control
concept discussed in Section 14.6.4, this system actually incorporates a greater
degree of task distribution. The tasks performed by the various sub-systems
are described in the following paragraphs.
The PLC monitors all measured plant variables and controls all major
plant motors. It carries out alarm checking, logic control for interlocking
of equipment, on/off control according to measured values and mill pebble
feed control according to a special program. It also produces reports once
per shift. A back-up PLC is available to take over in the event of failure
of the primary PLC.
The ELCON supervisory system provides colour graphic displays of the
process which can include dynamically up-dated indications of motor states,
measured variables and the status of alarms. It also provides displays of con-
trol loop details, plots of measured variable trends and hourly, once per shift,
daily, weekly, monthly and yearly reports.
The PROSCON system shown in Figure 14.20 is a process computer
which is dedicated to the task of multi-variable control of the three milling
circuits. This is an implementation of the multi-variable controller referred
to in Section 14.3 .3.
The control room of the East Driefontein plant is shown in Figure 14.21.
The operator's coiour video terminals can be seen in the foreground. This
is typical of the kind of operator interface found in many modern systems.
The old control panel, showing a mimic diagram of the plant, can be seen
in the background.
Figure 14.22 shows the overall structure of the control system for the
Cooke gold plant of Randfontein Estates. This is a PROSCON proprietary
825
J!L,
Operators' terminals
& printers
High speed
serial links
Milling Gold Uranium Services

and recovery recovery
thickening
PM - Process management computer

MC - Measurement and control computer
Figure 14.22. Structure of the control system for the Cooke gold
plant.
5 Operator terminals
and 3 printers
Back-up
CYGNUS
computer
High speed
serial links
Process
Figure 14.23. Structure of control system for Vaal Reefs No. 9

Shaft gold plant.
826
PROCESS CONTROL
system and includes six process computers. Four of the computers interface
directly with the process instrumentation and actuators and have relatively
low processing power. These perform averaging and processing of measured
variables, alarm checking, PID control and logic control. The remaining two
computers are much more powerful and provide colour graphic displays for
the plant operators, data collection and storage functions, reporting, event
recording and a process study package. Note the regular structure of the
system. This is typical of systems supplied by a single manufacturer. Again
there is a moderate degree of task distribution in this system and redundan-
cy to provide fault tolerance.
The process study package provides comprehensive facilities for interac-
tively manipulating the stored historical data. There are data preparation
facilities (plotting, selection and rejection, manual input, mathematical func-
tions and filtering), statistical analysis facilities (histograms, autocorrelation,
crosscorrelation and power spectra) and static and dynamic model identifica-
tion facilities. In addition, there is a facility for designing multi-variable
controllers.
The CYGNUS process control system installed at No. 9 Shaft gold plant
of Vaal Reefs is shown in Figure 14.23. Note the close correspondence with
the structure shown in Figure 14.18. The eight PLCs each interface with a
section of the plant. The division into sections is as follows: ore handling,
Mill 1, Mill 2, thickeners, leaching, CIP A, CIP B, and residue. The func-
tions performed by the centralized supervisory computer are as for the PRO-
SCON system discussed above. An interesting feature of the CYGNUS system
is the special keyboard shown in Figure 14.24. This allows very easy access
to comprehensive system functions by the plant operator.
An interesting aspect of the Vaal Reefs No. 9 Shaft system is that a KEN-
W ALT process simulator was used for testing of the process control system
and for operator training during plant commissioning. A special-purpose com-
puter which runs the simulator is connected to the instrumentation wiring
in place of the real plant instruments to allow this. The KENW ALT simulator
can also be used during design and provides the following:
• A facility for the plant designer to optimize flowsheets by allowing the
quick evaluation of the effects of a large number of plant variables.
• A facility for sizing equipment and designing a maintenance schedule
to minimize the effects of equipment down-time on plant performance.
• A facility for the control system designer to design and pre-tune the
control loops using the dynamic models inherent in the simulated pro-
cess flowsheet. Any problematic loops which may entail plant redesign
will become apparent at this stage.
e A facility for designing and testing the stop/ start sequence logic.
This illustrates the opportunities for optimizing the interaction between
process design and control system design that are now available.
827
00 >-3
OVERVIEW
N No 9 SHAFT GOLD PLANT
G
00 ::r::
( WATE~]
trJ
11111111111111111:i1111111
ELECTROWINNING trJ
111111111111111111111111111
><
• I ·.·I
88
>-3
~
D-r_•n
THICKENING
n
•
D MILL 1
•
r:j I l
•
>-3
.....,
~
DO MILL 2
~ REGEN $::
trJ
~
t-'
t-'
8?i)'
- c:;o
~
0
"TI
0
0
t-'
u
08~)00808008 000
0000000080 000
00~)08008000 80GJ
(0 0 0)8 008
Figure 14.24. Example of special keyboard used in CYGNUS process control system.
PROCESS CONTROL
14.7 References
Astrom, K.J. and Wittenmark, B. (1984). Computer Controlled Systems. Prentice-Hall,
Englewood Cliffs.
Bennett, S. and Linkens, D.A. (1984). Real-Time Computer Control. Peter P,eregrinus, London.
Bentley, J.P. (1983). Principles of Measurement Systems. Longman, London.
Brandt, P .J ., van Dalen, J.H. and Wessels, F.W. (1980). The automatic determination of trace
amounts of gold in gold-plant barren solutions. J. S. Afr. Inst. Min. Metal!., 80, 197 - 203.
Chang, J.W. and Bruno, S.J. (1983). Process control systems in the mining industry. World
Mining, 36 (5), 37 -41.
Considine, D.M. (1985). Process Instruments and Controls Handbook, 3rd edn. McGraw-Hill,
New York.
Doebelin, E.O. (1983). Measurement Systems, 3rd edn. McGraw-Hill, New York.
Dorf, R.C. (1986). Modern Control Systems, 4th edn. Addison Wesley, Reading.
Fertik, H.A. (1975). Tuning controllers for noisy processes. Trans. Instrument Society of America,
14, 292- 304.
Fletcher, G.W. (1972). Process Control. In Gold Metallurgy in South Africa, edit. R.J. Adam-
son. Chamber of Mines of South Africa, Johannesburg.
Hulbert, D.G. and Braae, M. (1981). Multivariable control of a milling circuit at East Driefon-
tein gold mine. Mintek Report No. 2113, Randburg.
Hulbert, D.G. (1983). Multivariable control and optimization of a milling circuit at East Driefon-
tein gold min.e. Mintek Report No. M98, Randburg.
Hutchinson, J.W. and Merwick, A.R. (1976). J.S.A. Handbook of Control Valves, 2nd edn.
Instrument Society of America, Pittsburgh.
ISO 5167 (1980). Measurement of Fluid Flow by Means of Orifice Plates, Nozzles and Venturi
Tubes Inserted in Circular Cross-Section Conduits Running Full. International Organization
for Standardization, Geneva.
Lenahan, W.C. and Murray-Smith, R. (1986). Assay and Analytical Practice in the South African
Mining Industry. Monograph Series M6. S. Afr. Inst. Min. Metall., Johannesburg.
Lynch, A.J. (1977). Mineral Crushing and Grinding Circuits. Elsevier, Amsterdam.
Owens, D.H. (1978). Feedback and Multivariable Systems. Peter Peregrinus, London.
Pauw, O.G., King, R.P., Garner, I(.C. and van Aswegen, P .C. (1985). The control of pebble
mills at Buffelsfontein Gold Mine by use of a multivariable peak-seeking controller. J. S. Afr.
Inst. Min. Metal!., 85, 89- 96.
Purvis, J.R. and Erickson, I. (1982). Financial models for justifying computer systems. Intech,
29 (11), 45 - 50.
Robert, R.V.D. and Ormrod, G.T.W. (1985). Automatic on-line analysis with atomic absorp-
tion spectrophotometry. Proc. MINTEK 50 Conference, Randburg.
Shinskey, F.G. (1979). Process Control Systems, 2nd edn. McGraw-Hill, New York.
Smith, C.A. and Corripio, A.B. (1985). Principles and Practice of Automatic Process Control.
Wiley, New York.
Stephanopoulos, G. (1984). Chemical Process Control. Prentice-Hall, Englewood Cliffs.
Ziegler, J.G. and Nichols, N.B. (1942). Optimum settings for automatic controllers. Trans.
A.S.M.E., 64, 759-768.
829
Chapter 15
The Chemistry of the Extraction of Gold

M.J. Nicol, C.A. Fleming and R.L. Paul
15.1 General Principles

15.1.1 The chemistry of gold compounds
Gold is the most noble of all the metals and this .is the key to both its eternal
romance and its many practical uses in industry. It is the only metal, for
example, that is not attacked in air or water by either oxygen or sulphur,
and its durability under the most corrosive conditions has led to its widespread
use in coinage and jewellery through the ages.
Gold is the only metal that is generally found in nature in the metallic
state, and the only gold compounds that occur in a natural state are the
telluride and stibnite species, AuTe2 and AuSb2 • Gold reacts with all the
halogens, most vigorously with bromine, which undergoes an exothermic reac-
tion with gold powder at room temperature to produce Au2 Br6 •
Of greater importance to the extractive metallurgist is the chemistry of
gold complexes in aqueous solutions, and this is a far broader subject. It
has been traditionally accepted that gold complexes in aqueous solution can
exist in one of two oxidation states, the aurous ( + 1) or the auric ( + 3), and
that all the gold complexes of hydrometallurgical interest fall into these two
groups. In more recent years, compounds with formal oxidation states of
- 1, + 2, + 4, and + 5 have been prepared and identified (Puddephatt, 1978;
Schutte, 1985), but these compounds are more of an academic curiosity at
present and will not be dealt with in this review.
Thermodynamics predicts that neither aurous nor auric cations will be
stable in aqueous solution, but will be reduced by water to metallic gold (see
Section 15.1.2). In order to stabilize these ions in aqueous solutions, it is
necessary to introduce a complexing ligand and to employ a suitable oxidiz-
ing agent, since no reaction occurs unless both are present in the solution.
For example, gold does not dissolve in either nitric acid (an oxidizing agent)
or hydrochloric acid (a complexing ligand), but dissolves quite easily in a
mixture of the two, aqua regia. Whether they are solid, in solution, or in
the gaseous state, gold compounds are invariably bound covalently, and by
far the majority are complexes.*
*NOMENCLATURE: A complex consists of a central cation with a number of ligands

bonded to it. Ligands may be ions (e.g. Cl-, SCN-) or uncharged molecules (e.g.
NH3 , H2 0, (NH2) 2 CS). The number of ligands bonded to the cation is known as the
co-ordination number of the cation. The stability constant {3 11 of a complex formed
831
A glance at any chemistry text sho~s that there are a large number of
gold complexes with a wide range of stabilities. Generalizations can be made
to show that the properties of these complexes vary systematically. These
allow many of the known thermodynamic characteristics of gold complexes
to be rationalized, and provide some basis for the prediction of their
behaviour.
A useful first generalization about the stability of gold complexes relates
to the nature of the ligand donor atom, i.e. the atom in the ligand that is
bonded directly to the gold. Both gold(!) and gold(III) are B-type metal ions,
which means that the stability of their complexes tends to decrease as the
electronegativity of the ligand donor atom increases (i.e. as the tendency of
the atom to attract electrons increases). This leads to stability orders such
as r > Br- > Cl- > F-, where the order of electronegativities is F- >
Cl- > Br- > r. More generally, it is apparent that electronegativity is a
periodic property of the elements, and for elements that would possibly form
complexes with gold, there is the following scheme:
Electronegativity
increasing
c N 0 F
p s Cl
As Se Br
Sb Te I
Stability
increasing
This scheme accounts for the stability orders noted above for halogen
complexes. It also accounts for stability orders such as SeCN- > SCN >
OCN or CN > NH3 > .!!zO (the atom bonded to gold isunderlined in each
case), and suggests why gold tellurides and stibnites are stable enough to be
found in nature.
For more accurate prediction of the stability of gold complexes it is
from a cation, Mz+, and n ligands, u-, is the equilibrium constant of the reaction
for its formation from the free cation and ligands:
Mz+ + nI!- = ML~- ny
[ML~-"Y]
(3 11
= [Mz+] [U'-t
[ ] denotes the activity of the species enclosed. If conditions are such that ML 11z-ny
precipitates from solution, the relationship
K, = [M'+] [I!-]"
holds. K, is known as the solubility product.
832
THE CHEMISTRY OF THE EXTRACTION OF GOLD
necessary to take this argument a step further. In general, whether the

oxidation of gold to the complex AuL; or AuL!+ occurs in the presence of
a particular ligand depends on the strength of the oxidizing agent and on
the relative standard reduction potentials for the reduction of gold(I) or
gold(III) to the metal in the presence of the particular ligand. It is also ap-
parent that the requirements for complex stability are different for gold(I)
and gold(III) (Finkelstein and Hancock, 1974), with less electronegative or
'soft' donor atoms preferring metal ions of low valency, and 'hard' donor
atoms preferring metal ions of high valency. Therefore gold(I) forms more
stable complexes with ligands containing soft donor atoms such as S, C, Se,
and P, while gold(III) forms more stable complexes with the more elec-
tronegative, or hard, donor atoms such as N, 0, F, Cl, etc. This indicates
that complexes of gold(III) with soft ligands will be easily reduced to the
gold(I) state, while complexes of gold(I) with hard donor ligands will pro-
bably disproportionate to gold(III) and gold (0). This accounts for the fact
that gold(I) preferentially forms aurous complexes with ligands such as
cyanide, thiourea, thiocyanate, and thiosulphate (which all have soft donor
atoms) whereas, in chloride solution (hard donor atom), the monovalent
Auel; species tends to disproportionate according to the following
reaction:
3AuCl; = Auel; + 2Au + 2Cl-.0
(15.1)
Gold(!) has the closed shell 4f 145d 10 electronic configuration, and
therefore gold(I) complexes are diamagnetic and adopt regular structures.
All known gold(III) complexes have the low-spin 4f 145d 8 electronic con-
figuration and are also diamagnetic.
The preferred co-ordination number of gold(I) is 2 and that of gold(III)
is 4, with gold(I) tending to form linear complexes and gold(III) tending to
form square planar complexes. For example, the cyanide complexes of gold(I)
and gold(III) are Au(CN); and Au(CN);, respectively, and both gold(I)
and gold(III) tend to take up extra ligands in an effort to attain their prefer-
red co-ordination numbers. Thus insoluble AuCN in contact with a solution
containing cyanide ions will tend to take up a second CN- to form the solu-
ble Au(CN);, whereas AuCN in the solid state attains the required co-
ordination number by forming a polymer of the type ...
= =
Au - C N -Au - C N... When additional cyanide is available, however,
Au(CN); is preferred to polymeric AuCN because the Au - C bond is
stronger than the Au - N bond in the polymer. Similarly, compounds such
as AuC13 form dimers in the solid or gaseous states to satisfy the co-
ordination number of 4 required by gold(III):
833
Table 15.1. Stability constants* for a selection of complexes of gold(I) and gold(III).
Gold(I) Gold(III)
Complex f32 Complex {34
Au(CN); 2 x 1038t Au(CN) 4- -1056

Au(S,0 3 )~- 5 x 1028 Aur; 5 x 1047
Au(CS(NHJX 2 x 102lt Au(SCN); 1042
Aur; 4 x 10 19 AuBr; 1032
Au(SCN); 1,3 x 10 17 AuC14- 10'6
1012
AuBr;
Auel; 109
*Values of stability constants used in this chapter are taken from the tabulations of Sil!en (Sillen
and Martell, 1964) unless otherwise indicated.
tHancock and Finkelstein (1970).
tCalculated from standard reduction-potential data reported by Groenewald (1975).
It should be noted that, although aurous and auric ions are generally
represented as Au+ and Au3 +, they do not occur in solution as the bare
ions, but in a hydrated state as complexes containing the number of water
molecules appropriate to the co-ordination requirements of the particular
oxidation state:
Since the atom bound to the gold in each of these ions is oxygen, these com-
plexes are of low stability. (Oxygen occurs at the top and towards the right
of the partial periodic table shown previously.) The simple ions tend to react
in solution to replace the water molecules by stronger ligands. Similarly, if
AuC13 is dissolved in water, it does not (as the chlorides of many other
cations do) give rise to a free hydrated ion and free chloride. Instead, the
chlorides remain bound to the gold, and dimerization takes place to satisfy
the co-ordination number requirements. If another ligand is present in solu-
tion, the dimer breaks down and the ligand takes up the vacant co-ordination
position.
The gold(I) and gold(III) compounds that are of the greatest importance
to the extractive metallurgist are listed in Table 15 .1, which also provides
information on the co-ordination numbers and stabilities of these complexes.
The gold(I) complex of greatest importance to the extractive metallurgist
is undoubtedly the extremely stable aurocyanide anion, which, for the reasons
outlined above, is the predominant gold species present in cyanide leach
liquors. There is no evidence for the formation of species with co-ordination
numbers greater than 2, analogous to the copper(!) complex anions
834
Cu(CN)i- and Cu(CN)!-. The insoluble gold cyanide compound AuCN

precipitates from acidic solutions of aurocyanide ions as follows:
Au(CN); + H+ = AuCN + HCN. (15.2)
This reaction takes place readily in strongly acidic solution and slowly in weak
acid solutions (pH 2 to 3). The reaction is of significance on South African
gold plants where uranium and gold are recovered from the same milled ore.
Uranium is leached in acidic solution (pH 1 to 2), and any aurocyanide pre-
sent in solution in the feed to the uranium-leaching plant will precipitate as
AuCN, and will be lost with the solid residue from the plant. This has been
a major impediment to the development of processes for the simultaneous
leaching of gold and uranium, which would have obvious economic advan-
tages over the conventional two-stage process.
The aurocyanide anion is readily oxidized by halogens to mixed-ligand
auric complexes of the type [Au(CN) 2X2] - (where X is c1-, Br-, or r).
Other cyanide complexes include the thallium(I) salt Tl[Au(CN) 2], in which
there is evidence of interaction between TI and Au (Stammreich et al., 1968),
the linear neutral molecule AuCN(PPh3) (Cariati et al., 1970), and the
complex Au(PPh;) 3 Au(CN) 2 , in which both cationic and anionic gold(I)
species are present (Cariati et al., 1970). Gold-cyanide-phosphine complexes
have been extensively studied, and were last reviewed in 1974 (Rigo and Turco,
1974).
Other complex aurous ions that are relevant in hydrometallurgy are the
thiourea, thiosulphate, and thiocyanate complexes. These are the only ligands
that are likely to compete with cyanide in the large-scale processing of gold
ores, and they are likely to become more important, since increased atten-
tion is being focused on the search for alternatives to the environmentally
undesirable cyanide process. The thiourea complex is unique in that it is the
only cation amongst the complexes listed in Table 15 .1, and this paves the
way to novel process options for the extraction of gold from thiourea leach
liquor (Groenewald, 1977).
The extensive chemistry of gold(I) complexes is almost matched by that
of gold(III) complexes, the vast majority of which have 4-co-ordinate square
planar stereochemistry. Examples include the complex anion AuCli and the
cation Au(NH3)!+, but the simple aquo-ion Au(HzO)!+ is not formed in
measurable concentration, even when gold(III) hydroxide is dissolved in nitric
or sulphuric acid (Bezzubenko and Peshchevitskii, 1962). All auric complexes
are strong oxidizing agents.
The halides are the most important gold(III) complexes from a
hydrometallurgical point of view. Gold dissolves readily in aqueous chloride
solution in the presence of an appropriate oxidizing agent (such as chlorine,
nitric acid, or ferric ions) to form the tetrachloroauric ion. This ion can be
isolated as a hydrated acid, such as H[AuC14]3H2 0, by evaporation of the
solution, or as a salt such as Na[AuC14]. Tetrafluoro and tetrabromo com-
plexes of gold(III) are also stable in aqueous solution, but salts of the AuI.;-
ion cannot be prepared from aqueous solution, since partial reduction to
Aul; takes place. Halide complexes of mixed valency such as
835
Csi[AuCl2] [AuC14] are also known. In aqueous solution both the Au Cl; and
AuBr4- ions are hydrolysed to some extent, forming species such as
[AuC13]H2 0 and [AuBr3]H2 0. These in turn act as weak acids forming
species such as AuX4 _ 11 (0H) 11 (where X is Cl, Br, or I, and n = 0 to 4 with
increasing alkalinity) in alkaline solution. The species Au(OH) 4 is stable in
strongly akaline solution.
The auricyanide complex, Au(CN);, like the corresponding aurous
complex, is extremely stable, its stability constant, which is known only
approximately, being of the order of 1056 • The complex K[Au(CN) 4] is
prepared by the addition of a solution of Na[AuC14] to a concentrated
potassium cyanide solution, and complex anions of higher co-ordination
number, such as Au(CN);- and Au(CN)!-, are formed in solutions contain-
ing excess cyanide.
A great many other gold(I) and gold(III) compounds have been prepared
and identified, mostly in the last 10 to 20 years. The most numerous are the
multitude of phosphine, arsine, and stibine complexes that tend to dominate
gold(I) chemistry. Gold(III), being a harder Lewis acid than gold(I), not only
forms complexes readily with soft ligands, but also with hard ligands such
as fluoride and nitrogen donors. Some of these compounds may undergo
reactions of importance to the hydrometallurgist, and interested readers are
referred to the recent reviews of Puddephatt (1978) and Schutte (1985).
15.1.2 An equilibrium approach to the hydrometallurgy of gold

Any chemical reaction, such as the reduction of a metal ion M2 + with
hydrogen
(15 .3)
has a thermodynamic driving force, the Gibbs free energy change of reac-
tion, AG. The value of AG is determined by the reacting species, their
activities, the temperature, and the pressure. If AG is zero, the reaction is
at equilibrium. If AG has a negative value, the reaction is driven in the for-
ward direction and reduction of the metal ions occurs, resulting in a decrease
in the activities of the reactants and an increase in the activities of the pro-
ducts with a corresponding decrease in the absolute value of AG, i.e. the
reaction moves towards equilibrium. If AG has a positive value, the reaction
is driven in the reverse direction, resulting in the oxidation of M to M 2 +,
and a similar decrease in the absolute value of AG.
The rate at which the reaction moves towards equilibrium is dependent
on a number of additional parameters that will not be discussed in this sec-
tion. It is important to note, however, that the rate of reaction is only in-
directly affected by the absolute value of AG.
Reaction (15 .3) can be considered to represent a galvanic cell consisting
of the two half-cell reactions:
M 2 + + 2e = M, (15.4)
and
H 2 = 2H+ + 2e (15.5)
836
The electromotive force of the cell, E, is defined by the equation

E = -1.::i.G/nF, (15.6)
where n is the stoichiometric number of electrons involved in the reaction
(2 in this case), and F is the Faraday constant (96 487 C mol- 1). Under
standard conditions of 25°C, 100 kPa, and unit activities of all reacting
species, Equation (15.6) can be written ?S
E 0 = -/::i.G 0 /nF, (15.7)
where E 0 and /::i.G 0 are the standard potential and the standard Gibbs.free-
energy change of reaction (15 .3) respectively.
The relationship between E and £ is given by the Nernst equation:
0
E = E 0 - (RT/nF) ln K, (15.8)
where R is the universal gas constant (8,314 J K- 1 mol- 1), T is absolute
temperature, and K is the equilibrium constant. At 25°C the constants in
Equation (15.8) can be evaluated, yielding
E = Eo - 0,0591 log K. (15.9)
n
K is defined as the arithmetic product of the activities of the reaction pro-
ducts raised to their stoichiometric numbers divided by the arithmetic pro-
duct of the activities of the reacting species raised to their stoichiometric
numbers. If the concentrations of all the reacting species are below about
0,01 mo11- 1, the activity coefficients are generally close to unity, and the
activities can be replaced by concentrations with little error. If the concen-
trations are high, the error introduced into the potentials can be as large as
50 m V. Because of the difficulties that are encountered in the determination
of activities, this discussion will, for the sake of simplicity, approximate all
activities by concentrations, and any possible errors incurred will be ignored.
In order that the Nernst equation can be applicable to half-cell reac-
tions such as that described by Equation (15 .4), a reference half-cell must
be defined. This reference is taken to be the hydrogen reaction, (15 .5), for
which both l::i.G 0 and E° are assumed to be zero. The activity of all solid
phases is defined as unity, and the Gibbs free energy of all pure elements
in their standard states is set to zero. Under these conditions, the free-energy
change for a reaction can be calculated from tabulated values of the stan-
dard Gibbs free energy of formation, l::i.G~, for all the reacting species.
In a reaction such as
Au3 + + 3e = Au, (15.10)
the standard Gibbs free-energy change for the reaction, l::i..G 0 , is given by
l::i.G 0 = l::i.G~(Au) - l::i.G~(Au3 +)
1
= -433 kJ mol- • (15.11)
(Unless stated otherwise, all free-energy data were obtained from Pourbaix,
1966 or Wagman et al., 1965.) The standard reduction potential for Equa-
tion (15 .10) is therefore
837
E' 433 x 103/(3 x 96 487) v

= (15.12)
1,50 v
=
The large positive value of E 0 (or the large negative value of AG indicates
0
)
that Au3 + is thermodynamically unstable, and should be readily reduced to

Au. With a value of -237 kJ mol- 1 for the Gibbs free energy of water, the
standard reduction potential for the reaction
0 2 + 4H+ + 4e = 2H20 (15.13)
is 1,23 V. Because this value is lower than the value of 1,50 V for the Au 3 +
- Au reaction, Au 3 + is unstable in water, and will undergo spontaneous
reduction to Au with the oxidation of water to oxygen. It can also be con-
cluded that oxygen, with a standard reduction potential of 1,23 V, is not
able to oxidize Au to Au 3 +.
With increasing pH of the electrolyte, the Au3 + ion will be precipitated
as hydrated auric oxide, Au 20 3 .3H20 or Au(OH) 3 , i.e.
Au3 + + 3H2 0 = Au(OH) 3 + 3H+. (15.14)
When reaction (15 .14) has attained equilibrium, AG = 0, and hence
AG = -RTin([H+] 3/[Au3 +]),
0
(15.15)
from which it can be shown that Au3 + of unit activity is at equilibrium with
solid Au(OH) 3 at a pH value of - 0,3. The reduction of Au(OH) 3 to Au,
Au(OH) 3 + 3H+ + 3e = Au + 3H2 0, (15.16)
can be shown to have a standard reduction potential of 1,48 V, and hence
the reduction potential at any pH value will be expressed by the equation
E = 1,48 - 0,0591 pH (15.17)
3
The reduction of Au + to Au (15.10), the precipitation of Au3 + as
Au(OH) 3 (15.14), and the reduction of Au(OH) 3 to Au (15.16) are examples
of three distinct classes of reaction, viz.
i) dependent on potential, but independent of pH,
ii) independent of potential, but dependent on pH, and
iii) dependent on potential, and dependent on pH.
Values of the Gibbs free energies can be found for a number of gold
compounds in the compilations cited above. Equations representing reac-
tions between these compounds can be written, and the potentials and critical
pH values that correspond to these reactions can be calculated from the free
energies. For the meaningful interpretation of this large amount of data, it
is convenient to plot the data in the form of a potential - pH diagram (also
referred to as a Pourbaix diagram), as shown in Figure 15.1. Equations of
class (i) above appear as horizontal lines, those of class (ii) as vertical lines,
and those of class (iii) as lines angled above or below the horizontal. Solid
phases are indicated by bold print, and soluble aqueous species by light print.
The potential - pH diagram in Figure.15 .1, which represents the r'eac-
tions of gold in the absence of any complexing agents, shows only the major
reactions. A complete diagram can be found in Pourbaix (1966). The two
838
2.0 3
Au+
Au(OH)3
----- --- --- --- ---

(15.10)
Os. 1'J7-.._ - -.._

Au
--- ---
--- --- -- --- -- -!
0
15
.f8J
-
--- ------ ---
----
-1.QL---'--'--1---'~'--.l..--'--'--l..--l~l..-.!-..L-L......I
- I 0 I 2 3 4 5 6 7 8 9 10 11 12 13 14
pH
Figure 15.1. Potential-pH equilibrium diagram for the system

Au-H2 0 at 25°C. The labels on the lines refer to the equation
numbers in the text. Concentrations of all soluble gold species
are 10- 4 M. P 0 = 1 atm. = PH (after Adamson, 1972).
2 2
dashed lines refer to the reduction of water,

2H20 + 2e = H 2 + 20H-, (15.18)
at E = - 0,0591 pH,
and to the reduction of oxygen,
0 2 + 4H+ + 4e = 2H2 0 (15.13)
at E = 1,23 - 0,0591 pH.
(The fugacities of H 2 and 0 2 are assumed to be unity.)
The extremely large stability domain of metallic gold, which extends
above the O? - H?O line, indicates that gold is stable with respect to oxida-
tion by oxygen at -all pH values, including those of the halogen mineral acids,
sulphuric, nitric, and phosphoric acids, and caustic alkalis.
In the presence of strong complexing agents, which can stabilize gold
in either the + 1 or + 3 oxidation states, the potential - pH diagram can be
significantly affected. As an example, the effect of cyanide ions on the
reaction
839
1.0
._,__ --- ----
---- ---- -.!.!.5, f ]J
>
L<.l"
:"§
;;;:-.__
---- --- ----

Au( CNF -
---- ---- --- ~ ~
--- ----
=
"
0
Cl.
(/S' ---- ---
·..?3.1 --- ---- ---
0
.::::"
u;:l ---- ---(!!.:_1 8) ---- ----
"Cl
" ----------
0::
(15.28)
Zn 2 +
-- ----
Au--..
-1.0
HZn02-
Zn
-2.0'--..L......--'--'------1.~L.-..L......--'--'---'---'~..L......_L__J____J
0 2 4 6 8 10 12 14
pH
Figure 15 .2. Potential-pH equilibrium diagram illustrating
features of the Au-CN-H2 0 and Zn-CN-H2 0 systems at
25°C. [CN-], 0 ,.1 = 10- 3 mol l- 1, [Au(CN);J = 10- 4 mol i-1,
[Zn2+l,otal = 10-1 mol 1-1.
Au+ + e =Au, (15.19)

E = 1,69 - 0,0591 log [Au+] (15.20)
is considered. Because of the strong complex between Au+ and CN-, viz.
Au+ + 2CN- = Au(CN); (K = 2 x 1038), (15.21)
Equation (15 .20) can be rewritten as
E = 1,69 -0,0591 log K - 0,0591 log ([cN-] 2/[Au(CN);J), (15.22)
which results in a negative shift in the E 0 value from 1,69 V to - 0,57 V.
This lower E 0 value applies to the new reaction obtained when Equations
(15.21) and (15.19) are added:
Au(CN); + e = Au + 2CN-. (15.23)
Because HCN is a weak acid (the dissociation constant for the reaction
HCN = H+ + Cl'r (15.24)
840
is 4,93 x 10- 10), a lowering of pH results in a reduction in the concentra-

tion of free cyanide ions. The effect of this hydrolysis reaction can be incor-
porated into Equation (15 .22), yielding
E = 0,53 + 0,059 log [Au(CN);J -0,118 log ([CNlt 0 t - 2 [Au(CN);])
+ 0,118 log (4,93 x 10- 10 + [H+]), (15.25)
where [CNL0 t is the total concentration of cyanide added to the solution.
Equation (15 .25) represents the relationship between the reduction potential
for (15.23) and the concentrations of all the species that may affect the value
of E.
The potentials corresponding to Equation (15 .25) are shown in Figure
15 .2. Because of the strong stabilizing effect of the cyanide ion, metallic gold
will be readily oxidized to aurocyanide ions by dissolved oxygen. It is of in-
terest to note, however, that the reduction of oxygen on a gold surface in
an electrolyte containing cyanide ions can yield an appreciable amount of
hydrogen peroxide (Habashi, 1966), cf. Section 15.1.3.3.
0 2 + 2H+ + 2e = H 20 2, (15.26)
E = 0,682 - 0,0591 pH (15.27)
rather than proceeding to water (15.13). Fortunately, the negative shift in
the equilibrium potential for gold in the presence of cyanide ions still re-
mains positive with respect to the reduction potential of zinc in the same
electrolyte. Zinc ions form fairly strong complexes with both OH- and
CN-, forming numerous compounds such as ZnOH+, Zn(OH) 2, HZnO;,
Zno;- and Zn(CN);-x, where x is from 1 to 4 (Finkelstein, 1972; Osseo-
Asare et al., 1984a). The effect of these complexes is to shift the reduction
potential of the reaction
Zn2+ + 2e = Zn, (15 .28)
E = - 0, 76 + 0,0295 log [Zn2+] (15.29)
to slightly more negative values at pH values above 7. The potentials cor-
responding to the reduction of zinc are shown in Figure 15 .2, from which
it is evident that dissolved aurocyanide ions will be reduced to metallic gold
by the corresponding oxidation of zinc.
It is of importance to note that the reduction potentials of zinc are below
those for both the reduction of oxygen (15.13) and water (15.18). Zinc is
rapidly oxidized by dissolved oxygen, and hence one needs to deaerate the
electrolyte prior to cementation in order to reduce the consumption of zinc.
However, zinc will still corrode spontaneously in water as hydrogen is evolved,
which accounts for the large stoichiometric excess of zinc required for the
cementation of gold.
Inspection of the electrochemical series reveals two further reduction
reactions that could be used for the precipitation of gold from cyanide, viz.
those of aluminium and borohydride:
Alo; + 4H+ + 3e = Al + 2H20 (15.30)
E = -1,26 -0,0788 pH + 0,0197 log [AlO;J, (15.31)
841
Table 15 .2. Standard reduction potentials for various reactions of gold (after Schmid and eurley-
Fiorino, 1975).
Reaction It' (V)
Au+ + e = Au 1,69
Aue1; + e = Au + 2e1- 1,154
Au(SeN); + e = Au + 2SeN- 0,662
Au(eS(NHJ,)7 + e = Au + 2eS(NH2), 0,352*
Au(S,03);- + e = Au = 2s 2 0;- 0,15
Au(eN); + e = Au + 2ew -0,57
Au3 + + 3e = Au 1,50
Auel; + 3e = Au + 4eI- 1,002
Au(SeN); + 3e = Au + 4Sew 0,623
AuH + 2e = Au+ 1,40
Au(SeN); + 2e = Au(SeN); + 2Sew 0,604
Auel; + 2e = Auel; + 2e1- 0,926
*Groenewald (1977)
and
H 2BO; + 5H2 0 + 8e BH,;- + 80H-
= (15.32)
E = - 0,413 - 0,0591 pH + 0,0074 log ([H2 BO;J/[BH,;-]). (15 .33)
However, because the reduction potentials of these reactions at a pH of 11
( - 2,13 V and -1,06 V respectively) are very negative compared to that for
the reduction of water ( - 0,65 V), the consumption of reagent due to the
evolution of hydrogen is high.
The hydrometallurgy of gold is not confined entirely to that of the
cyanide system, since gold also forms strong complexes with a number of
other ligands such as thiourea, thiosulphate, thiocyanate, and chloride. The
standard reduction potentials for these reactions are shown in Table 15.2.
The effect of chloride ions on the potential - pH behaviour of gold has
been discussed in detail by Finkelstein (1972). It is evident from the poten-
tials in Table i 5 .2 that the stable oxidation state of gold in the presence of
chloride ions is + 3. The reduction poti;ntial for auric chloride (1,002 V) is
below that for the reduction of oxygen to water (1,23 V), and hence it would
appear that gold should be dissolved by oxygen in strong chloride electrolytes.
The rate of oxidation is extremely low, so that gold appears to be inert under
such conditions. In practice, stronger oxidants such as dissolved chlorine (E
= 1,395 V) or hypochlorite (E = 1,715 V) are used for the oxidation of
metallic gold. Solutions of auric chloride are readily reduced to metallic gold
by soluble reductants such as sulphite, nitrite, ferrous, and stannous ions.
Gold is stable in the + 3 oxidation state in electrolytes containing unit
activities of dissolved gold and thiocyanate ions. However, the equilibrium
potential is shifted negatively as the concentration of dissolved gold is decreas-
ed, and this shift is greater by a factor of 3 for gold in the + 1 oxidation
state. As a result of this shift, gold in the + 1 oxidation state becomes the
stable species when the logarithm of the concentration of dissolved gold falls
842
below - 0,98 + log[SCN-], i.e. when the concentration of gold is less than
0,1 of the thiocyanate concentration.
Gold in the presence of thiocyanate ions is thermodynamically unstable
in the presence of oxidizing agents such as H 2 0 2 , 0 2 , and Fe3 +. It should,
however, be noted that thiocyanate ions can be oxidized to thiocyanogen
according to the reaction (Latimer, 1952):
(SCJ\!) 2 + 2e = 2SCN- E 0 = 0,77 (15.34)
Very strong oxidants, such as hydrogen peroxide, may therefore destroy the
complexing ligand, resulting in poor dissolution of the gold.
The neutral ligand thiourea forms a strong complex with Au+, thereby
forming a soluble gold compound exhibiting a positive charge. Because of
the low E° value (0,352 V) of the gold-thiourea complex, gold is readily
oxidized in the presence of thiourea. However, thiourea is readily oxidized
to formamidine disulphide (Preisler and Berger, 1947):
NHi{NH)CSSC(NH)NH2 + 2H+ + 2e = 2CS(NH2) 2 , E° 0,420 V
=
(15 .35)
so that a marked loss of thiourea would be expected even in the presence
of Fe3 + (E° = 0,77 V). The stability of thiourea towards oxidation is reduc-
ed with increasing pH.
Thiosulphate co-ordinates strongly with Au+, yielding a standard
reduction potential of 0,15 V. Unfortunately, thiosulphate is unstable in acidic
as well as alkaline environments and will be oxidized to tetrathionate, sulphite
or sulphate.
15.1.3 Electrochemical aspects

The thermodynamic characteristics of gold and its compounds of relevance
to the hydrometallurgy of gold have been outlined above. Of perhaps greater
importance are the kinetic considerations that apply to the leaching and
precipitation of gold. As will be discussed in detail in the next section, both
of these processes are electrochemical in nature, and it is therefore of some
importance that the electrochemistry of these systems should be briefly review-
ed as it pertains to these processes. The dissolution of gold involves an
oxidative (or in an electrochemical sense, anodic) reaction, whereas its
precipitation by cementation requires a reductive (or cathodic) process. Each
of these will be discussed in turn.
15.1.3.1 Anodic reactions of gold

The electrochemical behaviour of the noble metals (the platinum-group metals
plus gold) under oxidizing conditions is dominated by the formation of ox-
ide films. This can have pronounced effects on the kinetics of dissolution
of these metals, and partly accounts for their refractory nature. In the case
of gold, the nature and kinetics of formation of oxide films have been ex-
tensively studied (Nicol, 1980a) and, while the results have shown that the
oxidation processes are complex, the important overall characteristics in acidic
and alkaline solutions can be demonstrated by the data in Figure 15 .3. In
these experiments, the current flowing at a gold electrode in the solution
843
Potential, V
0.2 0.4 0.6 0.8
0.2
NE
u
"<t: 0
E
.~ ........
cg \
"
"O -0.2 \
Q
"........ \
;:l
u -0.4
\ I
- - - 0 , l MKOH
I I ---IMHCI04
I\,I
1.2 1.4 1.6
Potential, V
Figure 15 .3. Cyclic voltammograms for the oxidation of gold

in 0,1 M potassium hydroxide and in 1 M perchloric acid. The
abscissa scales are of different sensitivities as well as laterally
displaced to allow the curves to overlap and to demonstrate their
general similarity (after Nicol, 1980b).
shown was measured, and the potential was first increased and then decreased.
Positive currents correspond to anodic reactions, and negative currents to
cathodic reactions. The oxidation of gold therefore begins at about 1,4 V
in acidic solutions, and at 0,4 V in alkaline solutions. The peaks at 1,25 V
and 0,35 V on the reverse sweeps are due to the reduction of the oxide films
back to the metal.
These observations imply that the formation of oxide films can be
expected to adversely affect the dissolution of gold at potentials above 0,4 V
in alkaline solutions and 1,4 V in acidic solutions.
In the presence of species (or ligands) that form complexes with gold
in solution, the anodic characteristics of gold can be significantly modified
as a result of the increased stability of the gold(I) or gold(III) complexes with
the ligands. Of these, only the behaviour in chloride, thiourea, and cyanide
solutions has been studied in any detail, and the major conclusions of the
published work will be summarized below, particularly where they relate to
the use of these ligands in the extraction and refining of gold.
(a) Chloride
The main features of the anodic behaviour of gold in acidic chloride solu-
tions are shown in Figure 15 .4, from which it can be seen that dissolution
occurs at potentials above about 1 V. The plateau at high current densities
is due to the limited rate of mass transport of chloride to the gold surface,
844
10- I
,.-- 0.5 M NaCl

10-2
E:u
'--
<!'.
10-3
·~
"
<I)
'O
;: 10_,
t
::I
u 10-5
1.0 1.2 l.4 l.6 l.8 2.0 2.2

Potential, V
Figure 15.4 Steady-state current versus potential curves for the

oxidation of gold in solutions containing 0,5 M sulphuric acid
and various amounts of chloride (after Heumann and Panesar,
1966).
while the rapid decrease in the rate of dissolution from 1,4 to 1,5 V is due
to passivation of the surface by oxide, as predicted from the curve for acidic
solutions in Figure 15.3. Dissolution in the active region (i.e. at potential
below 1,4 V) is complex in that, depending on a number of variables such
as chloride concentration, temperature, and degree of agitation, both gold(I)
and gold(III) can be formed in varying proportions. Work at the Council
for Mineral Technology (Mintek) reported by Nicol and Schalch (1976b)
demonstrated that the formation of gold(I) at the anodes of the chloride-
based electrorefining process can lead to a number of operational problems
in the refining of gold. Some aspects of this will be discussed in Section
15.1.3.3(c).
(b) Thiourea
Gold(I) forms the strong cationic complex Au[SC(NH2) 2J; with thiourea,
and this has been the basis for a number of schemes proposed for the recovery
of gold from ores and concentrates (Groenewald, 1977; Pyper and Hendrix,
1981). Steady-state measurements of the current at a gold electrode at various
potentials in an acidic thiourea solution have shown that gold dissolves as
gold(I) with a current efficiency of 100% at potentials below about 0,6 V,
as shown by the data in Figure 15 .5. Comparison of the limiting current with
that calculated for the reaction
Au + 2SC(NH2) 2 = Au[SC(NH2) 2J; + e (15.36)
shows that dissolution is largely diffusion-controlled in the plateau region.
The increase in current at higher potentials is accompanied by a decrease
in the current efficiency, due to the competitive anodic process of the oxida-
tion of thiourea to formamidine disulphide and, at the highest potentials,
to further oxidation products. This irreversible loss of thiourea by oxidation
845
----"\ 100
10-2 \
\
\ 80
\
"'a ~
u
"~ \ ~
u
c:
60
.~ \ (I.)
·o
~
(I.)
'O
10-4 \--- s
(I.)
;::: \ ;:::
(I.)
........
\
(I.)
....
.... 40 ;;
;;
u u
\
\
\ 20
10-6
\
\
0.4 0.6 0.8 1.0 1.2

Potential, V
Figure 15 .5. Effect of potential on the anodic dissolution of

gold and the current efficiency of its dissolution into a solu-
tion containing 0,1 M sulphuric acid and O,lM thiourea at 30°C
(after Groenewald, 1977).
is a significant factor in the economic viability of its use in the leaching of

gold. The most important point to be learnt from Figure 15.5 is that gold
will be leached at a significant rate in thiourea solutions only if the potential
can be maintained above 0,4 V, and this requires oxidants such as ferric ions
or hydrogen peroxide. The reduction of oxygen is slow at these potentials,
and this precludes the use of oxidation by air.
(c) Cyanide
The electrochemical nature of the dissolution of gold in aerated cyanide solu-
tions is well established, and much work has been published on the anodic
behaviour of gold in alkaline cyanide solutions. This subject was recently
reviewed (Nicol, 1980b), and therefore only the most important features will
be discussed.
Detailed coulometric studies (Kirk et al., 1979) have revealed that the
stoichiometry of the anodic dissolution reaction in cyanide solutions is:
Au + 2CN- = Au(CN); + e (15 .37)
throughout the potential range - 0,65 to Q,55 V. At higher potentials there
is evidence of the formation of gold(III) species. A recent review (Nicol,
1980b) demonstrates that the anodic characteristics of gold in cyanide solu-
846
10
11
N
E
8 1'
u ,1
"<!'.'.
s
E 6
2
I
·v;
i:::
<!)
(\ 11
"'O
c::
4 I I II
...
<!)
I-< U1
::;
u 2 I
\
0
-1.0 0 1.0
Potential, V
Figure 15.6. Current versus potential curves for the oxidation

of gold in alkaline cyanide solutions (after Nicol, 1980b). 1,
0,077 M CW, pH 12; 2, 0,1 M CN-, 0,1 M OH-.
tions, as reported by different workers, exhibit significant variations.

However, examination of the curves (two of which are shown in Figure 15.6)
reveals that three peaks occur at potentials of about - 0,4 V, 0,3 V, and 0,6 V
and are associated with the formation of three passivating films. Those at
0,3 and 0,6 V are associated with the formation of oxide films, as comparison
with the data in Figure 15.3 will show for the peak at 0,3 V.
During the cyanidation of gold, the potential does not exceed zero, and
therefore it is appropriate to discuss in greater detail the anodic behaviour
in the region of the peak at - 0,4 V. As was recently pointed out (Nicol,
1980b), the unusually wide scatter in the results obtained by various workers
for the activity of gold towards dissolution in this region can probably be
traced to the presence of different amounts of heavy-metal impurities in the
solutions used. Of these metals, trace amounts of lead, mercury, bismuth
and thallium can have profound effects on the anodic characteristics of gold
in this region, as shown by the results in Figure 15. 7.
On the basis of measurements of the current at various potentials on
the active portion of the peak (i.e. at potentials below - 0,4 V), and of the
variation in current with the concentration of cyanide at constant potential,
it is generally accepted that the dissolution of gold occurs by way of the
following mechanism:
Au + CN- = AuCNads + e, (15.38)
AuCNads + CN- = Au(CN);. (15.39)
Passivation is therefore associated with an absorbed layer of AuCN. An
interesting clue to the possible nature of this layer is to be found in the
macromolecular structure of solid AuCN, which consists of linear chains in
847
---0.lM KOH, O.lM KCN

- - - 0.1 M KOH, 0.1 M KCN, and 2x 10- 6 M Hg(II)
6
10.0 - · - · O.lM KOH, 0.1 M KCN, and 1x10- M Pb(II)
"'au
"< .nI. I I\I
E I
i} I
.~ V . I
:g
Q.) 5.0 Ii/ \ I
"<:: I . I
...
Q.)
I11· \. I
5
u I . \
/./ .
-0.6 -0.4 -0.2 0 0.2 0.4 0.6
Potential, V
Figure 15.7. Effect of the addition of lead and mercury ions
on the anodic behaviour of gold in an alkaline cyanide solu-
tion. Before each anodic sweep, the electrode was left under
open circuit conditions for 5 minutes (after Nicol, 1980b).
which the cyanide ion functions as a bidentate ligand (Puddephatt, 1978).
It is conceivable that this type of structure could develop in two dimensions
to form a passive layer. The presence of foreign heavy-metal atoms (such
as Pb, Hg or TI) on the surface would be likely to disrupt the formation
of such a layer, thus reducing the extent of passivation.
It should be pointed out that the above observations apply only to the
behaviour of pure gold. In actual ores, gold is generally alloyed with other
elements such as silver, copper, and in some cases, even mercury. Problems
due to passivation of the gold surface under plant leaching conditions could
therefore be expected to be minimal. However, much additional research is
required to establish the effects of alloying on the anodic behaviour of gold.
15.1.3.2 Anodic reactions of zinc

The precipitation of gold from cyanide solutions by metallic zinc is an elec-
trochemical process in which the zinc dissolves anodically. It is therefore rele-
vant to discuss some of the major aspects of the anodic behaviour of zinc
in alkaline cyanide solutions. As pointed out in Section 15.1.2, several zinc
species such as Zn(CN)~- and Zn(OH)~- can be formed under these condi-
tions, and the electrochemical characteristics of the anodic process are
therefore complex, as shown by the current-potential curves shown in Figures
15.8and15.9. Therefore, in the absence of cyanide (curve a in Figure 15.8),
zinc shows little anodic activity at the pH values generally encountered dur-
ing precipitation. The addition of cyanide results in enhanced currents at
potentials above about - 1, 1 V (curve b). In some instances, a plateau is ap-
parent on the curve before the current decreases due to passivation of the
surface by a zinc oxide (or hydroxide) species. The plateau occurs because
the rate of the reaction
848
c
3.0
"'au
'<i:
2.0 E
~
·o;
c
0)
'O
1.0 =.......
0)
;::;
u
Potential, V
Figure 15 .8. Current-potential curves for the anodic dissolu-
tion of a rotating zinc-disc electrode in alkaline and alkaline
cyanide solutions. (a) 0,005 M Ca(OH) 2 , 1000 rev/min; (b)
0,005 M Ca(OH) 2 , 0,005 M NaCN, 250 rev/min; (c) solution
as for (b), 1000 rev/min (after Nicol et al., 1979).
(15 .40)
becomes controlled by the rate of mass transport of cyanide ions to the sur-
face of the zinc, as shown by the increased height of the plateau at the higher
stirring speed (curve c).
As could possibly be expected from the nature of the above reaction,
the rate (as measured by the current at.a constant potential) increases with
increasing cyanide concentration, and this results in both a cathodic shift
of the curves and an increased height of the plateau. Furthermore, an in-
crease in the pH of the solution results in similar trends in the curves, due
to the formation of soluble zinc hydroxy complexes such as Zn(OH)~- at
pH values above about 12, and to the formation of mixed hydroxycyano
complexes at lower pH values. The implications of these characteristics for
the efficiency of precipitation of gold will be discussed in detail in Section
15.3.1.
15.1.3.3 Cathodic processes

In addition to the obvious and important cathodic reduction of the auro-
cyanide ion to metallic gold, it is important to review briefly the other cathodic
processes that are relevant to the extraction and refining of gold, namely
the reduction of oxygen in alkaline solutions and the reduction of the gold'
chlorides in acidic solutions.
(a) Reduction of Au(CN)2-
Published information on the reduction of the aurocyanide ion under con-
849
0.5
a b c d
0.4
"'E
(.)
'-< 0.3
E
JS
·;;;
=
<lJ
"O
c: 0.2
<lJ
....
....
::;
u
0.1
0-1-~~--.'---+-~+r-ff-~~~~~~~~~~~~~~
-1.2 -0.9
Potential, V
Figure 15.9. Current-potential curves for the anodic dissolu-

tion of a rotating zinc-disc electrode in solutions of NaCN and
NaCl. a, 2 x 10- 2 M; b, 10- 2 M; c, 4 x 10- 3 M; d, 2 x
10- 3 M; e and/ 10- 3 M NaCN. All solutions contain 10- 2 M
NaCl at pH 11,0. The speed of rotation of the electrode was
1000 rev/min except for curve/, where it was 250 rev/min (after
Nicol et al., 1979).
ditions relevant to the processing of gold ores is not extensive, since most
workers have concentrated on the somewhat different conditions used in the
gold-plating industry. The following will therefore be a summary of work
carried out at Mintek (Nicol et al., 1979; Paul et al., 1983) in connection
with the cementation and electrowinning of gold.
Figure 15.10 shows current-potential curves for the reduction of
Au(CN); at a rotating gold-disc electrode in various deoxygenated solu-
tions. In the absence of Au(CN); (curve a), the cathodic current at poten-
tials below - 1,0 Vis due to the reduction of water to hydrogen and, as has
been discussed above, the small anodic current at potentials above about
- 0,4 V is due to the dissolution of gold. The addition of Au(CN); to the
solution results in curve b, which displays cathodic currents at potentials below
- 0,5 V due to the reaction
Au(CN); + e = Au + 2CN-. (15.41)
The rate of reduction increases with decreasing potential until it becomes
850
Peak
15.5mA/cm2
0.9
0.6
Potential, V
a
-0.6
-0.9
-1,2
Figure 15.10. Current-potential curves for reactions at a rotating

gold-disc electrode ( r = 1000 revI min) in alkaline. cyanide
(0,005 M NaCN and 0,005 M Ca(OH) 2) solutions (a, no gold
in solution; b, 1.6 x 10- 1 M gold; c 1.6 x 10- 1 M gold and
1,6 x 10- 5 M lead (after Nicol et al., 1979).
controlled by mass transport of the aurocyanide to the surface at potentials

close to - 1,0 V. The increased current at more negative potentials is due
to the simultaneous evolution of hydrogen. The addition of as little as 3 mg
of lead nitrate per litre of solution results in a significant shift of the curve
to more positive potentials, i.e. the reduction of aurocyanide is considerably
enhanced by the presence of lead in the solution. Similar effects are observ-
ed for mercury and thallium, and the catalytic effect of these metal ions on
both the anodic and cathodic characteristics of the Au/ Au(CN); couple
should be noted.
Variation of pH in the range 10 to 12 has no effect on the curve for
the reduction of Au(CN);. However, increasing cyanide concentration
results in a shift of the curve to more negative potentials by approximately
0,1 V for each tenfold increase in cyanide concentration. The addition of
small amounts of sulphide ions has no noticeable effect on eithei' the anodic
or cathodic reaction.
(b) Reduction of oxygen
The reduction of oxygen by the reaction
0 2 + 2H2 0 + 4e = 40H- (15.42)
851
- - - - N a C N 5 mmol/dm 3
200
--1111$~--NaCN 5 mmol/dm 3
Pb 2mg/dm 3
-1.0 0
-200
Figure 15.11. Cyclic voltammograms for the reduction of oxygen

(2,5 x 10- 4 mol l- 1) at a rotating gold disc electrode in solu-
tions of NaCN.
is thermodynamically possible at potentials lower than 0,63 V at pH 10.

However, the well-known kinetic inertness of this process, even at precious-
metal electrodes, results in large overpotentials for this reaction. In alkaline
solutions, therefore, the above reaction proceeds at a significant rate only
at negative potentials and, depending on the conditions, becomes mass-
transport controlled at potentials below about - 0,4 V.
In the absence of cyanide, the reduction of oxygen in alkaline solution
has been shown (Damjanovic et al., 1967) to proceed through the intermediate
formation of peroxide by the reaction
0 2 + 2H2 0 + 2e = Hp2 + 20H-. (15.43)
The extent to which peroxide is further reduced to hydroxide ions is depen-
dent on a number of factors such as pH, agitation, potential, and the presence
of trace amounts of heavy-metal impurities in the solutions used.
In the presence of cyanide, the rate of reduction of oxygen is generally
slower, as is shown by curve a in Figure 15 .11. The addition of small amounts
of heavy-metal ions such as those of lead, mercury, or thallium results in
a considerable enhancement of the rates of anodic dissolution of gold and
cathodic reduction of oxygen (curve bin Figure 15.11), with the peak for
the anodic process now superimposed on that of the cathodic reaction.
(c) Reduction of Auer;
As mentioned in Section 15.1.3.l(a), the anodic dissolution of gold results
in the formation of gold(I) and gold(III) complexes in various proportions,
852
~ 0.05 M Au(III), 2 M HCI

1111 0.04 M Au(III), 0,01 M Au(!), 2 M HCI
10-4
0.8 0.9 1.0

Potential, V
Figure 15.12. Current-potential curves (obtained from the

anodic sweeps of cyclic voltammograms) for a rotating (1000
rev/min) gold-plated platinum-disc electrode at 21°C (after
Nicol and Schalch, 1976b).
depending on the relative rates of the two consecutive reactions

Au + 2 c1- = AuCl; + e, (15.44)
and
Auel; + 2 c1- = Auel; + 2e. (15.45)
On the other hand, the dissolution of gold by the use of chlorine in concen-
trated chloride solutions results in the formation of AuCl; alone. The
reduction of gold(III) chloride to the metal can be expected to involve the
formation of gold(I) as an intermediate species, and an extensive investiga-
tion (Nicol and Schalch, 1976a) of the cathodic behaviour of gold in chloride
solutions has shown that the quality of the cathode deposit is strongly in-
fluenced by the relative amounts of gold(I) and gold(III) present in the
solution.
An explanation for this observation is to be found in the results in Figure
15 .12, which shows that the reduction of Auel; occurs at more positive
potentials than that of Auel;. The plateau in the curve for the former at
potentials below about 0,65 V occurs because the rate of reduction becomes
controlled by the rate of mass transport of Auel; to the surface. Under
normal electrorefining or electrowinning conditions, gold will be deposited
853
from both the gold(III) and gold(I) species. It is well known (Despic and
Popov, 1972) that metals deposited at their mass-transport-controlled rates
tend to form highly dendritic and powdery deposits. The fact that the cathodes
produced during the electrorefining of gold in chloride solutions are generally
of poor quality, in that they tend to be bulky and porous, is therefore not
unexpected.
The presence of gold fines in the solution as a result of the dispropor-
tionation reaction
3AuCl; = 2Au + AuCl; + 2Cl- (15.46)
also creates considerable handling problems in the electrorefining process.
From the above it should be apparent that process conditions should be ad-
justed so as to minimize the amount of gold(!) produced. Detailed studies
(Nicol and Schaich, 1976a, 1976b) of the anodic and cathodic reactions have
shown that the use of low temperatures and periodic interruption of the cur-
rent are the major factors that can contribute to reduced gold(I)
concentrations.
15.2 The Dissolution of Gold

15.2.1 The mechanism of cyanidation
The chemistry of the dissolution of gold (and silver) in alkaline cyanide solu-
tions has been the subject of considerable investigation for over a century.
Although the overall features of the reactions involved are well established,
there is still uncertainty regarding the details of some aspects of the mechanism
of the process. Undoubtedly the major advance in the understanding of
cyanidation was made by Kudryk and Kellogg (1954), who conducted ex-
periments to demonstrate that the dissolution of gold in cyanide solutions
is essentially an electrochemical process.
A simplified schematic diagram, which illustrates the 'mixed-potential'
electrochemical model (Figure 15.13), shows that the anodic dissolution of
gold in the presence of cyanide ions is coupled to the cathodic reduction of
oxygen. Each of these half-reactions has been discussed in detail in previous
sections. In the absence of any externally applied current, a gold surface im-
mersed in an aerated alkaline cyanide solution will undergo dissolution at
a rate governed by the requirement that the rate of the anodic process is equal
to the rate of the cathodic reduction of oxygen. As shown, this will occur
at a potential Em, known as the mixed potential.
The curves depict the situation generally observed under conditions of
'high-purity'. Data obtained by Kudryk and Kellogg (1954) under conditions
that probably resemble more closely those pertaining to plant operation, are
shown in Figure 15.14. It is noteworthy that there is no passivation of the
anodic reaction, and also that cathodic currents are treated as positive. The
plateaus in the anodic curves at potentials above about - 0,3 V and low con-
centrations of cyanide occur because the rate of dissolution becomes limited
by the rate of mass transport of cyanide ions to the gold surface. Similarly,
the plateau in the cathodic curve at potentials lower than - 0,5 V is the
854
Current
Anodic
Potential
Em
I
I
I
Cathodic
Figure 15.13. Simplified schematic diagram of the mixed-

potential model for the dissolution of gold in cyanide solutions.
limiting diffusion current for the reduction of oxygen from dissolved air.
Application of the mixed-potential model to these data results in the follow-
ing predictions.
i) At low concentrations of potassium cyanide (lower than about 0,017% ),
the rate of leaching will increase with increasing concentration of
cyanide, and the mixed potential will decrease from - 0, 18 to - 0, 5 V
at 0,005 and 0,017% cyanide respectively.
ii) At higher concentrations of potassium cyanide, the mixed potential will
decrease from - 0,55 V to - 0,75 V at 0,025 and 0,50Jo respectively,
but the rate of leaching will become independent of the cyanide
concentration.
iii) An increase in the concentration of oxygen by the use of pure oxygen
will result in a positive shift of the curve for the reduction of oxygen
(shown as the dashed line) with a resulting positive shift in the mixed
potential. Under these conditions, it is apparent that, at higher con-
centrations of potassium cyanide, the rate of dissolution will increase
with increasing cyanide concentration, and will become independent
of the cyanide concentration only when it exceeds about 0,0750Jo.
iv) For a given oxygen concentration such as that provided by an air-
saturated solution (2,5 x 10- 4 mo11- 1 at25°C), there will be a par-
ticular concentration of cyanide at which the rates of both the anodic
and cathodic reactions will be diffusion-controlled. This concentration
can be defined as follows.
The maximum rate (R.) of the anodic process
Au + 2CN- = Au(CN); + e (15.47)
will be given by Fick' s first law of diffusion, and can be described by
an equation of the form
855
§!. ~ 0
~
0 v; !;" v; ~
v;
0 0
N
~
0 0
c:i
0J
~
~·
z z z
u u u u<
~ ~ ~ ~
6
-<
\
3.5 \
\ KCN 0.01750Jo
3.0
2.5 \02
<8 \
2.0 \
~ KCN 0.0IOOJo
OJ
I-<
I-<
1.5
::i
u 1.0
0.5
0
-0.6 -0.4 -0.2 0 0.2
Potential, V
Figure 15.14. Anodic and cathodic current-potential curves for
gold (after Kudryk and Kellogg, 1954).
(15 .48)
where k is a constant, DcN the diffusion coefficient of the cyanide ion,
and [CN-] the concentration of cyanide. The factor 2 is included
because two cyanide ions are required for each electron produced.
Similarly, the maximum rate (RJ of the cathodic process
Oz + 4H+ + 4e = 2Hz0 (15.49)
will be given by
Re = 4kD 0 [Oz], (15.50)
where the factor 4 is included because four electrons are accepted by
each molecule of oxygen. Equating of these two rates and rearrange-
ment gives
[CN-] = 8Do .[02] (15.51)
DCN
which, on the substitution of values for the diffusion coefficients and
the above concentration of oxygen, yields a potassium cyanide con-
centration of 2,5 x 10- 3 M (or 0,016%). The practical implication of
this conclusion is that the rate of leaching in air-saturated solutions
will be unaffected by an increase in the cyanide concentration above
this level.
Increased rate of cyanidation can be achieved by the use of oxygen in-
stead of air, under pressure if necessary, increased agitation, and elevated
856
temperatures. These variations are employed in the 'intensive cyanidation'

process (Davidson et al., 1978), which is practised on several South African
plants for the treatment of gravity concentrates. In this process, concentrates
with a gold content as high as 20 kg per ton are subjected to cyanidation under
conditions of vigorous agitation with oxygen aeration and cyanide additions
as large as 50 kg per ton of concentrate. Under these conditions, over 97%
of the gold can be dissolved in 10 to 20 hours. An interesting observation
made during the development of this process is that the use of sodium cyanide
instead of potassium cyanide results in poorer recoveries from high-grade
concentrates, and this was shown to result from partial passivation of the
gold surface by the precipitation of the sparingly soluble sodium aurocyanide
salt. The use of potassium cyanide obviates this problem because of the greater
solubility of the potassium salt.
In practice, it is generally difficult to observe passivation of the gold
as described in Section 15.1.3.l(c), and therefore this is not normally con-
sidered. However, the literature contains several references to such observa-
tions, which indicates that passivation is at least partially responsible for
leaching rates that are lower than expected. As a result of testwork with
calcine, Cathro (1963) concluded that passivation of the gold surface occur-
red at high oxygen concentrations in the absence of salts such as those of
lead or thallium. In a more extensive study of the leaching and electrochemical
behaviour of gold in pyrite concentrates, Filmer (1982) found an inverse rela-
tionship between gold extraction and residual sulphur in the calcines. He pro-
posed a mechanism that involves the passivation of gold as a result of the
enhanced reduction of oxygen on the pyrite and pyrrhotite surfaces in con-
tact with the gold particles. This galvanic interaction results in a positive shift
of the potential of the gold surface, which can result in a mixed potential
that is in the passive region, as shown in Figure 15.6. Lead nitrate is general-
ly added to the leaching circuit on a number of plants in an attempt to
minimize the deleterious effects of sulp'hide ions (Section 15 .2.2) on the
leaching of gold. An additional function of the lead could be to reduce the
possibility of passivation by the mechanism outlined in Section 15 .1.3 .1 (c).
15.2.2 The effect of other constituents of the pulp

The poor extraction of gold by cyanidation can, in many cases, be attributed
to mineralogical factors, such as incomplete liberation of the gold particles
(Feather and Koen, 1973), the presence of tarnished or coated gold (Viljoen
and Mihalik, 1968), or galvanic contact between gold and pyrite particles,
which can result in passivation of the gold surface (Filmer, 1982). The first
two of these effects are discussed in other chapters.
Chemical effects that have an adverse influence on either the rate or
extent of cyanidation can usually be attributed to one of the following factors:
i) the depletion of cyanide and oxygen in the pulp due to reaction with
the constituents of the pulp,
ii) the adsorption of poisons on the surface of the gold particles, which
can retard anodic dissolution of the gold, and
857
iii) the presence in the pulp of carbonaceous material which can absorb
aurocyanide ions.
The literature contains a large amount of information on the cyanida-
tion of gold, which requires careful analysis if erroneous conclusions are to
be avoided. For example, the deleterious effect of nickel on cyanidation
reported by Beyers (1936) is probably due rather to the consumption of
cyanide that occurs as a result of the formation of the very stable complex
Ni(CN);- than to any possible effect of the complex itself (Osseo-Asare et
al., 1984a). Furthermore, much of the data has been gathered from ex-
periments in which large samples of gold (as leaf, foil, or in polished sec-
tions) are contacted with relatively small volumes of cyanide electrolyte for
short periods (1 to 2 hours). In such experiments the preparation of the gold
surface can influence the rate of cyanidation during the time of measure-
ment, and may not be representative of actual leaching practice.
15.2.2.1 Iron sulphide minerals
The common iron-containing sulphide minerals that are usually associated
with gold are pyrite, marcasite, and pyrrhotite. These minerals decompose
in aerated alkaline cyanide solutions to various extents, their reactivity in-
creasing in the order pyrite < marcasite < pyrrhotite. In the absence of
cyanide ions, the reaction products that have been observed include ferric
hydroxide, sulphide, elemental sulphur, thiosulphate, polythionates, and
sulphates (Hedley and Tabachnick, 1968). The formation of the higher ox-
idation states 9f sulphur consumes both lime and oxygen, e.g.
2FeS + 2Y202 + 2H2 0 + 20H- = 2Fe(OH) 3 + SzOi-, (15.52)
and
2FeS2 + 7 Yi 02 + 80H- = 2Fe(OH)3 + 4so;- + H20. (15.53)
As a result of the high rate of oxidation of pyrrhotite, the dissolved oxygen
is rapidly consumed and the concentration of oxygen may become very low.
Under these conditions, pyrrhotite may dissolve non-oxidatively, producing
sulphide ions and ferrous hydroxide (Gardiner, 1933):
FeS + 20H- = Fe(OH)2 + s 2-. (15.54)
The rates of all these reactions are enhanced by high alkalinities.
In the presence of cyanide ions, both pyrite and pyrrhotite are unstable
with respect to oxidation, yielding soluble ferrocyanide and thiocyanate ions
as the major products, e.g.:
FeS + Yi02 + 7CN- + H20 = Fe(CN)~- + SCN- + 20H-, (15.55)
although thiosulphates, polythionates, and sulphates are also produced. It
has been reported that ferrocyanide ions inhibit cyanidation (Beyers, 1936;
Taylor, 1983), although other workers (Holloway, 1982; Osseo-Asare et al.,
1984a) have stated that ferrocyanide has no effect on cyanidation. The lat-
ter view is probably correct, but the consumption of oxygen and cyanide re-
quired for the formation of ferrocyanide could appreciably affect the cyanida-
tion of gold.
858
Although ferrocyanide ions are thermodynamically unstable with respect

to their oxidation to ferricyanide ions (E° = 0,46 V) by dissolved oxygen
(E = 0,58 V at a pH of 11), the rate of oxidation is very low (Hedley and
Tabachnick, 1968). Strong oxidants such as permanganate ions and ozone
oxidize ferrocyanide very rapidly, and the resulting ferricyanide can oxidize
gold in the presence of cyanide ions (Putnam, 1950).
There is no general agreement as to the best procedure to be used in
the cyanidation of ores containing iron sulphide minerals. In many cases,
the addition of sufficient quantities of air and cyanide may be all that is re-
quired for adequate dissolution of the gold. If the consumption of cyanide
is excessively high, aeration in alkaline solution prior to the addition of
cyanide may be required (Gardiner, 1933; McLachlan et al., 1946). The aim
of this treatment is partial oxidation of the surface of the sulphide particles
and the production of a surface film of ferric hydroxide that will inhibit fur-
ther leaching of the sulphides upon the addition of cyanide ions. It is impor-
tant to note that ferric hydroxide is practically insoluble under the condi-
tions of cyanidation, since the concentration of soluble ferricyanide in
equilibrium with ferric hydroxide at a pH of 11 and a free cyanide concen-
tration of 10- 3 mo11- 1 is 1,3 x 10- 4 mo11- 1:
Fe3 + + 30H- = Fe(OH) 3 K 1 = 4 x 1038 (15.56)
Fe3 + + 6CN- = Fe(CN)~- K 2 = 5 x 1043 (15.57)
[Fe(CN)i-] = K1 [CN]6 .
6 Ki [OH]3
As a result, the consumption of cyanide is greatly reduced. During pre-
aeration, the pH value should be maintained below 12 so that the formation
of excessive amounts of sulphide ions, which have a deleterious effect upon
cyanidation, can be avoided (see below).
15.2.2.2 Copper sulphide minerals

The most common copper sulphide minerals associated with gold are
chalcopyrite and chalcocite. In a typical cyanide electrolyte, chalcocite is
somewhat more reactive than chalcopyrite, although both are oxidized to
soluble copper cyanide complexes and thiocyanate (McLachlan et al., 1946).
These minerals may therefore retard the cyanidation of gold, due to their
consumption of oxygen and cyanide ions.
The only stable oxidation state of copper in the presence of cyanide ions
is + 1, since any cupric ions are reduced to cuprous ions by the oxidation
of cyanide ions to cyanogen. An increase in the concentration of cyanide
ions results in the formation of the cpmplexes
Cu(CN);-, Cu(CN)i-, Cu(CN)!-,
Cu(CN);- being the predominant species under the conditions of cyanida-
tion (Osseo-Asare et al., 1984a; Shantz and Reich, 1978).
Hedley and Kentro (1945) investigated the effect of soluble copper in
cyanide solutions on the dissolution of gold from ores and pure gold. The
859
rate of dissolution in an electrolyte containing cyanide and copper at a molar

ratio of 3 to 1 was about 60% of the value obtained with an electrolyte with
a pure sodium cyanide content of 0,01 to 0,05%. At a molar ratio of 4 to
1, the rate of dissolution approached that obtained with the pure electrolyte.
Similar results have been reported by McLachlan et al. (1946).
The conclusion to be drawn from these results is that the cupro-cyanide
complexes have no effect upon the cyanidation of gold provided that suffi-
cient excess free cyanide is available. Stamboliadis et al. (1976) concluded
that the presence of soluble cupro-cyanide complexes retards the dissolution
of gold, even when sufficient lime, cyanide, and oxygen are present. However,
the titration method they used to determine the concentration of free cyanide
ions may have yielded erroneously high values (Shantz and Reich, 1978).
15.2.2.3 Arsenic sulphide minerals

The common arsenic sulphide minerals associated with gold are arsenopyrite
(FeS2 .FeAs2 or FeAsS), realgar (AsS), and orpiment (As2S3). These minerals
are all decomposed in oxygenated alkaline solutions, the reactivity increas-
ing in the order FeAsS < AsS < As2 S3 (Hedley and Tabachnick, 1968).
The stable oxidation products are HAso;-, so;-, and Fe(CN)~- (in the
case of arsenopyrite), e.g.
2AsS + 80H- + 5Y202 = 2HAso~- + 2so~- + 3H2 0. (15.58)
2FeAsS + 40H- + 6Yz02 + 12CN- = 2Fe(CN):- + 2HAso;- +
2SO~- + Hp. (15.59)
These reactions consume large quantities of oxygen, lime (AsS and As 2S3),
and cyanide (FeAsS), and may therefore retard the cyanidation of gold to
a large extent. During the cyanidation of ores containing these minerals, many
other compounds, in addition to those mentioned above, are produced, e.g.
s 2 - , SCN-, Spi-, Ass;- and HAsoi- (Nagy et al., 1966). If an adequate
supply of oxygen is available, however, the intermediate products such as
thioarsenite, Ass;-, and arsenite, HAsoi-, will be oxidized to arsenate
ions.
Apart from the consumption of oxygen, there appears to be no evidence
to indicate that any of the arsenic-oxygen compounds has any chemical effect
on the cyanidation of gold. Hedley and Tabachnik (1968) have suggested
that the adsorption of thioarsenite ions onto the surface of the gold particles
may retard dissolution. The deleterious effect of sulphide ions on
cyanidation is generally acknowledged (see below). The addition of lead
nitrate has been shown to enhance the oxidation of both sulphide and
thioarsenite ions.
15.2.2.4 Antimony sulphide minerals
The only antimony sulphide mineral associated with gold is stibnite, Sb2 S3 ,
the chemistry of which is very similar to that of As2S3 • In alkaline solutions
stibnite is oxidized to antimonites (HSbOi-), antimonates (HSbo;-), and
thioantimonites (Sbs;-), which consume lime and oxygen. The rate of
oxidation of stibnite, like that of arsenic minerals, is enhanced by a high
860
pH value (i.e. above 12), (Muir et al., 1984).

Hedley and Tabachnick (1968) have suggested that thioantimonites can
retard the dissolution of gold by being adsorbed onto the surface of gold
particles. Aeration in alkaline electrolytes (pH 10 or lower) prior to the ad-
dition of cyanide ions can result in improved dissolution of goltj. Lead nitrate
can be added to accelerate the oxidation of sulphide and thioantimonite ions.
If the antimonial ore is roasted for the removal of volatile antimony
oxides, the temperature and air flow must be carefully controlled if the pro-
duction of metallic antimony, which can coat the surface of the gold par-
ticles, is to be avoided (Nagy et al., 1966). Ores containing silver may form
'antimony glasses' (xAg20.ySb2S3) during roasting (Matsukawa and Sakai,
1956).
15.2.2.5 Anions containing sulphur

The sulphur anions that are known to be strong 'poisons' in the cyanidation
of gold are sulphide, thioarsenite, and thioantimonite ions. Fink and Put-
nam (1950) showed that concentrations of dissolved sulphide as low as
0,5 mg 1- 1 can halve the rate of leaching, and attributed this retardation to
the formation of insoluble aurous sulphide on the surface of the gold par-
ticles. Hedley and Tabachnick (1968) suggested that sulphide, thioarsenite,
and thioantimonite ions could be strongly adsorbed onto gold surfaces, thus
rendering the surfaces inert towards cyanide and oxygen.
Although these sulphur anions are thermodynamically unstable towards
oxidation by dissolved oxygen in alkaline solutions, the rate of oxidation can
be remarkably low. The addition of lead (as the nitrate, acetate, or litharge)
can enhance the rate ·of oxidation markedly. In the case of sulphide ions,
this effect has been attributed to the presence of precipitated lead sulphide,
which presents a large surface area for the reduction of oxygen and the
oxidation of the residual sulphide (Paul, 1984).
Thiosulphate and polythionate ions are produced during the oxidation
of many sulphide minerals, and are unstable towards further oxidation if
dissolved oxygen is present, e.g.
s 2oi- + 202 + 20H- = 2so;- + H 10. (15.60)
They also react slowly with cyanide in the presence of dissolved oxygen,
forming thiocyanate:
2sp~- + 02 + 2CN- = 2SCN- + 2so~-. (15.61)
Although these compounds can consume lime, oxygen, and cyanide ions,
they do not appear to exert any direct influence on the dissolution of gold.
Thiocyanate ions can be produced in relatively large amounts by the
oxidation of some sulphide minerals. These ions do not appear to have any
effect on the cyanidation process.
15.2.2.6 Cyanide complexes of various base metals

The presence of the cyanide complexes of Fe2+, Ni2+, Cu+, Zn2+ and Mn2 +
have no direct effect upon cyanidation, although the formation of these corn-
861
plexes may deplete the leach liquor of free cyanide, and retard the dissolu-
tion of gold. Ferricyanide is an oxidant that can actually leach gold (Put-
nam, 1950).
Low concentrations of Pb 2 +, Hg2 +, Bi3 + and TI+ can enhance the
leaching of gold under certain conditions (Fink and Putnam, 1950; Cathro,
1963). The electrochemical aspects of this effect have been discussed by Nicol
(1980a). It is generally agreed that the anodic dissolution of gold in cyanide
solution is a multistep reaction that involves the reactions
Au + CN- = AuCNacts + e, (15.38)
AuCNacts + CN- = Au(CN);. (15.39)
If the rate of the electrochemical production of the adsorbed AuCNacts is
higher than the rate at which the adsorbed species is removed as Au(CN);,
a macromolecular layer of AuCN may be formed on the surface of the gold.
The presence of small amounts of the heavy-metal ions cited above may in-
hibit the formation of this passivating layer, thereby enhancing the rate of
the overall reaction.
15.2.2.7 Calcium ions

Calcium ions are introduced into the cyanide pulp in the form of lime, which
is the most commonly used protective alkali, and as calcium cyanide. Bar-
sky et al. (1935) reported that the rate of cyanidation decreased when the
pH value was raised above 11 with lime, and that almost no leaching of gold
occurred at a pH of 12.2. However, if the pH value was adjusted with sodium
hydroxide, the rate of cyanidation was unaffected for pH values up to 12.5.
On the other hand, the addition of calcium chloride.to an alkaline cyanide
solution had no effect on the dissolution of gold.
It was concluded that high concentrations of both calcium and hydrox-
ide ions must be present for the effect to be observed. The reason for this
effect is not clear. It has been suggested that the retardation is due to the
precipitation of calcium peroxide at the gold surface (Kameda, 1949), but
this is most unlikely.
15.2.2.8 Carbonaceous material

The deleterious effects of carbonaceous material or 'preg-robbing' constit-
uents in the pulp, is well known, and only some pertinent review papers are
cited (Osseo-Asare et al., 1984b; Nagy et al., 1966; and Eisele et al., 1983).
Carbonaceous material can reduce the recovery of gold by restricting the
release of gold from the carbonaceous matrix, or by adsorbing dissolved gold
from the leach liquor (Osseo-Asare et al., 1984b). The treatment of these
ores involves roasting, addition of kerosene or heavy oils, flotation, and
aqueous chlorination.
Carbonaceous matter appears to consist of heavy hydrocarbons, which
do not interact with gold, an elemental carbon component, which behaves
like activated carbon and therefore adsorbs gold, and organic acids (notably
humic acids), which can form complexes with dissolved gold (Osseo-Asare
et al., 1984b).
862
The functional groups on the surface of the elemental carbon compon-

ent consist of carboxylic (R- COOH), phenolic- hydroxyl (R- OH), and
q uinone-type carbonyl (R - C = 0) groups (R being an aliphatic or cyclic com-
pound). In alkaline solutions, these groups exhibit a slight negative charge,
and it is not clear how interaction with the negatively charged aurocyanide
ion occurs. The adsorption of gold is enhanced by the presence of calcium
ions, and may therefore require the formation of ion pairs, as has been sug-
gested for the adsorption of gold by activated carbon (Davidson, 1974).
Humic acids also contain functional groups of the type found in the
elemental carbon component, and exhibit similar behaviour. The mechanism
of the interaction between aurocyanide ions and humic acids remains
unresolved.
15.2.2.9 Flotation reagents and other surface active reagents

Gold bearing ores are often subjected to flotation before cyanidation. When
the flotation concentrate is cyanided without prior calcining, the dissolution
of gold is often found to be poor. Whether this phenomenon is due to the
'lock-up' of finely divided gold in the sulphide minerals, the consumption
of oxygen or cyanide by the corroding minerals, the presence of sulphide
ions in the leach liquor, the presence of residual flotation reagents, or any
combination of the above factors, is difficult to elucidate. It is not surpris-
ing, therefore, that relatively little has been published on this subject.
In a review of this subject, Finkelstein (1972) concluded that thiol-type
collectors (xanthates, dithiophosphates, etc.) can be strongly adsorbed onto
gold, thus rendering the surfaces passive and retarding the anodic dissolu-
tion process.
15.2.3 Other leaching processes

In recent years, increasing emphasis has been placed on protection of the
environment from pollution, on the treatment of refractory gold ores, and
on underground processing. This has led to renewed interest in the develop-
ment of leaching processes that are as efficient and as cost-effective as
cyanidation, but more environmentally acceptable and less susceptible to the
deleterious effects of some constituents of refractory ores.
The thermodynamic behaviour of some of the more important gold com-
plexes is summarized in Figure 15.15, which compares the standard reduc-
tion potentials of the gold(III) - gold and gold(I) - gold couples (Nicol,
1980b). The points below the diagonal line correspond to ligands, L; that
form complexes with gold(III), AuL~- which are more stable than the
corresponding gold(!) complexes, AuL~-. The latter are the preferred species
for those ligands above the line. The standard potentials for the ferric-ferrous
couple in acid solution and the oxygen-water couple in acidic and basic solu-
tions are also shown.
The more important of these potential lixiviants, excluding cyanide, will
be discussed in relation to this figure.
863
15.2.3.4 Chloride
Chlorination was widely used for the leaching of ores during the second half
of the nineteenth century. With the advent of the cyanide process, and the
steadily decreasing grade of the ores being treated, this method has essen-
tially disappeared, and is at present employed only in the refining of gold.
As can be seen from the data in Figure 15.15, the oxidation of gold in chloride
solutions occurs at potentials above 1,2 V, and this requires strong oxidants
such as chlorine or ozone, the former generally being employed. The use of
acidic solutions during chlorination overcomes any passivation problems,
since the mixed potential is below that at which oxide films form on a gold
surface under these conditions (see Figure 15 .4). As shown in Section
15 .1. 3.1 (a), the oxidation of gold in chloride solutions proceeds through the
intermediate formation of gold(I) and, in the presence of excess chlorine,
gold(III) is the final product:
3
Au + -2c12 + c1- = Auel;. (15.65)
The rate of dissolution of gold by chlorination is about two orders of
magnitude greater than that in cyanidation mainly because the solubility of
chlorine is greater than that of oxygen in aqueous solutions.
Under the strong oxidizing conditions employed during chlorination,
the dissolution of other metals in the ore is appreciable, even sulphur, in
the form of sulphide minerals, being oxidized to sulphate. The excessively
high consumption of chlorine by these unwanted reactions renders chlorina-
tion uneconomical except under exceptional circumstances such as the
recovery of gold from antimonial slags (Muir et al., 1983). An additional
disadvantage is that silver chloride is only partially soluble in concentrated
chloride solutions, and therefore reports to both the leach residue and the
pregnant solution.
15.3 The Recovery of Gold from Solution

15.3.1 Cementation
The zinc-cementation process for the recovery of gold and silver from cyanide
solutions was introduced in 1890. During the following thirty years, three
major modifications were introduced that improved the efficiency of the pro-
cess. These innovations involved the addition of lead salts to the pregnant
solution (1894), the use of zinc dust in place of zinc shavings (1897), and
de-aeration of the pregnant solution prior to cementation (1916). The
historical review compiled by Leblanc (1942) on the recovery of gold from
pregnant liquors gives an excellent account of these early developments.
Once the major metallurgical and economic shortcomings of the pro-
cess had been resolved, the incentive for further research in this area largely
fell away. Fortunately, the rekindled interest in electrochemistry which began
in the 1960's led to research into the mechanism of cementation, and much
of the mystique has been stripped away.
The thermodynamic and kinetic aspects of the cementation process have
866
-1.5 -1.0 -0.5

Potential E, V
Figure 15 .16. Current-potential curves for the major reactions

which occur during the cementation of gold onto zinc.
already been introduced. The major reactions are the cathodic deposition
of gold onto the surface of the corroding zinc particles:
Au(CN); + e = Au + 2CN-. (15.66)
Zn + 4CN- = Zn(CN);- + 2e. (15.67)
Other reactions that can influence the cementation process are the reduction
of water and dissolved oxygen:
2H2 0 + 2e = 20H- + H 2 , (15.68)
and
(15.69)
The flow of electrons (i. e. electrical current) required to sustain the three
cathodic reactions (15.66), (15.68) and (15.69) is supplied by the oxidation
of the zinc. The current-potential curves corresponding to these reactions
are shown in Figure 15.16.
The potential at the surface of the zinc particles (the mixed potential,
Em) is uniquely determined by the currents of all four reactions. Cementa-
tion usually occurs with a mixed potential between - 1,0 and - 1,2 V (Nicol
et al., 1979). Over this range, the rates of reactions (15.66) and (15.69) are
controlled by the transport of cyanide ions and oxygen to the surface of the
zinc particles (Kakovskiy and Shcherbakov, 1967), whereas reactions (15 .67)
and (15 .68) are under activation control.
867
An increase in the concentration of cyanide ions causes the curves for

both the dissolution of zinc and the deposition of gold (and thus, Em) to
shift in a negative direction by 0,12 V for every tenfold increase in the cyanide
concentration. This shift increases the evolution of hydrogen (Equation
15 .68), thus increasing the consumption of zinc. Furthermore, the genera-
tion of gas bubbles can disrupt the flow of solution 'through the zinc bed
and reduce the overall efficiency of the process (Paul, 1985). A decrease in
the concentration of free cyanide causes the potential to shift in a positive
direction, and may cause the dissolution of zinc to become controlled by the
transport of cyanide ions to the zinc surface. Under these conditions, the
concentration of cyanide ions at the zinc surface may be so low that zinc
hydroxide is precipitated, which prevents any further dissolution of the zinc.
The deleterious effects of excessive and insufficient concentrations of free
cyanide were reported by Leblanc (1942), who recommended that the con-
centration of free cyanide should be maintained at between 0,006 and 0,015 OJo
sodium cyanide.
The pH of the pregnant solution in a conventional gold plant usually
lies in the range 10,5 to 11,0. Nicol et al. (1979) have shown that the current-
potential curve for the anodic dissolution of zinc is shifted in a negative direc-
tion with rising pH. In addition, at pH values above about 11, the mass-
transport peak is increased, owing to the formation of soluble zincates and
mixed hydroxy-cyanide complexes. The resulting negative shift in Em can be
beneficial if the deposition of the gold was not entirely under mass-transport
control. Leblanc (1942) reported that the cementation of gold was practical-
ly constant over the pH range 8 to 11, but that a marked improvement could
be obtained by an increase in pH to a value between 11,5 and 11,9. At very
high pH values (i.e. above 13), which may be encountered in carbon-in-pulp
(CIP) eluates, the negative shift in Em results in the evolution of excessive
hydrogen, which has a deleterious effect on cementation. Plaskin et al. (1948)
reported that the recovery of gold was maximized at a sodium hydroxide
concentration of 0,025% (a theoretical pH value of 11,8).
Concentrations of Pb2 + as low as 1 to 2 mg 1- 1 have been shown to
shift the current-potential curve for the deposition of gold in a positive direc-
tion by as much as 0,20 V (Nicol et al., 1979; Kirk and Foulkes, 1984; Mcin-
tyre and Peck, 1976). The mechanism associated with this considerable
enhancement is not clear. Similar effects have been found with thallium, mer-
cury and bismuth ions (Mcintyre and Peck, 1976). Other benefits arising from
the addition of lead include reduced zinc consumption (Leblanc, 1942), pro-
bably due to an increased overpotential for the evoluti'on of hydrogen, and
the precipitation of any sulphide ions as PbS (see below). The addition of
lead to above 10 mg 1- 1 may actually retard the cementation of gold (Nicol
et al., 1979; Leblanc, 1942), as the result of a combination of factors: the
effect of the lead on the dissolution of the zinc and the difficulty with which
gold nucleates on a lead surface.
Inefficient de-aeration of the pregnant solution accelerates the corro-
sion of zinc. If insufficient free cyanide is present in the solution . zinc hydrox-
ide will be produced, which results in passivation of the zinc s.urface. Even
868
if passivation does not occur, a positive shift in Em due to the additional

cathodic current may adversely affect the deposition of gold.
The zinc metal used for cementation must clearly have a large surface
area, because the overall rate of any electrochemical solid-liquid reaction is
directly proportional to surface area. Zinc dust fulfils this requirement, but
is readily oxidized by atmospheric oxygen during storage. The coating of zinc
oxide must be dissolved before cementation can occur,
ZnO + H 2 0 + 4CN- = Zn(CN)~- + 20H-, (15.70)
which accounts for the frequent necessity for circulation of the pregnant t>olu-
tion for considerable lengths of time at the start-up of a Merrill or Stellar
filter. High concentrations of free cyanide may be beneficial during this
period.
The most common impurities including sulphite, sulphate, thiosulphate,
ferrocyanide, zinc cyanide, thiocyanate, copper, nickel and cobalt, appear
to have little or no effect on the cementation process (Nicol et al., 1979;
Leblanc, 1942), although deleterious effects can be observed if the concen-
trations are very high. Plaskin et al. (1948) reported that copper adversely
affected the cementation of gold, which almost ceased at a copper concen-
tration of 200 mg 1- 1 • However, the deleterious effect of copper decreases
as the concentration of cyanide is increased, from which it can be inferred
that the effect of copper may be due to a reduction in the concentration of
free cyanide as a result of the formation of cyanide complexes. Hancock and
Thomas (1954) reported that nickel at concentrations higher than
200 mg 1- 1 has a slightly retarding effect on cementation.
The only species that have a marked deleterious effect on cementation
appear to be sulphide ions and soluble compounds of arsenic and antimony
(Nicol et al., 1979; Leblanc, 1942; Plaskin et al., 1948). The effects are observ-
ed even at very low concentrations (1 mg 1- 1 and lower). The poisoning ef-
fect of sulphide ions is thought to be due to the precipitation of insoluble
zinc sulphide on the surface of the zinc particles. No satisfactory mechanism
has been proposed for the effect of arsenic and antimony.
Aluminium is sufficiently electronegative to reduce aurocyanide ions to
gold, and its use as a cementing agent was patented by Moldenhauer in 1893.
The factor mitigating against its use is the dissolution reaction
Al + 40H- = Alo; + 2H2 0 + 3e, (15.71)
which requires the pH value of the solution to be above 12 in order that
Al3 + can be maintained in a soluble form. Furthermore, calcium aluminate
has a low solubility, and the use of sodium hydroxide or sodium carbonate
is required to adjust the pH; the use of lime must be avoided. A historical
review of the use of aluminium for the cementation of gold has been com-
piled by Nagy et al. (1966).
15.3.2 Extraction by activated carbon, ion-exchange resins or organic

solvents
These three processes have very similar broad metallurgical objectives. In
869
each case, an aqueous solution containing a mixture of ions is contacted

countercurrently with an immiscible extractant phase. The extractant phase
either contains chemical functional groups or a heterogeneous surface, which
enable the desired metal ion to be extracted (selectively, if possible) from
the aqueous phase into the extractant phase. The reactions are reversible so
that, after separation of the extractant from the aqueous phase, the desired
metal ion can be subsequently stripped into another aqueous phase. The strip-
ping step should result in the production of a purified and more concentrated
aqueous metal-ion solution, and should allow the extractant to be recycled
to the aqueous feed solution.
In ion exchange, the extractant is a solid organic polymer; in solvent
extraction, it is a water-immiscible organic liquid; and in activated carbon
the raw material is granular coconut shell, fruit pits, coal, peat, lignite, or
wood, which is first charred at 300 to 400°C and then activated (usually in
an atmosphere of steam) at 800 to 950° C. In principle, all three techniques
could compete favourably with the traditional zinc cementation route for
the recovery of gold from cyanide solutions. It is possible, in theory at least,
to achieve greater selectivity than can be achieved with zinc, and it is also
practical, in each case, to recycle the extractant after the gold has been
recovered, because the extractant generally does not undergo chemical change
during a loading-stripping cycle. This is not the case with the zinc route.
Another major advantage is that, in principle, each of the three techniques
can be applied directly to pulps - the carbon-in-pulp (CIP), resin-in-pulp
(RIP), and solvent-in-pulp (SIP) processes - which enables the costly filtra-
tion and clarification stages to be dispensed with, and the attendant losses
of soluble values in the filter cake to be avoided. As discussei:l elsewhere in
this chapter, the CIP process has been introduced in all the major gold-
producing countries of the world and is now the preferred route for the
recovery of gold in all new gold plants in the western world. The RIP pro-
cess has been designated the preferred technique for use on new plants in
the Soviet Union (Guchetl and Lezgintsev, 1968), and is currently attracting
interest in the West. The major problem with SIP is the loss of organic solvent
due to its entrainment in the pulp, but a promising development is the SIP
contactor in which these losses are minimized (Byerlee, 1980). At the time
of writing, however, SIP is regarded rather as an engineering curiosity than
a serious competitor with the zinc-cementation, CIP, or RIP processes.
The chemistry of these processes is complex and can be treated here only
in the barest outline. For further details, the reader is referred to works
devoted specifically to these subjects (Helferrich, 1962; Marcus and Kertes,
1969).
15.3.2.1 The activated-carbon process

(a) The adsorption of gold from cyanide solution
The use of activated carbon in the recovery of gold has undergone a major
revolution in the past ten years. In particular, the CIP process, which in-
volves the use of granular activated carbon for the extraction of dissolved
gold directly from pulps or slimes, has been shown to offer significant cost
870
advantages over the conventional gold-recovery process in most applications,

and the CIP process is now used in all the major gold producing countries.
With CIP technology, as with most new technologies, a fundamental
understanding of the mechanisms involved in the adsorption and elution of
aurocyanide onto carbon has lagged somewhat behind the engineering and
implementation of the process. For all the first-generation CIP plants,
therefore, it was necessary to adopt empirical or semi-empirical approaches
in the selection of activated carbon products and the derivation of optimum
operating conditions. In the last few years, however, a number of publica-
tions dealing with the fundamental aspects of the process have appeared and,
although the various research groups active in the field have not reached con-
sensus, a coherent picture is beginning to emerge.
Most of the studies have focused on the interaction between the auro-
cyanide ion and activated carbon, and a number of thermodynamic and
kinetic characteristics of the reaction have been identified. Some of these
characteristics contradict one another when viewed from the perspective of,
for example, the simple electrostatic interaction between metal anions and
anion-exchange resins, and no mechanism that adequately accommodates
all the observations has been developed. The reaction appears to be fairly
unique, and efforts to develop a mechanism in terms of other well-known
heterogeneous processes have failed.
A number of experimental observations relating to the aurocyanide-
carbon interaction have been drawn from Davidson (1974), Davidson and
Duncanson (1977), Dixon et al. (1978), McDougall et al. (1980), Fleming and
Nicol (1984), and Tsuchida et al. (1984), and are summarized below.
Kinetic characteristics. During thermal activation, carbon develops a
porous graphitic structure of molecular dimensions, which creates an enor-
mous internal surface area of the order of 800 to 1200 m2 g- 1 • The internal
pore~ are classified as macropores ~500 to 200 OOO A), mesopores (100 to
500 A), and micropores (8 to 100 A). Evidence suggests that at least 90%
of the total surface area of the activated carbons used in gold processing
is in the micropore category, and that this surface is either inaccessible to
the fairly large aurocyanide molecule, or is accessible only after extremely
slow diffusion. This property of activated carbon accounts for many of the
observed kinetic phenomena in the processing of gold cyanide.
Plant experience and numerous laboratory investigations have shown
that the initial rate of adsorption of aurocyanide onto carbon is rapid, and
is controlled by the hydrodynamics in the adsorption contactor. This initial
film-diffusion-controlled reaction, which presumably involves adsorption in
the macropores and mesopores, results in the establishment of a pseudo-
equilibrium in 4 to 48 hours*. Subsequently, gold cyanide continues to be
*The exact time taken for the adsorption reaction to reach this pseudo-equilibrium
is governed by factors such as pulp density and viscosity, mixing efficiency, carbon
particle-size and pore-size distribution, and the presence in the pulp of species that
adsorb onto and poison carbon.
871
adsorbed onto the carbon almost indefinitely and, in practice, it is difficult

to establish a true equilibrium. During this period of pore-diffusion-controlled
adsorption, gold cyanide presumbly diffuses slowly into the micropores of
the carbon and, as the cross-sectional area of the micropores approaches that
of the aurocyanide ion, the resistance to mass transfer becomes infinite.
Obviously, gold that diffuses very slowly into the micropores will diffuse
out only very slowly, and it is probably this fact that has led to confusion
regarding the reversibility of the reaction.
From a practical point of view, therefore, the interaction between auro-
cyanide and carbon can be considered to possess two thermodynamic regimes,
the macropore-mesopore equilibrium and the total equilibrium, and the value
of the former could be smaller than the value of the latter by a factor of
at least 10. What is most significant from a practical point of view, however,
is that the pseudo-equilibrium apparently responds to its chemical environ-
ment, and is influenced by the thermodynamics of the adsorption reaction
in much the same way as a true equilibrium would be. It is not a well-defined
entity, however, due to the unhomogeneity of the pore-size distribution from
one carbon particle to another and from one batch of carbon to another.
Therefore, since the pseudo-equilibrium is not well-defined theoretically, and
the true equilibrium is not readily determined experimentally, it is difficult
to model the aurocyanide-carbon interaction accurately on the basis of the
intrinsic thermodynamic properties of the system. Hence, empirical or semi-
empirical approaches based on the rate of adsorption are generally adopted
when modelling this reaction. (Hussey et al., 1978; Cho and Pitt, 1979; Nicol
et al., 1984a and 1984b; Van Deventer, 1984; and Williams and Glasser, 1985.)
It is apparent that the plant-operating parameters should be set in such
a way that the carbon loading is maintained within the macropore-mesopore
equilibrium range. The rate of extraction is fast and responsive to good mix-
ing efficiency in the adsorption contractors, and the rate of elution and the
ultimate efficiency of elution, under a given set of conditions, are also max-
imized. Most plants do, in fact, operate predominantly in the pseudo-
equilibrium range, with average carbon residence times of about 24 hours
per contractor, but there is an obvious economic trade-off between carbon
loading and operating costs in elution and reactivation (Nicol et al., 1984b).
Thermodynamic characteristics. The chemical nature of the surface of

activated carbon has not been firmly established, mainly because of the lack
of a suitable physical technique for the direct study of carbon surfaces. Ac-
tivated carbons have traditionally been divided into two types, those formed
at 700 to 1000°C, which extract acid from water, and those formed at 300
to 400° C, which take up base. Carbons used for gold adsorption are usually
activated between 700 and 800°C, and show both acid- and base-extracting
properties. These properties are consistent with the presence of carboxylic
acid and phenolic groups on the surface and, possibly, in addition to these,
quinone, chromene, and peroxide groups.
The gold-loading capacity of activated carbon increases with increasing
concentration of cations in solution in the order Ca2 + > Mg2+ > H+ >
872
Li+ > Na+ > K+, but decreases with increasing concentration of anions
in solution in the order CN- > s2 - > SCN- > so;- > OH- > c1- >
NO;. A salt such as CaC12 therefore enhances loading, whereas potassium
cyanide inhibits it.
In the presence of aurocyanide or silver cyanide, cations such as Ca2 +,
Mg2+ and Na+ are co-extracted by activated carbon, but they are not load-
ed in the absence of the metal cyanide complexes. However, the concentra-
tion of cations on the carbon is generally too low to accommodate a simple
ion-pair adsorption mechanism.
The adsorption of Ca2 +, Mg2+, Na+ and K+ increases with increasing
pH, whereas the adsorption of aurocyanide decreases. For Ca2 + and Mg2+,
this is associated, to a certain extent, with precipitation of the carbonate com-
pounds within the carbon matrix.
A neutral complex such as Hg(CN) 2 is adsorbed onto activated carbon
as strongly as anionic complexes such as Au(CN)~ and Ag(CN);, but the
adsorption of neutral compounds is not influenced by the ionic strength of
the solution.
The adsorption of copper from cyanide solution onto activated carbon
increases as the pH and the concentration of free cyanide are lowered, i.e.
the loading increases as the degree of cyanide co-ordination and the charge
on the complex anion decrease, in the order
Cu(CN); > Cu(CN)i- > Cu(CN)!-.
The adsorption of aurocyanide onto carbon is accompanied by a shift
in the equilibrium pH of the solution from between 5 and 6 to between 10
and 11, depending on the extent of loading. This could be due to the release
of either OH- or HCO;. There is no shift in pH when Hg(CN) 2 is
adsorbed.
The reduction potential of most commercial activated carbon products
lies between 0,10 and 0,40 V. This is sufficiently low for the reduction of
the gold chloride complex Auel; to the metal (E' = 1,002 V) but riot for
0
the reduction of the aurocyanide complex (£ = - 0,57 V). Nevertheless,
a correlation has been observed between the loading capacity of the carbon
for gold cyanide and its reduction potential, although the correlation for auro-
cyanide is not as well defined as it is for Auel;.
Oxidation of the surface functional groups of activated carbon with,
for example, chlorine or nitric acid, results in a marked decrease in its capacity
for gold cyanide and gold chloride.
The capacity of carbon for gold cyanide is greater in an aerated or
oxygenated solution than in a solution through which nitrogen is bubbled.
The nitrogen content of carbons loaded with gold from aurocyanide solu-
tion is reasonably consistent with the presence of Au(CN);, at gold loadings
of lower than 30 kg t- 1 • At higher loadings, the nitrogen-to-gold
stoichiometry is somewhat lower than 2.
The gold-loading capacity of activated carbon decreases with increas-
ing temperature.
The adsorption of aurocyanide onto activated carbon is a reversible pro-
873
cess in which the equilibria between the concentrations of gold in solution

and on the carbon are equal, whether arrived at from above or below. There
can be a degree of kinetic hysteresis, however, for the reasons outlined in
the section above.
In general, an anion need not be a complexing agent for gold in order
to displace the gold from the carbon. For example, gold that has been fully
loaded onto the carbon from cyanide solution can be completely stripped
with hot sodium hydroxide, and appears in solution as the aurocyanide anion.
The rate of elution is retarded by the addition of salts to the eluate, and
enhanced by increases in the temperature or the addition of polar organic
solvents such as methanol, ethanol, acetone, or acetonitrile to the eluate.
Mechanism of adsorption. Factors that affect the kinetics of adsorp-

tion are the hydrodynamic conditions in the adsorption stages, the particle-
size and pore-size distribution of the carbon, the viscosity and density of the
solution or pulp in contact with the carbon, and the presence of species that
can adsorb onto and poison carbon (Fleming and Nicol, 1984). These fac-
tors would have the greatest influence on the efficiency of gold extraction
in the initial stages of the adsorption process. Factors that influence the
equilibrium loading capacity are the pH of the solution in contact with the
carbon, ionic strength, temperature, and the concentration of free cyanide
(Fleming and Nicol, 1984), and these factors will have an influence on plant
performance only if the loading of the gold on the carbon is allowed to ap-
proach its equilibrium isotherm.
The various mechanisms of gold adsorption from cyanide solution that
have been proposed in the literature generally fall into one of the following
categories:
• the reduction of Au(CN); to metallic gold,
• the adsorption of M+ Au(CN); ion pairs,
• the electrical double-layer adsorption of Au(CN); and cations onto a
charged surface, with partial reduction of the Au(CN); to cluster-type
species, or
• the adsorption of Au(CN);, with subsequent degradation to the in-
soluble AuCN species.
In terms of these mechanisms, gold is present on the carbon surface in
one or more of the following forms:
• aurocyanide, Au(CN); (valency 1),
• gold cyanide, AuCN (valency 1),
• a cluster compound, Au_iCN)Y (valency 0 to 1), or
• gold metal, Au (valency 0).
The presence of sub-stoichiometric gold species Au, AuCN, or

AuJCN)Y on the loaded carbon is countered by the observation that gold
can be eluted from carbon with sodium hydroxide, and appears in the eluate
as aurocyanide. Moreover, it is significant that it is possible to partially elute
carbon with sodium hydroxide even after treatment of the loaded carbon
874
in boiling hydrochloric acid for several hours. These conditions would nor-
mally be more than adequate for the decomposition of aurocyanide to gold
cyanide and the liberation of hydro cyanic acid (Equation 15 .2). The fact that
this does not occur, or occurs only very slowly, when aurocyanide is loaded
onto carbon suggests that the carbon stabilizes the aurocyanide under acidic
conditions, probably as the acid, HAu(CN) 2 • In addition, the thermo-
dynamic evidence tends to counter mechanisms in which gold is reduced to
a valency of less than 1. Analyses of carbons with high gold loadings have
indicated a ratio of N to Au of less than 2. This suggests some chemical change
but, under the sort of loading conditions typical of CIP operations, the weight
of evidence favours a mechanism in which aurocyanide is adsorbed without
chemical change. Moreover, the sensitivity of the adsorption mechanism to
temperature and to the concentration of cations in solution suggests that ad-
sorption does not occur by simple coulombic ion exchange. This is supported
by the fact that a neutral molecule such as Hg(CN) 2 competes very effect-
ively with Au(CN); for adsorption sites on the carbon, and that cations
such as Ca2 + and Mg2+ are co-extracted with aurocyanide. These are
phenomena that would not normally be associated with ion exchange pro-
cesses, but are more consistent with a mechanism in which aurocyanide is
adsorbed as an ion pair of some sort. The evidence suggests that, at high
pH, aurocyanide is adsorbed predominantly as an ion pair of the type
Mn+[Au(CN;ln, (where Mis Na, K, Ca, or Mg, etc.), whereas, at low pH,
the predominant adsorbing species is the acid, H[Au(CN) 2]. Moreover, the
strong dependence of gold adsorption on the pH of the solution indicates
that H[Au(CN) 2] is adsorbed more strongly than M11 +[Au(CN) 2-ln· A
mechanism by which both H[Au(CN) 2] and Mn+ [Au(CN;ln are adsorbed
simultaneously over a fairly wide pH range would explain the shift in pH
that accompanies the adsorption of aurocyanide from neutral solutions, and
would also explain the fact that the ratio of Mn+ to Au on the loaded car-
bon is generally lower than the stoichiometric amount. Such a mechanism
is reasonable in view of the fact that the thermodynamics of adsorption on-
to carbon are more favourable for neutral compounds than for charged
species. Examples include the adsorption of mercury (McDougall et al., 1980)
and copper (Fleming and Nicol, 1984) from cyanide solution, and the ad-
sorption of molybdenum from acid solution (De Wet, 1985). On this basis
it would be predicted that H[Au(CN) 2] will be adsorbed more strongly onto
carbon than Au(CN);, and it is likely that the favourable thermodynamics
of this interaction will permit the uptake of protons from solution and the
formation of H[Au(CN) 2] at higher pH values than might normally be
predicted for this reaction.
It is likely that future significant advances in the understanding of this
mechanism will be made only as a result of the successful application of direct
physical techniques such as X-ray diffraction and infrared spectroscopy.
(b) The adsorption of gold from leach liquors other than cyanide
Alternative lixiviants for gold are becoming increasingly important in most
gold-producing countries and, since activated carbon is likely to play an in-
875
tegral part in any process for the recovery of gold in the future, the loading
and stripping characteristics of these gold compounds with respect to car-
bon will need to be studied in detail. One of the earliest producers to use
this technology is the New England Antimony Mine in Australia, which
recovers gold by thiourea leaching and carbon extraction (Hisshion and
Waller, 1984). Techniques for the elution of gold have not been developed,
however, and the carbon is burnt so that the gold can be recovered.
The standard reduction potentials for a number of alternative lixiviants
are shown in Table 15.2, where they are compared with data for the cyanide
system. On the basis of these reduction potentials, one would predict that
activated carbon with a reduction potential of 0,10 to 0,40 V (McDougall
et al., 1980) would readily reduce the gold chloride and thiocyanate com-
plexes on the carbon surface to the metallic state whereas the gold-thiourea
complex would be a borderline case, and would probably be reduced to the
metal only at thiourea concentrations lower than about 0,1 M. The gold-
thiosulphate complex would probably be too strong to be reduced by carbon.
Recent unpublished results from the laboratory at Mintek indicate, in
fact, that both the gold thiourea and gold thiocyanate complexes are adsorbed
onto carbon without chemical change, from typical leach solutions. Analysis
of loaded activated carbons for sulphur and gold indicates that the ratio of
these two elements is, over a wide range of loading conditions, close to the
theoretical value of 2 that would be expected for adsorption of the complex
ions, Au(CS(NHJJ; and Au(SCN);, without chemical change. Moreover,
the adsorption of the gold thiourea complex on carbon is genetally enhanc-
ed by the presence of anions in solution whereas the adsorption of
Au(SCN); is enhanced by the presence of cations in solution, which in-
dicates that these two species are adsorbed onto carbon by a similar ion-pair
mechanism to that postulated for aurocyanide. The fact that these two ions,
which are relatively unstable compared with aurocyanide, are not reduced
to the metal by activated carbon is further evidence that the very stable auro-
cyanide ion is unlikely to be reduced.
( c) Elution of gold from carbon

Since the adsorption of gold cyanide onto activated carbon is a thermo-
dynamically reversible process, chemical and physical factors that inhibit ad-
sorption will enhance elution, and those that enhance adsorption will inhibit
elution. A number of these factors are utilized in elution on all industrial
CIP plants.
The most important factor is temperature, since the kinetics (activation
energy 66,5 kJ mol- 1) and the thermodynamics (exothermic heat of reaction
40 kJ mol- 1) of elution improve with increasing temperature. All CIP elu-
tion plants operate at elevated temperature, and the modern trend is towards
elution at temperatures in excess of 100° C, with the use of pressurized
equipment.
Another important factor is the influence of ionic strength on the per-
formance of all CIP elution plants. The fact that the rate of elution increases
with increasing anion concentration but decreases with increasing cation con-
876
centrationmeans thatthe rate of elution passes through a maximum as the

concentration of the eluting salt (NaCN, for example) is increased. Optimum
performance is therefore achieved at the concentration of eluting salt that
corresponds to this maximum, but there are complicating factors. One of
these is that the anion that gives the best results, viz. the cyanide ion, decom-
poses fairly rapidly at elevated temperatures (40% in 6 hours, 90% in 24
hours at 95°C). This effect is best minimized by the addition of the cyanide
in increments throughout the elution cycle rather than in one dose at the start.
A second complication arises in the electro-elution process, in which gold
is continuously extracted from the eluate by electrolysis before the eluate is
recycled back to elution. The requirements for good electro winning efficiency,
especially with regard to the ionic strength (conductivity) and the recirculating
flowrate of the eluate, are generally quite different from the requirements
for good elution efficiency, and a measure of compromise is necessary.
In seeking to overcome some of these obstacles, the Anglo American
Research Laboratory (AARL), which developed the AARL elution method
(Davidson and Duncanson, 1977), has moved from a continuous to a batch
operation in which the carbon is first soaked in a hot concentrated solution
of sodium hydroxide and sodium cyanide and then treated with hot deioniz-
ed water to elute the gold. The water is passed through the column only once,
and the rate of elution is sufficiently fast for good elution efficiency to be
achieved with reasonably small volumes of water. The conductivity of the
eluate is too low for effective electrolysis unless the pre-soak solution is com-
bined with the eluate after elution. Alternatively, this method is particularly
suited to different methods of gold recovery such as chemical reduction (zinc
cementation), which allows a portion of the pre-soak solution to be re-used
several times, thus conserving the chemical in that solution. This approach
has been adopted on several plants.
Another factor that can be utilized in enhancement of the rate of elution
is the catalytic effect of polar organic solvents such as acetonitrile, acetone,
methanol, and ethanol. This effect is attributed to an increase in the activity
of the cyanide ion and a decrease in the activity of the aurocyanide ion in
polar solvents relative to the situation in water (Tsuchida et al., 1984).
It was shown recently in laboratory tests at Mintek that carbon loaded
from either gold-thiourea or gold-thiocyanate solution can be eluted effect-
ively with sodium sulphide or sodium cyanide solution, and that the rate of
gold elution with sodium cyanide is enhanced by addition of polar solvents
such as acetone or ethanol. Reasonable rates of elution can, in fact, be achiev-
ed even at ambient temperatures and pressures and it is likely that future
process technology for these alternative leaching systems will incorporate car-
bon elution rather than the measures employed at the New England Antimony
Mine.
15.3.2.2 The ion-exchange process

(a) Principles
Modern synthetic ion-exchange resin consists of an inert polymeric hydrocar-
bon matrix (usually of polystyrene cross-linked with divinyl benzene) to which
877
functional groups are chemically attached. The polymer matrix is a three-

dimensional network that swells in contact with aqueous solutions, absorb-
ing water and allowing ions present in the aqueous solution to diffuse through
the resulting gel to the sites of the active groups.
The active groups can be anionic or cationic in character, and the in-
teractions between these groups and cations or anions in solution are almost
invariably coulombic or electrostatic in nature. The rates of the exchange
reactions are therefore high, reversible, and generally diffusion-controlled,
and the selectivity of the resins for one ion over another is predominantly
a function of the size, charge, and polarizability of the ions.
The active groups in anion-exchange resins are either quaternary am-
monium groups for strong-base resins, or primary, secondary, or tertiary
amines for weak-base resins. The active groups in cation-exchange resins are
either carboxylic (weak) acid or sulphonic (strong) acid. Strong-acid and
strong-base resins are completely ionized in the pH range 2 to 12, and ex-
hibit maximum ion-exchange capacity under these conditions. Weak-acid
resins are ionized only at pH> 5, and weak-base resins at pH <9. This pro-
perty of weak-base resins can be a limitation in the processing of gold.
As shown in Table 15 .1, gold in aqueous solutions of hydrometallurgical
interest is generally present as an anionic complex. Consequently, anion-
exchange resins are more important than cation-exchange resins in the field
of gold metallurgy, and will be dealt with in more detail. One important ex-
ception is the aurous thiourea complex, which has a single positive charge
(thiourea being a neutral ligand) and is strongly extracted from aqueous solu-
tion by cation-exchange resins. The use of thiourea in the leaching of gold
ores is gaining momentum overseas and could conceivably become impor-
tant in South Africa in the future. The thiourea leaching reaction takes place
in acid solution (pH 1 to 2), and therefore strong-acid resins are required
for good extraction efficiency. Since the oxidant is generally the ferric ion
(although ozone and peroxide are also effective), ferric and ferrous cations
are present in solution in considerable excess. However, strong-acid resins
are fairly selective for the aurous thiourea cation over iron cations, and good
extraction and separation of gold is possible (Becker et al., 1983). The ex-
traction by strong-acid resins is so powerful, in fact, that desorption with
acids according to the equation
(15.72)
(where T is NH2 CSNH2 , and I - represents the organic matrix), is in-
complete even at high concentration of acid (Becker et al., 1983). Experience
with the technology of anion-exchange resins indicates that the elution of
cation-exchange resins will probably be most effective either coupled with
continuous electrolysis of gold or in the presence of mixed aqueous-organic
solvents.
Because of the importance of the cyanidation process in gold metallurgy,
anion-exchange resins have received far more attention than cation-exchange
resins. Strong-base resins extract aurocyanide anions (and other singly charged
anionic gold complexes) according to the equation
878
(15.73)
The equilibrium in Equation (15.73) is defined as the mass-distribution ratio,
K = [Au(CN);][x-1
(15.74)
sB [Au(CN) 2-][x-1
(where bars denote ions in the resin phase), the position of which is deter-
mined by the relative affinity of the resin for the counter-ion, Au(CN);,
and the co-ion, x-.
Weak-base resins are in the free-base (non-ionized) form at pH values
higher than 9 to 10 and, under these conditions, possess no ion-exchange
characteristics. In neutral or acidic solution the amine is protonated, and
ion-exchange occurs according to the following equations:
I- R2N + HX = I -R2NH+x-, (15.75)
(15. 76)
The equilibrium constant for reaction (15.76) is written
KwB = [Au(CN);][x~ ' (15. 77)

Ka[Au(CN);][x-1
where Ka is the equilibrium constant for the resin protonation reaction

(15.75). In this case the valu,e KwB is influenced not only by the affinity of
the resin for the respective anions, but by the pH of the solution, and this
is the only significant difference between strong- and weak-base resins.
The pKa of weak-base resins is determined both by the basicity of the
amine group and by the nature of the anion in solution. The strength of the
ion pair formed between the anion and the active group on the resin increases
with the charge and the polarizability of the anion, and the driving force
for ion pair formation permits uptake of the proton from far more alkaline
solutions when strong ion pairs are formed. For example, the strong ion pair
formed between an amine and the large, highly polarizable aurocyanide anion
yields a pKa of about 9 to 10 (Fleming and Cromberge, 1984a), whereas the
pKa of the same resin in chloride solution is some 2 log units lower. This
has important implications in the extraction of gold from cyanide leach li-
quors, which normally have pH values between 10 and 11.
An important property of an anion-exchange resin is its selectivity for
aurocyanide, because many complex cyanide anions and other anions are
usually present with aurocyanide in cyanide leach liquors, and the competitive
extraction of these anions can have a profound influence on the efficiency
of gold extraction. The competitive ion-exchange reactions that take place
can be represented by the general equation
z( I - R 3 N)YA + yBz- = y( I - R 3 N)zB + zN'- (15.78)
(where the anion A has a negative charge of y and the anion B has a negative
879
Table 15.3. Selectivity coefficients, kAu(CNJi IA of Au(CN) 2- against various anions, A - , on the
strong-base resin Amberlite IRA 400 (after Aveston et al., 1958).
Anion A kAu(CN)21A
Co(CN)~- 8 x 109
Fe(CN)~- 6 x 109
F- 5,5 x 103
Cl- 6,9 x 102
cw 3 x 102
Ni(CN)~- 2,5 x 102
CNS- 9,6
Ag(CN),- 5,8
c10; 4,6
Au(CN); 1,0
Mn04- 0,53
charge of z). The selectivity of the resin for anion B over anion A is defined
by the coefficient
[BY [A]' (15.79)
kB/A = -- '
[A]z [Bf
which has the same form as the equilibrium constant in Equation (15.74).
Table 15.3 shows values of the selectivity coefficient of a strong-base
resin for aurocyanide over various other anions that are commonly present
in cyanide leach liquors. The concept of a selectivity coefficient is useful for
theoretical studies, but is unfortunately of limited value when the relative
extractions of anions are predicted. In fact, the predictions that might be
made intuitively on the basis of selectivity coefficients bear little relation
to what is found in practice, as illustrated in Figure 15 .17. The isotherms
for various metals shown in Figure 15 .17 were obtained by the equilibration
of a pregnant cyanide solution from a gold plant with a commercial strong-
base resin. It could have been predicted, on the basis of the selectivity coef-
ficients in Table 15.3, that an anion-exchange resin would extract gold
preferentially from a solution containing cobalt(III) and nickel, whereas
Figure 15.17 shows that nickel and cobalt are preferentially extracted.
From a practical point of view, a far more valuable concept is the separa-
tion factor. This is defined as
[B] [A] (15.80)
CXB/A = -
[A] [B]
which is simply the quotient of the concentration ratios of the two counter-
ions in the resin phase and in solution. The numerical value of the dimen-
sionless separation factor is not influenced by the choice of concentration
units or by the charge on the anion, but it does vary with the concentration
of anions A and B in solution.
880
Copper
16 Conditions:
Ratio of solution to resin 500: 1
pH 11.5
Temperature z20 °C
~ 12
.-"<
Gold
0.1 0.2 0.3 0.4 0.5 0.6

Metal in solution, mg/I
Figure I5. I 7. Equilibrium loading of various metal cyanide com-

plexes from a Grootvlei pregnant solution onto a strong-base
resin (AIOI DU) (after Fleming and Nicol, 1984).
Since the preference of the resin for one anion over another is a func-
tion mainly of the properties of the anion, there is not much that one can
do to influence the selectivity in favour of aurocyanide. However, it has been
shown (Clifford and Weber, 1983) that the 'selectivity of anion-exchange resins
for monovalent anions over polyvalent anions is a function of the distance
that separates the fixed charges on the resin matrix. The uptake of a divalent
anion requires the presence of two closely spaced positive charges, whereas
no such constraint exists for monovalent anions. On this basis, the propor-
tion of sites having suitable locations for the adsorption of multivalent ions
will decrease as the capacity of the resin decreases or, in the case of weak-
base resins, as pH is increased. The influence of this phenomenon on the
selectivity of anion-exchange resins for aurocyanide over multivalent metal
cyanide complexes was first reported nearly thirty years ago (Aveston et al.,
1958). Table 15.4 shows that the application of this theory results in greatly
improved separation factors between aurocyanide and a range of multivalent
metal cyanide anions on a weak-base resin.
The data for two weak-base resins show that the separation factor be-
tween aurocyanide and the various multivalent metal cyanide anions increases
by a factor of ten to one hundred upon increasing the pH value of solution
from 6 to 11. In contrast, the separation factors on a strong-base resin are
largely unaffected by pH, as might be expected. It is also interesting to note
881
Table 15.4 Separation factors* between aurocyanide and various metal cyanide anions on two
weak-base resins (A7, Duolite International; IRA 93, Rohm and Haas) and one strong-base
resin (AlOl DU, Duolite International). Concentration of metal in solution 10- 4 M.
aAu(CN); /B
A7 IRA 93 AlOl DU
Anion B pH6 pHll pH6 pHll pH6 pHll
Ag(CN); 7,6 7,9 12,3 15,4 8,8 8,6

Ni(CN)~- 1,9 5,9 0,7 19,0 1,1 0,9
Cu(CN)!- 1,8 35,5 3,3 32,3 8,2 4,9
Co(CN) 63 - 0,8 71,0 0,9 6,3 1,6 0,8
Fe(CN)~- 0,7 6,2 0,8 13,5 2,0 4,4
*Calculated from data reported by Fleming and Cromberge (1984a).
that the separation factor between aurocyanide and the only other monovalent
anion in solution, silver cyanide, is unaffected by a change in pH.
Most weak-base resins have a low and variable concentration of per-
manently charged functional groups. These pseudo-strong-base groups arise
via the cross-linking of adjacent amine groups and have the properties' of
the quaternary amine groups in a strong-base resin. Russian workers
(Laskorin et al., 1977) have observed that certain weak-base resins with a
low strong-base content extract gold very selectively from cyanide solutions
of high pH, but that the selectivity disappears when the strong- base content
exceeds about 20% of the total resin capacity. The mechanism of rejection
of the multivalent complex cyanide anions in this situation is presumably
also based on charge separation.
The elution of weak-base resins is achieved most effectively and cheap-
ly by the treatment of the resin with a fairly dilute solution of sodium
hydroxide. This shifts the equilibrium in Equation (15.75) to the left-hand
side, and restores the amine functional groups to the free-base form. When
electrolysis is coupled with elution, 99% of the aurocyanide can be eluted
from a weak-base resin with 1or2 bed volumes of0,5 M sodium hydroxide
solution in less than 4 hours (Fleming and Cromberge, 1984b). The effec-
tiveness and simplicity of this method is a major factor in favour of the use
of weak-base resins in gold metallurgy, but this advantage has to be weighed
against the fact that their capacity is low compared to that of strong-base
resins, particularly at the pH values (10 to 11) that are usually encountered
in cyanide leach liquors.
The elution of aurocyanide from a strong-base resin is more difficult,
and is achieved either by ion-exchange mass action, i.e. by the reversal of
the equilibrium shown in Equation (15.73), or by the chemical destruction
of the aurocyanide anion. Because aurocyanide is so strongly adsorbed, its
elution by ion-exchange requires a large excess of the co-ion, x- . Anions
such as chloride, bisulphate, nitrate or cyanide can be used, but the rate of
aurocyanide elution is extremely slow unless the activity of these anions is
increased by, for example, the addition of polar organic solvents such as
882
acetone, acetonitrile, or methanol to the eluate. Anions such as thiocyanate

and the zinc cyanide complex, on the other hand, compete favourably with
aurocyanide, and much higher rates of elution can be achieved. When elec-
trolysis is coupled to elution, either of these anions can elute 99% of the
gold under ambient conditions in 6 to 24 hours (Fleming and Cromberge,
1984b).
Because thiocyanate and zinc cyanide themselves load strongly onto
anion-exchange resins, the resin must be regenerated before being recycled
to adsorption. In the case of the zinc cyanide complex, this is achieved ef-
fectively and cheaply by the destruction of the complex with acid (Fleming
and Cromberge, 1984b).
( I- NR3 ) 2 Zn(CN)~- + 2H2S04 = ( I- NR3 ) 2 So~- + ZnS04 + 4HCN.

(15.81)
The zinc sulphate and hydrocyanic acid produced in this reaction are col-
lected in alkaline solution so that the zinc cyanide anion can be reconstituted
for recycling.
The removal of thiocyanate from a strong-base resin can be achieved
effectively by treatment with a ferric salt (Fleming, 1984 and 1985). The
cationic complexes FeSCNz+ and Fe(SCN); are formed, and these are
readily displaced from an anion-exchange resin:
+ - + 2-
2 [ - NR3 SCN + Fez(S04) 3 = ( [ - NR3) 2S04 + Fe(SCN)S04 • (15.82)
Thiocyanate for recycling can be recovered from this solution by the precipita-
tion of the iron from solution as ferric hydroxide. The ferric hydroxide
precipitate can then be redissolved in acid and recycled to regeneration. In
this way the thiocyanate and ferric ions are conserved in the elution-
regeneration cycle (Fleming, 1985). The method is efficient, cheap and non-
toxic, and would probably be preferred to the zinc cyanide method in most
applications.
The alternative method for the stripping of gold from a strong-base resin
involves the chemical conversion of the aurocyanide anion to the aurous
thiourea complex:
+ +
[ - NR3 Au(CN); + 2(NH2CSNHz) + 2H2S04 = [ - NR3 HSO;
+ [Au(NH2CSNH2)2J+ Hso; + 2HCN. (15.83)
This method of elution is favoured in the Soviet Union, but suffers from
the drawback that the other metal complexes that are co-extracted with auro-
cyanide during adsorption are generally poorly eluted, necessitating multi-
elution procedures. This problem is particularly serious when cobalticyanide
and ferrocyanide are formed since, in acidic solution, they are converted to
species that are very difficult to strip from the resin (Fleming and Cromberge,
1984b; Fleming and Hancock, 1979).
(b) Applications
Although the zinc precipitation process is simple and cheap and the technology
is well developed, an anion-exchange concentration and purification step may
883
be preferred in certain instances. For example, the chemical composition of

the leach liquor occasionally does not lend itself to the quantitative or
economic recovery of gold by zinc cementation (e.g. when the concentration
of gold is very low or the concentration of impurities is very high), but the
major potential application of anion-exchange resins in the processing of gold
is in the recovery of aurocyanide direct from the pulp in the RIP process.
The filtration and clarification steps are costly components of the convent-
ional gold process, especially with certain ores that are not amenable to filtra-
tion, and it is likely that the RIP process would offer significant cost advan-
tages in most applications. In this regard, the RIP process is no different
from the CIP process, but in some situations RIP may be preferred to CIP,
and these are discussed below.
The use of anion-exchange resins for the extraction and concentration
of aurocyanide in dilute solutions or pulps was first proposed in the early
1950's, and a number of experimental studies were carried out at that time
(Hussey, 1949; Burstall et al., 1953; Burstall and Wells, 1955). Since then,
further work has been done in all the major gold-producing countries (Davi-
son et al., 1961; Demidov et al., 1967; Tataru, 1974; Stamboliadis et al.,
1978), and a recent study in South Africa (Fleming and Cromberge, 1984a,
1984b and 1984c) concluded that, in many cases, resin-based processes for
the recovery of gold are probably superior to carbon-based processes from
economic, chemical, and metallurgical points of view. This would seem to
be confirmed by reports emanating from the Soviet Union, since a 1978 arti-
cle indicates that one of the largest gold mines in the U.S.S.R. is using an
anion-exchange resin for the extraction of aurocyanide by an RIP process
(Anon., 1978). This is believed to have been the world's first application of
RIP to the extraction of gold, although the article states that 'RIP is now
used widely for gold extraction in the Soviet Union'. The questions that re-
main to be answered, before the large-scale application of RIP to the recovery
of gold becomes a reality, are all related to the engineering of RIP on a large
scale. For example, the mean particle size of resins is smaller than that of
carbon, which means that interstage screening in RIP will have to be carried
out at a finer cut than in CIP (probably 0,60 mm in RIP as against 0,85 mm
in CIP). This gives justifiable cause for concern, since interstage screening
has tended to cause bottlenecks on most large CIP plants. Probably of lesser
concern is the question of resin life in an RIP circuit, since the available
literature on the subject (Laskorin et al., 1978; Hosking, 1984) indicates that
resin losses are likely to be somewhat lower than carbon losses.
RIP will probably make its greatest impact in those areas where the treat-
ment of pulp by CIP poses problems. For example, it has been shown that
RIP is far less responsive to the presence of poisons in the pulp than CIP.
These poisons include organic compounds such as flotation reagents, solvents
and lubricants, and inorganic compounds such as hematite, calcium car-
bonate, shales, and clays. Each of these species retards the adsorption of
gold onto carbon, but apparently has little effect on the extraction proper-
ties of resin.
It is likely that operators treating small tonnages ( < 20 kt per month)
884
will find RIP an.attractive alternative to CIP. The capital costs for a carbon
elution-regeneration plant are high and generally cannot be justified on small
CIP plants. In these cases gold is recovered by incineration of the loaded
carbon, which imposes high operating costs on the process. RIP, with its
potential for relatively inexpensive elution and regeneration (in regard to
capital and operating costs), would be a viable alternative in these situations.
In the South African context, a potential advantage of resins over car-
bon is that they offer a means for the simultaneous recovery of gold and
uranium. Gold and uranium are, in fact, co-extracted on many uranium plants
in South Africa where, more by accident than design, uranium plants are
often used to scavenge gold that would normally be lost as a soluble residue
from the gold plant. This is because aurocyanide is relatively stable in the
uranium-leaching circuit, and both gold and uranium are present in solution
as anionic complexes. Various possible options for the translation of this
interesting chemistry into a viable process have been explored over the years
(Anglo American Corporation, 1977; Fleming and Cromberge, 1984d).
In addition, a number of uranium mines in South Africa have stockpiles
of old, discarded strong-base resin that has lost its ion-exchange capacity
as a result of poisoning or degradation. Although the value of the gold on
the resin often exceeds the value of the resin itself, recovery of the gold by
incineration of the resin has never been practical because of the very toxic
and noxious fumes that are produced. Consequently, the resin has been
stockpiled pending the development of more suitable gold-recovery
technology. One example of such a stockpile is the resin at the Stilfontein
mine, which was last used in the early 1950's in a fixed-bed uranium plant,
and was removed from the circuit owing to extensive poisoning by cobalti-
cyanide. The stockpile contains about 100 t of resin with an average gold
content of 900 g t- 1; a plant for the recovery of the gold from this resin was
commissioned in 1984. The method adopted, which is based on technology
developed at Mintek (Fleming, 1985), involves electro-elution with ammonium
thiocyanate, and has enabled gold recoveries of over 90% to be regularly
achieved. In addition, cobalt is partially stripped from the resin, and much
of the original ion-exchange capacity of the resin is restored after ferric
regeneration. This offers the possibility that the resin in this stockpile can
be returned to active service.
15.3.2.3 The solvent-extraction process

Solvent extraction is a well-established method of separation in extractive
metallurgy. This technique was first applied to the extraction of uranium
for nuclear purposes in the 1940's, and is currently widely used in the recovery
of uranium, copper, zinc, cobalt, nickel, vanadium, tungsten, and
molybdenum, as well as in separations involving the platinum-group metals.
The solvent-extraction process involves the transfer of a solute from an
aqueous to an organic phase, and the subsequent extraction of the solute
back into another aqueous phase under a different set of chemical or physical
conditions. For a species present in aqueous solution to be able to pass into
an organic phase, it must react with an organic component. Metallic ions
885
can occur in aqueous solutions in a variety of forms, either free or as com-

plexes and, in general, only a particular ion, or at most a limited number
of these ions, will react with the solvent and be extracted. The distribution
of the metal between the two phases will therefore be greatly influenced by
the composition of the aqueous phase. The solvent may be a pure organic
liquid or, more usually, a solution of an organic reagent in an inert diluent.
The organic reagent may undergo association, solvation, or other reactions
in the organic phase, and these will also affect the extraction equilibrium.
Solvent-extraction systems can be conveniently classified according to
the nature of the reaction involved in the extraction process, of which the
following types can be identified.
Extraction by physical distribution involves a simple molecular distribu-
tion of the solute between the aqueous phase and an inert solvent (i.e. one
containing no atoms with electron donor properties). Such a situation arises
only for solutes that are poorly solvated in the aqueous phase, and few such
systems are encountered in extractive metallurgy.
Extraction by salvation involves the replacement by the extractant of
some or all the co-ordinated water molecules on a metal cation to form a
species that is soluble in the organic phase. Solvating extractants are most
characteristically oxygen-containing organic molecule such as ethers, esters,
alcohols, ketones, and neutral organophosphorus compounds. This kind of
extraction is restricted to neutral species because it is necessary to preserve
the electrical neutrality of the phases. The extraction of auric chloride from
acid solution (as HAuC14) would be one example of such a process (Walker,
1956).
Extraction by anion exchange is characterized by the use of a suitable
cation for the extraction of an anionic metal compound in tlie form of an
ion pair, after a simple (non-metallic) anion has been exchanged between
the organic and aqueous phases in the reverse direction. Often the cationic
component of the ion pair is formed by the protonation of a compound of
weak basicity, most typically a nitrogen-containing base such as an amine.
Alternatively, the cation may be of the quaternary ammonium, phosphonium,
or arsenium type. Recovery of gold from cyanide, thiocyanate, or thiosulphate
leach liquors would fall into this solvent extraction category.
Extraction by cation exchange involves the formation, by a hydrated
metal cation or complex metal cation, or an electrically neutral co-ordination
complex with the extractant by the displacement of another cation (most com-
monly a hydrogen ion). The formation of the extractable complex can involve
chelation and chemical bonding between the extractant and the metal ion,
or it can involve simple electrostatic interaction between an anionic extract-
ant (such as a sulphonic acid) and a metal cation. When the cation displaced
from the extractant is a hydrogen ion, the extractability of the metal shows
a distinct dependence on the pH value of the aqueous phase. Thus gold could
be recovered from a thiourea leach liquor by cation-exchange solvent
extraction.
The anion- and cation-exchange reactions that take place in the solvent
are the same as those in ion-exchange resins. The major differences between
886
the two techniques arise as a result of the difference in the inert supporting
matrix, which is an organic liquid such as kerosene in solvent extraction,
and a solid organic polymer such as polystyrene in an ion-exchange resin.
This difference can give rise to differences in the rates of anion or cation
exchange, and also in the order of selectivity of the extractant for various
anions and cations but, in general, the metallurgical performance of the two
techniques can be expected to be very similar.
The grade or value of the aqueous solution being treated is probably
the most important consideration in the choice between resins or solvents
for a hydrometallurgical application. Solvent extraction is generally prefer-
red for the treatment of small volumes of high-grade solution, whereas resins
are preferred for the treatment of large volumes of low-grade solution. This
is because solvent is lost in solvent extraction, and this becomes relatively
important with respect to operating costs when large volumes of solution
are processed, whereas resins are effectively insoluble in water and therefore
do not suffer from this drawback. Another important factor is that resins
are able to treat metal solutions containing high concentrations of suspend-
ed solids, whereas costly solution clarification steps are indispensable in any
solvent-extraction process.
For these reasons, solvent extraction is unlikely to find applications in
the large-scale processing of solutions resulting from the cyanidation of low-
grade gold ores, although the process could possibly contribute to the pro-
cessing of high-grade ores or concentrates. Its greatest contribution, however,
is likely to be the final purification and refining of concentrated gold solu-
tions resulting from the leaching of materials such as jewellery and electronic
scrap, anode slimes, and platinum-group-metal (PGM) concentrates.
The traditional route for the industrial separation of the platinum-group
metals comprises a complicated series of operations based largely on selec-
tive precipitations, often of low efficiency (Edwards, 1976). A multitude of
solid-liquid separations and recycle operations is required. The poor once-
through yield of refined metals results in a high inventory of process in-
termediates, which is particularly undesirable in view of the high value of
the final product. Solvent-extraction methods, with their potential for effi-
cient separations and rapid throughput were found to represent an attrac-
tive alternative to established refining procedures, and have been used by
several major PGM refiners since the late 1970's. (Edwards, 1979; Cleare
et al., 1979; Barnes and Edwards, 1982.)
The existence of the PGM and gold in the form of complex anions in
chloride media has been used to advantage in the design of several anion-
exchange solvent-extraction processes. The reagents of importance are
primary, secondary, tertiary and quaternary amines and these are direct
analogues of the weak-base and strong-base resins referred to earlier. Like
weak-base resins, the primary, secondary, and tertiary amines undergo pro-
tonation reactions, and therefore their activity is also dependent on pH. The
extraction of the PGM by amine salts was first reported some thirty years
ago (Zeigler and Glemser, 1956). Shortly afterwards, a process for the
recovery of the PGM by solvent extraction was patented in the Soviet Union
887
(Bobikov et al., 1961). The results of pilot-plant operations for the recovery
of gold and PGM from copper- and nickel-containing refinery slimes have
also been described (Dolgikh et al., 1967), and it is probable that such pro-
cesses are currently in full-scale commercial use. Further extensive studies
by Soviet workers were described in a recent review (Gindin, 1981).
Other solvents that are used for the separation of gold chloride from
the PGM include dibutyl carbitol, which is used by Inco (Barnes and Ed-
wards, 1982), and methyl isobutyl ketone, which is used by Matthey Rusten-
burg Refiners (Reavill and Charlesworth, 1980). In both instances, metallic
gold of high purity can be recovered directly from the organic phase by the
addition of solutions of oxalic acid, iron(II), or other reducing agents. In
addition, the use of dialkyl sulphides for the extraction of gold and palladium
from chloride solution has been patented in South Africa (Edwards, 1979)
and, in ~ process employing isodecanol as the extractant (Fieberg and
Edwards, 1978), it has been shown that gold can be stripped from the load-
ed organic phases by contact with water at 50° C. The strip solution produc-
ed has a gold content of 50 g 1- 1•
Amines have also been used for the extraction of gold from alkaline
cyanide solution and, with ti:ioctylamine, the order of extraction was found
(Plaskin and Shivrin, 1963) to be
Au(CN); > Ag(CN); > Cu(CN); > Zn(CN)~- > Fe(CN):- > CN-.
This order is similar to the order of affinities of ions for anion-exchange
resins (Table 15.3). As with weak-base resins, good extraction of aurocyanide
by tertiary amines occurs under conditions in which the solutions are so
alkaline that a large percentage of the amine is in the unreactive, free-base
form, i. e. the driving force for the formation of strong ion pairs permits
the uptake of protons by the amine at higher pH values than would be ex-
pected for weak ion pairs. For example, the extraction of uranyl sulphate,
which forms relatively weak ion pairs with amines, is efficient only at pH
values lower than 3 to 4, whereas aurocyanide is extracted efficiently at pH
values up to 6 and 7. Nevertheless, the basicity of tertiary amine solvents
in aurocyanide solution (pKa ~ 6) is significantly lower than that of weak-
base resins (pKa ~ 9), and this limits the use of amine solvent-extraction
processes for the treatment of gold cyanide. Primary (pKa ~ 6,5) and
secondary (pKa ~ 7 ,5) amine solvents are somewhat stronger (Mooiman and
Miller, 1983), but are still too weak to extract gold efficiently at pH values
that would be compatible with free cyanide in solution (pKa ~ 7). The need
for the pH value to be reduced from about 10, which is normal for cyanide
leach liquors, to about 6, which would be necessary for amine solvent ex-
traction, is an added disadvantage.
Recently, Mooiman and Miller (1983 and 1984) showed that the basici-
ty of amine solvents can be increased significantly by the addition of tri-n-
butyl phosphate (TBP) or other organic phosphorus oxides to the organic
phase. They showed that the pK. for the extraction of aurocyanide can be
increased from between 6 and 7 to about 10 by the addition of TBP to the
organic phase at concentrations in the range 20 to 80% (by volume). The
888
Table 15.5. Standard reduction potentials of reductants for precipitation of gold.
Reaction E'(V)
Fe3+ + e = Fez+ 0,77

SO!- + 4H+ + 2e = H 2 S03 + H 2 0 0,17
Sn4 + + 2e = Sn'+ 0,15
2H2 C03 + 2H+ + 2e = (COOH), + 2H2 0 -0,39
Bo;- + 7H,O + Se = BH; + lOOH- -0,75
increase in amine basicity was attributed either to salvation of the amine by

the modifier or to an increase in the water content of the organic phase.
Evidence for the latter effect is the fact that the order of increasing amine
basicity in the presence of a range of modifiers parallels the order of increasing
solubility in water of the pure modifiers, and is consistent with the fact that
the pK. values of amines are generally 2 to 3 orders of magnitude higher
in water than they are in organic solvents.
TBP and other donor solvents extract aurocyanide fairly efficiently in
the absence of amine by a u1echanism that is probably similar to the
mechanism of extraction of auPJcyanide by activated carbon. Extraction from
alkaline solutions (pH 7 to 14) occurs only at high ionic strength, and prob-
ably takes place via salvation of the ion pair M"+ [Au(CN);J,,. However,
Russian workers (Zvyagintsev andZakharov-Nartisov, 1960) have shown that
aurocyanide, after acidification with hydrochloric, nitric, or sulphuric acid and
conversion to H[Au(CN) 2], can be extracted with donor solvents even from
solutions ofrelatively low ionic strength. They showed that extraction by solva-
tion decreased in the order ethers> esters> alcohols> ketones > aldehydes,
and the extraction coefficient decreased in the order hydrochloric acid >
nitric acid > sulphuric acid.
15.3.3 Chemical precipitation

The reduction of dissolved gold to the metallic state can be achieved by the
use of various (non-metallic) chemical reducing agents. For any reductant
to be effective, the standard reduction potential must be lower than that of
the corresponding half-reaction for the reduction of the gold complex. The
reduction potentials of several reductants that are used or are potentially
useful are given in Table 15 .5.
Comparison of these data with those in Figure 15.15 for the reduction
of the gold( III) and gold(I) complexes of various ligands, shows that all the
reductants should be effective in the precipitation of gold from chloride solu-
tions and that all, except ferrous ions, should be useful if the gold is leached
with thiourea or thiocyanate. On the other hand, it can be expected that gold
in solution as the thiosulphate complex will be reduced by only oxalic acid
or borohydride, while gold in the form of the aurocyanide ion can be
precipitated only by the borohydride ion.
889
These predictions are generally borne out in practice, the reduction of

gold from chloride solutions by ferrous ion, sulphur dioxide, or oxalic acid
being the most widely applied, notably in the refining of gold. The precipita-
tion of metallic gold from thiourea or cyanide leach liquors by chemical
reductants is not used commercially, because metallic reductants such as zinc
dust are less expensive and more effective than reagents such as sodium
borohydride. It is interesting to note, however, that the latter reductant can
be used for the precipitation of gold from concentrated solutions such as
those produced by the elution of activated carbon or by the intensive cyanida-
tion of gravity concentrates. Unpublished work at Mintek has shown that
90% of the gold (and silver) in a CIP eluate with a gold content of 250 g C 1
can be precipitated as an easily flocculated gold powder by the use of a ten-
fold excess of sodium borohydride at 80°C. The loss of borohydride by
decomposition, which is catalysed by the fine particles of activated carbon
in the eluate, was identified as the major practical problem in the use of this
reagent.
Finally, mention should be made of the precipitation of gold as insolu-
ble AuCN by the acidification of aurocyanide solutions. Such solutions are
metastable at low pH values (below about 3) because the precipitation of
AuCN by the reaction
Au(CN); + H+ = AuCN(s) + HCN (15.84)
is relatively slow. However, in the presence of other metal cyanide complexes,
notably those of copper, precipitation is rapid, and this has formed the basis
of a patent (Sorensen and Davidson, 1981) for the recovery of gold from
CIP eluates.
15.3.4 Electrowinning
Electrolysis was first used for the recovery of gold on the Witwatersrand in
1894 (Johnson, 1912). For some years the Siemens-Halske electrolytic pro-
cess competed with zinc cementation as the preferred method for the recovery
of gold, but was finally displaced by cementation in 1899, owing to the mark-
ed improvements in recovery that were obtained by the use of zinc dust.
The development of the CIP process (Zadra et al., 1952) rekindled in-
terest in electrowinning, and a number of new electrolysis cells were design-
ed. Although these cells have much higher space-time yields than the Siemens-
Halske electrolytic cell, almost no improvement in the final concentration
of gold in the barren solution (typically 0,2 to 1 g t- 1) has resulted. For this
reason electrolysis can be considered as a technical alternative to cementa-
tion only when some method for concentration of the gold, such as the
adsorption of soluble gold onto carbon or resins, or the use of gravity con-
centration prior to leaching (Davidson et al., 1978), is employed prior to the
recovery stage.
15.3.4.1 Cell design

The recovery of any electroactive species by electrodeposition is measured
in terms of a single-pass extraction efficiency, E, where
890
E = 1 - Cou/Cin•
where cin and cout are the concentrations of electroactive species entering
and leaving the electrowinning cell respectively. As the electrolyte has a finite
residence time in the electrowinning cell, E can be increased only if the rate
of deposition, R, is increased. The maximum possible value of R is given
by the equation
R = kAC,
where k is the mass-transport coefficient, A the electrode area, and C the
concentration of reducible species. The value of E can therefore be increas-
ed by an increase in the available cathode area and/ or the mass-transport
coefficient.
In terms of operating simplicity, an increase in E is most readily im-
plemented by increase of the electrode area. In the Siemens-Halske cell this
was achieved by the installation of a large number of planar lead foil cathodes
(Clevenger, 1915). Although a high value of E was attained, the cell was ex-
tremely large in relation to the mass of gold recovered (i.e. the space-time
yield was very low). In a later development of the cell, the lead was cut into
shavings that were packed into the space between the anodes (Clevenger, 1915)
in order to increase the cathode area in the smallest possible volume. It is
most probable that this development constitutes the first industrial use of
what is now termed a 'packed-bed' reactor.
The Zadra (Zadra et al., 1952; Elges et al., 1984), AARL (Davidson
et al., 1978; Young et al., 1984) and Mintek (Paul et al., 1983; Paul, 1985)
electrowinning cells are modern examples of packed-bed electrowinning cells
that are commercially available for the recovery of gold. Very high surface
areas are attained in these cells by the use of steel wool as the cathode material.
The development of electrowinning cells for the recovery of gold in the Soviet
Union has also followed the approach involving increase of the surface area.
This increase is achieved by the use of fibrous carbon-graphite, woven into
cloths, fabrics, or felts, as the cathode material (Maslii et al., 1976; Varent-
sov et al., 1984).
Many attempts have been made to improve the space-time yield of elec-
trowinning cells by improvement of the mass-transfer characteristics of the
reactor (Marshall and Walsh, 1985). The fluidized- bed reactor (Goodridge,
1977) was developed in an attempt to increase both the mass-transport coef-
ficient and the surface area of the reactor. Very few of these designs have
managed to overcome the problems associated with scale-up from the
laboratory to industrial units (Marshall and Walsh, 1985), and none has found
application in the gold-mining industry. The rotating tubular bed reactor
(Kammel and Lieber, 1981) and the Chemelec cell (Tyson, 1983) are used
for the recovery of precious metals from electroplating rinse water. The ap-
plication of the Chemelec cell for the recovery of gold from eluates has been
suggested (Tomlinson, 1984).
All packed- bed electro winning cells can be classified into two broad
groups according to their design. The cells in the first group operate with
the direction of the flow of electrolyte at right angles to the direction of the
891
(a) (b)
Figure 15.18. Schematic representation of packed-bed electrodes

with the flow of solution, U, at right angles (a) and parallel
(b) to the flow of current, I.
flow of current (a of Figure 15 .18). Those in the second group operate with
parallel flows of solution and current (b of Figure 15.18). Although the Zadra
and AARL cells incorporate a cylindrically shaped cathode surrounded by
a single tubular anode, these cells belong to the first group. Cells of Soviet
and Mintek design (i.e. a rectangular tank containing alternately spaced
anodes and cathodes positioned across the width of the tank) belong to the
second group.
The major design problem in the construction of an electrowinning cell
with the flows of current and solution at right angles to each other is the
elimination of any path that would enable the solution to by-pass the cathode
bed. Inspection of Figure 15. l 8a reveals that the vertical flow of solution
can freely enter the gap between the anode and the cathode, thus by-passing
the cathode bed. The problem can be alleviated if the electrolyte is pumped
down a feeder tube that is positioned vertically in the centre of the cathode
bed and has orifices along its entire length that allow the electrolyte to be
distributed horizontally. This technique (or a modified version of this techni-
que) is employed in both the Zadra and AARL cells. A further improve-
ment can be made by the use of a cation ion-exchange membrane to separate
the anode and cathode compartments (a feature of the AARL cell), thus
restricting the flow within the cathode compartment.
However, it is evident that by-passing of the cathode is a feature in-
herent in this type of packed bed design. The result is low single pass extrac-
tions in relation to the cathode volume and to the flow velocity through the
cathode.
In addition to the problem of the flow of electrolyte 'mechanically' by-
passing the cathode bed, 'electrical' by-passing of the electrolyte within the
bed must be prevented. Eiectricai by-passing occurs as a result of the finite
conductivity of the electrolyte, which causes the electrode potential to become
more positive towards the centre of the packed bed. The shift in potentials
occurs simply because the difference in potential between any two points in
the cathode is equal to the current flowing across those points multiplied
by the effective resistance of the electrolyte. The positive increase in poten-
892
-2.0 r - - - - - - - - , - - - - - - - - ,
Mass transport
1.5 control
>. -1.0
kl
-0-.5
No deposition
O'-------~'--~~---___,
0 0.5 1.0
Distance across cathode
Figure 15.19. Typical distribution of potential across a packed-
bed cathode of unit thickness.
tials therefore increases with decreasing conductivity of the electrolyte. A

typical potential gradient across the packed-bed cathode of unit thickness
is shown in Figure 15.19.
The deposition of gold from an alkaline cyanide electrolyte does not
begin until a potential of about - 0, 7 V is attained (the exact potential
depends on the composition of the solution, temperature, etc.). Over the
potential range - 0, 7 to - 1, 1 V the rate of deposition is kinetically controll-
ed, but becomes mass-transport controlled at potentials more negative than
-1,1 V. It is evident from Figure 15.19 that solution passing vertically
through the region of the bed marked AB will not be subjected to potentials
that are sufficiently negative to cause the gold to be deposited. It is this inac-
tive area of the packed bed that accounts for electrical by-passing.
Packed-bed electrowinning cells designed with parallel flows of current
and electrolyte do not suffer from the effects of electrical by-passing, since
the electrolyte is forced to pass through active regions of the bed upon entering
and leaving the cathode. Mechanical by-passing is also readily prevented in
this design (Paul et al., 1983).
Recent improvements to the Micron solvent-distillation procedure (Muir
et al., 1985) for the elution of loaded carbon have enabled the eluted gold
to be concentrated in a very small volume of eluate, and concentrations of
gold as high as 20 g 1- 1 have been obtained. The recovery of gold in a cell
containing a sufficient number of simple planar cathodes has been found
to be satisfactory (Muir et al., 1985).
15.3.4.2 Electrode reactions

The electrochemical reactions that occur at the cathode surface are almost
.identical to those that take place during cementation with zinc dust. The
predominant reactions are therefore the reduction of oxygen, water (i.e. evolu-
tion of hydrogen), aurocyanide, and argentocyanide ions. The cyanide corn-
893
plexes of mercury, lead and copper will also be reduced, whereas the com-
plexes of iron, nickel and zinc are not generally deposited. As is the case
in cementation, where a large stoichiometric excess of zinc is required for
a high recovery of gold, an excess of electrical current is required for the
electrowinning of gold. Current efficiencies of 1to5% are typical, although
the efficiency increases with increasing concentrations of gold. The reduc-
tion of oxygen and water consumes the excess current.
Temperature has a marked effect on the electrowinning of gold. As the
conductivity of the electrolyte increases with increasing temperature, and,
hence, the potential gradient across the cathode (Figure 15 .19) becomes flat-
ter, the active area of the cathode increases. Furthermore, the solubility of
oxygen decreases with increasing temperature, and the current consumed by
this parasitic reaction is reduced.
The only inorganic impurity that has been shown to affect the electrowin-
ning of gold is soluble hexavalent chromium (i.e. chromate ions), which is
reduced to insoluble chromic hydroxide at the cathode surface, and passivates
the cathode. Chromate ions can be produced by the transpassive dissolution
of stainless-steel anodes if the pH value of the electrolyte is much below 12,5.
Concentrations as low as 5 mg 1- 1 can reduce the single pass efficiency for
the recovery of gold to less than 10%.
The gold deposited on the cathode material is generally very fine grain-
ed and adheres poorly to the cathode. This effect is common to any electro-
chemical deposition process in which the cathodic process occurs under mass-
transport control, and results from the very high rate of nucleation, which
does not permit the existing nuclei to coalesce into a smooth deposit. It has
been suggested (Paul et al., 1983) that the increase in micros9opic surface
area of the cathode due to the presence of the gold particles could enhance
the evolution of hydrogen (a reaction that is under kinetic control) at the
expense of the reduction of aurocyanide ions (a reaction that is under mass
transport control), which would account for the reduced single pass extrac-
tions as the cathodes become loaded with gold.
The predominant reaction at the anode is the oxidation of water to
oxygen, although some cyanide may be oxidized to ammonia and carbon
dioxide. If the electrowinning cells are not equipped with a ducting system
to remove the gases, the odour of ammonia can become intense. The pH
of the anolyte must be maintained above about 12,5 if the anodes are made
of stainless steel, so that no corrosion of the alloy, and, hence, contamina-
tion of the electrolyte with chromate ions, will take place.
15.4 Clean-up, Calcination, and Smelting

The bullion produced in South Africa before the CIP process was introduc-
ed typically contained 88 to 90% gold and 7 to 11 OJo silver, together with
small amounts of copper, lead, zinc and iron (Adamson, 1972). The bullion
is assayed for its gold content at the mine before being dispatched to the
Rand Refinery for further processing. Upon receipt of this bullion by the
refinery, the bullion is weighed, melted and sampled while molten. The melt
894
is cast into bars and stored while the samples are assayed. When the assay
and mass have been agreed with the mine, the refinery will process the bullion
and the mine will eventually be paid for the gold content of the original bullion
by the South African Reserve Bank.
With the advent of the CIP process, this somewhat lengthy procedure
has been complicated by the presence of high concentrations of iron (5 to
8 OJo) in the bullion produced on certain mines. The refinery has found that
this bullion undergoes a degree of liquid separation while molten, and that
the gold content of the two molten layers is different. The gold content of
the samples taken from the melt are therefore not representative of the average
gold content of the bullion, and disagreement between the Refinery and the
mine concerned is inevitable.
The source of this iron contamination is the steel wool used as cathode
material by those mines that recover gold from the eluates by electrowin-
ning. Although the physical metallurgy of these metals has not been studied
in any detail, it appears that the liquid separation is due to the immiscibility
of liquid copper and silver with iron (Hansen, 1958), resulting in partition
of the gold between layers rich in iron and those rich in copper and silver.
If the iron content of the bullion is below about 3 OJo, the extent of the separa-
tion is insignificant.
The method used for the treatment of zinc-cementation slimes is com-
mon to most mines, and involves acid treatment to dissolve the excess zinc,
calcination at 600 to 800°C to oxidize the base metals, and smelting at 1200
to 1400°C with a borosilicate flux. In comparison, the treatment of steel-
wool cathodes differs markedly from one plant to another, and may involve
acid digestion followed by smelting of the residue, calcination followed by
smelting, or even smelting without any prior pre-treatment of the cathodes
(Hinds and Trautman, 1983). Digestion of the steel wool in hy:drochloric acid,
followed by dissolution of the residue in aqua regia, filtration to remove silver
chloride, and precipitation of the metallic gold with oxalic acid or sulphur
dioxide. has been suggested as a possible process for smaller operators who
may wish to avoid smelting (McClelland et al., 1985).
15.4.1 Acid treatment

The slimes obtained from zinc cementation typically contain gold (30 to 40% ),
silver (2 to 4%), lead (5 to 15%), zinc (15 to 30%), sulphide (1 to 4%) and
small amounts of copper, iron, nickel and silica. The sulphide content is due
to the presence of insoluble zinc and lead sulphides. Mercury is not present
to any extent in South African ores, but often attains high levels in some
U.S. ores.
The primary objective in acid treatment (usually with sulphuric acid)
is the dissolution of excess zinc. Ventilation is required to remove hydrogen
sulphide, which is formed by the reaction of the acid with the lead and zinc
sulphides. Most of the lead is converted to insoluble lead sulphate. If arsenic
or antimony is present in the zinc slimes, the evolution of gaseous arsine,
AsH3 , and stibine, SbH3 , can occur because the potentials required for the
reduction of these metals
895
As + 3H+ + 3e = AsH3 , E' = -0,61 v (15.85)

Sb + 3H+ + 3e = SbH3 , E' = -0,51 v (15.86)
are positive with respect to the potentials at which zinc corrodes in sulphuric
acid ( - 0, 7 to 0,8 V). These gases are highly toxic. Some zinc sulphate may
remain in the acid-treated slimes.
The steel-wool cathodes typically contain gold (60 to 70%), silver (5 to
10%), copper (5 to 10%), and iron (10 to 25%). Mercury and lead may also
be present if these metals were present in the eluate. The steel wool is readily
dissolved by sulphuric and hydrochloric acids, although a higher rate of reac-
tion is achieved with hydrochloric acid. Heating to moderate temperatures
(i. e. 50 to 60° C) markedly increases the rate of reaction. If the concentra-
tion of silver on the cathode is very high, some dissolution may occur in
hydrochloric acid, due to the presence of chloride and ferric ions (produced
by the oxidation of ferrous ions by dissolved oxygen)
Fe c1; + e = Fez+ + 2c1-, E' = 0,65 V (15.87)
Agel; + e = Ag + 2CI-, E' = 0,49 v (15.88)
The equilibrium potentials for the reduction of ferric ion and the oxidation
of silver in sulphuric acid indicate that silver will not be dissolved.
Fe3 + + e = Fez+, E' = 0,68 V (15.89)
Ag+ + e = Ag, 0
E = 0,80 V (15.90)
15.4.2 Calcination
Calcination of the acid-treated residues at 600 to 800° C in an adequate supply
of oxygen results in the conversion of the base metals to their oxides. The
free energy per mole of oxide produced (Pankratz, 1982) is negative for all
the base metals (Cu, Pb, Zn, Fe) at the temperature of calcination (Figure
15.18). The formation of silver oxide is thermodynamically favourable only
at low temperatures, and becomes unfavourable at the temperatures usually
employed for calcination. No thermodynamic data exist for the oxidation
of gold at these temperatures, but the free-energy change is certain to be large
and positive. If the cathodes are not treated with acid prior to calcination,
care must be taken to ensure that an unrestricted flow of oxygen is available
for the oxidation of the steel wool. Cathodes that are packed too tightly into
the calcining trays will undergo oxidation only at the exposed surface. Poor
oxidation of the steel wool prior to smelting is certainly the major cause of
high iron values in the bullion.
Zinc sulphate is decomposed to zinc oxide and sulphur trioxide during
calcination, but lead sulphate is decomposed only at temperatures ap-
proaching 1000° C. This result is to be expected from the free energies shovm
in Figure 15.20.
15.4.3 Smelting
If the calcination process has been efficiently performed, the gold and silver
will be in the metallic state and all the base metals will be present as oxides
896
200
'<)'
'O
"><
0
'S -200
0
g
....,
""'
~ -400
-600
-800
Figure 15.20. Free energy of formation of some metal oxides

from the elements and by decomposition of the metal sulphates
(P503 = 0,001 atm).
(with the exception of lead, which is present as lead sulphate). Under these
conditions, the smelting process merely requires the bullion to be melted in
the presence of a suitable flux that will react with the oxides to form a stable
free-flowing slag. Lead sulphate is decomposed to lead oxide and sulphur
trioxide at the smelting temperature, which is usually in the range 1200 to
1400°C (the melting point of gold is 1065°C).
The principal component of almost all metallurgical slags is silica, i. e.
Si02 (Elliott, 1984). Silica consists of units of Si04 in which each silicon
atom is tetrahedrally surrounded by four oxygen atoms, and each oxygen
atom is bonded to two silicon atoms (Turkdogan, 1983). The addition of
metal oxides, "MO, to molten silica results in a gradual breakdown of this
structured silicate network, as represented schematically by the reaction:
I I I I
- Si - 0 - Si - + MO = - Si - 0 - M2 + - 0 - Si - (15.91)
I I I I
The result of this reaction is the formation of a structure in which the oxygen
bridge between two silicon atoms is broken. As the mole percentage of MO
is increased, the number of non-bridging oxygen atoms is increased, and the
network becomes progressively disordered.
A large number of structural units have been deduced from Raman spec-
tra of silica-oxide melts of varying composition. The prominent units are:
897
Si02, Si20;-, Si2 0t, Si2 0~-, Sio:-,

corresponding to an increase in the number of non-bridging oxygen atoms
from 0 to 4. Attempts have been made to correlate the physical properties
of the metals in terms of the formation of these structural units (Turkdogan,
1983).
The reaction between metal oxide and silica in the melt is analogous to
the neutralization of an aqueous acid with oxide. A slag is therefore termed
'acidic' if it can react with 'basic' metal oxides. In terms of this loose defini-
tion, molten silica is very acidic, and remains acidic until reaction with metal
oxide has broken all the Si-0-Si bonds and the only structural unit remain-
ing is Sio!- .
Silica alone is not a suitable flux for the smelting of gold, because of
its high melting point (in excess of 1700° C) and high viscosity. Both the
melting point and viscosity can be reduced by the partial neutralization of
silica with soda at a ratio of N~O to Si02 of 50 mole per cent. However,
the capacity of this slag for reaction with the base metal oxides in the calcine
is then greatly reduced.
The solution to this problem is the use of a ternary slag consisting of
soda, silica, and borax, B20 3. Borax is almost as acidic as silica, but has a
melting point of 450°C. Much uncertainty still exists concerning the struc-
ture of B20 3, and reference will be made only to the critical review by
Greenwood (1973). The reaction of borax with fused metal oxides results
in the formation of non-bridging oxygen atoms in a fashion analogous to
that of the silicates. An increase in the molar ratio of metal oxide to borax
produces the following structural changes (Greenwood, 1973):
B203 - B4o;- - B3o!- - Bzo~- - Bo;- - Bo~-.
A flux consisting of equal proportions (by mass) of silica and sodium
borax (N~B4 07 or Na0.2B20 3) is considered to be the starting point for the
determination of the flux requirements for a particular smelt. The liquidus
for such a mixture consisting of N~O-B 2 03 , and Si02 at a mass ratio of
15:35:50 is about 800°C (Levin et al., 1964). This flux is strongly acidic, and
the melt is characterized by low viscosity.
After the base metal oxides have reacted with the fluxes, the slag should
be neutral or slightly acidic in order to protect the lining of the furnace (usually
alumina-silicates) from attack by a basic slag. This requirement is achieved
by adjustment of the total mass of flux used for the smelting process. Acid-
treated, calcined zinc slimes typically require the mass of the fluxes to be
30 to 40% of the mass of the calcine to be smelted. If the slimes are not
treated with acid prior to calcining, the mass of the fluxes may have to be
increased to at least 50 per cent of the mass of the calcine. Owing to the
increased mass of slag in these circumstances it may be advantageous to reduce
the liquid us of the slag ·by the use of a larger amount of borax. If the viscosi-
ty of the slag is too high, calcium fluoride can be added to the flux, since
the fluoride ion is a strong network breaker:
898
I I I
- Si 0 - Si + 2F- = 2( - Si - F-) + 0 2 - (15.92)
I I I
However, the production of oxygen ions effectively increases the activity of
the oxides in the slag and reduces the acidity of the slag.
Poor calcination, which does not convert the base metals completely
to their oxides, usually requires the addition of an oxidant to the slag.
Manganese dioxide is commonly employed if the problem is not severe:
Mn02 + M = MO + MnO. (15.93)
For the treatment of steel-wool cathodes, however, where large amounts of
metallic iron may be present, sodium nitrate is preferred:
2NaN03 + 3M = Nap + 2NO + 3MO, (15.94)
because of its higher content of available oxygen. The danger associated with
the use of oxidants is that silver can be readily oxidized into the slag. This
problem can be avoided by acid treatment or effective calcination of the
cathodes prior to smelting. The composition and mass of fluxes for any par-
ticular smelting operation must be determined experimentally.
15.5 Acknowledgement
This chapter is published by permission of the Council for Mineral
Technology (Mintek).
15.6 References
Adamson, R.J. (1972). Gold Metallurgy in South Africa, chap 5, pp. 120-151. Chamber of
Mines of South Africa, Johannesburg.
Anglo American Corporation of South Africa (James, G.S. and Davidson, R.J .) (1977). Gold
and uranium extraction. S. Afr., Pat. 77/1778.
Anon. (1978). Gold from Russia's Muruntau deposit. Coal-Gold Base Miner., 26, 75 81.
Aveston, J., Everest, D.A. and Wells, R.A., (1958). Adsorption of gold from cyanide solu-
tions by anionic resins. J. Chem. Soc., 231-239.
Barnes, J.E. and Edwards, J.D. (1982). Solvent extraction at Inco's Acton precious metal refinery.
Chem Ind. (Land.), 5, 151-155.
Barsky, G., Swainson, S.J. and Hedley, N. (1935). Dissolution of gold and silver in cyanide
solutions. Trans. Am. Inst. Min. Metal!. Eng., 112, 660-667.
Becker, E., Knothe, M. and Lobel, J. (1983). Gold recovery from non-metallic secondary raw
materials by leaching with thiourea and adsorption on ion-exchangers. Hydrometallurgy, 11,
265-275.
Berezowsky, R.M. and Sefton, V.B. (1979). Reprint of a paper presented at 108th AIME Meeting,
New Orieans.
Beyers, E. (1936). Some of the factors which influence the rates of dissolution of gold and silver
in cyanide solutions. J. Chem. Metal!. Min. Soc. S. Afr., 36, 37 89.
Bezzubenko, A.W. and Peshchevitskii, B.I. (1962). The question of the existence of the aquo-
ion of trivalent gold. Chem. Abstr., 56, 2039b.
Bobikov, P.I., Gindin, L.M., Kouba, E.F., Sokolov, A.P. andDolgikh, V.J. (1961). Recovery
of platinum metals from solutions. USSR Pat. 142, 768.
899
Burstall, F.H., Kember, N.F., Forrest, P.J. and Wells, R.A. (1953). Ion exchange processes
for the recovery of gold from cyanide solutions. Ind. Eng. Chem., 45(8), 1648-1658.
Burstall, F .H. and Wells, R.A. (1955). Recovery of gold from cyanide solution by ion exchange.
In Jon Exchange and Its Applications, pp. 83 - 88. Society of Chemical Industry, London.
Byerlee, H.W. (1980). A contractor for solvent extraction from slurries. Reprint of a paper
presented at Symposium: Ion exchange and solvent extraction in mineral processing. Council
for Mineral Technology, Randburg.
Cariati, F., Gallizzidi, D. and Naldini, N. (1970). Thiocyanato and cyano derivatives of gold(I)-
tertiary phosphine compounds. Chim. Ind. (Milan), 52, 995.
Cathro, K.J. (1963). The effect of oxygen in the cyanide process for gold recovery. Proc. Aust.
Inst. Min. Metal!., 207, 180-205.
Cho, E.H. and Pitt, C.H. (1979). Kinetics and thermodynamics of silver cyanide adsorption
on activated charcoal. Metal!. Trans., lOB, 165-169.
Cleare, M.J ., Charlesworth, P. and Bryson, D.J. (1979). Solvent extraction in platinum group
metal processing. J. Chem. Technol. Biotechnol., 29, 210- 224.
Clevenger, G.H. (1915). Electrolytic precipitation of gold, silver, and copper from cyanide solu-
tions. Metal!. Chem. Eng., 13, 803 - 860.
Clifford, D. and Weber, W.J. Jr. (1983). The determinants of divalenUmonovalent selectivity
in anion-exchangers. React Poly, 1, 77- 89.
Damjanovic, A., Genshaw, M.A. and Bockris, J. O'M. (1967). Hydrogen peroxide formation
on oxygen reduction at gold electrodes. Electroanal. Chem., 15, 173-180.
Davison, J., Read, F.0., Noakes, F.D.L. and Arden, T.V. (1961). Ion exchange for gold recovery.
Trans. Inst. Min. Metal!., 70, 247 - 290.
Davidson, R.J. (1974). The mechanism of gold adsorption on activated carbon. J.S. Afr. Inst.
Min. Metal!., 75(4), 67 -76.
Davidson, R.J. and Duncanson, D. (1977). The elution of gold from activated c11rbon using
deionized water. J.S. Afr. Inst. Min. Metal!., 77, 254-261.
Davidson, R.J., Brown, G.A., Schmidt, C.G., Hanf, N.W., Duncanson, D. and Taylor, J.D.
(1978). The intensive cyanidation of gold-plant gravity concentrates. J.S. Afr. Inst. Min. Metal!.,
78, 146-165.
Demidov V.I., Kreines, R.Z., Ivanova, M.I. and Kartasheva, A.S. (1967). Pilot plant tests on
absorptive extraction of gold from waste cyanide solutions. Sov. J. Non-ferrous Met., 8(8),
53 61.
Despic, A.R. and Popov, K.1. (1972). In Modern Aspects of Electrochemistry, edit. B.E. Con-
way and J. O'M Bockris, no. 7. Plenum, New York.
De Wet, H.F. (1985). The adsorption of Mo(VI) from acid solution onto activated carbon. MSc
Thesis, University of Stellenbosch, Stellenbosch.
Dixon, S.N., Cho, E.H. and Pitt, C.H. (1978). The interaction between gold cyanide, silver
cyanide and high surface area charcoal. AIChE Sym. Ser. 174(173), 75 -83.
Dolgikh, V.I., Bobikov, P.I., Borbat, V.F. and Ferberg, M.B. (1967). Recovery of platinum
metals from slimes of nickel and copper electrolysis by extraction with amines. Tsvet. Metal.,
40, 27-30.
Edwards, R.I. (1976). Refining of the platinum-group metals. J. Met. (N. Y.), 28(8), 4-9.
Edwards, R.I. (1979). In Proceedings of the International Solvent Extraction Conference (ISEC
77), vol. 1, pp. 24-31, edit. B.H. Lucas. Can. Inst. Metal!., Toronto.
Eisele, J.A., Colombo, A.F. and McClelland, G.E. (1983). Recovery of gold and silver from
ores by hydrometallurgical processing. Sep. Sci. Technol., 18, 1081-1094.
Elges, C.H., Wroblewski, M.D. and Eisele, J.A. (1984). Direct electrowinning of gold. In In-
ternational Symposium on Electrochemist1y in Mineral and Metal Processing, edit. P.E.
Richardson, pp. 501 512. Electrochemical Society, Pennington, N.J.
Elliott, J.F. (1984). In Proceedings of the Second International Symposium on Metallurgical
Slags and Fluxes, edit. H.A. Fine and D.R. Gaskell, pp. 45-61. Metal!. Soc. AIME,
900
W arrendale.
Feather, C.E. and Koen, G.M. (1973). The significance of the mineralogical and surface
characteristics of gold grains in the recovery process. J.S. Afr. Inst. Min. Metal!., 73, 395-405.
Fieberg, M. and Edwards, R.I. (1978). The Extraction of Gold from Chloride Solutions. Na-
tional Institute for Metallurgy, Randburg. Report No. 1996.
Filmer, A.O. (1982). The dissolution of gold from roasted pyrite concentrates. J.S. Afr. Inst.
Min. Metal!., 82, 90-94.
Fink, C.G. and Putnam, G.L. (1950). The action of sulphide ion and of metal salts on the dissolu-
tion of gold in cyanide solutions. Min. Eng. (N. Y.), 187(9), 952 955.
Finkelstein, N.P. (1972). In Gold Metallurgy in South Africa, edit. R.J. Adamson pp. 284-351.
Chamber of Mines of South Africa, Johannesburg.
Finkelstein, N.P. and Hancock, R.D. (1974). A new approach to the chemistry of gold. Gold
Bull., 7(3), 72-77.
Fleming, C.A. (1984). Regeneration of thiocyanate-loaded resins. S. Afr. Pat. 8419620.
Fleming, C.A. (1985). Extraction Metallurgy 85, pp. 757-788. Inst. Min. Metall., London.
Fleming, C.A. and Hancock, R.D. (1979). The mechanism in the poisoning of anion-exchange
resins by cobalt cyanide complexes. J.S. Afr. Inst. Min. Metal!., 79, 334-341.
Fleming, C.A. and Nicol, M.J. (1984). The adsorption of gold cyanide onto activated carbon.
III. Factors influencing the rate of loading and the equilibrium capacity. J.S. Afr. Inst. Min.
Metal!., 84, 85-93.
Fleming, C.A. and Cromberge, G. (1984a). The extraction of gold from cyanide solutions by
strong- and weak-base anion-exchange resins. J.S. Afr. Inst. Min. Metal!., 84, 125 -137.
Fleming, C.A. and Cromberge, G. (1984b). The elution of aurocyanide from strong- and weak-
base resins. J.S. Afr. Inst. Min. Metal!., 84, 269-280.
Fleming, C.A. and Cromberge, G. (1984c). Small scale pilot plant tests on the resin-in-pulp
extraction of gold from cyanide media. J.S. Afr. Inst. lvfin. Metal!., 84, 369-378.
Fleming, C.A. and Cromberge, G. (1984d). In Pro(:eedings of the International Conference
on Recent Advances in Mineral Science and Technology (MINTEK 50), edit. L.F. Haughton,
vol. 2, pp. 663 - 676. Council for Mineral Technology, Randburg.
Gardiner, B.L. (1933). Further contributed remarks on "Reduction works practice at Morro
Velho, Brazil". Trans. Inst. Min. Metal!., 42, 251-254.
Gindin, L.M., (1981). In Ion Exchange and Solvent Extraction, edit. J.A. Marinsky and Y.
Marcus, chap 8, pp. 311-368. Marcel Dekker, New York.
Goodridge, F. (1977). Some recent developments of monopolar and bipolar fluidised bed elec-
trodes. Electrochim. Acta, 22, 929-933.
Greenwood, N.N. (1973). In Comprehensive Inorganic Chemistry, edit. J.C. Bailor, H.J.
Emeleus, R. Nyholm and A.F. Trotman-Dickenson, vol. 1, pp. 665-992. Pergamon, Oxford.
Groenewald, T. (1975). Electrochemical studies on gold electrodes in acidic solutions of thiourea
containing gold(I) thiourea complex ions. J. Appl. Electrochem., 5, 71-78.
Groenewald, T. (1976). The dissolution of gold in acidic solutions of thiourea. Hydro metallurgy,
1, 277-290.
Groenewald, T. (1977). Potential applications of thiourea in the processing of gold. J.S. Afr.
Inst. Min. Metal!., 77, 217-223.
Guchetl, I.S. and Lezgintsev, G.M. (1968). Basic approaches towards developing a technology
of processing gold containing ore. Sov. J. Non-Ferrous Metals, 9(7), 2-6.
Habashi, F. (1966). The theory of cyanidation. A/ME., Trans. Soc. Min. Eng., 235, 236-239.
Hancock, H.A. and Thomas, G. (1954). The effect of nickel and hydrogen overvoltage on the
precipitation of gold. Can. Min. Metal!. Bull., 57, 337- 346.
Hancock, R.D. and Finkelstein, N.P. (1970). E'IE' diagrams: their uses in estimating unknown
standard reduction potentials with particular reference to d 10 electronic configuration metal
ions. National Institute for Metallurgy, Johannesburg, Report No. 1153.
Hansen, M. (1958). Constitution of Binary Alloys, 2nd ed. McGraw-Hill, New York.
901
Hedley, M. and Tabachnick, H. (1968). Chemistry ofCyanidation. American Cyanamid Com-

pai;iy, Mineral Dressing Notes 23.
Hedley, N. and Kentro, D.M. (l 945). Copper cyanogen complexes in cyanidation. Trans. Can.
Inst. Min. Metal/., 48, 237 -251.
Helferrich, F. (1962). Ion Exchange. McGraw-Hill, New York.
Heuman, T. and Panesar, H.S. (1966). Z. Phys. Chem., 299, 84-92.
Hinds, H.L. and Trautman, L.L. (1983). Refining of precious metal cathodes. Min. Eng., 35,
1545-1546.
Hisshion, R.J. and Waller, C.G. (1984). Recovering gold with thiourea. Mining Mag. (Lond.)
150, 237 - 243.
Holloway, J. (1982). Solving gold extraction problems. Chamber of Mines Journal, 47-51.
Hosking, J .W. (1984). The recovery of gold from ores by ion-exchange resins. Preprint of a
paper presented to the Perth and Kalgoorlie Branches of the Aust. Inst. Min. Metall. Regional
conference on "Gold-Mining, Metallurgy and Geology".
Hussey, S.J. (1949). Application of ion-exchange resins in the cyanidation of a gold and silver
ore. U.S. Dept. of the Interior, Bureau of Mines, R.I. 4374.
Hussey, S.J ., Salisbury, H.B. and Potter, G.M. (1978). Carbon-in-pulp silver adsorption from
cyanide leach slurries of a silver ore. U.S. Dept. of the Interior, Bureau of Mines, 8268.
Johnson, E.H. (1912). In A Textbook of Rand Metallurgical Practice, vol. 1, edit. R. Stokes,
chap 7, pp. 245 - 254. Charles Griffin, London.
Kakovskiy, I.A. and Shcherbakov, O.K. (1967). Kinetics of precipitation extraction of noble
metals from cyanide solutions. Izv. Akad. Nauk. SSSR, Metal., l, 76-86.
Kameda, M. (1949). Fundamental studies on solution of gold in cyanide solutions. III. Effects
of alkalies, lead acetate, and some impurities contained in foul cyanide solutions. Science Rep.,
Research Inst., Tokoku Univ., Ser. A, 1, 435 -444.
Kammel, R. and Lieber, H. (1981). Electrolytic recovery of precious metals from dilute solu-
tions. J. Met. (N. Y.), 33, 45 -48.
Kirk, D.W., Foulkes, F.R. and Graydon, W.F. (1979). Electron stoichiometry of anodic dissolu-
tion of gold in aqueous alkaline cyanide. J. Electrochem. Soc., 126, 2287-2288.
Kirk, D.W. and Foulkes, F.R. (1984). A potentiodynamic study of metal affecting precious
metal recovery from alkaline cyanide solutions. J. Electrochem. Soc., 131, 760-769.
Kudryk, V. and Kellogg, H.H. (1954). Mechanism and rate-controlling factors in the dissolu-
tion of gold in cyanide solution. J. Met. (N. Y.), 6(5), 541-548.
Laskorin, B.N., Sadovnikova, G.I., Petrova, L.N. and Fedorov, V.D. (1977). Gold selective
ion-exchangers. J. Appl. Chem. USSR, 47, 1794-1796.
Laskorin, B.N., Moskvicheva, G.I. and Skityaeva, E.I. (1978). Comparison between activated
carbons and resins for gold extraction. Tsvet. Metal!. 3, 15 -17.
Latimer, W.M. (1952). The Oxidation States of the Elements and their Potentials in Aqueous
Solutions, 2nd ed., p. 138. Prentice-Hall, Englewood Cliffs, N.J.
Leblanc, R. (1942). Precipitation of gold from cyanide solution by zinc dust. The influence
of certain factors and impurities. Can. Min. J., 63, 213-219, 297- 306, 371-379.
Levin, E.M., Robbins, C.R. and McMurdie, H.F. (1964). Phase Diagrams for Ceramists.
American Ceramic Society, Ohio.
Marcus, Y. and Kertes, A.S. (1969). Ion Exchange and Solvent Extraction of Metal Complexes.
Wiley-Interscience, London.
Marshall, R.J. and Walsh, F.C. (1985). A review of some recent electrolytic cell designs. Surf
Technol., 24, 45 - 77.
Maslii, A.I., Zamyatin, A.P., Varentsov, V.K., Krapivin, V.M. and Frolov, Yu. I. (1976). Using
electrodes with extended surfaces in hydrometallurgy. Sov. J. Non-Ferrous Met., 17, 40- 41.
Matsukawa, T. and Sakai, T. (1956). Metallurgy of antimonal ores bearing gold and silver.
J. Min. Inst. lap., 72, 123 -128.
McClelland, G.E., Wroblewski, M.D. and Eisele, J.A. (1985). Production of high-purity gold
902
from zinc precipitates and steel wool cathodes by hydrometallurgical refining. U.S. Dept. of
the Interior, Bureau of Mines, I.C. 9002.
McDougall, G.J., Hancock, R.D., Nicol, M.J., Wellington, O.L. and Copperthwaite, R.G.
( 1980). The mechanism of the adsorption of gold cyanide on activated carbon. J.S. Afr. Inst.
Min. Metal!., 80, 344- 356.
Mcintyre, J.D.E. and Peck, W.F. (1976). Electrodeposition of gold. Depolarisation effects in-
duced by heavy metal ions. J. Electrochem. Soc. 123, 1800-1813.
McLachlan, C.G., Ames, H.L. and Morton, R.J. (1946). Cyaniding at Noranda. Trans. Can.
Inst. Min. Metal!., 49, 91- 122.
Miller, J.D. and Mooiman, M.B. (1984). A review of new developments in amine solvent ex-
traction systems for hydrometallurgy. Sep. Sci. Technol., 19(11), 895-910.
Mooiman, M.B. and Miller, J .D. (1983). In Proceedings of the International Solvent Extrac-
tion Conference (!SEC 83), pp. 530- 531. AIChE, New York.
Muir, C.W.A., Stett, R.D. and Priddle, C.J. (1983). In Hydrometallurgy-Research, Develop-
ment and Plant Practice, edit. K. Osseo-Asare and J .D. Miller, pp. ·825 - 838. Metal!. Soc.,
AIME, Warrendale, Pa.
Muir, C. W .A., Hendriks, L.P. and Gussmann, H. W. (1984). The treatment of refractory gold-
bearing flotation concentrates using pressure leaching techniques. In Precious Metals: Min-
ing Extraction and Processing, edit. V. Kudryk, D.A. Corrigan and W. Liang, pp. 309- 322.
Metal!. Soc. AIME, Warrendale, Pa.
Muir, D.M., Hinchliffe, W.D. and Griffin, A. (1985). The elution of gold from carbon by the
Micron solvent distillation procedure. Hydrometallurgy, 14, 151-169.
Nagy, I., Mrkusic, P. and McCullock, H.W. (1966). Chemical treatment of refractory gold
ores. Report 38, National Institute for Metallurgy, Johannesburg.
Nicol, M.J. (1980a). The anodic behaviour of gold. I. Oxidation in acidic solutions. Gold Bull.
13, 46-55.
Nicol, M.J. (1980b). The anodic behaviour of gold. II. Oxidation in alkaline solutions. Gold
Bull. 13, 105 - 111.
Nicol, M.J. and Schaich, E. (1976a). An electrochemical and kinetic investigation of the behaviour
of gold in chloride solutions. I. The cathodic deposition of gold. Report no. 1848, National
Institute for Metallurgy, Johannesburg.
Nicol, M.J. and Schaich, E. (1976b). An electrochemical and kinetic investigation of the behaviour
of gold in chloride solutions. II. The anodic dissolution of gold. Report no. 1844, National
Institute for Metallurgy, Johannesburg.
Nfcol, M.J., Fleming, C.A. and Cromberge, G. (1984a). The adsorption of gold cyanide onto
activated carbon. I. The kinetics of adsorption from pulps. J.S. Afr. Inst. Min. Metal!., 84,
50- 54.
Nicol, M.J., Fleming, C.A. and Cromberge, G. (1984b). The adsorption of gold cyanide onto
activated carbon. II. Application of the kinetic model to multistage adsorption circuits. J.S.
Afr. Inst. Min. Metall., 84, 70-78.
Osseo-Asare, K., Xue, T. and Ciminelli, V.S.T. (1984a). In Precious Metals: Mining, Extrac-
tion, and Processing, edit. V. Kudryk, D.A. Corrigan and W. Liang, pp. 173-197. Metall.
Soc. AIME, Warrendale, Pa.
Osseo-Asare, K., Afenya, P.M. andAbotsi, G.M.K. (1984b). In Precious Metals: Mining, Ex-
traction, and Processing, edit. V. Kudryk, D.A. Corrigan and W. Liang, pp. 125 -144. Metal!.
Soc. AIME, Warrendale, Pa.
Pankratz, L.B. (1982). Thermodynamic Properties of Elements and Oxides. U.S. Dept. of the
Interior, Bureau of Mines. Bull. 672.
Paul, R.L. (1984). The role of electrochemistry in the extraction of gold. J. Electroanal. Chem.,
168, 147 - 162.
Paul, R.L. (1985). In Design and Operation of Plants for the Recovery of Gold by Activated
Carbon, chap. 25, SAIMM Johannesburg.
903
Paul, R.L., Filmer, A.O. and Nicol, M.J. (1983). In Hydrometallurgy, Research, Development
and Plant Practice, edit. K. Osseo-Asare, K. and J .D. Miller, pp. 689- 704. Metal!. Soc.,
AIME, Warrendale, Pa.
Plaskin, I.N., Suvorovskaya, N.A. and Budnikova, 0.K. (1948). Precipitation of metals from
solution by metals. Izv. Akad. Nauk. SSSR., Otdel, Tekh. Nauk., 131-138.
Plaskin, I.N. and Shivrin, G.N. (1963). Mechanism for the extraction of amines in an alkaline
system. Dok/. Akad. Nauk. USSR, 150(3). 870- 873.
Pourbaix, M. (1966). Atlas of Electrochemical Equilibria in Aqueous Solutions. Pergamon,
Oxford.
Preisler, P.W. and Berger, L. (1947). Oxidation-reduction potentials of thiol-dithio systems:
thiourea-formamidine disulphide. J. Amer. Chem. Soc., 69, 322-325.
Puddephatt, R.J. (1978). The Chemistry of Gold, edit. R.J.H. Clarke. Topics in Inorganic and
General Chemistry, Monograph 16. Elsevier, Amsterdam.
Putnam, G.L. (1950). A laboratory study of the influence of ferricyanide on the extraction of
gold. J. Chem. Metal!. Min. Soc. S. Afr., 50, 228-231.
Pyper, R.A. and Hendrix, J.L. (1981). In Extraction Metallurgy '81, pp. 57-75. IMM, London.
Reavill, L.R.P. and Charlesworth, P. (1980). The application of solvent extraction to platinum
group metal refining. In Proceedings of the International Solvent Extraction .Conference (!SEC
80), Liege, vol. 3, pp. 80-93.
Rigo, P. and Turco, A. (1974). Cyanide phosphine complexes of transition metals. Coard. Chem.
Rev., 13, 133.
Schmid, G.M. and Curley-Fiorino, M.E. (1975). In Encyclopedia of Electro-chemistry of the
Elements, vol. IV. edit. A.J. Bard, pp. 87-178. Marcel Dekker, New York.
Schutte, C.J .H. (1985). Aspects of the chemistry of gold. Chemsa, 11, 86- 90.
Shantz, R. and Reich, J. (1978). A review of copper-cyanide metallurgy. Hydrometallurgy, 3,
99-109.
Sillen, L.G. and Martell, A.E. (1964). Stability Constants of Metal Jon Complexes, 2nd ed.,
pp. 542 - 552. Charles Griffin, London.
Sorensen, P.F. and Davidson, R.J. (1981). Gold Recovery. S. Afr. Pat. 81/8894. Appl.
Stamboliadis, E., McHardy, J. and Salman, T. (1976). The effect of copper ions on the cyanida-
tion of a gold ore. Can. Min. Metal!. Bull., 69, 128- 130.
Stamboliadis, E., McHardy, J. and Salman, T. (1978). Ion exchange techniques for the recovery
of gold from cyanide solutions. Can. Min. Metal!. Bull., 71, 124-127.
Stammreich, H., Chadwick, B.M. and Frankiss, S.G. (1968). Vibrational spectrum and struc-
ture of solid thallium dicyanoaurate. J. Malec. Spectrosc., 1, 191.
Tataru, S.A. (1974). In Proceedings of the Tenth International Mineral Processing Congress
1973, edit. M.J. Jones, pp. 1135-1145, IMM, London.
Taylor, P.R. (1983). Copper and iron impurities in the cyanide leaching of gold. Hydrometall.
Copper, Its Byprod. Rarer Met., edit. L.A. Haas and D.R. Weir, pp. 86-94. Am. Inst. Min.
Metal!., Petrol. Eng., New York.
Tomlinson, D. (1984). Chemelec - The novel electrochemical metal recovery process and its
various applications in the mining/refining field. TMS paper selection A84-55. Metall. Soc.,
AIME., Warrendale, Pa.
Tsuchida, N., Ruane, M. and Muir, D.M., (1984). Studies on the mechanism of gold adsorp-
tion on carbon. International Conference on Recent Advances in Mineral Science and
Technology (Mintek 50), vol. 2, edit. L.F. Haughton, pp. 647 -656. Council for Mineral
Technology, Randburg.
Turkdogan, E.T. (1983). Physicochemical Properties of Molten Slags and Glasses. Metals Society,
London.
Tyson, A. G. (1983). The Chemelec cell for recovery and recycling of metal in the electroplating
industry. In Int. Conf Heavy Met. Environ., pp. 988 - 991. CEP Consultants, Edinburgh.
Van Deventer, J .S.J. (1984). Kinetic models for the adsorption of gold onto activated carbon.
904
International Conference on Recent Advances in Mineral Science and Technology (Mintek

50), vol. 2, edit. L.F. Haughton, pp. 487 - 494. Council for Mineral Technology, Rand burg.
Varentsov, V .K., Luk'yanov, V .0. and Blaginina, N. V. (1984). Electrolytic extraction of noble
metal from alkaline eluates on graphitic carbon cathodes. Sov. J. Non-Ferrous Metal., 25,
39-40.
Viljoen, E.A. and Mihalik, P. (1968). Determination of the Nature of Coatings on Gold Par-
ticles. Report no. 385. National Institute for Metallurgy, Johannesburg.
Wagman, D.D., Evans, W.H., Parker, V.B., Halow, I., Bailey, S.M. and Schumm, R.H. (1965).
Selected Values of Chemical Thermodynamic Properties. U.S. National Bureau of Standards,
Washington, D.C. NBS Technical Notes 270-3 and 270-4.
Walker, G.A. (1956). The chemical treatment of gold slimes for the recovery of high purity
gold and silver. Proc. Australas. Inst. Min. Metal!., 180, 21-54.
White, H.A. (1905). The solubility of gold in thiosulphate and thiocyanate. J. Chem. Metal!.
Min. Soc. S. Afr., 6, 109-111.
Williams, D.F. and Glasser, D. (1985). The modelling and simulation of processes for the recovery
of gold from solution using activated charcoal. J.S. Afr. Inst. Min. Metal!., 85, 237 - 244.
Young, G.J.C., Douglas, W.D. and Hampshire, M.J. (1984). Carbon in pulp process for recover-
ing gold from acid plant calcines at President Brand. Min. Eng., 36, 257-264.
Zadra, J.B., Engel, A.L. and Heinen, H.J. (1952). Process for Recovering Gold and Silver
from Activated Carbon by Leaching and Electrolysis. U.S. Dept. of the Interior, Bureau of
Mines, R.I. 4843.
Ziegler, M. and Glemser, 0. (1956). Extractable halogenoid complexes of the heavy metals.
Angew. Chem., 68, 620-623.
Zvyagintsev, O.E. and Zakharov-Nartissov, O.I. (1960). The distribution of dicyanoauric(I)
acid between aqueous solutions and certain alcohols and ketones. Chem. Abstr. 55, 19436i.
905
Chapter 16
Plant Design and Commissioning

G.G. Stanley
16.1 The Importance of Good Plant Design and Punctual

Commissioning
Two very important aspects of extraction metallurgy for which the
metallurgical engineer is responsible are plant design and plant commission-
ing. Their importance arises from the fact that both are closely related to
the financial success of the overall project of which they form part. In the
long term, good design is imperative if the profit on the capital invested in
the plant, and indeed in the whole undertaking of which the plant is part,
is to be maximized. Furthermore, good design can itself contribute greatly
to easing commissioning problems, and together with a carefully-planned
and executed start-up can make possible the crowning success of a plant
brought into production on time, at designed capacity and efficiency and
within budget. When each day of lost production can cost thousands of
Rands, commissioning delays can become extremely costly.
16.2 The Design and Construction Organization

Before a plant of any significance can be designed and built, an organiza-
tion for carrying out these tasks must be set up. Several types of organiza-
tion are possible, and some of these are:
(1) Design and construction done entirely 'in-house', i.e. entirely by the
owner company w!th its own design and construction resources. In
South African practice where the vast majority of mines are founded
by one or other of the mining groups, the term 'owner company' would
include the founding group.
(2) Design and construction done entirely on contract by an outside firm
on a 'turn-key' basis, i.e. the plant is handed over to the owner as a
going concern on completion of commissioning.
(3) Design done by the owner company and construction by outside con-
tractors. This is the system most frequently used for South African gold
plants.
(4) A joint owner-contractor design and administration organization where
both parties supply personnel and resources on an agreed basis, but
construction is usually done by the contractor or a sub-contractor.
There are several other possibilities, and the owner company will choose
whichever best suits its circumstances. Where an outside contractor is involved
907
as a prime contractor on design and/ or construction, it is highly desirable

that such contractor be thoroughly familiar with local conditions, and
preferably be able to demonstrate the successful execution of several local
contracts.
16.3 Status, Responsibilities and Qualities of the Design Metallurgist

Regardless of the type of organization chosen, it is absolutely essential to
success that a metallurgical engineer be included in the Project Team and
that in matters affecting metallurgical design, the metallurgist' s decision shall
be supreme. All other disciplines must accept that in this instance they are
there to serve the metallurgist' s purpose and that the metallurgist is not simply
there to act as an adviser to them; his requirements must be paramount, even
though, as is likely, he will not be in command of the design and construc-
tion organization. The metallurgist carries the responsibility for the technical
success of the project, and consequently must insist that he shall have the
ultimate say regarding metallurgical matters. This, of course, is not to say
that the metallurgist is to be totally inflexible in his attitude. All engineering
is a compromise, and it is the metallurgist' s obligation to keep an open mind,
to ensure that whatever compromise has to be made is the best possible one
from the metallurgical viewpoint, and with imagination, even to seize some
benefit from it. The metallurgist assigned to a design team must also realize
that the achievement of a successful design is probably more dependent on
him than any other member of the team; it requires of him much hard work,
meticulous attention to detail at all stages, experience, imagination, deter-
mination, the ability to communicate, and above all, a profound knowledge
of the technical and operating aspects of the subject. Armed with these
qualities, he will have the authority to ensure that the metallurgical aspects
of the plant design will receive the priority they must have if success is to
be achieved.
16.4 General Procedure for Plant Design

16.4.1 The procedural plan
The efficient execution of all design and construction projects requires a
proper plan of procedure, that is, the setting-up at the initiation of the pro-
ject, of a list of the steps that must be completed, arranged in the sequence
in which they must occur for the project to proceed in a logical and orderly
fashion. Such a plan, in the case of an ore-treatment plant, could be as
follows:
(1) Ore testing
(2) Process definition
(3) Production of basic flowsheets
(4) Production of piping and instrumentation diagrams
(5) Production of general arrangement drawings and conceptual models
(6) Equipment selection and specification
(7) Costing and preparation of definitive budget
908
PLANT DESIGN AND COMMISSIONING
(8) Production of final flowsheets

(9) Production of detailed design drawings and models
(10) Construction
( 11) Commissioning
The steps in the plan frequently overlap, and furthermore, there are many
other activities involved in the creation of an operating plant, but those listed
are the ones normally involving the metallurgical engineer, and these will
be dealt with in the following discussion.
16.4.2 Ore testing
The first step in designing a successful ore-treatment plant is to obtain the
most comprehensive information possible regarding the ore to be treated.
The more comprehensive the information, the less will be the likelihood of
unpleasant surprises regarding the nature of the ore after the plant has been
commissioned, surprises which can be extremely inconvenient and costly to
deal with once the plant has been built. In other words, the more and better
the information, the more likely will it be th<J.t the design will be 'right the
first time'. The relatively small cost of obtaining the maximum possible in-
formation regarding the ore before plant design commences is usually money
well spent, and should not be begrudged.
Generally speaking, the object of testing Witwatersrand and Free State
gold ores is simply to check their amenability to the known processes of gold
recovery such as comminution, concentration, thickening, cyanidation at am-
bient pressure and temperature, liquid/ solid separation, and recovery of gold
from solution; and to obtain design data for the application of these pro-
cesses. In other words, known processes are capable of yielding acceptable
recoveries so that process development is not necessary. The same remarks
apply in general to the Barberton sulphide ores, which although more refrac-
tory than Witwatersrand and Free State ores, can usually be made to give
acceptable yields by suitable application of known techniques. The subject
of Process Development is therefore not dealt with in this book, although
it is conceivable that some South African gold deposit will yet require the
development of completely new techniques to enable profitable recovery of
its gold.
16.4.2.1 Obtaining the sample

Orebody samples
At this stage of the exploitation of the Witwatersrand conglomerate reefs,
practically any new area developed is likely to be covered by at least a thou-
sand metres of younger rocks. Thus, unless the area is rapidly accessible to
already-existing development, the first bulk sample of ore available is that
obtained from a shaft intersection. But the time necessary for the design and
construction of a large new plant can be five years or more, and the amount
of capital tied up in shaft-sinking and development for a new Witwatersrand
mine is enormous. This capital cannot become productive until the plant is
commissioned. Consequently plant design cannot await exposure of the
orebody in shafts or other development. Thus the only samples normally
909
available when a new Witwatersrand plant is designed are a very limited

number of borehole cores.
In the case of South African gold deposits outside the Witwatersrand-
Free State conglomerate system, the mining depths involved are very much
less and the scale of operations is normally far smaller. Consequently the
amounts of money involved .are vastly less. There is tµus a greater possibili-
ty in the case of these non-Witwatersrand deposits that plant design can be
delayed until good orebody exposures have been obtained in underground
development. But even in the case of non-Witwatersrand deposits, it frequent-
ly is the case that the plant must be designed on the basis of borehole cores
only.
The amount of sample contained in a single borehole core is of course
very limited, and it should be standard practice to drill several deflections
from the original hole whenever values are intersected.
Regardless of the source of the initial samples, it is highly unlikely that
they will be fully representative of the ore body. The design metallurgist must
attempt to compensate for this by drawing on all available information regard-
ing deposits as close as possible in nature and location to the one he is con-
cerned with, and also by allowing sufficient latitude in his design to accom-
modate any likely variations in ore characteristics. In this regard, Witwaters-
rand metallurgists are fortunate in having to deal with a deposit which is
almost unique in the consistency of its properties over several hundred
kilometres, and about whose characteristics so much is already known.
Sampling surface dumps and dams

In contrast to unmined ore, surface dumps and dams are usually reasonably
easily accessible. They do however, present considerable sampling difficulties
of their own, and special techniques have to be employed for each type of
dump.
Rock dumps can be sampled from shafts sunk through them, adits driven
into them or channels cut through them, any of which techniques really calls
for mining expertise. However, the problem of the accurate sampling of coarse
rock produced by these primary operations is well-known, and involves a
large-scale splitting operation ranging from automatic sampling to selection
of a certain proportion of loaded conveyances. Failing such splitting, the
whole primary sample must be submitted to a pilot-scale plant operation
designed to produce the required design information. Care should be taken
when obtaining the primary sample from a surface rock dump containing
a mixture of fine and coarse material (i.e. almost any dump containing
material that has not previously been washed), to avoid inaccuracy due to
the downward migration of the fines resulting from the percolation of rain
water. Dense constituents of the dump, particularly gold and uranium
minerals, can be carried to considerable depths into the underlying soil.
Interest in sampling sand dumps, usually old tailings dumps, has increas-
ed of recent years with the rising gold price, while sampling of slimes residue
dams has frequently been undertaken in connection with the uranium and
pyrite recovery programmes in South Africa, and for obtaining stability data;
910
(see Chapter 12). The method usually adopted in all these cases is auger sampl-
ing, the resulting hole being lined with piping of suitable diameter to pre-
vent collapse of the hole and contamination of the sample with wall material.
For holes of any appreciable depth, some form of lifting tackle has to be
provided for removal of the charged auger from the hole.
16.4.2.2 Mineralogical examination

Mineralogical examination is normally carried out by a specialist mineralogist,
and in the case of ore samples is done on polished sections of the rock. Techni-
ques used include optical microscopy, electron microscopy, microprobe
analysis and gangue dissolution (Lamos, 1982). The object of mineralogical
examination is twofold: first to establish the composition of the rock, the
mineral species present and their proportions, and second to determine the
association of the gold, that is, the proportions that are 'free', that are lock-
ed within other minerals, particularly sulphides, and the size and range of
the gold particles. Knowledge of the mineralogy identifies minerals which
might cause difficulties in metallurgical treatment, and can also reveal the
presence of minerals of potential economic value in addition to the gold,
silver, sulphur and uranium for which samples are normally assayed. (For
example, nickel and cobalt are known to occur in the Witwatersrand con-
glomerates and could become of economic significance in future.) Minerals
which can cause difficulties in gold-silver extraction are those which adversely
affect the cyanidation process, those which adversely affect other phases of
the treatment process such as grinding, concentration, thickening or filtra-
tion, and those which can cause environmental problems. In the first category
are sulphides and minerals containing arsenic, antimony or copper (see
Chapter 15) and carbon, while in the second are very hard minerals which
could cause crushing and grinding difficulties, clay minerals and
phyllosilicates which can cause transport, thickening anci filtration problems
and also gold adsorption problems from cyanide solution, and finally, in
the third category, mercury and its minerals.
Mineralogical methods can examine individual pieces of rock only. Hence
it is important that a large number of pieces, as representative as possible
of the likely feed to a treatment plant, should be examined in order to avoid
any bias in the overall picture. Because of the fact that gold normally occurs
in extremely low concentrations, it is very difficult to derive arty useful in-
formation about it by' mineralogical examination of polished sections. (In
South African gold ores, silver occurs as a solid solution in the gold and hence
is seldom reported on although present to the extent of 10% or more of the
gold.) However, mineralogical investigation of a new ore should include ex-
amination of a concentrate of the heavy constituents (most easily obtained
by heavy medium separation of a milled ore sample using bromoform as the
medium); this can give some information regarding the nature of the gold
and uranium mineralization, particularly with regard to particle size and sur-
face condition and the extent of locking within other minerals, all of which
can be extremely useful from the plant design viewpoint.
911
16.4.2.3 Sample preparation

The primary sample
The type of preparation necessary depends on the type of sample available
and the testing to be done. In the case of borehole cores, preparation before
crushing can consist of little less than selection of appropriate lengths of the
core to include the correct proportions of reef and waste, followed possibly
by grouping of the cores into combinations representative of specific areas
of the orebody if it is suspected that the ore properties may vary from area
to area. After crushing and mixing, actual metallurgical testing can com-
mence, although the extent of the testing possible on such a limited sample
as borehole cores is obviously restricted. Priority should be given to batch
cyanidation tests as these give a good overall picture of the amenability of
the ore to treatment; thereafter, depending on the amount of sample available,
testing can be progressively extended to other aspects of treatment. But it
must be borne in mind that even a single rolling bottle cyanidation test will
require several kilograms of sample, allowing for the necessary quantities
for assay, if any reasonable accuracy is to be achieved.
Bulk samples permit much more scope in metallurgical testing than
borehole cores. The necessity for splitting or 'sampling the sample' if the
original bulk sample is too large, has already been mentioned in Section
16.4.2.1; accurate splitting of run-of-mine samples is always difficult, and
should be avoided, if possible, until at least primary crushing has been
performed.
Methods for splitting or sampling coarse materials include:
(1) Use of an automatic sampling plant specially designed for handling run-
of-mine feeds. This can be single- or multi-stage, depending on the ex-
tent of reduction and the accuracy required.
(2) Conveying the material in a large number of small batches, as in front-
end loader scoops or railway trucks, and the selection and separate
dumping of a certain proportion of the batches to form the sample.
(3) Forming the primary sample into a homogeneous bed by depositing
it in successive layers, one above the other, in an elongated heap, and
then removing a transverse slice from the heap to form the sample.
(4) Splitting the material into size fractions, determining the relative pro-
portions of the fractions and then recombining some of each of the
fractions in the correct proportions to form the sample. This method
obviously yields the size distribution of the original sample, which is
invaluable where it represents the feed to a crusher station or to a run-
of-mine milling operation.
Other methods of splitting coarse bulk samples can no doubt be devis-
ed, but these examples indicate the sort of operation involved. In all cases,
great care must be taken to avoid loss of fines, since the fines are normally
enriched by already-released valuable components; where the possibility of
dusting losses exists, it may be advisable to dampen the sample slightly.
The relationship between sample size and accuracy is discussed in
Chapter 13.
912
At this stage, unless run-of-mine milling tests or possibly sorting tests

are to be done, size reduction of the selected portion of the primary sample
can commence. This is usually done by means of pilot-scale jaw crushers
or some other type of crusher capable of handling wet fines, if no washing
of the sample has yet been done. Where a considerable degree of size reduc-
tion (say greater than 4: 1) is required, successive passes through a progressive-
ly more tightly adjusted crusher may be necessary. The opportunity should
not be missed of collecting data on crushing energy requirements, although
this will require the measurement of both ore mass and crusher energy input.
Primary crushing can bring the maximum particle size into the 25 to
30 mm range, at which size mixing and splitting are considerably easier than
on run-of-mine material. Suitable splitting methods at this stage include: (1)
coning and quartering, (2) riffling, and (3) use of single- or multi-stage
mechanical splitting plants, and various others.
The secondary sample

Splitting of the primary sample as just described yields a secondary sample
which will normally be the source of material for actual metallurgical testing.
The size of the secondary sample will depend on whether the subsequent
testing is to be at laboratory or pilot plant scale; if the former, the secon-
dary sample can be of the order of 100 kg, while for pilot plant testing,
anything up to 100 t or more could be required. In the case of pilot-scale
testing, crushing of the primary sample can be done on a continuous basis
and form part of the actual testing.
16.4.2.4 Laboratory testing

Metallurgical testing of gold ore that can be carried out at laboratory scale
includes:
(a) Gold, silver, sulphur and uranium assays.
(b) Conventional grindability determination.
(c) Pebble competence testing and autogenous grindability determination.
( d) Thickening.
(e) Cyanidation, including reagent consumption and effects of fineness of
grind and contact time.
( t) Filtration.
(g) Concentration by gravity methods and flotation.
(a) Assaying
Gold, silver, sulphur and uranium assays should be done as a mat-
ter of routine on new Witwatersrand ore head samples. For Barberton
area ores, arsenic and antimony assays should be included. Silver can
usually be omitted once the head gold: silver ratio has been determin-
ed. Uranium can be omitted when gold-only treatment is envisaged,
although its behaviour in any concentrating operation for other con-
stituents should be checked on. 0,5 kg of sample is sufficient for a head
assay and 1,0 kg for a residue assay, both in triplicate.
913
(b) Conventional grindability determination

The determination of grindability (i.e. the energy necessary per unit
mass to produce a given size reduction) in rod, ball and pebble milling
at laboratory scale is usually done by the Bond method, which pro-
duces the Bond Grindability Index for the sample. The method is well
described by Pownall (1962), who also describes the use of the result
in mill size selection, a subject that is more fully dealt with in Appen-
dix 3.4 of Chapter 3.
25 kg of sample is required for determination of the Bond Index
at several different mesh sizes with some check determinations if
necessary.
In connection with grindability testing, it should be borne in mind
that batch laboratory tests do not produce the same size distributions,
particularly of minerals and metals, as does continuous plant grinding.
Continuous pilot plant testing is perhaps the answer here.
( c) Pebble competence testing and autogenous grindability determination
The suitability of an ore for autogenous milling is partly deter-
mined by the competence of the large sizes to form grinding media,
i.e. their ability to experience the forces in a tumbling mill with a low
probability of shattering. Three methods of testing this characteristic
have been employed, namely simple dropping of representative pieces
onto a hard surface, tumbling a sample of the large pieces in a rotating
drum, and pendulum crushing of sample pieces. Apparently only the
first of these methods has been used in South Africa; it results in a
'number of drops versus per cent remaining mass' curve (Figure 16.1),
which is compared with the curve for an ore of known milling
characteristics.
Autogenous grindability is measured at batch scale by the use of
a simulated closed circuit test, similar to the Bond grindability test. Both
pebble competence and autogenous grindability tests are discussed fur-
ther in Appendix 3.3 of Chapter 3.
( d) Thickening
Thickening tests are usually done by measuring the settlement rate,
i. e. the rate of fall of the slime/ clear water interface of the pulp in a
transparent measuring cylinder. The method of performing the test and
of using the information obtained for the calculation of thickener
capacity is given in an appendix to Chapter 18(Appendix18.3). Papers
by Moncrieff (1963), Barnea (1977) and Wildhelm and Naida (1979)
deal with the subject. Thickening tests can be done on the products
from grindability testing, so should require no additional quantities of
original sample.
( e) Cyanidation
Since, in the case of Witwatersrand-type ores, cyanidation alone
is normally capable of dissolving well in excess of 90% of the gold,
cyanidation is the most significant test of the ore's amenability to treat-
ment that can be made. Barberton ores, however, usually require some
form of pretreatment to render the locked gold accessible to cyanide
914
Number of drops 10 20 30 40 50
ll7o Mass remaining 57,5 47,0 37,1 31,3 26,1
Drop height = 2,74 m
20+-~~~~~~~~~~~~~~~----t-
o 10 20 30 40 50
Number of drops
Figure 16.1. Number of drops versus per cent remaining mass

for Leslie ore.
and to eliminate arsenic and antimony, but thereafter cyanidation is

usually capable of making a significant extraction of gold, so that again,
the cyanidation test is of crucial significance.
Cyanidation testing at laboratory scale is normally done by the
rolling-bottle method. Where no pretreatment is required, testing can
be done directly on the product of gi;indability determination, although
if a fineness of grind versus extraction relationship is to be obtained,
a number of grinds in addition to those for grindability testing will be
required. If at all possible tests should also be done for different con-
tact times to determine a contact-time versus extraction relationship
since this is very important information for plant design. The cyanida-
tion test can also give useful information regarding cyanide and lime
consumption. Assuming that 10-litre bottles are used about half full
at pulp density of 1,46 (i. e. 1: 1 liquid: solid), about 3,6 kg of solids will
be required for each bottle charge. If 4 grinds at a standard treatment
time of 30 hours are tested together with 3 other treatment times in
the range 25 to 40 hours at say 80 OJo - 75 µm, a total of 7 tests will
be done. These should be duplicated, so that the quantity of sample
required for reasonably comprehensive testing will be 14 x 3 ,6 = 50,4 kg.
Refer to Chapter 18, Appendix 18.5, for details of the cyanidation
test.
(f) Filtration
Testing for filtrability can, and should, be done on the products
915
from the various cyanidation tests. The tests usually comprise measure-
ment of the rate of the removal of filtrate from a feed slurry by batch
vacuum filtration under standard conditions in a Buchner funnel, or
using a specially constructed miniature filter leaf submerged in the pulp
sample. Refer to Chapter 18, Appendix 18.4, for details of filtration
testing. Methods of using the results in the calculation of filter duty
are given by Osborne (1975) and Dahlstrom and Purchas (1957). The
tests should also include a determination of the dissolved gold losses
to be expected at various applied wash ratios (i.e. ratio of applied wash
to solids filtered) above and below about 1: 1 by mass.
(g) Concentration
Because of their cheapness, the preferred plant-scale methods of
concentrating gold and other dense constituents of South African ores
are gravity methods .. However, it is difficult to simulate gravity con-
centration satisfactorily in batch-wise bench-scale operation with the
type of equipment used in full-scale plants, such as Johnson drums,
continuous riffle belt concentrators and strakes such as plane tables.
But laboratory equipment specifically designed for this purpose, e.g.
the Haultain Superpanner and the micropanner, and even bench-scale
batch heavy medium separation using bromoform (r.d. = 2,83) diluted
as required with xylene (r.d. = 0,86) or carbon tetrachloride (r.d. =
1,59), can give some indication of the concentrate quantities and grades
to be expected at various grinds. 2 kg of sample should be sufficient
for these tests.
Flotation is used in some Witwatersrand and Free State gold plants
for concentrating sulphides for the production of sulphuric acid and
with the secondary aim of enabling further intensive treatment of the
sulphides for the recovery of locked gold (see Chapter 5). In the case
of Barberton-type sulphide ores, a gold-sulphide concentrate is made
which is roasted to render the gold-enclosing sulphides porous and to
eliminate arsenic and antimony to enable successful recovery of the gold
by cyanidation (see Chapter 8). Where cyanidation tests have given poor
recoveries owing to a high proportion of gold locked in pyrite or because
of the presence of arsenic or antimony, flotation tests should be in-
cluded in the laboratory testing. These will indicate the recoveries of
sulphides and gold that can be expected by this method, and may yield
some information about required flotation contact times and reagent
consumptions. If sufficient concentrate can be made, information about
its subsequent treatment for recovery of the gold can be obtained. 10
kg of sample will be sufficient for flotation tests alone, but anything
up to 100 kg, will be required if adequate concentrate is to be made
for exploratory work on concentrate treatment.
The performance of all the laboratory tests indicated in Section
16.4.2.4 will require something approaching 6 tons of sample (200 kg
if the pebble competence and autogenous grindability tests are exclud-
ed), but will yield a picture of the ore's treatment requirements which
in most cases will be entirely adequate for the design of a successful
916
Table 16.1. The quantities of sample required for the performance of the various types of
laboratory test.
Sample mass required

Type of test (kg)
Head assay (triplicate) 0,5

Residue assay (triplicate) 1,0
Ball, pebble or rod mill grindability (Bond test) 25
Cyanidation 50
Heavy medium concentration 2
Flotation 10
Concentrate roasting JOO
Pebble competence and autogenous grindability 6000
flowsheet. Table 16.1 summarizes the quantities of sample required for

the various tests.
16.4.2.5 Pilot plant testing

Generally, laboratory scale testing, together with data from existing plants,
will yield all the data necessary for the design of plants to treat non-refractory
ores such as the Witwatersrand and Free State conglomerates. However, in
the case of relatively unknown and refractory ores, it may be advisable to
include pilot-scale testing as well. Pilot testing and the reasons which could
make it advisable are discussed in Chapter 18. It must, however, be borne
in mind that pilot plant testing is expensive and should not be done
unnecessarily.
16.4.3 Process Design

16.4.3.1 Process Design Criteria
The design process proper commences with the drawing up of the Process
Design Criteria. This is essentially a statement of what the plant will be re-
quired to do and the framework within which it will have to do it. It includes
a statement of the capacity the plant is to have, what material it is to treat,
the sources of its feed, what it is to produce, the time schedule for the com-
missioning of the various stages and general information regarding the ex-
ternally imposed parameters of the plant design. It is normally prepared by
the mining and financial consultants of the company that will own the plant.
It deals essentially with what the plant is to achieve, and as such is the basic
directive to the plant designers, setting the limits within which they must
operate, and the targets they must attain. The design metallurgist must in-
sist that he be given the Process Design Criteria as part of the essential
documentation of his commission.
16.4.3.2 Flowsheet design

Whereas the Process Design Criteria lay down what is to be achieved by the
plant, the flowsheet has to do with the means by which the objectives are
917
to be attained. It is a diagrammatic definition of how the requirements

specified in the design criteria are to be achieved, and its preparation is essen-
tially the business of the metallurgical engineer. Flowsheet design is a major
and vital part of the design process, and the correct choice of flowsheet is
crucial to the technical and financial success of the plant to be built.
The design process

Flowsheet design is the arranging in diagrammatic form of the necessary
equipment, installations and interconnections to achieve the goals specified
in the design criteria, while complying with the treatment requirements in-
dicated by test work or any other source of information. Initial flowsheet
design usually takes the form of a drawing-up of very rough freehand block-
type conceptual diagrams based on the design metallurgist' s knowledge of
the subject. These preliminary diagrams are subjected to a process of con-
sideration, discussion, analysis and development by the designer and his
associates, and as a result gradually evolve to a stage where they can either
be rejected or handed over to a draftsman for drawing up in a more standar-
dized format. It is usually the case that several technically feasible treatment
routes will emerge from this initial design activity. Almost invariably, an
estimate of the capital cost of the proposed plant will be required at an early
stage of the project, particularly if a decision has to be made among several
possible alternatives. Accordingly the aim of this first stage must be to pro-
duce a flowsheet of each possible treatment route in sufficient detail to per-
mit the preparation of a capital cost estimate. However, great detail is to
be avoided, since generally all that is required at this stage is an Order of
Magnitude estimate as quickly as possible. This estimate is based on rough
prices for the major equipment in each plant area (e.g. the mills), increased
by rule-of-thumb factors to cover the ancillary equipment and buildings. For
example, the cost of a complete milling unit including foundations, ore
storage, conveyors, sumps, pipes, buildings, etc. is approximately five times
that of the mill and drive itself. Another useful rule-of-thumb is that the
capital cost of gold extraction plants varies as (capacity)'13 , while Ruhmer,
Svoboda and Wilson (1984) give relationships between price and capacity
of individual types of machine. Greater detailing and progressively more ac-
curate costing are called for as the project proceeds and the design becomes
increasingly more defined, passing from the Order of Magnitude estimate
to the Preliminary Estimate (-15% + 25%) to the Definitive Estimate
( -10% + 10%) and finally the Detailed or Revised Estimate ( -5% + 5%).
16.4.3.3 Quantified flowsheet

For a flowsheet to be of use in subsequent costing, evaluation and design
stages, it must be quantified, i.e. include the necessary information regar-
ding the flow streams throughout the plant, and the equipment to be instal-
led in the plant. The preparation of a quantified flowsheet is greatly facilitated
by the use of a standardized form of drawing sheet, the AO being a conve-
nient size, on which are printed tabulations for the entering of flow (or stream)
data and of equipment data. The flow data table is printed at the bottom
918
Table 16.2. Example of a flow data table.
Water:
Stream Descrip- Solids Water Pulp % solid Other
No. tion (t/h) (t/h) (t/h) Solids r.d. ratio m3/h m3/min data
2
3
etc.
of the sheet, and may be headed as shown in Table 16.2.

This tabulation can be repeated across the whole width of the sheet, ex-
cept for the title space in the right-hand corner; it can occupy 10 to 200Jo
of the top-to-bottom dimension of the sheet.
A flowsheet of any appreciable size will probably have to be divided
into sections that can conveniently be accommodated on the printed sheets
described, so that in fact 'the flowsheet' will become a set of flowsheets. A
section of the overall flowsheet is then drawn on each sheet of the set in the
appropriate detail, using stylized representations of the various pieces of
equipment and lines of different types for the interconnecting flows, e.g. solid
lines for ore and pulp, dotted for water, chain dotted for compressed air,
and so on.
In the case of duplicated streams, it is necessary to show only one stream
on the flowsheet, provided it is made clear under the heading 'Description'
of the flow data table that the stream shown is 'one of X', X being the number
of parallel streams; the quantities under the other headings will then be for
one stream only.
Each stream in the flowsheet is numbered and entry of data into the
flow data table commences with the insertion of a description and all available
primary data for each stream against its number in the table. Primary data
are those data available at this stage; they are the data on which the flowsheet
design to this stage has been based, and are usually obtained from the des~gn
criteria and test results.
Initially, all flow rates must be either monthly flow rates or must be
based on the full number of time units in the length of month specified in
the design criteria, taking no account of the fact that for various reasons
actual running times will be less than lOOOJo of calendar time. They should
be pencilled in at this stage. These initial flow rates are subsequently cor-
rected to actual flow rates when the actual running time for each section of
the plant has been estimated as explained in the section 'Estimating Run-
ning Times' further on in this chapter.
When all primary data have been entered, the secondary, or derived,
data are next calculated or estimated and inserted in the tabulation. The secon-
dary data are frequently derived from flow balancing around sections of the
flowsheet, and an example of this follows.
919
50 t/h
150 t/h W/S = 1,2
2 x 50 t/h
W/S = 0,53
Inlet dilution ~ Discharge dilution
'$t~~~
,,_11,ll)o~ Ball mill
~A
D
B
Figure 16.2. Flow diagram with primary data.
Flow balancing
Suppose we have that the new feed rate to a ball mill-classifier circuit is 50 dry
t/h of crushed ore containing 80Jo moisture, that the mill discharge w/s is
to be0,42 and the cyclone underflow w/s is to be0,53. Experience has shown
that with the size of cyclone to be used, the overflow w/ s will have to be
about 1,20 to give the required overflow size distribution, and that a cir-
culating load ratio of 2: 1 is to be expected. To complete the flow rate tabula-
tion, we are required to calculate all the flows and associated data around
the circuit, including dilution water flow rates. The method is firstly to draw
a rough flow diagram with all the primary data entered on it (Figure 16.2).
Next draw up, on the bottom of the flowsheet, a flow rate table headed
as in Table 16.2 and enter on it, in the sequence in which tliey appear on
the rough flow diagram, all the flows into and around the circuit. The tabula-
tion will then look like Table 16.3.
Then, in the tabulation,
A2 (i.e. line A, column 2) = (100 x Al)/ A4 - Al 4,4
A3 Al + A2 = 54,4
A7 = Al/2,7 + A2 = 22,9
A8 = A7 x (1000/60) = A7 x 16,67 = 382
B5 (B6+1)/(B6 + 0,37) (by formula)= 1,7
B2 Bl x B6 = 53
B3 Bl + B2 = 153
B7 (Bl/2,7) + B2 = 90
D5 (D6 + 1)/(D6 + 0,37) (by formula) 1,8
D2 = Dl x D6 "" 63,0
D3 Dl + D2 = 213
D4 = (Dl/D3) x 100 = 70
D7 = (Dl/2,7) + D2 = 118
D8 = D7 x 16,67 = 1974
C2 = D2 - (A2 + B2) = 5 ,2
C7 C2 5,2
CS = C7 x 16,67 = 86,7,
920
Table 16.3. The primary data of Figure 16.2 entered into a flow data table. The line letters and column numbers would normally be omitted; they
are included here to identify quantities mentioned in the text.
2 3 4 5 6 7 8 9
Water:
Stream Descrip- Solids Water Pulp O/o solid Other
A 1 Circuit feed 50 92
B 2 Cyclone underflow 100 0,53
c 3 Mill inlet dilution
D 4 Mill discharge 150 0,42
E 5 Mill discharge dilution 'V
F 6 Cyclone feed 150 1,20 l'
G 7 Cyclone overflow 50
z>
>-l
tl
Table 16.4. The completed flow data table for the flowsheet of Figure 16.2.
tI1
(fJ
.......
0
2 3 4 5 6 7 8 9 z
ztl>
Water:
Stream Descrip- Solids Water Pulp O/o solid Other
(]
A 1 Circuit feed 50 4,4 54,4 92 - 22,9 382 0
$;
B 2 Cyclone underflow 100 53 153 1,70 0,53 90
$;
c 3 Mill inlet dilution 5,2 - 5,2 86,7 .......
(fJ
D 4 Mill discharge 150 63,0 213 70 1,80 0,42 119 1969 (fJ
.......
E 5 Mill discharge dilution 118,4 - - - 118,4 1974 0
F 6 Cyclone feed 150 180 330 45 1,40 1,20 236 3943 z
>-<
"'....
1:-.)
~
G 7 Cyclone overflow 50 128 178 28 1,215 2,56 147 2442 z
0
and so on. The mill discharge dilution is calculated by difference between

cyclone feed and the mill discharge, and finally cyclone overflow is the dif-
ference between cyclone feed and cyclone underflow. The calculations are
obviously well suited for performance on a programmable calculator. The
completed tabulation will be as shown in Table 16.4.
A 7 and AS are included as they provide a useful check on the accuracy
of the calculation, for the sum of all the flows entering the circuit should
equal the only flow leaving it, i.e. A7 + C7 + E7 = G7 and likewise for
the corresponding figures in column 8.
This example shows how a large amount of useful output data can be
inferred from a small amount of input data in the case of a closed circuit.
The final step would be to transcribe the data into the flow data tabulation
on the flowsheet.
Estimating running times

The flow rates initially entered in the flow data tabulations are based on 100%
running time. Before these can be corrected to actual flow rates, estimates
of actual running times must be made.
The factors involved in estimating running times are:
(1) the scheduled running time in the section of the plant concerned, that
is, the number of hours per month during which it will be manned and
is planned to run,
(2) the proportion of lost time to be expected as the result of random
unplanned stoppages such as breakdowns, power failures, chute
blockages, etc., as well as of routine maintenance where this has to take
place within scheduled running time as defined above.
Of major significance to the setting of scheduled running times are the
provisions of the South African Mines and Works Act (Act 27 of 1956). Sec-
tion 9(1)(c) of this Act states 'no person shall perform or cause or permit
any other person to perform, at any mine or works, any work in connection
with the operation of a mine or works, on a Sunday, Christmas Day, Day
of the Covenant or Good Friday, or cause any other person to perform, at
any mine or works, any such work on Republic Day, unless the work is
operating any continuous chemical, metallurgical or smelting process, if a
stoppage thereof during the whole of any such day would either prevent its
immediate resumption on the next succeeding day or diminish the effectiveness
of the process.' The effect of this is that ore hoisting and operation of crusher
plant may not take place, without special permission, on Sundays, Christmas
Day, Day of the Covenant or Good Friday, but may take place on Republic
Day, which is a non-compulsory working day. Milling and cyanidation, on
the other hand, may be carried out on any and every day of the year. In
addition to the limitations imposed by the Mines and Works Act, crusher
station operation is frequently limited by company policy to less than three
shjfts per day for reasons which include synchronization with ore hoisting
schedules and reducing maximum demand on the electric power supply
system. Thus whereas scheduled milling and cyanidation plant running times
922
are normally 24 hours every day of the year, scheduled running time for
crusher stations has to be calculated by the formula:
Scheduled running hours per month = (Calendar days in month - No. of

Sundays - statutory holidays) x
(number of shifts per day) x 8
For plant design purposes, the calendar days per month would be the
same as those used in the plant capacity specification in the Design Criteria,
and the incidence of statutory holidays would be disregarded. Thus in a 31-day
month, with four Sundays, the scheduled running time for a crusher station
operating a 2-shift day would be
(31 - 4) x 2 x 8 = 432 hours,
whereas milling and cyanidation would be schedule for
31 x 24 = 744 hours.
The fraction of scheduled running time that can be expected to be lost
due to unscheduled random causes (or, more usually, the percentage of
scheduled running time to be expected) depends on the type, design, loading
and condition of the equipment involved, the standard of plant design and
the competence of management and operation, to mention but a few of the
factors involved. In other words, it is the result of many complex factors,
and Table 16.5 summarizes average experience in the South African gold min-
ing industry with regard to running time percentages (i.e. availability percen-
tages) to be expected from various types of machine. Table 16.5 is based
on the assumption that routine maintenance of crusher equipment is done
out of scheduled operating time, but for all other equipment it assumes that
routine maintenance such as planned maintenance, liner renewals, etc., are
done in the scheduled operating time of the plant section concerned. Hence,
for crusher station equipment, operating time in hours per 31-day month
would be estimated as:
. (31 calendar days - 4 Sundays) x number of operating shifts
percentage availability from Table 16.4 h
per d ay x 8 x ours (16.1)
100
and for all other equipment except that intended for intermittent operation,
would be estimated as:
percentage availability from Table 16.4
31 x 24 x hours (16.2)
100
Estimating actual flow rates
When the actual monthly running times for the various types of equipment
in the plant have been estimated, the initial flow rates entered in the flow
rate tabulations on the flowsheet can be corrected by the factor:
Hours in month used for initial flowrate calculation (16.3)
Estimated actual running time in same length month
for the type of machine involved
923
Table 16.5. Percentage availabilities of various types of machinery in South African gold ore
treatment plants.
Percentage availability
Number
of plants
Type of machine in sample Highest Lowest Average
Jaw crushers 11 99 74 90
2 Cone/ gyratory crushers II 99 75 89
3 Vibrating grizzleys 4 99 89 94
4 Vibrating screens 16 99 77 93
5 Photometric sorters 2 95 83 89
6 Radiometric sorters 4 97 70 86
7 Rod mills 10 96 83 91
8 Ball mills 20 96 79 89
9 Pebble mills 24 96 82 90
10 Run-of-mine mills 9 94 83 88
11 Pulp pumps 15 99 63 93
12 Thickeners 27 99 89 96
13 Leach vessels, air agitated 23 99 87 96
14 Leach vessels, mechanically agitated 2 99 95 97
15 Drum filters 31 95 74 85
16 Belt filters 75
For example, in a circuit using Symons crushers, Table 16.5 indicates

that for 2-shift per day operation, the actual running hours to be expected
would be (31 - 4) x 2 x 8 x 0,89 = 384,5 hours. If the initial combined
secondary crusher discharge flowrate entered in the flowsheet tabulation was,
say, 100 t/h, the estimated actual flow rate would be 100 x (744/384,5) =
193 tons per hour.
Bear in mind that Table 16.4 assumes that routine maintenance of crusher
station equipment is done outside scheduled running time, but that
maintenance is done of all other equipment within scheduled operating periods
and therefore causes lost time. Where these assumptions are not valid, the
estimated running times will require appropriate adjustment.
When the original flow rates entered in the flowsheet tabulations have
all been corrected for estimated actual running times they can be inked into
the tabulation.
The flow data tabulation is completed by entering under 'Other Data'
such data as size distributions, pH, temperatures and reagent concentrations.
The completed tabulation then forms the basis for equipment selection as
described in the following section.
16.4.3.4 Equipment sizing and selection

The design procedures so far described have provided some of the essential
data on which equipment sizing and selection can be based, namely the flow
data pertaining to each stream in the plant. The next step is to determine
924
with the help of this data what capacity in terms of area, volume or energy
input is required to bring about whatever change is required in each stream,
whether of position, size distribution, chemical state, moisture content, etc.
That is, using data or design formulae given in this book or elsewhere, it
can now for instance be ascertained what volume has to be allowed for to
provide any required retention time at any point of the circuit, or, what
amount of crushing or milling capacity in terms of kilowatts has to be pro-
vided at each size reduction point, or what amount of screening or filtering
area has to be installed to achieve the necessary separations. Having deter-
mined the total amount of processing capacity to be provided at each point,
together with a factor of safety allowance, it is then relatively simple, by
consulting manufacturers' literature, or using design formulae, to decide on
what combinations of numbers and capacities of machine are required at
each point.
Generally, there will be several combinations of available sizes and
numbers of machine that will fulfil each requirement. For instance, in a mil-
ling plant, the necessary work can be done either by a large number of small
machines or vice versa. The decision as to which is the correct combination is
essentially an economic one, that is, determination of the relative prof-
itabilities of the various alternatives. The capital cost of providing one or
more machines (or some means of overcoming the undesirable effects of stop-
pages, for instance by providing storage capacity) and the cost of incor-
porating these in the plant must be compared with the estimated reduction
in loss of earnings resulting over the life of the plant in each case with due
regard to the time value of money. This essentially amounts to the calcula-
tion of the Net Present Value (NPV) or the Internal Rate of Return for each
possibility, as described later in Section 16.4.3.5. The difficulty always is the
estimation of the effect on future earnings, but with the help of experience
some reasonable estimate can usually be made and a satisfactory determina-
tion made of the relative profitabilities of the various alternatives.
Generally speaking, however, the result of the above calculation will
usually indicate that, in the case of major equipment at any rate, 'big is
beautiful', that is, the use of the least number of large machines is usually
the most profitable choice, with due regard to the limits imposed by technical
feasibility and mechanical reliability. This is because, in general, installed
capital cost per unit of capacity decreases rapidly with increasing machine
capacity. In particular cases, however, the choice might not fall on the big-
gest machine available for various reasons such as anticipated high breakdown
frequencies, departure from standard or incompatibility with desired plant
modularity and loss of too high a proportion of plant capacity coupled with
plant balancing difficulties if a very small number of machines is installed.
But, eventually, by using the NPV method tempered with discretion, a deci-
sion can be reached regarding the question of the numbers and sizes of
machines to be installed to carry out each duty in the plant.
The final steps in the preparation of the quantified flowsheet can now
be carried out, namely the allocation of an identifying letter or group of let-
ters to each machine symbol on the flowsheet and the entering of the rele-
925
vant data against that letter in the Equipment List.

The Equipment List is arranged as a tabulation similar to Table 16.6
at the top of each flowsheet section:
Table 16.6. Example of an equipment list for flowsheet of Figure 16.2.
Item Description No. off Size, capacity, etc. Material
A Ball mill feed belt 750 mm, 70 t/h

B Ball mill 2,4 m.d. x 3 m, 19,6 rpm
c Ball mill discharge sump 1 3 m3 Concrete
D Cyclone feed pumps 2 5000 l/min, r.d. = 1,4 Hard iron
E Cyclone classifiers 2 450 mm, 20° Natural
rubber lined
Like the Flow Data tabulation, the Equipment List tabulation can oc-
cupy 10 to 20% of the top-to-bottom dimension of the sheet and can be
repeated across its whole width.
The flowsheet has now been completely quantified and specified and
can form the basis of further stages in the design process such as flowsheet
choice and civil engineering design.
Choice of equipment supplier

The metallurgical engineer member of the design team must have the final
say in the choice of equipment supplier. The practice of choosing equipment
solely on the basis of lowest price must be avoided.
The choice of which make of machine to be installed depends on such
considerations as:
(1) Suitability as regards performance characteristics and dimensions
(2) Competence of design
(3) Reputation of machine and manufacturer
(4) Price
(5) Delivery time
(6) Back-up facilities and service
(7) Standardization within the plant or larger organization
The prospective purchaser is well advised to investigate these and all other
relevant points very thoroughly, for poorly designed and manufactured equip-
ment, incapable of achieving the maker's claims, can be very expensive in-
deed, regardless of first cost. The old adage 'Quality is the cheapest thing
one can buy' should be kept well in mind.
16.4.3.5 Flowsheet evaluation and selection

Confronted with several possible flowsheets each of which will fulfil the design.
criteria, the metallurgist will require to apply some test for selecting the best
alternative. He may, for example, select the flowsheet with the lowest
estimated operating cost, or the lowest capital cost or the one that requires
926
the least operating labour or the one that gives the greatest recovery. But,
while each of these criteria is admirable in itself, it may not lead to the selec-
tion of the best flowsheet, for in fact there is an even more basic criterion
than those suggested, namely the maximization of the company's profitability,
with which all other criteria must comply. The essential, fundamental, reason
for the existence of the company is to make the maximum possible profit
and if a criterion has any other objective than this, it is invalid. 'Maximum
profit', taking all relevant factors into consideration, including such matters
as environmental preservation, is the basic criterion of flowsheet selection.
The basis for selecting the most profitable alternative

Given that the flowsheet to choose is the one that will maximize profitabili-
ty, the problem is to decide which one will do that. For profit is the
difference between income and expenditure, both of which comprise a number
of factors which can be combined in different ways. For instance, income
can be increased by selling more product, and this can be achieved by treating
more ore per unit time or by extracting a higher proportion of the contained
values. Expenditure can be reduced by reducing the running costs of the plant
and also by reducing the interest payable on the capital cost of the plant.
Lower capital costs and therefore lower interest charges can be achieved by
simplifying the plant design, but this might adversely affect the recovery ob-
tained. Also, the matter of interest raises the question of the time value of
money or how to balance present expenditure against future income. The
problem is to find a means which will take account of all these and the many
other factors involved, and enable a rational choice of the most profitable
design.
Several solutions to this problem are in use, but any valid method must
be based on estimating the profitability over the whole expected life of the
project and not just the first few years or even a single instant; looking at
too limited a period can be totally misleading. In other words, the method
used must take into account both the probable losses in the early years of
the project and balance these against the hoped-for profits in later years and
finally produce a single 'figure of merit' for each alternative.
The discounted cash flow (DCF) method

The method which complies with the requirements just stated is called the
Discounted Cash Flow (DCF) method. It comprises the estimation for each
year of the project's life, including the time spent on design, construction
and commissioning as well as actual production, of the difference between
income and expenditure. 'Income' means the after-tax income from the sale
of the plant's product(s), and 'expenditure' includes capital expenditure as
well as operating costs. The difference between income and expenditure for
each year is the 'cash flow' for that year. Losses are designated as negative
cash flows, and profits as positive. Each of the cash flows so calculated is
then 'discounted' to its Present Value (PV), that is, its value at the start of
the project (taken as 'time zero').
'Discounting' is done by calculating what sum of money invested at time
927
zero at compound interest at a determined rate (discussed later) will amount

to the nett profit or loss estimated for each year by the time the end of that
year is reached. The calculation makes use of the compound interest formula:
A = P[l + (vllOO)]" (16.4)
where A is the amount, or sum of money, that P (the sum invested) will
amount to at v per cent per year over n years. Rearranging,
P = [Al[l + (vllOO)]"], (16.5)
which says that P is the amount of money to be invested now at v per cent
per year to become A at the end of n years. That is, P is the Present Value
of A at v per cent per year for n years.
The two methods of using DCF

There are two main methods of using the DCFs calculated as described in
the previous section to produce a single figure of merit for the flowsheet alter-
native concerned. In the first, the interest rate used for discounting is that
(usually found by trial and error) which causes the sum of the negative DCFs
to equal the sum of the positive DCFs over the life of the project. This rate
is called the Internal Rate of Return (IRR), and the flowsheet chosen is that
which will give the highest IRR. In the second method, the interest rate used
is the highest rate of return the company could obtain by investing its money
in any other way open to it (called 'the opportunity cost of money'). The
difference between the sum of the inflows discounted at the opportunity cost
of money and the sum of the outflows discounted at the same rate is the
Nett Present Value (NPV) of the project, and that flowsheet is chosen which
will give the highest NPV.
The objective of the IRR method is to determine a rate of return, whereas
the objective of the NPV method is to calculate a monetary value. Both of
these quantities represent a figure of merit, or rating, for the flowsheet for
which it is calculated.
Example of the application of the DCF principle
Suppose that, to handle the increasing output from an already producing
mine, a new plant is to be built with a capacity of 200 OOO tons of hoisted
ore per month. The ore grade is estimated to be 8 g Ault. Pfoduction is ex-
pected to commence in the second half of the third year of the project and
it is estimated that 600 OOO tons will be fed to the plant in that year, and
that the designed plant feed rate of 200 OOO t/month will have been attained
by the end of that year. The gold price is taken as R20 per gram. The com-
pany's opportunity cost of money is 18% per year, and the life of the plant
is expected to be 30 yea.rs.
Preliminary test work has shown that if no sorting is practised, run-of-
mine milling can be used and will give a recovery of 96%, i.e. residue assay
value will be 0,32 git. The capital cost of this plant would be RlOO million,
of which R35 million would be spent in the first year of the project, R50
million in the second and R15 million in the final year. Working costs in
this case would be R5 per ton of feed. But if automatic electronic ore sorting
928
is practised, 25% of the ore can be rejected at an assay value of 0,35 git.
This will raise the assay value of the sorted ore to 10 ,5 5 gl t, which, by applying
the square root relationship (Chapter 6, Section 6 .3), will increase the residue of
the treated ore to 0,37 git, and the overall residue including sorted waste will rise
to 0,365 git, i.e. recovery will drop to 95,44%. However, if sorting is
employed, washing and screening of the ore will have to be done. Also, since
sorting will remove a large proportion of the coarse material from the ore
supply, run-of-mine milling will not be possible, and the sorted ore will have
to be crushed to feed a conventional milling plant. The necessary washing,
screening, sorting and crushing plant and provision for disposing of the sorted
waste, will cost R30 million, but because the remainder of the plant will have
to handle less tonnage it can be made smaller and its cost can be reduced
by R40 million, i. e. the overall cost of the plant will be reduced by Rl 0 million
to R90 million. Of this amount, R40 million would be spent in the first year,
R42 million in the second and R8 million in the third year. Overall working
costs per ton of feed with waste sorting are estimated as R6,00. Which of
these two alternatives will be the more profitable for the company?
To answer this question, using either of the DCF methods, it is advisable
to draw up a table for each alternative (Tables 16.7 and 16.8).
Note, in column 1, that the numbering of the years begins with zero.
This is because the start of the project (time zero) is taken as being when
the first expenditure is made and because no interest charges will be incur-
red during the first year from that time. The labour of calculation can be
greatly reduced by using 'Present Value of 1 per year' tables or by use of
the formula:
Present Value of 1 per year over n years at v per cent
[1 - (1 + vllOO)-n] (16.6)
vllOO
in obtaining the totals at the bottoms of columns 5 to 8 rather than calculating
the quantities in those columns for each individual year. The figure obtain-
ed from tables or by the use of the formula gives the sum of present values
of n units (say Rl) invested singly at v per cent per year at yearly intervals
commencing one year from time zero, i.e. they give:
n
I; 11(1 + vllOOY.
l
Note that this excludes the value of n = 0, which must be included in
the Nett Present Value. By starting the DCF tabulation with year 0, we en-
sure that the value for n = 0 is included. Note further that the tables and
the formula assume equal annual nett cash flows, and also that the figure
obtained from either of these sources must be multiplied by the value of one
of those equal cash flows to give the correct total Present Value. Where,
as is usually the case, the nett cash flows during the first few years are not
equal, the figure obtained by Equation (16.4) or table must be divided by
(1 + vllOO)ct where dis the delay in years before equal annual cash flows are
shown in the tabulation (d = 2 in Table 16. 7); to this corrected figure are then
929
'°0w s-· .....
~ ::i 8:
$ll -· = $ll >-3
~
S::&o 2<D tI1
0(1)§>-li:P..
p (") p.. ~ e:.. e:.. tI1
><:
8~ro£g. >-3
'B~~~;;.J
Table 16.7. IRR and NPV calculations. All quantities in R million.
Alternative I. No waste sorting and run-of-mine milling. '~ s:;;·:...i ....,;:;·
. . . . . _ ...-ll1
~
(')
(1) (2) (3) (4) (5) (6) (7) (8)

"'. § g ~
0 0.. - r.n
.z
1"""1-
>-3
......
<
Discounted Discounted Discounted
>o § Sll ::r
(l)
>o .... ., .... (]) tI1
Discounted
0 .... ::i (]) p..
cash flow cash flow cash flow cash flow :4. 0 00 (1) u:;· ~
Year of Capital After-tax Nett cash at 600/o at 700/o at 68,60/o at 180/o c""'v.:><n tI1
::i ...... >o 0
project payments income flow .per year per year per year per year -· 0
~;:r!llgg
0\ (]) = ~
'" ::i (]) (]) r
g O\p..P..P.. r
0 35,00 -35,000* -35,000 35,000 -35,000 -35,000 rn ...oo -J .-.... n
0\ :--· p:i
c
50,000 -50,000 -31,250 -29,412 -29,656 -42,373 rt-
~
2 15,000 38,559 +23,559 +9,203 +8,152 +8,288 + 16,920 g,~~0' ~ 0
3 110,016 + 110,016 +26,859 +22,393 +22,955 +66,959
•
s:;:::;:: (]) 0
0 ~ :::i -<
4 110,016 + 110,016 + 16,787 + 13,172 + 13,615 + 56,745 O:;:r:;:r!ll:E
~ ·~ 0 ~ en ~
5 110,016 + 110,016 + 10,492 +7,748 +8,075 +48,089 ~ (]):;:: 0
(') rt-...
O'
1--1 0
C:~~~M 0
'°=
~::I
(i)
...... Sll
::r
(])
r
0
30 110,016 110,016 0,000 0,000 0,000 +0,767 0 ., ~ ::i -·
>o::i~o..g.
Nett present value: + 14,578 -1,877 +0,050 +375,440 <D 00' 0 N -·
.... " $ll
~~-·
~ 0 ..... ::i ~
*Because the mine is already in production, the capital cost can, in South Africa, be regarded as part of the mine's working ~:;::~.....,~
~ p:i en
'.._,;,,,,' r.n
costs. This would reduce the tax payable so that in practice the cash outflow would be less than that shown. .... rt- ...... er
oo· ~a (D ~
~ ..... :::i - ::i.
wS:o ?"::l
- i (]) :;:: -.:::J~
~z~· ::r
.j::.. '""O (]) -·
~<~ g.
:::i :::i
(!)
-· "Cl
0
$:>) • "'
...., 0 "'
$:>) :::i (j 0:....,
c:nrt-o:=.:::r-
o..=r~~·<t>
(!) (!) i:: '<
~-< § --~
ri~c:nrn~
P.. i:: vi"
- (!) • ...., ;!;.
Table 16.8. IRR and NPV calculations. All quantities in R million. OQ\°'l'l (!)
N ool'l 2'..'"d
Alternative 2. Waste sorting and conventional milling. :Jl ~g.(!) t;;·
.o ;:,9--l~-
(!) (2) (3) (4) (5) (6) (7) (8) -·. 0
::;: "Cl :::i 00
Discounted Discounted Discounted Discounted ;:r' :Jl ....,. ~
:=.:~
$:>) $:>)
cash flow cash flow cash flow cash flow (!) (!) g ::;:
Year of Capital After-tax Nett cash at 650/o at 700/o at 68,40/o at 180/o (") $:>) (!) '"Ci
i::
project payments income flow per year per year per year per year 0...., -
S"' °' "Cl
~
s ;::.-·
:::i ;:r'"'
Oo $:>)
z
....,
0 40,000 -40,000 -40,000 -40,000 -40,000 -40,000 "'
42,000 -41,ooo -25,455 -24,706 -24,941 -35,593
-· ;:r'
00"' 0
"'(")::;: §~ tl
iii" trJ
2 8,000 33,077 +25,077 +9,211 +8,677 +8,904 + 18,010 ;:r' ~ ;:r' ....,
- -
r;/)
0 (!) 0 ......
3 108,722 -108,722 +24,203 +26,334 +22,912 +66,172 ::;: ' - ' ::;: 0
4 108,722 -108,722 + 14,668 + 13,017 + 13,605 +56,078 "'::;:
- ;:r'P"
$:>)
a"
(D"
z
5 108,722 + 108,722 +8,890 +7,657 +8,079 +47,523 >
- a
;:r' (")
$:>) -· (!)
......
- P":;:o O'
....,
- ::;: :::0
~ :=.:--
........ "' ; :r' (j
30 108,722 + 108,722 0,000 0,000 0,000 +0,058
z'"Ci s O' (!)
0
Nett present value: +5,194 -2,286 +0,013 +373,183 $:>)
< "'"':::ip..
(!)
0_, P"<.<
- a"
i::
-
0
; :r'
....,
(!)
~
~
......
r;/)
~("D r;/)
._ ::::!l
0 z
::;: '"Ci 0
;:r' 0
::;: 0
"'~<s·s z
"'
w
~·
(!) $:>)
'""'"Cl>
(Jq
I
"'
z
;:r'
- 0
(!)
(!)
Table 16.9. IRP and NPV calculations on the difference between alternatives 1 and 2.
(1) (2) (3) (4) (5)

Capital After-tax Nett cash
Year of payments* income flow Discounted cash flow
project Rm Rm Rm at 45,5% per year
0 -5,<lOO +5,000 +5,000

+8,000 -8,000 -5,498
2 +7,000 +5,482 -1,518 -0,717
3 + 1,294 + 1,294 +0,420
4 + 1,294 + 1,294 +0,289
5 + 1,294 +1,294 +0,198
30 + 1,294 + 1,294 0,000
+0,128
*A negative quantity in this column means that the capital payment for alternative 1 is less
than that for alternative 2, which is equivalent to a positive cash flow for alternative 1. And
vice versa.
2 at the company's opportunity cost of money is R373,183 million. That is,

both the IRR and the NPV indicate that flowsheet 1 is a better economic
proposition than flowsheet 2.
Choice of evaluation method

The IRR and the NPV methods will generally assign the same relative ratings
to projects to which they are applied. However, in the case of mutually ex-
clusive possibilities, such as the two flowsheets just considered, the project
with the highest IRR is not necessarily the most profitable, because the capital
involved might be smaller. In other words, by investing a larger amount of
money in a project with a lower rate of return, the ultimate total profit made
by the company might be greater; this would be indicated by a higher NPV
in the second case, and on these grounds, the NPV method is preferable.
In practice, because so little additional effort is involved, it is recommended
that both methods be used. Alternatively, the IRR on the difference between
the two projects can be calculated. If this is greater than the company's op-
portunity cost of money, then the additional capital of the more expensive
project earns a higher return than it could do elsewhere, and the additional
capital expense is justified. Table 16.9 shows how the IRR on the difference
between alternatives 1 and 2 is calculated, and indicates that the return on
the additional capital of alternative 1 is about 45,5%, well above the com-
pany's opportunity cost of money and therefore justified.
Further references on the use of the DCF method of project evaluation
are given in Stermole (1982 and 1984), Wild (1976 and 1977), and Manssen
(1983).
932
Sensitivity analysis
In cases where the DCF analysis indicates very little difference in the NPV
or IRR of the alternatives being considered (as, in fact, in the example given),
a Sensitivity Analysis would be carried out. This is done by calculating the
effect of changes in the various factors on the final NPV or IRR. For in-
stance, the capital expenditure might be rescheduled, or a range of estimated
working costs might be tried for each year. In this way it can be found which
are the most significant factors in the NPV or IRR, and special attention
can then be given to ensuring that the estimates used for them are the best
possible.
Precautions in using the DCF method

During the early stages of plant operation, teething troubles have usually
to be overcome, and results are not as good as in a plant that has been
operating for some time. The use of over-optimistic performance and cost
figures in the cash flow calculations for the early stages must therefore be
guarded against. Also, every effort must be made to ensure that the best possi-
ble estimates of cost and price escalation, and of the gold price, are used
in the calculation of future cash flows.
Another quirk of the DCF method to be wary of arises when cash flows,
after being negative in the project's early years, and then positive for the
bulk of the remaining years, again become negative in the final years, possibly
as the result of capital expenditure on securing residue dams before the com-
pany ceases to be active. In this case, two different IRRs can be found which
will cause the sum of negative and positive DCFs to be zero over the pro-
ject's life. The correct IRR is that value which the IRRs calculated for suc-
cessive years in the vicinity of 75% of the project's life, approach
asymptotically.
16.4.3.6 The conclusion of the process design phase

The selection of the most profitable process route by application of the NPV
method concludes the first stage of the metallurgical design engineer's task.
He must now await a decision by the company concerned as to whether or
not to proceed with the project. If it is decided to proceed, detailed plant
design can commence, in which the first stage is the development of the
General Arrangement of the plant.
16.4.4 General Arrangement

Sufficient information is available in the flowsheets used for process evalua-
tion to enable the laying out of the main sections and equipment of the plant
in their physical positions relative to each other to form an integrated, func-
tional, whole. The arrangement thus arrived at is termed the General Ar-
rangement of the plant. But, before the layout of the plant can be done in
anything other than the most generalized, or conceptual, terms, informa-
tion regarding the site on which it is to be built is essential. This implies that
the site should have been chosen and a detailed survey made, before plant
layout commences. A soil survey to establish the suitability of the surface
933
layers for supporting the plant structures, is also essential.
16.4.4.1 Site selection

The siting of the plant has to be considered in relation to ore supply and
residue disposal, to reagents, stores, materials and energy supplies, to exter-
nal access, to the avoidance of nuisance creation and hazard, to drainage,
to the advantageous use of natural topography, and finally to what is
available.
If the plant is to be supplied from one shaft only, then obviously the
best site is as close as possible to that shaft, with direct transport of ore by
belt from the headgear bin to plant. If, however, the plant is to be supplied
from several sources, a decision has to be made as to whether to site the plant
adjacent to one or other source and deliver ore to it from the remaining
sources by overland transport, or to place it at some central location to whiCh
ore will require transport from all sources. Some of the factors involved in
this decision are the availability of suitable sites, the cost of transport, the
relative amounts of ore to be delivered from the various sources both initial-
ly and in the future and accessibility to waste and residue disposal areas.
Any site chosen must of course be adequately accessible to personnel
and to the supply of the necessary operating stores, water and energy, but
must also be adequately drained and avoid hazards such as sinkholes, cav-
ing and heaving ground, dykes, waste dumps and residue dams. Further, ad-
vantage should be taken of the natural topography to minimiz~ conveyor
belt lifts and pumping requirements, both within and outside the plant. Con-
sideration should be given to the direction of prevailing winds to minimize
nuisance caused by or to the plant.
The siting of waste rock dumps depends partly on the means to be used
for supplying them; they must be adequately accessible by whatever transport
method is used, but in addition should be founded on sound, well-drained,
reasonably flat ground, preferably with a minimum depth of soil to prevent
the formation of a 'ground wave' at the toe of the dump. Adequate clearance
must be allowed around the dump to prevent it from becoming hazardous
to neighbouring structures and objects.
Residue dams should be sited on well-drained ground with a slope of
between 1 and 10 per cent. One essential requirement for a slimes dam is
that water must drain away from the bottoms of the walls and the underly-
ing soil to prevent their shear strengths from becoming too low to resist 'slips'.
From this point of view, the top of a hill is an excellent site for a slimes dam,
but this is seldom practicable and a site sloping in one direction only has
usually to be chosen. However, it is imperative that the dam be kept suffi-
ciently far from waterways and areas where water collects to ensure that its
toes never become water-logged, even in flood conditions. Sufficient area
must be available for the deposition of all residue produced within the fore-
seen life of the mine, although not all of this area need be used initially. Fur-
ther information on areas required is given in Chapter 11, Section 11.2.2.
Slimes dams should be so sited that the necessity for intermediate pum-
ping stations ·between plant and dam is avoided; the supervision of such in-
934
termediate stations is always difficult, and in addition they require extensive

emergency spillage and overflow arrangements, frequently amounting to a
small slimes dam of their own.
16.4.4.2 Elements of good layout

There are certain basic principles to be observed when striving for good plant
layout. These are:
(1) The layout must be clear and logical. Each step of the process should
occupy a clearly-evident area, and these areas should follow each other
in the logical sequence of the process. Not only will this make for simpler
plant control and maintenance, but it will also enhance the plant's
aesthetic appeal, making it a pleasanter working environment.
(2) Transportation requirements must be minimized, whether horizontal
or vertical. This applies to everything that has to be moved to, within,
or away, from the plant, including ore, residue, reagents, stores,
materials, energy, people, and of course, gold bullion.
(3) Ease of operation, supervision and maintenance must be maximized.
(4) Safety and well-being of personnel must be maximized.
(5) Security must be maximized.
(6) Adequate provision must be made for plant expansion.
These requirements are frequently conflicting, so that the final layout
always represents a compromise among them; the best design is the one that
achieves the best compromise.
16.4.4.3 Minimizing transportation requirements

Transportation is a totally profitless activity; unfortunately, it is unavoidable
because the various operations required in the treatment of the ore cannot
be carried out at a single point in space. Obviously it should be kept to the
very minimum, and the very best way to do this is to use the simplest possi-
ble flowsheet. However, some transportation is unavoidable, but some ways
to minimize it are these:
Coarse ore, that is, ore too coarse to be transported as a pulp, is usually
transported within the plant by conveyor belts and chutes (see Chapter 2).
The minimum possible number of belts should be used, which implies that
wherever possible, a single belt should be used between original loading and
final delivery points, with no duplication either in series or parallel. Changes
in elevation should be kept to the absolute minimum commensurate with
headroom, storage and process requirements. Closed circuits, for instance
in crusher plants, should take off from and return to the main stream as
directly as possible, although it is often advantageous to bend the main stream
itself into a right angle or U-shaped configuration. Reducing the number of
belts also reduces the number of transfer points, which are always high
maintenance areas; chute lengths should be kept to a minimum and the max-
imum possible use made of the dead-box principle to minimize maintenance
requirements.
Pulps are transported by launder or pipeline, and as with coarse ore,
935
Figure 16.3. A bulk lime unloading, storage, slaking and

distribution plant.
transportation requirements are minimized by adhering as closely as possi-

ble to straightline layouts and minimizing changes in elevation. Where possi-
ble, of course, it is preferable to use gravity flow rather than pumping, but
because of minimum gradient requirements for gravity flow and the fact that
feed and discharge points of the various processes are at different levels, some
pumping is inevitable. Water pumping is much less arduous than slurry pump-
ing, so that it is logical to lay out the plant to allow slurry to gravitate
downhill and water to be pumped uphill, for instance in the mill-thickening
plant circuit.
Reagents should be stored as close as possible to the point of addition
to the process stream, with easy access to the storage facilities by delivery
transport, and with the minimum possible of manual handling. Hence
reagents should be delivered by bulk transport rather than in small containers
requiring individual handling (Figure 16.3). Where a solid reagent has to be
added at several points to the process stream, it is usually preferable to con-
vert it into a slurry or solution and distribute it by ring-main.
Stores and materials should be delivered as close as possible to the point
of use in the plant. Mechanized means of off-loading and subsequent hand-
ling should be provided. This is facilitated by providing storage areas to serve
those sections of the plant consuming significant quantities of stores. For
example, crusher station and mill should each be provided with a storage
area or areas served by mechanical handling equipment such as gantry cranes
for both off-loading and conveying stores out of storage areas. There should
be easy access to all storage areas by incoming transport, and easy access
936
for the mechanical handling equipment from these areas to the points of
consumption.
A special category of material requiring particular attention is the samples
generated in the course of plant operation. Difficult transport of these (e.g.
the necessity for negotiating long stairways) can contribute very significant-
ly to rendering them valueless by spillage or container breakage. Easy access
should always be provided to sampling points and the containers (usually
buckets or bottles) in which samples are accumulated. Automatic continuous
removal of water from pulp samples in these containers is possible and greatly
reduces the hazard of loss of sample by spillage. It can frequently be arrang-
ed to deliver samples by pipe or chute to convenient collecting points, but
care must be taken to prevent sample holdup on the way. Mechanical aids
to sample transport such as hoists and trolleys all help to increase sample
reliability. Samples should never be transported in open vessels.
Energy is chiefly supplied to metallurgical plants in the form of elec-
tricity, with additional minor amounts in the forms of compressed air and
steam. Where electricity is delivered by high voltage overhead transmission
lines, these lines will have to terminate in a switchyard and substation which
are best located outside the plant area. Power at reduced voltage can then
be delivered from the outside substation into smaller substations within the
plant area by cables. Such cables are frequently located in trenches covered
with concrete slabs, but care must be taken to avoid providing illicit access
to the plant area by these trenches. Better practice is to carry cables in above-
ground cable racks. Electrical reticulation within the plant is normally by
cables which should be carried in easily accessible but unobstructive overhead
cable racks. Great care must be taken to avoid positioning electrical equip-
ment where it can be wetted.
Main trunk service pipelines carrying water, compressed air or steam
should be restricted to ground level piping reserves or to overhead pipe racks.
Main process stream pipelines can frequently share the same facilities. In
all cases, adequate access must be provided for maintenance, and great care
must be taken to avoid obstructing travelling ways with pipe columns or
cables.
16.4.4.4 Maximizing ease of operation
Ease of plant operation can be maximized by the following:
(1) Choosing the simplest possible flowsheet, compatible with efficiency.
For instance, use the minimum possible stages of crushing and milling
(or use run-of-mine milling and eliminate crushing altogether), keep
the number of parallel circuits to a minimum, use a small number of
large units in each step of the process, rather than vice versa. Consider
very carefully whether complicating the flowsheet to achieve a small
additional recovery really is economic.
(2) Providing easy accessibility to checking, sampling and control points.
Try to keep these as far as possible on one level and arrange them on
a continuous route through the plant which·can be easily patrolled by
operators and from which direct, unobstructed sight can be had of each
937
section of the plant (Dillon, 1970). Personnel lifts should be provided

for access to very elevated levels such as cyclone classifier floors and
the tops of leach vessels.
(3) Keeping the plant compact, and keeping the separate stages of the pro-
cess in well-demarcated areas. A sprawling, straggling plant with stages
intruding on each other is very difficult to control. However, the de-
mand for compactness must be balanced against that for accessibility.
(4) Providing adequate surge capacity at strategic points throughout the
process to enable reasonable flexibility in operation. Such surge capacity
permits a degree of independence of operation of the sections of the
plant which can greatly simplify the task of the operator in balancing
throughput rates and making equipment available for maintenance.
(5) Using simple, easy-to-control, robust, well-engineered plant and pro-
cesses. Automatic control should be limited to simple, easily-understood
loops which really do take over part of the operating burden.
(6) Ensuring good lighting and ventilation.
(7) Providing adequate facilities in each area to enable any spillage that
occurs to, be quickly, easily and reliably cleaned up.
16.4.4.5 Maximizing ease of maintenance

Good maintenance is supremely important to high plant efficiency. Design
features which promote good maintenance are:
(1) Easy access for transport to and from the plant, within the plant and
to and around individual pieces of machine and equipment. Ease of
access to pipelines and cables is very important.
(2) The provision of mechanical handling equipment such as gantry cranes,
lifting tackle on crawl beams, service lifts, etc.
(3) Adequate headroom over machinery.
(4) Good lighting and ventilation.
(5) Convenient placement of section workshops and associated storage
areas.
16.4.4.6 Maximizing ease of supervision

Good supervision results from good awareness of what is happening in the
plant and good communication between operating staff and supervisors.
Hence it is facilitated by:
(1) Ease of observation of plant by good placement of supervisors' offices
and good access from them to the plant sections. The laying out of
a continuous route providing good visibility of the plant will promote
both good operation and good supervision.
(2) Siting supervisors' offices adjacent to operators' offices or control
rooms; this enables easy communication with operating staff, and pro-
vides the same degree of accessibility to the plant as is available to the
operating staff.
The optimum siting of senior supervision (e.g. Plant Manager) of-
fices is a difficult question, because the people involved normally re-
938
Figure 16.4. A reduction works guardhouse.
quire easy access to both the plant and to other departments of the mine,
and also need to be reasonably easily reached by officials from such
other departments and also visitors from outside the mine. Thus the
question of plant security arises. Probably the best solution is to site
these offices outside the plant security fence, but as close as possible
to the plant entrance.
16.4.4.7 Maximizing security

Good security can be greatly simplified by suitable plant design. Security can
be enhanced by:
(1) Keeping the plant area plan shape as close to rectangular as possible,
avoiding salients into or out of the main plant area.
(2) Enclosing the plant area in a strong mesh fence at least 3 m high and
topped with a barbed wire barrier. For even better security, a double
fence should be provided with a spacing of at least 10 m between inner
and outer fences, the area between the two being kept completely free
of anything that could provide cover, including long grass and vegeta-
tion. The perimeter fence must be floodlit at night, and nothing which
could provide cover should be positioned within at least 10 m of it,
either inside or outside.
(3) Encircling the plant area with a roadway immediately inside the
perimeter fence. This will ensure that clearance between inner fence
and any permanent cover is maintained. This perimeter road will be
entered at the plant gate and branches leaving it will give access to off-
loading points within the plant itself. Such branches should continue
through the plant and rejoin the peripheral road to eliminate the necessi-
ty for reversing or turning.
(4) Providing one gateway only through the perimeter fence, comprising
a vehicle gate and adjacent pedestrian gate. These should be of heavy
construction to prevent their being charged through by any vehicle likely
to approach them. The gates should be under surveillance from an ad-
jacent guardhouse so arranged that pedestrian traffic has to pass
through it. (Figure 16.4).
939
If other gateways have to be provided (e.g. for rail spurs) they

should be sufficiently close to the main gateway to be under surveillance
from the latter, and should be under direct guard whenever open; they
should not be opened except for immediate passage of traffic and should
be kept very securely locked at all other times. Care must be taken to
avoid providing illicit means for entering and leaving the plant, such
as by conveyor gantries, cable, pipe or conveyor tunnels, or stormwater
drains.
(5) Providing unobstructed visibility and easy access to all plant sections.
(6) Enclosing major security areas within the plant (such as extraction house
and smelthouse), within brick walls at least up to a height of 3 m and
mesh or sheeted walls above that height, depending on whether the walls
are internal or external. Entrances to the smelthouse should be kept
to the minimum. One should be a vehicle entrance to enable bullion
and other high-grade material to be loaded within the smelthouse behind
locked doors. Good surveillance of the interior of the smelthouse
through the mesh portion of the walls and internal doors should be
possible. Change-room, ablution and toilet facilities should be provid-
ed within the smelthouse for smelthouse staff; these should be arrang-
ed in two sections so that personnel can leave 'street' clothing in one
section and pass through the ablution facilities to don working clothes
in the other. Lunchroom facilities should also be provided within the
smelthouse so that personnel can remain within the smelthouse for the
entire working shift.
16.4.4.8 Maximizing safety

One of management's major responsibilities is to provide a safe working en-
vironment. In general many of the principles of good plant layout already
stated in this section, particularly the provision of easy access, unobstructed
travelling ways, and good lighting, visibility and ventilation, will automatically
result in a safe plant layout, and their observance therefore becomes even
more essential. In addition, provision should be made in the plant layout
for at least one first aid station easily accessible by both pedestrian and
vehicular traffic, and possibly for a small fire station as well, even if only
for the storage of fire-fighting equipment. These are usually best located
within the security area in the vicinity of the plant entrance.
16.4.4.9 Provision for expansion

The essence of adequate expansion provision is the allowing of sufficient space
within or adjacent to each plant section, not forgetting electrical substations,
when the original plant layout is made. The reservation of some area out-
side the original plant area for expansion is usually most unsatisfactory as
it forces at least some of the section expansions to be separated from the
original portions of the same sections; probably the best solution in this case
is to build a second plant on the area allowed for expansion.
But provision for expansion does not mean only the provision of space.
It is frequently more profitable to make portions of the original plant equip-
940
ment sufficiently large to handle foreseen expansions than to increase their

capacity later. This applies particularly to single items of plant which can
be made sufficiently large to handle foreseen increased tonnages, for exam-
ple receiving bunkers, stockpiles and main conveyor belts. Application of
the DCF method will enable a choice to be made between whether to include
the larger equipment in the original plant or to provide smaller capacity in-
itially and increase it later.
Waste rock and residue disposal facilities should be so sited that they
can be expanded without rerouting conveyor belts or pipelines supplying them;
such rerouting can necessitate the introduction of additional transfer points
and/ or pumping stations,. either of which is to be avoided.
16.4.4.10 The use of CAD and models

CAD (computer-aided design) can be extremely helpful in developing plant
layouts, but perhaps even more helpful is the use of plant models.
The use of models in laying out plants cannot be too strongly recom-
mended. The amount of money saved by their use through the avoidance
of even minor errors can make the sum expended on them trifling by com-
parison. Furthermore, a model can subsequently serve as a very useful training
aid.
The model used for developing the plant General Arrangement, i.e. the
overall layout of the whole plant, can be very simple, comprising a contoured
plan of the plant area to a scale of about 1: 1OOO attached to a flat baseboard.
The plan should clearly indicate areas to be avoided such as caving ground,
poorly drained areas, dykes, and areas where the soil strength characteristics
are inadequate for structure support. Blocks of wood or other suitable
material representing the main sections of the plant to the same scale as the
plan are then moved about on the baseboard until the most satisfactory layout
has been achieved. The arrangement thus arrived at is then used as the basis
for the General Arrangement drawings. A similar principle can be used in
developing the layouts of increasingly smaller sections of the plant as design
proceeds to increasing detail.
Models are particularly useful for developing chute, piping, and cab-
ling layouts that avoid obstructing travelling ways, and for ensuring that there
is adequate access'to all equipment in the plant (Sproesser, 1980). Figure 16.5
shows the type of model used in plant design.
16.4.5 Detailed design

Mention of all the details that make up a good plant design is obviously not
possible within the space available, but some pointers to good detail design
are given in this section. Good design requires continuous close interaction
between metaliurgist and draftsman.
16.4.5.1 Plant layout and design details

Clearances
Adequate access to plant equipment is vital to good operation and
maintenance. In general, a minimum of 1 OOO mm clearance should be allowed
941
Figure 16.5. A reduction works design model.
between smaller items of equipment such as pumps and between them and
adjacent walls. A minimum of 1 200 mm should be allowed behind centrifugal
pumps to permit withdrawal of the shaft barrels from beneath drive motors.
As the size of the machine increases, so should the space around it increase.
This is particularly so in the case of grinding mills, where modern 5 m diameter
mills should have at least 12 m between them to accommodate walkways and
sufficient floor space for relining activities, i.e. they should be laid out on
a minimum of 17 m centres. Mill spacing, however, is also influenced by
the dimensions of the storage arrangements for the mill feed; if for example
storage silos of greater than 17 m external diameter are used, mill spacing
will have to be greater than 17 m. As a general rule, there should be clearance
all around any machine equal to its own width plus 20%.
Conveyor belts should not be built against walls; a travelling way at least
900 mm wide should be provided on both sides of every belt.
Headroom over machinery should be adequate to permit any components
to be lifted completely clear of obstructions and easily moved horizontally.
The headroom must allow for the distance taken up by the lifting equipment
itself such as the crawl beam, the crawl, lifting tackle, hooks and slings.
Generally, there should be at least 2 300 mm clearance between the bottom
of crawl beams and floor level where such beams have to pass over or along
travelling ways; the same clearance should be allowed over obstruetions.
Access levels
Where the operation and maintenance of a machine requires access at dif-
942
ferent levels, platforms or galleries are usually provided, particularly when

several similar machines are grouped together in one building. Examples of
this type of machine are crushers and mills.
Machines of any appreciable mass have to be mounted on foundations
sunk into the natural ground, so that normally ground level is the lowest
access level. But except in the case of pumps and similar smaller equipment,
operation is usually more convenient at one or more levels above ground.
Crushers require at least one additional level above ground, and for the larger
gyratory and cone crushers it is frequently convenient to provide two
operating levels, the first being that at which crusher adjustment is carried
out and the second that at which the feeders are situated. Likewise, operating
and maintenance convenience of grinding mills and their associated classifiers
normally require several platforms, the first slightly below mill centreline
and completely encircling the mill to provide access for discharge sampling
and also for relining operations, and for large mills a second at a little below
the highest point of the mill but extending backwards from the mill feed end
for easy observation of feed belt discharges and the interiors of feed hop-
pers. Local control panels can also be situated at this level, which is normal-
ly sufficiently high to allow good surveill;mce of the surrounding area as well
as the mill itself. Cyclone classifiers can usually be located so that their spigots
are accessible from this level. But good acccess to cyclone feed piping, valves,
overflow samplers and vortex finders generally requires several additional
platform levels. Where the installation comprises several adjacent units, the
platforms would obviously be extended laterally to provide continuous levels
serving the whole installation.
Frequent convenient stairway interconnections between levels must be
provided, usually one complete set interconnecting all levels as directly as
possible per mill. A personnel and materials lift should also be provided in
a plant of any appreciable size.
The main operating levels for the various sections of the plant should
be as nearly as possible at the same elevation, and should be interconnected
to form a continuous patrol route through the plant. Control rooms should
be on this level or easily accessible to it; alternatively, it is often possible
to combine all control rooms into one from which all the main indoor and
some outdoor installations can be observed.
Thickeners
The essential requirements of good thickener layout are simplicity of feeding
and overflow arrangements, simplicity of draining and good access. In South
African gold metallurgy practice, thickening follows milling and the milled
pulp can be either pumped or gravitated between mill and thickeners. The
final mill product is normally cyclone classifier overflow, so that the pulp
usually leaves the milling process at a sufficient elevation to permit direct
gravity feeding of the thickeners, even allowing for sampling weirs and trash
removal on the way. Direct gravity feeding has the obvious advantage of
avoiding an intermediate pumping installation, which because of the large
volumes involved has itself to be large. However, pump feeding does have
943
I
VERTICAL! SPINDLE
SPILLAGE PUMYS
No. 2 I Thickener
~-~No.) I0,;ok<"<~r ___,.,.__.' IT-,,IH,..-,1
CAGED BALL
I I
NON-RETUIN VALVES
'
PLAN
CUT BACK TO PERMIT

HANDLING OF STANDARD
PIPE LENGTHS
SECTION A-A
Figure 16.6. Conceptual layout of a thickener drainage and

spillage system.
944
several advantages, including greater freedom in siting the thickeners and

the possibility of including guard cyclones or other processing steps before
the thickeners, and should be given due consideration at the design stage.
Thickeners can be laid out in parallel straight lines in which the feed
for each is tapped off a supply launder or pipe extending the entire length
of each line, or in a 'grouped' arrangement fed from a central distributor;
obviously, it is also possible to combine the two systems. The straightline
arrangement will probably require less head between mill cyclones and
thickeners and generally speaking will be simpler than the 'grouped' arrange-
ment from the feeding and overflow collection point of view when more than
four thickeners are involved. However, with modern large thickeners, more
than three or four are seldom required for the whole thickener installation,
and this number can conveniently be accommodated in a single group which
can be fed from a central distributor. Guard cyclones, if employed, can be
accommodated in the distributor support structure.
Thickener underflow removal and drainage systems require careful design
if they are to be satisfactory in operation. As thickeners increase in diameter,
suction columns become impractically long if underflow pumps are placed
outside the perimeter of the thickener tank, particularly if low underflow
w/s is to be maintained. Consequently, modern practice is to place the
underflow pumps, usually centrifugals, below the centre of the tank in the
access tunnel, so that very adequate arrangements have to be made by way
of spillage sumps and self-starting vertical-spindle spillage pumps to prevent
underflow pumps from being flooded in the event of leakage into the tunnel.
Very careful attention must be given to the design of the thickener and
tunnel emergency emptying systems to ensure that they will not themselves
be put out of action in an emergency. Thickener tunnels should be isolated
from each other either by making each a completely separate entity with its
own emptying arrangement or by partitioning it off from any interconnec-
ting trench or sump by a watertight wall. The bottom of such interconnec-
tion should be below the bottoms of the thickener tunnels so that the latter
can only drain outward into it through non-return valves of the caged floating
ball type; drainage columns should pass through the watertight wall if a com-
mon drainage pump is used for all thickeners. The walls of trenches and
drainage pump chambers must extend well above ground level to prevent
flooding by surface water, and important valves should be operable from
ground level through extended spindles if necessary. Figure 16.6 shows a
layout serving a cluster of four thickeners, although the same design prin-
ciples would apply to a straightline layout using an interconnecting trench.
Note that provision is made for the drainage of thickener contents through
the side as well as the bottom of the tank; this is to enable the separate
drawing-off of supernatant water and partially thickened pulp through the
side drain and their disposal within the remainder of the thickener plant,
while the thickened pulp can be transferred forward to the leach plant.
Thickener tunnels must be sufficiently wide to accommodate all the
necessary pipe columns on one wall and leave enough width for the conven-
ient movement of the largest pieces of equipment, probably the underflow
945
pumps, that will have to pass through the tunnel. For the same reason, the
tunnels should be equipped with crawl beams spaced at least 2 300 mm from
the tunnel floor. The edge of the tunnel roof at the entrance to the tunnel
should be 'chamfered' back to permit the handling of standard length pip-
ing into the tunnel. Underflow pump suction connections and thickener
drainage columns should both be equipped with two cut-off valves in series
and should have permanent high-pressure water connections for easy clear-
ing of chokes.
Thickener bridges, like tunnels, must be sufficiently wide to permit easy
movement of machinery components and should also be equipped with crawl
beams.
Thickener overflow collection systems tend to reflect the feed systems;
that is, when the thickeners are laid out on the straightline system, the
overflow collection system generally comprises a main collection launder for
each line, while in the case of grouped thickeners, the overflow channels
generally converge as a fan directly into the thickener overflow collecting
tank or 'return water' tank.
In the straightline layout, it is possible to arrange for thickener feed,
overflow and underflow systems all to be carried at different levels on the
same structure over the thickener tunnel interconnecting trench, but care
should be taken that no launder can overflow into another beneath it. Using
this system, it is sometimes possible to arrange for underflow transfer pumps
and return water pumps to be housed in the same building.
The best material for thickener feed and overflow and underflow col-
lection systems seems to be concrete piping with compression ring type flanges
and inspection slots running the length of each section to permit easy descal-
ing of the pipe interior.
The use of one or more thickeners to serve as return water collectors
can be strongly recommended; continuous operation of the rakes and
underflow pumps is necessary to avoid damage to the rakes, but totally
eliminates the necessity for the very labour-intensive and unpleasant manual
cleaning of return water collectors.
Leach plant
The current tendency is towards the use of propellor-agitated leach vessels
of depth: diameter ratio of 2 or more, and of depths up to 22 m. To avoid
excessive short-circuiting of pulp in continuous leaching, at least six vessels
should be used in series, and preferably eight (see Chapter 6). Thus modern
leach plants are arranged in parallel strings of eight or more tanks, and the
layout must provide sufficient head between tanks to ensure that peak flow
requirements can be met, and also allow the by-passing of one or two tanks
in each string when necessary.
To reduce the physical distance between feed and discharge points of
each string and thus the difference in elevation between these points, it is
usual to locate alternate tanks on two parallel centrelines, with pulp passing
back and forth between the two half-sets as it flows towards final discharge.
The tanks can be arranged on either a square or a triangular grid, as shown
946
Inflow Inflow
Outflow Outflow
SQUARE TRIANGULAR
LAYOUT LAYOUT
Figure 16.7. Two types of leach vessel layout.
in Figure 16. 7. The triangular arrangement is recommended as being more

versatile. Normally, corresponding tanks in the two half-strings are located
at the same elevation, but succeeding pairs are located at successively lower
elevations. Alternately, the tanks can all be sited at the same elevation;
although this results in a progressive diminution of effective volume of the
successive tanks, it does have the advantage of preventing overflows as the
result of power failures.
Open launders and simple drop-in cut-off gates are preferable to pipes
and valves for tank interconnections from the point of view of ease of clear-
ing chokes, but obviously they must be sufficiently deep to accommodate
the maximum change in pulp level likely to occur in operation. Interconnec-
tions must be kept as short as possible, and where relative movement bet-
ween tanks is expected, should incorporate suitable means to accommodate
947
this. A by-pass column running the whole length of the installation and con-
nected to each tank by a valve is sometimes provided, but is very expensive
in valves, and is unnecessary if sufficient intertank connections are made as
shown in Figure 16.7.
Leach vessels should be contained in a bunded area graded at 2 Yi OJo
towards a sump connected to drainage pumps outside the bund. Joints in
bund concrete should be sealed with polyurethane before machinery installa-
tion. Emergency draining of tanks is done directly into the bund which should
be capable of containing at least two leach vessel volumes. Provision should
be made for reaching and operating drain valves even when the bund is full.
Non-emergency emptying of tanks can be through a drainage pipe system
delivering directly to the above-mentioned drainage sump although the large
change in suction head can cause pump operation difficulties, and drainage
via the bund might be preferable. Leach vessels should be provided with
manholes on hangers through which the tank can be entered from ground
level when empty, for inspection and clearing.
In the case of mechanically-agitated vessels, some means must be pro-
vided for raising and lowering heavy components such as impellers, shafts,
gearboxes and motors between ground level and tank top. In the past this
provision has taken the form of a crawl beam and crawl on each tank but
recent large installations have incorporated tower cranes for this purpose
(Anonymous, 1985).
Carbon-in-pulp plants
The carbon circuit in a carbon-in-pulp plant is a closed loop with the carbon
moving countercurrent through the pulp in part of the loop, then being
separated, eluted, regenerated and finally returned to the tail absorption
vessel. Carbon-in-pulp absorption vessels can thus be laid out on a straight
line with the elution and regeneration plants somewhere alongside or at either
end of the absorption line, or the latter can be bent into a U with the carbon
treatment section at the top of the U. Several U-shaped strings of absorp-
tion vessels, starting and finishing at the central carbon treatment building,
can radiate from this building in which all pulp transfer pumps and ancillary
equipment, as well as the carbon treatment equipment itself, can be housed.
A control room can be situated at the top of the central building with a good
view over all the absorption tanks, so that a very compact and easily super-
vised plant results, as seen in Figure 16.8.
The disadvantage of the U layout is that it is less flexible as regards ex-
pansion than the straightline arrangement, in which space can be provided
at the ends or alongside the original lines for further tanks.
As with leach vessels, there must be sufficient fall between succeeding
tanks to cater for peak flow rates and also the by-passing of tanks as re-
quired, but absorption vessels normally require additional head room to ac-
commodate the intertank screening arrangements. Adequate lifting and
transport arrangements must be provided over the tanks for handling agitator
mechanisms, shafts and propellors, and the interstage screening equipment
whether static or mechanical. The trend in large plants is towards tower cranes
948
Figure 16.8. A carbon-in-pulp plant using the U-type layout.
for this purpose.

Carbon-in-pulp absorption vessels should be located in a bund with ade-
quate spillage and drainage facilities. Bund floors should slope at 2% or more
towards spillage pump intake sumps, the pumps being located outside the
bund. Joints in bund concrete must be sealed with polyurethane. Bearing
in mind the relatively high concentration of gold in loaded carbon, and the
easy transportability of the carbon, the carbon-in-pulp plant should be regard-
ed as a high security area and be enclosed within its own security fence with
guarded entrance and self-contained facilities so that movement of person-
nel in and out of the security area is minimized. The plant for recovery of
gold from the.carbon-in-pulp eluates, whether by zinc duct precipitation or
electrolysis, should be located adjacent to the smelthouse, and the eluates
piped to it, rather than incorporating the recovery section in the carbon-in-
pulp plant and transporting very high grade solids from it to the smelthouse.
Filter plants
With the advent of carbon-in-pulp technology, it seems unlikely that any fur-
ther large filter plants will be built for recovery of gold bearing solution from
leached pulp. However, it is possible that filter plants might still be built
for recovering water, cyanide and lime (and possibly even some gold) from
carbon-in-pulp tailings, and the following pointers are offered with this
possibility in mind.
Disc and drum filters are normally mounted on concrete foundations
sunk into the natural ground level. However, the foundations should be suf-
ficiently high to bring the machines up to a convenient distance above an
949
operating floor which will be the first level above ground. Drum filter pans
are frequently of concrete cast integral with the foundations. Siting the filters
above ground allows for the installation of drainage sumps, conveyor belts
or repulpers, and ancillary equipment such as filtrate and vacuum pumps,
at ground level, preferably where they can be observed from the operating
level, leaving the latter uncluttered. If a conveyor ,belt is installed for the
removal of discharged filter cake, it should be possible to walk along its en-
tire length between filter foundations and belt stringers in order to reach con-
veyor idlers for servicing; similarly, any other cake handling equipment should
be easily accessible for maintenance. Any weightometer installed on the belt
must be mounted well clear of the nearest filter, which generally means siting
the repulper well away from the filters and locating the weightometer in the
gap between filters and repulper. The repulper tank should be of concrete
rather than steel, and it should discharge by overflowing a weir, never by
direct connection to a pump, except for drainage purposes. Weir overflow
also provides a sampling point, and should be designed as such. The section
of belt between the last filter and the repulper should be built over a narrow
extension of the main filter drainage sump which reaches to the repulper so
that the latter can overflow into it in emergency. The belt must be well washed
and the washings should gravitate directly into the repulper.
If wash sprays or wash distribution troughs are fitted to filters, they
should be visible from the valves controlling them to facilitate wash adjust-
ment. Access must be provided to the sprays or distributors for cleaning pur-
poses, and the valves, together with all other valves and controls, must be
conveniently located on the operating floor.
Idlers on conveyor belts carrying filter discharge cake should have can-
tilevered closed-end rollers to provide increased protection against entry of
pulp into bearings. The rollers should preferably be rubber covered. Crawl
beams and lifting equipment must be provided over all filters.
Belt filters do not require any drainage or emergency emptying ar-
rangements, but do tend to produce large quantities of low-density spillage
resulting mainly from leakage of sealing water. They must therefore be sited
over impervious floors graded towards spillage pumps. Foundations for belt
filters are normally very much less massive than for drum and disc filters,
comprising essentially footings for the steel framework. The necessary
headroom for the repulper and the return belt rollers has therefore to be pro-
vided by making the framework columns sufficiently long. This headroom
requirement is usually several metres on all except the smallest filters. Con-
sequently an operating floor above ground level and surrounding each filter
is usually required. As with drum and disc filters, ancillary equipment should
be placed on ground level, but well clear of the filters themselves because
of the very wet conditions pertaining. However, to minimize duct lengths,
belt levitation fans are best sited on the operating floor adjacent to the filter.
The main belt on a large belt filter is an extremely massive item, and
good provision must be made for handling it and also the filter cloth itself,
during renewal operations. This generally implies easy access to the feed end
of the filter, good mechanical handling aids and plenty of space between
9:'i0
filters. The changing of rollers and shafts is also facilitated by adequate inter-
filter spacing; the length of these rollers approximates the width of a filter,
so that the spacing between filters should be filter width plus 20%. Lifting
services in belt filter bays are best supplied by a single gantry crane covering
the entire area.
Smelthouses
The main criteria in smelthouse design .are:
(1) minimization of manual handling,
(2) provision of a pleasant environment, particularly with regard to heat,
fumes and lighting,
(3) high security.
The chief source of feed to the smelthouse is normally a precipitation
stage, either zinc dust or electrolytic, while a frequent secondary source is a
gravity concentration stage located within the milling circuit. The smelthouse
should be so situated that the products from both these sources, which are
very high grade, can be delivered into it without manual handling.
Generally the smelthouse personnel are responsible for carrying out the
'clean up' or removal of accumulated gold precipitates from the vessels in
which the precipitation takes place and its transfer to the smelthouse. This,
together with the fact that in many plants the precipitate is the only feed
to the smelthouse, has resulted in the smelthouse generally being located in
the cyanide plant area rather than in the mill area.
It is usually possible to site the precipitation units immediately adjacent
to the smelthouse and at a sufficient elevation to permit gravitating the
precipitate through pipes directly into receiving vessels in the smelthouse.
In fact, electrolytic precipitation plant can be sited within the smelthouse
itself, but visible and controllable from outside by shift operating staff. Plac-
ing the smelthouse within the cyanide plant almost automatically ensures that
concentrates from the mill will be delivered by pumping, so that both
precipitated gold and concentrate will reach the smelthouse without manual
handling.
The smelthouse equipment should be so laid out that the gold bearing
materials being processed follow as logical and direct a sequence as possible
through it, with as little manual handling as possible. Slurries should be
pumped or gravitated between stages, and cakes and other products which
cannot be pumped should be handled by chutes or in containers travelling
on roller tracks or wheels or suspended from overhead tracks. In larger plants,
such as those generally encountered on the Witwatersrand and Free State
goldfields, slag handling is facilitated by the fact that smelting is now in-
variably done in electric submerged arc or induction furnaces, pouring of
the melt being done through a cascaded train of moulds which overflow from
one into the next so that the slag leaves the cascade in a continuous stream
while pouring is in progress. This arrangement is very well suited to con-
tinuous granulation of the slag by allowing it to fall into water flowing down
a launder into a pump sump below floor level; from this sump the granulated
slag and water are pumped continuously to a conical bottom holding tank
fitted with an overflow launder and sited within the smelthouse itself; the
settled slag, after draining off the water, can be run out into bags for
transport.
Pleasant working conditions within the smelthouse are ensured by mak-
ing it as spacious as possible so that personnel are not obliged to remain con-
tinuously in the immediate vicinity of hot furnaces, and by providing good
ventilation and lighting. The main task of the ventilation system will be the
efficient removal of dust and fumes from furnaces, and this will usually re-
quire the provision of collecting hoods over furnaces and other sources of
dust and fume, and exhausting these to atmosphere through filters and fans.
A smelthouse can be a very cold place in winter when no furnace is operating,
and some means of heating it at those times should be provided.
The smelthouse should be illuminated during daylight by glazed south-
facing sides of sawtooth roof sections; windows in walls should be avoided.
Artificial lighting at night must provide sufficient illumination to enable the
entire interior of the smelthouse to be easily visible from external observa-
tion points.
Security is enhanced by constructing all smelthouse walls, at least up
to 3 m above ground level, of reinforced concrete or brick, by providing as
few entrances as possible and equipping these with double-locked stout steel
external doors and steel grille internal doors. One external door should be
sufficiently large to admit road transpott so that this can be loaded and off-
loaded within the completely locked smelthouse. The entire interior of the
smelthouse should be observable through suitably guarded openings from
the adjacent building. These observation openings should preferably be
located on the continuous patrol route of the plant shift operators. Change-
house, toilet and eating facilities should be provided within the smelthouse
so that movement in and out of the smelthouse can be minimized. The
strongroom, of which walls, floor and roof should be of heavily reinforced
concrete at least 300 mm thick, should have a single top-grade door equip-
ped with treble combination locks and 7-day time lock. If any of the walls
of the strongroom forms part of the outer wall of the smelthouse itself, care
must be taken that the exterior of such wall is kept completely unobstructed
and under frequent surveillance, and above all that it should not be a com-
mon wall with some relatively unfrequented and closed-off area such as a
substation or storeroom from which an attempt could be made to gain ac-
cess to the strongroom. Smelthouse acid treatment vat vent ducting should
be of timber, and definitely not of plastic on which static electric charges
can be generated by the passage of gas.
Spiilage handling
This is a very important aspect of plant design, but is frequently overlook-
ed. A clean plant is essential to efficiency and cannot be achieved if spillage
handling is not designed for. Care must be taken to ensure that spillage can
be dealt with in all areas of the plant by hosing-down of adequately sloped
floors to spillage launders and/ or pump sumps and directed to the appropriate
re-entry points in the circuit. This is especially true in reagent make-up areas.
Drainage
In terms of Government Notice R287 of 20 February 1976, paragraph 10,
the discharge of any spillage into any evaporation dam, stormwater drain
or other waterway on the mine property is illegal. Therefore provision must
be made in the plant layout to keep plant spillage or overflows of any descrip-
tion, including pulp, water, solutions and reagents, from entering the storm-
water drainage system. This usually calls for the siting of tanks within bund-
ed areas, and the grading of floors of machinery areas so that all spillage
is directed inwards into internal spillage systems. Overflows of water, par-
ticularly those from the mill water system which can be of considerable
volume, are usually best dealt with by the provision of an overflow dam sited
at the lowest point of the plant area, and from which the water can be reclaim-
ed for re-use in the plant.
16.4.5.2 Construction details

Buildings
Walls - The preferred materials for walls at least up to first floor level are
brick, cement block, or reinforced concrete for high-security areas. Above
this level, corrugated sheeting of galvanized steel has been usual in the past
although the recently introduced 3CR12 steel is probably more cost-effective
over its service life despite the fact that it should be painted. 3CR12 is however
not strongly resistant to chloride ion attack, and in such conditions corrugated
plaster sheeting should be used.
Roofs - Possibly the most cost-effective material is corrugated or IBR
sheeting in 3CR12 steel, even though this will usually require painting exter-
nally for appearance's sake. In areas of high chloride ion concentration, rein-
forced plastic is preferable, although the possibility of hail damage must be
kept in mind.
Windows - Areas not requiring high lighting levels can usually be adequately
illuminated in daytin1e by translucent fibre-reinforced plastic corrugated sheet
panels in walls and roofs. Where high lighting levels are required, glazing
the steeper flanks of sawtooth roof sections will usually be adequate for top
floors or single-floored build,ings, but vertical strip fibre-glass wall panels
will be required on lower floors. Where damage by hail is possible, wire rein-
forced glazing should be used, and ease of cleaning and repair must be pro-
vided for.
Main structural members - In the past, structural steelwork has normally
been of carbon steel. However, a selective range of structural sections in
3CR12 chromium-bearing corrosion resistant steel is now available, and
although this material is considerably higher in first cost than carbon struc-
tural steel, its 'life cycle cost', i. e. the cost per unit of service life taking into
account the much reduced maintenance costs, could well show it to be com-
petitive with the traditional material; it should be considered for any new
plant construction.
Floors - Main building floors and operating levels should be of reinforced
9J<'.'lJ
concrete. They should preferably not be level, but should be graded at one
per cent to drainage points. Under and around pulp handling equipment,
floors should be graded 6% towards spillage channels. Auxiliary floors and
interconnecting galleries less used than main operating floors can be of rein-
forced concrete if wet, with egg-crate-covered openings if necessary to per-
mit surveillance of equipment on floors below. If dry, auxiliary floors can
be of steel egg-crate, but these should be raised about 30 mm above suppor-
ting steelwork by angle iron welded to the latter to prevent the formation
of dirt-collecting pockets.
Stairway treads - These are best made of steel egg-crate 300 mm wide, and
200 mm rise per tread.
Handrails - Stanchions should be solid castings and the rails solid rod, not
piping. Top .rails should be 1 OOO mm above floor level with a second rail
about 500 mm above the floor. Steel kicking plates 150 mm high should be
provided under handrails.
Sumps - The essential function of a sump is to provide storage capacity
to absorb differences between inflow and outflow rates. In the case of a sump
drained by a centrifugal pump, the changes in fluid level resulting from dif-
ferences between inflow and outflow result in some degree of automatic level
regulation by adjusting the total head on the pump and hence its pumping
rate. The primary problem in sump design is therefore so to configure and
size the sump that changes in fluid level, in conjunction with the pump head
capacity characteristics, can bring about the necessary changes in pumping
capacity without allowing the sump either to overflow or to run empty.
One solution to the primary problem is to make the sump very large
in relation to the inflow so that even a very big difference between inflow
and outflow can be at least temporarily accommodated. This, of course, is
the situation in the case of drainage sumps and bunds, but in the case of
sumps for continuously-operating pumps it is not a practicable solution, firstly
because it is uneconomic of space and materials and secondly because, in
the case of pulp handling sumps, settlement and possible subsequent break-
ing away of the solids cause major operational problems. A secondary pro-
blem is therefore to design the sump with as little dead volume and oppor-
tunity for solids settlement as possible.
The general procedure for sump design is therefore to size the sump to
handle the maximum surges anticipated (and this can in some cases be
minimized by interlocks between pumps drawing from sumps and equipment
delivering to the sump) and then to decide on an effective depth which will
allow, in conjunction with the pump characteristic, some measure of
automatic pulp level control. To avoid solids settlement, sumps should be
restricted in horizontal cross-section area, and their bases should be graded
towards pump suctions at not less than 40° to the horizontal. The suction
column should continue at this slope through the suction valve and then sweep
to the horizontal immediately before meeting the pump. All corners, both
vertical and horizontal, of sumps handling suspensions should be well rounded
to minimize solids lockup. Pulp pump suction columns should have high
pressure water connections on both sides of the suction valve, and all sue-
tion columns should incorporate at least one angled flange to facilitate valve
removal or disconnection of the pump.
The bases of sumps must be sufficiently elevated above pump intakes
to ensure a net positive suction head on the pump under all flow conditions
to prevent pump cavitation.
The preferred materials for sumps handling fluids only are steel and con-
crete, if necessary protected against corrosion by suitable linings. For pulp
sumps also, either steel or concrete can be used, but steel almost always re-
quires rubber lining internally and frequent painting externally, and from
the viewpoint of minimum maintenance, concrete is preferable. Even con-
crete, however, must be protected against erosion by some form of lining
if pulp is allowed to impinge directly on it, and concrete sumps are of course
very difficult to modify once built. Suction columns connected to concrete
sumps should not be bolted to sump walls by studs; they should be a snug
fit inside a flanged stainless steel nozzle set in the concrete and pass through
this nozzle to the interior of the sump, being retained by a flange bolted to
the nozzle flange. All sumps should have drain valves connected to the lowest
point.
Launders - Materials used for launders are usually carbon steel plate, car-
bon steel pipe or wood plank. When necessary, these materials can be pro-
tected from abrasion by 5 mm of 50 shore natural rubber. However, all the
construction materials mentioned require painting for external protection,
and concrete.piping has been found to be a very satisfactory low maintenance
alternative which does not require painting although it needs more support
structure than the traditional materials.
Concrete piping for launders should be joined with compression type
flanges rather than concrete collars, and should be of the slotted variety for
ease of inspection and descaling. Any closed type of launder should have
frequent inspection openings, at least of the same cross-section dimensions
as the launder itself.
As regards launder shape, a flat-bottom configuration is obviously easiest
to construct from flat materials such as plate or planks. However, the Chezy
formula
Ve = C(Rh.S)y, (16.7)
where Ve = pulp velocity,
Rh = hydraulic radius ( = stream cross-section area/wetted perimeter)
S = the launder slope
C = a constant for a given Rh,
shows that pulp velocity is increased both by increasing hydraulic radius and
by increasing launder slope. C varies almost linearly with Rh, so that Ve
varies nearly as Rh312 • Therefore, for maximum economy of materials, the
optimum launder shape is that having the greatest Rh with the smallest other
dimensions. This condition is met by a semi-circular channel running com-
pletely full. However, some freeboard or means to prevent overflow must
be provided, and this can conveniently be done by using a pipe running half
full at maximum flow rate.
An excellent launder design method is given in a paper by Green, Lamb

and Taylor (1978), although it must be pointed out that the method is aimed
at deriving the necessary launder dimensions and slope to maintain a pulp
velocity sufficient to prevent solids settlement; it does not deal with the in-
itial period when the pulp is accelerating from virtually zero to design velocity.
Steeper slopes than those given by this method are necessary to prevent solid
settlement during the acceleration phase.
Piping
Water, solution and pulp pipes are normally made of carbon steel with wall
thicknesses selected to accommodate the required maximum pressure. High
density polyethylene piping is, however, being increasingly used for fluids
and fine pulps where the danger of damage resulting from external events
is small, e.g. in buried pipelines. Columns carrying coarse, abrasive solids
will require protection by rubber lining, or they can be made of armoured
rubber hosing if sufficient support and adequate pressure rating can be pro-
vided. Pipe lengths should be joined with flanges, and not welds, to enable
easy assembly and dismantling. Pipelines subject to scale build-up, for ex-
ample those carrying mill circuit water and residue pulp, should incorporate
means for easy descaling. Quite extensive pipe systems, such as the mill water
reticulation, can be descaled by taking them off-line, and filling them with
a 3 OJo inhibited hydrochloric acid solution, taking care to provide safe escape
routes for the evolved gas. For long columns such as residue pipelines, descal-
ing is freqµently done by means of go-devils or hedgehogs. These are wooden
spheres carrying a large number of partly inserted screws on their surfaces,
or they can be made of serrated steel semi-circles.welded together along their
straight edges to simulate a sphere. Go-devils or hedgehogs are inserted into
the empty pipeline through tee branches which can be closed by means of
blank flanges or valves. Alternatively, if the branch is equipped with two
valves with sufficient distance between them, the go-devil can be inserted
into the column with the latter onstream. The fluid flow forces the go-devil
along the column while its surface projections cut away scale accumulations.
Pipeline design - In the case of pulp pipelines, the design objectives are
firstly, given the required volumetric flow rate and solids size distribution,
to calculate that pipe diameter which will result in the necessary minimum
pulp velocity to prevent solids settlement. The second objective is then to
determine the pressure loss per unit pipe length, in order to determine the
necessary pipe slope for gravity flow or the necessary pump delivery pressure
when designing pump delivery columns. Obviously, in the case of pipelines
handling fluids only, the first objective does not apply.
A procedure for pipeline design is given in Appendix 16.1 to this chapter.
Pump selection
The problem in pump selection is first to calculate the total dynamic head
on the pump at the necessary velocity to prevent solids settlement (as deter-
mined in Appendix 16.1), and then from makers' catalogues, to select a pump
956
capable of delivering the required flow against this head. The pump speed
and drive power requirements will then also be obtainable from the maker's
data. A method for pump selection is given in Appendix 16.2 to this chapter.
16.4.5.3 Piping and Instrumentation Diagrams

Concurrent with detailed plant layout drafting, Piping and Instrumentation
Di'agrams (P & IDs) are prepared. These show each and every piece of equip-
ment which is connected to any other piece by piping or launder, and all
interconnecting, inc·oming and outgoing pipelines and launders throughout
the plant; that is, duplicated items of equipment and interconnecting flow
are all shown. This contrasts with the flowsheet, where duplicated equip-
ment and flows are represented by one symbol only. All instrumentation is
also shown as well as its connections to the plant equipment, using a stan-
dard representation. Each pipeline and launder is then allocated an identify-
ing code and number, and is labelled on every P & ID sheet on which it ap-
pears; this identification enables easy tracing of the pipeline through any
number of sheets. The P & IDs thus become very useful aids to detailed
design; in fact, in the design of any except the simplest plant, they are
indispensable.
16.4.5.4 The Definitive and Revised Cost Estimates

When process design, P & IDs and detailed plant layout drafting have been
completed, it will be possible to call for tenders for the supply of all but minor
equipment and for the performance of all the significant construction work.
This will make available cost estimates for these two major components of
the overall project, and so enable preparation of the Definitive Cost Estimate,
that is, an estimate accurate to within ± lOOJo. The subsequent completion
of all detailed design work will enable refinement of the Definitive Estimate
to within ± 5 OJo to produce the Revised Estimate.
16.5 Plant Construction and Commissioning

16.5.1 Metallurgical involvement in the construction phase
It is highly desirable that the official who will be in charge of plant opera-
tion, and, if possible, his second-in-command, should be involved in the
design, construction and commissioning process at as early a stage as possi-
ble, preferably as part of the metallurgical component of the Project Team.
This will ensure their complete familiarity with the design background and
operating philosophy of the plant. But as the design process gives way to
the construction phase, the plant manager designate and his assistant should
move onto site, or at least make frequent site inspections. They should have
regular contact, possibly at daily meetings, with the construction supervisor
to report and obtain correction of any design and construction errors they
observe and generally to act as process advisors to the construction team to
ensure that the plant will be capable of being safely and efficiently operated
when startup commences.
957
16.5.2 Preparation for commissioning

Concurrently with the construction phase, or even earlier if possible, the
manager designate will have to devote much of his time to preparing for plant
startup, for this will be uniquely his responsibility. In conjunction with the
design and construction personnel he will have to prepare the startup
timetable; in conjunction with maintenance and buying departments he will
have to ensure adequate supplies of stores, reagents and spares at startup
time; and above all, he will have to ensure adequate manning of the plant
by trained personnel. As part of the training process, operating manuals will
be required, and the manager designate will be involved in the preparation
of these and also in the arranging of training programmes, possibly in ex-
isting plants, for the plant personnel. The drawing up of log sheets and ar-
ranging for proper record keeping and reporting of plant operations will also
be his responsibility. The manager designate will also be required to establish
a procedure for takeover of sections of plant from the construction person-
nel as they are completed; this is essentially a safety precaution to ensure
that there is no misunderstanding as regards who, at any particular time,
is in charge of any particular section of the plant and consequently is respon-
sible for safety in that section. A very comprehensive exposition of the duties
of the plant manager in relation to startup has been given by Fulks (1982).
Fulks deals with American chemical plant practice and uses the term 'plant
engineer' rather than 'plant manager', but his recommendations are of direct
relevance to South African gold plant construction.
16.5.3 Commissioning
16.5.3.1 Personnel
Commissioning is best carried out by a specially-assembled commissioning
group under the plant manager and comprising metallurgists, engineers with
artisan backup to carry out minor alterations and trouble-shooting ex-
peditiously, and experienced operating personnel under a plant foreman. No
attempt should be made to start up a new plant with inexperienced personnel.
16.5.3.2 Cold commissioning

The construction crews normally hand over each section of the plant when
it has been 'mechanically completed', i.e. when the buildings, structure and
equipment are all in place and connected up mechanically and electrically
according to the drawings and the equipment has been correctly lubricated,
checked for operation in the sense that it will start and run when required
to do so, and motors will turn in the correct direction. The commissioning
team should not accept any portion of the plant from the construction crews
until it is satisfied that the section has been completed at least to the stage
where commissioning can commence with a reasonable hope of success and
above all that it is safe. Having taken over a plant section, the commission-
ing team then is responsible for that section and can proceed to the first stage
of its task, namely 'cold commissioning'. Cold commissioning means run-
ning the section without process material in it. For example, in commission-
ing a mill circuit, the mills, feed belts, etc. would be run empty at normal
958
operating speeds, but the mill water reticulation services would be complete-
ly functional (which implies that the thickeners and return water system would
already be operational and filled with water) and the normal flows of dilu-
tion water would be passing through the mill-classifier circuit, pumps, pipes,
launders, etc. Cold commissioning is the stage at which unsuspected con-
struction debris in pipes and elsewhere is discovered, leaking pipe joints are
corrected, conveyor belts are trained, automatic controls can be roughly set,
overheating machinery detected and rectified and so on.
In short, it is the stage in which the plant section is brought to the state
where it appears to be capable of handling the process stream reasonably
efficiently, safely and continuously, but without actually having handled nor-
mal process material. The plant manager must keep in direct contact with
the commissioning process and should have at least a daily meeting with the
construction supervisor to report progress, and by means of 'punch lists' re-
quest rectification of construction faults which cannot be dealt with by the
engineering personnel attached to the commissioning team. At these meetings
also, any necessary temporary handing back of sections for construction fault
rectification can be arranged.
16.5.3.3 Hot commissioning

After all obvious faults which would prevent the safe and reasonably effi-
cient handling of the process stream have been eliminated, 'hot commission-
ing' can commence. This is the crucial stage at which the actual process
material begins to pass through the plant and at which it becomes evident
whether or not the effort of the preceding months and years is to be crown-
ed with success. Normally, feed rates are kept low at the commencement of
hot commissioning and gradually built up as faults are eliminated and the
plant settles down to normal operation. The importance of careful record
keeping should not be overlooked, even at this stage. Inevitably, design faults
will come to light during hot commissioning, but if the design work has been
careful and guided by experience, the faults should be minor and relatively
easy to correct. However, if, unhappily, major faults become evident, they
must be faced and dealt with boldly to reduce to the minimum the loss of
production and profit while the fault is corrected. Quick decision-making
by the plant manager and his advisors is necessary, but decisions must be
tempered by due recognition of the fact that at this stage the plant will pro-
bably not have attained designed conditions at all points; alterations must
not be rushed into which would be made unnecessary by correct functioning
of the process at some other point. In other words, treat causes, not
symptoms.
The plant manager will obviously be heavily involved in hot commis-
sioning but must maintain his close liaison with the construction supervisor
and must also keep his fellow heads of department and company manage-
ment informed. Should major alterations become necessary, then of course
further work will be required by the design section of the construction
organization, and the plant manager will have to make an input to this.
959
16.5.3.4 Some practical commissioning tips

( 1) If possible, commissioning should be carried out on waste rock to reduce
the value of lockup and loss due to incorrect processing.
(2) A void having ore, reagents, etc. in storage for extended periods before
plant startup. The properties of these materials can be adversely af-
fected during storage so that eventually startup has to be commenced
with material for which the plant was not designed. Also fines can set
hard and become extremely difficult to move after extended storage.
(3) Only partly fill storage facilities such as stockpiles, bins and tanks before
startup. Stockpiles, in particular, segregate badly as they are filled, so
that unless drawoff occurs reasonably concurrently with filling a large
core of fines can form which can seriously affect plant operation and
require a long time to eliminate. Furthermore, if storage has to be emp-
tied for fault correction, obviously the less material to be handled the
better.
(4) Crushers should be set somewhat coarser than designed to begin with
and gradually 'pulled up' to correct setting to avoid choking and damage
if they are not able to handle actual operating conditions.
(5) Commence commissioning on manual control and gradually introduce
automatic control as operation settles down.
(6) Run-of-mine mills should initially be fed dry (i.e. without discharge)
at the highest rate at which rock can be got into them. This is in order
to build up a pebble load as quickly as possible and to avoid pipeline
blockage with coarse discharge. When the power draft reaches a max-
imum and commences to decline, the feed rate should be reduced to
hold the power at maximum and dilution water opened. Steel grinding
media should not be added until a satisfactory pebble load, both as
regards quantity and size distribution, has been built up. In particular,
avoid adding steel if the initial feed is fine, as the steel will simply retard
the buildup of a pebble load. It is general experience that large run-of-
mine mills require as much as six months before they achieve efficient
operation.
(7) Have a range of sizes of cyclone spigots and vortex finders available
to enable quick changes for rapid optimization. This applies particularly
to spigots, whose size is more critical than that of vortex finders. Start-
up vortex finders need not be rubberized as they will probably be chang-
ed before wearing out.
(8) Thickeners should be filled to overflowing with water before startup
otherwise the incremental water lockup before they overflow can ex-
ceed the drawdown of the return water tank and the mill water system
can run empty.
(9) Thickeners should not be circulated during startup. Because of the
higher settling rate of the coarser particles, circulation can cause the
concentration of sand in the settled pulp which in turn can cause rake
overload and tripout. It is better to keep the underflow pumps com-
pletely stopped with occasional short spells of running (without cir-
culating) to avoid underflow system blockage, until underflow
960
water: solids has diminished sufficiently to permit continuous drawoff.

(10) Remember to fill tanks and sumps which would normally contain recir-
culated solutions required in the process, with a suitable temporary
substitute to enable the process to get started. Normally clean water
is satisfactory.
16.5.3.5 Acceptance runs
Where the design and/ or construction of the plant have been carried out by
some organization other than the owners, it is usual to include in the con-
tract some form of' acceptance run'. During this, inputs and operating con-
ditions of the plant are held as close as possible to those specified in the Pro-
cess Design Criteria (see Section 16.4.3.1), and a determination is made as
to whether the plant is then able to attain the specified operating and output
targets. Obviously, time has to be allowed for overcoming possible initial
faults and shortcomings, and a mutually agreed commissioning period may
be specified in the contract. The contract will state the penalties to which
the design/ construction organization becomes liable if the plant does not meet
the Design Criteria within a specified time, and possibly also contain 'incen-
tive clauses' which lay down rewards for beating specified requirements,
usually as regards date of acceptance. A very important point in drawing
up the acceptance clauses is that the acceptance criteria should be capable
of being measured and that they be very carefully specified and understood
by both parties to the contract. For example, it is useless to specify what
the characteristics of a certain process stream shall be, when in practice it
is impossible to determine them, at any rate to the necessary degree of ac-
curacy. Also, differences of interpretation can result in conflict situations
between the parties, and great efforts should be made to avoid them by careful
and thorough statement of the acceptance criteria.
16.6 Conclusion
This chapter has given a necessarily very condensed scheme for the systematic
design of a South African gold extraction plant, together with design infor-
mation and procedures and some pointers to the actual commissioning pro-
cess. If the design metallurgist conscientiously follows the route laid down,
the probability of achieving the objectives mentioned at the commencement
of the chapter, namely a plant brought into production on time, at designed
capacity and efficiency, and within budget, will be very high, and the achieve-
ment will be one of which the metallurgist may be justly proud.
16. 7 References
Anonymous (1985). Cranes cut costs at Ergo. South African Mining World, 4(10), 26-29.
Barnea, E. (1977). New plot enhances value of batch thickening tests. Chemical Engineering,
84(18), 75-78.
Dahlstrom, D.A. and Purchas, D.B. (1957). Scale-up methods for continuous filtration equip-
ment. Institute of Chemical Engineers, London.
Dillon, V.A. (1970). Special features of the Kinross Mines, Ltd. reduction plant. In Mineral
96i
Processing and Extractive Metallurgy, edit. M.J. Jones, pp. 485-508. I.M.M., London.
Fulks, B.D. (1982). Planning and organizing for less-troublesome plant start-ups. Chemical
Enginering, 89(18), 96-106.
Green, H.R., Lamb, D.M. and Taylor, A.D. (1978). A new launder design procedure. Mining
Engineering, (A.I.M.E.), 30, 1210-1216.
Lamos, A.W. (1982). Liberation of gold grains from ores by chemical fusion. Research report
12/83, Chamber of Mines of South Africa Research Organization, Johannesburg.
Manssen, L.P. (1983). Financial evaluation of mining projects: is common practice enough?
Mining Engineering, (A.l.M.E.), 35, 601 - 606.
Moncrieff, A.G. (1963). Theory of thickener design based on batch sedimentation tests. Trans.
I.M.M., 1963 -1964, 73, 729 759.
Osborne, D.G. (1975). Scale-up of rotary vacuum filter capacities. Trans. l.M.M., 83,
C158- C166.
Pownall, J.H. (1962). Grindabilit,y test procedure and ball-mill size selection. Mineral Process-
ing Information Note No. 3. Warren Spring Laboratory. Department of Scientific and In-
dustrial Res~rch, London.
Ruhmer, W. T., Svoboda, 0. and Wilson, B.M. (1984). South African costs of equipment for
the metallurgical industry, 1984. Report M212, Council for Mineral Technology, Randburg,
South Africa.
Sproesser, W.D. (1980). Models speed plant design. Chemical Engineering, 87(6), 113-116.
Stermole, F.J. (1982). Economic Evaluation and Investment Decision Methods, 4th edn. In-
vestment Evaluation Corporation.
Stermole, F.J. (1984). Personal skills: economic analysis of mineral and energy investments.
Mining Engineering (A.l.M.E.), 36, pp. 143 - 146.
Wild, N.H. (1976). Return on investment made easy. Chemical Engineering, 83(8) 153-154.
Wild, N.H. (1977). Program of discounted cash flow return on investment. Chemical Engineering,
84(10), 137 -142.
Wildhelm, J.H. and Naida, Y. (1979). Sizing and operating continuous thickeners. A.I.M.E.
Society of Mining Engineers, preprint 79- 30.
962
APPENDIX 16.1
Pipeline Design
The following procedure for pipeline design has been supplied by P.R. Bailey,
Consulting Metallurgist, Gencor Ltd, to whom grateful acknowledgement
is made:
(1) Determine solids and pulp relative densities (rds and rdP), and solids
size distribution down to 30 µm.
(2) Calculate required pulp volumetric flowrate Q, in litres per second.
(3) Calculate the solids concentration in the pulp by mass, using
C = rd/rd 2 - 1) (16.8)
m rdp(rds - 1)
(4) Determine the transport medium relative density, rd,m. For quartzitic
ore of rds = 2,7 it is assumed that all particles smaller than 30 µm are
in permanent suspension and may therefore be regarded as part of the
transport medium. The corresponding size for solids of relative densi-
ty other than 2, 7 is given by
27
Suspension size ds ( µm) = d' x 30. (16.9)
r s
rd/mass fraction < ds>(rdP - 1) + (rds - rdp)

Thenrd = ~~~~~~~~~~~~~~~~~- (16.10)
tm (mass fraction< ds>(rd - 1) + (rds - rdp)
(5) Calculate solids concentration by volume in pulp, Cv. Cv is normally
given by
CV= d (16.11)
r s
but for the purposes of step 7, the following formula which corrects
for rd,m is used:
cm (16.12)
cv(corrected) = rd s - cmrs-rtm
(d d )
(6) Determine d50 , the size in micrometres passing 50% of the solids, from
the cumulative size distribution.
(7) Determine Durand's F factor from Figure 16.9, using Cv(correctectl and
dso-
(8) Estimate a suitable pipe diameter D in millimetres.
(9) Calculate the limiting pipeline velocity (i.e. minimum velocity to pre-
vent solids settlement) using
V = 0,03162 F (2gD rds - rdtm)v'm/s (16.13)

rdtm
(g = gravitational acceleration, m/ s2).
963
10 20 30 40 60 80 100 200 300 400 600 800 1000 2000

t-------'---'-----'---L--'----+-------'----'---'-----L--'----+----~
Cv = solids concentration by volume :

1,3+------
1,2
1,0
~
~ 0,9
i:i
.9
:g
8"'
'.§, 0,8
"'
0,7
0,6
0,5
0,4
JO 20 30 40 60 80 100 200 300 400 600 800 1000 2000

d 50 = size passing 50% of sample (µm)
Figure 16.9. Durand's 'F' factor at various pulp densities as

a function of particle size distribution (Durand and Condolios,
1952).
964
Table 16.10. Hazen-Williams friction constants C.
Material c
Mild steel, unlined, new 130
Mild steel, unlined, old 100
Cast iron, unlined, new 130
Cast iron, unlined, old 70
Asbestos 150
Concrete, spun 150
Copper 150
Lead 140
Brass 140
Galvanized steel, new 140
Galvanized steel, old 100
Polypropylene 160
Mild steel, rubber lined, new 150
Mild steel, rubber lined, old 110
(10) The actual pipeline velocity is calculated from

_ 1273,24 Q I
V act- 2 ms (16.14)
D
(11) Compare Vact with V. If Vact is smaller than V, steps (8), (9) and (10)
must be repeated with other values of D until a maximum value is found
which will ensure that vact is greater than v for all operating condi-
tions.
(12) Calculate the hydraulic gradient for water (S,J (i.e. the pressure drop
per unit distance due to friction), using the Hazen-Williams formula
= (279203,7 Q)l,85185
SW c x D2,63 (16.15)
The friction constant C for various pipe materials is given in Table

16.10.
(13) Correct the friction loss for water as obtained in step (12) to that for
pulp, Sm. If Vact is less than V (a situation to be avoided), the flow is
likely to be heterogeneous, and the following correction due to Newitt
as given by Madigan (1972) should be applied:
S
m
= S
w
(l + 1100 X Cv(correctect> x g x D x Vs x (rds - 1))
( vact)
3 m/m
(16.16)
where Vs = 'average' terminal settling velocity of the solids.
Vs can be found from Stokes' Law:
2
V = (d50 ) X g X (rds - rdtm)
s 9 ~~ (16.17)
18 µ x 10
965
.
Particle d50 size (µm)
60 100 300
0,100 20 30 40 50 80 150
0,080
0,060
0,040
0,030
,...._
~ 0,020
~
·~0
~
·;; 0,010
5 0,008
8
§: 0,006
<!'.
0,004
0,003
0,002
0,001-l""=----~--~---~----,----,.-----+-
o 10 20 30 40 50 60
Percentage of solids by volume
Figure 16.10. Apparent pulp viscosity as a function of solids

concentration for various particle size distributions (Thomas,
1965).
where µis the apparent viscosity in pascal-seconds (Pa.s) obtained from

Figure 16.10.
If vact is equal to or greater than v, the flow is _likely to be
homogeneous, and the alternative Newitt correction should be applied:
Sm = SW (1 + cv(corrected)(rd,-1)). (16.18)
The maximum pipe diameter and the pressure loss per unit of pipe length
(the hydraulic gradient) have now been determined. If the diameter falls
between standard pipe diameters, a decision will have to be made as
to whether to use the next larger or next smaller standard pipe diameter.
The hydraulic gradient should be recalculated for the diameter chosen.
966
References
Durand, R. and Condolios, E. (1952). The hydraulic transport of coal and-solid inaterials in
pipes. Proc. Colloquium on Hydraulic Transport of Coal, National Coal Board of Great Britain,
Paper 4.
Madigan, C. (1972). Slurry pumping. Amdel Bulletin 14. Australian Minerals Development
Laboratory, Adelaide, South Australia.
Thomas, D.G. (1965). Transport characteristics of suspensions: VIII. A note on the viscosity
of Newtonian suspensions of uniform spherical particles. Jn!. Colloid Science, Vol. 20, pp.
267-277.
APPENDIX 16.2
Pump Selection
(1) Determine from drawings or models, the total pipeline length, i.e. suc-
tion plus delivery length.
(2) Determine the numbers and types of fittings that will be incorporated
in the pipeline, and by reference to Table 16.11, calculate the equivalent
additional pipe length due to these fittings. Then add this equivalent
length to the pipeline length determined in step (1) to give total
equivalent length (metres).
(3) Calculate friction head by:
Friction head= total equivalent length x Sm, metres of water, where
Sm is the friction loss per metre of pipe as determined in Appendix
16.1.
(4) Determine the static head on the pump. Static head is the difference
in elevation (in metres) between the highest point reached by the delivery
column and the surface of the pulp in the pump sump. (Do not multi-
ply this difference by the pulp relative density.)
(5) Determine the delivery head, which is the pressure (in metres of water)
required at the end of the delivery column. If the column is open-ended,
delivery head is zero. If the delivery column supplies a cyclone, delivery
head is the feed pressure of the cyclone.
(6) Calculate total dynamic head on pump:
Total dynamic head = friction head+ static head+ delivery head
(metres of water). (16.19)
(7) Use the pump manufacturer's head ( = total dynamic head) versus quan-
tity ( = pulp flow rate, Q, II sec) curves to select pump and pump speed
capable of giving the required flow rate against the calculated total head.
Interpolating between maker's curves for various speeds can be made
by using the Law of Similitude:
2
Head2 = ( rpm2 ) (16.20)
Head 1 rpm 1
967
....,
"'
0\
00
Table 16.11. Table of pipe lengths of equivalent friction loss to various pipe fittings.
::r::
tI1
Nominal pipe size, mm 25 50 75 100 150 200 250 300 350 400 450 500 550 600 tI1
><
....,
Equivalent length of strai~ht pipe in metres :;::l
Standard bend 0,8 1,6 2,4 3,2 4,8 6,4 8,0 9,6 11,2 12,8 14,4 16,0 17,6 19,2 >
("')
Medium sweep bend
Long sweep bend
Close return bend
0,6
0,3
1,9
1,3
0,5
3,7
2,0
0,8
5,5
2,6
1,1
7,5
4,0
1,6
11,0
5,2
2,2
15,0
6,5
2,8
18,5
8,0
3,5
22,0
9,0
4,0
26,0
10,5
4,5
30,0
12,0
5,0
34,0
13,0
5,5
38,0
14,5
6,0
42,0
16,0
6,5
46,0
-
....,
<
tI1
Square elbow 1,6 3,5 5,0 7,5 10,0 13,3 16,7 20,0 24,0 28,0 31,0 34,0 37,0 40,0 ~
45° elbow 0,4 0,7 I ,I 1,5 2,2 3,0 3,8 4,5 5,6 6,0 7,0 7,5 8,0 9,0 tI1
....,
Ordinary entrance 0,4 0,9 1,3 1,8 2,7 3,5 4,5 5,0 6,0 7,0 8,0 9,0 10,0 11,0
Borda entrance 0,7 1,4 2,2 3,0 4,5 6,0 7,5 9,0 10,5 12,0 13,5 15,0 16,5 18,0
>
l'
l'
Sudden [ d/D = Y. 0,4 0,7 1,1 1,5 2,2 3,0 3,8 4,5 5,6 6,0 7,0 7,5 8,0 9,0 c:;::l
contraction d/D = Yi 0,3 0,5 0,8 1,1 1,6 2,2 2,8 3,5 4,0 4,5 5,0 6,0 6,0 6,5
d/D = % 0,2 0,3 0,5 0,7 1,0 1,4 1,6 2,0 2,4 2,7 3,0 3,5 3,8 4,1 0
~
Sudden [ d/D = Y. 0,8 1,6 2,4 3,2 4,8 6,4 8,0 9,6 11,2 12,8 14,4 16,0 17,6 19,2 0
enlargement d/D = Yi 0,5 1,0 1,5 2,0 3,0 4,0 5,0 6,0 7,0 8,0 9,0 10,0 11,0 12,0 '"I1
d/D = % 0,5 0,9 1,4 1,9 2,9 3,8 4,6 5,5 6,5 7,5 8,5 9,5 10,3 10,8 0
0
Through standard tee 0,5 1,0 1,5 2,0 3,0 4,0 5,0 6,0 7,0 8,0 9,0 10,0 11,0 12,0 l'
tJ
Standard tee split 1,6 3,5 5,0 7,5 10,0 13,3 16,7 20,0 24,0 28,0 31,0 34,0 37,0 40,0
Standard tee side outlet 1,6 3,5 5,0 7,5 10,0 13,3 16,7 20,0 24,0 28,0 31,0 34,0 37,0 40,0
Angle valve open 4,0 8,0 12,0 16,0 25,0 33,0 40,0 50,0 58,0 65,0 73,0 80,0 90,0 100,0
Gate valve open 0,2 0,3 0,5 0,7 1,0 1,4 1,6 2,0 2,4 2,7 3,0 3,5 3,8 4,1
Y. closed 0,9 1,8 2,8 3,8 6,0 7,5 10,0 12,0 14,0 16,0 18,0 20,0 22,0 24,0
Yi closed 5,0 10,0 14,0 20,0 30,0 40,0 50,0 55,0 70,0 80,0 85,0 100,0 120,0 150,0
% closed 21,0 40,0 60,0 85,0 120,0 170,0 210,0 250,0 290,0 330,0 350,0 400,0 450,0 500,0
Globe valve open 8,0 16,0 24,0 32,0 51,0 70,0 85,0 100,0 115,0 130,0 145,0 160,0 180,0 200,0
Saunders KB valve 0,6 1,1 1,8 2,7 3,1 3,6 5,1 6,3
(From "Planning Centrifugal Pumping Plants" by Sulzer S.A. Ltd)

Table 16.12. Maximum pump impeller tip speeds for some typical materials.
Maximum tip speed

Material (m/s)
Moulded natural rubber (pulp pumping) 25

Neoprene, nitrite and chlorinated butyl
rubbers (pulp pumping) 28
Hard iron (pulp pumping) 28
Hard iron (water pumping) 40
(8) Check that the impeller tip speed of the selected pump (i. e. the tangen-
tial velocity at the greatest diameter of the impeller) is within the limit
for the impeller material. Some suggested maximum tip speeds are given
in Table 16.12.
(9) Determine the clear water pump power requirements from maker's flow
rate versus power curves for the pump and speed selected. Interpola-
tions can be made by using the relationship:
Power oo (speed) 3
(10) Multiply the clear water power by pulp relative density (rdµ) to give
the power that will be drawn by the pump.
969
Chapter 17
Loss Control and Safety Management

F.K. Vice and D.J. Nel
17 .1 Introduction
Loss control is defined as the intentional management action directed at the
prevention or reduction of loss resulting from exposure to the perils associated
with the pure risks of business. To achieve a high level of loss control, manage-
ment must identify all areas where such losses can occur and manage the
control of these losses. This can best be achieved by the implementation of
a formal loss control programme.
The introduction of loss control or safety programmes as an integral
part of a complete management system was started in the South African min-
ing industry in 1975. Today, the programmes are well entrenched and have
been successfully implemented on most of the gold mines in the industry.
For example, the programme introduced in the metallurgical division of one
of the largest gold mining complexes in the world has achieved encouraging
improvements not only in casualty rates but in physical condition standards
and plant availability and utilization.
The programmes consist of a number of activities which direct manage-
ment's efforts at critical areas with past loss experience or which have the
potential for loss. Emphasis is placed on the activities being predictive and
preventive rather than reactive. Many of the activities are interrelated and
supportive of each other. For example, meaningful accident analysis is depen-
dent upon effective accident investigations, and subjects for communication
to employees emanate from accident investigation and analysis. Personal pro-
tective equipment standards are established from historic accident data, from
planned task observations or the critical task inventory.
Managing the programme is a function of operating management, and
the greatest potential for control exists at the point where the action takes
place. Only line management, from the supervisor to the most senior manager,
can ensure a successful loss control programme. Loss control decisions can-
not be separated from operational decisions and are thus the responsibility
of line management.
17.2 Loss Control Department

In order to provide the required motivation, expertise and education of line
management in the principles and accepted practices of loss control and to
provide for the vital auditing function, the formation and staffing of a loss
control department is necessary. For maximum effectiveness staff should be
971
>-l
~k=:==t
'D
::r::
d
-.)
N
sol I trJ
trJ
><
>-l
~(')
....,
>-l
<
trJ
Physical condition
rating percentages
601
I I I I ~
trJ
>-l
501- I I I I >
t:""'
t:""'
16 c
?::I
40 14 a
>-<
12 0
Disabling injury
'I1
30V I I I II I I I I II I r---~ I f-10 frequency rate
0
0
8 t:""'
ti
6
0
1981 1982 1983 1984 1985
Figure 17 .1 . Metallurgical Division: physical conditions and disabling injury trends.

LOSS CONTROL AND SAFETY MANAGEMENT
drawn from experienced metallurgical personnel. Loss control officers are

administratively and operationally responsible to senior metallurgical per-
sonnel and technically to the loss control co-ordinator.
17 ,3 Physical Condition Rating Programme

The physical conditions in working areas are of fundamental importance in
loss control. The Physical Condition Rating Programme is an effective system
to enable management to upgrade physical conditions in working areas and
to provide an unbiased measure of compliance with predetermined standards
regarding physical conditions. It is a relatively inexpensive programme and
it soon develops employees' pride in their working environments. Loss con-
trol staff normally form the nucleus of condition rating teams, and other
team members include senior and middle managers from the operating and
engineering disciplines. Benefits accruing from such a programme include
inter alia:
Identification of substandard conditions.
Elimination of accidents and fire causes.
Reduction of wasted energy.
Maximization of space usage.
Minimization of stores and equipment inventories.
Guaranteeing of good working place appearance.
Encouragement of better work habits.
An illustration of the improvement in physical conditions in a metallurgical
plant over a five-year period is reflected in Figure 17 .1. Note the concomit-
ant decrease in the lost time injury rate.
A detailed checklist which directs the attention of rating teams or in-
dividual auditors to specific areas of known loss or to those areas which
possess the potential for loss should be used during the inspections. The
significant aspects covered by the checklist are:
Premises and surroundings.
Electrical, mechanical and personal safeguarding.
Fire protection.
Loss control programme knowledge.
First aid equipment.
The same comprehensive checklist is used for both physical condition ratings
and general planned inspections; the items listed and requirements stated are
extracts from metallurgical standards, managt:;rial instructions or Mines and
Works regulations (Appendix 17.1).
To upgrade the level of programme implementation and to encourage
additional participation by all levels of supervision and management as weli
as operating staff, expansion into other activities should be considered such
as formal planned inspections, group meetings, planned task observations
and objective accident/incident investigations.
973
17 .4 The Manager's Policy

A loss control or safety programme without top management support will
fail. To affirm management's full commitment to and indeed involvement
in a safety programme, the most senior manager should issue a statement
clearly specifying his safety philosophy and policy. The statement would in-
clude the reduction or elimination of incidents involving personal injury,
damage to property, environmental health, fires, and unplanned interruption
of operations. It must be clear to employees that all levels of supervision
have a high degree of commitment to and involvement in the loss control
programme. A programme steering committee under the chairmanship of
a senior manager should be formed to develop the policy and set objectives
to guide the programme through the various phases of implementation.
17 .5 Training
Formal training of all levels of supervisors in the principles of safety and
loss control is necessary. Practical and theoretical training is required for
line personnel to enable them fully to understand and appreciate their respons-
ibilities in respect of the programme. These must not be seen as additional
to their management or production responsibilities but rather to complement
and assist in their occupational duties. This training can be done either by
in-house courses or by using outside organizations. It is necessary for addit-
ional and special training to be provided for loss control staff.
17 .6 Planned Inspections
The planned inspection is one of the most widely used of the loss control
activities available to management. It is an excellent tool for detecting
potential incidents before loss of any kind occurs. It highlights the emphasis
of the programme on prediction rather than reaction. Inspections can be
classified into two main categories, namely general and critical parts.
17 .6.1 The general inspection

This inspection is similar in many ways to the physical condition rating and
is normally conducted by line personnel in their own areas of responsibility.
It provides a formal system of inspection and follow-up, or overinspection,
by a more senior line official to ensure a consistent identification of substan-
dard conditions. The use of a checklist to assist the person conducting the
inspection in knowing both "what to look for" and "what to look at" is
highly recommended (Appendix 17.1).
All areas and structures in the plant should be inspected at pre-
determined intervals depending upon the criticality of the area and the nature
of the work being performed therein. Inspection of plant areas and workshops
should be conducted monthly by the metallurgical foreman and engineering
foreman. Substandard conditions should be noted during the inspection and
recorded, prioritized and finally remedial action responsibilities should be
specified. To ensure that inspections have been judiciously conducted, over-
974
inspections should be done in the same areas by the plant superintendent

or engineer on a random selection basis as soon as possible after the initial
inspection. An example of an inspection record sheet is shown in Figure 17 .2.
17 .6.2 Critical parts inspections
These inspections provide a formal system of inspection, service and overhaul
of plant, machinery and equipment which will restrict downtime to an ac-
ceptable level. The inspections are normally conducted by engineering per-
sonnel and controlled by the Planned Maintenance Department.
17.6.2.1 Pre-use inspection of equipment

These inspections ensure detection of faults in power-driven equipment, which
if not detected could result in loss. They are in fact an extension of the Critical
Parts Inspection and are usually conducted by the equipment operators using
a check list containing the critical items. As with the general inspection it
is imperative that overinspections be conducted in both these areas on a
random basis by senior plant personnel.
Regular analysis of substandard conditions and other deficiencies will
provide clear indicators to plant managers of repetitive problems which could
require action by them to bring about change in design or operational
procedures.
17. 7 Accident/Incident Investigation and Analysis

Investigation of accidents is the responsibility of line management, and an
effective safety management programme must include an efficient and for-
mal system of accident and incident investigation. The reason for investigating
accidents is basically to prevent a recurrence of the event. This can only be
achieved by the clear identification of the sequence of events that resulted
in the accident.
It is essential for the manager or supervisor to under.stand what he is
trying to control with accident/incident investigation, and to provide some
clarity the following definitions are provided. "An incident is an undesired
event that could or does result in loss" and" an accident is an undesired event
that results in physical harm to a person or damage to property. It is usually
the result of contact with a source of energy above the threshold limit·of
the body or structure." The source of energy could be kinetic, electrical,
chemical or thermal.
Investigations should not be restricted only to events which have resulted
in loss due to some source of contact but shquld include other events which
could have caused losses or interruptions to operations or processes. Examples
of these are: shortage of strategic materials; unusually high incidence of wear
and tear on equipment components; near-loss or no-loss incidents that under
slightly different circumstances could have resulted in loss.
It is essential that incidents are reported to supervisors and managers
with the minimum of delay so that prompt action can be taken and investigat-
ions initiated before conditions alter or important evidence is destroyed.
975
\0 -J
-.J
::r::
°' tTJ
Figure 17.2. Sample record sheet for planned inspections. ~

-J
METALLURGICAL DIVISION ~
(')
NAME: -J
PLANNED INSPECTIONS: INITIAL AND OVERINSPECTIONS (Plant Foreman)
~
Estimated Actual ~
Hazard Job Immediate remedial Permanent remedial completion completion Inspectors tTJ
Date Area Sub-standard condition class Req. no. action taken action to be taken date date sign ~
t""'
t""'
c:;;c
0
>-<:
~
0
0
t""'
Note: Hazard potentials: A Extensive or permanent disability Immediate action required (hazard to be rectified or area to be made safe) t:l
B Disruptive, serious injury Schedule to be rectified as soon as possible (within one week)
c Minor or no-loss Schedule for a later date
17. 7 .1 The investigation

Investigations must include a visit to the accident scene by the investigators
and the interviewing of casualties and witnesses to obtain as much insight
as possible into the conditions and attitudes which prevailed at the time of
the accident.
Accident causation is somewhat complex and could require broad con-
siderations of personnel, equipment, materials, the environment and in par-
ticular management's decisions which impact on design, installation and
maintenance, operational practices and training. Only with relatively simple
incidents is it generally feasible to expect line supervisors to be proficient
in the causal determination and the recommendation of appropriate remedial
actions. For obvious reasons, however, their active participation in the in-
vestigation of incidents which occur in their areas of responsibility is im-
perative. The loss severity potential and frequency of the incident will deter-
mine the level of management involvement. Loss control officers in their
capacity as safety professionals should participate in the investigation to pro-
vide guidance, advice and to encourage probing into the less obvious areas
which could have had some influence on the incidents. To obtain meaningful
and unbiased information investigations should be conducted on the basis
of fact finding rather than fault finding and blame fixing.
The recording of facts by the investigators must be accurate and com-
plete. Identification of the immediate and basic causes should be comprehen-
sive and remedial actions and those responsible for the implementation of
these actions must be clearly specified. The immediate causes (or symptoms)
of the accident include both substandard acts and substandard conditions.
These result from deviations from accepted standards or practices and
originate from the basic or underlying causes and the lack of management
control.
A sample investigation form is shown in Figure 17 .3. The form is design-
ed to lead the investigators to recommend permanent remedial actions. When
completed, it should contain all the necessary information for meaningful
accident analysis.
It is imperative to ensure that the specified remedial action is taken and
it is in,eumbent upon the loss control staff to provide this follow-up service.
The accident/incident investigation is a highly significant component
of an effective safety programme. Safety professionals agree that all pro-
gramme activities are to some degree complemented or downgraded by the
quality of these investigations and the action taken to prevent the recurrence
of the accident/incident. To assist investigators in identifying the immediate
causes of substandard acts and conditions, and more importantly the basic
causes of personal and job factors, the classification reference guide shown
at the bottom of Figure 17 .3 is provided, usually on the reverse side of the
investigation form.
17. 7 .2 Analysis
17. 7 .2.1 Statistical data
It is normal practice for managers to obtain performance statistics on
977
\0
-..J Figure 17.3. Sample record sheet for accident/incident investigation. ...,
00
METALLURGICAL DIVISION
@
tii
ACCIDENT/INCIDENT INVESTIGATION REPORT NO. ...,><
~
I. BASIC DATA (Describe where applicable)
I
3. IMMEDIATE CAUSES: (what acts, failure to act
and/ or conditions contributed most directly to the
...,n
Department . . . . . . . ... Date of Occurrence ~I
DayofWeek = SunOl Sat07
Plant ................... Time (01-24 hours)
incident)
~
~
Section tii
Name and Coy/P.F. No. of injured:
s;
l'
l'
Occupation
4. BASIC CAUSES (reasons for the existence of these
c
Age (years)
?:I
acts/ conditions) 0
Ethnic Group
--<
0
'"rj
Occupation and Name o·f Person in Most Control:
0
0
l'
tJ
5. TEMPORARY REMEDIAL ACTION
Probable Recurrence Rate: Freq. Occ. Rare
I
Action Comp.
Loss Severity Potential: Major Serious Minor By Date
Part of Body Injured: COSTS
Property Damaged: MAT.

LAB.
Task Performed:
PROD.
Extent of Damage or Breakdown: .... 6. PERMANENT REMEDIAL
DAYS
LOST ACTION
Agency: [TOT.
Nature of Injury:
Witness Names/Numbers .... 7. PLANT SUPT/DIV ENG. COMMENTS, ADDI-

TIONAL INSTRUCTIONS
2. DESCRIPTION: INCLUDE SKETCH IF APPLIC-

ABLE (use back of page)
r'
PLANT ENG.
0
\./)
\./)
8. NAMES/NO'S OF INVESTIGATION TEAM (1

0
z--l
:;a
0
r'
~
z
d
\./)
LOSS CONTROL OFFICER
~
trJ
~
1. Basic causes g) Inadequate emergency preparedness (including first aid training and
a) Insufficient or inadequate inspections (planned inspections, critical equipment). ~
~
parts inspections, pre-use inspections, four-or five-step plan). h) Inadequate procedures, standards and rules (including signs and
b) Inadequate follow-up on inspections. colour-coding).
c) Inadequate maintenance.
d) Substandard accident/incident investigation or remedial action.
i) Inadequate communication and/or enforcement of procedures, ~
standards and rules. $';
\0 e) Inadequate incident reporting programme. j) Insufficient or inadequate planned task observations. trJ
-....!
\0 t) Inadequate accident/ incident analysis. k) Inadequate skill training. z--l
(Continued overleaf.)
\0 I) Inadequate training in hazard recognition. d) Failure to warn. >-3
00
0 m) Inadequate job induction. e) Failure to secure. ::i::
trJ
n) Inadequate standards for enforcement or provision of protective 1) Operating at improper speed.
trJ
equipment. g) Making safety devices inoperable. ><:
o) Inadequate standards or performance in giving of instructions (PTI h) Using equipment improperly (misuse/abuse). >-3
;:o
and safety tips and hints). i) Failure to use protective equipment. ;l>
p) Inadequate follow-up on instructions. j) Improper loading or placement. (')
q) Inadequate purchasing standards.
r) Inadequate procurement/ delivery of material or equipment.
s) Inadequate mining or engineering design and/or methods.
k) Taking improper position.
I) Alcohol or drugs.
m) Unauthorized entry.
-tii
>-3
t) Improper placement {physical capability, heat tolerance, acclimatiza- n) Failure to recognize hazard. ~
tion, eye-sight, hearing, aptitude). o) Other. trJ
u) Inadequate propaganda programme - group meetings and/ or
Conditions
~
L'
promotions.
a) Inadequate guards. L'
v) Other. c:::
b) Defective tools, equipment or material. ;:o
Suggested list - personal factors: c) Inadequate tools, equipment or material. 0
a) Inadequate capability - physical, mental, aptitude. d) Inadequate warning system. ~
b) Stress - physical, mental. e) Hazardous atmospheric conditions (gases, dusts, fumes, vapours).
c) Lack of knowledge. 1) Other fire and explosive hazards. Sil
d) Lack of skill. g) Substandard housekeeping. 0
e) Language/ communication problem. h) Inadequate ventilation. 0
L'
1) Improper motivation - negative, etc. i) Inadequate illumination. tJ
g) Other.
2. Immediate ca111ses 3. If any cause is selected from the above list, the reason therefor must
Acts be given.
a) Operating/working without authorization.
b) Working in dangerous area/situation. 4. Remember: problems are seldom, if ever, the result of a single cause.
c) Working without stopping equipment. Look for multiple causes.
occupational injuries to employees on a daily, weekly, monthly and annual

basis to facilitate the monitoring of trends and to initiate corrective action
in problem sections or areas. These statistics enable comparisons to be made
between various sections and supervisors on an equitable basis in ratios or
rates. The standard formulae for the determination of the more commonly
used rates are:
Frequency rate
The disabling injury frequency rate relates the injuries to the hours worked
during the period and expresses them in terms of a million-hour unit by use
of the following formula:
Number of disabling injuries x 106
Employee-hours of exposure
A disabling injury is defined as an injury arising out of and in the course
of employment which prevents the person from performing his or her own
job or a similar job for one or more shifts.
Reportable disabling injury rate

The reportable injury rate refers to the injuries and deaths which are report-
able in any month to the Chief Inspector of Mines under Mines and Works
Regulation 25.1. These are related to the number of persons at work during
that month by use of the following formula. Rates are expressed per annum
per 1000 persons.
Number of injuries and deaths x 12
Number of persons +- 1000
Fatality rate
The fatality rate relates the deaths which are reportable in any month to the
Chief Inspector of Mines under Mines and Works Regulation 25 .1 (a) to the
number of persons at work during that month by use of the following for-
mula. Rates are expressed per annum per 1000 persons.
Number of deaths x 12
Number of persons +- 1000
Additional analytical data

From the accident investigation records, identification of certain key facts
about each injury and the accident which resulted in the injury can be ob-
tained. Summarization and analysis of these facts will generally show a pat-
tern of injuries and accident occurrences. These patterns serve as a guide
to areas within plants, qmditions, acts and circumstances at which accident
prevention efforts should be directed.
The frequency of analysis will to some extent depend upon the number
of accidents which occur. In metallurgical sections where the accident ex-
perience is low, analysis should possibly cover an annual or six-monthly period
981
so that meaningful trends can be obtained.

Salient items which should be extracted for analysis include:
Occupation of injured.
Wark place: the specific place or area where the persons were injured.
Nature of injury: the type of physical injury incurred.
Part of body: the part of the injured person's body directly affected by the
injury.
Agency: the object, substance or exposure which directly produced or inflicted
the injury.
Task perfarmed: the task the injured was performing when the injury
occurred.
Immediate-causes: what acts, failure to act and conditions contributed most
directly to the accident.
Basic causes: the real problems and the reasons for the existence of the
immediate causes.
Other facts such as time, age, ethnic group, experience can be included
in the analysis.
This section has mainly dealt with occupational injuries. Loss in the other
areas such as damage to equipment, loss and wastage of materials and the
environment should be subjected to the same management corrective and
preventative action.
17. 7 .3 Benefits derived from accident investigations and analysis

Facts highlighted during investigations and the analysis of these and other
relevant data pertaining to accident trends have a direct impact on other pro-
gramme activities. Management deficiencies relating to supervision, discipline,
training and communication are identified. Faults are found in task pro-
cedures, plant design and layout and material and equipment specifications.
All of these cause management to react and apply the required remedial action
to tighten the control function.
17 .8 Communications
Effective lines of communication are essential to successfully manage a safety
system. These can be established on a group basis whereby communi,cations
from top management descend via the various levels of supervision down
to individual groups of workers and also provide an avenue for communica-
tions from the workers to top management. In addition personal communica-
tion systems on a one-to-one basis must be formulated.
17.8.1 Group meetings

All plant employees, except clerical staff, attend monthly group meetings
where topics supporting critical subject promotions are discussed. The ma-
jority of these subjects are selected from the incident investigation and analysis
activities and are based on actual case histories of accidents or high potential
near-loss incidents. The plant production superintendent and general engineer-
982
ing supervisors in consultation with the loss control officers normally select
the topics and jointly document the salient points for promotion. Obviously
the topics must be applicable to the section in which they are promoted and
supported by visual aids in the form of posters or objects such as tools or
equipment. The duration of the meetings is generally no longer than 10
minutes and meetings are held by the various levels from middle managers
down to the workers. The plant production superintendents and engineering
supervisors hold meetings with the next line of supervision below them, name-
ly the metallurgical foremen and engineering foreman. These foremen hold
meetings with the plant foremen and artisans who in turn communicate down
to the technical officials and the remaining work force. Random checks are
necessary to ensure that all employees have an understanding of the topic
and that an increased awareness of the potential loss and prescribed action
has been achieved. Brief notes of the meetings are kept, and points raised
for upward communication and action taken are documented for discussion
in senior management safety meetings.
17.8.2 Personal communication

Personal communication is effectively used to control losses by assisting
employees to learn certain aspects of their jobs, providing individual job
orientation, prbmoting and using proper task instruction techniques and key
point tipping (more commonly known as 'safety tips and hints').
Newly engaged employees or employees who are transferred to sections
of the plant with which they are unfamiliar are given formal on-the-job in-
duction by their immediate supervisors. The foremen are responsible for in-
ducting technical officials and artisans as soon as possible after the person
arrives in the section. To ensure that no important points are overlooked,
guidelines are provided for use by the foremen. These are in most instances
specific to sections. An example of an induction format for the milling section
is given in Figure 17 .4
As is the case in most programme activities, a follow-up on the induction
is done by a metallurgical training officer with the new employee within a
month to evaluate the effectiveness of the induction. Supervisors are en-
couraged to use proper task instruction techniques whenever an unfamiliar
task is assigned to an employee to ensure complete understanding of what
is to be done.
17 .9 Emergency Preparedness
Even in the most efficiently run operations the risk of emergencies exists,
in the form of fires, incidents involving chemicals, and accidents resulting
in multiple casualties. To ensure prompt and effective action to control the
losses which are possible from these emergencies, comprehensive procedures
for gold plants should be established. Persons who are responsible for tak-
ing action in the event of an emergency occurring must be fully instructed
and trained in their duties. Wherever practical, unscheduled drills or mock
emergencies should be staged at intervals.
983
Figure 17.4. Sample record sheet for on-the-job induction of new employee.
ON-THE-JOB INDUCTION OF THE NEW EMPLOYEE

The employee will complete this form with his immediate Supervisor and return it to the Train-
ing Officer.
SECTION: MILLING
NAME: ......................................... . DATE: .......................................... .
DO YOU KNOW? Explanations I would like

understood more information
I . Where the fire alarm is situated?
2. Where the assembly point is?
3. Where the medical station is?
4. What personal protective equipment should

be worn in your section?
5. How to identify drinking water on the

plant?
6. The regulations pertaining to machinery?
7. The regulations pertaining to conveyor

belts?
8. The regulations pertaining to machine

guarding?
List below any further information you require regarding your job.
EMPLOYEE SIONA TURE: ...................................................... .
REMARKS BY SHIFT FOREMAN _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
SIGNATURE OF SHIFT FOREMAN:
c.c. L C 0
984
Emergencies with high loss potential are grouped into three categories:
accidents resulting in injury to persons; fires; miscellaneous.
17 .9 .1 Care of the injured and m

In addition to the qualified medical staff and the modern medical services
provided on the gold mines, a high degree of first aid training of employees
is maintained. This ensures prompt and competent treatment of the injured
and ill to alleviate pain and suffering and to prevent or reduce loss. Specialized
training is required in the treatment of cyanide poisoning and the use of
breathing apparatus and resuscitation equipment.
Medical stations, with qualified medical nursing staff in attendance, are
situated in close proximity to the plants. Within the confines of the gold plants
first aid equipment, including breathing apparatus and resuscitation equip-
ment, is provided at strategic positions. In areas where acid and cyanide are
handled, safety showers and eyewash fountains should be provided. The
locality of all these facilites must be clearly identified by international sym-
bolic signs. During induction employees should be briefed on the availability,
and shown the exact locations, of the facilities. To promote interest and main-
tain proficiency in first aid treatment of patients employees should be en-
couraged to participate in annual competitions.
17.9.2 Hazardous chemicals

The handling of hazardous chemicals is a consfant source of potential loss
in chemical process plants and management recognizes its obligation to in-
form employees of the dangers involved and the treatment in case of
accidental exposure.
Data sheets containing the basic information on all toxic substances used
in gold plants are used to communicate the information to employees. The
data include inter alia the description, usage and first aid treatment of each
substance, and should be displayed on notice boards in the areas where the
substance is used. It is essential that the information is conveyed to illiterate
employees by their supervisors. An example of a data sheet for cyanide ap-
pears in Figure 17. 5.
Similar data sheets are available for: hydrochloric acid, caustic soda,
sulphuric acid, degreasing fluid, lead nitrate, aluminium sulphate and zinc
dust.
17 .9 .3 Fires
The incidence of fires in gold plants is low but the potential is relatively high
owing to the large quantities of rubber conveyor beltiµg used. Stringent fire
prevention standards should be enforced when arc-welding or oxy-acetylene
equipment is used. No-smoking zones should be defined and open fires within
the confines of the plants prohibited. Fire protection measures should receive
close attention during physical condition ratings and planned inspections
(refer to section 3 of checklist in Appendix 17 .1), and detailed surveys should
be conducted by fire protection consultants to identify exposures which have
possibly been overlooked by plant and loss control personnel.
985
Figure 17 .5. Hazardous chemical or substance data sheet.
Name Cyanide
Appearance NaCN: Normally clear, colourless to pale straw-coloured liquid but
may sometimes be darker.
Ca(CN) 2 : Slightly straw-coloured liquid but changes to yellowish-
brown colour during storage.
Ca(CN) 2 : Steel grey to black flakes.
Usage Gold dissolution.
Health Hazard Extremely poisonous and can be absorbed through the skin. Symp-
toms of poisoning: General weakness, heaviness of limbs, difficulty
in breathing, headache, dizziness, nausea and vomiting followed by
death.
Emergency & First Send for medical assistance.
Aid Treatment Ingestion: Give cyanide antidote followed by amyl nitrite. Apply
artificial respiration if breathing stops.
Absorption through the skin: Remove contaminated clothing and
drench the skin with water.
Inhalation of fumes: If patient is still breathing break a vial of amyl
nitrite and give to patient by inhalation for 20 seconds. Repeat every
2-3 minutes.
Personal Protective Full face shield
Equipment to be Worn Elbow length PVC gloves
During Handling Acid resistant overalls or coats
Boots
(Airline respirator to be available)
Storage Store in closed tanks equipped with breather pipes.
Spill or Leak Prevent spillage from entering drains or water courses.
Procedure Detoxify with ferrous sulphate or hypochlorite.
Flush residue away with copious quantities of water.
17 .9 .4 Miscellaneous
The incidents in this category which have the greatest potential and probability
for loss include:
Major electric power failure.
Significant acid spills from a storage tank or pipe column.
Acid or cyanide spillage during bulk off-loading operations.
An example of the emergency procedure to be followed when a major
electric power failure occurs is:
Ciose master valves. (All master valves are clearly identifiable.)
Drain pipe columns.
Close air valves under pachucas.
Ensure that emergency power plant is started to drive thickener rakes.
Drop load in filter pans.
Arrange for equipment to clear chokes in pachucas.
986
Check slime filter pumps before restarting to ensure that slime has not
solidified.
Start diesel generator to ensure adequate water supply in case of a fire.
17 .10 Physical Examinations of Employees

17 .10.1 Medical examinations
In terms of section 13 of the Occupational Diseases in Mines and Works Act
of 1973 and the amendment thereto dated 28 January 1983, certain areas
in gold plants are classified as risk areas owing to the presence of potentially
harmful silica dust. These areas are: crusher plants, grinding mill plants,
rotary filter plants, smelt houses, surface rock tipping bins and change- houses
for persons employed in dusty atmospheres.
Before commencing duties employees allocated to these areas must be
medically examined by the Medical Bureau for Occupational Diseases or the
mine medical staff and declared fit to work in dusty atmospheres. They must
be re-examined annually.
17.10.2 Eyesight examinations·
The eyesight of all employees is assessed on engagement and at yearly inter-
vals thereafter. This will ensure that those employees for whom good eyesight
is important have the visual ability to enable them to work safely and
efficiently. Where possible, deficiencies are corrected by medical treatment
or by the wearing of spectacles. To determine visual acuity, employees
undergo the Snellen E chart test which assesses the person's ability to see
correctly the orientation of the letter E symbol. It can be effectively used
for illiterates. Snellen E charts of different manufacture differ from one
another in the size and width of the lettering; for this reason, it is important
to conduct eyesight tests at distances which suit the chart used. The
616 - 6160, Curry & Paxton Ltd (London) chart, has been found to be most
suitable for routine eyesight testing.
The Snellen chart should be kept free from dust and grime as the ac-
cumulation of such substances changes the contrast ratio of the chart resulting
in incorrect visual acuity measurements. The Snellen chart (reduced size) is
shown in Figure 17 .6.
Following the assessment of visual acuity, colour perception should be
determined on men employed in work categories requiring unimpaired col-
our vision. The test is done on binocular vision. As it is not necessary for
this defect to be quantified, and for reasons of simplicity, the Ishihara Col-
our Blindness Test is recommended. Apart from the pseudoisochromatic
plates, no other special equipment is required (Van Rensburg, Strydom and
Kielblock, 1981). Included in these categories are: electricians and electrical
aides, vehicle drivers, operators, and technical officials.
In addition to the visual acuity and colour blindness tests, the depth
perception, peripheral vision and the rapid adaptation to low light conditions
and glare recovery of all vehicle drivers are assessed using an orthorater (or
similar) vision tester.
987
24
12
Figure 17 .6. Snellen chart (reduced size).
988
17.10.3 Audiometric testing

In recent years noise pollution leading to deafness has been recognized as
a major industrial hazard. The nature of the product handled in the rock
delivery section of gold plants, and the nature of the processing in the
crushing, sorting and milling sections results in a large percentage of the work
force being subjected to noise levels in excess of 85 dB A scale for periods
of 8 hours or longer per day. Areas in the plants where noise levels in excess
of 85 dB A scale occur should be identified, and employees working in these
areas should be required to wear hearing protection devices for the duration
of the shift. Casual visitors to these sections should be supplied with protective
devices for use whilst in the areas.
Noise-induced deafness caused in the work situation is compensatable
in terms of the Rand Mutual Assurance rules. Although ear plugs and ear
muffs afford some protection against deafness, the solution lies in engineer-
ing the noise out at the source. This, of course, is a long-term solution and
considerable work by design engineers is necessary to reduce noise emission
from machinery.
The hearing ability of all gold plant employees is assessed on engage-
ment. The information is used to provide a data base against which periodic
assessments are compared and thus ensures timeous warning of hearing im-
pairment. An audiometer is used to measure objectively the amount of hearing
loss in decibels at various frequencies.
Audiometric instruments and soundproof cubicles are used for the tests
and approximately 25 persons can be tested per day on each instrument. The
information is stored in the employees' computerized personal records to
facilitate access.
The person being tested is seated in a soundproof cubicle and reacts by
pressing a button to tones relayed through earphones by a trained operator.
The results are printed on an audiogram for record purposes. These are com-
pared with medically established norms to determine into which of four
categories they are placed. Results of the tests are filed in the employees'
personal records.
Category 1 - Cases where hearing loss is below warning level.
Category 2 - Cases where hearing loss is above warning level but below
referral level.
Category 3a - Cases where the sum of the hearing loss for either ear
at high or low frequencies is greater than referral level.
Category 3b - Cases where the difference in the sums of the hearing
loss between the two ears exceeds 45 dB in the low frequency range or
60 dB in the high frequency range.
Category 3c - Cases where the sum of the hearing levels for low or
high frequencies shows an increase of 30 dB or more when compared
with the last audiogram or an increase of 45 dB over the last 3 years.
After categorization, the following action is taken for the various
categories:
Category 1 - Audiograms are stored for record purposes.
989
Figure 17. 7. Audiometric testing in progress. The audiometer

is on the left of the picture, and the door of the cubicle is in
the open position.
990
Category 2 - Persons in this category are warned to use hearing

protection devices. Their audiograms are stored and they are retested
after six months to establish whether any further hearing loss has
occurred.
Categories 3a, 3b and 3c - People in these categories are referred for
medical treatment or specialist examination (Nairn 1985).
Figure 17. 7 shows audiometric test equipment in use.
The eyesight and audiometric testing is done in accordance with the
recommendations of the Industrial Hygiene Branch of the Chamber of Mines
of South Africa.
Figure 17 .8 shows an environmental survey record sheet.
17.11 Personal Protective Equipment

Protection of employees by mechanical means and engineering design is
generally more reliable than protection which is dependent upon human
Figure 17 .8. Sample record sheet for results of dust, noise and illumination survey conducted
at a gold plant on 23 January 1986.
ENVIRONMENTAL CONTROL DEPARTMENT
WASHING, SCREENING, SORTING AND CRUSHER PLANT
Sample Ave. Ave. Ave.

point Description dust noise illum. Remarks
Recommended levels:
Dust - Less than 200p/ml
Noise - Less than 85 dB (A)
Illumination - Walkways - above 5 Lux
Moving machinery - above 20 Lux
Observed by: Environmental Superintendent:
Sect. Envrm. Officer: Gold Plant Superintendent:
991
behaviour. A machine designed so that it effectively confines flying particles

eliminates a cause of accidents and the need for an operator to wear eye pro-
tection. Similarly, mechanical removal of hazardous fumes should be pro-
vided, rather than respirators for operators in hazardous environments. Not-
withstanding the progress made in providing for safety in engineering design
of machinery and equipment the wearing of certain personal protective equip-
ment is mandatory in gold plants.
Each task performed by employees should be critically examined to
identify those having a potential for injury which can be eliminated, or reduc-
ed in severity, by the use of protective equipment. Information from accident
analysis (Section 17. 7 .2) is used to identify these tasks or occupations. For
example, if the analysis shows high accident trends in toe injuries amongst
grinding mill relining workers, the issue of steel toe-cap safety shoes would
be justified. Generally a number of basic protective equipment items are sup-
plied to all employees. These are: safety helmets, safety boots or shoes,
overalls, and wrist-length PVC gloves.
Other items specific to particular occupations would be supplied, such
as respirators for sand blasting operators and spray painters, and welding
helmets and spats for boilermakers.
It is essential that stringent standards in respect of items to be worn whilst
specific tasks are performed are clearly defined, and supervisors must en-
sure complete compliance with these standards. Protective equipment should
be maintained in good condition. Safety goggles with badly scratched lenses
could impede an employee's sight and he may be reluctant to wear them and
thus risk an injury.
Most items of protective equipment are of a personal nature and
employees should not have to share such items. These include safety helmets,
goggles, hearing protection devices and dust respirators. Items which could
be shared by a number of employees include safety harnesses and lifelines.
Employees must receive instruction and demonstrations in the use of the more
sophisticated equipment such as respirators and safety harnesses.
A number of factors should be considered when specifications are defin-
ed or selection of equipment is made. These are: comfort, provision of the
necessary protection, correct equipment for specific application, and ease
of use.
Comfort - If the item is not comfortable to wear, it will generally not
be used. Examples of this are gloves which are too rigid and footwear
with incorrect pitch of the sole.
Provision of the necessary protection - Safety goggles having low im-
pact resistance, but supplied for use at pedestal grinders, provide no
protection.
Correct equipment for specific application - Particulate filter
respirators supplied as protection against solvent vapours are of little
value.
Ease of use - If a safety harness is difficult to put on or assistance
is required to fit the harness, there is a chance that it will not be used.
Figure 17 .9 shows personal protective clothing in use.
992
Figure 17 .9. Personal protective equipment worn whilst off-

loading cyanide solution from a tanker.
17.12 Task Analysis and Procedures

Supervisors will agree that there is a correct way to carry out all critical tasks
and that most accidents result from substandard acts because an established
procedure was not followed. A critical task can be described as any sequence
of steps or activities carried out in the performance of a specific work assign-
ment and which has a history of, or the potential for, significant loss. This
could be in the form of an accident which results in injury to persons, damage
to equipment, wastage of material or pollution of the environment.
It is necessary at an early stage in the safety programme development
that all critical tasks be identified and an inventory of these compiled for
each individual occupation within the gold metallurgy discipline. Respons-
ibilities and organizational relationships which are normally included 'in a
position charter or job description are not classified as tasks and should not
be included. Supervisory functions such as giving instructions should not be
included, and no benefit is derived from including management functions
like planning and forecasting. Both line management and loss control staff
should be involved in compiling the inventory. Members of line management
have an intimate knowledge of the tasks performed by.each occupation but
experience has shown that guidance is required to enable them to classify
tasks as critical or non-critical.
17 .12.1 Identification of critical tasks

To identify tasks which are critical and the ranking of these in order of
criticality, the following should be examined:
Can the task, if not carried out correctly, result in significant loss whilst
being performed?
993
Can the task, if not carried out correctly, result in significant loss after
having been performed?
How serious is the loss likely to be? (What is the severity of injury,
cost of damage or cost of production loss likely to be? Are other persons
or departments likely to be affected?)
What is the expected frequency of occurrence?
Frequency of occurrence is governed by a number of factors, of which
the most important are:
The number of times the task is performed in the organization in a
specific time period (repetitiveness).
The chance that there will be a loss as a result of performing the task
(probability of loss).
It must be recognized that there are many degrees of criticality and, in
fact, every task worth doing is critical to some degree. It follows, therefore,
that a system which develops a scale of criticality is likely to result in fewer
differences of opinion than one which merely classifies the task as critical
or not critical. It is suggested that the above parameters be converted into
three scales relating to severity, repetitiveness and probability of loss.
Although much subjective judgement is still required, the fact that each
parameter is given due consideration results in a more consistent and logical
ranking and subsequent classification of tasks according to criticality.
Severity is derived from the injuries and costs of the losses being incur-
red or the loss most likely to be incurred as a result of wrong performance
of the task. In many cases a whole range of losses could occur but only the
most likely result should be considered. For example, if the lock-out pro-
cedure for a tube mill which is being relined is not followed correctly and
this results in an accident, it is likely to be serious, whereas an incorrect
shovelling technique used to clean up a slimes spill is more likely to result
in a small loss rather than a large one.
17.12.1.1 Severity
A scale of from zero to six is suggested as follows:
0 No injury or a loss of less than RIOO.
2 Minor injury without lost time or a loss of RlOO to R500.
4 A lost time injury without permanent disability or a loss of
more than R500 but not exceeding R2 OOO.
6 A fatality or permanent disability or a loss exceeding R2 OOO.
The degree of severity corresponding to the various values on the scale
and the number of points on the scale can be varied to suit requirements.
17 .12.1.2 Repetitiveness
Repetitiveness can be assessed from the following tabulation according to
a scale of 1 to 3.
994
Number of times task is performed by each person
Number of persons Less than Few times Many times

performing task daily per day per day
Few 1 2
Moderate number 2 3
Many 2 3 3
The scale used can be extended if desired and, in this example, a scale
of 1 to 5 could have been adopted.
17 .12.1.3 Probability
The probability of loss occurring each time a particular task is performed
is influenced by the following factors:
Hazardousness, i. e. how dangerous is the task?
Complexity of the task.
The chance that there will be loss if the task is performed incorrectly.
It is not suggested that these factors should be evaluated separately but
they should be borne in mind. The key question is, "How likely is it that
things will go wrong as a result of the performance of this task?'' For exam-
ple, there is a higher than average probability (chance) of injury when
manhandling heavy material whereas there is a less than average probability
of injury when stacking small items in an equipment store. Since, for the
sake of simplicity, only the most likely loss is considered when evaluating
severity, it follows that only the probability of that particular loss should
be considered.
A scale of from - 1 to + 1 is used as follows:
- 1 Less than average probability of loss.
0 Average probability of loss.
+ 1 Greater than average probability of loss.
The scale can be extended if desired. The points allotted to each of the
three parameters are then added to indicate a scale of criticality ranging from
0 to 10. It is, in effect, an order of priority. Management may decide that
all tasks allotted less than, say 3 points, will be disregarded from a loss con-
trol point of view and not be listed as critical tasks, whereas tasks allotted
8 or more points will be regarded as the most critical tasks requiring immediate
and constant attention.
17 .12.1.4 Identification of tasks requiring procedures, standards or rules

Once all critical tasks have been identified the next step is to identify those
for which procedures, standards or rules should be established to control
potential loss. A standard is a description or specification of the completed
task. A procedure is a step by step description of how the task is accomplished
and it is written in sufficient detail to suit the purpose for which it is intend-
995
ed. It is a tool for teaching employees the most systematic way to do a critical
task consistently with maximum efficiency. Some tasks, particularly those
performed by a skilled artisan, may not require a standard or a procedure
but only a few task rules. Other tasks may require both a standard and rules.
If a procedure is required, all necessary rules and standards are included since
the procedure describes every step (Chamber of Mines of South Africa, In-
ternational Safety Rating Manual, February 1984).
17.12.1.5 Task analysis

Having identified the tasks which require procedures, all important aspects
of a task must be considered and evaluated in order to determine one unified
procedure for doing the task the correct way. This is achieved by task analysis.
The four steps representing the basic framework for doing a task analysis are:
Determine the critical task to be analysed.
Break the task down into an orderly sequence of steps.
Determine the potential for incidents (loss exposures).
Make an efficiency check of each task step. (Is there not a better, safer
or more effective method?)
Develop recommended controls.
If required, a complete and detailed task procedure can now be drawn
up and used to develop lesson plans for skill training purposes. Experience
has shown, however, that the information contained in the task analysis
worksheet will usually suffice for task observation, information and lesson
plan development. A specimen of a task analysis worksheet for adding cyanide
to a pachuca is shown in Figure 17.10.
In the initial stage of the programme introduction, large numbers of
tasks will be identified which require procedures to be written or revised.
This work should be scheduled and priorities set for procedures. Loss con-
trol staff must provide the necessary advice and guidance to line manage-
ment when this is done. Whenever a serious incident occurs, the appropriate
procedure, standard or rule should be reviewed and where necessary revis-
ed. It is important that procedures, standards and rules and the changes
thereto be communicated to supervisors and workers. The training depart-
ment plays a vital role in this regard and many lesson plans used for formal
training courses originate from task procedures. Procedures, standards and
rules are available to literate employees in all control rooms, and illiterate
employees are briefed by their supervisors on what is relevant to their jobs.
17.12.1.6 Critical task inventory summary

Some of the tasks performed by a gold plant technical official appear on
the inventory summary in Figure 17 .11.
The criticality of each task has been determined. It is convenient to make
use of the summary to include other factors which have an influence on safety
and efficiency whilst the task is performed. These are training requirements,
personal protective equipment, eyesight and hearing tests.
996
Figure 17.10. Typical task analysis worksheet.
PROTECTIVE EQUIPMENT
TO BE WORN
Full face shield
Elbow length PVC gloves
Task Analysed: Adding Cyanide to Pachuca Acid resistant overalls
Occupation: Technical Official Boots
Note: Air line respirator to be available
STEP SEQUENCE OF TASK LOSS EXPOSURES RECOMMENDED

NO. STEPS (Safety, Health, Equip- CONTROLS
"How the task is being ment Damage, Produc- ''What can be done to
done" tion Loss, Grade) prevent it from going
"Only record what can wrong?"
go wrong"
1. Person performing the Personal injury. Ensure that person per-

task to wear appropriate forming the task is clad
protective equipment. with all the appropriate
protective equipment
before commencing with
task.
2. Start cyanide pump (see Damage to pump if Adhere strictly to start-

Standard Procedure). started incorrectly. ing procedure.
3. Fill metering tank with No loss exposure (tank No loss exposure.

cyanide. totally enclosed).
4. Check that cyanide Waste of reagents as Ensure pump has stop-

pump stops when meter- contents of metering ped. Check visually.
ing tank is full and reset tank will overflow into
automatic stop switch pachuca if pump is not
after pump stopped. stopped.
5. Ensure correct amount Waste of reagents. Ensure pachuca level is

of slime in pachuca. Insufficient dissolution correct. Re-check
of gold. measurement.
6. Check slime pH (must be Waste of reagent and Ensure pH is correct.

10,5 or more). personal injuries through Check and re-check.
cyanide gas if pH is
below 10,5.
7. Ensure agitation is Inadequate agitation will Check agitation visually.

adequate. result in inadequate
dissolution of gold.
8. Open valve to empty Insufficient dissolution Ensure that correct

cyanide contents from of gold if all cyanide in amount was added by
metering tank into metering tank is not conducting a titration
pachuca. added. test (see procedures for
cyanide titration).
Figure 17.11. Typical summary record of a critical task inventory.
Eyesight Test
....
<l)
Occupation Task '§ ....

0 ::>
0
-""
t 0
0 u
Technical Decant a
Official Merrill filter
Recover amalgam
Prepare a cascade
trolley for gold
pour
Weigh gold bar
Acid treat gold

slime
Acid treat Merrill

filter
Add cyanide to
pachucas
Coat Merrill
filter
*Refer to table below for particulars of items and codes

PERSONAL PROTECTION EQUIPMENT CODE
Protective Equipment Code Protective Equipment Code
Hard hat A Standard overall J

Clear goggles/face shield B Rubber suit K
Shaded goggles/face shield c Acid resistant overalls L
Welding helmet D Red reflective jackets M
Sand blast helmet E Numo jacket: Black PVC N
Gloves: Elbow length PVC FI Apron: Full length leather 0
Short length F2 Safety belt and chain p
Long leather F3 Ankle spats Q
Short leather F4 Respirator R
Long rubber F5 Hearing protection device s
Short rubber F6 Dust coats T
Approved boots/ shoes G Butcher aprons u
Rubber gumboots H Air line respirator v
Rubber leg guards
998
17.13 Planned Task Observations

Planned task observation is a proven technique which enables a supervisor
to know whether or not a worker is performing all aspects of a specific task
with maximum efficiency. Maximum task efficiency means greater and safer
production at lower costs. Observations of substandard performance pro-
vides warning to take corrective action before downgrading incidents occur.
There are basically two types of task observations, namely the informal
observation and the planned observation. The informal observation is done
by a supervisor as he goes about his routine activities. He naturally looks
at tasks being performed by the workers and will on occasions notice unsafe
acts, but in fact these informal and casual observations are usually quite
haphazard and are of little value. The planned observation, however, is a
purposeful activity which requires time to observe a task in a systematic way.
It is not done informally nor simultaneously with $Orne other duty or activity.
An effective planned task observation programme will indicate to
management the degree of worker compliance with the authorized task pro-
cedures with special emphasis on personal injuries. It will highlight if ex-
isting training procedures no longer meet the practical on-the-job re-
quirements. The programme must ensure systematic coverage of the tasks
which have been included in the critical task inventory summary, starting
with the most critical and proceeding on a diminishing priority basis to the
less critical. It must be flexible to enable observations on problem areas reveal-
ed by the incident analysis to be done whenever necessary. Priority must be
given to observing the task performance of the new employee as well as the
poor performer and the known risk-taker.
The observations should be done by the foremen and technical officials
in charge of workers and should not as a matter of routine be delegated to
a staff department such as industrial engineers. Details of how a planned
task observation should be conducted are:
Select the worker and task. This could be influenced by accident trends,
a new worker engaged in the section or a programmed routine
observation.
Prepare for the observation. The supervisor should refer to the previous
observation done on the worker and the task and must be completely
familiar with the sequenced steps to be followed, special tools, safety
or other equipment to be used. This information is available in the pro-
cedure and on task analysis worksheets. The worker should be advised
that he will be observed whilst doing the particular task.
Making the observation. The supervisor must give the observation his
undivided attention and should make notes of any deviations from the
authorized procedure. Only if the worker is doing something that
seriously endangers himself or others or if a process is likely to be
affected must he be interrupted.
On completion of the observation the supervisor should discuss any
deviation and other salient points with the worker and complete the
planned task observation record shown in Figure 17.12.
999
-
0
0
0
>-3
:r:
tI1
~>-3
Figure 17.12. Sample record sheet for observations on planned tasks. ~
n
METALLURGICAL DIVISION >-3
......
NAME:
<
tI1
(Plant Foreman) $:
t'rJ
Suggested improve- ~!:"""
Reasons for deviation ments in procedures,
!:"""
Working Person and action to ensure training, tools, equip- Superior c::::
Date Task observed place observed Observer Deviations future compliance ment and material sign :;o
0
~
0
'T1
0
0
!:"""
\::)
To enable management to direct corrective action at the most important

areas, a quarterly analysis of deviations from authorized procedures, the
reasons therefor and the suggested improvements in procedures, training,
tools and equipment is done for the discussion at the senior steering com-
mittee meetings.
17.14 Engineering and Purchasing Controls

The prime objective of this activity is to ensure that in the design stage of
equipment, plant or processes, exposures to hazards and injuries are
eliminated or controlled. In addition, good design will optimize maintenance
of plant and equipment and facilitate production at the planned rate.
A checklist is used which itemizes the important points to be considered
by competent persons and is designed to review the most common hazards.
It does not preclude other hazards. The intent is to focus attention on the
potential risks to employees who will operate the equipment or work in the
plant area. Engineering provisions will be made in the design to eliminate
or reduce to a minimum any dangers to the health and safety of the employees.
Before final approval by the Manager Engineering, designs of new or
modifications to existing buildings, structures and equipment will be scrutiniz-
ed by the following persons: design engineer, plant engineer, chief engineer
plants, firemaster, security officer, environmental officer, and loss control
officer.
Designs are checked to ensure compliance with mechanical, electrical,
civil and structural engineering standards and Mines and Works regulations.
Other areas considered are acoustics, radiation, ergonomics, fire protection
and security.
Construction of new plants and major modifications is usually under-
taken by outside contractors rather than mine labour. A procedure to be
followed by contracting firms whilst working in the metallurgical plants is
included in the tender documentation. Before construction work begins, the
senior site engineers should be briefed by the loss control officer on their
responsibilities in respect of the safety programme.
The purchasing function must be included in the programme to ensm:e
that no items are issued which do not have established purchasing specifications.
Rigid quality control examinations of items both at the manufacturing stage
and on delivery to the users are necessary to ensure that losses which can
arise from using inferior equipment, tools or chemicals are minimized.
17 .15 References
International Safety Rating Manual, Advanced Level (1984). Chamber of Mines of South Africa,
Johannesburg.
Nairn, W.A. (1985). Noise-induced deafness. Association of Mine Managers of South Africa,
Circular No. 1/85, Johannesburg.
Van Rensburg, A.J., Strydom, N.B. and Kielblock, A.J. (1981). A guide to the assessment of
visual acuity, colour perception and hearing in the mining industry. Research Organization,
Chamber of Mines of South Africa, Johannesburg.
1001
APPENDIX 17.1
Checklist for
General Planned Inspection and Physical Condition Ratings
Hazard Potentials
"A" - Extensive or permanent disability = Immediate action required
"B" Disruptive, serious injury Scheduled to be rectified as soon
as possible
"C" Minor or near losses Scheduled for a later date
***
The following is a brief description of what the rating team will scrutinize during
inspections:
1. Premises and surroundings

1.1 Building Clean and in Good State of Repair
1.1.1 Gutters, sheeting, doors, hinges and windows in good order.
1.1.2 Broken windows removed completely to eliminate immediate danger.
1 .1.3 Broken window panes near electrical equipment must be replaced imme-
diately.
1.2 Walking Surfaces, Elevated Walking Surfaces and Roads Clean and in Good
State of Repair
1.2.1 Floor openings covered.
1.2.2 No tripping and slipping hazards, e.g. loose boards, holes, grease or oil or
any other substance that may cause slipping.
1.2.3 Trenches of up to 150 mm in depth must be curved on both sides and trenches
in excess of 150 mm must be covered or handrails erected.
1.2.4 Elevated walking surfaces in excess of 0,5 metre above normal immediate
surroundings must be guarded with handrails.
1.3 Good Lighting - Natural and Artificial
1.3.l Windows clean.
1.3 .2 The standards of illumination in all sections of the plants will be determined
by the Environmental Control Department in conjunction with the Plant
Superintendent in day and night conditions. The Environmental Control
Department will be required to do an illumination measurement during April
and October, each year and produce a report of the results. Evaluation will
be based on the results of the above reports against the set standards.
In case of any deviation from the set standards, remedial action must be
taken followed by a remeasurement of the specific area.
1.3 .3 Surveys conducted by competent persons appointed in writing by the Plant
Superintendent will be acceptable.
1.4 Ventilation, Heating and Dust Control
1.4.1 Extraction fans available where required.
1.4.2 The maximum permissible dust limit is 200 particles per ml. The Environ-
1002
mental Control Department will be required to take dust measurements

during January, April, July and October and produce a report of the results.
Evaluation will be based on the results of the quarterly report against the
set standards. In case of any deviation from the set standards, remedial action
must be taken followed by a remeasurement of the specific area.
1.4.3 Steam Pipes - Leaking steam pipes, valves and flanges having the potential
to cause injuries to persons must be barricaded off and made safe until repairs
can be effected.
1.5 Noise Control
1.5 .1 The maximum equivalent noise exposure is 85 dB.
1.5 .2 The Environmental Control Department will be required to take noise level
measurements on all sections of the plant during April and October and
produce a report of the results.
1.5 .3 Areas above 85 dB must have the appropriate symbolic signs displayed at
all main entrances of the buildings.
1.5 .4 Any person working in such areas permanently must wear ear protection
at all times.
1.5 .5 Surveys conducted by competent persons appointed in writing by the Plant
Superintendent will be acceptable.
1.6 Aisles and Storage Demarcation
1.6.l Aisles, pathways, roadways, storage areas and areas in front of switchgear
and remote controls must be suitably demarcated.
1.6.2 These areas must be kept clear.
1.6.3 A "No Parking" area of lm2 in front of all open element type heaters to
be demarcated with yellow lines.
1.6.4 Colour Codes
(a) White
Indicates travelling areas; normally no work should be done nor should
any stacking or storing be permitted in these areas.
(b) Green
Indicates storing a~eas. Should only be used where material and equip-
ment is stored on a permanent basis.
(c) Yellow
Indicates "No Parking and No Storing Areas" and is normally used
where access is required in an emergency, e.g. Fire Extinguishers,
Emergency Exit Doors, Remote Switches, Electrical Switchgear and
Distribution Switch Boxes and also in front of open element type heaters
to prevent the storing of any combustible materials in front of the
heaters.
(d) Grey
Indicates working areas and is normally used in workshops.
Please Note: The above colours are specific for demarcations according to
Managerial Instructions and should not be used for decorative
painting on floor areas inside the plants.
1003
1.7 Good Stacking and Storage Practices

1.7 .1 Corners of stacks to be bonded.
1.7.2 Stack stable and height of stack restricted to three (3) times the base.
1.7 .3 Non-compatible materials stored separately.
1.7.4 Heavy articles to be stored on lower shelves.
1.7 .5 Stacking of materials must be done in such a manner as to avoid the necessity
for climbing over stacked articles. (Easily accessible from walkways.)
1.8 Plant and Yard Clear of Superfluous Materials
1.8.1 Scrap sent to salvage yard.
1.8.2 Rubbish sent to dumping site.
1.8.3 Re-usable items sent to storage area.
1.9 Scrap and Refuse Bins Removal and Disposal
1.9.1 Sufficient bins for scrap and waste.
1.9.2 Mantles changed regularly when full.
1.9.3 Mantles not overfilled.
1.10 Colour Coding
1.10.1 Colour coding board depicting all the various colours in words and in colour.
1.10.2 Boards prominently displayed.
1.10.3 Colour codes adhered to.
1.11 Guarding Around Tips, Tanks, Platforms and Pebble Bins
1.11.1 Effectiveness of guards.
1.11.2 Platforms or landings above 2 m high fitted with kickboards of 100 mm.
Kickboards must be flush with the walking surface.
1.11.3 Handrails installed to a minimum height of 900 mm above the walkway sur-
face and kneerails at 450 mm.
2. Electrical, mechanical and personal safeguarding
2.1 Machine Guarding
2.1.1 Pinch points guarded.
2.1.2 Gears, couplings and "V" Belts totally enclosed.
2.1.3 Insides of guards and areas that are to be protected painted ORANGE (where
applicable) so that when a guard is removed it will be obvious.
2.1.4 Guards securely bolted in position.
2.1.5 All drives securely guarded (handrails not acceptable).
2.2 Lock-out System and Usage
2.2.1 Switches fitted with effective locking-out devices and padlocked while
employees work on equipment, regardless of being isolated in substations.
2.2.2 Tripwires fitted to both sides of conveyors for the full length and in good
working order.
2.2.3 To prevent unauthorized use of machinery in workshops, all units must be
locked out in the absence of the artisan.
2.3 Labelling of Switches and Isolators
2.3.1 Switches and isolators clearly labelled to eliminate any possibility of
confusion.
2.3.2 Isolators labelled on both back and front of the panel.
1004
2.4 Ladders and Stairs - Numbered and Recorded

2.4.1 Portable ladders numbered and records kept of monthly inspections.
2.4.2 No portable wooden ladders are allowed in any section of the plant other
than smelthouses and electrical substations.
2.4.3 Vertical ladders above 2 m enclosed with a cage.
2.4.4 Vertical ladders without landings barricaded at the top.
2.5 Lifting Gear, Crawls, Hoists, Slings and Records
2.5.1 Coffin hoists, chain blocks and rope tackles clearly numbered and records
kept of monthly inspections.
2.5.2 No manilla rope will be used.
2.5.3 Slings marked with their safe working loads, either by means of •paint or
ferrules (if painted, a notice board should be erected at sling storage areas
indicating colour for SW L).
2.5.4 Crawl beams painted ORANGE and "Safe Working Load" painted on web
of beam, e.g. SW L 2 Ton. Safety stops fitted. All crawls to be numbered
and records of inspections kept.
2.5.5 Hooks pop marked (both sides), checked, records kept of these checks and
latches in order where applicable.
2.5.6 The standard checklist for scaffold planks must be used. Checks carried
out monthly and records kept.
2.6 Hand Tools - Condition of Hammers, Chisels, Etc.
2.6.1 No split hammer handles or burred heads.
2.6.2 No mushroom head tools.
2.6.3 Tools deal). and in good condition.
2.6.4 Standard of tool storage.
2.7 Welding Equipment, Cutting Torches, Hoses and Flash-Back Arrestors
2.7 .1 Welding machines numbered, checked and records of inspections kept.
2.7.2 Welding plug and feeder cable in a good state of repair (no joints).
2.7.3 Welding cables free of bolted joints and in a good state of repair.
2.7.4 No open wires protruding from the holder.
2.7.5 Oxygen and acetylene trolleys equipped with fire extinguishers (9 kg). The
portapack oxygen and acetylene sets require a 2,5 kg fire extinguisher.
2.7.6 Oxygen and acetylene bottles including portapacks fitted with flash-back
arrestors, as near to the bottles as possible.
2.7.7 Oxygen and acetylene bottles on trolleys must be secured.
2.7.8 Condition of hoses and gauges. (Only approved brass connectors permitted.)
2.7.9 Approved oxygen and acetylene bottle trollies to be used for transporting
purposes.
7.10 Only AFROX recommended pipe clips to be used for tying oxygen and
acetylene pipes together.
The key to close off gas bottles must be kept in position on acetylene bottle
whenever this apparatus is in use.
.12 Controls of oxygen and acetylene bottles must be closed when not in use
or when left unattended for any extended length of time.
. 13 Welding protection screens available in workshops and in good condition .
1005
2.8 Operating Equipment - Valves, Handwheels and Master Valves

2.8.1 Operating equipment in good order, e.g. handwheels, chute handles, launder
isolators, etc.
2.8.2 Master valves identified and records kept of monthly inspections as per
standard checklist.
2.8.3 Master valve location plans displayed at strategic places.
2.8.4 Master valves easily accessible.
2.8.5 Bulk hazardous chemicals, steam and compressed air lines to be fitted with
master valves.
2.9 Grinding Equipment
2.9.1 Notice at the grindstone indicating the RPM and the appropriate symbolic
sign pertaining to the wearing of safety goggles.
2.9.2 Goggles available at grindstone at all times.
2.9.3 Tool rest parallel with the centre of the stone and a maximum of 3 mm away
from the stone.
2.9.4 Stone free of glazing and loading. No evidence of side grinding. Stones not
in standard condition to be removed from grinder.
2.10 Hazardous Chemicals and Old Explosives Storage
2.10.1 "A Class" - Hazardous Chemicals
Requirements NaCN Na OH HCl
(a) Bund Walls Yes Yes Yes

(b) Security Fence Yes No No
(c) No Smoking Signs Yes No Yes
( d) No Naked Flame Signs Yes No Yes
( e) Two x 9 kg Fire
Extinguishers Yes No Yes
(f) Appropriate Symbolic Signs
WW4, WW5 Yes Yes Yes
(g) Notices re Protective
Clothing to be Worn Yes Yes Yes
(h) Demarcation where applicable Yes Yes Yes
(i) Persons to be Conversant
with First Aid Treatment for
these Chemicals Yes Yes Yes
(j) All Areas that require a
security fence must be locked
when not in immediate use Yes No No
(k) The following symbolic signs to be displayed whilst offloading hazard

chemicals:
No Entry Sign PV6
Warning of Corrosion WW 4
Warning of Poisonous Substances WW5
1006
(I) The appropriate symbolic signs must be displayed at all possible entrance
routes to hazardous chemical installations.
(m) All containers containing chemicals used for tests (solids and liquids)
must be clearly labelled.
(n) Flanges on delivery lines of all bulk hazardous chemical pumps to be
covered to prevent spraying out of contents in the event of a leak
occurring.
Notes
The bund wall requirements for cyanide, caustic soda and hydrochloric acid
area:
(a) Cyanide
Must be able to hold the total contents of the bulk installation.
(b) Caustic Soda
Must be able to hold the total contents of bulk installation.
(c) Hydrochloric Acid
Must be able to hold the total contents of bulk installation.
2.10.2 Old Explosives Storage
2.10.2.1 A plan to be posted up indicating the following:
(a) Position of boxes.
(b) The number of fire extinguishers required and distances away from
boxes.
( c) Type of extinguishers to be used.
(d) The number of "No Smoking" and "No Naked Flame" symbolic signs
required and location thereof.
2.10.2.2 Detonator and cartridge explosives boxes must be of robust construction
and manufactured from non-metallic material.
2.10.2.3 Kept locked at all times.
2.10.2.4 Boxes must be numbered and records kept.
2.10.2.5 Boxes must be painted RED.
2.10.2.6 Boxes must be clearly marked "Old Explosives" or "Detonators".
2 .11 Flammable Substance Stores
Flammable
liquid Gas Diesel- Paraf- HTH Meth a-
Requirements Stores stores stores ene fin stores no!
(a) Bund Wall NIA NIA NIA Yes Yes NIA Yes
(b) Security Fence NIA NIA NIA No Yes NIA Yes
( c) No Smoking
Signs Yes Yes Yes Yes Yes Yes Yes
( d) No Naked Flame
Signs Yes Yes Yes Yes Yes Yes Yes
(e) No Unauthorized
Entry Signs Yes Yes Yes Yes Yes Yes Yes
( f) 2 x 9 kg Fire
Extinguishers Yes Yes Yes Yes Yes Yes Yes
1007
Flammable
liquid Gas Diesel- Paraf- HTH Met ha-
Requirements Stores stores stores ene fin stores no!
(g) Drip Trays

(where applicable) Yes Yes NIA Yes Yes NIA Yes
(h) Trestles for
Drums (where
applicable) Yes Yes NIA NIA NIA NIA NIA
(i) Aisles and
Storage Areas to
be Demarcated Yes Yes Yes NIA NIA Yes Yes
(j) Stacking and
Storing Neat
and Safe Yes Yes Yes NIA NIA Yes NIA
(k) New Gas and
Flammable
Liquid Stores to
be Constructed
in Compliance
with Managerial
Instructions Yes Yes Yes NIA NIA Yes NIA
(I) Oxygen and
Acetylene Bottles
must be stored
separately NIA NIA Yes NIA NIA . NIA NIA
(m)AII Cylinders to
be secured
in upright
position (both
empty and full
bottles) and
clearly marked
"full" or
"empty" NIA NIA Yes NIA NIA NIA NIA
(n) All Stores and
Areas which
require a fence
must be locked
when not in use Yes Yes Yes NIA Yes Yes Yes
Notes
1. Bund wall requirements are as follows:
(a) Dieselene
Holding the total contents of the bulk installation.
(b) Paraffin
( c) Methanol
1008
2. Oil pump stations at Crusher Plants to be equipped with symbolic signs

and fire extinguishers as for oil stores.
3. All permanently fitted diesel and paraffin storage tanks will be classed
as bulk storage areas.
4. Acetylene cylinders used at On Line Gold Analyser (OLGA) machines
to be stored as per gas store standards, i.e. enclosed with fire ex-
tinguishers available and the appropriate symbolic signs displayed.
5. Appropriate symbolic signs to be displayed at all OLGA machine rooms,
for warning against flammable gases, as well as the required number
of fire extinguishers.
6. Oxygen and acetylene sets normally used in instrumentation workshops
must be stored outside the building as per gas store specifications.
7. Handy gas bottles at acid plants and laboratories must be stored as per
standard.
8. Rubber Lining Stores and Workshops
8 .1 Only Class 1, Division 2 electrics will be allowed in rubber lining stores
and workshops.
8.2 No heaters will be allowed in the rubber lining stores and workshops.
8.3 Offices inside a rubber lining workshop can be equipped with a heater
providing it is not an open bar type.
2.12 Portable Electrical Equipment
2.12.1 All electrical equipment to be fitted with pinplugs, numbered and records
of monthly inspections kept in a record book (i.e. drilling machines, disc
grinders, extension leads, etc.). Inspections to be done according to a checklist
which must be available in the book.
2.12.2 Urns, heaters, etc. in good order (no joins in flexible cords) and records
kept of monthly inspections. Heaters to be permanent fixtures.
2.13 Earth Leakage Relays - Permanent and Portable
2.13.1 Records of monthly inspections to be kept.
2.13.2 Where permanent, the office location must be recorded in the monthly in-
spections book.
2.13.3 Where portable, the unit must be numbered and recorded in the monthly
inspections book.
2.14 Vehicles, Dump Trucks, Forklifts and Rail Transport
Front
end Fork- Dumper Trac- Shunt-
Requirements loaders lifts Cranes cars tors LDVS ers Locos Jumbos
Roll Over Bar Yes Yes No Yes Yes NIA NIA NIA NIA
Seat Belts Yes Yes Yes Yes Yes Yes NIA NIA NIA
2,5 kg Fire
Extinguisher Yes Yes Yes Yes Yes Yes Yes Yes Yes
Pre-use
Checklist Yes Yes Yes Yes Yes Yes Yes Yes Yes
1009
(a) Safety belts must be worn on all vehicles which are equipped with them.
(b) Notices to be fitted to vehicles if passengers are permitted. Number of
passengers to be specified.
( c) If passengers are permitted, seats and seat belts to be fitted and used.
( d) Pre-use checklists available on vehicles and correctly completed by
drivers.
2.15 General Electrical Installations
2.15.1 Cables properly suspended. Disconnected cables must be starred and taped.
2.15.2 Cable glands secure and armouring well earthed in gland.
2.15.3 Sprag tubing intact and not unravelled.
2.15.4 Double adaptors being used correctly, i.e. not exceeding 15 amps.
2.15.5 Instrument panels locked (regardless of voltage).
2.15.6 Holes where circuit breakers and cables have been removed closed with
grommets.
2.16 Substations
2.16.1 Notices to be displayed:
(a) No unauthorized entrance (on all doors).
(b) Procedures in case of fire.
( c) Treatment for electrical shock.
( d) Unauthorized persons prohibited from working on electrical apparatus.
( e) "Danger" symbolic sign (Flash signs) on all doors.
2.16.2 Tap switches on transformers locked.
2.16.3 Transformers fenced or enclosed as per regulation, with no unauthorized
entrance and ''danger'' symbolic signs posted.
2.16.4 A plan to be posted up outside the substation indicating the types of fire
extingµishers, the number required and their position.
2.16.5 Substations with two doors or more must have an emergency door which
can be easily opened from the inside.
Outside locking devices on emergency doors must be removed.
2.16.6 Substation emergency exit doors to be demarcated a "No Parking" zone
to cover at least the width and radius of travel of the door.
2.16.7 Inspection records of substations available.
2.17 Personal Protective Equipment
2.17 .1 Wearing of personal protective equipment enforced.
2.17.2 Respirators used where necessary.
2.17.3 Hard hats to be worn in the plant area at all times except in offices.
2.17.4 Records of equipment issues available.
2.17.5 Symbolic signs used to indicate particular equipment or clothing to be worn
in specific areas or workshops.
3. Fire Protection
3.1 Plans
3 .1.1 Fire hydrant plans posted up in conspicuous places and including positic
of fire station and first-aid stations.
3.2 Fire Hazard Control
3 .2.1 Flameproof equipment to be used in flammable areas.
iOlO
3.2.2 Emergency telephone numbers displayed at PT telephones.

3.2.3 Outside contractors supplied with fire extinguishers by the General Engineer-
ing Supervisors.
3 .2.4 No unauthorized fires allowed in the plant area.
3.2.5 No tapping-off of any fire service to be allowed.
3.3 Overinspections by Senior Personnel
Overinspections are to be conducted by:
3.3.1 Metallurgical Superintendents and Plant Engineers - quarterly.
3.3.2 Plant Superintendents, Plant Production Superintendents and Engineering
Supervisors - monthly.
3 .4 Area: Demarcated and Clear, Extinguishers and Hydrants Accessible
3 .4.1 Extinguisher positions indicated.
3.4.2 Extinguishers properly maintained, i. e. 1,5 m frDm floor to base of ex-
tinguisher.
3.4.3 Fire hydrants demarcated.
3.5 Maintenance of Extinguishers, Hydrants and Fire Carts
3.5.1 Extinguishers and hydrants checked monthly and records kept. (Hydrant
pressures 4 to 5 bar.)
3.5.2 Extinguishers screened three-monthly and inspection records updated, both
on extinguishers and in a record book.
3.5.3 Foam and equipment available on fire cart.
3 .5 .4 C0 2 cartridges weighed every 3 months and records kept.
3.5.5 Checklists available on fire carts and in fire boxes and contents checked daily.
3.5.6 Only B.C.F. type Fire Extinguishers to be used in OLGA machine rooms.
3.6 Fire Alarm System
3 .6.1 Fire ~!arm clearly understood by all employees.
3.7 Fire Drill and Instructions on the Use of Extinguishers
3 .7 .1 Employees must know how to o~erate a fire extinguisher.
3.7.2 Fire drills to be held monthly and records kept of the results.
4. Loss Control
4.1 Knowledge of·
Loss Control Policy Statement
Safety Promotion Topic
Coiour Coding
Symbolic Signs
Quebec Safety System - 5 Point
2 Notice Board Indicating Injury Experiences
2.1 Strategically positioned notice board indicating the previous best perform-
ance and the current position of injury experience.
Poster Programme - Safety, Statutory and Symbolic Signs
.1 Posters must support current safety topic .
2 One topic promoted at a time.
3 Loss Control Policy - prominently displayed.
4 Only the appropriate (international) Safety and Symbolic Signs used unless
1011
otherwise specified by Management.

.4.4 Bulletin Boards
4.4.l Boards illuminated and uncluttered and the following displayed:
Manager's Policy
Plant-wide Rules
Major Loss Announcement
Safety Promotion Topic
Poster Promoting the Topic
5. First-aid Treatment
5 .1 Availability and Stocks
5 .1.1 First-aid stocks checked daily according to a checklist and replenished as
required.
5 .1 .2 Stretchers checked daily according to a checklist.
5 .1.3 First-aid boxes positions clearly demarcated.
5 .1.4 Cyanide antidotes available where applicable.
5 .1.5 Emergency rescue and first-aid equipment available at ammonia storage
areas.
5 .1.6 Acid, cyanide, ammonia areas equipped with at least one safety shower,
with an eyewash fountain and checked weekly.
Safety shower standards are as follows:
(a) The shower must be automatically activated when a person steps under
it.
(b) The eyewash fo.untain should be so situated that the lever is easily
accessible.
5 .1. 7 All safety showers and eyewash fountains must be readily accessible and
equipped with appropriate safety symbolic signs, viz:
Safety Symbolic Sign
Eyewash Fountain Symbolic Sign.
5 .1.8 Notices pertaining to the treatment of the following must be displayed in
both official languages in all change-houses and first-aid stations.
Gassing
Drowning
Heat Illness
Electric Shock
1012
Chapter 18
Laboratories and Pilot Plants

C.W.A. Muir
18.1 Introduction
It is generally accepted in the gold mining industry in South Africa that effi-
cient, ongoing metallurgical support should be provided for plant operations.
At the same time, it is essential that research be undertaken to improve cur-
rent process steps and.to develop new techniques. The recent progress in the
use of carbon-in-pulp (CIP) technology is a prime example of the latter.
The discussion in this chapter is concerned with metallurgical testwork
and evaluation, rather than with analytical chemistry and assaying. The lat-
ter subjects are dealt with comprehensively in a companion volume to this
book, entitled Assay and Analytical Practice in the South African Mining
Industry (Lenahan and Murray-Smith, 1986). Metallurgical laboratories may
be conveniently divided into two types, the first of which are control or sup-
port laboratories which are situated on mines, and are largely concerned with
day-to-day operations. These on-mine laboratories vary considerably in size
and scope, as later discussion will show. The second are the laboratories
operated by the mining houses or groups. Longer-term research and the in-
vestigation of new concepts are carried out in the group laboratories of most
of the major mining houses in South Africa. Five of the six of these do in
fact run group laboratories; Anglo American Research Laboratories (AARL)
is the largest, and primarily dedicated to research on gold and uranium ex-
traction.Eighty-four people are employed in the metallurgical section of this
laboratory alone. The smallest of the group laboratories is that operated by
Anglovaal with a total staff of about forty.
New work in research on local ores is carried out at Mintek, The Coun-
cil for Mineral Technology, which although largely funded by the central
government, carries out investigational programmes for the Chamber of
Mines, groups and individual mines. Research at Mintek in the national in-
terest is carried out free of charge, and the cost of other research is propor-
tional to its confidentiality. It should be pointed out that Mintek, and most
of the laboratories of the major groups, are not only concerned with the
metallurgy of gold, but their interests cover a wide range of work on other
precious metals such as platinum, as well as base metals and industrial
minerals. For example, a considerable amount of research is carried out on
the flotation and smelting of platinum group elements at J.C. I.' s group
laboratory, as well as on subsequent processing steps. Gold Fields' laboratory
is directly concerned with operations at Zincor (Zinc Corporation of South
Africa Ltd).
1013
In addition to the facilities which have been mentioned, certain local

equipment contractors and reagent suppliers run their own laboratories, and
can provide information on diverse matters such as the effect of flocculants
on pulp filtration, and CIP circuit design. Some work is also carried on at
the universities; for example, one of them is at present carrying out, on a
contract basis, a study on the use of bioleaching to enhance gold extraction
from refractory ores in subsequent cyanidation. It is not the intention in this
chapter to consider such laboratories, and the discussion will be restricted
to the control or support laboratories on gold mines and the research at group
laboratories and Mintek.
The second topic of interest is the operation of pilot plants. The main
reasons for using a pilot plant to test a new process or unit operation are
to evaluate the risks inherent in a new process and to operate in a continuous
processing mode, which is usually not possible in bench-scale tests since these
are normally done on a batch basis. Continuous operation in a pilot plant
allows a study to be made of the effect of recycled streams on the overall
process, and shows the build-up of impurities in the overall circuit. Differences
in water qualities in bench-scale test work and in the pilot-scale operation
are also important. It is usually essential to pilot a process to provide reliable
design data for the full-scale plant and to produce the necessary informa-
tion for a feasibility study. Another reason is to convince management that
a novel process is feasible; a successful pilot plant is usually the best way
of achieving this objective.
As would be expected, pilot plants, even when operated on gold mines,
are usually the responsibility of the research, rather than the service,
laboratories. Some overlapping in staffing can exist however, particularly
if shift operation is necessary. The advantages that accrue when a pilot plant
is operated at a mine relate particularly to the availability of typical
feedstocks, taken as bleed streams in the full-scale plant. In certain cases,
for example, when it is not necessary to have a continuous flow of typical
feed into a pilot plant, pilot-scale research can be undertaken more conve-
niently at the group laboratory, where support facilities are usually more com-
prehensive. The siting of a pilot plant therefore depends on the nature of
the research that is being undertaken.
It has been pointed out that the main gold-rich regions in South Africa
are highly concentrated in the Witwatersrand area or its vicinity, which makes
for a healthy interaction between research organizations and the mining in-
dustry. This close location assists research, in contrast to the situation in the
United States, Canada, Australia and the Soviet Union, where gold-rich
regions can be thousands of kilometres apart.
The ore of the Transvaal and Free State mines is generally very amenable
to processing by cyanidation. It is nevertheless remarkable how much effort
has been applied in recent years to research on the cyanidation process, which
has been used since before the end of the nineteenth century, and which was
not the subject of any significant efforts in research and development until
about the mid 1970's. While the annual output from amenable gold sources
comprises approximately 99% of the total South African production, never-
1014
LABO RA TORIES AND PILOT PLANTS
theless significant amounts of gold are obtained from refractory deposits in

the eastern Transvaal. Obviously, the amount of research and development
that is applied to these reserves is disproportionate to their economic impor-
tance but relates rather to the difficulties that are experienced in extracting
gold from them. A very minor amount of gold is recovered from anode slimes
in copper refining, as well as from the two platinum group metal refineries
in South Africa. These sources are considered to be outside the scope of this
discussion.
18.2 Characteristics of Metallurgical Laboratories of the Gold

Mining Industry
In this section, the layouts of both on-mine and group laboratories are dealt
with, together with their equipment and their staffing, followed by the func-
tions which they perform.
18.2.1 On-mine laboratories

18.2.1.1 Layout
The scope of the laboratories situated at gold mines in South Africa ranges
from the simplest grading room to extensive, well-equipped facilities. On some
of the smaller mines, the metallurgical testing function is assumed by the
assay offices.
In the Gencor group all routine analytical work for process control is
undertaken by assay laboratories on each mine. Metallurgical process develop-
ment work is undertaken by a Metallurgical Services Department with area
offices in Springs serving group mines in the central and eastern Transvaal,
and in Stilfontein serving the western Transvaal and the Orange Free State.
The Anglo American group has a very large operation at Vaal Reefs
and at the Joint Metallurgical Scheme in the Orange Free State. The scale
of the metallurgical test facilities that are available at these locations falls
between those found in group and on-mine laboratories.
The advent of the East Rand Gold and Uranium Company Limited in-
troduced new concepts into gold metallurgy on the Witwatersrand. Here the
unit operations include flotation, roasting and CIP, all on a vast scale. They
do not include crushing or grinding. The metallurgical laboratory that deals
with plant operation is also more elaborate than that provided on the average
mine. Rand Mines Milling and Mining have recently started treating tailings
dumps, but at the time of writing the facilities for on-site metallurgical
testwork were still quite limited.
Dump leaching in 1985 is the preserve of quite small operations, usual-
ly less than 50 OOO t/month. It may be expected that this method of gold
recovery wili be practised on an increasing scale. In the present small opera-
tions, facilities for testwork are extremely few and far between, and all these
operations obviously need better provision than they have at present.
The smaller the group that controls a mine the greater will be the object-
ion to having specialized manpower and equipment spread in on-site
laboratories, so that the central or group laboratory will tend to take over
1015
BC BC
A \Vet floor area
CTB B Filter press
B
BC Bench with cupboard
B CTB Concrete-topped bench
with storage shelves
B below
A C Drying oven
CTB [f] D Infrared drying hood
E Vented sample preparation
CTB [f] cupboard
F Sink
n
8,9 m G Locker
CTB c
- - - - - - - - 8,5 m _ _ _ _ _ _ __,_
Figure 18.1. Layout of a metallurgical research laboratory

sample preparation room.
more of the work of the on-mine laboratory.

A typical, well-appointed metallurgical laboratory on the mine will con-
sist of a grading room, a small laboratory where shift titrations and other
determinations such as pH measurements can be carried out, and a separate
room for an atomic absorption analyser, used mainly for the rapid deter-
mination of gold in solution. Adequate facilities for sample preparation will
be provided, and entirely separate locations for high and low grade materials
will be essential, in order to prevent contamination. Provision will be made
not only for preparing samples for the assay office, but also for moisture
determinations necessary for evaluation purposes. The floors in both the high
and low grade sample preparation rooms should be carefully cambered to
drainage systems, so that high standards of housekeeping can be maintained.
Adequate provision in the form of racks for the storage of samples
should be provided. The importance of the provision of adequate storage
facilities in an on-mine service laboratory cannot be overemphasized, and
generally at least three separate storerooms should be available. One of these
will be used to store laboratory glassware and reagents, as well as diverse
items such as assorted filter papers and retort stands. An adequate stock of
spare sieves used for grading analyses should be on hand, stored separately
with calibrated sub-master screens. Specialized equipment of a non-routine
nature can well be stored in this area, glass electrodes for pH meters and
spare hollow cathode lamps for the atomic absorption analyser being ex-
amples. In a third storage area, equipment of a more robust type as used
in sample preparation would be found. This store will contain enamel buckets,
1016
aluminium pans and the like. It will of course be necessary to provide an

office for the use of the person in charge of the metallurgical laboratory where
managerial functions can be carried out. Provision for basic toilet facilities
will also be necessary for the laboratory staff.
Facilities at on-mine metallurgical laboratories therefore depend on a
number of factors, not least being the philosophy that exists among senior
management in each of the six major mining houses. Obviously, it is to be
expected that greater emphasis will be placed on on-mine facilities if group
laboratory facilities are either non-existent or limited, and vice versa. Draw-
ings of what is considered to be a well-appointed on-mine metallurgical
laboratory appear in Figures 18.1 and 18.2.
18.2.1.2 Equipment
It is impractical to provide an equipment list that will suit the needs of every
on-mine laboratory, for as has already been stressed, such laboratories vary
widely in scope. However, the following is an inventory for what is considered
to be a well-equipped facility:
Controlled temperature baths.
Jaw crusher.
Rolls for grinding tests.
Rod mills.
Full range of sieves including spares.
Sub-master set of sieves.
Sample splitters.
Rotap-type machine for screening tests.
Laboratory flotation machine.
Range of balances.
Bulk scale.
Drying oven.
Microwave oven.
Thermometers.
Equipment for the preparation of all types of samples, both high
and low grade.
Dust extraction facilities.
Pressure filters.
Rolling bottle assembly.
Oxygen measurement equipment.
Facilities for CIP efficiency determination and carbon elution ef-
ficiency determination.
Vacuum filters
Hot plates.
A range o-f glassware: beakers, filter funnels, measuring cylinders,
Buchner funnels and flasks, burettes, pipettes, desiccators, etc.
Buckets.
Pressure pumps.
Vacuum pumps.
1017
--
0
00
.fil ili F
26 m .....J
:r:
tr1
'Ti
O'Q"
~v GG ~
.....J
i::
~
(D Room 5 ~
(1
......
r=i1
I sell
@] v
00 Rr I JJ .....J
......
tv
r
Po
til
'<
0
~
tr1
,...
i::
....,
0 13 m INJ BC (M
811[§] Room 2
~
t""'
Po
s
!]11 Room I
s
~VI
(D
?O
[
~
II[]'[] I E
2"
.... J L
()q
(')"
e.. m Room 1 Wet laboratory
0
'TI
....
(D Room 2 Dry laboratory I Vented sample preparation cupboard Cl
"'
(D Room 3 Balance room J Sieve shaker 0
Po
.... K II Room 4 Research officer's office K Storage shelves
t""'
()
Room Assistants' office tl
:::- 5 L Work bench
S"' M Variable-speed rolls
cr' A Wet floor area N Pulp density balance
0
.... B Small James table 0 Cupboard
Po
BC Bench with cupboard below p Balance
0.... c Heated rolls Q Safe
':< CTB Concrete-topped bench R Desk
D Filter presses s Scale
E Sink T Drawing desk
F Wet grading spray v Vacuum manifold
G Flotation cell w Water distiller
H Drying oven
Chemical reagents.
pH and redox buffer solution and meters.
Cleaning equipment.
Miscellaneous items : filter papers, filter stands, retort stands, sta-
tionery, protective clothing, etc.
First aid equipment, including the provision of a cyanide antidote.
A desk-top computer for the evaluation of results would be
desirable.
18.2.1.3 Functions
The list that follows indicates typical examples of testwork that is carried
out in mine laboratories:
Laboratory milling and gold leaching using a batch rod mill follow-
ed by cyanidation in bottles on rolls.
Cyanidation tests for gold to determine the effects of leaching time
and reagent concentration.
Filtration tests to determine filtration rates and to compare dif-
ferent types of filter cioth and the effect of different flocculants.
Grading tests and the determination of gold in the different size
fractions.
S~ttling tests to measure thickener efficiencies and the effect of
feed pulp densities, fineness of grind and flocculant additions.
Optimization of sample preparation procedures.
The determination of the percentage of moisture in various
samples.
The determination of the specific gravities of various pulp samples.
Preparation of all routine samples for chemical analysis and
assaying.
Preparation of composite samples.
The calibration of mass flowmeters.
The development of procedures for plant control.
Trouble-shooting investigations.
Flotation testwork.
The determination of carbon loading and elution parameters.
Pollution control.
Tracer tests to identify short-circuiting in leaching operations.
18.2.1.4 Staffing
On-mine laboratories are staffed by plant metallurgists as well as laboratory
assistants and labourers. The person in charge is usually appointed on a per-
manent basis, but employment of other staff is much more flexible, depend-
ing on needs as they occur. If operational problems were to arise, other plant
metallurgists and assistants could be allocated to help in the laboratory to
find solutions. The commissioning of new equipment can place a greater load
on a laboratory; an evaluation or sampling exercise can make similar
demands.
1019
When possible, personnel participating in training schemes can be useful-

ly employed in the laboratories. Last but not least, representatives of com-
panies who are engaged in testwork to promote their products can provide
very valuable information at little or no cost.
It would be expected that the supervisors of such laboratories would
be qualified, being either diplomates or graduates. The plant metallurgists
who make use of the facilities would be similarly qualified. Recent graduates
and diplomates in training schemes would gain valuable experience at such
on-mine facilities.
18.2.1.S Comment
Some of the advantages in carrying out on-mine metallurgical research are:
(a) Fresh ore pulp is readily available, and this is particularly important
where flotation is involved, since ageing alters flotation response.
(b) A quick response to operational demands and better communication
are possible.
( c) Testwork is conducted under actual conditions relative to water purity.
Disadvantages are:
(a) The central laboratory is sometimes seen as a competitor, and on-plant
metallurgists may sometimes be diverted to the on-mine laboratory to
perform long-term research. This may lead to deteriorating plant oper-
ation and a lack of attention to urgent plant problems.
(b) The effectiveness of group resources in technical manpower and equip-
ment may be diluted.
18.2.2 Group la~oratories

It was pointed out in the introduction that five of the six large mining houses
in South Africa find it worthwhile to finance their own group laboratories
to carry out not only research investigations, but also testwork of a routine
nature to back up metallurgical operations at their mines. The confidentiali-
ty that is possible in in-house projects is an attractive feature of this system.
The group laboratory provides a pool of expertise, made up of graduate
metallurgists, chemical engineers, chemists and mineralogists as well as
diplomates in some or all of these fields. Mineralogists provide a service to
mining and prospecting as well. The combined knowledge in group
laboratories of the metallurgical operations in each group has been built up
over many years, as the staff turnover is usually less in the stable conditions
that exist in these organizations than in metallurgical plants where transfers
are common. The staff at group laboratories are usually part of head office
staffs, so that the advantages that are consequently enjoyed make for this
stability.
In addition to graduates and diplomates who lead the project teams and
who direct work in metallurgy, analytical chemistry and mineralogy, these
laboratories employ matriculated laboratory assistants and technicians as well
as tradesmen, administrative staff and labourers. A great emphasis is placed
1020
LABORATORIES AND PILOT PLANTS
on training, and matriculants are encouraged to study further to improve

their qualifications. The director or manager of a group laboratory usually
reports directly to the Chief Consulting Metallurgist, and ideally, regular
meetings are held with him together with those consultants and assistant con-
sultants interested in particular projects. Metallurgical managers and other
senior plant personnel would attend these meetings as well. Project research
and testwork is the subject of regular reporting, when the overall status of
the work that is being undertaken is featured at regular intervals in sum-
marized form. Technical aspects are not the only reported activity; it is ob-
viously necessary to maintain strict financial control, and the finance depart-
ments at the head offices are regularly advised of the state of annual budgets.
The way in which a group laboratory is financed is a vexed problem.
Costs are usually shared between the group mines and head office. In South
African tax law there is an incentive to carry out research and development,
so that the direct cost of operating a group laboratory is considerably less
than that indicated by its annual budget. A workable financing method is
to ensure that the operations in the group that make use of the facilities of
a group laboratory contribute in a reasonable way, perhaps on the basis of
before-tax profit. It would then be up to each mine to ensure that it gets
its proper share of the group laboratory's output in proportion to its con-
tribution. Priorities would be decided by the Chief Consulting Metallurgist,
particularly if a number of metals or minerals were mined in the group. Alter-
natively, it could be the responsibility of the group laboratory to promote
its interests by demonstrating success in certain projects. However, by the
very nature of research, some projects will fail. For this reason, the good-
will of the sponsors must be maintained by service work which may be of
a routine nature, but best provided by a central facility.
The titles of the different sections that make up the laboratory vary from
group to group, but in most cases they accord with the disciplines of ore-
dressing, hydrometallurgy, analytical chemistry and mineralogy. Pyro-
metallurgy is important at some of the group laboratories but is not included
here since, in relation to gold, it is not the subject of any significant research
effort at the present time. Obviously, staff are involved in administration,
and workshops and maintenance are essential. Without exception, pilot-plant
facilities exist at each location, their extent depending mainly on current and
previous work. A further facility which is common is a library of text books
and journals. It is most useful to researchers and innovators to have selected
journals circulated on a regular basis so that they are able to maintain their
current awareness. Attendance at conferences and symposia is another ex-
cellent method of maintaining awareness, local meetings usually being well
attended by group laboratory personnel.
In addition to the group laboratories, Mintek provides a very impor-
tant input into research in the gold mining industry. Mintek will be discuss-
ed separately in Section 18.3.
In dealing with on-mine laboratories, layout, equipment, functions and
staffing have been discussed. The same categories are used for group
laboratories in the sub-sections which follow.
1021
18.2.2.1 Layout
All of the group laboratories have been in existence for a considerable time,
and several make use of defunct mine offices or adapted buildings. Without
exception, additions have been carried out to the buildings over the years,
and there is considerable variation in layout. All of them have separate
metallurgical testing facilities which involve hydrometallurgy as well as ore-
dressing. Support is provided by analytical and mineralogical sections.
Classical fire assaying methods for gold and silver are still used by the
assayers, but increasing use is made of instrumental methods such as atomic
absorption. The mineralogical sections of several of the group laboratories
have advanced microprobes which are useful in studying aspects of gold
metallurgy.
Pilot-plant facilities are common to all of the group laboratories, and
will be discussed in a later section. Provision for crushing and milling of
samples exists, and certain of the larger laboratories have well-stocked
libraries, while inter-library loans involving the Johannesburg Public Library,
Mintek and the Council for Scientific and Industrial Research (CSIR) are
common among all of them. At several of the group laboratories canteens
supply meals, usually at subsidized rates.
18.2.2.2 Equipment
The equipment that a well-equipped group laboratory has is discussed in this
section. Obviously, since this chapter deals specifically with metallurgical
testwork, only the equipment used for ore dressing and in hydrometallurgy
need be listed.
All of the equipment listed in Section 18.2.1.2 for a typical on-mine
laboratory will be found at a group Metallurgical Laboratory. In addition,
the following items are included, in view of the greater capability that is
required:
Ore dressing
Cone crusher or roll crusher.
Pulverizer.
Sieve shakers.
Sample splitters including spinning rifflers.
Cyclosizer.
Infrasizer.
Laser beam particle sizer.
Flotation machines.
Magnetic separators, wet and dry.
Jig for gravity separations.
Shaking table.
Corduroy table.
Superpanner.
Heavy media cone_.
Rod mills on rolls.
Bond Index mill.
Flotation reagents and chemicals.
1022
Dust extraction equipment.

Binocular microscope.
Centrifuge.
Pycnometer for density determination.
Desk-top computer terminal.
Hydrometa!lurgy
Constant temperature baths with rolls.
W aterbaths.
Autoclaves.
CIP test rig.
Pumps, metering pumps, valves, solenoid valves.
Rectifier.
Electrowinning cells.
Shaking tables for flasks.
Columns for carbon/resin loading and elution.
Sample drying facilities.
A glassed-steel reactor (Pfaudler type).
Columns for heap leaching simulation.
Stirred tanks and storage vessels.
Pachucas.
Chart recorders.
Combined magnetic stirrers/hot plates.
An extended range of glassware including ground glass jointed
apparatus.
Mineralogy
The mineralogical section at a group laboratory provides a service
for the group mines, geological exploration and the laboratory itself.
The equipment required is listed below:
Diamond saw.
Sample splitting device.
Grinding equipment.
Diamond polishing equipment.
Small superpanner.
Heavy liquid separation apparatus.
Frantz isodynamic separator.
Binocular and ore microscopes.
Modern photographic equipment.
Carbon-coating equipment.
Electron microprobe.
X-ray diffraction apparatus.
X-ray fluorescence apparatus.
18.2.2.3 Functions
Ore dressing
The work of an ore dressing section at a group laboratory deals routinely
with testwork to support operations at the group mines as well as work of
1023
a more investigative nature.

The list below indicates the sort of work that is performed:
Sample preparation, large and small, for further testwork and chemical
analyses.
Comminution testwork, Bond Work-Index determinations, relative
grindability tests, liberation tests using rod and ball mills and drop tests for
autogenous and semi-autogenous milling.
Gravity separation testwork using shaking tables, jigs, spiral concen-
trators, duplex table, Haultain superpanner.
Sizing and classification using screens, sieves, the cyclosizer, water and
air elutriation, laser sizing devices, hydrocyclones.
Flotation testwork, rougher flotation, cleaner flotation and recleaner
flotation using a variety of laboratory-scale flotation machines.
Evaluation of flotation reagents such as activators, collectors,
depressants and frothers.
Development of ne'Y flotation processes, making use of pH effects, and
study of zeta potential.
Magnetic separation testwork using wet and dry magnetic separators,
permanent magnets, electromagnets, WHIMS, Jones magnetic separators,
Carpco separators, magnetic filters, Franz separators.
Sink-float test using the Eriksen cone.
Pilot-plant operations on a continuous and semi-continuous basis, with
grinding mills, classifiers, flotation, magnetic separation, gravity separation,
etc.
Thickening and filtration studies for sizing and design of equipment,
and testing of flocculants.
Improvements to existing ore-dressing flowsheets, and the design of new
flowsheets.
Testing of new ores and ore-dressing techniques.
Hydrometallurgy
Typical investigations in a department dealing with hydrometallurgy at a
group laboratory will be as follows:
The cyanidation response of ore from new mines and prospects. This
would include borehole cores prepared by the ore-dressing section.
The solving of problems at group mines as and when they arise, and
general trouble-shooting.
The effect of such parameters as fineness of grind, agitation times, and
varied reagent additions for operating plants.
The use of new reagents.
Intensive cyanidation of gravity concentrates.
Heap leaching tests in columns.
Methods for the control of water and air pollution.
Filtration testing and the evaluation of new flocculants.
Optimization of thickener performance.
Testwork on carbon-in-pulp to obtain design parameters.
Evaluation of carbon samples.
1024
Loading and elution tests on carbon.

In-stope leaching of gold.
Acidic pressure leaching of refractory gold ores to liberate locked gold.
Bioleaching with species such as Thiobacillus ferrooxidans to liberate
locked gold from sulphides.
Roasting to liberate locked gold from sulphides.
Gold extraction from solution using activated carbon in columns.
Mineralogy
The following are typical of the activities in the mineralogy section of a group
laboratory:
Rock identification and the ore minerals present from samples provid-
ed by exploration geologists.
Establishing modes of occurrence of valuable constituents.
Determining likely liberation characteristics.
Performing microprobe analysis of gold and uranium minerals.
Examination of plant products, for example, leach residues and carbon
from CIP circuits.
Establishing the reasons for losses in gold plants.
Monitoring various beneficiation processes.
Assisting mine geologists.
Monitoring borehole intersections.
Studying reef channel samples.
Identification of likely cyanicides.
In most group laboratories, the mineralogical section is not only con-
cerned with the occurrence of gold, but with many other elements as well.
18.2.2.4 Staffing
The group laboratory represents the store of metallurgical expertise within
that group, so that the senior staff are of necessity carefully chosen. The
laboratory director has probably graduated in one of the disciplines that have
been discussed, while those in charge of the different sections have degrees
in those particular fields.
The laboratory staff are almost invariably attached to the group head of-
fice, and the advantages that result from this association mean that turnover
is minimal. This is a very important requirement if continuity is to exist. To
build detailed technical knowledge of all of the operations in a group re-
quires the passage of several years, and it is therefore an expensive procedure.
Minimal staff turnover, however, can mean that fresh ideas are not introduced
with new people. The staff in the different sections are made up of graduates,
diplomates and matriculants, supported by less well-educated labourers and
assistants in the lower grades. Administrative facilities are essential, and must
include a team operating word processors to produce the technical reports
which, after all, are the tangible evidence of the research that has been done.
The availability of workshop and maintenance services is also very necessary.
The ever-increasing role that desk-top computers are playing in
laboratory research has revolutionized experimental design and the process-
1025
ing of data in metallurgical research. Since large masses of data can now
be rapidly processed, the interaction of parameters involved in extractive
metallurgical processes can be determined with increased accuracy. Person-
nel at the group laboratories in the future will probably include staff skilled
in mathematics, statisticians and an increasing proportion of computer
programmers. The biggest section at a group laboratory is usually the
analytical chemistry section but its activities are outside the present discussion.
18.3 Mintek
In the sections that follow, the early history and the present capability of
the Council for Mineral Technology to contribute specifically to gold
metallurgy in South Africa will be discussed. Much of this information has
been provided by Mintek's Information and Liaison section (S.A. Allison,
personal communication).
18.3.1 Introduction
The identification of mineral resources in South Africa had advanced suffi-
ciently in 1934 for the state to establish the Minerals Research Laboratory
with the object of making exploitation possible. Even at that time, several
metallurgical laboratories were operated by the mining houses and devoted
to those projects likely to yield profits in the medium to short term. The results
of the testwork, however, remained confidential, and the thought was that
work in the national interest should be undertaken by a state-financed research
organization. A state-financed institution is also able to undertake work of
a long-term nature aimed at the fundamental understanding of. economical-
ly beneficial new processes or modifications to existing processes.
In 1934 the research complement of the Minerals Research Laboratory
was two, and it was located in the Department of Metallurgy at the Univer-
sity of the Witwatersrand. Ten years later the organization had its own
premises and had its name changed to the Government Metallurgical
Laboratory. By an Act of Parliament, it developed into the National Institute
for Metallurgy, which earned a world-wide reputation for its practical ap-
proach to extractive metallurgy. The current site was established in 1974,
and in 1981 a further Act of Parliament reshaped the body into the Council
for Mineral Technology, which has at the present time an annual budget of
some 25 million rands, and employs about 350 engineers, scientists and techni-
cians, and 450 people in supportive roles, largely concerned with
administration.
The phenomenon of the occurrence of South Africa's mineral wealth
within a relatively small geographical area allowed a very large complex to
be built on a single site, in contrast to the tendency overseas to decentraliza-
tion. The U.S. Bureau of Mines' scattered locations are a good example of
this. The location of the group laboratories all within easy reach of Mintek
is particularly useful, since the Mintek library is very comprehensive and pro-
bably one of the best in the world. In addition, the comprehensive facilities
and the expertise available are also very beneficial.
1026
The relationship between the gold mining industry and Mintek is ex-
cellent, and many cooperative efforts have enabled Mintek to avoid duplica-
tion of the research efforts of the various group laboratories, and vice versa.
Today, Mintek defines as its primary aim the promotion of mineral
technology, including research into the properties, composition, recovery,
extraction, processing, refining and utilization of minerals and mineral pro-
ducts. Each of the divisions at Mintek specializes in a particular branch of
mineral science and technology. Most of the research on the metallurgy of
gold is conducted within the Hydrometallurgy and the Mineral and Process
Chemistry divisions, and where appropriate, in the Ore Dressing division.
The objectives in research on gold metallurgy as stated in the most recent
Mintek report (1984) are as follows:
(1) To increase the recovery of gold (and associated elements), and to reduce
the operating costs of existing gold plants.
(2) To determine the occurrence of gold (and associated elements) in ores
and processing products so that their behaviour during processing can
be characterized.
(3) To investigate the development of less conventional flowsheets, with
special consideration of such techniques as carbon-in-pulp, resin-in-
pulp, the use of carbon and resin in continuous adsorption columns,
run-of-mine milling, gravity concentration, flotation, and wet high-
intensity magnetic separation.
(4) To investigate the economic recovery of gold (and associated elements)
from sand dumps and slimes dams.
(5) To develop instrumentation, on-line analytical techniques, and control
strategies for ore-dressing and hydrometallurgical processes.
(6) To develop improved extraction processes for the recovery of gold from
refractory gold ores, with special emphasis on improved gold recoveries
and pollution considerations.
(7) To investigate improvements to existing processes and the development
of alternative processes for the refining of gold and silver.
(8) To study the techno-economics of gold production, both primary (from
mines) and secondary (from dumps), with the object of developing com-
puter models to assist in the comparison of processes and the selection
of optimal flowsheet configurations.
The Council of Mintek is advised by two specially established commit-
tees. One is the Technical Advisory Committee and the other the Gold and
Uranium Research Advisory Committee; both are made up of senior members
of the mining industry as well as members of university staff.
18.3.2 Categories of sponsorship and the conditions pertaining to them

The categories in which projects can be undertaken are as follows:
Category A - Projects of national importance fully paid for by the
State.
Category B Partially sponsored projects.
Category C Fully sponsored projects.
1027
Table 18.1. Conditions of the various Mintek project categories.
Sponsors'
Period of Costs Rights to Licence rights to use
Category confidentiality recoverable inventions conditions expertise
A At Mintek's Nil Belong to At Mintek's

discretion Mintek discretion
B 12 months By arrange- Mintek's but Free to Free:
ment negotiable (in sponsors with- unrestricted
advance) in RSA; else-
where may be
negotiated
c Five years Full As for B As for B As for B
but may be
extended
D At Mintek's Nil Belong to
discretion Mintek
Category D - Exchange of services with other institutions.

Although all information is confidential to Mintek; work in any of the
above categories can be classified as secret or confidential to ensure a greater
degree of confidentiality.
The conditions for the various Mintek project categories are shown in
Table 18.1.
18.3.3 The facilities at Mintek

Mintek is much larger than any of the laboratories operated by the mining
houses, and it has.extremely comprehensive facilities, not only in the form
of laboratories and pilot-plant equipment, but also support facilities. In-
vestigations in gold metallurgy are only a part of the overall picture in which
studies of many other metals and minerals are concurrently undertaken. The
technical expertise, space and equipment that are available nevertheless mean
that any conceivable project connected with gold can readily be undertaken.
The space provided for pilot plants is vast, and even the most elaborate pilot
plant would occupy only a part of one of the six enormous bays, usjng
whatever equipment necessary. The equipment available for sample prepara-
tion can also deal with any possibility. In perspective, however, the Anglo
American Research Laboratory (AARL) probably has more research
personnel working on gold metallurgy at any one time.
Mintek' s greatest contribution to progress in gold metallurgy in South
Africa has been in the development of the carbon-in-pulp (CIP) and carbon-
in-leach (CIL) processes, which have revolutionized the flowsheets of many
of South Africa's gold plants. This has led to more recent work on resin-in-
pulp methods, which have application where base metal cyanide contamina-
tion in solution is low. A great deal of development work continues at Mintek
on further refinements.
1028
Table 18.2 Typical research programme steps.
Step Purpose
Origin of concept. Initiate ideas and concepts to be investigated.

Basic research. Determine basic feasibility of the project.
Preliminary economic evaluation. Determine if economic incentive is sufficient to pro-
ceed with the project.
Laboratory research. Develop basic data.
Product evaluation Evaluate the suitability of product.
Process development Produce preliminary design of a full-scale plant.
and preliminary engineering.
Pilot-plant study. Prove basic reliability of the proposed
Demonstration unit. process and demonstrate feasibility.
Full-scale plant. Full-scale production.
The equipment that is available at Mintek is probably greater in extent

and variety than that in any group laboratory, particularly in the case of
the smaller ones, so that joint pilot-scale investigations are sometimes under-
taken at Mintek. The general scope of the research undertaken is covered
by the list relating to group laboratories in Section 18.2.2.3 above, and
Mintek' s staffing arrangements are covered by the requirements in Section
18.3.1.
18.4 Pilot Plants

18.4.1 Reasons for piloting
It was observed in Section lS.l that the main reasons for using a pilot plant
to test a new process or unit operation are.to limit the risks that are inherent,
and also to be able to operate continuously and in closed circuit, rather than
batchwise as in most bench-scale testwork. These observations are considered
worthy of repetition since they are the basic reasons for building a pilot plant.
In a novel process, as in carbon-in-pulp in the South African context,
the design and construction of a full-scale plant would represent a very large
investment in both time and money. While the rewards can be high in the
pioneering of a novel process route, the risk is a factor to be considered very
carefully. The operation of a pilot plant definitely reduces such risk.
The sequence of steps that should be followed in a research programme
was listed by Van Sweringen (1982). The list in Table 18.2 shows an adapta-
tion of that sequence.
It must be stressed that the basic concept must have been proved before
a pilot plant is built. The tendency to make adaptations to the concept during
the erection of the plant should be avoided, unless any suggested improve-
ment can be justified by laboratory testwork. The aims of pilot-scale testwork,
other than to reduce risk and provide the opportunity to operate continuously
in closed circuit, include:
1029
Testing materials of construction.

Estimation of maintenance requirements.
Assessing hazards that may exist, e.g. in the use of chlorine as a sol-
vent for gold.
The solution of problems relating to waste disposal.
Training operating staff.
Testing different control routines.
Obtaining data for the design of the full-scale plant.
Feasibility and cost studies.
Demonstrating to senior management that a concept is viable.
The production of by-products other than gold.
18.4.2 Types of pilot plants

The most obvious variable in pilot-plant classification is its size, which can
vary from a bench top to a prototype unit. Obviously, the costs and the time
that are involved are proportional to the size of the unit. The complexity
of the process usually defines the size that is required to test its feasibility.
Since metallurgical operations within the gold mining industry are com-
paratively simple when compared to, say, the production of synfuels, it is
to be expected that pilot plants will generally fall within the smaller size range.
However, the fact that there is a lower limit to the size of metallurgical equip-
ment such as, for example, mills or flotation machines, imposes a limit on
how small a pilot plant can be.
The amount of instrumentation installed in a pilot plant also has a bear-
ing on costs. In the simplest type of pilot plant, data gathering is usually
by manual means and the operators are required to optimize the operating
conditions and to set them manually by means of control valves and the like.
Some form of automatic data capture in the form of recording instruments
is normally also used. Capital costs increase with increasing complexity in
control philosophy. Since sophisticated computer control of plants is as yet
relatively undeveloped in the gold mining industry, proportionately less money
is required for a hydrometallurgical leach plant. The situation could, however,
rapidly change with the availability of such systems as Proscon and Centum.
18.4.3 Scale-up factors

Pilot-plant design with the eventual goal of building a full-scale plant should
use commercially available equipment and not place reliance on makeshift
apparatus. The use of commercial equipment ensures that in the pilot-scale
operation the same effects, such as short-circuiting in reactors, will be found
as in the commercial plant. The transfer of a novel project from pilot-scale
to full-scale conditions is the period of greatest risk, as large sums of money
are then required. The risk lies in unknown factors, chemical or physical,
which are different in the two situations. Typical of this situation was the
development of the carbon-in-leach process.
In the case of conventional gold plants, however, experience gained over
many years allows of design after only basic laboratory testing has been
performed.
1030
Good design at the pilot scale ensures that the scale-up factor can be
high, much higher than the generally accepted statement in previous years
that a scale-up factor should not be higher than 10. Pilot-scale carbon-in-
pulp rigs at Grootvlei have recently been translated to full-scale operations
where the scale-up factor was about 1000 (P.A. Laxen, personal communi-
cation).
18.4.4 Design of pilot plants

The specific purpose for which the pilot plant is intended must first be defined.
If this is not done, considerable modification during its operation is likely.
The development of a mathematical model when unit operations at reduced
scale are used allows high scale-up factors and reproduces process variables.
It also reduces the time required for the engineering .design. Sometimes,
however, equipment must be designed to overcome a particular problem when
novel concepts are being investigated. Mintek' s screens for use in CIP cir-
cuits are an example of this, as are the electrowinning cells used on eluates
derived from loaded carbon.
When a pilot plant is to be constructed, those items with long delivery
times should be ordered as soon as the decision to go ahead has been taken.
CAD (computer-aided design) systems will greatly speed up pilot-plant design
and allow metallurgical flowsheets to be constructed rapidly and accurately.
18.4.5 Operation of pilot plants

The operating cost of a pilot plant is directly related to the time period over
which the plant is operated, so that it is very necessary to decide on a pro-
gramme to meet the objectives as part of the design of the plant. The results
obtained in the laboratory during the earlier part of the investigation are us-
ed to define operating conditions. If operating problems develop at the pilot
stage, it is usually less costly to return to the laboratory to overcome them,
rather than persevere with pilot-scale work.
Statistical planning of pilot-plant operation as a function of the effects
of variable parameters by using previous laboratory data can save much time
and provide more meaningful results.
It is usually necessary to operate a pilot plant over a wider range of con-
ditions than would be the case at full scale in order to define the effect of
variables on the process.
Apart from establishing the viability of the process that is under study,
the operation of a pilot plant is used to provide information on operating
costs, which can be used in a feasibility study. Other benefits relate to the
training of operators for full-scale operation, and those concerned with plant
maintenance, as well as the establishment of safety procedures. It is to be
expected that much more detailed information will be gathered in pilot-plant
operation than on an operating plant, so that the sampling and monitoring
procedures would be more intensive.
18.4.6 Pilot plant location

The operation of a pilot plant very often requires a continuous supply of
1031
feed material which can best be supplied at an operating plant. Carbon-in-

pulp or carbon-in-leach processe~ are examples of these, and pilot facilities
are most appropriately established where 2, continuous source of milled pulp
is available. Pilot plants using batch processing procedures, or where the
technology is more complicated, are preferably carried out at a research
laboratory. Pressure or bacterial leaching comes to mind in this instance,
as does investigational work on new ore bodies where no continuous supply
of feed material is possible.
In continuous operation, the disposal of wastes is a problem more easi-
ly dealt with at a mine, where slimes dams are available. In a 'greenfields'
operation this facility would not be present. Operating a pilot plant at a mine
also means that the transport of feed material is avoided, as are the legal
aspects, including the necessity for obtaining police permission for transport
of gold-bearing material, particularly when significant values are present,
as in gold-rich flotation concentrates. The problem of oxidation of pyritic
concentrates while transporting them over long distances is also overcome
by immediate processing at an on-mine pilot plant.
A concept which is receiving attention at the present time is that of mobile
pilot plants, mainly for ore-dressing operations. A budget estimate in 1985
for a 10 ton/hour mobile pilot plant consisting of a crushing and screening
unit, a milling and classification unit, as as well as a flotation unit, was Rl ,4
million (J. Dear, personal communication).
18.4. 7 Type of investigation carried out in pilot plants

Some of the aspects of gold metallurgy that have been studied in recent years
at pilot scale are listed below:
CIP and CIL processing.

Elution of gold-loaded carbon.
Elution of gold using methanol (the Micron system).
Development of electrowinning cells (the AARL, Mintek and Micron
systems).
Regeneration of carbon in kilns or in the 'Rintoul' furnace.
Loading of gold onto carbon from "regen-aqueous" solutions from
uranium solvent extraction.
Pinned bed carbon adsorption.
Continuous gold elution.
Pressure cyanidation in a pipe reactor.
Acidic pressure preleaching to liberate locked gold in refractory ores.
Bacterial leaching with Thiobacillusferrooxidans to liberate locked gold
in refractory ores.
Sulphation roasting of pyrite concentrates to liberate locked gold.
Flotation of pyritic concentrates.
Run-of-mine milling.
Dump cyanidation.
Intensive cyanidation.
Chlorination to dissolve gold in refractory slags.
1 f)')....,
lVJL..
Reverse osmosis for the treatment of tailings darn water.

In-stope leaching with thiourea and thiocyanate.
18.5 References
Lenahan, W.C. and Murray-Smith, R. (1986). A&Say and Analytical Practice in the South African
-Mining Industry. Chamber of Mines of South Africa, Johannesburg.
Van Sweringen, R. (1982). Encyclopedia of Chemical Technology, 3rd edn. 17, 890. John Wiley,
New York.
1033
APPENDICES
In most laboratories in the gold mining industry standard procedures for
routine metallurgical testing are in use. An edited selection of these follows.
The permission of the Joy Manufacturing Company to use a publication
entitled 'Determining Thickener Capacities', by H.J. Gisler, is gratefully
acknowledged.
APPENDIX 18.1
Determination of Relative Grindability using a Laboratory
Batch Rod Mill
1. Introduction
( 1) Given a mill feed charge of acceptable size, the amount of milled product
passing a stated fine size increases linearly with time up to about 90%
passing.
(2) If x = milling time,
y = mass of product passing stated size expressed as percent of feed
mass,
c original mass percent of feed passing stated size,
r milling rate = mass ground through stated size per unit time,
f feed mass,
100 rx
then y = --f- + c
(3) The Milling Curve, which is a plot of y against x, is a useful tool in com-
paring grindabilities of different ore bodies and therefore the power re-
quirements and costs of milling.
(4) Laboratory batch milling rates are commonly expressed in units of g/min
of -75µm.
2. Procedure
(1) Reduce representative 10 kg sample of ore to - 3 mm using jaw crusher
and roll crusher.
(2) Using a stationary or rotary splitting technique, obtain at least five
representative samples each weighing 1 kg (plus/minus 1 g). Determine
per cent - 75 µm according to the standard dry-screening method.
(3) Charge a stainless steel batch rod mill as follows:
1. To the clean, dry 200 mm d x 200 mm mill, add 1 kg sample of
- 3 mm material.
2. Add 540 ml of domestic tap water (equivalent to 35 percent moisture
1000 ~ 540
5
since x 100 = 35)
3. Add suitable grinding rods to half fill the mill. It is essential that
1034
the same mass and number of rods are always used.

4. Close and secure the mill.
(4) Place the mill on rollers and roll at 75 r.p.m. for 5 minutes.
(5) Remove the mill from rollers. Open the mill carefully and wash all solids
adhering to the lid down into the mill.
(6) Remove rods singly, carefully washing all solids from each rod down
into the mill. Use the minimum amount of wash water.
(7) Using a wooden frame, support a clean 75 µm deep-frame screen in a
clean, empty bucket. Slowly and carefully wet screen the total contents
of the mill on this screen, ensuring that no solids are left and retaining
oversize and undersize separately. Undersize may be discarded if no
subsequent mineralogical investigations or ore treatment steps are
necessary and the sample is plentiful.
(8) Filter oversize carefully to avoid loss and (a) oven dry if no mineralogical
examination is required or (b) dry in the open-air in a sheltered posi-
tion. Weigh oversize (plus/ minus 1 g).
(9) Repeat steps 3 to 8 for milling times of 10, 15, 20 and 30 minutes. Con-
vert masses of oversize to percent of original feed by dividing by 10.
(10) Draw a milling curve as shown:
100
y = Percentage y = rx + c
passing 75 µm
..
50
x = milling time (min)
Figure 18.3. Graph of milling time versus percentage passing 75 Jlm.
3. Notes
(1) The milling rate or relative grindability remains constant according to
the equation y = rx + c. If different samples are roll crushed under slight-
ly different conditions to produce different amounts of - 75 µm in the
- 3 mm fe,ed, then c varies in the above equation. So whilst r remains
constant, different values of y (percent - 75 µm in product) will be ob-
tained for a fixed time x.
(2) If sample size is limited to say 1 kg, both oversize and undersize from
first milling may be recombined (dry) afth weighing oversize and remilling
four more times. However, errors may be introduced through repeated
filtration and recombining operations.
1035
APPENDIX 18.2
Standard Procedure for Routine Size Analysis of Finish-ground
Products
1. Introduction
Total grading analysis error is an accumulation of errors from various in-
dependent sources such as sampling, sample preparation, screen mesh ac-
curacy and repeatability of screening technique.
To reduce this error the following conditions must be observed:
(i) Sample preparation must be done meticulously.
(ii) Good quality screens must be used and be maintained in good condition.
(iii) Every operation in the standard grading analysis procedure must be
carried out, and the taking of shortcuts must be avoided.
2. Equipment required
1. Balance of at least 200 g capacity and resolution to 0,1 g (an electronic
balance with auto-zero facility will reduce weighing time considerably).
2. Spatula with 200 mm to 250 mm blade.
3. Deep aluminium, enamelled or stainless steel dishes, 200 mm diameter.
4. Paint brush, 50 mm. Bristles must not be cut short.
5. Water spray with steady head arrangement.
6. Water hose connection with 6 mm tube.
7. A 63 µm aperture full height test sieve of 200 mm diameter for wet screen-
ing. This sieve should be in good condition and be regularly inspected
for holes, tears, etc.
8. Sets comprising one standardized 75 µm and one 150 µm diameter test
sieves for dry screening. (The number of sets required will be dictated
by the needs of the mine.) These sieves should be kept in immaculate con-
dition and used for monthly composite, residue, cyclone O/F and thickener
U/F samples only. (See screen calibration diagram.) Each set should be
clearly marked with an identification number.
9. A laboratory sieve shaker.
3. Grading procedures
1. Sample preparation
1.1 Each shift must submit dry samples of residue, cyclone overflow and
thickener underflow of at least 1 kg each from each plant unit to the
grading room for sieve analysis. Figure 18.4 shows the residue sample
preparation procedure; other samples need only be dried before sub:
mission to the grader.
1.2 Each entire sample must be deagglomerated by gently breaking it through
a 1,0 mm aperture screen onto a 1 m x 1 m rubber mat to break up lumps
formed during drying.
1.3 Coarse material remaining on the sieve after all the lumps have been
disintegrated must be combined with the - 1,0 mm portion.
1036
Residue samples (3 Shifts)
i
Cuprous chloride added and stirred
(if T.G. values are being determined)
t
Filtered by shift concerned on filters at the sampling point
(if U.G. values are being determined, repulp with water and filter six cycles)
t
Total sample dried in stainless steel pans in ovens heated with infrared
elements
t
Deagglomerated through 1,7 mm screen
i
Rolled at least 10 times and 2 x 1 kg samples taken by dip sampling
One sample to assay office One sample for grading analysis

for T.G. analysis (or U.G. (1 OOO g)
where applicable)
i
Deagglomerate (1 mm)
t
Grade according to standard
procedures
t
Report
T.G. = Total gold

U.G. = Undissolved gold
Figure 18.4. Schematic diagram of residue sample preparation procedure.
1.4 Roll the sample on the rubber mat from alternate corners at least ten
times.
1.5 With the spatula, spread the heap to an even depth of about 15 mm over
the mat.
1.6 Sample 100,0 g from the material on the mat with a small spatula or tea-
spoon by taking small dips over the entire area of the flattened material.
(If more 100 g lots are required, the sample should be re-rolled, again
spread out and sampled as before.)
2. Wet screening
2.1 Place the sample prepared as above in an aluminium or enamelled dish
and add approximately 150 ml of water.
2.2 Play a gentle spray of water onto the pulp until the dish is about half
full. Slowly decant the slime through the 63 µm aperture sieve reserved
for wet screening. The passage of slime through the screen can be assisted
by gently tapping the side of the screen.
Repeat the procedure until most of the slime has been transferred
1037
to the screen. Finally, wash all the material from the dish onto the screen
using the rubber hose.
2.3 By holding the screen under the spray in one hand, wash the material
remaining on the screen in a rotary action whilst tapping gently on the
side of the screen with the other hand.
2.4 Wash the solids on the screen back into the dish using the rubber hose.
Allow the solids to settle and drain off the excess water. Dry slowly to
avoid spitting.
3. Dry screening
3 .1 When the sample is dry, remove it from the drier to cool.
3.2 Clean the appropriate standardized 75 µm and 150 µm screens using a
soft paint brush. Brush the bottom of the screen cloth with the screen
inverted on the table, then brush the top and finish by brushing off the
bottom.
3 .3 Arrange the nest of screens with the 150 µm screen on top and a pan
at the bottom and transfer the sample from the dish to the screens. Fit
the nest of screens tightly in the sieve shaker so that no lost motion oc-
curs after the motor is started.
3.4 Shake for 10 minutes plus/minus 10 seconds. (Check the time switch
on the shaker with a stopwatch and mark the clock face to give a run
of exactly 10 minutes.)
4. Weighing
4.1 When the shaker stops, remove the nest of screens. Invert each screen
on top of a large clean dish.
4.2 Brush the bottom of the screen gently using a soft paint brush covering
the whole area of the screen.
4.3 Holding the inverted screen about 100 mm above the dish gently tap the
side of the screen frame.
4.4 Transfer the material from the dish to the balance pan and weigh to the
nearest 0, 1 g.
5. Reporting
5 .1 Each set of standardized screens has factors which must be added to the
oversizes to obtain the corrected oversize mass on each screen.
The screen sets must not be separated but always used together.
For example:
In set No. 1 the 150 µm sieve has a factor of + 0,3 and the 75 µm
sieve has a factor of _.:._ 0,2.
Results are calculated as shown in Table 18.3.
5 .2 To standardize a set of screens the average of jive replicates is compared
with the average of five replicates obtained from the same sample using
a set of sub-master screens. The sub-master screens are used only for
standardizing and are checked at intervals against a sample of known
size distribution to give the correction factors shown in column (2) of
Table 18.4.
1038
Table 18.3. Application of screen correction factors.
Screen Aperture Actual mass Correction Cale. mass

no. (µm) oversize (g) factor oversize (g)
IA 150 4,1 +0,3 4,4*

1B 75 20,3 -0,2 20,1*
*Values to be reported.
5 .3 The grader would report the corrected, not the actual, weights. Alter-
natively, to minimize the possibility of using wrong screen factors, the
metallurgical clerk could have a list of the screen set numbers and their
corresponding factors. The grader would then merely report the over-
size weights and the screen set number and the clerk would make the
required adjustments.
5 .4 Calculation and screen calibration should be checked on a regular basis
by senior personnel. Figure 18.5 shows a scheme used by one South
African mining group in the standardization of its screens.
Table 18.4. Calculation of screen correction factors.
Sub-master set Working set
(1) (2) (3) (4) (5) (6)

Average Corrected Average Cale. Corrected
over- Correc- over- over- correc- over-
size tion size size tion size
Aperture (g) factor (g) (g) factor (g)
(µm) (1) (2) (3) (4) (5) (6)
150 4,5 +0,3 4,8 4,4 +0,4 4,8

75 21,3 nil 21,3 21,6 -0,3 21,3
(1) and (4) are the averages of five replicates in each case.
(3) = (1) + (2)
(6) = (3)
(5) = required correction factor = (6) - (4).
1039
Group Laboratory Master Screen Set (M) (kept in safe at Laboratory)
t
Lab. Sub-Master Set (Mi) (kept by Head of Mstallurgy Section at Laboratory)
I
Mi used to calibrate
j
Calibrates mine Assay
l
Calibrates mine Master
working master screen Office Master Screen kept by Asst.
sets for the Screen AOl Plant Superintendent
laboratories (MMl)
(SMl to SM3)
!
Calibrates needed
i
Calibrates monthly
number of working composite master set
sub-master sets (MM2)
Al to Ax
i
Calibrates working
sub-master sets:
SI - Residue samples
S2 Cyclone O/F and
Thickener VF/samples
Residue sample screens should be renewed monthly, the new set being used
for residue samples and the old residue screen moved down one position.
Residue screens should be clearly marked.
Figure 18.5. Screen calibration scheme.
1040
APPENDIX 18.3
Determination of Thickener Area
1. Introduction
The thickener capacity required to handle a predetermined tonnage of acer-
tain pulp, by overflowing a clear solution and obtaining the desired pulp den-
sity of thickener discharge, depends upon the settling rate of the solids in
that particular pulp. The settling rate may be determined by simple laboratory
tests using measuring cylinders and can then be used in a simple formula
for calculating the necessary thickener area.
2. Procedure
The procedure used by the Joy Manufacturing Company, as reported by H.J.
Gisler, is as follows:
Place a measured quantity of pulp at a known density in a beaker or
glass cylinder. Fix a narrow strip of paper on one side of the container. Mix
pulp thoroughly. Draw a line on the paper at the top of the pulp and mark
"O" minutes. For five minutes, at one-minute intervals, mark the point to
which the solids have settled. Graph distance settled versus time. The settling
rate R is the slope of the straight portion (i.e. the pre-compression portion)
of the graph. This determines the free settling rate of the solids at the initial
density.
Usually readings are taken at three different densities of the pulp
corresponding approximately to densities which will exist in the various zones
in the thickener.
Decant sufficient clear water or solution to establish a pulp with inter-
mediate density. For instance, if initial pulp densify was 4: 1, water to solids,
the removal of one-fourth of the water would establish a density of 3: 1. Mix
thoroughly. Repeat readings of settlement as above.
Then decant again to obtain a pulp at one-third density. The pulp just
tested was at 3:1 dilution, so decanting one-third of the water will give 2:1
dilution, water to solids. Mix thoroughly. Repeat settling measurements at
one-minute intervals for five minutes.
The settling rate per minute should be uniform during the testing at each
dilution, until compression is reached, at which time the amount of settling
will decrease during each succeeding minute. Measure the pre-compression
settling marks in millimetres, thus determining the settling rate in millimetres
per minute for each pulp density, and convert this to millimetres per hour.
The distance settled versus time graphs for all dilutions tested should be in-
cluded in the test report.
Determining final density
Final density is then determined. Thoroughly mix the pulp remaining after
the test at 2: 1 dilution and allow to settle for 19 hours. Mark the position
of settled pulp and let stand for a few hours to see if final density was reached.
If pulp continues to settle, mark its position at hourly intervals until settling
stops. Decant off all clear water or solution. Then determine moisture con-
1041
4:1 3:1 2:1

dilution dilution dilution
.,.•· ..
•;
....
.....
..
'
,
·: , ·..
......
.•.
Figure 18.6. The measurement of slime settling rates.
tent and density of pulp by weighing and drying.

Calculating thickener area
Thickener area required is then calculated by applying the above-determined
data in the following formula:
41 67 (F-D)
A = ' R (Coe and Clevenger)
A Thickener area in square metres per metric ton of dry solids thickened
in 24 hours.
F Initial w/ s.
D Final w/ s to which pulp will settle cir w/ s at which you want to discharge
pulp from thickener. Usually it is desired to discharge pulp from the
thickener at its final w/ s as shown in the above test. However, if you
want to discharge pulp more diluted than the actual final w/ s, the w/ s
desired should be used in above formula rather than the final w/s to
which the pulp will settle.
R = Settling rate in mm per hour.
Calculations of indicated thickener area from each of the three settling rates
obtained in tests will indicate any change in settling rate in the different zones
of the thickener, and the largest area obtained from the three calculations
should be used.
Assume the following data were obtained from the above tests:
1042
/Feed
4:1 zone
R = 92
3:1 zone
2:1 zone R = 46
Compression zone
Figure 18.7. The change of settlement rate with decreasing dilution in a thickener.
At 4:1 dilution, R 153 mm per hour

At 3: 1 dilution, R 92 mm per hour
At 2: 1 dilution, R 46 mm per hour
Final density D = 1: 1.
Applying these data to above formula, one obtains:
41,67 (4-1)
A 0,82 m2/ton/24 hours.
153
41,67 (3 -1)
A 0,91 m2/ton/24 hours.
92
41,67 (2-1)
A= 0,91 m2/ton/24 hours.
15
Factor of safety
Normally a 25 per cent factor of safety is allowed in determining the thickener
area. Thus, in the above case, 0,91 m2 (largest of three figures) plus 25 per
cent factor of safety, equals 1, 13 m2 area required per metric ton of dry
solids fed to the thickener per 24 hours.
If the pulp reached its final density during the 19-hour test, a standard
depth thickener is considered adequate. However, if additional time was re-
quired to reach final density, the thickener volume would have to be large
enough to retain the pulp for this extra time.
Storage capacity
Another factor to be considered is the storage capacity desired in a thickener.
For example, it may be necessary to shut a filter down for repairs and at
the same time keep the remainder of the plant in operation. In this case the
thickener will act as a storage reservoir. Where storage capacity in a thickener
is of importance such as in many of the non-metallic flotation plants, a larger
factor of safety, say 2, should be applied in calculating the area. Greater
depth should not be used unless the lift of the thickener mechanism can also
be increased.
1043
Duplicating thickener action

Many times a pulp standing without agitation will not settle to as low a final
density as when aided by movement of thickener rakes in actual operation.
The DECO Test Laboratory developed a miniature thickener mechanism
which is used in determining final density under actual operating conditions.
APPENDIX 18.4
Filtration Rate Determination
Buchner funnel
Valve A
To vacuum
Retort stand
Buchner flask
Figure 18.8. Apparatus for measuring filtration rates.
Apparatus required
(1) 500 ml Buchner flask and stopper.
(2) 120 mm dia. Buchner funnel if filtering medium is filter paper. Alter-
natively, a filter funnel 120 mm dia. x 70 mm deep with a plastic grid
base over which a filter cloth is fitted.
(3) 200 ml measuring cylinder.
(4) Vacuum hose.
(5) Vacuum gauge.
(6) Beakers.
(7) Pipette for flocculant addition.
(8) Retort stand.
(9) Stopwatch.
(10) Constant vacuum supply and valves.
Method
(1) Set up the apparatus as shown in the Figure 18.8.
(2) Prepare representative 200 ml or 250 ml samples of the pulp to be tested
in beakers.
(3) Open vacuum valve A and while keeping B closed pour a stirred sample
of pulp into the funnel, which is held vertically so that the base is
horizontal and the pulp spreads evenly over the cloth or filter paper.
(4) Open valve B; start the stopwatch and note the vacuum gauge reading.
(5) As soon as the solution disappears from the surface of the filter cake,
record the time required for the forming of the filter cake.
1044
Table 18.5. Form for reporting filtration test results.
Pulp tested:
Pulp W/S:
Filtration area of funnel: m2
Flocculant tested:
Cake Calculated
Pulp Floe. form Wash Wash Dry Total Cake Cake Mois- filter
Test vol. addtn, time, water, time, time, time, wet, dry, ture, Filtrate duty,
no. ml Vacuum git sec. ml sec. sec. sec. g g o/o ml t/m2/h t""'
>
t:O
200 0 38 38 210 155 26,2 90 0
2 200 0 38 100 20 15 73 206 154 25,2 152
~
.....,
Examples of calculations: 0
IC!
.......
(1) Filter duty (t/m2 /h) ~ x .!_ x Dry cake wt (g) tT1
(Filter area 0,01131 m2) Total time a 1000 OOO (/)
3600 x __ l_ x 155 >

z
38 0,01131 1000000 tl
1,298. ""d
......
t""'
(2) Filter duty (t!m2/h) 3600 x - - - x ~ 0.....,
73 0,01131 1000 OOO
""d
0,67.
~
~-
0
z
.....,
~::..
(/)
Vi
(6) Remove and weigh the wet filter cake.

(7) Dry the filter cake and weigh again.
(8) Measure the amount of filtrate collected.
(9) This method is used to compare the filtration rates of different pulps
or pulps with added flocculant.
If wash and drying time is to be included in the filtration tests, step
5 is altered and a standard amount of water, say 100 ml, is poured over
the filter cake when the last of the pulp solution is filtered, and the time
for this water to be filtered is measured. A standard drying time may
also be added to the total filtration time if required.
(10) When testing flocculants the desired amount is gently stirred into the
pulp before the latter is poured into the funnel.
Table 18 .5 is a suggested form for tabulating the results.
APPENDIX 18.5
Cyanidation Tests using Rolling Bottle Method
1. Sample preparation
Samples can be received as pulps, mill feeds or boreholes cores. Generally,
pulps will require no grinding and can be air dried before splitting, or split
as a pulp. Where samples require grinding, a grinding curve may be con-
structed using the procedure described in Appendix 18.1.
2. Procedure
2 .1 The normal residual cyanide and lime values after a leach test should
be about 0,020% both for lime (CaO) and cyanide (NaCN) in solution.
Values in the range of 0,015 - 0,025% are acceptable but values outside
this range indicate that the test should be repeated with a suitable ad-
justment in cyanide/lime additions to correct the residual values. Before
tests are performed, the percent active ingredients of the reagents should
be determined by titration.
2.2 Clean, dry, 5 litre cyanidation bottles are weighed and the empty mass
recorded on the bottle.
2.3 Weigh a sample and if necessary grind it at 50% solids to the required
degree of fineness. 1 g of lime is usually added to the grinding mill prior
to grinding the sample.
2.4 The ground pulp is filtered, the filtrate being used to wash out the grind-
ing mill, and the filter cake is transferred to the cyanidation bottle. Tests
are normally performed at 50% solids on 1 kg of sample and thus, the
cyanidation bottle is placed on a scale and the filtrate used to make the
pulp up to 50% solids, 2 kg of pulp being the norm.
2.5 If the sample was received as a pulp, then the required volume of p"ulp
for 50% solids by mass is transferred to the bottle, or if the pulp was
1046
dried, the required mass of sample is transferred to the bottle and pulped
with an equal mass of water.
2.6 More lime is now added to the pulp, and the bottle swirled carefully to
ensure that all solids are in suspension before carefully placing the bottle
on the bottle rolls. For 1 kg of sample the lime addition will be approx-
imately 2,5 g or if the sample was received as a pulp, 3,5 g of lime.
2.7 Cyanidation tests are usually carried out at ambient temperatures, but
can be performed in a water bath at a controlled temperature (normally
20 - 25 ° C). The pulp is first aerated on the rolls for the required period,
usually between two to six hours, the bottle being left uncorked.
2.8 The bottle is removed from the rolls after the aeration period has elapsed
and the pH of the pulp is checked. The ideal is a pH of about 11, and
if less then some lime should be added. Sodium cyanide is now added
to the pulp. As a guideline, for 1 kg of solids approximately 0,30 g of
sodium cyanide will be required, equating to approximately 15 ml of a
2 OJo sodium cyanide solution. Addition as a solution will ensure good
mixing of the cyanide throughout the pulp. Swirl the uncorked bottle
Table 18.6. Calculation of the gold recovery in a cyanidation test.
Mass of bottle + pulp at end of test = 2302, 1 g

Mass of bottle empty 295 ,3 g
Mass (g) Assay (Au g/t) Au (mg) Recovery (%)
Pulp 2006,8 4,16

Residue 997,7 0,41 409,06 4,90
Solution 1009,1 7,86 7931,53 95,10
Calculated head 997,7 8,36 8340,59 100,00
Assayed head 997,7 8,77 8749,83
% Extraction based on calculated head and washed residue = 95,10%

% Extraction.based on assayed head and washed residue = 95,32%
Grinding curves and worksheets should be appended to the report.
and place carefully on the rolls for the required cyanidation period, which
can be from sixteen to forty hours.
2.9 After the required cyanidation period has elapsed, the bottle is wiped
to remove any moisture on the exterior, and then weighed, the total mass
of the bottle and pulp being recorded. The pulp is now filtered, the filtrate
being used to wash out the cyanidation bottle. The residue is now washed
three times with tap water, dried, weighed and then assayed for gold.
The volume of the filtrate is determined, some of it is transferred to a
sample bottle, this being assayed for gold and the remainder kept for
residual cyanide and lime concentration determinations by titration. The
methods may be found in the book entitled Assay and Analytical Prac-
tice in the South African Mining Industry (Lenahan and Murray-Smith,
1986).
1047
3. Calculation of results
Table 18.6 shows the method of calculation of the percentage recovery of
gold in a cyanidation test.
APPENDIX 18.6
Muffle Roasting of Sulphide Concentrates
(1) Place the 100 g sample to be roasted in a silica dish 145 mm long, 85 mm
wide and 20 mm deep. This leaves 3 to 5 mm of freeboard above the
sample.
(2) Put the sample in the cold furnace, set thermostat at 320°C and switch
on. The furnace door must remain about 25 mm open throughout the
roasting period, to ensure an unrestricted supply of oxygen.
(3) Rabble the sample every 15 minutes. Use a rabbler consisting of2 toothed
steel plates fastened to a steel rod in order to obtain efficient mixing.
(4) Raise the temperature by about 100°C every hour until the roasting
temperature is reached. This can be done in four steps, e. g. about 25 ° C
every 15 minutes. It takes about 7,5 hours to reach 750°C.
(5) Leave the sample in the furnace at target temperature for a further two
hours, while still rabbling every 15 minutes.
(6) If the calcine is to be cyanided, it must be washed very thoroughly by
repulping. Grinding the calcine before cyanidation usually improves gold
extractions.
NOTE: If several dishes are used simultaneously in one furnace, the
dishes should be swapped, front dish going to the back, about
one hour before the end of the test, to ensure that all dishes have
a reasonable time at the target temperature, as indicated in Figure
18.9.
Calcine dishes
/~
[~J [DDJ
I I
or
DO
Figure 18.9. Rotating the positions of dishes in laboratory muffle roasting.
1048
APPENDIX 18.7
Determination of Activated Carbon Parameters
1. Determination of the equilibrium gold adsorption capacity constant

Because gold adsorption onto activated carbon is very sensitive to such
parameters as temperature, pH, cyanide concentration and to ionic strength,
any method used for measuring the equilibrium adsorption capacity of carbon
must involve the control of these parameters. The method as used at the Anglo
American Research Laboratories (AARL) for the adjudication of commer-
cial products is as follows:
(1) Riffle out approximately 10 g of the carbon sample and crush the sample
using a pestle and mortar before screening within the range 0,15 -
0,43 mm. The screened carbon is then boiled in a similar volu~e of 3
percent hydrochloric acid (vol/vol) for 10 minutes. Decant and wash the
carbon twice with equal volumes of cold distilled water before oven dry-
ing at l l0°C overnight.
(2) Prepare a borate buffer solution as follows:
Dissolve 15,46 g H 3B03 and 18,64 g KCl in approximately 4 litres of
distilled water (solution A). Dissolve 10,00 g NaOH in 0,5 litres distilled
water and add to solution A while stirring until pH 10,0 is reached (check
with commercial buffer standard). Add 0,740 g KAu(CN) 2 and make up
to 5 litres with distilled water, i.e. 100 git Au. To every 1 litre of this
bulk solution add 20 mg NaCN prior to use.
(3) Weigh out five portions of the prepared carbon as in (1) following cool-
ing in a desiccator, such that the masses are 0,10, 0,20, 0,30, 0,40 and
0,50 g respectively. The masses of each portion must be recorded to four
decimal places.
(4) Place each sample of carbon in a 250 ml conical flask. Add 100 ml of
the buffered gold solution as prepared above (2). Stopper the flasks.
Agitate the flasks in a thermostatically controlled shaking machine at
25°C for 20 hours at a medium speed.
(5) Filter off the carbon quickly using Whatman No. 541 filter paper. After
discarding the initial 10- 20 ml filtrate, determine the concentration of
gold in the filtrate in each case.
(6) Calculate the mass of gold adsorbed by the carbon by means of the dif-
ference in the gold concentration between the head solution and the
filtrate obtained from each sample. Determine the k value of the carbon
by plotting the empirical Freundlich isotherm, X/M = kc 11 ", on log-log
graph paper.
i.e. log X/M log k + l/n log c.
where X mass of gold loaded on carbon (mg)
M mass of carbon sample (g)
c equilibrium concentration of gold in filtrate (g/t)
k constant defined as the carbon loading (mg Au/ g C)
in equilibrium with a residual solution gold concentra-
1049
tion, c, of 1 g/t. This constant may be better expressed

in kg Au/tC as the expression is more meaningful when
related to a large scale CIP operation.
l/n = slope of plot as described above.
The relative standard deviation of the above method was shown to be
3 ,5 percent.
2. Determination of gold adsorption rate

As the measurement of adsorption rates is entirely empirical, many different
methods may be used provided good reproducibility is attained. In this regard
strict attention needs to be given to the particle size of the carbon sample
and to the preparation of the sample, while the control of temperature is
also necessary. The method developed at the Klipfontein Organic Products
laboratories and adapted by the AARL is as follows:
(1) Screen out 2 g of wet carbon in the size range 1,40- 1, 70 mm. Remove
carbon fines from the screened sample using ultrasonic cleaning until no
further fines are evident (several decant washings may be necessary). After
oven drying at 110° C weigh out 1,000 g carbon sample for kinetic test.
Moisten sample with 1 ml standard pH 10 borate buffer solution at least
1 h prior to use.
(2) Prepare 11 of a fresh solution containing the following:
10 mg/I Au added as K (Au(CN) 2)
100 mg/I Ca added as CaC12
100 mg/I CN added as KCN
Adjust the temperature of the test solution to 25°C before use.
(3) Place 1 litre test solution in a stoppered oblong plastic container (150 mm
x 150 mm x 140 mm) and add the prepared carbon sample while noting
the time of addition. Immediately commence shaking on a linear shaker
adjusted to 130 oscillations/ min. The linear shaker is housed in a thermo-
statically controlled glove-box at 25°C. Withdraw 5 ml aliquots after 0,
15, 30, 45, 60, 75 and 90 minutes for gold analysis.
The relative standard deviation of the above method was shown to
be 3 ,4 percent.
3. Determination of attrition resistance

Carbon hardness or attrition resistance is a most important characteristic
in any gold recovery circuit. Certain standard methods for the measurement
of the attrition resistance of carbon, such as the A.S. T.M. test, involve the
crushing action of a ball bearing on predried carbon. However, as these tests
bear little resemblance to actual conditions in a CIP circuit and are perhaps
too drastic, different methods for testing carbons for CIP have been
developed. Two test procedures in common use are presented below.
The abrasion resistance method as developed by the Council for Mineral
Tec,lmology involves the use of a laboratory float machine in order to simulate
accelerated plant conditions. The method is as follows:
1050
(1) Sieve carbon to between 1,7 - 2,0 mm, wash and dry in an oven at 100°C
for 16 hours.
(2) Place 30 g of this washed and sized carbon in 3 litres deionized water
contained in a 4 litre capacity float cell. A Denver laboratory flotation
machine, size D-12, is fitted with 9,5 cm stainless steel impeller and a
12,5 cm stainless steel hood.
(3) With the air inlet in the open position and at a constant speed of
1200 r.p.m. agitate the carbon and water for 60 minutes.
(4) Wash carbon onto a 0,850 mm screen. Dry the + 0,850 mm fraction in
an oven at ll0°C for 16 hours. Establish mass of this fraction and ex-
press as a percentage of the total carbon mass.
The measurement of carbon attrition as developed at the AARL makes
use of a tumbling bottle as follows:
(1) 50 g of oven dried carbon (overnight at l 10°C) which was initially wet
screened to 1,00 2,36 mm is agitated with 100 ml water in a 1 litre
"Consol'' preserve jar or bottle for a period of 24 h. Agitation is carried
out by rotating the bottle, end over end, at a rate of 40 r.p.m. (this is
easily effected by wedging the bottle into a larger open ended plastic
beaker or bottle and rolling on a set of rolls).
(2) Following the agitation for a period of 24 h, the carbon sample is wet
screened through a 1,00 mm screen and the undersize carbon recorded
as a percentage mass distribution of the original carbon sample. The test
is carried out in triplicate and the results averaged for specification
purposes.
The relative standard deviation of the above method was 4, 7 percent.
4. Miscellaneous tests
Other carbon specifications which may be required in certain instances in-
clude the following:
(1) Particle size distribution. While the size distribution may be dictated by
the efficiency of the interstage screens on a particular circuit, + 1, 18 -
2,36 mm carbon is generally used in South Africa at present (1986).
(2) Bulk density. Bulk density data may be required in establishing plant
balances.
(3) Solubles and moisture content. For payment purposes the weight of car-
bon will be calculated on a water washed, dry solids basis.
(4) Volatile content. The measurement of the carbon volatile content (B.S.
1016: part 3: 1965: 3.4.2.) provides a most useful method of quality con-
trol with regard to carbon regeneration.
1051
Tables and Formulae
W.R. Flook
CONTENTS
Page
Note on conversion tables .............................. . 1054
Table 1 Conversion table: length ................................ . 1055
Table 2 Conversion table: area ................................... . 1056
Table 3 Conversion table: volume ............................... . 1057
Table 4 Conversion table: velocity .............................. . 1058
Table 5 Conversion table: rate of flow ........................ . 1058
Table 6 Conversion table: mass .................................. . 1059
Table 7 Conversion table: pressure .............................. . 1060
Table 8 Conversion table: energy ................................ . 1060
Table 9 Conversion table: power ................................ . 1061
Table 10 Conversion table: density ............................... . 1061
Table 11 Some fundamental constants ........................... . 1062
Table 12 Moisture content, density, etc. of pulps ............ . 1062
Table 13 Formulae .................................................... . 1064
Table 14 Base units of the International System .............. . 1067
Table 15 Derived units of the International System .......... . 1067
Table 16 S I prefixes .................................................. . 1068
Table 17 Critical speed of mills .................................... . 1068
Table 18 Testing-sieve series ....................................... .. 1069
Table 19 The periodic table ......................................... . 1071
Table 20 The Greek alphabet ....................................... . 1072
1053
TABLES AND FORMULAE
Conversion tables - explanatory notes

In each of the conversion tables which follows, quantities in each horizontal
line are equal when they are expressed in the unit defined at the top of their
respective vertical columns. To convert from units of "x" to any other, the
horizontal line containing unity in column "x" should be used. All values
of other units in that line will be equal to I unit of "x".
In each tabulation where the number of units is equal to ''n'', the number
of primary conversion factors required to calculate the remainder is equal
to n-1. These primary or basic factors are in italics and marked with an
asterisk and are, where possible, exact conversions, definitions of the unit
or derived from two or more of these.
A list of the source conversion factors follows. The more obvious and
well known, such as 12 inches = I Imperial foot, 1000 g = I kg, etc. are
excluded.
I Acre 4840 square yards
I Atmosphere 101,3250 kilopascals
I Atmosphere 760 torr (mm Hg)
I Bar I 00 kilo pascals
I Barrel (oil) 42 U.S. gallons
I Bdtish thermal unit 251,9958 calories
I Calorie 4,1868 joules
I Carat 0,2 grams
I Foot (Cape) 1,033 Imperial feet
I Gallon (Imperial) 4,54609 litres
I Gallon (Imperial) 160 Imperial fluid ounces
I Gallon (U.S.) 128 U.S. fluid ounces
I Gallon (U.S.) 231 cubic inches
I Hectare 10 OOO square metres
I Horsepower (Imperial) 550 foot-pounds per second
I Horsepower (metric) 75 kilogram-metres per second
I Inch 2,54 centimetres
I Kilogram per centimetre squared 0,980665 bars
I Kilowatt-hour 3,6 x 106 joules
I Kilowatt-hour 1,341022 horsepower
I Mile (nautical) 6080 Imperial feet
I Mile (statute) 1760 yards
I Mining Claim (S.A.) 60 OOO Cape square feet
I Morgen 86 400 Cape square feet
I Ounce (troy) 20 pennyweights (dwt.)
1 Pennyweight (dwt.) 24 grains
I Pound (avoirdupois) 0,45359237 kilograms
I Pound (avoirdupois) 16 ounces (Av.)
I Pound (avoirdupois) 7 OOO grains
I Pound (troy) 12 ounces (Tr.)
I Ton (Long Imperial) 2240 pounds (Av.)
I Ton (short) 2000 pounds (Av.)
1 Ton (metric) 1000 kilograms
1054
TABLES AND FORMULAE
An Imperial gallon is defined as 10 lbs of water and the accepted conversion

into the metric system is 4,54609 litres. These conditions are only
simultaneously possible at 22,l 7°C at which temperature the relative densi-
ty of water is 0,997 763 7. Since this temperature is reasonably close to that
of average plant solutions, it is assumed for both water and mercury in the
pressure table. The relative density of mercury at this temperature is
13,540492. However, torr is defined as 1 mm of mercury at 0°C (relative
density 13,5951) and under standard gravitational force.
Table 1. Conversion table: length.
Miles Feet
Kilo- Centi-
Nautical Statute metres Metres Yards Cape English Inches metres
1,85318
1,15152 1,85318 1853,18 2026,67 5885,77 6080 72 960 5
xl0
1,60934
0,868 421 1,60934 1609,34 1760 5111,33 5280 63 360 5
xl0
0,539 612 0,621 371 1000 1093,61 3176,03 3280,84 39370,1 100 OOO
5,39612 6,21371
x 10- 4 x 1(}-4 0,001 1,09361 3,17603 3,28084 39,3701 100
4,93421 5,68182 9,144

0,9144 2,90416 3 36 91,44
x 10- 4 x 10- 4 x 10- 4
1,69901 1,95644 3,14858

x 10- 4 x 10- 4 0,314 858 0,344 333 1,033 12,396 31,4858
x 10- 4
1,64474 1,89394 3,048
x 10- 4 0,3048 0,333 333 0,968054 12 30,48
x 10- 4 x 10- 4
1,37061 1,57828 2,54 2,77778 8,06712 8,33333

x 10- 5 x 10- 5 x10-s 0,0254 x 10- 2 x 10- 2 x 10- 2 2,54
5 ,39612 6,21371 1,09361 3,17603 3,28084

x 10- 6 x 10- 6 0,00001 0,01 x 10- 2 x 10- 2 x 10- 2 0,393 701
Original data Calculated data
~
1"""
VJJ
C::> Table 2. Conversion table: area. >--l
Vi
Cl'\ >
to
Square feet l'
Square tI1
(/).
Square Square Mining Metres Square
Miles Kilometres Hectares Morgen claims Acres (centare) Yards Cape English >
z
u
2,61257 2,78784 '"Tj
2,58999 258,999 302,380 435,428 640 2 589 988 3 097600
x 107 x 107 0
~
247,105
1,00872 1,07639 ~
0,386 102 1 100 116,750 168,120 1 OOO OOO 1 195 990
x 107 Xl0 7 el'
3,86102
x 10-J 0,01 I 1,16750 1,68120 2,47105 10 OOO 11959,9 100 872 107 639 >
tI1
3,30709 8,56533
x 10-J x 10-J 0,856 533 I 1,44 2,11654 8565,33 10244,I 86400 92 196,5
2,29659 5,94815
x 10- 3 x 10-J 0,594 815 0,694444 I 1,46982 5948,15 7113,93 60000 64 025,3
1,5625 4,04686
x 10-J x 10-l 0,404 686 0,472 469 0,680 356 I 4046,86 4840 40821,3 43 560
3,86102 1,16750 1,68120 2,47105

x 10- 7 0,000 001 0,0001 x 10- 4 x 10-• x 10-• I I, 19599 10,0872 10,7639
3,22831 8,36127 8,36127 9,76176 1,40569 2,06612

x 10- 7 x 10- 7 x 10- 5 x 10- 5 x 10- 4 x 10-• 0,836127 I 8,43416 9
3,82766 9,91358 9,91358 1,15741 1,66667 2,44970 9,91358

x 10- 8 x 10- 8 x 10- 6 x 10- 5 x 10- 5 x10- 5 x 10- 2 0,118 565 I 1,06709
3,58701 9,29030 9,29030 1,08464 1,56188 2,29568 9,29030

x 10- 8 x 10- 8 x 10- 6 X 10- 5 x 10- 5 x 10- 5 x 10- 2 0,111111 0,937 129

Table 3. Conversion table: volume.
Gallons Fluid ounces

Mega- Cubic Barrels Cubic Milli-
litres metres (oil) feet Imperial U.S. Litres U.S. Imperial litres
3,38140 3,51951 109

1000 6289,81 35 314,7 219 969 264 172 1 OOO OOO Xl07 x 107
0,001 1 6,28981 35,3147 219,969 264,172 1000 33814,0 35195,1 1 OOO OOO
1,58987
x 10- 4 0,158 987 1 5,61458 34,9723 42 158,987 5376 5595,57 158 987
2,83168 2,83168
X 10-s x 10- 2 0,178 108 1 6,22884 7,48052 28,3168 957,506 996,614 28316,8
4,54609 4,54609 2,85940

x 10- 6 x 10- 3 x 10- 2 0,160 544 1 1,20095 4,54609 153,722 160 4546,09
3,78541 3,78541 2,38095

x 10- 6 x 10- 3 x 10- 2 0,133 681 0,832 674 1 3,78541 128 133,228 3785,41
0,000 001 0,001

6,28981 3,53147
0,219 969 0,264 172 1 33,8140 35,1951 1000
~
to
x 10- 3 x 10- 2 t""'
l:'I'1
2,95735 2,95735 1,86012 1,04438 6,50527 7,8125 2,95735 [/)
x 10-s x10-s 1 1,04084 29,5735
x 10-• x 10- 3 x 10- 3 x 10- 3 x 10- 2
>
2,84131
x 10-s
2,84131
x 10- 5
1,78713
x 10- 4
1,00340
xl0-3 0,00625
7,50594
x 10- 3
2,84131
x 10- 2 0,960 760 1 28,4131 s
'"Tj
6,28981 3,53147 2,19969 2,64172 3,38140 3,51951 0
10-9 0,000 001 0,001 1 :;;:;
x 10- 6 x 10- 5 x 10- 4 x 10- 4 x 10- 2 x10-'
~
c::::
Original data Calculated data t""'
0·~
(..fl
·-I
>
l:'I'1
TABLES AND FORMULAE
Table 4. Conversion table: velocity.
Metres Miles Feet Kilo- Metres Feet

per per per metres per per
second Knots hour second per hour minute minute
1,94260 2,23694 3,28084 3,6 60 196,850

0,514 773 1,15152 1,68889 1,85318 30,8864 101,333
0,44704 0,868 421 1,46667 1,60934 26,8224 88
0,3048 0,592 105 0,681 818 1,09728 18,288 60
0,277 778 0,539 612 0,621 371 0,911 344 16,6667 54,6807
1,66667 3,23767 3,72823 5,46807
x10- 2 0,06 3,28084
x 10- 2 x 10- 2 x 10- 2
9,86842 1,13636 1,66667
0,00508 x10-1 0,018 288 0,3048
x 10- 2 x 10- 2
Table 5. Conversion table: rate of flow.
Mega- Cubic Mega- Cubic Gallons/min

litres/ metres/ litres/ Litres/
min sec hour Feet/sec Metres/min second Imperial U.S.
16,6667 60 588,578 1000 16 666,7 219 969 264 172

0,06 3,6 35,3147 60 1000 13 198,2 15 850,3
1,66667
x 10- 2 0,277 778 9,80963 16,6667 277,778 3666,15 4402,87
1,69901 2,83168
x 10- 3 x 10- 2 0,101 941 1,69901 28,3168 373,730 448,831
1,66667
0,001 0,06 0,588 578 16,6667 219,969 264,172
x 10- 2
3,53147
0,00006 0,001 0,0036 0,06 13,1982 15,8503
x 10- 2
4,54609 7,57682 2,72765 2,67573 4,54609 7,57682
x 10- 6 x 10- 5 x 10- 4 x 10- 3 1,20095
x 10- 3 xl0- 2
3,78541 6,30902 2,27125 2,22801 3,78541 6,30902
x 10- 6 x10-s x 10- 4 xl0- 3 Xl0-3 x 10- 2 0,832 674

For mining and metallurgical purposes
w
1 n.:11
...l..V..JV
Table 6. Conversion table: mass.
Tons Pounds Ounces

Penny-
Kilo- Avoir- Avoir- weights
Long Metric Short grams dupois Troy Troy dupois (Troy) Grams Carats Grains
1,568
1,01605 1,12 1016,05 2240 2722,22 32 666,7 35 840 653 333 1 016 047 5 080 234
xl0 7
1,54324
0,984 207 l 1,10231 1000 2204,62 2679,23 32 150,8 35 274,0 643 015 1 OOO OOO 5 OOO OOO
xl0 7
0,892 857 0,907 185 1 907,185 2000 2430,56 29 166,7 32 OOO 583 333 907 185 4 535 923 1,4 x 107
9,84207 1,10231
x 10- 4 0,001 x 10- 3 1 2,20462 2,67923 32,1508 35,2740 643,015 1000 5000 15432,4
4,46429 4,53592
x 10- 4 x 10- 4 0,0005 0,453 592 1 1,21528 14,5833 16 291,667 453,592 2267,96 7000
3,67347 3,73242 4,11429

x 10- 4 x 10- 4 x 10- 4 0,373 242 0,822 857 1 12 13,1657 240 373,242 1866,21 5760
3,06122 3,11035 3,42857 3,11035 6,85714 8,33333

x 10- 5 X 10- 5 x 10- 5 x 10- 2 x 10- 2 x 10- 2 1 1,09714 20 31,1035 155,517 480
2,79018 2,83495 3,125 2,83495 7,59549

~
l:C
X 10- 5 x 10- 5 x 10- 5 x 10- 2 0,0625 x 10- 2 0,911 458 1 18,2292 28,3495 141,748 437,5 r
tI1
C/J
1,53061 1,55517 1,71429 1,55517 3,42857 4,16667 5,48571
x 10- 6 x 10- 6 x 10- 6 x 10- 3 x 10- 3 x 10-J 0,05 x 10- 2 I 1,55517 7,77587 24 >
z
9,84207 1,10231 2,20462 2,67923 3,21508 3,5-2740 tl
x 10- 7 0,000 001 x 10- 0 0,001 x 10- 3 x 10-J x 10- 2 x 10- 2 0,643015 1 5 15,4324 "!1
0
1,96841 7 2,20462 4,40925 5,35846 6,43015 7,05479 ~
2 x 10- 0,0002 0,128603 0,2 1 3,08647 ~
x 10- 7 x 10- 7 x 10- 4 x 10- 4 x 10- 3 x 10- 3
c
·-
0
V1
\0
6,37755
x 10- 8
6,47989
x 10- 8
7,14286
x 10-s
6,47989
'x 10- 5
1,42857
x 10- 4
1,73611
x 10- 4
2,08333
x 10- 3
2,28571
x 10- 3
4,16667
x 10- 2
6,47989
x 10- 2 0,323 994 1
r
>
tI1
Original data For mining and metallurgical purposes Calculated data
TABLES AND FORMULAE
Table 7. Conversion table: pressure.
Kilo- Metres Pounds/ Inches Feet mm of

Atmos- grams/ of square of of Kilo- mercury
pheres Bars sq cm water* inch mercury* water* pascals (torr)t
1,01325 1,03323 10,3095 14,6960 29,8020 33,8237 101,325 760

0,986 923 1,01972 10,1746 14,5038 29,4123 33,3814 100 750,062
0,967 841 0,980 665 9,97792 14,2233 28,8436 32,7360 98,0665 735,559
9,69983 9,82835
0,100 221 1,42548 2,89074 3,28084 9,82835 73, 7187
x 10- 2 x 10- 2
6,80460 6,89476 7,03069

x 10- 2 x 10- 2 x 10- 2 0,701 517 2,02790 2,30156 6,89476 51,7149
3,35548 3,39994 3,46698

x 10- 2
0,345 932 0,493 120 1,13495 3,39994 25,5017
x 10- 2 x 10- 2
2,95651 2,99568 3,05475

x 10- 2 x 10- 2 x10-2 0,3048 0,434 487 0,881 097 2,99568 22,4695
9,86923 1,01972
0,01 0,101 746 0,145 038 0,294 123 0,333 814 7,50062
x 10- 3 x 10- 2
1,31579 1,33322 1,35951 1,35651 1,93368 3,92131 4,45048

x 10- 3 Xl0-3 x 10- 3 x 10- 2 x 10- 2 x 10- 2 x 10- 2
0,133 322
-2
Original data *at22°C tat density of 13,5951 & g = 980,665 cm s Calculated data
Table 8. Conversion table: energy
Horse Kilo- Horse British Kilo-

power- watt- power- thermal gram- Foot-
days hours hours units metres Calories* pounds Joules
1,53885 4,752 6,44285

17,8968 24 61066,4 6 569 874 7 7 7
xl0 x 10 xl0
5 ,58759
1,34102 3412,14 367 098 859 845 2 655 224 3 600 OOO
x 10- 2
4,16667
0,745 700 2544,43 273 745 641 186 1 980 OOO 2684519
x 10- 2
1,63756 2,93071 3,93015

107,586 251,996 778,170 1055,06
x 10- 5 x 10- 4 x 10- 4
1,52210 2,72407 3,65304 9,29491

2,34228 7,23302 9,80665
x 10- 7 x 10- 6 x 10- 6 x 10- 3
6,49837 1,163 1,55961 3,96832

0,426 935 3,08803 4,1868
x 10- 8 x 10- 6 x 10- 6 xl0-3
2,10438 3,76616 5,05051 1,28507

0,138 255 0,323 831 1,35582
x 10- 8 x 10- 7 x10-1 xl0-3
1,55211 2,77778 3,72506 9,47817

x 10- 8
0,101 972 0,238 846 0,737 562
x 10- 7 Xl0-7 x 10- 4
Original data •International table Calculated data
1060
TABLES AND FORMULAE
Table 9. Conversion table: power.
Horsepower B.T.U. kg Foot Joules/

Kilo- per metres/ pounds/ sec
watts British Metric min sec sec (watts)
1,34102 1,35962 56,8690 101,972 737,562 1000

0,745 700 1,01387 42,4072 76,0402 550 745,700
0,735 499 0,986 320 41,8271 75 542,476 735,499
1,75843 2,35809 2,39079
x 10- 2 x 10- 2 x 10- 2 1,79310 12,9695 17,5843
9,80665 1,31509 1,33333

x 10-J x 10- 2 x10-2 0,557 695 7,23302 9,80665
1,35582 1,81818 1,84340 7,71040

x 10-J x 10- 3 x 10- 3 x 10- 2 0,138 255 1,35582
1,34102 1,35962 5,68690

0,001 0,101 972 0,737 562
xl0- 3 xl0- 3 x10- 2

Table 10. Conversion table: density
Pounds/ Short Grams/ Pounds/Gallon Pounds/

cubic tons/ cubic cubic Grams/
inch cubic yard cm U.S. Imperial foot litre
23,328 27,6799 231 277,412 1728 27679,9

4,28669
x 10- 2 I, 18655 9,90226 11,8918 74,0741 1186,55
3,61273
x 10- 2 0,842 778 8,34541 10,0221 62,4280 1000
4,32900
x 10-J 0,100 987 0,119 826 1,20092 7,48052 119,826
3 ,60475 8,40917 9,97792

x 10- 3 x10- 2 x 10- 2 0,832 698 6,22901 99,7792
5,78704 1,60185
x 10- 4 0,0135 x 10- 2 0,133 681 0,160 539 16,0185
3 ,61273 8,42778 8,34541 1,00221 6,24280

x 10- 5 x 10- 4 0,001 x10-2
x 10- 3 x 10- 2

1 n.e:-1
.J.VV-1
TABLES AND FORMULAE
Table 11. Some fundamental constants.
Name Symbol Value Unit
Velocity of light c 2,99792548 x 10s m s- 1

Standard gravity g 9,80665 m s- 2
Absolute temperature 273,16 oc
ice point
Pi 7r 3,141592654
Naperian log base e 2, 71828182846
Radian 57 ,2957795131 degrees
Table 12. Moisture content, density, etc. of pulps.

Relationship between units of measurement of moisture content of pulps. ·Specific gravity of
dry solids = 2,7.
Pulp LIS % by mass LIS % by volume

density m 3/kg by mass liquid solid by vol. liquid solid
2,700 0,370 0,000 0,00 100,00 0,000 0,00 100,00

2,655 0,377 0,010 1,00 99,00 0,027 2,65 97,35
2,611 0,383 0,020 2,00 98,00 0,055 5,22 94,78
2,569 0,389 0,031 3,00 97,00 0,084 7,71 92,29
2,528 0,396 0,042 4,00 96,00 0,113 10,11 89,89
2,488 0,402 0,053 5,00 95,00 0,142 12,44 87,56
2,450 0,408 0,064 6,00 94,00 0,172 14,70 85,30
2,413 0,414 0,075 7,00 93,00 0,203 16,89 83,11
2,377 0,421 0,087 8,00 92,00 0,235 19,01 80,99
2,342 0,427° 0,099 9,00 91,00 0,267 21,08 78,92
2,308 0,433 0,111 10,00 90,00 0,300 23,08 76,92
2,275 0,440 0,124 11,00 89,00 0,334 25,02 74,98
2,243 0,446 0,136 12,00 88,00 0,368 26,91 73,09
2,211 0,452 0,149 13,00 87,00 0,403 28,75 71,25
2,181 0,459 0,163 14,00 86,00 0,440 30,53 69,47
2,151 0,465 0,176 15,00 85,00 0,476 32,27 67,73
2,123 0,471 0,190 16,00 84,00 0,514 33,96 66,04
2,095 0,477 0,205 17,00 83,00 0,553 35,61 64,39
2,067 0,484 0,220 18,00 82,00 0,593 37,21 62,79
2,041 0,490 0,235 19,00 81,00 0,633 38,78 61,22
2,015 0,496 0,250 20,00 80,00 0,675 40,30 59,70
1,990 0,503 0,266 21,00 79,00 0,718 41,78 58,22
1,965 0,509 0,282 22,00 78,00 0,762 43,23 56,77
1,941 0,515 0,299 23,00 77,00 0,806 44,64 55,36
1,918 0,521 0,316 24,00 76,00 0,853 46,02 53,98
1,895 0,528 0,333 25,00 75,00 0,900 47,37 52,63
1,872 0,534 0,351 26,00 74,00 0,949 48,68 51,32
1,851 0,540 0,370 27,00 73,00 0,999 49,97 50,03
1,829 0,547 0,389 28,00 72,00 1,050 51,22 48,78
1,808 0,553 0,408 29,00 71,00 1,103 52,44 47,56
1,788 0,559 0,429 30,00 70,00 1,157 53,64 46,36
1062
TABLES AND FORMULAE
Table 12. Moisture content, density, etc. of oulps (continued).
Pulp LIS % by mass LIS % by volume

density m3/kg by mass liquid solid by vol. liquid solid
1,768 0,566 0,449 31,00 69,00 1,213 54,81 45,19

1,749 0,572 0,471 32,00 68,00 1,271 55,96 44,04
1,730 0,578 0,493 33,00 67,00 1,330 57,08 42,92
1,711 0,584 0,515 34,00 66,00 1,391 58,17 41,83
1,693 0,591 0,538 35,00 65,00 1,454 59,25 40,75
1,675 0,597 0,563 36,00 64,00 1,519 60,30 39,70
1,657 0,603 0,587 37,00 63,00 1,586 61,33 38,67
1,640 0,610 0,613 38,00 62,00 1,655 62,33 37,67
1,624 0,616 0,639 39,00 61,00 1,726 63,32 36,68
1,607 0,622 0,667 40,00 60,00 1,800 64,29 35,71
1,591 0,629 0,695 41,00 59,00 1,876 65,23 34,77
1,575 0,635 0,724 42,00 58,00 1,955 66,16 33,84
1,560 0,641 0,754 43,00 57,00 2,037 67,07 32,93
1,545 0,647 0,786 44,00 56,00 2,121 67,96 32,04
1,530 0,654 0,818 45,00 55,00 2,209 68,84 31,16
1,515 0,660 0,852 46,00 54,00 2,300 69,70 30,30
1,501 0,666 0,887 47,00 53,00 2,394 70,54 29,46
1,487 0,673 0,923 48,00 52,00 2,492 71,37 28,63
1,473 0,679 0,961 49,00 51,00 2,594 72,18 27,82
1,459 0,685 1,000 50,00 50,00 2,700 72,97 27,03
1,446 0,691 1,041 51,00 49,00 2,810 73,75 26,25
1,433 0,698 1,083 52,00 48,00 2,925 74,52 25,48
1,420 0,704 1,128 53,00 47,00 3,045 75,28 24,72
1,408 0,710 1,174 54,00 46,00 3,170 76,02 23,98
1,395 0,717 1,222 55,00 45,00 3,300 76,74 23,26
1,383 0,723 1,273 56,00 44,00 3,436 77,46 22,54
1,371 0,729 1,326 57,00 43,00 3,579 78,16 21,84
1,360 0,736 1,381 58,00 42,00 3,729 78,85 21,15
1,348 0,742 1,439 59,00 41,00 3,885 79,53 20,47
1,337 0,748 1,500 60,00 40,00 4,050 80,20 19,80
1,325 0,754 1,564 61,00 39,00 4,223 80,85 19,15
1,315 0,761 1,632 62,00 38,00 4,405 81,50 18,50
1,304 0,767 1,703 63,00 37,00 4,597 82,13 17,87
1,293 b,773 1,778 64,00 36,00 4,800 82,76 17,24
1,283 0,780 1,857 65,00 35,00 5,014 83,37 16,63
1,272 0,786 1,941 66,00 34,00 5,241 83,98 16,02
1,262 0,792 2,030 67,00 33,00 5,482 84,57 15,43
1,252 0,799 2,125 68,00 32,00 5,738 85,16 14,84
1,243 0,805 2,226 69,00 31,00 6,010 85,73 14",27
1,233 0,811 2,333 70,00 30,00 6,300 86,30 13,70
1,223 0,817 2,448 71,00 29,00 6,610 86,86 13,14
1,214 0,824 2,571 72,00 28,00 6,943 87,41 12,59
1,205 0,830 2,704 73,00 27,00 7,300 87,95 12,05
1,196 0,836 2,846 74,00 26,00 7,685 88,49 11,51
1,187 0,843 3,000 75,00 25,00 8,100 89,01 10,99
1,178 0,849 3,167 76,00 24,00 8,550 89,53 10,47
1,169 0,855 3,348 77,00 23,00 9,039 90,04 9,96
1063
TABLES AND FORMULAE
Table 12. Moisture content, density, etc. of pulps (continued).
Pulp LIS Ofo by mass LIS Ofo by volume

density m 1/kg by mass liquid solid by vol. liquid solid
l, 161 0,861 3,545 78,00 22,00 9,573 90,54 9,46

1,152 0,868 3,762 79,00 21,00 10,157 91,04 8,96
1,144 0,874 4,000 80,00 20,00 10,800 91,53 8,47
1,136 0,880 4,263 81,00 19,00 11,511 92,01 7,99
1,128 0,887 4,556 82,00 18,00 12,300 92,48 7,52
1,120 0,893 4,882 83,00 17,00 13,182 92,95 7,05
1,112 0,899 5,250 84,00 16,00 14,175 93,41 6,59
1,104 0,906 5,667 85,00 15,00 15,300 93,87 6,13
1,097 0,912 6,143 86,00 14,00 16,586 94,31 5,69
1,089 0,918 6,692 87,00 13,00 18,069 94,76 5,24
1,082 0,924 7,333 88,00 12,00 19,800 95,19 4,81
1,074 0,931 8,091 89,00 11,00 21,845 95,62 4,38
1,067 0,937 9,000 90,00 10,00 24,300 96,05 3,95
1,060 0,943 10,111 91,00 9,00 27,300 96,47 3,53
1,053 0,950 11,500 92,00 8,00 31,050 96,88 3,12
1,046 0,956 13,286 93,00 7,00 35,871 97,29 2,71
1,039 0,962 15,667 94,00 6,00 42,300 97,69 2,31
1,033 0,969 19,000 95,00 5,00 51,300 98,09 1,91
1,026 0,975 >20 96,00 4,00 >60 98;48 1,52
1,019 0,981 >20 97,00 3,00 >60 98,87 1,13
1,013 0,987 >20 98,00 2,00 >60 99,25 0,75
1,006 0,994 >20 99,00 1,00 >60 99,63 0,37
1,000 1,000 >20 100,00 0,00 >60 100,00 0,00
Table 13. Formulae. The formulae in the following two tabulations can be used to derive any
of five groups of variables defining a pulp or slurry, provided two are known.
These variables are as follows: Abbreviation

used
1. Solids content Mass (tons) s
Volume (m1) Vs
2. Water content Mass (tons) or volume (m1) w
3. Pulp Mass (tons) p
4. Pulp Volume (m1) Vp
5. Relative moisture content
Pulp density Pd
Liquid/Solid ratio Ls
Fraction moisture M
The constants in the formulae apply to a mixture of liquid of specific gravily 1 and dry solids
of specific gravity of 2, 7.
For dry solids of any other specific gravity equal to D, these constants should be changed
as follows:
I D - 1
2,7 = D 1,7 = D - 1 o,37 = n 0,63 =
D
1064
TABLES AND FORMULAE
Table 13. Formulae (continued).
s Vs w p Vp
I
Pd&S I 0,37.S
I S.(1-0,37.Pd)I 1,7.S.Pd I 1,7.S I
I I Pd - 1 I 2. 7. ( Pd-1) I 2. 7. ( Pd-1) I
_____ l _____ I I I I
I I I I I
Pd&Vs 2,7.Vs I I Vs, ( 2, 7-Pd) I 1, 7. Vs, Pd I 1, 7. Vs I
I I Pd - 1 I Pd - 1 I Pd - 1 I
____I I I I I
I I I I I
Pd&W W.(Pd - 1) I W.(Pd-1) I I 1,7.W.Pd I 1,7.W I
1 - 0,37.Pd I 2,7 - Pd I I 2,7 - Pd I 2,7 - Pd I
____I I I I I
I I I I I
Pd&P 2,7.P.(Pd-l) I P.(Pd-1) I P.(2,7-Pd) I I p I
1,7.Pd I 1,7.Pd I 1,7.Pd I I Pd I
____I I I I I
I I I I I
2,7.Vp.(Pd-l)I Vp.(Pd-1) I Vp.(2,7-Pd) I Vp.Pd I I
Pd&Vp 1,7 I 1,7 I 1,7 I I I
____I I I I I
I I I I I
Ls&S I I I I I
I 0,37.S I S.Ls I S. (Ls+l) I s. (0,37+Ls) I
____I I I I I
I I I I I
Ls&Vs 2,7.Vs I I I I I
I I 2,7.Vs.Ls 2,7.Vs.(Ls+l) J Vs.(2,7.Ls+l)I
1
____I I I I I
I I I I I
&J Ls&W w I w I I 1 I 1 I
Ls I 2, 7. Ls I I W' ( 1 + LS ) I W' ( 1 + 2, 7. Ls )I
~
H
____I I I I I
I I I I I
~ Ls&P p
Ls+l
I
J
P
2, 7. (Ls+ 1 )
I
I
P.Ls
Ls+ 1
I
I
I P. (0,37+Ls) I
I Ls+ 1 I
I Ls&Vp
____I
Vp
0, 37 + Ls
I
I
I
V
1 + 2, 7. Ls
I
I
I
I
Vp. Ls
Ls + O, 37
I
I
I
I
Vp. (Ls+ 1 )
Ls + 0, 37
I
I
I
I
I
I
I
I
____ I I I I I
I I I I I
M&S I 0,37.S I ~ I s I S.(1,7.M+l),i
I I 1 - M I 1 - M I 2, 7. (1-M) I
____I I I I I
I I I I I
M&Vs J 2,7.Vs J I 2,7.Vs.M I 2,7.Vs I Vs.(1+1,7.M)I
I I I 1-M I l'"="'M I 1-M I
1_____ 1 I I I I
1 1 I I I I
M&W I W.(1 - M) I W.(1 - M) I I w I W.(1+1,7.M) I
I M I 2,7.M I I M I 2,7.M I
I I I I I I
I I I I I I
M&P I P.(1-M) I P.(1 - M) I P.M I I P.(1+1,7.M) I
I I 2, 7 I I I 2, 7 I
I I I I I I
I I I I I I
M&Vp I 2,7.Vp(l-M) I Vp.(1-M) I 2,7.Vp.M I .?....2.:..Y.P_ I I
I 1,7.M+l I 1,7.M+l I 1,7.M+l I 1,7.M+l I I
I I I I I I
1065
VARIABLES TO BE DERIVED
......
0
0\
0\ I
s Vs w p Vp Pd Ls M
I ~
to
I I I I I I s+w I
S&W I - I 0,37.S I I S+W I 0,37.S+W I 0,37.S+W I sw II W+S w I
I t"""
tI1
I I I I I I I I I r:n
I I I I I I I I I
>
S&P I
I
I
I
-
I
I
I
I
0' 37 · 8 I
I P- 8
I
I
I
I
I
I
-
I
I
I
I
I
I
P I P
I
I
I
I P - 0' 63 · 8 I P - o. 63. sI s - 1 I 1 - i'
I
I
s I
I
I
I
a
'Tj
S&Vp I - I 0,37.S I Vp-0,37.s I Vp+0,63.s I - I 1 + 0,63.SI YE - 0,37lvp - 0,37.sl 0
I I I I I I Vp I s IVp + 0. 63. sI l'::i
I I I I I I I I I $:
vs&W
I
I 2 , 7 .vs
I
I _
I
I _
I I
I 2, 7.Vs+W I Vs + w
I .
I 2,7.Vs +WI __
I I
W_ I _w_ I
I ...., c
I I I I I I Vs + W I 2, 7. Vs I W+2, 7.Vs I '"(>"er ~
.,"'
""~
I
I
Vs&P II 2'7. Vs II
I
I
I
I
I
I
I
I
I
I
I
I
I
I P
I
I
i P
I
I
I 2 7. vs I
I
I -
w
tI1
I p - 2' 7. Vs I I P-1, 7. Vs I p - 1, 7. Vs I 2'7. Vs - 111 - "-'-F I >-rj
~
0
~
I I I I I • I I I I
I I I I I I I I I
;;;-
~ Vs&Vp II 2, 7.vs /I _ I Vp _ Vs I Vp+l, 7.Vs I
I I I
I 1 + 1,7.Vsl ~ I Vp - Vs I
I Vp I 2, 7. Vs I Vp + 1, 7. Vs I "'
;a ~
I I I I I I I I I ()
0
I
I r w
I
I P- w I
I I
I
I I
I 1,7.w + P I 2,7.P I
I
w
I
I w
I
I
a
s·
W&P I - I 2,7 I I - I 2. 7 I 1, 7. W+P I p- w I p I c
I I I I I I I I I ~
I I I I I I I I I
W&Vp II 2,7.(Vp-W)II Vp - w I
I
-
I
,2,7.Vp-l,'ZWI
I
-
I 1,7.wl w I
I 2,7---vp-12,7.(Vp-W) l2,7.Vp-l,7WI
w I
I I I I I I I I I
I I I I I I I I I
P&Vp 12,7.(P-Vp) I p - Vp I 2,7.Vp-P I - I I _!'. I Vp-0,37.P 12,7.Vp - p I
I 1, 7 I 1•7 I 1•7 I I I Vp I p - Vp I 1 • 7. p I
I I I I I I I I_· I
I ~
l Pd = ~ I Pd= --2.,2_I Ls = 1 _0,37.Pdl Ls = _M_ I M= 2 7 - Pd I M= ~
I Ls+0,371 1,7.M+ll Pd-1 I 1-M I 1,7.Pd I l+Ls
I I I________ I I I _ _ __
TABLES AND FORMULAE
Table 14. Base units of the International System.
Quantity Name of unit Unit symbol
Length metre m
Mass kilogram kg
Time second s
Electric current ampere A
Temperature kelvin K
Luminous intensity candela cd
Amount of substance mole mol
Table 15. Derived units of the International System.
Unit expressed
Unit symbol, in terms of
where base or
different from supplementary
Quantity Name of unit basic form units
Area Square metre m1

Volume Cubic metre m3
Frequency Hertz Hz s-1
Density Kilogram per cubic metre kg m- 3
Velocity Metre per second m s- 1
Angular velocity Radian per second rad s- 1
Acceleration Metre per second squared m s- 2
Angular acceleration Radian per second squares rad s-•
Volumetric flow rate Cubic metre per second m3 s-1
Force Newton N kg.m s- 2
Surface tension Newton per metre,
joule per square metre
Pressure Newton per square metre,
pascal
Viscosity, dynamic Newton-second per square
metre, pascal. second
Viscosity, kinematic Metre squared per second
Work, torque, energy, Joule, newton-metre,
quantity of heat watt-second J,N.m,W.s kg m 2 s- 2
Power, heat flux Watt, joule per second W, J/s- 1 kg m2 s- 3
Heat flux density Watt per square metre W m- 2 kg s- 3
Volumetric heat release rate Watt per cubic metre W m- 3 kg m- 1 s- 3
1067
TABLES AND FORMULAE
Table 16. SI prefixes.
Factor Factor in words SI prefix SI symbol
18
1 OOO OOO OOO OOO OOO OOO or 10 trillion exa- E
1 OOO OOO OOO OOO OOO or 10 15 billiard pet a- p
1 OOO OOO OOO OOO or 10 12 billion tera- T
1 OOO OOO OOO or 109 milliard gig a- G
1 OOO OOO or 106 million mega- M
I OOO or 103 thousand kilo- k
100 or 102 hundred hecto- h
10 or 101 ten deca- da
0,1 or 10- 1 tenth deci- d
0,01 or 10- 2 hundredth centi- c
0,001 of 10- 3 thousandth milli- m
0,000 001 or 10- 6 millionth micro- µ
0,000 OOO OOO 001 or 10- 9 milliardth nano- n
0,000 OOO OOO 001 or 10- 12 billionth pico- p
0,000 OOO OOO OOO 001 or 10- 15 billiardth fem to- f
0,000 OOO OOO OOO OOO 001 or 10-1s trillionth atto- a
Table 17. Critical speed of mills.
General formula
Critical speed = ±
where g 9,80665 m s- 2
D diameter in metres
Specific formula
Critical speed = -42,29.D-v, r.p.m.
1068
Table 18. Testing-sieve series.
TYLER A.S.T.M. STANDARD BRITISH STANDARD SABSl97
Nominal Mesh Aperture Mesh Wire Aperture Wire Toler- Aperture Wire Toler-
Aperture '12 '12 diameter, diameter, ance diameter, ance Aperture
in. series in. mm. series in. Mesh in. mm. in. ±%a Mesh in. mm. in. ±%a mm
2,97 ........ ........ 2,207 ········ 3,00 76,2 0,19 to 0,32 2

2 ········ 2,10 ........ ........ 0,192 ········ 2,00 50,8 0,16 to 0,245 2
1 y, ········ 1,48 ········ ........ 0,162 ········ 1,50 38,1 0,145 to 0,210 2
1 ········ 1,050 26,67 ........ 0,148 ........ 1,00 25,4 0,135 to 0,177 3
7/8 ........ 0,883 22,43 ........ 0,135 ........ 0,875 22,2 0,127 to 0,166
314 ........ 0,742 18,85 ........ 0,135 ........ 0,750 19,l 0,122 to 0,154
5/8 ........ 0,624 15,85 ........ 0,120 ........ 0,625 15,9 0,108 to 0,135 3 ........ ........ ········ ........ ········ 16,00
1/2 ........ 0,525 13,33 ........ 0,105 ........ 0,500 12,7 0,094 to 0,122 3 ········ ........ ········ ........ ........ 13,20
7116 ........ 0,441 11,20 ........ 0,105 ········ 0,438 11,l 0,088 to 0,122 3 ........ ........ ········ ........ ········ 11,20
3/8 ········ 0,371 9,423 ........ 0,092 ........ 0,375 9,52 0,083 to 0,112 3 ........ ........ ········ ········ ········ 9,50
5';16 ........ 0,312 7,925 2Y, 0,088 ........ 0,312 7,93 0,073 to 0,093 3 ........ ........ ........ ........ ........ 8,00
114 3 0,263 6,680 3 0,070 3 0,250 6,35 0,063 to 0,083 3 ........ ........ ........ ........ .. ...... 6,7-0
µm
5 660 ........ 0,221 5,613 3y, 0,065 3Y, 0,223 5,66 0,050 to 0,075 3 ........ ........ ........ ........ 5,60
~
········
4 760 4 0,185 4,699 4 0,065 4 0,187 4,76 0,045 to 0,066 3 ........ ........ ........ ········ ........ 4,75
4 OOO ........ 0,156 3,962 5 0,044 s 0,157 4,00 0,039 to 0,058 3 ........ ........ ........ ········ ········ 4,00 t:C
L'
3 360 6 0,131 3,327 6 0,036 6 0,132 3,36 0,034 to 0,052 3 5 0,132 3,34 0,068 3 3,35 trJ
2 830 ········ 0,110 2,794 7 0,0378 7 0,111 2,83 0,031 to 0,047 3 6 0,1107 2,81 0,056 3 2,80 'JJ
a>
2 380 8 0,093 2,362 8 0,032 8 0,0937 2,38 0,0291 to 0,0433 3 7 0,0949 2,41 0,048 3 2,36
2 OOO ........ 0,078 1,981 9 0,033 10 0,0787 2,00 0,0268 to 0,0394 3 8 0,0810 2,05 0,044 3 2,00
1 680 10 0,<165 1,651 10 0,035 12 0,0661 1,68 0,0244 to 0,0354 3 10 0,0660 1,67 0,034 3 1,70
1 410 ········ 0,055 1,397 12 0,028 14 0,0555 1,41 0,0220 to 0,0315 3 12 0,0553 1,40 0,028 3 1,40 "I1
1 190 14 0,046 1,168 14 0,025 16 0,0469 1,19 0,0197 to o:0276 3 14 0,0474 1,20 0,024 3 1,18 0
I OOO ........ 0,039 0,991 16 0,0235 18 0,0394 1,00 O,Cll69 to 0,0244 5 16 0,0395 1,00 0,023 3 1,00 ~
840 20 0,0328 0,833 20 0,0172 20 0,0331 0,84 0,015 to 0,0217 5 18 0,0336 0,85 0,022 5 0,85 a::
710 ········ 0,0276 0,701 24 0,0141 25 0,0280 0,71 0,013 to 0,0189 5 22 0,0275 0,70 0,018 5 0,71
c:::
L'
590
500
28 0,0232
0,0195
0,589
0,495
28
32
0,0125
0,0118
30
35
0,0232
0,0197
0,59
0,50
0,0114
0,0102
to
to
0,0165
0,0146
5
5
25
30
0,0236
0,0197
0,60
0,50
0,0164
0,0136
5
5
0,60
0,50
>
trJ
········
'"""'3
Table 18. Testing-sieve series (continued).
>
to
t""'
TYLER A.S.T.M. STANDARD BRITISH STANDARD SABS 197 tI1
Cll
Nominal
Aperture
Mesh
.J2
Aperture Mesh
.J2
Wire
diameter,
Aperture Wire
diameter,
Toler-
ance
Aperture Wire
diameter,
Toler-
ance Aperture
>
z
in. series in. mm. series in. Mesh in. mm. in. ±"lo a Mesh in. mm. in. ±"lo a mm tl
'Ti
420 35 C>,0164 0,417 35 0,0122 40 0,0165 0,42 0,0091 to 0,013 5 36 0,0166 0,42 0,0112 5 0,425 0
350 ........ 0,0138 0,351 42 O,OJOO 45 0,0138 0,35 0,0079 to 0,0114 5 44 0,0139 0,35 0,0088 5 0,355 :::0
297 48 0,0116 0,295 48 0,0092 50 0,0117 0,297 0,0067 to 0,0100 5 52 0,0116 0,30 0,0076 6 0,300 ~
250 0,0097 0,246 60 0,0070 60 0,0098 0,25 0,0059 to 0,0087 5 60 0,0099 0,252 0,0068 6 0,250 c::::
········ t""'
2JO
177
65
........
0,0082
0,0069
0,208
0,175
65
80
0,0072
0,0056
70
80
0,0083
0,0070
0,21
0,177
0,0051
0,0045
to
to
0,0074
0,0061
5
6
72
85
0,0083
0,0070
0,211
0,177
0,0056
0,0048
6
6
0,212
0,180
>
tI1
149 JOO 0,0058 0,147 JOO 0,0042 JOO 0,0059 0,149 0,0038 to 0,0049 6 JOO 0,0060 0,152 0,004 6 0,150
125 ........ 0,0049 0,124 115 0,0038 120 0,0049 0,125 0,0031 to 0,0041 6 120 0,0049 0,125 0,0034 6 0,125
105 150 0,0041 O,J04 150 0,0026 140 0,0041 0,J05 0,0025 to 0,0034 6 150 0,0041 O,J05 0,0026 8 0,106
88 ........ 0,0035 0,088 170 0,0024 170 0,0035 0,088 0,0021 to 0,0029 6 170 0,0035 0,088 0,0024 8 0,090
74 200 0,0029 0,074 200 0,0021 200 0,0029 Q,074 0,0018 to 0,0024 7 200 0,0030 0,076 0,002 8 0,075
62 ........ 0,0024 0,063 250 0,0016 230 0,0024 0,062 0,0015 to 0,0020 7 240 0,0026 0,065 0,0016 8 0,063
53 270 0,0021 0,053 270 0,0016 270 0,0021 0,053 0,0014 to 0,0018 7 300 0,0022 0,053 ........ 8 0,053
44 ........ 0,0017 0,044 325 0,0014 325 0,0017 0,044 0,0012 to 0,0016 7 ........ ········ ........ ........ ........ 0,045
37 400 0,0015 0,037 400 0,001 400 0,0015 0,037 0,0009 to 0,0014 7 ........ ........ ......... ........ 0,038
Base = 200 mesh = 0,074 mm ~2 Base = 18 mesh = 1,0 mm ~2
a permissible variations in average openings.

Table 19. The periodic table.
-
1 2
The Chemical Elements
1
1,008 1,008 4,003
H H He
3 4 5 6 7 8 9 3 4 5 6 7 8 9 10
6,941 9,012 10,81 12,011 14,007 16,0 18,998 6,941 9,012 10,81 12,011 14,007 16,0 18,998 20,179
Li Be B c N 0 F Li Be B c N 0 F Ne
11 12 13 14 15 16 17 11 12 13 14 15 16 17 18
22,!190 24,305 26,982 28,086 30,974 32,06 35,453 22,990 24,305 26,982 28,086 30,974 32,06 35,453 39,948
Na Mg Al Si p s Cl Na Mg Al Si p s Cl Ar
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
39,102 40,08 44,956 47,90 50,941 51,996 54,938 55,847 58,933 58,71 63,546 65,37 69,72 72,59 74,922 78,96 79,904 83,80
K Ca Sc Ti v Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Br Kr
37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54
85,468 87,62 88,906 91,22 92,906 95,94 98,906 101,07 102,91 106,4 107,87 112,40 114,82 118,69 121,75 127,60 126,91 131,30
Rb Sr y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn Sb Te I Xe
55 56 57+ 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86
132,91 137,34 138,91 178,49 180,95 183,85 186,2 190,2 192,22 195,09 196,97 200,59 204,37 207,2 208,98
Cs Ba La Hf Ta w Re Os lr Pt Au Hg TI Pb Bi Po At Rn ~
ttl
87 88 59• 90 91 92 t"""
226,03 140,12 231,04 238,03 tI1
Fr Ra Ac Th Pa u VJ
>
+ 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71
s
'Tj
138,91 140,12 140,91 144,24 - 150,4 151,96 157,25 158,93 162,50 164,93 167,26 168,93 173,04 174,97 0
Lanthanides; La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
90 91 92 93
~
c:
....... 89 94 95 96 97 98 99 100 101 102 103
~
140,12 231,04 238,03 237,05
0
-..)
..... Actinides Ac Th Pa u Np Pu Am Cm Bk Cf Es Fm Md No Lvv tI1
T:
TABLES AND FORMULAE
Table 20. The Greek alphabet
The letters of the Greek alphabet, frequently used in technical terms, are given her_e for pur-
poses of convenient reference.
A alpha a N v nu n
B
°'{3 beta b z ~ xi x
r ')' gamma g 0 0 omicron 0
d 0 delta d IT 7r pi p
E epsllon e p p rho rh, r
z l zeta z r; <J<; sigma s
H '1 eta e T T tau
e (){} theta th T v upsllon ii
I iota i <I> <P phi ph
K K kappa k x x chi kh
A /... lambda 1 'I' psi ps
M µ mu m fl
"'
w omega 0
1072
Index
Note: pages 1-614 are in Volume 1, pages 615-1072 in Volume 2
A
Aandenk Activator, in flotation, 245-6
gold prospecting, 2 Activity
Abrasion breakage, 136, 182, 195 chemical, 831-2
Absolute potential for sorting, 113 coefficients, 837
Accept fraction Acts, substandard, 977
in waste sorting, 109 Actuators, 817, 818
Access electrical, 818
levels, 942-3 electro-hydraulic, 818
roadways, 700 pneumatic, 818
Accidents Adjustable diverters, 284
analysis of, 971, 975, 977-82 Adsorption
causes, 982 gold losses by, 32, 304
classification reference guide, 979-80 isotherms, 388
investigation of, 971, 973, 977, 982 Adsorption process, 32, 304, 379-81,
report forms, 977, 978-80 383-7' 458-9
statistical data, 977, 982 chemistry, 383-93, 870-7
Acetonitrile, 482, 486 economic evaluation, 413-17
Acid effect of acidity, 423, 425, 426
adsorption, 386 effect of base metals, 426-7
flotation, 263-4 effect of calcium carbonate, 435-9
pressure, preleaching, 1032 effect of calcium ions, 421
production, 723 effect of carbon size, 440-2
spillage, 986 effect of cations, 421
Acid treatment effect of cyanide concentration, 422-3,
hazards, 356-7 424-6
of carbon, 570-5, 603 effect of flotation reagents, 434
of steel-wool cathodes, 373, 896 effect of ionic strength of the
of zinc-gold precipitate, 356-7, 895-6 medium, 421, 422
vat, 356, 357 effect of mixing rate, 439-40
Acid washing effect of number of stages, 401-5
effect on elution process, 493-7 effect of physical factors, 439-43
effect on gold absorption, 427 effect of pulp solids concentration and
Acidity size, 443
effect on gold adsorption, 423, 425, effect of temperature, 422-5
426 equilibrium effects, 417-21
Aero!, 255 experimental studies, 388-93, 1049-50
Acrylamides, 287 film diffusion control, 396
Activated carbon in gold recovery gold lock-up, 401-5, 407
historyI application/ general, 379-83 inter-particle diffusion, 396
chemical and physical properties, mass transfer coefficient, 396
383-7 mathematical models, 393-9
hydrophilic properties, 385-6 branched pore, 396-7
hydrophobic properties, 385-6 equilibrium effects, 398-9
solubility of adsorbed species, 386 Freundlich, 398
stages of formation, 385 "kn", 394-5, 399, 401, 429-31
see also Carbon; Gold recovery Langmuir, 398, 422
1073
INDEX
Adsorption process (cont.) Albion fault, 41, 45

Van Deventer, 396-9 Algae, 24
mechanism, 874-6 Aliphatic alcohols, 242-3
organic molecules, 386 Alkaline flotation, 263-4
theories, 386-7 Alkalinity
Adsorption rate, gold on carbon,
measurement of, 1049-50 on-line measurement, 328, 815
Adsorption tanks, 381, 458, 461-3 Alkoxy-substituted paraffins, 242, 243
carbon distribution, 468 Allanite, 33
carbon transfer, 478-9 Allis-Chalmers Hydrocone crusher, 117
interconnections, 479 Allogenic detrital-oxide suite, 27
interstage screening, 470-8 'All-sliming' process, 121, 122
air supply, 476-7 Alphabet, Greek, 1072
capacity, 473-4 Alteration products, 15, 16, 33
cloth, 473, 476 Alternating magnetic field
layout, 471-3
of induction furnace, 367-8
surges, 477
Aluminium
layout, 479-80
cementation by, 869
mixing mechanism, 463-9
mixing rates, 439-40, 469 Aluminium sulphate, 985
Amalgam barrel, 230
requirements, 468-9
Amalgam, gold
return to service after stoppage, 468
retorting of, 352-4
Aeration
weighing of, 377
during flotation, 248-9
Amalgam plates, 121, 122, 220, 230
inadequacy of, during leaching, 304
Amalgam press, 230
Aerofoil type mixer impellers, 469
Amalgamation, 228, 229-30, 232
Aeropromoter flotation reagent, 255
Amine solvent-extraction process
Afrikander Lease gold plant, on-stream
treatment of gold-cyanide solutions,
cyanide analysis, 328
888-9
Agitation
air, 264, 295-6, 301-3 Ammeter, 192
Ammonia
mechanical, 264, 301-3
Agitator tanks, generation in electrowinning, 894
Ammonium hydroxide, in refractory ore
flat-bottomed, 301
treatment, 59
pachuca, 300-1
Analogue hard-wired control system,
Agnes Mine, 47, 53-4, 56
821-2
Air-cleaned screens, 470-7
air supply, 476-7 Analyser, Microtrac, 807
Air eliminator, 808 Analysis, of data, 272, 827
Air flow rate control, Analysis, on-line, 266
in flotation, 265 of cyanide, 814-15
of gold, 811-13
in leaching vessels, 302-3
Analysis, spectrophotometric, 814-15
Air lifts
carbon transfer, 478-9 Angle of repose
Air pollution of ore, 77
control of by use of carbon, 379 Anglo American Electronic Laboratories,
see also Pollution 328
Air Pollution Prevention Act, 698 Anglo American Joint Metallurgical
Air-sparged leach vessels, 301, 303 Scheme, 715, 716, 731-6
Air supply to leaching vessels, 303 calcine treatment, 735-6
Akins classifier, 89, 123, 150 feed sources, 733
Alarm monitoring, 822, 825 flotation plants, 733-4
Note: pages 1-614 are in Volume 1, pages 615-1072 in Volume 2.
1074
INDEX
Anglo American JMS (cont.) South African Mining Industry,

flotation tailings, 734-5 652, 1013, 1047
history, 732 Assay office
metallurgical test facilities, 1015 metallurgical testing, 1015
pyrite concentrates, 735 Assaying
Anglo American on-mine laboratories of gold, 266, 652
layout, 1015 difficulties of, with laboratory
Anglo American Research Laboratories, samples, 270
270, 328, 460-1, 1013 in ore testing, 913
electrowinning cell, 554, 556, 891, 892 instrumental methods, 1022
elution method, 481-3, 491, 492-503, precision of, 787
561, 723, 877 A.S.T.M. testing sieves, 1069-70
Anglovaal Research Laboratory, 1013 Atomic absorption analyser, 1016, 1022
flotation testing, 271, 272 Atomic absorption spectrophotometer,
Angular velocity, 205 811-13, 816
Anionic metal cyanide complexes, 331 Atomization, electrothermal, 811-13
Anions Attrition breakage, 136
adsorption of, 386, 387-8 Audiometric testing, 989-91
Ankerite, 52 Auger sampling, 710, 911
Anodic-stripping voltammetry, 816-17 Auric complex, 831, 834, 835, 836
Anolyte, 557-8 Auricyanide, 836
Anti-creep bar, 170 Aurocyanide
Antigorite, 52 elution from ion exchange, 882-3
Antimonial ore, 50, 57, 345 extraction by ion exchange resins,
Antimonates, 860-1 878-82
Antimonites, 860-1 extraction by resin-in-pulp process,
Antimony, 52, 58, 219, 346, 861 883-5
Antimony sulphide minerals, 860-1 extraction by solvent extraction
Apron feeders, 83-84, 186 process, 885, 889
Arc furnaces, 364-6, 369, 372 Aurocyanide ion, 387, 422, 834, 873,
Archaean granite-greenstone terrains, 5, 874-5, 876
6, 24, 35-7' 51, 60, 345-6 interaction with carbonaceous
Archimedes spiral, 150, 163 material, 863
Armco Autometrics Particle Size oxidation by halogens, 835
Monitor, 131, 807-9 reduction, electrochemical, 849-51
Armoured rubber hose, 188 reduction by aluminium, 841
Arsenic, 219, 236 reduction by borohydride, 842
-bearing minerals, 345, 346 reduction by zinc, 841
removal of, 350 Amous complex, 831, 834, 835
Arsenic oxide, 346, 349, 350-1 Autogenous grindability testing, 211, 914
Arsenic sulphide minerals, 860 drop tests, 210, 914, 1024
Arsenic trioxide, 348, 349, 351 Autogenous mills, 160-2, 178
Arsenides, 32 automatic control, 200-1
Arsenites, 58 design, 212-6
Arsenopyrite, 16, 20, 25, 50, 51, 52, 53, power draft, 161-2, 215
54, 57, 236, 237 power maximization, 192-3, 200-1
effect on gold leaching, 860 types of size reduction in, 195-6
roasting, 345, 346-7, 348, 350 Auto-I et filter, 341
treatment of Automatic control, 328, 815
off-gas from roasting, 351 levels, reagent, in leaching, 291
Ashing, 381 levels, solution, in precipitation, 337
Assay and Analytical Practice in the of grinding mills, 199-201
1075
INDEX
Automatic on-line reagent origin of gold in, 60

concentration measurement, 328, 815 production, 2, 4
Automatic on-line size analysis Barberton mines, mining and treatment
instruments, 131 crusher stations, 74
Automatic reagent addition control, 328 flotation methods, 235, 236-7
Automatic sampling, 766-77 mining methods, 70
Automatic spectrophotometric analysis, refractoriness of ores, 219, 228-9, 237
814-15 roasting of ores, 351-2
Autopact cleaning process for Stellar sulphide ores, 71, 127
filters, 341 transport methods, 76
Availability Barberton Mountain Land, 40-9
of machinery in gold ore treatment Barberton Sequence, 37
plants, 923-4 Barb rook fault, 41, 47, 48
Axial flow mechanical circulator system, Barge decant systems (tailings dam),
301 682-3
Barren solution gold value (in carbon-in-
B pulp treatment), 410, 411-12,
Backing plates 413-17, 419
for end liners, 167 Barring gear, 157
for grid liners, 173 Barrow, John, 1
Backfilling, 70, 112, 655, 708 Basal Reef, 2, 12, 30
Backlashing, of grinding mills, 199 Base metal oxidation, 355
Bacteria, 24 Batch flotation tests, 268
Bacterial leaching, 229, 237, 741-3 Batch leaching, 295, 296
ideal conditions, 742 Batch milling, 137
research in, 1032 Batch thickeners (collectors), 277-8
see also Leaching Batea pan, 230
Bafokeng disaster, 1974, 677-8 Beatrix gold mine, 2
Bag filters carbon-in-leach plant, 449-55, 479
dust extractor, 118, 376 elution procedure, 503, 505, 508, 509,
fume recovery, 651-2 511, 513
on storage tanks, 292 interstage screens, 471, 472
Bag houses, 349, 351 Beisa gold mine, 2
Ball mills, 123, 124, 158-60 carbon elutriator, 512, 513
automatic control, 200 reverse leach process, 436
design, 212-16 Belt conveyors, 76, 87-91
power maximization, 191-3 Belt feeders, 83, 184, 186, 819
type of size reduction in, 195-6 Belt filters, 307-8, 309, 311, 312, 317
Balls for gold slime dewatering, 360
description, 159, 179 plant design, 950-1
mass, in SAG mill load, equation, 217 Belt scrapers, 89
rationing, 183-4 Belt trippers, 73, 74, 79, 90
size in relation to optimum milling, Belt washing, 88-9
143-6, 183-4 Bentonite materials, 322
wear, equation, 193-4, 217 Berm penstocks, 684-5, 701-3
wear, theories, 182-3 Betty, W.K., 333
Bi!rberton, eastern region, 47-8 Bias
Barberton greenstone belt, 39-60 in sampling, 746, 751, 759
Barberton mines, 35-60 in sub-sampling, 780
discovery, 1, 35 tests of, 791
mineralogical aspects, 49-60 Bimodal size distribution, 136
Note: pages 1-614 are in Volume 1, pages 615-1072 in Volume2.
1076
INDEX
Bins, 80, 100 Bowl liner (cone crusher), 102, 103

size of openings in, 185 Bracken ore, 21
Bioleaching, research, 1014, 1025 Braggite, 34
Biotite, 52 Brannerite, 21, 33
Bird Amygdaloid, 12 Breakage
Bismuth, 52 as a first order process, 147
Blaauwbank area causes, 135-6
gold prospecting, event, 133, 146
Black Reef, flotation of, 23 7 related to ball size, 143-4
Black Reef Quartzite Formation, 13, 14, British Standard
25, 26 testing sieves, I 069-70
Black sand concentrate Brittle fracture, 128-9, 131, 133-5
in mill liners, 173, 176, 178, 228 Bromoform, 916
Blake-type jaw crusher, 98-100 Bronkhorst, J.G.S., 1
Blast furnaces Brown's tanks, 122
capacity, 648 see also Air-sparged leach vessels
description, 648 Bubble tube, 265
furnace charge, 648 Bucket-wheel excavator, 716-17
matte, 648 Buffelsfontein gold mine, 257, 259
modernization, 652 Building ·construction details, 953-6
slag, 648 Bulk density
smelting, 644-8 of mill loads, 191, 194, 195
Blending, 643-4 Bulk flotation, 238
Blinding, 313 Bulk tankers
Blow-off air pressure, 311 for delivery of cyanide, 292, 293
Blyvooruitzicht gold mine, 2, 15, 21 Bullion
residue retreatment plant, 715, 716-19 bars, identification & marking, 374
waste rock retreatment, 738-40 bars, weighing, 377, 619
'Boiling' (in flotation), 250 composition following calcined slime
Bolted block liners, 171-3, 175 smelting, 372
renewal, 177-8 composition following direct smelting,
Bolted grid liners, 173-4 369
renewal, 177-8 crude, casting, 620
Bond, Fred, C. crude, sampling, 619-20
Bond Index for Witwatersrand ore, fineness, 373, 619
212-13 melting, 366-7, 619, 621
Grindability Index, 212-13, 216, 914 poured into ingots, 368-9, 620
Third Theory of Comminution, 140-3, production from calcined slime, 372
212, 213 production from steelwool cathodes,
Bond Work Index determinations, 1024 372-3
Booysens Shale, 12 refined, assaying, 623
Borates refined bars, labelling, 623-4
formation of, in slag, 371 refined bars, monetary fineness limits,
Borax 625
flux, 230, 231, 369, 371, 373, 898 refined bars, monetary mass limits,
slag, composition of, 354, 370, 898 624-5
Borehole cores refined bars, sampling, 623, 778
as ore testing samples, 910-11, 1025 refined, casting, 623
limitations for ore testing, 910 Bunds, 948, 949, 953
preparation for ore testing, 912 Bunkers, 76-8
Bowl (cone crusher), 102, 103 Buoyancy
Bowl classifier, 124 effects, 219
i077
INDEX
Buoyancy force adsorption, 435, 863

equation, 147 build-up in carbon-in-solution, 603
Butler-Volmer equation, 560-1 levels on elution, 505-7
Button, Edward, 1 loss during thermal regeneration,
Button moulds, 369, ·377 574-5
Buttons, bullion, recording, 377 source of contamination, 437-8,
Buttressing, tailings dam, 692-3 571
By-passing, in electrowinning cells, 892-3 Calcium carbonate scale
By-products removal by acid treatment, 317, 319,
blending, 643-4 323, 956
cupellation, 648-9 removal by go-devils, 956
pan furnace smelting, 649-51 Calcium cyanide, 291, 292, 305
sampling, 641-3 Calcium fluoride (fluorspar), 371
sintering, 643-4 Calcium hydroxide, 291
smelting, 645-8 Calcium ions, 32, 246
By-Products Ltd., 616 Calibration
of instruments, 803
c of massmeter, 115, 187
Cable torque thickener, 282 Camel-back dumping frame, 112
Cable trenches, 937 Candle pressure (Stellar) filters
Cake discharge systems, 309-11 for clarification, 320-1, 322
filter tailing, 309-11 for zinc-gold slime recovery, 333, 335,
gold slime, 341-2 337, 340, 341-2, 355
Calcination, chemistry, 896 Capacitance probes, 806
Calcine, pyrite, 235, 236 Capital cost
cyanided for gold recovery, 253, 255 comparison, for carbon-in-pulp, 382
effect of quenching, 346 effects of number of adsorption
recovery from roaster off-gases, 350-1 stages, 401-2
treatment, carbon-in-pulp, 455-6 estimates, 918
treatment, cyanidation, 735-6 in relation to machine size, 925
water on pyrite flotation, 248 in relation to plant size, 918
Calcined zinc-gold slime Carbon
composition of, 370 abrasion loss, 401-5, 411
fluxing of, 371-2 effects of impeller speed, 469
production of, 361-3 effect of impeller wear, 448
smelting of, 371-2 acid treatment, 537, 570-5
use of submerged arc furnace for calcium build-up, 570-1
smelting, 364 conditions, 571-2
weighing before smelting, 377 contact time, 572-3
Calcining, of steel-wool cathodes, 372-3 effect on adsorbed gold & silver,
Calcining process, for pyrite (roasting), 572
31, 59, 345-52 neutralization of excess acid, 574
furnaces, 348-50 process position, 573-4
Calcining trays, 361, 372, 373 strength, 572
Calcite, 46, 52, 53 temperature, 573
Calcium activity, 429-39
analysis of deposits, 439 effect of calcium carbonate on,
build-up, rate of, 436, 570-1 435-9
contamination of carbon, 381, 405, effects of carbon size, 440-2
426, 435-9, 570-1 effect of grind fineness, 443
effects on carbon effect of impeller speed on, 448, 469
1 ()"70
iVIU
INDEX
Carbon, activity (cont.) kinetics, 389, 871-2

effects of mixing rate on, 439-40 measurement of loading capacity,
effect of poisoning on, 431-5 1049-51
effect of pore loss on, 431 mechanism, 387-93, 872
effect of pulp density, 443 thermodynamics, 389, 872-4
effect of regeneration on, 431 gold recovery by, outline, 379-83
effect on operations, 429-30 granules, 387
meter, 815-17 in ores, 14, 21, 24, 32, 33, 59, 305
adsorption of organics, 385 internal area, 384, 385
ash content, 537 inventory, factors affecting, 399-417
ashing, 381 loading
attrition resistance, assay value, effects, 406-13, 417-21
measurement of, 1050-1 effect of cations in solution, 872-3
bed equilibrium loading, 399, 417-21
expansion, 523-4 factors affecting capacity, 874
flow resistance, 523-4 kinetics, 872-4
friction loss, 523-4 loading rate
breakage, 575-84 effect of missing rate, 439-41
decrease with time, 581 effect of carbon size, 440-2
determination of, 575 effect of pulp solids, concentration
during mixing, 581-3 & size, 442-3
during regeneration, 583 loss
effect of carbon type, 581 distribution in plant, 583-4
in carbon-in-solution systems, moisture content
601-2 determination of, 1051
in eductors, 581 oxides, surface, 385-6
in pumps, 575-81 oxygen, effect on adsorption, 427-9,
main causes, 584 437
minimizing, 581 particle size distribution
points of loss from carbon-in-pulp determination, 1051
process, 57 5 poisoning
size distribution in pulp streams, by calcium carbonate, 435-9
575 by organics, 381, 405, 411, 431-5
size reduction mechanisms, 575 effect of adsorption contact time,
bulk density, measurement of, 1051 405
calcium contamination, 381, 421, 426, in dump retreatment, 457
435-9 pore structure, 384-5
catalytic ability, 386-7 powdered, 387
complex organic polymer, 383-4 use in scavenging dissolved gold
concentration meter, 815 from residue pulp, 257
contaminants, 381, 405 reducing ability, 386, 387, 392
counter-current movement, 381 regeneration, 381, 431, 530-50, 575
density, 465 chemical, 530
determintion of adsorption thermal, 531-50
parameters, 1049-51 desirability of, before elution,
distribution in adsorption vessels, 575
445-7 dewatering before, 542
elution chemistry, 876-7 equipment, 541-50
elution profiles, 509 furnace atmosphere, 534-7
see also Elution inorganic constituents, effect,
gold adsorption properties, 387-93 537, 574
history of use, 380 kinetics, 534-6
i079
INDEX
Carbon, regeneration (cont.) equipment, 457-80
principles, 531-4 fine carbon recovery, 480
residence time, 537-41 interstage screening, 470-8
Rintoul kiln, 547-50 layout, 948-9
rotary kilns, 541-7 mixing in adsorption tanks, 463-9
relative density, 465 oversize particles, difficulties caused
residence time in adsorption vessels, by, 459-60
407-11 prescreening, 460-1
screening, 511 spillage handling, 480
settling rate, 468 treatment of calcine by, 735-6
soaking, prior to elution, 498-9 Vibro-Energy mills in, 735
solubles Carbon-in-pulp process, 379-83
determination of, 1051 advantages of, 382, 383
spillage, effect on activity, 449 calcine treatment by, 455-6
surface capital & working cost comparisons,
oxides, 385-6 382
positively charged sites, 389 carbon loss in, 575-84
Zeta potential, 389 chemistry of, 870-7
testing, 1019 design and operation, 399-417
transfer, 447-8 Discounted Cash Flow calculations,
use for initiating arc in smelting, 372 413-17
volatile content, measurement of, 1051 distribution & settling rate of carbon
see also Elution; Regeneration in adsorption tanks, 468
Carbon-bearing mica, 52 dump retreatment by, 457, 727-31
Carbon-in-leach plant economic evaluation, 413-17
East Rand Gold & Uranium, 253, 723 effect of flotation reagents, 434-5
Carbon-in-leach process, 381, 449-55 effect of number of stages, 401-5,
advantages & disadvantages, 455 416-17, 463-4
carbon inventory, 452 effect of pulp solids, concentration &
carbon transfer, 452-4 size, 443
effect of leaching rate, 454-5 effect of retention time, 461-3
mathematical modelling, 449 effect of throughput rate on
number of stages, economics, 449-52 efficiency, 383
research in, 1028, 1032 engineering maintenance requirements,
Carbon-in-pulp plant 383
adsorption tanks, 461-3, 479-80 Internal Rate of Return, 413-17
carbon transfer, 478-9 inter-stage screening, 381, 470-8
computer flowchart, 400 kn model, 394-5, 399
design and operation, 399-417 mathematical modelling, 393-9
effect of carbon distribution, 445-7 pilot plants, 393-4, 1028-9, 1032
effect of carbon poisoning, 433-9 pre-screening, 459-61
effect of carbon transfer method, rate constant, effect, 429
447-8 rate equation, 394-5
effect of head grade, 444-5 research, 1028, 1032
effect of impeller blade wear, 448-9 silica grit, removal, 460-1
effect of loading rate constant k on tail solution assay value, 413
operations, 429-30 treatment of flotation concentrates,
effect of number of stages, 401-5, 456, 729
413-17, 463 treatment of underground slime, 457
effect of spillage return, 449 working cost comparison with
equilibrium effects, 417-21 conventional process, 382
11"\Q{\
lVOV
INDEX
Carbon-in-pulp testing, 1024-5 in adsorption theory, 386

Carbon-in-solution, 584-608 Caustic cyanide solution, 381, 481, 498
acid treatment, of carbon, 603 Caustic soda
adsorption of competing species, 603-8 in gravity concentrate treatment, 230,
adsorption systems, 585-99 232
carbon breakage, 601-2 precautions, 985
carbon fluidization, 600-1 Cementation, 331-43
closed tank system, 588-90 by aluminium, 869
deep fluidized beds, 590-1, 592-3 chemistry of, 866-9
design considerations, 599-608 effect of cyanide concentration, 868
effect of calcium build-up, 602-3 effect of dissolved oxygen, 868
effect of flow rate on adsorption of effect of impurities in feed solution,
competing species, 604 869
effect of pH on adsorption of effect of insufficient free cyanide,
competing species, 605, 607, 868-9
608 effect of lead ion, 868
fluidized bed expansion, 601 effect of pH, 868
gravity cascade system, 585-7, 590 Central goldfields, 5
NIMCIX column system, 594, 595-8, Central launder screens, 471-3
603-6 Central Rand Group (of sediments)
packed bed system, 591-2, 595 geology, 5, 6, 7, 12, 13
plants, 584-5 uranium minerals, 33
pressure drop across packed beds, Centrifugal acceleration, 204-5
599-600 Centrifugal classifiers, 150-2
process comparisons, 598-9 design, 207-8
relative viscosity of suspensions, 600 modelling, 205-7
types of solution treated, 584-5 theory, 204-5
Carbon Leader Reef, 2, 12, 14, 15, 21, Centrifugal force
108-9, 237 in classifier, 151, 204
Carbon tetrachloride, 916 Centrifugal pumps
Carbonaceous material for classifier feed, 187
as precipitants of gold, 58 for thickener underflow, 283
constituents, 862 selection, 967-9
effect of, on gold leaching, 862 Centrifuging, in grinding mills, 135, 191
recovery by flotation, 237 Chain curtains
treatment of ores containing, 862 of chute feeders, 80
Carbonium sites, 388, 391 Chain feeder, 80-1
Carboxylic surface oxides, 385 Chalcocite, effect in gold leaching, 859
Carletonville, 12 Chalcopyrite, 13, 14, 16, 20, 23, 25, 26,
Cascading, of grinding mill charge, 135 52, 54, 57' 237' 859
Cash flow, 927-33 Chamber of Mines of South Africa
Casing, of conveyor belting, 88 Industrial Hygiene Branch, 991
Casting, continuous, of gold coinage Safety Code, 91
alloy, 638-9 Chamber of Mines flotation cell
Cataracting, of grinding mill charge, 135 (COMRO cell), 272-3
Catchment paddock, 701 Check-in/check-out system, of
Cathodes accounting, 789-90
electrowinning, 554-67 Checklist, general inspection, 974-5
smelting of, 372-3 Check sorting, 111
Catholyte, 557 Chemelec electrowinning cell, 891
Cations Chemical analysis, precision
effect in adsorption, 421 of gold-bearing materials, 787
i081
INDEX
Chemical composition configuration for flotation, 263

of gold, 28-30, 55-7 Circular screens, 471-2
Chemical treatment Circulating load
to remove coatings from gold, 59 in crushing, 104
Chemicals, hazardous in grinding, 197-9
handling, 985, 986 in plane table concentration, 233
safety measures, 1006-9 in ROM autogenous milling, 193, 198
Chemwes ratio, formula, 209
flotation plant, 236, 260 Clarification
residue retreatment plant, 715 efficiency of, effect on zinc-gold slime
Chert, 39, 44, 45, 47, 49, 52 composition, 370
blocks, as mill liners, 121 necessity for in precipitation, 333, 334
Chezy formula, 955 of precipitation feed solutions, 317-23
Chipping breakage, 136, 182 types, 318-23
Chloride leaching, 866 use of thickeners for, 279
Chloride solution Clarity
anodic behaviour of gold in, 844-5 of filter filtrate, determination, 313
electrodeposition of gold from, 853-4 Classical comminution laws, 139-40
Chlorination Classification, 147-52
aqueous, for treating carbonaceous control, 200
ores, 862, 866 efficiency, formula, 209-10
testing of, for leaching, 1032 formulae, 209-10
Chlorine tests, 1024
gold leaching, 866 types, 149-51
gold refining process (Miller process), Classifiers, 147-52
615, 621-3 Akins, 123
Chlorite, 15, 20, 25, 31 bowl, 123
Chloritoid, 15 centrifugal, 150-51
Chocks, mill liner, 170 cone, 149
Choke feeding, of crushers, 105, 117 cyclone (hydrocyclone), 124, 150-1,
Chrome steel (for mill liners), 170, 171, 152, 159
175 access, 943
Chromene groups, 386, 391 amalgam recovery, 230
Chromite, 16, 20 design, 207-8
Chromium, 27 modelling, mathematical, 205-7
adverse effect in electrowinniug, 567, 894 start-up, 960
Chute tests, 1024
dead-box, 935 theory, 204-5
feeders, 185 desliming, 73, 74, 94
for ore transport, 85-7, 935 rake, Dorr, 149
movable, 115 spitzkasten, 149
rectangular, dimensions, 85 Cleaner cells, 255, 263
washing in, 92 Cleaning table, for bullion bars, 373
Chute feeders, 80-1 Cleanliness
radial doors of, 80, 81, 184 in the smelthouse, 377
Chute slopes, 86 Clean-up
CIL of zinc-gold precipitate, 340-2, 355-6
see Carbon-in-leach Clearances, access, 941-2
Cinnabar, 50 Clinker, gold adsorbed by, 32
Circuit Closed circuit
complexity, 222 tertiary crusher, 73, 74, 104
Note: pages 1-614 are in Volume 1, pages 615 1072 in Volume 2.
1 ()Q')
.!VO""
INDEX
Closed-circuit milling, 143, 147, 159, Combings, grizzley, 94

197-9 Comminution
automatic control, 199-201 energy, 129, 139-43
Closed tank systems laws, 139-41
for carbon-in-solution, 588-90 testwork, 1024
Clothing, protective, 971, 991-2, 997, 998 third theory, 140-1
in smelthouse, 375, 377 Communications, 982-3
Cloths group meetings, 982-3
filter, 306, 307, 311-13, 317 personal, 983
acid treatment, 313, 317 Complex, chemical, 831-2
brushing, 317 Complex mass balance (flotation), 267
screen, 96-7 Composite loads, 123, 193
Coagulants, 287, 322 Composite milling, 161, 162
Coal Composite Reef, 27
absorption of gold by, in dumps, 32 Compound interest, formula, 928
Coarse gold particles, 219, 222 Compression breakage, 136
in laboratory flotation testing, 271 Compression pulses, 128
Coarse ore Compressive forces
feeders, 80-5 in brittle fracture, 128
storage, 76-80 in crushers, 98-9, 102, 105
transport, 75-6, 935 in grinding mill, 135-6
Coated gold, 31-2, 58-9, 219, 304, 857, in thickeners, 285
865 Compressive strain energy, 128-9
Cobalt, 33 Computer-aided design, 941, 1031
adsorption on carbon, 426 Computers
poisoning of resin, 885 control systems, 822-8
recovery, solvent extraction process, centralized, 822-3
885 distributed, 823-8
Cobaltite, 16, 20, 52, 237 digital
Coconut shell carbons, 384, 388 for automatic control, 822-8
Cocopans, 75 for automatic mi!iing control, 201
waste disposal by, 111-12 laboratory, 1025-6
Coe and Clevenger method mini, 822-3
in thickener design calculations, 324, process control, 802
326, 1042-3 use in plant evaluation, 790-1
Coffinite, 16, 33 Concaves, crusher, 101, 102
Coin blanks Concentrate regrinding, 160
casting, continuous, 639 Concentrate treatment, 288-31, 232
grinding, 639 by amalgamation, 229-31
history, at Rand Refinery, 637-8 by direct smelting, 369
inspection, 640 Concentrating process of gold in ore
production of, 616, 638-9 formation, 55
strip rolling, 639 Concentration tests, 916
weighing, 639 Concentrator
Collectors, flotation, 236, 239-42, 555 primary, 221, 224, 232
blended, 253 secondary, 221
synergism, 242 see also under specific types
Collectors, ore (batch thickeners), 277-8 Conc~ete bins, 80
Colloids, 336 Condenser
Colour codes, 1003 of retort furnace, 352, 353
Colour perception test, 987 Condition rating programmes, 973
Column flotation cell, 273, 274 benefits of, 973
1083
INDEX
Condition rating programmes (cont.) ore treatment plants, 793-4
checklist, 973 historical background, 794
teams, 973 PID, 800-1
Conditioning of pulp, 257, 264-5 proportional, 800-1
mechanical, 264 real-time automatic, 793
Conditions, working, substandard, 977 sequencing, 793
Conductivity probes, 806 supervisory, 793
Cone classifiers, 121, 122, 149 system design, 794, 795, 797-801
Cone crusher, 102, 103-6, 117 architecture, 819-28
secondary, 73, 74 bottom-up approach, 797-8
tertiary, 73, 74 cascade configuration, 798
Confidence Reef, 1 feed back, 797-8
Conglomerates, 13-17, 25, 26, 27 feed forward, 797, 798
Conical moulds, 372 Inverse Nyquist Array method,
Caning and quartering, 913 799-800
Consort Bar, 45-6 multiple actuator, 798
Construction multivariable controllers, 799-800
buildings, 953-6 philosophies, 819-28
fault rectification, 959 top-down approach, 797, 798-800
materials, 953-6 systems
Contact time analogue, hard-wired, 821-2
see Residence time computer, centralized, 822-3
Contaminants computer, distributed, 823-4
in carbon-in-pulp process, 381, 405 Cooke gold plant, 825-7
Continuous casting machine, 638-9 CYGNUS, 826, 827
Continuous leaching East Driefontein, 824-5
see Leaching, continuous V aal Reefs, 827
Continuous milling Controllers
see Milling, continuous multi-variable, 799-800, 827
Continuous steel belt calcining furnace, PI, 801
361-3 programmable logic, 823, 825
Continuous thickener three-term (PID), 798, 800-1
see Thickener, continuous tuning, 801
Continuous vacuum filtration, 305-17 tuning method, Ziegler and Nichols,
Continuously stirred tank reactor theory, 801
261-2 Controls
Control on engineering, 1001
actuators, 817-19 on purchasing, 1001
automatic, 820-1 Conversion factors
of crushers, 116-17 common, 1054
of milling, 199-201 Conversion tables
continuous, 793 area, 1056
objectives, 795-6, 797 density, 1061
principles, 795-801 energy, 1060
variables, 796-7 length, 1055
derivative, 800-1 mass, 1059
economics, 794 power, 1061
final elements, 817-9 pressure, 1060
flotation plant, 265-6 rate of flow, 1058
logic, 793 velocity, 1058
manual, 793 volume, 1057
Note: pages 1-614 are in Volume 1, pages 615-1072 in Volume2.
1084
INDEX
Conveyor belts, 87-91 Countercurrent decantation, 279

loading and unloading, 90 Countercurrent flow, carbon-in-pulp
pebble transport, 115 process, 381, 393-4, 458-9
transport by, 76, 935 Covalent bonding in gold compounds,
walkways, 942 831-2
washing on, 92 Covariance, 791
waste disposal, 111, 112 Crack network, 128-9, 140, 182, 195
Conveyor belt guards, 91, 117 Crack tips, 128, 140
Conveyor, pan, 186 Crackle breccia, 45
Conveyor tripper, 76-7, 90 Cratonization, of Kaapvaal craton, 39
Conveyors Creosote, use in flotation, 255
intermediate, 112 Cresylic acid, use in flotation, 242, 255
Cooke gold plant Critical level of stress (in comminution),
control system, 825-7 129
Cooper, R.A., 33 Critical parts inspections, 975
Cooperite, 34 Critical-size particles, 196
Copper, 30, 52, 56 Critical speed of mills, 135
adsorption on carbon, 426-7 formulae, 135, 1068
behaviour during silver chloride Critical task
leaching, 632 analysis, 996
deposition of, during electrowinning definition, 993
of gold, 552 identification, 993-4
in gold bullion, 369 inventory, 991, 993, 996
in gold particles, 17 lesson plan, 996
in zinc-gold slime, 370 probability of loss, 994-5
mine tailings, size distribution, 668-9 procedures, 995-6
recovery, by solvent extraction repetitiveness, 994-5
process, 885 rules, 996
selective elution of, 488-9, 497 severity, 994
Copper cyanide complexes, 59 standards, 995-6
Copper sulphate worksheet, 997
in flotation 245, 246, 253, 257 Crookes, Sir William, 33
in on-line cyanide monitor, 814-15 Cross-stream sampling cutters, 771-3
Copper sulphide minerals, 16, 20, 32, 50 Crowe-Merrill precipitation process, 333-7
effect of leaching, 859-60 on-line gold analyser for, 811
Corduroy blankets (strakes or tables), 59, practice, 338-41
220-1, 223-4, 229, 233 Crowe tanks, 333-6, 341
Coronation Reef, 7 Crown Mines (RM3)
Correlations residue retreatment plant, 236, 724
among gold and heavy minerals, 27 Crowning
Council for Mineral Technology (Mintek), of conveyor belt pulleys, 88
50, 1013, 1021, 1022, 1026-9 of metal surface during induction
divisions, 1027 melting, 368
facilities, 1028-9 Crucibles, 364
Gold and Uranium Research Advisory capacities, 364
Committee, 1027 liners, 364
project categories, 1027-8 smelting in, 371
research objectives, 1027 Crude bullion
sponsorship conditions, 1027-8 see Bullion
Technical Advisory Committee, 1027 Crushed rock sampling, 770, 913
Council for Scientific & Industrial Crusher bin, secondary, 73, 74
Research (CSIR), 1022 Crusher, gyratory, 100-6, 121, 122
1085
INDEX
Crusher, Hydrocone, 117 concentration, effect on adsorption,

Crusher products 422-3, 424, 426, 427
sampling, 770 concentration, effect on calcium
Crusher station, 72, 73, 74, 91, 98 fouling of carbon, 437-8
capital cost compared with ROM concentration, effect on cementation,
milling, 161, 929 868-9
running time, 923 concentration, effect on elution,
Crushers 485-6, 499-502
access levels, 943 consumption rate, 295
capacities, 106 data sheet, 986
control of, 117 decomposition, 486-8, 502
protection, 106-7 levels in precipitation, 337
start-up setting, 960 on-line analyser, 328, 814-15
types, 98-105 safety precautions, 292-3
Crushing, 72-4, 98-107 solution, anodic behaviour of gold in,
boundary with grinding, 127-8 846-8
closed-circuit, 104 solution, dissolution of zinc in, 848-9
pendulum, 210 spillage, 986
primary, 73, 74, 100 Cyanide monitor, 294, 328, 814-15
secondary, 73, 74, 103 CYCAD monitor, 294, 328, 815
tertiary, 73, 74 Cyclone classifiers, 150-2
Crushing breakage, 195 access, operating, 943
Crushing, sample, 913 capacity, 207
Cumulative size distribution, 131-5 design, 207-8
see also Size distribution feed dilution, effect of, 205
Cupe! furnace, 648-9, 650, 651, 652 feeding, 18 9
Cupellation, 648-9 gold particles, size in underflow, 21
Cushioning, in grinding mills, 198 mathematical models, 200, 205-7
Cusum techniques (for quantifying open top, 208
accumulations), 790 operation, 198-9
Cyanamid Aeropromoter 208, 255 overflow density, for flotation, 266
Cyanidation, 121, 122, 235, 277, 291-305 theory, 204-5
calcine, 228-9, 257 two-stage, 151
chemistry, 836-49, 854-63 underflow, effect on mill operation,
design calculation, 326-7 198
gold losses, 304-5 see also Hydrocyclone classifiers
intensive, 230, 857 Cyclone off-gas treatment system, hot, 351
mechanism of, 854-7 Cyclone system of tailings dam
on-stream analyser, 328, 811-15 construction, 671-3, 677, 683, 687
plant layout, 946-8 Cyclone, thickener guard, 945
residues, flotation of, 239 Cycloning
research, 1014-15, 1032 preparation of backfill, 708
tests, 914-15, 1024-5 Cyclosizer, 1024
pressure leaching, 1032 CYGNUS control system, 826-8
rolling bottle, 1046-8 Cylpebs, 161
short-circuiting, 298-9, 1019
Cyanide D
additional methods, 293-4, 997 Daggafontein, flotation plant, 235
bulk specifications, 292 Dart valve, 265
bulk, transport, 292 Data
concentration, 291-2, 331-2 adjustment, 791
1086
INDEX
Data (cont.) flowsheet, 917-26

analysis, 746, 791 pilot plants, 1031
highway, 823-4 plant
logging, 823 detailed, 941-57
preparation, 827 general arrangement, 933-41
redundant, 791 process, 917-33
use in non-linear optimization, 791 Design criteria, 917, 923
Davcra cell, 273, 274 Design metallurgist
Davis, E.W. status, responsibilities and qualities,
ball wear theory, 182-3 908
Davy McKee contactor, 463 Detergent, use in amalgamation, 230
Dead-box Detonators
chute, 87 regulations for disposal of, 106, 117
mill feed hopper, 166 Detrital gold, in Witwatersrand reefs, 17,
Dead load, 77, 79, 80 25
De-aeration Development, mining, 69
prior to cementation, 333-6, 841, 866 Development waste, 70
Decant system Dewatering
tailings dams, 657, 678-9, 680-3, 684, of pulps, 277-90
703-5 of zinc-gold slime, 357-60
Decantation Dewatering screen, 94
dewatering of gold slime by, 357 Dextrin, as depressant in flotation, 247
slime treatment process, 707-8 DO nett (nett dissolved gold loss)
Decanter head, in slime collectors, 27 8 determination of, 328-30
Decking, conveyor, 87, 88 Diameter wear rates
Decks, vibrating screen, 93, 95 of grinding balls, 193-4
Deelkraal gold mine, 2 Diamonds
Deep fluidized bed system in Witwatersrand reefs, 16
for carbon-in-solution, 590-1, 592, 593 Diamond mine tailings
Degreasing fluid, 985 size distribution, 668, 669
Deister table, 233 Diaphragm pumps, 283
Delay time Diaphragm seals (for mill feed hoppers),
in particle flow through mill, 138 165-6
Delivery head, calculation of, 967 Diaphragm valves, 188, 817
Delkor Linear Screen, 460 equivalent friction loss, 968
Density Diapirs, 41
cyclone overflow, effect on underflow, Diatomaceous earth (filtration aid), 319,
198 359, 370
Density meters, nuclear, 786-7, 807 as component of smelting flux, 371
calibration, 786-7 Diesel locomotives
Denver flotation cell, 260 on mines, 75
laboratory cell, 271 Differential size distribution, 131
Department of Minerals & Energy see also Size analysis
Affairs, permit required from, 666 Digicon controller, 201
Department of Water Affairs, Digital computers, 201
permit required from, 666 in process control, 822-7
Depressant, flotation, 246-7, 255, 257 Digital electronic techniques
Derrick screens, in CIP plants, 460 in mechanized sorting, 110
Descaling pistols, for cleaning bullion, in plant control, 822-7
373 Digital terrain model, 714
Design Dimethylformamide (for gold elution),
computer-aided, 941 482
1/"\0"7
!VO/
INDEX
Dips, of Witwatersrand reefs, 69 Drag rake thickener, 280, 281, 282
Dipstick, for mill load level measurement, Drainage grid
192 of rotary drum filter, 311
Discharge cone, of thickener, 283 Drainage systems, of continuous vacuum
Discharge, grinding mill filters, 311-12
density, 191 Drainage systems, tailings dam
mass flow rate, 207 sides, 693, 695, 701
peripheral, 139, 153, 169 underwall, 657, 685, 700
pumping of, 187 upper surface, 701
screens, 139, 153, 168-9 Drilling, rock, in mining, 69
aperture sizes, 169 Drive (mine development), 69
open area, 169 Drives
types, 153-4 grinding mill, 152
Discharge, plough, of conveyor belts, 90 thickener, 281
Discharge rate variable speed, 819
of crushers, determination, 116 electric, 819
Discounted Cash Flow, method, 927-8 hydraulic, 819
application, 928-32 Drop testing method (for pebble
precautions in using, 933 competence), 210, 914
Disintegration cloud, 129 Drum feeders
Displacement force for controlling ore flow, 84, 185
in hydrocyclone, 204 for grinding mills, 164
Disposal of reject rock, 111-13 Drum filters, 305-7
Dissolution rate cake discharge systems, 309, 311
of gold, 32, 295-6, 304, 846-7, 854-63 drainage systems, 311
Dissolved gold loss feed systems, 309
determination of, 328-30 filtrate systems, 309
Distomat, 712 plant layout, 949-50
Distribution box, thickener feed, 284 zinc-gold slime, 358-60
Distribution modulus, 132, 133, 137, 141 Dry solids mass balance
Distributors, pulp, 188-9 flotation, 266-7
Disturbances, in plant operation, 796, milling, 920-2
797-8 Dump cyanidation
Dithiophosphate, 25 3 testing of, 1032
Diverters, thickener feed, 284 Dump leaching, 740-3
Dominion Reef, 5-8, 14, 25, 27, 33 Dump retreatment operations, 236, 253,
Donor atoms, 832-3, 836 457, 715-40
hard, 833 Dump volume, calculation of, 711-15
soft, 833 see also Residue dumps; Sand dumps
Door covers, use in mill relining, 177 Durand' s 'F' factor (in pipeline design),
Dorr agitators, 295 963-4
Dorr rake classifiers, 94, 122, 123, 149 Dust control, 117-19, 1002-3
Double-deck vibrating screen, 96 Dust extraction systems
Double vee clarifier, 321-2 for crusher plants, 117-19
Downflow sand bed clarifier, 318 for cyanide storage facilities, 292
Draft tube circulators, 301-3, 463, 465, for roasting plants, 349-50
468, 469 for smelthouses, 376, 952
Draft tube recirculator tank, 302-3 Rand Refinery, 651-2
Drag force Dust loss in sampling
equation (Stokes' Law), 147-8 precautions against, 912
in centrifugal classifier, 150-1, 204 Dust suppression, 117-19
mm
1.l.VVU
INDEX
Dusty atmosphere on mines, 75
medical examination for workers in, Electric power
987 failure of, emergency action, 986-7
Dykes, geological, 7 restarting adsorption vessels after,
Dynamic head, on pumps 468
calculation of, 967 reticulation in plant, 937
Dynamothermal metamorphism, of Electrical equipment, safeguarding of,
Barberton greenstone belt, 39 1004
installations, safety, 1010
E portable, safety, 1009
Ear muffs, 119, 989 Electricity supply, to plant, 93 7
Ear plugs, 119, 989 Electrodes
East Driefontein gold plant in arc furnace, 364, 365, 372
control system, 824-5 in electro winning process, 554-63, 567,
East Rand Gold & Uranium Company 568, 570, 633, 635-6, 891
(Ergo), 236, 252-3, 260 Electrolytes
carbon-in-pulp plant, 479 effect on adsorption of aurocyanide
metallurgical laboratory, 1015 on carbon, 421, 422
residue retreatment plant, 715, 719-24 in electrowinning, 551-2, 555-9, 566,
East Rand goldfield, 12, 13 567, 569-70, 627-8, 634-6, 890-3
Eccentric (crusher), 98, 101-4 Electrolytic gold refining, 626-8
see also Crushers Electrolytic precipitation, 332
Economics, of waste sorting, 113-14 see also Electrowinning
Eductors, for carbon transfer, 581 Electrolytic silver refining, 633-7
carbon breakage in, 581 Electromagnets, 107
Edwards roaster, 59, 228, 253, 348-9 Electromechanical belt massmeters, 782-5
Eersteling Electron microprobe analyses, 28-30,
first South African gold mine, 1 34-5, 51, 56, 911
Effective cell volume, 261 Electronic distance measuring, 712
Effective rougher cell mean residence Electrostatic precipitators
time, calculation, 261 dust extraction, 118, 349, 350-1, 376,
Effluent trench, tailings dam, 628-30
maintenance, 700-1 fume extraction, 350-1, 376, 628-30
Egoli residue retreatment plant, 716 Electrowinning, 230, 381, 481, 550-70,
Elandsrand gold mine, 2 890-4
ball milling after ROM milling, 124 anode reactions in, 894
Elastometer-lined pumps, 187, 188 application in carbon-in-pulp process,
maximum impeller tip speed, 969 550-70
ELCON supervisory system, 825 cathode reactions in, 894
Electric furnace chromium, effect of, 894
bullion melting, 621 efficiency, effect on Zadra elution, 507
bullion refining, 621 electrochemical reactions, 551-2
by-product smelting, 652 electrode reactions in, 893-4
calcining, batch, 361 extraction efficiency, 890-1
calcining, continuous, 361-3 generation of ammonia, 894
coinage strip production, 639 history of use, 332, 550
holding, refined gold, 623 passivation in, 894
induction, 367 redox system, effect of, 567
power supply, 625 research on, 1032
reverberatory, 364 temperature, effect of, 894
silver melting, 633, 637 Electrowinning, cell design
Electric locomotives AARL cell, 556, 891
1089
INDEX
Electrowinning, cell design (cont.) geometry, 497, 508

allowance for base metals, 570 gold plating in, 520
anode materials, 562 heat insulation, 517
cathode materials, 561-2 materials of construction, 518-22
cathode thickness, 560-1, 569 rubber lining, 518
cell comparisons, 559-61 upflow system, 522-3
cell sizing, 569-70 volume, 513-14
characteristic length, 569 copper, selective, 488-9
Chemlec cell, 891 costs, 502-3
classification, 555 electrowinning efficiency, effect of,
current for gold reduction, 564-6 507
electrical by-passing in, 559, 561, 892 equipment, 511-30
fluidized bed cell, 891 heat exchangers, 526-7
inter-electrode clearance, 570 instrumentation & control, 527-30
limitations on cross-section area, 570 measurements for control, 529
materials of construction, 561-2 operating sequencing, 527-8
mechanical by-passing in, 559, 892 factors affecting
Mintek steel wool cell, 558, 559, acid washing, 493-7, 506
561-2, 891 calcium levels on carbon, 505, 507
modelling, mathematical, 552-3 cyanide concentration, 485-6
NIM graphite cell, 557-9 cyanide decomposition, 486-8
packed bed, 891 electrowinning efficiency, 507
packed bed cathode, potential ionic strength, 483-5, 876-7
distribution across, 559-60 organic solvents, 486
parallel flow, 555, 892 pretreatment reagents, 498-502
right-angle flow, 555, 891-2 solution flow rate, 502, 508-9
steel wool, packing density, 563 temperature, 483, 484, 497-8, 500,
tubular bed, 891 505, 876
Zadra, 555, 891 gold and silver profiles, 509
Electrowinning cell, factors affecting kinetics, 489-92
chromium in solution, 567 mathematical modelling, 489-92
flotation reagents, 568-9 process
flow rate, 566 Anglo American Research
gold concentration, 563-5 Laboratory, 481-2, 492-503,
gold deposited, 566-7 877
hydrophobic organics, 568 by organic solvents 482, 877
impurities in solution, 567 by polar solvents, 877
solids in suspension, 568 by thiocyanate, 877, 883
solution conductivity and pH, 566 by thiourea, 877
temperature, 566 from strong base resin, 882-3
Elements, periodic table, 1071 Zadra, 481-6, 491, 503-11
Elution, 481-530 rate equation, 489
carbon profiles, 509-11 research in, 1032
column, 381 silver, behaviour in, 481, 503, 509
configuration, 515-18 system pressure, 522-7
corrosion, 518-19 control, 527
design, 515-18 Elutriator, carbon cleaning, 512-13, 514
discharge arrangements, 516-17 Emergencies
downflow systems, 522-4 preparedness for, 983-7
downflow v. upflow comparison, mock, 983
522-3 Emergency pump, thickener draining, 283
1090
INDEX
Emulsifier Ethanol, use in elution, 481-2

in zinc-dust precipitation, 338, 341 Eureka Syncline, 43, 44
Enargite, 52 Euxenite, 33
Encasing minerals, 31, 219 Evaluation
End liners, grinding mill, 153-5 metallurgical, 909, 1013
inlet, 167 process, 745-91
outlet, 153-5, 167-9 Evander goldfield, 2, 5, 12, 13, 16, 124
rod mill, precautions, 158 Examinations,
Endless riffle belt concentrator, 221, audiometric, 989-91
226-7, 231, 232 eyesight, 987-8
Endless rope haulage, 75, 111-12, 115 medical, 987-91
Energy Exhaust fans, 118
in-plant reticulation, 937 Exothermic reaction (pyrite roasting), 346
requirements for comniinution, 129, Explosives
139-43, 145, 212-16 in reduction plants, regulations, 106,
transfer in milling, 189, 196-9 117
Environmental impact, of tailings dams storage of, safety, 1007
assessment, 666-7 Exposed machinery, regulations, 117
monitoring, 667-8 Exsolution phenomena, 52, 57
remedial measures, 667 Eyesight, examination of, 987-8
see also Pollution, control Eyewash fountains, 985
EPAC screens, 470-7
Epigenetic gold occurrences, in Barberton
F
greenstone belt, 41
Equilibrium approach, gold Fairview gold mine, 43, 44
hydrometallurgy chemistry, 836-43 carbon-in-pulp plant for calcine,
Equilibrium ball loads, 183-4 393-4, 436, 455-6
Equilibrium effects, gold adsorption, elution procedure, 481, 493
417-21, 439 flotation plant, 253-5
Equipment residue retreatment, 716
choice of supplier, 926 Famatinite, 52
clearance around, 941-2 Fatality rate, 981
headroom over, 942 Feedback
list, 925-6 control, 797
pre-use inspection, 975 in flotation circuit, 263
selection, 924-5 Feed dilution
sizing, 924-5 cyclone classifier, 151, 198-9, 205, 207
Equipment-related problems grinding mill, 198-9
in laboratory flotation testing, 270 control, 200
Erfdeel-Dankbaarheid gold mine, 2 thickener, effect on capacity, 285
Ergo Feed pressure, cyclone, 208
see East Rand Gold & Uranium Feed pulp, thickener, 284-7
Company Feed rate, ROM mill, effect of feed
Eriksen cone, for sink-float tests, 1024 sizing, 196
Erlichmanite, 35 Feed size
Estimate, capital cost, 918 ball mill, 159
Definitive, 918, 957 rod mill, 157
Detailed, 918 ROM mill, effect on, 195-6
Order of Magnitude, 918 Feed 'systems, filter, 309, 314
Preliminary, 918 Feeder box, grinding mill, 163
Revised, 918, 957 Feeders
Estimation of ore stock, 116 automatic control, 819
1091
INDEX
Feeders (cont.) candle pressure, 320-1
cyanide, 293 leaf, 320
grinding mill, 153, 163-6, 184-6 pre-coat, 319-20, 359
ore, 80-5, 184-6 sand, 318-19
Feedwell, thickener, 277, 278, 282-3 leached gold pulp, 305-12
Fence belt, 307-8
plant, 939 rotary drum, 305-7
slimes dam, 698-700 pressure, sample drying, 773-4
Ferric hydroxide, generated in vacuum leaf (Merrill), 333
iron sulphide leaching, 858, 859 Filtrate
Ferricyanide in iron sulphide·leaching, 859 clarity, 313
Ferrocyanide, in iron sulphide leaching, extraction system, 306, 307, 309, 310
859 pumps, 309
Ferrous grinding media, 159, 161, 178-9 receiver tank, 309, 359, 360
wear of, 181-2 Filtration
Ferrous sulphate, formation in roasting, composition with carbon-in-pulp
347 process, 382-3
Fick' s first law of diffusion, 855-6 continuous vacuum, 305-17
Fig Tree Group, 37, 38, 44, 45, 46 practice, 312-16
Filter rate, measurement of, 1044-6
aids, 313, 319 tests, 915-16, 1019, 1024
effect on composition of zinc- types, 305-12
gold slime, 370 Fineness
bags, 320 gold
cake, 314 composition, 56-7, 232, 369, 372,
discharge systems, 309, 311 373, 619
pressure drop across, 312, 313 sizing, 20-1, 51, 125-7, 237
washing, 306, 307, 314-15 mill product, 127, 138, 195, 196, 216,
cloth, 306, 311, 312-13, 360 237
acid treatment, 313 refined bullion, 625, 626
brushing, 313 Fire assay
washing (belt filter), 311 only method for gold ore, 266
duty, 312 precision, 787
factors affecting, 312-13 Fire station, 940
elements (dust extraction), 118, 376 Fires
maintenance, 316-17 prevention of, 985
medium, 312 protection, 1010-11
requirements, 312 First aid
paper use in filtering acid-treated equipment, 985
precipitate, 360, 373 training station, 940, 985
recovery, effect of feed grade, 445 treatment, 1012
speed, effect on recovery, 315-16 First order rate process
Filter plants definition, 147
belt handling, 950-1 Fixed jaw (of crusher), 98, 99
conveyor belt, 950 Flammable substances storage
layout, 949-51 safety, 1007-9
repulper tank, 950 Flaring (cyclone underflow), 198
spillage system, 950 Flat-bottomed agitator tanks, 301
weightometer installation, 950 Flaws (in brittle substances), 128-9
Filters Flint
clarifying blocks, 121
1092
INDEX
Flint (cont.) application, 786-7

pebbles used in grinding, 121 calibration, 787, 1019
Flocculants d.c., 805-6
feeding, 284, 288 installation, 188, 806
mixing, 288 principle, 805-6
theory of action, 287-8 Flow rate, measurement of
use in hopper clarifier, 321 liquids, 804-6
Floor space requirements (of gravity solids, 780-6
concentrators), 222, 228 Flow rates
Floors estimating, for plant design, 923-4
construction, 953-4 Flowsheet, 917-33
materials, 954 balancing, 920-3
Flotation (in gold ore treatment), 21, 59, design, 917-18
127, 220 estimating for, 923-4
carbonaceous material, removal by, evaluation, financial, 926-33
862 flow rate, tabulation, 918-22
cells, 260, 271, 272, 273 quantified, 918-24
chemical aspects, 239-48 selection, 926-33
circuit configuration, 263 Fluid bed roasters, 228-9, 347, 349-50,
cleaning and transport phase, 239 723
concentrates Fluidized-bed reactor (electrowinning),
treatment by CIP, 456, 729, 732 891
treatment by roasting, 253, 257, Fluorspar
345-52, 735 component of smelting flux, 371, 898
control, 265-6 Flux, smelting
feed, 253, 255, 257 chemistry, 896-9
feed density, 265-6 for calcined zinc-gold slime, 371
history of use in South Africa, 235-6 for cathodes, 373
new applications, 272-4 for direct smelting of gold
physical variables, 248-51 concentrates, 369
plant design, 261-5 Fly ash, 349
plant practice, I 51, 251-60, 722-3, Flywheel (jar crusher), 99
727, 729, 733-34 Footwall, 69
processing, 236-60 Forbes Reef area, 41, 49
rate constant, 261, 262, 268, 271, 272 Forbes screen (for CIP), 471
reagents Forward leach plants, flotation in, 264
effect on adsorption in CIP, 434-5, Fracture of brittle solids, 128-9
456 Free milling ores, 15, 236, 237
effect on cyanidation, 239, 863 Free State Geduld mine
effect on electrowinning, 568-9 flotation in milling circuit, 273
feeding, 266 member of JMS residue retreatment
testing of, 734, 1024 scheme, 731
scale-up, 261-3 Freundlich model
separation phase, 239 of adsorption, 398
tests, 266-72, 916-17, 1019, 1024, 1032 Friction constant, Hazen-Williams, 965
analysis of data, 272 Friction head, calculation of, 967
checking of results, 271-2 Friction loss, of pipe fittings, 967, 968
pyrite treatment, 1032 Froth
trouble-shooting methods, 266-72 depth, 249
Flow data table, 920-2 modifier, 246
F lowmeters, electromagnetic Frother, 242-3, 257
a.c., 805 Fuchsite, 52
1093
INDEX
Fullerton Creek, 48, 49 measurement, 105

Fume collection setting, 117
at Rand Refinery, 628-30, 651-2 start-up setting, 960
mercury, 354 Garnet, 16
smelthouse, 376, 952 Gas adsorption, on carbon, 379, 385
Funda filter, 341 Gas cleaning systems, 376, 628-30, 651-2,
Furnace 952
blast, by-product smelting, 644-8 Gates, plant, 939-40
calcining, 361-3 Gates, slide
continuous steel belt, calcining 361-3 in ore feeders, 184
cupel, 648-9, 650, 651, 652 in launders, 947
electric Gate values, 188
calcining, 361 pressure drop through, 968
continuous steel belt, 361-3 Gaudin-Meloy size distribution equation,
induction, 367-8, 619, 621, 625, 134
639 Gaudin-Schuhmann size distribution
reverberatory, 364 equation, 133
single-pot, 367 Gear boxes, speed-reducing, 157
submerged arc, 364-7 Geduld mine, 26
fusion, 363-8, 619, 621, 639 Geluk Subgroup, 37, 38
gas-fired, 633 Geofabric, 695
induction, 367-8, 619, 621, 625, 637, Geology
639 Barberton area, 35-41
pan, 649 Witwatersrand depository, 5-13
reverberatory, 364, 369, 370, 371-2 Gersdorffite, 16, 20
roasting, 348-50 Geversite, 35
fluid bed, 349-50 Gibbs free energy, 836, 837, 838
rabbled hearth, 348-9 Gloves, PVC, 992
single-pot, 366, 367, 369, 370 Go-devils, 956
submerged arc, 364-6, 372, 373 Goethite, 16, 51
Fuses, blasting, disposal regulations, 106, Gold adsorption on carbon
117 economic evaluation of plant
Fusion furnaces, 363-8, 619, 621, 639 alternatives, 413-17
factors affecting
G acidity, 423, 425
Gabions, 701-3 base metals, 426-7
Galena, 13, 14, 16, 20, 23, 52, 54, 57 calcium carbonate, 435-9
Galvanic carbon size, 440-2
cell, 836-7 cyanide concentration, 422-3, 424
interaction between gold and other equilibrium effects, 417-20
minerals, 32 flotation reagents, 434-5
Gamka River ionic strength of the medium, 421,
gold occurrence, 1 422
Gangue mixing rate, 439-40
dissolution, in mineralogical oxygen, 427-9
examination, 911 pulp solids, concentration and size,
entrainment, in flotation, 248-50 443
in gold ore processing, 125, 219, 236, temperature, 422, 425
239, 243-4 from non-cyanide solutions, 875-6
minerals, 52, 53, 55 gold lock-up, 401, 403, 405, 407-11,
Gaps (cone crusher) 429-30
1094
INDEX
Gold adsorption on carbon (cont.) Gold compounds

history, 379-83 chemistry, general principles, 831-54
in extraneous ash, 32 co-ordination numbers, 834
ion pair theory, 392-3 covalent bonding, 831
kinetics, 389, 393-9, 429-43, 448 dimerization, 834
loading capacity, measurement of, stability, 834
1049-50 Gold, co-ordination numbers, 833
mechanism, 385-93, 874-6 Gold, correlation with heavy minerals, in
process, description, 458-9 Witwatersrand ore, 27-8
rate, measurement of, 1050 Gold-cyanide complexes
thermodynamics, 389 chemistry, general principles, 833-6
Gold amalgam formation in leaching, 846-8, 854-7
recovery process, 229-30 reduction of, 849-51
retorting of, 352-4 effect of leach, mercury and
Gold analyser thallium, 851
Olga Mark III, 811 Gold deposits, geology
On-Line, 811-13 Barberton Mountain Land, 35-41
Tell-tale, 811 Witwatersrand Triad, 5-13
Gold analysis, on-line, 811-13 Gold disc electrode, 850
Gold, anodic dissolution, 854-7 Gold, discovery
Gold, anodic reactions, 843-8 historical background, 1-2
in chloride solution, 844-5 Gold, dissolution
in cyanide solution, 846-8 by intensive cyanidation, 230, 856-7
in thiourea solution, 845-6 by leaching, 291-305
Gold assaying chemistry, 836-48, 854-66
fire method, 266 inhibition of
at Rand Refinery, 652 by coatings, 31, 32, 59, 219, 304,
precision of, 787 857
Gold, associated with radioactive by passivation, 843-4, 847-8, 857
constituents of ore, 21-4 by sulphide ion, 858-9
Gold bars relation to fineness of grind, 125, 127
cleaning and sampling, 373-5 Gold electrochemistry, 843-8, 849-54
Gold-bearing solutions Gold, electrowinning, 550-70, 890-4
analyses, 332 at Rand Refinery, 626-8
characteristics, 331 cell design, 552-61, 569-70, 890-3
recovery methods, 331-42 extraction efficiency, 563-9, 890-1
Gold by-products from chloride solutions, 853-4
handling of slag in smelthouse, 375 see also Electrowinning
treatment, 615, 616, 641-52 Gold, elution from carbon, 481-530,
Gold call, 788 876-7
Gold carrier, pyrite as, 20 see also Elution
Gold, chemical composition of, 28-30, Gold extraction
55-7 chemistry of, 831-905
Gold, chemical precipitation of, 889-90 plants
Gold chloride, 835-6, 842, 844-5, 852-4, commissioning, 957-61
866 control, 793-829
Gold-chlorine complexes design, 907-57
reduction of, 852-4 operating disturbances, 796
Gold complexes operating strategy, 795-6
diamagnetism, 833 performance optimization, 796
oxidation states, 831 Gold, fineness (composition), 56-7, 232,
stability, 832-3 369, 372, 373, 619
1 ()Q<;
..l.V..J-1
INDEX
Gold, fineness (composition) (cont.) Gold, oxidation
of refined bullion, 625, 626 by dissolved oxygen, 841
Gold, fineness (sizing), 20-1, 125-7, 237 Gold, particle size, 20-1, 22, 51, 125,
Gold, flotation, 236-7, 238, 264 237
see also Flotation Gold, passivation, 843-4, 847-8, 857
Gold, formation of oxide films, 843-4, Gold, precipitation process, zinc dust,
847-8 331-42
Gold halides, 835-6 Gold, production figures
Gold, hydrometallurgy, equilibrium Barberton area, 36
approach to, 836-43 Witwatersrand-0.F.S., 2-4
Gold in residues, 30-2 Gold, recovery
Gold in solution by chemical precipitation, 889-90
measurement of, 811-13 by electrowinning, 550-70, 890-4
Gold, leaching, factors affecting by ion exchange, 870, 877-85
antimony sulphides, 860-1 by resin-in-pulp, 884
arsenic sulphides, 860 by solvent extraction, 870, 885-9
base metal complexes, 861-2 by zinc-dust precipitation, 331-43,
calcium ions, 862 866-9
carbonaceous material, 862-3 formulae, 788
cqpper sulphides, 859-60 from poisoned resin, 885
cyanide concentration, 855-7 relation to comminution energy, 142
flotation reagents, 863 Gold refining
heavy metals, 847-8, 852 chloride process (Miller), 615, 621-6
iron sulphides, 858-9 degolding of chlorides, 626
lead nitrates, 857 electrolytic, 616, 626-8
operating conditions, 304-5 fume recovery, 628-30
oxygenation, 303 furnace charge, 622
pulp constituents, 857-8 platinum metals, behaviour in, 627-8
pulp dilution, 303-4 sweep treatment, 630
pyrrhotite, 858 washing of cathodes, 628
sulphur-containing anions, 861 Gold Refining Branch, Rand Refinery,
Gold, liberation, 123, 125-7, 219 618-30
Gold, lock-up Gold, scavenging in carbon-in-solution
in CIP plant, 401, 403-5, 407, 408, plants, 584-608
411-13, 429-31, 450-1 Gold, simultaneous recovery with
in grinding mill, 790 uranium, 885
in ore storage, 77 Gold slime (zinc-gold slime)
in reduction plant, 789 acid treatment, 356-7
Gold loss calcining, 361-3
in CIP, 443-9 composition, 354, 370
in filtration, 312, 313, 314-16, 328-30 dewatering, 357 -60
in leaching, 291, 304-5 treatment, 356-63
monitoring, 811-13 Gold slime, belt filter, 360
Gold mine tailings Gold-thallium electrode, 815
characteristics on impoundment Gold theft
method, 685-6 precautions against, 376-7
effect of, 668, 669 Golden Quarry mine, 44, 54
mineralization of, 30-2 Government Gold Mining Areas
Gold nuggets, 26 flotation plant, 235
Gold, occurrence in nature, 831 Government Reef, 7
Gold ore, radioactive constituents, 21, 23-4 Grab samples (sorted waste), 116
Note: pages 1 - 614 are in Volume 1, pages 615 - 1072 in Volume 2.
1 ()Ot;:
-1..V--'V
INDEX
Grading energy requirements, 139-42

procedure, 129-31, 1036-40 regrinding, 216
tests, 1019 history, 121-5
Granite media, 178-9
granite contact, factor in gold balls, 179
mineralization, 49 pebbles, 179
surrounding Barberton greenstone rationing, 183-4
belt, 39 rods, 178-9
Graphite, in New Consort ore, 52 ROM ore, 179
Grate, grinding mill discharge, 153, 159 size, matching to feed size, 179-81
Gravity cascade contact system for wear, nature of, 181-3
carbon-in-solution, 585-90 wear, Davis theory, 182-3
advantages & disadvantages, 221-2 wear, Prentice theory, 183
Gravity classifiers, i47-50 mills, general description, 152-7
Gravity concentrate autogenous, pebble, 160-2
amalgamation, 229-30, 352 autogenous, ROM, 160-2
direct smelting, 369 ball, 158-60
for mineralogical examination, 91 design methods, 212-16
treatment of, 228-31 design, of rod, ball, pebble and
Gravity concentration composite mills, 212-14
advantages/ disadvantages, 221-2 design of ROM and SAG mills,
circuits, 231-3 214-16
definition, 219 feeders, 163-6
development, 233-4 fine ore, 184-5
equipment, 223-8 coarse ore and pebbles, 185-6
historical background, 220-1 lifter pans, 168-9
philosophy, 221 liners, end, 166-9
use for recovery of coated gold, 304 liners, shell, 170-5
Gravity concentration circuits, 231-3 operation, automatic, 199-201
Gravity separation testwork, 1024 operation, principles, 189-201
Gravity settling clarifiers, 321-3 pebble, 160-2
Gravity take-ups (conveyor belt), 89-90 power drawn by, formulae, 214-15
Greek alphabet, 1072 relining methods, 175-8
Greenstone belts, 5,6 rod, 157-8
Barberton, 35-41 ROM, 160-2
Greywacke, 39, 44, 47 particle breakage in, 125-9
Grid liners, grinding mill size distributions, 129-35, 136-7
inlet end, 167 stage, 143-6
outlet end, 168 theory, 125-47
shell, 173-4 transport mechanisms in, 137-9
Grindability determination Grizzlies, 73, 74, 91, 94-5
autogenous, 211, 914 mechanized, 95
Bond, 914 pebble selection, 114-15
for fine materials, 216 static, 94
relative, 1034-5 washing on, 92
Grinding (see also Milling) Grizzly bars, 94
Bond theory, 140-1 Grootvlei screen (CIP interstage), 471,
breakage 472
mechanisms, 135-6 Group meetings, loss control, 982-3
types, 135, 195 Guar derivatives
closed circuit, 147 flocculants, 313
control, automatic, 199-201 flotation reagents, 247, 255, 734
1097
INDEX
Guardhouse, 939 117-18
Guidelines, for induction of new Heap-leaching, 111, 113, 740-1
personnel, 983 tests, 1024
Gypsey Queen mine, 48 see also Leaching
Gyratory crushers, 100-6, 121, 122 Hearing loss categories, 989, 991
Gy's fifty-piece experiment, 753-6 Hearing protection devices, 989, 992, 998
Gy's formulae for sampling precision, Heat-exchangers, elution plant, 524, 527,
751-3, 759, 765-6 529
Heating, building, inspection of, 1002-3
H Heavy medium separation
H-carbons, 384-5 tests, 916
Hadsel mill, 124 Heavy minerals
Half-cell reactions, 836, 837 distribution in Witwatersrand
Handrails, construction of, 954 conglomerates, 26-7
Hangingwall, 69, 70 Hedgehogs (pipe-cleaning), 956
Hard-wired analogue control systems, Helmets, safety, 992
821-2 Hematite, 50, 51
Harrison, George, 1 effect on carbon in CIP, 456
Hartebeestfontein gold mine, 17, 27 formulation in roasting, 59, 346, 347
flotation plants, 235, 255-7, 258, 263, Henry's Law, 333-4
264 Heterogeneous flow
radiometric sorting tests, 738 Newitt correction, 965, 966
Hastelloy C276, for elution column Hhohho area, Swaziland, gold in, 41, 49
construction, 519 High-rate thickeners, 290
Haulages Hinged rakes, thickener, 281-2
ore transport, 75 Hoisting, 70, 75
pebble, 115 Hold-up, grinding mill, definition, 153
waste disposal, 111-12 control of, 191, 198-9
Haultain Superpanner, 916, 1024 effect of, 138-9, 193
Hazardous chemicals washing-out of, 176
data sheets, 985 Holidays, statutory, 922
cyanide, 986 Hollingworthite, 35
handling, 985 Homestake
storage safety, 1006-7 elution system, 481
Hazen-Williams formula, 965 interstage screening, 470, 478
Head Homologous series
cone crusher, 103, 105 for adsorption of organic molecules,
gyratory crusher, 101, 102 386, 393
Head drum (conveyor belt), 87, 88 Hopper (rotary filter), 305, 306, 307, 309
Head grade Hopper clarifiers, 321-3
effect on carbon-in-pulp plant, 444-5 Hopper feeders (grinding mill), 165-6
effect on cyanidation residue, 291 Hoppers, storage, 78
effect on filter plant, 445 construction of, 78
in flotation tests, 271-2 Horizontal belt filters, 307-8, 309, 311
Header valve cake discharge, 311
rotary drum filter, 306, 307 capacity, 317
Headgear bin, 73, 74 drainage system, 312
Headroom feed system, 309
over equipment, 942 filtrate system, 309
Health Act, 664, 698 maintenance, 317
Health regulations, dusty atmosphere, zinc-gold slime dewatering, 360
Note: pages 1-614 are in Volume!, pages 615-1072 in Volume 2.
1098
INDEX
Horizontal outlet branches I
on pulp pumps, 187 Idlers, conveyor belt, 87, 88
Hot cyclone arsenic recovery system, 351 for filter discharge belts, 950
Humic acids, 24, 862, 863 impact, 88
Hydraulic classifiers, 147-52 spacing on weightometer
Hydraulic gradient, calculation of, weighlength, 783
for pipelines, 965 Ill personnel
in continuous leaching plants, 297 care of, 985
Hydrocarbons in Witwatersrand Impact breakage, 135-6
conglomerates, 14, 24, 862-3 Impact idler, 88
Hydrochloric acid Impellers, flotation machine
carbon treatment for calcium salt speed of, 249-50
removal, 381, 493, 496-7 effect on flotation results, 270-1
coated gold treatment, 59 Impellers, mixing
data sheet available, 985 effect on carbon breakage, 469, 582-3
filter cloth treatment, 313 effect on carbon consumption, 448-9
gravity concentrate treatment, 231, requirements for, 468-9
232 Impellers, pump
pipeline, descaling, 189, 956 speed, calculations, 967
precipitation filter treatment, 355 maximum speeds for various
sand bed clarifier treatment, 319 materials, 967, 969
steel wool cathode dissolution, 373 Incident
Hydrocyclone classifiers, 124, 150-2, definition, 975
197-8 investigation and analysis, 975, 977-82
design, 207-8 loss severity potential, 977
mathematical modelling, 205-7 preparedness for, 983-7
tests, 1024 Inclinometers, tailings dam, 697
theory, 202-5 Induction furnaces, 367-8
see also Cyclone classifiers bullion melting, 621
Hydrocyclone system of tailings dam coin alloy melting; 639
construction, 671-3 gold refining, 621, 625
Hydrofluoric acid leaching techniques in silver melting, 637
mineralogy, 17 steel wool cathode smelting, 367
Hydrofoil impeller blades, 469 Induction, of new employees, 983, 984,
Hydrogen 985
bonding adsorption on carbon, 385-6 Inertia, 219, 227
evolution in acid treatment of zinc- Infinite time recovery constant (flotation),
gold slime, 357, 895 268, 271, 272
evolution in cementation, 868 Information display, 822
evolution in electrowinning, 551, 566-7 Information processing, 745
Hydrometallurgy Injector sampler, 777
gold, physical chemistry, 836-43 Injuries, disabling
investigations, 1024-5 care of, 985
laboratories, 1022 definition, 981
equipment, 1023 frequency rate, formula, 981
Hydrophobicity Inlet area, cyclone, 208
in flotation process, 239-42 Inlet end liners, 167
Hypersthene, 52 renewal, 178
Hysteresis Inlet trunnion, 15 3
in pinch valves, 817 In-motion train scale, 781-2
Inputs, control, 797
interfering, 803
1099
INDEX
Insoluble alloys effect on silver adsorption, 390-1

of gold, 58 Ions
Inspections, planned, 974-5 ion-pair theory of adsorption, 392-3
checklist, 1002-12 theory of adsorption by carbon, 386
critical parts, 97 5 Irarsite, 35
equipment pre-use, 975 Iridarsenite, 35
general, 974-5 Iridium-osmium-ruthenium alloys, 34
record sheet, 976 Iridosmine, 16, 34
Inspector of Mines Iron
dusty atmosphere monitoring adsorption on carbon, 426
requirements, 117 cast, mill liner blocks, 121, 170, 175
Instrumentation diagrams, 957 cause of bullion sampling errors, 620,
Instrumentation, industrial, 801-2 895
accuracy, 803-4 in Barberton gold, 56
calibration, static, 803-4 in bullion from direct smelting of
costs, 802 gravity concentrates, 369
dynamic response, 804 in Witwatersrand gold, 30
terminology, 802-3 in Witwatersrand ores, 18, 27
measurand, 802 removal from gravity concentrates,
primary sensing element, 802-3 230-1
sensor, 803 Iron arsenates, formation in roasting, 346
signal processing, 802 Iron oxide
transducer, 802 coatings on gold, 32, 58, 219, 222
Integrating massmeter, 115, 782-7 formation of, during sulphide
Intermediate conveyor (waste disposal), roasting, 346-7
112 Iron sulphides, 14, 15, 16, 17, 20, 32, 33,
Intermig impeller system, 468 47, 48, 49, 50, 51, 52, 53, 54, 55,
Inter-molecular bonds, 128 56, 57, 58, 59, 60
Internal Rate of Return, 925, 928-33 cause of retractoriness, 219, 237
International System, of units effect on leaching, 858-9
base units, 1067 recovery by flotation, 235, 236, 237,
derived units, 1067 239, 241, 253-60
Interstage screening methods (CIP), 470-8 recovery by gravity concentration,
'Iodine number', of carbon, 384 231-3
Ion exchange Irrigated float (for pulp level sensing),
applications to gold recovery, 883-5 265
comparison with solvent extraction, Ishihara colour blindness test, 987
887 Isoferroplatinum, 35
principles, 877-83
resins J
elution, 882-3 Jacking screws (for conveyor pulleys), 88,
regeneration, 883 89
selectivity, 880-2 Jacks
Ionic radius of gold in solvent energy in mill relining, 176
theory, 391 James shaking tables, 227-8. 229, 231,
Ionic solvation-energy theory of gold 233, 369
adsorption, 390-1 Jamesonite, 345
Ionic strength of media Jamestown area, 41, 43
effect on gold adsorption, 421-2, schist belt, gold in, 45-6
872-4 Jarosite, 32
effect on gold elution, 483-5 Jaw crushers, 74, 98-100, 121
Note: pages 1-614 are in Volume 1, pages 615 -1072 in Volume 2.
1100
INDEX
Jaw crushers (cont.) Kromdraai

eccentric, 98, 99 gold discovery,
jaw liners, 98 Krugerrand, blank production, 616, 617,
pitman, 98, 99 637-40
sample crushing, 766 Krugersdorp Formation, 12
toggle plates, 98, 99
JCI Mineral Processing Research L
Laboratory L-carbons, 384-5
flotation research, 270-1 Laboratories, 1013-29
on-line gold analysis, 813 control, 1013
Jigs, 226, 1024 group, 1013, 1020-6
JMS flotation plants, 260, 263, 733-4 equipment, 1022-3
Job orientation, 983 facilities, 1021, 1022
Jockey, rope haulage, 75 financing, 1021
Johannesburg Subgroup, 12 functions, 1023-5
Johnson drum concentrator, 221, 224-5, layout, 1022
231, 232 reporting, 1021
Johnson press, 357-8 staffing, 1020-1, 1025-6
Mintek (see under Mintek)
K on-mine, 1013, 1015-20
Kaapsehoop advantages & disadvantages, 1020
gold discovery, components, 1016
Kaapvaal craton, 35, 37, 39 equipment, 1016, 1017, 1019
Kalgoorlie goldfields, 121 functions, 1019
Kambalda screen, 477-8 layout, 1015-17, 1018
KENWALT process simulator, 827 staffing, 1019-20
Kerogen, 14, 21, 23-4, 33, 237, 247 storage facilities, 1016-17
Kerosene, 862 university, 1014
Key point tipping, 983 Laboratory flotation, 269-72, 1019, 1024
Keyed block liner (grinding mill), 170-1, analysis of data, 272
175 checking of results, 271-2
Keys difficulties
in block liner, 171, 176-7 avoiding, 271
in Osborn bar liner, 170 equipment-related, 270
Kick, Friedrich sample-related, 269-70
Comminution Law, 139-40 parameter-related, 270-1
Kilowatt meter (for grinding mill), 192 reproducibility, 271
Kimberley-Elsburg series, 12 methods, 269-71
Kimberley Reefs, 13, 16 Laboratory mills (for flotation tests), 270
Kinross gold mine Laboratory-scale methods (flotation),
flat-bottomed leach vessels, 301 269-72
Kinross screen (CIP), 471, 472 Laboratory testing
Kitchens (for arsenic recovery), 351 of gold ore, 913-17
Klerksdorp goldfield, 2 Ladders, use, safety, 1005
Klimpel flotation model, 261-2, 268, 271, Lamella clarifiers, 323
272 Langlaagte
Kliptek screen (CIP), 471, 472 chute, 81, 86
"kn" model feeder, 186
carbon-in-pulp process, 394, 399-401 gold discovery, 7
Komati valley 41 Langmuir adsorption equation, 398, 422
Komatiites, 37, 39 Lap screen, 471, 472
Kromberg fault, 49 Lashing (ore), 70
1101
INDEX
Laser beam bunds, 948

in Microtrac Analyser, 809 layout, 946-7
in photometric sorter, 109 mechanical handling equipment,
Launder screen, 471, 472 948
Launders spillage systems, 948
construction, 955-6 rate (see kinetics)
design, 956 reagents, 291-5
peripheral overflow, 149, 277, 321 reagents other than cyanide, 863-6
shape, 955 residue retreatment, 727
Laurite, 34, 35 reverse, 31, 257, 264
Law of similitude (for pump speeds), 967 simultaneous, gold and uranium, 835
Leaching gold, 277, 291-305 tests, 1019, 1025
bacterial, 741-3 rolling bottle, 1046-7
batch, 295-6 thiocyanate, 865
chemical effects, 855-63 thiosulphate, 864
chemistry, 854-6 thiourea, 864-5, 878
chloride, 866 vessels
conditions, 303-4 air-agitated, 301, 463
continuous, 297-300 air requirements for, 303
design calculations, 326-7 drainage system, 948
dump, 740-3, 1015 flat bottomed, 301
factors affecting interconnections, 947-8
antimony sulphides, 860-1 layout, 946-8
arsenic sulphides, 860 mechanically-agitated, 301-3
base metal complexes, 861-2 spillage system, 948
calcium ions, 862 Lead
carbonaceous material, 862-3 behaviour during silver refining
copper sulphides, 859-60 process, 632
cyanide concentration, 855-7 bullion, cupellation, 649
fineness of grind, 127, 142-3, 151 effect on reduction of gold cyanide
head grade, 291 complexes, 851
iron sulphides, 858-9 refractoriness caused by, 58
lead nitrate, 857, 860, 861 salts
lead salts, 861 in cementation by zinc, 333, 336-7,
nickel, 858 866, 868
oxygen, 58, 291, 295, 302, 303 in leaching, 857, 861
pulp constituents, 857-63 sulphate
pyrrhotite, 858 behaviour in calcining, 896-7
sulphur-containing anions, 861 behaviour in smelting, 897
surface active agents, 863 Lead-bearing ore (Barberton), 50
feed pulps, oversize constituents, Lead nitrate
459-60 data sheet available, 985
flotation tailings, 238, 723, 727, 734-5 in calcine leaching, 59
forward, 264 in cementation by zinc, 336-7
gold losses, 30-2, 58-9, 304-5 in gold ore leaching, 857, 861
heap, 113, 740-1, 1015, 1024 in zinc-gold slime, 370
in situ, 740 Lead-zinc couple in cementation, 333, 336
kinetics, 295-6, 326-7, 449, 450, 454-5, Leaf clarifiers, 320
843 Leco sulphur analyser, 266
plant Leeds & Northrup Microtrac Analyser,
air requirements, 303 131, 807, 809-10
1102
INDEX
Leeds flotation cell, 271 SAG mill, 162

Lenses, ore, 70 grid liners, 167, 168
Lepidocrocite, 51 inlet end, 167
Leucoxene, 15, 16, 25, 33 outlet end, 167-9
Levels, access, to machinery, 942-3 pocketed, 173
Lewis acid, 836 renewal, 176-8
Lewis, Henry, 1 rod mill, 158
Liberation of gold, 123, 125-7, 219 shell, 170-5
Libraries, laboratory, 1022 Linnaeite, 20, 52
Lifter arms (grinding mill), 168-9 Liquid level, measurement of, 806
Lifter bars, 172, 173, 174 air bubblers, 806
spacing, 17 5 capacitance probes, 806
super-elevation, 175 conductivity probes, 806
Lifter pan, 139, 153, 154-55, 159, 168, differential pressure, 806
169 float displacement, 806
Lifters, 174-5 ultrasonic, 806
Lifting devices, thickener rake, 282 Liquid phase
Lifts, 943 effect on thickener capacity, 285
Ligands, 831-3, 835, 836, 842, 844 Liquid: solid ratios
donor atoms of, 832, 833 formulae, 1064-6
Lighting, 938, 940, 951, 953 tables, 1062-4
inspection of, 1002 Liquid-solid separation, 305-23
Lily fault, 41 Liquids
Lily mine, 55 flow rate, meaurement of, 804-6
Lily syncline, 46 sampling of, 778
Lime Liquor, gold-slime dewatering, handling
concentrations of, 360
in leaching, 291-2, 293 Lithium hydroxide, as tracer in flotation,
in zinc-dust cementation, 337 267
consumption rate in leaching, 295 Live load (of ore storage), 77, 78
distribution (in-plant), 294, 936 Load, grinding mill, 153
effect on gold adsorption by carbon, bulk density, 161, 191, 193, 194, 214
426 configuration as function of mill
handling and storage, 292, 293, 294, speed, 134, 135
936 disappearance rate, 196
in flotation, 246, 255 maximizing transfer of energy in, 196-9
in thickener feed, 285, 287 volume related to mill power, 191-3,
on-line measurement, 328, 815 200-1
safety precautions, 294 Loading
slaking, 294, 936 of conveyor belt, 90
types available, 292 Lock-up
Linear screen gold
fine carbon recovery, 480 in CIP plant, 399, 401-13
loaded carbon separation, 511 in mill, 789-90
trash removal, 284-5, 286, 460-1 in plant, 790
Liner bolts, 174 ore
Liners, grinding mill, 153, 166-78 in bins, 80
as gravity concentrators, 228 in bunkers, 77
ball mill, 159 in silos, 80
consumption, 162, 175 in stockpiles, 79
ball mill, 159 water
rod mill, 158 in thickeners, 288-9
1103
INDEX
Locking-out, power supply 1004 in loss control, 974

of conveyor belts, 91 Manganese
of grinding mills, 176 compounds as coatings on gold, 58
Log sheets, 958 dioxide in smelting flux, 369, 370,
Longitudinal divergence of grizzley bars, 371, 899
94 Manganese steel
Loss control, 971-1012 crusher liners, 98
accidents grinding mill liners, 167, 170, 171,
investigation and analysis, 975-85 172, 173, 174, 175
prevention, 973, 974, 977, 982, Manhattan Project, 33
985, 986, 993-8 Manholes
definition, 971 in grinding mills, 154
department, 971, 973 in leach vessels, 948
Manager's Policy, 974 Manifold pipe, for tailings dam cyclone
officers, 973, 977 feed, 677
physical condition standards, 971 Mantle (crusher), 101, 102, 103
programme, 971 Manual pebble selection, 114
steering committee, 974 Manual waste sorting, 107-8
staff, 971-2 Marcasite, 52, 858
training, 974 Marievale plant, 123
Low-level discharge, grinding mill, 139, Mass
153, 168 conversion tables, 1059
operating mills, measurement of, 201
M Mass balance, 759
McArthur, J.S., 333 in flotation, 266-7
McKelvey concentrator, 223, .224 in CIP plant, 393, 394
Machine guarding, 117, 1004 in sampling accuracy determination, 759
Magaliesberg Mass flow measurement, 780-7, 804
gold occurrence, 1 in crusher station, 115-16
Magmatic cycles, 39 in milling plant, 187
Magnetic drum separators, for separating moisture determination, 785-6
iron from concentrates, 230 see also Measurements
Magnetic flowmeters, 786-7, 805-6 Massmeter, integrating
calibration, 786, 787, 1019 electromechanical, 782-5
installation, 188, 806 calibration 782-5
Magnetic separation testwork, 1024 in crusher station, 115
Magnetic tramp-iron removal equipment, in milling plant, 187
107 magnetic, 804-6
Magnetite nuclear, 785
formation in roasting, 16, 52, 346, calibration, 785
347 Materials handling, 936-7, 93 8
Magnets safety, 1005
conveyor belt, 107 Mathematical descriptions of size
separation of iron from concentrates distribution, 131-5
with, 230, 369 Mathematical models
Main-Bird series, 12 CIP, 373-99, 400
Main Reef, 1, 9, 12,13 cyclone classifier, 205-7
Main Reef Leader, 9, 12, 13 elution, 489-92
Maintenance, ease of flotation, 261-2, 268-9
factors in, 938 Matrix, of Witwatersrand reefs, 14, 15,
Manager's Policy 16, 17, 20
1104
INDEX
Matte, 648 Merrill-Crowe zinc dust precipitation

Mauch, Carl, (cementation), 333-42
Maucherite, 52 chemistry, 866-9
Mean residence time circuit, 335, 336, 388
in CIP, 463 clean-up, 341, 342, 355
in flotation, 261-2 cyanide and alkalinity levels, 337
in grinding mills, 137-9, 198 de-aeration, 333-6
in leach vessels, 297-8, 327 feed clarity, 333-4
Measurement errors, 803-4 operation, 337-42
Measurements Stellar pressure filters, 341-2
basic physical, 804-10 vacuum leaf filters (pits), 338
chemical, 810-15 zinc and lead nitrate feeding, 336-7
conversion tables, 1054-70 Metal-cyanide complex adsorption, 426-7
distribution of, 747-50 Metal detectors, 107
liquid level, 806 Metal plate decks (vibrating screen), 96
mass of mills, 201 Metallurgical accounting, 787-90
mass flow rate, 745, 780-7, 804, 806 call system, 788-9
precision, 746 solids, 788
process, 801-17 solution, 788
repeatability of, 745, 746, 804 gold call, 788
static calibration, 804 primary, 789
Mechanical agitation secondary, 789-90
in adsorption vessels, 463, 468, 469 Metallurgical testwork, 909-17, 1019,
in flotation conditioning, 264 1023-5, 1027-9, 1032-51
in leach vessels, 301-3 Metamorphic grades, 13
Mechanical conditioning (flotation), 264 Metamorphism
Mechanical draft tube recirculator tank, cause of gold recrystallization, 17
301-3, 465, 468, 469 contact (in Barberton greenstone belt),
Mechanical handling equipment, 936, 39
938 Metasilicate, sodium
safety measures, 1005 formation of, in slag, 371
Mechanized grizzlies, 95 Methanol elution, 481-2, 877
Mechanized waste sorting, 108-11 Methylisobutylcarbinol (flotation reagent),
Medical Bureau for Occupational 734
Diseases, 987 Mica, 52
Medical stations, 985 Michenerite, 35
Megamines, 260 Micron solvent-distillation process, 893
Mercaptobenzothiazole collector, 255 Micropanner, 916
Mercury Microprobe analysis, 51, 911, 1025
condensation, 352-4 of Barberton gold in-situ, 56
distillation, 352, 353 of PGM metals in Witwatersrand ore,
effect on gold-cyanide complex 34-5
reduction, 851 of Witwatersrand gold in-situ, 28-30
gold recovery with, 220-3, 230, 232 Microprocessors, 201, 823
hazards of vapour, 353 Microscopy
in Witwatersrand gold, 28, 30 in gold distribution studies, 26- 7
mineralization, in Barberton ore, 50 in mineralogical examination, 911
precautions, 354 Microtrac Analyser, 131, 807, 809-10
weighing, 377 Middelvlei Reef, 12
Merriespruit Migmatites, 39, 41
flotation plant, 235 Milk of lime, 294, 305
residue retreatment plant, 715 Mill gold, analysis, 370
1105
INDEX
Mill, grinding stage, 145

autogenous, 160-2 tests, 210, 211, 914, 1019, 1024, 1034-5
ball, 123, 158-60 theory, 125-47
breakage in, 135-7 three-stage plants, 123
capital and operating costs transport in mill, 137-9
conventional v. ROM, 161, 162 two-stage plants, 123, 213
design, 212-16 Mineralogical examination (in ore
feeders, 163-6 testing), 911, 1023, 1025
grinding media, 178 Mineralogy
balls, 179 Barberton ores, 49-5 8
pebbles, 179, 182 Witwatersrand ores, 15-30
rationing, 183-4 Mineralogy, equipment, 1023
rods, 178-9 Minerals
wear, 179-83 troublesome, identification, 911
hold-up, 138, 153 Minerals Research Laboratory, 1026
liners, 166-78 Mines & Works Act, of 1911, 122-3
as concentrators, 228, 232 Mines and Works Act (Act No. 27 of
load, 153, 161, 189 1956), 91, 106, 117, 665, 666, 698
culling, 58 statutory holidays, 922
density, 161, 191, 193, 194, 214 Minicomputers, 822-3
motion, 135 Mintek, 50, 1013, 1021, 1022, 1026-9
volume, 191, 193 see also Council for Mineral
volume, control of, 191-3, 200-1 Technology
low-level discharge, 153-4 Mintek carbon activity meter, 815-17
motion of ore through, 137-9 Mintek carbon-in-pulp pilot plant, 445-6,
overflow-type, 153 449, 470
pebble, 121, 160-1, 162, 163 elution, 489-92, 505
power draft, 161, 175, 189-95, 198, elutriator design, 512
212-16 Mintek on-line instrumentation, 265
formulae, 214, 215 carbon and resin concentration, 815
rod, 157-8 gold in solution, 813
speed, 139, 157, 158, 159, 162, 190-1 pulp level, 265
critical, 135, 1068 Mintek steel wool cell, 554, 559, 891
types, 152-3 operating experience, 561-70
Miller, F.B., 615 Mixed potential electrochemical model of
Miller chlorine process, 615, 621-3 gold dissolution, 854-6
Millerite, 52 Mixing
Milling, 121-201 in tumbling mills, 138
autogenous, 160-2 theory of, in stirred vessels, 297-9
automatic control, 199-201 Mixing equipment for CIP, 463-9
continuous, 137 air mixing, 463
design methods, 212-16 carbon breakage by, 581-3
multi-stage, 162-3 draft tube circulators, 463
objectives, 189 mechanical mixing, 463-9
operating principles, 189-99 open tank, 463, 468
pebble, 121, 160, 161, 162, 163, 192 Modder Deep
rod, 157-8 flotation plant, 235
run-of-mine (ROM), 124, 160, 161, Modderfontein, 74
162, 163, 192, 194, 195, 196 residue retreatment plant, 716
semi-autogenous (SAG), 161, 162, Models
193, 194, 195, 196 as aids in plant design, 941
Note: pages 1-614 are in Volume I, pages 615-1072 in Volume 2.
1106
INDEX
Modified guar Ni-hard, 171, 175

depressant, 247, 255, 734 Niccolite, 52
flocculant, 313 Nickel, 17, 30, 33, 56, 497
Moisture adsorption by carbon, 426, 603, 604,
content, effect on screening, 97 608
determination, for mass measurement, effect on cyanidation, 858
785-6, 1019 elution from carbon, 481
Moisture trap recovered by solvent-extraction
of filtrate system, 309 process, 885
'Molasses number', 384 Nicrofer, 519
Monarch Reef, 12, 15 NIM graphite cell, 557-9
Monazite, 16, 33 NIMCIX column adsorption system, 594,
Moncheite, 35 595-8, 603-7
Mono-modal size distribution, 136 advantages & disadvantages, 598
Moodies operation, 595, 598
group, 37, 39 shortcomings, 595
Hills, 46-7 Nitre
Motor truck transport in smelting calcined cathodes, 373, 899
ore, 76 in smelting flux, 371
waste, 112 Nitric acid
Mould trolley, 372 treatment of osmiridium concentrates,
Moulds 230, 232
bullion, 366-7, 368-9 Noise
'button', 364, 368, 372 control, 119, 1003
Movable chutes (for pebble selection), 115 -induced deafness, 989-91
Moving belt feeder, 83, 184 Noise level target, 119
Moving jaw (jaw crusher), 98, 99 Noise pollution, 989
Muffle roasting laboratory, 1048 Noise suppression, 119, 989
Multi-deck shaking tables, 228 North Reef, 12
Multi-stage milling, 162-3 Nuclear conveyor belt massmeters, 784-5
Multiple tray calcining furnace, 361, 362 Nuclear density gauge, 266, 284, 807
Murchison Range Nuclear Development Corporation, 33
gold occurrence, 1 Nuclear Fuels Corporation of South
Muscovite, 15, 20, 27, 52, 257 Africa, 723
N 0
Natalia Reef, 1 0 bservation, task
National Institute for Metallurgy, 50, informal, 999
1026 planned, 999-1001
Needle valves, 817 Occupational Diseases in Mines and
Nernst equation, 551, 837 Works Act, 987
Nett Present Value Odendaalsrus district
as affected by waste sorting, 113-14 gold prospecting, 2
calculation of, 928-33 Off-gases
Neutron activation analysis of from roasters, treatment of, 350-1
Witwatersrand gold, 30 Oleum, 723
New Consort gold mine, 45, 46, 52, 54, OLGA Mark III Gold Analyser, 811
56 Olifants River
New Machavie mine, 26 gold find, 1
Newitt correction On-line determination
for heterogeneous flow, 965 of alkalinity, 328, 815
for homogeneous flow, 966 of carbon concentration, 815
1 1 """
.l.lV/
INDEX
On-line determination (cont.) adsorption onto carbon, 386, 431-5,

of cyanide, 328, 814-15 456, 531
of gold, 811-13 Organic solvent, in elution, 481-3, 486
of pH, 328, 815 Orifice plate
of sulphur, 266 for control of leach-vessel air, 302
On-line measurement for fluid flow measurement, 804
of reagent concentrations, 294-5 Orpiment, effect in gold leaching, 860
Onverwacht group, 37, 38, 44, 45, 46, 48, Osarsite, 35
49 Osborn bar liner, 159, 170, 175
Open circuits Osborn liner bars as grizzley bars, 94
flotation, 263 Osmiridium, 34-5, 173, 222, 228, 230,
milling, 143 232, 366, 369, 370, 619, 620, 626
Open-end discharge, 153, 154, 155, 169 Outlet end liner, 167-9
Open tank mixers, 303, 463, 468 Outlet trunnion, 153
Open top cyclones, 208 Outokumpu
Operation, ease of flotation cells, 260
factors in, 937-8 Outputs, control, 797
Opportunity cost, of money, 928 Overalls, 375, 377, 992, 997, 998
Optimization, non-linear, 791 Overflow classifier, size distribution, 159,
Optimum load volume, grinding mill, 200, 207, 920
191-2 Overflow dam, 953
Orange Free State Overflow density, cyclone, 151, 198, 920,
gold mines, production, 2-4 922
Orangite, 33 Overflow launder, peripheral
Ore classifier, 149
crushing, 98-107 thickener, 277
feed sizing Overflow-type tumbling grinding mill, 153
effect in ROM milling, 192, 195-6 Overflow water, thickener, 284, 289
significance for grinding media collection system, 946
size, 179-81 Overflow weir
feeders, 80-5, 184-6 classifier, 148
fracture, 128-9 tailings dam, 681, 684, 701, 703, 704
lock-up, 77, 79, 80 thickener, 279, 284
run-of-mine, sizing, 179, 180, 181 Over-grinding, 151
storage, 76-80 Overhead ropeways, 76
stream Over-inspection, 974-5
deflection of, for pebble supply, Overload protection, thickener, 281-2
115 Overloading, grinding mill, 198, 199
testing, 909-17, 1019, 1024-5, 1027 Oversize particles
transport, 70, 75-6, 87-91 in CIP plant feed, 459-60
washing, 72, 74, 91-4 Oxidation
see also Refractory ore of base metals during calcining, 361,
Orebody 896
sampling of, 909-10 of base metals during smelting, 363,
Ore-dressing, laboratory equipment, 899
1022-3 role of activated carbon, 387, 389,
functions, 1023-6 392-3
Ore-pass system, 70 state, gold complexes, 831-3
Organic cellulose fibres (as filter aids), Oxides of iron
319 as coating on gold, 32, 58, 59, 219,
Organic molecules 222, 304
1108
INDEX
Oxides of iron (cont.) 128, 129, 135-6, 139-43, 181-3,

in Barberton ore, 58 195-6
in Witwatersrand ore, 15 Passivation, 847-8, 857, 865, 866
Oxidized gold ores in electrowinning, 894
Barberton, 54-5, 58 Pay shoots, 27
Witwatersrand, 32 Peak-seeking technique
Oxygen in mill power control, 200-1
adsorption by carbon, 437 Pebble competence testing, 210-11, 914
adsorption during roasting, 347 Pebble consumption rate, 179, 182
enrichment, in blast furnace, 646 Pebble feed control
in carbon activity, 385, 427-9 automatic, 200-1, 825
in gold cementation, 333-4 Pebble feeders, 185-6
in leaching, 291, 295, 303, 851, 854, Pebble milling, 72, 113, 121-3, 124, 144,
856-7 160-2, 179, 182, 192-6
in pyrite roasting, 347 design, 212-14
partial pressures, 230 operation, 192, 196-9
Oxygen-consuming constituents of gold see also Milling
ores, 58, 857, 858, 860 Pebble selection, 72, 73, 74, 114-15
Oxygen, reduction, cathodic, 851-4, 867 Pebble storage, 115
in electrolytic cell, 894 Pebble transport, 115
Ozone, 866 Pebble wear rate in milling, 179-82,
192-4, 196, 199
p
Pebbles
Pachuca tanks, 122, 295, 298, 299, 300, in conglomerates, 14, 15
301, 321 Pendulum crushing, 210
Packed-bed electro winning cells, 554-61, Penstock decants, 671, 680-2, 683, 684,
891 701-5
Pad dock system, of tailings dam Pentlandite, 13, 16, 20, 23, 52, 53, 57
construction, 658, 669, 670-1, 677, Periodic table, of the elements, 1071
682-3, 686 Peripheral discharge, grinding mill, 139,
Pan conveyor, 83-4, 186 153-4, 169
Pan furnace, 649 Peripheral overflow launder
smelting in, 649-51 classifier, 149
Paragenesis thickener, 277
of Barberton ore minerals, 57-8 Periphery screens, CIP, 471
Parallel rod screening surface, 95-6 pH value
Particle Size Monitor, 131, 807-9 automatic control of, by CYCAD
air eliminator, 808 system, 328, 815
Particles, gold changes at electrodes in
size distribution, 21, 22, 125, 126, 237 electrowinning, 551-2
Particles, ore effect on calcium fouling of carbon,
size distribution, 129-35, 179, 180, 181 437-8
determination of, 129-31, 1036-40 effect on electrowinning, 566
on-line determination of, 131, on anode corrosion, 566
807-11 effect on cementation of gold, 868
relation to grinding media size, effect on elution column corrosion,
143-4, 179-81 519
size effect on settlement rate (Stokes' effect on flotation pulp conditioning,
law), 147-8 264-5
in thickeners, 287 effect on gold adsorption, 423, 425, 426
size of flotation feed, 127, 250 effect on gold dissolution, 838-42, 862
size reduction in milling, 125, 127, of flotation pulp, 243-5
1 1 [){)
_I__l_V/
INDE X
pH value (cont.) milling plant, 188, 189

measurement of, 265, 328, 815 rubber lining, 188, 676, 956
potentia l-pH diagrams (Pourba ix siting, 937
diagrams), 519-21, 838-41 tailings, 976-7
Phenoli c surface oxides, of carbon, 385 transpo rtation, 935-6
Photogr ammetr y velocity
aerial, 712-14 actual, 965
terrestrial, 714 limiting, 963
Photom etric sorting, 109, 110 Piping, acid treatment, 189, 956
Phreatic surface Piping and Instrum entation Diagrams, 957
Pitman (jaw crusher), 98, 99
of tailings dam, 685, 686, 687-9, 691,
Placers, 5, 13-16
692, 694-5
Plane table concentrator, 221, 225-6, 232-3
Phyllosilicates, 13, 15, 20, 24, 25, 26, 27,
31 Planned Maintenance Department, 975
Planned task observations, 971, 973,
Physical Conditi on Rating Programme,
973 999-1001
Physical Conditi on Ratings Plant acceptance tests, 961
checklist, 1002-12 Plant commissioning
cold, 958-9
Pier-mo unted thickener drives, 281
hot, 959
Piezometers, 696-8
Pigg's Peak importa nce of punctuality, 907
area, 41, 49 personnel, 958
mine, 49 practica l tips, 960-1
Pilgrim 's Rest safety, 958
gold prospecting, 1 Plant constru ction
Pilot plants, 917, 1014, 1029-33 metallurgical involvement, 957
design of, 1031 organizations for, 907-8
computer-aided, 1031 Plant control procedures
instrum entation , 1030 development, 820, 1019
Plant design
investigation, types of, 1032-3
mobile, 1032 computer-aided, 941
details, 941-57
operations, 1024, 1031
reasons for, 1014, 1029 general arrangement, 933-41
scale-up, 261, 1030-1 importa nce of good design, 907
siting, 1014, 1031-2 metallurgical involvement, 908, 917
staffing, 1014 organizations for, 907-8
testing, 917 procedure, 908-33
objectives, 1029-30 use of models in, 941
types, 1030 Plant drainage, 953
Plant expansion
Pinch valves, 188, 817
Pine oil, 242-3 provision for, 940-1
Plant layout
Pinions, grinding mill, 155, 157
Pipe racks, 937 principles, 935
Pipe sampler, 776-7 Plant lighting, natural, 953
Pipeline Plant manager designate
descaling, 189, 956 involvement in design and
design, 956, 963-6 commissioning, 957-9
equivalent length, 967 Plant security, 376-7, 939-40, 985-6
of fittings, 967 Plant site
friction loss, 965, 967, 968 selection, 934-5
materials, 685, 956 survey, 933-4
1110
INDEX
Plant structural steelwork, 95 3 Power availability maximization, 189-91

Plate-and-frame filter press, 355, 357, 358 autogenous (ROM) mills, 192-3, 200-1
Plate-and-frame heat-exchangers, 527 ball mills, 191-2
Plate-type screening surface, 96 composite load mills, 193-5
Platforms, operating, 943 pebble mills, 192
Platiniridium, 34 rod mills, 191-2
Platinum group metals, 16, 20, 26, 33, automatic, 200-1
34-5 effect of feed size in ROM and SAG
cause of bullion assay errors, 620 mills, 195-6
in electrolytic gold refining, 627 Power draft
refining by solvent extraction method, tumbling mills in general, 161, 189,
887-8 190-1, 198, 214
research, 1013 autogenous mills, 161-2, 215
Platinum plant tailings, ball mills, 159, 214
size distribution, 668-9 composite load mills, 161, 214
Plies, conveyor belting, 88 pebble mills, 214
Plough discharge, conveyor belt, 90 rod mills, 214
Plough feeders, 84-5 Power utilization maximization, 196-9
Plug flow, in grinding mills, 138 automatic, 201
Plug-type valves, 188, 817 Pre-aeration to oxidize reducing
Pocketed liners, 173 substances in pulps, 59, 304, 859
Pollution Precipitated gold, in Witwatersrand ores,
control, development, 1019, 1024 24
control, tailings dams, 663-5, 666-8 Precipitation
from sulphide roasting, 346 kinetics, 843, 849, 854, 867-9
see also Air pollution Precipitation, chemical
Polyacrylamides, 287, 313 of gold, 889-90
Polyglycol ether, 734 Precipitation, electrolytic, 332, 550-70,
Polymer matrix 890-4
in ion exchange, 877-8 Precipitation feed solutions
Polymers, 24, 287-8 characteristics, 331, 332
Polymetron system, 294 Precipitation process, zinc dust, 333-42
Polypropyl glycol ethers, 242-3 monitoring, 342
Polypropylene pressure candle filter precipitation,
in electrowinning cell construction, 341-2
561-2 solution preparation, 333-7
Polythionates, 858, 861 alkalinity level, 337
Polyurethane clarification, 333
in reciprocating plate feeder, 82-3 cyanide level, 337
on screen decks, 96, 97, 119, 460 de-aeration, 333
on shaking tables, 228 lead addition, 336-7
Pongola Sequence, 37 zinc addition, 336-7
Potassium-bearing phyllosilicates, 27 vacuum leaf (Merrill) precipitation,
Potassium cyanide 338, 339
in elution, 481 Precipitation units
in intensive cyanidation, 857 siting of, 951
Potassium nitrate (nitre) Pre-coat filtration
in smelting flux, 369, 371 for solution clarification, 319-20
Potential, mixed, 854, 866 for zinc-gold slime de-watering, 358-60
Potentiometric titration system Pre-coat mixing tank, 359
CYCAD, 328, 815 Premises
Pourbaix diagrams, 519-20, 838-41 planned inspection checklist, 1002-4
1111
INDEX
Prentice, T.K. outputs, 797

ball wear theory, 182-3 Product size distributions, mill, 196
Present Value, calculation, 927-3 3 automatic control, 199
President Brand carbon-in-pulp plant Profile
AARL elution procedure in, 493 of mill liner, 171, 172, 174-5
calcine treatment, 455-6 Profitability, maximization, 927-33
President Brand gold mine Programmable logic controllers, 823, 824,
residue retreatment, 731-5 825
President Steyn gold mine, 2, 21, 124 Project team, plant design and
gravity concentration circuit, 231-2 construction, 908, 957
residue retreatment, 731 Promise Reef, 7
Pressure candle filters PROSCON supervisory system, 824, 825,
in clarification, 320-1, 322 826, 827
in zinc dust precipitation, 341-2 Protective clothing, 375, 377, 991-2, 998
Pressure precipitation, 341-2 Protective equipment
Pressure pre-oxidation (as alternative to personal, 971, 991-2, 997, 998, 1010
roasting), 229, 237 Proterozoic basin, 37
Pressure sand filters, 318-19 Pulleys, conveyor, 87, 88
Pressure transducer, 818 Pulp, 301-3
Primary concentrator, 221, 224, 232 aeration, 301-3
Primary crushing, 73, 74, 100 conditioning, for flotation, 264-5
Primary mill discharge pulps, pumping density
of, 187 conversion tables, 1062-4
Primary mills, 157, 159, 161, 162, 197 effect on cyclone size splits, 151,
Primary screens, 73, 74, 93 205-7
Prince Albert district in flotation, 265-6
gold occurrence, 1 in leaching, 303-4
Probability of loss, 995 in rod milling, 158
Process control, 745, 793-829 in tailings dam construction, 675-6
analogue, 821-2 measurement of, 807-9
automatic, 820 thickener feed, 284
computer, 822-4 density control
history, 794 in flotation, 265-6
manual, 820 in milling, 191, 197-9
objectives, 795-6 in thickening, 283-4
principles, 795-801 depth in flotation, 249
system design, 819-28 measurement of, 265
Process evaluation, 745-92 dilution (see density)
components, 745-6 filtration rates
definition, 745 measurement of, 1044-6
mass measurement, 780-7 filtration system for extraction of
metallurgical accounting, 787-90 solution sample, 328, 813-14
sample preparation, 778-80 mass flow measurement, 786-7
sampling mean residence time
practice, 759-78 in flotation, 249, 261, 262
theory, 751-9 in grinding mill, 13 8, 198
statistics, 746-50 in leaching, 297-8
use of computers in, 790-1 moisture content, conversion tables,
Process measurements, 801-17 1062-4
Process variables, classification of, 796-7 paramaters
inputs, 797 formulae, 1064-6
111"l
..l.1.1.t,..
INDEX
Pulp (cont.) preferential grinding of, 123

pH control, 265, 810, 815 Pyrometallurgy
pre-aeration, 59, 304, 859 definition, 345
residence time (see mean residence of gold, 345-77
time) Pyrophyllite, 15, 27
sample preparation, 778-80 depression of, in flotation, 246, 247
sampling of, 771-7 effect of pH on flotation of, 243-4,
settlement rates, 324, 325 245
measurement of, 1041-4 Pyrrhotite, 13, 14, 16, 20, 23, 25, 26, 32,
size distribution, 129-35, 157, 159, 50, 51, 52, 57, 237, 345
199, 207 effect on leaching, 858
determination of, 1036-40
of cyclone products, calculation, Q
205-7 Quartz, 14, 15, 20, 25, 26, 31, 33, 43, 46,
transport of, 935-6 47, 51, 52, 53, 54, 55, 246
trash removal from, 284-5, 459-61 Quartzite, 13, 25, 43, 54, 70, 71, 179
viscosity of, 283 Quinone-hydroquinone couple, in
water:solid ratio reducing property of carbon, 386,
conversion tables, I 062-4 392
measurement of, 807 Quinone-type carbonyl on carbon surface,
Pump selection, 956-7, 967, 969 863
Pumping
water, 936 R
Pumps Rabble arms, 348
carbon transfer, 479 Rabbled hearth roaster, 348-9
breakage in, 57 5-81 Radial doors
recessed impeller, 479 of chute feeders, 80, 81, 184
impeller tip speeds for various Radioactive constituents of Witwatersrand
materials, 967, 969 gold ores, 21-4
milling plant, 187-8 Radioactive tracer tests
tailings, 676, 677 in flotation, 267
thickener underflow, 283 Radioactivity in gold ore
vacuum, 313-14 use in waste sorting, 110-11
Punch lists, 959 Radiometric sorting, 110, 111
Punched plate screening surfaces, 96 in waste rock retreatment, 738
'Purple of Cassius' test, 338, 342 Rail liners
PVC gloves, 992 in grinding mills, 170
Pyrite in silos, 79
as cause of refractoriness, 31, 58, 219, Rail transport
237, 345 on mines, 75
concentrates Rake classifiers, 94, 122, 123, 124, 149,
bacterial leaching, 237, 743 150, 163
CIP treatment, 455 Rakes, classifier
roasting, 253, 257, 345-52 reciprocating, 149
effect on gold leaching, 858-9 spiral, 150
flotation of, 251-66, 345-6 Rakes, thickener, 279, 280-2
gravity concentration of, 219-28, 231-3 drives, 281
in Barberton ores, 44, 47, 48, 50" 51, power, 281
52, 53, 54, 57 torque rating, 281
in Witwatersrand ores, 13, 14, 15, 20, hinged, 281-2
25, 26, 27-8, 31, 33, 35, 50, 51, lifting devices, 282
52 monitoring, 290
1113
INDEX
Rakes, thickener (cont.) furnace charge, composition of, 643-4

overload protection, 281-2 modernization, 652
Rand Mines Milling & Mining, tailings pan furnace smelting, 649, 651
retreatment plant, 715-16, 724-31, sampling, 641-3
1015 sintering, 643-4
flowsheet, 724-6 smelter, 641-52
operating problems, 730-1 Rand Refinery, coin blank ·production,
operating results, 731 637-40
solution balance, 730 continuous casting of strip, 639
Rand Refinery, 363, 367, 369, 371, 373, grinding, 639
374, 375, 894-5 history, 637-8
bullion acceptance procedure, 619-20, inspection, 640
894-5 punching, 639
bullion, crude, initial weighing, 619 strip rolling, 639
bullion, crude, melting and sampling, weighing, 639-40
619-21 Randfontein Estates gold mine
by-products treatment, 641-52 Composite Reef, 27
(see also Rand Refinery, by-products Cooke gold plant control system,
treatment) 825-7
chlorides, degolding of, 626 Rate constant
coin blank production, 637-40 adsorption, 393, 394, 395, 396, 399,
(see also Rand Refinery, coin blank 401, 419-20, 422, 425, 429-31,
production) 433, 434, 439-43, 444, 448,
computerization, 652 871-2
functions, 616-17 desorption, 399
furnaces, 621, 623, 625, 626, 637, 639, elution, 489, 490, 491
648, 649 flotation, 261, 262, 272
Gold Refining Branch, 618-30 in regeneration, 534-7
fume recovery, 627, 628-30 leaching, 449
history, 615-16 Rattle, of grinding mills, 198
metallurgical control, 617 Reactor clarifier, 321-2
metallurgical surplus, distribution of, Reactor mixing theory, 261-2, 297-9, 490
617 Reagents, flotation, 239-47, 253, 255,
percentages of gold and silver credited 257, 734
to depositors, 617 effect on adsorption, 435, 456
refining process, gold effect on electrowinning, 568-9
chlorine (Miller), 621-6 evaluation, 734
electrolytic, 626-8 flow control, 266
refining process, silver, 630-7 Reagents, leaching, 291-2
sampling, 619-20, 623, 639, 641-3 addition methods, 293
Silver Refining Branch, 630-7 consumption, 295
smelter, 641-52 distribution, 936
sweep treatment, 630 interruption of supply, 305
see also Gold refining monitoring of concentrations, 294
Rand Refinery, by-products treatment; on-line measurement, 294, 328, 814, 815
641-52 storage, 936
bag filters, 651-2 strength, 291
blast furnace smelting, 644-8 types available, 292
blending, 643 Realgar, effect on leaching, 860
cupellation, 648-9, 650, 651 Reciprocating exhaust pumps (filter
fume collection, 651-2 vacuum), 313-14
Note: pages I - 614 are in Volume 1, pages 615 - 1072 in Volume 2.
1 1 1 A
111 '-t
INDEX
Reciprocating mechanical rakes temperature and residence time,

(classifier), 149 537-41
Reciprocating plate feeder, 82-3, 185, 186 Regrind mills, 159, 216
Recleaner cells, 255, 256, 258, 263 Reichert spirals, 233
Record keeping Reject fraction
in rod and ball mill load level control, in waste sorting, 109, 110
192 Reject rock
of gold-bearing products, 377 disposal, 111-13
of new plant, 958-9 Relining methods, grinding mills, 175-8
Recrystallization Remedial action, responsibilities, 974
of gold, 17, 20, 25, 27 Replacement
Rectangular chutes of matrix material by gold, 20
dimensions, 85 Reproducibility
Redox system, 386, 567 of laboratory flotation results, 271
Re-dressing, of concentrates, 228 Repulper, 311, 718, 719, 950
Reduction ratios Repulper feed conveyor, 309, 718-19, 950
in crushing, 72, 102 Repulper tank, 311
in rod mills, 158 Research programme,
Reef picking, 111 steps in, 1029
Reefs, definition, 5, 69 Reserve Bank, 617, 618
Refined bullion Residence time
see Bullion carbon-in-pulp adsorption, 405, 407
Reflectivity in continuous leaching, 297-99, 305
in photometric sorting, 109 in flotation, 248-9, 257, 260, 261, 262,
Refractory ore 267, 268
Barberton area, 51-4 in grinding mills, 137-9, 198
causes of refractoriness, 51, 219, 237, of ore in plant, 789
304 Residue dams
definition, 50, 345 area required, 934-5
gold particle size in, 237 siting, 934
methods of treatment, 58-9, 220, 237, see also Residue dumps; Slimes dams;
345 Tailings dams; Tailings dumps
see also Ore Residue dams, reclamation
Regeneration, chemical bucket-wheel excavator, 716-19
carbon, 530 monitor guns, 721
ion exchange resins, 883 monitoring, hydraulic 716, 721, 726
Regeneration, thermal, of carbon, 531-50 operations, 720-4, 726
carbon breakage in, 583 Residue dumps
furnace atmosphere, 534-7 retreatment
oxidation of carbonaceous residues, at Blyvooruitzicht, 715, 716-19
533 at Chemwes, 715
principles, 531-33 at East Rand Gold (ERGO), 715,
response of adsorbates, 532, 533 719-24
Rintoul kilns, 547-50 at Joint Metallurgical Scheme
rotary kilns, 541-7 (JMS), 715, 731-6
carbon temperature, 542-4 at Merriespruit, 715
heat transfer, 545 at Rand Mines Milling & Mining,
materials of construction, 546-7 715, 724-31
quenching, 546 role of mine surveyors in assessing,
retention time, 545-6 709
stages in, 531 sampling, 709-10, 720, 910-11
structural changes during, 539-41 Sandrill sampling, 710, 711
1115
INDEX
Residue dumps (cont.) Residues
surveying disposal, 655-706, 934
tacheometric, 711-14 pumping, 676-7, 934-5
values, determination of by Digital use of mine backfill, 112
Terrain Model (DTM), 714 Resin concentration meter, 815
volumes, determining of, 712, 714, Resin-in-pulp process, 884-5
720, 721 research in, 1028
Residue dumps, reclamation Resins, ion exchange, 877-8, 882-3
bucket-wheel excavator, 716-19 Resuscitation equipment, 985
hydraulic monitoring, 716, 719, 721, Retention time
726 see Residence time
pipelines, 720 Retort boats, 352, 353
pump stations, 720 Retort furnace, 352-3
overflow catchment area, 720 Retorting
repulper, 718-19 of gold amalgam, 220-1, 352-4
sand, 726-7 of iron, 231
trash removal screens, 719, 720, 726 weighing of the amalgam, 377
Residue retreatment, 707-43 Return dams
acid production, 716, 723, 731, 735 recycling of gold, 328, 584
bacterial leaching, 741-3 Return idlers, 87, 88
carbon-in-pulp, 455-6, 457, 716, 723, Return water, tailings dam
726, 727-9, 731, 735-6 management, 677-80
leaching, 723, 727, 734-5 system, 657, 660, 680-3, 706
operating problems, at RM3, 730-1 Return water tank, 277, 289, 946
operating results, at RM3, 731 Reverberatory furnaces, 364, 369, 370,
plants, 715-36 371-2, 649
Blyvooruitzicht, 715 Reverse leaching, 31, 257, 264
Chemwes, 715 Rhenosterhoek Subgroup, 6, 7, 8
City Deep, 716 Rhodesian craton, 35
ERGO, 715, 722-4 Rhodium, 34
Joint Metallurgical Scheme, 715, Rhodium sulphide, 35
716, 731-6 Riding ring, grinding mill, 154
M erriespruit, 715 Riffle belt concentrator, 221, 226-7, 231,
Rand Mines Milling & Mining, 232
715-16, 724-31 Riffler, spinning, 762, 778-9, 1022
pyrite concentrates, 723, 729, 735 Riffles, 220, 224, 225, 226, 227, 228
roasting of pyrite concentrates, 723, Riffling, 913
735 Ring main, 294, 936
calcine treatment, 723, 735-6 Rintoul kilns, 547-50
sand Risk areas, dust, 987
process development, 724-6 Road transport
reclamation, 726-7 concentrates, 735
treatment, 727 reagents, 292, 293
solution balance, at RM3, 730 rock, 76, 112
sources of feed, 707-8 stores, 936, 939
treatment of flotation concentrates, Roads, in-plant, 939
723, 727, 729, 733-4, 735 Roasters
treatment of flotation tailings, 723, duplex, 348
727, 734-5 Edwards, 59, 228, 253, 348-9
uranium extraction, 715, 716, 723, fluid bed (fluosolids), 228-9, 347,
731, 732 349-50, 723
Note: pages 1-614 are in Volume l, pages 615-1072 in Volume 2.
1116
INDEX
Roasters (cont.) cake discharge system, 309-11

rabbled hearth, 348-9 capacity, 317
simplex, 348 cloths, 306, 312-13, 317
Roasting, muffle (laboratory), 1048-9 drainage systems, 311
Roasting process feed system, 309, 314
adsorption of oxygen, 347 filtrate system, 309
chemistry, 346-7 maintenance, 316-17
physical reactions, 347 soluble loss, 313
refractory ore, 58, 59, 235, 345-6 calculation, 316, 328-30
testing, 916, 1025, 1032 effect of wash ratio, 315, 316
treatment of gases, 350-1 vacuum, 313
two-stage, 350 pumps, 313-14
types of furnace, 348-50 washing, 314
Rock dumps waste Rotary drum precoat filter, 358-60
construction, 111-12 Rotary filter cycle, 314-15
reclamation, 738-9 Rotary header valve (filter), 306
retreatment, 708-9, 736-40 Rotary kilns, carbon regeneration, 541-7
sampling, 910, 912 carbon temperature, 542-4
siting, 934, 941 heat input, 544
Rock mattresses, 701 heat transfer, 545
Rod mills quenching of carbon, 546
automatic control, 200 retention time, 545-6
description, 157-8 slag formation in, 545
design, 212-14 temperature profile, 543-4
history on South African goldfields, temperature zones, 542
123-4 tube material, 546-7
liners, 158, 167, 172, 175 Rotary sampler, 773-4
maintenance requirements, 157-8 Rotating bowl,
power pulp distributor, 188
draft (formula), 214 Rotating tubular bed reactor, 891
maximization, 191-2 Rotoscoop dewaterers, 94
Rod samples (of by-products), 642, 777 Rougher cell concentrates, 255, 257, 263
Rod tangles (in rod mills), 158, 214 Rougher cell sizes, 253, 254, 256, 260
Rods, grinding Rougher flotation banks, 252, 253, 254,
description, 178 256, 257, 258, 259, 263
specification, 179 Routine plant control, sizing
Roll feeders, 84 analyses for, 131
Rollers (idler), 87 Rubber casing (of conveyor belting), 88
Rolling bottle cyanidation test, 915, Rubber coatings
1046-8 on screen decks, 96
Rolling mill, gold alloy, 639 Rubber lining
Roofs, materials, 953 feed hoppers, 166
Rope haulage of mill end liners and discharge
ore transport, 75 screens, 174, 175
pebble transport, 115 of mill shell, 172, 174
waste disposal, 111-12 of pipes, 188, 676, 956
Roping (of cyclone classifier), 198 of pumps, 187, 188
Rose, Sir T.K., 615 of valves, 188
Rosin-Rammler equation, 133 Running time meter, 192
Ross chain feeder, 80-1 Running time, estimation, 922-4
Rotary converter, 652 available, 923-4
Rotary drum filter, 305-7, 309-12, 317 losl, 922
1117
INDEX
Running time (cont.) precision of, 774-5
scheduled, 922 rotary, 773-4
Run-of-mine milling, 72, 113, 124, 157, dispenser, 811, 813
160-2, 163 increment, 746
costs, 161 mass, in relation to maximum particle
effect of feed size distribution, 195-6 size, 765
feeders, 185-6 preparation, 745, 778-80, 912-13,
history, 124, 162 1016, 1024, 1036-7
start-up, 960 equipment, 778-80
testing, 1032 high grade materials, 778-9
Run-of-mine ore procedures, 912-13, 1019, 1036-7
constituents, 15-16 pulp, 778
crushing, 100, 106 run-of-mine ore, 780, 912-13
feeders, 80-5, 185-6 primary, 912-13
sampling of, 116, 761-70, 909-10 secondary, 913
automatic, 766-70 storage, 269, 1016
manual preparation, 780 Samplers
mechanized preparation, 780, cross-stream, 771-3
912-13 injector, 777
size distribution, 179, 180, 181, 213 non-probabilistic, 775-6
storage, 78-80 pipe, 776-7
see also Ore rotary, 773-4
Ruthenarsenite, 3 5 spinning disc, 770
Ruthenium, 34 Sampler's picks, 373
Rutile, 16 Samples
composite, 746, 764
s obtaining for ore testing, 909-11
Saaiplaas plant, 124 quantities required for tests, 916-17
Safety transport of, 937
maximization of, in plant design, 940 Sampling
training, 974, 983, 985 accuracy of, 746, 758-9, 760-1
Safety boots, 375, 992 auger, 710, 911
Safety factor automatic, 766-7
allowance in equipment selection, 925 tests for bias, 766-70
Safety helmets, 992 t-tests, 766-70
Safety management, 971-1012 bias in, 746, 751, 759, 766-70, 780
Safety Regulations, 91, 106, 117, 973 bullion, 374, 617, 619-20, 623, 778
Safety shut-down, 822, 983, 986-7 by-products, 641-3
St Helena carbon-in-pulp plant, 418, 419, coarse rock dumps, 738, 910, 912-13
470, 478 confidence limits, 746-7
elution procedure, 503, 508, 510, 511, crusher products, 116, 770
512 data adjustment, 791
St Helena gold mine, 2 error, fundamental, 751, 752, 753
Sala unit flotation cell, 27 3 flotation
Sample laboratory, 269-70
analysis, 745 plant, 269
compositing liquids, 778, 811, 813, 814
effect on sampling precision, 764-5 non-probabilistic, 776
crushing, 765, 780, 913 orebody, 909-10
cutters, 771-6 practice, 759-78
cross-stream, 771-3 precautions against dust loss, 912
Note: pages 1-614 are in Volume 1, pages.615-1072 in Volume2.
1118
INDEX
Sampling (cont.) Scatter angle

precision, 746, 751-8 in Microtrac Analyser, 809
calculation of, 751-8 Scavenger flotation banks, 254, 255, 256,
geostatistical methods, 756-8 257, 263, 274
Gy's 50-piece experiment, 756 Schmitt feeders, 164
Gy's formula, 751-5 Scintillometers, 110
probability theory, 751-3 Scoop feeders, 163
probabilistic, 776 Scoops, mill discharge, 153
probe, 814 Scotsman fault, 41
Rand Refinery, 617, 619-20, 623, Scraper
641-3, 777-8 of rotary drum filter, 309, 359
residue dumps, 709-10, 720-1, 910-11 Screen
channel, 709-10 calibration
grab, 709 laboratory procedure, 1038-40
groove, 709-10 cloths, 96-7, 473-6
rock dumps, 736, 738, 910 correction factors, determination,
rod, 642-3, 777 1038-40
run-of-mine ore, 761-70, 780 decks, 95-7, 460
automatic, 766-70 Screen-and-lifterpan mill discharge, 139,
manual, 780 153, 154, 168-9
mechanized, 780 Screening
sand dumps, 710, 910-11 carbon-in-pulp process, 379, 381,
slimes dams, 709-10, 910-11 458-61, 470-8, 479, 480, 511
statistical concepts, 746-50 dry, laboratory procedure, 1038
stop-belt, 761-6 in sorting and crushing plants, 72, 73,
accuracy, 765 74, 93, 95-8
crushing, effect of, 765 interstage, 381, 470-8, 479
mass, of sample, 765 laboratory procedure, 1036-40
precision, 762-5 loaded carbon, 480, 511
surface dumps, 709-10, 910-11 oversize particles, 97, 459-60
theory, 751-9 prescreening of CIP spillage, 480
variance, fundamental, 753 sorting by, 108-9
Sand surface
filtration, 318-19 parallel rod, 95-9
in clarification, 317-19 plate-type, 96
in smelting flux, 371, 373, 897-8 woven wire, 96-7
in thickeners, 289 theory, 97
reclamation, 715, 716, 726-7 waste sorting by, 108-9
residue retreatment, 724-6, 727-9 wet, laboratory procedure, 1037-8
treatment, 122 Screens, Linear, 284-5, 286, 460-1, 480,
Sand bed clarifier, 318 511
Sand dumps, 707-8 Screens, mill discharge, 139, 153, 154,
assay values, 707-8 168, 169
sampling, 709-10, 910-11 Screens, vibrating, 73, 74, 93, 95-8,
Sand-slime 108-9, 114-15, 284, 286, 381, 460
ratio, 122 capacities, 97-8
retreatment operations, 715-36 primary, 73, 74, 93
separation, 121 secondary, 73, 74
Sand tank clarifier, 318 testing, 1024
Sandrill, 710-11 theory, 97
Scale-up method Screw feeder, 293
for flotation plant, 261-3 Screw take-ups, 89
INDEX
Seals in gravity classifiers, 147-8

hopper feeder, 165-6 rate of, in thickeners, 285, 324-5
liner bolt, 174 measurement of, 1041-2
pump shaft, 187 Settling velocity, terminal, 147-8, 965
Seatings (jaw crusher toggle plate), 98, 99 Shaking tables, 221, 227-8, 229, 1024
Secondary concentrator, 221 Shale, 7, 12, 13, 39, 44, 47, 70, 108, 179
Secondary cone crusher, 73, 74 grinding media, 179
Secondary crusher bin, 73, 74 Shear
Secondary grinding circuit, 74, 123, 163, breakage of rock particles by, 135,
213 136
Secondary screens, 73, 74 Shears (tectonic), 47
Security Sheba Fault, 43, 44
maximization of, 939-40 Sheba Hills, 40, 41, 44
risk, in gravity concentration, 222 Sheba mine, 35, 36, 42, 43, 44, 52, 54,
smelthouse, 376-7, 940, 951-2 56, 232-3
Security fence gravity concentration circuit, 232-3
plant, 939 Shell and tube heat-exchangers, 527
tailings dam, 698-700 Shell, grinding mill, 152, 153
Sedimentation, 277 ball mill
rate, measurement of, 1041-2 liners, 159
Stokes' law, 147-8 liners, 170-8
Sediments rod mill, 157, 158
of the Witwatersrand Triad, 5, 7-9 liners, 158
Self-aligning idlers, 88 liners, 170-8
Self-generating flaws (in brittle material), Shield
128 Southern African, 35
Semi-autogenous run-of-mine milling Short-circuit, in leaching, 296, 297, 298,
(SAG milling), 161-2, 187, 193, 299
194, 195 minimization of, 299
automatic control, 200-1 Shower, safety, 985
maximum steel content of load, 194 SI prefixes, 1068
see also Autogenous mills; Milling Siderite, 52
Semi-open flotation circuits, 263, 264 Siemens-Halske electrowinning process,
Semi-variogram, 756-8 332, 890, 891
calculations, 762-3 Sieve series, 129, 130, 131, 1069-70
de Wijsian model, 758 Sieving
estimates of variance from, 757-8 to determine size distribution, 129,
nugget effect, 757, 762 131, 1036-40
range, 757 Silex blocks, 121
sill, 757 Silica, 18, 364, 370
spherical, 758 component of flux, 231, 371, 898
Senmin R55, 255 component of slags, 363, 897-8
Sensitivity analysis grit removal, 459-60
in Present Value estimation, 933 reactions in smelting, 897-8
Separators, magnetic, 230, 1024 Silicates, 20, 25
Sericite, 25, 52, 53 depression, 246
Serpac clarifier, 323 formation of, in slag, 371
Set-points, in process control, 796, 797, Siliceous chemical precipitates
798 (geological), 39
Settlement Silicosis
in gravity clarifiers, 321 precautions against, 117-18
11 "){)
i..1...i:..v
INDEX
Sills bi-modal, 136

geological, 7 crusher products, 116
semi-variogram, 757 determination and reporting, 129-33,
Silos, 79-80 1036-9
Silver, 17, 27, 28, 29, 30, 51, 52, 56, 57 feed, effect in ROM milling, 196, 199
adsorption, effect of ionic strength, gold particles in ore, 21-2, 125-6, 237
421, 873 mathematical descriptions, 131-5
chloride milled products, 136-7, 157, 159, 196,
behaviour of copper and lead in 199
leaching of, 632 mono-modal, 136
coating on gold in ore, 58 run-of-mine ore, 180-1
crude, analysis, 630-1 on-line measurement, 131, 807-10
de-golding of, 626 overflow, cyclone, 200, 207
leaching of, 630-3 Size fraction, definition, 131
reduction by zinc, 632 Size of separation in cyclone classifiers,
cyanide 205
adsorption onto carbon, 389-91, Size reduction
426, 873 mechanisms in milling, 135-7
determination by X-ray spectrometer, types in autogenous milling, 195-6
623 Size segregation of coarse ore, 79, 960
elution, 481, 497, 503, 509 Sizing
gold-silver alloy, 363 equipment, 926
in direct-smelted bullion, 369 flotation cells, 261
in mill gold, 370 ore, 72, 94-8
recovery by chloride leaching, 866 testing, I 024
recovery from by-products, 615, Skips
641-52 rock hoisting, 70, 75, 79
refining, electrolytic, 633-7 waste dumping, Ill, 112
analysis of electrolytic silver, 637 Skirt plates, 86
melting furnace, 637 S kirtings, 86
Moebius cells, 633 Slag
operation, 634-6 blast furnace, analysis, 647
Thum cells, 633 composition, 354
Silver Refining Branch, Rand Refinery, formation during smelting, 363, 648-9,
630-7 897-9
Simpson's Rule granulation, 366, 375, 951
in calculating dump volume, 712 handling, 372, 375, 951-2
Single-pot furnace, 366, 367, 369, 370, 373 pouring, 366, 372
Single-stage milling, 124, 213 removal from bullion bars, 373
see also Milling sampling, 375, 641-3
Sink-float tests, 916, 1024 smelting of, 641-8
Sintering, 643-4, 652 storage cone, 375, 952
Site selection Slags, antimonial
dumps, 934 treatment by chlorination, 866
plant, 933-4 Slaking
tailings dams, 660-8, 934-5 lime, 294
Size analysis, 129-31 Sleeve bearing (gyratory crusher), 101,
procedure, 1036-9 102
reporting, 1038-9 Slide gates, 184
Size distribution Slime treatment
ball mill loads, 184 history, 121-2, 707
1121
INDEX
Slimes dams Sodium cyanide, 291, 292, 422, 481, 482,

assay values, 708 483, 488, 493, 498, 499, 501, 502,
bacterial leaching of, 741-2, 743 503' 857' 877
construction and operation, 655-706 Sodium hydroxide, 481, 493, 498, 501,
retreatment, 31, 253, 715-36 502, 503
sampling, 709-10, 910-11 Sodium isobutyl xanthate, 255
siting, 660-8, 934 Sodium mercaptobenzothiazole, 240, 253,
see also Tailings dams 734
Slimes tailings, delivery pipeline, 676-7 Sodium nitrate
closure, 701 in smelting flux for steelwool
Sliming, cathodes, 899
in thickeners, 284 Sodium silicate (flotation depressant), 246
Slip, in grinding mills, 134, 135, 191, 198 Sodium sulphide (gold elution by), 481
Slipper bearings, 154, 155, 169 Soil
Sludge valves, on hopper clarifier, 321-2 survey, 662-3, 933-4
Smelter department, Rand Refinery, Solids, flow rate, measurement of
641-52 see Mass measurement
Smelthouse Soluble gold loss (in filtration), 313-16
change rooms, 377, 940, 952 calculation of, 328-30
construction, 376, 377, 940, 951, 952 comparison with CIP, 382-3
design, 951-2 see also under Gold
illumination, 376 Soluble lead salts
precautions against gold theft, 376-7 in cementation, 333, 336-7, 866, 868
safety hazards, 375-6 in leaching, 860, 861
security, 376-7, 940, 952 Soluble sulphides, 331, 860, 861
strongroom, 952 Solution trench, 656, 660, 702
ventilation, 952 maintenance, 699, 700-1
Smelting Solvent extraction, 885-9
blast furnace, 644-8 by anion exchange, 886
oxygen enrichment, 646 by cation exchange, 886
slag analysis, 647 by physical distribution, 886
by-product (at Rand Refinery), 641-52 by salvation, 886
calcined slime, 370-2 comparison with ion-exchange, 887
chemistry, 645, 646, 896-9 in OLGA gold analyser, 811
gold bearing material, on-mine, 363-76 in platinum-group-metal refining,
gravity concentrates, 228 887-8
direct, 369 Sorted waste bins, 73, 74, 111, 112
steelwool cathodes, 372-3 Sorting, waste, 70, 72, 74, 107-11
Snellen E test chart, 987, 988 absolute potential for, 113
Snub pulleys (conveyor belt), 87, 88, 89 by screening, 108-9
Socket (gyratory crusher), 102, 103, 104 economics, 113-14
Soda manual, 107-8
as a component of smelting flux, 371, photometric, 109-10
898 radiometric, 110
Soda ash (in gravity concentrate smelting Sound-proof cubicles (for audiometric
flux), 369 tests), 989-90
Sodium carbonate South African Mint, 615, 617, 637, 640
in de-golding of silver chloride, 626 South Reef, 12
in smelting flux South Reef Leader, 12
for calcined slime, 371 Soutpansberg
for steelwool cathodes, 373 gold mining, 1
Note: pages 1-614 are in Volume 1, pages 615 1072 in Volume 2.
1122
INDEX
Spacer bars (in Osborn bar liner), 170 Starch (flotation depressant), 247
Species balance (in flotation), 266 Static grizzlies, 94-5
'Spectator' cations, 388 Static head, determination of, 967
Spectrophotometric analysis, 811-15 Statistical analysis facility
Speed-reducing gearboxes in PROSCON system, 827
in conveyor belt drives, 87 Statistics, concepts in, 746-51
in grinding mill drives, 157 confidence intervals, 747-8
in thickener rake drives, 281 distribution of values, 746-50
Sperrylite, 34 effect of particle size, 750
Sphalerite, 13, 14, 16, 20, 23, 25, 26, 52, kurtosis, 749-50
53, 57, 237 log-normal, 750
Spider (gyratory crusher), 101, 102 normal, 748-50
Spigot (cyclone classifier), 151, 152, 198, skewness, 748, 749, 750
208 histogram, 750
Spigot system of tailings dam population, 746
construction, 673-5, 677, 683, 686 sample increment, 746
S piling bars, 186 standard deviation, 746-7
Spillage handling, 449, 480, 945, 952-3 variance, 746-8
Spillage system, 89 variation, coefficient of, 746-7
CIP plant, 480 Statistical technique
crusher station, 89 for calculation of bias in sub-
filter plant, 950 sampling, 780
leach plant, 948 Steady-head tank, 189
thickener, 944, 945 Steam locomotives
Spillways, 701, 703 for ore transport, 75
Spinning disc sampler, 770 Steel
Spiral concentrators, 1024 for building construction, 953-4
Spiral ramp for elution column construction,
thickener rake lifting, 282 518-22
Spitzkasten, 121, 149 for regeneration kiln construction,
Sponge gold, 230, 352 546-7
melting of, 364, 369-70 Steel addition
see also Gold in autogenous milling, 193-5
Spontaneous disintegration, of Steel bars
stressed brittle solids, 128-9, 140 grizzley, 94
Spout feeders, 164-5 mill liners, 121, 170
Spray nozzles, 93 Steel cord construction (conveyor belting),
Spreader conveyor, 112 76, 88
Spur-rings, 155, 158 Steel-frame bins, 80
Stage grinding, 123, 143-6 Steel grinding balls, 179
Stairways, 943 Steelwool cathodes, 554, 555, 556, 558-9,
handrails, 954 561, 566, 891
treads, 954 acid dissolution of, 373, 896
Stamp mills, 121, 122, 161, 707, 708 smelting of, 372-3
Stamp weights, 121 Stellar candle filter
Standard deviation, 746-7 cla.rification, 320-1
properties of, 747-8 zinc-gold slime recovery, 333, 335,
Standard flotation circuit configuration, 337, 340, 341-2, 355
263 Steynsdorp goldfield, 41, 48-9, 50
Standard reduction potentials, 838-43 Stibiopalladinite, 35
Standby pump sets, milling circuit, 188 Stibnite, 52, 57, 236, 345, 831, 832, 860
Stand-pipe, in mill water system, 189 Stilfontein gold mine, 236, 255, 260
1123
INDEX
Stockpiles, 73, 74, 78-9, 960 Sulphide flotation, 235-74, 345-6, 916
Stock solution Sulphide ores, Barberton, 51-5
of flocculant, 288 treatment for gold recovery, 58-60
Stokes' law, 148, 287 Sulphides
drag force, 204 bacterial leaching, 23 7, 741-3
terminal settling velocity, 147, 965 effect in gold leaching, 858-61
Stop-belt sampling, of run-of-mine ore, flotation, 235-74, 345-6, 916
761-6 mineralogy, 13, 14, 15, 16, 20, 21, 23,
accuracy of, 765 25, 31, 32, 33, 35, 44, 45, 49,
precision of, 762-5 51, 53, 56, 58, 71, 123
Stopes, 69, 70 pressure leaching, 237
support of, 69, 112 roasting, 345-52
Stoppages, plant, 78, 922 soluble, 331, 336
adsorption vessels, 466, 468 Sulphidic-type reefs, 44
Storage Sulphur
start-up considerations, 960 oxidation of, during chlorination, 866
safety, 1003-4 Sulphur analyser, 266
Storage areas, in-plant, 936-7 Sulphur concentrations
Storage tanks in Witwatersrand ore size fractions, 71
for cyanide, 292 Sulphur-containing anions
Storm water effect on gold leaching, 861
catchment paddocks, 657 Sulphur dioxide, recovery from, 346, 349
diversion system, 657 roaster off-gas, 350-1
management, 683-5, 689-90, 701-5 Sulphur grade
Stratification of deck load (of vibrating of flotation concentrates, 253, 255,
screen), 97 257
Streaks (pay-shoots), 27 Sulphur trioxide, in roaster off-gas, 351
Stress (in rock breakage) Sulphuric acid
level of, 136 data sheet available for, 985
tensile, 128 manufacture, 33, 229, 235, 236, 253,
Stringers (conveyor belt), 87, 88 255, 350, 351, 916
Strip production, gold alloy, 639 treatment of coated gold, 229
Struben, Frederick Pine Theophilus, treatment of steelwool cathodes, 373
Submerged arc furnace, 364-6 treatment of zinc-gold slime, 355,
smelting in, 372-3 356-7
Submerged hopper clarifier, 322-3 Sumps
Sub-sampling, 778-80 capacity, 954-5
quantification of bias in, 780 construction materials, 955
Substandard performance, identification, design, 954-5
999 Superelevation, of lifter bars, 175
Substandard safety conditions, Supergene gold enrichment, 55, 57
identification, 974-5 Svpervision
Sudburyite, 35 maximizing ease of, 938-9
Sulphates, decomposition products of Surface breakage of rock particles, 195, 196
iron sulphides, 858 Surface dumps, waste rock
Sulphide concentrates construction, 111-12
laboratory muffle roasting of, 1048 retreatment, 736-40
roasting, 345-52 sampling, 910
Sulphide ions siting, 934
effect on gold electrochemistry, 851 Surface energy, change in, during
effect on gold leaching, 861 comminution, 129
Note: pages 1-614 are in Volume I, pages 615 - 1072 in Volume 2.
1124
INDEX
Surface oxides, of carbon, 385 procedures, 698

Surge capacity, of ore storage, 78 return water dam, preparation, 706
Surge tank, pregnant solution, 322 solution trench, maintenance,
Sutherland, James, I 700-1
Swaziland, gold in, 41, 49 statutory requirements, 698
Swaziland Supergroup, 37 stormwater control, provision for,
Sweep treatment, 630, 641 701-3
Swing Lift thickener rakes, 282 tailings delivery columns, sealing,
Swing stock (jaw crusher), 98, 99 701
Syferfontein Subgroup, 6, 7 top surface, preparation, 703
Symons Short Head cone crusher, 103-4, underdrainage, maintenance, 700
106 walls, preparation, 701, 703
Symons Standard cone crusher, 103, 104, components, 656-7
105 access roads, 656-7, 700
T containment wall, 656, 657, 670,
Table concentrate 671, 674, 690
treatment of, 228, 232, 233 decant system, 657, 671, 680-3
Tacheometric survey, 711-12 delivery system, 657, 675-7, 678,
Tail drum (belt conveyor), 87 693' 701
Tail pulleys, 87 return water system, 657, 677-80,
crowned, 88 706
Tail values, in laboratory flotation, 271 stormwater catchment paddocks,
Tailings 656, 657, 660, 701, 702
delivery to dam, 675-7 stormwater diversion system, 657
pumping, 676-7 toewall, 656, 657, 670, 671, 692
surface disposal, 655 underdrainage system, 656, 657,
underground disposal, 655 670, 671, 674, 687-9, 690
see also Tailings dams condition monitoring, 695-8
Tailings dams, 311, 655-706, 934-5 factors to be monitored, 696
see also Residue dams; Slimes dams formal procedures, 696
abandoned instrumentation, 696-8
problems, 691 record-keeping, 698
area requirements, 657-60 senior management involvement,
for deposition, 657-9 696
for return water system, 660 surveyor's duties, 696
for services, 659-60 construction, 668-75
total, 660 cyclone method, 669, 671-3
barge decant systems, 680, 682-3 factors affecting choice of method,
berms, 684-5 668-9
berm penstocks, 684 mechanical method, 675
buttressing, 692-3 paddock method, 669, 670-1
cycloned, 693 spigot method, 673-5
hydraulic, 693 decant systems, 680-3
mechanical, 693 barge system, 682
catwalks, 682 catwalks, 682
characteristics of settled solids, 685-7 penstock system, 681-2
closure, 698-706 pool walls, 682-3
access roadways, maintenance, 700 delivery to, 675-7
boundary fence, 698-700 distribution, 677
catchment paddocks, preparation, pipelines, 676-7
701 pulp density, 675-6
penstocks, sealing, 703-5 pumping, 676-7
111C'
.l.l,L,,..J
INDEX
Tailings darns, delivery to (cont.) remedial measures, 693-5

valves, 676 repair, 691-5
design, 668-90 return water
effect of operation method, 669 evaporation system, 678
pumping and piping systems, 676-7 management, 677-9
slope stability, 685-90 systems, 680-3
environmental impact, 666-8 treatment by reverse osmosis, 1033
assessment, 666-7 ring main slimes delivery, 677
control, 667-8 rock buttressing, 687, 692, 693
remedial measures, 667 safe slope angle, 688, 689-90
freeboard, 670, 674, 679, 684 seepage, 679, 680, 687, 691
improving, 683 settled solids density, 657, 721
in-wall delivery branches, 677-8 site selection
instrumentation, 696-8 evaluation, 662
inclinometers, 697 exploration, 662-3
piezorneters, 696, 697, 698 fatal flaw analysis, 661
licensing, 666 pollution control, 663-5
life, 659, 934 regional screening, 660-1
method of formation, 668-75 soil survey, 662-3
cyclone system, 671-3, 677 sizing, 657-60, 934
factors affecting, 668-9 slope drainage, installation in existing
mechanical system, 675 wall, 693-5
paddock system, 670-1 slope geometry, 688, 689-90
spigot system, 673-5, 677 adjustment of, 693
observation and c~ntrol, 695-8 influence of, 689-90
old slope stability, 685-90
problems, 691 analysis, 690
operation, 668-75 foundation soils, influence of, 685
cyclone system, 671-3 phreatic surface, influence of, 685,
effect of pulp characteristics, 668-9 687-9, 691
mechanical system, 675 rate of rise, influence of, 657-8,
paddock system, 669-71 687, 688, 689
penstocks, 681 spigot system of construction, 673-5,
problems, 691 677, 683, 686
remedial measures, 692-5 statutory requirements, 665-6
return water management, 677-80 storrnwater management, 683-4
spigot system, 673-5 on side slopes, 684
storrnwater management, 683-5 regulations, 683
penstocks, 680, 681-2 surge capacity, calculations, 684
access to, 682-3 surge capacity, provisions, 684
berm, 684 tailings delivery, 675-7, 678
permeability, effect on slope stability, pipelines, valves, 676-7
686 pumps, 676
permits, 666 relative density of tailings, 675-6
planning, 665-6 siting, 677
pool, 656, 670 spacing, 677
relocation, 693, 694 tailings product characteristics,
pool wall, 656, 670, 682-3 influence of, 685-7
push-down drains, 695 underdrains, 656, 657, 687-90
rate of rise, 657-8, 687 water
reinstatement, 691-5 balance, 679-80
Note: pages 1 - 614 are in V olurne 1, pages 615 - 1072 in Volurne 2.
1126
INDEX
Tailings dams, water (cont.) A.S.T.M., 130, 1069-70

management, 679 British Standard, 130, 1069-70
network, 680 D.I.N., 130
storage, 679 S.A.B.S., 130
see also Residue dams; Slimes dams Tyler, 130, 1069-70
Take-ups, conveyor belt, 89-90 Tests, plant acceptance, 961
Talmage and Fitch method Testwork, metallurgical, 909-17, 1013-51
in thickener design calculations, 324-6 Tetrahedrite, 345
Task analysis, 993-8 Thallium
worksheet, 997 effect on aurocyanide reduction, 851
Task instruction techniques, 983 Thermal regeneration
Task observation, planned, 971, 973, of carbon, 381, 531-50
999-1001 Thickener area
Task procedures, 993 determination of, 324-6, 1041-4
Tell-tale gold analyser, 811 Thickeners, batch (collectors), 277-8
Tellurides, gold, 831, 832 Thickeners, 278-90
Tellurium compounds, as cyanide access tunnels, 945-6
consumers, 219 bridges, 279, 946
Telsmith Gyrasphere crusher, 104-5 capacity, feed, 285
Temperature factors affecting, 285, 287
as factor in filter duty, 312 specific, 287
as interfering control input, 803 capacity, storage, 285, 1043
effect on acid treatment of carbon, circulation, 289
573 crusher station, 73-4
effect on cyanide decomposition, design calculations, 324-6, 1041-4
487-8 drainage systems, 945
effect on electrowinning, 894 draining of, 290
effect on elution, 483, 484, 490, 500, feed pulp, 284-5
505, 506, 876 feed systems, 943, 945
effect on flotation, 251 materials, 946
effect on gold adsorption on carbon, feedwells, 282-3
422-3, 424, 425 final density of settlement
effect on regeneration, 537-41 determination of, 1041-2
effect on thickener capacity, 285, 287 flocculants, 287-8
of carbon in regeneration, 542-4 feeding, 288
Temporary door covers in mill relining, mixing, 288
177 guard cyclones, 945
Tensile stresses as cause of rock high rate, 290
breakage, 128 layout, 943-6
Tensioning liquid/ solid interface level,
of screen cloths, 97 measurement of, 806
Terminal settling velocity operating procedures, 288-90
equation (Stokes' law), 147-8 overflow system, 946
Tertiary ball mills, 159 overflow water, 284, 288-9
Tertiary cone crusher, 73, 74 rakes, 280-2
Tertiary crusher bins, 73, 74 drives, 281
Tertiary cyclone underflow, construction, 281
particle size of gold in, 21 overload protection, 281-2
Testing speed, 281
audiometric, 989-91 torque rating, 281
eyesight, 987-8 sliming, 284
Testing-sieve series spillage systems, 945
1127
INDEX
Thickeners (cont.) of thickener drives, 281

start-up, 288-9, 960-1 Total work (in comminution)
tanks, 279 definition, 140
underflow density control, 283-4 Tourmaline, 16, 20, 46
underflow systems, 283, 945 Toxic substances, data sheets
use as return water collectors, 946 availability, 985
Thickening tests, 324-6, 914, 1019, 1024, for cyanide, 986
1041-4 Tracer tests
Thioantimonites, 860 in flotation, 267-8
Thioarsenite, 861 in leaching, 298, 1019
Thiocyanate Traction rake drive (thickener), 281
as cyanicide, 58 Training
complexes, 835, 876 belt conveyors, 88
adsorption by carbon, 876 Training idlers, 88
elution from carbon, 877 Tramp iron
eluant in resin-in-pulp technology, as cause of gold coatings, 32
883 removal equipment, 107
formation of, 861 Tramp metal detectors, 107
ions, 842-3, 858 Tramp oversize material
leaching, 865, 1033 in cyclone overflow, 204
Thiol-type collectors, 240, 241 in mill liner concentrates, 232
effect on cyanidation, 863 Tramp steel
Thiosulphates prevention of settlement in pumps,
chemistry of, 842, 843, 858, 858 187
effect in cyanidation, 58, 861 Transducers, piezoelectric, 815
leaching with, 835, 864 Transport of ore particles
Thiourea within tumbling mills, 137-9
aurous complex, 835, 842, 843 Transportation, in-plant
adsorption by carbon, 876 minimization of, 935-7
leaching, 864-5, 878, 1033 Transvaal Chamber of Mines, 615
solution, anodic behaviour of gold in, Transvaal Gold Mining estates
845-6 flotation plant, 235
Third Theory of Comminution, 140-2 Transvaal Sequence, 13
Three Sisters region, gold in, 47 Transverse divergence
Three-stage milling plants, 123, 163 of grizzley bars, 94
Throw-away tails (flotation), 273-4 Trapdoors
Thrust bearing (gyratory crusher), 101, for pebble selection, 115
102 Trapezoidal Rule
Thucholite, 14 in calculating dump volume, 712
Timber bins, 80 Trash removal
Tipping-face, of waste dump, 111-12 from pulp stream, 284-5, 459-61
Titanium-uranium minerals, 21 Traube's rule of adsorption, 386
Tjakastad Subgroup, 37, 38 Travelling ways, safety
Toewall (of tailings dam), 656, 657, 661 inspection of, 1002
Toe zone Triethoxy butane (flotation frother),
of mill load, 135, 136, 137 242-3
Toggle plates (jaw crusher), 98, 99 Trip wires, conveyor belt, 91
Tools, hand Trippers
safety, 1005 belt unloading, 90
Topaz, 46 for pebble delivery, 115
Torque rating for pebble selection, 115
Note: pages 1 614 are in Volume 1, pages 615-1072 in Volume 2.
1128
INDEX
Trammel density, 283

grinding mill, 122 density control, 283-4, 290
washing, 92-3 monitoring, 289
Troughing idlers, conveyor belt, 87, 88 pumps, 283, 945
Trunnion systems, 945
drum filter, 305 Underground disposal of waste rock, 70,
grinding mill, 152, 153, 163 111, 112-13
liners, 166-7 Underground slime
Trunnion bearings treatment of, by CIP, 457
drum filter, 305 Underground waste sorting, 70
grinding mill, 152, 153 U nderloading
Tube mills effect on screen efficiency, 97
introduction to Witwatersrand, 121, U ndersize particles
122, 708 'difficult' particles, 97
Tumbling mills effect on efficiency, 97
automatic control, 199-201 in screening, 97
descriptions, 152-63 Undissolved gold tailing, effect of head
design, 212-16 grade
operation, 189-99 equation, 291
power draft U nisel gold mine, 2
factors controlling, 190-1 Unit cell (flotation), 273
formulae, 214, 215 University of the Witwatersrand
maximization of, 189-96, 198 Minerals Research Laboratory, 1026
speed, 135, 190-1 Unloading
critical, 135, I 068 conveyor belt, 90
Tungsten UNS N10276, alloy for elution
recovery, solvent extraction process, column construction, 519
885 Upflow bed contactors (for carbon
Tungsten carbide (in osmiridium treatment of solutions), 381
concentrates), 230 U pflow sand bed clarifier, 318
Turbidity meters (thickener overflow), 284 Upper Monarch Reef, 12
Turffontein Subgroup, 6, 12, 13 Uraniferous kerogen, 14, 16, 21, 23, 24,
Turning tube divider, 778-9 31, 33
Two-stage cycloning, 151 Uraninite, 14, 16, 21, 23, 25, 27, 31, 33,
Two-stage milling plants, 123, 163, 213 236
Two-stage pumping, 187-8 Uranium
Tyler correlation with gold, 27, 110
testing sieve series, 130, 1069-70 flotation, 237, 264
see also recovery from gold plant
u tailings
Ultrasonic leaching, 719, 723, 735, 835
carbon concentration meters, 815 on-line analysis, 266
level detectors, 265, 806 radioactivity
particle size meters, 807-9 in Witwatersrand gold
Ulundi Syncline, 42, 43, 44-5 metallogeny, 24
U nderdrainage system, tailings dam, 657, in radiometric sorting, 110
670, 674, 687-9, 690, 700 recovery from gold plant tailings, 253,
Underflow 715, 716, 719, 722, 723, 731,
classifier, 149, 150, 197, 198, 204, 732, 733, 735
206-7 simultaneous recovery of gold, 835, 885
density, 197, 198 solvent extraction process, 885
thickener, 278, 283 titanium uranium minerals, 21
1129
INDEX
Uranium oxide, recovery from South of adsorption, 396-8

African gold ores, 235 Vapour adsorption by carbon, 379
Uranium plant Variable-speed belt feeders, 83
gold recovery from the primary Variable speed drive, 819
filtrate, 584-5 of pumps, 187
U ranothorite, 3 3 Variables, process, classification of, 796-7
input, 797
v manipulated, 797
Vaal Reef, 8, 12, 23, 24, 27 output, 797
Vaal Reefs mine true value, 803
Elsburg reef treated at, 13 Variance
elution procedure, 493, 499, 503 analysis of, 774-5
laboratories, 1015 calculation of, 751-3, 757-8
No. 9 Shaft control system, 827-8 properties of, 747-8
Vaal Reefs screen (CIP interstage), 471, nested analysis ANOVA experiment,
472 780
Vacuum box properties of, 747-8
in horizontal belt filter, 307, 312, 360 Variation, coefficient of, 747
Vacuum de-aeration calculation in stop-belt sampling,
of precipitation feed solution, 333-6 762-5
Vacuum extraction system, filter, 306, Vehicles, safety, inspection, 1009-10
307, 309, 310 Ventersdorp Contact Reef, 8, 9, 13, 14,
Vacuum leaf filters (Merrill filters), 333, 16, 20, 25-6, 27', 126
337 Ventersdorp lava, 7
Vacuum leaf precipitation, 338-41 Ventersdorp Supergroup, 5, 6, 13
Vacuum pumps, filter, 309, 310, 313 Ventersdorp System, 5
sizing, 314 Venterspost mine, 2, 17, 27
Valve hysteresis, 817 Ventilation
Valve positioners, 818 crusher station, 117
Valves inspection of, 1002-3
control, 817-18 lime slaking plant, 294
ball, 817 plant, 940
diaphragm, 817 Rand Refinery, 628, 651
diverter, 817 smelthouse, 357, 354, 375, 376, 952
globe, 817 Vermiculite (as filter aid), 319
needle, 817 Vezin sampler, 766
plug, 188 Vibrating feeder, 82, 184, 186, 293, 819
flow characteristics, 817-18 Vibrating grizzlies, 95
gate, 188 Vibrating screens, 93, 95-8
in mill pulp circuits, 188 capacity, 98
in mill water circuits, 188 CIP plant
in tailings delivery systems, 676 carbon sizing, 57 5
pinch, 188, 817 interstage, 381, 470
plug, 188 fine carbon recovery, 480-575
rotary header, filter, 306 pre-screening, 459-60
sludge (hopper clarifier), 321-2 pebble selection, 114-15
Vanadium screening surfaces, 95-7
recovery, solvent extraction process, theory, 97
885 trash removal, 284, 286, 459-60
Van der Waals' forces, 385 washing on, 93
Van Deventer's model Village Main, residue retreatment, 716
Note: pages 1-614 are- in Volume 1, pages 615-1072 in Volume 2.
1130
INDEX
Virginia gold mine pumping, in-plant, 936

calcine plant, 455, 456 quality, in flotation, 247-8, 270
flotation plant, 235, 255, 256 treatment with carbon, 379
Viscosity Water Act (Act 54 of 1956), 665-6, 698
apparent, 965-6 Water drainage
of flocculant solutions, 288 from washed ore, 93
of pulps, 147, 205, 283, 965-6 from tailings dams, 657, 662, 686,
units, 1067 687-9, 693-4
Vogelstruisbult mine see also Drainage
flotation plant, 235 Water pressure
Von Maltitz, Count (inventor of stamp at rock-washing sprays, 93
mill), 121 Water reticulation, 189, 937
Von Rittinger, P.R. Water-seal pump shaft glands, 187
comminution law, 139, 140 Weathered ore (Barberton area), 51, 54-5
Vortex finder, 150, 151, 197, 199, 208 Wedge bars (grinding mill liner), 172
super-elevation of, effect on mill
w power, 175
Walls, building, materials of construction, Wedge block
953 jaw crusher gap adjustment, 100
Wash box, conveyor belt, 89 Wedged block liner (grinding mill), 170-1
Wash solution, filter, 306-7, 314, 316, renewal, 176-7
329 Weighbridge, in-motion train weighing
Wash sprays, 89 system, 781-2
conveyor belt, 89 Weighbridge, static, 781
filter, 306, 311, 315, 950 as calibration standard for other
rock washing, 73, 74, 91, 92, 93 weighing equipment, 781
Wash troughs, drum filter, 950 Weighfeeder, 819
Washing Weightometers, 115, 187, 782-5, 950
conveyor belt, 88-9, 950 accuracy, 187, 784-5
rock, 72, 74, 91-4 calibration, 115, 187, 783-4
Washing-out, grinding mill, 176 bulk load, 783-4
Washing tower, 93 chain, 783
Washings handling system, 89, 94, 118 deadweight, 783
Waste rock, 69, 70, 71, 116 electronic, 783
dumps nuclear, 785
construction, 111-12, 655 effective weighlength, 783
reclamation, 708, 736-40 nuclear, 785
retreatment, 736-40 Weights, stamp, 121
sampling, 738, 910 Welding equipment, safety, inspection,
siting, 934 1005
Waste sorting, 70, 72, 74, 107-14 W elkom gold mines, 2, 13
economics of, 113-14, 928-9 Wemco flotation cells, 253, 255, 260
gold loss in, 114, 929 West Rand goldfields, 5, 7, 12, 15
manual, 107-8 West Wits Line goldfield, 2, 12, 13
mechanized, 108-11 Western Areas mine, 20
photometric, 109 Western Reefs mine, 2
radiometric, 110, 111, 738 Wet drilling, in stoping, 69
screening, 108-9 Wet scrubber
Water dust extractor, 118, 292, 349, 376
in electrolytic cell for arsenic recovery, 351
oxidation, 894 White iron (in mill liners), 170, 175
reduction, 867, 894 White Reef, 12
1131
INDEX
Wick drains (tailings dam), 695 491, 503-11, 526, 527, 528, 529,
Williamson controller, 201 553, 561
Windows, materials, 953 Zand pan
Winze-raise connections (in mine flotation plant, 235
development), 69 Zinc Corporation of South Africa Ltd,
Witwatersrand Basin, 7, 10, 25, 33 1013
Witwatersrand Co-operative Smelting Zinc
Works, 615, 616 anodic reactions, 848-9
Witwatersrand gold mines cementation of gold
history, 1-2, 3 chemistry, 866-9
location, 11 Merrill-Crowe process, 331-41
production, 2, 3, 4 pressure candle filter process, 341-2
Witwatersrand Triad, 5-35 -cyanide complex
gold in, 5-33 adsorption by carbon, 426
PGM in, 34-5 elution from ion-exchange resin,
sulphides in, 35 883
uranium in, 21, 23, 27, 31, 33 loading onto ion-exchange resin,
Wohlwill electrolytic process, 627-8, 629 883
Wood chips, 284, 459-60, 480, 512-13 data sheet available, 985
Worcester gold mine, 45 dissolution in cyanide solution, 848-9
Woven wire screening surfaces, 96-7, -gold slime
473-6, 477 acid treatment, 356-7, 895-6
calcining, 361-3, 896
x clean-up, 355
Xenotime, 16, 33 composition, 354
X-ray photoelectron spectroscopy de-watering, 357-60
investigation into gold cyanide smelting, 370-2, 896-9
adsorption, 392 -lead couple, 333, 336, 866, 868
XRF techniques, for on-line uranium passivation, 848-9
determination, 266 recovery by solvent extraction process,
Xylene, 916 885
shavings, 333, 866
y sulphate, decomposition in calcining,
Yuba jigs, 224 896
ZINCOR
z see Zinc Corporation of South Africa
Zadra Ltd
electrowinning, cell, 554, 555, 559, Zircon, 16, 20, 33
891, 892 Zirconium, 27
elution procedure, 481-2, 485, 486, Zwartkoppie Formation, 38, 44, 45
1132

The Extractive Metallurgy of Gold in South Afri (B-Ok - Xyz)

Uploaded by

Copyright:

Available Formats

The Extractive Metallurgy of Gold in South Afri (B-Ok - Xyz)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Extractive Metallurgy of Gold in South Afri (B-Ok - Xyz)

Uploaded by

Copyright:

Available Formats

THE EXTRACTIVE METALLURGY

OF GOLD IN SOUTH AFRICA

THE MONOGRAPH SERIES

Ml Lognormal-De Wijsian Geostatistics for Ore Evaluation

THE SPECIAL PUBUCATIONS SERIES

SPl Proceedings, Underground Transport Symposium (1986)

THE SYMPOSIUM SERIES

S1 Mathematical Statistics and Computer Applications in Ore

This book was commissioned by the Sub-Committee of Consulting

©The Chamber of Mines of South Africa, 1987

ISBN 0 620 10944 0 Volume 2

Typeset by Scientific and Technical Typesetting

Printed in the Republic of South Africa

Chapter 1 Geology and Mineralogy of the Principal Goldfields

Tables and Formulae . .. .. .. .. . .. .. .. . .. .. . . . . .. .. .. .. .. . . .. .. . . . .. .. . . . .. .. .. 1053

Chapter II Disposal of Residues, by G.I. MacPhail and

Components of a residue disposal system ...... .

11.6.1 General ................................................... . 698

Chapter 12 Retreatment of Residues and Waste Rock, by

The Anglo American Corporation Joint Operating

Chapter 13 Process Evaluation, by H.E. Bartlett and

13 .5 .1.1 Calculation of precision of manual stop-

Chapter 14 Process Control by I.M. MacLeod and H.E.

Chapter 15 The Chemistry of the Extraction of Gold, by

15 .1. 3Electrochemical aspects . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 843

Chapter 16 Plant Design and Commissioning, by G.G.

Chapter 17 Loss Control and Safety Management, by

Chapter 18 Laboratories and Pilot Plants, by C. W .A. Muir 1013

Tables and Formulae, by W.R. Flook ........................... 1053

Table 7 Conversion table: pressure . . .. .. . .. .. . . . .. . . . . .. . .. . . 1060

Index ...................................................................... 1073

Chapter 1 Geology and Mineralogy of the Principal

1.3.4 Mineralogical aspects of the Barberton gold ores 49

Chapter 2 Sm1ing and Crushing, by G. G. Stanley . . . . . . . . . 69

Chapter 3 Milling and Classification, by G. G. Stanley . . . 121

3.4.2 Drum feeders ........................................... . 164

Chapter 4 Gravity Concentration, by D.W. Penman ...... 219

Chapter 5 Flotation, by B.R. Broekman, L.A.C. Carter and

Chapter 6 Cymmidation, by G.J.C. Young .................... 277

Chapter 7 Recovery of Gold from Solution by Cementation,

Chapter 8 Pyrometallurgy of Gold, by G.I. Gossman .... 345

8 .2.3 Roasting furnaces .... .. .. .. .... .. .. .. .. .. .. .. .. .. .. .. .. 348

Chapter 9 Application of Activated Carbon to Gold

9.3.6 Operating considerations ............................ . 443

Refining of Gold at the Rand Refinery

10.2 Historical Development

Table 10.1. Throughput of the Rand Refinery

Bullion Fine gold Fine silver

1922 133 230 110 362 6 710

Bullion: 99,85% of the gold

Periodically a pro rata distribution of the value of any metallurgical

GRB 86%1Au ~~~~~~ ~;___ ,

: 1 : ,--- _§!ll3___ -; Sitr :

Not shown for simplicity are:

10.4 Metallurgical Operations

10.4.1 Summary description

10.4.2 Gold Refining Branch

Assay Dip Melting Fume Electro-

Figure 10.3. Weighing mine bullion on arrival at Rand Refinery.