Flotation Theory Reagents and Ore Testing PDF

flotation
THEORY, REAGENTS AND ORE TESTING
Rala D. Crozier
752
rr<n
PERGAMON PRESS
flotation
THEORY, REAGENTS AND ORE TESTING
Flotation has become one of the most important techniques
available for mineral processing. This work provides a unique
and authoriative review of sulphide mineral collector
properties, their manufacture and use with specific ores. Special
emphasis is placed on the different flotation mechanisms
involved in particle capture of sulphide and non-sulphide
minerals and the effect of antagonal mechanisms on reagent
selection.
The author provides details, some previously unpublished, of
the chemical properties, manufacture methods and possible
surface-active impurities of commercial collectors and frothers.
In addition, the chemical composition of a broad number of
North American and European commercial reagent designations
are usted.
Ore sampling, sample preparation, testing machines and
routines are covered as practica! guides to mine laboratory staff.
Suggestions to testing procedures, equipment selection and
graphical data evaluation methods to multivariable problems are
provided.
Arecommended text for practising flotation mili metallurgists,
technicians and professional engineers working in the mining.
and mineral processing industry. Also of interest to laboratories
serving geologists, mining exploration and water treatment
reas.
ISBN O 08 041864 3
9780080418643
FlOTA1'~ON
Theorv, Reagents and Ore Testing
\
Related Perqamon Titles of ~I"lterest
Books
DOBBY and RAO
Processing of Complex Ores
FINCH and DOBBY

Column Flotation
GILCHRIST
Extraction Metallurgy
WILLS
Mineral Processing Technology
Journals
Acta Metallurgica et Materialia
Canadian Metallurgical Quarterly
International Journal of Rack Mechanics and Mining Sciences &

Geomechanics Abstracts
Minerals Engineering
The Physics of Metals & Metallography
Scripta Metallurgica et Materialia
Full details of all Pergamon publications/free specimen copy of any

Pergamon journal available on request from your nearest Pergamon office.
Fl TI
Theory, Reagents and Ore Testing
by
RONALD D. CROZIER, C.Eng., F.I.M.M., Ph.D.

Consulting Engineer, Santiago, Chile
PERGAMON PRESS
OXFORD . NEW YORK . SEOUL . TOKYO
1el .e c';~ q ,
Q4C ./ .....
o;J~/q~
U,IC Pergamon Press pie, Headington Hill Hall,

Oxford OX3 OBW, England
U.5A Pergamon Press, lnc, 395 Saw Mili River Road,
Elmsford, New York 10523, U.s.A,
I<OREA Pergamon Press I<orea, I<PO Box 315, Seoul 110-603, I<orea
JAPAN Pergamon Press Japan, Tsunashirna Building Annex,
3-20-12 Yushirna, Bunkvo-ku. Tokyo 113, Japan
Copyright @ 1992 Pergamon Press pie

AII Rights Reserved. No part of this publication may be
reproduced, stored in a retrieval system or transmitted in any
form or by any means: electronic, electrostatic, magnetic
tape, mechanical, photocopying, recording or otherwise,
without permission in writing from the publishers.
First edition 1992
Library of Congress Cataloging-in-Publication Data
Crozier, Ronald D.
Flotation: theory, reagents, and ore testing/by Ronald D.
Crozier.
p. cm.
Includes bibliographical references.
1. Flotation. 2. Flotation reagents. 1. Title.
TN523.C76 1992 622'.752--dc20 91-33446
ISBN 0-08-041864-3
British Library Cataloguing in Publication Data

A catalogue record for this book is available frorn
G-UIM the British Library
Printed in Great Britain by BPCC Wheatons Ud, Exeter

CONTENTS
lntroducton 1
Chapter 1. Flotaton Fundamentals 5

Definition of terms 5
Reagents employed in flotation 9
Chapter 2. The Mechansm oC Flotation 11

Coursing bubble flotation 12
Contact angle and Hallimond tube measurements 13
Nascent bubble flotation 17
Electron transfer to mineral surface 18
Chapter 3. Sulphide Mineral Flotation 25

Stereochemistry of xanthates and dithiophosphates 25
Six coordination complexes 25
Seven coordination complexes 26
Dithiolate chemistry and surface compounds 29
Thiol collector bonding with sulphide minerals 36
The nature of the bubble-particle bond 40
Frother-collector surface complexes 44
Chapter 4. Thiol Collector Chemistry 46

Properties of xanthates 47
Manufacture of aIkali metal alkyl xanthates 49
Manufacture of thiophosphates 55
Manufacture of xanthogen formates 58
Manufacture of dialkyl thionocarbamates 62
Miscellaneous collector structures \ 64
Chapter 5. Cornrnercial Sulphhydryl Collectors 66
v
_""
Chapter 6. Propertles of Flotaton Froths 85

Flotation frothers and froth hyclrodynamics 85
. Dynamic froth stability 85
The effect of the collector on froth properties 86
The influence of the frother on the rate of flotation 88
Over oiling 92
Chapter 7. Chemcal Properties o Frothers 94

Partially soluble frothers 94
Completely water miscible frothers 98
Chapter 8. Flotation Modifiers 101

Activating agents 102
Inorganic depressants 103
Organic depressants 105
Chapter 9. Sulpbides - Depression and Activation 107

Sodium sulphide 107
Chemical properties of sulphides 107
Molybdenum concentrate cleaning 111
Copper depression in the molybdenite separation 116
Sodium sulphide as a depressant 117
The chemistry of the Nokes reagents 120
Sulphidisation of refractory ores 122
Chapter 10. Mili Tests - Case Histories 127

San Manuel - a non-sulphide moly separation process 127
CODELCO Andina - slime sulphidisation 136
CODELCO El Teniente - molybdenite recovery 147
Chapter 11. Sulpbide Mili Practice 174

Reagent consumption in the flotation industry 174
Copper mill data 176
Flotation process design 200
Copper mill flotation practice 205
Non-sulphide copper ores 208
Flotation of complex ores 209
vi
Chapter u. Noa-metallc !\/1!i!ll!eJrall 1Fliot,il!l1:imn 212
Non-sulphide mineral flotation 212
The nature of oily collector flotaton 213
Structure of mineral surfaces 216
Mechanism of collector adhesion 217
Collectors for industrial minerals 220
Fattyacids 221
Anionic collector chemistry 224
Industrially important fatty acid collectors 225
Cationic collector chemistry 225
Factors affecting selective flotation 228
Metallic oxide minerals mill practice 229
Industrial minerals mill practice 230
Chapter 13. Flotaton Testing 258

Planning a test 258
Experimental design 260
Testing procedures 276
Equipping the flotation laboratory 278
Comparison of laboratory flotation machines 285
Sampling and sample preparation 287
Sampling working concentrators 291
Testing routines 295
Test procedures used by a reagent supplier 297
Recommendations on setting test conditions 303
Batch flotation tests 306
Standard test procedure used by Sherex on coal 310
Column testing 313
Flotation column testing examples 317
Bibliography 324
Subject Index 336
Author Index 343
vii
lIST OF TABlES
Table 1.- Classification of polar minerals 6

Table 2.- Col1ector structures 7
Table 3.- Frother structures 8
Table 4.- Modifiers 10
Table 5.- Xanthate contact angles for selected sulphide mineral 16
Table 6.- Properties of dixanthogens and xanthic anhydrides 31
Table 7.- Dithiolate reagents 33
Table 8.- Reduction potentials for dithiolate /thiol couples 34
Table 9.- Products extracted from thiolated surfaces 35
Table 10.- Mineral electrochemical properties 38
Table 11.- Band-gap energies of sulphide mineral semiconductors 40
Table 12.- Decomposition of Na ethyl xanthate on drying at 50C 51
Table 13.- Properties of alkyl xanthates 51
Table 14.- Xanthate assays, slurry suspension process 52
Table 15.- Commercial stability data on xanthate solutions 55
Table 16.- Density and viscosity of Aerofloat col1ectors 58
Table 17.- Properties of alkyl xanthogen ethyl formates 61
Table 18.- Structures, and suppliers of alkyl xanthates 67
Table 19.- Structures, and suppliers of xanthic esters 68
Table 20.- Structures, and suppliers ofxanthogen formates 69
Table 21.- Structures, and suppliers of thionocarbamates 70
Table 22.- Structures, and suppliers of thiophosphate col1ectors 70
Table 23.- Structures, and suppliers of miscel1aneous col1ectors 72
Table 24.- Properties of alcohols used as frothers 94
Table 25.- Properties of terpenes and ketones used as frothers 96
Table 26.- Properties of water soluble frothers 99
Table 27.- Critical pH for flotation 101
Table 28.- Effect of col1ector on critical pH foc flotation 102
Table 29.- The effect of sodium cyanide on critical pH 102
Table 30.- Sodium sulphide saturation solubility 108
Table 31.- Effect of pH on the equilibrium values of sulphide ions 110
Table 32.- Effect of temperature on the pH of NaSH solutions 111
Table 33.- Reagent practice, moly recovery from porphyry ores, Cu circuit 112
ix
-
Table 34.- Reagent practice, moly recovery'from porphyry ores, moly circuit 113
Table 35.- Regional mine performance 115
Table 36.- Moly plants using the sulphide process - Shirley (1979) 119
Table 37.- San Manuel copper and moly assays 129
Table 38.- Molybdenite plant reagent consumption 135
Table 39.- Microscopic analysis of the moly plant streams 138
Table 40.- Lab results of sand/slime tests, without added reagents 138
Table 41.- Lab results of sand/slime tests, with split after 6 minutes 140
Table 42.- Lab results of sand/slime tests, with split after 12 minutes 142
Table 43.- Plant test of the effectiveness of the sand/slime separation 144
Table 44.- Screen analysis and copper distribution - Andina 145
Table 45.- The effect of NaSH conditioning time on copper recovery 146
Table 46.- Andina - results of test for addition point of NaSH 148
Table 470- Material balances for El Teniente 152
Table 48.- Floatability of samples from different mine sectors 157
Table 490- Material balance and screen analysis retreat regrind 167
Table 50.- Effect of Nokes reagent on rougher recovery 171
Table 51.- Effect of Nokes reagent on first cleaner recoveries 171
Table 520- Moly/copper separation cleaner tests with columns 172
Table 53.- Froth flotation in the UoSAo; source: Ll.S. Bureau of Mines 175
Table 540- UoS. consumption of collectors in 1975 and 1985 176
Table 55.- Consumption offrothers in the total U.S. mining industry 177
Table 56.- Industrial copper mili flotation performance data 179/18
Table 570- Industrial copper mili flotation reagent data 185/191
Table 58.- Average porphyry copper mine performance in 1975 192
Table 590- Collector use frequency comparison for copper milis 194
Table 60.- Frother use frequency comparison for copper mills 196
Table 61.- Collector consumption in copper sulphide ore flotation 198
Table 620- Power and water use in U'S, flotation milis 199
Table 63.- Screen size equivalents 201
Table 64.- Principal sulphide minerals 206
Table 650- Secondary minerals in Pb-Zn complex ores 210
Table 66.- Collector reagents for non-metallic minerals 220
Table 67.- Fatty acid content of oils and fats (weight %) 222
Table 680- The structure of sorne anionic collectors 222
Table 69.- Critical miscelle concentration for anionic collectors 224
Table 700- Nitrogenous cationic flotation agents 226
Table 71.- Cationic amine collector structures 227
Table 72.- Molecular species solubility, pK a , and polar head diameter 228
Table 73.- Commercially important boron minerals 232
Table 740- Glass fiber colemanite specifications 233
Table 75.- Typical composition of spruce pine alaskite 235
Table 76.- Reagents employed in the flotation of feldspatic ores 237
x
Table 77.- Reagents for fluorspar, sulphide and barite flotation 240
Table 78.- Fluorspar grade specifications 242
Table 79.- Specification for commercial quartz and feldspar 242
Table 80.- Reagents employed in the flotation of glass sand in the USA 243
Table 81.- Experimental design and results - glass sands 245
Table 82.- Experimental design - glass sands flotation 246
Table 83.- Reagents employed in the flotation of phosphate rack in the USA 250
Table 84.- Reagents used in potash flotation 252
Table 85.- Reagents employed in the flotation of potash in the USA 253
Table 86.- Scheelite flotation results 253
Table 87.- Mineralogical analysis of Kings Mountain spudomene ore 255
Table 88.- Feldspar, quartz, spudomene minerals in N.e. pegmatites 256
Table 89.- Flotation testing variables 259
Table 90.- Influence of EtX and alpha-terpineol consumption on grade 268
Table 91.- Influence of EtX and octanol consumption on concentrate grade 269
Table 92.- Three-variable five-Ievel Box-Wilson design layout 273
Table 93.- Flotation cell dimensions and standard flotation test conditions 285
Table 94.- Reproducibility of time-recovery curves for Leeds cells 287
Table 95.- Parameters to determine sample size for different minerals 288
Table 96.- Column flotation variables used in reground pulp cleaner study 318
Table 97.- Overall experimental plan for the RRC samples in the 2" column 320
Table 98.- Optimum flotation conditions for the 2" column 321
xi
UST OF FIGURES
Fig. 1.- Cavitation rnechanisrn of bubble forrnation 18

Fig. 2.- The electrical double layer rnodel 21
Fig. 3.- The structure of ice 22
Fig. 4.- Localized bond structure for water 23
Fig. 5.- Cation and anion hydration showing water orientation 24
Fig. 6.- General stereochernistry for [M(bidentateh]x:!: 26
Fig. 7.- [CuII(xanthate)z] or [{Cu(xanth)Z}xCuIIx_z(HzO)z] 27
Fig. 8.- Pentagonal bipyramid and capped octahedron geometry 27
Fig. 9.- [PbII(ethyl xanthatejjlone pair)] 28
Fig. 10.- [PbII(di-isopropylphosphorodithiolate)z(lone pair)] 29
Fig. 11.- [M(bidentateh(lone pair)]X:!: 30
Fig. 12.- Structure of chalcopyrite and enargite 39
Fig. 13.- Wurtzite and sphalerite, from Pauling 39
Fig. 14.- Leja and Schulman (1954) interpenetration theory 43
Fig. 15.- Frother-collector interaction complex 44
Fig. 16.- Frother volume versus alcohol chain length 87
Fig. 17.- Effect of frother concentration 88
Fig. 18.- Iso-recovery curves, Hallimond tube 88
Fig. 19.- Iso-recovery curves for chalcocite 89
Fig. 20.- Rate of flotation of copper ores 90
Fig. 21.- Performance of frothers on recovery of coal/ash 91
Fig. 22.- Illustration of two laboratory time-recovery profiles 93
Fig. 23. Ionization of sodium sulphide solutions 109
Fig. 24.- Relative effect of the addition of NaSH on pH 111
Fig. 25.- Relationship between pH, NaSH depression and floatability 118
Fig. 26.- San Manuel molybdenite recovery circuit 128
Fig. 27.- Andina rougher circuit in 1976 136
Fig. 28.- Andina rougher circuit in 1977 137
Fig. 29.- Rougher circuit for Andina's sand/slnne mili test 141
Fig. 30.- Overall molybdenum. recovery, El Teniente 149
Fig. 31.- Simplified flotation schematic for El Teniente 151
Fig. 32.- Sewell ore CuT, CuNS and % oxide copper content 154
Fig. 33.- Sewell, Colon and combined ore non-sulphide assay 155
xiii
Fig. 34.- Mineral content of Teniente concentrate - 1984 155
Fig. 35.- Monthly moly assays and tonnage produced by sectors 156
Fig. 36.- Monthly moly assays for 1989 for Colon and Sewell 156
Fig. 37.- Floatability of samples from different sectors of the mine 158
Fig. 38.- The alkaline c1eaner flotation circuit at Colon 160
Fig. 39.- Performance of the copper c1eaner circuit 161
Fig. 41.- The effect of pH on pyrite and molybdenite recovery 163
Fig. 42.- Effect of pH on Cu and moly recovery from second c1eaner 164
Fig. 43.- Molybdenum content of Teniente tailings water 164
Fig. 44.- Potential-pl-l diagram for moly-sulphur-water 165
Fig. 45.- First cleaners and copper circuit flow diagram 166
Fig. 46.- Effect of grind on copper and moly recovery 168
Fig. 47.- Moly content vs particle size - Colon 169
Fig. 48.- Moly content vs particle size - Sewell 170
Fig. 49.- Proposed new Colon cleaner circuit 170
Fig. 50.- Redox measurements of rougher flotation - Teniente 172
Fig. 51.- Overall mill performance, world survey 193
Fig. 52a.-Collector frequency graphs, world survey 195
Fig. 52b.-Frother frequency graphs, world survey 197
Fig. 53.- Complex ore differential flotation circuit 204
Fig. 54.- Effect of grind on copper recovery from ores 208
Fig. 55.- Forces in adsorption and chemisorption 216
Fig. 56.- Water clathrates 218
Fig. 57.- Effect of brine composition on ratio of oleate chemisorption 228
Fig. 58.- Feldspar, mica and quartz flotation, Spruce Pine, NC 236
Fig. 59.- Weathered decomposed pegmatite processing, N.e. 236
Fig. 60.- Fluorspar milI, Reynolds Mining Corp., Texas 238
Fig. 61.- Mill circuit for a fluorspar-zincheavy media concentrate 239
Fig. 62.- Pilot plant flowsheet for sand beneficiation 244
Fig. 63.- Lock cycle tests and pilot plant - glass sands 244
Fig. 64.- Flow diagram for a phosphate rock plant 249
Fig. 65.- Potash scrubbing-desliming-reagent conditioning system 251
Fig. 66.- Foote spudomene flotation circuit 254
Fig. 67.- Spudomene and feldspar-quartz-mica flotation, Kings Mt, NC 255
Fig. 68.- Results of 2 kg laboratory flotation tests, single variable search 261
Fig. 69.- A two-variable factorial design 262
Fig. 70.- Experimental design based on a square pattern 264
Fig. 71.- Cubic experiment for a truncated 3 by 3 design 265
Fig. 72.- Effect of frother and collector on copper recovery 266
Fig. 73.- Collector/frother interaction on recovery and grade 267
Fig. 74.- Interaction of octanol and EtX on recovery and grade 270
Fig. 75.- Box-Wilson rotatable two-variable five-level design 272
Fig. 76.- Box-Wilson rotatable three-variable five-level design 273
xiv
Fig. 77.- False optimum in a one variable at a time search 274
Fig. 78.- Steepest ascent experimental procedure result 275
Fig. 79.- SSDEVOP: climbing a response surface, from Mular 276
Fig. 80.- Standard 5.5 liter (2 kg) flotation cell from ASTM 281
Fig. 81.- Flotation cell proposed by ISO committee on coal 282
Fig. 82.- Flotation cell impeller and diffuser 283
Fig. 83.- Comparison of the performance of flotation machines 286
Fig. 84.- Minimum representative sample size 289
Fig. 85a.-Standard 2 kg ball milI used in Chile (body) 301
Fig. 85b.-Standard 2 kgball mill used in Chile (cover) 302
Fig. 86.- Schematic of a flotation column 314
Fig. 87.- CIMM's column flotation pilot plant 317
Fg. 88.- CIMM recommended flowsheet for cleaner column flotation 322
xv
ACKNOWlEDGEMENTS
The author acknowledges with thanks the assistance given by the folIowing companies
and publishers in permitting the reproduction of illustrations from their publications:
John Wiley & Sons Inc., 605 Third Ave, New York, NY 10016
Stephen J. Lippard, Editor, Progress in inorganic chemistry, Volume 23, copyright -
1977; containing: D.L. Kepert, "Aspects of the stereochemistry of six-coordination",
Figures 6.7,8 & 9; and Michael G. B. Drew, "Seven-coordination chemistry", for
Figures 10 and 11.
Gordon and Breach Science Publishers, 270 Eight Avenue, New York, NY 10011,
Janusz S. Laskowski, Editor, "Frothing in flotation", copyright: 1989; containing Ronald
D. Crozier, and Richard R. Klimpel, "Frothers: plant practice", Figures 15, 20, 21, and 22
Plenum Publishing Corporation, 233 Spring Street, New York, NY 10013

Jan Leja, "Surface chemistry of flotation", copyright: 1982, Figures 2,55 and 57
Cornell University Press, 124 Robert Place, Ithaca, NY 14850

Linus Pauling, "The nature of the chemical bond", copyright: 1960, Figures 3, 12, and 13
Australasian Institute of Mining and MetaIlurgy (Inc), Clunies Ross House, 191 Royal
Parade, Parkville, Victoria 3052, Australia
K.L. Sutherland, and I.W. Wark, "PrincipIes of flotation", copyright: 1955, Figure 25
The Society of Mining Engineers, AIME-SME, Littleton, ca

Leja, J., & Schulman, J.H. (1954), Flotation theory: molecular interaction between
frothers and colIectors at solid-liquid-air interfaces, Trans. AIME., 16, pp.221-8. Figure
14, depicting the interpenetration theory
Mular, A.L. (1989), Modelling, simulation and" optimization of mineral processing
circuits, Challenges in mineral processing, Eds., K.V.S. Sastry and M.e. Fuerstenau,
SME-AIME, Littleton, ca, pp.323-349. Figure 77, 78, and 79
Mular, A.L. (1976), Optimization in flotation plants, Flotation, Ed. M.e. Fuerstenau,
SME-A1ME, Littleton, ca, pp. 895-936. Figures 75, and 76
Redeker, Immo H. and Bentzen, E.H., (1986),Plant and laboratory practice in non-
xvii
metallic flotation, Chemical reagents in the mineral processing industry, Editors, Deepak
Malhotra and W.F. Riggs. SME of AIME, Littleton, ca, pp. 3-20 Figures 58, 59, 61,
and 67
'rile Institution of Mining and Metallurgy, 44 PortIand Place, London, W1N 4BR
Grainger-Allen, T.J.N. (1970), Bubble generation in froth flotation machines, Trans.
I.M.M.;]2, pp. C15-C22, figure 1
Lekki, J., & Laskowski, J. (1971), On the dynamic effect of frother-collector joint action
in flotation, Trans. I.M.M., Section CM, C174-80. figures 16, 17, and 18
Mathieu, G.L and Sirois, L.L. (1984), New processes to float feldspathic and ferrous
minerals from quartz, Reagents in the minerals industry, M.J.Jones and ROblatt, Eds,
(London: IMM), pp. 57-67; Figure 62, and 63
Mine and Quarry Engineerng, 19 Fairlie Road, London, UK

Wrobel, SA, (1953), Power and stability of flotation frothers, governing factors, Mine
and Quarry Engineering; 19, pp. :?63-7, Figures 16
Pergamon Press, Headington Hill Hall, Oxford OX3 OBW

Beas-Bustos, E., and Crozier, RD. (1991) Molyfcopper separation from concentrate of
the combined acid and basic circuits at El Teniente, Reagents in Minerals Engineering,
Camborne School of Mines, September 18-20. Figures 31-51
Crozier, RD., (1991) Sulphide collector mineral bonding and the mechanism of
flotation, MineralsEngineering, vol 4, Nos 7-11, pp.839-858, Figure 5
The author is also grateful to Mr. Farlow Davis for permission to use unpublished
material on molybdenum depressants; to CODELCO-Chile division management for
providing the data used in the case histories on the Andina and El Teniente mines; to
the Director of Chile's Centro de Investigation Minera y Metalurgica for providing the
column testing data; and to Sherex Chemical Company, Inc, which provided flotation
testing procedures and details of the test equipment. Sabine Slotta is thanked for the
many tedious hours spent copy editing the manuscript.
xviii
INTRODUCTION
This text is primarily addressed to the mill metallurgist and to technical

service and development engineers working for reagent suppliers, rather than
to flotation researchers or mill process designers. With this audience in
mind, flotation theory is covered so as to emphasise the existence of distinctly
different mechanisms in the flotation of simple metallic sulphides, complex
sulphides, metallic non-sulphides (loosely known as oxides), industrial
minerals, and coals.
As a primary target of flotation testing is correction of a chemical problem in

a working mill, be it the flotation media (water) or the selection of a
collector, frother or modifier to improve the recovery efficiency of a mill, a
considerable amount of space has been devoted to attaching the proper
chemical composition to commercial trade names, as well as providing
relatively detailed information on the properties of the industrial collectors,
frothers and modifiers in normal use. Reagents produced by former major
collector suppliers such as Minerec Corporation (out of business), and the
Dow Chemical Company, no longer in this niche business, are also identified,
to aid in the interpretation of older published mill data and because many
laboratories still have samples of Dow and Minerec reagents which can be
used to broaden a particular flotation study. A brief description of the
synthesis of collectors is given in case special samples need to be prepared,
and their physico-chemical properties are summarised, as these data are
generally not available in standard handbooks. Water quality, per se, is too
vast a subject to try and summarise, but as the interaction between collectors
and cations in solution is a common problem in mineral pulps, sorne of the
coordination chemistry of collectors is touched on, as is the semi-conductor
properties of the principal sulphide minerals. .
\
The technology to treat low grade ores has evolved by trial and error in the
operating mills, and very little is based on fundamental research. Most of the
published fundamental research has been of relatively little value in resolving
practical problems, primarily because it has been carried out using synthetic
mixtures of pure minerals. In addition, test equipment normally used in
- 1-
2 FLOTATION - Theory, Reagents and Ore Testing
micro-flotation laboratories usually mimics the obsolete pneumatic flotation

machines of the turn of the century. Thus, the results of much of the
published fundamental research cannot be applied to the design of sulphide
mineral flotation circuits. Because of mechanistic incompatibility of agitation
flotation and pneumatic flotation, some of the conclusions and mathematical
relations proposed in the literature on the rate of flotation of pure minerals is
particularly dangerous if it is used in the development of scale-up algorithms
and in choosing the key process variables in the design of experiments. This
criticism of simplified flotation testing equipment does not mean that there is
not a crying need for an accurate micro-flotation apparatus capable of
obtaining scale-up data and accurate processability responses from gram-
sized samples obtained from core drilling programmes. BUT, just as the
grade and mineralogical data must be inserted into a representative 3-D grid
of an ore body to generate reliable geologic modelling of a mining project's
reserves, the same broad analysis must be applied to the floatability response
to provide useful recoveries and operating costs. The statistical confidence
range of the floatability model based on bench flotation data is poor, and if
based on micro-flotation experiments it is considerably poorer than the
reserve estimate because chemical analysis of exploration cores is more
accurate. Thus, designing a mill based on current theoretical knowledge
using micro-flotation data could well be disastrously more expensive than
investing in adits, tonnage samples, and pilot plant tests.
The deve!opment of an optimum metallurgical sequence to pro cess a

complete!y new ore is always expensive, therefore it must be carefully
budgeted in all new projects. An improper geologic evaluation of the
probable ore grade and reserves can have serious financial consequences for
a mining company if the ore runs out before all the capital invested has been
recovered, so valuation of an ore deposit justifies a large expenditure in
drilling and exploration. Similarly, unexpected processing difficu1ties, which
result in a recovery reduction of 3 or 4% on predicted efficiencies, will
affect the cash flow values used to calculate the debt service, and can also
have a crucial effect on corporate existence, considering the mammoth
investments required for today's low grade ores. So the allocation of
resources, human and financial, to testing budgets during the pre-investment
decision period must be carefully weighed.
Much has been written on how to design and manage the proper exploration
of an ore body and the design of an optimum drilling programme.
Computer programmes are capable of processing drill, core and tunnelling
data, and presenting these in terms of three dimensional topologies of the
mineral species, and spewing out the proven, probable and inferred reserves
of the future mine, including an indication of the confidence leve! of the
Introduction 3
quantity of ore that will be available. However, no similar computer

capability has been standardised in the more mundane area of testing ores
for processability, particularly if flotation is involved.
The chemistry of mineral surfaces is dealt with systematically, and is readily

available in standard publications, though not necessarily for conditions
applicable to flotation. The review of the mechanisms of flotation
summarised in this text is slanted to emphasise factors that affect the
selection or design of flotation reagents, not equipment and process circuits.
The greater emphasis and breadth of information on sulphide reagents
reflects the writer's expertise and not the relative size of the industrial
minerals reagent market. The testing procedures designed for sulphide ores,
whose behaviour is more complex, can all be easily adapted to non-sulphide
ores and also to non-metallic flotation, cleaning or separation processes, so
sulphide testing is generally covered more fully.
As will be apparent from the text, the selection of appropriate reagents for
the optimum flotation of a particular ore is more an art than a science; but
the use of a systematic approach to obtain laboratory flotation data can
reduce the hit or miss aspects of this particular art. Information on the
performance of flotation reagents is frequently contradictory because non-
identical flotation circuits may be conforming to a different flotation regime,
and therefore the reagents in the full-scale plants act different1y from the
case histories described in the literature. Even if we limit our review to
current1y fashionable flotation equipment, we must qua1ify any reagent
recommendations, as frothers and collectors will not react in the same
manner if the flotation is carried out in small subaeration machines, forced
air units, modern giant cells or flotation columns. Quite obviously, there are
also operational differences between the generically similar machine designs
of different manufacturers, which can affect the choice of the reagents.
This mechanistic factor, i.e. that a particular reagent's flotation response is

different, depending on whether the flotation occurs because of bubble
capture, or mineral surface catalysed micro-bubble formation, is glossed over
in most texts on flotation. In these notes it will be over-emphasised, because
it is so important in diagnosing prablems in cleaner or scavenger circuits, and
in extrapolating applicability of a particular reagent from one ore to another
or from one type of flotation equipment to ~other (i.e., from cell to column
flotation, where only bubble capture occurs).
To add complexity to our information matrix, there are at least four different
types of minerals that are upgraded by flotation: (1) metallic sulphide
minerals, (2) metallic oxides or tarnished sulphide minerals, (3) non-metallic
insoluble minerals, and (4) non-metallic soluble, or semi-soluble minera!s. In

each case the mechanism of flotation can be bubble capture or nascent
bubble formation, or a combination of the two.
Sub-groups of these four minerals are ores whose mineral dissemination

requires that they be ground to colloidal sizes to liberate the valuable species;
minerals where differential flotation is required to separate complex ores that
include mixtures of hydrophyllic minerals; and those that are naturally
hydrophobic and amenable to flotation without the use of collectors. It is also
obvious that the milI metallurgist, the flotation equipment designer, and the
reagent designer have quite different goals, and therefore possibly conflicting
points of views, on how to apply the technology used in flotation. Hardware
designers, for example, should be aware that designations such as a "strong"
or a "weak" collector do not describe the strength of adhesion between a
mineral and a bubble; rather, it is an inverse description of selectivity, or the
effect of a particular reagent on the rate of flotation of individual minerals, so
a change to a 'stronger collector' in a formu1ation is unlikely to compensate
for poor over-all recovery because of an overly deep cell. When industrially
proven flotation machines are involved in a side-by-side mill test and one is
not performing competitively, its inadequate performance cannot be
compensated by a reagent suite change, as the effect of the reagent change
could affect the operation of both machines identically, in which case the
cause of the difference is more likely to be due to a physical operating
variable, such as pulp density, percent solids, air feed rate, impeller rpm,
mineral particle size distribution or sorne other operating detail that has not
been fully controlled during the test periodo
Froth flotation, the surface dependent process par excellence, consists of a

deceptively simple procedure - the separation of valuable constituents o a
two or three phase mixture by levitation - but, in fact, flotation embodies a
very broad gamut of non-obvious sequential micro-processes, the
combination of which have as a final result the levitation of a desired mineral
species, leaving other minerals and gangue in the tailings. The same reagent
and micro-process sequence does not act the same when applied to different
minerals, i.e. non-metallics, metal oxides and metal sulphides can be
depressed or promoted by a particular reagent. This behaviour results in
what appear to be completely contradictory results reported in research
published by different reputable metallurgists. The reason for this
inconsistency is the great diversity in mineral morphology of commercially
important raw materials. Thus, there is no single valid theory of the
mechanism of flotation which can be used to design universally applicable
testing procedures, and therefore no 'cookbook' mineral processability test
recipes.
Chapter 1
FLOTATION FUNDAMENTAlS
Definition of terms
Frotlt flotation is a process used to separate minerals, suspended in liquids, by

attaching them to gas bubbles to provide selective levitation of the solid
partic1es. It is the cheapest and most extensively used process for the
separation of chemically similar minerals, and to concentrate ores for
economical smelting.
Floatable minerals can be c1assified into non-polar and polar types, according
to Wills (1988). The segregation into these two types of minerals is based on
their surface bonding. The surfaces of non-polar minerals have relatively
weak molecular bonds, difficult to hydrate, and in consequence such minerals
are hydrophobic. Non-polar minerals inc1ude graphite, sulphur, molybdenite,
diamond, coal, and tale, all naturally floatable in the pure state. The ores
containing these minerals usually require the addition of non-specific
collectors to their pulp to aid the natural hydrophobicity of the floatable
fraction, i.e. oily collectors, such as fuel oil, kerosene, coal distillates, etc.
Polar minerals have strong covalent or ionic surfacebonding, and exhibit

high free energies at these surfaces. Therefore, surface hydration is rapid due
to the strong reaction with water molecules which form multilayers on the
mineral. Thus these species are hydrophyllic. The listing in Table 1 is
adapted from Wills (1988). The minerals are listed in groups of increasing
polarity, divided into c1asses dependent on the magnitude of the polarity.
Minerals in group 3 have similar degrees of polarity, but this polarity can be
changed easily in the case of the minerals grouped under 3 (a) because they
are susceptible to sulphidisation.
\
The minerals in group 1 are all sulphides, with the exception of the pure
metals. Their flotation characteristics are not uniform, and, based on
Finkelstein and Poling (1977), must be further divided into three more sub-
-5-
groups: those rninerals that form a xanthate, dithiophosphate, mercaptan,

thionocarbamate, or xanthogen frmate dimer (dithiolate) on their surface
on contact with the collector (covellite, chalcopyrite, stibnite, arsenopyrite,
pyrite, sphalerite, orpiment, realgar, and the native Au, Pt, Ag and Cu); those
that form metal xanthates (galena, bornite, chalcocite);
1
and the others, where
the analytical data is ambiguous on whether the collector bond is with the
metal cation or with the surface sulphur atoms in the mineral.
TabLe 1.- CLASSIFICATION OF POLAR MINERALS
Group 1 Group 1 continued
GaLena PbS Orpiment As2S3

ChaLcopyrite CuFeS2 Pent Landite CFe,Ni)9S8
CoveLL ite CuS ReaLgar AsS
Bornite CUS FeS4 Cinnabar HgS
ChaLcocite CU2S ALabandite MnS
Energite CU3AS,Sb)S4 Native Au, Ag, Pt, Cu
Argentite A92 S
MiLLerite NiS Group 2
CobaLtite CoAsS
Arsenopyrite FeAsS Barite BaS04
Pyrite FeS2 Anhydrite CaS04
Pyrrhotite FE7S8 Gypsum CaS042H20
SphaLerite ZnS AngLesite PbS04
Stibnite Sb2S3
Group 3a Group 3
MaLachite CU2C03(OH)2 FLourite CaF2
Azurite 2CuC03CuCOH)2 Witherite BaC03
ChrysocoLLa CUSi03 2H20 Magnesite Mg C03
WuLfenite PbMo04 DoLomite CaMg(C03)2
Cerrusite PbC03 Apatite CaSCCF,CL)CP04)3
ScheeLite CaW0 4
Smithsonite Zn si Li cate cLay
Rhodochrosite MnC03
Siderite FeC03
Monazite CCe,La,Di)P04
Group 4 Group S
Hematite Fe203 Zircon ZrSi0 4

Goethite FeO(OH) Hemimorphite Zn4Si207COH)2.2H20
Chromite FeCr204 BeryL Be3AL2Si6018
PyroLusite Mn02 Garnet Ca3AL2CSi04)3
Borax Na2B407
WoLframite CFe,Mn) W04
CoLumbite CFe,Mn)CNb,Ta)206
TantaLite FeTa206
RutiLe Ti0 2
Cassiterite Sn02
source: WiLLs (1988)
Flotation Fundamentals 7
Table 2.- COLLECTOR STRUCTURES

Type Structure Water Solubility
Soluble Collectors
S
Alkyl xanthates or R-O-C-SNa Soluble

Alkyl dithiocarbonates
Sodium 0,0 Dialkyl R-O S Soluble
dithiophosphates
R-oIP(s-Na+
Insoluble Oils R-O-8H very sltly

Mercaptans soluble
2-Mercaptobenzothiazol ~Je-s 8
-+
Na
Soluble
O,O-Aryl,O-aryl Ph-O S
dithiophosphoric
acids Ph-oIP(s-Na+ Insoluble
Dialkyl xanthogen s O
formates
R-O-C-S-C-O-R' Insoluble
Dialkyl thiono- S
carbamates
R-O-C-NHR' Insoluble
S
Xanthic Esters Insoluble
R-O-C-S-R'
Metal oxide and non-sulphide mineral collectors
O
Slightly soluble
Fatty acids R-C-OH in water
O
Sulphonates R-S-O-OH Soluble
O
Alkyl amines R-HNH2 Soluble

Quaternary ammonium \
compounds RR'R"R"'NCl Soluble
The reason to make this differentiation is that the mechanism of flotation is

unique when the collector reacts with the mineral surface to form a
dithiolate. Nascent bubble flotation is the rate dominant mechanism, when
s FLOTAnON - Theory, Reagents and Ore Testing
dimers are Iorrned, while coursing bubbles dominare when there is true oily
collector flotation, With the exception of pentlandite, ambiguous minerals
are less often encountered in the flotation laboratory,
TabLe 3.- FROTHER STRUCTURES
Type Structure Water SoLubiLity
ALiphatic ALcohoLs R-OH s Light

R = ALkyL gp with 5 to 8 C
MISC CH3CHCH2CHCH3
CH3 OH
Pine OiLs ell

TerpineoLs HC
/~~ ""-:::CH HC I
/CH" CII
C( 3
I I ZI CII I % sLight
Z
HC, )H HZC I CHOII
CH ""-ClJ/
I I
"3 C-
1-CH
OH
3 CIJa
~ - te~pifll!(Jl fl!flchyl alcoool
CresyLic acid
sLight
ALkoxyparaffins /OCH2CH3
1,1,3-triethoxybutane CH3CHCH2C-H sLight
or TES I \CH 2CH3
OCH2CH3
PoLygLycoLethers
Dowfroths R(OR' )xOH miscibLe at Low 'x'
to partiaLLy
Aerofroths R(OR' )xOH soLubLe at high 'x'
The largest volume flotation operations are those that involve the upgrading
of commodities: iron ores, coal, phosphates, limestone, potash, copper... but
the technically most interesting are the separation of similar metals from
complex ore, such as mixed lead, zinc, silver and copper ores, or
copper /molybdenum concentrates.
Flotation Fundamentals
Bu/k flotation is a sornewhat imprecise term that covers nearly a1l the normal
rougher or scavenger flotations, where a single mineral or a group of related
minerals are separated from gangue and other low value minerals in a single
flotation step. An exarnple would be the recovery of a mixed copper
sulphide concntrate from an ore containing pyrite and gangueo Others
inelude the separation of KCI from sodium chloride, upgrading phosphate
rock, desulphurising iron ores, etc.
Differential flotation is the terrn norma11y used to describe the separation of

complex ores, and is generally restricted to describing the separation of
similar minerals from each other (e.g. copper, Iead, zinc, silver and gold from
a single ore, or molybdenum from copper in concntrate, etc.) where the
successful and economic recovery of each component involves the
sophisticated use of collectors, depressants and flotation activators.
Reagents employed in flotation
The reagents employed in flotation are generally interfacial surface tension

modifiers, surface chemistry modifiers, and/or flocculants. Usua11y they are
elassified under five headings: co11ectors (sometimes known as flotation
promoters), frothers, modifiers, activators and depressants. The designation
co11ector or promoter, interchangeably, reflects the opinion of the individual
metallurgist, who may consider the primary function of the collector to be a
flotation rate accelerator, or a fine particle agglomerator, in a particular
flotation system.
Collectors are reagents that coat and/or react with mineral surfaces and
make them water repellent or attachable to air bubbles. The chemical
structures of most commonly employed collectors are shown in Table 2.
Sulphide ore collectors a11 contain sulphur and are thiols or can hydrolyse to
a thiol. Non-sulphide and non-metallic minerals are normally floated
employing collectors such as fatty acids, amines, quaternary ammonium
compounds, sulphonates or petroleum oils.
Frothers are surface active reagents that aid in the formation and
stabilisation of air-induced flotation froths. The commonly employed
frothing agents are aleohols which are only slightly soluble in water, or the
more modern frothers, which are generally varieties of polyethers or
polyglycol ethers that are completely miscible with water. Examples are Iisted
in Table 3.
Modifiers, activators and depressants The boundaries between the functions

of a specific in<?rganic flotation aid are rather fuzzy. For example, for pH
la FLOTATION - Theory, Reagents and Ore Testing
control, an environment modifier such as lime may be used; but lime

contains caleium, and the caleium cation is a known depressant for pyrites in
copper flotation or for quartz and tale in the flotation of silver ores. So lime
would need to be c1assified under both headings. The silicates and the
phosphates act as modifiers when they are used to control the effect of
cationic and anionic impurities in flotation water, but can also act as
activators or depressants in different mineral systems. Another example of a
dual function is that of the sulphhydryl anion which can be a depressant for
sulphide copper, especially in the copper/ moly separation, but, with very
careful concentration control, this anion can be used as a sulphidiser to
activate non-sulphide or oxide copper.
TabLe 4.- MODIFIERS
pH Modifiers
Lime:
Soda ash:
Caustic soda:
Acids:
Resurfacing agents
Cations: Ba++, Ca++, Cu++, Pb++, Zn++, Ag+
Anions: Si0 3, P04-- CN-, C03' S=
Organic CoLLoids: dextrin, starch, gLue, etc.

Chapter 2
THE MECHANI5M OF FlOTAT~ON
The morphology of froth flotation is not well defined because full-scale

systems are normally opaque. In macro terms, froth concentrates form by
selective capture of mineral particles in the bubble generation zone, followed
by a second zone where there is coagulation of the discrete bubble stream
into a loose froth. These first two zones are norma1ly identified as the "pulp"
in a flotation cell, and can be fixed in height by a level control mechanism.
The third region appears when the gas phase predominates over the liquid
phase; here, the froth starts to condense and returns water to the pulpo
Final1y a structurally stable mineral-laden froth forms, and is removed over
the lip of the flotation machine. The height (or thickness) of this stable froth
can be control1ed by the designer and operator by fixing the total froth height
and the pulp height. The operating variables, that can be fixed by the
foreman within the flotation bank, are: pulp mineral content (pulp density)
and flow rate; aeration gas rate; in some cases agitation intensity (in
conventional cells); pulp/froth interface height; and total overflow height,
which fixes the froth drainage height and the froth removal rateo
A material balance algorithm for a single cell can quantify both the pulp mass
flow rate in and out of the cell, and also the concentrate mass rate, if assays
are available for the different streams. At a lower confidence level, the rate
of return of water to the pulp, and the rate of flow of the different
components of the concentrate, can also be calculated. Under steady
conditions, the height of the interface between the concentrate froth and the
draining froth can be inferred from froth physical property measurements,
such as density or metal concentration profiles,
The key mechanistic variable that is undefined is exactly how a bubble

se1ectively captures a mineral containing ore particles. Observation confirms
that for selective capture the desired mineral must be natural1y floatable or
be preferential1y coated with a collector. Once the desired particle is tagged,
there appear to be two main, non-exclusive, bubble loading mechanisms in
operation: a) multiple particle bubble collisions in the pulp which eventually
result in particle adhesion, and b) micro-bubble formation directly on the
mineral surface within the bubble formation zone of the pulpo There is
consensus that the forces that join the mineral particle to the bubble, and
-11-
that define the rate of capture of the mineral by the bubble, involve electron
transfers when the mineral is conductive, and surface tension forces when
the mineral is an insulator.
Coursng bubble Ilotation
The froth flotation literature labels these two particle capture processes as
the coursing bubb1e process, and the air precipitation 01' nascent bubble
process (the latter wiIl be used in this text). Bubb1e capture, 01' coursing
flotation, predominares when oily collectors are used to concentrate non-
metallic minerals or non-sulphide metallic ores, 01' when collectors
containing double bonded sulphur atoms are used with metallic sulphide ores
which do not form dithiolates on the mineral surface, but ionise and form
metal salts. As we see from the data in Table 53 (page 175), the bubble
capture mechanism occurs mainly with minerals where the concentration
ratio is less than 5:1, and selectivity is usually obtained by employing specific
depressants rather than making collector adjustments.
In these cases, the flotation efficiency of different reagents can be measured

by simple tests such as Hallimond tube experiments. though even with these
simple ores one would hesitate to extrapolate the se1ectivity results of
Hallimond tube flotation tests (which normally are carried out using only a
collector) to a specific industrial flotation, as the concentrate obtained is
primarilydependent on the rate of flotation. As is weIl known, the flotation
rate depends more on the frother employed, rather than the collector.
With oily collectors, contact angle measurements, and the corresponding

calculated adhesion force between a bubble and a mineral particle, may be a
guide to the mineral carrying capacity of a bubble, at least when very dilute
pulps are involved; but under normal industrial flotation conditions, where
pulps contain at least 30% solids, the rheology of the inter-bubble fluid is the
property that fixes the mineral-carrying capacity of the froth, not the strength
of the particle to bubble bond, though the differential value of this force for
two or more minerals and the gangue may help predict the se1ectivity of a
reagent.
KIassen and Mokrousov (1963) describe the fundamental difference between

the collectors for these two types of flotation by pointing out that: "In physical
adsorption, the reagent maintains its chemical individuality, whereas, in
chemisorption, [when e1ectron transfers occur] it forms new compounds".
They confirm other researchers, such as Gaudin (1957), that when the
collector acts as a micro-bubble nucleator, the dosage must be such that only
a small proportion of the mineral surface is covered by the collector, which
Mechanism o' Flotation
means that the surface of a Iloatable sulphide mineral is j<rOT hydrophobic,

i.e., in highly selective flotation with a concentration ratio greater than 7:1,
for bubble capture to occur with a sulphide mineral, the bubble contact angle
ar the optimurn reagent dosages must be zero or near zero,
Sorne of the effects of physical variables on bubble capture flotation can be

deduced from the principal characteristics of oily reagents, which are: 1) a
tendency lo form miscelles when the concentration exceeds a critical
concentration (Clvlt,') and temperature (Krafft point); 2) a significant surface
tension reduction in dilute solutions; and 3) dissociation, and, as a function
of pH, ionisation through hydrolysis. The nuances of the application of non-
sulphide collectors to flotation of minerals is covered extensively by
Sutherland and Wark (1955)0 As in the case of thiol collectors, the oily
coliectors also show a depressing effect, due to "over-oiling", if the dosage is
such as to produce a multi-molecular layer on the particle surface, Oily
coliectors are generally less selective than the thiols and have a serious
drawback, which is the need to heat some flotation pulps, as their coliector
action is related to miscelle formation and for the longer chain alkyl groups
their Krafft temperature is considerably aboye ambiento
Contact angle and Hallimond tube measurements and their meaning
Contact angles, as used by Sutherland and Wark (1957) in their studies to

predict the effect of critical pH and depressant concentrations on floatability,
provide go-no-go data on whether a particular coliector will or will not form
a monolayer under given pulp conditions, The authors recognise that they
only measure whether this mono- to multi-layer of coliector has or has not
adhered to a polished sample of a pure mineral, as can be seen from the
foliowing passage from page 97 of their classic text:
"No matter from what source they [the minerals] came, each
gave a constant maximum contact angle with ethyl xanthate,
and moreover, this angle was invariably 60 oo. Metals gave
the same angle, and it was immaterial whether the sodium
or potassium salt was used. This angle of contact is
therefore characteristic of ethyl xanthate; any surface that
has responded fulIy to the colIector .exhibits it. There can
only be one explanation for this fact, namely, that the
mineral becomes coated with a film ~f xanthate, and that
the effective surface is, in fact, the xanthateitself."
Other authors have pointed out the static nature of contact angle
measurements and the danger of extrapolating contact angle results to
Ilotation systems in general, be they sulphide or non-sulphide. For example,

Laskowski (1974) cites Klassen and Mokrousov (1963) saying:
"These authors point out that flotation is typically a non-

equilibrium process. The phenomena that occur on the
surfaee of particles and bubbles in agitated, aerated mineral
pulps, when these particles strike bubbles, are dynamic-non-
equilibrium processes which eannot be analysed based on
contact angle values obtained at equilibrium".
Leja (1982, p.7) says:
"in the opimon of the author, the contaet angle is an

indieator, but is not a measure of the hydrophilic character
[of an ore]".
The seminal work on the signifieance of contact angles in fiotation has been
done by Gardner and Woods (1973, 1974, and 1977), and summarised by
Woods (1976, 1977, 1981). The key to their work was the design of a cell
which allowed the simuitaneous measurement of contact angles and a surface
potential on an eleetrode, and a miero-fiotation device equipped with an
electrode capable of transferring an electrieal potential to a bed of mineral or
metal particles whieh permitted dynamic measurement of electrochemical
potentials and flotation reeoveries. The initial work was with gold and
platinum, which was then extended to pyrite and galena surfaces.
Based on contact angle measurements, their major eonclusions were:
"The presence of dixanthogen on the [platinum or gold]

surface clearly renders the surfaee strongly hydrophobie....
when xanthate ions are adsorbed on platinum at potentials
cathodic to dixanthogen formation, the contact angle in this
region remained close to zero and henee specifical1y
adsorbed xanthate ions do not make the surface
hydrophobic" .
In addition they found:
"It is interesting to compare the results of investigations of

the interaction of xanthate with a galena surfaee and with
lead. Lead methyl, ethyl and butyl xanthates on the metal
surface were found to be hydrophilic even at low eoverage.
Only at high potentials, where lead xanthate is oxidised to
Mechanisrn of Flotation 15
dixanthogen, were significant angles observed. H the

potential was decreased after the dixanthogen was Iormed,
the immediate contact angle rernained that characteristic of
dixanthogen. However, dixanthogen slowly reacts with the
surface to reform lead xanthate, the contact angle decreased
and the surface became hydrophilic".
And finally very pertinent to the thesis in this text:
"It has been generally considered that chemisorbed xanthate

will be attached to a metal atom in the sulfide surface.
However, it has been pointed out by Winter (1975) that
xanthates can form bonds with sulfur, compounds of the
type ROCSS-S-SSCOR being quite stable. Also, such
compounds cannot readily be distinguished from
dixanthogens by spectroscopic analysis and could weIl be
formed at sulfide surfaces. The possibility of involvement of
such compounds, and the instability of surface species which
give rise to the release of monothiocarbamates into
solution, add complications to the general picture of the
interaction of sulfides with coIlectors...u
Data on contact angles fOI xanthates from which many of the generalisations
on the properties of coIlectors have been based are reported by KIassen and
Mokrousov (1963), and surnmarised in Table 5. Extending the range of the
data in Table 5, Taggart (1947) cites a series of contact angle measurements
where the nurnber of carbon atoms in the xanthate alkyl group went from 1
to 16, and the contact angle increased from 50 for methyl xanthate, steadily
to 80 fOI amyl, 87 fOI hexyl, 90 fOI 7 carbons, 94 for 8, and 96 for 16
carbons (cetyl).
Statements in text books that longer chain alkyl group xanthates are "stronger
collectors'' are based on these contact angle values. In fact, beyond sorne
minirnum bond strength between the mineral and the bubble, persistence of
adhesion is onIy a function of the froth rheology and turbulence. What does
correlate with contact angles for xanthates is that longer chain alkyl groups
coincide with lower selectivity, so amyl xanthates wiIl result in the production
of a larger volume, lower grade, rougher concentrate.
\
Xanthates with alkyl chains longer than hexyl provide lower flotation
recoveries. Decyl and longer chain xanthates are good depressants. A weIl
known, but generaIly unpublished, property of xanthates is that they may be
used as emergency copper depressants in moly/ copper circuits. Operators
habitually correct for a momentary ineffectveness of the inorganic Cu

depressant by the addition of more primary copper collector, be it a xanthate
or xanthate derivative. Curiously, the amount of additional colIector that will
act as a depressant with a particular moly circuit pulp corresponds to that
amount which will generare a measurable bubble contact angle in the lab.
An nteresting experiment (which 1 believe has not been performed) would
be to see if a colIector concentration which generates any contact angle
corresponds to an over oiling dose industrially,
We have commented on "over oiling", a term which was first used around the
time of World War 1 to describe the phenomena that excessive amounts of
colIector in a flotation systern will result in a decrease in recovery. This
decrease in recovery with increased dosage tends to be ignored in published
papers on flotation research.
The reason for this omission seems to be that Hallimond tube and other
micro-flotation tests do not detect the drop in recovery with excess collector
and/or frother, and for many of these papers the only flotation
measurements carried out have be en on a micro-scale.
TabLe 5. XANTHATE CONTACT ANGLES FOR SELECTED SULPHIDE MINERALS
N of Carbon GaLena ChaLcopyrite Bornite Pyrite

atoms in xanthate
1 O O O O
2 59 60 60 60
3 68 69 68 67
4 74 78 73 74
5 88 90 86 82
6 100 94 95 95
Effect of Branching
ButyL Xanthate
NormaL 74 78 73 74
Branched 77 78 78 78
The Hallimond tube has been very popular in research laboratories because it
eliminates the cost of assaying head, tailings and concentrate. The need to
assay is avoided by studying artificial ores made by mixing pure minerals.
The most common system has be en a mixture of galena and sand (or pure
quartz). To perform a flotation experiment in a Hallimond tube, a very
dilute pulp (usually about 1% solids) is placed in the bubble generating unit,
which is the size of a test tube. The tube has a side arm welded to the upper
Mechanism of Flotation 17
regin; to operare, the unit is tilted until the liquid is level with the side arm,
and air is turned on to the bubble generating device (usually a glass frit).
The concentrate overflows into the side arm, is collected for a fixed time, and
weighed. As a frother normally sequesters collectors and interferes with the
operation of Hallimond tubes, they are usually omitted. Thus we have the
anachronism that a device without a controllable froth is used to simulate
froth flotation.
Because Hallimond tube flotation does not require the use of a frother, the
effect of frothers on flotation performance also tends to be ignored in artic1es
reporting fundamental research, In addition, many of these articles only
report the collector concentration in the water phase, and do not inc1ude the
pulp density. Therefore, it is important to translate the published reagent
dose shown in recovery curves into mill terms (grams or pounds per ton of
dry mineral), using a best estimate of pulp density (1 - 30%?) if unspecified,
when evaluating whether the information reported has any possibility of
being pertinent to mill problems. Typical sulphide ore flotation dosage of
collectors used industrialIy are about 0.02 to 0.2 pounds per short ton of ore,
or 10 to 100 gm per metric tonne of ore. These ranges translate to
approximately 5 mgjl to 50 mgjl or 0.029 to 0.29 mol/litre of the water
phase, for the molecular weights of the typical xanthates studied.
Nascent bubble flotatioo
Klassen and Mokrousov, in their "Introduction to the Theory of Flotation",

postulate that the firstbubble to adhere to a mineral particle is a micro-
bubble (formed by desorption of dissolved gases) which serves as a bridge for
the adherence of a bubble large enough to be capable of levitating the
particle. Klassen proposed this mechanism based on a thermodynamic
analysis showing that the bridging micro-bubble accelerates macro-bubble
attachment rates. Taggart, already in 1930, had observed that it was probable
that micro-bubbles were involved in flotation, and that they were formed on
the mineral surface by precipitation from dissolved gases in the cavitation
region of the pulpo Summarising work published over the past 20 or 30 years,
flotation is thought to occur because bubbles form on the surface of a
valuable mineral partic1e at points selectively covered by frother-collector
molecules complexes. Simultaneously, secondary bubbles are formed at
miscellaneous electrical discontinuities that do not provide strong bonding
between the partic1e and its bubble; i.e. in the case of gangue partic1es.
Possibly, if Klassen is right, the mineral particles first nuc1eate micro bubbles
generated from dissolved gases, and these micro-bubbles then act as bridges
for the formation of larger bubbles or the capture of larger bubbles that
could be recirculating in the impeller zone. The important point is that, in
18 FLOTAnON - Theory, Reagents and Ore Testing
the flotation process, there is arate limiting step, which is the velocity of the
formation, or attachment, of a micro- or macro-bubble to the mineral
surface. Once the bubbles have attached themselves to ore partic1es, the
bubbles rise in the pulpo First they form a lattice which appears like a loose
sponge, then later the bubbles tighten into a froth. This tightening process
results in the squeezing out of a great part of the occ1uded water, which
returns to the pulpo This returning water mechanically c1eans the froth and
detaches the gangueo During this c1eaning stage there may be sorne
secondary capture of valuable mineral that has fallen from bubbles in the
stable froth region, but this secondary capture is not a controlling step in the
overall mineral recovery process.
In an agitation flotation machine, air is normally drawn into the impeller

through a hollow shaft by mechanical suction induced by the rotating
impeller, or with a positive air supply. After the air reaches the impeller
region, the flotation bubbles form by ,a two-stage process. First, cavitation
"sausages'' form at the trailing edges of the impeller blades. These break up
to form bubbles on the surface discontinuities of the mineral partic1es. These
discontinuities can be chemisorbed frother-collector complexes adhered to
the mineral surface, or naturally hydrophobic areas.
Grainger-Allen (1970) studied the formation of froths in a flotation machine

equipped with flat rotor blades. He employed a cryolite pulp where the
refractive index would allow photography of the bubble formation process.
His high speed movies confirmed that the bubbles are formed from cavitation
induced air sausages which break up into bubbles onto the mineral partic1es
(shown schematically in Fig. 1).
Fig.1.- Cavitation mechanism ofbubble formation
Electron transfer to mineral surface
Sulphide collector adsorption on to a sulphide mineral requires the transfer

of two electrons to the surface of the mineral partic1e. This electron transfer
mechanism is very similar to the electrochernical process of metal corrosion,

During the corrosion process, metal goes into solution at a point on its
surface in contad with water, where the electrolyte is lacking in dissolved
oxygen; to go into solution the metal atom becomes a cation by losing the
number of electrons corresponding to its normal valance. To maintain a
constant surface potential, these electrons must travel through the conductive
metal to an o),.)'gen rich region where they take part in an oxidation reaction,
the result of which is that metal cations in solution, that have diffused from
the point of dissolution, are converted into insoluble metal oxide; so, in iron
corrosion, the rust is deposited far from the point where a pit is being
generated. In flotation, the electrochemical process is very similar. The
collector molecule approaches the metal surface and is adsorbed only after it
donates or has donated electrons to the mineral partic1e. So as not to
generate a charge on the partic1e, the excess electrons migrate to an oxygen
rich point on the partic1e surface where they are consumed by reduction of
adsorbed oxygen to form OH- ions, or H 20, or H 202, raising the pH, a
phenomena very easily measured during the adsorption of collectors on
minerals. Under most practical flotation environments, deaerated pulps will
DOt float, and it is thought that sorne depressants (sodium sulpbide, for
example) may block flotation by scavenging oxygen. The reaction involved in
xanthate adsorption on a mineral, see Woods (1976), are therefore:
ELECTROH TRAHSFER ro A "IHERAL SURFACE
step 1 s . .s
step 2
su sn
2 (RO C-S >ads---+ (RO e-s>2 + 2 e-
ol'
s" _
su ..s
(RO e-s ~ds + RO C-Ss~l;;-+ (RO e-s>2
These reactions are ba Ian ced electrochenically by:
~02 + 2H+ + 2e- ~ H 20
~02 + H20 + 2e ~
20H-
+ 2H+ + Ze
- \
02 ~ HZOZ
Indirect evidence of this last reaction has been provided by Jones and
Woodcock's (1978) discovery of perxanthates in tailings pulp from flotation
20 FLOTATIN - Theory, Reagents and Ore Testing
mills in Australia. They postulare the reaction to form perxanthates in a

flotation eell to be as follows:
s s
2ROC-S + H202 --~ ROC-S-O- + H20
and point out that the only source of hydrogen peroxide in a pulp would be
the last reaetion aboye.
The nature of the dixanthogen forrned will depend on the metal cation
involved. In the case of galena the infrared speetra show that the anodie
reaction involved is
s s
PbS + 2ROC-S + 1/202 + 2H+ ----~ PbCROC-S)2 + So + 2e
The electrical double layer theory, originally proposed by Helmholtz in 1879,

evolved from a postulate that when a erystalline solid surface is generated by
a fracture, the eleavage process severs the ionic bonds and leaves an
electrieal imbalance on the surface, and, if the solid is in contact with water,
this imbalanee is compensated by a eharge gradient. This primitive
Helmholtz model was refined by Perrin in 1904, and is shown in Fig. 2 as case
(a), labelled as the compact layer theory; illustration (b) is based on Gouy
(1910) and Chapman (1913) who refined the model to inelude dilute
solutions, based on the assumption of the existence of a diffuse continuum of
ions in a non-structured die1eetrie layer around the mineral partiele; (e) adds
ion interaction as proposed by Debye and Huekel (1923) and Stern (1924),
which ineludes the premise that the individual ions can approach the mineral
surface up to a minimum distance "d". This theory in fact is a triple layer
theory, as can be seen from the illustration. While (d) corresponds to a case
where a specific ion in solution has entered the double layer and resulted in a
charge reversal.
IHP OHP IHP= Inner- He Isoho ltz P Ieme

OIHP= oueee He Iraho ltz p In ne
Solvated cations
-1-+
.L~
'" sx Water dipole in
edsorbed anion in IHP
the down So Ioa't.ed cat ion:

position
.'".
"
.S
z::
Soluated anian
Water d ipo Ie in
the up position ELECTRICAL DOUBLE LAYER
110DEL
DlP OHP
b'" I
la4 I
I _I L '1'
,1 I
~ I
---r--- 'I'd
I I
I I .,
I I '1'5
Helmholtz-Perrin Gouy-Chapl'l1an d iffuse Stern doub le la!Jl'r

c:onpact layer layer
(a) (b) (e)
Charge reuersa 1
because of
adsorp-t ion af
~-:--::;; ~ counterions
(a)
Fig. 2.- The electrical double layer model

The structure 01 water is also a factor. Most theories 01' the rnechanism of
flotation stick to the electrical double layer theory as an adequate description
of the influence of the water phase on collector adhesion to a mineral
surface. Possibly it is an over-simplification, as the structure of water is
controversial. Linus Pauling suggests that for pure water, at ambient and
lower temperatures there is considerable regularity to the structure; between
O and 4 he suggests that the crystalline regularity 01' water molecules
corresponds to a mixture of ice 1 crystals (Fig. 3), which has a density of 0.92,
mixed with ice Il, a structure similar to quartz, which has a density of 1.17. As
the temperature approaches the maximum density of water, at 4C, the
relative proportion 01' ice II rises. Above 4C, water becomes more
amorphous, the crystalline regularity becoming more and more discontinuous
as the temperature increases.
Fig. 3.- The structure of ice, showing the orientation of the hydrogen
bonding, adapted from Pauling (1967).
Lone e lect.ron
pairs
Fig. 4.- Localised-bond structure for H 20 with sp3 orbitals for oxygen
Fig. 4 shows the spatial arrangement of the orbitals in water, and the
explanation for the abnormally high dipole that this molecule exhibits. In
mineral pulps, the aqueous phase is alkaline and contains common saline
salts in solution, usually including calcium, magnesium, silicates, sodium
chloride and excess hydroxyl groups. Schematically, the orientation of the
water molecules in the hydrated anions and cations will be as shown in Fig. 5.
Laskowski proposes that the hydration of a mineral surface in an impure
aqueous solution will consist of multilayers of water molecules ordered into
ice type crystals with stray water molecules and small hydrated ions captive
within the clathrate structure. Presumably the value of the measured
electrical double layer charges reflects the energy necessary to penetrate this
hydration layer.
The topology of the layer may well be even more complicated than
Laskowski's model, as concentrated transition metal salt solutions are known
to have peculiar solubility characteristics . which reflect a crystal-like
regularity. For example, a 60.2 weight per cent zinc chloride solution, a
concentration which corresponds to 5 molec~les of water per molecule of
zinc chloride, becomes a perfect solvent for polyacrylonitrile polymers, while
a solution with less than a 45% zinc chloride (a nine to one mole ratio) is a
non-solvent (see R.D. Crozier, et al, U.S. Patent 3,346,685, 1967). X-ray
diffraction data on the concentrated solutions gives diffuse crystalline
24 FLTATIN - Theory, Reagents and Ore Testing
scattering, though the liquid ordering involved has not been determined,
The electrical behaviour of metal sulphide minerals is more complicated;

rnost are semi-conductors, so their electro-potential properties can vary
considerably from point to point on the particle surface as a function of the
cation, the crystal structure, and the homogeneity of the ore. For these
minerals the flotation process is complex and very difficult to predict from
simple laboratory experiments.
sodium cation chlorine anion
Fig. 5.- Cation and anion hydration showing water orientation
Jan Leja (1982) provides an extensive review of the mechanism of adsorption

of fatty acids, sulphonates and amines on inorganic and metallic ores, and
probably the most extensive, easily available reference on the adsorption
mechanisms. In addition, there is considerable recent literature on zeta
potential and other electrochemical sweep experiments, with interpretation
of the results in terms of flotation conditions. Though, as virtually all of the
data is based on pure, or at least selected mineral samples, the results of
laboratory measurements are at best indicative rather than quantitative, if
extrapolated to full scale equipment.
Chapter :3
SUlPHIDE MINERAL FlOTATION
Stereochemistry of metal xanthates and dthiophosphates
An essential factor in the evaluation of spectroscopic data of the compounds

formed by collectors on mineral surfaces is the stereo-chemistry of the
complexes that the transition metals can form with thiolates in solution.
These complexes are particularly important in the interpretation of the
existence of dithiolates on the mineral surface, as the complexes themselves
have sulphur-to-sulphur bonding, and many have active lone pairs.
Six coordination complexes
Iron, nickel, cobalt, chromium, Mn IlI , and Cu Il can form six-coordination

complexes with tris(bidentate ligands), with geometries as illustrated in Pig. 6
and 7. Structural parameters for xanthates and dithiophosphates are reported
by Kepert (1977). Por the ferric compound, the constants are b = 1.22 or
1.23 and G = 20.5 or 21.8, according to three different investigators.
The reaction of the formation of the aboye bidentate complex is:
But the reaction is known to continue, through self reduction, to produce

another, more massive, six-coordination complex and dixanthogen:
8[Fe I I I(EtX)3] -------~ 2[FeII(EtX)3]3FeIII"+ 3EtX-XEt

\ dixanthogen
A pseudo four-coordination dithiolate complex is formed by CuIl. In the case

of copper xanthate, it forms a puckered square planar equatorial girdle of
two xanthate molecules and either water or a bridging ligand in the axial
position. It is considered pseudo four-coordination because the axial bond is
- 25-
Xanthate
planar
Edge length S and

height h
e)
"Z
for [lHb identat 3 1
CompLex b
1.34 24.7
1.32 23.8
1.24 19.9
1.22 21.1
1.21 18.7
1.19 16.8
1.21 20.3
1.19 20.1
Fig. 6.- General stereochemistry for [M(bidentateh]X
much longer (3.5 A) and weaker than the sulphur bonds. The cupric
reactions are:
which react further, in solution:

[CUI(EtX)2J2cuII + EtX-XEt
dixanthogen
This complex will hydrate to form a bridging six-coordination cuprous

complexo The water molecules, though, can be substituted by copper ions,
For both the cupric and the cuprous bidentate complex, the geometry is
shown in Fig. 7.
Seven coordination complexes
Seven coordination complexes, in all cases of interest in flotation, involve

lone electron pairs in one of the coordination axis. These electrons can be
transferred to the mineral surface or can be used to form polymeric
complexes in the pulpo The most common geometry of the seven
coordination complexes are the pentagonal bipyramid and the capped
octahedron shown in Fig. 8.
Sulphide mineral flotation 27
)('antlha-t:e
e ..
/:.:. ;"'::::"':::" ~_-l/
<S>
s
II
r ce (xanthate ) 2] POLYMER Cu
According to Mumme and Winter (1971), as reported by Drew (1977),

bivalent lead forms a seven coordination complex with a lone electron pair
which can polymerise. The opinion that polymer formation can occur is
based on examination of samples where "close to the axial position
presumably occupied by a lone pair an equatorial S atom of another
molecule" was found. Drew describes the complex as a pentagonal
bipyramid. The five positions in the equatorial plane surrounding the metal
atom are at a shorter bond distance from the central metallic atom than the
two axial locations. Certainly the existence of a complex will be sufficient to
confuse surface infra-red spectra, and definitely it will show up as the lead
Pentagonal bipYl'a~id Capped octahedl'on

Fig. 8.- Pentagonal bipyramid and capped octahedron geometry
xanthate salt if leached from the concentrate by an organic solvent. Similar

structures, with lone pair bridging, are produced for [Sb(dedtchOone pair)],
[Bi(dedtc>J(lone pair], and [Te(ethyl xanthatej-flone pair)], where dedtc is
diethyl dithiocarbamate.
Dithiophosphates are more likely to produce a polymeric structure with lead,

where the bridging bonds between the sulphur atoms in two complexes are
considerably longer than the equatorial plane Pb ones (3.232 A, 3.175 A,
compared to 2.761 A, 2.772 A, see Fig. 9). The sketch of a trimer (Fig. 10)
shows that this complex wiIl have more available electron pairs to form bonds
with the mineral surface, which may explain the preference for
dithiophosphates as coIlectors for galena.
Note that for the trivalent metals the geometry of the complexes are capped
octahedrons. The evidence on polymer formation is tenuous, which probably
means that xanthates reacting with lead cations in solution will usually only
form dimers and trimers at the mineral surface. The complex wiIl have
available Ione electron pairs to transfer to the mineral particle.
s~~~;----
~ L-.~~
XAHTHATE
~~
~) _~ep'
~.,.
Sulphur ai:ol"ls .------...
[PbIICci:h!:>ll xAntnato)3 Clono pair)] DltlER
Fig. 9.-[Pb II(ethyl xanthate)3(lone pair)]
Trivalent arsenic, antimony, telurium, bismuth, and vanadium form a very

similar complex with xanthates and dithiophosphates, except that they have
seven coordination because there is a lone electron pair protruding out of
one of the faces in the axial position. Fig. 11 shows the capped octahedron
structure in a diagram where dirnensions are easier to visualise, and sorne
typical key dirnensions for arsenic and antirnony xanthates and a bisrnuth
dithiophosphate reported by Kepert.
lone pail:'
Seven Coordination COMPlex

units fOPMing polYMep
Fig. 10.- [Pb Il ( di-isopropylphosphorodithiolateh (lone pair)]
Dithiolate chemistryand surface compounds
Dixanthogens: before going on to analyse the Finkelstein and Poling (1977)

review of the nature of the compounds formed on the surface of minerals
when they react with collectors, it is important to c1arify the probable identity
of the dixanthogens mentioned in the older flotation literature. Sixty years
ago Cambron and Whitby (1930) were pointing out that reacting xanthates
with iodine or chlorine produced impure liquid dixanthogens, while a more
selective oxidiser, such as sodiurn tetrathionate, will produce solid
dixanthogens without the need to recrystallise. More recent reviews, among
them Reid (1962), have stressed that the melting point is variable, and an
indicator of purity. Table 6 contains representative melting point and boiling
point data on common dixanthogens, mostly from Reid, but rechecked
against unpublished data taken by Professor J.C. Vega de Kuyper at the
Catholic University of Santiago, Chile, in 197~/81.
5
StructuraL Parameters of Tris(bidentate) (Lone pair) CompLexes
CompLex M- a M- b aMa bMb aa ab
[AS(S2CNEt~)1] 2.35 2.85 90.1 106.0 54.8 67.2

[As(S2COEt 3 2.28 2.94 92.0 107.6 56.2 68.7
[Sb(S2COEt)3] 2.52 3.00 87.5 113.1 53.0 74.5
[Bi{CS2PCOPr)2}3] 2.70 2.87 91.2 99.9 55.6 62.1
Fig.11.- [M(bidentateh(lone pair)jX:!:
All the flotation literature surveyed, including recent compendia such as Leja
(1982), always refer to dixanthogens as oily liquids. This suggests that impure
products have been used in flotation experiments, because methyl, ethyl and
isopropyl dixanthogens are solids at room temperature, as is the
mercaptobenzothiazole dithiolate. The ethyl and mercaptobenzothiazole
dithiolates, which are readily available commercially, are yellow solids, in all
grades.
The 1983 Merck Index, under dixanthogen (compound 3402) lists as alternate
names thioperoxydicarbonic acid diethyl ester, or dithiobis[thioformic acid]
O,O-diethyl ester, and describes the product as yellow needles, mp 28-320 ,
onion like odour, solubility in alcohol 2g/100 mI, freely soluble in benzene,
ether, petr. ether, oils. Almost insoluble in water. Use: insecticide
formulations, herbicide, topical parasiticide. Under dithiobis[benzothiazole]
(monograph 3389) it also lists the dithiolate of mercaptobenzothiazole as
pale yellow needles, density 1.5, mp. 1800 , (minimum commercial
specification, mp 168) insoluble in water, use: rubber accelerator. Under

dixanthogen, the latest edition of Hawley's Condensed Chernical Dictionary
lists ethyl dixanthogen (CAS: 502-55-6) uneler the trade name "Sulfasan", as
yellow crystals, mp. 30, insoluble in water, and used as a herbicide,
insecticide, parasiticide, It also lists the mercaptobenzothiazole dithiolate
with CAS: 120-78-5, yellow salid, properties, d= 1.54, mp. 168, use: rubber
accelerator.
TabLe 6.- PROPERTIES OF DIXANTHOGENS AND XANTHIC ANHYDRIDES
s S
Xanthic anhydrides = R-O-C-S-C-O-R
MethyL ye LLow so Lid 54-55

EthyL yeLLow soLid 52-55
IsopropyL yeL Low soLid 54-55
PropyL brown oiL 557 1367
ter-ButyL brown oiL 7
IsobutyL yeLLow oiL 7 d=1.126
AmyL yeL Low oi L 7
HexyL yeLLow oi L 7 ND=1.541
(MBT)2 unknown
S S
Dixanthogen R-O-C-S-S-C-O-R
MethyL R=CHrO yeLLow soLid 23-23.5 907

EthyL R=CH3CH2-0 yeLLow soLid 32 1077 d=1.2604
IsopropyL R=(CH3)2CH2-0 yeLLow soLid 57.5-58.5
PropyL R= CH3(CH 2)2- 0 yeLLow oi L 117
ter-ButyL R= (CH3)3C-O yeLLow soLid 58
IsobutyL R=(CH3)2CH2CH2-0 yeLLow oi L 165 d=1.0S
AmyL
HexyL yeLLow oiL ND=1.5569
(MBT)2 yeLLow soLid 180
A caveat in reviewing very old work on elittiolates is that in pre-WW 11

Iiterature there is a confusion between eliethyl elixanthogen and Minerec A.
Por example, Taggart, in his Hanelbook of Mineral Dressing (1945), says on
page 12-11: "Minerec is principally dixanthogen, formed from treating
xanthate with ethyl chlorocarbonate, which is a strong oxieliser"; but Minerec
A is not a dixanthogen, it is an impure xanthogen formateo Taggart's error

arises from reading a rather deceptive U.S. patent of AH.Fischer's (1928),
that was a consequence of a patent priority problem between Guggenheim
Brothers and du Pont, Guggenheim Brothers had hired Dr. AH. Fischer to
develop an improved collector for their El Teniente mine. Dr. W. A
Douglass (1927), working for du Pont, was simultaneously carrying out a
general programme on testing xanthate derivatives as improved col1ectors.
He obtained a clear priority on a xanthogen formate patent in the U.S.A, so
the only U.S. patent issued to Guggenheim Bros. did not cover xanthogen
formates, ergo the misinformation of Taggart and others. Minerec then
licensed the du Pont patent, and later was the only commercial producer of
xanthogen formates.
Physical or chemical adsorption of a col1ector to a solid surface involves an

interaction of e1ectrical forces; each type of bonding involves a reduction in
free energy of the system, and the production of a definite amount of heal.
The difference is that in physical adsorption the bond is amorphous, while in
chemical adsorption there is a directional bonding involved. Metal1urgical1y,
the differences between the two bonds is that, in physical adsorption the bond
is weak and easily reversible, so a reduction in the concentration of the
absorbent in the liquid phase will desorb the species from the surface, while
chemical adsorption is irreversible and normally localised on active sites of
the mineral surface.
Experiments performed over half a century ago demonstrated that, when

thiol collectors are employed to float sulphide minerals, the collection
mechanism involves the transfer of electrons to the mineral particle and the
formation of dithiolates. What has remained controversial has been whether
dithiolates are always involved in the collection mechanism or whether in
sorne cases the metal xanthate is formed and deposited on the mineral
surface. Based on the premise that the active collector is a metal xanthate, a
number of metallurgists have suggested that the effectiveness of a particular
xanthate for the collection of a specific mineral could be deduced from the
solubility of the metal xanthate.
The role of alkyl dithiolates in flotation systems, and the type of surface
bonding involved, was extensively reviewed by Finke1stein and Poling (1977).
The relevant dithiolates they studied are shown in Table 7.
They devoted much effort to evaluating the considerable volume of publishecl

information on the flotation of galena with xanthates. Among the more
bizarre theories put forward that they evaluated was that the interaction of a
xanthate with galena resultecl in the formation of elemental sulphur on the
Table 7. - DITHIOLATE REAGENTS: N0t1ENCLATURE AND ABBREVIATIOI~S
DithioLate Corresponding Thiol
Name FormuLa Abrv Name Abrv
DiaLkyL dixanthogen ROCCS)SSCS)COR )(2 ALkyL xanthate )(
Bis CdiaLkoxyphosphono-
thioyL) disuLphide CRO)2PCS)SSCS)PCOR)2 CDTP)2 DiaLkyL dithio- DTP
phosphate
2,2' Dithiobis
benzothiazoLe s S CMBT)2 Mercaptobenzo- MBT
\ / thiazoLe
CSSC
/ \
N N
mineral surface, and that it was the natural floatability of the sulphur that
caused the flotation of galena. More important was the controversy that, in
all cases, for a xanthate to act as a collector, it must have reacted to form
dixanthogen. The analysis is best summarised in their own words:
"As Granville, Finkelstein and Allison (1972) noted, the case

for dixanthogen as the essential hydrophobic entity rested
primarily on the electrochemical evidence, in essence, that a
bubble only adheres strongly to galena when the potential of
the surface is hi;h enough to allow dixanthogen to be
formed. The same is true at the present time. .... Woods
(1977) identified two stages in the reaction between the
xanthate and galena. At potentials between -0.2 V and + 0.2
V on the hydrogen scale, a reaction that he identified as
chemisorption took place. At potentials more anodic than
0.2 V, a reaction took place that increased in rate steadily as
the potential was raised, until it was limited by the
accumulation of producto This reaction is the formation of
dixanthogen....."
After the evaluation of considerable contradictory data on the collector

composition found on galena surfaces, including infra-red analysis of galena
treated with dixanthogens, they concluded thatxhe active species on galena is
the lead xanthate. The work they reviewed on dixanthogen adsorption
showed that galena reacts as follows:
PbS + )(2 ~~ PbX2 + So

This explains why sorne 01' the early research suggested that elemental
sulphur could be the active collector on galena al'ter exposure to xanthates,
Finkelstein and Poling also reviewed the available literature on surface

reactions 01' dithiophosphates and mercaptobenzothiazole with differenr
minerals. They noted that dithiophosphates are weaker collectors than the
corresponding xanthates, and that they are less readily oxidisable, as shown
by their dithiolate./thiol reduction potentials (Table 8), and that their lead
complexes are more soluble than the corresponding lead xanthates. They
also concluded that the active coUector on the galena surface is the lead
thiolate, as is the case with the data they could examine for
mercaptobenzothiazole. The reduction potential data for the xanthates and
dithiophosphates is 01' interest because, at least theoreticalIy, it can be
compared with the rest potential 01' the mineral in water to see if the reaction
wil1 occur. The general conclusions 01' Finkelstein and Poling on the coUector
species present on the mineral surface at flotation is summarised in Table 9.
TabLe 8.- REDUCTION POTENTIALS FOR DITHIOLATE/THIOL COUPLES
HomoLogue Xanthate DTP HomoLogue Xanthate DTP

methyL -0.004 0.316 isobutyL -0.127 0.158
ethyL -0.060 0.255 amyL -0.159 0.050
propyL -0.091 0.187 isoamyL n/d 0.086
isopropyL -0.096 0.196 hexyL n/d -0.015
butyL -0.127 0.122
Finkelstein and Poling's most interesting conclusion in their overal1 survey 01'
published data on infra-red surface analysis (which predates Woods review)
is that the active species on galena is alead xanthate. As there is considerable
evidence that the lead atom in galena is labile, there should be Pb+ + present
in solution in flotation pulps, so the lead xanthate identified on the mineral
surface is likely to be the seven coordination complexes depicted in Fig. 9.
These complexes have lone pairs which can be donated to the mineral
surface.
Woods' 1977 analysis, based on voltamograms for galena, and the concept
that significant contact angles are essential to flotation, is as foUows:
"The more important question is whether the presence 01'

dixanthogen is essential for flotation to occur. It is apparent
that flotation takes place with a number 01' mineral-coUector
systems without disulfide formation. A problem to be
considered for the interaction of xanthates with galena les

in the faet that bulk lead xanthate is hydrophilic and
multilayers of this compound on galena itself give rise to a
low contact angle.
The electrochemical measurements indicate that the initial

chemisorbed xanthate layer is hydrophobic and that
flotation commences at potentials where this species is
formed with only small quantities of dixanthogens being
present. The chemisorbed layer on galena, therefore, has a
different effect on the superficial tension from the lead
xanthate and could be the important species in flotation of
galena; this does not exclude the possibility that
dixanthogen also plays an important supporting role."
TabLe 9.- PRODUCTS EXTRACTED FROM THIOLATED SURFACES

(FinkeLstein and PoLing (1977) )
MineraL Product formed on mineraL surface by coLLector
Name FormuLa EthyL PropyL ButyL AmyL HexyL Dithiophos MBT
Orpiment AS2S3 +
Stibnite
ReaLgar
AS2S3
AsS
@
@
@
e
@
@
@
@
+ @
@
e
@
Cinnabar HgS @ @ @ @ +
SphaLerite ZnS @ @ @ @ + @ @
Antimonite AS2S2 @ + + + + @ @
ALabandite MnS @ + + + +
Bornite CuSFeS4 @ + + + + Cu(DTP)2 Cu(MBP)2
ChaLcocite cU2S @ + + + + Cu(DTP)2 Cu(MBP)2
Galena PbS + + + + Pb(DTP)2 @
Pyrrhotite F~S8 I!l I!l I!l I!l I!l @ @
Arsenopyrite FeAsS I!l I!l I!l I!l I!l @ e

Pyrite FeS2 I!l I!l I!l I!l I!l (DTP)2 (MBT)2
+Fe(MBT)
ChaLcopyrite CuFeS2 I!l I!l I!l I!l I!l Cu(DTP)2 CU(MBP)2
CoveLL ite CuS I!l I!l ++ I!l + + I!l + + I!l++ Cu(OTP)2 Cu(MBP)2
Mo l i bdeni te MoS2 I!l +? I!l + ? I!l + ? I!l + ? I!l Mo(DTP)x Mo(MBP)x
\
+ = metaL xanthate I!l = dixanthogen @ = no positive identification

36 FLOTA'flON - Theory, Reagents and Ore Testing
Thol colector bondng with metallic sulphde minerals
To evaluate the validity of f1otation theories, an important factor is to define

the type of bonding that exists between the surface of a mineral and the thiol;
i.e. is this a sulphur bond with the metal? The evidence is that the collector-
mineral bond is a sulphur-to-sulphur bond and not a sulphur-to-metal bond,
as is generally assumed. The following quote from Ronald Woods (1977)
provides a corroborating opinion on this very controversial subject:
"It has generally been considered that chemisorbed xanthate

will be attached to a metal atom in the sulphide surface.
However, it has been pointed out by Winter (G. Winter,
(1975)"Xanthates of Sulfur: Their Possible Role in
Flotation", Inorganic and Nuclear Chemistry Letters, Vol.
11, pp. 113-118) that xanthates can form bonds with
sulphur, compounds of the type ROCSZ-S-S2COR being
quite stable. Also, such compounds cannot readily be
distinguished from dixanthogens by spectroscopic analysis
and could well be formed at sulphide surfaces."
Strictly speaking, from the point of view of differentiating between flotation

mechanisms, it is more important to demonstrate that the chemisorption
bond between the collector and the mineral surface is not a metal-sulphur
bond, rather than that it is a sulphur-sulphur bond. For example, Clark
(1974) points out that chemisorption on the surface of either an n or p type
semiconductor results in a bond with hybrid covalent and ionic characteristics
which makes it non-specific to a particular atom. The electron availability
from the double bonded sulphur atom of the thiono group depends on the
following equilibrium:
e 00
S.. . h S:
'e v
I t-~---I>
:0:
'R
As the species is the anion of xanthic acid, it is obvious that the single bonded
sulphur can provide an electron for an ionic bond. A very similar picture can
be given for the dithiophosphates and the > mercaptans, while the
thionocarbamates, xanthogen formates and xanthic esters exhibit resonance
in their sulphur-to-carbon bonding, as well as a similar resonance in the
nitrogen or oxygen atom.
The authority for conc1uding that electrons donated by a collector to a

sulphide mineral form a sulphur to sulphur bond is Linus Pauling, In his "The
Nature of the Chemical Bond", pages 442/448, he says:
"Many sulfide minerals have structures closely related to

those of sphalerite and wurtzite. Chalcopyrite, CuFeSb is
an example (Fig. 12). Its structure is a tetragonal
superstructure of sphalerite, with the copper and iron atoms
in the zinc positions of sphalerite. Energite, CU3AsS4' has a
structure that is a superstructure of the wurtzite
arrangement, The sulfur atoms are in the same positions as
in wurtzite, and the atoms of copper and arsenic replace
those of zinc in an ordered way, so as to give discrete AsS4
groups (Fig. 13). The observed As-S bond length, 2.22 A,
agrees exactly with the calculated value for a single bond,
2.22 A (from the covalent radii, with the correction for
electronegativity difference). The Cu-S bond length, 2.32 A,
corresponds to a bond number about 0.7 (the appropriate
single-bond radius of copper is 1.23 A). Approximately the
same Cu-S bond length is found in other copper sulfide
minerals. The copper-sulfur bonds have only a small
amount of ionic character, and the conclusion may be drawn
that the electric charge of the copper atom is negative,
probably close to -1."
Pauling makes a similar analysis for a number of other sulphide minerals. In

all cases he reaches the same conclusions, that in transition metal sulphide
crystals the metal is negatively charged, and the sulphur atom is positive and
an electron acceptor. In a 1970 article, with the title "Crystallography and
Chemical bonding of Sulfide Minerals" (Mineral Soco Amer. Spec. Paper J,
125-131), he put forward a more controversial concept: bonding
characteristics of "argononic or trans argononic sulfurs", i.e. sulphur atoms
with the same number of electron pairs as argon, and trans argononic where
the atom has more electron pairs than argon. In his conclusions in this paper
he says:
"The foregoing discussion of the structures of sorne sulfide

minerals, especially the observed bond lengths and bond
angles, leads to the conclusion that \he sulfur atom may
have either an argononic or a transargononic structure with
formal charge O, + 1, or (argononic only) +2, sometimes
with resonance between two structures."
Sorne specific calculated charges given in this paper are: for molybdenite,
average charge for Mo, -0.74; for S, +0.37; for proustite (Ag 3AsS3 ) , formal
charge for Ag -1; for sphalerite and wurzite, average charge for zinc -0.67;
and for sulphur, +0.67; while for galena, which has a sodium chloride
strucrure, the bonds are resonating, and the average charges are + 0.43 for S,
and -0.43 for Pb. These are listed in Table 10, and compared to the mineral's
semi-conductor properties and flotation critical pH.
The general conclusion is that, for the majority of the sulphide minerals, the
negative charge on the metal forces the collector- mineral electron transfer to
be from the collector su1phur atom to the mineral su1phur atom.
TabLe 10.- MINERAL ELECTROCHEMICAL PROPERTIES
MineraL Surface compound CriticaL pH **

Name FormuLa EthyL AmyL DEDTP Type EthyL AmyL DEDTP
SphaLerite ZnS ? MX ? P none 5.5 none
pyrrhotite Fe7s8 X2 X2 ? P 6.0
Arsenopyrite FeAsS X2 X2 ? N 8.4
GaLena PbS MX MX MX P 10.4 12.1 6.2
Pyrite FeS2 X2 X2 X2 N 10.5 12.3 3.5
ChaLcopyrite CuFeS2 X2 X2 MX N 11.8 >13 9.4
CoveL Lite Cus X~. X2+ MX MX 13.2
Bornite CU5FeS4 MX MX P 13.8
ChaLcocite CU2S ? MX MX >14
MoLibdenite MoS 2 X2+? X2+? MX
MineraL Charge on Atom* Semiconductor
Name Cation SuLphur Type Band-Gap eV
SphaLerite -0.67 +0.67 P 3.8

Pyrrhotite P
Arsenopyrite N 0.16
GaLena -0.43 +0.43 P 0.38
Pyrite - 1 + 1 N 1.1
ChaLcopyrite - 1 + 1 N 0.33
CoveLLite - 1 + 1
Bornite P 0.82
ChaLcocite
MoL ibdenite -0.74 +0.37
* = Charge on atom from PauLing
DEDTP = Diethyl dithiophosphate
** = critical pH for fLotation aboye which mineral
depressed (Sutherland and Wark)
o-:
~:d = Cu
~1\!1l = As
E/'IARGITE
= Cu
e= Fe 0=8
CHALCOP'iRITE
Fig. 12.- Tetragonal crystal chalcopyrite and orthorhombic enargite
= Zn
IJURTZITE 8PHALERITE
Fig.13.- Wurtzite and sphalerite, from Pauling
The exact points to which col1ectors will attach themselves on mineral

partic1es wil1 depend on the micro semi-conductor topology of the surface.
As will be recalled, the n-type semiconductor c~nducts through e1ectrons in
the conductance band, and thus is an electron donor, while the p-type
conducts through the movement of vacancies in the valency band and thus is
an e1ectron acceptor. Ideal1y for rapid adsorption of col1ector molecules, the
mineral partic1es should have n and p regions distributed on the surface, the
p-type to provide a point of adherence for the collector, the n-type for the
transfer of electrons to oxygen dissolved in the pulp to thus maintain
electroneutrality for each mineral particle. The bulk semi-conductor
properties (Baleshta et al, 1970) of the more important sulphide minerals are
Iisted in Table 110
TabLe 110- BAND-GAP ENERGIES OF SULPHIDE MINERAL SEMICONDUCTORS
MineraL Type FormuLa Band-Gap Energy

eV
Arsenopyrite N FeAsS 0.16
Bornite P cU5FeS4 0.82
ChaLcopyrite N CuFeS2 0.33
Digenite P cU1. SS
GaLena P PbS 0.3S
Marcasite P FeS2 0046
Pyr i te N FeS2 1.1
Pyrrhotite P Fe7SS
Zinc suLph. P ZnS 3.S
Germanium Ge 0.75
Sil i con Si 1.2
The nature of the bubble-particle bond
We have looked at how collectors adhere to mineral partic1es but not at the
nature of the collector-bubble bond, or for that matter the overall mineral
partic1e-bubble bond. We know that mineral capture by bubbles is a
reluctant phenomena, In all cases studied, an induction delay in gas-solid
contact has been detected, One can visualise the time delay as the bubble
girding its loins and battering its way past the defending forces to finally
reach the mineral surface. In this case the defending forces resisting the
bubble approach are the water of hydration around the particles and the
oriented water layer at the liquid/gas interfaces.
The theory and mathematics of these surface tension phenomena and their
effect on the dynamics of flotation are complexo KIassen and Mokrousov
(1963) extensively reviewed the classic thermodynamic information available
prior to 1960. Laskowski (1974) published "Particle-bubble attachment in
flotation", a thorough up-date, His synopsis, which agrees with KIassen and
Mokrousov, says:
"The hydrophobicity, defined by the contact angle, is not

enough to describe the flotation properties of a given solid,
since the conditions connected with the speed of
displacement of the liquid layer between the solid particle
and the bubble must be also fulfilled."
The disqualification of the value of the contact angle in predicting flotation

performance does not go Iar enough, though Laskowski puts his finger on the
key factor that determines the flotation kinetics in bubble contact; i.e., that
the rate controlling mechanism is the thickness and tenacity of the liquid
layer surrounding the bubble and the particle, which must be parted before
there is bubble contact and adhesion.
A very pertinent analysis of the alternative methods of obtaining bubble-

particle adhesion is contained in Glembotskii's (1963) 1961 text, Flotation.
Here the magnitude of the energy required for a bubble to attach to a
mineral surface is compared to the energy necessary to generate a bubble in
situ on the surface. His conclusion is that, particularly at large contact
angles, the energy required to generare a new bubble on the surface is
considerably lower than the energy expended in attaching a bubble to a
mineral particle, postulating therefore that collection of mineral particles will
preferably occur through micro-bubble formation on the mineral (nascent
bubble flotation) rather than attachment of existing bubbles to a suspended
particle (coursing bubble attachment). This is an obvious conclusion when
one considers that the energy of adhesion is proportional to the surface
tension and the area of liquid surface that must be displaced from the solid,
and that the contact area of a micro-bubble is much smaller than that
required to attach a preformed bubble. This relation also explains why the
preferred frothers used in flotation barely affect the liquid surface tension, as
otherwise, if they increased surface tension significantIy, particle capture
would become more difficult.
The forces holding a collector to the surface of a sulphide mineral are due to
bonding through the transfer of lone electron pairs as well as ionic bonding
by van der Waals forces. This bonding is obvious, but due to the use of
contact angles to evaluate whether bubble attachment occurs at given
conditions, an implicit subconscious assumption is present that the alkyl
group of col1ectors is hydrophobic and therefore that a bubble attaches itself
to this hydrocarbon layer through surface tension forces. In fact, the use of
contact angle measurements to evaluate positive floatability is dangerous, as
there is incontrovertible evidence that successful sulphide mineral recovery
by flotation requires that the amount of collector present be considerably less
than a monolayer of collector, while stable contact angles are measured when
there are multilayers on the sample surface. \
It would appear that to obtain a measurable contact angle on a sample, an

area of the sample proportional in size to the bubble used to measure the
angle must be covered with collector (or be hydrophobic), while, if a collector
covers this much area of a sulphide mineral particle in a flotation cell, this
results in depression rather than flotation. Therefore, if one detccts a contact

angle in a flotation system, the dosage of the f1otation reagent is too great,
and corresponds to the conditions at which over-oiling is present rather than
those Ior optimum flotation.
Contact angle measurements are successfully used in industriallabs to screen

depressants for any type of mineral. This may be significant in evaluating the
mechanism of depression.
To determine what force, other than surface tension, accounts for collector-
frother adhesion, one can analyse the bonding of each link in the chain
joining the bubble to the mineral particle. First, what is the collector to
mineral bond: it is obvious that as the collector has transferred electrons to
the sulphide mineral, there is chemical adsorption, and a covalent or ionic
bond between the collector and the mineral.
Second, what is the bond between the exposed side of the collector and the
bubble: here the same reasoning as that used for the validity of the contact
angle indicates that the surface tension forces that come into play with
hydrophobic surfaces cannot be major contributors to bubble adhesion; but
there is independent evidence that, in sulphide mineral systems, frothers
adsorb onto collector coated minerals, so chemical adsorption between the
frother and the collector is probably present. Supporting this is the fact that,
to adsorb collectors, the sulphide mineral particle must maintain
e1ectroneutrality by oxidation of oxygen dissolved in the pulp; therefore, there
must be a charge on the collector adhered to the mineral, and this e1ectron
deficit will be available to form a chemical bond between the frother and the
collector.
Third, what is the bond between the collector-frother complex and the
bubble: a simplified mechanism of nascent bubble formation on the mineral
surface posits that this will occur at electrical discontinuities on the sulphide
particle surface; this charge anisotropy will attract surface-active species in
solution, which will then stabilise the growing bubble through normal
surfactant action. In effect, an e1ectron bond is formed at the same time as
the bubble, and bubble stability is produced by independent frother
molecules, with either the same or different chemical structures than the
species forming the collector-frother complexo
Summarising, to obtain bubble-particle adhesion, a sulphide mineral particle

must first become coated with a partial layer of collector molecules. The
coated particle will then abstract frother molecules from solution. These will
intermingle with the collector molecules aIready on the mineral surface; and
then the nascent bubble (generatcd by hyelroclynamic mechanisms) is

nucleatcd by organic discontinuities on the mineral surface. This description
has been limited to sulphide minerals because it only describes the flotation
mechanism when starvation quantities of reagents are employed. Manyoxide
and salt systems require orders of magnitude greater quantities of reagents,
due to mechanistie differences in the flotation process,
~
~c'" 41'd /.l4llill water \
c:::Jo
""!E':l
frother
~+-'"0 lec u les
~ \ o?? \ oC;:::;> rP II t
Step 1
bubble approaeh ~ UJ!!lU ~ ~:~I.~I:~
~'ol(\
/~~~0~{; Mineral particle f\.\ 0(/
J
1:\ solid phase /)",oJ.i
bubble
air phase
Step 2
bubble adhesion
~
~
I
II
co 1 lector frother
Fig. 14.- Leja and Sehulman (1954) interpenetration theory
Leja and Schulman (1954) studied the interaction between colIectors and
frathers in 1954, and demonstrated that a colIeetor eoated mineral will
remove frother from a solution. Their experment was quite simple. They
took a sample of ground galena, plaeed it in a xanthate solution, drained off
the solution and then added water containing a known amount of a flotation
frather. Analysis of the liquid phase demonstrated a reduetion in the frother
content and therefore the formation of sorne form of colIeetor-frother
adduct, The experiment was repeated with a gamut of collector-frother

species, and gave the same results. As there were no bubbles present, the
attraction between the collector and the frother molecules could not be due
to surface tension, In this paper they also proposed their interpenetration
theory; Fig. 14 reproduces their original diagrams. They assume that the
attractive forces involved are ionic.
Frother-collector surface complexes
Fig. 15, taken from Crozier and Klimpel (1988) depicts a xanthate-alcoholic
frother interaction complex on the surface of a sulphide mineral. It shows the
chemisorption of the demised sulphur atoms and the physical adsorption of
the double bonded sulphur atom. Note that these double bonded sulphur
atoms are on the same plane as the single bonded atoms, The frother
molecules within the collector matrix are oriented so that the hydroxyl groups
are pointing away from the mineral surface, so, when bubbles adhere, these
groups are located at the interface with the air water boundary to bond the
bubble to the mineral surface.
priJTIary 0]("
If secondary Air-water interface
(J ,/
\ frother BUBBLE /'
"OH-"O / /'
Hydrogen oer-t-ted
H, ) .,/ excep't; for OH
.......0 H
H H .....
,....P~ H frot.her posit.ion
assulI"es H bond
with collect.ol' Oxygen
collect.o
ti\.
sulphur
1.84 A Mineral
Mineral surface
surface
Fig. 15.- Frother-collector interaction complex

The bond between the frother alkyl group and the collector could be a
hydrogen bond with the oxygen atom attached to the CSz group, See page
36 for the possible resonance structures for xanthate anions which would
accommodate this sort of bonding. Note that the collector's single bonded
sulphur atoms transfer electrons to form a bond with the mineral's sulphur
atoms, not the metal atoms. Though the illustration assumes adsorption of
the double bonded sulphur atom, it can chemisorb to the mineral sulphur
atom because the double bond can resonate between the S =S and S = O
positions in the collector molecule. This resonance can also provide the
electrons for the frother-collector hydrogen bond. The atomic radii employed
in the ilIustration are based on Pauling (1967).
There is evidence that the formation of collector-frother complexes is

catalysed by minerals. The catalytic effect of the mineral in a flotation system
is frequently overlooked when analysing the possible reactions occurring on
the concentrate. The author once published the comment that
thionocarbamates and xanthogen formates reacted vigorously, and received a
letter from the chemist recognised as a world authority on thionocarbamates
saying that he had tried the reaction and nothing happened, at any
temperature, over a number of days of mixing. He was quite right, of course;
in a glass beaker there was no reaction, but in a metal container there is a
slow reaction, and in the presence of mineral there is a very vigorous
reaction. At that time the author's company was marketing such a mixture as
a collector. It was delivered to the sole client in tank cars, and the change in
composition with time in the tank car had been monitored for a number of
years trying to see if the active species that provided the unique, mineral
specific, properties of the collector could be identified.
Sometimes industrial data can elucidate these types of theoretical probIems,

even when they are difficuIt to test experimentally in a Iaboratory with very
sophisticated instruments. An interesting case was experienced over a
number of years with the Anaconda, Butte concentrator. Because of changes
forced on the reagent supplied, by competitive factors and periodic sharp
mineralogical changes, we were abIe to deduce that for dialkyI
thionocarbamates, the alkyI group Iocation on either the e =S side or the N
side did not affect metallurgicaI properties.
This was discovered because pressure for prompt suppIy forced the use of a
temporary synthesis route, which on being \modified into an economic
industrial process reversed the Iocation of the alkyI groups. Later bench
flotation showed that, for that particular ore, a methyI group on the nitrogen
side was as effective a collector as the traditional ethyI alkyI thionocarbamate,
and much cheaper 1.0 manufacture. In the manufacture we ran into the
problern that imrniscible collector and the by-product brine densities were
such as 1.0 malee separation impossible, so the collector was extracted with a
lower density alcohol - MIBC. The mixture was tested as a collector in the
concentrator and was found 1.0 he as good or better than the pure
thionocarbamate, Interestingly, Butte was employing MIBC as its frother at
that time, al. a dosage of about half that of the collector. It was found that,
even though the MIBC contained in the collector mix was as much as the
MIBC fed independently as a frother, it did not reduce the amount of frother
required for goodoperation. The gross amount of collector added also
remained unchanged, so in effect, concentrate and recovery grade was
unchanged at half the collector consumption and double the frother dose.
The revival of the Leja and Schulman collector-frother interpenetration

complex is the result of practical experience which has resulted in the greater
(and more carefully chosen) use of collector/frother mixtures fed directly to
the ball milI, plus the separate evaluation of frother structures to tailor these
to specific hydrodynamic froth problems that are independent of the
optimum coIlector from the mineral recovery standpoint. This choice of the
second component should certainly be taken into account when a
recommendation is made to dilute an insoluble oily collector with a
surfactant to "emulsify" it and improve dispersion.
Chapter 4
THIOl COllECrOR CHEMISTRY
Properties of xanthates
Xanthates are available commercially as solutions, powders or pellets.

Pellets are the most desirable, as there are fewer dusting problems (xanthate
dust is irritating and toxic) and better storage stability. All xanthates
decompose in the presence of moisture, giving off carbon disulphide as one
of the decomposition products, so that they should be treated as potentially
highly flammable. There is a considerable range in the purity of commercially
available xanthates. Quality control is somewhat of a problem, as accurate
analysis requires the availability of an experienced chemist because of the
rapidity of decomposition of solutions in contact with airo The commercial
availability of xanthates, and their trade designations, are shown in Table 18,
page67.
Potassium xanthates crystallise as anhydrous salts, while the sodium salts

have two molecules of water of crystallisation and are therefore less stable.
Thus, commercially available sodium xanthates are rarely over 85% active.
The stability of xanthate solutions cannot be predicted a priori, as it is a
complex function of solution concentration, storage temperature, excess or
deficiency of the alkali content (too much or too little will reduce xanthate
stability). At a pH of below 3, the half life of all xanthates is reduced to
minutes. With these caveats the decomposition of xanthates in solution, as a
function of temperature and concentration, shown in Table 15, can be used as
a practical guide to the stability of good quality xanthates.
The reaction employed in the production of xanthates is:
R-OH + NaOH + cS2 = R-O-CSSNa + H20
- 47-
As most sodium alkyl xanthates crystallise with two molecules of water, very
pure xanthates can only be obtained by recrystallising frorn a non-aqueous
solvent or by preparing the alcoholare with metallic sodium.
Recrystallisation is done by dissolving the xanthate in alcohol and salting it
out with ether, The anhydrous xanthate has a crystalline cream colour, and is
easily soluble in water. If kept cool and dry (in a dessicator) the salt is
relatively stable, but on exposure to moist air it will darken and acquire a
disagreeable odour.
In the presence of water the following secondary reactions can occur, either
during the synthesis reactions or the dissolution of the solid xanthate:
S
6R-0-C-SNa + 3H20 ---~ 6 R-OH + 2Na2CS3 + 3CS2 + Na2c~
6NaOH + 3CS2 ---~ 2Na2CS3 + Na2C03 + 3H 20
and in the presence of air the xanthate can be oxidised to dixanthates:
2R-0-CSSNa + ~2 + cO2 ---~ R-O-CSSSSC-O-R + Na2C03
The reaction will occur without COz if there is a ,mineral surface present to
catalyse it. With dissolved copper cations in the pulp the reaction is:
4EtXK + 2cu++ ---~ 2CuCEtX)2 + 4K+ ---~ cU2CEtX)2 + EtX-XEt
The oxidation ofaxanthate solution with iodine dissolved in an iodide

solution is quantitative and is employed as an analytical method, using starch
as the end point indicator.
2EtXK + K13 + 21- -----~ EtX-XEt + 3K13
Sodium peroxydisulphate (Na2S20S) reacts very similarly to iodine, as do

ferric salts.
The trithiocarbonates can continue hydrolysing, so commercial xanthates may

contain dixanthogens and free carbon disulphide, which make them highly
inflarnmable, plus having the fol1owing inorganic impurities:
Another method for the synthesis of xanthates is based on reacting ethers

with NaOH and carbon disulphide:
Et-O-Et + cS2 + NaOH ---~ 2Et-0-CSSNa + H20

Thiol Collector Chemistry 49
Tertiary aleohols are difficult to react with caustic to form alcoholares, bur do
react readily with sodium metal, thus this is a good laboratory route lo pure
xanthates:
Manufacture of alkali metal alkyl xanthates
The Iiterature on the methods of xanthate manufacture consists almost

exclusively of patents. AII other published information on xanthate
production appears to be descriptions of laboratory recipes and methods of
purification. The author is not aware of any real patent on an industrial
xanthate process, rather they aII appear to have been written for defensive
reasons and not to teach the arto Thus, descriptions of the processes in this
section are based on the author's limited manufacturing experience, rather
than on literature reviews.
An example is the sodium ethyl xanthate process described in the Kirk-

Othmer "Encyclopedia of Chemical Technology". DI. Guy Harris, at the time
a chemist at the Dow Chemical company, a major xanthate manufacturer,
wrote the xanthate article, but quite obviously could not reveal Dow
proprietary information. He chose to quote from a German process
described in the Department of Commerce, Office of Technical Service, P.E.
Rept. 74736:
"Por the manufacture of potassium ethyl xanthate, one

German plant used 400% excess alcohol and equimolar
quantities of carbon disulphide and 50% aqueous potassium
hydroxide. The reaction temperature was maintained at
40C by cooling. This step took a half hour. The product
was dried in a vacuum drier, and water was removed from
the recovered alcohol by distillation. The xanthate was
obtained in almost quantitative yield and assayed 95%."
From the author's experience, to make sodium ethyl xanthate using exactly
the same process, it would be described as folIows: in a steel kettle, equipped
with agitation and cooling, a 300% molar quantity of recovered alcohol (a
96% azeotrope of alcohol and unreacted CS2) , plus an equimolar amount of
fresh alcohol is charged. To this a slight exeess of 50% caustic is slowly
added. The rate of addition wiII depend on the cooling available, and should
be adjusted to such arate that the alcoholate reaction temperature does not
exceed 60C. The alcoholate is then cooled to about 35C (but not less than
25C) and a slight excess of CS2 added through the bottom of the reactor at
such a rate so that the cooling system maintains the reaction at, or below,
50 FLTATION - Theory, Reagents and Ore Testing
40C. The reaction time will depend on the geometry of the reactor and its
eooling eapacity. Usually this is designed for about a 40 minute overall
reaction time. Yields, in solution, are on the order of 95 - 98%, but at 40C,
in the reaetion mix, xanthate loss is about 0.25% per hour. It is very unlikely
that a commercial dried xanthate could be obtained by this method, because
of decomposition and degradation during the drying step. The alcohol can be
recovered by vacuum andjor steam distillation, and the solution used on site
as a raw material or as a col1ector. From the description of the process,
there are three critical steps involved: 1) the preparation of the alcoholate, 2)
the xanthate reaetion, and 3) the drying step. The sodium ethylate reaction is:
3EtOH + NaOH ~--------~ NaOEt2EtOH + H20
The sodium ethylate, or ethoxide, is a solid, very soluble in ethyl alcohol,

whieh in the presenee of water rapidly hydrolyses baek to alcohol plus eaustie.
This is the reason for the 400% exeess alcohol, as it is neeessary to keep the
equilibrium, in the aboye hydrolysis reaction, to the right. The yie1d can be
improved by using solid eaustie to limit the water present, but the alcohol
exeess must be inereased to keep the reaetion within bounds.
The seeond reaetion also has side reactions tlrat oeeur as a funetion of
temperature and water eontent. The main reaetion is:
s
NaOEt.2EtOH + cS2 ------~ EtO-C-SNa + 2EtOH
The side reaetions, in addition to the hydrolysis of the alcoholate, are:
6NaOH + 3cs2 ----~ 2Na2CS3 + Na2C03 + H20

EtOH + Na2CS3 -----~ EtO-CSSNa + NaHS
Higher reaction temperatures will favour both hydrolysis and the aboye
reaetions, redueing yields. The third, and always most eritieal step, is the
reeovery of the dry xanthate. In the presenee of exeess water, there is a
signifieant deeomposition of the xanthate to volatile components (CS2 and
alcohol) even at drying temperature as low as 50C. Based on a 1959
Russian artic1e, Harris, in the Encyclopedia artic1e, reports the data in Table
12 on the decomposition of dry sodium ethyl xanthate,
In the presenee of water the decomposition rate is mueh higher. The a1kali
metal alkoxides with alkyl group ehain lengths of three or more carbon
atoms, and partieularly those prepared from secondary or tertiary alcohols,
are considerably more unstable than ethyl alkoxide, as are the xanthates
thcmselves, As they are considerably less soluble in their own alcohol than
the ethyl xanthates (see Table 13), the excess alcohol process is only
commercially interesting for the ethy! xanthate, and then only if it can be
used as is. The economics of the solvent recovery and the drying step are
questionable, when compared to suspension processes.
TabLe 12.- DECOMPOSITION OF Na ETHYL XANTHATE ON DRYING AT 50C
Time of heating, hours Xanthate content, %
O 79.32%
40 74.38
64 65.66
88 65.06
TabLe 13.- PROPERTIES OF ALKYL XANTHATES

chain SoLubiLity g/100 9 soLvent
xanthate In water In its own
MoL. Length aLcohoL
Wt. angstr OC 30C OC 30C
K MethyL CH 30CSSK 146.3 7.1
K EthyL C2H50CSSK 160.3 8.4 80.1 123.8 5.6 12.5
K PropyL C3H70CSSK 174.3 9.7 43 58 1.9 8.9
K Isoprop C3H70CSSK 174.3 16.6 37.1
K ButyL C4H80CSSK 188.4 10.9 32.4 47.9
K Isobut. C4H80CSSK 188.4 10.7 47.7 1.6 6.2
K AmyL C5H110CSSK 212.4 12.2
K IsoamyL C5H110CSSK 212.4 - 28.4 53.3 2.0 6.5
Na EthyL C2H50CSSK 144.2 8.4 81.6 128.5 43.5 99.6

Na PropyL C3H70CSSK 158.2 9.7 17.6 43.3 10.2 22.5
Na lsoprop C3H70CSSK 158.2 12.1 37.9 19.0
Na ButyL C4H80CSSK 172.2 10.9 20.0 76.2 39.2
Na Isobut. C4H80CSSK 172.2 11.2 33.4 1.2 20.5
Na IsoamyL C5H110CSSK 186.3 24.7 43.5 10.9 15.5
Na AmyL C5H110CSSK
Xanthates are commonly manufactured in a b'ltch process, using solid caustic

and an inert diluent to reduce the effect of the water on the alcoholate
reaction. Solvent suspension simplifies agitation. The solvent chosen must
be unreactive, and a solvent for alcohol and CS2, but not for the water and
xanthate. Its volatility should be such as to reduce the drying energy and
temperature required. Most solvents commonly employed are petroleum
52 FLOTAnaN - Theory, Reagents and Ore Testing
ether, hexanc, benzene, toluenc or xylene.
In the suspension process a much larger reactor is charged with the recycled
solvent and a moderate excess of alcohol (contained in the recycled solvent),
and caustic (preferably beads, though flakes may be used) is dumped into the
reactor at a rate compatible with the eooling system available. The peak
temperature during the alcoholate reaetion should be kept to less than 60C.
Adjusting the reactor pressure, this can be aeeomplished by refluxing the
inert solvent. After the alcoholate reaetion is complete, an equimolar amount
of CS2 is added, again at arate dependent on the eooling system eapaeity.
The reaetion temperature depends on the alkyl ehain length, but norrna11y
will be under 50C. The produet is a slurry of solid xanthate containing,
hopefu11y, only a portion of the water of reaetion. After the reaction is
eompleted, the final xanthate slurry is filtered and the solid xanthate is
vacuum dried. The recovered solvent is also dried, by any convenient
method, and recycled to the reactor. At equilibrium eonditions, an alcohol
excess builds up in the recycle stream; thus, insufficient feed will reduce
yields and excess will keep increasing the excess alcohol, so the optimum for
each alkyl chain length must be determined by operating experience. The
required excess usually is in the 10 to 40% range. For a11 the xanthates, a
slight CS2 excess is required to maximise yields, usually in the 2 - 5% range.
Raw material speeifications have dryness as the top priority, and purity
second. Typical xanthate compositions obtained from the suspension
reaction system are shown in Table 14. The higher moisture content of the
sodium xanthates is due to their containing up to two molecules of water of
erystallisation. Overall reaction yields will be somewhat lower than the dry
basis xanthate purity, due to reactor losses and decomposition losses in the
drying step.
TabLe 14.- XANTHATE ASSAYS, SLURRY SUSPENSION PROCESS
ALkyL group Xanthate % Moisture % Dry basis %

K ethyL xanthate 88 - 92 O - 1% 89 - 93
Na ethyL xanthate 82 - 86 3 - 5.5 87 - 91
K sec amyL xanthate 75 - 85 4 - 6 80 - 88
K amyL xanthate 88 - 93 0-2 88 - 94
Na isopropyL xanthate 75 - 80 12 - 18 88 - 94
Na sec butyL xanthate 70 - 77 10 - 15 80 - 88
Only the Dow Chemical Company is believed to have operated a

continuous process for the manufacture of xanthates (Dow ceased co11ector
manufacture in 1980). Their products were consistently of a higher xanthate
purity and lower moisture content than competitive material, which suggests
that a1l other producers employ batch processes, The Dow continuous
process can only be surrnised, as there is no public information on the
process or even on whether it was non-batch. Their Na isopropyl xanthate
averaged 90 - 93% purity, compared to a norrn of 75 - 80% in the market,
and moisture content ranged from 2.5 to 5% compared with a norm of 12 -
18%. Por sodium sec-butyl xanthate, the figures are 80 - 85% purity,
compared to 70 - 77%, and moisture 2 - 5%, compared to 10 - 15%.
The dry xanthate powder is unacceptably irritating, so nearly all comrnercial

xanthates are pelletised before shipment.
Liquid xanthates: manufacturers consuming their xanthate direct1y, either in

a flotation process or as a raw material for another process, will by-pass the
drying step and wil! filter, then dissolve, the xanthate out of the inert solvent
slurry with water and use the xanthate solution directly. The capital savings is
considerable, as the vacuum drying equipment, plus pelletiser, can equal the
cost of the rest of the plant. Alternately, a third process can be employed in
which the xanthate is produced in a high intensity mixer vessel. In this case,
the reactor is similar to a bread dough mixer, except that the container is
equipped with a mechanically refrigerated cooling jacket. To manufacture a
given xanthate, the reactor is charged with t1ake caustic and a modest excess
of alcohol, 10 to 30%, depending on the alkyl group involved. After the
alcoholate reaction is completed, using the same maximum temperatures as
those stipulated in the suspension process, the paste is reacted with 1.02 to
1.05 molar ratio of CS2, with the carbon disulphide fed at a rate compatible
with the cooling available, but generally maintaining the temperature below
30C (considerably lower if very high purity xanthate is required). After the
reaction is completed, the paste is dissolved with water and the solution
employed immediately in the mill or further process. The quality of the
xanthate produced varies considerably as a function of the mixing intensity
and the efficiency of the cooling system. The use of mechanical refrigeration
in these units can be justified by the improved xanthate yields and the smaller
equipment volumes per unit of production. The drawback of the process is
that it is difficult to recover the excess alcohol fed, so if alcohol interferes
with the subsequent reaction, the process cannot be used. The advantage of
the process is that it can produce xanthates with alkyl groups of any chain
length.
\
Stability of xanthate solutions: industrial experience on the decomposition

rates of good quality xanthate solutions, in storage, are shown in Table 15. At
ambient temperature, these rates normally are below 1% per day. In a pulp,
the stability of the xanthate cannot be predicted. It is quite obviously a
54 FLOTATlN - Theory, Reagents and Ore Testing
function of thc pH of the environmcnt, but the decomposition is very

significantly affected by the catalytic effect of transition metal sulphides, as
well as, to a lesser degree, all the different constituents of the gangueo In
storage, the alkali content is crucial, as both an excess or a deficit reduce
xanthate stability. Most published stability data is unreliable because it has
been obtain in vitro. They must be employed with considerable caution when
inferring or analysing results in flotation testing.
Commercial xanthates are normally supplied as pellets, which can be

dissolved in water and then used in flotation mills in concentration of 5 to
25%. At normal concentrations, the solution viscosity is on the order of
1.3 cps.
Xanthate handling and safety: in powder form it is a strong irritant, so

operators should employ chemical safety clothes, gloves, goggles, and a dust
respirator, when handling. Its toxicity is moderate, with an LD sD of 500 to
2,000 mg/kg of body weight, dependent on the alkyl group. Xanthate
solutions, particularly concentrated, are asevere eye irritant, so eye baths
should be provided in the handling areas, as well as the use of the
appropriate safety clothes.
Xanthate powder has been known to ignite spontaneously, though this does
not seem to have been observed with pellets. Pelletised material must be
kept dryand cool to maintain long term stability in storage. Protection from
moisture is particularly important, so drums should be stored on their sides,
if they are in the open. Because decomposition products are extremely
flammable, storage precautions should be those usual for flammable
materials. If stored in bulk, bins should be nitrogen purged.
Xanthate solutions may be handled in steel or plastic pipes without problems,

but not in contact with copper, copper alloys, zinc or aluminium, etc., so care
should be taken that valves do not have copper alloy components. Xanthate
solutions must not be agitated with air sparges because oxidation can result
in hydrolysis to CSz, with the consequent tire hazard and ambient toxicity
problems. Combustion of solid xanthates is similar to sulphur, so the same
fire fighting methods apply.
When cleaning storage tanks that have contained a xanthate solution, toxic
gas precautions should be taken as there is always the possibility of
decomposition to CS2, so suitable equipment for dealing with a poisonous
explosive atmosphere must be used. The maximum worker exposure limit to
CS2 vapours in air is 20 ppm, average, during an 8 hour work day.
Thiol Collector Chernistry 55
Xanthate analysis: the most generally accepted quality control method 01'
analysis for xanthate pellets is the Dow Chemical Company procedure,
described in its "Mining Chemical" trade publication. Summarised, the fine1y
ground powder is dissolved in acetone (0.5 g in 35 ml of acetone), the
solution is filtered and neutralised with about 40 ml of 0.1 N standard HCL
solution and promptly back titrated with 0.1 N NaOH solution, using 3 to 5
drops of methyl red as the indicator. This method is preferable to titration of
an aqueous solution, as, in water, impurities such as thiocarbonates and
sulphides will report as xanthate. Numerous instrumental methods are
recommended in the literature for pulp, or dilute solution, analysis.
TabLe 15.- COMMERC1AL STAB1L1TY DATA ON XANTHATE SOLUT1ONS

xanthate Trade name soLution Specific gravity crystaL StabiLity
Cyan- conc. Hydrometer 60/60 Point Avg. Loss dai Ly %
Dow amid % reaL 15c 20c 25C oC 20c 30C 40C
K ethyL Z-3 303 10% 1.0481.046 1.045 -3 1 .1 2.7 4.6
25% 1.1221.120 1.118 -9 1.4 2.1 4.4
35% 1.174 1.172 1.170 -15
Na EthyL Z-4 325 10% 1.0481.0461.045 -3 1.1 2.7 4.6
25% 1.122 1.119 1.117 -11 0.7 2.0 4.3
35% 1.174 4.171 1.168 -12
Na iso- Z-11 343 10% 1.041 1.040 1.039 -4 0.3 0.6 1.8
propyL 25% 1.1031.101 1.098 1* 0.2 0.7 1.7
Na 1so- Z-14 317 10% 1.036 1.034 1.032 -3 0.6 1.8 3.6
butyL 25% 1.087 1.084 1.081 -9 0.6 1.6 4.0
35% 1.1181.1171.112 -13
Na sec- Z-12 301 10% 1.0391.0381.036 -3 0.2 0.6 1.2
butyL 25% 1.0991.096 1.094 -8 0.1 0.5 1.4
35% 1.1361.133 1.130 12*
K iso- 350 10% 1.0371.0361.035 -3 0.8 2.1 4.2

amyL 25% 1.0881.0861.084 9* 0.7 2.0 4.8
35% 1.1221.119 1.116 19*
Thiophosphates
O,O-Dialkyl dithiophosphate acids: are prepared by reacting phosphorous

pentasulphide (P 4SlO) with an alcohol in an inert media. The reaction is
dangerous because of the production of H 2S, a gas which is deadlier than
hydrogen cyanide at low concentration. The pentasulphide reaction has an
irregular incubation period, so start-up is usually he1ped by leaving a heel
from the previous synthesis in the reactor. The reaction by-products inelude
dialkyl sulphide, trialkyl dithiophosphates, sulphur, and other rrunor

substances. The acids are unstable, ancl must be converted into alkali metals
salts for use as collectors.
RO S
P + 2H 2S
RO SH
The dithiophosphoric acids are clear liquids with a characteristic smell,

soluble in organic solvents such as ether, chloroform, benzene, ethyl alcohol,
etc., but insoluble in water. They react with alkalis to form their salts.
Sodium and ammonium salts are soluble in water and are employed as
collectors. Zinc salts are employed as lubricant additives or polymerisation
stabilisers, while those of Hg, Pb, Co, and Ni can be used to purify the
dithiophosphates,as these are insoluble in water. The acids can be purified
by vacuum distillation. Boiling points of sorne acids are:
S.P. oc at mm Hg
O,O-dimethyLdithiophosphoric acid 65 - 68 3
O,O-diethyLdithiophosphoric acid 80 - 82' 2
O,O-dipropyLdithiophosphoric acid 8 - 10 2
O,O-diisopropyLdithiophosphoric acid 82 - 85 1.5
These acids react, in many cases, in a like manner to xanthic acids, for
example they form dithiols when treated with oxidising agents such as iodine:
s s
--------~ CRO)2P-S-S-P(RO)2 + 2HI
bis(O,O-diaLkyLdithiophosphate)
With chloracetic acid (Houben-Weyl (1973)) they form esters:
s S
2(RO)2P-SH + CLCO,oc2H5 ------~ CRO)2P-s-co,oc2H5 + NaCL
S-ethoxycarbonyL
O,O-diaLkyLdithiophosphate
With chloramine the reaction products are structures similar to the

thionocarbamates:
s s
O,O-Dil1lY! dithiopliosphoric acids: c1iffer from the dithiophosphates in that

they are stable acids, generally insoluble in water, with most of the
commercially available forrnulations resuIting in colIectors with frothing
characteristics, The most common types are based on phenols or cresols,
partially reacted with phosphorus pentasulphide. The reaction, as in the case
of the alkyl acids, usually is carried out using an inert atmosphere and a heel
in the reactor from the previous synthesis.
In flotation they behave similarly to the alkyl equivalents, with the exception
that as the acids are stable, they need not be converted to the alkali salts
prior to use as a collector. They will decompose in an alkaline media with
the evolution of H 2S, so may act as mild sulphidising agents during the
flotation process. Their collecting properties are very dependent on the
cresylic aeid purity used in their manufacture, so the technology of the
manufacture of reliable collectors, based on dithiophosphates, includes
having a reliable source of raw materials. Today's availability of synthetic
cresylic aeid has ameliorated the problem. As will be noted in the
descriptions of the aryl dithiophosphate collectors marketed by Cyanamid,
the usual manufacturing pro cess consists in reacting an excess of cresylic acid
with p 4SlO' so the exact composition of the collectors is unknown. The
amount of unreacted cresylic acid fixes the frothing properties of the
collector.
Aerofloat 15 consists of the reaction product of cresylic acid plus 15% by

weight P 4S 10' Theoretically the mixture will contain a 42.9 % solution of
dithiophosphoric acid in cresylic acid.
Aerofloat 25 similarly is produced by reacting cresylic acid with 25% by

weight P4SlO; it theoretically contains 72.6% collector dissolved in cresylic
acid, and therefore has a lower frother strength than AF 15.
Aerofloat 31 consists of AF 25 mixed with 6% of thiocarbanilide.
Aerofloat 33 appears to be similar to AF 31.
Aerofloat 241 is an aqueous solution of AF 25, neutralised with ammonia.
Aerofloat 242 is AF 31 neutralised with ammonia; it is similar to AF 31, but a

slightly stronger frother. \
Manufacture of dithiophasphoric acids: according to M. Oktaweic (Chem.

Abstr. 52, 12786 (1958)), the best yields (about 80%), are obtained when
cresylic acid (8 moles) is reacted with P4SlO while heating to 115C, for about
15 minutes. He does not mention the need for a heel to start the reaetion.
TabLe 16.- DENSITY AND VISCOSITY OF AEROFLOAT COLLECTORS
AerofLoat No Density Viseosity, eps

g/mL @ 20c 15C 30c
15 1.10 50.9 18.0

25 1.19 129.9 40.3
31 1.19 127.5 62.5
33 1.19 1004.7 209.5
242 1.13 157.0 57.5
The most eommon method of carrying out the reaction is to suspend the
phosphorus pentasulphide in an inert solvent, heat to the right temperature
and carefully add the alcohol to the agitated reaction mix, The most
common inert medias employed are benzene, toluene or xylene; the choice
depends on the reaction temperature.
An example of the preparation, from a patent, is:\ to 440 g of technical P4SlO

(75% grade) suspended in 500 mI of benzene, maintained at a temperature of
4QoC, 380 g of absolute ethyl alcohol are added over a 2 hour periodo Large
quantities of H?S are given off. The mixture is allowed to decant for an hour,
and then heated to 70 - 75C for 3 hours. The reaction mix is cooled, filtered,
and the filtrate vacuum distilled. The yield is 80%. Using the same method,
the yield on production of O,O-dsopropyl dithiophosphoric acid is 85%.
The aryl dithiophosphoric salts are used as flotation reagents. The solutions
are corrosive and irritating, and must be handled in stainless steel piping and
pumps, or cautiously selected plastics. They are toxic and shouldbe handled
using normal chemical safety precautions. They are more corrosive than the .
alkyl equivalents, and should be handled with greater careo They are
especially corrosive to plastics, so hoses should be avoided unless previously
tested for durability. They are highIy flammable and toxic, especially for fish.
They will produce strong allergenic reactions on handling, so complete
chemical protection should be used by the operators. They are smelly, and
can give off H 2S, which is highIy poisonous, so should be stored in well
ventilated places.
Xanthogen formates
These are the product of the reaction of a sodium alkyl xanthate with ethyl
chloroformate:
s s s o
Et-O-esNa + Et-O-eeL ----~ Et-O-e-s-e-O-Et + NaeL
Dialkyl xanthogen formates are yellow, odorous, oily liquids, essentially

insoluble in water. The diethyl xanthogen formate was used formerly as the
collector for cement copper in Leach Precipitation Float processes, but in
recent years has only been employed in the El Teniente acid circuit. The
other xanthogen formates listed are used in the copper circuit at Lepanto in
the Philippines, in Flin Flon in Canada, and Inspiration in the U.S.. The
collector formula employed in El Teniente until recently contained 30%
gasoline and 10% MIBC, which is said to have improved moly collection
slightly and reduced collector costs significant1y because the total dosage was
equal to that used with pure XF, with no loss in recovery.
Commercial xanthogen formate contains only about 75% XF. The main
impurities are active collectors: xanthic anhydride (10 to 20%), diethoxy
carbonyl sulphide (less than 10%), while diethyl carbonate, the only inert, is
present at the 2 to 5% level. The probable by-product reactions are:
s o s s s o
Ro-e-s-e-OR' + Ro-e-s ----~ Ro-e-s-e-OR + R'e-s-
xanthic anhydride
o o o o
R'o-e-s- + eL-e-OR' -----~ R'o-e-s-e-OR'
diethoxy carbonyL suLphide
o o
eL-e-OR' + R'-OH ------~ R'o-e-OR' + HeL
diethyL carbonate
------~ R'-OH + e02
Xanthogen formates hydrolyse slowly in the presence of water, giving off

COSo Normal decomposition rates are less, than 1% per year, except in the
presence of catalysts, such as Zn, Cu, Fe, or mineral surfaces. In stee1
containers or in pulps, but not in glass, xanthogen formates will react with
thionocarbamates and dithiophosphates. They react readily with a1ky1 amines
to give thionocarbamates. With tertiary amines they react to quaternary
ammonium salts,
s o S S
+ - 11 + _ 11
2CH3HOCSOCH2cH3 + 2N(CH3)3 -~ CCH3)3N-S-COCH3 + (CH3)4N-S-C3

i
CH3cH2
which could be interesting collectors,
Like the xanthates in solution, xanthogen formates can be handled in steel

pipes, and plastics, but not in the presence of copper, or copper alloys, etc..
Plastics must be used with caution as the reagents are good solvents for PVC,
but polyethylene and propylene stand up very well. If stored in galvanised
drums, these should be vented at Ieast once ayear. Steel storage tanks that
are agitated will show considerable corrosion where fluid velocities are
significant, especially if the xanthogen formate has been diluted with an
alcohol, such as MIBC. U nagitated standard steel storage tanks have been
used for over 30 years without a failure.
Like all organic compounds, xanthogen formates are flammable; due to

decomposition they may contain 1% CSz, which will reduce their flash point
into the red label range of less than 70F. The pure product has a high flash
point. Because the structure is pesticide-like, the reagent is toxic to fish and
a human irritant, with occasional subjects showing high sensitivity to skin
contact and nausea in its presence. Normal chemical safety equipment is
recommended to handle it; untreated leather boots have a limited life in
contact with reagent.
Manufacture of dialkyl xanthogen fonnates: the sole producer in the world of

xanthogen formate collectors, between 1929 and 1977, was the Minerec
Corporation, in Baltimore, U.SA.. At this plant there was a dedicated plant
for diethyl xanthogen formate, plus a number of small units for the
production of the longer chain alcohol xanthogen formates reacted only with
ethyl chloroformate.
The diethyl xanthogen formate was manufactured in the Baltimore plant by a

batch process. First the sodium ethyl xanthate was made by the process
described in the xanthate section. The xanthate, in solution in its excess
alcohol, and the water remaining from the original 50% caustic solution,
were loaded into the xanthogen formate reactor and cooled to less than 35C
with mechanically refrigerated coils. Agitation was started up and an
equimolar quantity of ethyl chloroformate was then added slowly to the
xanthate batch, at arate that kept the temperature around 35C. After all
the ECF was added, it was stirred at the same temperature for at least one
hour. The oil produced was then transferred to a wash tank where it was
Thiol Collector Chernistry 61
decarued Irom thc brine. The brine was sent lo distillation lo recover the
excess ethyl alcohol, and the oil was washed lo rernove the rernaining brine.
The process is much more sensitive than would appear from the description,
as impurities in the ethyl chloroformate can upsct the reaction, producing
unstable by-products which remain in the oil phase. The final product cannot
be neutralised and cleaned up with soda ash or caustic, as it degrades.
In 1977, Tecnomin Ud. started up a pilot plant in Chile to develop a better

process for the production of xanthogen formates. The new process is
covered by U.S. Patent 4,605,518, issued in August 1986. The improvement
consisted in converting the xanthogen formate reaction into a semi-
continuous process. This considerably improved product quality and allowed
the by-product concentrations to be controlled, which permitted tailoring the
reagent to improve recoveries on ores containing non-sulphide copper and
clay fines.
In the patented process, the xanthogen formate reactor is operated with a

high volume circulating loop, without agitation. At start-up after finishing a
xanthate batch, the reactor is loaded with about 50% of the ECF required to
fully react the xanthate batch, and the loop pump started up. Normally, the
budgeted ECF amounts to about 95 to 97% of the theoretical. Important also
is that the water content of the xanthate solution be such as to fully dissolve
all the sodium chloride that forms in the xanthogen formate reaction. The
reaction sequence then is based on the fact that the reaction is occurring in
two phases: an aqueous xanthate phase, and an oil phase in which the ethyl
chloroformate is soluble. The geometry of the piping and the reactor is
designed so that flows of the raw material and the two phases can be
controlled and dosed to the reaction zone, which is the interior of the main
circulating pump. By adjusting the oil and brine phase ratios, one can drive
the different by-product reactions and tailor the final composition of the
collector. The details of the process are contained in the patent.
TabLe 17.- PROPERTIES OF ALKYL XANTHOGEN ETHYL FORMATES
ALkyL Group MoL. Wt Density Refractive Viscosity Minerec

g/ml 20C nD 25C cps 20C name
EthyL 194.0 1.165 1..523 2.55 A

IsopropyL 208.2 1.511 2048
ButyL 222.2 1.123 1.515 3.04 B
IsobutyL 222.2 1.126 1.516 2.89 898
sec-butyL 222.2 1.113 1.511 2.61 201
2-isopentiL 363.2 1.026 1.490 2.90 27
After the xanthate batch is fully reacted, the XF reactor is circulated, at

temperature, for an hour or so, to complete the reaction, The mix is sent to
a decantation/wash tank. The alcohol-laden brine is distilled to recover the
excess alcohol, and the oil phase thoroughly washed with water, containing a
small amount of soda ash.
Dialkyl thonocarbamates
Dow Chemical's Z-ZOO (ethyl isopropyl thionocarbamate) was the first

selective collector for sulphide ores marketed. Dow's patented
thionocarbamate production was promptly followed by a somewhat similar
product made by the Minerec Corporation. The Minerec equivalent was
produced by reacting a xanthogen formate with an alkyl amine, a reaction
which produces a series of contaminants as co-products. In the early
seventies, Minerec Corporation built a plant based on a catalytic process.
The process was not as low cost as the Dow route, but did produce a higher
purity product. This plant was then operated by the Essex Chemical Co.,
which purchased Minerec Corporation, but is now shut down and the plant
dismantled due to ecological problems in Baltimore. More recently Shell, in
Antofagasta (Chile), has been producing thionocarbamates by a similar
process, apparently based on a different catalyst, as Minerec has a valid
patent in that country. There is also production in Yugoslavia, Canada,
J apan, Spain, and possibly Mexico and East Germany.
The Dow process is based on reacting a xanthic ester with an alkyl amine:
s s
IsopropyL-o-C-S-CH3 + Et-NH2 ---~ IsopropyL-O-c-NH-Et + CH3sH
Due to better yields, the Dow process is lower cost than the direct xanthate
plus amine reaction, but only if the methyl mercaptan by-product can be
disposed of economicaUy. Since Dow closed down its plant (and its process
patent expired), the xanthic ester process has not been used because the
mercaptan disposal problem has not been solved. The traditional laboratory
synthesis is based on refluxing an alkyl isothiocyanate dissolved in a great
excess of the corresponding alcohol, a process first reported by Hoffman in
1870, and probably the process employed by Douglass (1927), at Du Pont, in
1928. The reaction is:
s
R-OH + R'-N=C=S ------~ RO-C-NHR'
The first practical industrial process was developed by Harris (1954) in 1950.
It consisted of reacting an alkyl xanthate with an alkyl halide, followed by
reacting the ester forrned by the first reaction with an alkyl amine to form the
thionocarbamate:
s s s
RO-C-SNa + R"X -~ RO-C-SR' + NaX + R'NH 2 -1> RO-C-NHR' + R"SH
A more recent process, in which thionocarbamates are produced catalytically

from the direct reaction of a xanthate with an alkyl amine, was developed at
Minerec by Crozier et al (1976). The catalyst employed is either nickel salts
or palladium salts. The process is simple and has the virtue of producing a
very pure thionocarbamate. It is difficult to carry out in vitro, as a slight
excess of CS z is a key, and usually must be present as an impurity in the
xanthate.
The alkyl thionocarbamates are colourless liquids, or solids, insoluble in

water, but very soluble in organic solvents, so, when employed as collectors in
flotation they can be mixed with frothers in any proportion to aid feeding to
the flotation cells. Due to their impurity content, they have the characteristic
smell of sulphur compounds. Alkyl thionocarbamates by the ester route
normally have a purity of 90 to 86%, while the product of the catalytic
process are 98 to 99% pure, with only excess alcohol as the contaminant, so a
water wash is sufficient to ensure a good final product.
The Dow industrial process for the production of Z-200 was carried out in a
single reactor. First the xanthate reaction was produced using a large excess
of alcohol, and caustic: 4 moles isopropyl alcohol, and 1.135 moles of NaOH
per mole of CS? The reaction temperature was maintained between 35
and 50C by cooling, The reaction mix was then reacted with 1.01 moles of
methyl chloride to esterify the xanthate. This reaction was carried out at a
temperature kept between 44 and 55e, by refluxing the reactor contents.
On completion of this reaction, the amine was added, while keeping the
temperature below 500 e. The final product was steam distilled to remove
the mercaptan and the excess alcohol employed in the xanthate synthesis.
The Minerec catalytic process can be describe by the reaction:

s s
NiS04 " 6H20
-------------.'~ RO-C-NHCHCCH3)2 + NaHS
To produce the isopropyl ethyl thionocarbahiate, at Minerec, the xanthate

was first prepared using a 20% excess alcohol, without alcohol recovery, and
making the alcoholate using 75% caustic in a high intensity mixer-reactor.
The xanthate was then dissolved in sufficient water to keep the sulphhydrate
that will be produced by the next step in solution, and reacted at 800 e with a
30% excess of ethyl amine. The reaction is slow, and usually needs to be left
overnight to obtain decent yields. The catalyst was filtered off (a somewhat
mcssy job as the nickel salts produced are very fine). The collector oil was
then decanted and washed. The resultant sulphhydrate brine is a serious
disposal problern, so the Minerec, as well as the Dow plants, are no longer in
operation. Yields were on the order of 88%, and purity over 98%. The
rnethyl thionocarbamates, normally produced by Minerec, form an oil whose
density is very close to that of the brine, so to effect the decantation, MIRe is
added to dissolve the organic phase, and the mixture sold, after washing, as
collectors,
Thionocarbamates are relatively stable compounds, but under acid conditions

they hydrolyse:
s
RO-C-NHR' + H20 ------~ R-OH + R'NH2 + COS
They can be transformed into their S-alkylthiocarbamate equivalents by

reacting with alkyl iodides:
s o
a structure that makes for interesting speculation on the mechanism of

attachment of thionocarbamates to mineral surfaces.
Thionocarbamates are less corrosive than xanthogen formates, but should

still be handled in steel equipment, avoiding copper and copper alloys.
Plastics also must be tested prior to use as hoses or as pump seals.
Thionocarbamates are flammable, and should be handled with chemical
protection equipment because of possible skin rashes, etc.. They are toxic to
fish. '
Miscellaneous collector structures
Xanthic esters and long chain mercaptans are being recommended as

specialty collectors capable of enhancing molybdenite and gold recoveries
when these minerals are co-products of copper sulphide.
Xanthic al/yl esters have been marketed by American Cyanamid for a number
of years as specialty collectors for co-product molybdenum in copper ores
(Re 3302). The use of these esters as supplementary collectors has be en
proven in a number of mines, and at this writing are the only structure that
has consistently improved molybdenum recovery in a gamut of copper mines.
Mercaptans were laboratory tested extensively in the early thirties, but wefe
rejected commercially because of odour problems. Pennwalt (U.SA.), and
Phillips Petroleum (USA.), with strong mercaptan by-product positions,
have bcen trying to introduce longer chain rnercaptans as molybdenum
collectors, with varying success,
For a numbcr of years Cyanamid has been marketing various grades of

mercaptobenzothiazole (Re 404, 407, 412) as supplementary for the flotation
of tarnished and oxidised ores. They were first developed for the flotation of
lead carbonate. A similar application is given to thiocarbanilide. These two
structures are:
lO - N
11 - +
-,VC-S Ha
mercaptobenzothiazole
thiocarbanilide
\
Chapter 5
COMMERCIAl SUlPHHYDRYl COllECTORS
Only recently have collector and frother manufacturers and marketers started
to identify the chemical composition of their reagents. A particularly good
overview of reagent use as well as elear identification of their trade marked
products is contained in the latest Cyanamid (1989) "Mining Chemicals
Handbook". The tabIes in this chapter are based on Crozier (1977), but now
inelude current commercial sources of collectors and frothers,
The application data have been updated using company literature, general
references such as the "SME Mineral Processing Handbook" (1985), review
artieles, and mill visits. The tabIes should be a good starting point for a data
base which can be continuously up-dated from trade literature and technical
publications. Products no longer available commercia1ly, such as the Dow
xanthates and Z-200, and the Minerec xanthogen formates, have been
ineluded because many flotation laboratories have extensive colIections of old
sample bottles which can be used in screening tests when current reagents
fail, and if a particular structure is spectacularly successful, custom
manufacture can generally be made available. The Dow Z series
nomenelature for xanthates, and the Cyanamid thiophosphate designations
are also of interest because much of the old literature on flotation identifies
reagents only by their trade names.
The trade designations and compositions of the collectors have been grouped
under xanthates (Table 18), xanthic esters (Table 19), xanthogen formates, or
Minerecs (Table 20), thionocarbamates (Table 21), dithiophosphates
(Table 22), and miscellaneous collectors, such as mercaptobenzothiazole, etc.
(Table 23). Tabulations for individual suppliers have been ineluded, as those
with extensive product lines do publish and distribute booklets. The old Dow
products are easy to identify as the Z series are xanthates, with the exception
of Z-200, which is a thionocarbamate. Cyanamid identifies its xanthates by
the trade name Aero Xanthate and a number, its dithiophosphates by the
designation Aerofloats and a number, and unusual compositions (such as
xanthic esters) by the designation Aero promoters. A very brief surnmary of
- 66-
Cornrnercial sulphhydryl collectors 67
how and in what quantities to apply collectors to the recovery of particular

rninerals is listcd, by collector type, in the adjunct tables designated 18a, etc.
The selection of a suite of reagents for effective processing of a particular

mineral by flotation requires a considerable knowledge of surface properties
and solution chemistry, and their effect on the differential rate of the valuable
TabLe 18.- STRUCTURES, AND SUPPLIERS OF ALKYL XANTHATES

R-O-CC=S)-S M Where M is sodium or potassium
Type Manufacturer Trade name
Potassium EthyL X Am. Cyanamid Aero X. 303

CH 3CH20CC=S)S-K+ Dow ChemicaL Z-3
Sodium EthyL X Am. Cyanamid Aero X. 325

CH 3CH20CC=S)S-Na+ Dow ChemicaL Z-4
UCB CBeLgium)
KerLey Chem.
Potassium Isoprop~L X Am. Cyanamid Aero X. 322

CCH3)2CHOCC=S)S-K Dow ChemicaL Z-9
UCB CBeLgium)
Sodium isopropyL X Am. Cyanamid Aero X. 343

CCH3)2CHOCC=S)S-Na+ Dow ChemicaL Z-11
UCB CBeLgium)
ZUPA
Hoechst PIPX
CANDINA FLex 415/P
KerLey Chem. KI-80L
SheLLflot SF-113
Sherritt-Gordon
IQM
RENASA
Sodium isobutyL X Am. Cyanamid Aero X. 317

CCH3)2CHCH2oCC=S)S-Na+ Dow ChemicaL Z-14
Hoechst PIBX
CANDINA FLex 215/P
SheLLflot SF-114
Stanchem-CIL
IQM
ZUPA
Sodium sec-butyL X Am. Cyanamid Aero X. 301
CH3CH2CHCCH3)OCC=S)S-Na+ Dow. ChemicaL Z-12
X = xanthate
Table 18 cont'd.- STRUCTURES, AND SUPPLIERS OF ALI(YL XANTHATES

R-O-CC=S)-S f<1 uher e 11 is sod i um or potassium
Type 'lanuracturer Trade name

Potassium sec-amyl X Dow Chemical Z-5
Methylpropyl carbinoL
CH3CCH2)2CHOCH3CC=S)S-K+
Potassium amyl X Am. Cyanamid Aero X. 350
CH3 CCH2)3CH20C C=S) S-I(+ Dow ChemicaL Z-6
Hoechst PAX
CANDINA FLe){ 315/P
Prospect Chem.
KerLey Chem. KI-SOL
Sodium isoamyL X Am. Cyanamid Aero X. 355

CCH3)2CCH2)2CH20CC=S)S-Na+ Stanchem-CIL
Sherritt-Gordon
Prospect Chem.
Industrias Quimicas
Mexi canas (IQM)
Potassium hexyL X Oow ChemicaL Z-10

CH3CCH2)50CC=S)S-K+
and the undesirable minerals in a panicular ore. But the choice of specific
equipment to employ - say selecting between the use of giant cells, flotation
columns, and different schemes of "flash" flotation - is still an art, and the
particular sequence of the steps used in a separation process direct1y affects
how collectors, frothers, modifiers, and depressants interact, so it is
impossible to give categoric recommendations on how or where to apply
given reagent. Collecting and collating case histories for flotation mills is a
monumental task that has not be en repeated since Professor Arthur Taggart
put out the last edition of his "Handbook of Mineral Dressing" nearly fifty
TabLe 19.- STRUCTURES, ANO SUPPLIERS OF XANTHIC ESTERS
R-O-CC=S)-S-R'

ALLyL AmyL Am. Cyanamid Aero P. 3302
Minerec Corp M-17S0
ALLyL hexyL Am. Cyanamid Aero P. 3461
Minerec Corp M-2023
Commercial sulphhydryl collectors 69
years ago. Bis recomrnendations on how unusual ores have been deah with is
stil! probably the only reliable source for suggestions on the initial reagents to
be uscd in an experimental prograrnme to provide process design data. But it
cannot cope with al! the perrnutations of combined minerals that have
contradictory flotation responses. Despite this, it certainly is the most prized
possession of the mil! manager of a boutique concenrrator purchasing ores
from small miners in a polymetal!ic district.
TabLe 20.- STRUCTURES, AND SUPPLIERS OF DIALKYL XANTHOGEN FORMATES

GeneraL structure R-O-CC=S)-S-CC=S)-O-R'
DiethyL XF Minerec Corp Minerec A
CH3CH2oCC=S)SCC=O)OCH2CH3 Tecnomin Ltda T-2001
SheL Lflot SF-203
Prospect Chem.
Dow ChemicaL DFC-404
EthyL isopropyL XF Minerec Corp M-2048

CCH3)2CHOCC=S)SCC=O)OCH2CH3 Dow ChemicaL DFC-411
EthyL butyL XF Minerec Corp Minerec B

CH3CCH2)30CC=S)SCC=O)OCH2CH3
EthyL isobutyL XF Minerec Corp M-898, 1995
CCH3)2CH2CH20CC=S)SCC=O)OCH2CH3
EthyL secbutyL XF Minerec Corp M-201
CH3CH2CHOCH3CC=S)SCC=O)OCH2CH3
EthyL amyL XF Minerec Corp M-104D
CH3CCH2)3CH20CC=S)SCC=O)OCH2CH3
EthyL methyLamyL XF Minerec Corp M-27
or EthyL 2-heptanoL XF
CH3CCH2)4CH20CH3CC=S)SCC=O)OCH2
TabLe 21.- STRUCTURES, AND SUPPLIERS OF DIALKYL THlONOCARBM1ATES

R-O-CC=S)-NH-R'
EthyL isopropyL TC Dow ChemicaL Z-200
CCH3)2CHOCC=S)NHCH2CH3 Minerec Corp. M-1661, 2030
SheLLfLot SF-323
ZUPA
KerLey Chem. KI-200
Hoechst X-23
Am. Cyanamid AP-3894
MethyL isopropyL TC Minerec Corp. M-1703
CH3)ZCHOCC=S)NHCH3 Dow ChemicaL ZM-ZOO
Methy L buty L TC Minerec Corp. M-1331

CH 3CCHZ)OCC=S)NHCH3 Am. Cyanamid Aero P. 6098
MethyL isobutyL TC Minerec Corp. M-1846
CCH3)2CHCH20CC=S)NHCH3
EthyL isobutyL TC Minerec Corp. M-1669
CCH3)2CHCH20CC=S)NHCH2CH3
TabLe 22.- STRUCTURES, AND SUPPLIERS OF THIOPHOSPHATE COLLECTORS
GeneraL structure - DIALKYL DITHIOPHOSPHATES

R-O S
V
R-O/ ""S-Na +
DiaLkyL dithiophosphates
Na diethyL DTP Am. Cyanamid Sod. AerofLoat
CCH3CH20)2CP=S)S-Na+ Hoechst HOSTAFLOT LET
CANDINA SPELD 3456
PhiLL ips 66 CO 500
Na diisopropyL DTP Am. Cyanamid AF-211
[CCH3)2CHOJZCP=S)S-Na+ AF-243
Hoechst HOSTAFLOT LlP
CANDINA SPELD 3457
PhiLLips 66 CO 510
Cornmercial sulphhydryl collectors 71
Table 22 cont'd.- STRUCTURES. AND SUPPLIERS OF THIOPHOSPHATE COLLECTORS

Na diisobutyl DTP Am. Cyanamid Aero P - 3477
[(CH3)2CHCH20(P=S)S-Na+ Aero P - 5430
ALL ied CoLLoids DPI-4560
Hoechst HOSTAFLOT LIB
CANDINA SPELD 4659
Minerec M-2044
Phillips 66 co 540
Na Di-sec-butyl DTP Am. Cyanamid AF-238
[CH3(CH2)2CHOCH3J2(P=S)S-Na+ Allied ColLiods Procol CA825
Hoechst HOSTAFLOT LBS
CANDINA SPELD 3458
Minerec M-2176N
PhiLLips 66 C0530
LubrizoL FLotezoL 100
Na DiisoamyL DTP Am. Cyanamid Aero P - 3501
[CH3(CH2)3CH20J2(P=S)S-Na+ Aero P - 5474
Na di-methyLamyL DTP Am. Cyanamid AF-249
[(CH3)2(CH2)3CH20J2(P=S)S-Na+
DIARYL DITHIOPHOSPHATES
These coLLectors are
manufactured by reacting
cresyLic acid with P2S5.
The Cyanamid number is the
percentage of theoreticaL P2S5
used in the reaction.
15% P2S5 Am. Cyanamid AF-15

CANDINA SPELD 1335
Hoechst PHOSPHOKRESOL C
25% P2S5 Am. Cyanamid AF-25
ALLied CoLLoids ProcoL CA825
CANDINA SPELD 1334
Hoechst PHOSPHOKRESOL B
Minerec M-2074
Mixtures of AryL DTP
AF-25 pLus 6% Thiocarbanilide Am. Cyanamid AF-31. AF-33
ALLied CoLLoids ProcoL CA832
CANDINA SPELD 1333
Hoechst PHOSPHOKRESOL A
AF-25 neutraLized with NH40H Am. Cyanamid AF-241
ALl ied ~oLLoids ProcoL CA833
CANDINA SPELD 2666
Hoechst PHOSPHOKRESOL E
AF-31 neutraLized with NH40H Am. Cyanamid AF-242

72 FLOTATlN - Theory, Rcagents and Ore Testing
TabLe 22 cont'd.- STRUCTURES. AND SUPPLIERS OF THIOPHOSPHATE COLLECTORS
Mixture of AF-238 pLus sodium Am. Cyanamid AF-208

AerofLoat CANDHIA SPELD 5658
Hoechst HOSTAFLOT LBS
MONOTHIOPHOSPHATES
R-O, s
Am. Cyanamid Aero P. 3394
Am. Cyanamid Aerophine 3418A
TabLe 23.- STRUCTURES, AND SUPPLIERS OF MISCELLANEOUS COLLECTORS
Mercaptans
n dodecyL C12H25SH PhiLLips 66 CO 100
t dodecyL or LauryL CO 200
MercaptobenzothiazoLe
rg-~'c=s Am. Cyanamid

Hoechst
Aero P. 400
FLotagen
~rr Captan
Mixtures of
mercaptobenzithiazoLe with Am. Cyanamid Aero P. 404
dithiophosphates Aero P. 407
Aero P. 412
Aero P. 425
Thionocarbamate
dithiophosphate mixtures Am. Cyanamid Aero P. 4037
ThiocarbaniLide
O - M
II -
,.g....... c-S H.
+
Am. Cyanamid Aero T. 130
A thumb nail summary of the properties of the more common collectors is

that the xanthates are work horse, relatively non-selective collectors. The
shorter the alkyl group the more selectivity, which is frequent1y referred to as
Comrriercial sulphhydryl collcctors 73
a "weaker" collector, while the longer chained xanthates, such as amyl or

hexyl, are less selective, or stronger collectors.
The phosphorous pentasulphide based collectors, III general, have

characteristics that parallel the xanthates, except that, for the same alkyl
group, they tend to be more selective (weaker), and are easier to tailor by
mixing with complementary collectors to enhance recovery of secondary
values in an ore, such as gold, silver and rnolybdenum. The cresylic acid
derivatives have frothing properties, which can be desirable or undesirable
depending on the ore type, and/or the size distribution of the pulp particles,
and the process water quality.
The thionocarbamates were first studied, in the late twenties, on a laboratory

scale by Dr. WilIiam Douglass, of du Pont. The Dow Chemical Company
first recognised the industrial value of ethyl isopropyl thionocarbamate (Z-
200) in the early sixties. They popularised its use as a general purpose, highly
selective collector for low grade porphyry copper minerals, and as a selective
adjunct in c1eaning more complex ores. They did not extend the alkyl groups
available commercialIy. In the late sixties Dr. Fischer, at Minerec, recognised
that methyl butyl and isobutyl thionocarbamates could replace Z-200 when
the mill requirements combined good recovery and selectivity.
Xanthogen formates were first developed by Minerec, in the late twenties, to

replace xanthates in acid circuits. They are unique in the copper-
arsenopyrite and copper-zinc circuits, where moderately low pH flotation is
mandated. Theyare also used when slimes and clays are a problem.
Mercaptans were mildly popular as colIectors in the early thirties, when mill
operators equated odour with colIecting strength. They faded from use
because of their smell, and, possibly more important, because the short chain
mercaptans have frother properties. Recently the very long chain ones
(lauryl mercaptan) have been touted as a replacement for fuel oil in the
primary and by-product recovery of molybdenum. Trithiocarbonates (i.e.,
xanthates based on mercaptans) were also popular sixty years ago. Sorne are
still used in South Africa.
A caution that cannot be repeated often enough when designing a test

programme for an operating mill is looking out for reagent interactions. A
common cause of anomalous results in a sid'e by side rougher bank test is
recycled scavenger concentrate, which will contain smalI quantities of the rest
of the mill's reagents.
-..
.:,.
Table 18a.- STRUCTURES, AND MINERAL USAGE OF ALKYL XANTHATES OR ALKYL DITHIOCARBONATES
R-O-(C=S)-S M Where Mis sodium or potassium
Type Mineral Dose gft Complementing Dose gft Flotation Conditions '"Tj
t""'
Potassium EthyL X a) Sphalerite 8 - 40 none - Natural pH 8 - 9 with 150 - 330 gft Cuso4 O
CH3 CH20(C=S)S-K+ b) Chalcopyrite 30 none - At pH 10.3 adjusted with lime ~
>-~
c)
d)
Cinnabar
other metal
sulphides
e) Malachite
45
8 - 80
66
Creosote
K amyl xanth.
60
66
Natural pH 8 with 300 gft CUS04 activation
pH 9.5 with lime at Kamoto, Zaire

-
O
'"7
"'-<
Sodium Ethyl X a) Sphalerite 45 - 60 none - Natural pH 8 - 9.5 or adjusted with Lime t-J
:r'
CH 3CH20(C=S)S-Na+ Coal tar 5 adding 330 - 500 gft Cuso4 for activation o
FueL oi L O
15 ....
b) GaLena 33 - 450 none - NaturaL pH 7.5 :<:
CoaL tar 1 ?:J
220 ButyL xanthate 80 For finely disseminated galena in ore with (1l
:o
high carbon, naturaL pH 8. (]q
(1l
c) Bulk sulphides 125 Creosote 70 SuLphuric acid to pH 5 ~
d) other metal 8 - 80 .....
C/l
sulphides :o
e) Bornite 17 K amyL xanth. 20 pH 10.7 with Lime at Mufulira, Zambia I:l
e,
Potassium Isoprop~l X a) Chalcopyrite 65 Barrett oiL 40 At pH 10.2 with Lime O

....
(1l
(CH3)2CHO(C=S)S-K b) Chalcopyrite, 12 Z-200 12 At pH 10.2 with Lime
bornite t-J
(1l
c) Free goLd, 60 Aerofloat 25 1.3 At natural pH 7.8 :4-
auriferous
suLphides
s
(]q
TabLe 18a cont'd.- STRUCTURES, ANO MINERAL USAGE OF ALKYL XANTHATES OR ALKYL DITHIOCARBONATES
Type MineraL Oose gft CompLementing Dose gft FLotation Conditions
Sodium isopropyL X a) SphaLerite 30 - 100 none - At pH 10.0 - 10.8 adjusted with Lime and
(CH3)2CHO(C=S)S-Na+ 25D - 370 gft cuso~ activation
b) Marmatite, 55 none - At pH 10.8 adjuste with Lime and 205 gft
Cuso4 activation
c) Marmatite, 50 none - At pH 11.0 adjusted with Lime and 110 gft
SphaLerite (j
CUS04 activation o
d) ChaLcopyrite 45 none - At pH 8.5 adjusted with Lime S
e) ChaLcocite 16 none - At pH 11.4 adjusted with Lime 8(1J
f) ChaLcocite, 100 AerofLoat 249 12 At pH 9.5 ...,
native Cu FueL oi L 165 \J.
g) GaLena 8 - 60 none - At pH 7.0 - 10.5, either naturaL e.:..
or adjusted with Lime
h) Stibnite 16 none - At pH 7.7 with Lime, using 2.7 kgft Lead
-
acetate activation =
eO
::r
i) pyrite v th 4 - 40 none At pH 5 - 8.5 with suLphuric acid ::r
;-
cobaLt vaLues 60 AerofLoat 25 8 At pH 9 with suLphuric acid and 80 gft '-<
CUS04 activation
...,e,
-s,
j) Pyrrhotite 40 none - At pH 10.0 with Lime
"
2-
Sodium isobutyL X a) pyrite 10 - 20 none At naturaL pH, powerfuL coL Lector Cii'
(CH3)2CHCH20(C=S)S-Na+ b) ChaLcopyrite 12 - 15 AerofLoat 238 12 At ph 9 with Lime "...,
O"
U1
_1
U1
-.
TabLe 18a eont'd.- STRUCTURES, AND MINERAL USAGE OF ALKYL XANTHATES OR ALKYL DITHIOCARBONATES 0\
Type Manufaeturer MineraL Dose gft CompLementing Dose gft FLotation Conditions
Sodium see-butyL Am. Cyanamid a) ChaLeopyrite 60 none - At pH 10.0 with Lime

CH3CH2CH(CH3)O(C=S)S-Na+ Dow Chemi ca L b) ChaLeopyrite, 70 none - At pH 12.0 with Lime and 300 gft
'Tj
ehaLeoeite NaSH for activation r-
e) ChaLeopyrite, 16 - 50 ThioearbaniLide 16 At pH 11.5 with Lime O
ga Lena with Au Aero Promoter 2 At naturaL pH 6.0 - 8.0 r--]
and Ag vaLues 404 ;J>
d) GaLena 20 none - At naturaL pH 8.0 ....-J
e) SphaLerite 8 - 16 none - At pH 11.0 - 11.5 with Lime O
Burner oi L 4 and 80 - 250 gft CUS04 activation Z
f) Cinnabar 125 none - At pH 8.0 with soda ash and 200 gft
CUS04 for aetivation -J
100 - ::;'
g) Free Au, auri- none At pH 8.6 with Lime, pLus 4 gft (1l
ferous pyrite fatty aeid as frother aid O
...,
:;<
Potassium see-amyL X Dow ChemieaL a) SphaLerite 32 none - At pH 10.4 with Lime,
:AJ
and 185 gft CUS04 (1l
~
MethyLpropyL earbinoL b) ChaLeopyrite 40 Creosote 16 At pH 9.6 adjusted with Lime [[q
CH3(CH2)2CHOCH3(C=S)S-K+ e) Stibnite 125 none - At naturaL pH 6.5, using (1l
:=t
1.2 kgft Lead acetate activation r-e
d) AngLesite 330 none - At pH 10.2 using 3.3 kgft NaSH f or
00
~
suLphidisation ;:l
p..
e) CobaLtite 95 none - At pH 8.8 with soda ash,
adding 500 gft CUS04 and O
t-t
450fgft NaSH to act,vate (1l
f) Pyrrhotite 82 AerofLoat 25 55 At naturaL pH 8.4, -J

(1l
adding 20 gft CUS04
~
.....
;:l
[[q
TabLe 18 eont'd.- STRUCTURES, ANO MINERAL USAGE OF ALKYL XANTHATES OR ALKYL DITHIOCARBONATES
R-O-(C=S)-S M Where M is sodium or potassium
Type MineraL Dose g/t CompLementing Dose g/t FLotation conditions
Potassium amyL X a) ChaLeopyrite 8 - 40 none - At pH 10.0 - 10.5 with Lime
CH3(CH2)3CH20(C=S)S-K+ Z-200 8 At pH 10.0 - 10.5 with Lime
8 - 30 Aero Prom. 404 1 - 12 At pH 11.5 - 11.8 with Lime
21 Z-11 or Z-9 6 At pH 9.5 with soda ash (J
b) ChaLeopyrite, 45 none - At pH 8.4 with Lime O
gaLena a
e) SphaLerite 220 none - At pH 12.1 with Lime, and 825 g/t CUS04 S
co
d) BuLk bismuthite, 20
native bismuth,
AerofLoat 208
AerofLotat 25
9
3
puLp aerated 15 min to enhanee seLeetivity
At naturaL pH 6.9 a.&...
C~
..o=
Sodium isoamyL X moLybdenite
(CH3)2(CH2)2CH20(C=S)S-Na+ e) Stibnite 165 none - At pH 8.0 with soda ash, adding
740 g/t Lead aeetate for aetivation ;:T'
;:T'
.. f) free goLd
g) Cerussite
80
80
none
none
-
-
At pH 8.6 with soda ash
At pH 10.0 with soda ash, adding
'-<
O-
\~
1 kg/t NaSH for suLphidisation
h) MaLaehite 50 PaLm oiL 16 At pH 9.0 using 900 g/t NaSH to n
suLphidise g.
i) Cuprite 1,650 none - At pH 9.0 with Lime, adding 500 gft i'P
n
NaSH to suLphidise O'
....
Potassium hexyL X a) ChaLeopyrite 8.3 Aerofloat 238 8 At pH 11 with Lime '"
CH3(CH2)50(C=S)S-K+
--..J
--..J
cxJ
"'l'j
TabLe 19a.- STRUCTURES, AND MINERAL USAGE OF XANTHIC ESTERS L'
O
R-O-(C=S)-S-R' ~
>-l
Type MineraL Dose g/t CompLementing Dose g/t FLotation Conditions
5
ALLyL AmyL a) ChaLcopyrite, 7.5 Z-11, AX-343 4.5 At pH 10.5 with Lime
z
chaLcocite, ""'3
moLybdenite ::T
(il
b) cu, zinc, moLy 2 - 20 FueL oiL 80 ...,
O
suLphides; :;::
suLphidised Cu
carbonate ~
(il
Pl
(]q
AlLyl hexyl a) SimiLar to 2 - 20 (il
AP-3302 ...,.
I:l
en
Pl
I:l
e,
O
...,
(il
>-j
(il
,.,.
~"
5
(]q
TabLe 20a.- STRUCTURES, AND MINERAL USAGE OF DIALKYL XANTHOGEN FORMATES .
R-O-CC=S)-S-CC=S)-O-R'
Type Mi nera L Dese gft CompLementing Dese gft FLetatien Cenditiens

DiethyL XF a) ChaLeepyrite 8 - 16 2-11, AX-343 4 At pH 10.5 with Lime
CH3CH2oCC=S)S(C=O)OCH2CH3 80 nene - At pH 5 with suLphurie aeid
16 AmyL xanth. 20 At pH 7.5 with suLphurie aeid in eLeaning
magnetite eeneentrate
b) Cement eepper 40 nene - At pH 4.5 with Lime in an LPF eireuit,
aLse in sea water pulpo
e) GaLena 40 nene - At pH 9.5 - 10.0 with seda ash (J
o
EthyL iseprepyL XF :3
(CH3)2CHO(C=S)S(C=O)OCH2CH3 :3
(1J
;:;
EthyL butyL XF a) Cu and Pb 8 - 80 ALkaLine eireuits :
CH3(CH2)30(C=S)S(C=O)OCH2CH3 suLphides en
EthyL isobutyL XF a) ChaLeeeite 12 - 16 nene - At pH 10.0 - 10.5 with Lime

So
"d
- ~
(CH3)2CH2CH20(C=S)SCC=O)OCH2CH3 b) ChaLeopyrite 100 nene At pH 10 with Lime
,~
e,
EthyL seebutyL XF a) SphaLerite 8 - 30 nene - At pH 9 - 9.5 (Lime) adding 400 -
'$..
800 gft Cuso4 fer aetivation
2".-
CH3CH2CHOCH3(C=S)SCC=O)OCH2CH3
(D
EthyL amyl XF o
CH3(CH2)3CH20(C=S)S(C=O)OCH2CH3 8
.....
en
EthyL methylamyl XF a) Chaleepyrite 20 nene - At pH 9 with lime
or Ethyl 2-heptanel XF b) Cu sulphides in 8 - 80
CH3(CH2)4CH20CH3(C=S)S(C=O)OCH2 Cu-2n ores
-.l
\0
ce
o
TabLe 21a.- STRUCTURES, AND MINERAL USAGE OF DIALKYL THIONOCARBAMATES

R-D-(C=S)-NH-R'
"Tj
Type MineraL Dose gft CompLementing Dose gft FLotation Conditions L'
O
>-]
EthyL isopropyL TC
(CH3)2CHO(C=S)NHCH2CH3
a) ChaLeopyrite
b) ChaLeoeite
14
5
Aero P 404
Isoprop. Xanth.
1
9
At pH 10.5 with Lime
At pH 11 with Lime

4
e) SphaLerite 33 none - At pH 9.3 (Lime) pLus
25
370 gft Cuso4 for
aetivation z
...,
MethyL isopropyL TC ::r
(1l
CH3)2CHO(C=S)NHCH3 ..,
O
~;;::,:
MethyL butyL TC a) ChaLeopyrite 31 EthyL xanth. 9 At pH 8.5 with Lime
CH3(CH2)O(C=S)NHCH3 b) ChaLeoeite 8 - 20 none - At pH above 12 ::J
(1)
e) Cement Cu 295 none - At naturaL pH 5 of dump :>l
q
Leaeh iron Launders (1)
d) SphaLerite 12 EthyL xanth. 12 At naturaL pH 8 with g
300 gft CUS04 en
:>l
MethyL isobutyL TC 5-
(CH3)2CHCH20(C=S)NHCH3 O
@
EthyL isobutyL TC a) ChaLoeite 8 ...;
(CH3)2CHCH20(C=S)NHCH2CH3 (1)
~
5"
:
TabLe 22a.- STRUCTURES, ANO MINERAL USAGE OF THIOPHOSPHATE COLLECTORS
GeneraL structure - DIALKYL DITHIOPHOSPHATES
R-O 5
"p/
R-O/ "'5- Na +
Type MineraL Dose gft CompLementing Dose gft FLotation Conditions

o
O
El
DiaLkyL dithiophosphates a) ChaLcocite 10 Z-200 3 At pH 11.8 with Lime ElCD
Na diethyL DTP b) MisceLL. 18 Z-14 4 At pH 11 with Lime ...,
(CH3CH20)2(P=S)S-Na+ Cu SuLphides Q.
c) SphaLerite 12 - 30 None At pH 9 - 9.5 (Lime) adding 160 - 205 gft e:..
Cuso4 to activate
40 Z-200 20 At pH 9.3 (Lime) adding 270 gft of '"
E..
"O
CUS04 to activate ::r'
e-
'-<::
Na diisopropyL ~P a) SphaLerite 12 - 50 None At naturaL pH 8 - 8.5 adding 410-600 gft ...,o,
[(CH3)2CHOJ2(P=S)S-Na+ CUS04 to activate. At pH 10.9 (Lime) 's,
and 250 gft CUS04 to activate ro
33 Z-6 12 At pH 11.6 (Lime) pLus 205 gft CUS04 S2..
to activate the zinc (D
ro
Na diisobutyL DTP a) cu, Zn, Ag 8 - 40 S
...,
[(CH3)2CHCH20(P=S)S-Na+ ores '"
DO
......
00
TabLe 22a eont'd.- STRUCTURES, ANO MINERAL USAGE OF THIOPHOSPHATE COLLECTORS N
Type MineraL Dose g/t CompLementing Dose g/t FLotation Conditions

Na Di-see-butyL DTP al ChaLeopyrite, 20 Z-11 4 At pH 10.5 with Lime
[CH3(CH2'2CHOCH3J2(P=S'S-Na+ ehaLeoeite
bl cu, Zn and pb 8 - 40 ALkaLine eireuits '"'f'I
suLphides l
O
Na DiisoamyL DTP al Cu, Zn and Pb 8 - 40 s:;
[CH3(CH2)3CH20J2(P=S)~-Na+
Na di-methyLamyL DTP
[(CH3)2(CH2'3CH20J2(P=S)S-Na+
suLphides
a) Cu suLphides 8 - 40
-
>-l
O
Z,
>-j
DIARYL DITHIOPHOSPHATES ;:r"
ro
These eoLLeetors are CH 3-"p/S ...,
O
manufaetured by reaeting CH -/ "<, 5- NQ + ':<
eresyLie aeid with P2S5. 3
The Cyanamid number is the :::o

(1l
:
pereentage of theoretieaL P2S5 I)q
used in the reaetion. (1l
g
Ul
15;( P2s5 a) Cu, Zn, Pb and 2 - 80 ;;
Ag suLphides ~
e,
25;( P2S5 a) Free Au, 80 Z-4 80 At naturaL pH 7.2 O
-t
auriferous pyrite (1l
b) GaLena 66 None - At naturaL pH 8.5 ~
e) Argantite 40 Z-6 40 At naturaL pH 8.8 (1l
d) DeseLoizite 80 Z-6 40 At ph 11 (Lime) pLus 1.6 kg/t .....

Ul
pb nitrate to aetivate S
I)q
e) MetaLLie Fe 100 Z-6 40 At pH 4 (HCl)
Table 22a cont'd.- STRUCTURES, ANO MINERAL USAGE OF THIOPHOSPHATE COLLECTORS
Type Mineral Dose gft Complementing Dose gft Flotation Conditions

Mixtures of Aryl DTP a) Galena 37 - 50 None At naturaL pH 7.5 - 8.5
AF-25 plus 6% Thiocarbanilide b) Cu, Pb and 20 - 80
Ag sulphides;
oxidized gold ores
AF-25 neutraLized with NH40H
AF-31 neutralized with NH40H a) ChaLcopyrite 20 - 29 Z-6 4 - 33 At pH 9 - 9.5 with soda ash
b) " , GaLena 50 Z-6 10 At pH 9.8 with soda ash
c) GaLena 16 None At naturaL pH 8
140 None At pH 9.5 with NaOH ()
e) Argentite 4 None At naturaL pH 8 o
3
Mixture of AF-238 plus sodium a) ChaLcopyrite, 29 Z-6 4 At pH 9.5 - 10.5 with Lime S
Aerofloat Au and Ag vaLues
...,lO
b) Sphalerite 23 None At naturaL pH 8.5 adding !J.
c) Native Au, 4 - 40 330 gft CUS04
2-
Cu and Ag '"
E..
"el
~
MONOTHIOPHOSPH.(TES .
R-o'p/S
R-o/ 'S-No.+
~
o
s.
DITHIOPHOSPHINATES
co
(")
R"P/'S 5'
....
R / 'O-No. ...
'"
o:
w
TabLe 23a.- STRUCTURES, ANO MINERAL USAGE OF MISCELLANEOUS COLLECTORS CO
""-
Type Mi nera L Dose 9/t CompLementing Dose g/t FLotation Conditions
MercaptobenzothiazoLe
O-Lc~~
'Tj
t'""'
o-r O
>-3

>-3
.-
O
Mixtures of Z
mercaptobenzithiazoLe with a) ChaLcopyrite 57 Z-11 6 At pH 9.8 with Lime ..-j
dithiophosphates i:r"
(1J
O
...
:<
Thionocarbamate ::o
(1J
dithiophosphate mixtures substitute for Z-200 ~
()q
(ll
ThiocarbaniLide a) Ag suLphides 33 AF-238 12 At naturaL pH 8 g
'l
O-M 11
-,S.....-c-S
~
H.
+
~
;:t
O-
O
....
(ll
ThiophosphoryL ChLoride a) BuLk SuLphides Used in acid circuits >-3
{1l
"-0",,/'
R-o/ "o
=.
;:t
()q
Chapter fa
PROPERTIES OF FlOTATION FROTHS
Flotatlon frothers and froth hydrodynamics
As we have seen, the choice of a frother, or a frother combination, for a

particular flotation application depends on the frothing mechanism involved.
For sulphide flotation it is not unusual that two or more frothers will be
blended, as one must complement the collector used to form interaction
complexes, and at the same time, flotation efficiency depends on a
mechanically satisfactory froth. In a majority of the cases a frother
containing at least a portion that is an alcohol, such as MIBC, is the best
choice to match the collector, and a higher molecular weight frother, such as
the pine oils or the polypropylene glycols, can be added to modify the
physical properties of the froth andjor the bubble size when coarser partic1es
are to be floated.
Dynamic froth stability
Froths break down through rupture of bubbles due to film thinning, as shown
by Ewers and Sutherland (1952), who state that the key factor is movement
of inter-bubble liquid away from the point of potential rupture: "The surface
moves from a region of low surface tension (high surface pressure) to a
region of high surface tension. When the surface tension is highest at the
centre of the disturbance, the film will be stable; when the surface tension is
lowest at this point the surface film and hence the substrate will move away
from this point, and the film will rupture". Thus, low molecular weight
frothers, such as aleohols, which tend to diffuse rapidly, will even out surface
tension differences along the bubble film, and.produce less stable froths. This
concept of film flow hindrance explains the better stability of three-phase
froths with inter-bubble mineralloads. Here tt.e mineral partic1es act as pegs
locking the bubbles in position, as do slimes, colloids, and other compounds
affecting inter-bubble liquid viscosity.
The stabilising effect of the mineral IS the reason that flotation in a
- 85-
Hallimond tube, in the absence of a frother, is possible. Because of this, great

care should be taken in extrapolating results based on Hallimond tube data,
because conclusions of the relative merits of a collector with a particular ore
are questionable in the absence of a frother.
Surface tension differences between frothers is probably of trivial importance

in selection of frothers, as in commercial flotation the normal frother
concentrations are such that the lowering of surface tension in a mili pulp is
hardly measurable. In fact, a significant lowering of the pulp surface tension
is probably deleterious, as many have found when a detergent got into their
flotation circuit by accident. The observed lack of load carrying capacity of a
Iroth, or foam, generated by a detergent, may well be due to non-interaction
of the collector with the detergent on the mineral surface; i.e. the detergent
coats the bubble surface to form a stable foam, but does not link to the
collector on the mineral surface.
The effect of the collector on froth properties
Frother Power: Wrobel (1953) defined this term as the volume of froth
generated in a standard machine under standard operating conditions. Fig.
16 shows froth voIume data for systems without a mineral pulp, reported by
Leja and Schulman (1954) for straight, branched chain and cyclic alcohols
having a varying number of carbon atoms, used alone or in conjunction with
straight chain xanthates. In all cases, reagent concentration was one
thousandth molar (M/1000).
As can be seen from Fig. 16, sorne general conclusions can be reached on
the effect of carbon atoms in the alkyl group, or chain length, on the froth
properties generated by aleohols. Froth volume, which Wrobel equated to
frother power, increases with the number of carbon atoms up to 6 or 7, and
then drops drastically when the alcohol has more than 8 carbon atoms. A
xanthate, with less than 6 carbon atoms in the alkyl group, alone does not
produce frothing, while lauryl xanthate wiIl produce a froth on its own, but
will kill the froth if mixed with an aleohol, possibly explaining why long chain
xanthates are depressants. With strong frothers, such as cresylic acid or pine
oil, lauryl xanthate produces a non-mineralisable foam.
The curve for froth volume versus number of carbons in the alcohol shows a
steady increase in froth volume, up to 7 carbon atoms. Mixing the alcohol
with longer chained xanthates, such as potassium amyl xanthate (KAX),
increases froth volume for the short chained aleohoIs and strongly increases
frother power, when combined with alcoholic frothers with up to 7 or 8
carbon atoms, while a shorter chain xanthate such as potassium ethyl
Properties of Flotation Froths 87
xanthate (KEX) dccreases frother power for aleohols with 5 or less earbon
atoms, and only moderately inereases the froth volume for the longer ehain
aleohols. Lekki and Laskowski (1975) repeated sorne of the above frother
power measurements, in the presenee of minerals, and, in general, confirmed
Leja and Schulman's pure frother /collecror data.
Fl'llthel' + lilil Fll'othell' +

HUI 100 IIAX
8111 8111
Fl'oth Gil Gil

lJolul'le
411 411
31a 21a
HUMhel' oC e a tDM5 in all<yl ehafn HUMhel' oC e a tOM5 in Ll'anched

chain alcohol
4 e atoOls-isobutanol 5 e atoOls- 2-Oleth!,lI-2-butanol
6 e atoOls=4-roeth!,lI-2-pentanol 7 e atoros=2-eth!,llhexanoI
11 e atoros= MIBe
Fig. 16.- Frother volume versus alcohol chain length
These data support the observation in operating mills that control over the
froth hydrodynamics is easier through changes in the collector than by
changing dosage or type of frother employed. It also explains the fact that
longer chain xanthates are "stronger" collectors, because their effect on froth
texture is to make it less selective due to retention of particles, which
automatically results in increased overall recovery coupled with reduced
concentrate grade. At least in sulphide flotation, the correlation between
contact angle and recovery could well be fortuitous, as under normal
operating dosages the contact angle for the different xanthates, if present,
does not vary a measurable amount. \
As noted previously, in the presenee of a mineral, froth stability is enhanced

and the froth character changes. For example, arnyl xanthate combined with
alpha-terpineol (the main constituent of pine oil) is a strong frothing
combination, but in the presence of a sulphide mineral these reagents will
88 FLOTATION - Theory, Reagents ami Ore Testing
tend lo produce Iittle or no froth and only a thick dry film of mineral will
float. Salts (clectrolytes) in solution in the pulp will generally significantly
reduce froth stability during sulphide flotation, even if present only in very
low concentrations. On the other hand, in coal flotation, sea water can be a
good self frothing media. These also are phenornena that can be explained
by Irother-collector interaction on the mineral.
The nfluence of the frother 00 the rate of flotaton
From two different papers by Lekki and Laskowski (1971, 1974) it is possible
to compare the effect of varying collector and frother dosages on flotation
rate, when the flotation is carried out in a HalIimond tube (Fig. 17 and 18)
or in a laboratory flotation machine (Fig. 19). Note that to obtain flotation
rate curves with a Hallimond tube, the xanthate dose must be an order of
magnitude greater than in a froth flotation celI, and that at the high collector
concentration required in the Hallimond tube the phenomenon of "over-
oiling" does not appear, although the interaction between the frother and
colIector on the recovery at constant fIotation time is obvious. But the same
experiment carried out in a laboratory machine produces iso-recovery curves
that very clearIy refIect the practical experience in milIs: i.e., that over feeding
either the colIector or the frother results in a rise and fall in the overalI metal
recovery. Because of this hill-Iike recovery response in fIotation, milI men
intuitively look for a reagent system where the peak of the hilI is fIat, to avoid
100 r-------------,
80
1.': ....s
lO
"
:>
8
60
~
~ "o
(,)
~
"e
" 40
~ -o
al
.!::
x llo
2;0 So
r"
II
O 20 40 60 00 100
Concentna t ion of IX tel'pineol EtX a.la it ion, l<ig/t
"!l/lite!'
Fig. 17.- Effect of frother Fig. 18.- Isorecovery curves varying

concentration on chalcocite both frother and colIector conc.
recovery in the presence of afier 1 min fIoat in a Hallimond
700 g/t ethyl xanthate tube
too sensinve a control requirernent on rcagent dosage as a function of

changes in the minera! fed to the rnill, An interesting complement to the
alcohol carbon chain versus Iroth volume data, shown in Fig. 20, is sorne
recent work by Klimpel and Hansen (1987), where they have shown that the
reaction product of a six carbon alcohol with a specified amount of propylene
oxide gives an unusually strong frother with an enhanced ability to float
coarse partic1es. A typical example on coal is illustrated in Fig. 21, and on
copper in Fig, 20. There is no question that sorne of the effect is due to
inherently enhanced collecting properties of the frother itself, in the case of
coal, and better frother-collector compatibility, in the case of minerals.
\ 105.
ttI
.....
o
o
...Po.
::
s..
o
+'
I
~
30 SO 70 90 110
EtX a.ddition, g/t
Fig. 19.- Iso-recovery curves for chalcocite as in Fig. 18

but using a batch flotation machine.
Besides the increase in coarse partic1e recovery, demonstrated by the alcohol-

propylene oxide adducts over the same alcohol alone, they also noted that
increasing the branching of the alcohol backbone gradually decreases the
effective particle size range floated by the six-carbon propylene oxide adduct.
Fig. 21 also shows the extreme of this logic: a highly branched triol (glycerol),
reacted with propylene oxide, actually is a more effective fine particle frother
than any of the alcohols alone. It is also i~portant to note in the work of
Klimpel and Hansen (1987) that each specific frother chemistry has a unique
particle size range; the benefits of blending frothers for improved overall
valuable recovery is obvious.
Eff@ct of Co Llec eor- Dos:et O!l1 Se lec hu i t!;J

100,-----------....,
" .
r& oO~
'~l
t'
Frotller DF-Z50 15 g/t
..m!I_----li'd
.;; :P 90 15 ym-----..-.-- 20 25
t-:- 60 '"
l
,',
'~
~ 80 ,1/
@ 40 MIBC DF250 r. ~'
glt glt ;!l. 70 " 10
-
30
/lO
/lO
'"
<'3 60 /
30 ~~ ~.'5
,
o 2 4 6 8 10 50
O 5 10 15 20 25
Rougher Bank Cell number
% Pyrite in CDllcelltrate
Eifellt of' Frother dosage Effect of Collector dosage and

and Hn~ dose on Selelltivity Particle size on copper Recovel:'Y
25 100
---0"
,jj
\ KA~ <lO gft 14 X 44 ""
--A A
l....J 20 ){RX 5 g/t \
\ ---~---~.....
~
8,\
, '<,
\
b,
S SO
149 X 14"" - - -
~----"'~-'~O
J
149lt--............
-
X
'J "
, _. J.
U
15 '\
'~.
''1
\ 70
...
.S 10
f.<
DF-t01a

\
"\
\ \
'o
a <44 JIM
~
~
o
40 g/t O".
30 g/t
~"
.-.........~ 8 60,
/ / 297
...
210 1lM------
8'
l(
5 6~
o 209/t 50, /
&'~
10 g/t
0,+--..--.--......---.-- /
40.......--.-~~~~--,-~~-
SO 60 70 eo 90 1 00 20 30 40 50 60 70 SO 90100
% Copper ReCOl'tlry
Colleetm- dosages, g/t
Fig. 20.- Rate of flotation of copper ores
Probably the most extensive combined laboratory and plant scale testing of
the influence of frother chemical structure and frother dosage on rate of
recovery is also the work of Klimpel and Hansen. As might beexpected, the
rate of industrial mineral flotation was found to be significantly slower than
that associated with non-mineral systems, regardless of the frother type used.
They emphasised the interaction of frother structure and dose with particle
size, and a number of industrially important trends were observed. The first
is the rate influence of increasing the molecular weight of polypropylene
glycol frothers (such as the Dowfroth frothers). A detailed study was
Properties 01' Flotation Froths 91
nle Pel'fol"~ance of var-ous fll'otl1ell's a t 1.9 lh/t Soltll'ol coUectOlc, en tlle

ll'eCOlJel'Y oC uall'ious sized coallasll pall'ticl!!!> as , functin of Fl'otllGlJ'
dO!;;age tRa tes Of coat l'ecou!!l'y indica tO!d b~' assocta ted nlmhelM;) _ 179 loesh
- 171:1 Mesh
r">.
r"--\ 3 x 1710 /O 12
UJ '1.9 9.9 1.0
I~
17
0.B 3 }{ 170 M 10
~ 5.7 4.7
a.a
~
27
~5M 14
23
... 3:> ..
~ 9.6
... 35 M 25 "-...,
~ 6.6 211
40
1:1.4
11 ~.9
29
a.a llil 9.2
0.0
1l h -~~~~-,---l.J...L.l.LL~-'--.a...l.II.."""""""'LL..R-.L..a~~~
n
l-OLJ_c.a..LR.l.- n1 lU
a.i 0.4 0.1 '1.4 '1.1 '1.4 0.1 0.4 '1.1 '1.4 '1.1 '1.4 0.1 11.4 0.1 0.4 a.i 0.4 l
0.2 11.2 11.2 11.2 11.2 11.2 0.2 0.2 0.2
MIDC MIDC MIDe DF-4BB DF-4BB DF-4BB DF-1B12 DF-HH2 DF-1912
FROIHER, poundsfton
Fig. 21.- Performance 01' frothers on recovery 01' coal/ash

particles as a function 01' frother dosage
conducted on compounds having the underlying structure
where 'n' was varied gradually so as to give molecular weights from 200 to
400. Increasing the value 01' 'n' clearly increased the rate 01' recovery 01'
minerals,coals and potash. However, almost always, this increase in rate 01'
recovery was accompanied by a decrease in selectivity. In mineral systems,
the alcohol frothers were almost always slower in rate than the polypropylene
glycol frothers, over all mineral particle sizes, especially the coarser sizes.
The alcohol frothers were 01'ten more selective, especially with the fmer
particles. In coal flotation, the alcohol frothers were found to have excellent
rate character, but almost always the rate associated with the alcohols was
much more sensitive to dosage than the rate associated with propylene glycol
frothers. A second important trend obsered is the strong influence 01'
frother dosage with particle size, as illustrated in Fig. 21. This is an
illustration on coal 01' the R/K trade-off that is frother dosage driven, rather
than collector dosage driven. The three frothers shown are among the three
most common coal frothers in commercial use today: MIBC, Dowfroth 400,
311d Dowfroth 1012. In the figure, the recoveries, as a Iunction of particle

size intervals, are shown in the appropriate columns labelled with the rare, K,
for each recovery indicated near that column. The peor coarse particle
recovery of MIBC is shown along with its superior fine particle selectivity.
The polypropylene glycol frothers are clearly superior coarse particle
frothers, The influence of the R/K trade-off with collector dosage is also
evident, with the rate slowing or even decreasing at higher dosages. As this
example implies, from the wide variation of rate associated with changing
frother dosage, the use of frother dosage change is a key control variable in
industrial f1otation operations. These data also show why it is common in
plant practice to first use frother dosage to "fix" short-terrn plant flotation
problerns, followed by changing collector dosage only if the frother
adjustment is not sufficient.
Over oiling
Seventy years ago, when frothers with collector properties, such as cresylic
acid, were commonly used as the sole flotation reagent, it was known that as
the dosage of the flotation "oil" was increased, mineral recovery first
increased and then decreased, This depression with excessive collector or
frother was known as "over oiling". As soon as a separate collector was
introduced, it was noted that the same behaviour was exhibited independentIy
by the collector and the frother; i.e., increasing the dosage of either the
collector or the frother results in a rise and then a drop in metal recovery.
Curiously, there is very Iittle published on this effect, and it seems to be
practically unknown among the researchers on flotation.
Klimpel (1987) has studied this over oiling extensively in industrial scale
sulphide and coal flotation processes, and has called the effect "the R/K
trade-off", and attributes the changes in recovery to changes in the rate of
flotation. This terminology is taken from the observation that as collector
dosage is increased from starvation dosages, both the equilibrium recovery,
R, and the rate, K, at which this recovery occurs, increase. However, as
collector dosage is further increased, the value of R continues to increase, but
the rate of flotation, K, passes through a maximum. Excess frother dosage
can also lead to the R/K trade-off, but the effect in plant environments is
most often due to excessive collector. If the rate is slowed sufficiently, the
associated plant capacity may be insufficient to correct for the time delay in
levitation, and the observed plant recovery actually falls at "excessive"
collector dosage.
Fig. 22 is one of the many examples which illustrates the R/K trade-off on a
copper mineral circuit. The K values indicated are the result of
mathcmatically fitting laboratory time-recovery profiles, Klirnpel has also

shown that the effect is very prevalent in many industria! plants, and that
industrial circuit performance can be mathernatically predicted with good
accuracy by simply using appropriate batch laboratory rate data and realistic
residence time distributions for the industrial cell.
The depressve effect of excess collector has been used ndustrally. In one
US mill in which a copper/moly separation was operated, it was not
uncommon to have problems depressing the copper with certain changes in
the feed ores. When this occurred, rather than increasing the dosage of the
depressant, additional copper collector was added to the moly circuit, and the
copper was successfully depressed at a lower cost.
LO
o Sy"teM 1 R = 0.80 J( = 3.0
o Systeh\ a R =0.711 K =5.0
CUh\ulative
Fractional
~-::o:==f~
-01-_ 1 - 1 " -
Necovery Nate or 1 Equilibriu..
K control or N control
"'-1-'"
1
tH
Lal> Ti..e <..inu tes>
Fig. 22.- Illustraton of two laboratory time-recovery profiles

run under different conditions
\
Chapter 7
CHEMICAl PROPERTIES OF FROTHERS
Reagents commonly employed as frothing agents can be divided into slightly

soluble in water, and those that are completely miscible, such as polyethers
and polyglycol ethers, which correspond to the more modern frothers. Table
3 (page 8) summarises the frother types and their chemical structures, and
Table 55 (page 177) the use in the US flotation market.
Partially soluble frothers
Aliphatic alcohols: a series of different mixtures of C to es

alcohols were
marketed, at one time, by du Pont and still are by Cyanamid, as tailored
frothers for different types of ores. The largest volume alcoholic frother
used today is methylisobutyl carbinol (MIBC or 4-Methyl-2-pentanol).
Sorne properties of alcohols used as frothers are shown in Table 24. Note
the generally low solubility, which resuIted in the assumption that all effective
frothers should be nearly insoluble. The synthetic glycol polymers reversed
this premise.
TabLe 24.- PROPERTIES OF ALCOHOLS USED AS FROTHERS
ALcohoL FormuLa Frz. Pt. B.Pt. Density SoLub.

Oc oC g/mL g/L
n-PentanoL CH3(CH2)3CH20H -78.5 137.3 0.8144 23.0

IsoamyL ALe. (CH3)2CHCH2CH20H -117.0 132 0.813 25.0
HexanoL CH3(CH2)4CH20H -52.0 156.5 0.819 6.0
HeptanoL CH3(CH2)5CH20H -34.0 176 0.822 1.8
MIBC (CH3)2CHCH~CHOHCH3 -90.0 132 0.808 17.0
CapryL ie ALe. CH3(CH2 5CHOHCH3 -38.6 179 0.822 12.0
4-HeptanoL CH3(CH2)2CHOH(CH2)2CH3 -41.2 161 0.8183 4.5
- 94-
Chemical properties of frothers 95
The aleoholic frother mixtures include:
Mixed C6 to C 9 aleohols: sp. gr. 0.856, viscosity 5 cps, solubility in water, low;
flash pt. 552C. A more selective frother than MIBC.
Mixed C4 to C 7 alcohols + hydrocarbon oil: amber liquid; sp.gr. 0.82;

viscosity 4.5 cps; solubility in water 5 gil; flash pt. 44C; used mainly on
copper Zmoly ores. Designed for copper-molybdenite ores, also
recommended for flotation of minerals such as tale, graphite, sulphur and
coa!. Mixed C4 to C 7 aleohols produce a livelier froth than MIBC.
Mixed Cs to Cs aleohols: sp.gr. 0.81-0.83; viscosity 6.9 cps; solubility in water

10 gil; boiling range 135-190oC; flash pt. 55C; produces a heavier froth
than AF 73 and AF 76, but less persistent than polypropylene glycols, pine
oils and cresylic acids.
Higher alcohols (still bottoms): strong froth, less persistent than

polypropylene glycols, used in coal flotation.
Natural oi/s: in the early days of flotation, eucalyptus oil was a popular
frother cum colIector (because the first industrial scale flotation milIs were in
Australia). Later, pine oil became the dominant natural oil frother, because
of its greater availability in the western world. The quality standard was set
by the Yarmour trade-named oils distilled by the Hercules Corp. (U.S.A.).
Essentially all the pine growing countries (Canada, Finland, China, etc.) have
supplied different grades of pine oils at one time or another, but from the
flotation point of view the problems that mill operators have faced is the
difficu1ty in specifying a particular grade and obtaining the same frother
performance from different suppliers. This problem, and the fact that the
natural oils all have colIector properties, have made them less popular with
modern mill operators who now favour MIBC and synthetic non-colIecting
soluble frothers, Those using pine oils remark on its smalI-bubble, closely
knit, froths that do not alIow excessive particle fallout, but break down readily
in the launder. The close-knit pine oil froth favours recovery but lowers
concentrate grade. Increasing the amount of pine oil fed to the cells tends to
flatten the froth, decrease its volume, and cause surface spatter. An
emergency supply of pine oil can be very useful at rnills with ores which tend
to have periods of over- frothing. It can also be used in blends with MIBC to
regulate the characteristics of the froth and ~ompensate for changes in ores
which may affect selectivity and rougher recovery.
A typical assay of a commerciaJ pine oil is:
AIpha Terpineol 60-70%

Tertiary alcohols 10
Borneo] and Ienchyl aleohol 10-15
Camphor 10-15
FellCllyl Alcohol
Typical properties are: sp.gr. 0.927 - 0.940, distillation range 200 - 235C,
flash point 172C (C.c.).
The eucalyptus oils have greater collector action because they are richer in
ketones than the pine oils, which contain primarily camphor, while eucalyptus
oil ketones are cineoles and piperitones rather than camphor, and borneols
and piperoteols, as well as terpenes. Typical eucalyptus oils available today
have a specific gravity of 0.921- 0.923 (25C), distillation range 174 -177C.
~3 ~3
lI.e/t~~1I
I '" e-e-ell] I
IIze/t~=o
I'~ e-e-CII.l I
IIze I alz IIze I CHi!:
""-- ca . . / ' " eH /
1,8 - Cineole Boroeol CaMPho~
TabLe 25.- PROPERTIES OF TERPENES ANO KETONES USEO AS FROTHERS
BoiLing Point oC Freeze Refrae~( lnd Oensity

760 mm 100 mm Pt. -c nO g/mL 20C
ALeohoLs
a-TerpineoL 219 149 36 1.4831 0.9336
BorneoL 212 210 1.01
FenehyL aLe. 201 133 39 1.4734 0.935
Ketones
Piperitone 233/5 1.4845 0.9324
Camphor 209 sb 179 1.000
Fenehone 193 122 6 1.41619 0.9452
Ethers
1,8-CineoLe 174 108 1.4574 0.9245
Chemical properties of Irothers 97
Typical specifications for pine oil to be used as a Irother are: a clcar, yellow
oil, with a typical smell, that boils between 170 and 228C, with density of
0.915 to 0.935 g/m!. At least 44% by weight of the composition should be
alcohols (expressed as terpineols). At least 78% must c1istils between 170 ano
220C. Its average solubility in water should be approximately 2.5 g/!.
Eucalyptus oil used in flotation has similar specifications, as Iar as alcohol
content is concerned. Its main odorous component is the ether 1,8-cineol,
rather than the ketonc, camphor. Its solubility in water is typically 1.4 g/l.
Cresylic acids: the same quality control problems as for pine oils apply to the
cresylic acids, as these proclucts are also distilled from impure starting
materials (coal tars), and in aclclition have strong collecting properties. Their
popularity declined when acid circuit mills went out of fashion, though they
are still popular in coal flotation because of their availability from coking
ovens, The froths produced by cresylic acids are generally similar in
structure to those of pine oil, but of a sornewhat larger bubble size. An
increase in frother dosage decreases froth volume and produces a tendency
to effervesce. The lower boiling fractions produce a more fragile froth and
are somewhat more selective.
Typical specifications for flotation grade cresylic acids were: distillation range
of 190 - 23SoC, with no more than 2% below 190C, no less than 7S%
distilling below 23So; density, 1.01 to 1.04 g/mI and solubility in water, 1.7
g/l. According to the Condensed Chemical Dictionary, today's commercial
grade has not less than SO% distilling over 204C. If it boils beIow 204 it is
called creso!. A typical commercial cut boils between 220 - 2S00C, and has as
typical compositions: meta, para-cresols 0-1%; 2,4- and 2,S-xylenols 0-3%;
2,3- and 3,S-xylenols 10-20%; 3,4-xylenols 20-30%, and C; phenols SO-60%.
Sorne properties of the individual isomers are:
Sp. gravo meLt pt Oc boi L pt oC

Meta cresoL 1.034 12 203 Liquid
Ortho cresoL 1.047 30.9 191 soLid
Para cresoL 1.039 35.3 202 soLid
Froth characteristics and collector properties of cresylic acids are very

dependent on the meta cresol and phenol content, so specifications of cresylic
acid purchases should be rigorous if froth control problems are to be avoided
in the concentrator. The purer cresylic acids are reputed to give a more
fragile froth and to require a stronger collector for good flotation. Impure
cresylic acids contain pyridines and sulphur impurities, which were much
sought after SO to 70 years ago in flotation milis when the frother was also
employed as a collector, but in conjunction with selective collectors these

cresylic acids can give process control problems.
Alkoxy paraffins: these structures were developed in South Africa by Dr.

Powell (1951). The best known products are TEB (triethoxybutane) and the
Powell accelerator, so named because it has a strong effect on the flotation
rateo They never have had extensive application outside South Africa, except
in Chile (El Teniente). The froths these reagents produce are similar to
those of pine oil, except that over-feeding does not affect the froth character
as mucho The structure of TEB is:
CH -CHCH CH ........ OCHZCH 3
3 I "'OCHZCH 3
OCHZCH 3
The Powell Accelerator has the following structure:

H OH
I I
C-CHZCHCH 3
O/ "O
I I
CH3CHCHz-C ~/CH-CHZCHC~
I I I
OH H OH
This is a more powerful frother than TEB, but otherwise has similar
characteristics: low water solubility, good stability, and a strong effect on
flotation rateo
Non-ionic polypropylene glycol ethers: The Dow Chemical company offers

Dowfroth 4082, and Union Carbide R 17 and R 48, which are partially
soluble. Their structure is:
CH3
I
CH3(O-CH-CH2) -o-c- CH3
I n I
CH3 CH3
Completely water miscible frothers
Polyglycol ethers: these frothers, which are completely miscible in water, were
first developed by Tveter (1952), at Dow Chemical Co., and Booth (1954) at
the American Cyanamid laboratories. Until they came on the market, mili
metallurgists assumed that good frothers could only be very slightly soluble in
water. These glycol ethers are produced as by-products of synthetic brake
fluids. They are marketed by the Dow Chemical Company under the name
Dowfroths, by Union Carbide as Ucon frothers, by Cyanamid as Aerofroths
and ICI as Teefroths. With MIBC they account for nearly90% of all frother
use in metallic ore flotation. The polypropylene glycols produce compact,
Chernical properties of frothers 99
lasting froth structures that break clown readily in the launders. Unlike the
froths Irom slightly soluble frothers, these frothers produce more tightly knit,
more selective froths, that do not spatter on over-feeding the reagent. The
Dowfroths are designated by numbers that are proportional to the polymer's
molecular weight. Dow informs its customers that the higher molecular
weight products are more powerful frothers; the strongest are Dowfroth 400
and 1400. Because of the high activity of the polyglycol ethers, al! can cause
persistent froths which can cause operating problems if over-dosed.
TabLe 26.- PROPERTIES OF WATER SOLUBLE FROTHERS
Dowfroth Union Carbide

200 250 1263 1012 1400 Ucon 200 UconR55
MoL. Wt. 206 250 348 400 400 400

Viscosity
cps 25C 7 12 19 27 33 75
Density
g/mL, 25c 0.97 0.98 0.978 0.988 1.007 1.009 0.9735
B.P. Start
Oc 760 mm 243 252 >250 293
Freeze Pt
oC < -50 -50 -41 -50 -45 -50 -50
pH 7.2 7.2 7.0
SoLubiLity totaL totaL totaL partiaL totaL totaL totaL
Chemi ca L
formuLa n= 3 4 5 6 6
Cyanamid
AF 65 AF 67 AF 68 AF 70 AF 73, 76, 77HP
MoL. Wt. 400 similar simiLar see mixed C4 to C8
Viscosity to to MIBC aLcohoL
cps 25C 71 DF-250 DF-1012 frothers
Density
g/mL, 25c 1.007
B.P. Start
Oc 760 mm 180
Freeze Pt
Oc < -50
pH
SoLubiL ity totaL
The Dowfroths are methoxy polypropylene glycols (polypropylene glycol

methyl ethers), with the general formula:
wuh Dowroth 200, 250 and 1012 having molecular weights 01' approximately
206, 250 and 400. Dowfroth 1400 is a diol of 400 molecular weight, with the
formula:
The American Cyanamid products (Booth, 1954) are condensations 01'

propylene oxide with propylene glycol. The formula for Aerofroth 65 is
approximately:
Union Carbide supplies similar products based on polypropylene glycols and

polyethylene glycols, undel' the trade names PPG and Ucon.
Polyglycol glycerol ethers: One 01' the two new families now being produced by
the Dow Chemical Company is 01' the general formula:
R-O-(CH-CH2-0H)
I n
CH3
where R is from 4 to 6 carbon atoms, The characteristics 01' such frothers

were described earlier in this section. An example 01' the second family is the
reaction product 01' glycerol with propylene oxide to a 450 molecular weight.
The products available in industrial quantities have the fol1owing properties:
Type XK 35004.00L XK 35004.01L XK 35004.02L
Avg. MoLecuLar Wt. 450 250 700

Viscosity, 25C, cks 510 600 220
Sp. gr. at 25C 1.05 1.09 1.03
FLash pt. (PMCC) -c +205 199 260
SoLubiLity in water miscibLe mise. mise.
XK 35004.1L and .02L are weak frothers that can be blended with other
polyglycolethers to fine tune froth properties. They are stable in the range
pH 3.5 to 12.3, and usual1yemployed in alkaline flotation.
Chapter 8
FlOTATION MOD~FIERS
The classical treatment of this subject is given by Sutherland and Wark

(1955) in their book "Principles of Flotation". By very elegant experiments
based on contact angle measurements, they were able to show that flotation
could occur only below a critical pH for a given mineral and reagent
combination and dosage. Based on this concept, a rational explanation for
the mechanism of activation and depression was developed which allows a
systematic approach to the subject. Their data on critical flotation pH for
potassium ethyl xanthate is shown at three temperatures in Table 27. In
Table 28 similar data is shown for ethyl dithiophosphate and xanthate, and
for iso-amyl xanthate for galena, pyrite and chalcopyrite.
TabLe 27.- CRlTICAL pH FOR FLOTATION
MineraL 10 Room Temp 35

Pyrrhotite 6.0
Arsenopyrite 8.4
GaLena 10.8 10.4 9.7
Pyrite 10.2 10.5 10.0
Marcasite 11.0
ChaLcopyrite 13.0 11.8 10.8
CoveLL ite 13.2
Activated SphaLerite 13.3
Bornite 13.8
ChaLcocite >14.0
If these type of data are to be used, the actualcurves published by Sutherland

and Wark should be consulted, as the critical pH increases quite sharply with
concentration for most collectors at the values'quoted aboye.
The effect of a depressant, such as NaCN, is to significantly reduce the

critical pH for flotation. Data reported by Sutherland and Wark (1955) for
the addition of 20 mg/I of NaCN to a pulp containing 25 mg/l of potassium
ethyl xanthate are shown in Table 29.
101
TabLe 28.- EFFECT OF COLLECTOR ON CRITICAL pH FOR FLOTATION
CoL Lector GaLena Pyrite ChaLcopyrite
Na diethyL dithio- 6.2 8.5 9.4

phosphate, 32.5 mg/L
Potassium ethyL 10.4 10.5 11.8

xanthate, 25 mg/L
Potassium iso-amyL 12.1 12.3 >13

xanthate, 31.6 mg/L
Sutherland and Wark (1955) provide a variety of graphs on the effect of the
cyanide ion on the critical pH, and also show that the effect of the cyanide
ion can be calculated from the ionisation constant for the different cyanides
as a 450 straight line on log-log paper.
TabLe 29.- EFFECT OF SODIUM CYANIDE ON CRITICAL pH
Temperature 10C 35C

ChaLcocite 12.8 10.7
with 20 mg/L NaCN 8.4 7.3
Pyrite 10.6 10.0
with 20 mg/L NaCN 7.1 6.9
GaLena 10.9 9.6
with 20 mg/L 10.9 9.6
Activating agents
These are chemicals that are used to permit flotation of a mineral that is
difficult or impossible to float with only the use of a collector and a frother.
Copper sulphate is the best example of an activator. It was first introduced by

Bradford in 1913 in Australia in zinc bIen de flotation, and later was applied
to the activation of gold bearing pyrite in that country. It is universallyused
in the flotation of sphalerite, which is otherwise impossible to float. It is also
used to reactivate minerals which have been depressed with cyanide, such as
chalcopyrite, pyrite, pyrrhotite and arsenopyrite. Thus, alternate use of
copper sulphate and cyanides can be used very effectively in differential
flotation. In mill use it should be remembered that the salt is quite corrosive
and is best handled in plastic equipment.
Flotation Modifiers 103
Lead nitrate or acetate are used to actvate stibnite and to reactivare copper
sulphides depressed with cyanide, They are sometimes employed to improve
recovery of tarnished gold. They are also activators for silicates and
carbonates in soap flotation, and Ior sodium chloride in fatty acid flotation to
separare potash from halides.
Sodium sulphide can act as an activator at low concentration for oxidised

and tarnished ores. Morenci has found that the ammonium salr is an even
more effective activator for their tarnished copper ore. Controlling the
effective concentration of the sulphhydrate ion is extremely tricky, as
sulphides are so easy to oxidise. Because of this, the few industrial
applications of sulphidisation have employed a circuit in which the floatable
sulphide minerals are taken off first, prior to the addition of sodium sulphide
or NaSH, to float the remaining oxide minerals. The choice of sulphide or
NaSH depends on the pH required in flotation, as the sulphide contributes
twice as much alkalinity as NaSH.
Hydrogen sulphide has been used in Butte to precipitate copper in solution

and allow recovery by flotation. It will depress gold and silver, and Cu-Fe
minerals in moly separations.
pH modifiers such as lime, soda ash, caustic soda, sulphuric acid, etc., can
also be considered activators because for all mineral/reagent combinations
there is a critical pH aboye which flotation will occur. Because theflotation
process is a surface phenomena that is extremely sensitive to the ion content
of the flotation water, the use of soda ash (caustic) versus lime as a pH
controller depends on whether the calcium cation will form insoluble
precipitates with naturally occurring anions in the pulpo
Generally, because of cost and availability, lime is the chosen pH modifier

except in the flotation of pyrite, particularly gold bearing pyrite, and to a
lesser extent in galena flotation, where soda ash is the first choice.
Inorganic depressants
These assist in the separation of minerals when the floatability of two or

more minerals are too similar for a particular. collector to effect a separation.
Besides lime, the other most used depressant is the cyanide ion.
\
SodiW71 cyanide is a strong depressant for iron sulphides: pyrite, pyrrhotite,

marcasite and arsenopyrite, and also for sphalerite. It will act as a less
effective depressant for chalcopyrite, enargite, tennantite, bornite and most
other sulphide minerals, with the possible exception of galena.
104 FLOTATIOl\J - Theory, Reagents and Ore Testing
Lune S used for depressing pyrite and other iron sulphides, galena,
marrnatitic zinc and sorne copper minerals, as well as cobalt sulphides, It will
depress gold in xanthate Ilotations, and interferes with sulphidisation
processes. For inorganic rninerals, it wiJI retard recovery of silicates when
cationic collectors are involved.
Zinc sulphate is used in conjunction with cyanide, or alone, for the depression
of sphalerite while floating 1ead and copper minerals.
Chromic acid and dichronuues will depress galena m copper-lead-zinc

separations.
Permanganates are used for the selective depression of pyrrhotite and

arsenopyrite in the presence of pyrite. It will depress sphalerite, and is also
used in copper/moly separation.
Sodium silicate is a complicated reagent because of its variable composition.

It is used as a depressant of quartz and other silicate minerals. Also used in
coagulation of slimes and as a modifiers in fine particle flotation. Generally
helpful if maintaining concentrate grade is a problem.
Sodium hydroxide is used for the depression of stibnite. It is helpful in gold

flotation, and will peptise slimes. It is a saponifier for fatty acids in inorganic
flotation, and is used in scrubbing ores and in scheelite flotation.
Sodium and calcium sulphites and Izyposulplzites and sodium hydrosulphite

are used in the flotation of lead and copper minerals while depressing pyrite
and sphalerite.
Ferrocyanide is used in the depression of copper sulphides in copper/moly

separation and in the separation of sorne copper sulphides from sphalerite.
Ferro and [erricyanides will help separate cobalt and nickel sulphides from
copper sulphides.
Sulphuric acid is used for quartz depression. It will reactivate pyrite

depressed with lime, and is used in gold flotation to clean iron salts that
interfere with gold recovery.
Hydrofluoric acid is used in the depression of quartz in the flotation of

feldspar with cationic collectors.
Sulphur dioxide, usually in conjunction with causticised starch, IS used to

Flotation Modifiers 105
depress galena Irom copper sulphides,
Organic depiressants
Organic depressants are large molecules usual1y with a molecular weight

aboye 10,000. The mechanism of their depressant action is not clear, though
Lovell (1982) thought that these are depressants because they contain a large
number of hydrated polar groups. The natural products are generally
polysaccharides, while new synthetic depressants inelude polyglycol ethers
and polyphenols. A curious extreme depressant is a xanthate prepared by
reacting sugar with caustic and carbon disulphide; it has not found a
practical use as it depresses everything.
The natural depressants inelude the fol1owing:
Quebracho and tannic acid are polyphenols used for the depression of calcite
and dolomite in fatty acid flotation of fluorite or scheelite and also pyrite.
The formula for polyphenols can be represented:
Structu.... of Tannin :Lnd Quebr;>cho
Starch and glues, either unmodified, as extracted from maize potatoes, or

partially hydrolysed to produce the more soluble dextrins, are used for the
depression of mica, tale and sulphur in sulphide flotation. They also act as
flocculants, particularly with haematite, and as elay dispersants. Starch is a
highly polymeric carbohydrate built up of glucose units with molecular
weights ranging from 50,000 to several million daltons. Corn starch consists
of two types of polymers: a linear amylose composed of 200 to 1,000 pyranose
rings bonded together through oxygen atoms in positions 1 and 4, and a
branched amylopectin with 1,500 or more pyranose rings with periodic
branching at positions 1 and 6.
..01,200
./h"--o
.-J./ ""'~ o~
'" Ofl 1+0 /11
~
H H
Glucose units
Corn starch is more strongly adsorbed on haernatite and geothite than on

quartz; its adsorption decreases with increasing pH. Adsorption on quartz is
minimal in the absence of calcium ions. In highly alkaline solutions, caleium
ion is adsorbed on quartz as a hydroxy complex, CaOH+, which is thought
to be responsible for the effectiveness of caleium ions in anionic silica
flotation.
Natural gll1llS, such as gum arabic, gum traganth, and guar, have strong
flocculent properties, but can be treated to be selective depressants for tale
and siliceous materials, Their che mical structure is:
-o-
Chapter 9
SULPHIDES - DEPRESSION AND ACTIVATION
Sodium sulphide
The sulphhydrate ion has been known as an activator of oxidised minerals

since the sulphidising process was patented [A.Schwartz (1905), US Patent
No. 807501], practically at the same time as the birth of flotation, and has
frustrated metallurgists ever since. Non-predictable performance is no
surprise if one considers that small concentration of HS- will depress the
flotation of metallic sulphides, and not activate carbonates or oxides; a
medium amount will erratically activate the flotation of non-sulphide copper
minerals such as ma1achite [CuC0 3.Cu(OH)z], azurite [(CuC0 3)z.Cu(OH)z],
and chrysocolla [Cu(Si401O)(OH)12.8H20]; more reliably it has be en used to
activate lead oxides, such as cerrusite [PbC0 3], and anglesite [PbS04], but
with generally high dose rates; a large amount will desorb collectors and
depress activated minera1s. The surface adsorption mechanism that provides
these contradictory qua1ities to sodium sulphide is still not completely c1ear.
In solution, it is very easily destroyed by dissolved oxygen, its activity is
strongly pH dependent, and it significantIy affects the system alkalinity, more
so if sodium sulphide is employed as the flotation modifier than if the reagent
is sodium sulphhydrate.
Historically, it was successfully employed in xanthate or thiophosphate

flotation of cerrusite using as much as 20 kg of sodium sulphide per tonne of
ore. As cerrusite characteristically has c1ay-like properties, and is a major
constituent of slimes, sulphidisation normally has been combined with the
use of sodium silicate as a peptising agent. Its main use as a depressant has
been to remove copper from molybdenite corcentrates during the c1eaning
step. When used as a rougher depressant for sulphides, it appears to operate
primarily due to its oxygen scavenging properties.
Chemical properties of sulphides
Sodium sulphide: Na2S; formula weight 78.04; ambient stable crysta1
107
108 FLOTATlON - Theory, Reagents and Ore Testing
Na2S.9H20, formula weight 240.18; commercially available form, Ilakes

which contain 60% by weight Na2S with an approximate density of 1.75.
Data available on the solubility of sodium sulphide is conflictive, due

primarily to the ease with which it is oxidised in solution in contact with air.
The best data based on a literature review is that reported by Seidell-Linke
(1965). Their interpolations and phase data are reproduced in Table 30.
TabLe 30.- SODIUM SULPHIDE SATURATION SOLUBILITY IN WATER AND PHASES IN

EQUILIBRIUM
tOe Density SoLubiLity SoL id phase
gm/mL gm/100g sat.soLn
o 11.0 Na2S.9H20
5 12.3 11
10 13.4
15 14.5
20 1.1682 16.5
25 17.5
30 1.1972 19.8
40 1.2287 23.4 Na2S.9H20
48.5 transition to Na2S.6H2o
50 27.6 Na2S.6H20
60 29.0 11
70 30.8
80 33.4
90 36.8
91.5 transition to Na2 S. 51;H20
VoLumetric soLubility, gm Na2S/Lt soLn
20 192.7
30 237.0
40 287.5
Ref: SeideLL-Linke (1965), page 1113
Ionisation as a [unction of pH: On dissolving Na2S or NaSH in water,

ionisation proceeds in various stages. First
Na2S + 2H2o 2Na+ + H2s + 20W or
NaSH + H20 ~ Na+ + H2S + OH- foL Lowed by
H2s H+ + HS-
where the equilibrium constant is .

(H+)(HS-)
9.1 x 10 -8 pK = 7.04
while the third stage is
and the equilibrium constant is

Sulphides - Dcpression ami activation 109
1.'] " 10-12; pI< = 11.96
These equilibrium constants indicate that, on a molar basis, below pH 5 there

are no sulphide ions present, and at pH 5 only 10% of the HzS has ionised to
HS-. At pH 7, half the HzS has ionised to HS-, while between pH 8 and 11,
over 90% of the sulphide ions are HS-. The next ionisation node is at pH 12
when the moles of HS- and S = are equal, while by pH 13, 90% of the
ionisation is to S=. For rough estimation of the ionic concentration, in
grams per litre, the curves in Fig. 23 can be multiplied by the NazS
concentration.
00
N
~
z: /-~
':s:! 9.4 I
H8
\
,,
z: .-l
I
- ...
o
1-<
~
~
:
~
~
o
0.3 \
\
S-
1-< ~
z: ~
~ :
u al
z: o
O : 0.2
u o
-- -......
u
z:
o
e
~
~ 0.1
-
:
0
4 6 8 10 12 14
pH
Fig.23. Ionisation of sodium sulphide solutions
Sodium hydrosulphide: NaSH, formula weight 56.06; ambient hydrate,

NaSH.2H2, formula weight 92.09; fused hydrate, NaSH.3H20 , formula
weight 110.1; melting point 22C.
Surprisingly, there is very little information available in the literature on the

solubility of NaSH in water. Seidell-Linke (1965) indicate that at 20C, in
water containing 2% NaCl, solubility of NaSH is 35.8 g/lOO g of saturated

solution. NaSH is available commercially as a solution containing 16%
sulphide; a hot solution containing 40 to 44% active NaSH; crystals (tri
hydrate) which are 50.9% active, and which on melting (at 22C) dissolve in
their own water of hydration; and flakes containing 70 - 72%, which melt at
55C.
When dissolving flake NaSH in water, caution is necessary, as the more

impure products can decompose with evolution of H 2S, a gas that is as toxic
as hydrogen cyanide (TLV for both is 10 ppm), and more dangerous because
it paralyses the sense of smell, the odour threshold being close to the fatal
concentration. It forms explosive mixtures in air in the range 4.3 to 46%
H 2S. "
The pH as a function of concentration is shown in Table 31. Fig. 24 shows

the effect on solution pH of the addition of equal weights of NaSH and Na2S,
As can be seen, at any given dose, the pH for NaSH is about a quarter unit
lower than for Na2S, The effect of temperature is shown in Table 32.
TabLe 31.- EFFECT OF PH ON THE EQUILIBRIUM VALUES OF SULPHIDE ION,

HYDROSULPHIDE ION AND HYDROGEN SULPHIDE
For a giL of Na2S the conc. of each ion is:

pH
H2S, giL HS-, giL S=, gl L
4.0 0.4363 3.8538 x 104 4.1085 x 1012

4.5 0.4354 1.2163 x 103 4.1005 x 1011
5.0 0.4328 3.8226 x 103 4.0000 x 1010
5.5 0.4245 1.1857 x 102 3.9972 x 109
6.0 0.4003 3.5356 x 102 3.7692 x 108
6.5 0.3391 9.4722 x 102 3.1933 x 107
7.0 0.2286 0.2019 x 102 2.1530 x 106
7.5 0.1126 0.3146 x 102 1.0605 x 105
8.0 0.0432 0.3819 x 102 4.0711 x 105
8.5 0.0147 0.4095 x 102 1.3806 x 104
9.0 0.0047 0.4188 x 102 4.4649 x 104
9.5 0.0015 0.4209 x 102 1.4191 x 103
10.0 4.7415 x 104 0.4188 x 102 4.4649 x 103
10.5 1.4660 x 104 0.4095 x 102 1.3805 x 102
11.0 4.3229 x 105 0.3818 x 102 0.0407 x 102
11.5 1.1259 x 105 0.3145 x 102 0.1060 x 102
12.0 2.2852 x 106 0.2018 x 102 0.2152 x 102
12.5 3.3885 x 107 0.0946 x 102 0.3191 x 102
Ammonium hydrosulphide: NH4SH, formula weight 51.11; melting point

118C; solubility at OOC is 128.1 g/lOO g H 20. The commercially available
Sulphides - Depression and activation 111
products are ammonium hydrosulphide crystals, and a 40 - 44% solution.

Ammonium sulphide is used more as a flotation activator, because the
ammonia ion appears to clean the mineral surface and enhance the
sulphidisation reaction.
lQQ
SIi\l
1 I
61i\l 20 g/l I
40
\ I
, I
39 19 g/l
29
5g/l \ ~ = NaSH
\ \
Mole/li t:re
29 g/l
19 \
x 19-5
8 I
6
f
\
4 1 \
n ! I
,I
1 5 gIl
3
2
\ I f
1
8 9 19 11 12 13
pH
Fig. 24.- Relative effect of the addition of NaSH and Na2S on solution pH
TabLe 32.- EFFECT OF TEMPERATURE ON THE PH OF SATURATEO

SOLUTIONS OF NASH ANO NA2S
soLution pH
Temperature 15C 20C
pure water 7.27 7.17 7.08
5 giL NaSH 10.48 10.29 10.12
5 giL Na2S 10.64 10.45 10.27
Molybdenum concentrate c1eaning
The most important application of HS- as a depressant is in the up-grading

of molybdenite concentrates obtained as a by-product of copper sulphide
flotation. It is the key reagent in the depression of chalcopyrite during the
------------------.,-------------~~-~ ...- .. -. __ .-
TabLe 33.- REAGENT PRACTICE IN MOLYBDENUM RECOVERY FROM PORPHYRY ORES, COPPER CIRCUIT
Heads %-200 pH Main - PRIMARY FLOTATION CIRCUIT - concentrate % SoLids
PLant & company % cu % Mo mesh MineraL CoLLectors Frothers % Cu % Mo Thickn Flt -'
-'
(--.)
United States
Bingham, Utah, Kennecott 0.7 0.025 60 8.5 Cpy/Cte Reco. F.O. Cres.Ac. ,MIBC 28 0.9 45 35
Ray, Arizona, Kennecott 0.9 0.018 60 11.5 Cpy Reco. F.O. MIBC(SheLL 10) 18 0.2 60 15
Chino, New Mexico, Kennecott 0.9 0.008 55 11.0 Cpy Reco. F.O. SheLL 10,P.O. 20 0.2 50 15
Mc GiLL, Nevada, Kennecott 0.9 0.016 52 10.8 Cpy Re404,Z-200 MIBC, P.O. 20 0.2 O 45
Sierrita, Arizona, DuvaL 0.3 0.03 47 11.0 Cpy Z-6,3302,F.O. MIBC 25 2.5 50 30 '"rj
MineraL Park, AZ, DuvaL 0.4 0.04 60 11.5 Cpy Re3302,Z200 MIBC 18 1.3 50 20 l'
Esperanza, Arizona, DuvaL 0.4 0.03 60 11.5 Cpy Re3302,Z-6 MIBC 22 1.5 50 20 O
MIBC(SheLL 10) 28 30
~
San ManueL, Arizona, Magma 0.7 0.018 60 10.5 Cpy/Cte Minerec,F.O. 0.5 55
Mission, Arizona, ASARCO 0.7 0.018 56 11.5 Cpy Re238,Z- MIBC,P.O. 25 0.4 60 35
SiLver BeLLs, AZ,ASARCO
Pima, Arizona, Cyprus
Bagdad, Arizona, Cyprus
Morenci, AZ, PheLps Dodge
0.7
0.5
0.5
0.8
0.008
0.013
0.015
0.007
48
50
50
55
11.0
11.5
11.5
10.5
Cpy/Cte
Cpy
Cpy/Cte
Cpy
Re238.Z-10
Z-6,F.O.
Z-4,Z-6
Z-200,Minerec
MIBC,Cr.Ac.
MIBC(SheL L 10)
MIBC,P.O.
Cres.Ac.
30
26
28
22
0.45
0.25
1.5
0.14
50
50
50
45
20
35
20
35
-
>-3
O
Z
Twin Buttes, Arizona, Anamax 0.6 0.03 53 11.0 Cpy Z-6 MIBC 28 1.0 50 35 >-]
Inspiration, AZ, Inspiration 0.7 0.007 55 10.5 Cpy Minerec 898 MIBC,AF-65 35 0.3 O 35 ::r'
Pinto VaLLey, Cities Service 0.5 0.011 35 10.5 Cpy/Cte Z-14,NaAerofL Dow250 26 0.2 30 15 Gl
o...,
Canada ::;<
Brenda, B.C., Noranda 0.2 0.05 40 8.0 Cpy Z-6 MIBC 22 3.0 65 18 ::<:i
Gasp, Quebec, Noranda 0.6 0.015 67 10.0 Cpy/Cte Re3302,Z-6 BHB 27 0.3 50 25 Gl
~
IsLand Copper, B.C.,Utah Int. 0.5 0.017 50 11.5 Cpy Z-6, Re238 MIBC 24 O. 65 35 q
Lornex, B.C., Rio ALgom 0.5 0.016 50 9.5 Cte/Cpy Z-6,Z-11 Dow250,P.O. 33 0.3 50 35 Gl
i:l
GibraLter, B.C., PLacer 0.4 0.01 70 10.5 Cte/Cpy IPX-343 MIBC,P.O. 30 0.5 55 30 ~
(/'
~
ChiLe ::;
Chuquicamata, CODELCO 2.0 0.06 46 11.0 Cte RE238,Z-11 Dow250,MIBC,PO 40 0.9 60 35 o...
EL Teniente, CODELCO 1.5 0.04 67 4.2 Cte/Cpy Minerec A Diwl012 42 0.5 - - O
...,
EL SaLvador, CODELCO 1.2 0.024 50 11.0 Cte Z-200,Z-11,F.O MIBC,P.O. 42 0.8 60 45 (1J
Andina, CODELCO 1.8 0.015 70 9.0 Cpy Minerec2030 MIBC,Dow250 28 0.25 55 40 >-]
(1l
Peru ~
ToquepaLa, Southern Peru 1.2 0.018 60 11.5 Cpy/Cte Z-11 Pine OiL 31 0.3 60 30 S'
q
COMECOM
BaLkhash, Kazakhstan, USSR 0.4 0.01 62 12.0 Cpy/Cte Z-11,Z-14,Kero Pine OiL 16 0.15 45 45
ALmaLyk, Uzbekstan, USSR 0.7 0.01 46 11.5 Cpy Z-7, Kero Pine OiL 17 0.2 60 24
Kadzharan, Armenia, USSR 1.2 0.05 60 9.0 Cpy Z-7. Kero Pine OiL 16 1.5 60 25
Medet, BuLgaria 0.3 0.008 60 8.5 Cpy Z-14, Kero Pine OiL 13 0.2 50 20
Table 34.- REAGENT PRACTICE IN MOLYBDENUM RECOVERY FROM PORPHYRY ORES, MOLY CIRCUIT
MolY/copper separation section practice Final CuFinal Molyoverall
concen- concen- Reco-
Feed Heat ----- Reagents used Number trate trate ver~
Plant & company pH Treatment Depressor Other Clners %CU %Mo %CU %CU l,f10
United States
Bingham, Utah, Kennecott 11.0 Roast Utah process NaCN,Nokes 6 30 0.1290 1.0 90 56
Ray, Arizona, Kennecott 11.5 - Na~Fe(CN)6 Nokes,Na2Zn(CN)4 9 20 0.0882 1.8 82 31
Chino, New Mexico, Kennecott 11.0Steam, roastNa H Nokes, 14 20 0.1 82 1.25 79 52
Mc Gill, Nevada, Kennecott 9.0 Steam Na~Fe(CN)6 NaCN 9 20 0.1560 1.5 78 15
Sierrita, Arizona, Duval 9.0 Steam No es - 8 26 0.2080 3.0 90 76
Mineral Park, AZ, Duval 7.0 Steam Na4Fe(CN)6 - 8 20 0.1090 0.3 76 62
Esperanza, Arizona, Duval 7.0 - Na4Fe(CN)6 Roast to Mo03 8 25 0.1283 1.5 87 74 (j)
San Manuel, Arizona, Magma 11.5Pres. Steam Na4Fe(CN)6 NaEZn(CN)4,H202 6 30 0.0685 1.0 92 70 e
Mission, Arizona, ASARCO 11.5 - Na4Fe(CN)6 Na N 10 28 0.0885 1.0 89 67 -S
Silver Bells, AZ,ASARCO 11.0 - Nas,NaSH - 8 30 0.1885 0.8 84 30 ::T
Pima, Arizona, Cyprus 11.0 Roast Na H,(NH4)2S Roast for talc 6 28 0.1042 1.8 83 33 P.:
(11
Bagdad, Arizona, Cyprus 8.5 - Nokes NaCN 12 30 0.1590 1.0 88 64 CI1
Morenci, AZ, Phelps Dodge 10.5 cook Na~Fe(CN)6 Na2Zn~CN)4,H202 8 26 0.0685 1.0 83 -
Twin Buttes, Arizona, Anamax 10.5 - Na H 8 29 0.2573 1.1 76 35 tJ
Inspiration, AZ, Inspiration 10.5autocLave Nokes - 7 38 0.1792 0.5 76 40 (1l
Pinto Valley, Cities Service 11.0 - Nokes - 7 27 0.1188 1.12 87 36 ...

'"'d
(1l
CI1
CI1
Canada
Brenda, B.C., Noranda 8.0 - NaSH NaCN 12 28 0.2092 0.1 88 81 o'
~
Gaspe, Qu:bec., Noranda 10.0 Steam NaSH Na2(CrOh) 20 30 0.0688 0.5 90 56 P>
Island Copper, B.C.,Utah Int 10.5 - NaSH Na2zn~c )4,NaCN 10 24 0.0675 1.0 88 67 i:l
p.
Lornex, B.C., Rio Algom 9.5 - Nokes 7 34 0.0683 1.0 88 64
Gibralter, B.C., Placer 10.5 - - - - 30 0.1580 1.0 85 45 :.>
P.
:~r
Chile :;.
Chuquicamata, CODELCO 9.6 - Anamol D NaCN 5 42 0.2090 0.1 90 56
o'
El Teniente, CODELCO - - Nokes NaCN 25 42 0.1595 0.5 83 42 ;:l
El Salvador, CODELCO 11.0 - Anamol D NaCN 7 46 0.1097 1.0 81 60
Andina, CODELCO 9.0 - NaSH - 8 28 0.0388 1.0 90 50
-------
Peru
Toquepala, Southern Peru 8.0 - Na4Fe(CN)6 Na(ClO4) 8 32 0.1287 1.2 88 37
COMECOM
Balkhash, Kazakhstan, USSR 11. O Steam Na2s - 7 23 0.0560 na 85 41
Almalyk, Uzbekstan, USSR 11.0 - Na2s - 6 17 0.0450 na 83 f.O
Kadzharan, Armenia, USSR 10.5 - Na2S - 6 16 0.2079 0.7 80 65
Medet, Bulgaria 8.0 Steam Na2S - 8 20 0.1041 1.8 82 26 i-'
~.J-)
114 FLTATlN - Theory, Reagents and Ore Testing
Index to abbreviations for TabLes 33 and 34

Z-4 Na ethyL xanthate
z-6 1< amyL xanthate
Z-7 1< butyL xanthate
Z-10 Na hexyL xanthate
Z-11 Na isopropyL xanthate
Z-14 Na isobutyL xanthate
I<ero kerosenes
F.O. fueL oiL
Z-200 ethyL isopropyL thionocarbamate

Miner 1661 11 11
Miner 2030 =
Miner 898 ethyL isobutyL xanthogen formate
Minerec A ethyL ethyL xanthogen formate
Reco Na dicresyL dithiophosphate
Re 404 mercaptobenzothiazoLe
Re 3302 aLLyL amyL xanthic ester
Creso ac. cresyLic acid

P.O. pine oiL
MIBC methyL isobutyL carbinoL
AF 65 poLygLycoL ether
Dow 250 poLypropyLene gLycoL ether
molybdenite c1eaning process, supplied from sodium sulphhydrate in the

western world and from sodium sulphide in Russia. Table 33 lists the
reagents employed in the seventies in moly/copper separations.
A very important factor in choosing the primary collector/frother system for

ores that contain molybdenite is the effect of the reagents on the subsequent
moly separation from the copper concentrate, as the maximum copper
allowed (0.5%), and molybdenum grade (50%) specified in the [mal moly
concentrate, is stringent. The flotation circuit used in moly recovery nearly
always is based on depressing the copper and other minerals and floating off
the molybdenite into a concentrate with less than 1% copper; if this is not
attained, the concentrate must be leached (usually with ferric and ferrous
chloride) to remove surplus copper. The copper depression processes
employed can be divided into chemical (as mentioned), and thermal, which
inc1ude roasting and steaming. Roasting will normally decompose the
collector on the mineral surface and vaporise the frother; it may also oxidise
the copper mineral surface to sorne extent and provide natural depression.
The concentrate, after thermal treatment, is then floated in up to 25 stages of
c1eaning, using fuel oil and MIBC as the flotation reagents.
Note that there are only six mills that do not use sorne form of sulphhydrate .
ion in the copper/molybdenite separation. Flotation recovery of
molybdenite from porphyry copper ores is governed primarily by copper
metallurgy, but is adjusted to accelerate and optimise molybdenum recovery
as well, particularly where rnolybdenum is an irnportant by-product or co-

product, such as at Sienita, Brenda or Chuquicamata. As can be seen Irom
the aboye tables, there is no consistent pattern in the relative efficiency of the
recovery of copper or molybdenum, except for the generally better recovery
from ores with a lower copper contento This improved recovery is probably
due to the greater attention paid to optimising molybdenum recovery and the
lesser crowding out effect of other sulphide minerals. The regional average
mine head assays, and corresponding overall recoveries, tend to confirm this
surmise, viz:
TabLe 35.- REGIONAL MINE PERFORMANCE
Region number of Heads Recoveries

mines % cu % Mo % cu % Mo
United States 16 0.66 0.02 83.9 50.3

Canada 5 0.42 0.021 87.8 62.6
ChiLe-Peru 5 1.62 0.031 86.4 49.8
COMECOM 4 0.67 0.02 84.8 43.0
TotaL average 30 0.78 0.022 84.8 50.8
(Note: these data are deliberately pre-1980 because they provide a more
comprehensive technology comparison)
The tables show .that in 67% of the plants producing molybdenite, some form
ofaxanthate is used as the collector. Half of them used it as a single
collector, with or without the addition of fuel oil as a moly promoter. Eight
plants used a dithiophosphate coupled with a xanthate and fuel oil; none
used only a thiophosphate. Five employed a thionocarbamate; only two of
thern did not add a second collector. Two mines - El Teniente and
Inspiration - used xanthogen formates. Final1y, five operations used a xanthic
ester (Cyanamid RE 3302), usuaIly combined with a xanthate, Xanthic esters
are the only family of collectors that show a statisticaIly significant correlation
with molybdenite recovery.
Reviewing frother practice, we find that 18 of the 30 operations employed

MIBC as the primary frother, 12 in combination with still bottoms (SheIl10),
which is a heavier alcoholic fraction. The next most popular frother was pine
oil, which is reputed to improve molybdenite (unconfrrmed statistical1y), and
which was used by 12 operations, In four cases it was combined with MIRC,
in 3 cases with a soluble frother (Dowfroth 250), and never with cresylic acid.
Cresylic acid was used in 3 mines, in two cases combined with MIBC.
Soluble frothers were used in 6 cases, of which only in one case - that of El
Teniente (Dowfroth 1012) - it was used as the sole frother.
116 FLOTATION - Thcory, Reagents and Ore Testing
An .analysis of all the data available for reagent practice in copper sulphide
concentrators (Crozier, 1978) showed no statistically significant correlation
between rnolybdenite recovery and either collector or frother selection, with
the cxception of a weak one with xanthic esters, despite the amount of
published opinions on improved moly recovery with fuel oil additions to the
collector suite, or the use of pine oil as a frother. The probable reason for
this contradiction is that in mills for which the data was analysed, the main
collector-frother combination was selected to enhance the recovery of the
most valuable mineral component - usually copper. The exceptions are
Sierrita in the U .S.A., with copper heads of only 0.3%, and Brenda in
Canada, with 0.2 Cu in the concentrator feed. In both cases the molybdenum
recovery is much higher than the norm (76% and 81%, respectively),
indicating that the circuit design optimised moly and copper recovery, It is
significant that in both mines a xanthic ester (Re 3302) is used as the
collector, and MIBC as the frother. Two mines in Canada and two in the
United States, which have low copper heads (0.4%) also have abnormally
high molybdenum recoveries (61%); again Re 3302 was used with a xanthate
as the colIector.
Copper depression in the molybdenite separation
In the recovery of molybdenite from a copper concentrate, it is the norm to

depress the copper and pyrite gangue, and float off the molybdenite, although
in very unusual cases the opposite has been opted for by the metallurgist.
The more important copper depressants employed in moly separation have
been:
1.- Sodium ferrocyanide and ferricyanide

2.- Sodium zinc cyanide
3.- Sodium cyanide
4.- Nokes reagent (P20S dissolved in NaOH)
5.- Arsenic Nokes, or Anamol D (AS 203 dissolved in Na2S)
6.- Sodium sulphide and sulphhydrate
7.- Arnmonium sulphide
8.- Sodium thioglycolate
In addition, according to Table 34, in the U.SA., 9 out of 16 mills employ a

thermal process to strip the rougher flotation reagents from the copper
concntrate prior to moly recovery, This luxury was not indulged in the rest
of the world except for the COMECOM countries, where 2 out of the 4 mills
listed strip the primary reagents. The stripping processes employed were:
1.- Steaming
Sulphides - Depression ami activation 117
2.-Pressure steaming
3.-Skin roasting (a light roast)
4.-Cooking (indirect heating of the pulp)
5.- Conditioning with acid and hydrogen peroxide
6.-Conditioning with soelium hypochlorite
Needless to say, the skyrocketing oil prices of the early seventies eliscouragecl
heat treatment of copper concentrates, and encourageel the elevelopment of
more sophisticated copper depression schernes, as also the use of
hydrometallurgy to effect the final copper removal from the moly
concntrate.
Post reagent stripping: the practice used in the removal of molybdenite

from the copper concentrate was to rougher float a molybdenite concentrate,
and then submit this concentrate to multiple cleaning steps (from 5 to 20).
This operation has been an obvious candidate for the substitution of multiple
flotation banks by flotation columns; but this substitution is not easy because
the specifications on the minimum copper, pyrite and insoluble content of
commercial molybelenite concentrates is stringent, anel the value of the co-
product to the mill does not allow the luxury of Mo losses to the copper tails.
Sodium sulphide as a depressant
Because hydrogen sulphide is a weak acid, prediction of mineral depression

by its ion s requires a knowledge of the system pH. The classic work
elucidating this subject was publisbed by Sutherland and Wark (1955). They
point out, based on contact angle data, that, just as in the case of the cyanide
ion, "there is a critical hydrosulphide ion concentration below which
adsorption of xanthate is possible, aboye wbich adsorption is impossible" (see
Fig. 25). They add that the critical hydrosulphide ion concentration for 25
mg/I of potassium ethyl xanthate is: 0.01 gm HS-/1 for galena, 0.30 mg/I for
chalcopyrite, 1.3 mg/l for bornite, 1.7 mg/I for covellite, 2.5 mg/l for pyrite,
and 6.4 mg/I for chalcocite. Similar curves can be measured for other
collectors and at different concentrations.
Tbe mechanism of depression of sulpbide minerals by the HS- ion is

controversial, Leja (1982) suggests that tbe .presence of colloidal sulpbur on
the mineral surface coulel be a elepressive factor, that thio anions and
sulphanes may be formed on the mineral surface, or tbat the sulpbbydrate
ion could destroy tbe collector coating on tbe bulk floated mineral, if
depression is during tbe cleaning process. Klassen and Mokrousov (1963),
writing in the late fifties, quote Russian data showing tbat tbe depressive
effect of sulpbides is considerably more effective, tbe shorter the alkyl group
118 FLOTA TION - Theory, Reagents and Ore Testing
100
(")
o Contact possible
e be lou cur\!es
80
--
ro
;+
ro
~ 60
>1.
I:ll
E
I
o
::tJ 40
C'\
en
"f'J
z:
20
4 5 6 7 8 9 10 11 12
pH
Fig. 25.- Relationship between pH and the concentration of sodium

sulphide necessary to prevent contact at the surface of various
sulphide minerals, in the presence of 25 mg/l of KEtX
is on a xanthate. They also emphasise the decomposition of the sulphhydrate

ion in solution by oxygen, suggesting that in many cases the depression of a
sulphide mineral may well be due to oxygen scavenging in the pulp, pointing
out that "sulphide minerals depressed by sodium sulphide regain floatability if
sufficient time is allowed for the dissolved oxygen to react with the sulphide".
Also of interest is their observation of the catalytic effect of mineral surfaces
on the decomposition of the sulphides, saying:
"...variations in the activating or depressing action of sodium

sulphide with time are due to the fast oxidation of
hydrosulphide and sulphide ions in the pulpo Studying the
rate of oxidation of these ions, Shvedov and Shorsher found
that, on agitation of an aqueous solution of sodium
sulphide, the bivalent sulphur ions disappear from solution
in 35 minutes. When fine1y ground pyrite is introduced into
the same solution (10%), the sulphide ions disappear from
the solution even faster: they cannot be detected after 20
minutes. In both cases, paralle1 with the disappearance of
bivalent sulphur ions, a quantity of S04 = ions appear.' ....
"Mitrofanov considers that the sulphide ion oxidation

proceeds catalytically on the surface of particles. The
average time of oxidation of monolayers of S = and HS- ions
on the surface of such minerals as galena is on the order of
a fraction of a second.'
Sulphides in the copper/moly process: the first use, in the western world, of
sodium sulphide in the depression of copper in a molybdenum by-product
plant appears to have been in 1934 at Cananea (Mexico). The process also
was patented in Russia in 1936. The Nokes reagents, a sulphhydrate variant,
were patented in 1949 by Kennecott for use in El Teniente.
TabLe 36.- MOLY PLANTS USING THE SULPHIDE PROCESS - ShirLey (1979)
Anamax NaSH onLy depressant used - hot water in the rougher.

Andina NaSH onLy depressant.
Mission NaSH used onLy in the rougher and first two stages of
eLeaning prior to roasting.
Pima Preeonditioning with ammonium suLphide prior to rougher

fLotation. A mixture of NaSH (85%) and (NH4)2S (15%)
used as primary depressant. Sodium eyanide and Zn eyanide
eompLex used in various of the eLeaner stages.
Brenda NaSH used a primary depressant. NaCN used in
finaL cLeaners.
IsLand Copper NaSH used a primary depressant. NaCN used in finaL eLeaners.
Lornex NaSH used a primary depressant. NaCN used in finaL eLeaners.

GibraLter Preeonditioning with (NH4)2S, with NaSH then used as the
primary depressant.
Gasp NaSH used a primary depressant. NaCN used in finaL eLeaners.

Russian Na2S used as a primary depressant. PLant feeds generaLLy
PLants steamed prior to the suLphide addition. Hot puLp fLoated in
the rougher. \
Inspiration Ammonium suLphide preeonditioner prior to rougher and
phosphate Nokes as primary depressant
Pinto VaLLey Ammonium suLphide preeonditioner prior to rougher and
phosphate Nokes as primary depressant
Air (oxygen) and water are the most important variables with respect to
sulphide consumption in the depression process. The effective amount of
sulphhydrate ion is based on its ionic concentration, therefore pulp dilution is
a major factor in consumption. Shirley (1979) reports (Table 36) that at a
Canadian mine treating chalcopyrite, where the molybdenite rougher is
operated at 25 to 30% solids, NaSH consumption is about 10 kgjt, while
other mines running at 35 to 40% solids have consumption rates just aboye
10 kgjt. A further reduction in sulphhydrate consumption can theoretically
be obtained by replacing the air in the flotation machines with nitrogen
sourced from an oxygen plant feeding a flash smelter. Apparently the
laboratory reductions in reagent consumption are not always duplicated
consistently in the full-scale operations,
In many applications the sulphide fed into the copper j molybdenite roughers
is supplemented with sodium cyanide or the ferro or ferricyanides in the
cleaners to try and reduce the number of stages required to meet concntrate
specifications.
The chemistry of the Nokes reagents
The Nokes patent (US Patent No. 2,492,936) specified a series of

formulations which produced HS- for the depression of copper in a moly
separation. Only two have been used commercially: LR-744, developed for
El Teniente in 1949; and arsenic Nokes, better known as Anamol D, as it was
developed by Anaconda for Chuquicamata, but fell under the c1aims of the
Nokes patent.
LR-744, or Plzosplzate Nokes is easily prepared, but involves a very dangerous

operation as HzS is evolved, and can be evolved copiously if the reaction runs
away. As has been noted aboye, hydrogen sulphide is as toxic as hydrogen
cyanide, as well as having a very broad explosive limit in air, so precautions
must be taken even in laboratory preparations. At an industrial scale it is
prepared in a jacketed stirred tank with adequate cooling water availability.
The apparatus and the area must be explosion proof, using gasoline refmery
standards for the e1ectrical installations. The recipe for the preparation is
to dissolve caustic soda in water (highly exothermic), cool the solution to 40-
45C, then slowly add phosphorous pentasulphide, keeping the tank stirred
vigorously and at a temperature lower than 75C. The ratio of NaOH to
PzPs should be around 3:2 (El Teniente uses 13:10), and amounts should be
such that the resultant solution will be 25% by weight.
The reaction has been studied by Castro and Pavez (1977), who postulate the
fol1owing main reaction
while Gutierrez and Sanhueza (1977) indicare that they have identified the
Iollowing hydrolysis products
Na3ps4 + H2 0 ------~ Na3PS3o + H2 S
Na3ps30 + H2 0 ------1> Na3PS202 + H2S
Na3PS202 + H2 0
Na3PS03 + H2 0
Castro and Pavez (1977) deterrnined that at a 13:10 ratio, the solution
assayed an equivalent 27% Na2S, and at 1:1, 30%. Using this reagent, El
Teniente is one of the very few mines that does not need to supplement the
depressors to meet concentrate copper specifications, nor do they heat the
concentrate of the rougher pulpo The improved effectiveness is attributed to
unidentified phosphorous compounds.
Anamol D or Arsenic Nokes is covered by the Nokes patent, but in use its
addition rate is adjusted to maintain a pulp emf aboye -200 mV, which is
covered by an Anaconda patent issued to J.P. Delaney (US Pat No 3 655 044,
1972). It was for many years the standard depressant in Anaconda's
Chuquicamata mine. Its use was discontinued by CODELCO because of
concern over handling arsenic trioxide on site. Anamol D is prepared by dry
mixing arsenic trioxide and sodium sulphide, or can be prepared by adding
arsenic trioxide to a sodium sulphide or sulphhydrate solution. Ease of
dissolution depends on the purity of the arsenic trioxide. The ratio of sodium
sulphide to arsenic trioxide employed varies between 3 and 4:1. In a mixture
containing 4 parts sulphide, only 27% of the Na2S reacts. Castro and Pavez
(1977) suggest that the fol1owing reactions occur:
AS203 + 3Na2S + 2H20 --~ Na3Aso2S2 + Na3Aso3 + H+
AS203 + 3Na2S + 2H20 ------~ Na3As04 + Na3Asos3 + H+
These reactions indicate that the depressant species are HS-, sodium
arsenate and a mixture of mono-, di-, and tri\thioarsenate. With Anamol D,
the HS- ion is more important as a depressant than the arsenic compounds,
which probably function as oxidation inhibitors to protect the HS-; while in
the LR-744 there is only a small residual HS- concentration, suggesting that
the phosphorous compounds are probably the more important depressants
for copper.
Laboratory testing to develop, or improve, a moly/ copper separation system

requires considerable experience with the process; but today it is obvious that
column flotation will provide a much lower capital hardware alternative, and
should change the depression steps to a simpler scheme. Because of the
aging effects on wet concntrate, it is prudent to plan the experimental work
on-site, even though al! the analytical and mineralogical work can be done at
a better equipped research facility.
Sulphdisation of refractory ores
In studying the processing of oxide ores, it is important to pay attention to the

composition of the gangueo There are two main varieties of gangue: siliceous
and carbonaceous. The first is mainly silica and the latter calcium carbonate.
In addition, pyrites will weather to limonite, and if alumina is present there
will be col!oidal days. The main oxidised copper minerals are: the two basic
carbonates, azurite and malachite; the oxide, cuprite; and various silicates, of
which chrysocolla is the most important. The principallead oxides are: the
carbonate, cerussite; and the sulphate, anglesite. The main zinc oxide
minerals are: the carbonate, smithsonite; and the silicate, calamine.
The first use of sulphidisation was with Australian lead-zinc-silver ores, the
pioneer flotation industry. The excessive consumption of sodium sulphide
and difficulty in controlling the flotation process has kept the Australian
CSIRO active in research on methods of improving the process. Veryearly
on, the stage addition of the reagents was instituted, and in the past decade
the use of ion specific electrodes to control the residuals of SH- and collector
molecules has been successfully applied. Jones and Woodcock (1979)
studied the recovery of values from oxidised lead minerals from old dump
material, which they treated by flotation of a sulphide concentrate (galena
and marmatite) followed by sulphidisation and flotation of an anglesite
concentrate. Sulphidisation by slug addition of Na2S gave relatively
uncontrolled sulphidising conditions, but potential controlled sulphidisation,
using a sulphide ion-selective electrode (ISE), resulted in good control and
improved metaUurgy. The optimum Es value for three stage flotation was
found to be -600 mV (the absolute value may not be reproducible because of
differences in electrode construction). They noted that sulphidisation time,
at -600 mV, over the of range 1 - 5 minutes, had little effect on oxide
recovery. Flotation time was important and three 10 minute stages (total 30
minutes) was needed for good lead recovery. A decant wash prior to
sulphidising improved metaUurgy and decreased sulphide consumption.
Excess S = in solution displaced xanthate from the mineral surface, the
amount displaced varying dependent on Es. At Es more negative than -600
mV, maximum displacement occurred.
The laboratory procedures followed are of general application for

sulphidisation studies, so will be reviewed in detail, Flotation tests were
made in a Denver DI lab flotation machine, using a 500 mI glass cell.
Impeller speed was fixed at 1800 rpm. The Es values were measured as the
potential between an ISE and a saturated calomel electrode (SCE). pH
measurements were made with a glass electrode/SCE pair. E p t values are
the potential of a platinum electrode versus SCE. Oxygen concentrations
were expressed as per cent saturation as measured with a Beckman 777
oxygen analyser.
Flotation tests were conducted on 500 gm portions riffled from a bullesample

of residue dump material supplied by The Zinc Corporation, Broken Hill,
N.S.W., Australia. The sample assayed 3.69% total lead, 2.74% total zinc,
and 76 g/t silver. About 3/4 of the lead was present as anglesite. The zinc
was mainly present as marmatite. Soluble salts, including zinc and
manganese, were presento Sizing of the material was 2.5% plus 0.417 mm,
26% minus 0.074 mm, with 80% passing 0.208 mm. Commercial reagents
were employed in testing. Flake sodium sulphide (62% NazS) was used as
water solution containing 5, 10, or 20% sodium sulphide, depending on the
Es values required. Potassium amyl xanthate (KAX) from American
Cyanamide and MIBC from Shell (Australia) were used as 1% water
solutions. Sodium metasilicate (BDH technical reagent) was used as a 10%
Na zSi0 3.5HzO solution, and copper sulphate was used as a 5%
CuS04.5H20 solution.
The flotation procedure used was based on a scheme provided by The Zinc
Corporation. A 500 g sample was wet ground in an iron ball mill for 20 min.,
giving a product sizing 5% plus 0.1457 mm, 58% minus 0.074 mm, and 80%
passing 0.104 mm. After transfer to the cell and conditioning for 1 mino with
225 g/t copper sulphate, the sulphide minerals (galena and marmatite) were
floated with 135 g/t KAX and 18 g/t MIBC, using 5 mino roughing and 10
mino scavenging. Conditions for this sulphide float were kept constant. The
rougher and scavenger concentrates from each test were assayed separately,
but the results for a combined sulphide concentrate are reported.
After sulphide flotation, oxide flotation was conducted either immediately,

or, as in plant practice, afier decantation of mineralised liquor. With no
decantation, flotation was accomplished by conditioning the pulp with 1800
g/t sodium silicate for 1 min., sulphidising ~th sodium su1phide as described
later, and then floating with KAX and MIBC. However, in sorne tests the
pulp was allowed to settle, and a constant volume (1200 ml) of slightly turbid
supernatent liquor was decanted to remove dissolved salts. This decanted
liquor was added to the final tailings for filtration. The settled pu1p was re-
pulped with 1200 mI of tap water, and, after conditioning with sodium silicate,
sulphidised and floated.
Two different sulphidising-flotation techniques were investigated.
1.- A single slug addition of NazS was made. After 1 min, 225 g/t
CuS04.5H zO was added, and after a further 1 min, 90 g/t KAX and
4.5 g/t MIBC were added, Rougher flotation of an "oxide"
concentrate (i.e., sulphidised anglesite) was conducted, After
addition of a further small amount of KAX and MIBC, a 10 minute
scavenging was conducted,
2.- Three sulphidising periods of 1, 3 or 5 min at controlled Es levels

were used with the air inlet on the cell closed. After each
sulphidising period, KAX and MIBC were added and a 5 or 10 min
flotation period was used. No Na-S was added during the flotation
periods, and the concentrates were kept separate. In most tests the
Es value was kept at the same level in each of the three stages. In
sorne tests, however, a different value of Es was used in each stage.
All solid products from each test were dried, weighed, and assayed. After
fusion with NazOz, cooling and dissolution of the melt with 1:1 HCI, and
appropriate dilution, total lead and zinc were determined by atomic
absorption spectrophotometry (AAS). Sorne products were assayed for silver
by AAS or fire assay.
Leach Precipitation Float (LPF): the process, where copper from partially
oxidised ores was recovered by a sulphuric acid leach followed by metallic
copper precipitation with iron particles in the pulp and flotation, is now
obsolete, though the term is used to describe certain sulphidisation schemes.
Currently, oxide copper ores are heap leached with sulphuric acid, the
dissolved copper separated from iron in solution by liquid-liquid extraction,
and metallic copper recovered by direct e1ectrowinning, a very low capital
and cost process; it rescued the financial1y crippled Arizona copper mining
industry in the early eighties by providing instant cash flow.
The classic LPF process was still used in the USSR into the sixties, but in the
post World War 11 western economy LPF had been displaced by vat leaching
and cementation with scrap iron. When the vast amount of cheap scrap iron
from the battle fields was exhausted, the LIX copper extraction process was
introduced. This process produces copper in an acid solution pure enough to
be directly electrowon. A precocious example of direct electrowinning was
the Anaconda Company operation in Chuquicamata, Chile (1913) , where the
oxide copper ore was low in iron, and therefore could be sent directly to an
electrolytic refincry, without elaborate pre-purification,
Extreme sulphidisation is now frequently referred to as an LPF process,

The closest to the original LPF process was the procedure used in the Butte
concentrator of the Montana Mining Division r-f the Anaconda company.
This plant was built in 1963; its design included a sand/slime separation to
a110w efficient recovery of the oxide copper in the slime circuito Palagi and
Stillar (1976) describe the process as follows: the slime flotation plant treats
mainly -200 mesh material sourced from the primary cyclones (5.6% +65
mesh, 18.2% +200 mesh, and 70% -400 mesh). In sweet ore there was
relatively little soluble copper in the slime, so it was floated normally. In sour
ore, the slime's grade ran up to 0.1% Cu higher than the sand grade, with up
to 40% of the acid soluble copper in the plant feed reporting to the slime
fraction. With sour ore, to recover the copper in the slime, this was acidified
to pH 4.0 and piped to the special slime treatment plant where the pulp was
stored for 20 minutes in an agitated leach and precipitation tank, where the
pH was lowered to 2.0. After the acid soluble copper was in solution, the
slime was reacted with hydrogen sulphide gas in an in-line static mixer, where
artificial covellite was precipitated. The amount of H 2S fed was
automatically controlled by an EMF probe held at O to + 150 mV. The pulp
was then floated with diethyl xanthogen formate and Cyanamid 3302 as the
collectors, and the concentrate upgraded in the sand float cleaner, as it was
essentially free of co11oidal fmes. Separate treatment improved slime copper
recovery by 15 to 20 points.
According to Bolles (1985), Morenci originally treated a monzonite porphyry

containing primarily chalcocite and pyrite. Later, considerably greater
amounts of "oxide" copper was encountered: mainly cuprite, azurite,
malachite, and brochantite. The bulk of the oxidised copper was/is formed
subsequent to the breaking of the ore as imperceptible coatings on the
chalcocite. Oxidation by weathering begins when the ore is broken, and
continues until it is delivered to the crushing plants. Typically, the proportion
of oxidised copper increases by two to three times between drilling for
blasting and delivery to the millo For 1972, he quotes an average assay of:
total copper, 0.83%; oxide copper, 0.16%; pyrite, 5%; quartz, 43%; feldspars,
21%; micas, 4.3%; molybdenite, 0.015%; gold, 0.0014 oz/t; silver, 0.04540z/t.
Bolles further says: "..the Morenci LPF process converts approximately 40%
of oxide copper minerals to a sulphide which can be recovered by
conventional flotation methods in an alkaline circuit", and describes it as
fo11ows:
"Acid soluble copper content of the Morenci ores often

process was devised to recover part of this copper bearing

mineral. The process involves leaching the acid soluble
copper minerals with dilute sulphuric acid in a revolving
drum. Precipitation of the dissolved copper by a unique
sulphide precipitant was done simultaneously in the same
vessel. The treated slurry was made alkaline and
subsequent grinding and flotation operations were
performed on the alkaline pulpo Flotation of the
precipitated cupric sulphide occurred together with the
naturally occurring sulphide copper."
The chemical reaction took place within a slowly rotating drum 6 ft in

diameter and 26 ft long. The drum was lined with acid-resisting brick and
was equipped with rubber-covered lifters for discharge of treated ore slurry.
Contact within the drum averaged 5.5 mino
Dry ore was fed into the drums by conveyor belts; water was then added to
dilute to approximately 78% solids within the drum, followed by sulphuric
acid to lower the pH to 1.6. Finely ground calcium sulphide in a water slurry
was drawn from a circulating loop pipeline and injected into the drum.
Control of this reagent is vital to the success of the process. This was
accomplished with an oxidation-reduction meter whose electro des were
immersed in a sample stream of the discharge solution. A negative millivolt
reading was maintained, which resulted in not-quite-complete precipitation
of dissolved copper.
Normal operation results in minor evolution of hydrogen sulphide gas, which

increases with excess use of precipitant. Therefore, fume control is essential.
This was accomplished by directing the flow of air that sweeps the drum
mixer into the ball mill, where lime additions completely react with the H 2S
in the gases during passage through the ball milI.
The calcium sulphide precipitant was an impure product produced locally

from pyrite and burnt lime. These components, mixed in proper proportions,
were finely ground (dry), then moistened and pelletised. After hardening by
gentle heating and screening to remove fines, the pellets were fed with fine
coal into a reducing kiln. Reduced pellets, containing both calcium sulphide
and metallic iron, were magnetically cleaned, then stored for wet grinding
before use.
Chapter 10
Mlll TESTS CASE HISTORIES

m
San Manuel a non-sulphde moly separation process

At San Manuel, the final copper concentrate, which contains bctween 1.0 and
1.2% molybdenite, is thickened to 50% solids and sent to two conditioners in
series. In the first conditioner, a sodium zinc cyanide complex is added to the
pulp, which is kept at a pH of 6.5 to 7.0, using sulphuric acid. The second
conditioner is fed oxide. Conditioning takes between 22 and 25 minutes.
From the conditioner the pulp is diluted to 20% solids with fresh water, and
fuel oil is added as the moly collector while it is sent to the rougher flotation
stage, where the pH is controlled with sulphuric acid. The feed to the first
cleaner combines the concentrate from the rougher and the tails from the
second and third cleaners. In the first cleaners and the scavengers, sodium
ferrocyanide is added as the copper depressant, and MIBC as the frother. In
the next two cleaning stages the reagents added are potassium ferricyanide
and sodium hypochlorite plus MIBC as the frother, following which
potassium ferricyanide is added at each stage. The froth in the final stages is
kept in check with Exfoam (Dearborn Chemical Co). The amount of each
reagent fed in the different flotation stages is controlled by measuring the
pulp's Redox potential. The San Manuel flowsheet thus consists of two
distinct stages: (1) collector removal in the conditioners and the rougher
flotation using hydrogen peroxide and sodium zinc cyanide, and (2) ferro and
ferricyanide copper depression in the cleaner and scavenger circuits. The
moly concentrate from the final cleaners may or may not be leached to meet
the copper specs. San Manuel's copper and moly concentrate assays are
shown in Table 37.
Because Farlow Davis's paper on the S~ Manuel moly plant was only
circulated as a mimeographed handout at the local section meeting of the
AIME in Arizona, and therefore difficult to obtain, the following pages are
reproduced verbatim, as they provide his explanation of the mechanisms
involved in the copper depression process at San Manuel.
-127 -
PLant feed thickener

1(1)
concentrate underfLow (1 ) Density measurement and controL
"lO) (2a) (2) Redox potentiaL measurement and controL
Conditioner number 1
I
(2a)= Redox potentiaL measurement
concentrate underfLow (3) = pH measurement and controL
v (3)(2)
Conditioner number 2
I
concentrate underfLow
1(3)
v
roughers
I I
concentrate v
I Tai L - - - - -..- - - - ,
CLeaner number 1 .. ---------------,

I
I I
v '"
concentrate TaiL---..- -
I (2a)(3)
v
CLeaner number 2
I
concentrate
I
v
Tai L - - - - -..
v v
cu thickener
I
. Tl
v
r------=----
CLeaner number 3
lnd Scavenger ~" - - - " ,I
concentrate
CLe:rn_e_r__n_um_b_e_r__
-1
concentrate
I
4_"~J'
TaiL ------
TaiL ~ I '"
.. Concentrate
__;'3;"
Concentrate
TaiL
Scavenger
I
v
l
I
A
TaiL----'I
r---"=Tl LJ
v
T'il~
CLeaner number 5
'00"0"'"
CLe:ner number 6 .. ~ I
-1 I '"
concentrate T:i L J
I
ALternate
v - - - - - - -..~
FiLter
--------------.. Dryer
Fig, 26.- San Manuel Molybdenite recovery circuit (Farlow Davis, 1976)
Mili tests - case histories 129
TabLe 37.- SAN MANUEL COPPER AND MOLY ASSAYS
Copper Coneent rate 110 Lybdenum Coneent rate

Raw FLoated
Cu 27.6% 27.9% 0.5%

MOS 2 1.12% 0.2% 96.2%
Re 0.0017% 0.08%
Au az/t 0.38
Instrumentation: The molybdenite plant thickener is equipped with a

density measuring device and vari-drive pump operating in conjunction with a
recorder-controller to maintain a constant density and stable flow to the
conditioning tanks, Manual control of the underflow can produce surges in
the conditioners and rougher, detrimentally affecting the copper-molybdenite
separation.
The pH of the conditioners, rougher and first scavenger are automaticaIly

controlIed by the addition of sulphuric acid with positive displacement
pumps.
Oxidation-reduction potential (ORP) measurements are taken from the #1

conditioner, #2 conditioner and second cIeaner. Using a platinum
electrode and a silver-silver chloride reference, the ORP measurement
depends on the ratio of the reducing or oxidising activities or concentrations
and not upon the absolute amounts presento Thus, in an operational sense,
ORP can be understood about as welI as pH.
Laboratory work has shown a strong response of ORP to the reagents used in
the conditioning steps of the molybdenite circuito This strong response of the
ORP to reagent conditions controIling the oxidation-reduction reactions in
the pulp suggested the use of this function in regulating rates of reagent
addition.
Sodium zinc cyanide: This is the reactant product of sodium cyanide and zinc
oxide. It serves two purposes: (1) to provide a desired level of HCN in the
conditioning step with hydrogen peroxide; and (2) to remove, either wholIy or
partially, the colIector coating on the mineral surface. The addition rate is
critical for an effective conditioning step with-hydrogen peroxide. As wilI be
explained later, too little or too much are equaIly unsatisfactory. It is added
to the #1 conditioner to prepare the coIlector on the chaIcopyrite for
nuIlification by the hydrogen peroxide in the second conditioning step.
The process of cyanide attacking the chaIcopyrite surface would be driven by

two factors: (1) the property of cyanide as a reducing agent; and (2) the
ready complexation of Cu + and Fe + + ions at the mineral surface with a
slow solution of both. On slightly altered chalcopyrite surfaces, cyanide
would reduce cupric sites to cuprous, at the same time losing fui equivalent
free cyanide as cyanate. A monolayer of adsorbed cupro and ferro cyanogen
complexes would form which, given time, would solubilise 1eaving vacant sites
for further cyanidation alld solution. The process is one of mobile
equilibrium.
Since at least sorne surface oxidation seems to be required for effective

mineral flotation by soluble xanthates or oily collectors, the reducing action
of cyanide plus the presence of a cyanide monolayer would. represent two
strong deterrents to mineral surface-collector interaction.
While the foregoing may apply to elean or slightly altered chalcopyrite

surfaces, the chemistry is complicated by chemisorbed and physically
absorbed reagents responsible for the hydrophobic condition of the mineral
surface. These would inelude the primary and secondary collectors in
addition to penetrations of the collector hydrocarbon members by fuel oils,
alcohols and perhaps other products of collector oxidation and
decomposition.
Admittedly, the action on the mineral surface under these conditions is not
elear. If the cyanide does penetrate the absorbed collector, then sorne
complexing and solution would be expected. It is possible that metal collector
bond eleavage might take place leaving the collector ions in proper condition
for reaction with the hydrogen peroxide; or, at least, leaving the surfaces
elean for reaction with the ferro-ferri cyanide complexo
Hydrogen peroxide (fed as a 50%.solution) serves two purposes. First, the

pulp is conditioned with hydrogen peroxide to transform all the water soluble
sulphhydryl type collectors present, such as the normal1y used isopropyl
xanthate, to insoluble materials. Secondly and possibly most important, the
residual peroxide in the pulp leaving the conditioning step oxidises a portion
of the sodium ferrocyanide fed to the rougher flotation to form ferricyanide.
At the inception of using hydrogen peroxide in the molybdenite plant, S-3302

started to be used in the copper plant as the primary collector with
satisfactory resuIts in the molybdenite plant. Since then, Minerec SM-S has
replaced the S-3302. Neither of these oily collectors seem to be acted upon
by peroxide in the conditioning step. Several other oily col1ectors have been
tested at San Manuel since hydrogen peroxide was put in use and in every
case there was no noticeable effect in the conditioning step. Since these oily
Mill tests - case histories 131
collectors are the primary contributors to the hydrophobic state of the

mineral surface, it fol1ows that any mechanism accounting for the depression
of chalcopyrite must also account for the removal from, or neutralisation of,
these collectors from the mineral surface, Straightforward oxidation of these
molecules is not a satisfactory explanation. If, as has been suggested, the
cyanide and the pH create an environment in the first conditioner whereby
the col1ector ions are removed from the mineral surface, then oxidation of
the col1ector may not be necessary if the mineral surface could be altered to
prevent collector ion reattachment,
The hydrogen peroxide dimerises the isopropyl xanthate into what is

commonly cal1ed dixanthogen. Laboratory tests have indicated that hydrogen
peroxide will react in a similar manner with all the commonly used
sulphhydryl type col1ectors. The dimerisation of these col1ector is a re1ative1y
slow one requiring minutes to take place even under the most ideal
conditions. The main factors directly affecting the conditioning time are the
pH of the pulp, the efficiency of the conditioner and the concentration of
hydrogen peroxide in the pulpo The size of the conditioners was chosen on
empirical1yderived data from the pilot plant tests and on the maximum
volume of pulp that could be expected to be treated. It was soon realised
that the large plant conditioners were not as efficient as the smal1er pilot
plant ones, and their capacity had to be increased.
To insure the dimerisation reaction going to completion within the time

available.ethe pH of the conditioner wherein the hydrogen peroxide is added
normally should be maintained between 6.5 and 7.0. Up to a point, the lower
the pH, the faster the reaction takes place. Although the plant can be
operated below 6.0 pH, it is more difficult because of the heavy metal ions
put into solution, which in turn acce1erate the decomposition of the hydrogen
peroxide. In addition, the corrosion and maintenance costs would increase at
lower pHs. With the conditioning time limited by the size of the
conditioners, the reaction time required can be adjusted to sorne extent by
increasing or decreasing the pH or the amount of hydrogen peroxide.
Other minor factors affecting the reaction time are temperature and the type
and amount of reagent being used. The process has operated satisfactorily
with pulp temperatures varying from 60 F: to 90 F throughout the year.
Different xanthates have been used with no necessary adjustments from one
to the other. Oily col1ectors, as has been stafed, apparently do not react with
hydrogen peroxide in the conditioning step. It has been found that over-
feeding the primary collector (SM-8) in the copper plant will adversely affect
the copper-molybdenite separation. Chemical changes in the molybdenite
plant will not remedy the effect. An increase in the secondary collector
(Z 11) in the copper plant will not result in the sarne situation, since
apparently a change in the hydrogen peroxide Ieed rate will result in a
satisfactory separation.
Another factor which indirectly affects the reaction time and directly affects
the amount of peroxide that must be added is the concentration of the
hydrocyanic acid in the feed to the second conditioner. Hydrocyanic acid ~"
needed to stabilise the hydrogen peroxide in the rather hostile atmosphere of
a copper concentrate pulpo Besides the normal catalytic decomposition of
peroxide in such an environment, hydrogen peroxide reacts with chalcopyrite,
San Manuel's primary copper mineral. This reaction will take place in
preference to the slow dimerisation reaction. Hydrocyanic acid can be used
to inactivate catalysts causing decomposition of the peroxide. It was also
pointed out that if there were not enough hydrocyanic acid present to
"poison" all the catalysts present, the catalytic decomposition would take
place at a faster rate. In other words, enough hydrocyanic acid has to be
present and anything less than enough is unsatisfactory. A sodium zinc
cyanide-sodium cyanide mixture is added to the pulp in the first conditioner.
It was found that the cyanide feed rate could be set at a fixed value which
would provide a satisfactory amount of hydrocyanic acid during normal
fluctuations in the amount of feed to the plant. Of course, drastic changes in
the amount of feed would call for an adjustment to the cyanide feed rate.
Also, it can be seen that lowering the pH would also call for an increase in
the cyanide feed rate.
The function of sodium ferrocyanide and potassium ferricyanide is to

produce a hydrophyllic surface on the copper and the iron minerals which, in
effect, depress these minerals while allowing the molybdenite to float. Even
though ferrocyanide has been effectively used to depress chalcocite in several
plants, its ability to depress chalcopyrite in most cases has been proven to be
unsatisfactory. The use of ferricyanide had not been seriously considered as
a commercial depressant because of its relatively high cost and the general
opinion among experimenters that there was little or no difference between
using it or ferrocyanide. In the case of copper sulphide minerals such as
chalcocite and covellite, there is probably little difference. However, in the
case of chalcopyrite where there are both copper and iron positions on the
surface, the problem becomes more complexo In all probability the
ferrocyanide and the ferricyanide absorb equally well on the copper
positions: however, absorption at only the copper positions would result in
only partial depression of the mineral. Partial depression refers to a state in
which the rate of flotation of a mineral is slowed down, but that the mineral
will float if given enough time.
H is believed that in order to obtain complete depression, ferrocyanide or

ferricyanide ions must also be absorbed at the iron positions as well as the
copper positions. Ir the iron atoms on the mineral surface are in the reduced
form, ferricyanide will be absorbed effectively; and if the iron atoms on the
surface are in the oxidised state, ferrocyanide will effectively absorb on them.
The iron in chalcopyrite is generally corrsidered to be in the reduced state;
therefore, ferricyanide would be an effective depressing ion. So, in the
treatment of a commercial copper concentrate pulp containing chalcopyrite,
two approaches can be taken. These approaches are: (1) the addition of
ferrocyanide with a small amount of ferricyanide to take care of the reduced
iron positions, or (2) the addition of ferricyanide with a small amount of
ferrocyanide to take care of the oxidised iron positions.
In San Manuel's process, the first method of approach is utilised in the

rougher and the first c1eaner. However, in both cases, it is unnecessary to
add ferricyanide as such. In the rougher feed there is residual hydrogen
peroxide which reacts with part of the ferrocyanide addition to form
ferricyanide. This is a reversible reaction. Residual sodium hypochlorite
existing in the second c1eaner tailing oxidises part of the first c1eaner
ferrocyanide to ferricyanide.
Although pyrite has not been mentioned in this discussion, it appears that the
proposed theory applies equally as well to its depression. The iron positions
in pyrite are also generally considered to be in the reduced state; and,
therefore, ferricyanide would be the effective depressant if the surfaces of the
mineral were fresh and c1ean. However, if the surfaces were subjected to an
oxidising atmosphere, ferrocyanide would be the effective depressant.
There are, of course, factors other than the ratio of ferricyanide to

ferrocyanide that have to be taken into consideration in obtaining an effective
depression. The pH of the pulp should be between 7.0 and 8.0 for the best
results. Also, the addition of too much fue! oil will have a tendency to
interfere with the depression.
In the five final c1eaners, potassium ferricyanide is added. In the pilot plant
test work, the results indicated that, if ferricyanide was added in the fmal
c1eaners, a satisfactory concentrate grade .could be produced without the
need of adding sodium hypochlorite; whereas, if sodium ferrocyanide were
used in the final c1eaners, sodium hypochlorite would have to be added to
produce an acceptable grade concentrate. As it turned out, hypochlorite had
to be added in both cases.
At times, ferrocyanide has been substituted for ferricyanide with excellent

results. However, when it is used, the addition rate of hypochlorite has to be

increased. The increased amount of hypochlorite required is probably due to
a portion of the hypochlorite reacting with ferrocyanide producing
ferricyanide,
The function of sodium hypochlorite addition to the second c1eaner is not

fully understood, It is believed that the sodium hypochlorite performs several
functions at once: name!y, it disperses the slirne faction and oils, dissolves a
great quantity of slimes, dissolves molybdenite that has been smeared on
copper mineral particles, and adjusts the pH of the pulpo In addition, it
tarnishes the surface of sorne iron and copper mineral partic1es, and it
oxidises sorne of the ferrocyanide added to the first c1eaners to ferricyanide.
In the pilot plant test work it was not necessary to use sodium hypochlorite in
the circuit. However, its use in the molybdenite plant is essential for the
continuous production of acceptable molybdenite concentrate. Various
substitutes have been tested to replace hypochlorite but, to date, none have
been successful.
Exfoam 636: The reagent is a formulation sold by Dearborn Chemical

Company. It has been found to contain approximate!y 60% kerosene and
sorne glycols. Its use in the final three c1eaning stages is essential in the
production of an acceptable grade of molybdenite. It is not used to kill the
froth but simply to loosen the froth somewhat to provide bubble drainage.
Fuel oi!: a 50-50 mixture of #2 diese! and kerosene is added to promote

molybdenite in the rougher circuit. The addition of fue! oil should be he!d to
a minimum because an excess seems to interfere with the depression of iron
and copper minerals. Also, fue! oil tends to flocculate the slime partic1es,
creating a difficult separation in the c1eaners. An emulsifying agent, Dowfax
9N9, is added to the fue! oil used in the flotation circuits.
Sulphuric acid is added automatically at several points to control pH.
Metlzyl isobutyl carbinol is used as a frother; it must be added at various

points to produce a satisfactory froth.
Water: The proper use of water is extremely important to the successful

separation of molybdenite from copper sulphide concentrate. At San Manue!
fresh (mine) water is used for dilution in the rougher, first and second
c1eaners, and scavenger floats. The water do es not vary in pH and soluble
iron content and, for these reasons, it is not suitable for use in the fmal four
c1eaners. Domestic water is used there for dilution. In a molybdenite
Table 38.- MOLYBDENITE PLANT REAGENT CONSUMPTION

Reagent gm/metric ton of
Copper Conc. Treated
Sodium cyanide 100 - 225
Zinc oxide 20 - 35
Sulphuric acid 750 - 975
Hydrogen peroxide 750 - 1,125
Sodium ferrocyanide 700 - 1,250
Fueloil 225 - 275
Caustic 350 525
Chlorine 300 - 550
Exfoam 636 150 - 250
Potassium ferricyanide 200 - 225
Methyl isobutyl carbinol 30 - 50
cleaner circuit, it is generally thought that the more dilution the better. This
has proven to be the case at San Manuel; however, the amount of dilution is
limited by the necessity of maintaining a satisfactory froth condition
throughout the cleaner circuito The addition of frother to the cleaner stages
to obtain a froth is not completely satisfactory because of its tendency to
build up in the final cleaners. Therefore, the solids density maintained in the
various cleaners is important to proper cleaning. This does not mean to
imply that a final value can be set for each cleaner since optimum values will
vary from day to day, but generallimits can be seto
Table 39.~ MICROSCOPIC ANALYSIS OF THE MOLY PLANT STREAMS
All figures Feed Rougher 2nd Scavenger Final Moly

are weight % Tail Tail Concentrate
Chalcopyrite 80.91 80.37 88.97 1.80

Covellite 1.45 1.18 0.66 0.01
Chalcocite 0.39 0.30 Trace
Bornite 0.51 0.29 0.12 Trace
Native Copper Trace Trace
Molybdenite 2.60 0.25 5.87 97.72
Gangue 6.56 8.38 1.90 0.31
Pyrite 7.58 9.23 2.48 0.05
Assays
\
Cu 27.61 27.92 31.48 0.50
MoS 2 1.12 0.2 3.72 96.20
Re 0.0017 0.080
Au oz/ton 0.38
COIDElLCOAndina - slime sulphidisation in a mmdi/slme separation
In 1977, Andina pushed up mill feed from 13,000 to 14,000 t /d, with a
consequent drop in what was alrcady an unsatisfactory recovery, so corrective
action had to be taken.
SAMPLE AND ASSAY POINTS
Group No 1 Group No 2 Group No 3

rloat -t Ime rloat time f10at ti.. e
Minutes 6 minu-t.es B fI1inutes
Bank feed transfer frofl1 transfer frol'l'l Rougher

group 1 group 1 to 2 group 2 to 3 tails
Fig. 27.- Andina rougher circuit in 1976, showing sampling points.
The starting point for the experimental programme was from the 1976
circuit, shown in Fig. 27. It consisted of two banks of 10 each Agitair No
120 flotation cells having a unit volume of 240 ft3, and a total rougher
residencetime, per bank, of 20 minutes. The cells were in three groups of 3,
3 and 4 cells. The effect of increasing the tonnage was a reduction in
residence time in the flotation bank, and a coarser grind with increased
middlings. The metallurgical department was asked to study seven different
schemes to improve mill performance:
1.- Increase rougher flotation time.

2.- Install a sand/slime separation of the rougher feed with separate
flotation in two banks.
3.- A 6 minute float of all the feed, then a sand/slime separation, and
individual flotation of the slimes and sands in the remaining cells.
4.- A 12 minute float of all the feed, then a sand/slime separation, and
individual flotation of the slimes and sands in the rernaining cells.

5.- NaSH addition lo the head of flotation.
6.- NaSH addition after 6 minutes of Ilotation,
7.- NaSH addition after 12 minutes of f1otation.
Laboratory f1otations were made to check the effect of a 3 minute increase in

f1otation time. To do this, samples were taken of the heads, concentrate and
rougher tails, every day for a period of two rnonths, The pulp sarnple of the
tails of the third group were then floated in the laboratory for 3 additional
minutes without adding collector or frother. Average recovery increased
from 85.59% to 87.82% for the periodo So two additional cells were added to
each bank, the maximum number that could fit in the underground cavern.
The new circuit is shown in Fig. 28.
Fronl grinding
and classitying
SAtlPLE AliD ASSAY POItns
Group Iio 1 Group Ho 2 Group No 3

float tiMe CJoat tine Cloa1: tir:ae
6 Minutes 6 ninutes 121'rdnutes
Dan!< Peed transCer fron transCcr fran Rou!Jher

group 1 gl'oup 1 te Z group Z -to 3 tails
Fig. 28.- Andina rougher circuit in 1977, showing sampling points
To evaluate the effect of a sand/slime separation, ten tests were performed,

Samples were taken of the rougher feed, concentrate and tails, under normal
plant operating conditions, to obtain a standard to test as a comparison. The
feed sample was screened at -200 mesh, to simulate a cyclone cut. To
approximate the cyclone, the coarse sample h~d about 15% of the -200 mesh
fraction added to it. The +200 mesh sample was repulped to 40% solids,
with the addition of 15 g/t of amyl xanthate as a kicker for the middlings, and
floated in the laboratory for 8 minutes. The -200 mesh fraction was floated
for 8 minutes, at normal solids, and without added reagents.
,.....
TabLe 40.- LABORATORY RESULTS OF SAND/SLIME TESTING WITH SPLIT AT THE BALL MILL AND WITHOUT ADDED REAGENTS W
O'J
From baLL miLL discharge - Standard operation SAND/SLIME simuLation resuLts
I ROUGHER SECTIONS ~~i"h"9' CycLone SpLit
Proauct % Wt % Cu
ResuLts % Cu >-j
+ 200 m 36.6 1.5 1
IGroup No 1 f--- :oncentrate
'~,~ads 1.34
28.90
- 200 m
TotaL
63.4
100
1.24
1.34
l'
O
TiLs 0.31 ~
I .-
i~ Recovery 77.7 ~
~
>-<
~ -- O
ResuLts % Cu sL ime Sand .....,.
!Group No 2 ~ Circuit Circuit -"-<
Heads 0.31 ResuLts % Cu Groups Groups ResuLts % Cu
Concentrate 8.00 No. 1, 2, No. 1, 2, >--1
TaiLs 0.22 Heads 1.24 and 3. and 3. Heads 1.51 i:l'"
(ll
% Recovery 6.7 Concentrate 20.41 Concentrate 20.37
I
~
TaiLs 0.20
% Recovery 49.8 I I
Tai Ls 0.18
% Recovery 36.6
O
;
ResuLts % Cu ?:J
IGroup No 3 ~ ~ (ll
'jEads
Loncentrate
0.22
3.11
GLOBAL Run #1 #2 ~
(1l
TliLs 0.20 ResuLts % Cu % Cu g
I ,; Recovery 1.5
Heads 1.34 1.40
Ul
~
GLOBAL Run #1 #2 Concentrate 20.39 6.50 5Q..

TaiLs 0.19 0.22
ResuLts % Cu % Cu These are resuLts % Recovery 86.43 87.0 O
...
lIithout B (1l
Heads 1.34 1.34 sBnd/sLime
Concentrate 21.33 21.33 circuit -J
(ll
TaiLs 0.20 0.20 Ul
% Recovery 85.9 85.9
I Further resuLts in TabLe 40b S'
(Jq
~ v
Tabte 40b.- LABORATORY RESULTS Of SAND/SLIME TESTING WITH SPLIT AT THE

BALL MILL AND WITHOUT ADDED REAGENTS.
Raugher resutts withaut sand/sLime - 20 minute fLatatian time

GLOBAL Run #3 #4 #5 #6 117 118 119 #10
ResuLts % Cu % Cu % cu % Cu % cu % Cu % Cu % Cu
Heads 1.45 1.54 1.62 1.59 1.55 1.54 1 .. 6 1.41

Cancentrate 11.89 9.91 12.41 9.80 16.50 13.74 5.67 12.26
TaiLs 0.27 0.22 0.22 0.32 0.28 0.28 0.25 0.21
% Recavery 83.3 87.7 88.0 82.6 83.4 83.5 86.7 86.6
Raugher resutts withaut sand/sLime - 24 minute fLatatian time

GLOBAL Run #3 #4 115 #6 117 118 #9 #10
Resutts % Cu % Cu % Cu % Cu % Cu % Cu % Cu % Cu
Heads 1.45 1.54 1.54 1.46 1.41

Concentrate 7.34 8.18 8.79 5.12 8.43
Ta'lLs 0.20 0.21 0.26 0.24 0.19
% Recovery 88.6 88.6 85.7 87.7 88.5
Sand/stime resuLts
GLOBAL Run #3 114 115 116 #7 #8 #9 #10
ResuLts % Cu % Cu % cu % Cu % Cu % Cu % Cu % Cu
Heads 1.45 1.54 1.62 1.. 59 1.55 1.54 1.46 1.41

Concentrate 9.62 12.68 11.87 10.13 15.57 13.07 6.95 13.14
Tai Ls 0.27 0.21 0.20 0.30 0.27 0.27 0.24 0.20
% Recavery 83.3 88.0 89.4 83.3 83.7 83.7 86.3 87.1
The resu1ts for the first two tests, where the overall flotation time was
unchanged, are shown in Table 40 and 4Ob; the detailed resu1ts are included,
with a summary for the extended flotation times at the bottom of the tableo
The third series involved a bulk float for the first 6 minutes (the first group of
rougher cells), followed by a sand/slime separation, and sand and slime
circuits. Here another ten experiment series was runo Samples were taken
of the plant feed, the concentrate and tails of the first group; a common
concentrate for the second and third group, and the overall rougher tail.
With these data, standard recoveries for the first group and the combined
group 2 and 3 were calculated. As in the first series of runs, the tails from
rougher group 1 was sampled and screened to simulate a cyclone split. The
+ 200 mesh fraction was repulped to 40% solids, 15 g/t of amyl xanthate was
added and a 6 minute float was runo The -200 mesh fraction was also floated
for 6 minutes without further addition of reagents. An example of the data
obtained is shown in Table 41.
, ,. r
1-'
TabLe 41.- LABORATORY RESULTS OF SAND/SLIME TESTING WITH SPLIT AFTER 6 MINUTE ROUGHER
5
miLL discharge
'"l'j
ResuLts % Cu TaiLs l'
O
!GrOup No 1 f--- Heads
Concentrate
1.46
28.10
CycLone SpLit
Product % Wt % Cu ~
Tai Ls 0.55 -J
1-'<
% Recovery 63.6 + 200 m 39.7 0.61
O
..I ResuLts x Cu
- 200 m
TotaL
60.32
100
0.51
0.55 Z
IGroup No 2 ~
Heads 0.55
.. -J
::r
(1J
Concentrate 5.60 O
TaiLs 0.34 SL ime Sand
% Recovery 14.8 Circuit Circuit ~
..I ResuLts % Cu
Results
Heads
x cu
0.51
Groups
No.2
and 3.
Grou~s
No.
and
3.
ResuLts
Heads
% Cu
0.61
::o
(U
IGroup No 3 ~ Concentrate 4.93 Concentrate 4.55 ~

(1J
Heads 0.34 TaiLs 0.26 TaiLs 0.23 g
Concentrate
Tai Ls
0.80
0.25
% Recovery 13.6
I 1
% Recovery 10.2
'"
..I x Recovery 8.32
GLOBAL
.. Run #9 #10
po
;:!
o..
GLOBAL Run #9 #10 O
ResuLts % Cu % Cu ~,
ResuLts x Cu % Cu ResuLts (1)
without a Heads 1.46 1.41 -J

C"O
Heads 1.46 1.41 sand/sLime Concentrate 5.59 12.78
Concentrate 5.67 12.26 circuit TaiLs 0.24 0.20 C.
=:l
TaiLs 0.25 0.21 % Recovery 87.3 87.2 [Jq
The Iourth series of ten experiments consisted of normal flotation for the 12
minutes corresponding to the first two banks of the unexpanded roughers,
folIowed by a sand/slime split and separare flotations in the rernaining cells
01' the roughers. Samples were taken of the plant feed, the combined
concentrate of the first and second groups, the tails from the second group,
concentrate from the third group, and the overall rougher tails. With these
data, standard recoveries for the combined groups 1 and 2, and for group 3,
were calculated. As in the first series of runs, the tails frorn rougher group 2
were sampled and screened to simulate a eyclone split. The + 200 mesh
fraction was repulped to 40% solids, 15 g/t of amyl xanthate was added and a
4 minute float was runo The -200 mesh fraction was also floated for 4
minutes without further addition of reagents. The 4 minute float is intended
to simulate the third group of cells. One example of the data obtained is
shown in Table 42.
19
12 9 1.4 ... 1.5
SampLe point No. Stream

1 Section A feed
2 Combined concentrate groups 1 & 2 section A
3 TaiL group 2 section A
4 CycLone discharge
5 Sand circuit feed
6 Sand circuit concentrate
7 CycLone overfLow
8 Concentrate sLime circuit
9 Concentrate section A
10 Feed section B
11 TaiL sLime circuit
12 Tai L section B
13 Feed section El'
14 Concentrate section B
15 Tai L section B
Fig. 29.- Rougher circuit for Andina's sand/slime mil! test

TabLe 42.- LABORATORY RESULTS OF SAND/SLIME TESTING WITH SPLIT AF~ER 12 MINUTE ROUGHER .....
Combined Groups 1 & 2
t:J
ResuLts % Cu
From baLL miLL discharge Heads 1.59
Concentrate 11.94
I ResuLts % Cu
TaiLs 0.34
% Recovery 80.9 f2
O
IGroup No 1 / - - Heads 1.59
Concentrate 16.30
Tai Ls
~F3
Tai Ls 0.53 CycLone SpL it
I
% Recovery 68.9 Proauct % Wt % Cu
O
+ 200 m 40.20 0.39 Z
"
IGroup No 2 1-- ResuLts
Heads
% Cu
0.53
- 200 m
TotaL
59.80 0.30
100 0.27 F3
I:T"
(ll
Concentrate 4.70
TaiLs 0.34 " O
....
% Recovery 12.0 ::<
I .-
SL ime
Circuit
Sand
Circuit
:;;::J
" (ll
!Group No 3 1-- HeadsResuLts % Cu

0.34
ResuLts
Heads
% Cu
0.30
Group 3 Group 3 ResuLts
Heads
% Cu
0.39
~
a'"
(JQ
Concentrate 1.00 Concentrate 0.84 Concentrate 3.66

TaiLs 0.32 TaiLs 0.23 TaiLs
I % Recovery 1.6 % Recovery 3.3 I I % Recovery
0.15
5.7
e,
GLOBAL" Run #6 #10 o....
GLOBAL "Run #6 #10 (ll
ResuLts % Cu % Cu ResuLts
without a ResuLts % Cu % Cu >-j
(1)
Heads 1.59 1.55 sand/sLime
Concentrate 9.80
Tai Ls 0.32
16.50
0.28
circuit Heads 1.59
Concentrate 9.23
1.55
15.34
g.
'"
(JQ
% Recovery 82.6 83.3 TaiLs 0.','1 0.27
To complete the sand/slime test series, a full plant test was carried out where
the B section was modified and the ten existing celIs repiped to conform to
the flow diagram in Fig. 29. A series of eight runs were made in the plant
during the A shift (8 am to 4 pm). Samples were taken from the points listed
in Fig. 29, using specially design air lift pumps for the pulp samples and
diversion launders for the concentrate samples. Simultaneously, the same
samples were taken on the A section, to evaluate each experimental runo
The testing shown in Table 43 completed the non-sulphidisation part of the

slime /sand study. The milI test showed quite clearly that a sand/slime
separation followed with separate flotation circuits, and the use of an amyl
xanthate kicker in the sands circuit, was a justified permanent investment,
and certainly paid the cost of a milI test, as copper recovery was improved by
over a point, and molybdenum more so. Examination of the distribution of
the copper in the different size particIes of the tails, shown in Table 44,
suggests that a sand regrind could improve recovery; but with an
underground concentrator, Andina did not have the space to add a ball milI
without a very large capital cost.
The second part of this study was the possibility of avoiding the investment in
a sand/slime modification of the plant to see if the non-sulphide copper can
be floated by sulphidisation. The experimental design for the sulphidisation
testing was based on preliminary trials to determine the range of the NaSH
dose and the conditioning time required to complete the reaction. The
laboratory flotations were made using the standard grind and flotation time.
The responses me.isured were sulphide and oxide copper recovery.
Another important variable was the optimum addition point for the NaSH
(Table 45).
To determine the optimum addition point for NaSH in the plant, a set of
runs were made in the laboratory with a plant pulp sample from the feed
distributor box; another set with a tailings sample from the first group of celIs
(i.e., after 6 minutes of flotation); and fmally a tailings sample from the
second group of cells (i.e., after 12 minutes of flotation). The results of
these experiments are shown in Table 46.
These tests did show a definite favourable effect of sulphidisation, but not
enough to discourage the installation of ~ sand/slime separation and
individual flotation circuits.
Table 43.- PLANT TEST OF THE EFFECTIVENESS OF SAND/SLIME SEPARATION 1-'
Group No 1 and 2 common Tails from Group 2 t
CYCLONES
Results % Cu %sol ids ..
II
----.J Str eam % solids % split % Cu %-200 m
Heads 1.67 39.3
Concentrate 19.00 40.3 Feed 38.6 100 0.41 58.2
Tails 0.41 38.7 bott om 62.4 51.4 0.55 25.1
% Recovery 77.1 over head 27.8 48.6 0.31 93.2 'Tl
l'
O
Slimes circuit Slime Sand Sands Circuit >-.l
Circuit Circuit ;J>
Results % Cu %sol ids Results x Cu % sol ids >-.l
Heads 0.31 27.8 Heads 0.55 35.8
5z
Concentrate 2.10 20.5 Concentrate 4.80 17.2
Tai ls 0.24 32.7 Tails 0.31 37.4
% Recovery 3.3 % Recovery 6.0 >-.l
;:l""
OVERALL RESULTS FOR SECTION

.. Pulp OVERALL RESULTS FOR SECTION Pulp
(lJ
e
....
~
Test No 2
Feb 29, 1977 Results % Cu % Mo % solids Results % Cu % Mo % solids j
raw t/h 640
Heads 1.67 0.015 39.3 Heads 1.67 0.015 39.3
g
!)t:
Concentrate 11.60 0.170 26.2 Concentrate 11.60 0.170 26.2 (1l
Tails 0.26
% Recovery 86.4
0.005
68.7
34.0
-
Tails 0.26
% Recovery 86.4
0.005
68.7
34.0
-
a
'.1'
;
i:l
Q..
SEcnON A % -200 mesh screen SEcnON B
o....,
Test # tonnage Recoveries % Cu % Mo Feed bottom overhead % sl ime Recoveries % Cu % Mo C'J
,..,
1 600 90.4 82.7 61.9 30.6 98.3 46.3 88.7 76.7 {1l
2 640 86.4 68.7 58.2 25.1 93.2 48.6 86.4 68.7 ~
3 600 86.6 68.8 62.2 27.7 96.9 49.9 86.2 64.5 5'
4 630 85.4 64.7 65.2 26.2 96.8 55.3 84.6 59.1 ao
5 650 84.9 66.3 56.4 28.0 96.9 41.1 84.5 64.3
6 670 84.0 60.2 54.9 27.5 85.7 47.2 80.3 56.3
7 600 87.3 68.6 63.7 30.4 95.8 50.9 87.0 55.4
8 630 86.6 78.1 53.5 24.1 93.9 42.1 84.6 73.9
627.5 avg 86.4 69.8 59.5 27.5 94.7 47.7 85.3 64.9
TabLe 44.- SCREEN ANALYSIS. corPER DISTRIBUTION, ANDINA A 14,000 t/d, 1977
HEADS CONCENTRATE FIRST GROUP
r1esh Wt % % Cu Wt % of %cumm. Mesh Wt % % Cu Wt % of%cumm.
on Cu Cu on Cu Cu Cu
+65 12.88 0.53 4.43 4.43 +65 0.4811.40 0.20 0.20
+100 10.44 0.65 4.40 8.83 +100 2.1416.10 1.25 1.45
+150 9.36 1.03 6.25 15.08 +150 6.2819.30 4.38 5.83
+200 10.00 1.57 10.18 25.26 +200 11.1822.10 8.93 14.76
-200 57.32 2.01 74.74 100.0 -200 79.9229.50 85.24100.0
TotaL 100 1.54 100.0 TotaL 100 27.50 100.0
TAILS FIRST GROUP CONCENTRATE SECOND GPROUP
+65 13.86 0.48 12.38 12.38 +65 0.90 5.50 0.46 0.46
+100 9.60 0.48 8.58 20.96 +100 1.749.80 1.56 2.02
+150 9.34 0.51 8.87 29.83 +150 3.6811.90 4.02 6.04
+200 8.60 0.50 8.00 37.83 +200 5.7812.90 6.84 12.88
-200 58.60 0.57 62.17 100.0 -200 87.9010.80 87.12100.0
TotaL 100 0.54 100.0 TotaL 100 10.90 100.0
TAILS SECOND GROUP CONCENTRATE THIRD GROUP

+65 16.70 0.44 24.77 24.77 +65 0.88 1.30 1.0041.00
+100 10.98 0.41 15.18 39.95 +100 0.96 3.70 3.18 4.18
+150 10.02 0.33 11.15 51.10 +150 1.503.60 4.83 9.01
+200 8.80 0.25 7.42 58.52 +200 2.94 2.70 7.1016.11
-200 53.50 0.23 41.48100.0 -200 93.72 1.00 83.89100.0
TotaL 100 0.30 100.0 TotaL 100 1.20 100.0
TAILS THIRD GROUP
Mesh Wt % on % Cu Wt % of Cu %cumm. CU
+65 13.58 0.41 25.19 25.19

+100 11.22 0.37 18.78 43.97
+150 10.00 0.25 11.31 55.28
+200 10.26 0.16 7.43 62.71
-200 54.94 0.15 37.29 100.0
TotaL 100 0.22 100.0
The first of the test series is an evaluation of the effect of adding the NaSH to
the flotation feed. For this, two pulp samplqs were obtained under normal
plant operating conditions. One sample is floated under standard laboratory
conditions that simulate plant conditions. The other sample is floated under
the same conditions as the first sample, except that 0.10 lb/t is added to the
flotation cell at the start of the test.
TabLe 45.- THE EFFECT OF CONDITIONING TIME, AND DOSAGE OF NASH, ON corrER RECOI
FIRST GROUP OF EXPERIMENTS - NaSH to PLANT FEED
NaSH Condition % recovery % recovery % recovery
Lb/ton -ing mino oxide Cu suLph. Cu TotaL Cu
o O 2.7 84.7 87.4
0.05 1 3.0 84.7 87.7
0.20 1 3.5 84.9 88.4
0.12 2 3.7 85.0 88.8
0.05 3 3.4 84.2 87.6
0.20 3 3.5 85.4 89.0
SECOND GROUP OF EXPERIMENTS - NaSH to TAILS GRP 1 CELLS

Lb/ton -ing mino oxide CU suLph. CU TotaL Cu
o O 2.9 84.6 87.5

O O 2.7 84.7 87.3
0.05 1 2.7 85.1 87.8
0.25 1 3.2 86.8 90.0
0.15 2 3.2 85.0 88.3
0.05 3 2.7 85.1 87.7
0.25 3 3.5 86.4 90.0
THIRD GROUP OF EXPERIMENTS - NaSH to TAILS GRP 2 CELLS

Lb/ton -ing mino oxide Cu suLph. CU TotaL Cu
o O 1.3 84.0 85.3
O O 1.4 83.9 85.2
0.05 1 1.7 85.4 87.1
0.25 1 1.7 88.0 89.7
0.15 2 1.6 85.3 86.9
0.15 2 1.6 85.0 86.6
0.05 3 1.8 85.2 87.0
0.25 3 1.9 86.3 88.2
The second series is intended to simulate the addition of NaSH in the plant,
after 6 minutes of flotation (i.e., to the tails of the first group of 3 cells).
Again two samples are taken of the pulp feed from the planto One is floated
under standard conditions, and the second has NaSH added to the flotation
cell 2 minutes into the test, which corresponds to 6 minutes in the planto
The third series simu1ates the addition of NaSH to the plant after 12 minutes
of flotation (i.e., to the tails of the second group of 3 cells). Again two pulp
samples of the plant feed were collected using the standard procedures. The
first sample was floated under standard conditions, and the second sample
had NaSH added to the laboratory flotation cell after 4 minutes of testing,
which corresponds to 12 minutes of plant flotation.
The sulphidisation experiment in Andina was less successful than the

sand/slime programme. Examination of the results in Table 45 indicates
that, to be effective as a sulphidiser, NaSH must be added at a high enough
concentration and given at least 3 minutes of conditioning prior to exposure
to the oxygen in the aerated flotation cells. As a consequence, the series
reported in Table 46 gives improved recoveries that are of the sarne order of
magnitude as the sand/slime separation and separate flotation. The data is
nonetheless interesting because of the extensive evaluation of the recoveries
for each of the size fractions. Andina has since increased its flotation
capacity to 40,000 t/d, and has been purchasing a large quantity of NaSH, so
it is probable that an operable sulphidisation pro cess has been worked out. It
also employs the sulphhydrate in the depression of copper in the
copper/ moly separation plant.
The copper porphyry ores in Chile all have about 0.25 to 0.30% Cu as non-
sulphide minerals. As a consequence, with the exception of Chuquicamata,
whose ore is unique, al1 the major concentrators have copper recoveries of
under 85%. The cerrusite content of the gangue contributes to this poor
recovery because it generates colloidal slimes that adsorb on the copper
mineral surface and retard flotation. In this case, if sulphides improve
flotation, the effect may well be to peptise the cerrusite rather than to
activate the oxide minerals or sulphides. There is clear evidence of this in the
Andina data reported in Table 46, where the NaSH addition was more
successful in activating sulphides, especially when added late in the flotation.
CODELCO El Teniente - molybdenite recovery improvement programme
A case history featuring a comprehensive data acquisition and test

programme to improve molybdenite recovery at El Teniente was published
by Beas et al (1991), from which these data are taken.
At El Teniente, partly because thedeclining ore grades and serious rock-

burst problems required pushing equipment capacity beyond its design limits
to keep on budget in copper output, molybdenum recoveries, which normally
ran at about 50%, started drifting down from 48% in 1985, to a low of 34% in
1989 (Fig. 30). The programme undertaken by the in-house metallurgical
staff to find improved reagents, and/or a rrlore efficient processing scheme,
included extensive ore sampling both for head grade scheduling and to
smooth out processability upsets, and provided a good exercise in test
planning and data evaluation. The period described in this section covers
"-'
Te b l e 46.- ANDINA - RESULTS OF TESTS ro DETERMINE ADDITION POINT OF NASH FOR SULPHIDIZATION
&
HEADS
Mesh % tu Mesh % Cu
Flotat1on tests without Flotation test with NaSH edded +150 1. 05 +325 1. 71
addition of NaSH te feed +200 1. 52 -325 1. 96
Flotation Time O te 6 minutes Total tu 1.5810
CONCENTRATE # 1 CONCENTRATE # 1 CONCENTRATE # 1 CONCENTRATE i 1
Mesh - ~ of toT-arper size--Cu Mesn-----"kLij % f--fotaTper size-Cu Mesh % -CU "lo of totaT pe r- size------Cu Me~U"~l p e r- 51:<::! Cu
+150 ~~-9.OIj
Cu recavo
12.5r--
r-ec cve r-v
54.0 +15~57
Cu r-ecov
12~~
r-eccve r-v
53.5
Cu recov.
+150 ~-- 9.58
r-eccver-v
12:07 54.-1 +150
+200
Cu r-ecov .
9.43
r-ec ove r-v
12;7J~--ST.8 ~
+200
+325
14.87
18.59
5.54
7.98
61.7
63.7
+200
+32518.41
14.88 5.61
7.87
61.3
63.9
+200
+325
14.63
18.53
5.48
8.03
62.4
62.8 +32519.49
14.55 4.68
8.50
l.3
61.8
O
-325
TOt.l
21.93
~rE>:i15
33.64
59.67
60.8 -325 22.03
Tata l-----ro. 54
33.58
59~50
51.0 -325 22.17
Toh1---15.49
33.73
59-:tf1
60.7 -325
lotal
22.80
16.67
34.22
60.1""J
60.5
~
,..,
Flotation time 6 to 12 minutes Flotation test with NaSH a.dded .....
CONCENTRATE # 2 CONCENTRATE # 2
to ta.ils of first group
CONCENTRATE # 2 CONCENTRATE i 2
O
~1e5-n-~ of--~r- -s~u Mesh % Cu -% 01 total pe~u ~ u U;;~- -per s t z e tu Mesh % tu % of total pe r- s t z e Cu Z
Cu recov.r-eccv e r-y Cu r-e ccv , r-ec ove r-v Cu r-eccv r-ecove r-v Cu r-ec ov r-ec cvej-y
+150 ~~5~ 5.37-- 23.2 + 150------0.19 - -:~ 23:-2 +150- 5.17 --"-42 -~ +150 6.10 5.47 24.9 ...]
+200 9.49 2.36 26.2 +200 9.53 2.39 26.4 +200 9.45 2. 19 25.9 +200 9.40 2.00 26.3
+325 11.89 3.41 27.2 +325 12.07 3.37 27.2 +325 12.28 3.53 28.4 +325 12.53 3.63 26.4
::r'
(1J
-325
Total
14.10
10.57
14.42
25.56
26.1
-----
-325
lotal
14.22
10.48
14.40
25.48
26.1 -325
lotal
14.33
10.63
14.35
25.49
27.2 -325
Total
14.59
10.71
14.67
25.77
25.9
..,
O
':;<
Flotation time 12 te 24 minutes Flotation test wth NaSH added
to t1111 s of second graup ~
CONCENTRATE # 3 CONCENTRATE # 3 CONCENTRATE # 3 CONCENTRATE # 3 (1l
Mesh % Cu % of total per ~ u Mesh- -------";-CU % of-fci"faT per S:fze-Cu Mesn-r.l:u % oftciEal -pe~u Mesh % tu "/., of total p e ~
Cu recavo t-ec over-v Cu recov. r-eccver-y
-----z;o
Cu r-eccv r-ecove r-y Cu r-ecov r-eccve r-v &(1l
+150 ----:r:lT 0.5~ 2.5 +1~09 0.61 +~3~1~ 2.2 +ISo 2. 0~3- ----0.85 3.9
+200 3.72 0.19 2.1 +200 3.75 0.22 2.2 +200 3.68 0.21 2.5 +200 2.47 0.35 4.6 g
+325 3.69 0.29 2.3 +325 3.66 0.27 2.2 +325 3.71 0.31 2.4 +325 2.28 0.35 2.6 Cfl
-325 3.83 1.97 3.6 -325 3.87 1.94 3.6 -325 3.85 2.07 3.7 -325 3.11 2.51 4.4
Total 3.71 3.02 Iot~_L _ 3 '~,.J.&4 Total -- 3.59 :r.1l7 total 2.73 4.07
Mesh % tu L/ of teta
IAI L5
j per size tu Mesh % tu % of total per- size tu Mesh
tAlLs
%" Cu "lo of tata 1 persfZel:u Mesh tu
TATCS
of tata 1 per- SlZC. tu
o.
+150 0.24
q
Cu r-ecov , r-ecover-y
4.72 ~- 20.3 +15U-<r:25
CU recov.
-.r:og-
r-ecove r-v
10~7
Cu r-eccv , r-ecove r-v
+1:;0-- 0.23 -----.m- ~
L/o 10
+150- -0~.-9o------r:r:4
Cu r-e c ov . recoverv o
l
+200
+325
-325
0.17
0.18
0.21
0.89
0.86
5.28
9.9
6.9
9.6
+2000.170.91
+325
-325
0.15
0.21
0.84
5.44
10.2
6.7
9.2
+200
+325
-325
O .18
0.14
0.20
0.93
0.75
5.15
9.3
6.5
8.48
+200
+325
-325
0.12
0.20
0.20
0.61
1.27
5.19
7.9
9.3
9.2
,..,
(1l
Total 0.21 11.75 Total 0.22 11.88 lotal 0.21 11.6::1 fotal O.la 10.03
~
RUN # 1 OVERALL RECOVERY 88.3% OVERALL RECOVERY 88.1% OVERALL RECOVERY 88.4% OVERALL RECOVERY 90. O~"
a'
rJC
RUN 8 2 OVERALL RECOVERY 88.5% OVERALL RECOVERY 88.4% OVERALL RECOVERY 88.2% OVERALL RECOVERY 89.67-
RUN 8 3 OVERALL RECOVERY 85.6% OVERALL RECOVERY 85.5% OVERALL RECOVERY 85.6% OVERALL RECOVERY 87 .6~"
RUN 8 4 OVERALL RECOVERY 88.2% OVERALL RECOVERY 88.3% OVERALL RECOVERY 88.2"/' OVERALL RECOVERY 89.0%
about 36 months of work between 1988 and 1991. First, a general review of
the plant operations is given, to provide a base line for the studies,
Over all Molybdenurn Hccovery %

5{j --~-_ ..._ --_.-._--- .... _-----_._ ..._ - - - - - - - ~ - - ~ - ~ - ,
ce
ce
40
~ --.--
~,.
o
35 ......
~
1985 1986 1987 1988 1989 1990 Jan Feb March
Fig. 30.- Overall molybdenum recovery at El Teniente
In 1990 the El Teniente Division of CODELCO produced 300,472 t of

copper and 2,633 t of molybdenum. To obtain this output, 32,000 t/d of ore
was mined from the weathered upper levels of the ore body, crushed and
ground in the eighty year old Sewell mill (altitude 2,100 m), and, after
conditioning with sulphuric acid to help recover copper contained in the non-
sulphide minerals, the pulp was transported for 10 km in the old Sewell
wooden tailings flume, which drops down the mountain 800 m to the Colon
concentrator (altitude 1,800 m), where it was floated in a rougher and cleaner
circuit kept at pH=4 with sulphuric acid. Under the same roof, an additional
54,000 t/d of ore from the primary area of the mine (Teniente levels 6 and
greater), were crushed underground, transported by rail to Colon, and then
ground and fed directly to a separate alkaline flotation circuit (maintained at
pH 10.5 to 11.0 with milk of lime) in this millo In 1990, all the ore treated at
Sewell came from the weathered upper levels of the mine, while a little over
60% of the Colon feed was from primary ore sources.
Both Colon and Sewell's flotation processe.s start with a rougher equipped
with 1,500 cu.ft WEMCO flotation cells, folIowed by one cleaning stage
equipped with 1,000 cu ft WEMCOs. The c~ncentrates from these cleaners
are cycloned, the sands re-ground in open circuit ball mills, both cyclone
slime and reground pulps are mixed (the pH obtained on mixing is 8.5 to 9),
then fed to a common final cleaning and scavenging circuit, where the pulp
undergoes a second cleaning (3 banks each with eight 500 cu, ft. cells),
followed by a scavenger bank (one bank with 7 large 1,500 cu.ft. cells), and
finally a re-scavenger (2 banks containing twelve modified 100 cu.ft. cells).
The final tails from this section are returned to the Colon primary rougher
head, and the concentrate is sent to the molyI copper separation planto The
primary tails from the acid flotation are mixed with the tails from the Colon
circuit and fed to the tailings re-treatment plant, where they are scavenged
using 3,000 cu.ft., 1,500 cu.ft., and 1,000 cu.ft, flotation cells without addition
of fresh reagents.
Fig. 31 is labelled "simplified", as pipes and valving are in place to route all or
a portion of the copper circuir tailings to the retreatment plant, as well as the
tails from the first cleaner banks of the Colon and Sewell sections. The
copper concntrate from the copper circuit is then processed in a
moly/copper separation plant using Nokes Reagent (phosphorous
pentasulphide plus caustic) to depress copper. The different streams are
number coded, and their pulp tonnage and Cu assays are shown in Table 47.
For the particular month documented (April 1989), global copper recovery
was 79.8%. Colon's alkaline circuit recovered 8227% of the gross copper
input, and from the acid circuit (Sewell), copper recovery was 71.6%. The
Oyeran cleaning recovery was 96% for copper.
The current copper collector in both the alkaline (Colon) circuit, and in the
Sewell acid flotation circuit, is diethyl xanthogen formate (now Shellflot 203;
xanthogen formates used previously were Minerec A and TM-200l), and the
frothers are Dowfroth 250 and MIBC. As a moly conector, diesel oil is added
to the pulp canal prior to coming down the hill from Sewell, and to the
conditioner in Colon. In both cases the dosage is in the 10 to 20 g/t range for
the diesel oil, 15-20 g/t for the frother, and about 35 g/t for the xanthogen
formateo Prior to 1989, 90 g/t of a mixture (designated T-3010 and patented
by Teniente), consisting of Minerec A or Tecnomin 2001 (60 parts), gasoline
(30 parts), and MIBC (10 parts), was the standard conector. The addition of
gasoline was justified as a molybdenite collector, and the MIBC as an
emulsifier. As well as improving recovery, the mixture resulted in a
significant cost reduction as, prior to 1979, 90 g/t of unmixed Minerec A was
the standard. As frothers, 25 g/t of Dowfroth 1012, or Powell accelerator or
TEB, were used. Because of a very leathery and dry overflow from the
flotation cells with DF-1012 after the changeover to the large WEMCO cells,
Dowfroth 250 was substituted, and is used in the 15 to 20 g/t dosage range.
The addition of MIBC as a modulator has resulted in a more manageable
froth when ore changes occur.
12
1 Colon Se~Jell
pH:: 10.5 -11 pH :: 4.0
---;:gher
flotation
'5\,l 3
,
J eleaner
fiotation
1\ 6 16 J Cleaner
\ flotaton
14
11 r 13
I Rougher
\ flota.tion
Colon Colon 1/ Sl!Wl!ll Sl!Itll!ll
15
4 11 19 111
24 . 29
COll1hinedl\
21
1)-
second Mol!;l/Cu
~
Cleaner V Circuit
21 15 23
Tailings
i'lotation )--
2!.l pH : 12+ 22 Copper
Circuit 1 25
Final Concentra tes

26
.,
Colon Copper Mo l!jhdenull1
28 l{ Scauenger
.-"""'""'-
q Re-
Scauenger ~:
24
3El po
FIHL TIL
Note:Reagent T-3010 = 60% diethyl xanthogen formate mixed with 30%

gasoline and 10% MIBC
Flotation Reagents: (1989) Colon - Collector Shell 203 at 25 g/t
Fue! Oil20 g/t
Dowfroth 250 plus MIBC
(1989) Sewell- Collector T-3010 70 g/t
Fuel Oil15 gJt
(1990) Sewell - Collector Shell 203
\
Fig. 31.- Simplified flotation schematic for El Teniente

TabLe 47.- MATERIAL BALANCES FOR THE FLOW SHEET - FIGURE 31
TENIENTE MATERIAL BALANCES ApriL 1989 ......
U1
t0
units:t/d COLON SEWELL Retreat Tai Li ngs OveraLL Primary Rougher Pri mary CLeaners
PRIMARY FLOAT
Stream number (1) (12) (19) (27) (1)+(12) (2) (13) (5) (14)
Tonnage 53520 35389 4684 86593 88909 60599 37210 8342 3274
% cu Tot 1.31 1.33 20.8 0.31 1.32 1.27 1.29 8.02 10.51
Cu content 701.1 470.7 974.5 266.9 1171.8 771 478.5 668.8 344.1 "'I'j
Concentrate l'
Stream number (6) (16) (21) (29) (25) (3) (14) (6) (16) O
Tonnage 3231 1453 2717 401 2700 7942 3274 3231 1453 >-3
% Cu Tot 19.73 23.19 34.45 7.76 34.65 8.03 10.51 19.73 23.19 ;J>
Cu content 637.5 337 935.9 31.1 935.5 637.7 344.1 637.5 337 >-3
.....
TaiLs O
Stream number (8) (18) (24) (28) (30) (4) (15) (7) (17)
Tonnage 52657 33936 1968 86192 96192 52657 33936 5111 1821 Z
% Cu Tot 0.253 0.394 1.96 0.274 0.274 0.253 0.396 0.612 0.394 ,..,;j
cu content 133.2 133.7 38.6 235.9 235.9 133.2 134.4 31.3 7.2 ::r"
(1J
RETREATMENT PLANT o....

OveraLL baLance :<
Stream number (19) (21) (24) Fraction of RecircuLation retreated :Al
Tonnage 4684 2717 1968 CoLon scavenger taiLs KO=1.00 (1)
~
% cu Tot 20.8 34.45 1.96 SeweLL scavenger taiLs K1=1.00 (JQ
cu content 974.5 953.9 38.6 Scavenger circuit taiLs to CoLon rougher 1<3=1.00 (1)
Copper circuit SeweLL taiL to TaiLings retreat K4=1.00 ,..,.

::J
Stream number (19) (21) (22) Combined CoLon + scav. taiLs to TaiLings retreat K2=1.00 [f1
~
Tonnage 4684 2717 4704 ::J
% Cu Tot 20.8 34.45 14.17 Copper recoveries Percent Q..
Cu content 974.5 935.9 666.5 OveraLL CoLon concentrator 82.69 O
Scavenger circuit OveraLL SeweLL concentrator 71.59 ....
(1l
Stream number (22) (23) (24) Retreatment pLant 96.04
Tonnage 4704 2736 1968 GLobaL for both miLLs 79.84 -J
(1l
% cu Tot 14.17 22.95 1.96 Rougher fLotation CoLon 82.72 ~
Cucontent 666.5 628 38.6 Scavenger - CoLon 95.32 S'
MoLy/cop~er PLant Rougher fLotation SeweLL 71.92 (JQ
OveraLL a Lance Scavenger - SewelL 97.91
Stream number (21) (25) (26) Copper circuit recovery 58.41
Tonnage 2717 2700 17 Scavenger circuit 94.21
% cu Tot or MoTot 34.45 34.65 2.52
Cu content or Mo 935.9 935.5 0.4
Note: CoLon recovery based on adding CoLon + TaiLings retreat concentrate + taiLs from retreat sections
"Stream number" = code shown by Lines in Figure 31
MilI tests - case histories 153
The unique acid flotation circuir ernployed in Sewell since 1926 was imposed
by the ore from the altered cap, which required an uneconomic amount of
lime to produce a stable alkaline pulpo This inherent acidity has been a
constraint on the choice of flotation reagents cver since. Due to the low
natural pH, and the use of sulphuric acid as a modifier, the initial mill
experience in the twenties was the need for 2 to 4 kilos of potassium ethyl
xanthate per tonne of ore to obtain an acceptable recovery, This prohibitive
collector consumption inspired the comrnercial development of acid stable
xanthogen formate as the preferred collector (in 1928) by DI. Arthur Fischer
at Minerec, The acid circuit's extraordinarily high xanthate consumption
justified the continued use of over 100 g/t of exotic Minerec A as the main
collector for the next fifty years.
Molybdenum losses: The experimental prograrnme devised to identify the

cause of the drop in the overall molybdenum recovery, and to help define the
steps necessary to improve it, concentrated on the following known trends:
a progressively Iower content of molybdenum in the ore treated

an increase in the amount of oxidised molybdenum in the ore
a reduced f1otation time generated by an increased ore tonnage
treated in the mill
low floatability of colloidal molybdenite particles
the need to operate with a very high alkalinity of the pulp in the
copper cleaning circuit (pH > 12), because excess pyrite in the
concentrate was reducing the Caletones smelter copper capacity,
requiring a more selective cleaning step.
changes detected in the molybdenite rate of flotation
The laboratory aspects of the experimental programme concentrated on

determining the grind required for good liberation of the valuable copper and
molybdenum minerals, and the re-grind required to adequately reject pyrite;
the skewed distribution of copper and moly as a function of particle sizes;
and the effect of pH on floatability of the mineral species,
The process survey concentrated on identifying mechanical and equipment

factors that could influence moly recovery.
Teniente ore mineralogy: The mineralogical trends in the average ore

received at Colon reflect the reduced importance of the tonnage of ore from
the upper part of the ore body which has traditionally supplied the old Sewell
concentrator. In future, as mining levels aboye Level 6 are exhausted, both
the copper oxides and the oxidised moly will decrease, reducing today's
recovery problems. Fig. 32 shows total copper and oxidised copper trends
Ior .Sewell ores for the period after 1983, and the longer term trends of the
non-sulphide content of the Sewell, Colon and composite Teniente ore are
shown in Fig. 33.
As can be seen from Figs. 32 and 33, the copper and molybdenum grade is
slowly decreasing, as is the oxide copper contento Partial data for the oxidised
molybdenum content has been:
Proportion of total moly as oxide

1988 1989
Sewell 23.2% 17.9%
Colon 27.0% 20.7%
The higher oxidised moly values for Colon, as compared to Sewell, are due to
feeding a significant amount of Teniente level 4 ore directIy to Colon, rather
than to the rougher flotation in the Sewell circuito Mine levels aboye
Teniente 6 are very high in oxides; in one case - Teniente 4 North Standard -
they run close to 40% non-sulphide. To minimise the effect of this non-
floating oxide, blocks with particularly refractive molybdenum (i.e. N-9, N-B,
and N-15) were identified and were dosed to Colon to be diluted with
primary ore.
159
1.48
14 . ......: 4~ H!L ~ 43 : 40 H9. 1-,35 L32 28
1.2 ,.:., ,."....... . . _R"1'-J,+f I,++- 24

.,... ,,
I
.,...
c.
c. ....
--.... 20 C.
C.
o ....
.... o
o1: ,.,}
.......
1Il
................. ............ 16 .,

.,e
0.8 . ~
..... -. ::
... "C
x
t 0.6 I:,:I..j:,f ..............1':'1 + 12 o
ll.. rf.
0.4 . I:::II:::Ij::f 1:::1+ 8
O:-j--~~~~~~III ~.~
. ~ ]mlll~ 1-- 4
o
1~ 1~ 1~ 1~ 1~ 1~ 1~ 1~ 1~
Year
I~ CuNS rn CuT ..... % Oxide Cu I
Fig. 32.- Sewell ore CuT, CuNS and % oxide copper content
0.<,---------------------,
Combinad
...............e; ;;;;::::::::::::::.:::::::: -..- : ".?.~~~.~ ,"", .
1987 1988 19851 1990 1991 1992 19513 199L 1995 1996
Yaar
Fig. 33.- Sewell, Colon and combined ore non-sulphide assay

1987 to 1996
120,------------------------,
100
80
~
w
60
o::
W
Il..
40
20
o
Dec Feb Apr Jun Aug Oet Dec
Jan Mar May Jul Sep Nov
Fig. 34.- Mineral content of Teniente concentrate - 1984
The mineralogy of the fmal concentrate w~s studied using monthly samples
that covered most of the year 1984. The main copper mineral recovered was
chalcopyrite, which as can be seen from Fig. 34, was predominant over the
secondary copper minerals: chalcocite, covellite and bornite.
DAILY TOI'iI'iAGE
PERCEMT MOL~BDENU~ t-thousands
O.!H 1[m , r - - - - - - - - - - ,
gross "01"
0.035
80
0.03
0.025 60
0.02
0.Oi5 '.10
0.01
20
0.005 non-sulphide 1001'1
0.0 o
Oct Dec Feb Oct Dec Feb Apr
Fig.35.- Monthly moly assays and tonnage produced by sectors - 1988/89
Colon Concentrator Sewell Concentrator

,------ --,0.035 , - - - - - - - - - - - - - - - - - ,
0.0
Jan Mar May Jul Jan Mar May Jul
Fig. 36.- Monthly moly assays for 1989 for Colon and Sewell
To be able to correctly evaluate the laboratory data collected during 1988

and 1989, it was necessary to extensively sample the different areas being
worked in the mine to obtain sorne idea of the short term variability in the
minerals reaching the two concentrators. Fig. 35 shows composite monthly
assays for two important sectors of the mine, and the monthly tonnage mined
in 1988/89, from all important mine sectors, during the main course of this
study. The average total molybdenum assays of the ore fed and tails obtained
each month in Colon and Sewell, covering the same time period, as well as
the non-sulphide moly concentrations, are shown in Fig. 36.
Mil! tests - case histories 157
The effect of processability of different sectors of the mine was also testcd
with sufficiently large samples to perform normal industrial size laboratory
flotations using Teniente's standard procedure. The raw data is shown in
Table 48.
TabLe 48.- FLOATABILITY OF SAMPLES FROM DIFFERENT SECTORS OF THE MINE

% Recovery on Roughing
Mine SampLe CuTot MoTot CuTot MoTot Acid - ALkaLine

Sector No Acid ALkaL ine CuTot MoTot
TTe4 1 81.61 46.23 77.57 44.56 4.04 1.67
Fortuna 2 67.80 47.15 59.05 31.57 8.75 15.58
3 89.45 34.07 84.06 32.89 5.39 1.18
TTe4 4 67.63 69.47 62.65 65.82 4.98 3.65

North 5 81.43 81.39 78.42 84.37 3.01 -2.98
Standard 6 89.16 86.39 88.86 87.56 0.3 -1.17
7 81.20 80.90 76.43 80.82 4.77 0.08
8 78.84 75.01 74.33 79.99 4.51 -4.98
9 51.86 28.44 48.62 24.29 3.24 4.15
10 33.39 24.90 35.25 35.25 -1.86 -10.35
11 33.08 24.58 36.47 30.28 -3.39 -5.7
12 18.97 20.31 20.61 35.87 -1.64 -15.56
TTe3 13 97.09 86.22 96.60 90.26 0.49 -4.04

IsLand 14 93.45 87.05 92.78 86.99 0.67 0.06
15 92.95 82.34 93.12 81.12 -0.17 1.22
16 96.84 92.18 97.24 90.21 -0.4 1.97
TTe5 17 92.64 90.59 92.10 87.87 0.54 2.72

PiLLars 18 94.71 81.82 92.54 81.80 2.17 0.02
19 93.00 86.26 92.14 89.70 0.86 -3.44
20 93.28 90.08 93.76 89.61 -0.48 0.47
Weighted avg.
recoveries 88.2 86.769 86.2 84.7415 2.00 2.00
The most interesting data obtained from this very extensive floatability test
programme based on mine sector samples is the difference between the
recovery obtained from an acid pulp and an aIkaline pulp, rather than the
absolute value of the recovery, whose representativeness is difficult to
evaluate a priori. These differences have been plotted in bar form in Pig. 37,
which also shows the sector source for the different groups of samples. Prom
these data it is obvious that Teniente level4\:lre contains extremely refractory
copper and molybdenum, while Teniente levels 3 and 5 are normal. Por these
sectors, Table 48 shows absolute recoveries for copper aboye 90%, and for
molybdenite aboye 80%. From the differential recovery values, the optimum
process for the easily floated ore in levels 3 and 5 is not obvious, though for
copper there is a slight tendency in favour of an acid environment.
158 PLTATION - Theory, Reagents and Ore Testing
RECOVER'l DIFFEilENCE EACH SAI1PLE

ACID-ALKALIHE FLOYATION TEST
COPPER RECOVER'l MOL'lBDENUH RECOVER'l
20~-----'-::":'-----------'
Acid recouep~ bettep TTe'1 TTe4 TTe3 TTe5
- 1 f r ! f - - - - - - - - - - - - - - - - / -11lJt-----t----ae- I - + - - - - - - "
AIka line ",ecoue"'!J be1:1:e",
- 1 s - 1 f - - - - - - - - - - - - - - - - / -'w--+------Ut---+---j
1 3 5 7 9 11 13 15 17 19 1 3 5 7 9 11 13 15 17 19
2 4 6 81E!12 14 1618 2E! 2 4 6 8 ra 12 14 16 18 213
Pig. 37- Ploatability of samples from different sectors of the mine
It is ironic that twenty years ago CODELCO management opted to design

Colon as an alkaline flotation mill to avoid dependence on xanthogen
formates as collectors because there was only one producer in the world, and
because, supposedly, xanthogen formates were useful only in acid circuits.
Today diethyl xanthogen formate is again the sole collector because it is
unique in an acid circuito The only selective competitor in alkaline flotation,
isopropyl thionocarbamate (Z-200), was used for a number of years in the
pH 10-11 Colon mill, but now has been judged not to match the recovery of
xanthogen formate collectors (at least with Teniente ores) at any pH.
Por Teniente level 4 samples, copper recovery is better in acid circuits, and a
higher pH (10-11) favours moly recovery. The moly recovery problem at
Teniente is losses in the cleaner circuits and not prirnarily molybdenite losses
in the rougher flotation. The month by month moly recoveries for Colon and
Sewell, shown in Fig. 36, suggest that moly recoveries from acid pulps could
be better than standard basic pulp flotation.
From Pig. 37, note that certain mine sectors are not floatable, particularly in
Teniente level 4.
Performance of the copper cleaner circuit: The final concentrate cleaning

plant operating conditions are chosen to minimise moly losses, and maximise
copper grade; but because of the design of the smelter, Teniente also has to
have very tight specification on insol and pyrite content of the final copper
concentrate, a problem that has become more serious because pyrite content
of the ore has increased. Meeting these requirements has forced regrinding
the primary concentrate to assure pyrite and insol liberation, and has had as
an unavoidable side effect the over-grinding of the moly fraction (as can be
seen in Fig.s 46 and 47). This operating procedure, causing over-grinding,
has been continued because the value of the moly lost, at current prices, is
lower than the potentially lost copper production. It has also resulted in
raising the flotation pH to a maximum (pH greater than 12), to help depress
pyrite, des pite the possibility that the very high a1kalinity of the concentrate
triggers soluble moly losses in the tailings as well as providing more calcium
to depress floatable moly (see Fig. 42). If the pH in the second cleaners is
dropped from 12 to 10, the moly lost in the tails halves, but the iron in the
final concentrate rises from 20 to 25%, and the copper grade drops from 32
to 28%.
To check the performance of the "copper circuit'' whose flow diagram and
equipment is shown in Fig. 38, particularly to obtain data on the rates at
which moly and copper floats in the Colon cleaner circuits, plant tests were
mn in which samples of the concentrate from each of the flotation cells in the
second cleaners, scavengers and re-scavengers were taken and analysed. The
results are shown in Fig. 39. These data can be interpreted assuming that, as
the residence time in each bank of cells is essentially the same for a11 the
flotation cells, the copper and moly content in the concentrate samples is a
measure of the flotation rates.
The first pair of graphs show the copper and insolubles assay in the
concentrates in the first panel, and moly grades in the second panel, during
normal operating conditions. From the shape of the copper curve, which does
not show a drop in concentrate grade by the last cleaner cell, it is obvious
that the rate of Cu flotation is too slow for the design cleaner flotation time,
so a very significant amount of copper is carried through to the scavengers
causing a large recirculating load (67% of the total recovered) from the
scavengers and re-scavengers, though overall copper recovery is not adversely
affected (94.2%), even when it continues to float vigorously through ce11 #5
of the scavenger bank. Molybdenum, on the other hand, is being severely
depressed in all the cleaner bank, and only starts to float after the fifth cell in
the scavengers. Note that the moly depression is such that the Mo grade of
the cleaner tails (i.e. the feed to the scavengers) is 1%, nearly double the
grade of the pulp fed to the first cleaner ce11 which contains 0.6% Mo. The
fact that moly only starts to float in the scavenger when the amount of copper
floating is dropping suggests that one factor, contributing to moly depression
is crowding out by the copper.
Colon 6 ~ 16~ Sewell ~PEn CI!!.CUIT - CO~

pH= 11 pH= 4 Reagents Colon Shel l 2lB 25 g/t F.O. 2B g/t
Seuel l T-3lJ11l, 70 g/t F.O. 15 g/t
Scavengers
1 Second
7x1,506 cu ft 2 Sca uenger-s :
ce lIs 12xlllll cu fi
Concenhate 3
3 lllJ500 ft 3 ce l ls , in 2
21 +------- 4 Second 4 banks
to He 2,41l0 ft3
Mol~ plant 5 Cleaners 5
3
llx15BB ft 6
6
cells
7 12JIlllll rt3 7
B 8
9
10
11
Reagent T-3010 = Shell 203 nxed uith
12
gasoline
~ 24
Fig. 38.- The alkaline cleaner flotation circuit at Colon
The second pair of graphs show the results of a plant test in which the Sewell
regrind ball milI was stopped. The rate of flotation of copper increased,
though the number of cleaner cells available appears to be barely adequate,
as the first scavenger cell still produces a concentrate with a specification
copper grade. The tail from the cleaners dropped from 18% to 11%,
signalling the higher flotation rate obtained.
The shape of the molybdenum flotation curves are improved, but because of
an ore change that doubled the grade fed to the second cleaners, the data
cannot be compared with the next plant trial. As the copper concentrate
coming off the head of the scavenger bank is nearly at specification, this
suggested the second plant trial in which, as well as shutting down the Sewell
regrind mill, the concentrate from the first three cells of the scavenger,
instead of being recirculated to the cleaners, was piped directly to the P-4
thickener which feeds the coppery'moly separation planto Under these
operating conditions the copper response is essentially normal, with a fully
functioning second cleaner bank. Note though that the second cleaner tail
now is at 4% Cu, as compared to 11% when cells 1 to 3 of the scavenger were
being recirculated, and 18% when the Sewell regrind was operating. The
moly flotation is marginally better, with recoveries impraved as the scavenger
Copper and insol concentrate grade Moly grade in concntrate
:f
55
2nd cleaner scavenger>
=
....-
""
Jr
""
~-l\
:LO
4b copper
M'
~~-r~ ..,'
~ :~_ cleancr 2-'
cleaner
~?Il -t e il ')1 .-
~
X
,
1
~~:~e
Insol
=_.. . ~~ __ ~copper . t.
o..;
head
grada
t-ail
cauenger
n __ ft. . . . _ . sCl\uenger ~\...- i:.ail
"tall
01234&67& (J.1234557 012~4557091CDn2 "
Cell nu.....her Cell nUMber
Sewell re-grind mill stopped - scavenger recycled to second cleaner

70
tl Il
8: Ineo I
.= : copper ~ 'ti.1""!li.~Qi1l
:;'R
-i'
O "" .'1-
J
K
" 35
..:; 00
'g :15
i'll
'" "
ax :~
l
~
Insol J
..e . . . c "
5
O
01 :i!30458TB 012:!461llT 012:J4!il$Tag10tl2 C12!4S67B 012345~7 01234liB7n.Gl0n2
Ce Lf -number- Cell nUlI1ber
Sewell re-grind mill stopped- scavenger cells 1- 3 to moly plant

7 5...,.----,--------,--------,
~ Holy
e
=
O
o Inso l
!:: <>'\,...... .ql!l A
.....
-..:;"
o
'"=
U
'"
:::l
Inso 1
p
rf copper 1.
x ....."...
lO'"
01 2 a 45 B 78 D1 234 5 B7
rr":
o f 2 3 450 7 o9111112 012345670 01234567 0123456HUlOn2
Cell nu.. ber- Cell 1l11 .. ber-
Fig.39.- Performance of the copper cleaner circuit

162 FLOTATTON - Theory, Reagents and Ore Testing
tail now is at 0.3% Mo, compared to 0.45% at normal operations with both
regrind milIs going.
Effect of lime as a depressant of molybdenite and pyrite: It is well known that,

particularly with by-product molybdenite, highly alkaline flotation conditions
reduce recoveries. Shirley (1979), in a review article, points out: "The
sensitivity of molybdenite to lime (CaO) usage was first noted at the huge
Utah Division concentrators of Kennecott. Their operational procedure was
modified to operate at the lowest pH possible consistent with maximum
copper recovery. Duriog a plaot test at Chuquicamata cooducted on half of
the flotatioo sectioos, it was ooted that wheo the normal operatiog flotatioo
pH was lowered to 9.6 from 11.3 a coosiderably higher recovery of
molybdeoite was obtaioed."
Chaoder and Fuersteoau (1972) have shown that in an alkalioe media,

washed molybdeoite has a high oegative zeta poteotial (about -43 millivolts),
with coocomitant low flotatioo recovery (circa 30%), while reducing the pH
to the 3.5 - 4.5 range will reduce the oegative zeta poteotial by over 10
millivolts (to -30 mV), at which point the recovery jumps to 70-80%. The
Tenieote data io Fig. 36 confirms that lower pH during flotatioo improves
moly recovery; recalculated and plotted versus the oxide moly cootained in
the ore (Fig. 40), it shows that oxide cooteot reduces recovery. An increase
io oxide reduces recovery more markedly in an aIkaline eovironmeot, but the
data show that moly recovery is better at all oxide levels in an acid circuit,
presumably because the molybdeoite surfaces are cleaner in a media
consistiog of very dilute sulphuric acid, andjor because adsorbed calcium
ions are removed from the surface, and because moly solubility is reduced.
The stroog effect of calcium ions 00 the depressioo of pyrite has beeo known
for sorne time, but surprisiogly little flotatioo data confirmiog this is
available. Cea aod Castro (1975) published Hallimood tube data that clearly
shows the multiplyiog effect of using lime rather thao caustic soda or soda
ash as the source of alkalinity. This confirms Glembotskii et al's (1963)
earlier observatioo "a comparisoo of the effect of Ca(OH)z with that of
NaOH 00 pyrite shows that lime is the more powerful depressant". lo Fig.
41, a comparisoo is shown betweeo the effect of pH 00 pyrite recovery, as
measured by Li Guoogming et al (1989), with that of caustic and lime 00
molybdenite flotatioo, measured by Castro and Paredes (1979).
Mill tests case histories 163
55
A
50
.. ~ ......... ,b., .Do.
'<>"'
e,
>
45 ... ,.,,;., ".!.- , ,
"jj""".""."~ ,~,;o""",.,,,,,
o 40
o
e
C<: 35 SEWELL ACID
30
e
e 25
] 20 ""'' .':.... ''''."'''''''".,,.,', ... ,
~ 15
:;;: COLON AlKAlINE
10
;;-<
5
O
o 5 10 15 20 25 30 35 40 45 50 55
Percent Oxide Content
Fig. 40. Moly recovery Sewell and Colon as a function of

oxide content
Effect of pH on P!:Ir-te and arseno- Effect of pH and CaJciu~ on

pyrite recouery by flotation ~olybdeni~e recouery
11111
......- ....-=,.;:-;.;-:.;:... - - ...... Bll

\
Pyrite
\
.
\ 611
411
211 1--
II
5 11 13 5 7 9 11 13
pH
Fig. 41. The effect of pH on pyrite and molybdenite recovery
The combined pulp feed to the copper circuit (which contains 25%
chalcopyrite and 30% pyrite) was submitted to bench scale flotation tests
performed at half pH unit intervals. Copper, moly, pyrite and insolubles
assays were made for all tests. The results are shown in Fig. 42. They
confirm that chalcopyrite is not significantly depressed up to a pH = 11. They
also confirm that pyrite starts to depress at .pH 11, but indicate that even at
pH = 12 there is far from complete depression, and that depression of
molybdenite, as a function of pH, in the pre~ence of calcium ions, is greater
than that for pyrite. The behaviour of the insolubles is indicated by the grade
recovered in the concentrate rather than by recovery. From these figures one
can conclude that there is no significant depression of insol in the pH range
Flotation Hecovery Versus pH
99-
98
"'"--- ,. --';;'--e--_
Unreground Sewell -1- Colon Concentrate
100,------------------------,
Copper
~--""--EJ
ce
97
96
""---"""""'=-
~1""'''' .... "" ~h"" .. ", ..
Moly IRecover'} I
....... --lt....... ... ""'~~,
e
Q)
95
>
o 94 Pyrite " ,'.,,'"
~
'-'
Q) 93
o: 92 ,, ~
e
o 91
.~
(5 90
w::: 89 16.4---15.9---16.3-16.5---15.8-15.9_15.7~16.1_15.7
88
87
Concsntr cla Percentooe Insoluble Content
86
85...L--,---,-------,----,-------,---,--,----,---,-----'
8 8.5 9 9.5 10 1 0.5 11 11.5 12
Flotatlon pH
Fig. 42.- Effect of pH on Cu and moly recovery from

second cleaner feed pulp
studied. The obvious alternative of testing soda ash or caustic as the modifier
reagent, to eliminate the effect of the calcium ion, is impractical because of
reagent carry-over from the rougher flotation at Colon where, with the high
ore tonnage treated, the cost of substitution of lime as the pH regulator is not
justified by the additional moly recovery. Similarly, opting to operate an acid
rougher in a concentrator designed for alkaline flotation requires expensive
acid proofing.
Moly Content of Tailings Water

2
,
... :1
, "
E
8
6
,,
" o, o'
" ,
-o .- 00
" ,-
"-
c.
4 - -
2
I
o
I I
Jan Fob Mar Apr MayJun Jul Aug Sop et NovDocJan F ob Mar Apr
1988 1989
I --~ Colon - . - Sawall I
Fig. 43.- Molybdenum content of Teniente tailings water

Lime, ancl high alkalinity, also reduce moly recovery because oxidised
molybdenite reacts at very high pHs with calcium ions to form powellite, or,
in the presence of limonite, the surface compound forrned could be
Ierromolybdite [3(FE:P3).Mo0 3.8HP]. Both of these are water soluble
and can be lost in the tailings water. Fig. 43 shows that this happened in
1989, when the moly content in the Colon tailings jumped up. At 12 ppm the
losses in solution in 1989 would have amounted to about 300 tonnes of Mo.
CIMM (Chile's Mining and Metallurgy Institute) developcd the

electrochemical phase diagram (Fig. 44) for the system molybdenum -
sulphur - water, which shows that pH 12 is beyond molybdenite's stability
region, It suggests that aboye pH = 11, molybdenite is not stable, thus,
partial oxidation at that pH can form a hydrophyl1ic surface making the moly
particles unfloatable in the presence of calcium. The reactions are:
Mo03 + Ca (OH) 2 .---------. CaMo04 + H20 Kps = 4 .2 X 10- 8

PoweLLite
for dissoLution, and
for the precursor reaction for the formation of ferromolybdite, the water avid
compound on the molybdenite surface.
1,----"""",---------:~----------------____,
ACTIVIT~ OF DISSOLVED SPECIES

0.8 -f,
113 M
0.6 HMoO~
0.4
0.2
-1 +---,-----,---,--+-----,------,-------,-----,-----,---''''+
3 5 7 9 11 13
pH
Fig. 44.- Potential-pH diagram for the moly-sulphur-water system

The CIM1\tf report further suggests that the mechanism of calcium

depression can be explained if the oxide layer on the pyrite consists of an
undercoat of Fe(OHh and FeC03 , on which there is a coating of a porous
layer of Fe(OHh The calcium ions would adsorb on to this latter porous
layer impeding adherence of the collector on the pyrite. A similar
mechanism has been put forward by Tsai et al (1971) and Shimoiizaka
(1972), in which only OH groups on the mineral surface are necessary for
calcium ion adsorption. Presumably a ferromolybdite surface can undergo
similar reactions and adsorb calcium ions.
Liberation 01 Copper and Moly: In June 1989, to check on possible over-

grinding, the Colon and Sewell concentrates were carefully sampled, as well
as all the major streams in the two regrind mills whose combined output feed
the copper circuit (second cleaners et al). The standard cleaner circuit and
operating conditions are shown in Fig. 45, and the results of the liberation
tests in Table 49.
STANDARD OUERALL CLEANER CIRCUIT PERFORMANCE

June 1998
First Cleaners COLON First Cleaners SEWELL
Vd = 2934 Vd = 2559
Slillles
682 Vd 1328
Vd = 5493
Cu = 23.87y. COPPER CIRCUIT
.. 23B7 Vd TAIL
9
MB=B.67::.o: Cu = 6.68x
=
Mo B.35::.o:
COPPER CIRCUIT RECOIJER lES x
Cu = 88 .By. Mo = 78 .Bx 3186 t/d CONCENTRATE
Cu = 35 .88y.
t10 = B .9Bx
Fig. 45.- First cleaners and copper circuit flow diagram

For both Colon and Sewell liberation of the molybdenite prior to regrinding
is practically complete - 99.0 and 99.4% respectively. In the case of the
copper minerals, there is a moderate to marginal improvement in liberation,
which could justify regrinding - 92.8% to 96.8% for Colon and 96.8% to
97.2% for Sewell - but the critical factor in concentrate specifications is pyrite
and insolubles. The greater sensitivity of the Sewell ore to form colloidal
slime, which are notoriously refractory to flotation, is due to the softer nature
of the ore. This shows up in the power consumption figures of Table 49 -
4.88 kWhjt for a reduction ratio of 1.68 for Sewell, as compared to 8.11
kWhjt consumed for a reduction of 1.79 at Colon.
TabLe 49- MATERIAL BALANCE AND SCREEN ANALYSIS RETREAT REGRIND

coLon CLeaner Concentrate
Feed to CycLone Regrind CycLone FinaL
CycLone underfLow MiLL out overfLow Product
Dry Tonnes/day 2,934 1,614 1,614 1,320 2,934

Fv PuLp (m3/d) 14,821 1,247 1,247 13,574 14,821
% soLids 17.32 67.11 67.11 9.08 17.32
% -325 mesh 57.35 26.64 51.37 90.77 69.09
D80 (micron) 93 143 80 33 59
% Liberated cu 92.76 96.84
% Liberated Mo 98.86 99.45
MiLL reduction ratio R80 1.79

MiLL Energy kWh/t 8.11
SeweLL CLeaner Concentrate

Feed to CycLone Regrind CycLone FinaL
CycLone underfLow MiLL out overfLow Product
Dry Tonnes/day 2,559 1,877 1,877 682 2,559

Fv PuLp (m3/d) 6,738 1,447 1,447 5,291 6,738
% soLids 29.58 65.92 65.92 11.75 29.58
-325 mesh 44.79 29.62 54.39 86.31 62.90
D 80 <micron) 102 124 74 40 65
% Liberated Cu 95.75 97.21

% Liberated Mo 99.36 99.16
Mil L reduction ratio R80 1.68

Mil L Energy kWh/t 4.88
In late 1988, CIMM had performed a similar size distribution study as a part
of the design of the moly plant rougher. The data showed that the moly
fraction of the concentrate fed to the moly plant contained 41.9% -13
micronparticles, while the CIMM study showed that the extremely fine moly
in the sccond cleaner concentrare amounted to 64 Gl -13 micron particles and

the copper fraction contained only 3St;if, -13 micron. Based on these studies,
in 1990, ro ameliorate over-grinding, the ball charge was reduced from 40%
to 30% of thc volume in both milIs.
10(h----~--~--~--~--~----,I~
90
A ......... \
................................'''".,''
I
80
e
~
8 70 ..- .
<Il
a::
60 1 .~ ;.
50 . . ------.-- ..__ .; ~. . - _--- .
60 65 70 75 80 85
% Passing 325 Mesh Tyler
Fig. 46.- Effect of grind on copper and moly recovery
The data in Fig. 46 show that regrinding second cleaner feed can improve
copper recovery, but at the expense of moly losses. Considering the
depressant effect of any type of slime on sulphide mineral flotation, picking
the optimum degree of comminution becomes a very difficult exercise. The
problem in choosing a moly flotation system is seen in the studies made of
the distribution of the molybdenite as a function of particle size. The result
for Sewell and Colon ores is shown in Figs. 47 and 48.
Sand slime separation: An obvious alternative studied in the flotation

laboratory, using pulps and cyclone overflows sampled during plant
operations, was simulation of a sand slime circuit. The reground sands, whose
assay was 21.14% Cu and 0.75% Mo, on flotation in the 2 litre laboratory
machine yielded a concentrate grading 33.3% Cu and 2.09% Mo, and tails
with 14.8% Cu and only 0.054% Mo; i.e. a Cu recovery of 54%, and a
95.28% moly recovery. The slimes, which assayed 23.86% Cu and 0.94%
Mo, yielded a concentrate with 29.55% Cu and 1.06% Mo, and tails with
11.7% Cu and 0.68% Mo; i.e. a Cu recovery of 84.37% and moly at 76.9%.
Globally these recoveries add up to only 95.94% on copper and 87.72% moly.
Compared to the standard circuit (Fig. 45) the sand slimes process
improved moly recovery significantly, but showed much poorer copper
recoveries because of losses in the sands cleaner. Regrinding the sands was
not followed up because of the rapid drop off in fine moly recovery.
Post moiy/copper separation pytite scavenger: A serious look has been grven
to removal of the excess pyrite after recovering the molybdenite at an
optimum pH throughout the Sewell and Colon flotation process. The
proposed final cleaner circuit is shown in Fig. 49. The advantage in moly
recovery of operating the second cleaners at the rougher pH, or lower, is
obviously a greatly enhanced molybdenite recovery. The disadvantage is also
obvious - the need for a considerably expanded moly/ copper separation
plant, as there are already problems of moly losses due to overloading the
existing plant.
Moly/copper separation plant studies: To compensate for the reduced

flotation time generated by the increased ore tonnage treated in Colon, an
additional circuit (A'), consisting of two banks of twelve 100 ft3 cells, was
added to the moly/copper separation rougher circuit (A). Studies were
made by outside consultants (CIMM) to check the operation of individual
cell banks in the moly plant. It was found that, in the roughers, 26 to 32% of
the fine particles of molybdenite were short-circuiting the flotation zone due
to the fact that the tank of the original 60 ft3 cells adapted for use in the moly
plant had been modified to 100 ft3 , after which they were prone to sanding.
Increased cell maintenance cleaning and raising the impeller speed from 230
rpm to 300 rpm improved the operation. A greater impeller speed, as
suggested in a papel' by Malhotra et al (1977), improves flotation of fine moly
particles. Increased processing time was also obtained by the addition of
two square 36"x36" column cells, and a 36" diameter round cell to the
molybdenite circuit.
70'----------------------,
o>e
60
'E
~ 50
o

,., 4Q . .
'O 30
:2 ..
Cij
20
' 1
u:
o-JL--.II-.,..L........u......,..Ib..,..I:J!I.,.IlJ!1,.llI.,..Il..I!I.,..IlJll,..I...llI.,..Il..IJlL,.IUlL-....,J
417 208 104 50 25 13
295 147 74 \33 18 -13
Particle size tracton - microns
[iIi"COlon Fresh Feed D_C_o_lo_n_c_on_c_. Colon Reground C_
Fig. 47.- Moly content vs particle size - Colon before and after regrind
70
o~
60
i 50
Eo
'-' 40
,-
"O ...... _-- ...-.......
:2 30
c:
e 20
o
10
u::'"
O
417 208 104 50 25 13
295 147 74 33 18 -13
Partid e size fraction - microns
Sewell Fresh Feed D Sewell Conc. lllIl!IlI SeweJ Reground
Fig. 48.- Moly content vs particle size - Sewell before and after regrind.
6E1x to Colon
rou!Jher
2nd Cleaners Scauenger Re-scauengel'

pH=9 pH=9 pH=9
Colon
Sewe II ' < : - - - - - - 7
Conc.
Copper/Moly Pyri1:e Cleaner

separat.ion Re1:urn to Tailin!J
pH=12" Re1:reat~ent plant
p Iarrt;
MOL!>' CONCENTRATE COPPER CONCENTRATE

TO SMELTER
Fig.49.- Proposed new Colon cleaner circuit

Depressant studies - Holees reagent optimization: The reagent employed in the

Colon moly plant is designated LR-744. It is made on site by the reaction of
phosphorous pentasulphide with caustic soda in a water solution, in the ratio
of 10 parts by weight pentasulphide to 13 parts NaOH. The mechanism of
depression of Nokes reagents and NaSH was studied by Castro and Pavez
(1977), who found that the direct effect of these reagents is a large reduction
of the pulp potential, Eh.
A May 1989 Teniente report on the effect of reagent dosage on moly

recovery and copper depression can be summarised as fol1ows:
Moly plant rougher flotation - increased dosage of LR-744 (1,500 g/t to 2,000
g/t) will improve total molybdenum flotation slightIy. In laboratory tests, with
higher levels (2,500 to 4,500 g/t), moly flotation is slightly impaired, possibly
because there is physical occlusion of moly with the increased depression of
the copper minerals, Moly. selectivity is directly proportional to LR-744
dosage, though the depressive effect of high doses restricts this route to the
production of a molybdenite concentrate with a low Cu content. Typical
laboratory data show:
TabLe 50.- EFFECT OF NOKES REAGENT ON ROUGHER RECOVERY

Feed rate LR-744 Rougher Recoveries
grams/tonnes CuT MoT
2,500 20.9% 91.1%

3,500 18,5% 89.5%
Moly plant first cleaner stage - up to 300 g/t of LR-744 can be fed without
reducing MoT recovery, but these amounts of depressant provide only
marginal improvements in CuT rejection. Greater additions of LR-744
increase se1ectivity, but at the cost of a significant drop in MoT recovery, viz:
TabLe 51.- EFFECT OF NOKES REAGENT ON FIRST CLEANER RECOVERIES
Feed rate LR-744 First cLeaner Recoveries

grams/tonnes CuT MoT
o 37.1% 81.8%
300 30.6% 81.2%
800 26.9?- 79.3%
These data underline that excess depressant should be avoided in flotation

stages where the primary function is molybdenum recovery.
172 FLOTATlN - Thcory, Rcagents and Ore Testing
Use offlotation colul1llls in the 1II0(v plant - Iaboratory tests of columns in the
cleaner stages gave the following results:
TabLe 52.- MOLY/COPPER SEPARATION CLEANER TESTS WITH COLUflNS

First CLeaner First CLeaner
%CuT %MoT %CuT %MoT
Feed 37.0 1.44 34.3 2.92

Concentrate 24.2 15.2 17.1 24.9
Recovery MoT 74.6% 93.3%
Depression CuT 95.4% 94.5%
Second CLeaner Thi rd CLeaner

%CuT %MoT %CuT %MoT
Feed 20.35 23.32 7.21 43.32

Concentrate 7.7 42.40 3.98 49.83
Recovery MoT 92% 83%
Depression CuT 81% 60%
Plant tests also have eonfirmed that the use of eolumn cleaner stages is
eeonomieally very attraetive in the moly/ eopper separation proeess.
Teniente is therefore going to modify the Cu/Mo separation plant.
Flotation recover~ us pulp potential Flotation recouer~ us pulp potential

Colon or-e , Dieth~1 XF, pH 12 Roughel:' Sewe l l , isobu t~ I xanth. pH 4
100,----------,--------,
00 00
llIJ 1lO
2:- ro ~ 70
o 1:
~ 00 ~ 00
u
v
JI:
8
eg so ~ .so
e
~
40
S .w
~
u, ;lO
Standal:'d ;JO
ro operating
20
conditions Standard
1~ 10 operatin!J
o " NaSH 1IMn04 conditions
-400 -:lOO -200 -'00 o '00 200 lOO ~ 500 00Il ~ 150 2511 35(] -450
Pulp potential (~U) Pulp potential (~U)
Fig. 50.- Redox mcasurements of rougher flotation - Teniente
As part of this molybdenum flotation improvement programme, a study of

the effeet of adding a reagent to ehange redox potentials of the pulp has also
been carried out. Results for rougher flotation of Colon and Sewell ore are
shown in Fig. 50. The colIeetor used for thc Sewcll ore at pH = 4 was isobutyl
xanthate, as there were problems adjusting the redox potential values with
the normal xanthogen formateo In light of the similar shape of the curves, the
aeid results should also apply to the xanthogen formares.
This work is being used to screen sulphidisation, as well as other inorganie

reagents for whieh there is no prior experience with Teniente ores. In
addition, a screening programme is on-going to evaluate pyrite selective
flotation reagents,
Chapter 11
SULPHIDE Mlll PRACTICE
Reagent consumpton in the otaton industry
The statistics and other data quoted in this section provide practical target
performance figures from operating mines (i.e., ore grades, concentration
ratios, concentrate grade, average mineral grind, power and water
consumption per tonne processed for the different types of ores, etc.), which
can be used as a guide in setting parameters for laboratory studies. The very
extensive average operating data, contained in Tables 56 and 57, should be
used, based on the predominant mineral of the ore to be tested, to select the
initial system conditions, and starting reagents.
The practical importance of the reagent families, and their evolution in

flotation, as well as the total volume of the reagent market in the U.S.A., can
be obtained from Bureau of Mines statistics, sorne of which are summarised
in Tables 52 and 53. Table 54 shows how the economic recessions have
affected the volume of ores processed by the mineral industry. Depletion of
mineral resources can be deduced from the drop in the grades treated, and
the evolution of the concentration ratios with this drop. CertainIy the
catastrophicdrop in metallic mining with the 1985 crisis is obvious; as is the
ecological pressure to clean up coal, reflected in the ore tonnage upgraded by
flotation, which has more than doubled.
The observed concentration ratios are a simple way to distinguish between

different types of flotation. If the concentration required to attain a
commercially saleable product is 30:1, the flotation regime is nascent bubble
capture. In the coal industry, and in the recovery of industrial minerals,
where the concentration ratio is also very low, the flotation regime is bubble
capture. When a mineral exhibits an intermediate concentration ratio, either
flotation mechaniosm may dominate, or both may be present simultaneously,
as is the case with complex metallic ores such as silver-zinc-lead, where the
task is to upgrade a 7 - 9% ore to a 25% concentrate to feed a smelter; i.e., a
ratio of roughly 3:1 to 4:1. The low concentration ratio in Table 53 for
- 174-
Sulphide Mili Practice 175
metallic carbonate and oxide ores, and the increase in tonnage treated since
1960, is due to iron ore, which the USBM puts in this classification.
TabLe 53.- FROTH FLOTATION IN THE U.S.A.; source: U.S. Bureau of 11i nes
Year 1960 1970 1980 1985
SuLphide Ores
Ore Treated tonnes 141,022,000 256,545,000 254,455,000 176,520,000
Concentration ratio 26.5:1 31.8: 1 38.1: 1 33.3:1
MetaLLi e Carbonate and Ox i de Ores
Concentration ratio 3.0:1 1.7:1 1.8:1 1.7: 1
NonmetaLLic Ores
Concentration ratio 3.0:1 3.4:1 3.5:1 4.2:1
Anthracite and CoaL
Concentration ratio 1.5:1 1.5:1 1.7:1 1.4:1
Grand TotaL ore
treated by fLotation 80,255,000 362,165,000 483,723,000 348,849,000
The quantities of coIlectors consumed in 1975 and 1985 (the only two years
reported in comparable detail in the Minerals Year Books), are shown in
Table 54. Note that the fatty acids dominate the oily coIlector market, and
the xanthates the metallic sulphide market.
The sulphide and non-sulphide ore tonnage treated is about equal, but the
oily coIlector consumption is thirty times sulphhydryl reagents because
industrial minerals require a unit dosage that much higher. The average
calculated reagent consumption is 589 g/t for the oily coIlectors, and only
23.5 g/t for the sulphide minerals. This is one reason why the inorganic and
oxide minerals' coIlector and frother market is financially much more
attractive to the reagent suppliers, despite the lower unit value of the fatty
acids and organic amines.
\
The breakdown by ore type for frother usage and consumption is less precise,
because the same reagents are used with all the different ore types. In
addition, there does not seem to be a consistent reporting basis between the
two latest years, as can seen from Table 55. Fuel oil and kerosene have been
separated in this table, beca use they modify frothers, and act as mild
collectors. But they can be a primary collector if the mineral is naturally
floatable, so are difficult to classify, and should be allocated to both the
frother and the collector tables.
TabLe 54.- U.S. CONSUMPTION OF COLLECTORS IN 1975 AND 1985
SuLphide CoLLectors Consumption, tonnes

1975 1985
ALkyL xanthates 3,699 58.6% 1,647 57.1%

Dithiophosphates 1,060 16.8% 611 21.2%
Xanthogen formates 507 8.1% 22 0.8%
Thionocarbamates 485 7.7% 586 20.3%
MercaptobenzothiazoL 185 2.9% 17 0.6%
5,936 2,883
Avg dosage gft ore 23.5 16.3
o i Ly CoL Lectors Consumption, tonnes
1975 1985
Fatty acids 59,509 77.0% 76,264 79.1%

TaLL oi L 10,953 14.2% *
Amines 4,276 5.5% 14,301 14.8%
800 series (suLphonates) 2,146 2.8% 5,411 5.6%
PetroLeum suLphonates 347 0.5% 484 0.5%
77,231 96,460
589.3 559.7
* incLuded in fatty acids
In 1975, frother consumption of 12,100 t compares with 83,167 t of sulphide

and oily collectors. This ratio widened considerably in 1985, reflecting either
improved frother chemistry, or more probably a shift to frothing collectors,
such as amines and sulphonates, in the industrial mineral applications. The
overall drop in both figures for 1985 confrrms the recession that hit the
mining industry in the early eighties.
Copper mill data
The material covered in Table 56 is a composite of a 1975 copper mines

survey (Crozier(1977, up-dated by data published in the SME Handbook,
miscellaneous plant visits, and reports in the trade magazines coveringmainly
the Pacific basin. It is therefore more representative of western porphyry
copper technology than the design norms followed in Europe and Africa,
Sulphide Mill Practice 177
where the geology of the deposits are somewhat different, The reagent use
and operating conditions reported also tend to be more representative of the
larger and newer mines.
Table 55.- eONSUMPTION OF FROTHERS IN THE TOTAL U.S. MINING INDUSTRY
1975 Tonnes Percent

MIBe 5,401 44.7%
DF 250 1,130 9.3
Barrett oil 1,078 8.9
Pine Oil 721 6.0
eresylic Acid 583 4.8
Aerofroth 71 345 2.9
Other Aerofroths 369 3.1
Other Dowfroths 92 0.8
NaLco frother 233 1.9
ueON 778 6.4
AL L others 1,367 11.2
TotaL frothers 12,099
FueL OiL 93,249

Kerosene 3,157
1985 Tonnes Percent

Aliphatic aLcohoLs 2,997 44.1%
PoLygLycoL ethers 763 11.2
PetroLeum based bLends 1,057 15.6
Pine oiL 495 7.3
PhenoL 353 5.2
Unspecified poLyoLs 1,125 16.6
TotaL frothers 6,791

FueL OiL 66,265
Kerosene 4,246
At the time the SME Handbook was published, over a million tonnes of
copper capacity had been removed from the market, mainly in the U.S.,
because of the mining industry crisis. Thus, temporarily or permanently, most
of the mines treating the lower ore grades. were shut down, or shifted to
leaching processes combined with low cost direct electrowinning of the
copper. AH these factors must be taken into account when generalising from
the data.
Nevertheless, it is interesting to see how predominant chalcopyrite and

chalcocite are in the ores of commercially important mines. But the types of
178 FLOTADON - Theory, Reagents and Ore Testing
reagents employed, and their point of addition to the flotation circuir,

certainly are not very standardised.
Tables 56 and 57 are both divided into ore characteristics, head grades, and
concentrate properties; and separately the collectors, frothers and modifiers
(where avaiIabIe) are shown with the quantities employed, and an attempt is
made to indicate the point in the flotation process where the different
collectorsare added. The completeness of the statistical data for the
different mines varies considerabIy. As there have be en significant ore
changes in the oIder mines over the years, the reagents reported are
consistent with the ore description, but do not necessariIy reflect current use.
Looking over Tables 56 and 57, it is interesting to note that there are
operating conditions and reagents listed that reflect quite arbitrary regional
preferences. Most noticeable is the tendency of U.S. designed plants (North
and South America) to grind their ore slightly finer and to operate at higher
alkalinity than the Pacific (mainly Philippines) mines. The only two mines
listed that employ an acid circuit, El Teniente in Chile and Lepanto in the
Philippines, do so because they have very specific ore characteristics that
require low pH operation. Lepanto uses an acid circuit because its ore is
very high in arsenic (enargite), while the El Teniente ore not only has a
relatively high non-sulphide content but also has acidic components in the
gangue that in the past have required very large amounts of lime to raise the
alkalinity to the habitual pH 9-11 (this may not be true in future when harder,
primary, ore will predominate).
From the mines listed in Tables 56 and 57, which represent over half of the
world's copper production (4,500,000 tonnes) in 1975, we can list average
operating conditions and performance for the different producing areas. As
can be seen in Table 58, between 1965 and 1975 the average copper
concentrator treated ores with 0.782% Cu, of which 85.35% was recovered.
Regional1y, head grades range from a low of half a percent copper in Canada,
to a high of 1.455% in Chile, while recoveries go from a high of 87.7% in
Canada, to a low of 80.4% in Europe.
TabLe 56.- INDUSTRIAL COPPER MILL FLOTATION DATA -
Principal "ineral CHALCOPYRlTE
DaiLy t % Oper Mi LL Feed Seeondary Gangue

miLLed time % Cu % Mo % Fe % Zn MineraLs composition
1B Twin Buttes, AZ 32,000 91 0.83 0.003 - Cu suLph Arkose, siLieated Limestone
1c Yerington, NV 13,500 98 0.60 - 2.8 - Ce Quartz, monzonite, granodiorite
3A Mission, AZ 23,000 96 0.66 0.024 3.7 0.07 Ce Quartz Ca, AL siLieates
3B Si Lver BeLL, AZ 11,000 97 0.64 0.008 3.4 Ce Quartz, pyrites
4 HighLand VaLLey, BC 16,000 97 0.51 - - - Bo Quartz, pLagioeLase, mico, eLays [j)
5 Bagdad, AZ 5,500 98 0.81 0.030 - - Ce Pyrites, siLieates S
'
7 Aitik, Sweden 6,500 95 0.49 Py Quartz, feLdspars, garnet, mica
8 Bouganvi LLe 90,000 0.48 0.010 Bo, Mag Quartz, feLdspars s:=-
(i>
9 Merritt, BC 5,000 98 1.00 0.029 Mag, Hem Quartz, feLdspars, eaLeite, kaoLinite
10 Pinto VaLLey, AZ 40,000 0.44 moLy ;.s;:
10A Miami, AZ 17,500 0.50 0.005 Cs
13 Sierrita, AZ 80,000 96 0.30 0.029 moLy Monzonite, diorite ...,"l:I
;l)
13A MineraL Park, AZ 17,000 0.40 0.04 moLy $2
13B Esperanza, AZ 15,000 0.40 0.03 moLy (-j.
14 Erstberg, W. rrian 7,500 2.80 - 35.0 Bo,Cov Magnetite, hematite (1l
15 MeLeese Lake, BC 30,000 0.40 0.020 Ce Quartz, serieite

16 Tide Lake, BC 7,000 1.50 Bo
17 Nairne, S. AustraLia 2,400 95 1.00 Ce AndaLusite, biotite, sehist
18 Lakeshore Casa Grande 8,200 1.50 0.001 5.0
20B MeGi Ll, NV 22,000 97 0.73 Ce SiLiea, aLumina
20c Ray, AZ 25,000 98 0.03 5.400 Ce Cup Quartz, serieite, monzonite, porphyry
200 Utah Mines 108,000 0.68 0.027 3.1 Ce,Bo,Cov Quartz, feLdspars, biotite
21 Noranda, Quebee 1,500 99 2.54 Sp Quartz, pyroxene, feLdspars
21A Logan Lake, BC 47,500 93 0.43 0.016 Bo Quartz, diorite, hornbLende
22A Superior, AZ 3,600 4.50 32.0 Bo
23 Marinduque, PhiL. 19,000 97 0.70 Bo Quartz, eaLeite, feLdspars
25 Hokkaido, Japan 1,100 97 2.35 1.10 Sp Quartz, eaLeite, ehLorite
26 Mount Isa, AustraLia 16,800 95 3.00 7.0 Quartz, doLomite
27C Rokana, Zambia 17,000 92 1.95 3.2 Bo, Car Quartz, doLomite 1-'
-...J
28A Geeo, Ontario 5,125 96 2.57 5.90 Sp, Ga Si Lieates \O
28B Horne, Quebee 1,860 2.45 Si Lieates
28C Brenda 26,000 95 0.19 0.051 MoLy Quartz-diorite, feLdspars, biotite, epi dote
TabLe 56a.- INDUSTRIAL COPPER MILL FLOTATION DATA (eontinued) -
Principal nineral CHALCOPYRlTE f-'
CXJ
o
Copper Coneentr % MoLy Coneen Preeious

%Cu %Mo %Fe %Zn InsoL %Reeov %Mo %Cu %Reeov Met. Ree
1B Twin Buttes, AZ 32 0.26 87 51 1.40 30 Ag "Tj
1C Yerington, NV 25 29 8.7 83 r-
3A Mission, AZ O
28 0.29 28 1.7 7.7 90 52 0.66 43 Ag >-3
3B SiLver BeLL, AZ 29 0.15 24 7.0 84 51 0.70 21 Ag ;:t>
4 HighLand VaLLey, BC 30 0.49 26 9.7 82 >-3
.....
5 Bagdad, AZ 32 1.45 22 15.5 74 54 0.75 64 Ag O
7 Aitik, Sweden 29 90 Au, Ag Z
8 BouganviLLe 30 0.13 85 Au, Ag
Merritt, BC >-3
9 27 ::r
(1l
10 Pinto VaLLey, AZ 27 0.05 29 5.6 53 1.25 Au, Ag o...,
10A Miami, AZ 27 0.06 87 1.25 54 ~
13 Sierrita, AZ 25 90 53 75 Ag
13A MineraL Park, AZ 20 0.1 76 0.30 62 ?d
(1l
l:ll
13B Esperanza, AZ 25 0.12 87 74 (JQ
(1l
14 Erstberg, W. Irian 28 95 Au, Ag
15 MeLeese Lake, BC 29 81 53 0.60 30 a
en
16 Tide Lake, BC 28 95 l:ll
:
17 Nairne, S. AustraLia 24 92 Au, Ag e,
18 Lakeshore Casa Grande 27 0.03 27 13.5 91 O
...,
20B MeGiLL, NV 18 0.30 30 12.5 76 Au, Ag (1l
20C Ray, AZ 21 0.41 25 15.2 85 49 1.40 37 >-3

(1l
20D Utah Mines 27 0.25 26 14.5 89 55 0.50 66 Au en
......
21 Noranda, Quebee 23 5.8 95 Au, Ag S'
21A Logan Lake, BC 33 90 56 0.65 75 (JQ
22A Superior, AZ 25 30 2.5 96 Au, Ag

23 Marinduque, PhiL. '..
27 6.0 87 Au, Ag
25 Hokkaido, Japan 22 4.0 94 Au, Ag
26 Mount Isa, AustraLia 25 29 8.5 96
27C Rokana, Zambia 32 18 9.6 89
28A Geeo, Ontario 27 8.0 94 Au, Ag
28B Horne, Quebee 16 96 Au, Ag
28C Brenda 30 0.20 86 55 0.30 78
TabLe 56.- INDUSTRIAL COPPER MILL FLOTATION DATA (eontinued) -
Principal Hineral CHALCOPYRlTE
DaiLy t % Oper MiLL Feed Seeondary Gangue

miLLed time % CU % Mo % Fe % Zn MineraLs Composition
28D Gaspe, Quebee 34,000 0.60 0.015 Ce CIJ
::
30A O'Okiep, S. Afriea 2,100 98 1.30 Bo PLagioeLase, biotite, hypersthene ..;j
30B Nababeep, S. Afriea 3,150 98 1.20 Bo PLagioeLase, biotite, hypersthene ;:r'
30c CaroLusberg, S.Afriea 4,200 98 1.25 Bo PLagioeLase, biotite, hypersthene s.:
C1l
31A Keretti, FinLand 1,600 95 3.30 23.0 0.60 Py,Sp Quartz, serpentine
31B
31c
VirtisaLmi, FinLand
PyhasaLmi, FinLand
750
?
95
95
0.65
0.88 29.6 4.20 Sp, Py
AmphibiboLe, pyroxene, garnet
Quartz, feLdspars, serieite ~
>j
32 PhaLaborwa, S. Afriea 53,000 0.55 Bo,Ce,Cup Carbonatite, phoseorite, doLomite ....
33A New CorneLia, AZ 33,000 99 0.68 2.6 Ce,Bo Quartz, feLdspars r:.:
n
r-r-
34A Sto Tomas, PhiL 5,000 97 0.39 3.2 Mag Andesite, diorite :;.
C1l
34B Benguet, PhiL~pi~es 16,000 97 0.39 3.4 Mag Andesite, diorite
35 Sahuarita, AZ 58,000 96 0.50 0.017 3.6 0.07 Ce Arkose, garnet, hornfeLs
35A Bagdad, AZ 40,000 0.70 0.03 Cs
37A Chambishi, Zambia 3,000 95 2.73 Ce,Bo Quartz, mica, feLdspars
37B Luanshya, Zambia 16,000 96 1.78 Ce,Bo Quartz, mica
40 Prineeton, BC 15,000 90 0.45 - ALbitized FeLdspars, pyrites
41 Niihama, Japan 1,400 98 1.53 17.0 Py,Prh ChLorite, aLbite, quartz
42 Kidd Creek, Ontario 9,000 97 1.90 8.90 Sp,Ga Quartz, feLdspars, taLe
43A Kipushi, Zaire 3,000 95 6.00 12.00 Bo,Sp Limestone, doLomite, sehist
44 Fierro, NM 3,200 95 2.00 2.7 0.70 Ce,Car,HetLimestone, doLomite, sehist
46 IsLand Copper, BC 38,000 90 0.48 0.017 3.7 MoLy Andesite, quartz, porphry
50 EL Teniente, ChiLe 105,000 1.48 0.047 2.2 Ce,MoLy
52 Andina, ChiLe 30,000 1.75 0.015 2.0 Ce,MoLy
53 Los Bronces, ChiLe 15,000 1.55 0.020 1.8 Ce,MoLy
l-'
54 EL Cobre, ChiLe 12,000 2.11 0.025 2.0 Ce,MoLy 00
I-d
TabLe 56.- INDUSTRIAL COPPER MILL FLOTATION DATA Ccontinued)
Principal .inerals Chalcocite, ltalachite, and IIomite
DaiLy t % Oper Mi LL Feed Secondary Gangue

mi LLed time % cu % Mo % Fe % Zn MineraLs Composition
1A Butte, Montana 44,000 95 0.69 Cp,Bo Quartz, compLex si Licates
1D Cananea, Mexico 21,000 83 0.00 3.500 Cov,Cp Quartz, cLays, sericite
3C Sacatan, AZ 11,000 93 0.72 3.0 Ma SiLicates, iron oxides, cLays e/)
19A Inspiration, AZ 20,000 94 0.50 0.008 1.4 Cp,Bo Sil i ca :::
19B Christmas, AZ 5,100 96 0.65 Andesite, diorite, Limestne ir
i:l""'
20A Chino, NM 23,000 94 0.91 0.020 13.2 Cp Ca siLicates, quartz, monzonite s;
C1l
22B San ManueL, AZ 43,000 98 0.75 0.015 3.5
27A ChingoLa East, Zambia 16,500 92 1.10 Cp,Bo,Ma Quartz, doLomite, feLdspars, mica <'
27B ChingoLa West, Zambia 25,000 97 1.50 2.2
27D KonkoLa, Zambia 5,100 90 3.00 Bo,Ma,Az Quartz, Ca, Mg, Fe, Mn carbonates '\J
....
33B Copper Queen, AZ 20,000 90 1.10 16.6
33C Morenci, AZ 61,000 99 0.85 0.008 3.7
Py
Ma,Cov Quartz, feLdspars '"
~
:;.
33D Tyrone, NM.. 29,000 99 0.85 2.6 Cov Quartz, feLdspars, sericite C1l
33E MetcaLf, AZ 30,000 97 0.80 0.010 4.5 Cp,Bo,Cov
36 GibraLter, BC 40,000 0.40 0.01 Cp
39 ToquepaLa, Peru 44,000 98 1.15 0.025 3.7 Cp SiLica, ALumina
45 White Pine, Mich 23,000 1.00 Si Licates
48 Chuquicamata, ChiLe 153,000 1.45 0.030 1.5 Cp,Bo
49 EL SaLvador, ChiLe 28,000 1.44 0.024 1.3 Cp,MoLy
Principal t1ineral ItALACHITE
43B Kamoto, Zaire 5,000 94 4.20 Cc,Car,Het Limestone, doLomite, schist

43C Kakanda, Zaire 2,400 96 5.00 Chr,Het
Principal t1ineral BORNlTE
i-'
00
37c MufuLira, Zambia 24,000 98 2.50 1.3 Cp,Cc Si Licates lJ.l
f-'
TabLe 56a.- INDUSTRIAL COPPER MILL FLOTATION DATA (eontinued) - OD
.;,.
Principal .inerals Chalcocite, Kalachite, and Bornite
copper Coneentr % MoLy Coneen Preeious 'Tj

%Cu %Mo %Fe %Zn InsoL %Reeov %Mo %Cu %Reeov Met. Ree r'
O
1A 8utte, Montana 26 19 - 3.7 75 Au, Ag S;..,
1D Cananea, Mexi eo 32 0.14 25 4.7 72 Au, Ag
3C Saeatan, AZ 32 20 11.7 84 Ag >-<
19A Inspiration, AZ 42 0.70 2 1.5 76 53 0.50 30 Au, Ag O
Z
198 Christmas, AZ
20A Chino, NM
23
21 0.50
21
32
10.0
5.4
72
77 50 1.15 29
Au, Ag
Au, Ag ..,::r-
228 San ManueL, AZ 28 0.54 28 7.0 86 56 0.90 69 Au, Ag o
27A ChingoLa East, Zambia o...,
278 ChingoLa West, Zambia 33 8 1.4 77 ':;<
27D KonkoLa, Zambia 39 85 ?J
(il
338 Copper Queen, AZ 13 33 7.0 80 Au, Ag po
()Q
33c Morenci, AZ 24 0.09 28 8.4 79 Au, Ag (Il
33D Tyrone, NM 23 0.15 30 8.0 79 Au, Ag i:l
.....
(1)
33E MetcaLf, AZ 260.15 27 10.0 61 Au, Ag Pl
36 GibraLter, 8C 300.15 85 1.00 45 e
Q..
39 ToquepaLa, Peru 28 0.40 27 6.0 86 53 1.20 30
45 White Pine, Mich 31 83 Ag O
...,
(Il
48 Chuquicamata, ChiLe 45 0.15 10 7.0 80 0.10 54
49 EL SaLvador, ChiLe 52 0.35 14 3.0 84 0.20 56 0-3
(Il
.....
~o
Principal nineral IlALACHITE S

()Q
438 Kamoto, Zaire 43 65

43C Kakanda, Zaire 24 80
--
Principal nineral BORNITE
37c MufuLira, Zambia 46 11 16.0 91

TabLe 57.- INDUSTRIAL COPPER MILL FLOTATION REAGENT DATA - COLLECTORS
Reag. gft POA Reag. gft POA Reag. gft POA Reag. gft POA
1B Twin Buttes, AZ Z-6 16.5 RF Z-200 1.7 CF
1C Yerington, NV Z-6 7.8 RF Z-11 3.7 RF Z-6 3.3 SdF Z-11 2.5 sdF
3A Mission, AZ Z-6 10.3 G, RF 238 2.1 G, RF 97A 6.2 G, RF CIl
::
3B Si Lver BeLL, AZ Z-10 8.3 G 238 8.3 G 97A 3.3 RF
4 HighLand VaLLey, BC NEX 12.4 RF KAX 15.3 Se eS
5 Bagdad, AZ Z-4 20.6 RF Z-6 4.1 RF 404 12.4 Se e:
e,
7 Aitik, Sweden KAX 18.6 C NIX 0.8 C NIX 4.1 (1l
8
9
BouganviLLe
Merritt, BC
3501
242 13.6 RF,Se Z-200 0.8 RF ;t::::
10 Pinto VaLLey, AZ Z-14 12.4 RF 238 12.4 RF
10A Miami, AZ Z-14 'i:l
-.
13 Sierrita, AZ KAX 2.1 RF 3302 6.2 G ~
13A MineraL Park, AZ Z-200 3302 ~
13B Esperanza, AZ Z-6 3302 :=;.
(1l
1} Er.stberg, W. Irian Z-200 16.5 G,RF Z-14 16.5 RF
1 MeLeese Lake, BC NIX 8.3 RF,Se Z-200 8.3 RF,Se
16 Tide Lake, BC Z-200 14.9 RF Z-200 1.7 CF
17 Nairne, S. AustraLia Z-200 24.8 RF
18 Lakeshore Casa Grande
20B MeGiLL, NV Z-200 6.2 G Z-4 20.6 G
20C Ray, AZ Rae 41.3 RF
20D Utah Mines Reeo 8.3 RF OiL 6.2 Se
21 Noranda, Quebee 404 792.4 G 343 466.4 RF CuS04 24.6 RF
21A Logan Lake, BC NIX 24.8 G,RF KAX 28.9 Se
22A Superior, AZ 404 24.8 RF NIX 4.1
23 Marinduque, PhiL. Z-200 16.5 G Z-6 20.6 RF,Se Z-11 20.6 RF,SC
25 Hokkaido, Japan 3501 11.1 3477 11.1 KAX 92.5 SC KEX 13.6 CF
26 Mount Isa, AustraLia NIB 206.4 G
27c Rokana, Zambia NIX 177.5 PKD 132.1 C
28A Geeo, Ontario 350 24.8 RF NIX 26.4 RF
28B Horne, Quebee KAX 44.2 PA KAX 102.4 CF NIX 69.8 PC '-'
28C Brenda KAX 31.8 G FO 12.8 C 00
U1
i-'
00
0\
TabLe 57.-INDUSTRIAL COPPER MILL FLOTATION REAGENT DATA - COLLECTORS (continued)

PrincipaL RineraL CHALCOPYRlTE
';j
l'
Reag. gft POA Reag. gft POA Reag. gft POA Reag. gft POA O
,..;
;;.
28D
30A
Gaspe, Quebec
O'Okiep, S. Africa
Z-6 24.8
238 24.8 RF
3302 11.2
NIX 24.8 RF .......,
30B Nababeep, S. Africa O
30c CaroLusberg, S.Africa NIX 12.4 G NEX 4.1 RF Z
31A Keretti, FinLand NAX 144.5 G >-3
31B VirtisaLmi, FinLand NAX 165.1 RF ::;
(1)
31C PyhasaLmi, FinLand NAX 66.0 C o...,
32 PhaLaborwa, S. Africa KAX 33.0 G ':;<
33A New CorneLia, AZ Z-11 4.5 G TP 1.7 RF 238 0.8 RF 25 0.8 RF ~
(1)
34A Sto Tomas, PhiL NIB 14.9 G 3501 4.5 G Z-14 111.4 SLF 2501 33.0 SLF ~
34B Benguet, PhiLippines Z-14 12.4 G Z-200 0.8 G (Jq
(1)
35 Sahuarita, AZ Z-6 7.4 RF 238 2.5 RF ::s
,..,.
35A Bagdad, AZ Z-6 en
Z-6 ~
37A Chambishi, Zambia NEX 8.3 RF KAX 4.1 RF ::sQ.,
37B Luanshya, Zambia NIX 28.9 G,RF
40 Princeton, BC NEX 20.6 RF KAX 24.8 Sc O
...,
41 Niihama, Japan KA 18.2 C Z-200 27.2 C 238 18.2 C (;l
42 Kidd Creek, Ontario 208 78.4 C NIB 12.4 Sc >-3

(1)
43A Kipushi, Zaire KEX 177.5 RF ,..,.
en
44 Fierro, NM Z-11 74.3 Z-200 18.6 G sr
(Jq
46 IsLand Copper, BC KAX 3.3 RF
50 EL Teniente, ChiLe MA 60.0 G
52 Andina, ChiLe
53 Los Bronces, ChiLe
54 EL Cobre, ChiLe
/
TabLe 57.- INDUSTRIAL COPPER MILL FLOTATION REAGENT DATA - COLLECTORS Ceontinued)
Principal .inerals Chalcocite. IIalachite. and 80mite
Reag. gft POA Reag. gft POA Reag. gft POA Reag. gft POA
1A Butte, Montana Z-6 2.1 RF Z-200 3.3 RF 1331 10.7 RF MA 16.5 RF
1D Cananea, Mexieo 238 7.8 RF 343 1.2 RF C/l
3C Saeatan, AZ Z-6 4.1 G Z-6 0.8 RF Z-6 0.8 Se e
19A Inspiration, AZ 898 12.4 RF -O
19B
20A
Christmas, AZ
Chino, NM
Z-6 10.3 RF
Rae 28.9 RF
404 7.8 G s:
o-
22B San ManueL, AZ Z-11 0.4 G 3302 G 4.5
_.~
(1l
0.4 SM8 G
27A ChingoLa East, Zambia
27B
27D
33B
ChingoLa West, Zambia
KonkoLa, Zambia
Copper Queen, AZ
NIX 177.5
NIX 239.4
Z-200 3.3
C
RF
G
PDK 132.1
Z-11 0.8 Se
-..,
""O
pO
33C Morenei, AZ Z-200 21.9 G ....
n
33D Tyrone, NM Z-11 18.6 G ('j'
331: Me'tea Lf, AZ Z-200 6.2 G NIX 3.7 Se (1J
36 GibraLter, BC Z-11
39 ToquepaLa, Peru Z-11 10.3 RF 3302 8.3 G
45 White Pine, Mieh NIB 90.8 DAP 9.9 Se
48 Chuquieamata, ChiLe SF-323
49 EL SaLvador, ChiLe SF-323
Principal ftineral IlAL\CHlTE
43B Kamoto, Zaire KEX 66.0 KAX 66.0 NaSH 1238.2
43C Kakanda, Zaire P 920.0 RF
Principal ftineral BORNlTE
37c MufuLira, Zambia NEX 17.3 RF KAX 17.3 Se
t-'
00
-J
t-'
00
00
TabLe 57a.- INDUSTRIAL COPPER MILL FLOTATION REAGENT DATA - FROTHERS
Reag. gft POA Reag. gft POA Reag. gft POA pH Reag. gft POA 'Tj
r
1B Twin Buttes, AZ R-23 10.3 RF 11.0 Lime 3,880 G O
1C Yerington, NV A-65 2.9 RF A-65 5.8 sdF A-710 3.7 SdF 11.0
3A
3B
Mission, AZ
Si Lver BeLL, AZ
PO 2.9 RF 11.5 Lime
11.0 Lime 825
G
G
~'""3
PO 37.1 G .....
4 HighLand VaLLey, BC MIBC 14.0 RF PPG 4.5 RF Lime 351 G O
5 Bagdad, AZ 1639 24.8 RF PO 28.9 RF 11.5 Lime 2,064 G Z
7 Aitik, Sweden PO 18.6 41G 8.3 Lime 289 G
8 BouganviLLe MIBC RF 250 RF A-65 8.5 Lime G e-3
9 Merritt, BC PO 5.8 RF Lime 413 e-
10 Pinto VaLLey, AZ 250 16.5 RF (1l
o>-1
10A Miami, AZ 250 PO 10.5
13 Sierrita, AZ A71 8.3 G MISC 33.0 M 11.0 Lime 1,032 G :.<
13A MineraL Park, AZ MIBC 11.5 :;;;
13B Esperanza, AZ MIBC 11.5 (1l
l:\:>
14 Erstberg, W. Irian PO 8.3 G,RF MIBC 24.8 G,RF 250 28.9 G,RF Lime 619 G ()Q
15 McLeese Lake, BC MIBC 8.3 RF,Sc 250 4.1 RF,Sc Lime 413 G (1l
16 Tide Lake, BC PO 14.0 RF PO 6.6 CF Lime 825 G ~

17
18
Nairne, S. AustraLia
Lakeshore Casa Grande
M300 2.5 RF Lime
Lime
495 G
G
..,
[f>
::l
20B McGiLL, NV 10.8 Q..
20C Ray, AZ 1639 74.3 RF 11.5 Lime 2,683 G O
20D Utah Mines MIBC 22.3 RF 8.5 Lime 743 G ....
(1l
21 Noranda, Quebec TEB 26.0 G 8.5 Lime 1,486 G e-3
21A Logan Lake, BC 250 11.6 RF PO G,Sc Lime 908 G o
22A Superior, AZ MIBC 20.6 RF Lime 2,064 G ;!;.
23 Marinduque, PhiL. MIBC 20.6 G A65 16.5 RF,CF 250 16.5 RF,CF 9.0 Lime 310 G ES'
25 Hokkaido, Japan PO 47.9 Lime 2,542 G (Jq
26 Mount Isa, AustraLia TEB 82.5 RF
27c Rokana, Zambia TEB 12.4 RF TEB 6.2 C
28A Geco, Ontario PO 27.2 RF NH3 118 G
28B Horne, Quebec PO 14.0 PA PO 38.4 PC Lime 2,022 G
28c Brenda MIBC 12.4 C 8.0
./
TabLe 57a.- INDUSTRIAL COPPER MILL FLOTATION REAGENT DATA - FROTHERS (continued)
Pr-incipal Hiner-al CHALCOPYRlTE
Reag. gft POA Reag. gft POA Reag. gft POA pH Reag. gft POA
C/)
28D Gaspe, Quebec MIBC 10. OLime G t::
30A O'Okiep, S. Africa CA .16.5 RF Lime 1,238 G >5"'
30B Nababeep, S. Africa g:
e,
30C CaroLusberg, S.Africa CA 16.5 RF Lime 619 G (il
31A Keretti, FinLand TEB 7.4 C Lime 2,064 G ~

31B
31C
VirtisaLmi, FinLand
PyhasaLmi, FinLand
PO 6.2
421 20.6
RF
C
Lime 2,683 G
Lime 1,403 G -........,
'i:J
32 PhaLaborwa, S. Africa TEB 8.3 G 8.0 :o
33A New CorneLia, AZ R-55 10.7 RF Lime 702 G ~
3~ Sto Tomas, PhiL 250 13.2 G 65 12.8 G Lime 248 G r=;'
(il
34B Benguet, PhiLippines 250 10.7 G Lime 318 G
35 Sahuarita, AZ MIBC 10.3 RF Lime 1,156 G
35A Bagdad, AZ MIBC 11.5
37A Chambishi, Zambia MIBC 16.5 RF
37B Luanshya, Zambia F1 4.5 RF 38Y 3.7 RF
40 Princeton, BC PO 24.8 41-G 8.3 Lime 1,032 G
41 Ni ihama, Japan PO 5.4 C MIBC 12.8 Lime 1,362 G
42 Kidd Creek, Ontario MIBC 14.9 C Lime 244 G
43A Kipushi, Zaire RO 99.1 RF Lime 2,270 G
44 Fierro, NM 250 8.3 RF MIBC 8.3 RF Lime 1,238 G
46 IsLand Copper, BC 71R 22.7 RF 1012 1.7 RF 10.5 Lime 619 G
50 EL Teniente, ChiLe 4.2 H2S04 G
52 Andina, ChiLe 9.0 Lime G
53 Los Bronces, ChiLe 10.5 Lime G >-'
54 EL Cobre, ChiLe Lime G co
\,Q
,....
\O
o
TabLe 57a.- INDUSTRIAL COPPER MILL FLOTATION REAGENT DATA - FROTHERS (continued)
Principal .inerals Chalcocite, ltalachite, and Bomite
Reag. gft POA Reag. gft POA Reag. gft POA pH Reag. gft POA
'"rj
Principal " CHALCOClTE t""'
O
1A Butte, Montana MIBC 4.1 RF 10.5 Lime 3,880 G
1D Cananea, Mexico 250 13.6 RF Lime 2,518 G ~
3c Sacatan, AZ MIBC 20.6 Lime 1,238 G ~
>-<
19A Inspiration, AZ MIBC 12.4 RF 10.5 Lime 1,238 G O
19B Christmas, AZ MIBC 22.7 RF 9.5 Lime 3,302 G Z
20A Chino, NM PO 51.2 RF MIBC 44.6 RF 11.0
22B San ManueL, AZ MIBC 23.5 RF 250 0.4 RF 10.5 Lime 702 G ~
27A ChingoLa East, Zambia Lime G I:T
27B ChingoLa West, Zambia TEB 12.4 RF TEB 6.2 C Lime G (1l
o....
27D KonkoLa, Zambia PO 27.2 RF NH3 118 G
33B Copper Queen, AZ 250 9.1 RF 11.5 Lime 2,889 G ~
33C Morenci, AZ CA 21.9 RF 10.5 Lime 2,229 G :::o
33D Tyrone, NM 250 16.5 RF 11.4 Lime 1,651 G (1l
(.ll
33E MetcaLf, AZ 250 8.3 RF Lime 3,508 G ClQ
36 GibraLter, BC MIBC 10.5 (1l
39 ToquepaLa, Peru
45 White Pine, Mich
PO 8.3
250 5.0
RF
F
73 8.3 RF 11.5 Lime
9.5 Lime
1,238 G
660 G
a
Ul
(.ll
48 Chuquicamata, ChiLe 11.0 Lime G i:l
49 EL SaLvador, ChiLe 11.0 Lime G e,
O
....
Principal" IlALACHlTE (1l
43B Kamoto, Zaire Lime 908 G >-J

(1l
43C Kakanda, Zaire 250 8.2 RF MIBC 8.2 RF Na2C031,068 G Ul
g.
Principal " DORNlTE ClQ
37C MufuLira, Zambia 25021.8 10.7 Lime 781 G
t
Abbreviations for Tables 56 and 57
Symbol identification Depressants
C conditioner OxRo oxide rougher REG regrind CN cyanide
CF cleaner feed PA pri'!1ary ~era~or Sc scavenger GG guar gum
CRT copper rou~her tailings PC pyr1te C1rcu1t SdF slime feed Nig nigrosine and dextrin
G grinding c1rcuit POA point of addition SRT sulphide
M middlings Reag reagent rougher tailings
Reagent Abbreviations
collectors Frothers (f)
3302
3501
DTP
Aeropromotor 3302
Aeropromotor 3501
dithiophosphate
NAX
NEX
NIB
sodium amyl xanthate
sodium ethyl xanthate
sodium isobutyl xanthate
R55
R23
RO
Union Carbide frother
Union Carbide frother
resin oil
CA
FI
73
cresylic acid
FLotol B
Aerofroth 73
ia
(1l
DAP di-isoamyl DTP NIX sodium isopropyl xanthate PO pine oil M300 PPG derivative
KAX
KEX
potassium amyl xanthate
potassium ethyl xanthate
Rac
SM8
Raconite
Minerec SM8 Powells acceLerator
~
PA S
25 Aerofloat 25 238 Aerofloat 238 Que quebracho
MA Minera A 242 Aerofloat 242 250 Dowfroth 250 ';I
....
244 Aerofloat 244
1638
1639
SheLl MIBC + 10% stiLlbottoms
Shell MISC + 15% stillbottoms p.
;:;.
Z-200
343 Aeroxan~ate343 (1l
350 Aeroxanthate 350
898 Minerec 898
1331 Minerec 1331
FO fueloil
P palm oil
FloccuLants and modifiers

GPG Aero-dri 100 202 Superfloc 202 TP tarpine
41G Palabora stabilizer 206 SU'lerfloc 206 WG water~lass
60 compLex 60 M-59 Po yhaII M-59 330 sU'ler loc 330
97A StepanfLote 97A Sep Separan 673 Na co 673
127 Superfloc 127 SPP sodium tripolyphosphate 1801 Na Lco 1801
2299 NaLco 2299
1-'
\O
1-'
Table 58.- AVERAGE PORPHYRY COPPER rUNE PERFORI,lANCE IN 1975
Region Number Copper Copper Production

of mines Grade Recovery Capacity, t/yr
USA 29 0.695% 84.2% 1,758,000
Canada 11 0.504% 87.7% 373,310
Chile 5 1.455% 86.3% 1,059,500
Australasia 8 0.663% 86.1% 508,780
U.S.S.R 7 0.815% 83.4% 391,270
Europe 3 0.632% 80.4% 72,000
Africa 2 0.939% 85.1% 195,000
The flotation data being analysed are biased towards U.S. design preferences,
as the statistical information represents very nearly a 100% of the U.S. mines,
90% of Chilean mines, only half the mines in Canada, U.S.S.R. and Pacific
area, and less than half of the European and African mines.
Fig. 51 shows four frequency bar charts listing flotation recovery for copper
and moly, ore treated grade, concentrate grade, grind minus 200 mesh
percent, flotation pH and co11ector consumption for a11 the mines in Tables
56 and 57. From these we see that copper recovery shows two distinct peaks
which probably reflect the difference in treatability of primary sulphide ores,
which even for very low copper grades show 90 + % recovery, and that of
partially weathered, high grade ores whose mean recoveries trend in the mid
eighties range. Mines with recoveries below 80% are obviously involved with
ores with special treatability problems. The greater complexity of the mixed
molybdenum ores is reflected in the very broad scatter in the moly
recoveries, which range rather indiscriminately between 30 and 80%. The
copper concentrate grades produced, bunch in the 25 to 34% range. Mines
with higher concentrate grades are treating relatively unusual ores, like those
found at Chuquicamata and El Salvador, where the weathered cap has rarer
mineralisation, such as atacamite. The numbers confirm that for most ores,
dissemination of the mineral is fine, and that a grind that gives 50 to 65%
minus 200 mesh is required for proper liberation.
As noted above, only two mines in the list operate with an acid pulp, but at
normal recoveries. Otherwise the listing shows that flotation at a neutral pH
is not attractive. There is no good explanation for this, other than to
speculate on the depressant effect of ions that are normally present in
solution in natural mine waters, but that could be precipitated by the lime
used to increase the pulp alkalinity. The scatter in the last graph underlines
that there are no generalisations that can be made on what sort of a co11ector
consumption should be considered normal, though not unexpectedly, there is
Sulphide Mili Praetiee 193
sorne eorrelation between collector consumption and copper grade. Also

noticeable is the use of multiple collectors to try and optimise by-product
reeovenes. The best eorrelation between eopper grade and collector
eonsumption is obtained with the data from the U.S. mines, where the linear
correlation eoefficient is about 83%.
Ouerall Recover~ Ore Grade
Cu Glt'ade Moly
813 89 8.88 1.6

x :::l
X :
~ ~
;,
III
."
69 ~
III ...38.86 1.2
III
'"g.
~ 69
t!l ~ o u
lE:
;, I:l:; +'
III ;, +'
: ~
e&: 49 49 ~ ~ 8.84
8.8 ~
u
u g. ;,
Q)
u ~
21i'l 29 8.82 8.4
"r--......_ ..... ...I..--l_....I-......

126 Ii'l 612 126 lit 612
NUI~hell of Mines NUlllhell of IIIines

Concenilt'aie Olt'e Feed Floiaiion Colleciolt'
Grade Grind pH ConsuMpiion
100 ~.,.......,.-.,.......,.-~-.100
29
"L........_ .... ..&.......l_...L.........

12 6 lit 6 12 " 1269612
NUI~hell of Mines NUlllhell of Mines

Fig. 51.- Overall mili performance, Tables 56 & 57
Table 59 shows the frequency of use of single coIlectors and mixes as

practised in 1975. The bar graphs in Fig. 52a iIlustrates this. The trend has
not changed significantIy, with xanthates being, by far, the most popular thiol
collectors. An attempt to correlate the 1975 coIlector data with flotation
showed a significant correlation between the use of xanthic esters and
enhanced recovery of molybdenite, and also a direct correlation between the
use of the xanthogen formate structure and acid pulps. There is some
support for the use of dithiophosphates as supplementary collectors in
TabLe 59.- COLLECTOR USE FREQUENCY COMPARISON FOR COPPER MILLS

% OF TOTAL MILLS SURVEYED
CoL Lector usage as

SoLe or Primary Gross
CoL Lector coL Lector TotaL
Xanthates 30.4% 78.3% 78.3%
Dithiophosphates 4.3% 5.8% 27.5%
Xanthogen Formates 1.4% 7.3% 7.3%
Thionocarbamates 5.8% 8.6% 18.8%
Xanthic Esters O O 8.7%
MercaptobenzothiazoL O O 5.8%
FueL oiL O O 17.4%
TotaL 41.9% 100% 163.8
Two or more coLLectors mixed:
Xanthate pLus ...
Dithiophosphates 15.9%
Xanthogen Formates 1.4%
Thionocarbamates 7.2%
Xanthic Ester 5.8%
FueL oiL 11.6%
Dithiophosphate pLus ...
FueL oiL 2.9%
FueL oiL pLus Xanthic Ester 1.4%
FueL oiL pLus MercaptobenzothiazoL 1.4%
Thionocarbamate pLus MBT 1.4%
Xanthogen Formate pLus ...
FueL oiL 4.3%
Thiophosphate pLus MBT 1.4%
Thionocarbamate pLus ...
Xanthic Ester 1.4%
Dithiophosphate 1.4%
Xanthates Dithiophosphates Allyl Xanthic Ester

Moly Coppe,.. ,loly Coppelt" Moly COPl1el:'"
HU; la
R
e
80 80 rIl:
=
rn
e
o 60 60
ll11l
V Bll
e
t:
y 40 41l
%
211 211
11 11
12 (, B (, 12 12 (, B (, 12 12 (, B (, 12
NUMBER OF MINES
Xanthogen forlUates T'bdorrocar-ba.rnates Mercaptobenzothiazole

Mol!! Coppel' Mol!! COpp@1' Mol!! COpp@1'
r
llillil Iillil
~ Izl

~
R 81il 81il
e a=
e i"
o I
61l 611
v
e
t:
y 49 49
%
21l 21l
Il Il
12 s 11 s 12 12 (, 11 (, 12 12 s 11 1> 12
NUMBER OF MINES
Fig.52a.- Collector frequency bar graphs
c1eaning circuits. Cyanamid has always pushed their thiophosphate and

mercaptobenzothiazole as good enhancers for the recovery of precious
metals and other secondary values in copper ores. It also introduced
AP 3302 (xanthic ester) as a special collctor for molybdenum, and the
data bears this c1aim out, though the ester is too weak to be used as a
primary collector for copper. Fuel oil is the other collector that is universally
used to enhance molybdenum recovery. The quantity used affects the
selection of the frother.
There is an inadequate amount of frother data to try to detect recovery or

selectivity trends, but some intuitive conclusions can be reached from the bar
graphs in Fig. 52b as to the way frothers are being applied. The data is
summarised in Table 60. Two trends are obvious from this table: an
increased tendency to fine-tune the froth by employing multiple frothers, and
a distinct increase in the popularity of MIBC. The driving force in these
changes between 1960 and 1975 has primarily been improved operating
stability, and enhanced recoveries; only secondarily have the changes come
because of reduced reagent costs.
TabLe 60.- FROTHER USE FREQUENCY COMPARISON FOR COPPER MILLS

% OF TOTAL MILLS SURVEYED
From TabLe 55 From Sooth (1962)

1975, 55 mi LLs 1960, 66 mi LLs
SingLe frother
Pine oiL 15% 17%
CresyLic Acid 2% 11%
MISC 29% 14%
Other Higher ALcohoLs O 3%
PoLygLycoL ethers 11% 21%
TES 2% 8%
SubtotaL - SingLe Frothers 59% 74%
Two or more frothers mixed
PoLygLycoL ethers pLus....
Pine oiL 13% 5%
MISC pLus .,.
Pine oiL 5% 5%
CresyL i e acid 4% 3%
PoLygLycoL ethers 11% O
Pine oiL pLus ...
CresyLic acid O 5%
PoLygLycoL ether pLus MISC 4% O
SubtotaL Mixed Frothers 37% 18%

MisceLLaneous Frothers 5% 8%
(Pyridine, etc)
PINE OIL CRESYUC Acro

~ol!l Cepp@" I~ol!l ColP!P@!i'
10i'l HIEl 1Ela :i.El0
~
R
88 1:10 e 88 ~ ea
e
o
f.O b8 v 6B 6B
e Il!
t:
y
48 48 48 48
%
1- "
2B 28 28 28
I I I I
8 8 8 E!
12 (, 8 b 12
Number of Mines
MIBC POLYGLYCOL ETHERS

Moly Coppel' Moly Coppel'
lE18 1111'1 11'11'1 1111'1
BI'I 81'1 R 88 88
e
e
bE! bE! o G8 GEl
v
e
r-
48 48 Y 48 48
%
2E! 28 28 2E!
El El El 1'1
12 (, 8 (, 12 12 (, 8 {, 12
Number of Mines Numher of Mines
Fig.52b.- Frother frequencybar graphs

198 FLOTATIOI\! - Theory, Reagents and Ore Testing
TabLe 61.- COLLECTOR CONSUMPTION IN COPPER SULPHIDE ORE FLOTATION

metric tonnes
TotaL Canada South Pacific USA USA USA

Mexi co Amer i ca area
CoLLector 1975 1975 1975 1975 1975 1965 1960
IsopropyL xanthate 4,054 69.5 2,156 369 1,459 1,082 823

AmyL xanthate 1,549 214 11 628 696 273 192
Sec ButyL xanthate 450 450 380
EthyL xanthate 251 106.8 99 45 370 396
TotaL xanthate 6,304 390.4 2,167 1,096 2,650 1,726 1,790
DicresyL DT 829 9.5 78 383 358 341 330

Disec butyL DT 579 234.1 118 87 140 13
diamyL DT 112 120 59
Di ethyL DT 120 120 466 516
TotaL dithiophosphate 1,640 243.6 196 470 730 880 846
DiethyL XF 1,286 1,286 32 232
EthyL isobutyL XF 105 105 89 118
EthyL methyLamyL XF 40 40
XF mixtures 98 98 68
TotaL xanthogen formate 1,528 1,286 40 203 189 350
EthyL isopropyL TC 983 119.1 164 299 400 254 258
MethyL butyL TC 153 153 102
TotaL thionocarbamate 1,136 119.1 164 299 553 356 258

ALLyLamyL xanthic ester 371 218.2 153
MercaptobenzothiazoLe 104 17 87 7 51
Sub TotaL CoLLectors 10,628 971.4 3,814 1,922 4,813 3,3303,132
Ore treated dry t '000 194,741 36,318 28,704 34,268 95,41473,926 3,290
Copper recovered, t '000 2,858 330 843 391 1,267 1,02 811
ro of reported Cu production 45% 40% 89 ? 82% 83% 83%
Table 61 shows Bureau of Mines data for sulphhydryl collector consumption

in the U .S.A. for the period 1960 to 1975, and is the only consistent published
data on the tonnage of individual collectors used in 1975 in copper flotation
world wide. It covers approximate1y the same mines as those canvassed in
Tables 56 and 57. The world wide data shows sorne regional preferences, or
possibly availability of the different types of collectors, while the comparison
on collector usage in the U.S., between 1960 and 1975, underlines how small
Sulphide MilI Practice 199
Table 62.- POWER AND WATER USE IN U.S. FLOTATION
Plants Energy Water

Minera l Number Capaeity Ore - t Cone Cone kw-hr gal/
1975 t/day Treated tonne Ratio tonne tonne
Copper 18 293,636 83,809,091 1,967,769 42.6 17.8 935
Copper-moly 15 467,273 132,209,364 2,400,717 55.1 19.4 842
Cu-Pb-Zn 13 26,636 6,658,300 500,620 13.3 18.2 743
Cu-Zn-Fe 3 7,727 1,966,700 735,889 2.7 22.3 671
Gold-silver 3 764 65,900 433 152.3 30.3 814
Lead-zine 21 28,091 7,510,000 865,993 8.7 24.0 924
Zine 10 17,818 3,139,800 193,066 16.3 20.4 682
Barite 3 w w w w w w
Feldspar-miea-quartz 13 9,455 2,331,091 1,032,900 2.3 27.5 2,640
Fluorspar 3 1,291 249,545 77,')85 3.2 75.1 1,870
Glass sand 22 36,364 6,680,818 5,516,727 1.2 na na
Iron ores 5 75,455 26,017,273 13,579,672 1.9 22.0 1,870
Kyanite 3 w w w
Limestone-magnesite 5 1,773 365,727 309,909 1.2 22.8 594
Mereury 1 w w w w w w
Molybdenum 2 w w w w w w
Phosphate 19 51,091 68,251,182 16,360,336 4.2 9.8 3,465
Potash 8 44,545 12,600,909 3,045,875 4.1 18.3 319
Tale 2 w w w w w w
Tungsten 2 w w w w w w
Coal 78 58,455 11,890,000 7,435,000 1.6 na na
Grand Total 252 1,465,273 384,113,636 54,788,472 7.0 17.4 1,397
1985
Copper 8 145,805 62,559,090 2,519,090 24.8 16.8 762
Copper-moly 7 474,200 114,332,730 2,790,91041.0 14.5 357
Cu-Pb-Zn 3 W W W W W W
Cu-Zn-Fe 1 W W W W W W
Gold-silver 3 470 118,182 4,550 26.0 72.6 616
Lead-zine 9 12,530 3,254,550 672,720 4.8 19.5 572
Zine 4 14,715 3,009,090 109,090 27.6 14.7 639
Barite 2 W W W W W W
Feldspar-miea-quartz 10 7,584 1,894,550 49.0 4,872
Fluorspar
Glass sand 8 11,600 2,521,180 2,162,730 1.17 NA 1,610
Iron ores 4 103,870 22,004,550 17,661,820 1.25 44.1 4,360
Kyanite
Limestone-magnesite 4 960 176,360 NA 610
Mereury 1 W W W W W W
Molybdenum 4 95,900 21,867,270 60,910 359.0 17.6 323
Phosphate 22 415,800 109,715,45024,619,090 4.5 12.4 3,480
Potash 5 41,170 10,090,9Q.o 1,516,360 6.6 12.0 78
TaLe 1 W W W W W W
Tungsten 1 W W W W W W
CoaL 77 93,80025,150,900 17,731,820 1.4 7.1 359
Grand Total 179 1,403,540 384,423,820 72,818,180 5.3
w = withheLd, but ineluded in total; na = not available
the shifts in collector use were during those years. The numbers do
underline the domination of this market by the xanthates, The frothers,
shown in Table 60, and analysed graphically in Fig. 52b, also indicate that
there was relatively little change in frother technology. Unfortunately there is
no more recent data which is cohesive enough to update the trends.
Certainly the copper industry took a bath in the early eighties, and, after
numerous bankruptcies, it has been reborn with new owners, a different
investment philosophy, and a fundamentally changed view on the most
economic technology. Flotation per se has not seen changes other than the
trend to reduced capital intensiveness through the use of giant flotation cells
and columns for upgrading rougher concentrates in cleaners circuits, or to
reduce the capital investment in moly-copper separations. Flotation is still
the lowest cost method for partition of two similar solids, but heap, or in-situ
leaching, followed by selective extraction of the metals from the solution, and
then direct electro-winning, has rescued from the scrap heap a number of
mines which had by-passed oxidised cap ores. The quantities of these
reserves are limited, so should become exhausted in a relatively short time,
but have given the industry a breathing spell to find new technology to treat
the massive, low grade, disseminated deposits that are now idle. The crisis
also spawned a gold-rush, which now has gradually tapered off.
The SME Handbook quotes comparable U.S. Bureau of Mines statistics on

power and water use in flotation mills in the U.SA., covering all types of
minerals. These are reproduced in Table 62. These data are important as
they provide a broader picture of the whole mineral industry, and how
flotation is used as a unit operation.
Flotation process design
In order to design an efficient laboratory test programme for a new ore, it IS

important to visualise the type of flotation circuits that would normally be
employed in treating a particular type of ore.
Circuit design normally starts with comminution. Allfloatable ores must first
be crushed and ground to a particle size that will liberate the valuable
componentes) from the gangueo The liberation grind will depend on the
dissemination of the ore components, but usually the optimum from the
recovery point of view is the 100 to 350 mesh range. At less than 50 mesh, the
particle is usually too heavy to remain attached to the bubbles, and at grinds
greater than 400 mesh the pulp will contain slime which can interfere with
efficient flotation. The equivalent micron sizes for the nominal screen sizes
normally found in a metallurgicallaboratory are listed in Table 63. From this
table we see that the approximate optimum particle size range found in a
Sulphide Mill Practice 201
normal f1otation pulp is between 30 to 150 microns,
Run of mine ore is crushed in jaw or gyratory primary crushers, followed by

secondary and tertiary cone crushers, to malee a grinding circuit dry feed.
Normally each stage of crushing is scalped with a vibratory screen. The feed
size to the grinding process depends on designo Autogenous grinding mills
are justified because of capital savings in the crushing circuito A two stage rod
mill/ball milI sequence also will tolerate a coarser feed. Modern grinding
circuits usually consist of an autogenous mill followed by one or more ball
mills. Both stages are operated in a closed circuit with a scalping cyclone
returning the oversize back to each stage of grinding. One equipment
supplier is recommending the installation of a "flash" f1otation cell into the
cyclone circuit to recover larger particle ore which contains malleable
components such as native gold or copper. Process water is added to each
stage of grinding. The final pulp can be discharged directly into the first
flotation cell, or more usually to an agitated conditioning tank where the pulp
density is adjusted to 30 to 40% solids, by addition of process water, and the
lime, collector, and part of the frother are added. For most ores the pH
adjusting reagent, conditioners, activators and depressants are preferably
added to the final ball mill feed. Modern plants control the ore feed rate to
the mill based on a particle size analyser located on the flotation pulp feed
line and the milI drive power demando This method tends to provide uniform
tonnage and particle size distribution of the feed.
TabLe 63.- SCREEN SIZE EQUIVALENTS
TyLer A.S.T.M British TyLer A.S.T.M British

microns Mesh Standard Standard microns Mesh Standard Standard
2,000 10 8 250 60 60
1,680 10 12 10 210 65 70 72
1,410 14 12 177 80 85
1,190 14 16 14 149 100 100 100
1,000 18 16 125 120 120
840 20 20 18 105 150 140 150
710 25 22 88 170 170
590 28 30 25 74 200 200 200
500 35 30 62 230 240
420 35 40 36 53 270 270 300
350 45 44 44 325
297 48 50 52 37 400 400
\
A major technology change during the past few decades is the great increase
in the size of the individual flotation cells, and the use of high capacity non-
mechanical flotation devices, such as columns. One consequence is that the
202 FLOTATIN - Theory, Reagents ami Ore Testing
traditional pulp distribution tank, which doubled as a conditioning tank, is

now no longer included in the flotation circuit. This trend in the size of the
equipment has not changed the basic unit operations in the pulp flow. AH
milis still have flotation banks dedicated to: 1) an initial rougher bank; 2)
scavengers; followed by 3) cleaners, and 4) re-cleaners. And the pulp is: 1)
conditioned with flotation reagents and air; 2) sometimes parted into sands
and slimes in cyclones or classifiers for separate treatment; 3) reground at
the cleaner and re-cleaner stage to further liberate the valuable minerals; and
4) thickened to recover process water.
The simplest flow scheme used with buIk sulphides and simple industrial
mineral upgrades is the rougher-cleaner arrangement:
Rougher-Cleaner Flotation Circuit

Final Conc. Feed Pulp
~
" ~
1 2 3 4 5 6 7 8
" .. .. ----Iotails
I I
~
CLEANER" ROUGHER FROIH
I
water
A more conventional circuit includes a scavenger after the
rougher:
Rougher-Cleaner-Scauenger Flotation Circuit
Final Conc. Feed Pulp 8CAVEHGER FROrH

" ...H
~_'H' 1234567 B
I I
~
CLEANER " ROUGHER FROTH
I
Water
Rougher bank pulp solids are normally maintained at 35% plus or minus 5%,
while the cleaner feed is normally diluted to around 25% to improve
selectivity.
If the mineral contains native gold, as is not uncommon with sorne copper
ores, the feed and tailings launder can be lined with sacking or corduroy, to
Sulphide Mill Practico 203
recover coarse gold particles that do not float, The cleaner can be cascaded
with re-cleaners, In a moly j copper separation, these may reach 15 stages.
The residence time in the rougher and cleaner varies according to the ore
involved. With typical copper ores, roughers in the Pacific basin are
designed for 8 to 12 minutes residence, while European and African mines
will provide up lo 40 minutes of roughing time. Cleaners typically are half to
two thirds of the rougher time. Complex ores (Zn, Cu, Pb, Ag mixtures)
normally involve European type flotation times.
All flotation processes involve differential flotation, as in all cases the

separation is valuablc mineraljgangue, or, in the case of copper ores,
valuable mineraljpyrite plus gangue; but the term differential fiotation
usually is reserved for complex ores, where more than one valuable mineral
is recovered, and mutually separated from each other and the gangue tailings.
The tonnage and monetary value of complex ores commercially floated is
considerably lower than the tonnage of bulk metallic and industrial minerals
milled, but technically it is more challenging as it normally involves
multiple flotation steps, multiplc collectors and frothers, in addition to
conditioners, activators and depressants; while bulk ores typically only
require conditioners, frothers and collectors. All milis employ flocculants and
drying agents in their tailings disposal and process water recovery systems, so
this technology is more generally applicable, though different for water
soluble and water insoluble minerals.
Complcx ore processing schemes can become very involved, but they all can
be simplified to a general set of circuits, consisting of a primary rougher-
scavenger flotation bank for the feed and each separable mineral, followed by
a cleaner-scavenger and recleaner - scavenger bank. The scavenger
concentrates are reground and returned to the primary feed, while the
recleaner scavenger concentrate is recycled. In each case the recleaner
concentrate is the final concentrate. The same circuit is used for the recovery
of the lead and zinc sulphides. The tails from the previous circuit is the feed
for the downstream systems.
The reagent suite used is different for each mineral recovered. Not shown in
Fig. 53 is a dewatering and reconditioning step placed between the copper-
lead and lead-zinc circuits to remove the collector, frother and other reagents
adhered to the ore in the previous circuitj and to add new conditioners,
collectors, frothers, depressants and activators for the next separation.
Reagent selection for the flotation of complex ores is dependent on the ores
COPPER CONCENTRATE lEAD CONCOOlIATE N
t o.j>.
~ 6
~
~
5
~
4
~
3
RECLEANER
~
2
~
1
la
f \ SCAVENGER 'TJ
r-
O
i"T ~
I
<l
==r-
CLEANER
'n
\ I
-~
f:-,--
l"'''"' I L -
+ ~LEANER
.. ,.. 1"
~
..
'1
'
..
1
1
1 ~
O
Z
GENERAL ~ COPPE.
~ ~~ >-l
~..
<ll
FEED ........
ROUGHER
6 --, SCAVENGER 1 2 .. 3 ..4 ..5 ..6 'I i:r'
(il
o....
lEAD ROUGHER SCAVENGER
<l~~ <l ';;<
r=~ ZINC CONCEH1"W\TE
:;o
<l SCAVENGER \
(il
~
<l-I 6 I 5 (il
g
'"
~
I:::l
e,
O
...,
- .. (il
>-3
7
==r-
(il
<l---- ;::?,.
S'
! <l, + _ (JI:
GENERAL TAIL <l I '-- L-!_ _ ~ ~

--~
ZINC ROUGHER SCAVENGER
I <l-------------------------------'
Fig. 53.- CompLex ore differentiaL fLotation circuit

and the exact mill circuit, and flotation conditions favoured by the designer.
As there are interactions between the inorganic and organic reagents, there is
little that can be said, a priori, about which set of reagents is an optimum.
Prediction of the circulating load from the scavengers is also difficult to
determine in the laboratory, and essentially impossible to do from first
principIes.
Copper mili otaton practce
As we have seen, by far the most common copper mineral treated by

flotation is chalcopyrite, followed by chalcocite, which is the primary mineral
in a quarter of the milIs; but essentially all the copper ores contain pyrite as a
significant secondary sulphide, and in sorne cases other iron sulphides, such
as pyrrhotite and magnetite are also present. The other copper minerals
accompanying chalcopyrite as the major component are: chalcocite, bornite,
covellite, cuprite, and native copper, more or less in order of importance.
When chalcocite is the major component, which occurs in one out of five
copper ores, the secondary copper minerals present are: chalcopyrite,
covellite, bornite, malachite, and azurite. In a significant number of copper
minerals, sphalerite is present as a less valuable secondary mineral. In Chile
and the Philippines, where molybdenite, gold and silver are important minor
minerals, arsenical ores, such as enargite and luzonite, are present. In
African mines, cobalt is an important by-product. Chilean porphyries tend to
contain around 0.3% copper as non-sulphide ores that are not readily soluble
in acid.
Table 64 summarises the most important sulphide ores, their nominal

formula, density and metal content. The copper minerals are listed in the
approximate order of their commercial importance, and the constituents of
complex ores are separated into primary and associated minerals.
The largest tonnage copper flotation mills treat low grade porphyry ores, with
chalcopyrite as the predominant mineral. As chalcopyrite contains less than
36% Cu, concentrate grades obtained contain about 30% copper, and low
head grades (0.4 to 0.9% Cu) mean that tailings with less than 0.1% Cu are
unusual. To maximise grade and recovery, the flotation plant should be
operated with powerful collectors having a high degree of selectivity. As the
chemists have not developed a collector that combines these two properties,
in most flotation plants a high recovery is obtalned using a strong collector in
the roughers, followed by a selective collector in the cleaners. If, as is usual in
low cost and low capital mills, the pulp fed to the roughers has a coarse grind,
a kicker (strong collector) is added to the main scavenger bank to recover
mineralised coarse particles, and the scavenger concentrate is re-ground in a
small ball mill to liberate copper and help reject gangue, cornbined with the
cleaner tails, and then returned to the rougher bank,
The flotation reagents employed in a specific operation will depend on the

primary and associated minerals, and the gangue contained in the ore. For
the great majority of Cu milIs, pyrite must be rejected, so flotation is carried
out at a pR above 9, generated with the addition of lime to the ball milis.
But if the ore contains recoverable gold, consideration should be given to the
use of soda ash or caustic soda for alkalinity control, as the calcium ion is
known to depress Au. When the important by-product is molybdenite, its
recovery is encouraged by adding hydrocarbons (kerosene or fuel oil), and
possibly a moly specific collector, such as Cyanamid Aero 3302 or long chain
mercaptans. The usual strong collector, habitually added to the ball mill, is
amyl xanthate or if more selectivity is desired, isopropyl xanthate.
In simple circuits, xanthate addition to the ball mill will result in the recovery
of a relatively low grade rougher concentrate. This concentrate is upgraded
in the cleaners by diluting the pulp down to around 25% solids, and adding a
more selective Cu collector, such as ethylisopropyl thionocarbamate (Z-200),
in combination with a mild depressant, which can be more lime added to the
cleaners to increase the pulp pR. When the copper minerals present contain
a large proportion of chalcocite, and bornite, the Z-200 in the cleaners can be
replaced by cheaper collectors - dialkyl dithio-phosphates. When high grade
concentrates are obtained, it is always an indication that the ore being
processed is from thealteration zone, and contains chalcocite, covellite and
bornite. Another sophistication is to separate the rougher launders into
sections, and stage-add the collectors: a dialkyl dithiophosphate or alkyl
thionocarbamate in the ball mill, combined with separate collection of the
concentrate from the first few rougher cells, followed by a selective xanthate
in the second half of the roughers, and a strong xanthate in the scavenger
section. Outokumpu Oy's "flash flotation" cells are a variant of this scheme,
where cyclone bottoms are scalped prior to return to the ball mill feed.
Selectivity, i.e. activation of the flotation of a copper sulphide mineral before

the promotion of the iron sulphides and gangue material, may be a function
of the chelant properties of the collector molecule. Attempts to use
molecular models and relate the length of the collector bite to the
interatomic distances on the mineral surface h\ave not be en fruitful, but the
rate of formation of a ligand bond may well affect the rate of flotation of the
mineral species. The formation of metal cation - collector complexes is
certainly important in the evaluation of depressants and pulp contaminants.
Alkaline circuits are practically universal in rougher flotation of porphyry

copper ores. The exceptions are the acid circuits of El Teniente in Chile,
Lepanto in the Phillipines, and Flin Flon in Canada. The ore of the latter
two contains a significant amount of copper in arsenic minerals, such as
enargite. The standard collector has been a xanthogen formate, usually ethyl
butyl XF. This is very stable in acid media and only decomposes at pH 10 or
greater. It also has been used in alkaline circuits successfully (Inspiration in
the U.S.), and is a good collector in sea water pulps.
100
e
~ ~
1)fi ~
~
<=
e "
90
-x
.
::l
tn
"
~
~ JU -"
r- ~ ~
:o.fi m ~. ~
j w ~ " o
Lo
::J
:>0
e ~
....
-lb
w
""
~
70
1O 1~,}- 20 25 SO $0 - 40 4.5 50
Grind. X .. inullO zaB naa:h
-- 60 613 70
::10 ,5 eo
:"1 = Chllllcopyrii;e ... = eh-. leos i-t.a
Fig. 54.- Copper recovery for chalcopyrite and chalcocite ores

- Effect of grind on recovery
Fig. 54 plots copper recovery as a function of grind for the mines in Tables 56
and 57. The lower recovery of copper from ores where chalcocite
predominates is obvious. Of interest is that, at least for gross statistical
relations, the correlation between grind and recovery is tenuous; possibly this
is due to the difficulty in floating fines, as undoubtedly liberation should be
improved. The probable explanation of the lower recovery of copper from
chalcocite ores is a greater fraction of the copper content corresponding to
non-sulphide copper minerals, which in tUfn implies the presence of sericite
and clays.
Non-sulphde eopper ores
Sulphidisation, in the laboratory, is a viable process to increase recovery of

oxidised copper minerals with flotation. This has very rarely been borne out
on an industrial scale. The exception that comes to mind is Morenci, where
at one time the ball mili feed was treated with ammonium sulphide which did
improve recovery in an erratic manner, Later a modified LPF process was
tried. In this process the ball mili feed was contacted with sulphuric acid in a
Sulphide Mill Practico 209
trommel which fed into the ball mills. In the ball mill the acid pulp was
neutralised with home made calcium sulphide, manufactured on-site by
calcining Iimestone with pyrite. The crude calcium sulphide contained
significant quantities of metallic iron which precipitated out cement copper
from the liquid phase of the pulpo The sulphide mineral and the cement
copper were then floated with ethyl isopropyl thionocarbamate. A similar
process, but using iron powder, was used at a mine in Baja California where
smelter slag was floated with sea water and Minerec A, after an acid leach. If
liberation problems have required a very fine grind which has resulted in the
formation of slime, a common solution used to improve overall recoveries
has been to cyclone the ball mill output and separate the pulp into sand and
slime circuits. The sand circuit is handled in the normal way, and the slimes
are agglomerated with additives, such as water glass (silicates).
Malachite, azurite and cuprite can be concentrated by flotation, though the

production of colloidal fmes during crushing and grinding is a major
problem. Fatty acid flotation was once a standard process, but today
sulphidisation with NaSH and the use of benzothiazole as a collector is more
in vogue. Cyanamid recommends their moly collector, Aero 3302, for copper
carbonate flotation. The sulphhydrate is the preferred sulphidising agent as
depression occurs at a pH of over 9.2 and sodium sulphide can contribute
sufficient alkalinity to push pulp pH aboye this critical Ievel, The new
popularity of sulphidisation is due to the availability of reliable electrode
probes that can assure the availability of sufficient sulphhydrate ion
concentration for surface sulphidisation, while keeping the total NaSH
concentration lower than the depression range.
Chrysocolla, and other copper silicates, do not float efficiently, They have
been recovered by acid leaching and the LPF process, but if available in an
economic quantity, the most profitable method of recovery now is heap
leaching followed by liquid-liquid extraction and direct electrowinning.
Flotation of complex ores
The terminology complex ores general1y is given to mixed sulphide ores

containing copper, lead, zinc and silver. Setting up a processing scheme for
this type of ore is complicated and empirical, as al1 of these ores contain non-
floatable sphalerite and most of them have si~ficant non-sulphide lead and
silver present. The flow diagram for the flotation of a simple complex ore
has been shown at the beginning of this section. The basic design consists in
a sequential flotation of the different minerals employing depressants,
activators, and selective collector-frother combinations to inhibit the flotation
.1:0
TabLe 6S.- COMMONLY OCCURRING MINERALS IN COMPLEX ORES CONTAINING GALENA (PbS) ANO SPHALERITE CZnS)
,....
o
Secondary (oxidized) MineraLs Associated SuLphide MineraLs Gangue MineraLs

name composition name compos i t i on name composition
'Tj
l'
Smithsonite ZnC03 Pyri te FeS CaLcite CaC03 O
ChaLcopyrite OoLomite >-3
CaLamine Zn4Si207(OH)H20 CuFeS2 CaMgCC)3)2 ;J>
Hydrozincite ZnS(OH)6 CC03)2 Marmatite ZnSCFeS)x Ankerite CaCFe,Mg) CC03)2 t4
WiLLemite Zn2Si04 Argentite A92 S Siderite FeC03 5z
Cerussite PbC03 Native Si Lver Ag Rhodochrosite MnC03
Stromeyerite CUAgS Quartz '-1
AngLesite PbS04 Si02 ::r
(1)
Pyromorphite PbS CP04)3CL Marcasite FeS FLuorite CaF2 o
....
WuLfenite Arsenopyrite FeAsS Barite ~
PbMo04 BaS04
:o
Vanadinite PbS (V04)3CL Tetrahedrite CCu,Fe,Ag)12 Sb4S13 (1)
PLumbojarosite Fe6CS04)4(OH)12 Tennantite CCu,Fe,Ag)12 As4S13 c

Cerargyri te AgCL Enargite CU3AsS4R a
ti>
EmboL ite Ag(CL,Br) pyrrhotite FeS ~

::l
Bromyrite AgBr Stibnite Sb2S3 o-
Iodyrite AgI
o....
(1)
Argentojarosite AgFe3CS04)2(OH)6 .-J

(1l
,..,.
ti>
~.
Sulphide Mili Practico 211
of each species. Native copper, as well as native gold ancl silver, can be
easily floated at slightly aboye a natural pl-l with a xanthate and a collector-
Irother such as pine oil or cresylic acid. The dithiophosphates and
benzothiazole mixtures are also good collectors. The minerals normally
associated with lead-zinc ores are listed in Table 65. They must be taken
into account in the selection of reagents, as many contribute soluble cations
to the flotation pulpo
Chapter 12
OXIDE AND NON-METAlUC MINERAL FlOTATION
Non-sulphde, carbonate, and metallic oxide mineral flotation
So far, we have dealt with minerals in which the flotation mechanism involved
is the selective adhesion of sulphhydryl collectors to sulphide surfaces; the
technology that has been described is restricted to nascent bubble flotation
(operative at concentration ratios aboye 15-30:1) where bubble attachment
occurs through bridging to frother / collector complexes chemisorbed on
active sites. As we saw, the principal minerals with this behaviour are the
porphyry copper sulphides and sorne complex sulphide ores, such as high
grade zinc-lead-silver sulphides, and sulphidised metallic non-sulphides, This
section will review the behaviour of non-sulphide metallic minerals,
industrial minerals and salts (insoluble, soluble and semi-soluble). And in
contrast to nascent bubble flotation, data taken using experimental
procedures and designs based on measuring contact angles, or using :
Hallimond tubes, and/or mechanical flotation machines, give acceptable and
credible results with these minerals.
The reason for this change in viewpoint is that in bubble capture flotation, or
oily flotation, there is no need for the transfer of an electron from the
air /water interface to the mineral surface to trigger the deposition of a
collector, and to provide the bubble particle bond. Por non-metallics,
adhesion between the particle and bubble is only (or possibly mainly) due to
the hydrophobic nature of the final surface of the particle encountering the
bubble.
Due to this, the choice of reagents is antipodal for non-metallic flotation -

depressants and activators are chosen for their effect on the environment and
not on the mineral surface. Por example, cations depress when they react
with the collector and take it out of the process by making it insoluble, and
not by changing the nature of the mineral surface. Thus the properties and
condition of the water phase are of the greatest importance in these flotation
processes.
- 212-
Non-Metallic Mineral Flotation 213
The nature 0[' uily collector flotation
Concentration of non-metallic ores involves a completely diferent

mechanism of selective particle attachment - bubble capture by collision -
where the mineral surface must have natural or induced hydrophobicity to,
allow particle capture by bubbles. Ores involved are generaIly non-sulphides. \
They contain minerals which range from metallic oxides, carbonates,
sulphates, or exotics such molybdenites, and tungstates, to mixtures of inert
and semi-soluble ionic salts, and pure elements such as sulphur or iodine, and
impure carbon, such as graphite and coals. The distinguishing characteristics
of the bubble capture flotation process is that it involves concentration ratios
in the 2:1 to 5:1 range. The economic importance of the reagent market for
each of these types of minerals is illustrated in Tables 54 and 55 (page
176/77), which also show the tonnage treated by flotation by the different
mineral industry sectors in America from 1960 to 1985, and the average
concentration ratios which were required in winning the different types of
minerals.
When concentration factors of two or three to one are involved, processes

that replace or supplement flotation include gravity, electrostatic, magnetic,
or heavy media separation processes. These are currently still popular in
processing heavy metals, such as gold and members of the platinum group,
titania sands, iron ores, etc. In the US flotation industry, only 40% of the
tonnage floated is industrial minerals, but the volume of the flotation
reagents employed is 80% of the market total, as ten times the grams per
tonne are used when compared to selective sulphide flotation. Of the
industrial minerals processed, the most important are phosphates (45%) and
iron ore (15.5%), foIlowed by potash (7%), coal (6.7%), glass sands (4%)
and feldspars (2%).
The industrially important non-sulphide copper ores are malachite and

azurite (carbonates), cuprite (oxide), chrysocolla (silicate), atacamite
(chloride), brochantite and chalcanthite (sulphates). The last three are water
soluble. At one time these were recovered by the LPF process, now they
usually are heap leached, so are of importance in flotation only as
contaminants during the early life of a mine when processing the mixed
minerals in the ore cap, where they can provide undesirable activating cations
in the pulps. For example, atacamite refractory surface properties are such
that, during the Spanish colonial period, prir to the existence of blotters, the
mineral was exported from northern Chile for use as ink-drying sands.
Industrially important non-sulphide lead ores include cerussite (carbonate),

and anglesite (sulphate); and for zinc, smithsonite (carbonate). Iron ores of
value are haematite and magnetite (oxides), goethite (hydroxide), and

siderite (carbonate). This last can be a source of limonitc, a bad actor in
secondary enriched copper ore flotation, where it acts as a colloidal
depressant for the copper minerals, and can form a coating of ferromolybdite
on by-product molybdenum at high flotation pHs, cutting recovery
significantly (see Beas et al (1991)).
Carbonate ores, and oxides such as cuprite, can be successfuIly recovered by

sulphidisation; i.e., by artificiaIly coating the valuable mineral surface with the
metal sulphide produced by reacting the pulp with an alkali sulphide or
sulphhydrate. The mechanism of flotation, and the reagents employed as
collectors, modifiers and frothers, in this case, are similar to those uscd with
copper sulphides, with the caveat that sulphidised surfaces are labile and
must be kept in the presence of an active sulphidising agent during the whole
flotation process. If the ores are high grade, and the required concentration
ratio s to obtain an acceptable concentrate are low, these can be treated as
non-mctallics, and economically floated with oils, fatty acids, amines,
carboxyl collectors, long chain xanthates, etc, Thus, when sulphidised, they
act in the flotation pro cess as nascent-bubble capture minerals, and un-
sulphidised or otherwise activated, they act as bubble capture minerals.
Note that when sulphhydryl collectors are tested with metallic minerals that
exhibit oily flotation characteristics, the quantities of xanthates et al needed
for successful recovery is such as to provide more than monomolecular
surface coverage (i.e., at least a fivefold increase in dosage is required as
compared to nascent or normal flotation). Laboratory tests are cited which
show that synthesised dixanthogens act as collectors on sulphide and non-
sulphide minerals; misinterpretation of these data can lead to erroneous
mechanistic conclusions, as when minerals have surfaces fuIly coated with
sulphhydryl collectors the dithiolate is adsorbing in the same way as a
hydrocarbon oil, while when dixanthogen is detected as the active species in
nascent bubble flotation, the xanthate has chemisorbed and polymerised in
situ.
Polar, non-sulphide minerals and ores that are preferentiaIly upgraded by

flotation include alkali earth compounds, such as calcite, fluorite, apatite,
phosphates, barite, scheelite, powellite, magnesite, dolomite, etc. Their ionic
lattice provides the energy for hydration. Weak acids, such as xanthates or
thiophosphate collectors, are useless in their flotation because of the soluble
salts formed with the salts' cations. Soaps, which are unreactive, and fatty
acids or other high molecular weight acids with a carboxyl group, can act as ,
collectors for these minerals, but in the case of fatty acids it is because they
wiIl react with cations in solution to form soaps. Selective flotation, as known
".
with the sulphide minerals, is not possible due to the generally

undifferentiated surface chemistry of the group. This lack of selectivity is a
general characteristic of oily flotation, where the concentration ratio is under
5:1; here depressants and activators are the more important tools that the
metallurgist can use in concentrare upgrading. Because of the sedimentary
genesis of these ores, one crucial step that is difficult to scale up from
laboratory testing is the degree of comminution and the washing intensity
required to remove clays and other colloidal slimes. This difficulty is
compounded because only clean samples are usually sent in from the field for
testing, so it is nearly impossible to obtain representative results without
truck-load quantities of ore for a pilot scale test.
Fortunately, at concentration ratios below 5:1, usually the valuable mineral

needs to be separated only from a homogeneous ganguc, so the degree of
rougher selectivity required is technically simpler, as the desired product and
the reject have a different chemical composition. A key factor, though, is the
mechanical removal of clays prior to the flotation step, as these will blanket
any selectivity available from collectors, as well as providing unacceptable
contamination of the concentrate. The roughing step, for a normal or reverse
flotation, consists of selectively coating either the valuable mineral particles
or the gangue, but not both, with a hydrophobic reagent so only the coated
particles will float. The process sophistication then comes in the cleaning
steps, where pH and selective regulators such as sodium silicate, or organics
such as tannins, dextrin, starch, etc., can generate differential recovery.
Flotation of soluble salts has become a serious alternative to differential

crystallisation in the chemical industry only in the post-World War periodo In
nearly all large scale applications the working fluid used is water, or, for
soluble salts such as KCI from NaCl, the media is a saturated brine, so the
overall process can be classified as semi-soluble salt flotation.
Glembotskii (1963) has a short chapter at the end of bis book on flotation
which summarises a Russian publication, "Concentration of Non-metallic
Minerals by Flotation", that suggests there has. been m-6ch wider use of
flotation in the purification of soluble salts in Russia than in the West. He
provides sparse details of the separation of nitrates from chlorides, and
sulphates from chlorides, as well as the well known technology of sylvite
purification. He underlines the use of lead nitrate
\
as an activator for sodium
chloride in some fatty acid reverse flotation systems. When it acts as a
depressant, he suggests this is mainly with chlorides, where there is the
formation of lead chloride which is relatively .insoluble, and thus can coat
crystal surfaces.
He also summarises a rule proposed by S.A. Kuzin which postulates that in a

mixture of two soluble salts with greatly different solubilities, the most
soluble can easily be floated off from the brine; but when the second salt's
solubility approaches that of the floatable salt, it is depressed until neither
salt can be floated if their solubilities are equal. The example cited is
polyhalite - K2S0rMgS04-2CaS04-2H20 - which is relatively insoluble in
water, but its solubility increases in a potassium sodium chloride brine;
concurrently its floatability increases with a change in composition of the
brines which increases the polyhalite solubility. On the other hand its
solubility is reduced in magnesium and potassium sulphate brin es, which
reduces its floatability accordingly.
Structure of mineral surfaces
In thiol collector adhesion to mineral partic1es, the key factor, which

mandates the strength of reagent adhesion to the surface and the rate of
capture of particles by bubbles, is the active sites where electron transfer to a
conductive matrix is thermodynamically possible. To understand and
quantify this phenomena, lit is necessary to become conversant with the
literature on catalysis. When a reagent adsorbs rather than chemisorbs on a
particle surface, the macro properties of the structure are controlling, and
bulk measurement of the mineral properties gives meaningful data which can
be used to predict selectivity.
<=
'"
~
....
:J
'"
s,
""
~
OJ
<=
OJ
-:;;
r ,
d ..
1
:;::
s::
....'"
'"
,,
/
""' I
.13
s::
.....'-' ,r
I
............'" 1
I
'" ~
,es:>
: 11
r CI>

~"
, ..._~,/ Chel'l'll isorp-t:. i.on
Fig. 55.- Forces in adsorption and chemisorption

For an in-depth survey of pertinent adsorption phenomena, the reader is

referred to Leja's (1982) trcatise on the surface chemistry of flotation, which
explains how to estimate, and legitimise, the forces involved in adsorption
and chemisorption. The latter is covered well by Clark (1974). Fig. 55
highlights two important adsorption energetic factors in flotation: the
potential energy curve versus distance from the surface, with minima at rp
and re' shows that for chemisorption, both the repulsive forces and the
attractive forces are greater and closer to the particle surface than is the case
in adsorption; and that for adsorption the total adherence energy has a lateral
component, which increases as the area of the individual particle covered
mcreases,
The key difference in morphology between minerals which float by bubble

capture and those that nascent bubble float, is their surface uniformity. The
mineral particle can be conductive and still react in the bubble capture mode
if its surface does not have active catalytic sites. We can thus generalise that
the cations of these types of minerals must not be predominantIy multivalent.
lonic and covalent crystal lattices will exhibit two distinct energetic levels.
lonic minerals will have charged surfaces that hydrolyse, and covalent
minerals will be naturally floatable because they are hydrophobic.
Mechanism of collector adhesion
In a simplistic bubble adhesion model for a naturally floatable mineral, the

particle surface, in an aqueous pulp, is completely unhydrated so that an
approaching collector molecule will not be impeded from reaching the crystal
surface. But with a hydrated solid, the surface is covered with water
molecules and cannot be approached without displacing water molecules
energetically attached to the solid, i.e. activation energy is required to
displace water of hydration. This surface is usual1y shown as packed with
spheres with dipole moments in a semi-oriented order (Fig. 2, page 21). A
more modern picture of a hydrated mineral is to postulate that its surface is
covered with water clathrates. .
There has been considerable research on the crystal structure of water. Post-
Pauling (1960), the molecular clusters postulated in liquid water at ambient
temperature generally are pictured as having the shape of a distorted basket
formed by an outer shell of five- and six-membered rings held together by
hydrogen bonding, with two or three water ~olecules and a proton in the
centre. CastIeman et al (1991,1989) have studied both gas phase and liquid
phase water clusters. In the gas phase they have found that the cluster which
can be identified as either H+(H20b or H 30+(H20)zo has a very stable
form, and described it as weakly bound aggregates whose properties lie
between those of water vapour and liquid water. Here the cage structure
consists of 12 five-mernbered rings, where the oxygen of the water molecules
forms the vertices. In the liquid phase they have identified anionic water
cluster ions with the formulae OH-(H 20 )n and O-(H 20 )n which can show a
different n, whose value is more diffuse than the gas phase n=20.
More interesting from the flotation point of view is that , with n =20, these
clathrates contain 40 hydrogen atoms, 30 of which are involved in hydrogen
bonding (forming O-H-O bonds), which means they have only 10 surface
hydrogens available for hydrogen bonding (if an eleventh is associated with
the proton in the centre of the protonated water formula). This structure
was detected as quite stable in the vapour phase. In liquid water the picture is
more diffuse, as it appears that n=20 is a preferred charge in the cluster, so
then the available surface hydrogen bonds are ten. When n is less than 20,
available hydrogen bonds are less than 10.
bonding hydrogens
Fig.56.- Water clathrates, maximum 11 hydrogen bonds available
Returning to our description of the collector molecule approaching the

hydrated mineral particle, the obstacle that needs to be surmounted for
adhesion is water that must be displaced with the collector's polar head. If
the hydrated particle is scaled with clathrates, then more than one water
molecule necessarily must be displaced by the collector. More than that, the
number of water molecules per collector molecule should be multiplesequal
to the clathrate structure presento It is interesting that in a very recent paper
by Yehia (in press), he measured heats of adsorption on apatite which
suggested that 8 water molecules were being displaced per collector molecule
adsorbed. He has interpreted this number to be related to the size of the
arca covered by the polar group of the Aerosol, but it is interesting to
speculate if it is not more likely to be related to the hydrogen bonds available
on the clathrate structures, and that this number could be fixed by the surface
charge originally on the unhydrated mineral.
With non-sulphide metallic minerals, and with non-metallic minerals, bubble

attachment does not depend on the transfer of electrons; rather, large areas
of the desired constituent of the ore must be hydrophobic, either because the
crystal lattice is naturally hydrophobic (covalent bonding), or because the
hydratable surface is coated with a hydrophobic reagent. In this case,
collector adhesion strength, and the induction times for coursing bubble
attachment, can be calculated using classical Gibbs thermodynamics, and the
energies involved can be unambiguously measured using bubble angles.
Reagent quantities employed need to be orders of magnitude greater because
selective complete coverage of the target mineral surface is required to attain
high recoveries.
To determine if a specific collector will adhere promptly to a hydrated

mineral, the electric charges on the surface must be determined, as must the
effect of changes in the environment on these charges. Zeta potential, pl-I,
electro-scans of the surface, etc., all provide pertinent information on surface
properties that affect reagent selection. Concepts that are not pertinent to
this type of flotation include dithiolate formation, frother-collector
interacti:ons, selection of modifiers, activators and depressants based on the
physico-chernical effects on the mineral particles' surfaces, etc. These
exception are important to keep in mind, as the same reagents normally will
have contradictory action in the two types of flotation; so those metallurgists
involved in non-metallic mineral flotation should try to compartmentalise
their tools of the trade to avoid confusion.
The fundamental factor to keep in mind in evaluating experimental data, and

applying these to selection of flotation reagents and operating conditions, is
the difference in the bubble attachment process. In non-metallic bubble
collision flotation there is no need for the transfer of an electron from the
air /water interface to the mineral surface.. Adhesion between the particle
and bubble is only due to the hydrophobic nature of the surface encountering
the bubble, while with sulphide minerals thre is an electron bond which can
be as strong as a covalent one.
Collectors for industrlal minerals
These are reageuts or oils that coat andz'or react with mineral surfaces ami
make the mineral selectively water repellent and attachable to air bubbles.
Those that react with the mineral surface can be c1assified as surfactants that
chemisorb, such as hydroxamic acids on haernatite, or oleates on fluorite or
calcite, or long chain ionic surfactants which absorb physically as counterions
in the electrical double layer, such as dodecyl sulphonate on alumina or
haematite, and dodecylammonium chloride on quartz; but over 90% of the
TabLe 66.- COLLECTOR REAGENTS FOR NON-METALLICS
Type FormuLa Cation Anion ELectroLyte
ANIONIC COLLECTORS
Soaps and fatty acids RCOOM weak
ALkyL suLphates RS04M Na+, K+ RS0 4- strong
ALkyL suLphonates RS03 M Na+, K+ RS03- strong
ALkyL hydroxamates RCONHOM RCONHO weak
ALkyL phosphates weak

CATIONIC COLLECTORS
Primary amine saLts RNH3X RNH3+ weak
ALkyL ether amines RONH3X RONH3+ weak
Quaternary ammonium RN( CH3)3C L RNCCH3)3C+ strong
saLts
SuLphonium saLt RSCCH3)2CL RSCCH3)2C+ strong
ALkyL pyridinium saLt "-<0>-""C1 strong
MorphoLinium saLts
SuLphonium saLts RR'R"SCL RR'R"S+
ALky L xanthates ROCSSM ROCSS- strong
AMPHOTERIC COLLECTORS Czwitterion form)
N-aLkyL-2-aminopropionic RNH2CCH2)2COOM RNH+2CCH2)2COO- Na+, K+

acids
n-aLkyL-N,N-dimethyL RCCH3)2NCH2COOM RCCH3)2N+CH2COO- Na+, K+
gLycines
ALkyL betaines CCH3)3NRHCCOOM CCH3)3N+RHCCOO- Na+, K+
rcagents used as collectors for industrial minerals are soaps and Iatty acids
(including tall oil frorn the naval stores industry).
Because the raw materials for the manufacture of these reagents are
generally natural oils and fats, products from different manufacturers are not
necessarily nterchangeable. It should also be emphasised that, as is the case
for detergents, cationic and anionics should not be used simultaneously, as
they can react and neutralise each other. Water quality, which is a problem in
sulphide ore flotation if heavy metal ions are in solution, is more of a
problem in industrial mineral flotation, which is normally associated with
soluble inorganic salts, as a11 cations and ions that are in solution in the
flotation pulp can interfere with the reagent action.
As a large number 01' anionic and cationic collector molecules used in

inorganic flotation are derived from natural oils and fats, an overview of fat
chemistry is helpful. Table 67 lists the approximate fatty acid content of a
few typical natural oils and fats. This assay cannot be taken as a specification,
as the use 01' caustic to remove acids from fats is ancient and continues in the
production 01' soaps; steaming to deodorise, and hydrogenation to control the
unsaturated acid content, are pro cesses now used for nearly a century, and
other purification processes are common, so the selling specification for
derivatives 01' these natural oils will vary for each manufacturer, Caveat
emptor!
Fatty acids
A typical process for obtaining useful col1ectors is treating crude vegetable

oils byalkali refining, water washing and drying, bleaching, hydrogenation,
and deodorising. The free fatty acids neutralised with caustic can be
recovered by centrifuging; bleaching and winterising results in removal of
stearins, while hydrogenation can go as far as a solid shortening, Animal fats
and oils are processed by rendering and hydrogenation, plus inter-
esterification and isomerization, to improve their quality. Classic detergents,
whose chemistry is similar to that of the flotation oils, consist of a
hydrophobic hydrocarbon chain containing 8 to 18 carbon atoms and a
hydrophyllic functional group which may be:
anionic, e.g., -os03- or -s03;

\
cationic, e.g., -N(CH3)3+ or CSHSN-+;
zwitterionic, e.g., -N+(CH3)2(CH2)2COO-;
semipoLar, e.g., -N(CH3)20;
or
Petroleum based linear alkylbenzene sulphonates, alkylbenzene-ether

sulphonates, alkylglycerol-ether sulphonates, alkyl esters of isothionate,
methyl alkyl laurates, fat based alkyl sulphates, fatty alcohol-ethylene oxide
sulphates and soaps are the most common anionic detergents. Cationic
surfactants inelude quaternary alkyl ammonium halides and other amines.
Table 67.- FATTY ACID CONTENT OF OILS AND FATS (WEIGHT %)

No. Soy Fish Coco TaL- Lard
carbon bean oiL nut Low
atoms Aeid FormuLa oiL oiL
8 Caprilie C7H15COOH 8.0
10 Caprie C9H19COOH 7.0
12 Laurie C11H23COOH 48.0
14 Myristie C13H27COOH 7.0 17.5 1.0 1.0
14 MyristoLeie* C13H25COOH 0.2
16 PaLmitie C15H31COOH 8.3 16.0 8.8 21.0 28.0
16 PaLmitoLeie* C15H29COOH 17.0 3.0
18 Steari e C17H35COOH 5.4 1.0 2.0 30.0 13.0
18 OLeie* C17H33COOH 24.9 27.0 6.0 43.0 46.0
18 LinoLeie* C17H31COOH 52.7 2.5 5.0 6.0
18 LinoLenie* C17H29COOH 7.9 0.7
20 Araehidie C19H39COOH 0.9 2.0
20 Araehidoni e* C19H31COOH 20.0
22 CLupanadonie* C21H35COOH 12.0
* = unsaturated aeids
Predicting the behaviour of a particular reagent suite in non-metallic

flotation is complex because the reagents ionise, dissociate, and hydrolyse, so
modifying agents become more important than in sulphide flotation, and
their action less easy to forecast; thus, the description of the activity of a
particular col1ector must always be qualified in terms of the environment.
Laboratory adsorption data, electrophoresis measurements, zeta potential,
etc., all seem, and are, very pertinent, but can be misleading because of the
interference of other ions.
The tendency for a col1ector to adsorb onto a mineral surface will depend on:
1.- The length of its hydrocarbon chain

2.- The number of chains in the structure
3.- The number of ionic groups in the structure and the type of
associated ions
4.- The pOSltlOI1 of the ionie groups on the ehain and the distance
separating the groups
5.- The nature of the ionic groups and the cross-sectional area of the
groups
6.- The nature, position, and number of nonionic hydrophyllic groups
These faetors affect the properties of the collector:
1.- The base strength, whieh can be quantified by the PKa values
2.- The solubility of the collector
3.- The critical miscelle concentration (eMe)
4.- The Krafft point
5.- The tendency of the collector to adsorb onto a specific mineral
TabLe 68.- THE STRUCTURE OF SOME ANIONIC COLLECTORS
Surfactant group Structure pKa
o
/
CarboxyLic soap R- C H+ or Na+/K+ 4.7 0.5
\
o
R-O-S-O-
11
SuLphate Na+ or K+
11
o
o
R-S-O -
11
SuLphonate Na+ or K+ :::: 1.5
11
O
Hydroxamate R-C-N
I Na+ or K+ :::: 9
o ol-
11
R\l9
Phosphoric acid P H+ or I'il:l+/K+ 7 2.5
H/ \_
224 FLOTATION - Thcory, Reagents and Ore Testing
Anionic colector chemistry
The most important anionic collectors are shown in Table 68. Mishra (1988)
points out that carboxylates are fully ionised at pH values aboye 10; the alkyl
sulphates do so at a pH aboye 3 (at a lower pH they hydrolyse to alcohol and
bisulphate); while alkyl sulphonates are stable even in strong acids. The long
chain Iatty acid carboxylates go through an ionisation process between pH 4
to 10. At pH around 8, the ionisation is to a 1:1 ion complex,
RCOOH-RCOO-. Similar intermediare species are present in the case of
alkyl sulphates and sulphonates in the half-ionisation pH range; as a
consequence, these compounds can act as a dual collector/frother at those
pHs. In fatty acid flotation it has been noted that adsorption, and recovery,
improve with the degree of unsaturation in the alkyl group. Collector
properties also change if aryl groups are substituted for sorne or aIl of the
hydrocarbon groups.
As in the case of sulphhydryl collectors, over-oiling will also convert a

coIlector into a depressant, so with the aboye surfactants there are
concentrations aboye which they wiIl not act as coIlectors. The phenomena
in this case can be explained by misceIle formation and the adsorption of
these macro-molecules on the mineral surface. A list of critical misceIle
concentration (CMe) values are shown in Table 69.
Note that, within the same species, the CMC occurs at progressive1y lower
concentrations as the chain length increases. There also is a significant
difference if the potassium salt is involved; i.e., the CMC value for the Na
oleate drops from 2.1 x 10-3 to 8.00 x 10-4 for potassium oleate.
TabLe 69.- CRITICAL MISCELLE CONCENTRATION FOR

ANIONIC COLLECTORS
Surfactant CMC. moLes/Litre
Na capryLate 3.5 x 10- 1
Na capriate 9.4 x 10- 2
Na Laurate 2.6 x 10- 2
Na myristate 6.9 x 10- 3
Na paLmitate 2.1 x 10- 3
Na stearate 1.8 x 10- 3
Na eLaidate 1.4 x 10- 3
Na oLeate 2.1 x 10- 3
Na dodecyL suLphate 8.2 x 10- 3
Na dodecyL suLphonate 9.8 x 10- 3
Na dodecyL benzene suLphonate 1.2 x 10- 3
The majority 01' the Iatty acids used industrially are of vegetable origin, rather
than animal fats, Tall oil Iatty acids are by far the most consumed organic
reagents in non-metallic ore flotation, Their main components are
unsaturated oleic, linoleic and linolenie acids, and the saturated palmitic and
stearic acids. All commercial products also contain a varying amount of rosin
acids, which are frothers and normally do not act as selective collectors.
At ambient temperature, the effectiveness of these fatty acid collectors is a

function 01' solubility, which favours the shorter chain alkyl groups. For
example, at a pulp temperature 01' 24C there is very low activity with
palmitic acid, which has a melting point of 63C; but if the pulp is
conditioned with the collector at 70 to 75C, recoveries rise to the 80 to 90%
tange. Note that it is the conditioning temperature that is important, as this
is the point in the process at which collector adsorption on the mineral
surface occurs, thus the flotation banks do not need to be heated if the
collector concentration and conditioning pH is controIled correctly. Varying
the conditioning temperature will affect recoveries more than concentrate
grade, thus it is an important variable to study in the optimisation of a circuit.
As has been noted previously, interfering ions are a greater problem in non-
sulphide flotation. The presence of calcium or iron in solution can result in
activation of the gangue with coIlector complexes. Here, pH can be an
important variable to correct these problems.
Sulphonate type reagents were first developed as anionic coIlectors for

selective concentration of low grade iron ores or for the removal of iron
contamination from glass sands, feldspars and other ceramic raw materials.
They now are employed in the flotation of barite, celestite, fluorite, apatite,
chromite, kyanite, mica, rhodochrosite, cassiterite, etc ..
Sulphonates are water soluble or dispersible liquids or pastes. They are

normally fed to the conditioner as a 5 to 30% water solution or dispersion.
Most of them have frother properties, so mixtures of sulphonates are a
possibility in obtaining grade and recovery without the addition of auxiliary
frothers. Total dosage is generally in the 130 tfl 2,000 g/t range.
Cationic collector chemistry
The only cationic collectors used in metallic mineral processing are the
amines. These reagents ionise by protonation. They are c1assified as
primary, secondary, tertiary or quaternary amines, depending on whether

they have one, two, three or four hydrocarbon radicals bonded to the
nitrogen atom. Prirnary, secondary and tertiary amines are weak bases,
whereas the quaternary amines are strong bases, and are compIetely ionised
at all vaIues of pH, whiIe the ionisation of the other amines is pH dependent,
so that the metallurgically active concentration is pH dependent. The next
tabIe lists a representative group of collector amines:
TabLe 70.- NITROGENOUS CATIONIC FLOTATION AGENTS
n-amyLamine CSH11 NH2 Primary aLiphatic amine

n-dodecyLamine C12 H2S NH2 Primary aLiphatic amine
Di-n-amyLamine (CSH11)2 NH Secondary aLiphatic amine
Tri-n-amyLamine ( CSH11)3 N Tertiary aLiphatic amine
AmyLamine
hydrochLoride SaLt of primary amine
TetramethyL
Ammonium chLoride CH3) 4N) CL SaLt of a quaternary amine
AniLine C6 HS'NH2 Primary aromatic amine
p-toLuidine CH3C6H4NH2 Primary aromatic amine
BenzyLamine C6HSCH2NH2 Primary aLiphatic amine
DiphenyLamine C6HSNH'C6HS Secondary aromatic amine
ALkoxypropyLamine R-O-( CH2)3NH2 Approx. moL. wt. 210
R = C8 - C10 mixture
TechnicaL coco RNH2 Approx. moL. wt. 203
Primary amine R = C8 - C18' SS% C12
TaLLow ami ne RNH3Ac Primary C18
Acetate R = approx, 96% C18
Representativeionisation constant for sorne typical amines are:
MethyLamine 4.4 x 10- 4 TrimethyLamine S.S x 10- S

DimethyLamine S.2 x 10- 4 DodecyLamine 4.3 x 10- 4
Miscellization: is a characteristic of hydrocarbon chain surfactants. It is

important in interpreting the molecular behaviour of fiotation systems.
Miscelles are aggregates of collector ions of colloidal size that form by van
der Waal's bonding between hydrocarbon chains. They form at a critical
miscelle concentration (eMC) because the hydrocarbon chain is non-ionic,
causing a mutual incompatibility between water complexes and the
hydrocarbon chains. These aggregates cannot be seen, but their presence is
detected because of light scattering when a Iight beam is passed through the
solution. The free energy decrease that the system experiences when this
phenomcna occurs is significant - 0.6 kcal per mol per CH z unit. For
dodecylamine, therefore, the free energy decrease is 7.2 kcal/rnole on
formation of miscelles,
Table 71.- CATIONIC AMINE COLLECTOR STRUCTURES

H
P~iMill'Y aIIlines n_IV
'H
R
Secooda~y aMines ni"'N-H
n
TelltiallY a~ines ni'N-Hu
HU
n t/
........ N el
Quate~nary i\MMoniW1
salts n''''''' 'ni
DiaMines It
" + ,"
R-N (CHa)x N,
H
H H
, t/
H N /H
TriaMines H_~t (CHZ'x-C-(CHZ
1 t
>x",
I H
H
PYl'idiniuJIl salts CN --
t
'H
/
Cl
/CH 2-CH2" _
MoX'pholiniuM salts R-NH 0+ CI
""-CH
2-CH2/
\
R'
1+ _
SulphoniuM salts R-S -CI
I
R"
Flotation systems that employ amines as eolleetors includc: siliea flotation in

cleaning phosphate rock; separation of KCI from other salts using primary
aliphatic amines; cleaning beryllium minerals, using coco or tallow amines;
upgrading spudomene in lithium mineral processing; etc.
The structures of typical eationic amine collectors are shown in Table 71.
Solubility: as the length of the hydroearbon ehain increases, the solubility of

the arnine is reduced. The effective ionie solubility, though, is pH dependent.
Table 72.- MOLECULAR SPECIES SOLUBILITY, pKA, AND POLAR HEAD

DIAMETER FOR CATIONIC COLLECTORS (from R.W. Smith, 1988)
Collector Solubi l ity Diameter of

moL/litre poLar grp A
n-dodecyLamine 10.63 2 x 10- 5 3.7
N-methyl-n- 11.0 1.2 x 10- 5
dodecylamine
N,N-dimethyl-n- 9.74 7.2 x 10- 6
dodecylamine
di-n-hexyl amine 11.01
N-(n-dodecyl)- 6.8 9.3 3.6 x 10- 7
propyL-1,3-diamine
n-decyLamine 10.63 5 x 10- 4 3.7
n-tetradecyLamine 10.62 1 x 10- 6 3.7
Factors afTecting selective flotation
Leja (1982) has pointed out that earboxylates, sueh as oleie aeid, can
ehemisorb, as well as adsorb, dependent on the ionic salt eontent of the pulp
(Fig.57).
58ft Sclution Concentration = 0.062;5 molal
0.9
'O 0.8
Ql FLUORITE =
.o
o 0.7
~ BARITE
'E 0.6
Ql NBF
.c ".J!I
o 0.5
!!J'" .'
~Ql 0.4 ,.,,0"'"
oe 0.3 ,.,1" Na~o4
o O""
t!::l"
'''".';1",.........~ ........"
olf
NBF
'8 0.2 .'
<1
u: 0.1 ......::::::::::::: ~~---~
o
o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Fraetion Oleate Adsorbed
Fig. 57.- Effeet of brine eomposition on the ratio of chemisorption

Note that sodium fluoride inhibits chemisorption of fluorite, and that the
sulphate ion does the same for barite, These data are interesting in that they
suggest the mechanism by which anions can influence collector selectivity, but
it must be noted that there is not clear evidence whether the oleic acid salt is
ionised or unionised on the mineral surface. Leja points out that co-
adsorption of non-ionised carboxylic acid contributes to an increase in overall
hydrophobicity of a particle, not so if the carboxylic acid is ionised.
As in the case of sulphide minerals, excessive amounts of collector will inhibit

Ilotation of inorganics. The mechanism visualised is the formation of a
second collector layer on the particle surface which is "upside-down'', with the
carboxylic anion exposed to the liquid phase, generating hydrophobicity.
Addition of an oil will neutralise this extra layer and help flotation. This is
probably one reason why fuel oil is a popular collector in copper roughers
when molybdenite is an important by-product, as the primary xanthate
collector can over-absorb on the molybdenite's natura11y floatable surface
layer and could well act as a depressant.
Carboxylic acids become more and more insoluble as the alkyl group
lengthens. In addition, the acids and soaps are powerful frothers, even more
so if the alkyl groups are unsaturated; therefore they norma11y can be used as
collectors without the need to add a separate frother, an ability which is
frequentIy a disadvantage. Because of their limited solubility, an increase in
pulp temperature reduces the amount of collector required for a given
recovery, and winter weather requires additional amounts of collector to
maintain recovery,
Metallie oxide minerals mili praetiee
According to Sutherland and Wark (1955), in the early days of iron flotation
oleie aeid salts were known to be promoters for haematite, as Sulman and
Edser (1924) patented their use of oleates or stearate soaps with sodium
silieate to separate haematite from feldspars. Based on later work of Keek et
al, they report that for 100 gjt oleic aeid and 50gjt of terpineol as the
frother, the optimum pH for reeovery is a littIe aboye pH=7. As recovery is
inhibited below pH = 5.5, they eonclude that oleic aeid itself is not the
collector, Calcium is a depressant at very low eoneentrations, as are metals
sueh as ferrie eations, and Pb 2 + , Mn2 + , Sn2 + , Cu 2 + , and Sn4 + , as these a11
form insoluble soaps with the fatty acid. \
The SME Mineral Proeessing handbook reports, in sorne detail, eurrent

flotation experience at the Republie Mine in Miehigan's Northern Peninsula,
where they are proeessing 22,000 tpd of a speeular haematite ore eontaining
36.5% Fe, and producing a concentrate with 65.4% Fe ancl 4.95% Si02, with
a recovery of 4%. The operator's experience is that conditioning is crucial
to minimising the amount of collector required. The collector is a distilled
tall oil containing 91% oleic acid, 6% rosin acid, and 3% unsaponifiables.
The amount of collector required is about 600 g/t of ore, ten times the
amount reported by Sutherland and Wark, and laboratory work at the mine
has shown that very intense conditioning will halve the amount of collector
required for a given recovery. The flotation circuit treats 18,000 tpd, with a
sand/slimes split using two conditioners, each consisting of a number of 4.5 x
4.5 x 5-ft cells equipped with 24" ship-type propellers driven at 290 rpm by
motors drawing 12.5 hp. The pulp density is important, and is maintained at
65 to 70% solids, which gives a residence time of 5 to 8 minutes. Recovery is
reduced in winter, so steam injection is used to maintain pulp temperatures
at 70F. Similarly the Hanna Mining Co.'s Groveland mine in Dickson
County, Michigan is processing 15,000 tpd of an ore containing magnetite,
haematite and iron silicates assaying 34.5% Fe, which they upgrade to 64.4%
Fe and 6.3% Si02, using flotation, elutriation and magnetic concentration.
Overall recovery is 42.2%. In their flotation section the deslimed elutriator
underflow is conditioned at 66% solids in a bank of ten 60 cu ft conditioner
cells equipped with ship-type propellers driven by 25 hp motors. Retention
time is 10 minutes, and the conditioning pH is held at 5.8 to 6.0. The reagents
used are:
Reagent gft Point of addition
PetroLeum suLphonate 400 1st conditioner

Fatty acid 150 1st conditioner
FueL oiL 200 1st conditioner
SuLphuric acid 500 2nd conditioner
Sodium sil i cate 150 1st rougher and 1st cLeaner
Industrial minerals mili practice
The use of flotation to upgrade industrial minerals has become more popular
as a consequence of declining reserves of high grade material. The products
usually treated include barite, beryllium, borates, calcite (limestone), clay,
graphite, gypsum, feldspars, fluorspar, glass sands, iodine, kaolin, kyanite,
mica, pegmatites, phosphates, potash, quartz, rare earths, spudomene, tale,
verrniculite, wollastonite, and others. The sources used for these summaries
are the SME Mineral Processing Handbook (1985), Taggart (1945) and
Redeker and Bentzen (1986).
Barite (BaS04) is found in combination with fluorite, celestite, quartz, galena,

sphalerite, and many other minerals, Its primary use is in the formulation of
drilling muds, for which the composition is important to obtain the maximum
specific gravity. If used for chemical processes, the specification is usual1y
+ 98%; while for drilling muds the specification is a specific gravity of 4.0 to
4.3.
Fine grained deposits, or the rejects from a gravity concentration operation,

are frequently upgraded by flotation. The material is ground to liberation
size (which can be minus 40 down to 325 mesh, depending on the ore). The
coarser material is deslimed in cydones, then floated using petroleum
sulphonates, or mixtures of sulphonates. The drawback to these flotation
reagents is that if the end product will be used in drilling muds, the col1ector
must be destroyed to avoid frothing during drilling. To do this, the
concentrate must be heated to 200 - 260C.
Outside the U.SA., barites are frequently associated with sulphide minerals
and are recovered from the tailings of the planto In this case the conector
used can be oleic acid, fatty acid sulphonates or sodium laurate in an alkaline
circuito Barite is easily depressed by acidifying the flotation circuit with
sulphuric acid. If quartz is present it can be depressed with the addition of
silicates to the pulpo
Bauxite is a rock rather than a specific mineral. It consists of hydrated

alumina minerals (gibbsite, boehmite, diaspore, etc.) contaminated with
quartz, iron oxide minerals and titania, Calcite, siderite, mica, kaolinite,
pyrite, and marcasite are also found as minor contaminants.
A case history of flotation is quoted by Norman Holme (1986): the ore's

minus 20 mesh washer tails, which amounted to about 45% of the original
plant feed, were discarded. The plus 20 mesh wasground to minus 35 mesh
and deslimed. The sands were conditioned for 2 minutes at 60% solids with
3.3 kg/t of natural petroleum sulphonate, a similar quantity of fuel oil, and
1.1 kg/t of sulphuric acid. The pulp was then diluted to 14% solids and
floated at a pH of 3.1. After 2 stages of c1eaning, about 86% of the alumina
in the deslimed feed was recovered in a concentrate representing 76.9% of
the deslimed feed weight.
Beryllium is obtained commercially fro~ beryl (3BeOAl z0 3'6SiO z),

bertrandite (4BeO'2SiOiHzO), phenacite (2Be'SiO z), chrysoberyl
(BeOAlz0 3) , and barylite (Be zBaSiz0 7) .
According to Paul Chamberlin (SME, page 29-7), beryl has not been treated
successfully industrially, though a complicated laboratory processing scheme

has been developed for the pegmatite type ore. The process starts with
desliming, then adjusting thc pulp pH to the 3 - 5 range with sulphuric acid
and removing the mica by flotation with tallow amine acetate, Tailings from
the mica float are dewatered, washed several times, and diluted to 25%
solids. Hydrofluoric acid is added to maintain pH 3, and the silicates (beryl,
feldspar etc.) are floated with coco or tallow amines away from the silica.
The silicate concentrate is then scrubbed with ca1cium hypochlorite to
rernove reagents, the clean pulp thickened, washed, and diluted to 60 to 70%
solids and conditioned with a fatty acid or petroleum sulphonate. The
flotation is performed at pH 6, if the collector is a fatty acid, or reduced to 3
to 4 with H 2S04 when using a sulphonate, The final beryl concentrate must
be cleaned several times to obtain a satisfactory grade. This step quite
obviously might be improved using column flotation.
Borates: the four most important boron minerals are listed in Table 73.
Colemanite is the only mineral that has been concentrated by flotation on a
regular basis.
TabLe 73.- COMM=RCIALLY IMPORTANT BORON MINERALS
MineraL FormuLa B203 % sp. gr. Water SoLubiLity
ULexite NacaB509'8H20 42.9 1.96 niL

coLemanite ca2B6011'5H20 50.8 2.96 niL
Borax Na2B4) 7'1 OH2 0 36.5 1.72 soLubLe
Cornet Na2B407'4H20 50.9 1.91 sLowLy soLubLe
Colemanite ores containing 20 to 30% B 203 must be upgraded. For flotation

concentration, they are ground to minus 20 mesh and deslimed with cyclones.
High clay ores may require various stages of desliming. The pulp is washed,
dewatered and conditioned at 20 to 40% solids with 500 to 1,000 g/t of a
mixture of sulphonate collectors. When carbonates are a problem, starch,
quebracho or dextrin can be added to depress them. No frother is required
when floated at the natural pH of 8 to 9. The rougher concentrate is then
normally cleaned to about 45% B203 . Erratic frothing can usually be
corrected by an improvement in the desliming step.
Table 74 shows the glass fiber market specification that marketable

colemanite must meet in the U.SA.. Flotation is also employed to separate
borax Irom sylvite in the borai process. The collectors that are normally
used are oleie and naphthenic acids, to which xylene, turpentine and kerosene
may be added. The flotation can be activated with barium salts, Boric acid is
naturally floatable. This property has been used for many years in the nitrare
industry, which has recovered boric acid from the iodine plant acidification
step when the world price and the ore was favourable. It also has been used
to process colemanite by decomposing the ore with saz, to form boric acid.
This is done by grinding the ore finely, suspending the solids in a saturated
boric acid solution, and then adding the gas. The reacted pulp is passed on
into flotation cells, where a small quantity of frother is added, and a boric
acid concentrate collected. The concentrate may be further deaned, or
dissolved in a hot, saturated boric aeid, from which pure acid is crystallised by
cooling.
TabLe 74.- GLASS FIBER COLEMANITE SPECIFICATIONS
Maxmum % Mnimum %
B203 44.0 36.0

Si0 2 2.5
Fe203 0.5
AL 203 1.0
CaO 32.51 31.0
MgO 1.0
Na20 &K20 1.0
SrO 1.9
S~ 1.5
H20 (free) 1.0
As (ppm) 100
MnO (ppm) <10
Cr203 (ppm) <10
Ni (ppm) <10
Co (ppm) <10
Cu (ppm) <10
V (ppm) <10
Chromite has been obtained by flotation from chrome ores employing a

mixture of natural petroleum sulphonates anJ crude tall oil with fuel oil as
collectors. The process consists of crushing and grinding the ore to flotation
size, followed by scrubbing and complete desliming. The sands are then
onditioned at 60 to 80% solids with the colIector mixture. The reagent dose
normally used is 500 to 1,000 g/t of a fatty acid-fuel oil mixture, plus 100 to
250 g/t of sulphonate, In sorne cases the addition of sulphuric acid to reduce
the pH to around 5 will markedly improve the selectivity. With sorne lode
ores, the addition of sodium fluosilicate, at the rate of 1 to 2 kg/t, wil! do the
same.
If the ores contain considerable amounts of carbonates, these must be floated

off first using sodium silicate and fatty acids in an alkaline circuit, fol1owing
which the feed is deslimed and floated as above. If excess magnetite is
present, this should be removed by magnetic separation, as it wil! float with
the chromite.
Cement rack, if of poor quality, can be upgraded by flotation. The following is

a good testing case history, reported by Holmes (1986), so is quoted in
extenso.
American Cyanamide studied a low grade limestone assaying 56% CaC03,

20% volcanic glass, 5% feldspars, and lesser amounts of quartz, opaline
silica, clay, biotite, hornblende and zircon. Liberation was only achieved after
grinding to 82% minus 200 mesh and was found to be impossible to deslime
due to losses (25 to 45%) of soft calcium carbonate to the slimes. On
flotation it was found that the use of a synthetic modified sulphonate
produced a very tough, difficult to break, froth, so testing was performed with
a 1:1 mixture of heavy fuel oil emulsified with water. The col1ector emulsion,
plus a frother, was then stage added in increments of 50 g/t for a total of 300
g/t. In addition, tests were run to determine the effect of preconditioning the
pulp with sodium silicate and soda ash. Stage concentrates were kept
separate to permit calculation of cumulative grades and recoveries.
All test runs produced acceptable concentrate grades (78% CaC03) and
recoveries after cleaning the rougher concentrate. The best results were
obtained at 3 minutes of conditioning with 1,000 g/t sodium silicate and
500g/t soda ash (pH = 9.3), followed by flotation with 300 g/t of sulphonate
promoter, 300 g/t fuel oil and 150 g/t frother, stage added to the six flotation
stages, obtaining an overall CaC0 3 grade of 81.4% and recovery of 92.6%.
Feldspar: glass and ceramic grade feldspars are produced from alaskite (a
uniform granite-like ore), or from more isotropic pegmatites.
The alaskite ores in Spruce Pine, North Carolina, according to Redeker and
Bentzen (1986), have the chemical and mineralogical composition shown in
Table 75. The run-of-mine is crushed and wet ground and then rid of its
minus 200 mesh fraction by cycloning and classifying. The fines, though high
in feldspars, are discarded as waste.
TabLe 75.- TYPICAL COMPOSITION OF SPRUCE PINE ALASKITE
Wt % Mi nera L \It. %
NazO 5.1 Soda FeLdspar 4Z.9

KZO 3.4 Potash FeLdspar 14.7
CaO 0.9 Lime FeLdspar 6.4
AL ZO:3 15.4 Quartz Z8.0
SiOZ 74.4 Muscovite 7.5
Fez0:3 0.4 Iron-garnet mineraLs 0.5
Loss on ignition 0.4 CLay very Low
The coarse product is differentially floated in three stages to remove the

undesirable iron-garnet minera1s and quartz. The first stage consists of 3 to
5 minute conditioning of a concentrated pulp (55 - 60% solids) at a pH of 2.5
to 2.7 with 750 g/t H 2S04 , 125 g/t of ta1low amine acetate, 500 g/t of fue! oil,
and 50 g/t of frother. The pulp is diluted to 20 - 30 % solids, which increases
the pH to 3.0 - 3.3, and floated for 2 to 3 minutes in wooden acid-proof
Denver DR-type flotation machines. In North Carolina these consist of 50
cubic foot cells in a trough with five flotation mechanisms. The mica
concentrate is c1eaned and screened, and plus 80 mesh sold to mica
producers for further processing. In the second stage the iron-garnet
impurities and the remaining mica is floated with anionic collectors. The
tails from the first flotation are cycloned to 65 - 75% solids and conditioned
for 5 minutes, at pH 2.2 to 2.5, with 200 g/t of sulphuric acid, 250 g/t of
petroleum sulphonate and 25 g/t of frother, then again diluted to 20 - 30%
solids, which increases pH to the 2.8 - 3.2 range, and floated for 5 minutes.
The concentrate of iron-garnet minera1sjoins the waste slime stream.
The third stage consists of the cationic separation of feldspar from quartz in a
hydrofluoric acid circuit. The tails from the second circuit are again
dewatered to 50 - 60% solids, and conditioned for 3 minutes at pH 2.5 with
600 g/t of HF, 200 g/t of tallow amine acetate, 50 g/t of kerosene, and 25 g/t
of frother. The dilution to 20 - 30% solids increases pH to the 3.0 - 3.5
range, and the product is floated for 5 minutes making a glass grade feldspar
concentrate ready for shipment. A small portion is further upgraded by dry
grinding in pebble milIs and iron scavenged with high intensity magnetic
separation, prior to selling it to the pottery rldustry. The quartz tails can be
c1eaned by floating off the impurities, sized and dried for sale as a high purity
product for use as a fused quartz (see Fig. 58).
to uaste
Reag, e
Rod Mili
trollHle 1screen
ea nesh
uaste
Ha net
filters~
__
Sep ~
Dry Ground Mica ~
Pebble IHU D~er
Pottery feldsp l' GI ss Gla s liaste!
Feldspar Quartz
Fig. 58.- Feldspar, mica and quartz flotation, Spruce Pine, NC
+1 inch .. a,...:s:....t_e ---.. a lka line
FIHE tUCA
H!1dr o
size:rs
Uast.e iron
.. inerals ...- ..J
COARSE MICA
HF - amine flotation
QUARfZ ~
FELDSPAR
filter
Fig. 59.- Weathered decomposed pegmatite processing, North Carolina

When the raw material is weathered pegmatites, the circuit used is similar,
except that in the first stage, after wet grinding in a rod milI, a very coarse,
fairly pure mica can be recovered with a trommel equipped with 1/8 "
screens. The minus 1/8 inch material is passed to Humphrey spirals, where,
because of the shape factor, the flat mica concentrates in the outer sides,
where it is removed by splitters. The impure mica is cleaned by acid circuit
flotation with olcylamine acetate as the collector (Fig. 59).
TabLe 76.- REAGENTS EMPLOYED IN THE FLOTATION OF

FELDSPAR-MICA-QUARTZ-LITHIA IN 1985
(aLL vaLues in thousands, except costs)
TOTAL Cost
FLOTATION pounds vaLue US$/t.ore
CoL Lectors
Unspecified amine 125 $117 0.483
Amine saLt 287 252 0.333
Fatty and rosin acid soaps 1,000 350 1.500
Fatty acid derivative 251 107 0.877
PetroLeum suLphonate 210 133 0.234
FueL oiL 886 96 0.668
TotaL 2,759 1,055
Frothers
AL iphati e aLcohoL 40 21 0.101
Pine oiL 39 24 0.037
PoLygLycoL 8 1 0.060
PoLygLycoL ether 35 25 0.095
PetroLeum based bLends 48 34 0.334
TotaL 170 105
Depressants
HydrofLuoric acid 525 202 1.144
Lignin derivatives 35 1 0.300
Activator
HydrofLuoric acid 599 271 0.813
pH reguLators
SuLphuric acid 1,743 56 1.543
Caustic soda 797 91 0.598
FLoccuLants
ALuminium saLts 351 10 0.140
Unspecified poLymer 33 36 0.040
EFFLUENT TREATMENT
pH reguLators
Caustic soda 719 14 3.420
Lime \1,299 55 2.421
Floccu Lants
Unspecified poLyacryLamide 5 6 0.040
PoLyacryLate 55 52 0.253
PoLyethyLeneimines 11 9 0.800
source:USBM
In 1985, according to the Minerals Yearbook, there were ten mills operating,
with a daily capacity of 2.65 million tonnes/year of ore. Actual operations
were at 71% of capacity. Concentrate produced amounted to 235,000 t of
quartz, 550,000 t feldspar, 60,000 t mica, and an undisclosed amount of lithia.
The reagents used are shown in Table 76.
Fluorspar: most fluorspars must be upgraded for marketing. Typically the

ore is washed and sized in trommels, Minus 10 mesh is floated, while the
coarser fraction usually is purified in a heavy media separation. The
flotation feed is wet-ground to 100 mesh, and the pulp heated to 55 to 85C
and conditioned 5 to 10 minutes with 500 to 3,500 g/t of NaZC03 and
NH 4C03 to disperse the c1ays and slimes, plus 250 to 500 g/t of quebracho to
depress carbonate minerals; at which point 250 to 1,000 g/t of fatty acid is
added as the collector, and the pulp conditioned for an additional 5 - 10
minutes. Then the pulp is diluted to 25% solids, and the fluorspar is floated.
The concentrate then needs multip1e c1eanings prior to reaching

specifications. The temperature is maintained during all the operation
Flotation Feed Final

1812B tpb bit/. -ZBB mesh tailh~s+-_-l
lBx
Soda ash + quebracho + _
CaFz fatty acid Concentrare
~~ ~J
C~---+---I
18- 48 ft 3 cells
I.m", :
COlICentrate
S'x4' Regrind nm FLUORSPAR COHCEtiTRATE
Fig. 60.- Fluorspar mill, Reynolds Mining Corp., Texas

Copper sulphate so!ution

tlerofloat 211 + ~IIlC
Plallt
Feed
2inc Roughel') B- 50 ft 3ce lls
Zinc tails G' )( 12'" Rod mil!
to th ickenee
and Fluorspar &' x O' Ball un
Circuit
'-- -.62:1. Zinc FINAL CONCENTRATE
to thickener and filter
Soda Ash +
SodiullI si licate
IFLUORSPAR cIRcUIrl
FINAL FLUORSPAR
CONCEIITRATE
No 1 Cleanel'
L- ---' 5 - 5Hrt3
celis
Fig. 61.- Mili circuit for a fluorspar-zinc heavy media concentrate

(Reagents and conditions sho~ in Table 77)
through the injection of live steam. Laboratory tests indicate that a

saponified fatty acid can be employed at a lower temperature if the gangue is
siliceous rather than carbonaceous.
TabLe 77.- REAGENTS FOR FLUORSPAR, SULPHIDE

AND BARITE FLOTATION
Reagent Consumption Reagent Consumption

gft gft
Lead-pyrite fLoat Zinc float

Na ethyL xanthate 45 Na diisopropyL di-
MIBC frother 20 thiophosphate 150
Zinc hydrosuLphide variabLe Copper suLphate 220
MIBC frother 25
FLuorspar fLoat Barite float

Soda ash 650 - 1500 Sodium cetyL suLphate 220
Sodium siLicate 700 Stepanflote 24,
Quebracho 700 ALcoLac EC111,
Starch 500 MIBC frother -, 45
OLeic acid coLLector 220
Fluorspar is supplied in three grades: acid, ceramic and metallurgical; they

must meet the specifications shown in Table 78.
TabLe 78.- FLUORSPAR GRADE SPECIFICATIONS
Acid Ceramic MetaLLurgica L

.....
#1 #2
CaF 2 97 95 - 96 80 - 90+ 60 (effective)

Si0 2 <1.5 <3.0 <3.0 <15
S <0.1 <0.3
Fe203 <0.12 <0.12
Pb trace <0.5
% effective CaF2 = %CaF2 - 2.5 x % Si02

Glass sands: sand is ubiquitous, but rarely of a suitable quality even for use as
aggregate in concrete, and certainly very rarely of the chernical quality
required in glass ma1cing, quartz melts or as a flux in smelting. It is also a
very large volume, low value, commodity, Normal processing requires
washing and grading as a minimum.
Sands of a quality suitable for glass making are frequently obtained by

purifying river or beach sand by reverse flotation. If the impurities include
iron and heavy metaIs, gravity and magnetic methods to remove the
impurities may be justified. When the impurities involved are carbonates and
feldspars, these can be removed by flotation. Depending on the impurities
present, glass sand flotation can be done either in acid or basic circuits. If
the sand is to be floated in an acid circuit, which is usually the case when it is
contaminated with high amounts of refractory heavy metals, it is normally
attrition-scrubbed at 60 to 80% solids, followed by desliming. The flotation
feed, also at over 70% solids, is conditioned with an acid at a pH of about 2
plus or minus 0.5, and with 250 to 1,000 g/t of petroleum sulphonate as the
collector, depending on the amount of impurities present, At the end of the
conditioning period 50 to 100 g/t of frother is added (mostly nowadays, a
polyglycol ether frother is used), and fue! oil is added when slimes are
excessive. The pulp is then diluted to 20 to 40% solids for flotation in
conventional cells. The overhead contains the iron minerals, heavily stained
quartz, and the heavy refractory minerals, and is usually sent to waste in most
plants.
The alkaline flotation route is chosen when high carbonate content is the
impurity, but again dependent on the impurities involved. The commonly
used collector is a fatty acid. Conditioning is done at 65 to 75% solids, with
the pH adjusted to about 9.5 plus or minus 0.5, using caustic soda or soda
ash, and with about 500 g/t of fatty acid. At the end of the conditioning, a
frother is added. Frothers used inelude pine oil, the water soluble glycols, or
MIBC, with dosage a function of the frother and the sand quality - usually 25
to 50 g/t, Fue! oil can be used to help disperse the fatty acid and control the
froth if the impurities tend to vary. Solubilizing the fatty acid with caustic has
allowed conditioning at a lower pulp density than that used with raw fatty
acid or fuel oil/fatty acid emulsions. Conditioning at too low a pulp density
has been one of the major reasons for unselective quartz flotation with fatty
acids, according to Redeker and Bentzen (1986). After conditioning, the
pulp is diluted to 20 - 40% solids and floate. The overhead contains the
carbonates, and is usually discarded.
If the sand is heavily contaminated, the quartz will be floated with a primary
amine. High solids conditioning, at a neutral or slightly acid pH, is carried
242 FLTATlN - Theory, Reagents and Ore Testing
out with SO - 200 g/t of the amine collector, followed by flotation.
The comrnercial sand specifications cover the size and the impurities
contained. For containers and float/flat glass, the sand should be less 20 plus
1S0 mesh. Closer screening is required for fused and optical grades of glass.
Chemical specifications are usually tolerances, rather than minimum values.
Undesirable are colour producing metals, such as chromium, nickel,
manganese, cobalt, and copper. Iron is less deleterious in small amounts for
clear glasses, and actually desirable for amber and green glasses. Refractory
heavy metals are always excluded,
TabLe 79.- SPECIFICATION FOR COMMERCIAL QUARTZ ANO FELOSPAR
Composition Quartz Sands FeLdspar Concentrate

and physicaL Foundry GLass GLass Ceramics
characteristics (1 ) (2) (3) (4)
TotaL Si02 % 95 - 99 97 - 99 65 - 68 65 - 68
Quartz, % 90 - 96 96 - 98 NS 8.0 max
AL2~ % 0.5 0.3 18.5 0.6 18.0 min

K2 + Na202' %
0 0.3 0.2 12.5 1.0 13.0 min
Fe2~ % NS 0.3 - 0.03 0.1 max 0.08 max

LOI, % 0.2 0.5 - 0.2 0.5 0.4
In 1985, in the USA, there were eight milis upgrading sand by flotation. They
processed 2.5 million tonnes of sand, and produced 2,163,000 t of clean sand;
a concentration ratio of less than 1.2:1. Reagent use is shown in Table 80.
The following case history from Mathieu and Sirois (1984), describing a
proposed Canadian operation for separating quartz from feldspars based on
sea sands, shows the effect of conditioning times, collector dose and pR on
the separation. Note that recoveries are highIy sensitive to pR. The
collectors used were petroleum sulphonate, tallow amine acetate, propane
diamine dioleate, and a Hercules, lnc. experimental iron mineral collector
identified as CX-62. Conditioning in a11 cases was done at 65% solids, and
the flotations at 25% solids. The specification targets are detailed in
Table 79.
Non-Metallic Mineral Flotaton 243
TabLe 80.- REAGENTS ENPLOYED 11'1 THE FLOTATION OF GLASS SAND IN 1985
<aLL vaLues in thousands)
TOTAL Cost
FLOTATION pounds vaLue US$/t.ore
CoLLectors
Diamine 91 $76 0.140
Fatty and rosin acid soaps 2,639 812 1.180
PetroLeum derivatives 596 70 0.885
TotaL 3,469 993
Frothers
Pine oiL 4 2 0.007
PoLygLycoL 15 13 0.150
TotaL 22 18
Depressants
Sodium siLicate 20 0.100
Activator
Sodium hydroxide 36 3 0.730
pH reguLators
SuLphuric acid 1,922 95 2.465
caustic soda 1,885 173 1.425
Dispersants
Sodium SiLicate 344 24 3.440
EFFLUENT TREATMENT
pH reguLators
SuLphuric acid 210 7 0.563
FLoccuLants
Source: USBM
The schematics show the laboratory lock tests employed and the flow sheet of
the pilot plant. Tables 81 and 82 detail the experimental design and results.
'\1
Poundry Screen
sand No 45 sieee
(').10 I'~
--!l'w
sand, Fine acreen
0.5-'- US tlo 00 acd leach
t,rt x
[1] ~Fe
~~~~M'
Silica
sand !
00 soln
i
2.2~
", Coarser---'---..
I~" [I]+-------------=~=-....J
Gidss sand
Flotation
SIMes tailings
( 0.1 wt x) Glass
feldspar
Uei - HGHS
Separatop
~
oQ--S
Ceranie - - - - -.... ~Fe Cone .~--_I Cross-rin9
Feldspar nagnetle separatnr
Fig. 62.- Pilot plant flowsheet for sand beneficiation
Fresh Uater
Filtration t-_-+,~rt:r: Coneentrate

1-_--,-_ _--' Ccylce 1)
"2 o + H 2 S04 residua 1 collector
Fe- Fe Idapar-
conl:entrate
(c~cle 11
Filt1'ation Quarb:
Concentrate
(e~cle 2)
Fe- Fe Idepar- oncentrate H O + H SO

Ccyele 2) .. Co Llectoes
Fe-Fe ldspar Concentrate

(e~cle 3)
Fig. 63.- Lock cycle tests and pilot plant - Glass sands
Table 81.- EXPERIMENTAL DESIGN AND RESULTS - GLASS SANDS

Process: Feldspar-iron flotation: single stage versus two stage
Reagents and Conditions
Operation Time Density Reagents kg/t
min % Solids HF H2S04 eollector pH
Test 1
Fe conditioning 5 65 1.2 1.2 PS 2.1
Fe flotation 4 25 2.4
Feldspar conditioning 5 60 1.0 0.2 1.3 TAA 2.2
Feldspar flotation 4 24 2.5
Test 2
Fe conditioning 5 65 0.5 ex 6.9
Feldspar eonditioning 5 60 1.0 1.1 0.5 POO 2.0
Feldspar flotation 3 24 2.5
Test 3 PS + POO
Fe eonditioning 5 65 1.3 0.3 + 0.45 2.9
PS = petroleum sulphonate; TAA = tallow amine aeetate
POO = propane diamine dioleate; ex = Hercules eX-62 iron eoLlector
Test Produets Analysis, % Oistribution
Wt% Feldspar Fe203 Quartz Feldspar Fe2~ Quartz
Fe Cone 1.6 10.4 4.90 79.7 1.7 34.0 1.4
Feldspar 16.0 55.2 0.49 42.7 89.1 34.0 7.7
Quartz 82.4 1.1 0.089 98.2 9.2 32.0 90.9
Feed (cale) 100 9.9 0.23 89.0 100.0 100.0 100.0
Fe eone 1.0 13.1 5.55 78.1 1.3 27.3 0.9

Feldspar 16.4 55.1 0.43 43.0 88.2 34.6 7.9
2 Quartz 82.6 1.3 0.094 98.1 10.5 38.1 91.2
Feed (cale) 100 10.2 0.20 88.9 100.0 100.0 100.0
Fe/feldspar 17.4 54.0 0.81 43.8 90.1 64.2 8.6

3 FeLdspar 82.6 1.2 0.096 98.1 9.9 35.8 91.4
Feed (caLe) 100 10.1 0.22 100.0 100.0 100.0 100.0
Graphite: graphite ores are c1assified as flake, vein and amorphous. The
most valuable, flake graphite, is found in metamorphosed feldspars. To
pro cess it, it must be carefully ground to liberate, but not degrade, the coarse
flakes. It is naturally hydrophobic, so was historically the first ore ever
concentrated by flotation (the Bessel brothers in the 1880s). Despite the
natural floatability, the separation of feldspar, quartz, mica and marble
gangue is normally improved by the addition of a small amount of kerosene,
and floated with pine oil or MIBC as the frother. Selectivity is improved by
adjusting the pH with caustic soda to the 7.5 to 8.5 range. Cleaning with
regrinding usually is ineffective because the graphite will smear over the
surface of the gangue 011 grinding, and make it floatable. Because of this,
and the natural interleaving of mica with the flake graphite, it is not possible
to produce grades of over 95% by flotation. Super purity graphite is made by
chemically leaching out the impurities from the concentrate.
TabLe 82.- EXPERIMENTAL DESIGN - GLASS SANDS FLOTATION - EFFECT

OF VARIABLES ON FELDSPAR-IRON FLOTATION FROM QUARTZ
Constants VariabLes ResuLts
Quartz Cone FeLdspar-iron eoneentrate

FeLdspar Fe203
Si0 2 Reeov Content Reeov fLoated
% % % % %
1 2.0 97.6 96.5 73.0 85.4 53.4

CoL Leetors 2.1 97.8 95.5 68.1 86.9 56.9
PS, 250 g/t pH 2.2 97.9 94.6 64.7 87.6 54.0
PDD, 500 g/t 2.3 98.0 92.7 58.3 88.6 59.0
Conditioning 2.4 98.1 88.7 50.2 90.8 65.7
time, 5 min
11 Condi-
PS, 150 g/t tion 3 98.3 89.1 50.6 92.0 60.0
PDD, 400 g/t min- 5 98.1 95.3 66.9 89.6
pH 2.2 utes 8 98.0 96.5 68.9 87.2
10 98.4 93.1 57.3 91.2 54.0
111 CoL- 0.45 97.4 97.3 76.4 86.7 55.2

Leet- 0.55 98.1 94.9 67.1 89.5 57.2
or 0.65 98.2 93.0 59.3 90.0 62.4
Conditioning dose 0.75 98.2 90.1 52.5 90.6 64.1
time, 5 min kg/t 0.85 98.3 87.7 46.6 91.7_ 68.6
with 0.75 98.1 91.5 55.9 90.1 59.6

re- 0.65 98.2 90.7 53.9 90.6 63.2
pH 2.2 eyeL- 0.55 98.0 91.0 54.9 90.6 60.3
ing 0.45 98.2 90.5 53.4 91.0 60.8
0.45 98.0 90.9 54.5 89.6 59.9
Gypsum rarely has been concentrated by flotation industrially. In the

laboratory it is easily floatable using sodium oleyl sulphate, fatty acid
sulphonates, sodium taurate, and primary and secondary amines. Alum is
reported to help selectivity. If gypsum is to be depressed, it can be done with
sulphuric acid, gelatin, or tannic acid, depending on the circuit employed.
Iodine is produced in Chile from the strong sodium nitrate mother liquor,
which contains over 500 gil of salts. The iodine is cut with S02 in a packed
absorption tower, followed by a reactor, where the effluent of the packed
tower is mixed with fresh mother liquor. The reaction between the iodide
Irom the SOz tower with the iodate in the original mother liquor produces
metallic iodine. The effluent from the reactor is a dilute iodine pulp which is
fed to standard small Denver rubber lined cells, without the use of reagents.
The concentrate is washed with water and the pulp melted under pressure to
purify the iodine. The only operating prablems encountered are when the
nitrate ores are high in boron minerals which also react with the SOz and are
converted to boric acid. Boric acid will float without reagents, and
contaminate the iodine. When the ores treated are very high in borax, the
effluent from the tower can be sent to flotation to recover boric acid, prior to
entering the cutting tank where the solid iodine is precipitated.
Kaolin: of the clay minerals, kaolinite, which is used as a filler for paper, as
well as in the manufacture of porce1ain, is the main componeiit that requires
extensive treatment to meet selling specifications. Clays are normally
washed, suspended in water, and "fractionated" into two sized fractions by
means of centrifuges or hydrocyclones. The fine white material is used for
paper filling, This fraction is frequentIy contaminated with 2 or 3 % of
anatase (Ti0 2) , which gives the product a yellow tinge which cannot be
bleached. To remove the anatase, a process known as ultra-flotation is
employed, where the anatase is removed by "piggy backing" on a second
mineral, which adsorbs it. The second mineral is usually ground limestone.
The mixture is floated with the usuallimestone reagents (soda ash, fatty acid,
and fue1 oil), and the resultant limestone concentrate removes all the anatase.
Kaolinite clays witha brightness of over 90 are produced by this method.
Kyanite is an aluminium silicate used primarily in the production of stee1

furnace refractories. To upgrade the mineral, it is wet ground in rod or ball
mills to 45 to 75 microns (200 to 325 mesh), deslimed, thickened, conditioned
and floated. To remove the pyrite, which is sold as a colouring agent in the
glass industry, a xanthate rougher and cleaner flotation is employed.
Conditioning is done at a neutral pH with 50 g/t of ethyl xanthate, and then
25 g/t of pine oil is added as the frother.
A second flotation step is used to remove the small amount of residual clay,
mica and pyrophyllite contained in the kyanite. The pulp is conditioned with
50 to 100 g/t of cocoamine acetate, at a neutral pH, and the contaminants
floated off.
\
The final step in the purification process is a bulk flotation of the kyanite. To
do this the pulp is thickened to 70% solids and given intense conditioning
with 1 to 1.5 kg/t of petroleum sulphonate at a pH of 2.3 to 2.6 (sulphuric
acid). The pulp is diluted to around 30% solids and the kyanite is floated and
given 2 or 3 stages of cIeaning, adding more sulphuric acid to keep the pH

below 2.6. The concentrate is conditioned with sodium silicate to remove the
flotation agents, dewatered and dried. Dry magnetic separation of the
kyanite removes iron minerals and iron stained material which the petroleum
sulphonate also activates. Overall kyanite recovery is 80 to 85% of the feed,
Phosphates: most of the phosphate used for fertiliser production is obtained

from phosphate rock which is found as unconsolidated or consolidated
sedimentary surface deposits composed of coarse calcium phosphate pebbles
mixed with fine grains, all intermixed with layers of sand and clay. Many
deposits contain slimes and contaminant iron phosphates, alumina and other
foreign matter. Hard rock deposits generally are of a sufficiently high grade
that they do not need to be concentrated for sale.
As is usual with industrial minerals after mining, which is frequently by

dredging, or hydraulically, the run of mine is first screened to eliminate inert
coarse material, then washed to get rid of as much slimes and cIay as
possible, and crushed and ground to liberate the mineral. The usual mineral
in Florida is collophanite, a fluorapatite carbonate with substitutions in the
lattice, which contains 31 to 36% P20S'
The phosphate rock is separated from the diluting quartz by selective

flotation with fatty acids, derived from tall oil, a by-product of the paper
industry. To properly disperse the fatty acid it can be presaponified with
caustic or ammonia, or this mix is added in high per cent solids to the
conditioning step. Fuel oil is also added to reduce the amount of high cost
fatty acid required to have good recoveries. The primary phosphate flotation
is not sufficiently selective to eliminate all the stained quartz and slime
contaminated gangueo The cleaner steps are more complicated than for
other minerals because of surface coating and activation of the gangue,
therefore it is necessary to strip off the collector from the primary flotation
by attrition with sulphuric or fluosilicic acid and float off the quartz in
cleaner, using an amine collector.
Reagent usage in the primary flotation is 400 to 1,000 g/t of fatty acid, about
a tenth of that amount of caustic or ammonia to obtain pH of 8 - 9 in the
conditioner, plus 500 to 2,000 g/t of fuel oil. Conditioning time is 5 to 8
minutes, and flotation 3 to 6 minutes at 20 - 30% solids. The de-oiling step
requires sufficient sulphuric or fluosilicic acid to wash at pH 3- 4. The
cleaner quartz flotation requires 50 - 100 g/t of amine collector, plus 25 to
150 g/t of kerosene at the head of the cells, and 3 to 6 minutes flotation time
at 25 to 30% solids.
PHOSPHATE
MINE ------- OVERBURDEN ---~ WASTE
I
'i'
PHOSPHATE ORE (Matrix)
v
I
MILL
v
I
PRIMARY DESLIMING -------------~ -0.1 mm WASTE
I
v
LOG WASHING - SCRUBBING
v
I
SIZING
----------~-0.1 mm WASTE
I
'1'
"
+1.5 mm -1.5 mm +0.5 mm
'1'
-0.5 mm +0.1 mm
I I
v
I
v
"
PEBBLE AGGLOMERATlON DOUBLE FLOTATlON
FLOTATlON
I I
"
FEED "
FEED
I I
v
"
CONDITIONER CONDITlONER
I I
v
"
AGGLOMERATlON FATTY ACID --~ -0.5mm WASTE
FLOTATlON FLOTATlON
I
-- - - - -~
"
DEOILING
I
AMINE "
FLOTATION --~ -0.5mm WASTE
I
"
PEBBLE COARSE "PHOSPHATE "
FINE PHOSPHATE
CONCENTRATE CONCENTRATE
Fig.64.- Flow diagram for a phosphate rock plant

TabLe 83.- REAGHITS EMPLOYED IN THE FLOTATION DF PHOSPHATE ROCK

CaLL vaLues in thousands)
TOTAL Cost
pounds vaLue US$/t.ore
CoL Lectors
Secondary amine 4,847 $1,600 0.223
Unspecified amine 4,800 2,880 1.200
Amine derivative 6,885 2,444 0.221
Quaternary ammonium saLt 1,644 553 0.319
Fatty and rosin acid soaps 151,447 23,015 1.419
Fatty acid derivative 10,770 1,268 0.121
FueL oiL 130,943 12,521 1.268
Kerosene 8,863 1,025 0.170
TotaL 275,410 47,950
Frothers
PoLygLycoL ether 82 41 0.019
Unspecified poLyoL 13 11 0.400
TotaL 145 84
Depressants
Sodium si Licate 1,170 139 0.190
Starches, ceLLuLose 374 52 0.350
pH reguLators
SuLphuric acid 134,866 3,398 1.3a0
Caustic soda 5,649 455 0.526
Ammonia 28,362 2,626 0.294
FLoccuLants
Anionic poLyacryLamide 19 17 0.003
Dispersants
Sodium siLicate 126 19 0.009
During 1985, in the USA, there were 22 flotation plants in operation at an

average of 75.4% of capacity. They treated 109,715,455 t of phosphate rock
grading 9.94% P 203, using 3,480 gallons of water, 12.3 kW-hrs of power, and
consuming 0.54 lbs of grinding b1alls per tonne of ore processed. Phosphate
produced was 24,620 tonnes of 31.21 % P 203 Reagent consumption is shown
in Table 83.
Potash: the most common potash ore is sylvite, which consists of a mixture of
potassium and sodium chloride, usually assayingabout 23% KCI, 72% NaCl
and 4% others, which includes clays, iron oxide and silica. The separation of
salt from KCI by flotation is relatively straightforward. In the early process
the NaCl, activated by lead or bismuth ions, was floated off with anionic
organic reagents and fatty acids, the KCI reporting to the tails. Concentrates
containing over 60% K20 could be obtained, but at a cost of rather high
reagent consumption.
Most current plants are doing the opposite, floating off the KCI (sylvite) and
leaving the halites in the tails. The run-of-rnine ore is crushed and classified
dry, then wet ground in a saturated NaCI and KCI solution, down to
liberation size (usually minus 4 plus 65 mesh). In the conditioning step, clay
depression reagents such as starch are added at a dosage dependent on the
ore grade, then the pulp is conditioned with the flotation reagents at 50 to
75% solids, The normal collectors are primary aliphatic amines of about C6
to C Z4 carbon chain lengths, used at arate of 250 to 1000 gil. Polyglycol
ether frothers are common, with fuel oil added to control the froth. Flotation
is then carried out at 20 to 35% solids, for 3 or 4 minutes, after which the
concentrate is cleaned. It is washed in a saturated KCI solution, and/or
water, and dried.
Rod MiLL screen undersize

(-8 mesh)
,.
Cyc Lones -----~l
(+100 mesh)
,. -100 mesh
Scrubbers
,.I -200 mesh
screens ~
I
+100 mesh
,.
Scrubbers
I
,. -200 mesh
screens ~
I
+100 mesh
I Hydro ~l
<4- - - - <41 + 250 mesh
,.<4
I
-270 mesh
Coarse
reagent
1
Conditioners
CycLones
1-<4
Fine +200 mesh
reagent
~l
I . -150 m Hydro --~
+150 mesh 1
~
-270 mesh
,.
to sLime
I Conditioners thickeners
L-~tO<4 I
FLOTATION
Fig. 65.- Potash scrubbing-desliming-reagent conditioning system
Most important in amine flotation of KCI is good scrubbing, desliming and

depression of the clays. The pulp is vigorously scrubbed in agitated tanks or
tumblers, to rernove c1ay particles from the crystal surfaces, and to disperse
the c1ay in the brine, Ores with over 3% slimes are scrubbed in stages, with
each stage followed by a screening to remove the slime-laden brine. The
screen oversize is repulped with c1ear brine at each stage. An optimum
power input must be determined and pulp densities are critical in effective
scrubbing, as dilute pulps reduce partic1e contacto The presence of carnallite
produces problems as, when it dissolves, it adds magnesium chloride to the
brine, and causes partial crystallisation of KCI and NaC!. If carnallite content
fluctuates when the low cycle occurs, the chlorides tend to crystallise on the
original crystal surfaces, while during the higher carnallite content period the
recrystallised chlorides tend to remain in the slurry as extremely fine crystals
that will be lost in the desliming steps, The slime removal circuit is
complicated, as can be seen from Fig. 65, which shows a typical procedure.
Potash collectors, according to the SME Handbook, are primary amines

produced from beef tallow, which is a mixture of palmityl, stearyl, and oleyl
amine, with alkyl groups containing 16, 18 and unsaturated 18 carbon-chain
atoms, Hydrogenation converts the oleyl to stearyl amines, saturating nearly
al! the groups, which increases hydrophobicity and surface adhesion.
Another quality on the market is distilled tallow amines. The final collector
is produced by converting the amines to acetates or hydrochlorides, which in
the flotation plant is neutralised to 90 - 97% with caustic,
Slime depressants in common use include starches, guar gum, dextrines, and
synthetic polymers, such as polyglycols and polyacrylamides. High molecular
weight polyglycols are effective, but affect frothing. Most amines in
commercial use contain smal! amounts of low molecular weight amines, such
as capryI, lauryl and myristyI, which provide frothing. MIBC, in potash
flotation, can be used when there is over-frothing, to provide more selectivity.
Typical reagents used (according to the SME Handbook), and their
consumption, are shown in TabIe 84.
TabLe 84.- REAGENTS USED IN POTASH FLOTATION
Dispersants, siLicates, etc. 50 - 100

SLime depressants, starches 450 - 1000
Others 70 - 120
CoLLectors, amines 50 - 150
There were five potash flotation pIants in the USA in 1985, operating at 70%
of capacity. Ore treated had 11.3% K20 and amounted to 10,090,900 tonnes,
from which 1,517,270 tonnes of 59.15% potash was produced. The flotation
reagents consumed are shown in Table 85.
TabLe 85.- REAGENTS EMPLOYED IN THE FLOTATION OF POTASH IN 1985

CaLL vaLues in thousands)
TOTAL Cost
pounds vaLue US$jt.ore
CoL Lectors
Primary amine 675 $454 0.138
Quaternary ammonium saLt 230 83 0.069
Fatty and rosin acid soaps 990 198 0.350
FueL oi l 1,953 213 0.425
TotaL 3,874 958
Frothers
ALiphatic alcohol 525 294 0.054
Polyglycol ether 1 1 0.002
Petroleum based blends 246 39 0.112
Total 772 334
Depressants
Gums and dextrans 1,732 332 0.208
pH reguLators
HydrochLoric acid 156 6 0.059
Floccu Lants
Anionic polyacryLamide 46 42 0.008
Unspec. poLyacryLamide 17 31 0.009
Unspecified polymer 424 215 0.192
Scheelite: flotation concentration of this tungsten mineral is typically with a

high grade fatty acid collector, though at several plants the fatty acid is
supplemented with synthetic modified sulphonates. Holmes (1986) reports
improved results with the use of 100 g/t of sulphonate collector to replace
the normal 400 g/t of oleic acid. The sulphonate collector was emulsified
with No. 6 fue! oil at a 1:1 ratio in water. A grind of 75% minus 200 mesh
was followed by a buIk sulphide float in which the pulp was conditioned at
30% solids with 2,000 g/t of sodium silicate followed by a two stage addition
of 100 g/t of emulsion per stage, as well as 7.5-15 g/t of glycol frother. The
pH remained at 9.7-10.0 throughout roughing and cleaning. The results are
shown in Table 86.
TabLe 86.- SCHEELITE FLOTATION RESULTS

Product Weight Assay W0 3 % Recovery
% % W03
\
Feed 100.00 0.54 100.00
Sulphide conc. 4.66 0.41 3.52
W03 concentrate 1.88 25.24 87.58
MiddLings 6.92 0.32 4.10
Tails 86.54 0.03 4.80
254 FLOTATION - Theory, Rcagents and Ore Testing
The competing concentrare recovered with 400 g/t of oleic acid was only 7 to
11% W0 3, at the sarne recovery leve!. Holmes notes that grinds as fine as
95% minus 200 mesh may be necessary to provide adequate liberation of the
scheelite. Sodium silicate additions of 500 to 2,000 g/t to the rougher and
each c1eaning stage may be necessary for gangue depression and slime
dispersa!. If carbonate gangue is present, quebracho (10 - 50 g/t) may be
needed to depress the carbonate in the c1eaners, though care must be taken
not to over-dose as quebracho affects sulphonate coIlection power. If the
process water is hard, soda ash should be used to reach the operating pH of
8.5 to 10.5.
Crushing
I
PebbLe '"
miLL grinding
I
-40 mesh
'"
cycLones ----~ to waste
Humphrey'" spiraLs
!
I I
'"
cLassifier '"
Spi raL Concentrate
'"
sands '"
tabLe
'"
Rougher fLotation ceLLs concentrate
,. I ~--,
concentrate '"
dryer
scav~;;&Le '"
Spudomene
con~entrate
concentrate
I
'" ngs
TaiLi 'I" '1- - - - ''"1
to Waste dewatering
'"
Spudomene
Concentrate
Fig. 66.- Foote spudomene flotation circuit
Spudomene and co-products (Pegmatite ores): lithium is obtained from its

ores only in the U.SA., otherwise the lithia concentrate is marketed directIy
for use in the glass industry by operations in Australia, Canada, and Africa.
There were two spodumene producing operations in America: Foote
Mineral, at King's Mountain, North Carolina, utilising a combination of
gravity separation and flotation to make a coarse spudomene concentrate

from spirals, and a fine concentrate from flotation (see Fig. 66 for a
simplified schematic); and Lithium Corporation of America, at Hill City,
South Dakota, recovering spudomene, mica and a feldspatic sand. A typical
mineral assay of the Kings Mountain (Fig. 67) spudomene ore is summarised
in Table 87.
TabLe 87.- MINERALOGICAL ANALYSIS OF KINGS MOUNTAIN SPUDOMENE ORE
MineraL Wt. %
Spudomene 15 - 25
Potash feLdspar 12 - 15
Soda feLdspar 28 - 33
Quartz 25 - 35
Muscovite and others 5 - 15
Ballllill
J1eagents
lJJl)-{;I:~ Quartz
Feld-
spar
.. -211B nesh
s l iaes
'------01-----1-+- Reagents
Fig. 67.- Spudomene and feldspar-quartz-mica flotation, Kings Mt, NC
In both processes the ore is wet-ground to flotation size and about 250 g/t of
NaOH added to the ball or pebble mili feed. The ball milis operate on a
closed circuit with cyclones delivering a -65 mesh product to the process.
The pulp is then diluted and recycloned to remove the -400 mesh fines.
t-J
V1
0\
Table 88.- FELOSPAR, QUARTZ, SPUOOMENE, MICA ANO CLAY MINERALS RECOVEREO ANO SOLO FROM NORTH CAROLINA PEGMATITES IN 1977
Feld
Glass Spar Pottery Spar spatic Spudomene Quartz Sand Clay Ground Mica
'"'rj
t"'"
Trade name F-20 K-40 NC-4 K-200 Foote sand Glass Ceramic Low Glass K-30 Hallosite KM-25 P-80 Oil O
Spar Chem. Iron dry Well
Chemical Analysis, weight %
~
..,;
......
O
Si02 68.00 67.10 68.15 67.10 68.70 79.20 63 64.12 64.80 92-98 99.03 47.90 53.8 44-48 44-48 Z
Al2~ 19.00 18.30 18.88 18.30 19.27 12.10 24.7 26.50 26.25 0.5-3.0 0.23 36.91 32.10 31-38 31-38
Fe203 0.07 0.07 0.07 0.07 0.06 0.06 1.7 0.9 max 0.10 0.1-0.15 0.20 0.30 1.40 5-max 5-max ...;
CaO 1.85 0.36 1.60 0.36 tr 0.52 tr tr tr - tr na tr - - ::r'
(1l
MgO tr tr tr tr tr tr tr tr tr - tr na tr - - ..,
O
K2 0 3.75 10.10 4.50 10.10 3.69 2.62 0.50 0.27 0.14 0.5-2.0 0.23 1.05 2.20 9-max 9-max ::;<:
Na20 7.15 3.80 6.70 3.80 7.91 4.80 0.30 0.32 0.35 - 0.08 0.25 0.13 0.5-2.5 0.5-2.5
~
Li20 - - - - 0.08 0.35 5.6-6.5 7.22 6.83 - - - 0.08 - - (1l
,>
Ign loss 0.13 0.26 0.10 0.26 0.25 0.35 0.30 0.50 na wet 0.11 13.50 10.30 4-5.5 4-5.5 ():
(1l
:
Approximate Mineral weight % ....
'"
;>l
Potash spar 22 60 26 60 22 16 3 2 1 2 - 3 5 I:
e,
Soda spar 61 32 57 32 67 41 3 3 3 4 - 2 1 ]5 ]5
Lime spar 9 2 8 2 - 2 - - - - - - - O
.....
Quartz 8 5 9 6 11 41 9 5 6 90-95 +9 9 2 10 5 5 (1l
Spudomene - - - - - - 85 90 90 - - - - - - ...;
Clay Minrls - - - - - - - - - - - 90 80 10 10 (1l
....
Muscovite - - - - - - - - - - - 3 4 80 80 '"
S'
Tyler mesh ():
+20 0.1 - 0.1 - - - -
-48 42.0 - na 1.5 - - 43.2
+100 85.5 56.15 35.1 20.0 0.5 - 82.1 - 2.2 1-2 10-70
+200 97.2 90.05 75.1 50.0 90.0 100 97.5 tr 14.4 13-18 70-90
+325 - - - - - - 1.0 30.1 78 10-30
After a second thickening, scrubbing and recydoning to remove fines, the

main produet is thiekened. The -400 fines are deslimed and sized in
hydrocyelones to a 15 micron cut. The -15 goes to waste and the + 15 is sent
to conditioning. Conditioning, with 750 g/t of tall oil fatty acid (having 5 -
7% rosin acid content), plus asma!! quantity of polyglycol frother, is done at
55% solids. This pulp is then diluted and floated at 30% solids, and eleaned
and reeleaned several times.
Flotation tailings are dewatered and conditioned for iron mineral and mica
removal by flotation with petroleum sulphonate in an acid circuit, The low
iron feldspatic sand is sold to glass makers. Flotation eleaned white mica is
also produced from the previous flotation's concentrate. Typical
specifications (Redeker and Bentzen (1986)) are shown in Table 88.
Tale is found in sedimentary roeks altered by regional or eontact

metamorphism, and in ultrafine igneous rocks which have been serpentinised.
Associated minerals inelude serpentine, quartz, calcite, dolomite, magnetite,
and in sorne cases niekel aad iron sulphides.
Tale ores are ground to 65 - 200 mesh for flotation. Although in apure form
tale has natural flotation properties, its ore requires differential flotation, so
it is usual to add a frother and fuel oil to the flotation eells in staged sma!!
quantities. If only carbonate gangue is present, small additions of amine
eollectors can inerease tale recovery. If the ore contains quartz, mica, etc.,
amines are to be avoided. When nickel and iron sulphides are present, these
can be depressed with citric acid. If graphite is present, the tale can be
depressed by raising the pH to over 12, and graphite floated.
Magnetic separation can upgrade the brightness of tale, and is necessary if

nickel and chromium is a contaminant.
Vennieulite is an interesting curiosity because it at times IS upgraded by

cationic table flotation - a turn of the century process.
Wollastonite is a calcium metasilicate with a density of 2.8 - 2.9, used as a

filler in paint and tiles. To produce a white fibrous material suitable for the
market, it is upgraded by:
1.- removal of heavy minerals by wet or dry m~gnetic separation

2.- removal of caleite by anionic flotation
3.- removal of quartz and silicates by cationic flotation
Chapter 13
FlOTATION TESTING
Plannng a test
The simplest studies of the floatability of an ore involve at least 25 clearly

identifiable variables. A full-scale study of the processability of a mineral will
require consideration of at least 100 variables. General knowledge of the
interaction between these variables is sparse, therefore care in formalising
the experimental design is essential, rather than just desirable. A flotation
recovery optimisation experiment, such as one in which the only change
studied is the collector dosage, is a waste of time, because the variable
studied is not independent of other variables held constant, so the effect on
mineral recovery detected can be dangerously deceptive. The key problem in
designing all good flotation experiments is to see that the important
interactions between the variables are reliably identified.
Tests to evaluate mineral dressing schemes are expensive. The number of

independent variables involved is vasto Table 89 lists typical flotation
affecting variables such as depressant, modifier, collector, etc. In fact, each
designation hides a group, rather than an individual variable, so each item
may involve 5 to 10 controllable conditions; therefore, what looks like a
relatively short list, is in fact well over a hundred independent variables.
Handling this great number of potential variables requires careful
experimental design to reduce the number of experiments required for
unambiguous results, so careful design is essential to detect all the inter-
actions; if this is not done, the data generated can calculate out "not
statistically significant" because of the noise in the system from the random
variations in the ore sample. Thus, the importance of designing the
experiments to maximise interaction information cannot be over emphasised.
The simplest interaction between variables, frequently overlooked in testing,

is the one between the frother and the collector. One reason is that it is
practically impossible to obtain reliable scale-up of data on the effect of a
frother change from any type of laboratory testing machine, or testing
- 258-
Flotation Testing 259
procedure, The result of this, in mines with years of operation, is that the
frother and collector doses are not effectively recheclced when the ore
gradually changes as the mining operations go deeper; so today many of
these mills are using an excess of collector and frother in their roughers, and
this excess is not only costly, but results in an undetccted lowcr than
optimum recovery.
TabLe 89.-FLOTATION TESTING VARIABLES

Optimisation of Reagent Suite DeveLopment of a dressing
- uniform mineraL scheme for a new mine
mineraL dissemination in the ore number of sampLes needed to
crushing and grinding conditions cover variabiLity of ore body:
size distribution (Liberation) ore hardness, abrasiveness,
age (time in contact with air) and grinding index
puLp percent soLids naturaL pH of deposit
reagent addition point ground water content
conditioning time mining methods
fLotation machine design appLicabLe concentration met~ods:
air rate gravity, Leaching,
impeLLer veLocity fLotation, or combination
residence time of the puLp determination of the optimum
puLp temperature treating circuit
ai r f Low rate optimum Liberation sequence,
miLL aLtitude (oxygen avai LabiLit y) i.e. regrinds and Locations
co LLector type within the circuit
coLLector quantity process water choices
frother type by-product recovery possibiLities
frother quantity waste disposaL
puLp pH pLus aLL the fLotation variabLes
modifier type (Lime, soda ash, etc.) aLready Listed
modifier quantity
activator type
activator quantity
depressant type
depressant quantity
water hardness
heavy metaL cations in soLution
operator skiLL
In considering the type of work that a flotation laboratory should be capable

of handling, if it is to be an adjunct to a working mine, it is important to note
that in an existing mill the easiest to change variables, which directly affect
mill through-put, and therefore economic performance, are: the reagents; the
pulp dilution; andjor an increase or decrease of the rate of feed to the
grinding circuit, which necessarily results in coarser or finer grinding.
General1y, after a mill is in operation, the pressure on the metal1urgist is to
increase tonnage without sacrificing recovery. He therefore is normally
routinely monitoring the replacement of the existing reagent suite, or the
step, if this is incorporated in the mili circuito There is somewhat less

freee!om in changing the type of reagents used, ane! in circuit changes such as
sandjslimes separations.
Experimenta! design
In planning an experimental programme, the following five steps should be

involved:
1.- a clear definition of the objectives of the experimental programme and of

the hypothesis that will be used to analyse the results;
2.- an experimental design that will quantify the hypothesis and provide the
data to reach the stated objectives;
3.- the performance of the experimental plan and the collection of the data;
4.- the analysis of the data obtained, and
5.- the formulation of the conclusions and recommendations.
Usually these five steps are a good framework for the organisation of the
final report on the technical study.
In planning the experiment, it is good practice to write out the main points of
the mechanistic hypothesis used as the basis for the design, and all the
identifiable variables. The variables that will be tested should be specifical1y
listed, indicating which will be varied and which will be kept constant. Those
that cannot be control1ed should be clearly identified. At this stage, a priori
information on which of the variables are dependent should be noted, as well
as any possible interactions. Written lists of variables are helpful when
evaluating results during the experimental programme, as too great a
familiarity with the milI operations result in not spotting the effects of
secondary or uncontrolled and unmeasured variables. These lists are also
important to determine the confidence level of the results that will be
required to arrive at a cogent management decision, and to confirm, by
measurement, that the testing programme has met these criteria. Although
these are the ideal factors to fIX the experimental design, in practice the
experiment will have to be trimmed to conform to the limited research funds
available.
A semi-empirical method of designing a flotation experiment can be best

described using a specific example. A copper mill, after an initial period with
recoveries of around 90%, experiences a change to a more refractory ore,
and recovery drops to 83%. The metallurgist is asked to determine if the
recovery can be improved by changing the frother and collector dose, or if
new reagents are required.
As a first step, most milI metallurgists' response to such an assignment will

be to run a series of standard flotations at a constant frother dose, varying the
collector amount. If the ore sample has been sent to an out side laboratory,
on the other hand, the first step will be to determine a standard grind and
flotation time for the test series. This is done by running a series of flotation
tests to determine the ball mill grind that results in the same rougher
recovery as that obtained at the plant. This exploratory series of tests
frequently gives the service lab float technician a better feel for the ore than
the in-house metallurgist. In both cases, the tests wiIl be run for a
standardised grinding time, with the length of the batch flotation time equal
to, or some arbitrary fraction of, the residence time in the plant rougher.
RECOVERY
T I I I I 90% T I 1 I 1
frother 40 gft 88% co LLector 50gft
.
T
86% .
T .
T
.
T
J.
J.
84%
J. J.
.
T
J.
.
T
J. 82%
.
T
J.
.
T
J.
80%
.
T
J. 78%
.
T
J.
I I I I I I I I I I
10 20 30 40 50 60 10 20 30 40 50 60
co LLector gft frother gft
Fig. 68.- Results of 2 kg laboratory flotation tests, single variable,

South American porphyry
To visually analyse the results for interactions, the data can be plotted as
grams per tonne frother versus g/t collector, with the recovery as a response.
Fig. 68 illustrates this stage in testing. It is based on scattered data taken
from a series of experiments carried out at el large, privately owned copper
mine in Chile, which had operated for a number of years with a coIlector feed
of 50 - 65 g/t, and around 40 g/t of a polyglycoI ether frother. It was shown,
after more than ayear of laboratory work, that, without changing the
reagents, a considerably improved recovery could be obtained at coIlector
feed rates of under 20 g/t and a frother dose of about 60% of this value.
As can be seen, this initial data gave no indication of the results that were
eventually obtained in the long testing programme; i.e., that the optimum
recovcry at this mill is attained at coIlector and frother doses of 20 g/t of
coIlector and about 12 g/t of frother. On the contrary, the preliminary
experiment suggests that the milI is operating at close to optimum recovery,
with 60 g/t of coIlector, and that, if any change should be made, it would be
to milI test an increase in the .frother dose to 50gjt.
As we know that invariably there is an interaction between the frother and

the coIlector in this type of test, the more meaningful way to plot these two
sets of data is in terms of the recovery as the main response. This is shown as
the first graph of Fig. 69. Visual examination is stilI ambiguous, and if there
is any trend visible it would support the idea of mill testing a slight increase in
the frother addition. Instead, if a full factorial had been done initially, the
results would have been as shown in the second graph of Fig. 69 (the data
shown is taken from scattered results which do not complete the factorial).
At a cost of about 30 flotation experiments, as compared to the 12 runs of
Fig. 68, the conclusion that equal, or better recoveries could be obtained at
less than half the reagent consumption would have been indicated ayear
earlier.
Numbers shown are the rougher fLotation recovery

Data from Fig. 68 Frother FactoriaL for same range
I I I I I
gft I I I I I
.79 60 79 80
86 50 74 79 82~83
. . . . .
79 83 85 86 83
40 79/83
-: 85 86 83
)
83
79
30
20
/3Q7
85 89 84 79 75
I I I
10 ~75I I I I I
20 30 40 50 60 20 30 40 50 60
coLLector gft coLLector gft
Fig. 69.- Results of Fig. 68 replotted compared to a two-variable factorial

design covering the same range
Although this two variable factorial design defines the interactions

adequately, it does require more experiments than a more sophisticated
design. These are some general rules on the design of experiments which can
provide more efficient information:
1.- symmetry in the experimental layout simplifies calculation considerably, as

well as helping in the visualisation of the responses;
2.- the number of levels used in the experimental design fixes the shape of the
response curve that can be detected:
a) two levels detects trends only;

b) three levels can detect the existence of a response curve, but not its
shape;
c) more than three levels progressively refine the shape of the
response curve.
Examining the factorial in Fig. 69, it is obvious that the experiment can be
simplified by using a repetitive unit cell based on an experimental group of
the four corners of a square. This reduces the factorial to a two variable two
level experiment, which can only detect linear responses; adding the centre
point converts it to a three level design yielding considerably more
information. This type of design, the corners and centre of a square, is a
popular layout for two variable designs, as is the corners of a cube plus a
centre point, in three variable problems, where tendencies can be seen clearly
by visual inspection.
In Fig. 70, the square with a centre point design is applied to the example.
The initial design consists of the square with a replicated centre point. The
existing plant conditions correspond to the left upper corner of the first
square (shaded). The unit difference in frother and collector dose is chosen
to yield a change in recovery of at least twice the experimental error. The
next group is picked, after the flotation results are in hand, based on: (a) that
the next experimental group has at least one common point with the prior
.group (doubles as a check of the experimental error of the next set); (b) that
the next centre point lies on an axial or diagonal line, starting from the prior
centre point, following the overall slope of the four corner points. Thus, with
five experiments, one generates a semi-factorial 2 variable 3 level design that
is easy to analyse, relatively quick to detect a maxima in a response surface,
but ambiguous if there is more than one maxima in the surface because only
three levels are studied. It is desirable that all the points be replicated; this
is not exorbitant in cost because the ten tests required are a convenient one
shift flotation laboratory load for a float technician and aide. A very
important point in the design is that all the flotations of each unit cell must
be in a random order to avoid unconscious bias of the operator.
60 60
82 86
50 50
40 I' I,
15 83
40
1-88-7[;,;;,;;;;,;~~,I
30
83
----1-4----
88
83 30
1_I~+
79
88
5
20 20
10 10
82 80
10 20 30 40 50 60
CoLLector gft
Numbers shown are the rougher fLotation recovery.
Ordinates are gft of frother
Fig. 70.- Experimental design based on a square pattern search path
If one is designing a three variable experiment, for example the effect on

recovery of changes in the frother and collector dose in combination with
changes in the system pH, it is obvious that the above procedure can be
converted into an empirical design based on a body centred cube layout. The
minimum number of experiments involved in this case are the 8 tests at the
corner conditions plus a replicated centre, i.e. again 10 flotations which can
be done in a single shift. If it is desired to replicate all runs, the unit
experiment takes 2 operating days (or shifts). This number of tests can be
conveniently carried out in one shift, while a complete 3 x 3 factorial involves
27 flotations, plus replications, which would take nearly a week's work, and
the results from tripling the number of experiments will provide only
marginally more complete a picture of the system reaction.
Fig. 71 shows a convenient form of recording the experimental design on

paper, which can also be used to plot the recovery response. The x and y-
axis can represent the frother and collector dose, respectively, and the z-axis
the pH. The recoveries recorded are from the prior experimental series. This
is a good point to indicate that pH is not normally a continuous function of
collector and Irother dose, as it will show a recovery discontinuity at the
natural alkalinity of the are (pH between 7 and 8.5), therefore a cube which
straddles this region will give ambiguous results.
Z
?8x
85x
.. y
)(
82xV. .-- >

?8x
Data: x y z Response
Frother coL Lector pH LeveL Recovery
Run # 9ft 9ft 9ft Lime percent
1 15 35 150 82%
2 15 85 150 74%
3 15 85 350 78%
4 15 35 350 82%
5 30 60 250 79%
6 45 35 150 82%
7 45 85 150 78%
8 45 85 350 80%
9 45 35 350 85%
Fig. 71.- Cubic experiment for a truncated 3 level by 3 variable test module
As can be seen, this simple experimental design method can give a surprising
amount of information from a very few experiments, particularly when one
considers the very high error estimated for the laboratory recovery
measurements; but if, say, the effect of varying the amount of a sulphidising
agent must also be studied (i.e. if one must estend the experiment to four
variables), experimental design for visual evaluation becomes much more
difficult. The obvious choice of repeating the cubic experiment at three
different levels of sulphhydrate is not easy to interpret because of the
possible inflections in the response surface.
As mentioned in Chapter 6, Lekki and Laskowski (1971) compared batch

flotation and Hallimond tube tests by measuring the copper recovery surface
for the chalcocite-ethyl xanthate - alpha terpineol. Their results are shown
in Fig. 72. An obvious conclusion from these data is that Hallimond tube
collector data cannot be scaled up, but that experiments employing 1 to 3,000
ml laboratory batch flotation machines can be sensitive enough to detect the
existence of a maxima, and that the operating conditions and dosages yield
recovery and grade results that approximate those obtained with plant tests in
which both the collector and the frother dosage is varied over a broad range.
Laborator~ batch flotation Ha Lf Inond tube r-ecouer-u after
copper recouery contours one n Inut.e
80 _
"-
t:ll
E
~ 105. _ 60 .~
~
+'

.9
III
l.
+'
l:
:!:: III
~0; 40 g
o
(l
10 30 123
EtX addition, kg/t
numbers by the data points = recovery %.
Fig. 72.- Effect of varying frother and collector dose on copper recovery
contours for chalcocite - batch tests vs Hallimond tube
Based on this preliminary work they later (1975) published the only detailed
results of a two level, two variable flotation factorial, induding its
mathematical analysis for the interaction between potassium ethyl xanthate
and three different types of frothers: one with collector properties, alpha
terpineol; another inert, diacetone alcohol; and a frother with distinct surface
active properties, octanol. The responses measured were recovery and
rougher concentrate grade.
The results of the recovery and grade measurement for alpha terpineol and
diacetone alcohol are shown in Fig. 73, and for octanol in Fig.74. Note that
for octanol the recovery assays were not completed because the concentrate
grades obtained with this strongly surface active frother were not of
"'- !ERP lilEOL AS FROTHER

Copper r-ecover-jj , % Concentrate grade. /.
1SB
+'
"-In
"
.S1BB
~
""-e
x'"
~ SB
lX-terpineol addition .. g/t
DIACETONEALCOHOL AS FROTHER
Recovery curves, Y. Concen1:ra1:e Grade, Yo
188 188
168 168
148 148
128 128
188 188
BEl 811
68 68
48 48
28 211
4~ B~ 12~ 16~ 2~1I 248 2B~ 48 88 128 168 288
Diace1:one alcohol addi1:ion, g/l

\
Fig. 73.- Cu recovery and concentrate grade as a function of the interaction

between the frother type and the collector (EtX), at varying dosage
Table 90.- INFLUENCE OF ETl( AND a-TERPINEOL CONSUMPTION ON GRADE OF

CONCENTRATE (B), AND COPPER RECOVERY (R).
"/ = 90+ 91)( + 92Y + 93)(2 + 94 1 + 95)(Y

Second order fit \./ = B \01 = R
Re9ression coefficients 90 = 2.17236 -40.9696
x - dosis of aT, 9ft 91 = 0.041448 1.941365
y - dosis of EtX, 9ft 92 = 0.0680189 1.069910
93 = -0.000026 -0.010184
94 = 0.0000326 -0.004522
95 = -0.0006911 -0.002463
Sum of squares removed by re9ression 15.6739 4442.935
Remainder of sum of squares 4.643579 134.502
Estimate of the population variance 0.4643579 13.4502
Value of Fcal c 6.75079 66.0649
F value for 10.5 de9rees of freedom FO.01 5.64 5.64
Canonical equation u1 = -0.00035 -0.004266
Z = u1t12+ u2ti u2 = -0.00034 -0.014404
Point of maximum response fo = 6.08756 91.41737
equal to the canonical xo = 103.34 83.7665
coordinates Yo = 52.14 95.4777
r/J = 13234' 10145'
IDEM - for DIACETONE ALCOHOL

w = 90 + 91x + 92Y + 93X2 + 9 41 + g5xy
Second order fit w=B w= R
Regression coefficients go = 60.79912 37.33319
x - dosis of aT, 9ft g1 = 0.2234259 -0.30250
Y - dosis of EtX, gft g2 = 0.1104235 -0.15490
g3 = -0.0004817 0.000935
g4 = -0.0000326 -0.004522
g5 = -0.0006911 -0.002463
Sum of squares removed by regression 126.338 127.4156
Remainder of sum of squares 23.888179 30.3678
Estimate of the population variance 1.137532 1.4318
Value of Fcal c 22.21278 17.7979
F value for 10.5 degrees of freedom FO.01 4.04 4.04
Canonical equation u1 = -0.00011877 -0.00090
Z = u1t12+ u2ti u2 = -0.00062041 0.00052
Point of maximum response f o = 86.77 5.78097
equal to the canonical Xo = 222.24 148.53
coordinates Yo = 20.745 117.32
r/J = 12143' 1337'
TabLe 91.- INFLUENCE OF ETX AND OCTANOL CONSUNPTION Of\1 GRADE OF

CONCENTRATE (Il).
w = 90 + 91)( + 92Y + 93)(2 + 941 + 95)(Y

Second order fit H =B
Re9ression coefficients 90 = 10.78667
)( - dosis of aT, 9ft 91 = -0.9606
Y - dosis of EtX, 9ft 92 =
0.43835
93 =
0.0033466
94 = -0.0021295
95 = 0.00024
Sum of squares removed by re9ression 15.697
Remainder of sum of squares 11.794879
Estimate of the popuLation variance 0.655272
VaLue of FcaLc 4.791
F vaLue for 18.5 degrees of freedom FO.01 4.25
CanonicaL equation u1 = 0.003349
Z = u1t12+ u2t22 u2 = -0.002132
Point of maximum response fo =
equaL to the canonicaL Xo = 139.5435
coordinates Yo = 110.786
tP = 143'
response contours shown in Fi9. 73 and 74
commercial interest. The calculated results for the three frothers are shown
in Tables 90 and 91.
The authors note that the ellipsoidal recovery contours, with a plateau at 80 g
EtX and 80 g a-terpineol per ton, are similar to those conditions found for
the Polish ore from which the chalcocite was obtained, and that the angle of
slope of the axes of the ellipsoid indicates the high interaction between the
two reagents. They also say:
"The relation between the grade of concentrate and the

consumption of EtX and a-terpineol is shown in Fig. 72. A
singular point is situated within the experimental range and
enables to postulate the hypothesis of a different
mechanism of cooperation between' the collector and the
frother at small dosages. An increase in the dosage of both
gives a remarkable increase in the copper content of the
concentrate. Under conditions of excessive concentrations
of both reagents, however, the grade of the concentrate
decreases and the flotation begins to be less selective. It is
to be pointed out that in flotation practice the grade of
concentrate is the main quality coefficient of the process

and its decrease usually leads to the addition of new doses
of reagents."
These data and their analysis is a more elegant presentation of the

experience obtained at EI Teniente, i.e., that nearly all established mills drift
into an overdosage of flotation reagents.
Rectangle
covers
+'
experimental ~ 150
region s:
.2
:::-e
-e
.. 100
x
+'
'"
50 100 150 200

Octanol addition. g/t
Fig. 74.- The interaction of octanol and EtX on copper recovery and
concentrate grade
Lekki and Laskowski's work is important in supporting this last conclusion, as

it is the result of sophisticated academic experimental work, which is only
rarely duplicated in operating concentrators. Reagent suppliers were among
the first to recognise the importance of trying to control the shape of the
ellipsoidal recovery response surface, as a flat ellipsoidal plateau recovery
response to changes in the frother and collector additions helps stable
operation and customer satisfaction with a particular reagent suite.
Box-Wilson semi-factorial experimental design: When studying a three

variable ore flotation problem, say the effect of collector dose, frother dose
and pH on the recovery as a response, we noted above that, in a random
order and with suitable replication to estimate experimental error for all
possible combinations of C1F1Ph1, we will have to perform experiments

C 1F1Phz, C 1F1Ph3 , .... , C3F3Ph3, or 27 tests (33) . The results of these tests
can be interpreted using regression analysis and an analysis of variance. If
the number of variables or levels that must be tested are greater than three,
the number of experiments required by a full factorial become exorbitant.
For example, a two variable experiment at 5 levels will require 32 tests, while
three variables studied at 5 levels require 35 = 243 tests for a complete
factorial. For these, five-level fractional factorials, or Box-Wilson
experiments, are an effective alternative.
The simplest example of a Box-Wilson experiment is a two variable example,

where the recovery response in a flotation is to be studied as a function of
collector dose and frother dose, with other factors kept constant. To achieve
this statistically, experiments are carried out and their results reduced to a
regression function of the form:
The general model (response function) includes three types of terms III
addition to the constant b o :
1.- Linear terms in each of the variables x1,x2,... ., ~

2.- Squared terms in each of the variables x12, x22, ..., xp2
3.- Cross-product or first order interaction terms for each paired
combination x1x 2, x1x3 , ... , XP-IXP
For the three-variable problem the response function would be
y = bo + b1x1 + b2x2 + b3x3 + b11x2' + b22x1' + b33x3' +

b12x1 x2 + b13x1 x3 + b23x2x3
The number of terms for p variables is therefore (p - l)(p + 2)/2.
Limiting the interactions to first order is justified in most industrial

experiments because the experimental accuracy does not require the higher
order interactions for adequate description of the response surface.
Box-Wilson rotatable experimental design provides an easy way to calculate a

5 level system which gives as much information as a full factorial, but with a
minimum of experiments. For a two variable experiment, the coordinates of
each experiment are located at a central point surrounded by two squares
rotated 90 degrees from each other. The arrangement is illustrated in Fig.
75, where the factor k equals half the range, and the range is the difference
between the extreme experimental points, which in turn are determined as
the approximate square root of p, where p is the number of variables to be

studied. If the sides of the squares are designated by 2a, the radius of the
circle (sphere, etc.,) in which all the outer points lie is given by r = lea, and
the maximum range of each variable is 2ka. If the centre point is replicated,
this two-variable design requires 10 experiments. For a three-variable
experiment k = 1.73, and, if we assume that a = 1, the coordinates or the
combination of levels for the three-variable design will be as shown in Table
92, and Fig. 76. The total number of experiments required are 15 plus
replicates.
Level
1
i 1
2
..~
l' : k.
l'ange : l' 3
~ l' : ka
t 1 4
5
when p : nUMbel' of val'iables I then k: JP
Fig. 75.- Box-Wilson rotatable two-variable five-level design
These design principles are applicable to more than three variables (at 5
levels), as they always include three types of combinations: the axial,
factorial, and centre points. Thus axial points include each variable at its
extreme levels, with the others at their centre point level. The factorial
points include all combinations of intermediate levels (indicated by the
plus and minus ones in Table 92). A centre point is a single test at the
average level of each variable. To estimate experimental error, the centre
points are usually repeated three to five times during the experiment.
Designs for any number of variables can thus be laid out using these
principles.
When testing has been completed, a regression analysis must be carried out
to determine the coefficients of the response function. For convenience, this
is done keeping the independent variables in coded form.
After the mathematical model has been determined, an analysis of variance

should be carried out to evaluate its statistical significance. Having
determined a model which is significantly representative of the system, an
analysis can be made of the shape of the response function. The analysis can
, ,
be contour plotting, or more sophisticated non-graphical analysis such as

ridge analysis which characterises the function's behaviour in p dimensions.
I 2a
I I ....-'
-. _ ~. _ ~ ~_l(_a-+C':l} ++_/a ~
I I
+ka ~

Fig. 76.- Box-Wilson rotatable three-variable five-level design
Most physical situations can be approximated by a quadratic function over a

reasonable range of the variables. This is the reason why the Box-Wilson
experimental design methods are industrially very papillar.
TabLe 92.- THREE-VARIABLE FIVE-LEVEL BOX-WILSON DESIGN LAYOUT

AxiaL points Factori a L Points Centre point
X = x1 Y . x2 Z = x3 X = x1 Y . x2 Z = x3 X=x1 Y . x2 Z=x3
1.73 0 0 1 1 1 0 0 0
-1.73 0 0 1 1 -1
0 1.73 0 1 -1 1
0 -1.73 0 1 -1 -1
0 0 1.73 -1 1 1
0 0 -1.73 -1 -1 " -1
-1 -1 1
-1 1\ -1
Other on-line search strategies: The simplest technique - varying one variable
at a time - usually consists of holding constant all manipulative variables,
while a search is made, variable by variable, for changes that result in a "rise
27-1- FLOTATION - Theory, Reagents and Ore Testing
in the desired response. It is assumed that when the response starts to

decrease, a maximum has been passed; then that variable is held constant at
the new value, and a search started with another variable.
3El
...
x
FALSE
OPTIMUM
X2
Fig. 77.- False optimum in a one variable at a time search experiment
Mular (1976) points out, using the response contours shown in Fig. 77 as an
example, that it is impossible to avoid "false maxima" using this procedure.
He believes that the steepest ascent method is sounder, as it is a steady-state
strategy based on orthogonal designs. The procedure should be obvious from
his example in Fig. 78. Here the calculated interactions from each orthogonal
experimental group are used to layout a series of individual runs designed to
find a maxima. A new orthogonal group is then run with the centre point at
the maxima. This sequence is repeated until the true maxima is found for the
system. The procedure can be generalised to more than 3 variables. A
cautious version of the steepest ascent method is EVOP, developed by Box
(1969). A variant, shown in Fig. 79, also taken from Mular, is SSDEVOP:
Simplex Self Directing Evolutionary Operations. Spendley et al (1962)
minimises the number of runs performed, and sacrifices statistical inference,
to speed up the experimental process. The experimental design requires
n + 1 runs for 'n' manipulable variables. Mular describes the process thus:
"To start a program, levels are selected for the starting

simplex design and runs at each vertex completed. ... The
vertex associated with the lowest measured response in a
simplex is replaced by its mirror image found through the
-..
.....
Flotation Testing 27S
face opposite the vertex to be reflected. Thus, there is a

continuous movement away from the low results 111
directions most likely to produce higher results. The move
is always to an adjacent simplex. The response at the new
vertex is measured and the lowest result in the new simplex
is dropped in favor of another mirror-image vertex and so
on. The procedure is repeated until an optimum is
approached. In the region of the optimum, successive
simplexes will tend to hover about the optimum, moving
around but seldom away from it."
Mular (1976, 1989) describes these experimental designs in more detail,

including the curve fitting mathematics, and the mathematical evaluation of
the statistical significance of the results.
design
00 runs
steps
Xz
Fig. 78.- Steepest ascent experimental procedure result
276 FLOTATI01\J - Theory, Reagents and Ore Testing
)(2
Fig. 79.- SSDEVOP: climbing a response surface, from Mular (1976)
Testing procedures
The type, size and investment in a mineral testing laboratory, and the
specification of the equipment and staffing required, varies according to the
testing objectives, and differs greatly if the work to be done is:
1.- concentrator performance monitoring and optimisation,

2.- supplier technical service - reagent selection or formulation,
3.- processability verification of a new ore,
4.- industrial development of new reagents, or
5.- the development of a mineral beneficiation process for a new mine.
The standard unit operations which must be provided in a flotation

laboratory include:
sampling and sample preparation services;
crushing, grinding and grindability determination facilities;
microscopy and mineralogical evaluation facilities and skills;
a fully equipped flotation laboratory;

data processing and experimental design resources.
Services that can be obtained from independent laboratories, if promptly

available, include:
chemical analytical services;
mineral surface physical and chemical characterisation;
equipment and skills; and
chemical synthesis capabilities.
Facilities that general purpose mineral testing laboratories usually have

available are:
equipment to perform gravitational, electrostatic or magnetomotive
separation testing;
classic gold processing facilities, including amalgamation, cyanide
leaching, zinc precipitation, carbon in pulp testing, etc.;
a leach laboratory capable of simulating agitation, percolation and heap
leach processes.
If the tasks envisioned include the development of the full mineral processing
scheme for a new mine, including its process design and construction, all the
above services must be available, and they must be supplemented with
equipment to handle various thousands of tonnes of ore samples, and a
complete line of crushing, grinding, screening and processing equipment of
such a size and design as to inspire confidence in the accuracy of the scale-up
data required for the detailed design of the mine.
The only purpose-designed facility capable of quantifying from scratch the

design data for a new mineral, that the author is familiar with, is the CIMM
in Chile. This mineral research and development complex was originally
funded from Brussels in 1969 with a 15 million dollar grant. Today, after
accumulating considerable equipment purchased to solve specific client
problems, its facilities probably could not be duplicated for under 50 million
dollars.
The existing installations of the "Centro de Investigacion Minera y

Metalurgica" cover 20 or 30 acres of a hillside with three main multistory
buildings. At the highest point, accessible to large tonnage trucks, there is a
full scale multistage crushing plant capable of handling various hundreds of
tonnes of ore per day, with a flexible screening circuit for the secondary and
tertiary cone and short head crusher stages. The grinding equipment includes
ball, rod and autogenous mills. Flotation facilities range from single cell
bench and floor units to a pilot unit capable of processing about a tonne per
hour of mineral, with cell banks that can be arranged so as to include

scavengers and cleaners. There are two concrete sloped floor tanks to
receive tailings, which can be used to pilot test ores with water reclaim, to
evaluate the effect of circulating flotation reagents. Test flotation columns
are available, as well as equipment to test different types of bubblers for
flotation columns. The leach pilot facilities have percolation tanks with a
comparable hourly ore treatment capacity, while the gold section includes
amalgamation drums. Obviously the support facilities complement the
larger scale equipment, with extensive analytical labs including all the
standard testing equipment, supplemented with electron beam analysis and
scanning electron microscopes.
All types of mineralogical studies are also performed at CIMM. The

Institute is essentially self financing, as it provides routine analytical service
to the mineral industry, as well as arbitral analysis for ore purchasing
contracts, and charges for all pilot plant work. International and domestic
mining companies have used its pilot plant and design services for essentially
all new mines built in Chile in the past 15 years. In addition, contract mineral
testing work has been done for mines in most of the South American
countries. As a service to the smaller, locally owned mines, CIMM operates
satellite analytical labs located within each of Chile's mining districts, to
provide assay services. The organisation could serve as a model for other
nations with large mineral wealth who need to develop indigenous mineral
processing technology.
Equipping the flotation laboratory .l", ,-,. "uY

SJ!-'J'J'_\.i..\tJ" ,5r"-
., <,/5.. J J'-
\ vJ"-..."
.-r-
In addition to sucli general facilities as are reception, quartering and crushing
areas, plus microscopes and other mineral identification and assay
instrumentation, the minimum equipment required in a flotation testing
laboratory is:
1.- Sample preparation, weighing and storage facilities, which should include
a plastic bag heat-sealing machine and a freezer.
2.- One or more stainless steel ball mills driven by a variable speed rubber
roller set. The mill is frequently home-made and consists of a
stainless steel tube, about 6 to 12 inches in diameter by 10 to 18
inches in length, equipped with a quick opening, gasketed, end
cover. The drive mechanism should be activated by an electric
motor equipped with a reliable timer.
3.- A standard laboratory flotation machine, fitted with square glass, 01'
transparent plastic, cells capable of handling samples of 500 gm to
1000 gm or more of ore. The bench area should have power, air and
water outlets, a large diameter timer and, optionally, a distilled
water supply.
4.- Adequate vacuum resources, as concentrate and tails collected during the
test are normally dewatered in a large Buchner funnel.
5.- A good sample drying oven. In cool climates the drier can be a steam
heated metal table, otherwise a home-made vented cabinet with
slide-out shelves will do the job adequately without overheating the
tailing and concentrate samples.
Flotation Cells: many different types of laboratory flotation machines have

been used during the past thirty or forty years, but today only three or four
types are in common use. All are mechanically agitated types with variable
speed impellers. Some are sub-aeration types, with limited capacity to
control the air flow to the agitation zone, and others require an outside
source of low pressure air to generate the froth. Most laboratories purchase
commercial machines with interchangeable glass or stainless steel test cells,
sized for flotation tests of 500 gm, 1,000 gm or 2,000 gm ore samples; i.e.,
effective cell volumes are about 1.5, 3 and 6 litres, as the usual test starts with
a pulp density of around 30 to 35% solids. In many cases the machines have
interchangeable impellers so that cell scale-up geometry can be maintained
reasonably constant for different types of industrial flotation cells. In most
cases, for both sub-aeration and blown flotation machines, the airflow can be
measured with a rotameter. Generally, rotor speed is adjusted to a higher
rate than that used industrially because of the poorer agitation in the small
cells, which can result in coarser particles in the pulp settling out. Because of
the need to monitor this condition, glass and transparent plastic cells are
preferred over stainless steel cells in most laboratories.
Attempting to make the flotation test independent of the skills of the

operator, many laboratories use metal cells and equip them with mechanical
paddles, and in some cases automatic pulp level controls. There are no good
statistical studies on whether there is any benefit to mechanical removal of
the froth. Klimpel, at Dow, will recommend this arrangement because he
also is depending heavily on analysing his results on reagent performance in
terms of the rates of flotation. As can be seen from Fig. 80, the US ASTM
standards group for the coal industry also try to improve accuracy by
specifying a mechanical paddle in their standard, while the Europeans,
Figs. 81 and 82, propose a hydrodynamics modification to the cell (deflector
block) which reduces froth skimming to a Jlurely rhythmic operation, and
possibly will not affect scale up.
Experience at laboratories specialising in metal sulphide ore flotation is

firmly against the use of rotating paddles and automatic froth level control
because it very significantly reduces the amount of information obtained from
each flotation. A non-transparent (metal) flotation cell and mechanical

paddling may be justified in two-component systems - mineral and gangue.
Presumably there are only three rate curves to be measured in this case:
mineral, gangue and water; and the three are linearly related by the material
balance. In a typical copper porphyry one encounters at least 4 major target
sulphide minerals for recovery (chalcopyrite, chalcocite, another copper
sulphide, and molybdenite, silver and gold), and a surface active
multicomponent gangue. Average rate of flotation curves for the gangue and
non-gangue do not provide sufficient information for an intelligent choice of
an optimum collector-frother reagent suite, and can be shown to have led to
erroneous comparative conclusions in collector substitution studies. A good
flotation technician, by observation of the colour of the froth, can
approximate differential rate of flotation for the sulphide minerals, and can
recommend, and check at laboratory scale, whether stage-added reagents
would improve recovery, and determine the approximate point down the
rougher flotation bank where the reagent addition should occur. Computer
analysis of the rate data for individual minerals is technically feasible, but
unlikely to inspire confidence in its results because of system noise.
Instrumentation: Absolutely essential is the availability of a rugged electronic

pH meter with a probe that can be kept in the froth or dipped in when
necessary. Also essential is a 10 or 12 inch diameter laboratory second-timer
to be set up in full view of the operator. Otherwise the laboratory should be
well equipped with accurate weighing machines, and at least one analytical
balance to calibrate reagent feeders. Redox and in-line infrared or atomic
absorption equipment is a luxury that is justified if a large amount of basic
research is routinely included in the duties of the float lab. The other item
that should not be skimped on is the type and quality of screen size
measuring equipment.
Water: is a controversial item in laboratory testing. Undoubtedly the

availability of deionised water for use in preliminary flotation testing trials is
desirable, but usually, unless the community has unusually hard water, tap
water is adequate for routine work. If the client has a water problem, and
particularly if reclaimed water, containing residual flotation and flocculation
reagents, is, or will be, the main source of process water, then storage
facilities for large quantities of process water sourced from the normal mine
supply should be made available at the laboratory.
Reagent Additions: water soluble reagents, such as xanthates, should be

made up daily, and are generally fed to the float cells as relatively dilute
solutions (decomposition rates for most xanthates are about 1% per day
and the decomposition products can be depressants). Inorganic salts also are
Meehan ica I padd Ie

and drive
Cell
Line for flotation level up water inlet

Line for conditioning level
Line wetting level
Fig. 80.- Standard 5.5 litre (2 kg) flotation cell from ASTM coal standards
proposal, equipped with a mechanical paddle
~ '\
/
Rubber pac
drive rtlechanisrtl
I
:
I --
./
I
I,
I
fi-i
~
L--' h- j
I '- ~~
I
II
,-
I
I
i
~
II
1
r: -I
~-
I
:
~-.\:.;,j.--
~~
TI
I
- -
~le ....eling screws ~ rtle-ta 1 base
256--1
1 - - 225-- "156
~; "I
'., 1
deflectol'
block
25B
215
typical
connect ion fol'
constant level ~158--1
control
hood diffuser
Fig. 81.- Flotation cell proposed by ISO committee on coal testing

Note that a deflector block replaces the mechanical paddles
~~
--
~) =n
1
~:
\ZIY-
~72~
27-----J1 ~
I 14--=11=--- .....
~.
t
43
~
W Impeller
I
Diffuser
all dimension in mm
Fig. 82.- Flotation cell impeller and diffuser dimension for coal flotation
machine, Fig. 81
normally handled as dilute water solution. Lime is handled as is usual ill the
mill, as a standard suspension.
Some texts recommend that non-miscible organic collectors should be

dissolved in an alcohol, or another organic solvent, to improve the accuracy
of feed in the lab. This very defmitely is not recommended, as all the
alcohols and many of the organic solvents have frother properties, and
frother-collector interactions are known to very significantly change the
flotation characteristics of the reagents. The phenomena does not
necessarily scale up from the laboratory results. As an example of the drastic
effects of organic additives, the EI Teniente mine in Chile, for 15 years, used
a mixture of lead-containing gasoline, pfus MIBC, with their traditional
xanthogen formate collector. This operating change was made because it was
found to reduce the collector consumption in half, without affecting copper
recovery, and in addition increased molybdenum recovery in the mill. But ore
changes reverted this effect, so Teniente is back to using a considerably
reduced amount of straight diethyl xanthogen formate.
Insoluble liquid organic collectors, such as the alkyl thionocarbamates,

xanthogen forrnates or the xanthic alkyl esters, can be added to the flotation
pulp using a hypodermic or any other method which produces calibrated
drops. In labs where a number of stable oily collectors are routinely used, it
has been found convenient to manufacture custom eye droppers by fusing
and stretching small diameter Pyrex tubes. With a little practice the diameter
of the stretched tube can be adjusted so that the weight of the drop produced
by the different collectors is the same. The labelled eye droppers equipped
with standard rubber bulbs are then kept in test tube racks. The fact that the
same number of drops gives the same dosage for the different collectors
reduces experimental errors during testing.
The point of addition for each of the collectors depends on usual industrial
practice, but for most of the oily reagents half to three quarters of the dosage
is usually added to the ball mill prior to starting to grind, and/or to a
conditioning stage. The use of a conditioning stage and reagent addition at
this point also conforms .to normal mill preference, unless it is a variable
under study. It is simulated in the laboratory by initial operation of the
flotation cell with the air cut off.
Selection of a float technician should be combined with his training. In the

copper metallurgical labs that the writer is familiar with, the normal
procedure has been to obtain a large sample of a well disseminated copper
ore, train the new technician in the standard batch flotation process used in
the lab for a day or so, and then subject the candidate for the flotation test
technician job to at least a hundred duplicate 500 gm or kilo flotations. A
competent technician, with a helper, can do at least 10 complete flotation
tests per day, so the selection cum training process must budget at least a
couple of weeks' time. The candidate is then considered ready to take on the
job when his tests per day meet the laboratory norm and his last 20 or 30
flotation test tails consistently check out to within 0.015% copper between
flotations, with a reasonably constant concentrate grade. Usually if the float
man does not have the knack, it will show up after the results of the first
couple of days of testing. This tails assay figure is based on the experience of
a good flotation laboratory, and means that in recovery terms an accuracy of
better than one unit of recovery is not attainable, with a reasonable number
of tests.
The data obtained from manual testing, by a good float man, will give
considerably more information than an automatic machine, or even one only
dependent on mechanical paddling, because, for example, the way the frother
has to be added to maintain a stable froth and the time changes in froth
colour can give important clues to the rate of flotation of the minor minerals
which can be very important in evaluating scale-up.
Comparison oft' the performance oft' commercial laboratory flotation

machines , ,t , ,)
:1
A recent paper from SouthAfrica, O'Connor et al (1987), provides data on
the comparative performance of Denver, Wemco, Agitair and Leeds
Autofloat machines (with square and round cells). The comparative testing
was done with a blended gold slime residue obtained from an old dumpliear .J'
Johannesburg. The slimes2-orl't'a'ined 2.6 wt% pyrite and screened 75% -200 ~'i ..
mesh (75 micron). The standard flotation conditions and the dimensions of
the machines and cells used are sho,~~;~~~~~:cLo .
TabLe 93.- FLOTATION CELL DIMENSIONS AND STANDARD FLOTATION TEST CONDITIONS
CoL Lector 90 g/t Na mercaptobenzothiazoLe
Frother 17 g/t Dowfroth 250
Aeration rate 6 L/min (STP)
SoLids concentration 1250 rpm
Froth height 20 mm
NominaL ceLL voLume 3 Litre
Machine Leeds Agitai r
ImpeLLer diameter, mm 80.0 90.0
PeripheraL speed m/min 314 353
Square CeLLs
dimensions, mm top 200 x 140 x 170 155 x 158 x 145
bot 140 x 60 x 120 155 x 158 x 145
Working voLume, cm 3om 4,000 3,550
CeLL surface area, cm 2 280 245
Froth crowding ratio surf/voL. 0.070 0.069
Round CeLLs
Machine Denver Wemco
ImpeLLer diameter, mm 72.0 50.0
PeripheraL speed m/min 282 196
Square CeLLs
dimensions, mm top 190 x 150 x 140
bot 160 x 90 x 65
Working voLume, cm 30m 3,590 . 3,350
CeLL surface area, cm 2 285 305
Froth crowding ratio surf/voL. 0.079 \ 0.091
Round CeLLs
dimensions, mm diameter 197
Recove:ry - Ti~1e cu:rves Concent:rate G:rade - Recove:ry

fOIl diffel'ent Milchines
r,jach ine types
-- ... --- ~ 20
Aqitair n 15
--
] 10
c.
j] 5
O-l--r-~-.-----,-___,_--,c--1
WemGO 10 30 50 70
Pe:rcent Recove:ry
TiMe in seconds
Reco uery - T i",e

80.,.---------,
RecD uery - Grade
70 30....,--------------,
60 '" 25
J
::J'l50 ~ 20
~
o 40 f15
~
U
~10
Q)
"= 30
x
20
'0
-
'&
:::s
CI:l
5
0 -1--~__._-__._____,_-__.____,_____1
10 30 50 70
~ Recovery
1 00 200 300 400
T i",e in seconds
Fig. 83.- Comparison of the performance of commercial laboratory flotation

machines
The results of time-recovery profiles and machine selectivity, expressed in

terms of the grade recovery curves, are shown in Fig. 83. The authors chose
to reflect the reliability of the machines by a comparison of the standard
deviations of the individual points in a time recovery curve for the Leeds and
the Denver machine. These data are in Table 94.
The authors' conclusions from these data are:
"When the results of laboratory flotation studies are

compared, the type of cell mechanism and the shape of the
cells must be considered. The Leeds cell has been found to
yield the most reproducible results over the entire flotation
time. This cell also performed best with respect to grades

and recoveries. The use of a round cell caused slower
flotation with poorer recoveries and higher grades."
It is unfortunate that the authors did not include a Hallimond tube in the
devices studied, as by far the largest number of theories on the mechanism of
flotation have been based on data obtained with Hallimond tubes and bubble
contact angle measurements. The comparison was also only between
different makes and designs of bench scale testing equipment, and not a
verification of scale-up capability. As the purpose of a laboratory test is to
simulate an industrial installation, the machine with the "best" kinetics and
recovery in the laboratory might or might not provide the poorest scale-up
data.
The data are useful as a guide to the selection of a machine for a specific
experiment, as most flotation laboratories own various different types of
machines, not because of confusion on which are the best, but because for
different ores it is necessary to pick a machine that most closely follows the
responses of the specific full-scale unit under study. This, and the ease of
getting into a testing rhythm, is the key to the empirical preference of good
flotation technicians between different machines.
As there is an incomplete description of how the different machines were

operated, the validity of the standard deviation data in Table 94 cannot be
evaluated. It is possible that it is more an evaluation of the float man, or
men, who performed the tests, rather than of the inherent properties of the
two machines.
TabLe 94.- REPRODUCIBILITY OF TIME-RECOVERY CURVES FOR LEEDS

AUTOFLOAT AND DENVER FLOTATION MACHINES
SuLphur recovery, % (standard deviation in parenthesis) versus
Time, seconds 15 30 60 120 240 480
Denver 12.7 24.6 35.9 47.8 58.6
(1.81 ) (1.98) (2.24) (2.63) (2.54)
Leeds 59.1 74.8 86.8 91.2 92.0 92.2
<0.62) <0.65) (0.99) (1.18) (1.31 ) (1.31 )
Sampling and sample preparation

c 0 ....\
f lq ~ (e
Normally the most difficult part of doing reli~ble ore testing is to obtain a \ \v\ C I'
durable representative sample of the ore to be tested. The most readable- c;.
reliable, practical and comprehensive treatise on this topic is contained in the
\
-!!" e-!c,dc
first 71 pages of section 19 (written by HA. Belue and M.D. Hassialis) of

Taggart's Handbook of Mineral Dressing. The SME Mineral Processing
Handbook, Section 30, updates the descriptions of recommended mechanical
sampling equipment, and is a good complement to the Taggart information.
Sample Size: the quantity required to provide a representative sample of a

particular section of an ore body is nearly always underestimated by the
client, so care to communicate the procedure for obtaining the test sample,
and its minimum size, must be clearly specified by the test lab. Luckily the
statistically required sample size usually is greater than the amount needed
for the routine analysis and processability tests.
Taggart's recommendation on minimum representative sample size is

approximated by the following relationship between sample size (m) and
largest particle diameter (d):
m = kd"
where the empirical parameters, based on Taggart's tables, are defmed in

Table 95 for sample sizes in pounds, and minimum particle diameters in
inches:
TabLe 95.- PARAMETERS TO DETERMINE SAMPLE SIZE FOR DIFFERENT MINERALS

No k n Ore Type
1 30,000 2 Gotd
2 3,000 2 SiLver
3 1,000 2 Low grade uniform distribution
4 9 1.5 Base metaL, high grade
5 1,100 2.13 Low grade, highLy variabLe
6 3,500 1.8 Average grade, variabLe
These values are plotted for the numbered curves on log paper in Fig. 84,
expressed as sample sizes in kilograms, or tonnes, and particle sizes in
millimetres.
The theoretical basis for this minimum sample size definition assumes that
an infrnitely large, inherently uniform, sample source is available with only
short scale anisotropy. In practice, these conditions do not describe a mineral
body, rather they are those of a specific shipment of ore, and only if the ore is
of a uniform size and the property to be measured is the main component
assay. Even under these very ideal conditions, the minimum size sample will
only be representative if the sample gathering procedure is random and
1000
(f.)
E
...
eo
C) 100
.2
~
10
0.1 -+----t---t--t""'l-'l-t-'!""t"'!'--+-~_;_'l_I_"'!"'t"'!"1"""'-......
1 10 100
average particle size, mm
Fig. 84.- Minimum representative sample weight as a function of particle size
systematically unbiased.
To provide reliable samples for valuable ores where the error involves large
sums of money, such as in the purchase of gold ores, Taggart describes
sampling plants whose construction costs would be comparable to the
investment required for a full size crushing plant. The designs include
.elaborate continuous sampling devices located at the different size mineral
transport points, just to provide reliably representative samples for assaying.
He quotes an amusing study made in the thirties which shows that, despite
these precautions, gold assays favoured the seller or purchaser, depending on
which party did the sampling.
If it is not really possible to obtain completely reliable samples of

homogenised streams from a crusher circuit, obviously there is no practical
basis for assuming that an ore-body can be sampled in any way that could be
considered representative of the processability to be expected during a mine's
lifetime. Not even a non-technical business planner underestimates this point

when the ore is sourced from a vein deposit, or has a weathered cap, with
secondary enrichment overlaying a primary ore. But even in the case of a
giant low grade porphyry, it is easy to overlook the potential unreliability of
samples, because even large disseminated low grade ore bodies are far from
mineralogically uniform, usually having a barren central pipe and a non-
uniform radial distribution of co-products, such as gold, silver, molybdenum
and arsenic.
The pattern used to collect samples for reserve calculations of an ore body
are determined by the need for geologic information on the formation, and
are carefully evaluated three dimensionally, using arbitrary cut-off grades, to
obtain a meaningful reserve figure for investment decisions.
In the preliminary evaluation of the processability of a particular ore, a

similar approach must be adopted in obtaining and evaluating mineral
samples for processability evaluation. Thus, the selection of samples to test
for treatability should be done in cooperation with the geologic team,
because changes in mineralogy will affect the efficiency of the recovery
process and should be taken into account in planning the experimental
programme that Will be the basis for the concentrator design. Furthermore,
the mining staff should also be included at an early date in planning the
metallurgical test programmes as, particularly in large mines, it is routine
practice to blend ores from different mine fronts to optimise processability.
So certainly there should be fluid and systematic communications between
the geologists, the mining planners, and the metallurgists. Careful attention
to these details are very important during the sample planning process as drill
cores, which are generally adequate to provide reliable data for reserve
calculations, have only a very limited value in metallurgical testing, while
adits, whose costs are high, are of little value in confirming reserves, of more
value to the mine planners to give an idea of geologic integrity and extraction
problems, but essential to the metallurgical study as these are the usual
source of the larger ore samples needed to test process ability. A poorly
selected adit location can therefore result in costly design mistakes in a
concentrator, while satisfying the information needs of the geologists and
mmers.
Although the most representative assay samples from an ore body are
obtained if a very large sample is crushed to a small uniform particle size and
then quartered systematically, finely crushed material is not ideal for an
extensive flotation study, as the sample should have the minimum of prior
crushing to assure that the material tested, on anyone day, has as close as
possible the same freshly generated surface history as the ore that will be fed
to the head end of the flotation cells in the future concentrator. Therefore,
for metallurgical purposes, it is important to choose a primary sample
particle size that wil1 not bias the results because of loss of high grade fines
during quartering and transport, as well as to provide a minimum of old
mineral surfaces. To keep stable samples, service labs usual1y will keep their
ore samples in an inert atmosphere and at as Iowa temperature as possible,
particularly when a sulphide ore is involved.
At a lab that the writer was associated with which dealt with repetitive work
on optimising reagents for a particular mine, it was possible to maintain a
large ore sample unchanged, in terms of its flotation characteristics, for a
year or more. Samples, on arriving at the lab, were crushed to half inch size,
quartered into convenient equal size samples, and stored in sealed plastic
bags at freezer temperatures. At the same time a number of attempts to
obtain untarnished samples from a mine in a humid location (Bouganville),
where transit time was a number of weeks, was never successful. Nickel ores
also seemed particularly prone to alteration, and had a very limited storage
life.
O'Connor et al (1990), whose study is an excellent model of an experimental

design and evaluation, provide evidence of the durability of pyrite and
arsenopyrite in an 18 month study done on an East Transvaal gold deposit.
They point out
"The reproducibility of the flotation procedure has been

shown to be excellent. In all tests the effect of the
parameter being investigated was compared to that of a
current standard run. This led to the discovery of the
importance of aging on arsenopyrite recovery ...,
The generally improved separation efficiencies using aged
ore led to an investigation of methods to artificially age the
'fresh' ore. Initially the fresh ore was leached in order to
see whether the aging process could be reversed ..... (this)
improved pyrite and arsenopyrite recovery and hence
reduced selectivity. The leached ore was then heated to
42C in an oven for 13 days. ... this treatment did not
simulate the aging process .. (and) .. caused a significant
reduction to pyrite recovery."
Sampling working concentrators
Sampling ore for flotation testing in an existing mine is easier because of the
number of points where mineral streams have automatic sampling devices
already installed for analytical purposes. In choosing where to obtain a

flotation sample, the source should always be the coarsest available
continuous stream of dry ore. Because of the aging phenomena, the source
point chosen should also minimise the time required to obtain a composite
that is representative of a particular period of operation or a particular
region in the mining area. There usually is enough available information
from the operating sampling system to make such a rational sampling plan
possible. If the investigation required is the processability of an ore that will
come from a future mine operation, the sample selection process and the
storage precautions become the same as those required in a service lab, i.e., a
larger than necessary sample of run of mine ore, which is then reduced in
volume by the use of the quartering procedures recommended by Taggart to
obtain a representative sample of the primary sample. This material should
then be crushed to no less than plus half inch, and stored in nitrogen purged
hermetic plastic bags in a freezer. A household freezer is the cheapest, and
usually is quite adequate to keep samples for a year or more.
Sampling an operating mill for immediate flotation testing is usually easiest,

and most representative, if periodic samples are cut by an automatic sampler
from the belt feeding dry ore to the rod or ball mills or from the belt leaving
the secondary or tertiary cone crushers, as long as the circuit does not have
by-passes (fines from a grizzly) or recycles of oversize. The standard period
chosen to make a composite will depend on the frequency of fluctuation in
the ore characteristics and the objectives of the test. For very special cases a
pulp sample from the ball mill discharge may be adequate, though normally it
is to be avoided as the aging of the mineral surface under normal pulp
conditions is impossible to stop and significantly affects floatability. It is not
unusual that if the metallurgical problem to be solved requires testing pulps,
then the laboratory flotation equipment must be set up in the mill building,
close to the final ball mill or conditioner tank discharge. Freshly fractured
mineral surfaces, in contact with water, will change their characteristics in
seconds, then undergo a second, less rapid, aging, which is also
uncontrollable, so results from pulp tests must be carefully evaluated.
Because the amount of information on the mill design and operating

conditions, as well as the amount of metallurgical and mineralogical
background information on the ore body, is vastly different in each case,
these factors must be taken into account in fixing the minimum sample size
that will be representative and adequate not only to complete the planned
test programme, but also to provide a reserve duplicate sample to confirm
results if third parties, such as banks, need convincing.
Summarising, the testing objectives in a flotation laboratory generally are to

determine:
1.- the general processability of a particular ore with conventional reagents,

2.- the amenability of the ore to being processed in an existing concentrator,
and the operating conditions required at start-up,
3.- the reagents and conditions required for optimum metallurgical results
when an ore's characteristics have changed, or are expected to
change as the mining fronts advance,
4.- optimisation of an existing operation, or processing circuit, and/or the
determination of the relative merits of a set of new reagents as
compared to existing reagents in use in a specific concentrator.
To meet these objectives, representative samples of the test ores or the

various ores that will be processed in future, must be obtained, and properly
protected from alteration. The minimum sample size required will rarely be
less than 5 to 10 tonnes, if extensive flotation testing is required. In the case
of a major investment in a gold mine, where finely disseminated are is the
norm, representative samples for plant design usually range in the 200 to
1,000 tonnes of run of mine are.
If the ore to be treated has been stockpiled, sourced either from low grade
overburden or waste, or, in the case of precious metals, an alluvial gravel,
there are statistical sampling models, such as those in Taggart, or Gy, which
relate the size of the representative sample with the particle size of the
disseminated metal and the average size of the blasted are, to the size of the
sample required. A representative sample of a gravel may be orders of
magnitude smaller than if the are is in the form of 10 inch lumps.
Preliminary sample treatment and characterisation: The evaluation and assays

that samples are submitted to, prior to flotation testing, depend on the
tester's familiarity with the particular mining operation involved, and the
scope of the study to be undertaken. In all cases the sample should be
checked for size uniformity and the existence of extreme fmes. If fines are
prevalent, they should be assayed independently to check if their metal
content is substantially different in composition to the bulk of the are. If so,
the sampling method used in the mine or by the geologists should be
specifically reviewed to assure that the sample has not lost fmes and is really
representative of the are body to be tested. \
If the bulk samples received at the laboratory are wet, and this is not
representative of the deposit, they should be rejected and replaced with fresh
dry samples. If the ore-body or the waste heap is naturally wet the samples
should not be dried, as long as the test work can be carried out promptly and
the samples can be kept sealed to maintain the moisture content constant
during the testing period. Otherwise a test flotation should be done with a
representative sample which has been split into two portions, and one portion
dried by the most convenient low temperature process available. If the
flotation test indicates that dried samples do not duplicate the tests
performed on the wet samples, the only alternative that might permit delayed
testing of the wet ore is freezing quartered and sealed ore batches.
Whenever naturally wet ore is to be tested, additional replicating standards
should be incorporated in the test programme to monitor mineral surface
aging during the test period.
On receipt, a normal sample should be screened and the plus half inch
material crushed separately, usually in a laboratory jaw crusher. The under
half inch fraction can then be screened to remove the minus quarter inch and
the oversize reduced to quarter inch with crushing rolls. The whole sample
should then be homogenised and split into about 10 kg portions using
standardised equipment. A rotary splitter (such as a spinning riffle) is
convenient for this step. The 10 kg samples are then sealed in plastic bags,
and, if the storage time will be long, placed in freezers.
Unknown samples from an exploration programme or from someone else's

property should be carefully sampled and a representative common sample
submitted to a reasonably complete chemical analysis; the exact elements
that should be included will depend on the nature of the ore. In addition,
physical properties such as density, hardness, abrasiveness, grinding index
etc., should be determined. Microscopic mineralogical examination should
include a number of random samples to evaluate crystallographic uniformity
of the ore. The report should quantify the minerals identified and give an
estimation of dissemination. This is also the stage at which some idea of the
grind that will give good liberation can be estimated by judicious choice of
metallographic samples.
Otherwise the extent and the type of primary non-metallurgical data needed
will depend on the objectives of the test planned, and therefore how
elaborate an experimental design will be required to accomplish the purposes
of the test at minimum cost. For most copper properties, the head assay
should include total copper and total non-sulphide copper, as a significant
oxidised mineral content would require exploring more complex flow
schemes and retreatment loops, as well as the possibility of sulphidising prior
to flotation.
Testing routines
The procedure used in a flotation testing series must take into account the
known surface instability of metallic sulphide ores and the possible aging
effects that can occur with all minerals; thus, it must follow a very strict
sequence and rhythm. Unless the concentrator will have an unusual location,
such as at extreme altitude or latitude, the tests are all run at the ambient
temperature of the laboratory. It is recommended, though, that a pulp
temperature record be kept for each test, and duplicates be included morning
and afternoon, as a long test series may involve a systematic rise in the pulp
temperature during the day, and produce a spurious trend in the recoveries.
1.- A day's supply of ore should be ground to minus 10 mesh using

double rolls, then homogenised and weighed into accurate sample
sizes for each flotation. Usual rougher flotation standards are 500,
1,000 or 2,000 gram lots.
2.- The clean mill is opened and placed on its end, and the ball charge
added. This charge normally consists of around 10 kg of ball
bearings in half to one and half inch size. Some operators like to use
a mixed charge to mimic the wear in the full scale mill. Whether
this helps the size distribution of the ground ore to be closer to the
plant material is a matter of opinion. In any case, whatever
procedure is used, it should be adhered to strictly. This is then
followed by the ore sample and enough water to make up a 50 to
70% solids charge. At this time usually a standard amount of lime
is also added for pH control, and about half the planned collector
charge, if the collector is a non-soluble oil. The motor timer is then
set to the previously determined standard grind, which usually will
be in the 3 to 10 minute range.
3.- For the next stages the use of a prominently visible large second
timer is essential as, for reliable flotation results, the need for
extreme consistency in the time between the different operations
cannot be over-emphasised. Usually the tests are carried out by the
flotation technician and a helper. The start of the grind should have
been coordinated with the flotation test, so that when the mill has
stopped, a clean flotation cell is available. The transfer of the
ground charge to the cell is usually carried out using a coarse mesh
screen-equipped funnel of a slightl~ larger diameter than the
grinding mill, which fits over the flotation cell. The mill is turned
over onto the funnel, which retains the ball charge, and the remains
of the ore charge adhering to the mill walls and balls is washed out
with the water that will be used in the flotation test. Usually the mill
pulp densities are in the 25 to 40% solids range. In the flotation
tests the pulp density varies during the time of the test and should
usually be set somewhat higher than that used in the mill at the start
of the batch flotation.
4.- The cell is placed on the machine stand and the agitator lowered
into position. The time between the end of the grinding cycle and
the start up of the flotation machine agitator should be kept as
uniform as possible during flotation tests in each series. To be able
to do this, each step of the transfer procedure should be such that it
is carried out in the same time at the start and the end of the day
when the operators are tired, so it should be consistent but
unhurried.
5.- The exact sequence followed after the start of the agitator will
depend on the ore and the experimental design. Usually there is a
short conditioning time which is used to adjust pH and add the
frother and any other of the reagents required for the particular test
being run. The reagent addition time should be consistent and
recorded, as should the time of each later addition of reagents.
Flotation machine settings, such as agitator speed and air rates,
should be standardised and recorded. Power consumption
measurements during the course of the flotation can be of value but
usually are not essential.
6.- After the start of aeration, the consistency of the rhythm and depth
of scraping, plus the pattern in which the froth from the rear of the
cell is recovered, is the secret of accurate reproducible flotation
results. The procedure is certainly not the same for every operator,
but always must be consistent. If there is significant change in pulp
density, the air flow rate valve is adjusted periodically to standardise
the air supply with the rotameter. Depending on the rate of flotation
of the different minerals contained in the ore being tested, the
timing of the scraping can consist of a sequence such as a stroke
every 15 seconds for the first 2 to 4 minutes, followed by 30 second
intervals for a similar period, and continuing at 60 second intervals
thereafter. The pulp and froth levels are kept constant by washing
down the agitator stem and the cell sides with flotation water from a
clip-controlled rubber tube fed from an elevated tank. This
washdown also must be at fixed time intervals, say every 60 seconds.
Frother is added as needed during the whole period of the test. A
common problem in testing is over-frothing. This preferably is
foreseen and controlled by limiting the amount of frother used in
each addition, or in some cases by varying the time in which stage
addition of the collector is done.
7.- Normally the concentrate is collected for equal time periods,
separating each batch. The collection process ordinarily covers
shorter periods at the start of flotation. During all the flotation

period the colour of concentrate corning over the lip should be
noted, as this can give clues to the rate of flotation of the different
species.
8.- The length of time that a given are will be floated will depend on the
plant design and the kinetics of the system. Usually tests run from
ten to twenty minutes each, but can be over an hour if the limiting
recovery for a number of the mineral species is being studied.
9.- Generally, concentrate samples are collected on circles of filter-
paper and dewatered on a large Buchner funnel by vacuum filtration,
and the full sample kept for assay or blending to obtain a common
concentrate assay sample. The pulp remaining in the cell is also
treated in the same way. All samples are dried in a low temperature
tray oven or on a steam table, and carefully weighed. Printed data
sheets are a must, as they help keep the data collection standard
between different operators and serve to remind the operator when
certain of the data must be recovered at fixed time intervals.
With a skilled two-man team, about ten flotation tests per day can be
completed. In laying out the experimental programme it is important to keep
this sort of limitation in mind as the unit - grinding, flotation, and sample
collection - cannot be broken up.
Tests in which concentrate cleaning has to be checked, or where the effect of

primary reagent effect on by-product separation efficiency such as moly
circuits is involved, usually require starting with 2 kg ore samples and testing
the cleaning step in smaller, 100 to 200 gram test cells.
Test procedures used by a reagent supplier and a mill
The following describes the flotation procedure used at the Dow Chemical
Company when it was active in the collector market; it is summarised from a
paper by David J. Collins, presented July 1970, at the 6th Annual
Intermountain Minerals Conference, Vail, Colorado:
Ore samples received from the mines, which range in size from 75 to 500 Ibs.,
are first crushed in a 4 by 6 inch laboratory jaw crusher which discharges on
to a 12 by 30 inch low head vibratory screen equipped with a 20 mesh deck.
Oversize is then further crushed in an 8 by 5 blch roll crusher which also
discharges on to the vibrating screen. Undersize is collected in large trays
placed under the assembly. Dust control is achieved using vacuum collectors
located on top of the screen and crushers.
After crushing, the samples are mixed for 30 minutes in a 2 cu. ft. twin shell
V blender, and then riffled down to provide suitably sized charges for
splitting into 500 g samples.
Following splitting, the samples are weighed to 500 g, and stored in heat
sealable plastic bags... For those ores which oxidise rapidly, the plastic bags
can be purged with nitrogen, sealed, and then stored in a freezer. The bags
are easily opened and empty cleanly.
All samples are wet ground in a 8 by 8 inch 304 type stainless steel ball mill
using a 304 type stainless steel ball charge. The ball charge is made up as
follows:
BaLL size No. of baLLs
1.25" 14
1.00 41
0.75 99
0.50 178
totaL baLL weight = 7.4 kg
The mill is rotated at 70 rpm using standard laboratory rolls. Following

grinding, the mill is dumped into an aluminium container which has an
aluminium rod screen on the bottom. The balls are easily washed using a
spray type nozzle and, due to the lightness of the assembly, the balls can be
agitated with a minimum of effort. The slurry is caught in a funnel and
discharged directly into the flotation cell. Speed is emphasised here: the
whole operation, discharging and washing, takes from 30 to 60 seconds.
Using two identical cells, this means that the next batch of slurry is ready for
flotation immediately after the previous flotation test has been completed.
Water generally employed is deionised water which is supplied to all washing
hoses. In addition, the laboratory is equipped with water storage and
pumping facilities which can be tied into the equipment water supply system,
and enables tests to be conveniently carried out using special water solutions.
Flotation Testing: The flotation cell that Dow employed had been completely
modified to substitute for the usual square glass unit. Their cell was
constructed from a 4 1/4 qt. 304 type stainless steel beaker with the
conventional 'quiescent zone' and lip. They felt it desirable to incorporate
the beneficial properties of a circular cross-sectional cell, and thus eliminate
dead zones which would allow material to settle and cause variation in
flotation results. The full cell width lip was retained in the modification.
Stainless steel punched plate baffles were installed on the standpipe
immediately above the agitator. These reduced excessive swirling,
particularly of the froth column. The cell was also fitted with a bracket,
mounted on the side of the unit, to which was attached a 10 rpm paddle used
for froth removal, as they thought that this device would greatly eliminate
human error or bias in a flotation test. Experience in other laboratories
indicates that this unbiased operation provides less reliable and complete
data than a good operator who can visually identify the different minerals in
the ore and their relative rate of recovery.
The machine was supplied with supercharged air. The rate of flow was
monitored using a glass rotameter, and maintained at a constant rate relative
to the cell surface area of 0.0381/cm 2Imin. (or about 7.71/min.). Water for
initial make up was added via a pipe which fed through the standpipe of the
machine. Pulp levels were measured at the beginning of each test and this
level was kept the same from test to test. Impeller speed was monitored, and
usually maintained constant at 1750 rpm.
The water, from a graduated plastic squeeze bottle, was used to wash down
the sides of the cell and standpipe during flotation, and was added at the
same rate as the volume of froth removed. The volume of froth removed was
monitored using a graduated receptor pan. It was felt that maintaining a
constant pulp level, and hence a constant froth column height, greatly
enhanced experimental reproducibility.
The speed with which tests could be effected naturally depended on the
conditioning and flotation times. As an example of what could be
done, Collins indicates that, with relatively short conditioning (grind 4
minutes, condition 1 minute, float 4 minutes), two men, one handling
grinding and sample filtration, and the other running the flotation, carried
out up to 66 flotation tests in one day.
Product handling: Tailing samples were filtered in a laboratory pressure filter

which had a quick release 1/2 inch inside diameter air valve on the lid,
through which slurry was added with the filter lid secured. This prevented
spillage of material if the filter shell was accidentally jostled. The filter shell
was made of 304 type stainless steel, and was lighter than the commercial
shell, it was non-rusting, thus non-contaminating to the samples, and released
more cleanly from solids than the conventional filters. The filter was
operated at 85 psi air pressure, which in general gave rapid filtration. Those
samples which filtered slowly were flocculated, Concentrate was then
recovered with a 6 inch diameter conventional vacuum filter.
\
This test sequence has been described extensively because it conforms

generally to good operating practice in competent flotation laboratories,
though the modified cell and the use of a mechanical paddle is not
recommended except for very special studies. As noted, in batch flotation
300 FLOTATIOl..J - Theory, Reagents and Ore Testing
tests, the float technician can obtain important information on the relative
rate of flotation of the different minerals when he hand strokes the froth out
of the cell. This added information probably compensates for any improved
uniformity of results which may or may not be obtained by using
mechanically driven paddles.
At the EI Teniente laboratory, the current procedure followed with in-house

flotation tests are to grind a one kilogram sample of ore crushed to 100%
minus 10 mesh in a 7 by 8 3/4 inch ball mill (Fig. 85a and b) loaded with 127
I" ball bearings. The grind target is 25% + 100 mesh, which is met by
grinding somewhat over 10 minutes. A sub-aeration machine is used, without
a metered air feed. The ball mill charge is dumped into a standard glass cell
using a perforated screen to retain the balls. These are washed with mill
process water, and this wash is made up to give a final flotation charge with
33% solids. Flotation is carried out for a fixed time, usually less than 10
minutes. The froth is stroked off once every 7 seconds, the stroke starting at
the rear of the cell and swinging around the spindle. Make up water is added
after each stroke, from a squeeze bottle, the amount regulated by a level
marked on the impeller housing. Both the concentrate and the tails are
dewatered by vacuum filtration in a porcelain Buchner funnel mounted on an
Erlenmeyer flask. A flocculant is added to the tails to reduce filtration time.
All samples are dried in an electric tray oven.
David Collins discusses the equipment used at Dow Chemical Company, but
not how the conditions at which tests should be carried out are chosen. We
should add that the preliminary screening of the processability of an ore, and
most mill optimisation studies as well as reagent replacement tests, are done
simulating the rougher stage, without attempting to determine the final
concentrate grade attainable. Batch flotation tests to determine optimum
conditions on subsequent separations, such as cleaners, scavengers, etc., and
the separation of more complex ores, such as moly/copper concentrates, or
lead, zinc, silver operations, require so called locked cycle tests, which are
very difficult to carry out reproducibly.
~
~ 6 3/4"
-,
Detail A
Detail A
1/4"..L. ,
I ;---:--r------ ------- ------,,
21/2" , ,,
,,
L ,, I
I
, I
I
-----: 0:-
,, ,
f-; :t/B , 83/4"
I
I
-;! F
)/S"
.I
H
5 V'S" I
- -
1 1'13/4"
I, 53/4"
~
IT I
,c--
: 1
f
:1 3/4"
I~-?"=l=73/4~
I
~1
- - - - - - - - - - _____________ J,
-S-*o ---- :
/1 =-t 13/10" : .
i
Fig. 85a.- Standard 2 kg ball mill used in Chile (body)

1"
1.0"
,j.
1
5/S" 19/16-+ 15/S" 4-- 1 5/S" ~1 7/16"
1 ; - - - - 3 1 3 / 1 6 " - - - + - - 3 13/16"----tI
1<----------j-75/S"-----------tI
54" Linatex lining --+l I+-
73~
13/4"
I < - - - - j - - 4"
Fig. 85b.- Standard 2 kg ball mill used in Chile (cover and retainer)
Recommendatlons on setting standard test iCOIIHJlHioi.1!s
The procedure to be followed in running a batch flotation test with a new ore
must start out with setting up appropriate grinding time and flotation test
conditions. If the test involves an existing concentrator, the first step is setting
the standard conditions which will be used as a comparison in the evaluation
of results. These are arrived at empirically because the continuous industrial
process will be approximated by a batch laboratory test. To do this, the exact
rougher flotation conditions must be obtained from the customer. The
information to be obtained from the customer (or mill superintendent if the
lab is in-house) must include:
1.- representative head, tails and rougher concentrate grades for the
type of standard ore which will be used during the test programme,
including any mineralogical information which is available;
2.- the standard mill and grinding section operating conditions which
resulted in the average tail, head and concentrate grades reported
have to be obtained in detail. This information should include all
reagents employed, dosage and addition points, as also pulp pH
wherever it has been measured, plus any information on observed
changes in the natural pH of the ore pulp;
3.- an annotated flow sheet, showing recycle streams, which will allow
the calculation of all the solid and liquid mass balances, pulp
densities, average pulp residence times in each of the operating
stages, starting with the different stages of wet grinding through the
conditioners, rougher flotation and scavengers. This is a good
document to use to record any unusual temperature conditions;
4.- the operating data should include all pertinent mineral assays done
during normal plant performance monitoring, and screen analysis of
the feed, cyclone recirculation, rougher pulp feed, tails and rougher
concentrate if available. Also the operators should be asked for their
opinion of the rate of flotation of the different minerals in the ore as,
from the colour of the concentrate overflowing the flotation cells
along the length of the rougher, a good observant operator has some
idea of how fast the different mineral species float.
The scope and nature of the data required should be set up a priori. Ideally
the ore sample which will be used for the flotation testing should be collected
concurrently with the operating data detailed ~bove, in such a way that at the
end of the data collection period a representative, stable, uniform sample is
made available for the flotation test programme. Experience indicates that
the sample size collected should be at least twice as much as the maximum
amount that the experimental design calls for. As a tonne of ore can be
packed in about three 55 gallon drums, which can be handled in a standard

pickup truck, it is better to have an excess of sample available from the start
of the test.
If the ores to be tested are the feed to an existmg mill, the grinding
conditions for the sample can be set as a given percentage retained by a 100
mesh screen, or better, they can be set up by trial and error to match the
flotation results obtained under the mill's current operating conditions. To do
the latter, a standard test flotation time is picked that is some ratio of the mill
rougher residence time modified from previous experience with a similar ore.
The pH, conditioning time, reagent feed quantities and conditions, including
location (mill or flotation cells), and the pulp densities used by the mill, are
generally duplicated as closely as possible. Usually, though, frother addition
sequence and dosage is adjusted arbitrarily to provide an "adequate" froth.
The amount of frother used in each test is of course recorded. Based on this
standard laboratory flotation sequence, a series of ore samples are ground in
the laboratory ball mill for different times and floated. From the tailings
assays obtained, and the behaviour of the rougher concentrate grade, the
standard grinding time and the standard flotation conditions are fixed. If the
mineral recovery results of the series do not approximate the mill conditions,
the grinding media load, or size distribution, or details of flotation conditions
may need to be varied, and another standardisation series run. Usually the
rougher recovery and concentrate grade in the lab is found to' be similar to
the plant, but the mineral particle size distribution and the frother dosage
rarely coincide with the plant averages.
Using this procedure to set the grinding and laboratory flotation machine
variables, such as air rates, agitator speed, sample times, froth dipping rate,
etc., an experienced flotation technician can scale up his collector dosage and
structure conclusions quite reliably. In all cases, though, the flotation results
of changing reagents, dosages, pH, or grind must be confirmed by duplicating
the experiments on a single flotation row (in older plants with smaller cells)
or in a pilot flotation cell row fed directly from one of the concentrator ball
mills. It is very rare that the effect of laboratory frother changes made in
batch bench level flotation test will correlate with the plant. Therefore nearly
all frother optimisation is done on a pilot row; followed by a full scale plant
trial. As a consequence, laboratory collector testing is usually done in the
lab with the standard mill frother, and if one fmds that a proposed new
collector requires a frother change to operate satisfactorily, scale up to the
mill will require optimisation of the frother in a full scale mill test. Because
of the danger of recovery losses, mill superintendents will be reluctant to do
this without a very significant potential recovery improvement.
If the ore to be tested is completely new (a case that is routine for contract
concentrators which purchase ore from small miners), or when a process
development project for a new mine is involved, the grinding conditions for
the test flotation are picked to approximate the size distribution obtained or
obtainable in the full scale plant. Matching the size distribution of mill feed
exactly is impossible, as industrial ball mills are invariably operated in a
closed circuit with the oversize material from a cyclone or rake classifier
returning to the mill feed. Under some circumstances, the use of a
laboratory scale rod mill could help to change the particle size distribution of
the laboratory grind to better approximate a full scale ball mill. Under no
circumstances, though, should dry grinding in a disk pulveriser be considered,
as the fines produced by pulverisers and the iron contamination make the
resultant flotation results very erratic.
Laboratory ball mills: Collins describes the Dow Chemical ball mill as a
commercial 8 x 8 inch laboratory unit. In Chile the standard mill is about 7
inches in diameter by 8.8 inches long, much lighter than an 8 x 8 inch mill,
and easily made in a normally equipped machine shop from standard thin
wall stainless steel pipe. The use of rubber lined stainless steel for the walls
does not seem to reduce the iron contamination as compared to plant mills,
as long as the balls used are standard steel ball bearings. CODELCO's
sophistication of adding a Linatex rubber lining makes for a quieter
laboratory, but otherwise probably has a negligible effect on the tests. The
drive mechanism usually is a table equipped with two rubber covered rolls
driven by a variablespeed electric motor. These tables do double duty in the
metallurgical labs and are used in agitation leach experiments or to prepare
and homogenise reagent solution by rolling the glass bottles. The rolls can
also be used to drive ceramic grinding mills if iron contamination is one of
the variables important to the flotation.
At El Teniente the ball charge is 127 1" ball bearings, weighing approximately
10 kg. Ore is crushed to 100% minus 10 mesh and the normal are charge to
the mill is 1,000 g with 500 cc of plant process water (67.7% solids). The
laboratory grinds flotation feed to 25% + 100 mesh, which for current ores
requires about 10 minute grinds.
One reason that flotation charges of less. than 1,000 grams are not too
popular in industrial laboratories is the need to use a smaller mill so as to
approximate the product size distribution t~ that observed in the flotation
mill. Wet grinding more than one float test quantity, and dividing the pulp
into various portions, is definitely not recommended, as surface aging will
vitiate the comparison between different test conditions.
The ball charge used varies with the operator. Usually it is a mixed charge of
1 1/4 inch, down to 1/2 inch ball bearings with a total weight of 7 and 10 kg.
A typical size distribution of the balls used in a standard charge for a 7 x 8.8
inch mill is:
baLL size number weight g*
1.25" 9 1,037
1.125" 16 1,530
1.00" 19 1,310
0.75" 103 2,304
0.50" 135 980
7,161
"the void volume of this charge is approximately 870 ml.
Batch flotation tests
Field laboratory practice in batch flotation tests: the most straightforward

flotation testing procedure is that used in the quality control laboratory of an
operating mill. Generally the grind, flotation time, ph, and reagent addition
points are all set by current mill practice. The test procedure consists of
weighing out 1,000 g of a representative mill feed, crushed. to 100% -10
mesh, and making up a laboratory ball mill pulp having around 70% solids
using industrial process water from the plant. Usually the grinding time is
set to give the same plus 100 mesh as observed in the flotation plant feed, and
the lime required to adjust the flotation ph is added at the start of the grind.
The ground pulp is separated from the grinding balls using a suitable coarse
screen, and the pulp is diluted to 33% solids with mill process water. The
flotation is carried out for a time that experience indicates properly simulates
the rougher banks in the mill (frequently this is less than the theoretical pulp
residence time in the plant). The concentrate is collected as a whole, dried,
weighed, and riffled for an assay sample. The same is done with the tails.
The heads are not assayed. The data is processed to give a calculated head
grade, which can be compared to the bulk sample assay, and thus provide an
index of the accuracy of the physical data, and the calculated recoveries for
the different tests run.
A very important caveat which must be kept in mind when evaluating

competitive reagents is that mill process water will have residual quantities of
the standard mill reagents. These frequently will interact with new reagents
in an unpredictable way, improving or impairing recoveries. Also, a bulk
concentrate sample gives very little information on the effect of new reagents
on the rate of flotation.
In a more sophisticated rougher flotation test programme, the grinding time

and ball charge would be determined by a series of standard flotations, and
the grinding conditions then chosen from the conditions that most closely
matched the plant performance on tails, recovery and concentrate grade with
the standard ore. Flotation tests required to evaluate new reagents are run at
the standard grind time, and the flotation concentrate sampled over fixed
time periods, with each sample weighed and assayed independently. The
overall flotation time can be set to that available in the mill, or until the
concentrate grade from the last sample drops to a specified level below the
bulk concentrate value. This procedure provides kinetic data, as well as
flagging differential flotation of the different mineral species in the ore.
Collector consumption, and the addition sequence in the test, usually

correspond to the amounts employed in the mill. Frother addition on the
other hand does not correlate well with the point of addition or the dosage
used in the concentrator. In the laboratory tests it is added during the course
of the flotation as a function of the changes observed in the character of the
froth. Normally the test frother used is the same as that in use in the mill, or
one that experience shows to be effective with the gangue and minerals that
are predominant in the ore. The frother type must be optimised for the
laboratory test programme, despite the fact that the optimum structure or
mixture and the consumption does not scale up reliably, so that most
definitive testing when a change in frother is involved is done on a test row in
the operating mill, if one exists, or on the full mill after pilot planting. The
frother is a key to controlling the rate of flotation of the different minerals,
and therefore strongly affects concentrator grades, so that for unknown ores,
chemistry and dosage should be optimised in the laboratory testing, but
unfortunately frothers tend to act differently in the batch laboratory tests as
well as being extremely sensitive to the water analysis, so the process of
optimisation must be repeated at full scale.
Cycle tests are performed when the circuit called for in a mill includes
returning a middling or cleaner tail to the rougher. This can be imitated in
the lab by flotation of the concentrate to simulate a cleaning step, returning
the tails from this flotation step to another rougher flotation of the same ore,
and repeating the cycle until the cleaner tail assay becomes constant.
Mill testing is essential when studying the possible replacement of a frother,

as batch lab tests are normally too insensitive'to differentiate between frother
types. Generally, if the mill has large ball mills feeding independent flotation
banks, the test is run on one bank, comparing the recoveries and operation of
the test bank with a similar bank with pulp fed from the same ball mill. If at
all possible it is desirable during the test to use fresh water only in the test
section, to avoid the effect of recycled reagents from the other flotation
banks. If this is not possible, testing can become quite involved as it requires
operating enough back to back tests, lasting at least a week each, to allow for
a steady state amount of the reagent under test to return in the flotation
water recovered in the tailings thickeners. Recycled water effects can be very
significant not only on how a frother performs, but also on collector
performance, and should always be considered in the experimental design.
At this point it is appropriate to re-emphasise the shape of the recovery

response surface obtained when frother and collector dosages are varied
(Fig. 73 page 267), and to note that pulp pH also interacts with the reagents;
and that it is a common mistake to set both frother and collector dosage
higher than required, as these dosages tend never to be lowered to check if
recoveries might not actually increase with a lower reagent feed rate. Any
mill that is using more than 30 to 40 gm/t of collector should at least run a 2
or 3 variable experimental design in the laboratory to check this point. For
testing at full mill scale, an EVOP programme is recommended. For a
description of this type of experimental design, see G.E.P. Box, "Evolutionary
Operation - A method of Increasing Industrial Productivity", Applied
Statistics (1957).
Designing batch tests: the simplest experimental study routinely faced by a

copper mill metallurgist is the evaluation of the reagent dosage rates at his
plant when the miners have noted a small change in the are that is being
delivered to the mill. His ore samples are obtained from the routine
sampling points for the head grade determination. The experimental
programme is a three variable experiment in which the collector, frother and
lime dosages are varied. The most efficient and economic experimental
design is cubic, with a symmetrical change in the dosages which can be
plotted as the apices of a cube (8 conditions) plus a centre point which has
been replicated to check experimental error. This is a ten flotation
experiment, which should require one day's service from his laboratory
flotation machine.
To minimise costs, the assays required can be limited to two or three head
assays to verify the uniformity of his overall sample, and ten concentrate and
tailings assays. Elements assayed would be total copper, non-sulphide
copper, and iron. The in-house incremental cost of this series would probably
run about US$1,000, including technician wages.
If the study required is to determine the processability and the optimum

flotation circuits for a new are, one still starts with an arbitrarily fixed
grinding time, and then has to design a more comprehensive flotation
Flotation Testing 30G
sequence to try and determine the exact circuit which will give optimum
recoveries for the particular ore. This can rarely be simulated in the lab.
The flotation conditions then depend on the bias of the particular float
technician and his standard laboratory procedures. Most labs try to maintain
a reasonably constant flotation environment by stage-adding reagents while
adding fresh water to keep a constant cell level; or water containing the
standard reagents, fed from an overhead container above the flotation
machine, can be used so as to keep a standard cell level and thus
approximately maintain a constant reagent concentration in the pulp.
Most tests include collecting timed concentrate samples during the test,
rather than a single composite. Assays, and screen analysis, of these partial
samples will provide flotation rate information and an idea of the floatability
of coarse fractions.
Normally the minimum assay done is to weigh the concentrate obtained and
assay only the tails, calculating the grade of the concentrate and recovery
from these data. Better practice is to weigh and assay both the tails and
concentrate and back-calculate the head assay to check the accuracy of the
test.
From this brief description of the test procedure there are a number of
points of non-duplication of the operating plant that should be obvious
sources of possible errors. First, the ball milling cannot produce heads with a
comparable size distribution to that obtained in the mill because there is no
oversize fed back to the mill during grinding. Secondly, in the full scale mill,
the rougher flotation is continuous and conditions in anyone rougher cell are
steady state and in no way comparable to a decreasing pulp density and
overall grade that occurs in the batch flotation test. Thirdly, the environment
is not identical to that in the mill as the dissolved air in the pulp and the age
distribution of mineral particles, etc., cannot be duplicated in a batch test.
Operator-affected factors include the rate of removal of concentrate, the rate
of addition of water to maintain cell level, and the visual froth control
through manual adjustment of air flows to the cell impeller. In practice the
difference between a really good float technician and a poor one is the
intuitive uniformity of flotation conditions maintained by the good technician
just by observing changes in the froth and other operating characteristics.
Most labs have tried mechanising the concentrate removal by installing
rotating paddles, but as the level must also be controlled, most labs have
found that their good technicians can obtain more reproducible results on the
standard test than can be obtained by the mechanical paddles.
One problem that can be solved by a very simple trick is the reliable addition
of fixed amounts of collectors and frothers to the float test. Most labs use
hypodermic syringes. The operator changes needles for the different
reagents, usually choosing a thinner needle for the less viscous liquids, trying
to maintain the same volume drops. A procedure that has been found more
accurate is the use of home made droppers made by drawing capillaries
between pieces of 6 mm glass tubing over a Bunsen flame and then cutting
the capillary at a diameter for each reagent that provides drops having the
same weight for the different reagents. If the reagent is a non-volatile oil, at
least a week's tests can be handled per filling of a particular dropper. Water
soluble reagents should be prepared fresh every day. Generally xanthate
solutions are fed as a 0.5% concentration solution, and conditioning agents
such as sodium cyanide, sodium sulphate, copper sulphate, etc. are fed as 5%
solutions.
Concentrate grade or copper recovery for a test series tend to fluctuate more
erratically than the metal content of the tails, so that experienced
metallurgists believe trends signalled by changes in the tail assays, more than
changes in calculated recoveries. It is also considered a good idea to slip in
at least one blind duplicate into each series of tests with a given ore for an
error estimate, plus routinely duplicate or triplicate tests with identified
samples. The reason for blind standards is not distrust of the tester, but
rather the human impossibility to avoid all subconscious bias in flotation
testing. Because of this human trait, and the lack of reliable automatic
flotation testing equipment, one research lab has had the habit of keeping the
flotation collectors and frothers unlabelled, but experienced technicians with
a good sense of smell can usually identify most common frothers and
collectors, so the use of unidentified reagents in a flotation lab is probably a
waste of time. The steps required in testing an ore for floatability will explain
the reason for the human bias problem involved.
Standard test procedure used by Sherex on coal tests
There is an extensive (40 page) standard for the flotation of hard coal issued
by the International Organisation for Standardisation, which includes
detailed examples of how to calculate all results. A similar standard is under
study by the ASTM in the U.S.
Sherex's test procedures, detailed below, exemplify one company's approach

to testing in this field.
Flotation apparatus required:

Denver subaeration D-12 machine
2,500 ml stainless steel cell
Pulp level - 2,4D0 ml
Impeller type - receded disk
Impeller diameter - 23/4 inch (70 mm)
Impeller speed - 1,250 rpm
Peripheral speed - 901.3 ft/rninute (4.6 m/sec)
Air flow - 3.0 - 6.0 litres per minute
A.- Rougher flotation conditioning time

For setting dry samples - 5 minutes with 1,500 ml of water. Then with the
impeller stopped, add water to the desired 2,4D0 mllevel.
Before reagent addition - without air - 10 seconds (to redisperse solids).
After reagent addition - without air - 4D seconds

with air - 20 seconds
Flotation time
Three minutes, unless plant flotation time is known, in which case, plant
flotation time is used.
Paddling intensity
Three two-handed strokes every 15 seconds. Paddle during the first five
seconds and allow froth to build up during next ten.
Washing down cell sides

After the first 30 seconds (after the first three paddling intervals) start
washing down the coal from the sides of the cell and the centre well shaft,
while maintaining the original 2,4D0 ml pulp level.
Percent solids
Simulate plant conditions. If not known, use 200 grams dry weight, or 209
grams "moist weight", or 8% solids.
Reagent consumption
Use the .same ratio ofstandard reagents (frother Zcollector) as used in the
plants, and, for a better comparison, maint~in this ratio to fuel oil or
kerosene with Shur-Coal reagents. Testing other ratios can be beneficial, but
the final decision on this will be made in the plant, where the effect of the
hydrocarbon oils on the froth structure and volume can be taken into
consideration in determining the optimum ratio.
When planning the test work, add the reagents in the ratio of their specific
density. In compiling data for consumption curves, attempt to hit the same
level of reagent feed rate as used in the plant, and at least two other levels,
one of which should be twice the standard to determine overdose effects.
Visual observations
Make notes on froth structure, abundant froth volume, tenacity of froth, end
of mineralisation, etc., because these often can determine the effectiveness of
a reagent under plant conditions.
Lab flotation test sheets

The data is recorded in lab books and in lab flotation test sheets.
B.- Stage reagent addition
If the effect of reagent stage addition is to be determined, or if such a

practice is used in the plant, the following procedure is to be used:
For wetting dry samples - 5 minutes with 1,500 ml of water. Then with the
impeller stopped, add water to the desired 2,400 mllevel.
Initial conditioning time

Before reagent addition - without air - 10 seconds (to disperse solids).
Add 60% of the total reagents with a microsyringe. The reagent should be
injected below the pulp surface.
After reagent addition - without air - 40 seconds

Initial flotation time - 1 minute

Stop impeller
Bring water level back to 2,400 mllevel by washing down the cell sides and
the impeller shaft.
Turn on the impeller and condition for 10 seconds. Add remaining 40% of
the total reagents with the impeller going.
Secondary conditioning time

5 seconds without air
15 seconds with air (to be subtracted from the flotation time)
Secondary flotation time

1 minute 45 seconds (plus 15 seconds of initial froth buildup 111 secondary
conditioning step)
c.- Cleaner flotation(if applicable)

Transfer the froth product from the rougher flotation into 2,500 ml stainless
steel cell. Add water till pulp is at 2,400 mllevel with impeller-diffuser centre
shaft submerged.
Impeller speed - 1,250 rpm
Conditioning time
Before reagent addition - 10 seconds
After reagent addition - without air - 5 seconds
Flotation time
1 minute 30 seconds
Note: to better assess the desired cleaner flotation time it is recommended to

take incremental cleaner concentrates at 0 - 15 seconds, 15 - 30 seconds, and
30 - 90 seconds.
Instructions for paddling intensity and washing down all sides remam the
same as for the rougher flotation.
Additional reagents to the cleaner flotation

Tests should be run with and without additional reagents. Recommended
cleaner flotation reagent consumption - 0, 12.5%, and 25% of the total
rougher reagent dosages.
Column testing
Flotation columns: The concept of flotation columns has been around for
nearly 30 years, but it has become a hot topic since the copper mining crisis
of the early eighties forced companies to drastically reduce capital
expenditures and operating costs. A schematic of a classical flotation
column is shown in Fig. 86. The obvious operating difference with flotation
cells is the lack of an impeller, or any other agitation mechanism, generating
a saving in energy and maintenance costs.\ Turbulence and mixing is
deleterious to column operation, and is usually avoided in large diameter
columns by incorporating internal baffles to maintain plug flow. The other
major difference in the operation of a column, when compared to standard
flotation cells, is that in most ore processing applications wash water is
sprayed into the froth at the top of the column, impossible to do in a

mechanical cell as it would kill the flotation froth. The amount of wash water
added is a major factor in flotation selectivity, recovery and column operating
stability. The ore pulp is fed into the column via a distributor located at
about 2/3 of the height of the column; the tails are removed from the
bottom, concentrate overflows at the top, and air bubbles are generated at
Wash water
x x
x
I xx x x~.. concentrate
rJn r-'I------ _
X X X x I CLeaning region
Froth height z 1 meter
PuLp feed
PuLp zone height z 10 m
o0 0 0
o 0
000
o Co
o 0 0 MineraL recovery zone
o L 0
o U 0
o .. 0
o N 0
000
000
Air injection BubbLer
1~1~IO
L . TaiLs
Fig. 86.- Schematic of a flotation column
the bottom of the column by a porous sparger. The design of this sparger is
critical to efficient operation, as, with most process waters available at mine
sites, plugging of the pores of the bubble generating device is a significant
maintenance problem. The sparger is also critical in setting up a laboratory
test programme, as bubble sizes and size distribution may affect plant
operating efficiencies, so must be evaluated during a test programme.
Industrial columns have a height of 9-14 m, and a diameter of not more than
2 m, without baffling. Usually they are operated with sufficient overhead
wash water to provide a net downward flow of water, a condition known as a
positive bias. Positive bias has been the norm in column operation because
the froth layer in a column is stabilised by the wash water. The greater the
flow of water down the column, the greater the selectivity and the thicker the
froth layer. The usual froth depth in stable operation is somewhat over a
meter. A negative bias will eliminate the froth altogether, very deleterious for
a process where the concentrate is the desired product, but a condition to
consider for coal cleaning when the reagent selection results in the lower
volume product (ash) going to the tails.
A major factor that must be kept in mind in the design of dressing

experiments with columns is that the rate controlling flotation mechanism is
always bubble capture, and usually with mineral that has been pre coated with
collectors in a prior flotation.
It is customary to describe the operating conditions of flotation columns in

terms of superficial velocities (J) of the fluids, to normalise the information
for different size columns. Typical values are as follows:
Jg = gas veLocity - 0.5 to 3.0 em/second

J p = puLp feed veLocity - 0.7 to 2.0 em/second
J w = wash water veLocity 0.1 to 0.8 em/second
Jb = bias water veLocity - 0.07 to 0.3 em/second
In addition, in scale-up equations it is also customary to normalise the gas

velocity for different height columns by the pressure correction:
(Pc)(Jg *)(Ln(PT/P c)
Jg =
PT - Pc
where:
J g*= gas veLocity at standard conditions at the top of the coLumn
Pc = absoLute pressure at the top of the coLumn
PT = absoLute -pressure at the sparger
For a 10 m column PT is approximately twice PC' so that J g ~ 0.69J ". The

effect of gas velocity on recovery and grade is dominated by the bubble size,
which is affected by both the absolute amount of gas fed to the column as
well as the pore size of the sparger. In practice, bubble sizes can range from
0.2 to 3.0 mm; industrially, a target range for normal pulp mesh sizes is 0.4 -
0.8 mm. Because bubble size is a function of the porosity of the sparger, the
type and quantity of frother present, and the gas rate, it is not possible to
generalise on the effect of the gas rate on recovery, although for a specific
system there is usually a reduction in mineral carrying capacity with an
increase in gas rate.
An increase in wash water rate will increase the height of the froth layer,
increase the concentrate grade, and reduce recovery. The optimum water
rate is an important variable to determine experimentally. Other variables,

such as column height, have a logical effect on flotation; i.e., the higher the
column, the greater the selectivity. A serious mill operating problem is
detecting and maintaining constant the height of the pulp/froth interface. In
the laboratory, where glass columns are prevalent, this level can be controlled
visually. The simplest arrangement in a plant is to put flow controllers on the
feed and the tails, and to vary the wash water as a function of the interface
height, which is approximated by a differential pressure measurement at the
bottom of the column. As this control method is inherently unstable
because there is a relation between the froth height and the wash water rate,
this is not the ideal operating procedure.
Flotation column pilot plant at Chile's CIMM: Chile's Mining and

Metallurgical Research Centre (CIMM) purchased a Deister/Flotaire 8"
diameter, 24 foot tall column and a Column Flotation Company of Canada
Ltd. (CFCC) 2" diameter by 26 feet tall cleaning unit. These columns can be
operated independently or as a rougher cleaner unit as shown in Fig. 87.
The Deister/Flotaire column consists of two 8" diameter by 12' metal

sections. When assembled, feed pulp is pumped by a Galigher centrifugal
slurry pump to a pipe at the top of the 11.34 meter column which ends in a
distributor located 2.5 m below the column lip. Tails are withdrawn from the
column bottom and pumped to concrete lined sloped side pools, and the
concentrate overflows to a stirred tank, from which it can be transferred to
the cleaner column by a Moyno pump. Air is supplied at the bottom of the
column using one of the following: a) a porous diffuser capable of producing
uniform micro-bubbles, b) an external bubbler, such as the standard Deister
air aspirator system supplied with the column, or c) the Turbo Air system
patented by the US Bureau of Mines, and currently sold by ARMCO
Chile, SA.
Under normal conditions the column operation can be stabilised after about
three hours of operation, at which point representative composite samples of
the streams can be taken. Usually for each run six or seven samples of each
stream are taken at 10 minute intervals, which means each test requires
about 4 hours running time, including feed preparation. Total feed required
per run is about 4 m 3. If additional reagents are required for each run, these
are added to the feed conditioning tank.
The 2" column consists of four glass tubes, each about 6.5' long, which when
assembled form a 26' (12.3 m) flotation column with a pulp feed about 2 m
below the concentrate lip. Operation is very similar to that of the 8" column.
llollghet'
Plant Feea Final

(fJresh pulp) 8" Concentr-a te
C 2"
0 ,C
,
...
Tlk1~~m,
- - PUMP
L
U
M
Ii
Moyno
PUMP
0
L
U
I'l
N
Ov",,-
flow
_ = 0
Ii> Tailings
Moyno pUMP
Fig. 87.- Chile's Mining and Metallurgical Research Centre

column flotation pilot plant
CIMM has successfully completed a number of laboratory scale, as well as

pilot plant, studies for CODELCO and the major privately owned copper and
silver mines, including EXXON's Disputada de las Condes concentrator at
San Francisco. A number of typical cases will be described, sourced from an
unpublished report by Mr. Leonel Gutierrez, Project Manager, Metallurgy
Division, CIMM.
Flotation column testing examples
Reground concentrate cleaning: the following cleaner test programme, on a

reground rougher concentrate (RRC), was performed in 1987. The objective
of the study was to determine if column. flotation would be capable of
replacing two cleaner stages in a copper/moly recovery circuit. The RRC
sample was collected from the ball mill cy~lone discharge, during normal
operation of the concentrator, at 17% solids, 96% minus 200 mesh; assaying
21.5% total copper, 0.43% MOT and 13.9% F~, with a screen analysis of
96% < 74 micron and 32% < 9 micron. The tests were carried out in the 2"
and 8" CIMM columns. The two inch column testing is detailed below:
Target - determine the effect on the recovery and grade of the cleaner
concentrate of the following variables:
Feed rate F (L/h)

Froth bed height He (cm)
Air feed rate Q
A
(L/h)
Bias (B = T-F > 0 B (L/h)
- Perform identical standard laboratory flotation tests, if necessary, as a

comparison with the results of the column tests.
- Determine empirical mathematical model of the data to be able to
recommend an optimum circuit for the plant, and optimum
operating conditions, based on a computer simulation.
- Size the column(s) recommended as a replacement for the existing cleaner
circuits in the plant.
Sampling - about 600 l/day of pulp was collected by means of a pump. This
was sufficient to perform 3 - 4 daily tests, which were carried out for a total
of 7 days (a total of 21 complete tests). To avoid changes in the sample it
was kept in a continuously stirred feed tank during all the testing period.
TabLe 96.- COLUMN FLOTATION VARIABLES USED IN REGROUND PULP

CLEANER STUDY.
LeveL Minimum MiddLe Maximum TypicaL

Operating VariabLe (-) (0) (+) VaLues
Feed VoLume F (L/h) 43 54 56
J F (cm/sec) 0.59 0.74 0.89 0.7-2.0
TaiLings voLume T (L/h) 46 60 74

J T (cm/sec) 0.63 0.82 1.01
Bias B = T - F B (L/h) 3 6 9
J B (cm/sec) 0.04 0.08 0.12 0.07-0.30
Air VoLume QA (L/h) 74 106 138

J A (cm/sec) 1.01 1.45 1.89 0.5-3.0
Froth height He (cm) 17 30 60

Assays - were done using a fast X-ray fluorescence equipment to allow

changing the experimental design on the run, and the final assays were
duplicated at the in-house laboratory using atomic absorption. All samples
were assayed for CUT, MOT' Fer, and insolubles. In addition, selective assays
were made on given screen fractions. The full experiment was repeated
during a period where the ore had changed.
Experimental design - the range of values of the variables chosen for this
study are shown in Table 96.
Later, a centre point was selected as a reference point, shifting the other
operating conditions to the extreme values, as shown in Table 97. These
decisions were taken on site as the assay results were obtained, and taking
into account how stable the column operation was at the previous conditions,
and evaluating visually the froth quality.
Supplementary laboratory batch flotation tests - batch flotation tests,

measuring the rate of flotation, were performed to simulate the cleaner,
scavenger and/or rougher flotations in the plant, assuming that the column
tails were sent back into the existing plant circuit, using the tails generated in
RRC - 14 and RRC - 17. These results were included in the concurrent
computer simulation made of the overall plant operations.
Review of the experimental results - the best results were obtained from tests
RRC - 09, RRC - 10, RRC - 11, and RRC - 12, with the following results:
Concentrate grade, % Average recovery, %

TotaL copper 39.5% 97.4%
TotaL moLy 0.52 84.4
InsoLubLes 5.6
These results were equivalent to those obtained in the plant with two cleaner
stages, with conventional recleaner sections. It would result in a significant
circuit simplification if an industrial size column were installed at the plant.
It is interesting to note that the best copper recoveries were obtained for the
-74 micron/ + 18 micron size range, and the best molybdenite recoveries
occurred in the -53 micron/ +34 micron fraction.
\
As the pulp characteristics varied somewhat, during the testing procedure, a

formal statistical analysis was made of the following variables' influence on
recovery:
Xl = % by weight solids in the pulp feed to the column

Xz = % copper grade (or molybdenum) in the feed
x3 = apparent average residence time of the pulp in the column;
minutes = 1.216(792.5-He) I (B + F)
~ = apparent average residence time of air in the column;
minutes = VA/Q A = 1.216 (HelGA)
While the dependent variables (Y) were:
% copper (or molybdenum) grade in the concentrate,

% recovery of copper (or molybdenum) in the concentrate
Multiple lineal polynomial regressions for each experimental condition

yielded r2 > 0.95 for most of the equations, so level curves of copper and moly
recovery, versus copper or moly grade in the feed, were obtained for the
different experimental ranges studied. These curves yielded the optimum
operating condition for the 2" column shown, in Table 98.
Computer simulation and final design of the proposed flotation circuit: Column
flotation can replace the two cleaner circuits used in the plant to upgrade the
rougher reground concentrate with one full size column, as demonstrated by
the fact that the test 2" column produced a concentrate grade of 39.3% Cu
and 0.69% Mo, as compared to an average of 35.3% Cu and 0.5% Mo in the
plant operations.
TabLe 98.- OPTIMUM FLOTATION CONDITIONS FOR THE 2" COLUMN WITH THE RRC SAMPLES
1.-Average specifications of the test puLp feed
Weight % soLids = 17.5; Copper grade = 21.5%;
MoLy grade = 0.43%; Screen size = 96% -200 # TyLer
2.-0ptimum coLumn operating conditions:
F = PuLp feed rate = 54 - 55 L/h
T = TaiLings voLume = 60 - 62 L/h
B = Bias = T-F 6 L/h
He= Froth height = 15 - 20 L/h
Tp= Average puLp residence time = 15.7 minutes
3.-Performance projected
Concentrate Grade, % Recovery, %
Copper 38.5 - 40.0% 97.0 - 98.0%
MoLybdenum 0.6 - 0.7 84.0 - 85.0
Based on these results CIMM proposed the flowsheet shown in Fig. 88.
Computer simulation of this circuit indicates that it is feasible to obtain a
final copper/moly concentrate with 38.6% Cu and 0.7% Mo, with overall
recoveries of 87.6% on Cu and 61.3% on moly. This performance would be
very similar to that obtained in the current plant. They recommended that,
prior to implementing their design, a further test be performed on-site with a
semi-commercial size column.
\
Plant Feed.
Wash water
Dilu. tion we. ter
Concentl!"ato~
tails
Reg~ound rougher
Concentrate
l ...
Final
Concentra te
Tails froM lotation colUMn
Fig. 88.- CIMM recommended flowsheet for cleaner column flotation
Design of the industrial column: Based on the experimental results, the key
design data used in the calculation of the recommended column flotation
flowsheet were:
Tp = 15.7 minutes = apparent average puLp residence time in 2" coLumn

Fe(T) = 1.5 = scaLe-up factor for Tp, then
(Tp)industriaL = 23.6 minutes
= RRC fed to cLeaners in the pLant = 9,788 t/day.

= puLp feed in the pLant = 2033 m3/h = current feed to the first
cleaner stage.
T = pLant taiLings voLume from the cLeaners = 2,251 m3/h of puLp
= (TP)industriaLx T/60 = 885.4 m3 of puLp = actuaL puLp voLume of
the fLotation coLumn
= voLume of a standard industriaL 12 m high coLumn, with a cross
section of 72"x72" = 40.1 m3
= design height of froth = 1 m
= effective puLp voLume of the above industriaL coLumn = 36.8 m3
= air consumption scaLe-up factor = 0.048 t conc./m3 of air
= air superficiaL veLocity in the coLumn = 1.44 em/second
Number of coLumns required = Vep/V pc = 88.4/36.8 = 24 coLumns
= Split-factor or concentration factor for plant, calculated from

the enr i chment of the final concentrate as compared to the f eed
to the cleaners in the experimental programme = 0.5148
= tonnage of final concentrate = 0.5148 x Fs = 5,039 t/day

= 5039/Fe(QA) = 5039/0.048 = 4,374 013 of air/hour
air per column = 4,374/24 = 183 m3 air/hour/column
Based on the superficial air velocity, JA' the calculated value

is 173 m 3 /h/column.
Therefore the recommended design for the final concentrate cleaners to

produce 1,800,000 t of concentrate per year requires 24 columns, 12 m in
height, and a square 1.83 m x 1.83 m cross-section. Each column will
operate at:
F = 85 m 3/h pulp feed rate

T = 94 m 3/h tailings discharge rate
B =9 m3/h bias flow of water
QAC = 183 m
3/h air feed rate
and these rates will result in a froth bed 1 m high, and a pulp residence time,
~ = 24 minutes.
\
BiBUOGRAPHV
Abramov, A.A. (1969) Role of sulphidising agent in flotation of oxidised

minerals, Tsvetn. Metall. 12(5), 7-13.
Agar, G.E. (1985) The optimisation of flotation circuit design from
laboratory rate data,XVth Int. Min. Proc. Congress, Cannes, June 2-
9,1985, vol. II, pp.l00-111.
Alferova, L.A., and Titova, GA. (1969) Study of the reaction rates and
mechanisms of oxidation of hydrogen sulphide, sodium
hydrosulphide and sulphides of sodium, iron and copper in aqueous
solutions by atmospheric oxygen, (in Russian), Zh. Prikl. Khim. 42,
192-196.
Arbiter, N., Young & Baker, (1951) U.S. Patent N 2,559,104.
Arbiter, N., Young & Baker, (1952) N 2,608,298.
Arbiter, N., Young & Baker, (1953) N 2,664,199.
Baleshta, T.M., and Dibbs, H.P. (1970), An introduction to the theory,
measurement and application of semiconductor transport properties
of minerals, Mines branch technical bulletin TB 106, Canada.
Ball, B., and Rickard, R.S. (1976) The chemistry of pyrite flotation and
depression, M.C.Fuerstenau, Ed. Flotation, AIME, pp.458-484
Beas-Bustos, E., and Crozier, R.D. (1991) Moly/copper separation from
concentrate of the combined acid and basic circuits at EI
Teniente, Reagents in Minerals Engineering, Camborne School of
Mines, September 18-20.
Bickerman, J. (1953) Foams: theory and industrial applications, Reinhold,
New York.
Bogdanov, O.S., Emlyanov, M.F., Maximov, I.I., and Otrozhdennova. L.A.
(1980) Influence of some factors on fine particle flotation, P.
Somasundaran Ed. Fine Particle Processing, voLl , Proc. of Int.
Symp. AIME, Las Vegas February 24-28,1980, pp706-720, (impeller
speed)
BoIth, F., Crozier, R.D., and Strow, LA. (1975) Dialkyl thionocarbamate
catalytic method, US Patent No. 3,907,854.
- 324-
Bibliography 325
Booth, R.B., & Freyberger, W.L. (1962) Froth and frothing agents, Froth
flotation 50th anniversary volume, D.W. Fuerstenau, ed., SME of
AIME, New York, pp. 258-276.
Booth, R.B. (1954) U.S. Patent 2,675,1Ol.
Born, CA., Bender, F.N., and Kiehn, O.A. (1976) Molybdenite flotation
reagent development at Climax, CO, Flotation, Ed. M.e.
Fuerstenau, SME-AIME, Littleton, CO, pp.1147-1184.
Box, G.E.P., and Draper, N.R. (1969) Evolutionary Operation, John Wiley &
Son, New York.
Bulatovic, S.M., and Salter, R.S. (1989) High intensity conditioning, a new
approach to improving flotation of mineral slimes, Processing of
complex ores, Proc. Int. Symp, CIMM, Halifax, August 20-24, 1989,
pp169-181, Pergamon Press, New York.
Cambron, Adrien, and Whitby, G.S. (1930) The oxidation of xanthates and
some new dialkyl sulphur- and disulphur-dicarbothionates, Can. J.
Research,]"I44-52.
Cambron, Adrien (1930) The mechanism of formation of thiuram and
xanthogen monosulphides, and observations on thiocarbamyl
thiocyanates, Canadian J. Research,2" 341
Castleman Jr., A. W., Yang, X, and Shi, Z. (1991) J. Chem. Phys. Vol. 94, p.
3268.
Castleman Jr., A. W., and Yang, X. (1989) J. Am. Chern. Soc. Vol. 111, p.
6845.
Castro, S., Goldfarb, .I., and Laskowski, J. (1974) Sulphidizing reactions in
the flotation of oxidized copper minerals, Part I, Int. J. Miner.
Process. 1(2), 141-149; Castro, S., Soto, H., Goldfarb, .I., and
Laskowski, J.(1974), Part II, Int. J. Miner. Process. 1(2), 151-161.
Castro, S., Gaytan, H., and Goldfarb, J. (1976) The stabilizing effect of NazS
on the collector coating of chrysocolla, Int. J. Miner. Process. ](1),
71-82.
Castro, S., and Pavez, O. (1977) Mechanisms of action of Anamol-D, Nokes
reagent and sodium hydrosulphide in MoS z flotation, Avances en
flotacion, yoU, pp. 111 -130, Univ. of Concepcion, Third Conference
on surface chemistry and flotation, 21 to 24 June, 1977, Concepcion,
Chile, (in Spanish).
Cea, Zocimo, and Castro, S. (1975) Depressant effect of the calcium ion on
pyrite flotation, Avances en flotacion, Vol.l, p.97-105 (in Spanish).
Chander, S.(1988) Electrochemistry of sulfide mineral flotation, Min. & Met.
Processing, p 104, August 1988. \
Chander, S., and Fuerstenau, D.W. (1972) On the natural floatability of
molybdenite,AIME-SME Trans., 252, pp. 62-69.
Chanturiya, V A., and Vigdergauz, V.E. (1989) Electrochemical modification
of sulphide flotation, Processing of complex ores, Proc. Int. Symp,
ClMM, Halifax, August 20-24,1989, pp.71-85, Pergamon Press, New

York.
Chudacek, M.W. (1990) A new quantitative test-tube flotability test, Minerals
Engineering, Vol.], pp. 461-472.
Chudacek, M.W., and Fichera, M.A. (1991) The relationship between the
test-tube flotability test and batch cell flotation, Minerals
Engineering, Vol...1-, pp. 25-35.
Clark, Alfred, (1974), The chemisorptive bond, Academic Press, New York,
NY.
Clingan, B.V. and McGregor, D.R (1987) Column flotation experience at
Magma Copper Co., with related experience of other mineral
processors, SME annual meeting, Denver, CO, Feb. 24-27, Preprint
No. 87-91.
Crozier, RD. (1957) Froth stratification and liquid mixing in a bubble tray
column, Ph.D. Dissert. Univ. of Michigan, Ann Arbor, Mich. 1956;
University Microfilms, Ann Arbor, Mich.
Crozier, RD. (1977) Reactivos de flotacion en recuperacion de minerales
sulfurados, Avances en flotacion,], Univ. Concepcion, Chile, pp.
301-344, Summarized in: Processing of Copper Sulphide Ores: Froth
Flotation reagents - A Review, Min. Mag. London, April, 1978.
Crozier, RD. (1978) Reactivos de flotacion aplicados a la recuperacion de
molibdenita con minerales de cobre porfiricos, Primer Congreso de
Ingenieros del Cobre, Chuquicamata, Chile, 4-8, December, 1978.
Summarized in: Flotation Reagent Practice in Primary and By-
Product Molybdenum Recovery, Mining Mag. London, February
1979.
Crozier, RD. (1979) La funcion del espumante en la flotacion de minerales
sulfurados, Avances en flotacion.A, Univ. Concepcion, Chile, pp.19
56, Summarized in: Frother function in sulphide flotation, Mining
Magazine, January 1980, pp.26 -36.
Crozier, RD. (1984) Changing patterns in the supply of flotation reagents,
Mining Magazine, London, June, 1984.
Crozier, R.D., (1991) Sulphide collector mineral bonding and the mechanism
of flotation, Minerals Engineering, vol 4, Nos 7-11, pp.839-858
Crozier, RD. and Klimpel, R (1989) Frothers: plant practice, Mineral
processing and extractive metallurgy review,j, pp 257-279, Gordon
and Breach Science Publishers, London.
Crozier, RD., and Strow, L.A. (1976) Process for the flotation of copper
sulfide ores with improved thionocarbamates, U.S. Patent No.
3,975,264.
Cyanamid, (1989) Mining chemicals handbook, Mineral dressing notes No.
26-1, American Cyanamid, Wayne, NJ,
Bibliography 327
Davis, Farlow e. (1976) San Manuel's process for molybdenite recovery,

Local Chapter Meeting, AIME, San Manuel AZ., October, 1976.
Dobby, G.S., and Rao, S.R., Editors (1989) Processing Comple.x Ores, Proc.
Int. Symp. on processing of complex ores, Halifax, Aug. 20-24, 1989,
Pergamon Press, New York
Douglass, W. (1927) US Patent No. 1,625,099.
Douglass, W. (1928) US Patent No. 1,659,396.
Down, R.F., and Turner, J. (1970) Concentration of oxide ores at Tynagh,
Metallurgy of lead and zinc, Eds. Rausch, D.O., and Mariacher, B.e.,
Vol. 1, pp.710-731, TMS of AIME, New York, NY.
Drew, M.G.B. (1977) Seven-coordination chemistry, Progress in inorganic
chemistry, Vol. 23, John Wiley & Sons, New York, pp. 67-210.
Drzmala, J., and Lekki, J. (1988) Application of flotometry for characterising
flotation in the presence of particle aggregation, Minerals
Engineering, Vol. 1, pp.327-336.
Dudenkov, S.V., and Abramov, AA., Discussion of Lekki & Laskowski
(1975).
Espinosa-Gomez, R., Finch, JA., and Jonson, N.W. (1988) Column flotation
of very fine particles, Minerals Engineering, Vol,j., 3-18
Ewers, W.E., and Sutherland, KL. (1952) The role of surface transport in
the stability and breakdown of foams, Austr. J. Sci. Res. A5, pp.697.
Filyk, A. (1976) A review of plant practice at Molycorp Inc., Questa Division,
Flotation, Ed. M.e. Fuerstenau, SME-AIME, Littleton, CO,
pp.1l85-1l99.
Finkelstein, N.P. and Allison, SA. (1976) The chemistry of activation and
depression in the flotation of zinc sulfide: A review, Flotation, Ed.
M.e. Fuerstenau, SME-AIME, Littleton, CO, pp. 414-457.
Finkelstein, N.P., Allison, SA., Lovell, V.M., and Stewart, B.V. (1975)
Natural and induced hydrophobicity in sulfide mineral systems,
/ advances in interfacial phenomena of particulate-solution-gas
systems; applications to flotation research, Somasundaran, P., and
Grieves, RB., ed.,AICHE, CEP Symposium Series No. 150, Vol. 71,
pp.165-175, AICHE, New York.
Finkelstein, N.P., & Poling, G.W. (1977) The role of dithiolates in the
flotation of sulphide minerals, Mineral Sci. Engng.,2,pp.177-97.
Fischer, AH. (1928) US Patent No. 1,684,536
Fleming, M.G. (1953) Effects of soluble sulphide in the flotation of lead
minerals, in Recent Developments in Mineral Dressing, Proc. First
Int. Mineral Proc. Congress, Londdn, 1952, IMM, London. pp. 521-
528.
Forssberg, K.S.E., and Hallin, M.L (1989) Process water recirculation in a

lead-zinc plant and other sulphide flotation plants, Challenges in
mineral processing, Eds., K.V.S. Sastry and I\t:LC. Fuerstenau, SME,
Littleton, CO, pp.452-466.
Fuerstenau, D.W. Editor (1962) Froth Flotation. 50th anniversary volume,
SME of AIME, New York.
Fuerstenau, M.C., Editor (1976) Flotation, AM.Gaudin memorial volume,
Vol. 1 and 2, SME of AIME, New York, NY.
Gardner, J.R., and Woods, R. (1973) The use of a particulate bed electrode
for the electrochemical investigation of metal and sulphide flotation,
Australian Journal of Chemistry, vol 26, pp. 1635-1644,
Gardner, J.R, and Woods, R (1974) An electrochemical investigation of
contact angle and of flotation in the presence of alkylxanthates, I.
Platinum and Gold surfaces, Australian Journal of Chemistry, vol 27,
pp.2139-2148.
Gardner, J.R., and Woods, R. (1977) An electrochemical investigation of
contact angle and of flotation in the presence of alkylxanthates. II.
Pyrite and Galena surfaces, Australian Journal of Chemistry, vol 30,
pp.981-
Gaudin, AM. (1939) Principles of Mineral Dressing, McGraw-Hill Book Co.
New York and London.
Gaudin, AM. (1957) Flotation, Second Edition, McGraw-Hill Book Co.,
New York.
Glembotskii, VA., Klassen, V.I., & Plaksin, LN. (1963) Flotation, Engl.
transl., Primary Sources, New York.
Glembotskii, VA., Klassen, V.I., & Plaksin, LN. (1972) Flotation, Engl.
transl., Primary Sources, New York.
Glembotskii, VA., et al (1963) Quoting Concentration of non-metallic
minerals by flotation, Izd. AN SSSR, 1952.
Grainger-Allen, TJ.N. (1970) Bubble generation in froth flotation
machines, Trans. I.M.M.:J2, pp. CI5-C22.
Granville, A, Finkelstein, N.P., and Allison, SA. (1972) Review of reactions
in the flotation system galena-xanthate-oxygen, Trans. I.M.M., vol.
81, pp. CI-C30.
Gray, H.B. (1973) Chemical bonds: an introduction to atomic and molecular
structure, WA.Benjamin, Inc., Menlo Park, CA.
Groppo, J.G., and Yoon, RH. (1982) Selective flotation of sulfide ores using
alkyl pyridinium collectors, Yarar,B. and Spottiswood, J.A Editors,
(1982), Interfacial phenomena in mineral processing. Proceedings of
the Engineering Foundation Conference, Rindge, New Hampshire,
Aug. 2-7, 1981. Engineering Foundation, New York, pp. 271-285.
Bibliography 329
Gutierrez, G., & Sanhueza, J. (1977) Efecto de carbon activado en la

selectividad del proceso de separacion molibdenita-cobre, Avances
en flotacion,], Univ. de Concepcion, Chile, pp.223-236.
Hansen, R.D., and Klimpel, RR (1986) The influence of frothers on
particle size and selectivity in coal/sulfide mineral flotation, Trans.
AIME, 280, pp. 1804-1811.
Hansen, R.D., Klimpel, R.R., and Bergman, R. (1986) U.S. Patent 4,582,496.
Harris, e.e. (1976) Flotation Machines, in Flotation, A.M. Gaudin memorial
volume, Vol. 2, M.e.Fuerstenau, ed., AIME, New York, pp. 753-
815.
Harris, Guy, (1954) Alkyl thionocarbamate collectors, US Patent No.
2,691,635.
Hayes, RA, Price, D.M., Ralston, J., and Smith, RW. (1987) Collectorless
flotation of sulphide minerals, Mineral processing and extractive
metallurgy review, voL]., pp.I-20.
Hayes, RA, and Ralston, J. (1988) Collectorless flotation and separation of
sulphide minerals by Eh control, Int. J. Min. Processing, vol. 45,
pp.55-84.
Heyes, G.W., & Trahar, W.J. (1953) The natural floatability of
chalcopyrite, Int. J. Mineral Dressing (London, Inst. Min. Metall.).
Heyes, G.W., & Trahar, W.J. (1977) The natural floatability of
chalcopyrite,Int. J. Min. Proc., vol..1-, pp.317-344.
Heyes, G.W., & Trahar, W.J. (1979) Oxidation-reduction effects in the
flotation of chalcocite and cuprite. Int. J Min. Proc., vol..1-, pp.317-
344.
Hoover, R.M., and Malhotra, D. (1976) Emulsion flotation of molybdenite,
M.e.Fuerstenau, Ed. Flotation, AIME, pp.485-505
Hosten, <;., and Tezcan, A. (1990) The influence of frother type on the
flotation kinetics of a massive copper sulphide ore, Minerals
Engineering, Vol.], pp. 637-640.
Houben-Weyl, (1973) Methoden zur organischen chemie, 12/2, 685 George
Thieme Verlag.
Huggins,DA., Wesley, RJ. and Jomoto, K. (1987) Column flotation, its
status and potential for Escondida, SME annual meeting, Denver,
CO, Feb. 24-27, Preprint No. 87-116.
Iwasaki.L, Malicsi, AS., Li Xiaoxei, and Weiblen, P.W. (1989) Insights into
beneficiation losses of platinum group metals from gabbroic rocks,
Challenges in mineral processing, Eds., K.V.S. Sastry and M.e.
Fuerstenau, SME, Littleton, CO, pp.437-451.
Jones, M.H., and Woodcock, J.T. (1979) Control of laboratory sulphidisation
with a sulphide ion-selective electrode before flotation of oxidized
lead-zinc-silver dump material, Int. J. Miner. Process.ti; 17-30.
Jones, M.H., and Woodcock, J.T. (1979) Perxanthates - a new factor in the
theory and practice of flotation, Int. J. Miner. Process.i S, pp.285-296.
Jones, M.J., and Oblatt, R, Editors (1984) Reagents in the mineral industry,
Proc. Symp. Rome, Sept. 18-21, 1984, IMM, London.
Kapur, P.e., and Mehrotra, S.P. (1989) Modeling of flotation kinetics and
design of optimum flotation circuits, Challenges in mineral
processing, Eds., K.V.S. Sastry and M.e. Fuerstenau, SME, Littleton,
CO, pp.300-322.
Kepert, D.L. (1977) Aspects of the stereochemistry of Six-coordination,
Progress in Inorganic Chemistry, Vol. 23, John Wiley & Sons, New
York, pp.1-66.
King, RP. Editor (1982) Principles of Flotation, South African Institute of
Mining and Metallurgy, Monograph Series No.3, Johannesburg.
King, RP. (1976) The use of simulation in the design and modification of
flotation plants, Flotation, Ed. M.e. Fuerstenau, SME-AIME,
Littleton, CO, pp. 937-962.
Klassen, V.I., and Mokrousov, VA. (1963) An Introduction to the Theory of
Flotation, Engl. transl. by J. Leja and G.W. Poling, Butterworths,
London.
KIimpel, Richard (1977) An engineering analysis of the effects of chemical
reagents in the laboratory evaluation of mineral flotation,
Avances en flotacion,], Univ, de Concepcion, Chile.
Klimpel, R.R. (1987) The industrial practice of sulfide mineral collectors
- Chapter 21, Reagents in the mineral industry, P. Somasundaran and
B. Moudgil, eds., Marcel Dekker, New York, pp. 663-682.
KIimpel, RR (1987) The use of mathematical modeling to analyze
collector dosage effects in froth flotation, Symp. on
Mathematical Modeling of Metals Processing Operations, Palm
Springs, Calif., The Metallurgical Society.
Klimpel, R.R., and Hansen, RD. (1987) Chemistry of fine coal flotation -
Chapter 4, Fine Coal Processing, S. Mishra and R Klimpel, eds.,
Noyes Publishing, Park Ridge, New Jersey, 78-106.
KIimpel, RR, and Hansen, RD. (1987) Frothers - Chapter 12, Reagents in
the mineral industry, P. Somasundaran and B. Moudgil, eds., Marcel
Dekker, New York, pp. 385-409.
Lager, T., and Forssberg, F.S.E. (1989) Current processing technology for
Antimony bearing ores, A review, part 2, Minerals Engineering, vol.
1, pp. 543-556.
Lager, T., and Forssberg, F.S.E. (1989) Beneficiation characteristics of
Antimony minerals, A review, part 1, Minerals Engineering, volJ, pp.
321-336.
Laskowski, J. (1974) Particle-bubble attachment in flotation, Minerals Sci.
Engng, vol.Q, no.4, pp. 223-235.
Bibliography 331
Laskowski, Janusz (1974) Fundamentos Fisico Quimicos de la Mineralurgia,

Translated to Spanish from the original Polish by Mariano Rawicz
and reviewed and up-dated by the author. University of Concepcion,
Chile, p. 503.
Laskowski, J. (1982) Redox conditions in flotation: oxides, Yarar.B. and
Spottiswood, JA. Editors, (1982), Interfacial phenomena in mineral
processing. Proceedings of the Engineering Foundation Conference,
Rindge, New Hampshire, Aug. 2-7, 1981. Engineering Foundation,
New York,pp.189-206.
Lefond, S.J., Editor (1983) Industrial Minerals and Rocks, Vol 1 and 2, SME
of the AIME, New York.
Leja, J. (1957) Mechanism of collector adsorption and dynamic attachment
of particles to air bubbles as derived from surface- chemical
studies, Bulletin ofIMM, 607, pp. 425-437.
Leja, J. (1982) Surface Chemistry of Froth Flotation, Plenum Press, New
York.
Leja, J., & Schulman, J.H. (1954) Flotation Theory: molecular interaction
between frothers and collectors at solid-liquid-air interfaces, Trans.
AIME., 16, pp.221-8.
Lekki, J., & Laskowski, J. (1971) On the dynamic effect of frother-
collector joint action in flotation, Trans. 1.M.M., Section C,lli), C174-
80.
Lekki, J., & Laskowski, J. (1975) A new concept of frothing in flotation
systems and a general classification of flotation frothers,
Proceedings of the 11th IMP Congress, Cagliari, M. Carta, ed.,
Instituto di Arte Mineraria, Universita di Cagliari, Cagliari, Italy, pp.
427-448.
Leroux, M., Rao, S.R, Finch, JA., Gervais, V., and Labonte, G. (1989)
Collectorless flotation in the processing of complex sulphide ores,
Adv. in Coal and mineral processing using Flotation, Chander S., and
Klimpel. R.R Eds., AIME, Littleton,
Li Guongming and Zhang Hingen (1989) The chemical principles of
flotation, activation and depression of arsenopyrite, Processing of
complex ores, Proc. Int. Symp, CIMM, Halifax, August 20-24, 1989,
pp.61-70, Pergamon Press.
Luttrell, G.H., Adel, G.T. and Yoon, RH. (1987) Modeling of column
flotation, SME annual meeting, Denver, CO, Feb. 24-27, Preprint
No. 87-130.
Malhotra, D., Hoover, RM., and Bender, F.N. (1977) An analysis of the
effect of some operating variables on flotation of molybdenite and its
implication on control, Preprint No 77-B-89, AIME annual meeting,
Atlanta, March 6-10,1977.
Malhotra, D., Hoover, R.M., and Bender, F.N. (1980) Effect of agitation on
flotation of molybdenite, Mining Engineering, p. 1392, September
1980.
Mathieu, G.!., and Sirois, L.L. (1984) New processes to float feldspathic and
ferrous minerals from quartz, Reagents in the Minerals Industry,
M.J.Jones and ROblatt, Eds, (London: IMM), pp. 57-67.
McKay, J.D., Foot Jr, D.G. and Shirts, M.B. (1987) Parameters affecting
column flotation of fluorite, SME annual meeting, Denver, CO, Feb.
24-27, Preprint No. 87-122.
Moys, M.H. (1978) A study of a plug-flow model for flotation froth
behaviour, Int. J Miner. Process.i S, pp. 21-38.
Mular, AL. (1989) Modelling, simulation and optimization of mineral
processing circuits, Challenges in mineral processing, Eds., K.V.S.
Sastry and M.e. Fuerstenau, SME-AIME, Littleton, CO, pp.323-349.
Mular, AL. (1976) Optimization in flotation plants, Flotation, Ed. M.C.
Fuerstenau, SME-AIME, Littleton, CO, pp. 895-936.
Mumme W.G.& Winter, G. (1971) Inorg. Nucl. Chern. Lett.,7, 505.
Nagaraj, D.R (1988) The chemistry and application of chelating or
complexing agents in mineral separations, Reagents in Mineral
Technology, Surface Science Series, vol. 27, P. Sumasundaran, Ed.
Marcel Dekker, New York, pp.257-333.
Nagaraj, D.R, and Gorken, A (1989) Potential controlled flotation and
depression of copper sulphides and-xanthate systems, Processing
complex ores, Proc. Int. Symp. CIM, Halifax, Aug. 20-24,
1989,pp.203-213 Pergamon Press, New York.
Nagaraj, D.R, Basilio, and Yo on, R.B. (1989) The chemistry and structure-
activity relationships for new sulphide collectors, Processing complex
ores, Proc. Int. Symp. CIM, Halifax, Aug. 20-24, 1989,pp.157-166,
Pergamon Press, New York.
Nokes, e.M., & Quigley, e.G., (1948), Differential froth flotation of
sulphide ores, U.S. Patent N 2,492,936.
O'Connor, C,']", Bradshaw, D.J., and Upton, AE. (1990) The use of
dithiophosphates and dithiocarbamates for the flotation of
Arsenopyrite, Minerals Engineering, Vot~, pp. 447-459.
O'Connor, e.T., and Mills, P.J.T. (1990) The effect of temperature on the
pulp and froth phases in flotation of pyrite, Minerals Engineering,
Vol,3, pp. 615-624.
Ozbayoglu, G., Atalay, M.U., and Basaran, B. (1985) Barite flotation: a
statistical experimental design approach, XVth Int. Min. Proc.
Congress, Cannes, June 2-9, 1985, vol. II, pp.337-345.
Palagi, c.o., and Stillar, S.S. (1976) The Anaconda c.e. Weed concentrator,
Flotation, Ed. M.e. Fuerstenau, SME-AIME, Littleton, CO, pp.
1029-1042.
Bibliography 333
Pauling, Linus (1970) Crystallography and chemical bonding of sulfide

minerals, Mineral. Soc. Amer. Spec. Pap.], pp. 125-131.
Pauling, Linus (1967) The chemical bond, Cornell University Press, Ithaca,
N.Y.
Pauling, Linus (1960) The Nature of the chemical bond, Cornell University
Press, Ithaca, N.Y.
Poling, G.W. (1976) Reactions between thiol reagents and sulphide minerals,
Flotation, Ed. M.e. Fuerstenau, SME-AIME, Littleton, CO, pp.
334-363.
Poling, G.W., and Vreugde, M.IA. (1985) Principles of mineral flotation,
The vVark Symposium, Australasian IMM, Parksville, Australia.
Potgieter van Aardt, J.R. (1951) U.S. Patent 2,561,251.
Powell, R.F. et al (1951) U.S Patent 2,591,289.
Ramsey, T. (1976) The Cyprus Pima concentrator, Flotation, Ed. M.e.
Fuerstenau, SME-AIME, Littleton, CO, pp.1079-1099.
Ranney, M.N. Editor (1980) Flotation agents and process: technology and
application, Noyes Data Corporation, Park Ridge, New Jersey.
Rao, S.R., and Finch, JA. (1989) A review of water re-use in flotation,
Minerals Engineering, Vol. 1, pp. 65-85
Rausch, D,O., and Mariacher, B.e. Editors (1970) Metallurgy of lead and
zinc, Vol. 1 and 2, TMS of AIME, New York, NY.
Redeker, Immo H. and Bentzen, E.H. (1986) Plant and laboratory practice in
non-metallic flotation, Chemical reagents in the mineral processing
industry, Editors, Deepak Malhotra and W.F. Riggs. SME of AIME,
Littleton, CO
Reid, E. Emmet (1962) Organic chemistry of bivalent sulfur, vol IV,
Thiocarbonic Acid and Derivatives, Chemical Publishing Co. New
York, N.Y.
Seidell-Linke, Solubilities of inorganic and metal-organic compounds, vol],
American Chern. Soc. Washington, DC.
Sharp, F.H. (1976) Lead-zinc-copper separation and current practice at
Magmont Hill, Flotation, Ed. M.e. Fuerstenau, SME-AIME,
Littleton, CO, pp.1215-1231.
Shashtry, K.V.S., and Fuerstenau, M.e. Editors (1989) Challenges in mineral
processing, SME of AIME, Littleton, CO.
Shimoiizaka, J. (1972) Fundamental studies on the flotation of complex ores,
Joint meeting MMIJ, AIME, Tokyo. .
Shirley, J.F. (1979) By-product Mo, in A. Sutulov, Ed., International
Molybdenum Encyclopaedia, Vol. 11, Intermet Publications,
Santiago, Chile, pp. 37-56.
Smit, FJ., and Foth, H.C. (1970) The 5000 TPD concentrator for milling the
complex oxidized lead-zinc ore at Tintic Division, , Metallurgy of
lead and zinc, Eds. Rausch, D.O., and Mariacher, B.C., Vol. 1, pp.
751-769, TMS of AIME, New York, NY.
Smith P.R (1976) Phosphate flotation, Flotation, Ed. M.C. Fuerstenau,
SME-AIME, Littleton, CO, pp.1265-1284.
Spendly, W., et a1., (1962) Sequential application of simplex designs in
optimization and EVOP, Technometrics.A, No.4, p.441
Steininger, J. (1968) The depression of sphalerite and pyrite by basic
complexes of copper and sulphhydryl flotation collectors, Trans.
SME/AIME, vol. 24, No.1, March 1968, p.34-42.
Strydom, PJ., Spitzer, D.P. and Goodman, RM. (1983) Fine coal flotation
with alcohol: dialkyl sulfosuccinate, Colloids and surfaces, ~, pp.
175-185.
Sulman, H.L. and Edser, E. (1924) U.S. Pat. 1492904.
Sutherland, KL. and Wark, LW. (1955) Principles of flotation Australasian
Institute of Mining and Metallurgy, Melbourne.
Sutulov, A (1963) El proceso de lixiviacion, precipitacion y flotacion,
Universidad de Concepcion, Chile.
Sutulov, A (1963) Flotacion de minerales, Universidad de Concepcion, Chile.
Sutulov, Alexander (1965) Molybdenum extractive metallurgy, Univ. de
Concepcion, Chile.
Sutulov, A (1974) Copper porphyries, University of Utah Printing Services,
Salt Lake City, Utah.
Sutulov, Alexander (1977) Flotation recovery of molybdenite, Paper
presented at 16th Ann. Conf. of Metallurgists, Vancouver, Canada,
August 1977.
Sutulov, Alexander (1978) Molybdenum and Rhenium, 1778-1977, Univ. de
Concepcion, Chile.
Taggart, AF. (1945, 1947) Handbook of mineral dressing, J. Wiley & Sons,
New York.
Tsai, M., Matsouka, I, and Shomoiizaka, J. (1971) The role of lime in
xanthate flotation of pyrite, J. Mining and Met. Inst. ofJapan, 87, pp.
1053-57, (in Japanese).
Tuwiner, S.B., & Korman, S. (1950) Effect of conditioning on flotation of
chalcocite, Trans. I.M.M.E.,187,pp.226-37, Min. Engng., Feb, 1950.
Tveter, E.C. (1952) U.S. Patent 2,611,485.
Tveter, E.C., & McQuiston Jr., F.W. (1962) Plant practice in sulphide
mineral flotation, Froth flotation, 50th anniversary volume, AIME,
New York.
Vartanian, K.T., & Gomelauri, LV. (1941) Depression of sulphides with
sodium sulphide,' U.S.S.R Patent N 48,010 (1936) and 63,803.
Bibliography 335
Weiss, A., Editor (1985) SME Mineral processing handbook, Soc. of Mining
Engineers, New York, N.Y.
Wills, BA. (1988) Mineral processing technology, 4th Edition, Pergamon
Press, Oxford.
Winter, G. (1975) Xanthates of sulfur: their possible role in flotation,
Inorganic and nuclearchemistry letters, VolJ1-, pp.113-118.
Woods, R. (1971) The oxidation of ethyl xanthate on platinum, gold,
copper, and galena electrodes. Relation to the mechanisms of
mineral flotation,!. Phys. Chem., vol.75, pp. 354-362.
Woods, R. (1972) Electrochemistry of sulfide flotation, Proc. Austr. IMM, No
241, pp. 53-62.
Woods, R (1972) The anodic oxidation of ethylxanthate on metal and galena
electrodes, Austr. J. Chem. vol. 25, pp.2329-2335.
Woods, R (1976) Electrochemistry of sulfide flotation, Flotation, AM.
Gaudin memorial volume, AIME, New York, N.Y.
Woods, Ronald (1977) Mixed potential mechanism in metallurgical systems,
Avances en flotacion,], Univ. de Concepcion, Chile.
Woods, R (1981) Mineral flotation, Chapter 11, Comprehensive treatise of
electrochemistry, Bockris, J. O'M., Conway, B.E., Yeager, E., and
White, RE., Eds., vol. 2, Plenum Publishing Corp. pp. 571-595.
Wrobel, SA. (1953) Power and stability of flotation frothers, governing
factors, Mine and Quarry Eng.,12, pp.363-7.
Yarar,B. and Spottiswood, JA. Editors (1982) Interfacial phenomena in
mineral processing. Proceedings of the engineering foundation
conference, Rindge, New Hampshire, Aug. 2-7, 1981. Engineering
Foundation, New York.
Yehia, Ahmed (1991) Adsorption of Aerosol O'I' on some synthetic
carbonate apatite, Minerals Engineering, In press.
Yianatos, J.B., Espinosa, R.G., Finch, JA. and Laplante, AR. (1989) Effect
of column height on flotation column performance, SME annual
meeting, Denver, CO, Feb. 24-27, Preprint No. 87-24.
Yoon,R.H., Luttrell, G.H., and Adel, G.T. (1989) Advances in fine particle
flotation, Challenges in mineral processing, Eds., K.V.S. Sastry and
M.e. Fuerstenau, SME, Littleton, CO, pp. 487-506.
\
SUBJECT ~NDEX
Activating agents 9, 102 Batch tests 295-313

Alaskite Bauxite 231
typical composition of 235 Beryl 231
Alcoholic frothers Beryllium 231
properties 94 Borates 232
Alkali metal alkyl xanthates Boron, important minerals 232
see also, xanthates Box-Wilson design layout 271, 272
manufacture 49 five-level 272
properties 51, 52, 55 three-variable 272, 273
stability 55 Bubble formation
suppliers and trade names 67 cavitation mechanism 18
Allied Colloids (UK) 71 Bubble-particle
Am. Cyanamid (USA) 57-72, 94, bond 40
98-100,115,123,125,207,209 contact 40-43
Amines interpenetration theory 43
as cationic collectors 226 Bulk flotation 9
critical miscelle concentration 226 CANDINA (Italy) 67-72
ionization constant 226 Carbonate ores 214
solubility 228 Carboxylates, see fatty acids
Ammonium hydrosulphide 116 Case histories
Anamol D, or arsenic Nokes 116 CODELCO Andina 136-147
chemistry 121 CODELCO EI Teniente 147-173
Anionic collector chemistry San Manuel 127-135
miscelle concentration 224 Cationic collectors 226
structure 223 chemistry 225, 227
Arsenic Nokes 116 miscellization 226
chemistry 121 PKa 228
Arsenopyrite solubility 228
lime as depressant 163
Barite 231
- 336-
Subject Index 337
Cement rock 234 Collector-mineral

Chalcocite bonding 4D-4D
depression of 132 interface 35
Iso-recovery curves 88, 89 Column testing 313
Chalcopyrite two inch laboratory column 321
depression of 132, 133 variables in cleaner study 318
structure 39 Complex ores
CIMM column flotation flotation circuit 204, 209
pilot plant 317 minerals contained in 206, 210
Chromic acid 104 Concentrate steaming 113, 116
Chromates Contact angle 13
depressant for galena 104 definition 37
Chromite 233 values for xanthates 16
CODELCO Andina 136-147 Copper sulphate 102
lab results sand/slime test 138-143 Copper-molybdenite separation 128-135
mill test sand/slime circuit 144-148 Copper mill flotation practice 179-191
rougher circuit 136, 137 Cresylic acid
CODELCO EI Teniente 147-173 frothing properties 97
column tests moly plant 172 Critical miscelle concentration
copper cleaner 158-161, 166-168 anionic collectors 224
dissolved Mo in tailings 164 cationic collectors 226
effect of pH on Cu/Mo/pyrite 164 Critical pH for flotation 163
lime, depressant effect on moly 162 Cyanamid, see Am. Cyanamid
mine sample floatability profile 157 Cyanide, depressant 103
mill material balance 151, 152 Cycle tests 307
moly recovery 1985/91149,169,170 Cyclic alcohols as frothers 96
moly recovery, acid vs alcaline 163 Depressants 103-105
Nokes depressant studies 171 quebracho 105
ore mineralogy 153-155 starch 105
sand/slime split 168 Depressants, copper moly circuits 116
Colemanite 232 ammonium sulphide 116
Collectors for non-metallics 220 Nokes reagents 116, 120-122
Collector suppliers and usage sodium cyanide 103, 116
alkyl xanthates 67, 74-77 sodium ferri, ferro cyanide 104, 116, 132
dialkyl thionocarbamates 70, 80 sodium sulphide and sulphhydrate 107, 116
dialkyl thiophos. 70, 71, 72, 81-83 sodium zinc cyanide 116
dialkyl xanthogen formates 69, 79 Dialkyl thionocarbamates
miscellaneous collectors 72, 84 manufacture 62-64
xanthic esters 68, 78 suppliers and trade names 69
Collectors 9 Dialkyl xanthogen formates
consumption 176, 198 manufacture 60
popularity 195 properties 61
structures 7 suppliers and trade names 69
_n .. _
338 FLOTATJOl'T - Theory, Reagents and Ore Testing
Differential flotation 9 Ferro and ferricyanides 104, 132

Dithiolate 35 Float technician selection 284
chemistry 29 Flotation
nomenclature 33 coursing bubble 12
reagents 73, 108 mill operating norms 193
DithiolatejThiol couples nascent bubble 17
reduction potentials 34 process design 200
Dithiophosphoric acids Flotation cells
boiling points 56 impeller and diffuser 283
manufacture 57 standard 5.5 liter (2 kg) 281, 282
Dixanthogen 29 Flotation column 314
adsorption 105 design of the indus. columns 315
chemical properties 31 pilot plant at Chile's CIMM 316
galena 20 testing examples 317-323
Double layer, 17 Flotation machines, laboratory
potential, 20 comparison of machines 285
frothers 99 Flotation of complex ores 209
Dow Chemical Co (USA) 49-55,62-70, 73, Flotation mechanism 13, 16
91,98-100,115,134, 150,177,285 cavitation 18
Electrical double layer theory 21 Flotation reagents
Electrochemical properties collectors 66-73
phase diag., see potential-pH diagram depressants 101-106
minerals 38 frothers 94-100
potential of sulphides, 103 modifiers 101-106
Electron transfer to min. surface 19 Fluorspar 238
Enargite grade specifications 240
structure 39 reagents for flotation 240
collectors for 73, 205 zinc heavy media concentration 239
Equipping the flotation lab. 278 Foote spudomene flotation 254
apparatus required 311 Froth flotation
ball mill 301, 302, 305 definition of terms 5
charge 306 in the U.SA. 175-191
flotation machines 279 Froth hydrodynamics
Exfoam '636' 134 effect of collectors 86
Experimental design stability, dynamic 85
false optima one variable 274 Frothers 9
Fatty acid collectors 221 alcoholic frother 94
see also anionic collectors alkoxy paraffins 98
composition 222 chemical properties 94-100
formulae 223 consumption 177
Feldspar 234 cresols 97
mica-quartz-lithia, reagents 237
Subject Index 339
Frothers (continued) Laboratory flot. machine 285-287

Cyanamid 99 comparison of performance 285
Dowfroths 99 Lead nitrate or acetate 103
effect of, on chalcocite rec. 88 activator for NaCI 215
ketone 96 Leeds Autofloat, reproducibility 28
popularity 197 Lime, see also depressants 104
power 86 depressant effect on moly 162
stability 85 LPFor
structures 8 Leach Precipitation Float 124, 125, 213
terpenes 96 LR-744, see phosphate Nokes
US consumption 177 chemistry 120
use in copper mills 185-190 Lubrizol (USA) 71
volume versus alcohol chain 87 Manufacture of
water soluble 99 thionocarbamates 62-64
Frother-collector thiophosphates 57
interaction complex 44 xanthates 49-54
surface complexes 43-46 xanthogen formates 60
Fuel oil 134 Mechanism of bubble formation 18
Glass sands 241 Mercaptans5,36,62,64, 72, 73,20
flotation results 245, 246 Mercaptobenzothiazole 65, 72
lock cycle tests 244 surface properties 146
reagents, flotation 243 Methyl isobutyl carbinol see MIBC
specifications 242 MIBC 134
Glue as depressant 105 Microscopic analysis moly plant 135
Graphite flotation 245 Minerals, electrochemical properties 38
Gums 106 Minerec (USA) 1, 31, 69-71
Gypsum 247 Modifiers 9,10,101-106
Hallimond tube 13, 86, 212, 266, 287 Molybdenite plants 112-116
Hoechst (Germany) 67-72 depressants 129-134
Hydrofluoric acid 104 mineralogy 135
Hydrogen peroxide 130 reagent consumption 135
Hydrogen sulphide 103 recovery circuit, San Manuel 128
Ice, structure of 22 recovery study EI Teniente 147-173
Interpenetration theory 43 the sulphide process 113, 119
Iodine flotation 246 Molybdenum cone, cleaning 111-126
10M (Mexico) 67, 68 Nascent bubble flotation 17
Isorecovery curves Natural oils 95
chalcocite 89 Natural'gums 106
Kaolin flotation 247 Nitrogenous cationic agents 226
Kerley Chemical (USA) 67, 68 Non-metallic minerals
Kyanite flotation 247 collectors for 230
Laboratory ball mills 310, 302, 305 fatty acid collectors 221
flotation mechanism 217
Nokes reagents, see LR-744, Anamol Properties of

chemistry of 120-122 alcoholic frothers 94
Oils and Fats alkyl xanthates 51, 52, 55
fatty acid content of 222 alkyl xanthogen ethyl formates 61
Oily collector flotation 213 dithiophospahtes 58 '::= p
ORP-Oxidation reduction potential 129 terpene and ketone frothers 96
Over oiling 16, 92 water soluble frothers 99
Oxide minerals 229 Prospect Chemical (Canada) 68,69
flotation model 213 Pyrite
Oxygen role in collector adsorption 19 depression 133, 162, 164
Particle size Pyrrhotite
effect on flotation 168 activation by copper sulphate 102
froth stabilization 85 Quartz and feldspars, specification 242
Pegmatite processing 236 Quebracho 105
ores 256 Rate of flotation
Pennwalt (USA) 64 copper ores 90
Perxanthates 20 effect of collector 90
pH effect of frother 88
critical pH for flotation 101 Klimpel curve 93
effect of collectors 102 RENASA (Peru) 67
effect of cyanide 102 Sampling and sample preparation 287-294
effect on molecular adsorption, 57 minimum representative size 289
effect on sulphide ion 120 Sampling working concentrators 291
Phillips 66 (USA) 64, 70, 72 San Manuel
Phosphate Nokes 116 copper and moly assays 129
chemistry 120 molybdenite recovery circuit 127-135
Phosphate rock process 248 Sand/slimes study - Andina 136-147
reagents 250 Scheelite flotation 253
Pine oil 96 reagents 253
Pilot plant flowsheet for sands 224 Screen size equivalents 201
Polar minerals Semi-soluble salt flotation'215
classification of 6 Shellflot (Chile) 67,69,70
Polyglycol glycerol ethers 100 Sherex Chemical Co,
Polypropylene glycol ethers 99 coal tests 310-313
Potash flotation 250 standard test procedures 303
reagents used 252, 253 Sherrit-Gordon (Canada) 67, 68
Potential-pH diag. moly-sulphur-Hg-, 165 Slimes
flotation recovery 139-146
sulphidisation sand/slime separation 136
Subject Index 341
Sodium cyanide 103 SUlphide minerals

Sodium hydrosulphide 109 band-gap energies 40
Sodium hydroxide 104 Sulphidisation
Sodium hypochlorite 133, 134 Non-sulphide copper ores 208
Sodium silicate 104 of refractory ores 122-126
Sodium sulphide, 107-111 Sulphonates 5
chemical properties 108 Sulphur dioxide 104
concentration effect on pH 111 SUlphuric acid 104, 134
ionization as a function of pH 109 Talc 257
ionization idem temperature 111 Tannic acid 105
phases in equilibrium 108 TEB98
solubility in water 108 Tecnomin (Chile) 69
Sodium sulphites and hyposulphites 104 Testing
Sodium thioglycolate 116 design 260
Sodium zinc cyanide 129 variables 259
Sphalerite Thiocarbanilide 65
structure 39 Thiol collectors
Spudomene 254 bonding with sulphide minerals 36
mineralogical analysis 255 stereochemistry of xanthates 25-30
and co-products (pegmatite ores) 256 idem dithiophosphates 29
and feldspars-quartz-mica flotation 235 Thiolated surfaces
SSDEVOP: a response surface 276 products extracted from 35
Starch 105 Thiophosphates 55-58
Stereochemistry 25-30 O,O-dialkyl dithiophosphates 55
dithiophosphates 29 O,O-diaryl dithiophosphoric acids 57
metal xanthates 25 manufacture 57
[M(bidentateh]X 26 properties 58
octahedron 27 suppliers and trade names 70-72
pentagonal bipyramid 27 VCB (Belgium) 67
seven-coordination 26 V.SA. flotation statistics 175-191
six-coordination 25 Vermiculite 257
Structure of water 23-25 suppliers and trade names 69
Structures, suppliers and mineral use of Water
alkyl xanthates 67, 74-77 at mineral surface 17
dialkyl thionocarbamates 70, 80 clathrate 218
dialkyl thiophosphate 70, 71, 72, 81-83 dilution 134
dialkyl xanthogen formates 69, 79, 136 structure 23, 218
miscellaneous collectors 72, 84 W ollastonite 257
xanthic esters 68, 78 Wurtzite
structure 39
Xanthates, see also thiols

analysis 52
analitical method 55
handling and safety 54
manufacture 49, 53
properties of 47-55
solubility 51
solutions stability 51, 55
suppliers and trade names 67
Xanthic acids, see xanthates
Xanthic anhydrides
properties 31
Xanthic esters
allyl ester 64
suppliers and trade names 68
Xanthogen formates 58
manufacture of 60
properties 61
Zinc sulphate 104
ZUPA (Yugoslavia) 67, 70
AUTHOR INDEX
Baleshta, T.M. 40 Klimpel, Richard 44,89,90,92

Beas-Bustos, E. 147 Laskowski, Janusz 14,23,40,87,88,266
Booth, RB. 98, 196 L~a,J. 14,24,30,43,46,86, 117,217,228
Cambron, Adrien 29 Lekki, J. 87,88,266
Castleman Jr., A W. 217 Malhotra, D. 169
Castro, S. 129, 162, 171 Mathieu, G.1. 242
Cea, Zocimo 162 Mular, AL. 274, 275, 276
Chander, S. 162 Mumme, W.G. 27
Clark, Alfred 36 Nokes, CM. 120
Crozier, RD. 23, 44, 63,66,116,176 O'Connor, C.T. 285,291
Davis, Farlow C 127, 128 Palagi, CG. 125
Douglass, W. 32, 62, 73 Pauling, Linus 22, 37, 39, 45
Drew, M.G.B. 24, 27 Poling,G.W.5,29,34
Ewers, W.E. 85 Powell, R.F. 98
Finkelstein, N.P. 5, 29, 32, 33, 34 Redeker, Immo H. 230, 234, 241, 257
Fischer, AH. 32,73,153 Reid, E. Emmet 29
Fuerstenau, D.W.162 Shimollzaka,J.166
Gardner, J.R 14 Shirley, J.F. 119, 120, 162
Gaudin, AM. 12 Smith, R W. 228
Glembotskii, VA. 41,162 Sutherland, K.L. 13, 38, 85, 101
Grainger-Allen, T.J.N. 18 Taggart, AF. 15, 17, 31, 68
Granville, A 33 Tsai, M., 166
Gutierrez, G. 121 Tveter, E.C 98
Hansen, RD. 89 Wills, BA. 5, 6
Harris, Guy 49, 50, 62 Winter, G. 15, 27, 36
Houben-Weyl56 Woods,R. 14,19,33,34,36
Li Guongming 162 Wrobel, SA. 86
Jones, M.H. 19, 122 Yehia, Ahmed 218
\
Kepert, D.L. 25, 28
Klassen, V.1. 12, 14, 15,17,40,117
343 -

Flotation Theory Reagents and Ore Testing PDF

Uploaded by

Copyright:

Available Formats

Flotation Theory Reagents and Ore Testing PDF

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

Flotation Theory Reagents and Ore Testing PDF

Uploaded by

Copyright:

Available Formats

flotation

THEORY, REAGENTS AND ORE TESTING

Theorv, Reagents and Ore Testing

FINCH and DOBBY

Canadian Metallurgical Quarterly

International Journal of Rack Mechanics and Mining Sciences &

The Physics of Metals & Metallography

Scripta Metallurgica et Materialia

Full details of all Pergamon publications/free specimen copy of any

RONALD D. CROZIER, C.Eng., F.I.M.M., Ph.D.

U,IC Pergamon Press pie, Headington Hill Hall,

Copyright @ 1992 Pergamon Press pie

British Library Cataloguing in Publication Data

Printed in Great Britain by BPCC Wheatons Ud, Exeter

Chapter 1. Flotaton Fundamentals 5

Chapter 2. The Mechansm oC Flotation 11

Chapter 3. Sulphide Mineral Flotation 25

Chapter 4. Thiol Collector Chemistry 46

Chapter 5. Cornrnercial Sulphhydryl Collectors 66

Chapter 6. Propertles of Flotaton Froths 85

Chapter 7. Chemcal Properties o Frothers 94

Chapter 8. Flotation Modifiers 101

Chapter 9. Sulpbides - Depression and Activation 107

Chapter 10. Mili Tests - Case Histories 127

Chapter 11. Sulpbide Mili Practice 174

Chapter 13. Flotaton Testing 258

Subject Index 336

Author Index 343

Table 1.- Classification of polar minerals 6

Fig. 1.- Cavitation rnechanisrn of bubble forrnation 18

Plenum Publishing Corporation, 233 Spring Street, New York, NY 10013

Cornell University Press, 124 Robert Place, Ithaca, NY 14850

The Society of Mining Engineers, AIME-SME, Littleton, ca

Mine and Quarry Engineerng, 19 Fairlie Road, London, UK

Pergamon Press, Headington Hill Hall, Oxford OX3 OBW

This text is primarily addressed to the mill metallurgist and to technical

As a primary target of flotation testing is correction of a chemical problem in

micro-flotation laboratories usually mimics the obsolete pneumatic flotation

The deve!opment of an optimum metallurgical sequence to pro cess a

quantity of ore that will be available. However, no similar computer

The chemistry of mineral surfaces is dealt with systematically, and is readily

This mechanistic factor, i.e. that a particular reagent's flotation response is

insoluble minerals, and (4) non-metallic soluble, or semi-soluble minera!s. In

Sub-groups of these four minerals are ores whose mineral dissemination

Froth flotation, the surface dependent process par excellence, consists of a

Frotlt flotation is a process used to separate minerals, suspended in liquids, by

Polar minerals have strong covalent or ionic surfacebonding, and exhibit

groups: those rninerals that form a xanthate, dithiophosphate, mercaptan,

TabLe 1.- CLASSIFICATION OF POLAR MINERALS

Group 1 Group 1 continued

GaLena PbS Orpiment As2S3

Hematite Fe203 Zircon ZrSi0 4

Table 2.- COLLECTOR STRUCTURES

Alkyl xanthates or R-O-C-SNa Soluble

Insoluble Oils R-O-8H very sltly

Metal oxide and non-sulphide mineral collectors

Alkyl amines R-HNH2 Soluble

The reason to make this differentiation is that the mechanism of flotation is

TabLe 3.- FROTHER STRUCTURES