Mineral Processing, 3rd Edition (1985)
Mineral Processing, 3rd Edition (1985)
Mineral Processing, 3rd Edition (1985)
MINERAL
PROCESSING
by
E. 1. PRYOR
A.R.S.M., D.S e., D.I.C., M.I.M.M.
THIRD EDITION
Chapter Page
A General Introduction
2 Acceptance Into The Mill 14
3 Primary Crushing 31
4 Secondary Crushing 58
5 Wet-Grinding Mills .. 72
6 Forces in Wet Grinding 105
7 Dry Grinding .. 127
8 Laboratory Sizing Control 140
9 Industrial Sizing and Sorting 179
10 Grinding Circuit Control 233
11 Methods of Separation 247
12 Dense Media Separation 263
13 Separation in Vertical Currents 295
14 Separation in Streaming Currents .. 324
15 Physics and Chemistry in Ore Treatment 369
16 Chemical Extraction 410
17 Principles of Froth Flotation 457
18 Flotation Practice 520
19 Magnetic and Electrical Separation 571
20 Testing and Research 600
21 Sampling and Controls 634
22 Unit Processes and Machines 657
23 Selected Ore Treatments 702
Appendices A Glossary 809
B Beaker Decantation .. 819
C Abbreviations 822
Index 824
v
PREFACE TO THIRD EDITION
32. Without these cannot a city be inhabited: and they shall not dwell
where they will, nor go up and down:
33. They shall not be sought for in publick counsel, nor sit high in the
congregation: they shall not sit on the judges' seat, nor understand the sen-
tence of judgement: they cannot declare justice and judgement; and they shall
not be found where parables are spoken.
34. But they will maintain the state of the world, and [all] their desire is in
the work of their craft.
E. J. PRYOR
DEVON,
September, 1965.
vii
PREFACE TO FIRST EDITION
The purpose of this book is to set out the basic principles which underlie
sound milling practice and to show their relation to standard commercial
operations. The sources named in the bibliography are but a few among the
number which have been studied in its preparation. Detailed acknowledge-
ment would be as impossible as would adequate recognition of private
correspondence and discussion with the teachers, research workers, consulting
engineers, and milling men who have so generously shared their experience.
The first draft of this book was made some twenty years ago. Revision
kept step with growing field experience and accelerated from the year
1944, from which time the author enjoyed the research and testing
facilities of the mineral-dressing section of the Bessemer Laboratory of the
Royal School of Mines. Generous grants from the Nuffield Foundation
and the Imperial College, and practical aid from industrial sources, made it
possible for him to visit a wide range of laboratories and plants in America,
Canada, Europe, and Africa and to study both the principles of mineral
dressing and their commercial application.
The author wishes to express his indebtedness to the following publishers
for permission to reproduce from the textbooks mentioned the various
illustrations indicated:
viii
Reinhold Publishing Corporation. From Riegel's Chemical Machinery
Figs. 6, 50, and 213.
Messrs. John Wiley & Sons, Inc. From Taggart's Handbook of Mineral
Dressing Figs. 7, 15,38,54,61, 113, 114, 138,186 and 233.
Messrs. Macmillan & Co., Ltd. From Truscott's Text-Book of Ore
Dressing Figs. 8,9, 10, 11, 12, 14, 16,20,28,29,30,31,32,33, 35, 37,
59, 60, 65, 66, 67, 68, 69, 100, 109, 111, 112, 115, 119, 133, 136, 225, 226,
227 and 232.
IX
CHAPTER 1
A GENERAL INTRODUCTION
This book is designed to describe the basic principles used in the concen-
tration of specific minerals from their ores. It further deals with the develop-
ment of these principles in industry and the requirements for effective process
control. The earlier editions were written in connexion with the teaching of
mineral technology at the Royal School of Mines in London, and were
designed to be used together with the laboratory study and plant experience
essential for a sound grasp of this science and its applications. This correlation
of principles and practice is continued here.
Physical science is operative in the whole field of human activity. In the
study of mineral dressing, which calls for special application of the laws of
physics, chemistry, and electricity, the term "physical science" is here used to
embrace these three categories. Its technology derives from fundamental
physical laws, and operates at two levels-basic and applied. The word
"technology" is here used in its sense of the combination of technical skill
and economic justification, since successful field operation is mainly judged by
financial criteria. "Basic technology" considers the general principles
applicable to a standard type of operation, while "applied technology"
connotes specially imposed variations designed to improve the treatment of
one specific ore body.
A mineral deposit normally contains at least two distinct types, or species,
of mineral. In the ore mined and sent to the mill these are closely inter-
locked. Mineral processing starts by freeing (liberating) the desired mineral
or minerals, by comminution. It then separates them.
Scientific application of the principles of physical chemistry to industrial
processing has led to rapid expansions in the chemical engineering industry,
with which mineral dressing has much in common. Indeed, the cyanide
process, used for the solution of gold from its ores, was among the earliest
continuously applied chemical processes, and was developed before chemical
engineering had become a profession.
The engineering crafts are called for in handling, transporting. segregating,
and disposing of large tonnages of rock with smoothness and precision.
Here, as with the chemical controls used in the flotation process, automation
is widely used in ensuring consistent quality of material at each stage of
treatment. Surface chemistry and physics, magnetic force and electrostatics
are important factors in modern milling. Leaching is reinforced by the use
of pressure and heat, and ion-exchange methods have been developed to deal
with the dissolved products, notably in uranium recovery.
It would be impossible to cover this wide field of scientific application in one
2 Mineral Processing-A General Introduction
After this division, which must be carried down to the resolution called for
by the problem, the possible techniques which can be used in developing the
desired effects can be considered and tested out in the laboratory, the pilot
plant, and the mill-in that order. These techniques embody methods of
applying fundamental scientific principles in such wise as to influence the
behaviour of the material under treatment. Preparation, including liberation,
is the first step in mineral engineering and is dealt with in the early chapters.
Separation, or concentration, comes next, and disposal last.
For easy reference some definitions of words or phrases which have a
special meaning in mineral dressing have been grouped in alphabetical order
in Appendix A (Glossary) at the end of the book. The definitions are
sometimes given in the text when they first arise.
Objectives
These are of two kinds, technical and economic. To bring the marketable
(or shipping, selling) product or "concentrate" into the technical condition
required by the customer, unwanted constituents in the original ore must be
removed or reduced below some specified percentage. The product may have to
conform to requirements as to particle size, assay grade, moisture content.
If more than one valuable mineral is present, the mineral dresser may be
obliged to separate them so that each can be marketed separately, or so that
the purchaser can handle them economically. The smelter or other purchaser
protects himself from financial loss by imposing a penalty on all concentrate
failing to reach the agreed grade, and it becomes the duty of the mill manager
to ensure that this grade is reached or exceeded. Broadly, the requirements
for a concentrate include its assay grade (with respect to the mineral or
minerals contained); its freedom from associated minerals; its moisture; its
sulphur content; and its grain size. Where the same element occurs in two
different combinations requiring different forms of subsequent treatment,
appropriate separation may be economically justified. An example of this
is ore in which copper occurs as sulphide and as carbonate. The former
concentrate is smelted and the latter leached at N'Changa, in Zambia.
Some of the valuable product is inevitably lost in the tailings finally rejected.
A second objective is to keep this loss as low as is economically justifiable.
The cost of treatment must therefore be balanced against the revenue
obtained by sale of concentrate.
A further requirement is that the plant shall handle an adequate tonnage
from the mine. Failure to do so might raise the overall operating cost or
otherwise complicate the orderly conduct of the enterprise.
Scope
The mill receives its raw material from the working places whence ore is
being mined or quarried. After concentration, the products pass from the mill
to a smelter or other consumer. These stages may be depicted in flow-sheet
form
4 Mineral Processing-A General Introduction
~
~r----- Mill Treatment -------,~ (2)
~
Further Treatment
t
Waste Dump
(e.g. Smelting)
Mineral Characteristics
The minerals associated in a piece of ore are given generic names according
to their composition-e.g., galena (lead sulphide), apatite (calcium phosphate),
quartz (silica). Samples of a given mineral from different sources usually
react similarly, but traces of impurity are sometimes present in quantities
too small to be discerned by ordinary assay, yet sufficient to modify the
response of the mineral to treatment. In theory chromite consists of FeO.
Cr.O., but it frequently contains aluminium or magnesium and has wide
variations in both Fe :Cr ratio and the Cr content. Again, malachite
(CuCO •. Cu(OH).) occurs in nature with a copper content far below that
indicated by its formula owing to intergrowth with silicates, dolomite, etc.
Cassiterite rarely approaches the 78·8 % Sn indicated by the formula SnO.,
its crystal lattice being intimately penetrated by other elements. It is there-
fore useful to remember that the generic name of a mineral does not guarantee
that specific behaviour of samples from two different deposits will be identical.
This will be further discussed when the physics of froth-flotation are co~
sidered.
Hardness, together with toughness, are the main determining factors in
crushing and grinding. Its hardness affects the brittleness and friability of a
mineral. Toughness determines its resilience and elasticity. The interplay
of these two qualities in a homogeneous piece of rock determines its response
to the compressive forces (crushing) and abrasive ones (grinding) compre-
hended in the word "comminution". A lump of ore is not homogeneous,
however, since it contains more than one mineral species. Assessment of its
6 Mineral Processing-A General Introduction
TABLE 1
HARDNESS SCALES OF MOH AND OF KNOOP
Toughness is influenced by the manner in which the crystals of the ore are
interlocked, and by their grain or size. A very fine (crystallite) structure is
usually more resistant than a coarsely crystallised rock. Many ores consist
mainly of valuable metallic sulphides and unwanted silica minerals, the
former often being friable and the latter tough. There is a tendency for the
smaller particles to be richer than the bigger lumps. and this can lead to
losses in treatment if it is ignored. It can also lead to sampling errors when a
heap of ore is being reduced to a small sample for assay purposes, if all the
rock is not systematically blended with respect to particle size.
Mined ore which is treated by mineral processing consists of two or more
minerals which vary in their resistance to applied crushing forces. The more
friable mineral or minerals are most likely to yield under crushing stresses and
strains. The weakening of the structure of a given particle by these stresses
depends partly on the percentage of each constituent mineral it contains.
Broadly, a micro-crystalline structure with the values disseminated evenly
Mineral Processing-A General Introduction 7
through the gangue minerals is far more resistant to crushing than one which
is coarsely crystalline. If in addition the values occur as occasional coarse
crystals instead of being minute and disseminated, these crystals may crush
even more readily. Grindability must be assessed specifically for each ore
complex. It cannot be predicted from knowledge of the behaviour of the indi-
vidual minerals of which a particular ore is composed. Various assessments.
such as the Hardgrove rating', based on the attrition resulting from dry grind-
ing of material under controlled conditions, have been published. This
extremely important subject, which affects both treatment cost and efficiency
of concentration, is developed in later chapters. One of the qualities affecting
grindability is the abrasiveness of the ore. This bears directly on the cost of
replacement of machinery worn during grinding contact and on the cost of
the power used in this unavoidable wear.
Methods of Treatment
When the mineral species in a given ore have been sufficiently freed from
one another by comminution, the process of concentration becomes possible.
The liberated particles must differ sufficiently in thf;ir physical, electrical, or
chemical properties to respond to an appropriate differentiating force.
Physical differentiation uses the shape, size, surface area, specific gravity,
porosity, colour, and gliding friction to sort out the mixed species. Electrical
treatment works on magnetic susceptibility (natural or induced), conductivity,
and radio-activity. Chemical attack depends on the reactivity of minerals in
a given bonded or lattice structure. All three types of attack shade into one
another and may be mixed in a given process. A list of the most important
exploitable characteristics, togt:ther with the mode of application, is given in
Table 2. These characteristics are developed by selected methods of grinding
and further treatment, which should always be based on preliminary study of
the specific ore.
Treatments in which chemical solutions attack mineral suspended in an
aqueous pulp are in growing use. The cyanidation of gold and silver ores is
one of the oldest continuous processes. The first stage of uranium recovery
is a chemical attack by acids or alkalis. Other elements treated by leaching
include copper, aluminium, manganese, nickel, tungsten, and zinc. Mineral
processing normally confines its operation to separation of ore constituents
without changing their physical state. In hydrometallurgy the chemical pro-
cesses are conveniently applied to ore in transit through the mill, and this leads
in practice to the entry of the metallurgist into the plant. These hydro-
metallurgical methods are discussed briefly in connexion with their use of
mill machinery in Chapter 16.
Pyrometallurgy is sometimes combined with mineral dressing. In one pro-
cess tin ore is chloridised and the metal chloride is then volatilised. In another
a mixture of smelted metal sulphides is separated by froth-flotation. Other
processes include the selective coating of one mineral in a mixture with
magnetic paint. the electro-chemical separation of antimony and zinc from
their ores, and other treatments, such as the extraction of magnesium from
8 Mineral Processing-A General Introduction
TABLE 2
PRINCIPAL EXPLOITABLE CHARACTERISTICS
Selective Type of
Mineral Separating Operation
Characteristic Force
Colour, Lustre. Visual, manual, Hand-sorting of graded ore, to
automated. remove detritus, waste rock,
special constituents. May use
fluorescent lighting, or impulses
triggered by reflected light.
Specific Gravity. Differential move- "Gravity" separation of sands and
ment due to mass gravels by D.M.S., jig, sluice,
effects, usually in shaking table, spiral.
hydraulic currents.
Surface Reactivity. Differential Surface Removal of relatively aerophilic
Tension in Water. mineral as froth from aerated
pulp by froth-flotation. Widely
used process.
Chemical Reactivity Solvation by appro- Hydrometallurgy. Ore exposed to
priate chemicals. solvating chemicals, perhaps with
heat and pressure, then filtered.
Dissolved element/s recovered
from filtrate, chemically, electro-
lytically or by ion exchange.
Ferro-Magnetism. Magnetic. Magnetic devices remove the pre-
ferred mineral. Also used to
remove "tramp" iron.
Conductivity. Electrostatic charge. Particles pass through high-voltage
zone. Rate of dissipation of in-
duced charge influences subse-
quent deflection. Differential
conductivity.
Radio-activity. IX or f3 rays. Emissions are signalled by G.M.
Valve which also activates a
separating or "picking" device.
Shape. Frictional. Sliding force is opposed by "cling"
of particle, resultant movement
depending on cross-section and
area, hence on shape.
Texture. Crushing. Screen- Characteristic shapes and surfaces
ing. Classifying. are developed during comminu-
tion.
sea water. In this book, which is concerned with the basic principles used in
standard methods of treatment, these specialised techniques are not discussed.
Choice of Method
A method of ore treatment must be chosen which is appropriate to a specific
ore as well as to a set of general principles. There are three main types of ore
complex-the massive, the intergrown, and the disseminated. Coal, and some
bedded deposits such as iron seams, are examples of the first-named type,
Mineral Processing-A General Introduction 9
Sequence of Operations
The flow-sheet is sometimes presented diagrammatically or in an embellished
form. Since it is functionally a chart showing the sequence of operations, it
should be simple. On the flow-sheet are marked any details required for the
operator's guidance. The author uses it in a form which can be typed and
cyclostyled, so as to make it a simple method of record, change and instruc-
tion. Thus considered, it can be presented (Fig. 2) as a block flow-sheet in
which all operations of one character are grouped. In this case "Prepara-
tion" deals with all crushing, grinding, and preliminary rejection. The next
block, "Separation", groups the various treatments incident to production
of concentrate and tailing. The third, "Product Handling", covers the dis-
posal of the products.
Preparation
sepJation
1
Product Handling
The main purpose of this flow-sheet is to analyse the general scheme, which
in a large mill can be very complicated, and to fix attention on the need for
"quality control" of the condition of the ore or pulp before it is transferred
from one stage of treatment to the next.
The second, and for many purposes sufficient, type is the simple line flow-
sheet (Fig. 3). On it can be shown details of machines, settings, rates, etc.,
or a figure may be added to each descriptive label for use in a separate
10 Mineral Processing-A General Introduction
reference file. This makes possible the recording of detail, without loss of the
essential simplicity and compactness which is so desirable. A large sheet of
paper can be handled in an office, but is a nuisance in the mill.
Run-of-mine ore is long-ranged in its sizes, and the biggest lumps may be
difficult to transport. It may be mixed with small quantities of waste rock
and debris. The primary purpose of all comminution is the unlocking of
values in the ore, but crushing may be combined with sorting, and performed
in stages. Screens are used to prevent oversized lumps of ore from entering
machines not built to handle them. Screens can reduce the size range of ore
fed to a concentrating appliance, if this improves its efficiency.
t t
Wet Grinding Plant
~
(+) .....- - - Classification --~) (-)
I
lr------- S +.
eparatIOn ----------.1
Concentrate Tailing
When dry crushing has produced pieces of rock so small that they can no
longer be gripped readily between crushing faces, in the way a nutcracker
grips a nut, wet grinding normally follows, with random blows rained upon a
passing stream of ore suspended in water. (For some purposes dry grinding
followed by dry treatment is practised. Minerals thus treated include talc,
cement, clinker, asbestos, and those treated in waterless country.) As the
wet-ground mixture, now called ore pulp or pulp, leaves the grinding device
it may be sent direct to the concentrating section. This form of transfer is
called open-circuit grinding. More usually, the pulp is passed through a
device which returns oversize for further attrition and sends on ore which has
been sufficiently ground. This arrangement is called closed-circuit grinding,
and is widely practised. Sometimes screens close the wet circuit (see Fig. 3),
but usually the solid part of the pulp is now too finely ground for their effec-
Mineral Processing-A General Introduction 11
tive use, in which case it can be sent through devices called classifiers. In
these the ore particles are separated into coarse and fine fractions relatively to
their rate of fall through water. Grinding can be done in one or more
stages of size reduction. The larger the tonnage treated the greater is the
justification for breaking down every operation into several stages, each
arranged to be applied within restricted limits. Mills which handle 50,000
tons of ore daily have not the limitations of the small 50-ton outfit, where one
machine must perform a wide range of duties at each stage.
The grinding section is controlled so as to deliver material in an appropriate
condition to the next stage of the treatment. Some easily won concentrate
or some undesirable ore-constituents may have been "scalped out" in the
grinding circuit, but the main work of concentration is done in a section of its
own. If more than one mineral is to be concentrated selectively, two or more
such inter-related yet independent sections are provided in series. Here the
work of separating a sufficiently high-grade concentrate is performed, to-
gether with the production of a tailing so low in value that further treatment
would not be profitable.
Each product leaves the machines either as moist gravel or sand, or as a
fluid pulp. In the latter case it can be run to waste if a tailing or de-watered
and disposed of (shipped). The various processes are controlled by sampling
and assay. Tailings must be disposed of either by dumping or running them
into a retaining dam, and the water used in the milling treatment may either
be reused or run to waste. Precautions are sometimes needed in the latter
case to ensure that no contamination of the district water supply results.
Products
Economics
In this chapter the economic factors which influence the choice of a process
and the efficiency with which it is to be operated can only be indicated in
general terms. They include reduction of transported bulk and weight,
standardisation of selling grade, and the nice balancing of production cost
against the market values of saleable products. Ore as it is mined can rarely
be smelted direct. Since pyrometallurgy is far more expensive than ore
dressing it must be reserved for material rich enough to bear that expense.
Mineral processing is the first stage of extraction metallurgy, applied to run-of-
mine ore because it provides the cheapest way to discard unwanted material.
Smelting is a later stage. The point at which processing hands over to smelt-
ing is determined by relative treatment cost, partly by the technical aid which
can be given to the smelter by evening the grade supplied it, and partly by
the question of transport. Freight to a distant smelter should only be paid
on what ought to be sent there, and if impurities and moisture are shipped,
it should be because that is cheaper than removing them in the concentrating
plant and because the smelting process can tolerate their presence. Loss in
the smelter varies, among other things, with the amount of slag, and the
more impurity the smelter must remove the more slag it is forced to pro-
duce. Hence pyrometallurgy benefits technically and economically by
receiving ore concentrates which have been processed to an optimum grade of
metal content and a permitted tolerance of undesired associated minerals.
The relative costs of mill and smelter treatment can be assessed with reason-
able precision, and a decision can then be made in each case as to the final
flow-sheet.
A large producer usually operates its own smelter, if labour, fuel, and flux
are available. The value to the mine of its ore is the selling value of the pro-
duct it ships. minus all the costs incurred in getting it to market. In addition
to mining these include concentration. transport and smelting costs. Whether
the mine runs its own smelter or ships concentrate to market, the cost of
smelting is determined by grade and tonnage ofthe mineral treated. Although
smelter schedules seem complicated, they are really technical documents
Mineral Processing-A General Introduction 13
relating (a) to treatment cost and (b) to transport and storage. (a) varies
with the assay of the mill concentrate; deductions covering slag and volatilisa-
tion losses; bonuses for additional desirable constituents; and penalties for
unwanted impurities which add to cost of smelter treatment. (b) affects the
concentrating plant since valueless tonnage shipped increases the freight
charge. As will be seen later, the higher the grade of concentrate the higher
will be the tailing loss, and the greater the treatment cost. These considera-
tions can be balanced out when deciding what should be the optimum plant
performance.
Milling costs include a proportion of capital cost and depreciation which
are affected by the ore reserves (a wasting asset); the nature and complexity
of the selected treatment for which the mill must be equipped; and the
working capacity, which limits the tonnage treated, and hence the rate of
mining, and the allocation of fixed overhead expenditure per ton mined or
milled. This last item, in practice, is apt to lead to quiet but persistent
pressure on the mill to try to exceed its designed and rated capacity. The
equation of all these considerations is fixed in the first instance by the con-
sulting engineers, but once the mill is "run in" and has trained its personnel,
it is common to find that alert management can adapt and modify the treat-
ment if the mine should wish to increase its output of profitable tonnage.
The domestic accountancy of the mill is analysed on some such basis as cost
of each part of the work per ton treated. A textbook is too static to go
into details on such matters, which are richly provided in the current technical
press. Economically, the plant is not isolated, but is intimately bound to the
prosperity of the mine on one side, and of the smelter or other market on the
other. It is a unit in a mass-production process that starts underground and
ends in some purchaser's factory. In its inter-company relationships the mill
is the "customer" of the mine and in turn "sells" to the smelter. If this
relationship is co-operatively understood, smoother running will result and all
concerned will benefit. Nothing so helps assembly-line techniques (and
modern ore treatment is work of that nature) as steady throughput, with clear
directives and objectives governing all mutually interdependent sections of
the plant.
Reference
1. Taggart, A. F. (1945). Handbook of Mineral Dressing. Chapman and Hall.
2. Anon. (1957). Mining Magazine. October.
CHAPTER 2
Preliminary
Run-of-mine ore contains not only rock which has been deliberately
severed for treatment but also a variety of valueless or even embarrasing
adulterants. Past generations of miners used personal craftsmanship to
select ore suitable for milling, so that the'iaborious handling methods of
those days characteristically produced a small but fairly clean tonnage for
treatment. The mining of large low-grade deposits by modern methods
relies on mechanisation and bulk handling, in which careful discrimination
between valuable lode-material and waste country rock is not usually feasible.
With the severed ore may be carried pieces of iron or steel lost from the
machines. These would damage the mill's crushing plant if fed in with the
ore. Wood, unexploded gelignite, lubricating oil, grease, and other unde-
sirable materials also find their way into the run-of-mine ore and should be
removed before they can become run-of-mill feed. Some deposits contain
clays and primary slimes of little or no economic value. These tend to cling
to the working parts of crushing appliances and to impede crisp action and
smooth passage. Slimes may give trouble in the concentrating and disposal
sections if allowed to pass that far.
When the magnitude of the operation justifies it, preliminary rejection of
waste rock and ore too low in grade to justify treatment may be undertaken.
This results in a reduction of the tonnage sent to the mill without the rejection
of an appreciable part of the values mined, and therefore is a concentrating
treatment. The special form of this dealt with in Chapter 12 (dense-media
separation) is an extremely important development of recent years which has
made it possible to up-grade many uneconomic deposits to a point where they
will repay more costly treatment.
"Acceptance" Operations
If the mine keeps records of tonnage dispatched to the mill, and the mill
records the tonnage accepted for treatment, any difference between these
should be brought into account as waste. Discrepancy between the two sets
of figures may arise from accounting error, but could be the result of planned
rejection en route of rock which on inspection in daylight proved too poor to be
treated. It is common to remove such material by methods described later
in this chapter. The result is to raise the head value of the ore accepted for
treatment as compared with that of the ore mined. This is therefore a con-
Mineral Processing-Acceptance into the Mill 15
centrating process, and can be expressed by the basic formula for weight
balance:
(H)
where F is the weight in specified units of feed (the milling term for material
sent to be treated), C the weight of concentrate (the valuable product), and T
the weight of tailing (the valueless rejected fraction). Suppose 1000 short
tons to have been mined and 100 tons of this to have been rejected as waste.
Then
and the tonnage accepted for treatment is 900. Useful as this figure is in
showing what quantity of the Company's money is spent in milling costs,
equation 2.1 is insufficient for control purposes. An "ingredient balance" is
essential to check the efficiency and justification of the process of rejection,
or sorting, to use the milling term. This, in equation form, is:
F/=Cc + Tt (2.3)
where /, c, and t are the assay values of feed, concentrate, and tailing respec-
tively, in appropriate units. In the case under consideration let the assays be
f=2'0% c=2'2% and t=0·2%. Then
The items which comprise this equation are not all measured in ordinary
working. Those which are usually checked are marked with an asterisk(*).
Although this preliminary operation has done but little to raise the grade of
the ore, it has been checked by an equation of metallurgical balance (2.3) and
the information yielded could be summarised as a metallurgical balance sheet.
TABLE 3
Sampling
Ore sampling falls into two categories, exploratory and control. When
prospecting for, proving, and developing a mine, the former type is used. In
routine exploitation of the ore and control of the mill operation, samples are
taken in order to check the efficiency of the work. This last type of sampling
is the main concern in mineral dressing. In a given sampling system the
quality of the material at the point under scrutiny is checked with sufficient
accuracy for the immediate purpose. This may include check of lump size,
moisture content, assay grade, purity, chemical state and other matters dealt
with later. A sampling nomogram has been constructed by P. Gy' which
indicates the best reduction treatment under correctly stated conditions. These
take account of the liberation mesh of the valuable ore constituents, the assay
grades and the densities of the minerals present.
Despite the obvious importance of good sampling as a means of technical
and economic control, it is on the whole one of the least satisfactory opera-
tions in milling. The physical difficulties involved in taking a truly represen-
tative sample and reducing it to the few pounds weight needed for assay are
formidable, and the chances that such a final sample will accurately represent
the thousands of tons from which it was drawn are best when finely dissemi-
nated ore of fairly even grade is being smoothly carried through the trans-
porting system. When, as in many gold ores, much of the value exists as
tiny scattered particles, or in irregularly distributed rich patches of ore, some
experienced workers consider the chances of accuracy to be so remote that
close daily check of recovery based on the entering tonnage is not possible.
In this section ideal practice will be considered.
The control sample cut from the original feed is called the head sample.
It may be taken by hand or mechanically, according to the way in which ore
comes to the crushing plant. An elementary method is to make a "grab"
sample, in which a small portion is taken from the contents of each passing
car-load of ore. Since the valuable constituents are frequently more friable
than the gangue, there is a risk in such cases that the richer ore has worked
Mineral Processing-Acceptance into the Mill 17
down out of reach, so that the sample thus taken from the top will not be
representative. With even distribution of values in a disseminated type of
ore this should not occur to a serious extent, but there are other objections to
hand sampling. If it is to provide reliable data for technical control it should
be foolproof. To take an adequate sample by hand and to reduce it accurately
to assay dimensions is work calling for conscientiousness, hard labour, and
monotonous repetition. Workmen combining such virtues are rare, and
would be better employed on other duties than routine hand sampling.
Methods of hand sampling are described later.
With mechanical sampling, the uncertainty of human partiality can be
avoided. In the ideal layout a point is chosen or contrived where the whole
ore-stream is falling freely under gravity. The sampling device cuts through
this falling stream at regular intervals and removes a percentage of ore
commensurate with (a) the importance of accuracy and (b) the facilities for
reducing a large sample to manageable size. The whole stream should be
cut for part of the time so as to remove a representative fraction from its
cated cones joined at their bases are mounted vertically and rotate so that at
each revolution an attached scoop (a) which projects from the upper cone cuts
through the falling stream of ore (b) and withdraws a sample proportional to
the opening in the sector of that cone. The scoop edges must be truly
radial and its sides vertical in order that an accurate cross-section of the ore
stream may be taken, and the sampler must run smoothly. A jerky motion
would result in an unrepresentative cut. For a 5 % cut the scoop must sub-
tend 18° of arc; for 10% 36°, and so on. Alternatively, two diametrically
opposite scoops may be used. The former arrangement is better since it
allows a larger aperture and reduces the risk of a piece of rock "bridging" the
scoop and thus deflecting into the main stream ore which should be included
in the sample. Several other stream-cutting samplers are standardised
manufacturing products. These samplers can only be inserted where there
is an appreciable fall of rock. They cannot work satisfactorily on feed
containing large lumps.
Straight-line cutters take several forms and can deal with a wide size-range
in the feed, beside taking but little headroom. One form consists of a samp-
ling conveyor set under and across the delivery from the main conveyor.
Two sprocket chains carry the sampling bucket, and normally the ore falls
between them. Periodically the bucket traverses the stream and cuts a
sample, which is tipped as the bucket turns round at the end of its journey and
Mineral Processing-Acceptance into the Mill 19
proceeds back, bottom up, through the stream of ore. The sample taken is a
proportion of the tonnage, being the total bucket area on the sprockets
divided by the area between the leading edges of successive buckets. At
Midvale, Utah, a 2 %cut of the lead-zinc ore is thus taken, and dumped to a
hopper. This discharges its contents in such a manner that a further cut
is made by the buckets, thus reducing the sample cut to I :2500 in one
mechanised operation. Typical of straight-line samplers are the Geco and
Geary-Jennings, which use chain-drive and screw-drive respectively to recipro-
cate a sample cutting device across the falling stream of ore. One mine uses a
pusher mechanism to "wipe" a sample sideways off the travelling conveyor-
belt at intervals.
Usually the mined ore contains lumps too large to be sampled before they
have passed through the crushing plant. This work is then done in the
conveyor system ahead of the fine-ore bins. If sorting has been practised
before this point, the fraction thus removed should be weighed and sampled
in order to check the efficiency of the operation and reconcile the mined
tonnage with that accepted for milling. This practice would give no protection
against "high-grading"* on the picking belts in the case of a spottily rich gold
ore. The closer the size range of the feed at the time the sample is cut, the
more representative it will be, since a big disproportion between the largest and
smallest lump of mineral is liable to upset accuracy of cut.
The size of the original sample cut from the ore stream should be as large
as can be conveniently handled. Even with a daily throughput of only 100
tons, a I %cut means that one ton of initial sample must be reduced to a few
pounds in weight for assay. This bulk reduction must be made in the simplest
possible manner, subject to working accuracy. The "grain" and distribution
of the minerals in the ore affects both accuracy of reduction and the decision
as to a safe routine procedure in cutting down the bulk. Where all or part of
the values consist of friable sulphide in a tough gangue, there will almost
certainly be a high concentration of value in the "fines". In such a case the
ratio between the weight of the largest lump of ore and the total weight of the
sample must be kept much lower than where a fairly homogeneous rock is
being handled. A 1000:1 ratio would mean that in a one-ton lott no piece
should weigh more than about two pounds. In Chapter 21 a typical procedure
for heap-sample reduction is described.
For routine work, a few empirical tests should be made to establish the
maximum ratio it is wise to tolerate. At the same time the amount which
can be rejected after each size reduction can be experimented with in order
to obtain adequate sampling accuracy with minimum effort. If the daily
original sample is of the order of ten tons, a simple crushing operation
followed by a 10 % cut would bring it down to one ton, which might be
* Where a term is italicised, its definition may be found in the Glossary,
Appendix 1.
t Throughout thi~ book "ton" refers to the short ton (2,000 lb. avoir.)
20 Mineral Processing-Acceptance into the Mill
Objectives in Sampling
The plant can be thought of as if it were a trading concern, buying its raw
material from the mine, processing it, and selling it to a smelter. Seen thus,
operating control is a form of technical accountancy, even though by its
nature it cannot achieve the complete precision of an arithmetical system.
To arrive at a metallurgical balance, the mill must be debited with the total
units of valuable mineral (or minerals) entering as feed accepted for treatment.
It is also credited with the total units of value in its products (concentrates and
tailings), in terms of equation 2.3. Although an exact balance is unattainable,
the closest possible agreement should be attempted. In the case of gold
ores, the high specific gravity of the metal causes it to accumulate in launders,
classifier beds, scoop-boxes and the crevices of such places as mill linings. Re-
tention of values in the circuit cannot be avoided and will nullify any attempt
to strike a short-term metallurgical balance. Over a long period, if there is no
theft, the head assay should be reasonably well accounted for from the known
recovery of gold and from the tailing assay, together with gold tied up in
cyanide solution in the plant. Although it is not always possible to relate the
head assay to the full working control of the mill circuit over a short period,
the information it gives is vital to long-term check on operations. The mining
engineer uses stope assay plans to calculate the value and tonnage of ore
mined and sent to the mill and the mill manager may be asked to account
for the difference between promise and performance. No stope assay sample
can compare in accuracy with an actual proportional cut made through the
moving stream of severed and hoisted ore entering the processing section.
Serious lack of reconciliation between two such sets of figures is prima facie
more likely to be due to technical difficulties in stope sampling than to returns
based on accurate sampling of the mill-head ore.
In the case of the base metal and non-metallic ores, theft is not an element
in the ordinary problem of striking a metallurgical balance. It is important,
however, to have a satisfactory check on what comes from the mine. A head
sample is used for this purpose. Provided it is properly taken, carefully
reduced to assay proportions, and accurately assayed, and that the tonnage
from which it was cut is known, the first essential in operational check-up
Mineral ProceJsing--Acceptance into the Mill 21
applied. In the Merrick Weigh to meter (Fig. 6) a short length of the moving
belt is picked up by a system of rods hanging from weighing levers, which bear
upon a beam held in balance by an iron weight floating in mercury. Move-
ments of this float are recorded on an integrating device which presents the
speed of the belt and the load per unit length in terms of total weight passing
over the weigh bridge. Sticky material on the belt can be compensated
by balancing an equal weight of return belt against loaded belt. The
weightometer can be calibrated either by passing known weights over it, or
by causing a length of heavy roller chain to trail from an anchorage over the
suspended section while the empty belt is running.
Approximate measurement of tonnage can be made by using some form
of revolution counter (such as could be improvised from a cyclometer or car
speedometer) to register the footage travelled by the conveyor, and combining
this with knowledge of the weight of ore resting on the conveyor per foot run.
This weight· is found by stopping the belt, removing the ore resting on a
measured length, and weighing it. This routine check must be made fre-
Mineral Processing-Acceptance into the Mill 23
quently. If ore is being drawn on to the belt from a guillotine form of gate
(Fig. 7), the clearance of the gate above the belt determines the height of the
"ribbon" of ore. If the gate is periodically adjusted in order to regulate the
rate of feed, the weight per foot run of belt must be ascertained for each gate
setting, and must be corrected periodically as the lip of the gate wears away.
The shiftsman in charge must book the time and cyclometer reading whenever
he alters the height of the gate. Another method of recording the footage
travelled by fixed-speed belt is to mount on the framework a portable recorder
of the type used to register the travelling time of a lorry. With this there
must also be an alarm system which gives warning when the belt is running
unloaded. This can be improvised from a trailing arm which is held off the
belt by the ribbon of ore. When the feed fails and the belt runs empty, the
arm falls and an alarm-circuit is closed.
Washing of run-of-mine ore removes obscuring dust and casual dirt from
the surface of the pieces and facilitates recognition when hand picking is to be
practised. A second effect may be the removal of primary slimes of no value.
Clays, colloidal fines, and similar valueless material can easily be taken out at
the earlier stages, but may be difficult to remove once fine crushing has liberated
a substantial quantity of finely shattered mineral. Such primary slime may
at times be a nuisance in the plant. It clings in dry-crushing machines and
interferes with their work. It chokes the apertures in screens, consumes
valuable reagents in flotation, and gives trouble in the thickening and filtration
of finished products because of its reluctance to settle and its tendency to
choke filter-cloths.
Some ores can receive a measure of beneficiation by the gentle disintegra-
tion at scrubbing or washing action. Loosely bound gravels, agglomerates,
and clayey ores containing limonite or manganese nodules are examples. In
the dredging of cassiterite in Malaya the richest tin-bearing gravel often lies
immediately above a softened kaolinised clay, which can embed and carry to
waste good tin if it is not dealt with early in the concentration. The angle o/nip
below which a crushing machine can seize and fracture a piece of rock is
increased when the system becomes slippery, as can happen with slimy ore in
the feed. Where fine material has been weathered down from the true ore
body, the valuable minerals therein may have oxidised to form more or less
soluble salts. It is sometimes desirable to wash such salts away in order to
prevent their entry into chemical reactions in the flotation or cyanide section
of the plant. This can be done while scrubbing the are.
The disintegrating forces used are of a gentle nature, not intended to frac-
ture the are but simply to tumble particle against particle until the desired
amount of cleansing has been accomplished. A water jet, or simple stirring,
may suffice to detach slime from the lump material so that is flows out of the
cleaning system as a dirty effluent while the cleaned are particles leave by
another channel.
In placer-mining the hydraulic monitor, gravel pump, hydraulic nozzle
24 Mineral Processing-Acceptance into the Mill
elevator, and dredge trommel help to disperse the clays they handle. The
interior of the trommel, a cylindrical heavy screen through which the pay-
gravel tumbles and is washed by jets of water, may be fitted with disintegrating
blades rotating about a central axis, which chop up the clay.
In ordinary hard-rock mining, a jet of water from a nozzle may play on ore
at any suitable point in transit for simple washing. More thorough in its
action is the drum washer (Fig. 8). Run-of-mine ore is fed in at one end of a
cylindrical or cylindro-conical shell as it rotates horizontally, and clean water
enters at the opposite end. Lifting plates may be used in the shell to help
raise the ore, which is thoroughly tumbled as it progresses toward the discharge.
There it is elevated by perforated scoops, or raised by a spiral and discharged.
This counter-current arrangement ensures that the entering ore is first washed
by the dirtiest water as the latter runs out. After travelling through the
machine, it receives its final washing from the cleanest newly-entered wash-
water. In various forms, this counter-current principle is used in many
mineral-dressing operations. Gentle agitation together wi.h washing is
sometimes performed on washing screens, which are specially adapted from
the standard screens discussed later and are fitted with spraying arrangements
or do their work in water. Some slime can be removed by washing ore on a
rising conveyor-belt. The belt must be troughed and have provision for
discharging the resulting pulp clear of the pulley mechanism. Removal is
incomplete, but the cleansing of the rock surfaces aids recognition and hand
picking further along the belt. A development from the relatively gentle
action of the drum washer is that of the Telsmith super-scrubber. This
machine rotates faster and gives a vigorous tumbling and rubbing action.
It accepts (in the 96" drum diameter size) material up to 8" in size and can
handle up to 180 tons hourly with a water consumption of 1100 gallons, using
a 125 H.P. drive.
For heavy work, where more strenuous tumbling action is required, the log
washer is widely employed. It is a shearing disintegrator, a tumbling device
and a washer, and can give a measure of classifying action. It consists of an
inclined tank (Fig. 9) in which one or two box girders rotate. These are armed
with stirring blades (about four per foot run), set at a pitch angle of about 65°,
so as to form part of two interrupted spirals. The blades rise inward and
force the ore up-slope against a down-current of water. The raw ore enters
from above, falling into a pool of water in the tank of the log washer, and
pulp laden with slime escapes over a weir at the lower end. The volume of
washing water must be so adjusted as to allow all the wanted ore to be worked
upward against the stream running clown to the discharge end.
Hand Sorting
Hand sorting or hand picking may be defined as removal from the run-of-
mine ore (as hoisted with its associated country rock and detritus) of material
which is not to be sent to the crushing and/or concentrating plant for tWlt-
.oent. The substances thus removed may be valueless gangue or waste rock
not worth milling; unusually rich ore which does not need up-grading by
Mineral Processing-Acceptance into the Mill 25
01/
:' 1
=I~
o •
~
~ :i 'I 0U
cJJ
;:l
....
F-
....
Q)
..::cU
......
'-'
..::::
~
'"
~
:::...
~
J
C:)
...
..
0
I
~
...& 00
"'~
~
"I
I
I
I
the M ill
26 sing-A cceptance into
Mineral Proces
1--'
I
I
I
I Z'
0
,
I (.)
(I)
::l
....
f-
.....
~. ....
qJ
.:::CI1
'-'
~
t
I ~
I a
;
OJ
~
I ~ .....;
I
~
~
I 0.:.
I ~
I
Mineral Processing-Acceptance into the Mill 27
ore-dressing methods; a specific ingredient of the ore which, for some reason,
is not provided for in the concentrating section of the plant; or such mining
detritus as timber, iron, and unexploded dynamite which might damage the
plant or interfere with the concentrating processes. In the sorting of coal
this hand picking was once the main treatment used to grade, to size, and to
remove pyrite, slate, and shale. Sorting alters the assay grade by producing a
richer and a poorer product and is to that extent a concentrating operation.
Since it is usually a minor process which takes advantage of easily contrived
display conditions in the earlier crushing stages, it is discussed at this point.
On some Rand mines up to 30 % of the mined rock is hand sorted. Removal
of such a high percentage is called "hard sorting". With the development of
mechanised methods of dealing with large tonnages, hand sorting has declined
in importance, but in a small plant anxious to produce something saleable by
cheaply contrived, though rather primitive, methods it can form a useful
part of the flow-sheet. With cheap labour or primitive conditions all hand
methods of beneficiation are important, as witness the large proportion of the
world's tin which is concentrated by panning and hand sluicing. Sorting of
some kind is done at most mines, if only ahead of the underground jaw-
crusher whenever its operator removes broken mine timber from the feed.
The mineral to be sorted out must have a sufficiently distinctive colour,
lustre, shape, or general appearance, so that it can be readily recognised and
removed. Since the task is monotonous working conditions should be made
comfortable, with good space and proper lighting. If the ore comes past the
worker in a steady stream, and is displayed in a single layer three feet or so
from the ground to minimise stooping, fatigue is lessened. Too large a lump
of rock is difficult to deal with, and too small pieces take a lot of finding
and removing, thereby slowing down the work. The size limits between
which hand sorting is considered most efficient are a maximum particle dia-
meter of 12" and a minimum of about 2t". Shadowless "daylight" lighting
from a diffusion lamp source is a good alternative to natural daylight. The
minor constituent should be handled by the sorter, and the major one left to
be transported mechanically.
In British and North American practice little or no hand breaking of the ore
is attempted during the sorting operation, so there is no need for the support-
ing surface to act as an anvil. In Latin America and some European and
Asiatic practice the ore may be broken, "cobbed", or "spalled" with hammers
to remove a desired portion. The most elementary arrangement is a level
floor where dumped ore can be picked and cobbed, the separated fractions
then being shovelled or swept away. Moving surfaces which bring the work
to the picker make for higher efficiency and bigger throughput. A pan
conveyor withstands shock loading and can handle large rocks. For all-
round convenience, the belt conveyor is the most adaptable arrangement and
is in wide use. Operators can stand or sit on one or both sides, from three
to six feet apart, and push or pull rock off the passing belt into appropriate
chutes. A wide and flat belt is best, fed so as to be clear of material for a few
inches from the sides. It can run up to a speed of 60 feet/minute. If the
feed is moist from washing, the belt should slope to aid drainage. An
advantage of belt sorting is that it can be combined with transport and eleva-
28 Mineral ProcessiIlK--Acceptaf1ce into the Mill
tion of the ore, thus giving this operating service at little beyond the cost of
the labour required to remove and dispose of the products. The revolving
sorting table has fallen into disuse, as its sole advantage over belt sorting is
centrali~ation of supervision when an unusually valuable material is being
handled. Youths, women, and convalescent labour are commonly employed
in sorting.
Consideration must be given to the relative cost of milling run-of-mine and
sorted ore, and the variation in total recovery with and without such sorting.
From time to time the rejected fraction should be assayed, to ensure that no
milling-grade ore is being lost, and the weight removed should be recorded
and brought into account when calculating the accepted tonnage. Hand
methods of sorting are sometimes assisted by the use of fluorescent light,
which cause some minerals to glow and thus reveal their presence. Geiger-
Muller tubes are also used, suspended above the picking belt so as to signal
the arrival of a piece of radio-active ore.
When the tonnage to be handled justifies the cost and the difference in
density between payable ore and that too poor to mill is sufficient, the dense-
media process (OMS) has replaced hand sorting. It can work cheaply on
a wide range of sizes.
Mechanized Sorting
(g) conductivity
(h) specific gravity
(i) reflection and refraction of a light beam
(j) light sorption
(k) radio-activity, natural and induced
(/) ferro-magnetism
These characteristics must be considered conjointly with those listed above
(a-f). For example, the quantity of radiation emanating from a particle
weighing, say, 10 grammes might be insufficient to activate an ejecting mechan-
ism which would act on a 20 gramme piece of the same mineral, with twice
the signal strength. Conductivity measurement can be made by allowing
suspended wires to brush the surface of passing lumps of rock. Density
differences are exploited in the OMS process (Chapter 12). A beam of light
can detect specific colour, lustre or reflecting angle, the last being used to
sort diamonds. Ferro-magnetism is considered in Chapter 22, and in the
next section of this chapter. Short-life radio-activity can be induced in some
minerals.
Although mechanized sorting-apart from the OMS process- is a new-
comer to mineral processing, the following industrial applications are now
well established. The La Pointe picker was developed in connexion with the
sorting of moderately coarse radio-active concentrates, prepared by jigging,
during which they are washed and sized, so that there is no random activation
by adherent radio-active dust. A single line of particles moves on a narrow
belt, and passes below a Geiger-Muller tube. If radio-active emission
exceeds a set level a sorting device is operated, which removes the signalling
particle from the moving line.
Another arrangement triggered by the particle's own radio-activity is the
K. & H. equipment. Run-of-mine ore of between 2" and 8" lump size is
washed and delivered piece by piece, so as to fall past three successive scanning
devices. The first is a neon light working on A.C. at high frequency, so
arranged that its beam falls on a photo-electric cell save when interrupted by
the falling particle. This interruption, by measuring the transit time, notes
the size of the lump. Next, the particle passes a scintillometer, which deter-
mines whether it is above or below the pre-set level of radio-activity for this
size of particle. Finally, it falls past a series of air jets which are electronically
controlled in accordance with the signal received from the previous scanners.
Pay ore continues to fall unhindered, but a piece which is below millhead
grade is blown sideways as an air blast operates, and falls to a waste con-
veyor. One of these units treated between 25 and 50 tons of Rand gold ore
hourly in a demonstration run. I n many of the banket ores there is a fairly
consistent ratio of gold to uranium, and this can be used to signal low-
uranium (and therefore low-grade gold-ore and waste rock) out of the feed.
At Eldorado in Canada, ore larger than 3" is similarly sorted.
Although iron and ste I which have inadvertently found their way into the
30 Mineral Processing-Acceptance into the Mill
ore stream may be removed by hand during such operations as have just been
described, more positive protection against their entry is usually made. If a
piece of steel enters a crushing machine damage may result and production
may be held up while repairs are made. "Tramp iron" is normally magnetic,
and can be removed by hanging an electro-magnet over the belt conveyor or
other suitable point. From time to time the magnet is swung to one side
and unloaded. This arrangement may fail to catch small items such as nails
and bolts which are held down by pieces of ore, so belt conveyors are sometimes
furnished with an electro-magnetic head pulley supplied with current from
a D.C. source. Iron thus held next to the belt is not discharged at the
delivery end of the conveyor but is carried round and dropped well clear of the
ore stream, after it has left the zone of magnetic attraction. These devices
are discussed in the chapter on magnetism.
Another protective device is the search coil. This is a sensitive device
which can detect a slight concentration of magnetic flux such as arises when a
feebly magnetic material passes through its coil. Some alloy steels used in
the mine are non-magnetic, and there is always the further danger that a piece
of metal may be pinned down or be too heavy to move. One mine uses three
search coils on its feed conveyors to deal with this possibility. When a piece
of iron or other metal passes the first of these coils, a current is generated
which can be amplified and caused to trip the motor of the feeder supplying
ore to the conveyor belt. The second search coil detects the same piece of
metal and stops the conveyor, the distance between these coils being such that
the crusher has worked itself empty before the conveyor stops. Thus the
inconvenience of a build-up of ore on a stationary conveyor is avoided. The
third search coil is an extra, which would operate in the event of either of the
first two failing, or if the shiftsman failed to remove all the iron. Normally,
he locates the tramp iron, picks it off, and then restarts the belt. In another
type of induced-current detector, flour is dropped on the belt as an alarm
sounds and the driving motor trips out, so that if the appliance continues a
short distance before stopping the trouble can be quickly located. The use of
guard magnets and hand picking is sometimes combined.
Reference
Gy, P. Minerais et Metaux. Paris
CHAPTER 3
PRIMARY CRUSHING
Preliminary
Crushing, grinding and other words or phrases associated with the size-
reduction of ore and other rock are all comprehended in the word "commi-
nution". This (Truscott) is "the whole operation of reducing the crude ore
to the fineness necessary for mechanical separation, or for metallurgical treat-
ment ... '" It is usual to make an arbitrary division of comminution into con-
venient stages. Primary crushing brings run-of-mine ore down to a maximum
size of the order 4" to 6" in average diameter; secondary crushing receives
feed at -6" and reduces it to below tn. "Dry" crushing includes work on
ore as mined, which may be somewhat moist when delivered .. It is succeeded
by comminution in water, arbitrarily called "grinding". Although a con-
siderable amount of fine grinding is done by dry methods, this book follows
usage by reserving the word "crushing" for an operation predominantly
dry and "grinding" for work on a suspension of ore particles in water. One
important difference between dry and wet comminution lies in the mode of
seizure of the particle. In the former case the particle is large enough to be
gripped between two solid steel members as they are pressed together by
mechanical forces. One or both of these members moves to and from in a
fixed-path cycle. The rock gravitating through the rapidly expanding and
contracting gap thus produced is nipped and crushed. In wet grinding the
bulk of the ore is already too finely divided for a particle to be seized in this
manner. It is therefore exposed while more or less free to move, to random
blows. There are exceptions to this generalisation.
Machines used in dry crushing must work in dusty conditions, even when
the main cause (escape of fine particles at transfer points) is dealt with. They
are usually worked intermittently, to fit in with the hoisting and delivery
programme of the mine. On completion of the dry crushing their product is
delivered to bulk storage (in the mill's fine ore bins). From these it is de-
livered at a controlled rate to the more continuous grinding and concentrating
processes.
The following classification of basic reduction steps is proposed by Hukki.·
Explosive shattering From infinite size to minus 1 metre
Primary crushing minus 1 m to minus 100 mm
Secondary crushing 100 mm" JO mm
Coarse grinding JO mm " 1 mm
Fine grinding 1 mm" lOOf1-
Very fine grinding 100 f1- JO f1-
Superfine grinding 10 f1- I f1-
increases with increased fineness of the ore being treated. From experimental
evidence Hukki suggests an apportionment of this power consumption of
0·35 kWh/ton in primary crushing of a brittle solid, rising to 0·6 in the
secondary crushing, 1·6 in coarse grinding and 10 kW/h in fine grinding in the
ratios stated above. These, in his view, suggest an important change in the
use made of the applied power through these stages. At the primary crushing
level, the results correspond statistically with the requirements of Kick's Law.
This phase is followed by approximate agreement with Rittinger's Law in the
intermediate stage, while Bond's formulation becomes increasingly good at the
finer end of comminution. These laws are discussed below. The enormous
rise in power consumption when a product well below a micron (p.) in size is
aimed at more or less rules out the use of standard grinding techniques on the
score of cost.
Crushing
Crushing Theory
The forces used to produce fracture of a perfect crystal are of two main
types. The structure is bound together by its inter-atomic forces of attraction.
Stress as considerable as 10" p.s.i. (pounds per square inch) is required to
disrupt this bonding or "theoretical strength", which can be calculated.
If tension is applied the crystal stretches elastically until it reaches its yield
point. and recovers if the stress is removed before this point is reached.
Once the elastic limit is exceeded a flaw is produced, usually in the form of a
minute crack. which becomes a focus for incipient fracture. In his classic
paper Griffiths' noted that the stress which the crystal can thenceforward
withstand is inversely proportional to the square root of the length of the
crack. This means that it will now fracture at a much diminished stress
loading. During the stressing which created the imperfection, work was done
34 Mineral Processing-Primary Crushing
to overcome the mutual bonding of the inter-atomic forces. This work was
stored as elastic energy, and was released as the atoms returned to their normal
positions. Since a crack existed, the atoms in its vicinity were able to shed
elastic energy while the crystal was being stressed. If this released energy was
sufficient to overcome the weakened inter-atomic bonds at the tips of the
crack, it grew rapidly (at a speed of about 15,000 ft/sec). The fact that stress
of the atomic bonding is focussed at crack-tips has been proved by polarised
light studies of plastic materials.
Such a crack in solid material may start as a scratch or a surface blemish (a
superficial discontinuity), as a minute fissure in a crystallite structure, or as a
defect in the atomic lattice of a crystal grain which, by yielding, permits the
start of plastic flow. Most rocks and man-made materials contain such foci
of weakness, so that practical strength falls far short of theoretical. Ideal
glass should withstand stresses up to 10" p.s.i., but in practice failure occurs
at about one ton p.s.i.
Where dislocations exist, crystals which would otherwise deform plastically
as one plane slides freely over another, have this plasticity blocked at each
such dislocation. Stress builds up at this point and the surrounding atomic
bonds are ruptured. The result is a minute crack which, given a little rela-
tively light further stressing, becomes a complete fracture. Once crack
propagation begins it proceeds at nearly the speed of sound. A corollary is
that once a crack is running at this speed, no further stress applied to the
main structure can catch up with the advancing fracture.
It is thus clear that two kinds of stress can operate-the reversible, which
is removed if the crystal is not loaded to its elastic limit, and the irreversible in
which surface discontinuities are formed or plastic deformation occurs. This
last is the result of the slipping of one plane of atoms over another. The
slip is local and so produces a deformed area, with one side in compression
and the other in tension. Energy is thus stored ready for use in further
deformation. Similar storage builds up in zones surrounding dislocations
which prevent the free sliding of one atomic plane over another. Naturally
occurring rocks contain random focal points where stresses due to defects are
stored till an initial crack is produced by a relatively small further stress.
Further weakening factors are the explosive shock during blasting of ore and
its chemical oxidation between ~everance and milling.
In a piece of ore at least two mineral species are inter-crystallised in various
patterns. The situation is far more irregular than that in the perfect and
completely pure crystal with which this discussion commenced. The prac-
tical application of crushing theory must therefore add statistical methods of
testing and an empirical approach to the considerations outlined above. No
two lumps of ore are precisely similar. It is common to think in terms of
such varieties of disruptive force as compression, tension, shear, torsion,
abrasion and shatter. The last four are compound forms of the first two.
When a beam suspended at its ends is centrally loaded till it bends, the lower
part is in tension while the upper part is in compression. When seizure of a
piece of rock occurs in such manner that the seizing forces move in opposing
directions wAile the rock is prevented from rolling, tensile shear predominates,
with local compression where its high points are gripped. Most crushing
Mineral Processing-Primary Crushing 35
Er =a (3.1)
Er = b
(3.2)
where a and b are constants and do the original particle size. If d p is the
product size, the Kick and Rittinger hypotheses can be reconciled by Dobie's'
equation:
(3.3)
Something of the difficulty facing the research worker in the field of commi-
nution may be appreciated when it is recalled that four types of binding force
are recognised in fundamental physics. Taking nuclear force (that binding
a proton to a neutron) as unity then electrostatic force (proton to electron) is
10-', nucleon decay force (emission of f3 particle) 10-1 I, and gravitational
attraction indefinitely weak. Nothing is at present known of the balance
between these forces which must be upset for cleavage to occur. Crystal
study shows. for a homogeneous material, three types of imperfection which
affect resistance to shear. Micro-defects are lattice imperfections due to
36 Mineral PI'(}cessiIlK--Priman' CrushiflK
"The total work useful in breakage which has been applied to a stated
weight of homogeneous broken material is invariably proportional to the
square root of the diameter of the product particles."
This is formulated thus:
"When F is the diameter that 80 % of the feed passes, P is the diameter that
80% of the product passes, and W is the work input in kw.-hr. per ton; then
WI, the work input required to reduce from an infinite feed size, is
WI ~ W J
[ \ r,;:F--VFIn'
vP
(3.4)
When F and P are in microns and Wi is the work index, or total kw.-hr. per
ton required to reduce from infinite size to 80 % passing 100 microns, or to
r/F--v=Jg
- ..
approximately 65 o/~ passing 200 mesh,
Wi=W
vlf: (3.5)
P 100
The Rittinger theory deals with measurement of surface areas, the Kick
theory deals with volumes of product particles, and Bond's "Third Theory"
with crack lengths formed. None of these theories can be regarded as
adequate, though to some extent they are complementary. A present weak-
ness of the "third theory" is that it considers only the linear dimension of a sur-
face rupture, whereas this is part of a new area. Against this must be set the
fact that a comprehensive set of Work Index figures has been prepared as a
result of extended tests on a variety of ores, and as its author says:
"The Third Theory and its work index permit closer predictions and more
accurate comparisons of all crushing and grinding installations than have
heretofore been possible and should result in more efficient operations." A
step towards reconciling these three theories has been taken by Holmes. H He
points out that the statement for new surfaces
fails to consider the energy absorbed during the process of elastic deformation.
The initial size of the particle undergoing deformation and fracture must be a
dominant factor. Kick's law, though applicable to homogeneous and
annealed material, is not valid for non-homogeneous rock including
various weakness zones. Bond's theory, being developed from much
testwork, provides an excellent basis for calculating the relative performances of
crushing systems. The "work index" values, however, are not constant for
fixed conditions of reduction, and the theory has the limits of empiricism.
Holmes propounds a modified form of Kick's law. suited to ore, in the
equation
B kD' r 0.6)
where Wi is the work index and r the "Kick's Law Deviation Exponent"
which expresses the degree of variation of particle strength with variation in
size, specific to the material and the mode of stress application.
In a discussion of these hypotheses Schumann" summarises his own con-
clusions thus: " ... readers may find it helpful to think of comminution from
the following point of view: Any actual comminution product may be re-
garded as a mixture of "finished" comminuted material and coarse material
not yet completely comminuted. Any comminution process simply converts
the coarse material into finished material but does not change the size distri-
bution or average size of the finished material itself. Only the relative
proportions of coarse material and finished material are changed, and it is
these changes in relative amounts of coarse and fine material that account for
the continually decreasing average particle size of the overall product during
comminution". This would seem as good a generalisation as can be made at
our present level of understanding of this complex subject.
If the ore sent from the mine is reasonably coarse, dry and free from slip-
pery and clayey matter, it can be gripped by fixed-path crushing devices and
the resultant fragments (or progeny) can gravitate rapidly through the system.
In most operations of any size there are two crushing stages or more, on the
general plan outlined in Fig. 16. In the first, or primary, section the run-of-
mine ore (which has been reduced to an agreed maximum size before dispatch)
is crushed to an agreed transfer size, at which it must pass through a screen to
the secondary system. This is usually between 4" and 6" maximum particle
size. It is determined by the set of the primary crusher. There is a working
relationship between the size a machine can accept (its gape) and the maximum
size which will pass through the discharge end-its set. This relationship is
called the reduction ratio. It is governed by the stresses and strains imposed
40 Mineral Processing--Primary Crushing
on the crusher during passage of feed through its crushing throat. The
quantity of rock being crushed at a given moment, the progeny created at
each nip, the percentage of voids between the particles, and the toughness
of the ore are main determinants of a safe reduction ratio for a given type
of machine.
To understand this clearly, consider a theoretical 24" cube of rock which
is to be crushed neatly into either 4" or 6" cubes (measured as length of
one side). The original surface is 24 sq. inches on each of the six sides,
,or 3,456 sq. inches. On reduction to 64 cubes 6" square the new area createcl
is 10,368"". On reduction of a similar cube to 216 cubes each 4" square the
new area is \7,280"2. The total crushing work in the two cases is in the ratio
1.7/1. Since the driving motor can only deliver a fixed maximum thrust
and only part of this force is able to create new surface, the actual crushing
ratio would be much higher than I.7 to I and the machine would either stall
on overload, or something would give way. Though reduction ratio is
conveniently stated as ratio of gape to set, the basic consideration is the
development of new surface. No known me~hod can evaluate this, as part
of the surface (e.g. internal fissuring) is invisible and, in any case, original
lumps and their progeny are of all shapes and sizes. The percentage of input
force ending up as new surface is probably much higher in the primary sec-
tion than later, but at present an empiric approach based on practical ex-
perience governs manfacture and use of equipment.
Primary crushing machines are of two main types-jaw and gyratory.
The largest can accept rock up to 84" by 60". At the discharge end the setting
or "set" is adjusted in ratio to the biggest piece the crusher is to receive,
usually from 4 to I up to 7 to I. The latter would limit a machine with a 6"
set to feed smaller than 42" tri-dimensionally. The bulk of the ore leaving a
typical underground stope is well below such an awkward size and weight,
which would be unwieldy in most mechani7ed handling systems. A 36" cube
of rock weighs over two tons. If through inadvertence tralllp ol'er.l'ize
(lumps too large to be seized by the crusher) arrives it obstructs the feed
arrangements. Such material can be arrested on a grid, made of old mine
rails flange up and suitably spaced, and be periodically broken down with a
sledge hammer. Primary hands over to secondary crushing somewhere
between lIIil/liS 6" and 4". In opencast mining, where the problem of handling
huge lumps of rock is not complicated by the need to get it through a mine
tunnel and shaft, larger maximum sizes may have to be dealt with.
PRIMARY CRUSH~.RS
Jaw Crushers l Feed range from 84" x 60" down.
Gyratory Crushers f Set, down to 4".
SECONDARY CRUSHERS
Gyratory Crushers I
Jaw Crushers ~ Feed from 6" down.
Cone Crushers JSet, down to t".
Hammer Mills
Rolls
Mineral Processing-Primary Crushing 41
FINE CRUSHERS
Rolls
Dry Ball Mills
Hammer Mills Feed from about I" down.
Disc Mills
Buhr Mills
Ring Mills Feed from about i" down.
Disruption by high pressure gases. Feed already fine
Explosive shatter}
Fire-setting From the solid rock.
At suitable stages, hand-operated hammering, pestle-and-mortar, etc.
Crushers which handle ore intermediate in size between some 4" and -!o',
using added water in their work, include:
Stamps. Receive wide size range (up to _2H), deliver down to -60 mesh
size.
Rod Mills. Rective -t", deliver between -14 mesh and -18 mesh.
:..,;--....,..---~ 7
8
Feed
,I
6 ~ -l
Dlstharge
Jaw Crushers
The Blake Crusher was patented by E.W. Blake in 1858. It was soon
improved to the final form in which the entering feed received the least and
the departing product the greatest crushing movement. Variations in detail
on this basic form are embodied in the bulk of the jaw crushers offered by
manufacturers today.
The functional drawing (Fig. 10) shows the essential details and Fig. II
gives a typical cross-section. Parts (2), (3), (4) and (5) form a loose linkage
which oscillates with a compound eccentric movement in the fixed framework
(I), (6), (7) and (8). The size of the largest escaping particle is governed by
the set, which is the horizontal distance between (6) and the tip of (5) when
at the widest opening. The back toggle plate (2) pivots loosely from a bearing
in (I). It is oscillated radially by the pitman (3), which is driven by the eccen-
tric (8). As this toggle rises it presses the lower end of the pitman forward, a
movement transmitted via the front toggle (4) to the swing jaw (5). The
horizontal displacement is greatest at the bottom of the stroke and diminishes
steadily through the rising half of the pitman's cycle. Thus, though the
driving force applied through the eccentric does not vary, the horizontal travel
of the swing jaw diminishes rapidly. Crushing force is least at the start of the
rising half-cycle when the angle between the toggles is most acute, and is
strongest at the top, when full power is being delivered over a reduced travel of
the jaw. Since the jaw (5) is pivoted from above, it moves a minimum distance
at the point where a large lump of ore has newly entered and a maximum
distance at the discharge end.
Consider a large piece of ore falling into the feed end or "gape" of the jaw
crusher. The swing jaw is moving to and fro at a rate depending on the size
of the machine and of the material it must crush (see Table 4). The running
speed should not be so high as to strain the moving parts, which must with-
stand reciprocal action, severe loading on the compression stroke and sudden
release on the return. It must also give time for rock broken at each "bite"
of the jaw to fall to a new position in the constricting space of the crusher
throat. The piece of ore falls till it is arrested, either above other ore, or by
TABLE 4
DATA ON BLAKE CRUSHERS
Reduction Loss of
Gape Set Ratio Tons/hr. R.P.M. H.P. Head
7" x 10" 3"
9·3 2 300 8 2' 6"
"
It" 4·7 4 300
being nipped between the fixed and swinging jaw. Within a fraction of a
second the moving jaw again closes on it, fast at first and then more slowly
but with increasing power to the end of the stroke. Though the jaw only
squeezes the ore for a short distance of its movement, this suffices to break the
big lump. The fragments now fall to a new arrest point where they find
themselves somewhat crowded, since the total cross-section is now less, while
the overall volume has been swollen by newly created voids. At this arrest
point another squeeze is delivered, this time with greater amplitude, since the
radius of the swing jaw from its centre has increased. Crushing continues
stroke after stroke until the crushed particles reach the lower end and fall
clear. At each arrested fall the crowding together of the fragments would
increase, owing to the combined effect of the increase in voids and the decrease
in cross-section. were it not for the steady increase in amplitude of swing.
44 Mineral Procl'.I'.linK Primary Cl'lI.I'hinK
This accelerates the discharge of finished material, which works down and out
at a rate sufficient to leave space for material arriving from above. Crushing
under these conditions, in which particles are relatively free to fall between
successive squeezes, is termed "arrested" in contradistinction to "choked"
crushing, in which the volume of material arriving at a given cross-section
would be greater than that leaving it if the rate of feeding was unrestricted.
In arrested crushing the main force exerted upon a particle is directly applied
by the jaws of the machine. In choked crushing a substantial amount of
comminution results from impact of particle upon particle. The character
of the product is different. Since in arrested crushing any particle small
enough can escape at the discharge area, much of the crushed rock is finally
delivered at a fairly coarse size. In choked crushing comminution continues
even when particles are smaller than the "set". The difference is analogous
to that of the orderly departure of a theatre crowd and of a panic in which
people are crushed by other bodies arriving at the exits at too great a rate.
The "set" of the Blake is the maximum opening between the jaw~ at the
bottom, measured with the "V" of the toggles at the steepest point, with the
eccentric full down. It is adjusted by using toggle plates of the desired
length. Wear is taken up when required by adjusting the back pillow on
which the end of the toggle bears. Since the toggles are loose in their sockets,
a tension rod is used to hold the system together and to aid the return stroke
of the swing jaw. A vertical spring may also be used to preserve smooth
Contact of the eccentric by acting upon the bottom of the pitman.
Arrested crushing can only take place if the rock broken during each
nipping stroke falls with reasonable fre<::dom during the return half of the
jaw's swing. Since the overall volume swells with each stroke (owing to
newly created voids), yet must drop into a decreasing horizontal cross-
section, there would be congestion if the rate of fall was not steadily accelerated
on the journey downward. This is made possible by a proper inter-relation
of the following factors:
Variations from the simple Blake crusher are numerous. They fall into
two main divisions:
(a) Variations in application of toggle motion
(b) Variations in slope of crusher throat
In the Telsmith crusher the movement is transmitted directly from an
eccentric to the moving jaw. There are no toggles and the need for imparting
reciprocating movement to a heavy pitman is avoided.
In the single-toggle crusher the swing jaw is hung on an eccentric and the
whole of its surface is in lateral and vertical motion. In another crusher both
plates move.
The effect of using curved jaws instead of straight ones is shown in Fig. 12.
The wearing plates which line the crusher throat are frequently cast with
vertical wedge-shaped corrugations, so as to impart a measure of beam-
loading to the applied force. Of the alloys used for these plates, manganese
steel is most favoured .
This crusher (Fig. 13) reverses the jaw action of the Blake, in that it applies
the maximum movement to the largest piece and the minimum to the smallest.
The fulcrum is below, and only slight variation of set occurs as the moving
jaw advances and recedes. Hence, for laboratory purposes where throughput
is less important than close control of rock sizes, such a crusher, if lightly fed,
can be made to do arrested-crushing work. If it is choke-fed, the choking
becomes serious as the cross-section decreases. Rock crushes rock and
there is not enough dilation as the moving jaw recedes to expedite departure
of the finished material. This results in over-crushing, in which the crusher
does work that would be better handled by grinding mills. and itself suffers
Mineral Processing-Primary Crushing 47
severe strains in doing it. The practical result is that the Dodge cannot be
built for heavy duty, and is rarely, if ever, incorporated in a flow-sheet.
Gyratory Crushers
I
I
I
I
I
I
I
I
I
I
I
h Eccentric
~ Sleeve
L II)
drop of under two feet and an output of half a ton hourly for 4 h.p., while
the largest model can receive 5-foot rock and reduce it to 12" or less. These
giants weigh up to 700 tons, need 500 h.p. and can drop over 3000 tons
hourly through their 32 feet of machine height when running at 175 r.p.m.
The throw, determined by the eccentriC design, is less with the gyratory than
with the jaw crusher, and is made greater for soft, tough rock than for brittle
50 Mineral Processing-Primary Crushing
ore. Somewhat different in its application of the breaking force is the Tel-
smith crusher. Here, instead of a spindle which sweeps out an acute cone,
the whole spindle remains vertical in its eccentric housing, thus applying the
same amount of lateral movement all the way down the breaking head.
First cost, installation, and ease of feeding favour the gyratory crusher;
which has not the problem of reciprocation of heavy and irregular loading
stresses to meet in such an acute form as with the Blake. In jaw crushers
half of each revolution is heavily stressed and half relaxed. This stress
alternation demands rugged design, fatigue-resistant metallurgy and con-
struction, and sturdy mill foundations. A crushing bowl can receive its
feed from any direction and work with the top buried in waiting rock. Power
consumption. whether idling, lightly loaded, or under full load, also favours
the gyratory. Maintenance is heavier than with the Blake, and where
transport is difficult. the machine is less easy to sectionalise. With clayey
material the greater amplitude of motion of the jaw crusher is advantageous
and its set can also be adapted more readily when there is a question of
adjusting the amount of size reduction done between the primary and secon-
dary crushing system. The most telling point in favour of the jaw crusher
where large tonnages are not in question is that its gape allows it to handle
more awkward oversize than the gyratory, the biggest machine, with an open-
ing 84" by 120" taking a 6-foot rock. A gyratory capable of handling this
would have a far higher tonnage capacity. With both types reduction ratio
varies in practice between 4: I and 7: I and tends to be raised as improved
engineering skill, metallurgy. and lubrication permit higher working
stresses.
The run-of-mine ore delivered to the crushing plant contains rock of all
sizes. A substantial proportion is already well below the set of the crusher,
and if fed to it may produce certain disadvantages. The first is that the
space in the crusher which could be occupied by ore in need of treatment is
taken up by material already undersized. If this is by-passed (Fig. 16, line (5)
to (7) from (- », wear is reduced and capacity increased. The second is
that it is cheaper to keep undersize outside the machine than to pass it through,
as this would cause wear and extra dustiness. The third is that undersize
tends to pack the voids between the big rocks and thus gives rise to choked
crushing. The fourth is that any natural stickiness in the ore tending to cause
hold-up in the crusher will be still more of a nuisance in the presence of
undersize. Against these considerations is one which is always in the mind
of the alert mill manager. Every manipulation introduced into the f1ow-
sheet is an added liability, and everyone which can sensibly be avoided tends
to simplify operations. In a mill where the crushing plant is more than
adequate for the tonnage treated, it will probably be better to send all the
milling ore straight through the primary crushing section. If, however, the
operation is so small that all the crushing is done in one machine, it is almost
certain that the undersize should be removed first so that the available force
will be applied to the coarser material.
The larger the tonnage treated, the more important it becomes to break
down the work of size reduction into definite stages and to use machines set
in series so that each deals with one limited range of size reduction. Thus a
jaw crusher might be handling all the primary work, and substantial relief
would be obtained by by-passing undersize. If ajaw crusher were used chiefly
to deal with the very big material and to provide more efficient feed to a
following gyratory, the by-passing might be done between the two crushing
machines. Layout with respect to subsequent handling problems would be
an inlportant factor. A typical layout is shown in Fig. 16.
The surge bin (2) receives dumped loads from skips, lorries, or mine cars,
and has enough storage capacity to maintain a steady feed through (3) to (4).
Some hand-picking may be possible on (4). The grizzly at (5) is spaced in
accordance with the ratio of reduction required. If no individual piece of
mined ore exceeds 24" in average diameter (or two-thirds of a ton in weight)
and a four-to-one reduction ratio is used, then the primary crusher (6) will be
set at 6" to give this ratio. If ore is delivered at - 20" and a five-to-one ratio
is satisfactory, then (5) and (6) can be set at 4". The screen (7) will similarly
be chosen to correspond in aperture with the set of (II), having regard to the
reduction ratio. Item (8) is required to establish the metallurgical balance
52 Mineral Processinf(---- Primary Crushillf(
~
Surge Bin (2)
t
Ore Feeder (3)
t
Selected Automatic Sampler
j
Minerals.
Secondary (II)
Crusher
1
Conveyor Belt (12)
Fine-Ore Bin (9)
sheet (total units of value accepted into the fine ore bin). Waste and selected
minerals removed at (10) and possibly at (4) are checked separately, to recon-
cile the mill's tally of ore with that sent from the mine.
Many operations involve "loss of head" where the feed gravitates through a
machine. At one time this was a serious matter, and mills were sited on
hi11sides or provided with considerable height to ensure adequate fall through
the process. Handling by modern methods has so improved in simplicity
and efficiency that if the machines used to move ore or pulp are correctly
chosen, properly installed and looked after it becomes a minor item, both for
cost and maintenance. Loss of head need not, therefore, be feared when con-
sidering by-passing of undersize. At the same time, it must always be
avoided where possible.
Mineral Processing-Primary Crushing 53
Two main types of appliance are used to sort the stream of ore into "over-
size" which is to be sent through the crusher and "undersize" which by-passes
it. Both are screens. The grizzly (Fig. 17) consists of an arrangement of
steel bars having a tapered cross-section. These are held apart by distance-
pieces and set with the thick end of the taper on the upper side of the sloping
deck they help to form. Thus any piece of ore small enough to pass between
two bars falls clear.
Old mine-rails set with the bulb of the rail down can be used. The grizzly
must be sturdily made, as it suffers shock from falling rock. If stationary it
must be set at a slope sufficient to ensure that ore slides freely. Slopes vary
between 25° and 50° according to the stickiness of the ore, a usual grade
being 40°. With a cantilever construction ore sets up vibration as it hits the
grizzly, sliding is helped and loss of head can be reduced. Sometimes an
arresting grid is set above the grizzly or between its discharge end and the
crusher. The purpose is to prevent the passage of an occasional piece of ore
so large that it might give trouble. This grid should be flat and stout enough
to serve as an anvil on which oversize can be sledged by hand. Moving
grizzlys have been developed in which alternate bars slide or otherwise move.
I I
I I
I
L .,
~-4
, ..:
--
I
I
I
.
I I
I I
• I
Feeding Arrangements
Practice varies, but the broad problem .of feed to the primary crusher must
deal with the following points. (a) The mine or the quarry does its hoisting,
transport of ore, etc., during part only of the twenty-four-hour day and
probably only on a five-day week, perhaps with an occasional three-day
holiday shut-down. The grinding section in many plants is feeding a process
which it is uneconomic to interrupt, and is usually arranged for continuous
operation. At some point, or points, therefore, a stock of ore must be held
which can be built up during mine working hours and drawn upon continu-
ously. The best place for this has been found to be storage immediately
ahead of the grinding section of the plant. Ore bins having the necessary
capacity are therefore planned between the crushing and grinding operations.
(b) Most mines buy their power, or generate it, with an eye to peak-loading
(maximum demand). The crusher house makes a substantial draught on
power supply periodically, while the grinding section and following operations
take their power continuously at a level demand. Hoisting is the other
periodic power consumer which can be balanced against crushing. The hoist-
ing rhythm is dictated by the needs of the underground management, and it
may therefore be economic to hold the hoisted ore temporarily in bins until
finished, when the crushing house can take over the power. If hoisting in its
turn is being balance~ against some other power-consumer such as air-com-
pression for underground mechanisation, this pushes crushing toward the
night shift. Ordinarily, ore is crushed as soon as possible after its arrival at
surface. Expensive binning is not desired in two places for one job, and
heavy loading near the shaft is sometimes undesirable. There is the further
consideration that the ore coming to the crusher house is long-ranged and
therefore awkward to restart after it has packed down, whereas if it is kept
moving the big lumps will go along fairly easily. (c) Skips, dumper lorries
and trucks, and other handling vehicles are intermittent in arrival, whereas
the crushing section, once started up, calls for steady feed. To provide this,
surge bins can be used. A surge bin is simply a convenient holding arrange-
ment able to receive all the intermittent loads normally to be expected,
and to feed them steadily through gates at controllable rates to the receiving
machine.
Mineral Processing-Primary Crushing 55
In the primary crushing stage large lumps of ore are handled and transfer
points must be tough and readily accessible in the event of blockage. A machine
widely used for smooth control is the Ross chain feeder (Fig. 18). This
consists of a curtain of heavy loops of chain, lying on the ore at the outfall
of the bin at approximately the angle of repose. It is controlled for rate of
feed as desired, and can either be started or stopped by the crusher attendant
or worked automatically. When the loops of chain move down, the ore on
which they rest starts to slide. The whole arrangement is easily accessible for
service.
Other feeding arrangements for coarse ore include the apron and a variety
of reciprocating mechanisms, which push forward the ore lying at the bottom
of the bin with strokes controllable for rate and amplitude. One heavy-duty
machine, the jar-bar grizzly (Fig. 19), is particularly effective in handling
clayey or sticky ore. It screens out undersize in the primary crushing circuit
and is built to withstand shock loading.
Protective Devices
Fig. 19. Illustrating the Operation of the Jar-Bar Feeder. The lumps in
contact with the bars are alternately lifted and dropped and at the same time
moved forwards as the tips of the elliptical sections rise and turn over.
breakage unless the weak link in the crusher fails, there is an important
secondary cause. The jaw crusher depends for normal operation on the fact
that broken rock contains some 40 % of voids. Near the discharge end,
however, there is a reduction in volume between the open and closed side
settings. Ifin this area feed conditions permit a substantially lower percentage
of void space, the conditions of strain increase. If all the feed goes to the jaw
crusher without a preliminary removal of fines, there can be danger when
there has been segregation of coarse and fine material in the bin or other
preceding part of the operation. Such fines could pass through the upper
zones of the crusher and drop into the final (sizing) zone so as to fill what
should be the normal voids. Should the bulk arriving at any level exceed that
departing, it is as though an attempt was made to compress solid rock. This
condition is commonly known as "packing of the crushing chamber". It is
just as serious as arrival of uncrushable material such as tramp iron and can
cause major breakage. Several speakers stressed the point that any protective
device designed to fail on overload should act behind the heavy flywheels and
not ahead of them. Such breaking devices as were originally used to give a
weak anchorage of the driving pulley to the flywheel therefore fail, as do weak
fuses in the drive motor, since they leave the stored energy in the flywheel to
continue its work on the jammed crusher and thus lead to breakage. The
usual breaking point is provided by a toggle plate. This is the least expensive
Mineral Processing-Primary Crushing 57
part of the crusher, and toggle failure stops all further crushing movement.
Various types of toggle for this purpose have been designed, the general
preference being for a shearing type with the scarfed joint held together by
tapered fitted bolts reamed home. Importance is attached to rigidity of this
fit. Some development has been made with alternative methods which are
not open to the objection of a failing toggle, that the crusher is shut down
during replacement. Hydraulic protection is making some headway. In
one form this is energised by the oil pump and accumulator system and
arranges for a release of the drive linkage at a given maximum stress. This
appears still to be on the drawing board. Fluid couplings, although useless
ahead of the flywheel, are helpful for start-up, according to one speaker.
Attention should be given to the toggle seats, which are at their best when
made of manganese bronze and well lubricated. Special lubrication systems
have been devised and are said to pay for their installation. A newly installed
toggle seating should be run in carefully. Restraining ropes fastened to the
toggles can prevent danger due to pieces being thrown into the crusher at the
moment of failure. A fluid drive together with a pre-stressed friction clutch
has been devised for use in the flywheel hub on the eccentric shaft. This
combination protects against overload and also allows the jaw to stop
immediately under excess loading. One speaker recommended development
of an air-operated clutch for this purpose. The advantage would be that
once the cause of overload had been cleared the crusher could immediately
be re-started and thus avoid the present delay while replacing a broken or
sheared toggle. Another recommended taking more care ahead of the
crusher so as to reduce the danger of undesirable material being fed to it. A
picking conveyor ahead of the crusher was suggested. This arrangement has
since gained ground. Among other suggestions were those of using a molyb-
denum-based pre-coating compound followed by high-pressure oil for lubri-
cation of the toggle seats, together with careful grinding and polishing of them
to a high finish. Reduction ratios also were commented on and preference
expressed for a series of low-ratio crushers rather than a dangerously high
single-ratio reduction.
References
I. Truscott, S. J. (1923). Text Book of Ore Dressing. MacMillan.
2. Hukki, R. T. (1962). Trans. A.I.M.E. 223.
3. Griffiths. A. A. (1920). Phil. Trans. Royal Soc.
4. Gaudin, A. M. (1939). Principles of Mineral Dressing. McGraw Hill.
5. Dobie, W. B. (1953). Recent Developments in Mineral Dressing. I.M.M.
6. Dyson, F. J. (1958). Scientific American. Sept.
7. Bond, F. C. (1952). Trans. A.I.M.E. 193.
8. Holmes, J. A. (1956). Trans. Inst. Chem. Eng. 35.
9. Schuhmann, R. (Jr.) (1960). Trans. A.I.M.E. 217.
10. Gaudin, A. M. and Hukki, R. T. (1946). Trans. A.I.M.M.E. 169.
II. Nordberg Mfg. Co.
CHAPTER 4
SECONDARY CRUSHING
In primary crushing the largest lumps of ore mined must be dealt with.
In secondary crushing the maximum sized piece is unlikely to exceed 6" in
average diameter and some of the unwanted material coming from underground
has probably been removed. The feed is therefore easier to handle. The
crushing machines need not have so wide a gape nor so sturdy a construction.
The transporting arrangements can be less robust, since the large pieces of
rock have now been reduced to more manageable fragments.
Washing and sorting, if practised, may be done in the primary section, but
is more usually combined with secondary crushing, where the rock is smaller
and more easy to handle. Secondary crushers are usually arranged in series
with the primaries, so they must be able to handle similar loads. Their main
task is to reduce the ore to a size suitable for wet grinding. It then goes to
the fine-ore bins, which must have sufficient storage capacity to receive all the
ore accepted for treatment and to keep the plant running continuously,
although the mine only delivers its ore periodically.
t
( + )~<---l/l Screen/s --_~ (-) 2
I
~
Waste to ......; - - - - - - BeIt Conveyor, .
and Sortmg 3
dump
t
Crusher/s (set f') 4
t
Conveyor or Chute 5
t
Fine Ore Bin 7
The purpose of the weightometer (6) and sampler is to record the tonnage
accepted for treatment and to sample it for assay grade, moisture and particle
size. Process control requires knowledge of dry tonnage treated and the
values contained in that tonnage.
If sorting or dense-media separation (Chapter 12) is also practised, it will
probably be introduced in this section. In the case of material requiring fine
crushing by dry methods, special equipment of kinds not considered in this
book may follow. The fixed-path crushing machines discussed in this chapter
normally deliver to the bins ore crushed to below between 1" and til. Closer
settings are possible, particularly if the secondary crushing is performed in
two stages so as to reduce mechanical strain by keeping each machine's
reduction ratio below 7 fo 1. The considerations which influence lay-out and
installed capacity include the crushability of the rock itself, the question of
waste elimination and the maximum rate of delivery of ore for immediate
handling. During the mine's transport period the fine ore bins are filled.
60 Mineral Processing-Secondary Crushing
though throughout the mill's working day (usually 24-hour) they deliver
steadily to the next stage of treatment.
Secondary crushing is today characteristically performed "dry". This
word is used relatively, as the ore may be moist, or may be wetted during
washing operations or when water is run in through a crusher to prevent the
build-up of clay. (This last is bad practice, and may lead to mechanical
trouble.) The machines mainly favoured are modified forms of gyratory
crusher, though other appliances, such as crushing rolls, are in limited use.
Beater mills, with their variants (hammer, whizzer, pin-disintegrating and
ring-roll types) deal with large tonnages of coal, asbestos, limestone and
easily broken mineral. The gravity stamp, though worked with added
water, is briefly described in this chapter, since it is a fixed-path machine. It is
obsolescent, its place being taken by the rod mill described later, but is still
in use in a few older plants.
I Spindle
2 Cone
3 Bowl
The springs then return the bowl to its correct clearance. While this happens
or when choked with clay, the Symons crusher is apt to let oversize escape.
Such clay is sometimes dealt with by introducing water with the feed. Better
practice is to remove it by washing (see Chapter 2). It is dangerous with
any dry-crushing machine to risk the entry of abrasives into its bearings or
bevel gears.
The springs which hold down the bowl yield when the load is too severe.
It is therefore usual to run the cone crusher in closed circuit with a screen,
thus ensuring that "tramp oversize" is returned for further treatment. This
recirculated material, which may include fragments of non-magnetic steel
not removed by the guard magnet, may call for removal by special methods.
With some ores there is a tendency for extra tough particles to "spring"
the crusher at a slight oversize to the set. This leads to accumulation in the
closed circuit of such particles, and if neglected can build up sufficient pressure
in the crushing throat to present a problem not met with in the types of
crusher considered in Chapter 3. (In passing, it should be noted that in all
closed circuits there is selective retention of one fraction of the ore stream
62 Mineral Processing-Secondary Crushing
which may call for special measures.) A simple solution is to use a screen
slightly oversized to the set of the cone crusher, and thus increase the severity
of the crushing action upon the most prominent (largest) particles.
Ratio of reduction is controlled by screwing the bowl up or down by
means of its capstan and chain. It is usually held between 3: I and 7: I,
though higher ratios can be worked with some ores.
A 2-ft. standard crusher receiving -2f" rock is rated to deliver 15 tons/hour
-i" in open circuit, or 60 tons/hour when reducing -4" feed to It". A 2-ft.
short-head crusher receiving I!-" feed and delivering -t" in closed circuit
has a capacity of 6 tons/hour. When reducing -2" feed to -t" product
the capacity is 20 tons/hour. The coarser 7-fl. standard crusher has a
capacity of 900 tons/hour when reducing -18" rock to It". Such general
figures, taken from manufacturer's literature, would require checking for
performance on samples of the specific ore if it was proposed to work a
crushing plant at full capacity.
In the Hydrocone crusher the cone is held up to the correct setting by a
hydraulic jack instead of by springs. An automatic method of reset (Fig. 24)
Minero' Processing-Secondary Crushing 63
allows the cone to yield when an uncrushable object enters, and to return to
the correct setting after this has passed through.
Gearless Gyratories
ROLLS
Although much work once done by rolls has been taken over by cone
crushers, these machines still handle a considerable tonnage. Standard
64 Mineral Processing-Secondary Crushing
spring rolls (Fig. 25) have two horizontally mounted cylinders. The set is
determined by spacing pieces (shims) which cause the spring-loaded roll to be
held back on its sliding mounting from the solidly mounted roll. Modern
rolls have both cylinders positively driven by separate motors, so that they
rotate inward and downward. Rolls crush by nipping the feed between the
approaching roller faces and it is essential that the entering material is seized
and drawn down.
One method used to calculate maximum size of feed is based on the angular
relations shown in Fig. 26. Tan B is the coefficient of friction (usually taken
as 0'3); A is the angle of nip, or wedge angle below which a particle is seized,
and above which it skids; R is the radius of each roll. 2X is the thickness
of the particle which can be gripped at zero setting, and S is the set, or distance
apart, of the roll faces at their nearest point of approach.
R-X
Then, cosB=-- (4.1)
R
2R+S
cosB=-- (4.5)
2R+d'
Hammer Mills
In these machines, which can be used either as primary or secondary
crushers, the breaking force is mainly due to a sharp blow applied at high
speed to free-falling rock. The moving parts are beaters (hammers, rec-
tangular plates, hanging bars or heavy metal rings). They move in a more or
less fixed circle of swift rotation, though they are loosely suspended from
pins on discs mounted on a driving shaft, inside a robust stationary casing with
a grid through which broken undersize leaves the mill. The beaters weigh
from a few up to 250 lb., and the larger machines can work on feed as coarse
as 8" cube. Fracture is chiefly produced by the flailing action as the beaters
hit the ore as they spin at from 500 to 3,000 r.p.m., though part of the com-
minution results from shatter of particle against particle or casing plate. The
hammers shown in Fig. 27 can be moved to pins nearer the disc's periphery
as they wear, and are readily reversible and removable. Wearing parts are of
high-carbon steel or tough wearing alloys with manganese, chromium or
molybdenum.
68 Mineral Processing-Secondary Crushing
Gravity Stamps
These machines, which crush by impact and work wet, are briefly considered
here, though obsolescent.
The stamp (Fig. 29) is an impact crusher. From the bottom up, the parts
composing it are the mortar box in which five stamps work; the die of each
stamp set in the mortar box; the stamp shoe which falls upon ore spread over
the die; the head into which the shoe is socketed; the stem which holds the head
and is part of the falling weight in this impact crushing; and the tappet, which
is keyed to the stem. Below the tappet is the cam, one of five revolving on a
camshaft. These cams lift a battery of stamps so that they fall in some such
sequence as 1, 3, 5, 2, 4. They are driven by a belt with a jockey pulley.
A lever system worked from a tappet on the middle stamp actuates a hori-
zontal rotating plate at the back of the stamps (the Challenge feeder) whenever
the ore-level drops below a set height above the central die. This feeds new
ore into the mortar box, thus maintaining a bed of rock undergoing attrition.
Water is piped in, and at each fall of a stamp water and ore splash outwards
from the die. A screen is placed along the front of the mortar box and when
a crushed particle rises high enough, is small enough, and also is carried truly
to an aperture in the screen, it escapes. Thus, through choice of a suitable
height of screen and the size of its aperture, and by regulating the volume of
transporting water available, the material stamped can be controlled as regards
its maximum escaping size, but not as regards its minimum size. The stamp
Mineral Processing-Secondary Crushing 69
has a very big reduction ratio, being able to accept 2" material and reduce it to
fine sand.
It will be realised when modern grinding methods have been considered
that this big reduction ratio is one of the reasons for the stamp's obsolescence.
In theory, nothing could be more efficient as a comminuting system than the
~------6·~.4~----~-
.
,0
'..
The gyratory crusher works continuously with half its surface, but the
jaw crusher works half the time with all its crushing surface. For the same
maximum-sized piece, the gyratory has considerably greater tonnage capacity
than the jaw crusher, but this is not necessarily an advantage. Where large
throughput is needed, and the need to avoid vibration is important, the gy-
ratory is better than the jaw crusher, but if transport to the mine is difficult it is
less easily delivered.
The running speed of a crusher is limited by the time needed for partly
crushed material to fall between squeezes. On slabs or flaky material the
gyratory is more efficient than the jaw crusher because of its more pronounced
beam action. In secondary work, the Symons cone has a higher ratio of
reduction than its nearest competitor, rolls. Hammer mills are rarely used in
hard-rock crushing.
The usual sequence of crushing machines is a jaw crusher, feeding to a
gyratory which sends its product via by-passing screens to cones or rolls, to
yield a final product anywhere from in to ~. maximum size. .The trend is to
take secondary crushing still smaller so as to ease the work of the grinding
circuit. Below this size, any further reduction is effected by tumbling mills
in the grinding section. When properly fed, jaw crushers, gyratories, and
rolls are all arrested crushers, but if the discharge arrangements become
obstructed or choke feeding is indiscriminately practised, rock grinds upon
rock, and waste of power, reduced throughput, and perhaps trouble in later
sections of the flow-line will result.
With wear, the capacity of the Blake crusher rises somewhat, since the
effect is to lower the pitman and increase the reciprocating stroke. Wear
of the working parts of the gyratory has an opposite tendency. It is easier
to alter the set of the Blake than that of the gyratory crusher, save with the
type having a hydraulically held spindle. Clearance of the crushing throat
after a stoppage under full load is also much easier with the Blake. Gyratories
have better protection against the entry of dust to bearings, and they are more
economical of power than reciprocating machines. Ore passing through loses
more height than with the jaw crusher.
CHAPTER 5
WET-GRINDING MILLS
Preliminary
Milling Action
The tumbling mill is a horizontal cylinder which is nearly half filled with
crushing bodies, usually cast iron or steel balls. Ore can be fed in at one end
and discharged at the other, continuously, and water added at the feed end
helps to flush it through the mill. When the cylinder is rotated, this mixture
of balls, ore, and water (the "crop load") is churned with, or battered by,
flying balls, according to the speed at which the mill is run. The kinetic
energy of this tumbling load is dissipated as heat, noise and comminution of
ore, grinding media and mill linings. Its useful energy is expended during
impact, abrasion, shear and compression of mineral particles, but it bears on
all moving surfaces in contact. The conditions are quite unlike those in dry
crushing. The inside of the mill is slippery with water and slime, and the
particles undergoing comminution are usually too small to be firmly gripped.
Further, the only fixed surface in movement is provided by wearing plates
(liners) which line the inside Of the cylindrical shell of the mill. These are
buried under the crop load during part of each revolution and then rise clear.
Chance contacts for each individual particle in a churning mixture of water,
Mineral Processing- Wet-Grinding Mills 73
,
r---- - -10'-:1'0'- - - - .,
.........
Types of Mill
pass through the fine screens lining the cylindrical part of the mill. Heavy
perforated plates protect these screens from injury, and a coarse screen is
mounted concentrically inside each fine one to give further protection.
External sprays provide water and the external casing can be flooded so that
the mill dips into water.
The critical speed of a mill is 76'\)d in r.p.m. when d is the mill diameter in
feet less the diameter of the largest ball in the crop load.
Mineral Processing- Wet-Grinding Mills 77
....
,-..
0U
'IJ
:::I
....
f--
....
.:::<Il
~
"'"'
~
f.. ~
§-
c::..
~
....;
.....
~
f
:
6
78 Mineral Processing-Wet-Grinding Mills
Critical speed being a function of peripheral speed, the rate at which these
mills can be run depends on the maximum diameter inside fully worn liners.
Centrifugal effect is strongest in the drum section, where the ability of the
mill to lift its crop load up the rising side of the mill is therefore highest. At
the same time the load inside the steep cone tends to work back down-slope,
so the resultant is a pronounced heaping-up of load in the drum and a tapering
off toward the discharge. Another differentiating action is also at work,
as in all ball mills. The largest balls tend to work to the top of the crop load
with the largest lumps of ore. They are then freest to roll or slide down-
slope on top of the turning load, where they then are most liable to be caught
between the bottom of the sliding mass and the downward-moving liners.
There is thus a tendency for the biggest balls and rock to work to the periphery
of the drum-section. The makers of this mill claim that the balls segregate
themselves somewhat, in such wise that the biggest are in the drum and so
disposed that when they rise in the turning load they fly or tumble the furthest
distance, while the smallest balls work up toward the discharge. Any such ten-
dency is useful, since it causes the biggest balls to work upon the newly-entered
and therefore biggest pieces of ore, while the smaller balls, with less interstitial
space, handle the partly finished material as it works its way out. At low
speeds the pressure of the heap piled up in the drum has crushing value, and
at high speeds balls break clear and rain hammering blows on the mass of
churning metal and rock several feet below, giving impact crushing. It is
obviously of value to be able, so to speak, to control one's punches, and to
use a heavy weight against a tough piece of ore with some selectivity of
target while keeping the light hits and tight jostling for the small stuff which
does not need such strong treatment. The effect of the conical shape of the
Hardinge mill on peripheral speed and kinetic energy on the crop load at
various cross-sections is illustrated in Fig. 30.
In the attempt to increase the tendency toward segregation of the largest
balls at the extreme radius, the Hardinge Co. also market the Tricone mill in
which the drum, instead of being truly cylindrical, has a slight back-slope
toward the feed end.
it is desired to pass worn balls out of the mill, and varies from t" to I
N. Grates
are cast with a slight flare outward to aV'Jid "blinding" and are procurable in
high-grade alloy steel. The size of the mill is limited at present by manu-
facturing possibility. 10' by 10' mills carrying a 45-ton ball charge and using
800 h.p. are in use. A characteristic feature of the low-discharge mill is that
the diameter is made as great as possible while length is either about the same
as the diameter, or at most twice as great. This follows logically when it is
appreciated that the purpose of low-discharge milling is to cut down the
dwelling time in the mill. A mill in which diameter and length are approxi-
mately equal is called a "square mill" in the United States.
Mill Liners
Mill liners perform two major duties. They protect the mill's cylindrical
80 Mineral Processing-Wet-Grinding Mills
Middlings
from flal
...
~
shell against abrasion by the tumbling mixture of ore, water, and crushing
bodies, and thus preserve it from wear. The shell not only provides the
frame of the mill, but must also be leak-proof. The liners also help to tumble
and rotate the crop load. The shell, geared to the driving mechanism,
revolves at a fixed peripheral speed. The liners, being bolted or otherwise
fixed, revolve without slipping. Resting on the liners is the "crop load", a
mixture of crushing bodies, ore at various stages of comminution, and water.
Ignoring transmission losses, the kinetic energy consumed in grinding is
that used to unbalance the crop load. This load is displaced by frictional
drag against the rising side of the rotating mill. Being somewhat fluid it
resists the displacing force by tumbling downward. The further the load is
lifted out of balance the greater is the counter-input of the kinetic energy
consumed during comminution. The grip of the liners transfers this grinding
force into the churning crop load. As the mill revolves, the crop load is
Mineral Processing- Wet-Grinding Mills 81
carried upward by the rising body-liners (the liners which cover the horizontal
inner surface of the cylinder, as distinct from the end liners which cover the
vertical circular ends). The binding forces causing the load to rise consist
of (I) the centrifugal action of the revolving mass; (2) its collective pseudo-
viscosity and (3) interlock produced by mutual adjustment of balls, liners,
and ore. If the liners are smooth, these three factors play the chief part in
determining how much of the mechanical energy being used to turn the mill
shall penetrate to the crop. If the liner surfaces have projecting ribs, positive
lifting force can also be applied to the load. Instead of losing contact soon
after it starts to rise and churning round quietly, the outside of the crop load
can thus be helped to climb higher, so that it falls back with sufficient violence
to add shattering action to the abrasion characteristic of churning motion.
Body liners receive heavier punishment than end liners, and wear down
about twice as fast. A contributory cause of wear is chemical corrosion,
either due to mine water or acid from decomposition of sulphide minerals.
Fortunately, it is frequently necessary to add lime in the mill circuit for
reasons connected with the concentration processes, and this gives protection
to the ironwork of the mill, neutralising any acids present. Although the
purpose of grinding is the comminution of ore only, the forces employed
also act on lining and contents of the mill, and all are subjected to the grinding
action. Liners and crushing bodies are therefore selected for their hardness,
toughness, and resistance to such wear. When made of alloy steel, they
must not introduce any element into the circuit which can cause trouble in the
concentrating section, since the steel abrades during grinding and is carried
forward to that section. Consumption of steel and cast-iron liners varies, and
is usually of the order of 0·03 to 0·3 lb. per ton of new feed.
Observations made by Bondi in which an experimental glass-ended mill
loaded with -in steel balls but no ore, showed that slipping in the layer of
balls next the smooth shell increased with the milling speed, in the 60% to
80% critical range, whether the mill was run wet or dry. This effect in aIr
diameter mill not containing mineral would not necessarily be repeated
under normal operating conditions. Abrasive grinding is proportional to
slipping of the charge. Bond's observations showed 15 % slip between the
outermost layer of balls and the smooth liners, and a further amount varying
from 5% to 10% with each next layer, radially inward to the fifth or sixth
layer where observation was not possible.
Rise in speed (and therefore in centrifugal pressure) increases the energy
of outward travel of the heavier components of the crop load. This should
cause a tendency towards strong keying of the solids near the shell and toward
mobile liquid conditions at the centre of the churning load where centrifugal
force is lowest. The keying effect on the solids of increased speed is, how-
ever, modified by any change in volume or weight of the crop load. The
behaviour of the upper portion changes with acceleration. Part of the charge
breaks loose and reduces the overall out-or-balance weight.
In a cylindro-conical mill the difference between the peripheral speed of
the drum liners and of the segmental ones which line the steep cone leads to
a change-over from pronounced impact grinding at the greatest diameter to
pronounced abrasive grinding toward the discharge.
82 Mineral Processing-Wet-Grinding Mills
Mill liners are made in many shapes, but they fall into three main groups:
(1) Smooth, having. simple frictional contact with the crop load.
(2) Grate (pocket, grid, El Oro, self-renewing) liners, in which the
outermost layer of ore and/or balls tends to wedge in the openings
(Fig. 34) and form the wearing surface.
(3) Lifting liners, with longitudinal waves, ribs, lifters or wedge bars
which to some degree grip the crop load on the rising side.
Coarse-grinding mills, which draw their feed from the fine ore bins and
reduce it mainly to coarse sand, usually have longitudinal projections on their
body liners which provide a straight-line key along which the rods can lie as
they are being lifted. Slip in zone AA-BB is thus prevented and impact
crushing dominates. Primary ball mills also have ribs, steps, bars, "shiplap",
or other projections, but these are not necessarily continuous horizontally.
Secondary ball mills usually have wave liners or smooth ones. The profiles
of some of these are shown in Figs. 36-37. Ribs and bars are shown in
a, b, c, i, j and stepped liners and shiplaps in f, g, h, and k. Some types of
wave liners are shown in d, e, I, and m. The coarse ore selected from the
crop load by Group 2 liners reduces consumption of steel. They are designed
• (a)
(t)
(e)
CIr)
(I)
(f) (m)
(go)
Zone of Turn.bling
(BB-CC)
:-----+-~~-Rolling Layer
of the liquid-solid ratio in the crop. Since the mill speed is usually fixed, no
running variation of this factor is possible.
Trunnion liners can be fitted from outside with- only slight delay. They
are either smooth or have a helical spiral. At the feed end this spiral, if used,
helps ore in. At the discharge end it retards departure by working against the
outflow. Liner bolts have an oval or square taper head. Liner backing is
often used in heavy-duty grinding to mitigate shock and danger of liner
fracture. Such backing may be I" planking, rubber sheet, old belts, or the
liners may be set in zinc. Circumferential grooves may be worn in the
liners by the frictional drag of the down-slipping crop load, the balls bearing
on the liners without spinning. As soon as the grooves are deep enough for
the balls to be in goO(~ contact with the liners this form of wear ceases.
The cost of liners is a relatively small item in grinding, so it is customary to
buy those best for the work, even if expensive. Time lost in renewal of
quickly worn out liners would more than offset the extra cost of a suitable
steel. Alloy steels are preferred for heavy work in primary mills using balls
larger than 2" in diameter. For secondary grinding, where impact is lower,
cast-iron or self-renewing liners are favoured.
Feeding
The grinding mill receives new ore from the fine-ore bins, returned classifier
sands if in closed circuit, new supplies of grinding balls as required, and water.
Usually, the feeding arrangement must elevate part of the load to trunnion
height in order to close the return circuit without the need for installing a
special device. The most widely used feeder is a scoop working in a feed box
that slopes downward to the gathering zone. The scoop (Fig. 39) turns as
part of the mill and its tip digs into the downward sliding ore. At the same
time it picks up mill-head water which is being fed to the box at a controlled
rate. The clearance between tip and box must exceed the thickness of the
largest piece of arriving ore to avoid risk of damage. The balls fed in
periodically to "top up" the mill charge enter through the discharge trun-
nion. Only the scoop's tip is armoured with abrasion-resisting alloy, and
no other part of the outside of the scoop should be in contact with ore.
Single scoops are most used, as they have the greatest capacity. The rate of
feed (new ore plus classifier returns) is limited by the handling capacity of
scoop and input trunnion. Ample clearance in the inside elevating spiral is
important, in order to avoid obstruction or bridging by large pieces of ore.
The spiral must be developed sufficiently slowly to ensure that feed is trapped
and sent to the trunnion with no risk of falling out.
Ir open-circuit work, or when the return circuit is closed by a hydro-
cyclone instead of a mechanical classifier, the scoop feeder (Fig. 39) is not
essential. Its place may be taken by a drum feeder (Fig. 40). When th,
crushing bodies are pebbles, or big pieces of rock, new supplies must be
added frequelltly. The!'e may be delivered via a feed hopper in light contact
with the feed trunnion, which must be suitably large.
Most modern plants are planned for continuous ore treatment on lines
Mineral Processillg- Wet-Grinding Mills 87
which call for a certain amount of chemical and physical control of the pulp
stream. Efficiency requires equable and smooth flow of a steady tonnage of
properly prepared pulp and the first step toward achieving this state of things
is taken in the grinding circuit. The primary mill receives (a) somewhat
long-ranged feed from the fine-ore bins, limited to perhaps j.N maximum size,
and (b) returned classifier sands. From its closed-circuit should be released
a steady tonnage of pulp of specified limiting mesh and water/solid ratio.
Classifier action is discussed later, but it is desirable at this stage to consider
the governing influence of the closed circuit used in wet grinding.
If a five-to-one circulating load is maintained in the primary circuit, this
means that of six units of ore entering the mill one consists of new ore and
five of material returned from the classifier after partial grinding during its
previous passage through the mill. If the maximum entering size is tN, and
the maximum discharge 60 mesh (12,700 microns and 211 microns respectively),
this represents a reduction ratio of about 60: 1. Since this reduction is not
achieved in one pass through the mill the true figure is not 60 to 1. For the
average particle it is 10 to I, since it is the progeny of rock which has passed
six times through the grinding zone b::fore obtaining release from the circuit.
Not all the particles are recirculated. Some are already undersize before
entering the mill, and these should be passed on when they reach the classifier.
Some, either through weight, toughness, or excessive size, may make more than
six journeys round, being broken down progressively to the finishing sand size.
88 Mineral Processing-WeI-Grinding Mills
The largest and heaviest particles tend to be retained in the mill longer than
the average pieces of new feed. This would lead to accumulation of over-
size and loss of capacity were it not countered by methods to be described in a
later paragraph.
The effect.on the total feed (new feed plus returns) of its heavy dilution by
partly ground ore is to reduce the average size of the particles undergoing
comminution. The predominating size in the return circuit is of ore which
has progressively been reduced toward a condition which will permit of its
release from the closed circuit. As the return feed/new feed ratio is 5: 1,
this "near-release" particle tends to dominate the behaviour of the ore in the
mill crop. The equalising action of these recirculating sands is a steadying
factor in the grinding work. So long as five-sixths of the total feed is steady, a
little "surging" from the average, either of tonnage or size of new feed, is
smoothed out. If, however, the character of the new feed becomes steadily
different, adjustment of its rate of admission is necessary.
The circulating load in a closed circuit can be calculated from the formula
F(b-a)
R=--- (5.1)
a-c
where R is the weight of returned oversize/hour; F that of the feed to the ball
Mineral Processillg-- Wet-Grinding Mills 89
New Feed
(one unit)
Return Sands
(five units)
--------,1
t 1.------- Water
ll,-------water
I
1
, - - - - - - - - Overflow
(one unit)
t
Secondary Ball Mill
(one unit of solids)
(two units)
1 tr-----water
'-------~----- Pump
mill, and a, b, and c are the percentages passing a specified screen (say 100#).
The a percentage is for R + F, b for the classifier overflow and c for the re-
turned sand R.
The mill's charge of crushing media is so composed as to cope with new
feed which is reasonably consistent in its proportion of various particle sizes,
held below a specified maximum dimension. There must be good working
liaison between crushing and grinding. A difficulty often encountered is
segregation of sizes in the fine-ore bin, which leads to alternating deliveries of
fine and coarse feed, even when no individual piece of ore exceeds the limiting
mesh. Obviously, the less work the grinding plant has to do on the entering
material, the greater is its throughput. Conversely, when the feed is "lumpy"
its rate of entry must be slowed down since the available grinding power
cannot be increased mechanically, and most mills have a fixed running speed.
Farrant 3 (Fig. 42) has worked out an optimum change-point from crushing to
wet grinding. 3" is a rarely used maximum feed size, the usual figure being
90 Mineral Processing--Wet-Grindinlr Mills
4~--~----------------~--------"~------+-~
.!.=
S
..
::I
0
.c
2
:.:
==
• 111
new ore to the primary circuit can be regulated. It is, however, simple to
vary the amount of ore circulating in the primary closed circuit. If the
secondary mill becomes overloaded, the primary circuit is set to grind finer.
This throws more grinding work on the primary mill and enables the skilled
operator to keep both loads in balance.
Two features of the flow-sheet (Fig. 41) are returned to later, but may now
be noted. In the primary circuit a higher circulating load is possible, because
the return sand is coarse enough to settle quickly in the classifier. The pulp
escaping from the primary circuit must not be sent direct to the secondary
ball mill, for two reasons. First, and most important, the solid-liquid ratio
must first be adjusted by removing part of the overflow water. Second, any
of the solids now sufficiently ground should by-pass the secondary mill. If a
rod mill is used ahead of the primary ball mill circuit, its discharge must enter
the primary classifier for de-watering.
In a simple operation where the ore does not vary much in quality, it may
be drawn direct from the fine-ore bin through one or more gates. The use of
multiple draw-off points helps to reduce size segregation. If there are no
big pieces, the gates may deliver down a steep slope on to a belt conveyor.
The height of the ribbon of ore is regulated by a vertical plate attached to the
gate. From time to time the conveyor is stopped and the weight resting on a
measured length is taken to provide a tonnage check.
A more robust arrangement is the apron feeder (Fig. 43). It has a variable
speed, and the height of the ribbon of ore moving along it is adjusted to give
the desired rate of feed by means of a vertical gate. The roll feeder is a
large cylinder revolving below the withdrawal opening of the ore bin. It is
simple, and readily accessible when obstruction occurs at the ore-bin discharge.
A cut-off gate controls the height of the ribbon of discharging ore, and the
rate of turn can be controlled if an adjustable driving mechanism is used.
Another method of checking the tonnage is provided by the Humboldt
feeder. Two hoppers alternately receive ore from the bin, one filling while
the other automatically tips its load. A further method, used for unusually
sticky ores, is to push ore forward into a chute by means of a reciprocating
ram. The length and speed of the stroke are adjustable, and the height of
the moving ribbon is regulated by a cut-off gate.
Where the ore differs in various parts of the mine, it is sometimes necessary
to keep the varying types separate. They may then be treated independently
or may be blended to produce a consistent grade of milling ore. This calls
for separate bins, each of which can deliver on to a blending conveyor. As
the proportion of each type of ore is an important factor in this treatment, the
draw-off feeder may be of a self-regulating type which delivers and regis-
ters its quota of tonnage to the main conveyor belt. One of these is the
Hardinge constant weight feeder in which a short conveyor belt is pivoted
below the ore bin. When the ribbon of ore upon it varies from the desired
weight, the whole belt swings up or down under the influence of its adjustable
counter-balance, thus altering the area of the delivery aperture a little and
correcting the rate of feed.
Another type of feeder uses electrical vibration. It works best on dry ore.
preferably fairly free from fine and clogging particles. The power unit
(Fig. 44) throws the feeder forward and upward. The ore flows in imper-
ceptibly short hops with little abrasive sliding against the conveying chute.
Feed rate is controlled by varying the amplitude of vibration, by means of a
rheostat. Any device operated by A.C. current to induce shaking must be
carefully protected against even a slight change in frequency, which has a
magnified effect upon the rate of feed.
Crushing Bodies
As the power transferred via the liners is dissipated in collision and inter-
facial friction, impact crushing and abrasive grinding occur. For impact
crushing to be effective there must be sufficient kinetic energy to disrupt the
particle on which the blow falls. It is applied either direct or by transfer
from the area of impact to the point where a particle is being gripped. The
relation between the particle and the crushing body must be such that the
Mineral Processing-Wet-Grinding Mills 93
energy applied exceeds the total resistance-or, in other words, the bigger
or tougher the rock, the larger the crushing force needed. It follows that
larger or heavier grinding media (balls or pebbles), capable of delivering more
powerful blows, are needed in the primary than in the secondary grinding
circuit. Each particle in movement in the crop load can interact with any
others and may at any moment be striker, anvil, or both simultaneously.
The liners, besides transmitting the driving force needed to generate grinding
energy, act as fixed anvils at parts of the revolution. An interesting point
is that so long as a free crop load exists the mill can absorb grinding power,
regardless of the efficiency with which that power is used. If only balls were
present, they would tumble and wear, and grind one another and the liners,
producing abraded iron and heat. In practice, this would be worse than use-
less. If only ore and water were in the crop, some grinding would result, but
the kinetic energy of the colliding and abrading components would be in-
sufficient to produce efficient grinding, unless an adequate percentage of
sufficiently large pebbles or pieces of ore were present.
The composition of the crop load influences the quality of grind. The pro-
portion of water to ore, and of ore to the grinding media, can be controlled.
The limiting mesh of ore is also controllable, but except for some hand picking
it is not possible to regulate the nature of the feed. The density, size, shape,
hardness, and toughness of the crushing bodies can be regulated when new,
but unless the mill is periodically stopped and its contents are dumped out
and sorted over, the shapes into which balls wear during their working life is
uncontrolled. Their shape is a factor in keying together the tumbling load.
There must be enough crushing bodies of appropriate weight to deal with the
variously sized particles being ground. Steel balls worn small cannot grind
by impact. Further, they reduce abrasive grinding by filling interstices
with metal instead of ore.
Impact crushing can best be applied by the unobstructed drop of a heavy
piece of steel, yet the stamp battery, which applies its energy in this way, is
obsolescent. The shattered products cannot be cleared away sufficiently
between blows. The rod mill applies part of its force as a dropping weight,
and can both shatter and abrade the ore the rods fall on or roll over. It
also discharges the crushed particles expeditiously and with a minimum of
further comminution.
Rod mills can be run at high speeds' in specially designed mills. Thelifting
projections of the liners must assist the rods to lie axially and unentangled,
and the rods should resist bending. The crushing load must not become
matted so that it falls irregularly and increases the already severe shock-
loading produced when each weighty steel rod crashes down from its break-
away point. The mill components must be stronger than in ball-milling, in
order to stand up to the punishment shell, liners, and bearings receive. Most
mills are not, however, "cataracted", as running under free-falling conditions
is termed, but "cascaded", the emerging rods rolling down the slope of the
crop load to the toe of the mill. In this action the rods act as multiple rolls,
the mill speed being well below two-thirds of critical. An important point
in this cascading roll-action is that the whole weight of the rod (reinforced
by that of all the rods pressing on it from above) bears upon a few particles
94 Mineral Processing-Wet-Grinding Mills
of ore as they hold the rods apart. Since rod mills are nearly always run in
open circuit, the biggest bridging particles will be near the feed end of the
mill, and the smallest near the discharge. This results in a fanning out of
the rods, visible as a tendency to work down and attempt to leave the mill by
reason of a definite slope toward the discharge end.
C ~C=====---i-_.a-J
I ]
Fig. 45. Particle Distribution in the Rod Mill
In Fig. 45 A the aperture between two rods formed by the bridging effect
of the largest newly entered particles (left) and the smallest (right) is seen to
form a tapering s10t. Collectively, the rods thus arrange themselves to act as
a screen through which the smallest particles can flush clear of the crop load
and then leave the rod mill if not again entrained by rods falling to the toe
of the charge as the particles are carried toward the discharge end. The
effect of a broken length of rod in upsetting this action is shown at B and the
exaggerated aperture produced by too large a piece of rock at C. The wedge-
shaped spacing presses oversized particles back toward the feed end and helps
to distribute the load. The particles at the discharge end determine the "set"
of the rod mill. All particles able to escape near this point leave the mill.
It has already been observed that every shattering blow produces a com-
plete range of sizes. The rod mill cannot give a finished product all of which
is just below a certain desired size. What it can do, given the necessary
mobility of the particles in the crop load, is to direct the crushing force
specifically toward the largest particles in the mill, and on the whole to avoid
acting upon the smaller ones which are being protected by the bridging and
screening action. This can happen so long as the voids between the rods do
not become choked with fine sands to the extent that they are sprung apart,
or choke fed. The analogy with crushing rolls under similar conditions
of loading would be satisfied if those rolls had so weak springs that they
yielded before sufficient crushing force could be directed upon the passing ore.
Mineral Processing-Wet-Grinding Mills 95
Rods, being lightly loaded, would yield if choke fed, grinding efficiency would
suffer, and the nature of the product would change. It has been found
that rod mills work best in open circuit, and suffer if a closed circuit is used
in which quantities of fine sand are returned. Loaded below choke conditions
with reasonably coarse material flushed by plenty of water, the rod mill can do
its own close-circuiting by reason of the wedge-screening action above
discussed. In one set of tests the water-solid ratio at feed and discharge ends
was 30:70 but it was found that in the mill itself a 70:30 ratio was operating.
The rest of the water was acting as a sluice, flushing out finished-mesh ore.
Rods wear down in use, and this affects the angle of nip. Tests have shown
that. regardless of diameter, rods circulate through the crop load without
marked segregation with respect to the turning centre of the load or of the
mill axis. High-carbon steel. hot-rolled and straightened, is the preferred
material for rods. This steel breaks when worn small, whereas mild steel
bends and tangles, matting the load undesirably. Efficiency is highest when
worn rods are removed periodically. Mills with big discharge openings
facilitate this. Rods can be charged in by crane in bundles fairly quickly and
worn ones removed. Trunnion-ended mills are also used, rods being charged
in through the discharge trunnion individually and the worn ones "grinding
themselves out".
Rod mills are not used for fine grinding, but, increasingly, to bridge the
gap between secundary crushing (to -i") and secondary grinding, the rod
mill discharging at -16 mesh. Because of their value in this transition range
they are said to give a "controlling grind". This means that their discharge
is of a regulated fineness, thus simp'ifying the loading of the ball mill with
appropriately sized grinding media. Rod mills are run at about 40% of their
volumetric capacity.
Many shapes have been tried in the search for the most efficient steel crushing
media. but the arproximately round ball is the only one in world-wide use.
It is cheap and efficient. A mill, when loaded to its full capacity, has nearly
half of its volume filled. In this load, the ratio between the weight of crushing
media available to exert pressure on the bottom part of the crop load, and the
ore packed into the voids between those media, depends on the space made
available by the shapes comprising the packing system. Rods give the
highest proportion of metal, consequently the highest static pressure. Balls,
for reasons discussed below, are random sized, but they tend to stay truly
spherical till worn small enough to become trapped in the voids between
larger balls, when they wear into tetrahedra with concave faces, which pack
the voids and reduce reservoir capacity and mill efficiency. Some primary
mills are stopped regularly, for removal of such material and broken balls.
The load varies in practice between 50'1.. and 40'1.. of mill volume, according
to the percentage of the grinding capacity being used. The percentage void
in a ball charge is about 38 ~,;" and the weight per cubic foot for an average
mixture of balls is about 280 lb., as against 390 lb. for rod!> (see Table 5).
A sphere has minimum surface area for a given volume or weight. It has
also equal mobility in all directions under crowded or uniformly restricted
conditions. It is better able to spin on any axis than is a non-spherical body,
since the latter can only turn freely on a limited number of axes, and hence
96 Mineral Processing- Wet-Grinding Mills
only when preferentially aligned. The ball thus has good spinning quality
until worn out of its spherical shape. Wear is caused by abrasion against
other balls, liners, and ore; breakage (of small balls) when hammered by
.large balls; damage when arrested after flight, due to brittleness without
adequate toughness; and chemical corrosion by acids derived from ore or
milling water, which can be lessened with an alkali such as lime.
Rate of wear has been the subject of considerable research and discussion.
It is said to vary as the cube of the diameter D of the ball by Davis', as
D 2 • • by Bond 6 and D2 by Prentice'. Wear is proportional to abrasive rub-
bing contact when a ball spins and grinds while cascading, or while slipping
down in a rising crop load. Under such conditions wear should be pro-
portional to surface, or to D2. When the ball is cataracting a greater amount of
impact crushing is introduced. Impact is proportional to the kinetic energy
of the flying object, or to its mass multiplied by the square of its velocity.
TABLF 5
DATA ON SIZES, WEIGHTS, AND SURFACE AREAS OF RODS, STEEL BALLS
AND FLINT PEBBLES. MILL DIAMETER (FT.) INSIDE SHELL LINING
I Area
per per
cuft. ton
RODS
(10' long) sq. in. sq. ft.
5" dia. 1100 3·28
4" 1384 4·09
3" " 1680 4·96 Approximate weight in rod mill
2" " 2760 8'15 390 Ib./cu. ft.
"
It" " 3680 10·88
BALLS lb.
5" dia. 1188 4·91 18·5 108
4" " 1487 6·14 9·5 211 In mill
3" 1980 8·19 4·0 505 about
2" " 2980 12·31 1·2 1670 280 Ib./cu. ft.
" 3950 16·32 0·5
I'y " " 4010
Since mass is proportional to D", the rate of wear varies with volume. This
seems due to greater abrasive power in the heavier ball rather than to impact,
which would tend to reduce wear as it work-hardened the metal surface. The
faster the mill is run the higher the charge rises and the greater is the loosening
of its upper portion. This increases slippage and wear.
Tests are made in order to evaluate the wear rates of mill liners and balls at
Mouni Isa have been reported. '4 They were made in an S-!, by 12' trunnion
overflow mill charged with 2" balls, to which test balls of various compo-
sitions were added in batches of fifty. These were of 31" diameter, making
them readily identifiable, and were removed at intervals of 448, 920, 1592 and
2504 hours and examined for wear. The results are given in a series of
Tables, starting with the analysis, hardness and specific gravity of each of the
ten types used. Some of the findings have been condensed in Table 6. The
wear rate factors were calculated from the formula
WO-WN
Wear Factor = - - - (5.2)
DO-ON
where W0 and Do were initial weight and diameter and WN and ON the final.
This factor represents the loss of weight sustained per unit surface area.
Comparison was with Ni-hard as 100 (column 3 of Table 6). A wear rate
factor for time taken to change in size from 3·45" diameter to 2·8 is given in
N
column 4.
TABLE 6
(I) (2) (3) (4)
Final Diam. Column
COMPOSITION diam. " loss wear rate
100 hrs. factors
I. Cast steel * * 183
2. White cast iron 2·57 0·0399 153 154
3. Austenitic Mn Steel 2·52 0·0391 153 154
4. Pearlitic Cr-Mo steel 2·75 0·0327 122 121
5. Heat-treated forged steel 2·63 0·0327 127 129
6. Climax 6-1 steel (Mn 6·4 Not
MoO·9j recovered 131
7. Martensitic Cr-Mo steel 2-80 0·0287 110 100
8. Ni-Hard 2·85 0·0268 lOOt 100
9. 28 % Chrome steel 2·85 0'026 97 95
10. 16% Cr 3% Mo steel 2·95 0·024 91 90
• Too small to be identifiable
t Ni-Hard was chosen as reference standard
The authors found the reduction in diameter of the test balls to occur at a
decreasing rate. Under the local conditions Ni-hard was shown to be the
most wear-resistant and economical liner material readily available, though
alloys 9 and 10 (Table) would wear longer. Martensitic chrome-moly steel
would be a favourable alternative to Ni-hard.
Thus far mill speed has been considered mainly as a percentage of critical
98 Mineral Processing-WeI-Grinding Mills
da 3 -db'
%W=--- (5.3)
t1j·-d,
where %W is the percentage (weight) of balls in the size range da-db , and d,.
and d, are the sizes of balls fed to and rejected from the mill. Though the d.
relationship for wear is not generally agreed, this formula provides a good
working guide for replacement in the crop load.
Mineral Processing-Wet-Grinding Mills 99
I I I I I I I I I
\ \ \ \ ~\ \*,l\~-+-+-+-+--1I--+-+-+-+-+-+--l---I"
\ \~ ~ [\\;1\
It \ \ .\ \ \ l,~:-';-\~~:+-t-+:+-t-+-t-t-+-t-+-+-+-I a.
\\\\\~~'\·':n.
\\ ' \ \ .\ l\.~~.:7.-"I,rt-t-;-+-t-+-+-t-+-t-+-~u
1\
\\ \ \
\ \ \ \ \)
~ ~':O; ~
~ --t-''k-r\.+-+-+-+-+-+-I-+--+-4--1 ao
10 II .2
Fig. 46. Relation between R.P.M., Diameter, Peripheral and Critical Speed
(Allis-Chalmers Mfg. Co.)
100 Mineral Processing- Wet-Grinding Mills
Unless pebble grinding media form part of the crop load and are correctly
allowed for in their relation to grinding capacity tramp oversize in the entering
ore feed is bad. A properly loaded mill contains no balls big enough to
shatter such material by free-falling methods. It remains in the crop,
cascading and wearing slowly down till it is either small enough to be nipped
and abraded efficiently, or has suffered random attrition sufficient to reduce it
to shatter-size. An occasional piece of oversize cannot be avoided, but an
appreciable quantity would build up in the crop and interfere with throughput.
Where an open-circuit rod mill is used between the secondary crushers and the
ball mills, the question of ball ratio is unimportant, since its discharge is
delivered at -14 mesh and size disparity between ore particles is then a minor
factor in grinding. An expression for the ratio between ball and particle
size had been worked out by Coghill and de Vaney'. If D is ball diameter in
inches, d - particle diameter, and K is a constant for rock grindability varying
from 55 for tough chert down to 35 for softer dolomite, then D" is Kd.
As the balls spin, ore is drawn between their surfaces, where it receives
abrasive grinding and shattering from any transmitted impact force generated
by the blows being rained down upon the crop load by flying balls. Thus the
ball needs to have surface hardness, expressed as a high "Brinell number", in
order to withstand abrasion. It also needs toughness in its core in order to
give resilient resistance. It is anvil to a descending ball, hammer when itself
a descending ball, and a transmitter of shock through the crop system. When
balls are worn out of shape they no longer spin freely and are less able to
coat themselves continuously with particles drawn from thz interstitial
reservoir between the balls. Then one of two things happens-either the same
coating of rock receives too much attention and is over-grou'ld because it is
not constantly changed and renewed, or ball rests against ball and metal
grinds metal. Either condition is inefficient.
Balls are made from chilled cast iron or forged steel, which may be alloyed.
If a delicate chemical concentration follows the grinding, care must be taken
that no endangering alloy is used, since ball wear can be as great as 4-!- lb. per
ton of ore in extreme cases. Ball hardness decreases wear rate, while ball
toughness decreases breakage. Wear is proportional to the mill diameter
and speed, the liner roughness and hardness, the ball diameter, the solid-
liquid ratio of the pulp, which affects its coating of the ball, and the acidity
derived from the ore and mill water. Wear is normally between I and 3 lb.
per ton milled, something between It and 2 being usual. Coghill and de
Vaney find ball wear to be proportional to the input of useful power, and to be
of the order of 0·15 lb./h.p. hour for steel balls.
In pebble mills sea-rounded flints or selected pieces of rock, between 2-!-"
and 4" in size, have long been used in grinding gold ore and in cases where
iron contamination must be avoided. At one time finely divided iron was
considered to be harmful to cyanidation of gold ore, and it was considered
that the use of large pieces of ore to complete the comminution of very small
pieces achieved several purposes-provision of crushing media, comminution
of the large pieces, and avoidance of iron in the cyanide section. When, as
with auriferous sulphides, the main purpose in grinding is to bring the
included gold to the surface of the sulphide particles, the metallurgical
Mineral Processing- Wet-Grinding Mills 101
considerations are quite different from those which precede efficient flotation,
and in many cases, plant economics favour their use. For many years the
Rand gold mines have screened 4" to 6" material (not necessarily of milling
grade) from the banket ore for this purpose. Where iron has been found
inadmissible, as in the final stage of grinding at Climax Molybdenum, pebbles
are used for the selective grinding of a relatively small percentage of the ore
after everything which could not benefit by this final grinding has been rejected
from the feed. This is a special case dictated by the metallurgy of the con-
centrating section. Another kind of special case is the preparation of iron-
free ceramics. The use of tungsten-carbide balls is reported, their high
initial cost being compensated by their high density, toughness and long
life.
Several factors have influenced the increasing use of autogenous milling in
current practice. The term, which is loosely used to include extraneous rock
as well as "pebbles" sorted from the mined ore, describes a crop load in
which selected larger pieces of ore are used to grind the more finely crushed
run-of-mine material. These are preferably of milling grade since they wear
down quickly and thus form part of the tonnage ground and sent for treat-
ment. Thus, they are part of the rated capacity and cost only the work of
selection (usually by screening), storage and controlled handling into the
mill. The general requirements for steel balls apply also to their efficient use.
Pebble size must be appropriate to the force required, in terms of the size
analysis of the mill feed. Since pebbles wear quickly, topping up of the
charge must be performed much more frequently than with steel, and handling
arrangements into the mill must be suitably designed. In one plant this is
today automated by a linkage between the meter which measures power
drawn by the crop load and the supply hopper gate where a reservoir of
pebbles is maintained.
The mill shape suited to rods and balls must be modified for the best use of
pebbles, and current developments are moving dramatically away from the
old concept of a long narrow mill to the opposite shape, as seen in
the Cascade or the Aerofall. Capacity and lift must take into account the
larger bulk occupied by pebbles of S.G. 2·7 than of steel (S.G. 7'9) and also
the logistics involved in moving a pebble of equivalent mass and much larger
size with the required force.
If the pulp is to receive chemical treatment the presence of abraded iron
may be undesirable. As much as 2 lb. or more of worn steel ball can join
the process feed with each ton of ore. Liner wear also can be considerably
reduced by the use of self-renewing liners which capture pebbles. These then
take the peripheral wear and the great bulk of the grinding is done by ore
on ore. Liner wear and lost time for renewal is also reduced. In grinding
with steel the cost of liners and balls is a major item."' I 0
Although the use of autogenous mills has not yet spread widely the savings
reported by industry show that where the ore is suitable for their somewhat
gentler comminution, they must be seriously considered in new developments.
Among recent studies of autogenous grinding are two Papers presented at
the 6th I.M.P.C. 11 . " The relatively gentle liberating action, lower power
consumption and wear, and elimination of preliminary crushing of a minus In
102 Mineral Processing-Wet-Grinding Mills
Capacity
This interdependence will be dealt with in detail when classifier action has been
studied (Chapter 10). Several other variables exist, but they need not now be
discussed. The interplay of these variables requires specific study in the case
of each ore milled.
The main cost items in grinding are for power and wear of steel. Power
consumption in a charged mill varies slightly between its fully loaded and
underfed condition, good and bad ratio of grinding media to ore and water,
good operation and bad. The economic concept of optimum capacity is
therefore associated with rate of wear of steel (balls and liners). From the
technical viewpoint capacity is qualified by product size (mesh of liberation
desired), which in turn is dictated by the requirements of the concentrating
section of the plant. The capacity is directly affected by the grindability of
the ore. Unfortunately, no fully acceptable definition of grindability has yet
been agreed on. It includes the qualities of hardness (or brittleness) and
toughness (resilient resistance) of the ore and the characteristic resulting
particle shape of the finished product (influenced by any tendency of rock
gangue to break along grain boundaries and of valuable metal-sulphides to
shear across them). To add to the difficulty, the purpose in grinding varies
considerably according to the process by which the ground product is to be
treated. This leads to diverse criteria of its suitability as a feed to the
concentrating process. In a given mine the association of the minerals in the
ore body and the grain of the ore may vary from stope to stope, with cor-
responding variations in the grindability. The effect in the mill is that some
ore "goes through" faster than other, thus calling for varying feed rate so as
to maintain equal loading. Bond's "Work Index" (see Chapter 3) gives a
useful empiric guide to grindability for an ore of consistent character.
Even if the mill is handled correctly it is still limited by design to a definite
order of flow capacity. "Flow" refers to the sum of new feed, return feed,
and water passing through the mill in given time. For good flow capacity,
the scoop must be generously proportioned and produce a pressure head
when revolving, so as to force material into the mill along the feed trunnion.
The latter should either be flared or given a helix at least half the projecting
height of the maximum particle fed in, in order to keep the feed moving briskly
into the grinding zone. At the discharge end, arrangements for evacuation
of products should be yet more generous so that a downgrade through the
mill is assured at all times. The combined area of the grate openings must
exceed the cross-section of the feed trunnion. The pick-Up scoops must be
able to draw the pulp-level well down toward the bottom of the grate openings.
In a high-discharge mill the discharge trunnion should be of greater diameter
than that at the feed end. It can be restricted when desired. It should also
flare outward to promote quick discharge.
There is a definite relation between mill capacity and speed. A speed of
just under 60% critical is favoured in many mills. Capacity falls off at high
speeds, markedly so in an observed case when 80% of the critical speed was
exceeded.
Much remains to be clarified in our understanding of the balance of the
forces at work. A Casc'\de mill at Vassbo' 0 which is equipped with variable-
speed drive has demonstrated the relationship between change in circulating
104 Mineral Processing-Wet-Grinding Mills
load and charge level. With increased toughness of ore the discharged sands
are coarser and the circulating load consequently increases, since the classi-
fying system sends them back. This increase is detected by a metering
device, and the speed of the mill is decreased. The charge level then rises
and the power drawn into the crop load increases somewhat. This produces
finer grinding from the extra kinetic energy available, and the mill discharge
product is correspondingly finer. The circulating load now falls and the
automatic control system speeds the mill up once more.
A limited amount of specialised wet grinding is still performed in pendulum
mills, such as the Huntington and Griffin, in roller mills like the Chilean
(developed from the primitive arrastre), and in grinding pans such as the
Cobbe. Descriptions of these obsolescent machines are given by Truscott. 18
References
I. Bond, F. C. (1959). Bull. Can. I. M. M. Aug.
2. Howes, W. L. Trans. A.I.M.M.E. 169.
3. Farrant, J. C. Trans. Inst. Chem. Eng. lIS.
4. Myers, J. F. (1953). Recent Developments in Mineral Dressing. I.M.M.
5. Davis, E. W. Trans. A.I.M.M.E. 61.
6. Bond, F. C. Trans. A.I.M.M.E. 153.
7. Prentice. T. K. J. Chem. Soc. S. Africa. 43.
S. Coghill, W. H. and de Yaney, F. D. (193S). Bull. Mo. Sch. Min. Tech. Ser. Sept.
9. Crocker, B. S. (1959). A.I.M.E. May.
10. Fahlstrom, P. H. (1962). World Mining. Sept. & Oct.
11. Jacobs, B. W. and Feik, J. (1963). I.M.P.e. Cannes. Pergamon.
12. Pasquet, M. and Joco, G. (1963). I.M.P.C. Cannes. Pergamon.
13. Truscott, S. J. (1923). Text Book of Ore Dressing. MacMillan.
14. Gilbert, I. and Wingham, D. W. (1963). Australian Bull. I.M.M. Proc. 207.
CHAPTER 6
Applied Power
Of the electrical power fed in, a loss of the order of 10% occurs in the
motor, and between 10 % and 15 % in the gears and mechanical friction of
the mill. The balance is available as "useful power"-as kinetic energy in the
tumbling crop load-but the fact that it is available does not by itself lead to
its efficient use. If the crop load is not properly constituted, part or all of this
kinetic energy will be wasted by conversion to avoidable heat, sound, and
ground-up metal.
When the mill is at rest the crop load lies more or less horizontally across
the lower half. If it slopes a little, the out-of-balance effect due to this slope
is being held by bearing friction. When the mill starts to rotate, much
extra power is needed to get the system in motion. The crop load must be
displaced in the direction of the rising side and its components must accelerate
from rest and build up their kinetic energy. As soon as normal running speed
has been reached and the system has settled down, the intake of useful power
falls to a steady level. The liners grip the load and carry it upward on the
rising side. Toward the limit of this rise, the upper part of the load breaks
~Load level
•. I I
Crftfca '-... --- at rest
-- -- -- --
·_·_·-·-:'--:::":"f .--~.-.-.-.
Li ne of movement of
. centre of mass
w
Fig. 47. MOI'ell/ent of Centre of" Mass of Crop Load
away and cascades or cataracts down to the toe of the load. This turning
mass is rotating about a centre W (Fig. 47). located somewhere along the path
Mineral Processing-Forces in Wet Grinding 107
shown at a point where it balances the useful power at work. The intake of
useful power is balanced by the displacement of the crop load. This power is
continuously converted to kinetic energy and from this into heat, sound, and
newly developed surface (of ore, balls, and liners). If R is the mill radius, W
the weight of the crop load, and a the horizontal displacement of the centre of
mass, then
Torque = WaR. (6.1)
Fig. 47 shows torque (in the sense of useful power) to be nil at zero and also
at critical speed. At critical speed (plus the extra speed necessary to compt.n-
sate slipping of the crop load and hold it by centrifugal force) grinding stops.
Such a position is largely theoretical, since it could not be reached in a
normally charged tumbling mill. The relationship between centrifugal
force and its radial tangential thrust at various depths along the mill's radius
passing through the crop load, precludes such seizure. The many inter-
acting factors at work in the churning charge tend to confuse the picture of its
dynamics. To clarify discussion, consider the case (admittedly over-simplified)
of a variable-speed mill with smooth liners, no ore or water, and a load of
steel balls of one size. As it starts very slowly from rest the load surface
tilts (Fig. 47) until the slope is reached where load stability fails and the
top layers of balls slide down. Neglecting slight irregularities in breaking
away, the power draft is steady for a given speed when the stable
displacement is at its maximum. This draft (ignoring mechanical loss
outside the mill shell) corresponds with the displacement of the centre of the
turning load from the vertical diameter by a distance a along the theoretical
path (zero to critical) in Fig. 47. As the mill speed is increased this dis-
placement reaches its maximum. With further speed increase the distance a
begins to recede and power input falls. The concept of freeze-up at critical
speed is not valid, in view of three main forces at work, of which centrifugai
fixation is only one. First, before this stage is reached any peripheral balls
rising clear of the down-slipping load after it has passed the plane of its
horizontal diameter fail to maintain tangential direction because they are
now acted on by gravity. Losing contact with the shell they take a falling
trajectory to the down-running side of the mill. In the course of this they
collide either with other balls, loosening the upper part of the charge, or with
the shell itself, thus transferring part of their kinetic energy back to the shell
from which they had received it. This acts against the input of new energy
once a balancing peak of flight has been passed. Second, the packing
structure of the charge changes steadily as mill speed increases. At rest, ball
rested on ball and voids between these spheres was at its minimum. With
rising speed the core of the charge, and also the upper layers, are loosened
so that the volume increases. For a given mill speed there is a critical
volume of charge at which the centre of mass a is at its maximum displacement
from the vertical diameter. If the volume of the charge had been sub-
critical at this speed it would have been possible to increase the out-of-
balance loading by adding more balls, and the power draft would have risen.
Similarly, it would have been possible to increase speed without increasing
charg~ to obtain the same effect. If, on the other hand more balls were
108 Mineral Processing-Forces in Wet Grinding
added, or speed were increased to the super-critical point, the combined effect
of the reduced out-of-balance dead loading and kinetic impact of falling balls
on the down side of the shell and the toe of the load would be to reduce the
useful input of kinetic energy to the system, and the current or wattage drawn
by the driving motor of the mill would fall. The third force, the consoli-
dation of the charge by centrifugal force, is modified by the first two.
The practical crop load is, of course, a mixture of grinding media of various
sizes and shapes, perhaps even of varied density, since steel and large pieces of
ore may form part of the crushing bodies. Next there is the ore, partly a
new feed from dry crushing and partly a return load of partly finished sands
from the closed circuit. Further, there is the make-up water added at mill
head, which modifies the plasticity of the tumbling load in accordance with
its specific surface and the percentage of water used. Finally, there are the
liners, of various types, shapes and degrees of wear, with grip modified by
changes in the amount of slime or slippery sulphide mineral anointing their
surfaces. Again, over-simplifying somewhat, the crop load can be pictured
as a loosely plastic body being continuously moulded into shape by tumbling
action and influenced in its mass cohesiveness by the frictional hysteresis of
its components. In most operating circuits the mill speed is fixed, but in all
the crop load varies slightly in volume with the grindability of the ore.
Since only "finished-grind" pulp is allowed to leave the grinding section
increased resistance of newly entering ore results in increased retention in the
closed circuit, part of which takes place in the crop load. If the load then
goes super-critical (in terms of the above discussion) less grinding power
is available, less grinding is done and the overcharge increases. If more easily
ground ore comes to the mill the charge is diminished and too much energy
seeks too little ore in the sub-critical loading which follows. The significance
of the Vassbo experiment, referred to in Chapter 5, in providing automatic
change of grinding speed geared to mill loading, can now be seen.
An equation for "best operating speed" (n) in terms of internal stability of
crop-load and frictional grip from the shell has been proposed by Da\ is.·
0.8158
n= revs/sec (6.2)
A[r: t{l+K"
where K = relr, (re being inner radius of charge and r the mill's radius). A
Paper by Guerrero and Arbiter' modifies this to include the effect of slip-
page.
In a ball mill the rate at which power is converted into kinetic energy is
fairly steady, but in a rod mill it varies somewhat abruptly owing to entangle-
ments, hold-ups, and momentary seizures of the rods as they turn.
It can be shown (the mathematical proof is developed by Rose and Sullivan) I
(6.3)
where D is the mill diameter. Gow et a'" find that in practice the exponent to
Mineral Processing-Forces in Wet Grinding 109
This mental approach enables him to see the problem of grinding realistically,
and to translate his vision into effective operational control.
time the specific surface of C decreases, though no change has been made
in the carefully controlled solid-liquid ratio which is being maintained in
relation to optimum specific surface. From this point there is cumulative
deterioration in the efficiency of comminution. Slowly the mill discharge
size analysis increases its percentage of coarse material. This discharge is
returned by the closed-circuit classifying system in increased volume though
there has been no increase in new feed of ore. Thus, both C's texture and
volume are changing from their optimum balance at an increasing rate. This
readily observed connexion between reduced power consumption and loss of
grinding efficiency (which may have adverse effects right down the flow-line
if it leads to overloading with wrongly ground material) underlines the vital
importance of a well-controlled grinding section.
Items a and b may now be related to this discussion. When mill speed is
increased without any other alteration C is reduced, and vice versa. This
fact was brought out clearly in the Vassbo experimental work referred to
earlier. There, the mill's best performance in overall terms was found to be
60% critical speed. Taking this 60% as the index of 100% efficiency, each
increase of I % of speed was accompanied by a drop of I % or so in efficiency.
At 80 %critical it was necessary to use 25 %more power to do the same amount
of grinding. This extra power was obtained by increased unbalance of the
crop load. The point is significant, since it shows that maximum unbalance
(accompanied by maximum draft of useful power) cannot be taken in isolation
as the criterion of efficiency. In this case it showed that the mill was too
big for its job and in consequence its diameter has since been reduced.
Maximum power draft must be related to the required finished grind, and
achieved by the correct composition and volume of C. In the Vassbo
operation autogenous grinding and a variable-speed mill were used, with
automatic linkage between speed and change in the volume of returned
circulating load. With ore as m change in size composition of m-components
is far faster than with steel balls, and variation is easier to arrange and ob-
serve. The work was done on a full working scale and tied in with the sub-
sequent treatment, so that effects on recovery could also be seen. The
inter-acting factors thus revealed should affect grinding research and develop-
ment.
A few further points may now be noted as accessory to the above discussion.
When the volume is steady, the net power is highest with the interstices
between balls full of ore and lowest with them full of water, which has a much
lower density and therefore reduces the total crop weight. Power used is
higher with plenty of "sharp" sand in the crop than with slimy sand only,
since the extra friction helps the liners to grip the load more firmly and raise
it higher. It is higher with a low-discharge mill, because pulp rises centrifugally
on the rising side but can escape near the periphery of the grate, whereas in
the high-discharge mill it can only overflow from the trunnion, so that a
larger volume of pulp must be retained on the falling side of the mill. A
drop in the ammeter reading of power input to a low-discharge mill might
show that the discharge grates were partly clogged.
The greater the diameter of a constant-speed mill, the higher is the centri-
fugal force and the stronger is the liner grip. More power is then used,
112 Mineral Processing-Forces in Wet Grinding
so available grinding capacity increases with wear of liners. The greater the
lifting grip of the liners, the more the power used. If the wave-contours of
the liners decrease with wear, the grip is reduced and this effect tends to
neutralise the one mentioned in the previous sentence. The higher the
density of the ore, the greater the power draft. With too little water in the
crop load a paste may be produced, lubricating the liners and reducing power
consumption because of increased slip. Power cannot be "pumped" into
tre crop load. It can only be drawn from the energising system, and then
only to the extent that the crop load is dynamically out of balance.
and so on. The figures in circles (Fig. 48) show the total number of cubes at
each stage of size reduction. For each halving of size, or mesh, the total
surface is doubled. With each reduction to cubes of half the previous linear
dimensions or mesh an amount of new surface equal to that already in
existence is produced. (It would be impossible, of course, to produce an
exact geometrical crushing of this kind by any milling procedure.) It will
be seen that when particles are reduced in even steps of mesh size then at
each size there is a geometric increase in the total surface area. Using the
accepted generalisation of Rittinger's law-that useful power consumed in
grinding is proportional to the amount of new surface created-there is
therefore a corresponding geometrical increase in the power needed from stage
to stage to give evenly spaced reduction of mesh size. In relating com-
minution to the effect it has produced. ohange in specific surface gives
Mineral Processing-Forces in Wet Grinding 113
valuable information. Specific surface is the total surface per unit weight of
dry ore, or for special purposes connected with transport and treatment,
per unit volume of a pulp.
In practice, the effect is that the power used to crush big rocks down to
gravel size is far less than that required to grind the same weight of gravel
T Q)
2
sizes to fine sand. The finer the finishing mesh required, the more rapidly
the cost rises and the greater must be the provision of fine-grinding equipment.
Homogeneous rock all drawn from one source usually behaves consistently,
but ores subjected to mineral processing are never homogeneous, as they
carry at least two different minerals. Occasionally these constituents are
loosely associated and can be separated by differential grinding, which frees
lightly bonded grains of one mineral from a semi-cemented adherence to the
other or others. Usually the minerals of the ore are interlocked. Each has a
toughness and hardness (or grindability) of its own, so that a solid lump of
ore is weaker at some points than at others. The percentage of each mineral
114 Mineral Processing-Forces in Wet Grinding
Grinding Objectives
Methods of concentration vary in what they demand as suitable preparation
of the feed. From one point of view, grinding spends money to produce new
surface, and from another, it produces specified sizes. As little new surface
should be generated as possible, consistent with the production of the correct
mesh of grind.
Broadly, ore concentration requires one of three types of grinding pre-
paration. For gravity concentration the desired constituent of the ore must
be liberated at the coarsest practicable mesh, and overgrinding or "sliming"
must be avoided. In chemical extraction the constituent to be dissolved
must be adequately exposed at the surface of each particle. Here over-grind-
ing does not hinder recovery, but usually improves it. For froth-flotation,
a particle not larger than 200p. to 300p. or srnaller than 5p. to lOp. is usually
required, with the desired mineral exposed at part of its surface. Hence,
the adopted grinding practice depends on the method of treatment which is
to follow. Where mixed methods are used, grinding may be done by stages,
linked with stage withdrawal of concentrate or gangue. In such a case a series
of grinding objectives may be pursued, each suited to a stage of treatment.
Comminution of Particles
The kinetic energy in the crop load can be applied to a given particle by:
(3) Shear and abrasion, the particle being dragged between spinning
balls.
(4) Impact of falling ball.
(5) Shock-wave transmitted through crop load by falling balls.
The particle may be seized and dragged between balls if the angle of nip
permits, in which circumstances it is subjected to pressure, shearing tension,
and abrasion. The particle may receive impact either through a direct blow
when it is resting suitably on a ball or liner, or when a shock-wave is propa-
gated at the toe of the charge by a descending ball and is transmitted deep
into the load.
In quiet conditions particles tend to settle on the most stable base. In a
fluid current there is a tendency to tum so that the minimum cross-section is
opposed to the stream. If a particle is hit when lying flat, any natural
tendency to become flaky is acrentuated. Should it be hit edgewise, so that
the maximum impact is applied to the minimum area, a higher shattering
effect results.
The mobility of the particle in the crop load is a determinant of its behaviour
in the grinding zone. When the mill is run with too dilute a pulp the ore
does not coat the metal surfaces properly, so grinding force is wasted, ball
hitting ball. At the right consistency the balls are properly coated by a
clinging layer of particles and a maximum amount of kinetic energy is directed
against the ore. If too little water is used, the crop load becomes sluggish
and pasty, kinetic energy is wasted in overcoming viscosity instead of being
available for comminution, and material remains sandwiched between two
grinding bodies instead of continuously making way for a fresh supply.
Too much fine material in the crop load, such as primary slime, clay, over-
ground friable sulphide, or gangue, is sometimes a nuisance, since the metal
surfaces become coated with a film which acts as a lubricant, increasing the
slip and reducing impact grinding.
The hardness of the crushing media is important, for on it depends the
crispness of delivery of the crushing blow. When pebbles or lumps of ore arc
used instead of steel balls they both grind and are ground. Their shape
tends to become spherical while they remain large, but it is unlikely that
sands are entrained between them by spinning, as occurs with balls. At what
is called their critical size, pebbles become too light and small to act as
crushing media, and for grinding purposes become tramp oversize until further
reduced. An autogenous charge is more angular than one of steel balls, .
and tends to lock and break away crisply rather than to slide under comparable
conditions.
One element in the conversion of torque into kinetic energy is the peripheral
speed of the mill. The rate at which the mill revolves is fixed for any given
gear ratio in the drive, but can be altered. The other determinant of peri-
pheral speed (27Trn where n = r.p.m.) is the internal diameter, which increases
slightly as the liners wear thin. Rotation of the mill causes the crop load to
116 Mineral Processing-Forces in Wet Grinding
climb to a slope exceeding its angle of repose. This angle of repose is higher
than it would be for a heap of static material, because centrifugal force is press-
ing the whole load outward. The higher the peripheral speed, the higher the
load is carried before its outer layers free themselves and fall, and the greater
is their kinetic energy on arrival at the toe of the charge.
The three types of tumbling action that can be produced at successive
increases of speed are called cascading, cataracting, and freeze-up. At
cascading speed the top of the load turns quietly over as it emerges, and
rolls down to the toe of the charge. At a higher speed (cataracting) some
rising balls are thrown clear of the charge and then fall parabolically toward
the toe, where they apply shock-loading at the point they hit. 8etween the
cascade and cataract speeds is an intermediate one, sometimes referred to as
avalanche speed, at which grinding is effected mainly by abrasion, but to
some extent by free fall of the crushing bodies. The final, but for all practical
purposes unattainable and theoretical, stage is freeze-up, which occurs when
the mill is turned at more than its critical speed (v' =gr), r being the radius
from the axis of the mill to the centre of a ball lying on the shell in the trough
of a liner depression. Slip prevents freeze-up from becoming complete under
working conditions. The concept of peripheral speed only applies to
material in contact with the body liners. Hence, only the outermost layer of
the charge is in a good position to attain full cataracting action. Some
differential action occurs in the Hardinge mill, since peripheral speed varies
along the cone (Fig. 30).
Not only is the freedom with which the ball descends important in deter-
mining the kind of grinding being done, but also the point to which it descends.
For efficient impact the flying balls must fall well inside the trough of the crop
load in the 8 o'clock area shown in Fig. 38. If the balance between crop
load, diameter, and mill speed is wrong, the balls may be thrown too far
horizontally and batter wastefully against the down-running breast of the
mill. Then metal grinds metal. Shatter produces more new surface than does
abrasion for the same input of energy since there is less frictional loss. In
Fig. 38 the action in the various parts of the crop load is shown. In the
segment A-A to 8-8, starting roughly at the 6 o'clock position and rising
to a little before breakaway point, the load rises quietly, with a certain
amount of ball-spinning due to slip, but with comparatively little grinding.
Next, from 8-8 to C-C (the latter is the area undergoing bombardment
at the toe) comes free fall during which the ball converts its potential into
kinetic energy. If the fall is cascading, much of this energy is used in abrasive
grinding during the descent. At the toe is a churning mass where the bulk
of the grinding is done. Not only is there direct contact between falling balls
and pieces of ore lying above balls and liners, but ore is packed between balls
and receives the shock transmitted through the C-C to A-A segment by
the continuous battering. At the extreme toe of the load the descending side
of the mill continuously folds the churning mass at the toe into the crop, and
carries it up to the breakaway point. The material next the liners is held with
more firmness than is any more central part of the crop load. The larger the
dropping ball or lump of ore, the less able it is to penetrate the charge near the
toe, and the more likely it is to be worked out to a point where it is seized and
Mineral Processing-Forces in Wet Grinding 117
under-run by the liners. Thus there is a tendency for balls and particles to align
themselves concentrically with respect to the centre of rotation of the crop load,
with the largest outermost and most stable during rise and the smallest most
loose at the charge centre. This alignment causes the largest balls to be lifted
at full peripheral speed and to be thrown outward from the breakaway point
with the greatest force. The dynamic conditions in the crop load can therefore
be pictured at a cross-section through a cylindrical mill. Outside, and travell-
ing at a little less than the peripheral speed, are most of the biggest balls and
pieces of ore. They are last to leave the shell at breakaway point because they
are held to it by the inside layers and also by the extra centrifugal force
working on them in this position of maximum mill radius and minimum load
slip. The cataracting effect required must be applied in terms of these
largest balls, which have the maximum individual mass, velocity, and inertia,
to ensure that they do not miss the proper target (line C-C at the toe) and
carry on to hit the liners. Next, by stages inward through to the more or
less oval-shaped part of the cross-section, come smaller and smaller balls
and pieces of ore. These have less superincumbent weight resting on them
and therefore slip more easily. They are circling more slowlY'and this lessen-
ing of peripheral speed is accompanied by increasing slip. Thus there is less
centrifugal force available to maintain the inside of the crop load in its climb,
so it sags away from the rising breast of the mill earlier than the outer part
of the load. Between A-A and B-B the whole load is in its most compact
condition and brings most of its weight and centrifugal force to bear on any
ore being abraded during slip or spin of the balls in this section. Above
B-B and also above the centre of turn the texture of the crop load opens
outward and upward, becoming progressively less dense as it gets away from
the centre. This again favours flight of the largest balls. In all this the effect
on the peripheral speed of the crop load produced by liner contour is obvious.
In plant practice despite some increase in slip and abrasion a relatively
high speed low-discharge mill is usual in primary grinding, followed by a
lower speed for secondary grinding. Mills can be geared for two speeds
where a variable tonnage input must be handled, though only one of these
speeds can be fully efficient. In the Hardinge cylindro-conical mill (Fig. 30)
the diminishing peripheral speed in the cone leading to the discharge trunnion
is responsible for a change-over from shatter to abrasive grinding. Partly
"finished" material (i.e. ore almost fine enough to be released from the closed
circuit) tends to get its final reduction here, an effect helped by segregation
of most of the smaller balls in this cone. This segregation has the further
effect that the large balls are more free to develop shatter in the drum, where
they can act at their best peripheral speed on the newly entering (and there-
fore coarsest) feed.
The foregoing discussion applies to mills run at below critical speed and
with a heavy crop load. Hukki 8 examines the effect produced when a lightly
loaded mill is run at super-critical speed. In Fig. 49 this is taken to 240%
critical with a 15 % bail load, and the maximum speed of the outermost layer
of balls remains constant. This means that the increas~d shell speed is
compensated by increased slip in the crop load, made possible by the use of
smooth liners, low discharge, and light loading. In the original research
118 Mineral Processing-Forces in Wet Grinding
i
speea-8S% of the Critical I Speed - 240 ~ of the Critical
8all Lood 50'7. I Boll Load 15"
I
I
I
I
I
I
I 200 em/sec B
u!'>OO I !>OOem/sec
III I . ~\\,
I
'"
........
E-
"~e"lfsec I
I 00\
, \,.",e ao"
'" ..
MInImum
u I ~ ?Qee Speed Difference
1l,)(JO I .9'r\e{'o between Liner 5 urf ace
III I ~ef\ and Ouler 8011 Layer
a. A~I~____~____________~~
II)
100
Maximum Speed of the Outer Layer of Balls
100
o 100 100
Mill Speed ,1;of the Critical
Fig. 49. Grinding Action at (A) Sub-Critical and (B) Super-Critical Speed
(after Hukki)
work this, while increasing abrasive grinding, had the drawback of heavier
wear on liners and balls. Despite the low crop load output of finished
material rose steadily in the experiments quoted (Table 7). More recent
tests with a large pilot mill have shown that when ore is used as grinding
media and mill lining, grinding at super-critical speed indicates a capacity
considerably above that of any sub-critical mill. Test work is still at the
experimental stage but it has been proved that no centrifuging takes place
during wet grinding at 150% critical speed.
NUMERICAL RESULTS OF GRINDING TESTS IN PILOT PLANT ROD AND BALL MILLS
Test No. 1 2 3 4 5
Type of mill used . Rod mill Rod mill Rod mill Rod mill Rod mill Ball mill
Mill speed, % ofthe cri(ical 48 72 96 130 130 200 s::
22·5 22·5 22·5 22·5 20 14 ~-
Tumbling load, % of mill vol. b
Dry feed, kg/h . 2030 2980 3920 5050 2000 ~
Motor, kVV--rpm . 11--1450 11--1450 11--1450 25--2940 11--1450 25--2940
Energy used, kWh/t 2·74 2·83 2·97 3·15 6·3 6·16 a ~
Screen analysis Pro- Diff. Feed Pro~ Diff. Feed Pro- Diff. Feed Pro- Diff. Feed Pro- Feedb Pro- Over- ~
Feed duct duct duct duct duct b duct flow ~-
Cumulative % undersize
--- I
6 mesh 33·2 99·5 66·3 40·6 98·6 58·0 38·2 97·5 59·3 42·2 98·8 56·6 34-4 99·7 99·8 ~
8 28-8 97·8 69-0 35·5 93·5 59·8 32·8 92·4 59-6 36·8 96·1 59·3 29·7 99-2 99·7 ;:;
10 25-3 93'7 68-4 31·0 87·9 56·9 28·3 83·6 55·3 32·3 90·2 57·9 25·5 97·9 98'9 ~
14 21·7 84·6 62·9 26·7 77-2 50·5 24·2 72·2 48·0 28·0 81·2 53-2 21-5 94-5 96·8 :-;-
20 18·7 72-9 54·2 22·8 66-2 43-4 22·2 61-8 39·6 24·3 70·3 46·0 18-2 87·7 92·1 ~
28 15-8 61-5 45-7 19·3 55·9 36-6 17·0 52·8 35·8 20'7 60·5 39·8 15·0 78-0 84·8 Cl
12·8 67·6 74·1 93·0 100·0 ....
:-;.
35 13-2 51·8 38-6 16·4 47·6 31·2 14·5 45·2 30·7 18·0 52·2 34·2
48 10·7 41·8 31-1 13-2 39·0 25·8 11·8 37·4 25·6 14·8 43·7 28·9 10·3 55·8 56·4 81·3 99·5 ~
~
65 8·4 33-7 25·3 10-4 31·6 21·2 9·2 30·6 21·4 11·8 36·2 24·4 8·0 44-8 33·2 61·0 95·7
100 6·5 26·2 19-7 7-7 24-8 17·1 7·1 24·3 17·4 9'4 29·2 19-8 6·0 34·2 13·7 38'5 81·9
150 4·9 21-0 16·1 5-8 19·8 14·0 5·2 19·4 14·2 7-5 23·8 16·3 4·5 26-2 5·7 26·8 65·4
200 4·0 17·8 13-8 4-5 16-6 12-1 4·0 16·5 12·5 6·0 20·5 14·5 3·6 19·1 3'7 19'0
I 51'7
a Figure is based on original feed of 2000 kg/h. b Classifier sands including circulating load.
-
::0
120 Mineral Processing-Forces in Wet Grinding
practicable and any attempt to reduce all the feed to finishing mesh in one
operation would result in wasteful over-grinding. The circulating load
built up by running the mill in closed circuit with a classifier solves the diffi-
culty.
The basic requirement of this arrangement is that the ore shall move
fairly rapidly through the mill, dwelling there for as short a time as a minute
if mill design permits. The shorter the passage (or dwelling time), the less
grinding work is done on it. Since each fall of a ball should produce some
finished material, the sooner this is removed the less it will be exposed to
over-grinding. Hence, low-discharge mills with generous feed and discharge
arrangements, and big diameters in proportion to their lengths, have supplan-
ted the earlier high-discharge long mills through which it was not possible
to push feed at a sufficiently fast rate.
Experimental work has shown that the faster the ore is rushed through
the mill, and the greater the classifier capacity, the more effective is the
grinding work, both in terms of output and in avoidance of over-grinding.
There is a limit to the practical application of this, since it costs money to
transport a heavy circulating load. This limit is usually found to be between
5 : 1 and 6 : 1 as a ratio between return sands and new feed in primary
grinding. It is lower in secondary grinding, being limited by the sorting
speed of the classifier.
When working at a 5 : 1 ratio, the mill receives one ton of new feed and
five tons which have already passed more than once. If the feed enters at
---i and is released from circuit at -60 mesh (= 1/ 120 the ratio of re-
H
H
),
duction for new feed is 90. Since the average particle passes through six
times (once as new feed and five times as return sands), the true ratio of
reduction is 6 : 90, or 15. Owing to the tendency of grates to hold back
large pieces, and for small sand to work out to the discharge end more easily
than coarse ore, the actual condition in the mill is not accurately given by
the~above figure. Much, if not most, of the return feed has been ground
nearly to finishing mesh. (It is characteristic of grinding in closed circuits
that what is called "near-mesh" builds up in the circulating load.) Only
a small proportion of the ore in the crop load consists of maximum-sized
pieces. The average size of feed being thus kept steady, grinding itself
becomes steadier and more controllable.
The effects of closing the grinding circuit with a large return load include:
(i) Reduction in the mean size of entering feed.
(ii) Marked increase in the circulation of nearly finished material.
(iii) Decrease in the retention and over-grinding of finished material.
(iv) Shorter dwelling time for finished material.
(v) Less need for ball ratio extremes.
(vi) Better interstitial loading of the ball charge.
(vii) Closer adjustment of the solid-liquid ratio.
must form a coating on liners and balls. The pulp must also be sufficiently
fluid to flow steadily through the mill. At a suitable ratio of particle surface
to water, the smallest particles settle slowy and behave as if they were part of
a heavy fluid, the proportions of water and ore determining its specific
gravity. This, combined with the effect of pseudo-viscosity retards the settling
rate of ore particles in the mill, although with the intense agitation in the
tumbling load there is very little settlement.
With too thin a pulp (too Iowa solid-liquid ratio) the solids tend to settle
and centrifuge outward, and coating of the balls by solid particles becomes
patchy to non-existent, while the voids between the outer layers become
overpacked. Even pulp distribution through the crop, and a clinging layer
of particles on the metal surfaces everywhere in the mill, are essential to good
grinding and reasonably high capacity.
Finely ground dry rock flows through a ball mill as though it were fluid,
and has good transporting power. When the moisture content rises beyond
8 % a stiff mud is produced, clogging the mill. This condition prevails with
moisture up to 15 %, beyond which fluidity begins to return, the material
behaving in a sticky, treacly fashion. From 20% onward efficient movement
begins to show. When a ratio of about 40 % water to 60 % ore is reached the
pulp is in danger of being too watery to coat the grinding media properly.
This is the character of the pulp in general terms, but it is modified by two
other factors-average particle size and density of ore. Solid-liquid ratio
is usually measured by the mill operator in terms of weight of a known volume
of pulp. The greater the density of the ore the smaller will be the volume
of solids needed to maintain a given percentage. The fluidity of pulp depends
on the amount of surface friction between the particles in a unit volume and
is therefore governed by its specific surface. Consider a pulp of solid-liquid
ratio 75 : 25 by weight, the solid fraction being ore of S.G.3.0. 75 grams of
this ore will occupy 25 c.c., and so will the accompanying 25 grams of
water. The volumetric ratio of this pulp is therefore 1 : 1. If the ore
consists of a stone weighing 75 grams, it will have very little surface and will
sink swiftly to the bottom of a vessel containing the 25 c.c. of water. If
this lump were cubic it would be nearly 3 cm. along each side and would have a
total surface of about 50 sq. cm. to rub against the water as it settled down. If
the ore were now pulverised into fine sand. the total surface of the resulting
particles would be measureable in tens or hundreds of square metres, and
when they were dropped into the same 25 c.c. of water the friction between
the faces and the cling between particle and water would slow down the
settling rate very markedly. Later in this book the word "pseudo-viscosity"
is used to described slow settlement due to such friction, which is a function
of the total surface per unit volume of pulp or the specific surface. True
viscosity refers to the internal friction in a fluid arising from its molecular
cohesion. When finely ground ore is moistened it coheres to form a sticky
mud. As more water is added the specific surface is reduced and particle
mobility increased. Thus, pseudo-viscosity varies with the area rather than
with the weight of solids involved. The flow characteristics of ore pulps
are factors in most processing treatments. "Pulp ratio" defines the per-
centage of solid by weight, as does "solid-liquid ratio" Finely ground
122 Mineral Processing-~Fo,.ces in WeI Grinding
particles in movement respond far more to specific surface effects than to those
produced by their collective or individual mass.
A coarse pulp might be very fluid at 70 : 30 ratio, but if the ends of the mill
containing it were closed and a prolonged grinding were given so that the
ore was reduced to extremely fine particles, it could on opening up be too
viscous to flow until it had received further dilution. Thus, as the average
mesh of the mineral decreases the solid-liquid ratio should be reduced in
order to maintain optimum fluidity. This effect is produced by an increase
in specific surface during grinding.
The author has made practical use of this change, which in the Hardinge
mill is continuous along the discharge-end cone. Extra make-up water was
piped in through the discharge trunnion and delivered halfway down the cone.
Thus two solid-liquid ratios were maintained, the higher being in the drum
where speciflc surface was lower than near the discharge end.
In ordinary continuous ore treatment, an optimum ratio is established
at which conditions in the crop load are best satisfied. In working out the
correct ratio, care is taken to suit it to the desired coarseness of grind, so
that the pulp gives enough frictional grip to the crop load to help it to the
desired amount of lift during rotation, and to ensure that the grinding media
shall be properly coated. The thicker the pulp, within limits imposed by
t/:lese considerations, the less wear of steel will take place. The higher the
pulp density, the more ore passes through the mill in a given time provided
the through rate is kept constant, and the greater is the lifting friction.
A rod mill is usually employed in open circuit and grinds at a mesh coarse
enough to make its grinding action very different from that in the ball mill.
Rods are kept apart by relatively large pieces of ore, and are not usually
coated. The ore is not pulped, but is coarse enough to permit of rapid
settlement of the solids till an appropriate solid-liquid ratio is reached.
Supernatant water serves to transport ore through the mill. Thus. while
most ball mills are worked at a solid-liquid ratio of between 70 : 30 and
80 : 20, a rod mill may be run at some such figure as 30 : 70 ~~ regards ratio
at discharge, though if it is stopped and the conditions in the crop load are
measured, the apparent density there may be much higher. An important
difference from ball-mill loading is that small particles are either excluded
from the feed or flushed through, so that the coarse ones on which the rods
are working cannot coat the metal at the working pulp density.
Control
(b) Ore in ratio to the charge and limited as to the maximum size of new
feed and the circulating load.
(c) Water in ratio to the ore, in such a proportion as will maintain both
fluidity and good coating of metal.
The first control is concerned with sampling ahead of the mill, to prevent
entry of an undue amount of tramp oversize. (This, like "critical size"
worn autogenous media, is ore too large to be expeditiously broken down
by the crushing bodies, but itself too small to act as such a body.) In a
small plant this can be done visually, but methodical cutting of a head sample
is better, and at the same time yields a sample for assay of entering feed.
Next comes the question of ball ratio. Here the problem is to impress
upon the crop load the right proportions of abrasion and shatter, as judged
by the end product. Usually it is not convenient to alter mill speed or liner
contour, except when periodically replacing liners, but it is possible to vary
the dwelling time in the mill, the solid-liquid ratio, and the kind of blow
struck. Dwelling time is a function of speed of feed, since feed displaces
a similar volume at the discharge end. Solid liquid ratio is changed by
altering the setting of a water cock in the feed launder. The blow struck
depends, inter alia, on ball size (or weight). With other conditions held
steady, if the percentage of heavy balls in the charge is increased, the impact
grinding is also increased. The space available between balls is greater, and
there are less points in the crop load where ball bears upon ball through the
coating of ore. Hence abrasive grinding is reduced and impact grinding
increased. The ball size required, or the blending of sizes, depends on the
finished mesh-of-grind. The finer this is to be, the more abrasive grinding
must be developed.
Bond' 0 has produced a formula which relates ball size and work index and
which amends an earlier one in important particulars. It reads
W 1
B= [KF] t [s.C {is
s
t j
,where (6.5)
It is thus obvious that the actual adjustment of a circuit must take simul-
taneous account of all these factors.
Returning to the question of proportionment of ball size to ore size, this
is best worked out by trial and error, with study of any recorded experience
of other operators working with similar ore. There is so much difference
between ores in grindability, that no safe generalisation as to ratio can be
=
made. The D2 Kd formula already given is an excellent starting point.
The mill discharge should be sampled and the sizes of its solid content ascer-
tained by screening (methods are given in a later chapter). If more impact
is indicated, the daily addition of new balls necessary to make up for wear
should be adjusted so as to include a greater proportion of bigger ones. If,
for instance, a particular mill is "topped up" daily with one ton of 2" balls,
t ton of 3" balls, and ten 4" balls, the effect of changing this to.t ton of 2",
.t ton of 3", and 124" balls might be tried, the effect on the solid-liquid ratio
being watched. If the next week or so showed that too much coarse grinding
was now being done and not enough fine grinding, some adjustment in the
reverse direction would be needed. Since the purpose of grinding is to
facilitate efficient concentration, the mill manager would watch the effect
of this change on recovery in the concentrating section, and would be careful
not to make any other kind of alteration in the plant while this one was under
observation, so that its effect could be properly judged.
The grinding mill does not respond rapidly to change. The high circulating
10"ld has a steadying effect, smoothing out surges. Time must always be
given before the full effect of a change is visible, and this time may be any-
thing Lip to a week. Each change of one factor necessitates adjustment of
the other factors affected, and since this is to some extent hit-or-miss, it takes
a certain amount of patience and skill to rebalance the crop load after each
change. The golden rules in this kind of operational research are: "Go slow;
only alter one thing at a time, if possible; wait after each alteration for the
plant to settle down; only make each alteration little by little." It is far
better to come slowly up to an optimum grind by a series of small adjustments
lasting several weeks than to "over-modulate" so that the whole flow-line
is upset. Grinding and concentration must be watched simultaneously
during this very important business of gradual correction.
Since total grinding effect becomes visible in the form of new surface pro-
duced, and new surface can be found by measurement of the diameters of
the particles in a sample (sizing analysis), the laboratory control of milling
efficiency is carried out by measurement of the sizes obtained. For this pur-
pose, samples are collected at regular intervals and sent for test. The solid-
liquid ratio is also maintained at the required setting by testing the pulp
density. If an ore with a specific gravity of 3·0 is ground at 75/25 solid-
liquid ratio, then:
75 grams of ore in pulp occupy 25 c.c. space
25 grams of water in pulp occupy 25 c.c. space
100 grams of pulp occupy 50 c.c. space
In other words, the pulp density required is 2·0. If a can of known volume
is filled with pulp from the mill discharge until it brims over, and is then hung
Mineral ProcessinK~ Forces in WeI Grinding 125
D
Fig. 50 Etf£'cl of Overloading a Low Discharge Mill
duction of capacity. Case C, where power intake has dropped to 295 amps
shows the pulp running from the centre discharge, clear evidence of overload
approaching a height sufficient to upset the desired out-of-balance. If the
mill were run with no ore or water, even higher readings would result, but
the mill would not then be running quietly. In the case illustrated, the mill
had a ratio of balls to ore of 6 to 1, was loaded rather more than half full,
and average dwelling time was about a minute. 0·9 ton of ore was finished
per horse-power hour.
In a large plant one operator can control from twelve to twenty primary
mills, or from twenty to thirty secondary ones, a helper dealing with the daily
addition of new balls. Repair work is the concern of the maintenance gang.
A form of control sometimes practised is to vary the rate of new feed in
order to compensate changes in the grindability of the ore. For this purpose
the Hardinge "electric ear" can be used. This is a microphone, set to respond
to the noise made by the mill. The microphone controls a feeder belt, and
holds the feed rate steady at the point where the volume of noise emanating
from the crop load remains constant. If the noise increases, feed rate is
quickened, and vice versa. The microphone setting is varied by trial and
error until optimum setting is achieved. Useful improvement in capacity
is reported by users of this device.
References
1. Rose, H. E., and Sullivan, R. M. E. (1958). Ball, Tube and Rod Mills,
Constable.
2a. Gow, A. M. (1930). Trans. A.I.M.M.E., 87.
2b. Campbell, A. B., and Coghill, W. H. (1934). Trans. A.I.M.M.E., 112.
3. Davis, E. W. (1919). Trans. A.I.M.E., 61.
4. Guerrero, P. K. and Arbiter, N. (1960). Min. Eng., May.
5. Bond, F. C. (1959). Bull. Can. I.M.M., Aug.
6. Hukki, R. T. (1960). I.M.P.e. )Lond.), I.M.M.
7. Pryor, E. J., and Heywood, H. (1946). Trans. I.M.M. (London), 55.
8. Fuerstenay, D. W., and Somasundaran, P. (1963). I.M.P.e. (France),
Pergamon.
9. Hukki, R. T. (1958). I.M.P.e. (Stockholm), Almqvist & Wiksell.
10. Bond, F. C. (1958). Trans. A.I.M.M.E., May.
CHAPTER 7
DRY GRINDING
Preliminary
Though most ores are reduced by wet grinding before being processed,
some can better be ground and treated dry. Many minerals and synthetic
substances require size reduction only. Other grinding problems arise in
which chemical instability, contamination, corrosion or risk of explosion
call for special precautions, such as milling in an inert atmosphere or one
where moisture is undesirable or must be removed. In an arid country the
chronic shortage of water may dictate the use of dry grinding methods.
Where a dry end-product is called for and can be processed up to the re-
quired state without the use of water, dry grinding is to be prefeFred. Among
the raw materials thus treated are asbestos rock and "crudy", coal for pow-
dered fuel, cement clinker, talc, metal powders, drugs, and chemical salts.
In addition to open and closed-circuit grinding, batch treatment is frequently
used. In this method, grinding media and feed are loaded into the grinding
mill and worked dry until the desired state of attrition has been achieved.
The product is then discharged.
In the treatment of ores by chemical methods, such as the cyanide process,
experimental dry grinding has shown promise. When comminution is
followed by froth flotation it is usual to protect the newly developed mineral
surfaces, and this is best done by grinding under water to which an)
required protecting chemicals can be added. The technical applications of
dry grinding in mineral dressing are at present limited by this consideration.
Fixed-path Mills
Taggare classifies dry mills into two groups: "in which the comminuting
elements are relatively few and follow definite paths (fixed-path mills)";
and those in which "the elements are multifarious, and not constrained as
to individual paths (tumbling mills)". The latter do the bulk of industrial
dry grinding, but the former, of which there are several types, handle an im-
portant tonnage.
Burr mills range from the old-fashioned grindstone, originally used in
grinding cereals, to vertical types. Two discs of stone, either horizontal or
upright, are rotated in opposite directions, or worked with the lower one
fixed and the upper revolving. Feed is central and finds its way along grooves
in the stone faces, maintained by stone-dressing, to a peripheral discharge.
The material is ground by attrition during its journey, being dragged between
the stone faces. Soft rocks such as clays, barytes, talc, lime and limestone
128 Mineral Processing-Dry Grinding
and gypsum, are treated in these mills. Feed is minus iN and discharge can
be as fine as minus 200 mesh. The mills are used for grinding material not
likely to be injured by frictional heat, and also where staining by iron must
be avoided. Care must be used to keep hard or uncrushable material out
of the feed. Developments of this principle include vertical-disc mills with
steel grinding-faces. The laboratory disc-grinder is widely used. Its discs
can be parallel or slightly offset to one another, the latter arrangement reduc-
ing choke and improving throughput. Capacity is low.
The colloid mill has a vertical spindle on which is mounted a conical
grinding unit, grooved vertically. This rotates at high speed in a fixed, close-
fitting bowl. The grinding elements can be of ceramic or metal. This mill
is used for grinding foods and soft minerals.
Hammer mills and rolls were discussed with intermediate crushers. They
can also be used for fine grinding. A special application of the hammer
mill is in the crushing of asbestos. The requirements are unusual, in that
the material, as mined, carries the valuable fibre sandwiched between layers
of shale. Hammer mills with heavy manganese steel plates are used to
"fiberize" the blocky fibrous rock. The beating action opens the fibres and
loosens attached shale. This is thus reduced to a fine grit which can be
screened away.
from which the Hardinge disc roll mill has been adapted. In this two
conic-section rolls ride above a revolving horizontal disc. This disc table,
on which the ore arrives centrally, runs at a speed just below that at which
peripheral discharge of crushed material begins. A dam forms round the
circumference over which the discharging material is pushed by arriving feed,
and falls into a classifying air stream which lifts finished mineral and returns
anything coarser to the grinding disc. In the Lopulco mill (Fig.52) there
are either two or three spring-loaded conic rollers. These mills range in
output from I to 50 tons/hour and are designed for low, medium or high speed.
There is external provision for adjustment of clearance between the rotating
table and the rolls, which cannot make direct contact. The feed receives
both loaded crushing and shearing attrition. An exhaust fan maintains air
sweeping and removes finished product. In addition to its wide use in pro-
ducing powdered coal this mill grinds a variety of softish minerals substan-
tially through 100 mesh. The list includes gypsum, lime, phosphate rock
and various industrial earths.
The older pendulum, or roller mills include the Huntington and the Griffin.
In both, one or more pendulums revolve inside a wearing ring against which
they bear owing to the centrifugal force set up by their rotation. Material
trapped between roller and ring is ground till it escapes through guarding
screens, set peripherally. In the Williams mill three to five rollers are pressed
outward in similar manner, but the mill is swept through by a current of air
which carries to a collecting cyclone or air-filtering arrangement all particles
small enough to be borne along. The air or gas can be preheated in a furnace,
and natural draught is aided by an exhaust fan above the grinding compart-
ment. In the Raymond bowl mill, which is used for producing pulverised
coal, the bowl rotates against spring-loaded muliers and finished material
is removed by a current of air. The Babcock and Wilcox machine has balls
rotating in a horizontal grinding ring, where they press on and pulverise
material fed down into the ring, the finished product dropping by gravity
to an external classifier. In a variation, the mill is swept by air or gas,
hot if necessary, discharge being upward. Where air-sweeping is used,
hot air can can be used for the purpose of drying the feed.
Intermediate between the fixed-path and the fully tumbling mill is the
vibrating ball mill. This has not changed importantly since its prototypes
were developed in the United States and Germany in the pre-war period.
An industrial model, now in increasing use, as described by Parici0 3 is nearly
6' high, 9t' long and 7' wide. It weighs six tons and is vibrated by the un-
balanced rotation of eccentrics driven by two 50 H.P. motors at 1,200 r.p.m.
The layout consists of a grinding cylinder rigidly attached to eccentric
mechanisms in independent horizontal cylinders parallel with it. This
assembly is mounted on four sturdy vertical springs and vibrates with an
amplitude of 3/4". In action the grinding media have a vibrating period
of about 1,160 r.p.m. and occupy about 80% of the mill's volume. The load
Mineral Processing-Dry Grinding 131
tends to rotate. For suitable ores this system has proved versatile and cheap
in standing space, cost of installation, weight and maintenance. A 30" dia-
meter mill reduces minus f' soft and friable feed to 99 % minus 325 mesh
at a rate of 2t tons/hour. Feed and make-up media enter through a dust-
tight flexible spout and are discharged via a retaining grate. This mill
can also be worked wet, and the retention time can vary from one minute
up. When used for batch grinding two receiving tanks are used and the
material is worked through from one to the other and then back until the
required fineness is reached.
Vibration milling has met the need for grinding metallic powders to sub-
micronic sizes in an inert atmosphere, and handles such inflammable elements
in this way as aluminium and magnesium. Capacities quoted range, on
limestone and using steel balls, from 1 ton hourly in the 15" diameter mill
to 15 tons in the 42", an 80% minus 4-mesh feed being reduced to 80% minus
200 mesh. Where iron in the product is inadmissible, alumina balls and
special linings are available. Materials handled commercially include
tungsten carbide, aluminous nickel, silicon carbide and iron oxide.
Tumbling Mills
Dry ball mills have much in common with those described in Chapter 5,
but there are marked differences in the method of operation. Continuous
and batch grinding are practised, sometimes with rods as grinding media,
but more usually with balls or autogenous crushing bodies.
Since batch-treatment mills are not limited in design to types which can
CYCLONE.
be fed at one end and discharged from the other, a variety of shapes are in
use--cylindrical, conico-cylindrical, oval, polygonal, and even cubic. The
liners and crushing media can be made of iron or ceramics, according to the
requirement of the work. Feed and discharge are usually made via an
aperture in the shell, a screen being placed over it to retain crushing media
during discharge. The mill is charged, run for a suitable period, stopped,
opened, and emptied.
In continuous work, the rod mill is sometimes used. Mineral tends to
move sluggishly through, and can be aided in its progress by the highe~t
possible gradient from feed to discharge end. This latter usually takes the
form of a grate with peripheral discharge. If the feed end is shaped to a
ftattish cone, a reservoir of new material builds there and is pressed toward
the rods. The rods are not displaced by this action, because they are steadied
by the combined effect of rotation and of bearing against the discharge liners.
The ball mill is the most favoured machine for dry grinding of hard rock.
A critical factor in good working is the moisture content of the feed. In dry
grinding the discharge product may have to satisfy a maximum moisture
specification. If the feed enters with a water content much above I % the
air circulated through the charge by the air-classifying and sweeping arrange-
ments has increasing difficulty in drying the load during its retention period.
By the time some 5 % of moisture in the feed is reached, this air is unlikely to
pick up enough heat from the frictional action of the crop load movement to
do much drying unless it is given external heating. Some moisture-saturated
air may have to be bled off from the closed air-sweeping circuit in addition to
that which escapes under working conditions. A typical air-swept arrange-
ment is shown in Fig. 53. Details of one type of air classifier are given in
Fig. 54. Dry ore has a high angle of repose, making it desirable that progress
through the mill should be assisted. Feed is often forced in by a screw con-
veyor, and a steep gradient through the mill is obtained by the use of a large
diameter, short length, and peripheral discharge. Another critically impor-
tant factor in operation is the f~d rate. Owing to its low mobility, partly
finished material tends to accumulate over the toe of the crop load, where
it cushions the falJing balls and uses their kinetic energy to redistribute the ore
instead of to crush by impact and shock-wave propagation through the mass.
It is therefore important to restrict loading so that the ore in the crop load
remains in the interstices. The ball soon becomes coated with a layer of ore.
This robs the blow of its crispness, and damps down transmission of shock.
On the other hand, the coating layer protects the metal parts of the mill
against wear, and considerably reduc(:s this item of cost, thus enabling dry
grinding to compare economically with wet in suitable fields. One of the
best ways of avoiding loss of efficiency is "air-sweeping", in which a current
of air is blown continuously through the mill. This removes the finest
material and thus takes out of circuit the principal cushioning and coating
fraction of the ore.
The ball load is kept lower in dry than in wet grinding-below 40%-to
avoid "over-carry" at cataracting speeds, which would cause the flying balls
to hit the down-running side.
One of the older methods of ensuring fineness in dry grinding is to divide
Mineral Processing-Dry Grinding 133
AIR AND
fiNAL PRODUCT
,.....f-9li::.o".-POR T
CONTROL
AIR
LOCK
the ball mill into two or more compartments. These are separated by grates
which retain the material in the compartment for which crop-loading is
most suitable until it can pass through the grate and be elevated by
scoops to the next (Fig. 55). The Hardinge Tricone Compartment mill
uses the back slope of it first (conic) section to promote media segregation
and makes it possible to dispense with the third section shown in Fig 55.
In an air-swept single compartment mill, which has gained considerable
favour in the preparation of pulverised coal, the difficulty of maintaining a
gradient through the long cone of the Hardinge is partly overcome by using a
retaining grate toward the discharge end. Behind this, lifters pick up the
undersize and drop it into the classifying stream of air sweeping through the
mill.
Another class of mill has peripheral discharge through the shell, the
openings being guarded by screens. This mill is also used for wet grinding.
The Krupp mill (Fig. 33) guards the fine screens, first by heavy grinding-
plates which take the full liner wear. and next by coarse punched-plate screens
134 Mineral Processing-Dry Grinding
which hold back material that might damage the more delicate fine screens.
The Aerofall mill (Figs. 56 and 57) is either fully autogenous or run with
a small percentage of balls as part of the grinding media .. ·• It receives a
wide-ranged dry feed, anywhere up to 18" in size, and reduces it to a fine sand.
One mill works on --42" ore. A ball charge varying up to 2t% of 5" or
larger balls, can be used where the natural feed does not contain an adequate
proportion of heavy rock, but usualiy none are needed. The mill varies in
Mineral Processing-Dry Grinding 135
diameter from 5' to 28' and in width from 2' to 8', and works at peripheral
speeds between 8' and 20' per second. It can be used in open circuit or
closed either by screens or pneumatic classification, finished material being
removed by air-sweeping and sorted in cyclones. Successful use is reported
with asbestos, iron ore and gold.· As with the pebble mill in wet grinding,
ore is used to break ore, but the dynamics of the crop-load are very different.
Shape and speed of the mill are designed to carry its -\" contents high on
the rising side, and the bulk of the free-falling action takes place near the
6 o'clock position. The full kinetic value of fall is developed, since no pool
of water exists to cushion impact.
..
,, ...
• .
•
...
.
,
•,
..
-..
N
o
:I.
,
"If[ ,
•
Pilot tests' show one-step comminution to use less power than a conven-
tional crushing and grinding sequence and to favour medium and fine-
grained ores which require finer finished grinding, where air-sweeping is
more effective. Ore tends to disintegrate selectively along grain boundaries
and to produce mineral species at their natural grain size. Any such effect
assists subsequent concentration. Liberation appears better than with
rod-milling and wear is less. If selective grinding accompanies quick passage
through the mill the later treatment benefits by having less over-ground
particles to deal with. The Swedish tests suggest that when working for a
coarser liberation mesh. air-sweeping is not fully effective. Tests at Doorn-
fontein' showed higher recovery both of gold and uranium after 24-hour
laboratory leaching of mi/1l1.\' 200 mesh banket ore crushed direct from run-
136 Mineral Processing-Dry Grinding
of-mine feed than of various samples drawn from the conventional flow-line.
With a steady increase in the chemical extraction of values from their ores
this improvement, which again suggests disintegration along grain bound-
aries, attracts the interest of process research.
Under normal weather conditions Weston considers a moisture c('ntent
in the feed tolerable up to 3t %, but in sub-zero temperatures external heat
to be needed from It%. Installed capital costs are lower than for com-
parable wet grinding and maintenance is much lower.
The Cascade mill (Fig. 34) was described in Chapter 5. It is used in dry
grinding in the same general way:
Operation
The moisture content of the feed is an important factor in dry milling.
If nothing is done to classify a circulating load, a very small percentage of
water (t % to 1t %) ruins fine grinding. If the circuit is closed through fine
screens, a little more moisture can be tolerated without clogging the screen
apertures, unless the material is soft and clinging. The temperature of the
circulating air rises as the ore is milled, and its humidity increases when it
picks up moisture from the passing feed. Enough air must be bled off to
avoid saturation and thus ensure that this transfer of moisture can continue.
The heat requirement for drying changes with the season, and must be supplied
by warming the air when necessary. The water content affects grinding,
cling in the crop load, cushioning, cataracting, coating of crushing media,
and mobility and classification of partly finished material in the closed circuit.
As with wet-grinding work, the mill capacity required for a given rate of
throughput depends on the grindability of the ore and on the sizing analyses
of feed and finished product. The finer the grinding is taken, the lower will
be the tonnage treated. Removal of finished material is very important to
high capacity. If it is allowed to remain in the mill it deadens the grinding
force by packing the interstices. It diverts what should be grinding energy
to the task of re-distributing the crop load and of overcoming its frictional
resistance and cling. This, though it adds to the heat of milling, contributes
little or nothing to the comminution of the larger particles. Where a fairly
coarse final grind is wanted (in, say, the 20-mesh zone), screening affords
positive separation since at this size a lively open load can be worked over
screens without "blinding". (Blinding connotes the wedging of an oversized
particle into a screen aperture, thus putting it out of service, and it becomes
serious with clinging material on small screens.) This coarseness calls for
strong air-currents to sweep finished material out of the mill, and a mill
design which includes retaining grates followed by lifters to the discharge
overflow may be needed.
When the finished material is required at minus 60 mesh, reasonably
gentle air-currents can be cheaply introduced into a circuit without giving
rise to a serious dust or abrasion problem, so that between 10 and 60 mesh
screening gives place to air-sweeping. In many cases tests have shown that
the introduction of one or other of these methods into what had been open-
Mineral Processing-Dry Grinding 137
Application
Industrial dry grinding is concerned mainly with homogeneous materials,
whether naturally occurring or synthetic. In such work the problem of
subsequent removal of a selected fraction does not arise. Increasing use is
now being made I)f dry comminution as a liberating stage in ore processing,
even where subsequent treatment is a vital factor. If chemical or magnetic
treatment is to follow grinding there is little risk of trouble, but when froth-
flotation is to be employed the delicate surface-active forces which this process
exploits may rule in favour of wet grinding. This is discussed in later chap-
ters. Flotation following dry grinding is already in use.
One manufacturer (Hardinge Co. Inc.) has over 5,000 commercial install-
ations currently at work on nearly a hundred kinds of material. These
include ores of chromium, gold, lead, manganese, molybdenum, phosphate
and platinum. Other substances ground vary from such hard and abrasive
materials as carborundum and cement clinker to soft or clayey feeds like
coal, talc and ball clay. A common factor is the need to avoid the use of
water.
Mill Capacity
This section applies to tumbling mills generally, wet or dry. The capacity
is related to the mill volume, and to that portion of it occupied by the crop
load. In a cylindrical mill the diameter can be reduced by interposing wood
battens between the shell and the liners. The length can be shortened by
inserting a timber packing behind the end-liners.
If only temporary reduction of capacity is desired, reduction of the volume
of the crop load is a simple expedient. Big pieces of ore can be introduced
deliberately to replace part of the feed. They grind down very slowly, thus
producing the effect of a reduced crop load. It is also possible to arrange
for a mill to run at more than one speed, so that when needed it is held to
its slower rate of turn. Reduction of capacity by such methods is accom-
panied by lower operating efficiency, shown either as higher cost or reduced
recovery of value.
Capacity can be increased by building up the crop load till unbalance
reaches its maximum. It is sometimes feasible to increase the mill speed.
138 Mineral Processing-Dry Grinding
The circulating load can be increased if the classifier and mill trunnions
are able to handle a higher tonnage, thus reducing the amount of over-grind
and thereby increasing throughput. In the primary circuit, the liner contour
can be accentuated so as to give higher lift and greater impact-grinding,
and the diameter of the mill can be increased a little if thin alloy-steel liners
are used. In most cases the simple practical course is to throw more of the
work of comminution on to the crushing section, by closing the set of the
machines until a finer product is delivered. It may be justifiable to introduce
a further stage of intermediate crushing, by rolls or rod mill, which can be
put in circuit to further reduce particle size when the grinding plant is
intermittently overloaded.
Generall Conclusions
What follows applies particularly to wet grinding after dry crushing,
and summarises the discussions of grinding up to this point.
The finer grinding is carried, the lower is the capacity of the plant and the
greater the wear and tear. It is therefore of major importance to control
all comminution in such a way that the products are released as close as poss-
ible to the required release mesh. Most dry-crushing machines can be
worked as "arrested" crushers. A distant approach to arrested crushing
can also be achieved in the low-discharge ball mill by working with a large
re-circulating load. High-discharge ball mills cannot release finished mat-
erial so easily, since it must be displaced by entering feed and water, and not
sluiced out down a gradient.
Dry-crushing plants are subject to bearing wear if dusty conditions prevail.
This can be avoided by using an air-swept sealed circuit. When crushing
rock, safety precautions are sometimes needed to guard the health of the
personnel. For humane reasons, it should be borne in mind that a mill
management drawing its labour from a drifting native population is unlikely
to be directly in touch with evidence of any bad after-effects of dust inhalation.
Metal sulphides, which comprise the bulk of values in ores subjected to
treatment, are usually far more friable than their gangue rock and are there-
fore liable to receive selective comminution. This can be bad for concen-
tration, and may make it advisable to remove some of these sulphides as a
rough concentrate in the grinding circuit itself. If gravity methods of con-
centration are being practised, stage grinding can be interlinked with stage
concentration to minimise such harmful over-grind of values. With flotation
to follow, the problem may not arise.
When the mill is receiving an ore carrying two or more valuable minerals,
it may happen that one of these is liberated at a coarser mesh than the other.
This could be an occasion for stage treatment. The usual practice calls
for a high-grade finished concentrate and an impoverished tailing, with
regrinding of insufficiently liberated middlings. In gold treatment, metallic
particles are frequently trapped out at the earliest possible stage in ore to
avoid locking them up in the interstices and quiet crannies of the milling
circuit-sometimes ironically called "gold-plating the mill". With extremely
Mineral Processing-Dry Grinding 139
References
1. Taggart, A. F. (1945). Handbook of Mineral Dressing, ChapmaI1 & Hall.
2. Robertson, R. H. S. (1960). Chemical Age, June.
3. Paricio, R. (1960). Mining World, June.
4. Weston, D. (1960). I.M.P.e. (Land.), I.M.M.
5. Fagenberg, B., and Ornstein, H. (1960). I.M.P.C. (Lond.), I.M.M.
6. Waspe, L. A. (1955). Mining Magazine, Nov.
7. Waspe, L. A. (1956). Mining Magazine, June.
CHAPTER 8
(a) Screening.
(b) Elutriation.
(c) Sedimentation.
(d) Infra-sizing.
(e) Microscopic examination (including electron microscopy and electron
diffraction).
(f) Turbidimetry.
Mineral Processing-Laboratory Sizing Control 141
(g) Permeability.
(h) Gas or liquid sorption.
The more important routine methods employed in mineral dressing are de-
scribed in this chapter.
Laboratory screening can be carried down to 37 microns. The micron
(fL) is IjlOOOth of a millimetre. The millimicron (mfL) is IjlOOOth of a micron,
or 10 angstroms (A). Routine sizing tests on laboratory screens usually
finish at about 70fL. "Sub-sieve" sizing is practised in the - 70fL +5fL range.
Microscopic inspection is applied at all sizes from sands down to 0·25
micron. Centrifuging of extremely fine suspensions is used down to 10
millimicrons, and the electron microscope can be employed in suitable cases
down to a millimicron (1mfL).
Purpose
The purposes of laboratory sizing control are to check the quality of the
grinding, the extent to which the values are liberated from the gangue at
various particle sizes, and to aid specific examination of ore constituents'
Size analysis of the feed transferred from the crushing to the grinding
section shows whether the former is doing its work correctly, and how much
comminution remains to be done by grinding. In the wet grinding section
samples from the mill discharge show the size distribution of the delivered
particles. Samples taken from the closed-circuit return show what sizes of
material are being sent back for further grinding. Those from the closed-
circuit discharge check whether properly liberated particles are being re-
leased to the concentrating section of the plant. In addition, sizing tests at
selected points help in checking the progress of the material through the
various stages of treatment. Study of the sizing analyses of all these samples
gives the plant manager clear information as to the way in which grinding
power is being used and how grinding affects recovery of the values from
the ore.
The concentration processes applied to the pulp exploit selected differences
between particles of valuable mineral and gangue. If two particles of the
same size are produced, one being composed of heavy mineral and the other
of light gangue, the heavy one has more potential gravitational energy than
the light one. If both start to fall together, the heavier one will convert
more of this potential to kinetic energy in a given distance. This difference
in behaviour is used in "gravity" methods of separation, which exploit
differences in mass and shape of mineral particles. By screening, the size-
variable can be removed, and the density-variable accentuated. In other
concentrating processes size control is used to find when the optimum size
had been reached, when to stop grinding, and when the pulp is ready for
treatment. In the mill the problem usually takes the practical form of
adjusting working conditions in order to obtain the maximum tonnage at
the required liberation size as cheaply as possible, while using the full capacity
of the grinding plant. This problem is solved by empiric alterations, guided
by laboratory sizing analyses of samples drawn from the mill circuit.
142 Mineral Processing-Laboratory Sizing Control
PLAN .~-".4.t.'
MM. J JI J 11111I1111111111111111111111111111
HEV. • • • ~ ~ . . . . .~ . . - . -
Laboratory Screens
Sieves used in ore testing are square meshed. The wires of the warp
are woven successively over and under those of the woof (Fig. 59). The
"mesh" is defined as the number of wires (or of openings) per inch, measured
along either warp or woof (not diagonally). Since the main systems used
in the English-speaking world differ in the thickness of the wires used, the
apertures framed by these meshes also vary appreciably. The chief American
systems are the Tyler and the U.S. Series (A.S.T.M.). The leading British
systems are the I. M. M. and the B.S. I n many laboratories an essential
operation is to separate the sample being tested into a graduated series of
sizes. This is done by assembling a "nest" of screens of diminishing mesh-
size. A widely used reduction ratio is provided by the Tyler series (Table 10,
col. I) which is characterised by an arithmetical progression (y2 =1·414).
The 200-mesh Tyler screen has a distance between wires of 0·074 mm. or
144 Mineral Processing-Laboratory Sizing Control
L':~4]'
74p. (Table 10, col. 4). The next larger screen in the series is therefore of 104p.
aperture (74 x 1'414) and the next smaller 53p. The wires used to
form the screen "cloth" as it is usually called are of a diameter complementary
to this spacing. A 4 y'2 series is also made for close sizing. The A.S.T.M.
series uses the y'2 ratio, but begins from I mm. or 18 mesh. The new U.S.
sieve series (A.S.T.M. E-II-6l) has the relationship of openings, wire
diameters and mesh sizes shown in Fig. 60. Dimensional details are listed
~=
...,9 J,~ ~~ ==
o.h N ~ ~~-=
Is NM - ",'!':l!"''''
II'tC"illN .. 9~,, _
~ 0'" 2
- - 1-- "'9 =- -
...':> ~:co t;-~~~
'" -
~ _ N-
....
..
...
.
u:»~ ,;.
N o co,u,
E 1·0 ~~ .... 0'"
E 0,8 ~- 0'"
ON
tI
on
0-6 =>
'" 0
~,
o - ·
... ... ...0 .;..
L-
a. IH ~
Qj
'" ~ 'illl
E
a 0·2 r-
~N
...
:'l '"'"
:c
0
U.S. Si eve Serie5 ASTM E-11-61
QI 0·1
0-08
- ::: - Cu r ve 5how ing r ela ti on5h i p
.... :;l. -~ = -
L-
;:!
:: 006 of
apen i ng5 to w i re d i ometer5
'~f,
OO~
(}{) 1
N
II I N
III o
I ,III
0 0
I111I1
0 0
o .:, N
6
in Table 8. It corresponds fairly well with the Tyler series, with which it
can be used interchangeably. The I.M.M. series is sometimes considered
as obsolete and superseded by the B.S. series, but it is in still wide use in the
British Commonwealth. The important difference from the Tyler system
is that in the I.M.M. series the wires have the same diameters as the spaces.
Thus, a 200 mesh I.M.M. sieve has 200 wires and 200 openings in one inch,
measured along either the warp or the woof, so each wire must have a di-
ameter of 1/400". This gives a spacing of 63 microns. It follows that the
description "200 mesh" can refer to 741-' or 631-' screen cloth, unless the system
is specified. The I.M.M. system has the convenience of allowing the user
to work out the mesh spacing by mental arithmetic. A series can be selected
which gives a rough approximation to a Tyler series. The B.S. series which
is taking the place of the older LM.M. conforms to the Tyler series, with
slight variations in wire diameter. The smallest woven-wire screen is 400
mesh or 37 microns, and a tolerance for weaving or wire-drawing error is
permitted. The old German system (D.LN. 1171 Standard) designated
sieves by the number of meshes per cm". This has been superseded by
D.I.N. 4188, the mesh dimensions for which are listed in Table 9, together
with those for France.
A source of error with the wire-cloth laboratory screen is irregularity of
weaving. Any displacement of wire or change of wire diameter results in a
number of meshes of the wrong size. Unfortunately, a row of meshes too
small (which would not matter much) is offset by a corresponding row of
oversized meshes when a wire is out of position. Given time, every particle
small enough could find its way through a very few such holes. Sieves,
particularly the finer sizes, are delicate instruments of precision and should
be carefully handled. When used for wet screening, they should be dried
gently, immediately after use. If they become "blinded" by particles stuck
in apertures, they should be freed by light brushing of the underside with a
camel-hair brush. The use of metal discs and rubber balls to force material
through a laboratory sieve is thoroughly bad, as is any other overloading or
rough handling. A master set of sieves should be used for periodic checking
of the work. Wear and tear will thus be revealed, and damaged sieves
can be removed.
Although described above as square-meshed, the effect of the weave (Fig.
59) is to produce a curved set of wires. Until a near-mesh undersized
particle falls centrally with its minimum cross-section normal to the aperture
it will bounce off repeatedly, unless it becomes wedged. "Blinding" is min-
imised by taut mounting of the cloth, which should be laid on a wooden block
and carefully soldered into its eight-inch frame, the soldering being finished
with a smooth down-slope to the sieve so that no particles are liable to lodge
in any crevice between cloth and frame. The fixing of a new cloth is best left
to experts, as a slack screen gives bad results, while wrong handling in
fitting may distort the meshes into parallelograms.
For most test purposes a series of screens such as is given in Table 10,
col. 1, or the nearest B.S. equivalent, is adequate. If particles are more
closely sized, prolonged shaking on the screen is needed, as many of them
will be so near the aperture size as to become wedged, or fail to find a way
TABLE 8
ASTM-£-1I-61
I I
in. in. Scale
(approx. (approx. Equivalent
Standard Alternate mm. equiva- mm. equiva- Designation
lents) lents)
• 11·2 mm. -(-8 in. 11·2 0-438 2-45 -0965 .441 in.
9·51 mm. t !n_
In.
9-51
8-00
0·375
0-312
2-27
2-07
'0894
'0815
.371 in.
2t mesh
• 8·00 mm. 111
6,;3 mm. ·265 in. 6-73 0-265 1-87 ·0736 3 mesh
6·35 mm. i in. (~ 6-35 0·250 1'82 ·0717
Standard I Alternate I
mm.
in.
(approx.
equiva- mm.
I in.
(approx.
equiva-
Screen
Scale
Equivalent
Designation
__________L -__________ -+________ __ ~ ~l~en~t~s)~_t------~--~le~n~ts~)--+_-------
"707 micron No. 25 0-707 0·0278 ·450 ·0177 24 mesh
595 micron No. 30 0'595 0·0234 .390 -0154 28 mesh
"SOO micron No. 35 0'500 0·0197 '340 ·0134 32 mesh
420 micron No. 40 0·420 0·0165 ·290 ·0114 35 mesh
• 44 micron
37 micron
No. 325
No. 400
0'044
0'037
0·0017
0·0015
I ·030
·025
'0012
·0010
325 mesh
400 mesh
• These sieves correspond to those proposed as an International (ISO) Standard. It is recommended that
wherever possible these sieves be included in all sieve analysis data or reports intended for international
publication.
(a) These sieves are not in the fourth root of2 Series, but they have been included because theyarein
common usage.
TABLE 9
CoMPARISON TABLE OF U.S., BRITISH, FRENCH, AND GERMAN SIEVE SERIES WITH TYLER EQUIVALENTS
BRITISH
TYLER (I) U.S. (2) STANDARD (3) FRENCH (4) GERMAN DIN (5) TYLER (II
Equiv. Size Opg. No. Opg. No. Opg. Opg. Opg. Equiv.
Mesh mm. mm. mm.
,
Microns mm. Mesh
t-·
4·24' 107·6
4'
3t'
101'6
I I
I
90·5
~.---
3" 76'1
I
2t'
2·12'
64'0
53-8 i!
2' 50·8 I
If' 45·3
It' 38'1
I
-~
It' 32·0
1'05' 1·06' 26·9 1·05"
1" 25-4 25.0
,883' i' 22-6 ,883"
,742' t' 19·0 20·0 ,742"
-- -.~---.
1
1---- ~ ~--- -----
18·0
- -- - - ---- -
i"
.' II
'624' 16·0 16·0 ,624"
'525' ,530" 13'5 ,525'
11,,, 12·7 12'5
I
'441' Ie 11·2 '441'
I - --"-----_ ..
.
---~ -~-- ---~--~--
10'0
,371' in 9'51 ,371"
,312' "1«r" I 8'00 8·0 '312"
,263" ,265' 6'73 ,263"
i" No.3 I 6'35 6·3
3t
--+
No.3t I 5·66
c------ --- --------
3t
5·000 5·0
4 4
i 4·76
38
-l
5
6
5
6 I
I
4·00
3'36 5 3·353
37 4·000 4·0 5
6
36 3'150 3'15
7 7 I 2·83 6 2·812 7
8 8 2'38 7 2-411 35 2'500 2'5 8
9 10 2·00 8 2·057 34 2·000 2·0 9
10 12 1·68
L 10 1·676 ! 33 1·600 1·6 10
14 16 1'19 14 1·204 I 14
I !
16 18 1'00 16 1·003 ' 31 1·000 1·0 16
20 20 0·841 18 ·853 20
I
TABLE 9 (continued)
BRITISH
TYLER (I) u.s. (2) STANDARD (3) FRENCH (4) GERMAN DIN (5) TYLER (1)
r:r:
Equiv. Size Opg. No. Opg. No. Opg. Opg. Opg. Equiv.
Mesh mm. mm. mm. Microns mm. Mesh
3u 'MUU MOO
24 25 0·707 22 ·699
29 '630 630 ·630
28 30 0·595 25 ·599 28
32 35 0'500 30 ·500 28 ·500 500
35
42
40
45
0'420
0·354
36
44
·422
'353
27
26
·400
·315
400
315
~I-~=--
·400
·315
35
42
48 50 0·297 52 ·295 48
Sieve Mesh and Apertures Aperture Dimensions in Mm. Surface Area in ....q. em. per Weight in grammes at Unity Weight
inmm. based on Tyler Sieves gramme at Unity S.G. S.G. to expose one sq. metre in gramme s
of Surface Surface at Unity
I relative S.G. to
to Unity cover 8-111
'bier Dis- at 200 Diam.
Vi Nearest ...¥earest tance Area Mean Average Average Sieve
B.S. I.M.M. between sq.mm. Diagonal Aperture Cube Oblong Plale Cube Oblong Plate Mesh
Series Particle Particle one par-
Wires ' ticle deel
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17)
1·05 in. 1·00 in.
25·40 mm. - 26·67 710 37·7 32·18 1·86 1·55 7·5 2·55 5,360 6,440 1,340 3,\120
0·742ln. 0·75 in. 0·003 -
19·05 mm. - 18·85 355 26·7 22·76 2·M 2·20 10·5 3·60 3,790 4,550 950 2,780
0·525 in. 0·50 in. 0·004 -
12·70mm. - 13·33 177 18·9 16·09 3·73 3·11 14·9 5·10 2,680 3,220 670 1,960
0-371 in. 0·375 in. 0·006 -
9·53mm. - 9·423 89 13·3 11·38 5·27 4·39 21·1 7·20 1,900 2,275 474 1,390 0·008
3mesb 0·25 in. -
6·35mm. - 6·680 44·6 9·45 8·052 7·45 6·21 29·9 10·2 1,340 1,610 335 982 0·011
4 0·1875 In. 170
. 4·76mm. - 4·699 22·1 6·64 5·690 10·55 8·80 42·2 14·4 948 1,140 237 694 0·016 120
6 .. 5 mesh
3·353mm. - 3·327 11·1 4·70 4·013 14·9 12·4 59·8 20·. 668 803 167 490 0·022 85
8 . 7 mesh 5 mesh
2·411 mm. 2·540mm. 2·362 5·6 3·34 2·845 21·1 17·6 84·3 28·8 474 569 118 347 0·031 60
10 .. 10 mesh 8 mesh
1·676 mm. 1·574mm. 1·651 2·63 2·33 2·007 29·9 24·9 119·5 40·8 335 401 83·6 245 0·044 42
14 . 14 mesh 10 mesh
1·204 mm. 1·270mm. 1·168 1·36 1·65 1·410 42·6 35·5 170 58·2 235 282 58·8 172 0·063
18 mesh 16 mesh 30
20 ..
0·853mm. Q·792mm. 0·833 0·69 1·18 1·001 60·0 50·0 240 81·9 167 200 41·7 122 0·089 21
28 .. 25 mesh 20 mesh
0·599mm. 0·635mm. 0·589 0·347 0·833 0·711 84·4 70·4 338 115 118 142 29·6 86·7 0·126 15
35 .. 36 mesh 30 mesh
O·422mm. 0·421 mm. 0·417 0·174 0·590 0·503 119 99·4 477 163 83·8 100 21·0 61·4 0·178 10·6
48 .. 52 mesh 40 mesh
0·295mm. 0·317 mm. 0·295 0·087 0·417 0·356 169 141 674 230 59·3 71·2 14·8 0·25 7·5
65 72 mesh 60 mesh 43·"
.. 0·~11 mm. 0·211 mm. O·!OS 0·043 0·294 0·252 238 199 952 325 42·0 50·4 10·5 30·7 0·35 5·3
100 .. 100 mesh 80 mesh
0·152 mm. 0·157 mm. 0·147 0·022 0·208 0·178 337 281 1,350 461 29·7 35·6 7·42 21·7 0·50 3·8
150 .. 150 mesh 120 mesh
0·104mm. 0·107 mm. 0·104 0·0108 0·147 0·126 476 397 1,910 650 21·0 25·2 5·25 15·4 0·71 2·7
BOO 200meBh 150 mesh
0·076mm. 0·084mm. 0·0055 0·105 0·089 674 1 562 922 14·8 17·8 3·71 10'9 1·00 1'9
300 mesh 200meBh 0.074 1 2,700 I
270 ..
0·063mm. 0·063mm. 0·0028 0·075 0·064 938 782 3,750 1,280 10·7 12'8 2·67 7'80 1·40 1'4
63
400 .. 0.088
0.0 1 0'0014 0'054, 0'046 1,300 1,090 5,220 1,780 7'66 9·20 1·92 5'61 1·94 1·0
End of Sieve Se ries. I
Micaons 0·034 1,760 1,470 7,060 2,410 5·66 6·80 1·42 4·15 2·63
20
I 2,400 2,000 9,600 ~,280 4·16 5'00 1·04 3·05 3·57
i I 0·025
I
10 I 0·015 4,000 3,330 16,000 5,460 2·50 3·00 0·625 1·83 5·96
5 0·0075 8,000 6,670 32,D00 10,900 1·26 1·50 0·312 0·915 11·9
1 0·0030 20,000 16,700 80,000 27,300 0·600 0·600 0·125 0·366 29·8
0.5 0.00076 1 80,000 66,700 320,000 109,000 0·126 0·150 0·031 0·092 119
I
Note.-These figures for surface area take no account of internal fissuring, or compacted aggregates. Surface area, as measured by gas-adsorption methods, may
show much higher values.
TABLE 11
DISTRIBUTION OF WEIGHT AND SURFACE IN GROUND SILICA
(After Heywood)
Medium Grade Fine Grade
Weight in Surface in Weight in Surface in
Size Undersize Undersize Undersize Undersize
% % % %
100 mesh B.S. 97-8 99·8 99'9 100·0
150 90·0 98·7
200 75·5 95·6 99·7 100·0
240 69·2 93·9
300 60·3 91·0 99·4 100·0
40 mi~ron~' 49·0 86·5 94'4 99·8
30 37-6 80·3 87'1 99·3
20 24·8 70·5 74'5 98·2
10 10·8 52-8 50·2 94·8
5 4·3 36·3 32-3 89·8
2 1·0 18·5 16·0 79·8
I 0·25 8·9 9'1 70·1
0·5 0·05 H 4·9 58·2
0·2 1·7 38·9
0'1 0'6 23-4
0·05 0'15 10·6
Note.-Percentage weight of undersize measured down to 2 microns size; subse-
quent percentages obtained by extrapolation.
solid surface, then practically the whole of such reagents are directed toward
the ordinary grinding mill undersize and, in the example cited in this Table,
more than half the reagent is consumed by the "-10 micron" fraction of
the ore. The significance of this will appear more fully later, when flotation
physics is studied.
Again, the -10 micron fraction, by providing more than half the surface,
is responsible for more than half of the pseudo-viscosity and internal friction
154 Mineral Processing-Laboratory Sizing Control
Particle Terminology
In studying the behaviour of a particle it is often useful to compare it
with a sphere, or to define some dimension common to a group. The mean
size, or "average particle size" may be defined as the dimension of a hypo-
thetical particle such that, if the comminuted material being examined were
composed entirely of such particles, this particle would have the same value
in respect to some stated property as is exhibited by the actual powder.
The projected dIameter of a particle is defined 7 as the diameter of a circle
which has the same area as the projected profile, which will be greatest when
the particle is examined from a direction perpendicular to its most stable
plane. The equivalent surface diameter is that of a sphere having the same
surface area as the particle. The equivalent volume diameter is that of a
sphere of like volume. Particles are open-pored when they contain cavities
open to the surface and close-pored when such cavities are internal only.
T he true d enslty . Ie IS
· 0 f a porous partlc . t hat 0
f Its - - exc Iud'mg pores.
. -mass
volume
Its apparent density is that which excludes open pores but makes no allow-
ance for closed ones. Specific surface is an important characteristic of a
mass of particles. It is usually expressed as cm, 2/g, of mass for a dry powder,
or as cm /1, at a concentration of n gil or n mill for mineral pulps or slurries.
2
fibrous when thread-like; flaky when plate-like; granular when irregular but
roughly equi-dimensional; irregular when non-symmetrical; nodular when
rounded but irregular; and spherical when globular.
large sample, screened with moderate efficiency, rather than a small one with
possibly a high sample-cutting error. Laboratory screening methods must
be adapted to the general rule that the degree of accuracy at which they aim
should be co-ordinated both with the accuracy required by the work in hand
and with that possible in sample cutting.
The nest of screens is assembled, with the largest-mesh sieve on top, and
a close-fitting pan at the bottom to receive final undersize. The sample is
placed on the top screen and a lid is fitted closely to minimise dusting loss.
The nest of screens is then shaken, and gently jarred. The point at which
the pan is bumped is changed every half minute for one 60° further round
the rim. Shaking, with jarring, is continued until very little further material
falls through the bottom screen. The time required for this is then made
standard for the particular class of material or work. When friable material
is being tested, care must be used not to work with such violence, or for so
long a time, that the sample is degraded on the sieves.
Most laboratory work is done on mechanically operated shaking appliances.
One widely used appliance is the Rotap shaker (Fig. 61). Up to six standard
8" round screens can be nested on this, and on most mechanised devices.
Vertical reciprocation of the sieves is accompanied by a rotatory and a tapp-
ing motion. The length of screening time can be controlled by an automatic
timer. Another six-nester is the Sherwen shaker, operated by A.C. current
(Fig. 62). The vertical displacement of the table at each vibrating stroke is
controlled by an electrical damping device, so that the amplitude of the
Mineral Processing-Laboratory Sizing Control 157
vibrating stroke can be varied to suit the material. Dusting loss is thus
minimised and the best kind of shaking action is selected for the job in hand.
Since no circular distributing motion is applied, the sieves must be taut, to
assure good use of the whole surface. In the Russell shaker (Fig. 63) the
nest is agitated by an out-of-balance coupling connected to a small electric
motor. The result is a "panning" motion which floats and churns the mater-
ial with good searching, mainly over the central part of the sieve.
With the Russell shaker, attachments are available which allow the nest
of screens to be sprayed by water, the undersize being flushed to a receiving
vessel.
Inaccuracy in screening arises from several causes. If the sample is pro-
perly dried, friction on the sieve leads to some electrostatic "bunching" of
the particles, so that several which are truly undersized to a given aperture
report as part of the "on" fraction. Such seizure increases with temperature,
and screening in a cool room therefore minimises it. Dampness is the worst
enemy of accuracy. It leads to the sticking together of particles, or their
clinging to the screen cloth when they should fall through. When dry-crushed
rock is being tested, and also When the test sample has come from a pulp
158 Mineral Processing-Laboratory Sizing Control
containing much soluble material care should be taken that no dried salts
are adhering to the particles. These would increase their effective cross-
section. In very exact work, microscopic examination may be used to check
whether particles have been aggregated by impact during grinding, so that
several are clustered together in such a way as to report as one. Dispersal
and free movement are aided by limiting the size range of the sample. This
can readily be done by removing the sub-sieve particles (say those finer than
200 mesh) before making the screen analysis.
The connexion between surface friction and mass mobility can easily be
observed. If a level teaspoonful of dry and wide-ranged sand (say - 65
+ 400 ~) is dropped from a height of about five feet it falls with but little
scattering. If a similar quantity of closely sized sand is similarly dropped
its particles will scatter evenly and over an appreciable area. In the second
case the specific surface has been reduced and with it the binding effect of
inter-facial friction. This simple experiment shows one of the many ways in
which solids moving through a fluid (gaseous or liquid) are influenced by the
E
~
u
10
5
9
8
7
6
# 300 100 72
P S3 104 152 211 295 422 599 853 Microns
Screen Aperture, Microns (log scale)
Fig. 64. Graphs of Screen Analysis in Table 12
Lower, direct plot. Upper, cumulative plot
To avoid distortion and possible misinterpretation, of the direct plot curve,
values of the weight increments should be plotted at equal geometrical
intervals on the size scale. To give equal prominence to all parts of the size
range, a logarithmic size scale is used, and the intervals taken normally
correspond to a y'2 sieve series. The same considerations apply to results
which are obtained by sub-sieve sizing methods, and the common practice
of plotting some arbitrary relationship between the increments on the size
scale should be avoided. If the graph is developed as a histogram, the areas
show the relative proportions of the material in each size interval.
The cumulative weight, or integral plot has the advantage that a constant
geometrical relationship is shown, even in the absence of the v'2 succession
of screens, since the logarithmic scale is plotted as apertures in microns,
and not according to any labelled sieving system.
Mineral Processing-Laboratory Sizing Control 161
The particular integral plot, in which log scales are used on both axes of
the graph, and in which percentage weight finer than a given size, as ordinate,
is plotted against particle size as abscissa, is known commonly as the Schuh-
manns plot (see also Fig. 64). As shown by Schuhmann and others,"·l.
the size distribution of fine powders produced by comminution tends to fol-
low certain empirical relationships. In the Schuhmann plot the distribution
approximates to a straight line over the greater part of its length, and may be
represented by the following equation.
(8.1)
where, =
y the percentage weight finer than size x.
m =a constant. The distribution modulus.
K =a constant. The size modulus.
tional to the rubbing of the total surface of the particle against the fluid
through which it is moving. This frictional retardation is of a complex
nature. One factor is the nature and degree of wetting of the surface by the
ambient fluid/s. Wetting is only complete when the air which was in contact
before immersion is displaced. and this is not necessarily complete. The
relative surface tensions of surface/air and surface/water (or other liquid
used) affects the balancing forces of attraction, and indirectly the flow pattern
of the liquid as it streams past the particle. Another retarding factor is the
spinning and swerving effect produced by unbalanced braking as the solid
moves through the liquid. At any moment the particle is drawing on its
gravitational energy in order to momentarily force its maximum horizontal
I;ross-section through the molecularly bonded liquid. On each side of its
axial plane of movement it has two equal masses in balance about a vertical
axis, but is simultaneously subject to frictional unbalance to the extent that
the corresponding surface areas are not similarly equal. Save for a perfect
sphere there can be no such balance. Differential braking therefore either
deflects the particle from a vertical path, or sets it spinning. This leads to
collision with other particles and loss or gain of momentum and kinetic
energy. Further, the path thus elongated calls for more conversion of gravi-
tational energy if the particle is to continue to fall then if it could have moved
vertically. In other words, gravitational effect is weakened, and the decel-
erating effect of friction is greater than can be accounted for by surface
braking alone. Thus, the determinants of the response to a given fluid
medium of a single particle are the resultant of (a) its specific gravity and
volume, (b) its total surface (result of shape plus texture) and (c) deviation
from linear drift. In screening, the selection into size grade depends on the
gauging effect of a system of rigid wires applied to the particle's cross-section.
The relative densities of heavy and light minerals do not enter into consid-
eration. When "equal-sorting" methods are applied, the criteria are mass
and shape. In dealing with the settling behaviour of homogeneous material,
volume and shape alone need to be considered. With ores, the fact that
more than one mineral is present, and that each may have been liberated
completely or may still only be partially unlocked, must be remembered in
assessing results.
In practice, excellent reproducible results are obtainable. Simple methods
can be improvised to allow size-control investigations to be applied to particle
sizes down to 15 microns or less. More complicated methods are needed
when dealing with very fine sizes. It is significant that several mill labora-
tories find it profitable to include routine tests down to the 5-micron zone
of investigation. Since a "jog in the data" is inevitable when two so diverse
techniques as sizing and sedimentation are applied to a sample, tests are
usually overlapped so that a curve can be constructed which smooths out
the differences.
In distilled water at 20° C, a quartz sphere 200 microns in diameter accei··
erates till it reaches a terminal velocity of 29·4 cm./min. At 15° C the rate
is 25·8 cm./min. If instead of being allowed to faIl unhindered through
static water it were introduced into a smoothly rising stream, this sphere
would, in theory, "hang" or remain stationary when the rising rate of the
Mineral Processinff-Laboratory Sizinff Control 163
water equalled the failing rate of the sphere. In practice, difficulties arise
through the setting up of small eddy-currents between the fluid and the
sides of the tube and through convection effects. so that the sphere would
"'dance" rather than remain stationary. Certain anomalies therefore arise
when a rising current of water is used to lift small particles while larger ones
drift down. although elutriation, which exploits this difference of motion,
is used in sub-sieve sorting.
A sphere is convenient for use in research work, since it has a calculable
area which is the minimum possible for a given volume, and which always
presents the same cross-section to the fluid through which it falls. When
an irregularly shaped particle falls through a fluid under similar conditions,
the total surface rubbing against that fluid is greater than for an equal
spherical mass and the rate of fall is therefore slower. Fall is also erratic
owing to changes in the cross-section as the particle spins during its descent.
When a large number of spheres of equal specific gravity move in a rising
column of water the smaller ones can be made to rise, the medium ones
to teeter or dance about in a dilated layer (called a teetering zone) and the
larger ones to fall through. The behaviour of any single sphere under these
circumstances is random, depending on the chances of collision, eddying of
the water. and of transfer of kinetic energy acquired during falling or rising
when combating such eddies or when taking part in such collisions. A
generalised behaviour can, however, be predicted. If instead of spheres a
collection of particles of homogeneous material but of random shape is
used. the individual variations become more complex, but a generalised
behaviour sufficient for some sorting separations can be obtained. When
the further complication is added of differential density between the particles,
there is further loss of accuracy. Nevertheless, this behaviour of particles
in rising currents of water permits of a very simple and convenient industrial
application called e/ass!fieatioll which is widely used in the large-scale control of
fine sizing or sorting. In the laboratory classification or elutriation of solids
in a rising column of liquid has limited use. With patience and care, even
better work can be done by sedimentation. a method based on the falling
rate of small particles through static liquids. The elements of the elutriator
(Fig. 65) include:
WATER,
WATER
\
/,11'
",
HYDRAULIC
WATER ~
SORT ING
CO LU MN
CONTROL
VALVE
FINER
PRODUCT
(OVERFLOW)
co""ur IIiIUI
",.,UlU'
. •• 0 -----!
K AU Of flU
4;
-
AIR IN
A
, - ,-----. -
AIR OUT
,......---, , - -
I~
! :q-
SA,.,PLE-
~
{{~ 'Vi
"
If)
I4j
.. ~Ot.
~::~.~J:
POliOVS_
PHI G ~ ~
- ~
I.£VEL.UNG
TI/'BE
....
• I
I I
II
HONOH£TER UJ l
2p. and 50p. the Gooden and Smith apparatus'8 has proved suitable for routine
work. The formula used in air permeability measurement at the Royal
Institute of Technology, Stockholm,'9 based on the rate of flow of a gas
through a compacted bed of powder, is:
1 I I £3 A 8v!z I
Qm=- .-.-.--.-LJpt--.-.
JRTl.~.~.~.
ko.k, p. S2 (1-£)2 L 3v'7T k,
M S 1-£ L Pm
It gives reproducible values within 2 % + or - on powders with specific
cm2
surface above 500 - - .
cm3
Here
pm = middle pressure of the gas in the bed (dynesjcm. 2),
Qm = volume rate of the gas flow through the bed measured at the
pressure pm (cm.3jsec.).
LJp = pressure drop across the bed (dynesjcm. 2).
A = cross-sectional area of the bed (cm. 2).
L length or depth of the oed (cm.).
£ = porosity or the fractional void space in the bed.
S = specific surface of the powder (cm. 2jcm. 3 firm measure).
f.L = viscosity of the gas (poise).
R = gas constant per mole = 0·8315 . 108 erg/deg. mole.
T = absolute temperature of the gas (degree Kelvin).
M = molecular weight of the gas.
ko and k, are numerical constants of which ko is supposed to be a real constant
and equal to 2'" while k, is supposed to depend upon the shape of the particles.
To be able to calculate the specific gravity one usually takes k, =2·5t.
These molecules are associated with the water molecules beyond them
against which they drag in a viscous manner. Thus, layer by layer, the zone
surrounding the falling sphere can be pictured as being disturbed by its
passage, the effect of this disturbance being calculable as loss of kinetic energy
through surface friction, or laminar flow. The effect has been embodied in a
classical formula by G. G. Stokes.
d2g (a-p)
v=---- (8.2)
181)
where v = velocity (free falling). .)
a = density of particle
p = density of fluid I
g = gravitational acceleration
1) = absolute viscosity
d = Stokes's diameter of particle
jr in C.g.S. units.
R = Reynolds number
The free-falling velocity for a particle moving through a still fluid is that at
which its effective weight is in balance with the drag exerted by its surround-
ings, so that its rate of movement is steady. The Stokes's diameter is the
equivalent free-falling diameter of a sphere moving at such a rate that its
Reynolds number is less than 0.2. The Reynolds number is a dimensionless
group of parameters which defines the flow pattern of a fluid surrounding a
particle
vdp
R=- (8.3)
1)
These definitions, together with the following example of a calculation of
Stokes's diameter, are taken from B.S.2955;9
EXAMPLE OF CALCULAnON OF
STOKES'S DIAMETER
To calculate the Stokes's diameter of quartz particles (density 26·5 g/m!)
corresponding to an observed time of sedimentation of 2 minutes over a
distance of 20 cm. in water at 15°C. (absolute viscosity 0·0113 poise).
Free falling velocity = 10 cm. per minute = 0·1667 cm. per second
Stokes's Law (8.2) above, is in this case
d2 (2'65-1) 981
0'1667=-x (8.4)
18 0·0113
18 x 0·1667 x 0·0113
d2 = = 2094 X 10- 8 (8.5)
1·65 x 981
d = 45·76 x 10-4cm = 45·76 microns. (8.6)
NOTE. For expl"nation of symbols see Stokes's Law. (Definition 328)
170 Mineral Processing-Laboratory Sizing Control
To check whether the fluid flow round the particle is of the streamline
type and the motion within the range of Stokes's Law. the Reynolds
Number is calculated as follows:
As this is less than 0·2, the motion is stream-line and the above cal-
culations are valid.
8 a-p
=v-.g--r (8.8)
3Q p
Rf = Kr n p /L2-n vn (8.9)
Mineral Processing-Laboratory Sizing Control 171
When a single particle falls through still water, it displaces its own volume,
thus creating a rising current. The effect is indefinitely small. If it falls
from a relatively warm upper layer it transports heat and increases any con-
vection at work. When a continuous stream of particles falls through still
water a definite but minute rising current is set up. (For the process to con-
tinue it would be necessary to withdraw the settled solids and water would
be carried out at the same time, thus more than offsetting this rise). Each
falling particle presses upon the water below it, and generates a current at
the cost of part of its own kinetic energy, which is thus transferred to the
surrounding fluid. To this loss of energy must be added that used in dis-
rupting the bonding of the liquid molecules through which it cleaves its pass-
age.
Consider next a vertical cylinder of uniform horizontal cross-section
through which a current of water is rising. Ignoring the effects of convection
and wall friction, the water rises everywhere at an even rate. (In mineral
processing vertical currents are called "'hydraulic water", and are not necessarily
172 Mineral Processing-Laboratory Sizing Control
If homogeneous particles close-ranged in size form the teeter bed, then the
behaviour of any mineral presented to it can be calculated.
This discussion is limited to the consideration of a batch of particles charged
174 Mineral Processing-Laboratory Sizing Control
Direct observation and measurement can be made down to less than one
micron (111) by use of a metallurgical microscope. For measurement of
particles, a linear scale engraved on glass is placed in the eyepiece (ocular)
and the instrument is then focussed on a stage micrometer. This is a trans-
parent engraved scale which shows either squares or other figures of known
dimensions, or better, one millimetre divided linearly into tenths and hund-
Mineral Processing-Laboratory Sizing Control 175
redths (lOOOfL, IOOfL, and IOfL). It is placed on the microscope stage and the
two scales can then be correlated for the ocular-objective system in use. The
stage micrometer is then removed and the particles are examined and meas-
ured. If they have been placed loose on the glass slide, they will settle in
their most stable position and the two major dimensions are presented. For
routine size evaluation special graticules such as the Patterson-Cawood4
are useful. The use of a drop or two of a dilute solution of Calgon helps to
disperse the material on the counting slide and to display the smaller particles
which might otherwise be masked by the big ones. Where a tedious amount
of particle-counting is involved various aids are available. A projection
microscope is less tiring to work at than a stand machine. Particles dispersed
in water can be made to flow through a small aperture and work a counting
mechanism as they interrupt a beam of light. The Zeiss particle size analyser
is semi-automatic.
"Liberation-Mesh"
10
I 2 J 4 5 6 7 8 9 10
9
II 12 13 14 15 16 17 18 19 20
B
21 22 23 24 25 26 27 28 29 30
7
JI 32 33
~~~ 37 38 39 40
4~!i;,~f:'~?
6
41 42 43 47 48 49 50
s
S4 i~
~~!.:·:::r
51 52 53 57 58 59 60
4
61 62 63
64 ::·:i~)
....':7' 67 68 69 70
J v
7' 72 73 74 7S 76 77 78 79 80
2
81 82 83 84 85 86 87 88 89 90
91 92 93 94 95 96 97 98 99 100
o 2 J 4 5 6 7 8 9 '0
control has to do with liberation-mesh and the mesh of grind at which optimum
unlocking of the desired mineral from those accompanying it in the ore is
reached. Thus far this chapter has discussed methods of sorting the particles
either by size or by settling characteristic. For some specific purpose such a
grading may be an end in itself, but comminution control is more usually
concerned with economic liberation of values from the crude ore.
Liberation is not achieved by grinding the ore down to the grain size of the
desired mineral particles. If Fig. 69 is taken as representing a stylised cross-
section through a cube of ore in which the values are 5 % and the gangue
95 %, then it will be observed that if the rectangle shown is cleft into 100
square subdivisions, 87 represent clean gangue (i.e. completely valueless
mineral) and 13 partly liberated material in which gangue and value are still
tightly bound. Though in Fig. 69 the value is 5 % and the L-shaped flake
should theoretically be completely liberated from gangue with a 20 to 1 reduc-
tion by comminution, the diagram represents a 100 to I crushing. The
incompletely separated squares or "middlings" (Nos. 26, 35, 36, 37, 44, 45,
46, 54, 55, 56, 65, 66, and 76) contain varying amounts of value from the tiny
fleck in 37 to the high loading of 55. Despite the theoretical heavy overgrind
(l00/1) instead of 20/1 no value has been completely liberated.
If the grain sizes of an actual ore specimen had been measured, they would
not correspond with the grinding treatment necessary for adequate liberation
of any specific mineral, as Fig. 69 shows. "Middling" is the name given to a
particle intermediate in composition between one of clean concentrate and
one of clean tailing. Usually it contains fractions of both minerals and can
only be treated after further grinding has liberated them. A particle which
contains firmly joined value and gangue is called a "locked middling". It
can only be separated into its constituent minerals by further comminution.
A variant is the middling composed largely of a third mineral species not
sought as a concentrate, but intermediate in its response to a given method
of treatment. The position in Fig. 69 is artificial, but the principle illustrated
by it is very important. Grinding commences by liberating, or freeing, part
of the major constituent at a relatively coarse mesh-ofgrind (usually abbre-
viated in our literature to m.o.g.). This may be above the maximum particle
size of the valuable constituent if the latter is the minor mineral present.
Even when grinding has brought all the ore down to the grain size of the minor
constituent, it is unlikely that any of this mineral will have been cleanly lib-
erated, though a good quantity of the major one is now ready for separation.
Grinding must be continued well below the measurable liberation-size of the
minor constituent before any substantial quantity of it is freed.
Consider, for example, particle 44 in Fig. 69. If a further 4 to 1 size reduc-
tion is made squarely one quarter-sized new particle of gangue will be freed,
two "locked middlings" carrying but little value, and one high-grade middling.
Yet, no completely gangue-free value has been liberated as the result of this
considerable (and costly) further grinding.
Occasionally the ore may have a weak boundary between gangue and value,
along which breakage occurs preferentially. In such a case, liberation may
be satisfactory at mineral-grain size. Usually the adherence of the minerals
in the crystalline matrix of the ore is strong, and during crushing and grinding
Mineral Processing-Laboratory Sizing Control 177
the various constituents are cleft across. This leaves the resulting particles
in the form of locked middlings, composed partly of gangue and partly of
value, with the more abundant phase (usually the gangue) far freer than
the scarcer one.
Loosely bonded gravels are easily disintegrated. A pebbly phosphate, or
a clay carrying nodules of manganese or iron, can be broken up to release the
value with but little comminution. Fracture planes (cleats) in coal and schis-
tose weaknesses in hard rock can be preferentially opened up by gentle
comminution appropriate to the material. If there is marked difference in
the hardness or toughness of the constituents of the ore, some differential
grinding may be possible using forces sufficient to break up the more friable
mineral while leaving the tougher one unaffected. In the case of coarse sands
or gravels which have become encrusted with some mineral of special value,
this can sometimes be washed off by tumbling treatment in log washers or
removed by gentle grinding action. These are special cases. Normally, the
ore must be ground down to an economic liberation point determined by test
work. Sizing control provides a valuable means of isolating the various
mesh-grinds for such testing.
It was noted above that the more abundant gangue is freed at a far coarser
mesh than the values. In order to avoid the cost of grinding all the ore to the
mesh where the "values" are fully liberated, it may be possible to arrange for
primary coarse grinding to be followed by a stage of treatment in which some
of the free gangue is removed. This reduces the quantity of ore needing
finer grinding and thus cuts down the grinding cost. If tests show that the
cost of such a treatment is lower than the cost of grinding all the ore to a fine
mesh, then this fact is considered during the construction of the flow-sheet
and the planning of the treatment plant. Again, primary crushing may
develop enough difference between the clean gangue and the partly liberated
value for the latter to be concentrated into a small bulk. This is a form of
stage' concentration used in dense-media separation. The best method of
solving the problem of optimum economic and technical liberation depends
less on the facilities at hand than on the nature of the ore and the way in
which its values are distributed through the gangue. For convenience,
consideration has been confined in this discussion to an idealised squared
area through a simple ore with only two components or phases. In mineral
dressing, a series of valuable products must frequently be separated and the
problems of liberation are then correspondingly complicated.
It is now clear that the crystal size of the desired mineral in the ore is not
the complete guide to the liberation-mesh or m.o.g. (mesh-of-grind) needed
for its concentration. This is arrived at by empiric laboratory testing, des-
cribed later. The optimum degree of liberation of the value is called the
"break".
References
I. Dallavalle, J. M. (1948). Micromeritics, Pitman.
2. Herdan, G. (1953). Small Particle Activities, Elsevier.
3. Cohen, E. (1957). I.M.P.e. (Stockholm), Almqvist & Wiksell.
4. Heywood, H. (1946). Trans. I.M.M. (Lond.), 55.
178 Mineral Processing-Laboratory Sizing Control
References-continued
5. Taggart, A. F. (1945). Handbook of Mineral Dressing, Chapman & Hall.
6. Heywood, H., and Pryor, E. J. (1946). Trans. I.M.M. (Lond.), 55.
7. British Standards Inst. (1958). B.S.2955.
8. Schuhmann, R. (Jnr.). A.I.M.M.E. (T.P. 1189).
9. Austin, J. B. (1939). Indust. Eng. Chem. (Anal.), 11.
10. Gaudin, A. M. (1926). Trans. A.I.M.M.E., 73.
11. Rosin, P., and Rammler, E. (1953). J. Inst. Fuel, 7.
12. Pryor, E. J., Blyth, H. N., and Eldridge, A. (1953). Recent Developments in
Mineral Dressing. I.M.M.
13. Kelsall, D. F. (1962). "R & D", Nat. Trade Press (June).
14. U. S. Bureau of Mines Report (1961). No. 5838.
15. British Standards Inst. (1963). Part 3, B.S.3406.
16. Carman, P. C. Trans. Inst. Chem. Eng .• 15.
17. Lea, F. M., and Nurse, B. W. J. Soc. Chem. Ind., 58.
18. Gooden, E. L., and Smith, S. M. Indust. Eng. Chem. (Anal.), 12.
19. Svensson, J. Royal Inst. Tech., Stockholm.
CHAPTER 9
Preliminary
(e) Separate ore into relatively coarse and fine fractions by exploiting
differences in settling rates.
(j) Split a long-ranged feed into fractions settling equally.
(g) Close grinding mill circuits so that no particles escape from them
into the concentrating section of the plant until they have been
reduced to the desired sizes.
(11) Remove or segregate slimes.
(i) Regulate size-range fed to a process.
180 Mineral Processing-Industrial Sizing and Sorting
TABLE \3
(a) Ratio between cross-section of particle and of mesh.
(b) Percentage of screen area open.
(r) Angle of incidence of feed.
(d) Efficiency of spread of feed over screen area.
(e) Kinetic energy of particle approaching screen opening.
(f) Moisture of ore feed.
(g) Stickiness of particle and of aggregated particles.
(h) Pressure of particles riding above those next the screen cloth.
(i) "Blinding" of screen apertures.
(j) Corrosion of screen material.
(k) Electrostatic "bunching".
(I) Shape of particle. .
(m) Percentage of "near-mesh" particles in the feed.
(n) Rate of feed, thickness of layer, tautness of screen.
(0) Shape of screen apertures.
(p) Motion imparted to particle by screen vibration.
Square-meshed cloth does its best work when set horizontal, but the shak-
ing movement imparted to it must then have a forward transporting com-
ponent. Oblong meshes are often used for feed which is moist or clayey,
and for needle-like particles. If the feed tends to "blind" the cloth, the long
axis of the rectangle is set in the direction of flow. Square mesh is usual
with tabular material. The percentage of screen area open (b) depends on
weave, diameter of screen wires, and shape of aperture. For a given mesh
various ratios of opening are available. (r) is concerned with the mode of
arrival on the screen. Ideally, a particle would fall with its minimum cross-
section normal to the aperture and at negligible velocity. In practice, it
competes with a crowd of other particles of random shape and size, falling
along various trajectories. Hence item (d) is most important, since the wider
the entering feed is spread, the easier will it be for a particle to find unobstruc-
ted passage to the screening surface. Since this is so (c) should be low.
A particle flying nearly horizontally toward the screen is most likely to hit
the layer of feed with its broadest dimension and to slide down on top, with
no early chance of burrowing its way down. (/) and (g) vary with climate
Mineral Processing-Industrial Sizing and Sorting 181
The grizzly (Fig. 17) was described in Chapter 3. Robustly built screens
(Fig. 70) are also available. The one shown can handle up to 1000 tons of
large rock hourly with far less loss of head than would be possible on the
static surface of a fixed grizzly. The machine must stand up to rough treat-
ment and the impact of heavy rock.
Punched screen is used for many purposes (Fig. 71). A variety of shapes
is manufactured, circular openings being recommended for coarse work and
slotted ones for fine.
Woven-wire cloth is widely used in the range between i" and 200 mesh.
Various shapes of aperture, crimps and weaves are available (Fig. 72). Steel,
stainless steel, monel metal, copper, and bronze are the chief metals from
which cloths are made. For delicate work at fine meshes, dry material is
sometimes sieved through silk or nylon. For special purposes very fine
square-meshed screens are made by electro-forming instead of weaving.
They are sold in 100' lengths three feet wide, with hole diameters in the range
25ft down to 2tft. The screen surface is nickel, on a copper base. Con-
ventionally woven materials of construction include mild steel, brass, phos-
phor-bronze, copper, copper-nickel, nickel-chrome, austenitic stainless steel,
galvanised and tinned mild steel. In milling, steel screen wire should not
be used to treat acidic and corrosive ores, unless it is first protected by being
182 Mineral Processing-Industrial Sizing alld Sorting
00000000000'-
000000000000
00000000000
000000000000
/#////#/A
00000000000
000000000000
00000000000
W/#////A
Round hole Diagonal 10\
=
=
=
=>
( )
( )
= =
=
=
=>
=
( )
=
= =
= ( )
( )
= =
= ( )
Rounded end
given a suitable plastic coating. Where corrosion is not serious, high carbon
steel is suitable, being strong and hard wearing. Maximum capacity of
screens with oblong apertures is obtained by using them with the long side
of the mesh set across the flow.
NOfto14,r.4W • ....-.
Il.... I ... ) r.,0"".",
'S4i!"U.
fI, .. 'W,I".
N
;;;
......
7; I
or:
U
'"
oL-------------------------~------~~~
o 0 5 I 0 I5 1.0 1.5
SCI'.EEN ... REA I'.EQUIREO (ft. Ton/Hou,'
TABLE 14
STANDARD COARSE SCREEN SPECIFICATIONS
Wire Diameters
Medium Medium
Light Medium Heavy Heavy
Opening
Wire % Wire % Wire % Wire %
Diamete Open Diamete Open Diametel Open Diameter Open
Inch Area Inch Area Inch Area Inch Area
4" ·500 79·0 ·625 74·8 ·750 70·9 1·000 64·0
3t" ·4375 79·0 ·500 76·6 ·625 72·0 ·750 67·8
3" ·4375 76·2 ·500 73·5 ·625 68·5 ·750 64·0
2.1"
4 ·375 77-4 ·4375 74·4 ·500 71·6 ·625 66·4
2!,' ·375 75·6 '43'75 72-4 ·500 69·4 ·625 64·0
2t" ·375 73-4 ·4375 70·1 ·500 66·9 ·625 61·2
2" ·3125 74·8 ·375 70·9 ·4375 67·3 ·500 64·0
In addition to the foregoing, rod-deck screens are used for coarse work.
The rods are sprung into place and can be changed individually. Wedge-
wire screening is employed for some purposes. It is strong and can be made
with small apertures. The blunt sides of the wedge strips which form the
separating surface are upward, so that material passing falls clear without
"blinding". Many other wire shapes are obtainable.
When a long range of sizes is being fed to a comparatively weak or light
screen, a robust coarse screen should be mounted above it to form a double-
decked system in which the heavier pieces do not reach the second deck.
A delicate screen can be supported by a coarser backing screen beneath.
The apertures in punched-plate screening may be round, square, rectangular,
or oblong with rounded ends, the last-named being less prone to blind than
holes completely circular. These screens are made of steel, steel alloy, brass,
monel, copper, or bronze. If openings are disposed in an equilateral tri-
angular pattern, the maximum ratio of opening to total surface is obtained.
The stouter the plate, the closer can be the openings and the longer the service
life. Against this must be set the fact that the thicker the plate the greater
the proneness to blind, and the higher the initial cost. This 6linding with
increased thickness is still more noticeable with woven wire.
The selected aperture depends on the working requirement, while con-
structional strength depends on the nature, size, and loading of the feed.
The ratio of aperture to screen area is not, therefore, a function of the mesh
in commercial screens. For instance, an 8-mesh screen woven from thin
wire can have 24 % more screening area than a 6-mesh weave in thick wire.
Hence, the opening required dominates specification as regards size of pro-
duct, but the total amount of screen capacity available at any point in the
flow-sheet depends partly upon the thickness of weave of the selected mesh.
Screens undergo such rough treatment that their working life may be meas-
able in hours in extreme cases. Flow of ore must be interrupted while a
broken cloth is being replaced. This consideration affects the choice as
regards robustness of wires.
Screening Machinery
Since separation must be made of all sorts of feeds, varying in condition
from the completely dry to the sticky, clayey, or "porridgy", each ore pre-
sents its own screening problems. Capacity, efficiency, mesh size, and wear
rate are relative to the specific ore and should dictate the choice of appliance.
The flat grizzly is used to retain large oversize which might cause trouble if
allowed to continue to the next point in the flow line. It is sturdy enough to
act as an anvil if such oversize is to be sledged down by hand, or can operate
in conjunction with any required breaking arrangement.
The inclined grizzly is ,sloped between 25° and 39° if it delivers its oversize
to a crushing machine, the gradient being chosen to allow sliding control by
means of hand or chain feeders. If the oversize is to run to the crusher free
of control, a slope well in excess of 35" from the horizontal is usual. For
dry quartz 45° should suffice, while sticky or moist rock might need 50° or
186 Mineral Processing-Industrial Sizing and Sorting
more. Grizzlys are simple and strong, but are wasteful of headroom.
The general types of separating device are presented schematically thus:
Wet and Dry Screens
I
I
Fixed
I
Moving
I I
r I I I I
Flat Inclined Revolving Shaking Conveying
Vibrating
I
I
Mechank:al
I
Electrical
The roll grizzly consists of a series of grooved rollers driven by sprocket
and chain unidirectionally in a supporting frame, the speed increasing from
entry end to discharge roll. The undersize drops between the grooves while
oversize moves flatly along. With this arrangement loss of head is minimised,
but power is needed for the roll drive.
. Grizzlys can be vibrated mechanically, electrically, or by the impact of
falling rock. They may also be shaken, or alternate bars may be moved by
eccentrics. They can be used as sorting tables or to control the rate of feed.
The mechanically vibrated bar grizzly illustrated in Fig. 70 is designed to
separate ore at I t" bar spacing or more, and handles up to 1000 tons/hour.
The curved bars tend to tumble the feed, and the vibrating action, powered
by a 7-!- h.p. motor, is effective on wet and sticky ore. The jar-bar feeder
grizzly is illustrated in Chapter 3.
One of the oldest screening devices is the trommel, which can be used wet
or dry. Its usual form is cylindrical. with the screening plates forming the
side walls, and a downward slope from feed to discharge end. Trommels
may be arranged in series (Fig. 74), with the coarsest discharge at the start,
Mineral Processing-Industrial Sizing and Sorting 187
in order to remove heavy oversize at the earliest point and by means of the
most robust screenplates. When "washing trommels" are used, the feed is
picked up by internal lifters made of angle-iron, and is sprayed. The water
runs out with the undersize and slime. A variation of the series of dimin-
ishing-mesh trommels is the compound trommel with concentric screens, the
coarsest inside, and separate discharge launders for the products. The dis-
advantage of this arrangement is that failure of a screen is hard to observe
and difficult to deal with quickly. Other variations include the polygonally-
sided trommel which gives positive lifting action to its contents and permits a
complete flat screen to be replaced as a unit. To overcome the problem of
providing a skewed bevel drive to sloping trommels, the conical trommel,
set horizontally and flaring down from feed to discharge, has had some use.
Trommels are chiefly used as sizers in gravel plants and stone-breaking work,
and on tin and gold dredges where they remove boulders and clay from the
gravels brought up by the buckets. In tin dredging a serious source of loss
of cassiterite is that due to embedding of the mineral particles in nodules of
clay, either because the richest alluvial ore lies directly on clay bottom in the
deposit, or through the jumbling of gravel and clay together during digging
and discharge from the dredge buckets. "Disintegrating trommels" have
been used, with cutting blades rotating inside the revolving cylinder. These
meet and slice up clay lump, so that the trapped cassiterite is released and can
run through as undersize. Trommels are simple, vibration-free, cheap,
strong, and economical of head loss through a series. Against this must be
set the fact that they "blind" easily, have poor capacity and cannot be re-
paired speedily or changed rapidly to a different mesh. Apart from the uses
mentioned above, the vibrating screen has replaced the trommel for most
ore-dressing purposes.
Shaking screens are usually worked dry, and chiefly in the sorting of coal.
A typical arrangement consists of an oblong box of which the screen forms the
bottom. This may be hung by chains or links, or supported from beneath.
In the latter cas' the Ferraris motion may be used. The Ferraris truss
carries a loaded deck by means of flexible battens set at a calculated angle.
When the deck is pressed forward the battens move through a rising arc,
lifting the load and throwing it forward. On release, the deck falls backward
to its stop-point, the effect being to toss the ore in the air, or at least to reduce
its clinging contact with the screen after the forward stroke has imparted to
it kinetic energy in the direction of travel. This loosening aids in the strati-
fication of the ore, and leaves the largest particles on top, where they press
upon the smaller ones which are trying to work through the meshes of the
screen. The upward tossing motion can aid in un blinding (a little), while the
jarring arrest as the deck falls back tends to loosen particles wedged in the
meshes. The vibration of shaking screens is a disadvantage, and they are
today but little used for hard-rock work.
Travelling-belt screens are typified by the Callow screen. In this, usually
made duplex, the screen cloth is bound along the selvedge to rubber strip
which forms a retaining lip on each side. Ore pulp is fed on, washed through
by sprays, and removed as undersize, while oversize is discharged at the far
end. They are little used today.
188 Mineral Processing-Industrial Sizing and Sorting
Vibrating Screens
These screens dominate modern sizing practice. They handle dry to moist
or sticky material as coarse as 10· in ring size, and as fine as 65 mesh. In
special cases they can work on dry feed down to far finer sizes. For most
mineral-dressing operations, screening stops at the point where the crushing
section delivers ore to the fine-ore bins for wet milling (say between t and
N
(a.) (C)
GENTLE STRAIGHT LINE MAXIMUM VIBRATION
VIBRATION OF FEED END AT DISCHARGE ENO
inertia of the framing, and the weight of the passing load. A motion devel-
oped in connexion with the screening of coal is shown in Fig. 77.
There are two main methods of producing vibration. For feed coarser
than a limiting size of the order of 1t", eccentric mo+ion is preferred. Below
this, and increasingly down toward a retaining mesh of t", the unbalanced
pulley is favoured.
Eccentric motion imparts a circular orbit. The typical arrangement con-
sists of a floating drive unit of which one end is shown in Fig. 78. From left
to right are (a) the drive unit, (b) a concentric bearing which carries a side
bar attached to the balance deck (c) a follower bearing connected to the
screen deck, and (d) the main shaft. When the revolving shaft attains its
operating speed the total throw of the eccentric (c) is divided between the
screening and balancing deck in inverse proportion to their weights. Many
variations in detail have been developed by the manufacturers, all being de-
signed to lesson the mutual strain between shaft and follower and to avoid
transmission of vibration to the mill structure by feeding back such impulses
in the form of useful work. Boxes can be tilted, sometimes while running,
between horizontal and 30'.
Where an unbalanced weight is revolved, vibration results. This can be
"Wet" Screening
is the "pool washing screen" (Fig. 81). The deck is interrupted at intervals
by a transverse "pool" into which water is sprayed. The feed arriving from
the previous section of screen-cloth is pulped in this pool, and the undersize
readily drains through the screen on the next section. This wet screen is
made with stepped series of decks, inclined deck, or a pool-interrupted flat
deck.
A stationary screen, the Dutch State Mines "Sieve Bend"l has gone into
considerable industrial use in the past few years. It handles satisfactory
volumes in a small mill space and makes good separation down to 100 mesh
or finer, with throughput as high as 50 tons per hour. The separating surface
flU Hili
I niH"
I'''''UI
is a stationary concave (Fig. 82) formed of wedge-wire bar screen set across
the pulp flow. Modified shapes are marketed in which the concaves extend
over 60°, 120 or 300' of arc. In the last of these feed enters vertically from
0
,
Mineral Processing-Industrial Sizing and Sorting 195
Efficient Operation
(c-t)
K= (9.2)
(I-I)
can also be used to show the relative amounts of sand discharge and overflow.
K is the ratio of weight in feed to the weight of concentrate. In this case
sand takes the place of the valuable product usually shown.
Square mesh exercises a restraining effect in two dimensions, whereas an
oblong mesh gauges the passing particle in one principal dimension. It thus
increases screen capacity at the expense of accuracy. The choice of mesh
shape must take into account three interrelated criteria-precision of sizing,
permissible tolerance of wrong sizes, and effect on overall operating profit.
Usually in mineral dressing, optimum liberation is finally regulated at the
classifier overflow, and it suffices in the screen-controlled sections if material
too large for efficient comminution in the next grinding section is held back
at any designated points. In this case, square-mesh accuracy is rarely of
vital importance. Rectangular mesh is favoured for acicular particles, and
for moist or clayey feed smaller than f'. Siabby particles are best handled
on square mesh. Material prone to blind the meshes should be treated on
oblong screens set with the long axis in the direction of flow. If the screen
product is to be delivered to mineral jigs the tighter size control possible with
square meshed cloth may be found important.
Suitable tensioning of the cloth in its securing frame is needed. The vi-
brating strokes should be distributed fairly evenly over the whole area, (a)
to avoid overstress at a point. line or node and (b) to ensure adequate tossing
of the passing stream of ore.
Good tension of the screen cloth is desirable to give efficient transfer of
the vibrating strokes from mechanism. via cloth. to load. Backlash and
slackness of the assembly are bad for efficiency. The combined effect of
vibration speed and amplitude. together with slope of screen. must be such
as will keep the material well stirred and running freely. If the amplitude is
too great, stratification will be upset and near-mesh particles will not be ade-
quately "ridden" into the meshes. If it is too feeble. the apertures will blind.
The moisture of the incoming feed may vary seasonally. in which case several
cloths of varied mesh can be kept ready. and changes made in aperture to
suit the altered condition of the feed.
Capacity of a section monitored by screening is higher when oversize is
not returned for retreatment but is sent to a different crushing system. This
is probably due to the lessening of re-circulated material not crisply dealt
with by the crusher from which it has already escaped to the screen. In a
large mill the cost of an extra crushing stage may be justified. but a small
plant would return the oversize in closed circuit for another pass. This
arrangement coulp be aided by the use of a screen rather larger than the set of
the crusher so as to keep down the volume of near-mesh circulating load.
Mineral Processing-Industrial Sizing and Sorting 197
The possibility would arise with spring-loaded crushers rather than those
having a rigid set.
The area round a dry screen is usually dusty, and may require hooding and
an extracting system. The main running cost in screening is for replacement
of cloth, subsidiary items being for power, labour, and loss of gravity head
through the appliance. Where the screens are set to gauge the size of an
important product it may be desirable to take special precaution against
delivery of oversize owing to the unnoticed rupture of a screen. This may be
done by duplicating the same mesh on a double-decked machine. If the
upper cloth ruptures, oversize will commence to come over from the lower
one, and this change can be caused to actuate an alarm.
Sands
Fig. 83. The Hukki Screening Cell
the force is determined by the speed at which this water passes upward through
the horizontal cross-section of the classifier at a given point. This in turn is
a function of the volume of rising water and the free area at the cross-section,
part of which is occupied by grains of mineral in teeter. If a particle is to
fall under these loading conditions it must overcome frictional drag and
collision in the teeter zone. Under these circumstances it is said to be sepa-
rated by hindered settling. If the classifying vessel also imposes horizontal
flow on its contents, a falling particle is displaced from the vertical to a dis-
tance proportional to the time it has taken in passing through the current.
A pulp of water and fine particles is the classifying fluid normally used in
mineral processing. Fine, dry powders can be classified in vertical or hori-
zontal air currents.
Classification deals with a mass of small particles in movement varying
from slight drift in parts of the mass to turbulence in other parts. An indi-
vidual particle is constrained by the packing density in its immediate neigh-
bourhood. If this is thought of in terms of the number of particles in a unit
volume of pulp (termed by the author the specific population in order to tie
up with the frictional factor of specific surface) it can be seen that a pulp has
a critical concentration below which unimpeded motion of individual grains
occurs and above which increasing intergranular interference is encountered.
The factors which influence movement of particles relative to the surround-
ing fluid may be summarised:
(a) The relative velocity of particles of the same S.G. and shape varies as
their sizes, a larger falling faster than a smaller one.
(b) With two particles of the same size and shape, but of different
densities, the heavier falls faster.
(c) With two particles of the same S.G. and size (displacement), but of
different shapes, the fall is retarded by skin-friction relatively to their
surface areas. (Maximum falling rate is developed by a sphere,
minimum by a thin plate.)
(d) Resistance to fall depends on the velocity of the falling particle
(Newtonian, intermediate, or Stokes ian) and varies directly as the
velocity when slow-falling through an intermediate zone of change
till it varies as the square of the velocity when falling rate is higher.
(e) Other things being equal, the velocity of fall varies as the squares
of the particle diameters when these are small, as the square roots of
the diameters when larger.
(f) Resistance to fall increases with the S.G. and viscosity of the fluid
medium through which fall occurs.
(g) Anomalies in behaviour may arise through flocculation, or the pre-
sence on the particle of minute air-bubbles.
(h) The degree to which individual particles can develop their shape
and mass properties is conditioned by the specific population and
spec((ic surface of their environmental pulp.
fragment, item (b) of the above list provides a simple means of separating
them. If the two particles are introduced into a vertical current of water,
which flows upward faster than the light particle would fall through still
water but slower than the falling rate of the heavy particle, then the one will
drift upward with the current while the other will fall slowly downward.
Here, if a feed were sized on screens and then presented to some such system,
efficient concentration would result. Removal of the variable of size would
develop a maximum difference in behaviour due to density. If the order
were reversed, it might be possible to cause small heavy particles to fall at
the same rate as big light ones, and then to separate them on a screen of
intermediate mesh, or by other methods if they were too small to screen. In
this case a classifying difference would be removed in order to develop maxi-
mum response to a sizing difference.
Since classification depends partly on frictional retardation, it cannot be
applied effectively to coarse material. It takes over from screening some-
where below 20 mesh, and is used on free-falling or "hindered-settling"
particles down to sizes of a few microns, the range being extended further
when required by the application of centrifugal force.
The mixture of fine particles with water acts as a heavy liquid. Its density
depends on (a) the specific gravity of the ore from which the fine particles
have come and (b) the solid-liquid ratio or percentage by weight of solids
in this fluid mixture. When a particle of relatively coarse size falls through
this fluid, it converts its potential energy to kinetic, its motive power being
the difference between its weight and that of an equal volume of classifying
fluid. Hence, the higher the ratio of solid to liquid the smaller becomes
the gravitational effect. In one form of classifier (the free-settling mechan-
ical type) fluid density is an important controlling factor in maintaining the
desired separation of undersize from oversize.
The kinetic energy of the particle, as it is generated, is used not only to
overcome viscous resistance in the fluid but also (if the particle attains
sufficient speed) to start vortexes and to displace other particles during
collision or frictional contact. It has been convenient to consider a single
particle dropping through the fluid, but myriads are moving in all direc-
tions in the classifier, and their individual collisions and reactions are far
too complex for mathematical resolution. Fortunately, the resultant of all
these collisions and rubbings is a generalised effect sufficiently controllable
to be of the greatest possible value in the fine grinding of ore. There is no
precise cut-off point in classification such as can be obtained in careful
screening, but at a given separation point the bulk of the particles will respond
in the desired manner. At the point where the separating cut is being made,
some particles will be diverted by drifting vortexes which throw them into
the wrong stream, but in milling the scheme of concentration is kept sufficient-
ly elastic to allow for this.
h was noted that large particles fall with some strength while small ones
fall gently. Two types of classifying treatment are available for exploiting
this differenc.e, and when sorting an ore the one chosen is that which will
select the most appropriate product for further treatment. For film sizing
or "stream action" (discussed under gravity treatment in sluices and on
Mineral Processing-Industrial Sizing and Sorting 201
shaking tables) the greater the difference between the sizes of two equal-
settling particles, the better will be their separation. Hence, if a sufficiently
coarse break or liberation grind to give Newtonian settling is called for,
classifiers which give maximum size differentiation in their sorting work are
used. These are hindered-settling classifiers. When the break point of
the ore must be taken at a mesh so fine as to bring the particles toward or
into the realm of Stokesian settlement the crowding of the particles through
a dense teeter bed, characteristic of the operation of hindered settlers, is too
violent to be practicable. A change is then made to a system which permits
free settling through a fairly quiet fluid pool, in which the specific population
is much lower. This form of separation is used in mechanical classifiers
and thickeners. Hindered settlement is used with separating "fluids"
carrying from 40 % to 70 % solids, while free-settling conditions operate
with between 3 % and 35 % of solids in the separating medium. Hydraulic
classification is applied to the sizing of homogeneous fine gravels and sands,
to settle a relatively fast-settling coarse fraction from a slower one. Class-
fiers (including thickeners, which exploit a more slow-moving variation of
the same principles) are used for a wide variety of purposes, including sizing,
sorting, desliming of foul effluents, dewatering muddy pulps, adjusting the·
solid-liquid ratio of a pulp, and development of greater response of ground
ore to concentrating processes.
" ... two main types, namely, Surface Classifiers, wherein the sizing
is effected at the water surface, by the water which brings the material
to the classifier; and Hydraulic Classifiers, wherein it is accomplished in
a restricted passage by fresh or added water introduced below. Surface
classifiers are employed for finer material, the discriminating velocity
being that which wells upward across the relatively extended surface at
the level of overflow; hydraulic classifiers are used for coarse material,
say above 80 mesh, the restricted passage permitting the requisite rising-
velocity to be obtained with no great amount of added water."
Feed
~/OW
Distrib"tion
Board
Y
........---."...--
~
OverflOw
Lip
acceleration, which may throw them out of the vortex in any direction from
vertical to horizontal, and this force acts without discrimination on large
and small, heavy and light particles. On the whole, the heaviest and coar-
sest sands gravitate downward to the bottom discharge while the lighter
material is crowded back to the top overflow. In its crude form, the spitz-
kasten does not make a particularly satisfactory separation, but it illustrates
the fundamental principles at work, and is therefore worth examination.
First. As with all classifiers, only two products are delivered. Every-
thing entering must either depart above or below. The greater the propor-
tion of solid withdraw.l below. the greater will be the amount of fine material
included in it. With feed and discharge rates steady and equal the total
number of particles, and their size distribution, is definite. All that can be
done in classification is to vary the ratio between the quantity of coarser
sand leaving below (via the spigot discharge) and the relatively fine overflow.
If the spigot were closed, all the pulp would eventually overflow without
classification, and this overflow would contain the maximum possible per-
centage of coarse particles. The greater the fraction removed as relatively
coarse oversize the finer must be the average size of both overflow and under-
Mineral Processing-Industrial Sizing and Sorting 203
flow, and vice versa. Separation does not mean that only coarse particles
will underflow and only fine ones overflow. This pitfall sometimes confuses
the learner, who imagines that if the coarsest particles are being removed
from one !,oint the result must be to give a much finer product at the other
withdrawing place. What happens is that classifier adjustment simply varies
the splitting conditions. Overflow and underflow are respectively fine and
coarse relative to this splitting.
Second. Since the volumetric capacity of the spitzkasten is constant, the
settling rate available to a particle varies as the volume of feed. Other things
remaining unchanged (including the ratio of solid to liquid in the feed), the
greater the fed volume, the less will be the dwelling time of each particle,
hence the coarser the overflow fraction since it has less time in which to settle.
Third. If the rate of feed is constant, but the solid-liquid ratio varies, the
specific gravity of the pulp varies with the increase or decrease in content of
solids. Since the rate of settlement varies inversely as the solid fraction, the
more watery the feed, the finer will be the average size of overflowing particles.
Fourth. It is customary to consider that classifiers separate particles by
virtue of a combined vertical and horizontal movement. To this end, the
maximum cross-section is usually provided in the horizontal plane of over-
flow, and in the surface area immediately adjacent to the overflow lip all
particles are deemed to move horizontally. Hence, the argument continues,
if in this area the particle is swimming or drifting in the overflowing current,
it is carried out of the classifying system. If its physical relation to the
system at this point causes it to sink, it is retained and may sink to the under-
flow zone for discharge in the coarse fraction. Teetering particles are dis-
cussed later in this chapter. Hence, the capacity of a classifier is related to
its horizontal area in the plane of overflow, and the concept of "surface
classification" or "surface-sorting" is closely bound up with pulp behaviour
in this plane. In theory, it should be possible to control the density of the
pulp at this surface, and also the horizontal speed toward the overflow which
influences the drifting rate of a particle. Operation based on these consid-
erations. together with those of mass and surface friction, should lead to an
accurate sorting system. In practice, only rough sorting takes place in the
spitzkasten. The entering particle undergoes random acceleration, in a
confused mixture of interfering vortexes. It passes from one solid-liquid
ratio to another in various parts of the box, with varying freedom of packing.
Finally it is delivered to the sorting surface with an unpredictable kinetic
energy directed at any angle by the vortex from which it is separating at the
moment.
Fifth. The shape of the spitzkasten provides quiet zones down the corners,
and beds of sand pack into them. Some particles work their way over these
beds to the overflow without undergoing sorting action. Toward the con-
tinuous underflow, periodical collapse of the packing chokes the discharge
orifice, causing surging and abrupt change in the downward flow of pulp.
Appreciation of the foregoing considerations will aid in understanding
the physical limitations of classification in the more elaborate appliances
discussed later. The first improvement in simple spitzkasten work is to use
a series of boxes. increasing in volumetric capacity, so that the coarsest sands
204 Mineral Processing-Industrial Sizing and Sorting
are removed first and the finest last (Fig. 85). To avoid settlement of ma-
terial on the sides, these are made steep. With the larger boxes this would
lead to high pressure on the underflow and the production of too watery a
discharge there. This is avoided by using the "gooseneck" discharge pipe
shown, with suitable provision for clearing it should "tramp oversize" from
a preceding spitzkasten settle in the pipe and choke it. At one time spitz-
kasten series, either in box or trough form, were widely used ahead of gravity
separation, but they are not much seen today. Baffles must not be used with
spitzkasten as they would interfere with surface selectivity.
To overcome the surging effect produced by the periodical sliding down
of settled sands from the corners of the box, cones are used. Feed is central
and overflow peripheral into a launder. If provided with a diaphragm to
prevent very fine material from dropping through with the coarser sands,
the appliance is called a diaphragm cone, and stilI further regulation of the
solid-liquid ratio at the bottom discharge may be provided by a float system
which varies the aperture in accordance with changes in the feed (Fig. 86).
Very little classification is done in these cones once sand has settled to a
fairly compact mass. The selective action occurs in the fluid layers of swir-
ling pulp above this bedding. Since feed is introduced centrally and the
pulp is then swept through a rising path toward the peripheral overflow, the
coarsest and heaviest particles tend to drop straight through or to settle out
from this eddying stream, while those smaller than a certain mesh are lifted
to the overflow lip. The solid-liquid ratio of the underflow depends on the
closeness of packing of the settled sands.
Under favourable conditions, the cone compares inefficiently in perfor-
mance with the far more costly mechanical classifier, save that it is unable
to restore the loss of head of the coarse sands dropping through it, and can-
not therefore close a milling circuit without the additiofl of a pump of some
kind. It can be used for desliming or dewatering. Simple devices based
on the spitzkasten find little application in modern plants, but have a limited
field of use in small operations.
Nests of boxes, with either pyramidal or V-shaped settlement zones, were
developed for thickening dilute suspension of fine sand, and for reclaiming
mill water. They offer a large quiet zone with a gentle motion over the
surface-classifying area. They afford a gentler application of the sorting
principles at work in the spitzkasten, in the same way that the settling cone
does in comparison with the cone classifier. Their place today has been
taken by thickeners of the type described in the next section.
The Thickener
Thus far the conditions which affect particles settling quickly from random
vortexes and under crowded conditions have been considered. In the thick-
ener, settlement is free and the particles are given hours to gravitate downward.
The classifiers hitherto considered were either trying to split a long-ranged
feed into coarse and fine fractions in a short time, to trap out coarse sands,
or to remove slimes from a fast-settling pulp. When the particles have been
very finely ground a prolonged dwelling time under quiet conditions is needed
for their settlement. The thickener is constructed with sufficient volumetric
capacity to give this time. It can be used:
(a) to reclaim water from a muddy effluent by allowing the silt to settle;
(b) to decant fouled water or chemical solutions;
(c) to change from one chemical wash-solution to another;
(d) to remove a dissolved mineral product from a pulp;
(e) to thicken (i.e. increase the solid-liquid ratio of) a pulp;
(f) to reclaim some mill water before discarding the solids from a
tailings pulp.
The essential features of a thickener are shown in Fig. 87. In a typical
operation, mill pulp carrying finely ground solids in suspension is fed in
centrally, through a "trash screen" which holds back any debris that has
accidentally entered. The entering pulp displaces part of its volume as a
206 Mineral Processing--lndustrial Sizing aild Sorting
OUWLHO M KAHn,"
ro Pf • .,L A ....
S\.lpu'n~'J'\I POA
SIJI.oIIL A_WI DIlIYINO
MICII4"o111N.
FUD ...t.
TmCI:L .U DUtlO_
DIKMUCt.
vided through which water or compressed air can be injected in the event of
choking. Instead of pumping, some large installations use bottom valves to
run slurry off. The piping discharge is then of ordinary diameters. The
thickened discharge is commonly led to a continuous filter which must period-
ically be shut down for servicing. During such a period, which may last
several hours, the thickener continues to receive feed and must store its slerry.
Provision is made for raising the rake mechanism to prevent overstrain under
such conditions. Since this is a safety precaution, the rakes must be gently
lowered as soon as normal running has been restored and the loading has
been reduced.
Choking around the well of the thickener is a serious matter that may lead
to the shutting down of the whole mill. Solid objects fall in, either through
sabotage, a kink in human nature, or through failure of workmen on the
thickener to tie loose tools to a safety belt. It is therefore a wise precaution
to have a run-off tank availj!ble into which the contents of the thickener can be
sluiced in emergency. If the thickener is "stalled", speedy repair is essential,
as there is rarely standby capacity to which feed can be diverted.
Where space is cramped, or where the risk of freezing entails protection,
tray thickeners having from two to six compartments are often used. The
pulp is divided into equal streams, and each is fed centrally to a compartment.
In one type each compartment rakes the settling slurry to a common well-
discharge. In another type the slurries can be kept separate. The water
overflow rises naturally from the lower trays to join that from the top peri-
pherallaunder, or, if desired, these overflows can be kept separate.
Failure of the settled slurry to come away from the well may be due to a
solid object obstructing the outlet. a choke in the piping system or a defect in
the diaphragm pump, such as a stuck valve or a ruptured membrane. Thick-
eners receiving a flotation concentrate sometimes build a thick scum of floating
froth on their surface. To minimise this, the feed should run in gently, since
splashing is likely to entrain air-bubbles.
Little power and attendance are normally required, but when a thickener
breaks down it can very seriously upset the running of the plant, since it
usually constitutes a "bottle-neck" in the flow-line. To simplify maintenance
and avoid a lengthy shut-down thickeners are sometimes built with an
approach tunnel below, ending in a pump room. Alternatively. underflow may
be pumped up through a central column large enough to permit a man to
enter.
Among recent developments in thickener designS are a two-stage raking
zone. The peripheral area has a relatively shallow gradient. while coarser
material which tends to settle more centrally falls to a steeper central portion.
The inner rakes which sweep this zone are attached to posts below the trusses.
These posts cut through the deposit and open channels through which water
can squeeze upward. The trusses themselves are not subjected to the strain of
shearing through this material. Another design is a flat-bottomed thickener
with peripheral discharge.
Where conti.nuous thickening is coupled with periodic filtration so that
storage capacity inside the thickener is required, automation has been suc-
cessfully used to raise or lower the rakes in accordance with changes in torque
Mineral Processing-Industrial Sizing and Sorting 209
signalled to a motor which lifts or lowers the raking mechanism. For sub-
zero working of exposed plant, electric heating of driving gear and pump
room can be used. For a thickener which is used to clarify a thin, but finely
divided and slow-settling feed, hydraulic disturbance where the feed enters
can be avoided by delivery through vertical screening which divides the flow
into a number of thin streams and checks these by baffling.
Thickening Theory
In 1916 Coe and Clevenger 6 produced a formula for calculating the thickener
area required in handling a known rate of loading. Their observations were
based on the zone sequence seen during the settlement of pulp~larification
of the uppermost layer and increasing pulp density down to the final stage of
compression ;(the critical point) below which no further settlement occurred.
Thus at any horizon in the transition zone there was a change in pulp density
which affected the subsequent settling rate and the specific population at that
horizon. Their calculation of the required thickener area was formulated in
respect of a transitional rather than a final or saturated concentration, and
within that limitation was dependable. It is
1·33 (F-O)
A (9.3)
RS
A (9.4)
where tu is the time in days, Co the concentration of feed as tons of solid per
ft3 of pulp, and Ho the height in feet.
Both formulae bring out the fact that (ignoring any effects of flocculation
or need for clarification to completely clean water) sedimentation is governed
by pool area rather than pool depth. One element in thickening is, however,
210 Mineral Processing-Industrial Sizing and Sorting
Hydraulic Classifiers
These appliances are also called "spitzlutten" and upward-current classifiers.
In its simplest form (Fig. 89) the hydraulic classifier resembles the spitzkasten.
with the essential differe'lce that. in addition to horizontal sorting at its maxi-
mum cross-section where the lighter particles overflow the discharge lip.
"""\
lSpl,ol d/,eltarp'
precision, but far more cheaply, than on the delicate fine screens which would
otherwise be needed. With sufficiently fine particles, classification can be
performed in columns of air instead of in liquid.
In mineral processing the main purpose of classification is to sort a feed
into two classes, settling and rising. If this feed contains heavy and light
mineral, small heavy particles will have the same falling characteristics as
larger light ones. They can therefore be dropped together and then separated
on a screen of an appropriate intermediate mesh.
Many shapes and varieties of hydraulic classifier have been developed.
Truscott 4 observes in his text-book:
The teeter column thus heightens the separating effect based on exploitation
of differences in mass and S.G. It also has a scrubbing action which aids in
cleaning th~ discharged product that runs the gauntlet. This scouring
arrests fine particles which otherwise might be carried, attached to larger
particles. to the lower discharge. The hydraulic column must be steady, and
is therefore fed from a constant-pressure source of water. Some water is
withdrawn with the solids leaving at the bottom discharge. and the rest is
discharged as the overflow. together with the water in the original feed.
Classifiers can be arranged in series. to give progressively finer underflow
products. They may receive feed from a desliming cone. or send out a final
overflow product for quieter sorting in such a cone. The coarser the desired
underflow. the more strong must be the rising current.
The teeter bed is intermediate in settling character. It tends to build up so
as to alter the character of the products discharged through it. To minimise
this defect. a class of hydraulic appliances having the general distinguishing
name of "hydrosizers" has been developed in which this teeter bed is
controlled as to composition. As the bed builds uP. the weight of its com-
ponent quicksand also increases. This can be measured by providing a water
connection in the form of a hydrostatic head which shows changes in the
back pressure exerted upon the entering hydraulic water (the "added water"
of Fig. 89). Appropriate control mechanism can be actuated by pressure
Mineral Processing-Industrial Sizing and Sorting 213
changes as in the Stokes hydrosizer (Fig. 90) in which the bottom discharge
aperture is opened or restricted in response to changes in the load of teetering
sands. This classifier is one of several made in multi-spigot form, to deliver a
series of graded products.
Another form of hydrosizer takes the shape of a miniature thickener.
Its rakes rotate fast enough to loosen up the settling sands and upward cur-
rents may be imposed upon the contents of the circular tank. By suitable
use of these combined forces very fine material is caused to overflow, while
denser particles drop and are raked to a central well-discharge.
Mechanical Classifiers
In the so-called "mechanical" classifier the pulp is fed into a rectangular
tank under conditions which allow the heavier and coarser solids to gravitate
fairly freely downward, while the lightest particles flow to a weir discharge.
The "mechanical" part of the classifier is usually a drag belt, a set of rakes,
or a spiral screw. Its functions are to stir the pool of pulp and remove
settling solids. In Fig. 91 a cross-section through the type of tank usual with
a Dorr (rake), Akins (spiral) or drag-belt classifier is shown, without its
mechanism. The pulp overflowing from the grinding mill carries 70%-80%
solids, and flows to the classifier through a short launder, or possibly after
elevation through a centrifugal pump. En route or in the tank more water
is added, so as to thin the pulp to an operating density at which overflow of
finer sands and settlement of coarser ones are most effectively produced. In
the launder the pulp stream is split into two or more channels which distribute
the feed fairly equably across the pool, approximately two-thirds of the
distance from the weir in direction of its "V" end.
Ignore for the moment the stirring effect produced in this pool by the
raking gear. The pulp, diluted to say 30% solids, falls from a slight height
into the shallow end of a wedged-shaped body of fluid. The height of the
feed launder above the pool determines the kinetic energy with which this
plunge is made. Rate of settlement through the pulp results from the inter-
play of complex forces. These may include thixotropy, adhesion and co-
hesion, which receive further discussion in Chapter 15. These forces influ-
ence shear in the body of the pulp under stated operating conditions (mainly
concerned with temperature, specific surface and turbulence). For con-
venience of operating control this complex is treated as a simple S.G. effect
for which the classifier pool is held at a specified operating density. This
determines the resistance to settlement which is offered at successive horizons
to the particle as it seeks to gravitate downward. There are two channels of
escape from the pool. A particle having sufficient mass to force a passage
to the bottom, or raked zone, falls and is quietly withdrawn by the mechanical
withdrawing device (not shown in Fig. 91). A particle which has insufficient
mass to descend can drift with the current from the feed launder to the over-
flow weir through the transporting zone. Equal-settling particles will report
together. A small heavy particle will settle with a larger one of lower
density.
214 Mineral Processing-Industrial Sizing and Sorting
If all particles in the classifier were equal in size, a clean separation into
heavy (sinking) mineral and light (overflowing) mineral would be possible. In
fact, the mechanical classifier tends to retain the heavier mineral in the ore
and return it, via the raking mechanism to the feed end of the mill with which
it is close-circuited even when the operator desires it to be sent on. The
particles comprising the solid fraction of the feed cover a wide range of sizes,
in an inclined tank and are actuated via an eccentric motion (8) and two sets
of link motion, 9 to 13, and 3 to 6. These links cause the rakes to moves
through an elliptical orbit, almost flat on its long axis which corresponds
with the slope of the classifier tank. The rakes start their climb at the lowest
setting, and gather settled sand. At the rising end of the stroke they lift
sharply. They then return, drop, and repeat the cycle. These classifiers
are described as simplex when the trough has one compartment, duplex and
quadruplex when there are two and four divisions separately raked. Pro-
vision (not shown) exists for raising the rakes. Spiral classifiers (Fig. 93)
is run at a speed varying with its maximum diameter, the shaft usually turning
at between 3 and 6 r.p.m. for a large spiral, and up to 20 r.p.m. for a small
one.
As settled sands emerge from the "V" end of the pool, they bring with
them some trapped or "drag-out" slime which should properly have overflowed
at the weir. Washing sprays are often used to return this material to the
pool. They offer a simple means of adjusting the pool density by providing a
region where extra water can be mixed in, but are not very efficient as deslimers.
With the rake classifier the emerging sands usually take an undulating con-
tour, and if the sprays arc allowed to play over a length of rising sand the
slime is trapped in the depressions, ploughed under at the next dragging
up-stroke, and thus smuggled through. To be effective the sprays must be
directed forcibly on to sands close to the point of emergence, and other
methods must be used for making final additions of water to the pool as the
sprays must run at full pressure. In this respect the spiral classifier is at an
advantage. It carries its sands upward on the rising side of the helix, so that
sprays can be used at any point and there is a good channel for running their
washed product straight down to the pool. The churning of newly emerged
sand by sprays causes it to slip back and interferes with the elevating function
of the rakes. Sufficient height must be gained for the sands to run by gravity
back to the feed scoop of the mill, after they fall from the classifier into the
return launder. The continuous and more gentle action of a spiral makes
possible a steeper slope in the classifier tank, settings of from 3" to 4" per
linear foot being common as against 2" to 3", with an occasional 31", for
reciprocating rakes. Removal of slimes reduces the tendency to slip, but a
certain amount of deliberate inefficiency is liked by some millmen, who
consider that the slime has a valuable lubricating effect in the feed trunnion
of the ball mill, and that without it the mill could not handle so large a cir-
culating load. In the "Overdrain" classifier a drag belt moves between
stationary longitudinal walls (called shrouds) so as to carry material trapped
between successive moving sections in bottomless compartments. From
these compartments supernatant slimes overflow through side openings and
back to the pool between shrouds and sides of classifier.
In the event of power failure, or shut-down, care must be taken that the
raking gear does not become buried as the sand settles compactly down.
Reciprocating rakes can be raised in some models, and spirals can be swung
clear. If a classifier is restarted with buried rakes, the mechanism may be
injured. Rakes and helixes wear. and are renewable.
If tramp oversize or steel from the mill overflows to the classifier, it may
lodge so as to project from the bedded zone. This causes rough action and
undue wear of the sand-moving mechanism. Spiral edges can be notched, or
given welded projections to dig such material up and get it back to the
mill.
The separating point can be up to 28 mesh, or anywhere between this and
minus 200 mesh. Several factors are used in setting the machine to deliver at
and around the required mesh. Since a particle must rise from the body of
the pool in order to overflow. the upward current in the vicinity of the weir is
one determinant. This weir. which is adjustable for height. is at the deep end
218 Mineral Processing-Industrial Sizing and Sorting
of the pool. The upward (hydraulic) current can be varied by the introduc-
tion of a baffle (see Fig. 91). The deeper this is set, the deeper the transporting
current is forced before an upturn in the flow stream of pulp can take place.
The distance between baffle and weir determines the cross-section of the rising
column, and therefore, the rising speed of the transporting current. For a
coarse-grained overflow the baffle may be set as close as I t" from the weir, and
for finer separations as far back as 24 H. It may be removed altogether for very
fine work.
The finer the release mesh, the longer must be the settling time given to the
entering feed and the gentler the stirring. The volume of pulp available for
sorting purposes is therefore an important factor. It is affected by the slope
of the tank bottom, the height of the weir, and the width of the classifier.
Usually the slope is made as steep as is possible without risking slip-back of
raked sands, in order to obtain maximum elevation. This aids their drainage
and allows the use of a stt:eper fall to the feed box of the ball mill from rake
discharge into the return feed launder.
Consider a classifier set with a slope of 3" in the foot, and with vertical walls.
The enclosed pulp has a wedge-shaped volume and a triangular longitudinal
section. When the weir overflow is raised, the progressive increase in the
volume of the pulp is proportional to change in longitudinal area (Table \5).
TABLE IS
The effect is to give each particle entering the pool a corresponding overall increase
in settling time.
The weir height affects the capacity in the pool. It also influences the
total area of horizontal cross-section in the plane of overflow. When the
weir is high, it damps the surging motion set up by the plunging movement
of the reciprocating rakes of the Dorr-type classifier. The greater the ton-
nage the classifier is handling. the bigger should be the pool. and hence the
higher the weir setting.
Lowering the weir helps when a denser overflow is needed. Care must be
taken, particularly when fine pulp is being sorted. not to reduce the capacity
of the pool too severely or cyclic surging may become excessive. This would
allow sands to escape from the closed circuit before they had been reduced to
the proper overflow-mesh. and lead to trouble in the concentrating section of
the plant. Consider the classifying zones shown in Fig. 91. When no baffle
is used, the transporting zone merges layer by layer into the sorting zone.
Mineral Processing-Industrial Sizing and Sorting 219
Think of the pool as it would be under conditions of smooth and steady flow,
with no stirring action from rakes to break up the stratification. It would
soon arrange itself in a series of thin layers, sorted from the vertical cross-
section of the feed and becoming more individualised as they flowed toward
the weir. The surface layer would have a density approaching that of water
and the bottom layer that of a high solid-liquid ratio. Further, the grain size
of each mineral would be highest at the feed end and lowest at the weir end,
because of the combined effect of gravity and rate of drift upon the particles
moving in or through the layer. (This discussion temporarily ignores two
factors-the upthrust of the layers at the weir end and the accumulation of
newly arriving particles at the bottom of the pool.)
An individual particle moving in such a layer can drop through, if its
gravitational force is sufficient to overcome its frictional retardation, during
the period of its travel from feed to weir. It can rise into a less dense layer
through eddying or if it is squeezed up and out by heavier particles. Each
of the lower layers is continuously receiving such particles as they drop down-
ward, and receives a greater load than it can part with by rejection upward, or
downward, of a corresponding weight of solids. Thus the near-release par-
ticles accumulate, and the sorting zone becomes too dense to work properly.
The generalized effect is that a rise in the overall density of the pool occurs,
accompanied by an increase in the pseudo-viscosity so that new feed finds it
increasingly difficult to fall through. This can be seen in operation particu-
larly in a circuit set for fine grinding such as is frequently closed by a bowl
classifier. The level of the pulp in the main tank rises and falls in a cyclic
rhythm, corresponding to a gradual rise in the overall specific gravity of the
pool. This is periodically relieved by a surge of heavy pulp over the weir and
a drop in the overall S.G. The word "overall" includes normal running
density and the heightened one of the high-density surge period, since there
need be no change in that of the overflowing pulp during the period of build-
up. This would be particularly the case if little or no use of a baffle was made,
as the lightly loaded surface layers of the pulp would be disproportionately
represented in the samples used to check the specific gravity. For this
surging to occur, more solids must arrive in the pool than are departing.
During the rise, the rakes become more lightly loaded although no change has
been made in the rate of new feed to the mill. If at this stage vigorous action is
taken to dilute the body of the pool (say, by directing a jet of water into it),
the rakes load up to normal, or beyond, and proper recirculation is restored.
This phenomenon only becomes serious when the mill has insufficient
classifier pool capacity for its input of ore. These layers can to some extent
be mixed by raking, or by setting up turbulence in other ways, e.g. by allowing
the feed to plunge into the pool from a slight height. When water is added
to reduce the density of the pool it should be mixed in, not simply used as a
spray wash on the raked-up sands. Unless it is mixed in, it may run as a lightly
loaded layer above the sorting zone, simply washing out some fine material
by its streaming action. The use of the baffle forces all surface water down
into the body of the pool and aids such mixing. Provided the diluting water
is properly mixed in-partly in the feed launder and partly in the body of the.
pool toward the weir end, the density of the overflowing pulp becomes the
220 Mineral Processing-Industrial Sizing and Sorting
vital operating control. If this mixing is not assured, the value of the control
is weakened.
There is a certain ratio of solid to liquid, called the critical density or the
critical dilution, above which the size of particle overflowed depends upon the
pulp density in the upper part of the pool. If the pulp is diluted below this
critical density, the particles no longer respond significantly to this form of
control. The critical density varies with the ore, temperature, and the
flocculating effect of chemicals such as lime which are frequently added in the
grinding circuit for reasons explained in the chapters dealing with flotation.
Broadly, the percentage of solids at which change is to be expected is some-
where between 3 %and 5 %. The most important factor in this critical density
is pseudo-viscosity or specific particle surface. A given ore has a critical
density proportional to the square of the particle mesh for a given solid-liquid
ratio since this governs the total area of surface in a given volume of pulp. If
the feed contains clay or other primary slime, the critical density is lower,
because the pseudo-viscosity is higher.
Provided the solids which should be returned to the mill for further grinding
are able to fall through to the raking zone, the speed at which the rakes are
run depend on:
(a) the degree of turbulence desired in the pool to break up the layers
as they try to form;
(b) the amount of raking needed to withdraw settled material.
Centrifugal Classifiers
The rate of fall of a particle varies as its effective mass. If centrifugal
force is applied, the effective mass is increased and, provided nothing hap-
pens to offset this effect, settling rate is higher. As particles are ground smaller
they reach a size where the surface drag against the surrounding fluid almost
neutralises the gravitational pull, with the result that the particle may need
hours, or even days, to fall a few inches through still water. This slowing
down of settling rate reduces the tonnage that can be handled and increases
the quantity of machinery and plant required. A lOp. particle of silica
settles through water at speed varying round 6 mm./min. which, for many
purposes, would be too slow. By superimposing centrifugal force the
gravitational pull can be tremendously increased.
The Hydro-Cyclone
During the past few years cyclones have replaced mechanical classifiers in
many grinding plants. The liquid-solid cyclone (this name, and the prefix
"hydro" usually being omitted) was introduced by Driessen in 1939, and the
references ll - 16 indicate only a fraction of the literature already published
concerning its principles, applications, and performance. When a pulp is
fed tangentially into a cyclone (Fig. 96) a vortex is generated about the
longitudinal axis. The accompanying centrifugal acceleration increases the
settling rates of the particles, the coarser of which reach the cone's wall.
Here they enter a zone of reduced pressure and flow downward to the apex,
through which they are discharged. The percentage of feed leaving as coarse
product depends on the aperture of the inlet and vortex finder provided the
underflow does not exceed some 30 % of the feed. At the centre of the
cyclone is a zone of low pressure and low centrifugal force which surrounds
an air-filled vortex. Part of the pulp, carrying the finer particles, moves
inward toward this vortex and reaches the gathering zone surrounding the
air pocket. Here it is picked up by the tube called the vortex finder, and
removed through a central overflow orifice. The vortex finder is so adjusted as
to project into the cylindrical section of the cyclone, and short-circuiting of
Mineral Processing-Industrial Sizing and Sorting 223
PRESSURE
GAGE
INCLUDED
ANGLE CONE
SECTION
APEX
OI'ENING
U DERfLOW
half the solids are discharged as the coarse fraction (apex discharge or under-
flow) and the rest ria the central overflow.
Q in this equation is a function of inlet diameter. overflow diameter,
pressure drop in transit and a constant K which varies with the included angle
of the apex.
Tarjan's equation 17 for the size of particle which revolves in equilibrium at
the circumference of the cyclone cylinder is
42e 2
d----
\ (Ps-P) h Q (9.6)
where h is the height of the cylinder in cm., e is given in cm. and Q in
I/s. son Fahlstrom observes that the hydro-cyclone has two sets of
characteristics. First are the ones fixed by construction-diameter, area of
feed entry, length of drum and vortex finder, and cone angle. Second are the
operating variables which include pulp concentration, feed pressure. diameter
of overflow pipe and of vortex finder. The effect of the classifying action
can be defined in terms of weight of yield and of percentage of solids in the
underflow. Separating size is given by his equation
8=ko (l-gu)lln (9.7)
I. Sharper classification.
2. Saving of floor space.
3. Less power consumption.
4. Less maintenance.
5. Ability to shut down the mill immediately under full load.
6. Ability to bring the circuit rapidly into balance.
7. Elimination of cyclic surging.
The cyclone is increasingly used for classification in the finer grinding ranges,
between l50fL and 5fL, although coarser separations are possible. Separating
efficiency is measurable as the percentage of misplaced product in either the
overflow or underflow. Confusion can arise in assessing operating efficiency
if the apex discharge is not under proper control. With a sprayed discharge
the issuing solids should not exceed some 70 /~ by weight of the total amount
leaving at the apex. It is sometimes set to give a thick underflow carrying a
higher percentage of solids. This means that the cyclone is overloaded and
that some oversized material is unable to report with the apex discharge,
and consequently is overflowed with the finer fraction. The correct adjust-
ment can be made by varying the diameter of the apex discharge. This can be
automated by pneumatic or hydraulic adjustment of the apex, signals being
initiated in the vortex chamber or feed zone.
The cyclone does not act as an effective substitute for the thickener in
dealing with material below about 5fL in size. It can, however, concentrate
the feed to a thickener or, alternatively, remove the bulk of the solids in the
underflow. The overflow then carries a relatively small percentage of the
finest solids, and can be led to any convenient settling area for further thick-
226 Mineral Processing-Industrial Sizing and Sorting
discharged by means of the screw, while the finer fraction over flows at the
other end.
The largest of the three spheres starting from rest at the point marked by
an arrow falls fastest. It is therefore exposed for the shortest period to the
displacing effect of the horizontal air stream. The smallest drifts the furthest
under the same conditions.
The sheltering, and the frictional drag upon air and particles touching the
containing walls of a vertical vessel, is sufficiently pronounced to require
special precautions. Build-up of particles on these containing walls may
occur, and machines used in air sizing and dust control sometimes contain
devices for preventing this from becoming serious. Friction in a dry atmos-
phere causes electrostatic forces to be generated upon the particles, and aids
their adhesion to one another and to the walls of the vessel. Apart from
effects of this kind, the physical characteristics which determine how a given
particle will behave in an air stream are its size, shape, density, and liability
to collide with other particles (dispersion or concentration). The effects
produced by the air stream depend on its velocity, humidity, viscosity (a
pressure effect), and the way in which it is constrained to move by the con-
taining walls of the appliance in which it operates. The settling velocity of a
particle in still air is quoted by Taggart for irregular grains of S.G. 2·5 as
3140 ft./min. for 5000fL particles; 470 ft./min. for 500fL; 50 ft./min. for 24fL
and 0·5 ft./min. for 0·OO32fL or 3·2mfL.
If the velocity of the transporting stream of air is suitably controlled,
relatively coarse particles are dropped while finer ones are carried onward.
Baffles and deflecters may be placed in the stream to sort out the coarse
particles in accordance with their inertia. In the cyclone a centrifugal
sorting action is set up, the dust finding its way to the sides and then falling
down while a vortex of comparatively dust-free air rises at the centre. The air
~=--
is usually moved by a fan, working either to push a clean column of air for-
ward or to draw dusty air through the system. To avoid wear of the blades,
fan-power should be applied to clean air, or, alternatively, after the coarse
particles have been trapped. A modified form of air classifier-the "gravi-
tational-inertial"-for removing minus 100 mesh material from crushed
limestone has been developed. 20 (Fig. 102.)
Despite the application of centrifugal force, the finest particles of dust
cannot be trapped in a cyclone. Such dust might be a menace to the health
of the workers. Various methods are used to minimise this risk. Cyclones
can be used to remove dust coarser than 5fL. Parts of the mill operations
230 Mineral Processing-Industrial Sizing and Sorting
S~condary
aIr
~ Coars~ product
molecules collide and ionise. The ionised molecules can then be attracted
toward a near-by grounded conductor. As they migrate they charge any
passing dust particles they may collide with, and influence them to travel
toward the grounded electrode. Dust can also be precipitated by passing it
through a duct in which supersonic waves are vibrating at a sufficient intensity
to cause the particles to flocculate into rapidly settling clots.
If two particles of the same shape and size, but of different specific gravities,
are fed into a mechanical classifier, the heavier one may sink and be returned
for further grinding while the lighter one overflows at the weir. Under such
circumstances the mechanical classifier becomes a concentrating machine.
This possibility was noted above when classifiers were referred to as sorting
devices, not sizers. It is often undesirable that particles should remain in
Mineral Processing-Industrial Sizing and Sorting 231
the closed circuit in this way. A metallic sulphide, once properly liberated,
can be caught in the concentrating section of the plant more efficiently when
it is still somewhat coarse than after it has been ground to an impalpable
slime. A particle of free gold is usually malleable and merely changes its
shape as the result of further passages through the mill. When necessary,
effects of this kind are partially countered by the introduction into the closed
circuit of a moderately efficient concentrating device appropriate to the
working conditions. This is used to remove as much as possible of the de-
sired value in the form of a rough concentrate, in order to minimise its over-
grinding or accumulation in the closed circuit. Such appliances are collo-
quially termed "scalpers". Jigs, flotation cells, hydraulic classifiers, and
corduroy strakes are among the devices used for this purpose.
Classifiers v. Cyclones
The closed circuit in wet grinding is rarely controlled by screening. The
choice lies principally between mechanical classifiers and cyclones, though
flow-sheets using gravity processes of concentration may have hydraulic
classifiers for reasons which are discussed in Chapter 14. The growth in use
of the cyclone has been rapid during the past few years, and data for adequate
comparison with the older appliances are still inadequate. The mechanical
classifier has the advantages over its rivals of tolerating and smoothing out
surges, of returning its oversize to the mill launder without need for an extra
machine, of being robust, long wearing, and easy to control. The screen
fails because it cannot handle fine sand efficiently, and is therefore unsuitable
for liberation of most of the low-grade and fine-grained ores which are treated
today. Use of the cyclone owes its phenomenal growth to several facts.
First, by its use of centrifugal force it can speed up settlement rate and there-
fore either handle larger tonnages with light equipment in a small space, or
make a separation at finer meshes than can the classifier. Next, running
costs are comparable but capital cost and installation are far cheaper.
Third, by returning its oversize direct to the feed trunnion, it does away
with the need for a scoop and feed box. Next, a point which specially
affects flotation, it only keeps a small tonnage in circulation and hence
reduces oxidising effects in the grinding circuit. Last, the limitation
in fine grinding when the circuit is closed by a bowl classifier is that
only a moderate circulating load is possible, the limit being dictated by the
free-settling speed of the near-release particles. Using the accelerated
settlement due to centrifugal force the cyclone makes possible a large circu-
lating load. Repeated passage through the secondary mill, as with the primary
one, lessens over-grinding.
Classifier Efficiency
This is checked by making screen analyses of properly collected samples of
(a)material entering the classifier, (b) coarse discharge, (c) fine discharge.
232 Mineral Processing-Industrial Sizing and Sorting
These sizing analyses, plotted as "direct" graphs, should show a marked peak
at the desired mesh of maximum separation, while the relation of the under-
size and oversize to this peak mesh is a guide to behaviour in the circuit
under the conditions ruling when the sample is taken. If operating changes
are called for, they must be made one at a time, empirically. It must be
remembered when judging efficiency that each change made anywhere in the
closed circuit affects the whole circuit. Classifier efficiency cannot safely be
considered in isolation from the conditions in the grinding mill and/or
concentrator.
The control used by the shiftsman during the hour-by-hour running of the
wet-grinding closed circuit corrects the density of the overflowing pulp.
Provided this pulp is truly representative of the conditions in the top few
inches of the classifier pool and is not upset by the presence of an overrunning
stream of diluting water, its two sorting components-pulp density and
hydraulic transporting energy-exercise effective control over the mesh
sizes in the overflow. The accurate control of pulp density is therefore a
major factor in operating efficiency. Haultain 21 observes that most bowl
classifiers are worked with a cyclic surge of between 15 min. and 30 min.
duration, during which the pulp in the rake compartment rises anywhere up to
9 above the normal working level, while corresponding variations in the
H
References
l. Hitzrot, H. W. (1957). I.M.P.e. (Stockholm), Almqvist & Wiksell.
2. Hukki, R. T. (1960). Aufbereitungs-Technik, 12.
3. Kuenhold, N. (1957). Min. Eng., June.
4. Truscott, S. J. (1923). Textbook pf Ore Dressing, MacMillan.
5. King, D. L., and Schepman, B. A. (1962). Trans. S.M.E., 223.
6. Coe, H. S., and Clevenger, G. H. (1916). Trans. A.I.M.E., 55.
7. Kynch, G. J. (1952). Trans. Faraday Soc., 48.
8. Fitch, B. (1962). Trans. S.M.E., 223.
9. Cross, H. E. (1963). J. S. AI I.M.M., Feb.
10. Gaudin, A. M., and Fuerstenau, M. C. (1960). I.M.P.e. (Lond.), I.M.M.
11. Bradley, D. (1960). Ibid.
12. Peachey, C. G. (1960). Ibid.
13. Kelsall, D. F., and Holmes, J. A. (1960). Ibid.
14. Cohen, E., and Isherwood, R. J. (1960). Ibid.
15. Stas, M. (1957). I.M.P.e. (Stockholm), Almqvist & Wiksell.
16. Fahlstrom, P. H. son. (1963). I.M.P.e. (Cannes), Pergamon.
17. Tarjan, G. (1950). Acta Tech. Hung., No. I.
18. Taggart. A. F. (1945). Handbook of Mineral Dressing, Chapman & Hall.
19. Dallavalle, J. M. (1948). Micromeritics, Pitman.
20. Anon. (1960). Min. Eng. A.I.M.M.E., Nov.
21. Haultain. H. E. T. Can. Mill. J., 67.
CHAPTER 10
Preliminary
This chapter is concerned with the routine control of the closed circuit
in the wet-grinding section of the plant. Sampling control, as designed
to check the state of the pulp at critical stages in its treatment and end-
product disposal, is discussed more fully in Chapter 23.
Control of the closed circuit has three aspects:
(a) Immediate running control.
(b) Long-term quality control, as it affects assay grades.
(c) Capacity control, which considers wear rates and economic through-
put.
The first of these is usually limited in its scope, and is designed to give
the shiftsman discretion to vary a few simple settings in order to maintain
the correct mesh-of-grind and solid-liquid ratio within permissible limits.
With the increasing use of automatic control, possibly linked by computer
with the overall operation, the scope and character of human shift-working are
changing. In some new plants an analogue computer integrates aspects
(a), (b), and (c) and restricts local intervention which might interfere
with the overall control plan. Subject to such considerations, which may
include operating elements outside the concentrator, items (b) and (c) include
controls only applied by the mill superintendent. They are rarely of a type
which can be entrusted to the workmen to vary.
Every alteration in the character of the ore pulp escaping over the classifier
weir into the concentrating section of the plant results in a definite change in
the working conditions there. It is therefore not possible to give an operator
in the grinding section discretion to make changes in his circuit which can
lead to such effects later on. The only person who can properly initiate such
major changes is the one who also can watch and foresee (or be prepared for)
their effects all along the flow-line. Since most of these effects are slow to
show themselves and are only visible as assay-return changes, the mill super-
intendent cannot delegate initiative in introducing changes in the grinding
section. After the bulk of the concentrate has been removed, and the pulp
approaches the final stages of treatment, greater discretion can be delegated
to the sbiftsman. The grinding n:quirement in all good plant practice is a
constant tonnage rate of pulp held closely to its optimum grind and solid-
liquid ratio. This is in control onhe shiftsman until the pulp overflows the
final classifier weir on its way to the concentrating section, and it is his job to
make sure nothing happens ahead of that weir which results in the overflow of
unsuitable pulp.
234 Mineral Processing-Grinding Circuit Control
It may be helpful at this point to recall the idea of the sectional or block
flow-sheet discussed in Chapter I. Transfer from grinding to concentration
is the most important sectional step in the whole series of operations. Mis-
takes made in the crushing section can be retrieved at the cost of efficiency by
grinding adjustment. Mistakes in grinding affect quality and quantity of
value recovered, and must therefore be appropriately guarded against.
Once the main work of treatment has been completed, risks may be taken so
toward the tailings end the shiftsman can be given more freedom to vary his
controls. Hence, the slightly paradoxical position is that the best workmen
may be needed in the grinding section, yet they may be given the least dis-
cretion as to their activities. Vigilance in this case is of a higher order,
concerned to check trouble at its source and not to restore order or take
chances. For such reasons as these automatic control is in increasing use in
the grinding circuit. The bulk of this chapter deals with control as a human
activity used to apply basic principles and maintain running norms.
Running Control
By "running controls" are meant the adjustments of the closed circuit
which can be made by the shiftsman in order to hold the delivery of finished
pulp at the required mesh-of-grind. Consider the closed circuit (Fig. 103).
The new feed varies in size, grindability and moistness. The returned
sand varies in grain size and tonnage rate. If the operator's instructions are
that the crop load is to be held to a solid-liquid ratio of, say 70:30, then he
must adjust the feed water (2) so that 30 units by weight are scooped in for
every 70 units by weight of ore.
Hence:
Moisture (1 +2+5) 30
If the specific gravity of the 70 units of mineral is 2'8, then the volume of 100
70 (weight) 100
weight units of pulp is - -+- 30 = 55 and the specific gravity is - -
2·8 (volume) 55
or 1·82. The density of the mill discharge must be held at this figure. To
check this. the shiftsman periodically weighs a known volume of pulp. and if
the weight is wrong he varies the feed water (2) slightly in order to restore the
correct solid-liquid ratio. The next control he may vary is the rate of new
feed. Since the total load of balls in the mill is constant and the total weight
of ore in the crop load is a definite fraction of this, the total weight of ore
«(I) + (5» fed to the mill should be constant. The amount returning at
(5) is the amount in the mill discharge (4) less the amount overflowed (6). It
therefore varies with the grindability of the ore, and adjustments are made
to (I) to compensate these variations. The shiftsman may be guided by the
Mineral Processing-Grinding Circuit Control 235
power intake of the mill (ammeter reading); the height at which the return
sands stand on the rakes: or the grinding noise of the mill. Any variation in
grindability. or in fineness of (4), alters the rate of settlement through the
j (1)
0)
Returned
Sands. - - t tr---- New
Feed
Ball Mill
~
Ball Mill Discharge (4)
pool of pulp in the classifier and in due course shows at the overflow as a
change in pulp density. The third control is therefore (3) the volume of
diluting water mixed intimately with the mill discharge pulp to form the
pool pulp. This must be varied so as to maintain the correct pulp density at
the weir overflow. The figure should never be allowed to vary when overflow
pulp is being sent to the concentrating section, no matter what emergency may
arise. Summarising, the shiftsman may vary any or all of the following:
(a) (1) Rate of new feed.
(b) (2) Flow rate of mill water.
(c) (3) Flow rate of classifier water.
As these rates react on one another, he must make any variation called for
with due regard to its effect on the others, and must hold the pulp density of
mill discharge and weir overflow at the specified values while maintaining
the optimum load of ore in the mill crop. When these controls are integrated
with reliable monitoring devices which check continuously on such running
signals as noise level in the mill, power draft, density and flowrate changes
in mill discharge and classifier overflow, tonnage rates of new and returned
feed, it is inevitable that the swifter and more comprehensive coverage thus
afforded must exceed human ability in detecting variance, and in aiding
corrective measures.
236 Mineral Processing-Grinding Circuit Control
!.
!.SO
70
~
:s
.: 1.500
1.600
1.550
"1
!.70-
~
....
~
~
0 '1)
~ ~
..
.~ 1.450
1:11
~
R.BO- _ei
c:s
0 ....
.
~
"
1.400 01
.
Q
Q
GO ...
~
-
1.350 c..
1.90 GO
1.300 c:
d'
:;
q; 8.00
1:250
1.200 8.10
1.150 8.20
1.100
8.80
The effect of varying each of these controls is set out in Table 16.
TABLE 16
Effect
Controlled
Variable Increase Decrease
(a) Rate of new Average particle size larger, Longer dwelling time, finer
feed (I). coarser discharge and faster grind, less return sand, more
throughput. Higher circulat- over-ground particles in weir
ing load. overflow.
--------
(b) Mill water (2). Flushes fines through mill. In- Increases coating on balls,
creases wear of steel. slows progress through mill.
May choke mill.
(c) Water to clas- Finer release mesh. Greater Coarser overflow. Lower cir-
sifier (3). circulating load. culating load.
When balance must be restored to a closed circuit which has become upset,
the shiftsman begins by adjusting the pulp density at the weir overflow.
The response is rapid, and it safeguards the release mesh while he finds what
has happened and decides what to do next. When a working condition has
changed slightly the circuit as a whole is slow to respond, and may even take
hours to show that it is upset, so the cause of the trouble is not necessarily
recent. Provided the classifier is not discharging tramp oversize in cyclic
surges, there is time to think before taking further action. With experienc~,
the shiftsman is unlikely to lose control. He watches the new feed, and if it
is coarser than usual he is prepared for the extra work being thrown on the
mill. Perhaps he has noticed a slight rise of the height of ore on the classifier
rakes, showing that the circulating load is building up, and he may have to
cut down the new feed a little to avoid an overload. He may see that the
ore-a friable sulphide, say-is richer than usual, and therefore knows that
it will grind more speedily. He will be alert to take advantage of this by
raising the feed rate a little.
How does the circuit become unbalanced when all controls are in working
order? The most likely cause is a change in the physical state of the new
feed. Coarse lump, and usually also low-grade ore, takes longer to reduce to
mesh-of-grind and must be fed in more slowly. Otherwise one of two things
will happen-a circuit overload, or a discharge of tramp oversize to the
concentratcr. Finely crUShed ore, and high-grade sulphide, is more quickly
finished and can be fed in faster. Since the sulphide is a better lubricant in the
crop load than sand, the experienced operator listens to the sound of the
mill and may decrease the solid-liquid ratio a shade. At the same time he
238 Mineral Processing-Grinding Circuit Control
watches his weir overflow in case this dense sulphide is failing to rise, and
needs a slightly higher pulp density to help it. He must, however, have
special authority if he is to be allowed to vary this control.
The unbalancing of the circuit is therefore due to a change in the amount
of work required to produce a given end-condition. Since the work input is
constant, the quantity of ore must be varied in accordance with its varying
grindability. Where the scale of mining and the nature of the ore body
justifies the cost, blending of the ore ahead of grinding may be practised.
One of the most serious emergencies that can arise is the plugging of the
grates of a low-discharge mill. If they have become partially clogged with
rock or small balls, the new feed is stopped, the mill kept running until the
return sands are negligible, and the mill is then stopped while the grates are
freed. It may be desirable to raise the classifier rakes at the same time, to
prevent an accumulation of return sands in the feed launder and scoop box.
If the feed rate has been too high or the mill water has failed, the remedy is
obvious. The worst position arises when the grates have become com-
pletely choked or "plugged". The mill runs silently, as the balls are no
longer free to move, and as the scoop goes round in the overfilled launder,
return sands are being thrown out on the floor. New feed must be stopped
immediately. If the mill is discharging at the central orifice, classifier sands
are still being returned, and for the moment nothing can be done about it.
The problem is to get the grates unplugged without starting a torrent of pent-
up material which may rush over the classifier weir and through to the con-
centrator. If the layout of the circuit permits, the safest thing to do is to open
the discharge port at the bottom of the classifier tank so that its contents
run out to the floor. This "emergency exit" is not shown in Fig. 92. It is a
simple orifice, closed by a valve or plug. When the classifier has been
emptied, the water supply to the head of the mill can be gradually restored.
If there is a violent outrush of mud, the mill bearings should be inspected
before returning to normal working. When things have been put straight,
the classifier port is closed, rakes are lowered, and all spilt sand and rock is
returned to the circuit before restoring new feed. The cause of this rare
trouble may be failure of feed water, but is most likely to arise from the
shiftsman's inexperience or inattention.
TABLE 17
I Possible Genera!
Item Phenomenon Associated Nature of
Changes Control
(I) (2) (3) (4)
New Feed from Ore finer or richer Less grinding needed Increase feed rate.
Ore Bins. than usual. to produce finished Watch pool overflow
m.o.g.* Pool density density.
may be upset by
over-grinding.
Mill Discharge. Solid-liquid ratio Feed rate too low, or Check feed and feed
has dropped, and feed blocked. Return water. See if classifier
pulp is too dilute. sands too low. Feed is surging, and pool
water too high. water on.
Solid-liquid ratio Feed too fast, feed Check feed rate. Per-
too high. water low, or failed. haps stop new feed.
Ammeter may drop Increase feed water
if mill overloads, cautiously. Check
classifier may send pool density.
out oversize.
Stopped. Grates plugged. Re- Stop new feed. Run
turn sands spilling out classifier. Then
from feed launder. check feed water and
restore it if failed. Be
prepared for surge of
mud.
Mill Noises. Noisy, rattling. Not enough feed. Check feeding.
Quiet, choking. Too much feed. Too Correct feed. Cut pool
high solid-liquid water if safe, to de-
ratio. crease return.
Ball Mill Amps dropping. Mill either over- Check feed and pool
loaded and too quiet water.
or underloaded and
too noisy.
-----.-----
Amps rising. If abnormal, mech- Check bearings and
ani ca I defect lubrication.
developing.
Classifier Density rising. Mill discharge finer. Gently increase water
Weir Too little feed water. to mill and/or c1assi-
Overflow. Pool dilution insuffi- fier.
I cient.
• m.o.g.-ml'sh-oJ~grind
240 Mineral Processing-Grinding Circuit Control
TARLE 17 (continued)
These are the chief control measures, and it will be noticed that they can be
reasoned out once the shiftsman has noticed at which of the three points the
quantity or quality of the new addition-new ore. feed water. pool water-
bas changed or become unbalanced. A steady deterioration in, say. through-
put of correctly finished ore would be due to maladjustments of a less tem-
porary nature. such as incorrect ball loading or a misplaced baffle.
In a well-controlled plant. the grinding mill rarely stops for any unexpected
reason except power failure. Routine maintenance overhaul is foreseen.
planned. and arranged to fit in with the mill superintendent's schedules for
such work.
Stoppages of a grinding plant are of two kinds. planned and involuntary.
The involuntary stoppage arises either from breakdown somewhere along the
flow-line which involves the mill. from power failure. or from a mechanical
defect in the grinding circuit itself. Planned stoppages can be divided into
two types-long stoppages for major overhaul. and brief stoppages for minor
repair and check-up. Take first the sudden and unpremeditated shut-
down. If the whole power supply to the system has failed. the first step is
to see that all switches are "off" so that a sudden restoration will not strain
the motionless machinery by throwing on full power. The place where
trouble is most likely to develop in the circuit is the pool of the rake classifier.
The moment the rakes stop. sand begins to settle. burying them. If this
happens. the rakes must be freed before restarting the classifier. If power
failure has been general there can be a standing instruction that everyone
available must start to turn the classifier over by hand until its load has been
raked up and out. The alternative is that the pool's contents be run out to the
floor and the rakes raised. With spirals. and some modern rake classifiers.
the lifting gear alone need be used and the rakes can be gently lowered after
Mineral Processing-Grinding Circuit Control 241
restarting. If the mill discharge is being pumped to the next point in the
flow-line, the power failure may lead to settlement of sands in the pumping
line, tanks and pump casings, which must therefore receive appropriate
attention. If part only of the grinding circuit has broken down it is probably
necessary to shut down the rest. New feed should be stopped at once. The
water can wait, but solids must not be allowed to pile up. The sections
receiving pulp from the grinding circuit should be warned of the interruption
and kept advised of developments.
Emergencies apart, it is sometimes necessary to shut down for a brief period
in order to make some running adjustment. A liner bolt may have worked
loose, a broken classifier rake may need removal, or some other small matter
may enforce a short stoppage. Before stopping, the necessary tools and
spare parts are made ready and anybody concerned is aSfigned his job.
This is important because the repair must be finished before the classifier has
had time to rake up a big load of return sands to the return launder. A good
shiftsman should be able to shut down, complete his repair, make a rapid
inspection of pinion, crown wheel, and shell, and have the mill restarted in
less than one minute. If he needs much longer than this a proper circuit
shut-down is necessary. This, unless the classifier is of a type which can be
restarted under load, is a more detailed business, and starts with the "grinding
out" of the circulating load until mill and classifier are only lightly charged with
ore. If the circuit is closed by a cyclone, the load undergoing classification is
small and shut-down is consequently simpler. The sequence of operations
is:
(a) Warn sections affected.
(b) Cut off new feed.
(c) When return load is small enough, stop the mill (this may be 20
minutes or longer after (b».
(d) Raise the classifier rakes.
(e) Stop the classifier.
(f) Stop mill and classifier water.
(g) Stop classifier overflow pump.
If reagents are being fed into the circuit, they should be shut down with
(c). At this stage stores required should be assembled and fitters ready. A
stop of this kind is usually made on the instruction of the mill superintendent
and arranged to take place during the day shift when the Company stores are
open. The supervising engineer should make a list of points he proposes to
have checked, including the condition of rakes, bearings, drives, submerged
classifier bearings, liners, and the wearing part of the scoop feeder. When the
work is completed, he should check over his list, see that loose tools have
been removed and that all bolts are secured. Before starting up, it may be
wise to make sure nobody has taken advantage of the occasion to take a siesta
in a dangerous dark corner. The usual order for restarting is:
(a) Warn the concentrator sections.
(b) Put on the lubricators, if shut off during repair.
(c) Start the classifier.
(d) Start the mill ("inching" it round to begin).
(e) Start feed and mill headwater.
242 Mineral Processing-Grinding Circuit Control
Efficiency Control
The controls dealt with thus far are operated by the shiftsman as he steadies
the grinding circuit. There are several other running adjustments beside
that of feed and water, but these should only be made under the instruction
of the mill superintendent. They have one or more of the following aims:
(a) Reduction of tailings loss.
(b) Improvement in concentrate grade.
(c) Reduction of ball and liner wear.
(d) Saving of power.
(e) Increase in throughput of properly liberated ore.
(f) Lowering of grinding costs.
All these items are interdependent. The increase in throughput of pro-
perly liberated ore can best be achieved by reducing over-grinding to its
minimum and increasing the inevitable slight rise in oversized particles to its
safe limit. Since ball and liner wear are proportional to power used and
tonnage ground, each improvement in item (e) at the same time improves
(c), (d), and (f). The approach to higher efficiency in an operating circuit is
usually made by trial-and-error methods, and is therefore made cautiously.
Any variation which causes a serious loss of values is expensive, and should be
avoided. The golden rules in bringing up the grinding circuit to peak
efficiency should be remembered. They are:
(a) Go slowly.
(b) Only change one thing at a time.
(c) Give each change plenty of time to make its effect felt right through
the plant.
(d) Observe the effect of one change for a substantial period before
making another.
It is cheaper in the end to wait a little than to rush enthusiastically ahead.
An alteration which seems completely logical and safe may have ignored some
detail in the layout of the plant which makes it impracticable. Multiple
variations can be made in test-work in a pilot plant where a limited tonnage is
exposed to experimental hazard, but are out of place in a full-scale commercial
operation. The methods used in bringing a circuit up to higher efficiency
are dealt with more fully in Chapter 20. They enable the mill superintendent
to select the best conditions for each important factor in the combined
operation. In closed-circuit grinding these factors include:
Mineral Processing-Grinding Circuit Control 243
Place Detail
Ball Mill Liner Contours
Mill speed
Solid-liquid ratio
Ball load
Ball size ratio
Classifier Rake speed
Weir height
Baffle depth
Baffle distance from weir
Circulating load
Water addition technique
to name some of the more important. It is obviously not possible to arrive
at accurate conclusions simply by varying any of these conditions in a com-
mercial circuit, since their mutual reaction would make complete observation
impossible and false conclusions would be drawn. Efficiency control in the
working circuit is obtained by holding all the selected factors (or elements)
interacting in that circuit to the optimum values laid down as the results of
tests made under laboratory control. The methods of maintaining this
control depend on careful sampling of the pulp passing through the circuit at
key points, in conscientious accordance with a routine procedure. This is
followed by a routine laboratory testing of these samples, devised to disclose
any shift-by-shift variations.
Such sampling may include:
Place Material Information Given
Mill Head New Ore Assay value. Size analysis, moisture.
Mill Discharge Pulp Solid-liquid ratio, pH. Size analysis.
Classifier Overflow Pulp Solid-liquid ratio, pH. Size analysis, assay
value.
Classifier Returns Return Sands Ratio of circulating load to new feed.
Sizing analysis.
After studying the returns for a given period, and correlating them with the
results reported from the concentrating section for the same lot of pulp,
information is gathered which suggests useful lines of process research.
These are carried out by laboratory methods where possible, and can then be
introduced into the plant in an orderly manner.
TABLE 18
(A) THE FEED.
Operation Method/s
I. Pre-blending Computer-controlled mine valuation and stoping
2. Blending from bins Constant-weight delivery to a common blending belt
3. Blending into bins Mines' controlled delivery
4. Preliminary pH *Lime to ore bin
control
5. Sorting (a) Radio-active check and rejection
(b) Colour check
(c) Conductivity check
(d) Detection of tramp iron
6. Sizing (a) Oversize warning by electric eye
(b) "Pebble" selection for autogenous grinding, by
screening
7. Feed rate See under "Mill"
(B) MILL
8. Crushing bodies (a) Power draft change
(b) Sound-level check + pebble feeder
*(c) Check on crop load level
9. Solid-liquid ratio (a) Flowmeter for mill water controlled by discharge
density
(b) Flow rate and density check at mill discharge. con-
trolling new feed and mill water
(c) Pulp temperature change. feed to discharge
(d) Sound-level check linked to new feed
*(e) Size analysis at mill discharge
(.f) Power draft variance signalled to new feed: with D.C.
mill varying mill speed
(e) ("LASSIFI("ATIO~
tunity for rejecting those which are broken or badly shaped should be seized.
The power factor in the main motor-circuit should be held at 1·0. Tramp
oversize in the new feed of a carefully ball-rationed mill will reduce the
capacity and a little care taken to avoid its delivery into the fine-ore bin is
justified. The mechanical state of the circuit affects the power bill, so good
maintenance and efficient use of the proper lubricants are called for.
Summary
At this point preparation of the ore has been sufficiently discussed. With
the completion of preparation, the run-of-mine ore should have been subjected
to some, or all, of these operations:
Sorting of true ore from detritus.
Washing to remove barren slimes.
Crushing.
Weighing.
Head sampling.
Grinding to desired m.o.g. at required solid-liquid ratio of discharge.
Smooth delivery to the concentrating section.
These operations should have been performed in accordance with a system
which has developed the solids in the pulp to the state in which they respond
most profitably to the treatment selected for their separation into the various
products made in concentration. Within practical limits such sampling
controls should have been imposed upon the work as will give the manage-
ment sufficient information to enable the efficiency and financial value of the
operation to be assessed.
Reference
1. Ramsey, R. H. Eng. Min. J., 146.
CHAPTER II
METHODS OF SEPARATION
Introductory
Ore
(Galena + Calcite)
~
Comminution
!
Concentration
1 I
. - - - - - - - - - (separation into - - - - - - - - - - ,
distinctivetuc\s)
I
(Galena (1» (Calcite
(3»
(Galena still
locked to Calcite
(2»
Ground Ore
l 1
First separation!
Concenlnlte
(Lead and Copper)
Ta;I'lo SecO;d Se"'"'tion]
r
Sulphides
~
Separation ~ Zinc Sulphide
r
Lead Copper
Sulphide Sulphide
r
Tails to Third Separation ~
Pyrite
--
Tails to F~>urth
SeparatIon
1
Barytes Final Tailing
Fig. 106. Ore Treatment at Rammelsberg
Mineral Processing-Methods of Separation 249
Particle Characteristics
If comminution of the ore is efficiently performed, each of the multitude of
resulting particles acquires distinguishing characteristics which can be
exploited by a suitable separating (concentrating) process. The treatment
chosen in the case of the mixture of galena and calcite (Fig. 105) is applied
to the pulp to make as complete a separation between the two minerals as
their degree of liberation allows. It is not necessarily completed in a single
step, because of the presence of incompletely liberated middlings. It may
produce
(a) finished galena (I) and calcite (3),
(b) finished galena (I) and middlings (2),
(c) middlings (2) and calcite (3),
(d) rich middlings (2) and poor middlings (2)
These (d) products could be disposed of as a low-grade concentrate and a
rich tailing, re-treated. or stockpiled, if better treatment facilities or higher
realisation prices were expected in due course.
The characteristics of the particle undergoing treatment may combine to
make the work of separation easy, or they may interfere with each other,
making concentration difficult. To be usable, they must be such that at the
same time and in the same pulp the two kinds of particles (concentrate and
tailing) do not, to any important extent, respond in the same manner to the
chosen treatment.
The physical, electrical and chemical properties of minerals which are most
commonly exploited in their concentration are listed in Table 19
Middlings
Separating Appliance
t t t + +
(3) (3) (2) (3) (5)
t
Concentrate
High-
Grade
t
Average
Low-
Grade
{-
Tailing
Middlillg Middling Middling
TABLE 19
lower-grade middlings. will be crowded out with the tail and lost, and the
higher grade particles will join the true concentrate, lowering its purity.
Since the cause of this is incomplete liberation, the remedy is to give the
middling a regrinding treatment. Fig. 109 shows how this can be done.
~
New Feed Return
Middling
1 -------.t
r-
Concentrate
SeparatingfPPliance
Middling
t
Grinding Appliance
t
Pump-----~------~
it is not necessary to begin by grinding all the minerals to their fully liberated
state in order to procure clean separation. Results can frequently be achieved
in stages. First comes grinding, next separation into clean concentrate, clean
tailing and "locked" middling (as incompletely liberated particles are called).
Finally, the middling is unlocked by grinding, and retreated.
Second, the middling, or any fraction of it, can be held in a closed circuit
if by doing so the work of separation is made more efficient.
Third, a true middling always needs special treatment not provided for in
the appliance which has sent it out as a middling. After this special
treatment it may be unsuitable for return to the sorting appliance. In the
case just considered, if treatment depended on the mass of the particle, the
ground fragments would probably be sent to a machine specially adapted to
deal with their smaller size.
Types of Middling
0 (a) (b)
0 D
(e) (d)
~ fe)
0 (f) (9)
PyrIte
Quartl
Gold
spond to electro-magnetic pull. Particles (b), (e), (f), and (g) might act as
gravity middlings. When the values are almost or quite masked by gangue,
they will be completely lost in flotation or chemical attack. Particle (g)
can be a special case. Suppose the gangue to be quartz, the outer enclosed
mineral pyrite and the centrally enclosed mineral gold (the last being the value
sought), then the grinding treatment must expose the gold. Regrinding
must therefore be far more elaborate than in usual cases, particularly if the
gold is segregated in the pyrite in specks only a few microns in size. Here
roasting treatment is usually preferred to further comminution.
Staged Concentration
Feed
t
Coarse Grind
Mineral liberated t
at coarse m.o.g. ~o(----Separation -----~>~Tailing ~
{'
Fine Grind
Mineral liberated
at fine m.o.g. ..
t
Separation ----~> Tailing
Coarsely mineralised ores suitable for such treatment are now rare. Further,
most minerals can be cheaply concentrated at a fine m.o.g. by flotation, so
the need to avoid fine grinding which dictated gravity-concentration methods
before 1920 no longer exists. If flotation is used, as it is today in most
mineral-dressing plants, instead of removing finished concentrates by stages
it may be feasible to discard liberated gangue at each grinding stage, as in Fig.
112.
This principle may be used to give selective treatment to the middlings
product at each grinding stage, as in Fig. 113.
This arrangement is a variation of the "cascade" principle, which is used
in chemical engineering to enrich or deplete an original feed by gentle stages
(Fig. 114).
A number of identical concentrating devices are arranged in series, new
feed being introduced midway. Each unit divides its feed into a slightly
254 Mineral Processing-Methods of Separation
Feed
~
Coarse Grind
1
Coarse Gangue ....;..~--- Separation ---_'" Upgraded Feed
t
Fine Grind
~
Tailing -oE"',---- Sepamti0l
Concentrate
Head Feed
~
Coarse Grind
Rejected t.
Tailing ~.----- Se~bonl
( +)
MediuJ Grind
l
Separation -----~~ Tailing ----'
l
Fine Grind
~
Tailing -oEEi---- Separation - - _...~F~i~m~·s~he~d~C;o~n~c;en~t~r~at~e
t
(+) = Upgraded Feed.
C T
Richer Poorer
C= Concentrate: T = Tailings
Fig. 114), while an impoverished tailing flows concurrent to the next unit
below (to the right in the figure). Thus concentrate C from unit 3 joins the
tailing from unit I in unit 2. The result is progressive enrichment in one
direction and corresponding impoverishment in the other. In mineral
dressing the usual arrangement is more compact (Fig. lIS), the separating
machines being marshalled into blocks, called roughers, cleaners, and
scavengers.
New Feed
~~~~~~~-_~----,Rou~her~r--II-~~~~~~~~~->:T-a-i-li-n-g---,I
r---c-o-n-c-e-n-tr-a-te-....
Tailing 1 Con"n!,"!,
~ Final
Concentrate
Final
Tailing
Panning
clean prospecting pan should be used. Scouring with sand usually suffices
to remove old rust. The pan should not contain oil or grease. If a gold
ore is to be tested, the pan should be dark in colour so as to show up the golden
specks clearly. Since a great deal of panning is done in pools and streams,
the beginner should learn to squat on his heels and pan from one gold pan
into another full of water, so that he can save and retreat the discarded
material until he has become expert.
An excellent practice material is a -20-mesh mixture of sand and galena.
A few hundred grams of this material are wetted down into water and worked
into a running pulp. This is next deslimed by gentle decantation, the pan
being held with its double riffie away from the operator. At no time should
pulp be allowed to stream over the riffie. It should be floated out into the
pool of water with a gentle swirling or rocking motion.
From time to time the pan should be tapped with the heel of the hand to
aid the heavy particles to burrow down to the bottom of the fluid body of
pulp. The top strata should then be panned off, using a jigging or swirling
motion. As soon as heavy particles show, the material still in the pan
should be repulped and rethumped. Successive barren strata can thus be
removed without loss of values. When the point is reached where most of
the sand has been rejected, the decision must be made as to whether a low-
grade concemrate and a clean tailing is to be produced, or a high-grade
concentrate and a middling. In the latter case, rejected sands which now
carry heavy mineral should be panned out into another holding vessel. It is
not possible to make a clean concentrate and a clean tailing in one operation.
The rejected tailings should be repanned to see how efficiently the work was
carried out. With practice, it is possible to use panning as a rough guide in
assessing efficiency of gravity treatment of sands.
The plaque, a white-enamelled concave disc II" in diameter, is used in a
similar manner in the examination of fine sands.
Gravity Separation
In gravity separation the combined effect of mass and shape of the particle
determines its movement relative to flowing water. In one development of
this effect the water flows vertically either continuously (classification) or in
oscillating motion (jigging). In a second method the ore is fed into a fairly
quiet pool of dense media (water mixed with slow-settling heavy minerals to
form a fluid which can be maintained at a high specific gravity). Ore enfering
this dense medium either floats or sinks. A third type of gravity treatment uses
flowing streams to effect separation. Here the pulp is carried horizontally
or down a slope, and separation depends on the rate of fall, and resistance
to displacement after the particle reaches the floor of the appliance.
Methods which exploit differences of gravity require that there shall be a
marked difference between the specific gravities of the value and the gangue.
The material must be sufficiently coarse to move in accordance with Newton's
law. Particles' so small as to settle in accordance with Stokes' law are
Mineral Processing-Methods of Separation 257
Chemical Methods
These depend on adequate exposure of the valuable ore mineral to the
action of a solvent. Solvation attacks the surface displayed. The new
surface developed as the result of comminution is doubled for each halving
of the mesh size of the mineral. Hence the time taken for dissolution of a
particle bears an important relation to its comminution. The steps taken in
separation by chemical methods are:
(a) Grinding for optimum exposure.
(b) Solvation of value.
(c) Separation of solute from residual solids.
(d) Re-precipitation of the value.
Step (d) may produce the value in a new compounded form, or as metal.
Flotation
Flotation is the term used generally to denote "froth-flotation" rather than
processes in which gravity provides the main separating force. In this book
the prefix "froth" is not employed. This extremely important process is
used in most of the mills at work today, and frequently provides the main
separating method. It must not be confused with "sink-and-float" processes
which are termed by the author dense-media separation (DMS).
Flotation is usually applied to minerals (sometimes values, sometimes
gangue) that cannot be liberated at a size suitable for treatment by all-gravity
methods. It is chiefly practised on feeds ranging between 60 mesh and a few
microns in particle size. Under the influence of appropriate chemicals, the
surfaces of some mineral particles can be physically and chemically modified.
Thus treated, some particles in the pulp remain attached to water
(hydrophilic) while others become air-avid or aerophilic. The latter cling to
air bubbled through the pulp. Loaded bubbles rise to the surface where
they form an unstable froth which overflows, carrying off the mineral thus
concentrated.
Amalgamation
At the turn of the century amalgamation was the main method used for
the separation of gold and silver (in their "free" or "clean metallic form")
from their ores. It is now subsidiary to cyanidation, with rare exceptions,
and is sometimes omitted altogether. Amalgamation exploits the fact that
mercury readily "wets" the surface of clean gold, silver, and electrum-metal.
Once particles of these metals have been absorbed into the mercury, the
resulting amalgam can easily be trapped and removed from the ore pUlp. The
mercury is then removed, and the residual gold is refined.
Exploitable Factors
these) applied forces are called physical, chemical or electrical one or more of
them is the disturbing factor. The general descriptive term "physical chemis-
try" embraces the study of these forces and their scientific use in such fields as
thermo-dynamics, electro-chemistry and electro-magnetism. In the follow-
ing chapters the role of physical chemistry is considered in many aspects.
The forces at work are sometimes mutually antagonistic and able to reduce
efficiency or even prevent the desired reaction unless they are recognised and
correctly dealt with. Basic and applied research, the latter specific to the
ore which is to be treated, should therefore isolate and study such interfering
factors. It should ascertain the strength and nature of both the forces relied
on for reaction, and of those inimical thereto. Only thus can successful con-
centration (separation of the previously balanced pulp into two differentially
balanced streams) become a science rather than an art or empiric skill.
When these matters have been thoroughly looked into, the flow-sheet is
worked out. Larger-scale tests in the pilot plant usually follow. The full-
scale plant is then designed and built and worked up to its operating norms
by the permanent staff. Since their work, like that of all who must deal
with an empirical element in the daily variations encountered, is "the art of the
possible" it is invaluable to the mineral engineers concerned to understand
something of the forces at work, so that process controls can be made to
strengthen their effective use.
Concentration Formulae
The basic two-product formula used in mill control was given in Chapter 2
under "Acceptance Operations". To this weight balance (F = C + T) and
ingredient balance (Fi = Ce + Tt) for weights FCT and assays iet of feed,
concentrate, and tailing, may be added the following:
(c-t)
F = C -- (lI.1)
(f-t)
(e-t)
F= T - - (11.2)
(c-f)
(f-t)
C=F-- (11.3)
(e-t)
(f-t)
C= T - - (11.4)
(e-f)
(e-f)
T=F-- (l1.5)
(c-t)
(e-f)
T=C-- (11.6)
(f-t)
260 Mineral Processing-Methods of Separation
T!l~se formulae, derived from the weight and ingredient balance, permit a
solution where four of the six quantities, including two assays, are known.
Where the mill operation is set to produce a concentrate rich in mineral a
reporting in CI , followed by one rich in mineral b reporting in C2, three-
product formulae can be used. That for weight balance is
(11.9)
F
Ka (11.12)
(11.13)
and recovery
(11.15)
Mineral Processing-Methods of Separation 261
E=~RnN
(11.16)
n
RaJb
(11.18)
(100-(Ra) (100-Jb)
(c-f)
Ec --.R (11.20)
(cmax-f)
262 Mineral Processing-Methods of Separation
General
In addition to the foregoing main methods of separating ores into value
(or concentrate) and gangue (or tailing), there are several specialised tech-
niques which lie beyond the general purpose of this book. The processes
used in ore treatment may combine more than one method in a given flow-
sheet. In gold milling, amalgamatIOn can be followed by flotation, cyan ida-
tion, or both. The lead of a lead-zinc ore can be concentrated on jigs and
tables, and the zinc by flotation. The method, or combination, chosen
depends on the break point of the values, the ore characteristics which can
most satisfactorily be exploited, and the cost of treatment. Such considera-
tions as early elimination of part of the gangue, simplicity of operation,
specific reaction of the sought mineral and a brisk rate of throughput are
factors which influence the choice of method.
References
\. Taggart, A. F. (1945). Handbook of Mineral Dressing, Chapman & Hall.
2. Diamond, R. W. (1928). Trans. A.I.M.M.E., 79.
3. Gaudin, A. M. (1939). Principles of Mineral Dressing, McGraw-Hili.
4. Douglas, E. (1961/2). Trans. I.M.M. (Lond.), 7\'
5. Stevens, J. R., and Collins, D. N. (1963). Trans. A.I.M.E., 226.
CHAPTER 12
Feed
!
Minerals -oE<;------ Fluid Bath ----_~ Minerals
S.G.-2·9 (S.G.2·9) S.G. + 2·9
(floats) (sinks)
Heavy liquids have long been employed for laboratory separations of this
kind, but have so far only limited industrial use. Several decades back the
possibilities of using a mixture of finely divided magnetite and water were
explored. The fluid made from this mixture is cheap and settles slowly
enough to act like a true heavy liquid. Commercial development failed at
that time for lack of an effective method of keeping it in clean working
condition. Bessemer, in 1858, patented the use of solutions of metal chlo-
rides, and in 1911 Ou Pont developed the use of chlorinated hydrocarbons to
obtain high densities. The first important success was achieved by the
• Following usage. the term "dense media" is used in this book, though in some contexts the plural
is less appropriate than the singular word.
264 Mineral Processing-Dense Media Separation
Chance process, patented in 1917. This is not strictly a simple DMS process,
since it employs a hydraulically dilated bed of sand. In 1928 Lessing re-
developed the use of calcium-chloride solution. In 1931 clay, gypsum, and
pyrite slurries were used in water, and in 1932 G.J. de Vooys established a
coal-cleaning process based on use of a barytes-clay medium. Mixtures of
such earths with water produce fairly stable fluids with specific gravities
ranging between 1·4 and 1·6. When run-of-mine material is fed into a sepa-
rating bath maintained at some such density, coal floats and the accompanying
shale sinks. The density range is, however, insufficient for treating the great
majority of ores which require a "parting density" of 2·7 or more. In 1935
the possibilities of a suspension of finely ground galena in water were investi-
gated and matured by Huntington, Heberlein and Co., in a pilot plant which
in due course led to a successful application at the Halkyn lead mine in
Wales. Meantime, DMS was being pioneered by the American Zinc, Lead
and Smelting Co. at Mascot, Tennessee, which started work with a galena-
medium in 1939. Later, magnetite and ferro-silicon took the place of
galena in the Mascot plant. The change was made after methods had been
developed for using the ferro-magnetic qualities of these substances to keep the
dense media fluid clean.
The DMS principle is extremely simple. The high specific gravity of the
bath may be produced by the use of heavy liquids, solutions of salts, or sus-
pensions of slow-settling solids in water. The bath may be almost static,
gently agitated, or in moderately brisk motion.
Two main applications are possible. DMS can remove lightweight rock
at an early stage in crushing. It is then a sorting treatment, used to remove
barren waste or ore of a grade too poor to warrant treatment in the con-
centrator. The second application of DMS produces a commercially graded
end-product. It is chiefly used to produce clean coal as a "float", and a low-
grade "sink" consisting of shale or of coal too high in ash for sale.
The various methods used in cleaning coal embody specialised applications
of mineral-dressing techniques, having the common factor of low specific
gravity in the mineral recovered as a "float".
Amenability
In mineral dressing the term "amenability" refers to the manner in which a
given ore responds to a given method of treatment. Broadly, DMS is
applicable to any ore in which, after a suitable degree of liberation by crush-
ing, there is enough difference in specific gravity between pieces of tock to
separate those which will repay the cost of further treatment from those
which will not. If this liberation is achieved at coarse gravel size or larger, a
difference in specific gravity of 0·1 or even less can usually be exploited
commercially. In the laboratory, where longer time can be given to small
quantities of material to move, much smaller differences in specific gravity are
sufficient to ensure separation. Parting densities above 3·4 are difficult to
maintain, and limit the applicability of the process.
If the values are irregularly distributed in the ore body, either as coarse
Mineral Processing-Dense Media Separation 265
crystals or aggregates, OMS may be possible. If the values are finely dis-
seminated through the ore, an adequate difference of density between the
crushed particles cannot be developed by coarse crushing. For the process
to be applicable the heavier fraction must be sufficiently liberated from the
lighter fraction at a mesh coarser than 48 mesh (preferably coarser than 5
mesh). Ore can be separated from waste if there is a sufficient difference in
specific gravity between the two minerals. This can be 0·1 or even less for
separation at 5 mesh or over, and at still finer grain sizes if the viscosity of the
media is low. If the specific gra vity of richer lumps of ore is higher than that
of poorer ones it is frequently possible to upgrade run-of-mine ore which
otherwise would not repay the cost of treatment, by using OMS to reject the
low-grade material.
TABLE 20
Value!
ton in Total Treatment
Tons Contents Value Cost Profit Loss
1J
Float 10 Is. lOs.
Sink 90 4s. ltd. 44Os. 450s. lOs.
Case 1
Feed 100 4s.6d. 450s.
Float 20 Is. 3d. 25s.
Sink 80 5s.4d. 425s. 400s. 25s.
} Case 2
Feed 100 4s.6d. 450s.
Float 30 2s.8d. 80s.
Sink 70 5s.3-!d. 370s. 350s. 20s.
} Case 3
Feed 100 4s.6d. 450s.
If this ore is treated as mined there is an inevitable loss of 50s. This can be
mitigated or even turned into a profit. Case 1 shows that by removing ten
tons of almost barren rock, lOs. worth of valuable concentrate is lost which
could only have been recovered by milling those ten tons at a cost of 5s. per
ton, or 50s. Thus the effect of this amount of pre-treatment is to reduce
the total loss. In Case 2 a more severe sorting has been applied, with the
result that more of the valuable mineral has been lost. The overall effect is
that a profit of 25s. is made on treating the tonnage. Case 3 shows the result
if too severe a OMS sorting is applied-the net profit is reduced to 20s. As
the result of tests it might be found that for maximum profit a specific gravity
266 Mineral Processing-Dense Media Separation
should be chosen at which more than 20 % and less than 30 % of the feed
would be discarded as the floating fraction.
This illustrates an important principle in mineral processing. Normally,
the plallt must be worked to produce maximum profit, not maximum recovery.
Note the qualifying word "normally". Under the stress of war the profit
criterion may be ruled out. Again, it may be the case that the mill is short of
ore, but costs nearly as much to run at 90 % capacity as at full load. The
foregoing type of calculation might not then apply.
In suitable cases, dense-media separation can sort a wide size-range of
ore cheaply. The tonnage of suitable feed must be sufficient to justify the
cost of installing a plant. Easily movable plants can be rented. They are
used to treat old dumps, upgrading their contents to a value which justifies
the cost of more expensive processing. They can also be used to make a
comprehensive field test on a few thousand tons of development ore while the
question of a permanent installation is being considered. The exploitation of
low-grade properties which otherwise were too poor to be mined has re-
peatedly been made possible by the use of this cheap up-grading method.
Where the ore mineral occurs in a lode too narrow to mine selectively, the
DMS plant can be placed underground, so as to sort out barren wall or
"country" rock for return as filling material, thus saving transport cost.
Forces at Work
gravitational push or pull with respect to the heavy liquid to start moving
upward or downward, has become a middling. In a media of given viscous
resistance there is a minimum size limit for a sphere of given S.G., below
which it lacks sufficient potential energy to move up or down. This may be
true viscosity (molecular shear) or pseudo-viscosity. DMS on the commercial
scale rarely uses true heavy liquids, but suspensions of appropriate heavy
minerals in water, sufficiently finely ground to have a slow rate of sedimenta-
tion. Hence the retarding force in such a bath is largely a function of the
specific surface involved. The main contributor to this specific surface is the
heavy mineral used in the media, but slimes, clays, and detritus from the
ore also exert an influence. (Physico-chemical influences associated with the
pH of the media will be more apparent when Chapter 17 has been studied.
They affect the dispersive and coagulative tendencies of the solids in the bath
as a whole). A further important factor arising from the effect of specific
surface is the size and shape of the particles used to constitute the dense
media. Spheres were considered above, in order to simplify the essential
picture, but a working bath is filled with angular mineral particles under
roughly controlled turbulent conditions, with a slow rise in media density
from top to bottom due to the tendency of the particles to settle. Despite
the complex of working forces it is helpful to start this study with its essen-
tials, the "weighing" of the ore particles in accordance with the amount of
media each of them displaces, and the generation of force in each ore particle
according to the amount it is out of balance with the media.
It now becomes clear that the particle size and shape of the mineral used as
dense media, together with such chemical effects as may originate with pH
of the water in which it is slurried, control its viscous resistance. This is
modified by the presence of contaminants and by temperature effects. These
therefore are control factors, since they help to determine the minimum size
of particle which can move sufficiently fast through a bath at a given parting
density. Leaving the hypothetical case of spheres, it is now clear that for
ordinary fragmented particle of ore this minimum size is also influenced by
shape. The main effect of shape is to develop a rubbing surface over which
the braking laminal flow of the media acts. A secondary effect is that, since
a particle cleaves its way relative to the media, the disturbance it generates is
proportional to the cross-section it presents.
Thus far consideration has been confined to a few pieces of ore or waste
rock. Continuing with the theoretical concept of a quiet bath (and ignoring
the disturbance due to removal of separated products from a plunging feed),
consider next a continuously arriving stream of dry ore which has been sized
by screening. Assume that of a hundred such particles forty-five will sink
and be withdrawn, forty-five float and overflow, and ten will form an equi-
poised teeter. At first separation is clean, but deteriorates as the teetering
population grows with the further arrival of middlings. The sinking fraction
is obstructed and dissipates part of its gravitational force in collision and
jostling through the crowded middlings and then tends to be retained. The
problem has some analogy to that considered in Chapter 9 under Mechanical
Classifiers. Left uncontrolled, the bath would soon choke with near-
middlings and separation would cease. The problem is handled by inbuilt
268 Mineral Processing-Dense Media Separation
The Process
jDoMo 1 Bath (b l
l
Floats to r Sinks to
Drainage Drained Drainage
Screen (c) Media Screen (c)
(+) (_) 1 1 (_) (+)
Sprays 1 lrsP'"Y'
Washing Screen ( - ) ...-.---Washing Screen
I
(+)
1
Dirty and
I
(+ )
L diluted media J
Washed Product Washed Product
1
Cleaning and
Reconstitution
(d)
Fig. 117. Stages in Dense Media Separation
These steps are merged into a continuous process on the lines shown in Fig.
117.
In operation, new feed and reconstituted media are fed continuously into
the bath. Floats and sinks are withdrawn, and drained of adherent or "drag-
out" media which may either be pumped back into the bath direct, or given a
cleaning treatment before return. The sinks and floats are next washed by
sprays which remove almost all the residual media, after which they are
Mineral Processing-Dense Media Separation 269
sent to the next stage of treatment. The medium removed from the finished
products by washing is collected, cleaned, adjusted to the working density,
and returned to the bath.
The feed must not contain cOlloidal material, primary slimes, or fine ore.
These are removed, as far as possible, before bath treatment. Colloidal
slimes and "fines" reduce the efficiency of treatment because they increase
the specific surface beyond the amount provided by the dense media. This
random variation in the pseudo-viscosity of the bath is undesirable if it in-
creases unduly. A controlled amount of such fine material may be needed in
order to aid stability of the dense media by decreasing its rate of settlement.
When the ore has been sufficiently washed it must be drained before entry to
the bath, to avoid undue dilution of the medium. Provided the feed rate is
fairly steady and its moisture content does not vary, the effect of such added
water can be compensated by adjusting the density of the re-circulating
media.
Normally dense media separation is applied to a feed ranging from about
3" down to + 10 mesh. For larger or smaller sizes special equipment is
needed.
A dense media composed of magnetic particles suspended in water is
easy to clean, and can tolerate a limited amount of slime from the ore without
serious deterioration of the separating action. New feed should, however,
be free from slime and undersize in order to aid the selectivity of the separating
fluid. This new feed should not include material liable to break down and
.thus produce contaminating slime and fine sand during treatment.
material treated. Ore structure affects this, granular and compact particles
being best for treatment. Those with voids raise the drag-out loss. Micaceous,
fibrous or acicular small particles are unsuited to treatment.
Dense media. One of the earliest modern processes to use a quiet bath for
upgrading coal was the Conklin (1922). The solid employed was magnetite
ground to -200 m.
Although magnetite is today widely used in dense media baths, it failed to
establish itself at that time, since a simple method of keeping it clean had not
been found.
In the 1930's the Barvoys process was developed in Holland. It continues
to enjoy a substantial success. The separating fluid is a mixture of clay
(S.G. 2'3) and finely ground barytes (S.G. 4'2), mixed in a ratio of 2:1 and
diluted with water to any desired density up to 1 :8. The barytes, at -200
mesh, together with the clay, forms an almost stable pulp in which great
accuracy of "cut" between sink and float is practicable. The viscosity of the
medium is an important operating factor. If too high, it impedes the move-
ment of particles in the separating bath. Froth-flotation is used to regenerate
fouled dense-media, by removing fine coal.
In another Dutch process, Tromp uses a less stable fluid, prepared from
finely ground magnetite or specially treated pyrite. The settling characteris-
tics of the medium cause the bath density to be somewhat lower at the top
than at the bottom. This aids in separation.
Another substance used in the Netherlands is loess, which differs from clay
in having a lower viscosity at the same pulp density. The processes using this
medium are discussed under "Coal" (Chapter 23).
In the treatment of ores the lightest mineral is usually quartz, at a density
exceeding 2·63. The substances used to form the dense media must be
far denser than for coal treatment, as the working density of the bath is twice
as high. Any substance used is chosen for the following main qualities:
Hardness. It. must not easily break or wear down into a slime under
working conditions.
Chemical Stability. It must not be chemically corrosive, or liable to react
with the ore minerals undergoing treatment.
Slow Settlement at Tolerable Viscosity. It must form a fairly stable
pulp without having to be ground very fine, or the medium will be too
viscous.
Specific Grm'ity. This must be high enough to give the required bath
density under reasonably non-viscous conditions.
Regeneration. In working, the dense media becomes foul, and must be
easy to clean before further use.
NOli-fouling. A certain amount of media lodges in cracks in the lumps of
cleaned and washed ore. It must not be of such a composition as to upset the
subsequent treatment of the ore.
Among the dense media best able to meet these requirements are those
listed in Table 22.
Galena was used in the dense-media process at Halkyn J following its
development in a pilot plant. It was also extensively used in North America
in lead-zinc plants. The density of PbS is 7-4-7'6. When pure it can be
Mineral Processing-Dense Media Separation 273
used to maintain a bath density of 4·3, but the practical working limit is of the
order of 3·3. Above this, movement of the ore is slowed down by the viscous
resistance to the medium. This leads to difficulties in dealing with small
particles of feed, and of middlings having a specific gravity close to that of the
bath. Galena is a soft mineral, and is prone to slime readily. Removal and
recovery of this slime by froth-flotation is somewhat difficult. The material
itself is expensive. The tendency of recent years is toward an increased use
of magnetic materials rather than galena. These can be more easily regener-
ated and are lower in primary cost.
TABLE 22
Method of
Material Regeneration
Galena Froth-Flotation
Magnetite }
Mill Scale Magnetic Separation
Ferro-Silicon
Pyrites . Used autogenously
Copper Pyrites ,
Hematite f Obsolescent
Magnetite (S.G. 5·0 < 5·2) is used where bath densities below 2·5 are
suitable. This is rather low for ores, but magnetite is increasingly used in
cleaning coal. It is regenerated by magnetic treatment.
Ferro-silicon is now the most widely used substance. With a silicon con-
tent of 10% its S.G. is 7·0, and at 25% Si it is 6·3. If the silicon content
exceeds 22 % it is only feebly magnetic, and if below 15 % the compound is
prone to rust. A 15 % ferro-silicon can be used to produce dense media with
specific gravities up to 3·5, but the more usual working range is 2·5-3·2.
Where the separation is between 2·65 and 2·9, the ferro-silicon is usually ground
to -100 mesh and blended with between 10% and 20% of Fe30 4 . At higher
working densities (2·8 < 3·0), -JOO-mesh ferro-silicon is used alone, or a
mixture of -65-mesh ferro-silicon with the Fe 30 4 • For bath densities above
3·0, -65-mesh FeSi is employed. FeSi containing 15 % Si is non-rusting
and has good magnetic qualities. A little lime is sometimes added to the
slurry, particularly when carbonate ores are treated. A dispersing agent,
such as tri-sodium phosphate is occasionally employed to reduce any tendency
for slimes in the bath to flocculate. Ferro-silicon is either prepared by grind-
ing lump material, usually in a wet ball-mill, or by air-blasting the molten
compound as it is tapped out from the furnace. In the latter case, the result-
ing particles are roughly spherical, and it is claimed that higher bath densities
can be used because the specific surface (and hence the pseudo-viscosity) of
these spheroids is less. A further advantage claimed for atomised FeSi is
that there is reduced loss through drag-out, the spheroids being less liable
to cling to the ore products. 5 One contributing factor to loss through such
adhesion is roughness of the ore surface. 6
It is further claimed that rusting tends to commence at sharp corners and the
spherical ferro-silicon is chemically more stable than that produced by
274 Mineral Processing- Dense Media Separation
crushing and grinding. Provided due care has been taken in manufacture to
avoid trapping of air into the molten globules and thus reducing their specific
gravity, this claim is interesting.
Mill scale is used as an alternative to magnetite. Flue dust from steel
smelters is used in Australia. Research has been active in seeking suitable
heavy powders to increase the commercial working limit of bath density.
Such metals as tungsten have had limited laboratory use in bath make-up.
Monitoring and automatic regulation of media density and viscosity are
practised in some installations.
In the Stripa process 7 an autogenous OM is segregated from the mill feed.
It is a fine magnetite-hematite sand, and its S.G. is regulated by rising hy-
draulic water which keeps it in teeter. In the Chance process, used in coal
preparation, a similarly dilated bed of sized sand is used.
The first Huntington-Heberlein sink and float plant came into use for ore
treatment in 1937, at the Halkyn Mine in North Wales. In its early stage the
method used was similar to that of the Barvoys coal cleaning process, save
with respect to type of medium and method of regeneration. As now worked,
the media is controlled for rate offlow through the open topped pyramidal bath
partly by use of several "steady head" flow tanks and partly by use of an
internal "steady head" outflow from the elevator casing, through which the
settled fraction of the ore is removed. At the top of the tank (Fig. 118)
rotating paddles (2) submerge the entering feed and impel the floating fraction
to the discharge. Rate of paddle rotation is controlled, a dwelling time
for float ore of between 15 and 25 seconds being normal. Sinking ore is
accelerated by flowing media through the pyramidal part of the tank to the
slowly moving elevator (3) where it is raised by perforated buckets which
allow non-adhering media to drain back down the casing (4) to the outflow
point (5). Further drainage is performed on the shaking screens (6) as they
convey floated and sunk products separately out of the system. This drained
media returns to the system. Wash-sprays remove the last of the adherent
media, which is dewatered. reconstituted and returned. An essential
feature of the process is its quiet bath, no mechanical agitation being used to
stIr Its contents. Media can be cleaned between overflow and return (for
example by froth-floating galena away from the entrained waste).
I I
I I
I
I
L,.-._-.J I
t
I
-_.I'L. . .,I\."'.,_
- -, .. _"-u,"'". oio- ..• .....
t
t
t .
I.
-e:I
I 0 -_.
-- .....
- - - --1tJI
... I
-- -=1
, -. -.--J~ --
Fig. 119. DMS with Magnetic Reclamation
a slight dropping of the spiral introduces more D.M. solid into circulation.
The operator thus has excellent control of separating density. The thickener
can be used as a storage tank when shutting down, as dense medium must not
be allowed to settle solidly in the cone, where it would be difficult to restore
it to a fluid condition. It is usual to provide a run-off sump to which all
circulating dense medium can be drained in an emergency, and from which it
can be returned to circuit either by pump or bucket elevator. In recently
constructed plants no run-off sump is provided. Instead, the bottom floor is of
concrete, and is surrounded by a curtain wall. The floor slopes gently to a
small gathering sump. This arrangement serves to catch spilt medium and
provides temporary storage if the contents of the cone must be dumped.
In another arrangement the contents of the cone can be piped to the thickener.
Medium of any required operating density can be introduced in any desired
volume at any depth or set of depths down the vertical axis of the cone.
Part escapes freely at the overflow launder and part with the underflow. The
Mineral Processing-Dense Media Separation 277
distribution of the outflow depends on the force with which the air-lift
{Fig. 119 (3» is worked, this air-lift being, in effect, a controlling pump which
varies the intensity of downdraw of the medium. A stirring mechanism
works inside the cone, but its action is indiscriminate so far as this hydraulic
effect is concerned. Since in operation a light ore fraction must overflow,
while a heavy one drops to the underflow, the inflow must always exceed the
underflow.
Recirculating media enters the cone gently at the required depth, and
then flows either toward the upper discharge, which is situated on the peri-
phery, or the bottom (apical) discharge. If the air-lift is run gently, the up-
ward current will be relatively strong, becoming gentler as the horizontal cross-
section of the cone widens. In its standard form the cone finishes above with
a cylindrical section. Some of the older designs had "closed" cones, with a
frustrum inverted upon the lower cone to give acceleration to the rising fluid
before discharge, but this is no longer favoured. If the entering media is
released near the top of the cone, much of it streams across to the overflow
after being distributed by the revolving paddle. This gives the upper part of
the cone a strong streaming force, while toward the apex there is an increasingly
heavy downward pull, controlled by the strength with which the air-lift is
being caused to withdraw fluid.
If the media is fed low down, streaming force gives place to lifting force.
Two simple controls are therefore available for adjusting the vertical and
horizontal components of the fluid motion to suit a given type of ore. Their
effect is intensified by the slight instability of the media, which by giving a
gentle density rise from top to bottom of the cone makes it more difficult for a
particle to fall the further it descends. Here two forces are acting against one
another. If the cone is imagined to be quiet, so that the only operative force
is fluid density, then a number of particles lying very near to the average
density would arrange themselves horizon by horizon with the lighter above
and the heavier below. If, however, the added effect of the continuous
withdrawal of medium at overflow and underflow be now considered, each
particle, at its particular horizon, is being moved by the frictional drag of
fluid running past its surface. The lighter material will be at the top and tends
to receive streaming friction, which moves it horizontally to the discharge of
the floated fraction; the heaviest will be well down in the drag toward the
underflow, where it is accelerated throught he continuously reducing section to
the apex. Other parti-;les will fail to generate a sufficiently strong character-
istic alignment and will build into a teeter bed, according to their mass and
surface.
In cone operation this teeter bed can be used as a controlling factor.
It has some points of resemblance to the teeter zone in a hydraulic classifier.
It "scrubs" the material passing through it, and its added resistance affects the
sorting action in the cone. The operator can see the level of fluid in the
sump from which medium is being pumped back. Since this sump is the
only "surge" or holding tank, and the total amount of fluid in the system
is constant, any increase of teetering ore in the cone displaces an equivalent
volume of media and causes a rise in the sump level. This can be used as a
teeter control by a skilled operator.
278 Mineral Processing-Dense Media Separation
into the plant, some clay or slimed ore should be added, so as to produce a
media having a consistency similar to that which characterises normal
operation. Clean ferro-silicon in a new bath usually settles rather too
quickly without this addition of slimes.
An interesting development of dense-media service is the Mobil-Mill
(Fig. 120). This is a compact cone plant, with single-stage magnetic cleaning,
fed direct from the washing hopper. Flocculation of the media takes place
in the magnetic separator itself. Sufficient permanent magnetism remains
with the material pulled out on the belt to ensure its settling during progress
through the densifier. The Mobil-Mill takes only 36 man-days to erect
or dismantle and stands on an area of 13' 10" wide by 29' 3" long and is
16' 2" high. It is sectionalised, on skids, and has a duty of between 5 and 15
tons per hour; it may be leased or bought in some countries. This ease of
installation opens interesting possibilities to the young concern, or the one
which is considering the possibilities of sink-float treatment but wants a run
on a pilot-plant scale before finally committing itself.
,.
Jj
"
.. •• _.:!_O
.
.-
To Co •• 5.".,,0''''
has the advantage in use that is can be stopped under full load and readily
re-started, since the Akins spiral is available for working settled medium up
to a suspended state. Where the DMS plant is required to work only during
hours of delivery of ore this can be an important aid to the operator.
A type of machine well suited to the gentle handling of large material
such as coal is the Link-Belt Co.'s Float-Sink Concentrator (Fig. 122). The
separating drum principle is also incorporated in such appliances as the Wemco
(Fig. 123) and the Hardinge (Fig. 124). Wemco machines are also made
with two separating drums (Fig: 125), operating at slightly different densities
so that a true middling can be continuously removed.
In coal preparation, which pioneered modern DMS, the feed is much
Mineral Processing- Dense Media Separation 281
lighter and the economic product is floated away from a minor sinking frac-
tion of shale and "dirt" . Operating requirements therefore differ sharply
from those in wh ich the same principles are applied to the treatment of
heavy minerals. Among the substances which can be used in the bath are
clay, barytes, loess, magnetite. blast furnace soot and ferro-silicon . The
moved to one end by the top part of the separator conveyor while sinks are
carried up to be discharged at the opposite end.
Both the Tromp and the Barvoys separators have provision for three-product
working. One modern form of this system is the Simcar (Fig. 128). Raw
coal enters at A and first-stage separation (clean coal from the dirt plus
middlings) is made in the low-density zone B at the top of the bath. The
float is moved by the scraper conveyor C to drainage screen D while the rest
of the feed sinks to the boot of the elevator F and out via G. Middlings
are checked between low- and high-density zones and carried up the com part-
ment H by gently rising medium, the flow rate of which is controlled from the
overflow box L. They are then moved by the middlings scraper conveyor K
to the middlings chute J, and delivered to their drainage screen. This
machine treats up to 200 tons per hour of ~ 8" + f' feed.
In the Barvoys process the density of the cIay-barytes suspension only
varies slightly from top to bottom of the bath. Middlings are lifted out by a
propeller-induced upward current. The Tromp process uses a horizontal
flow to yield a top float, a middling, and a sink discharge of dirt. In the
Ridley-Scholes process float and middling are withdrawn at points along a
V-shaped bath and sinks are removed on an endless belt.
A comparatively new type of coal washer, the Norwalt lO has been designed
(OINO • COAl " "'DOliN() 0 _(lUll lOW GlAvjn M(OIUM HIOH GlAY", MIOIVM
to unite the best features of both "deep" and "shallow" baths. The bath
is a cylindrical drum in the centre of which is the frustrum of a cone. Feed
enters at the apex and is forced into the separating fluid via an annular
curtaining hood. The heavy fraction sinks and is swept to the discharge
section by slowly rotating plates.
The Chance process differs from the DMS methods in which a relatively
stable slurry flows through the system. It more nearly approaches hydraulic
classification, but the shape of the cone and the method of introducing the
'hydraulic water is designed to maintain a dilated bed of sand which for the
most part remains at work in the separating zone. The large volume of
water overflowing with the clean coal (Fig. 129) is fed to the upper part of the
cone and has not been used to dilate the main separating zone as would be the
Fig. 126. The Drewboy Separator (Automatic Coal Cleaning Co., Ltd.)
case in straight hydraulic classification. The sand used is silica sized between
30 ~ and 100 ~. When this is dilated by water rising between f' and -V I
second (according to packing voids of the sand), the quicksand density is
2::
:;"
'"
~
01' PI.'U ~
(OINSITT
TO/' 01' .... ~ UOVIO L1vn CONTaOLUA)
"~
1"o
ROAn
OI$CHAAG( _-.J
a'"
2::
~
v,
DRAIN {5
I:>
i3
~.
;:,
Fig. 127. Arrallgement of Shallow-Bath Two-Process Separator (Automatic Coal Cleaning Co., Ltd.)
N
00
Vl
286 Mineral Processing-Dense Media Separation
'"
<l)
....<1l>
U
c:
o
E
i:ii
00
N
Mineral Processing-Dense Media Separation 287
COAl >
. . ··-··~···d , '1,'
->{,-
··CI<ANCE· SIlN~t=r
flOTATION CLEANE~
: SlURAT
: CISTERN
~:=:
~EFUIE
SeftUN!
-lO B 5\
CLEAN COlll
IOBSS.
AEFUS!
easily held at a point between 1·4 and \·7. Gravity control is effected by
varying flow-rate in the high-gravity and low-gravity zones. A particle of
coal \·676 mm. in diameter (10 ~ 8.S.S.) and S.G. \·55 has a terminal falling
velocity of about 6"/second in a liquid of S.G. \·5. It thus has no difficulty
in sinking through a current rising J"/second. The stirring mechanism in the
cone rotates slowly and assists the coal outward to the overflow weir. Over-
flow coal and shale discharged through the automatically controlled refuse
gates are desanded, the sand being returned to the top of the cone. The
normal cone does not have the additional middlings discharge shown.
Operation is checked by testing the products with heavy liquid of the correct
separating density. Too dirty a float is corrected by increasing the propor-
tion of water to sand, or by reducing the rate of feed.
In Fig. 130 the flow-sheet for a two-process treatment is shown. Dry or wet
screening removes the -10 ~ material from the raw feed for froth flotation.
and sends the - 10" -+ 10 # coal to the Chance cone.
A still more pronounced step away from "weighing" the feed into a quiet
bath is applied to the separation of heavy minerals in the Stripa process,
shown in Fig. 131. As in the Chance process a hydraulically dilated quick-
sand is used, water rising into the shaking trough (I) through the valves (F).
In a typical Swedish operation a fairly coarse magnetite sand is concentrated
on shaking tables from ore undersize screened through, say. 20 mesh, and
passes along the trough at a teeter density of 3·4. Coarse feed entering at (9)
either floats on this teeter or sinks into it, the two fractions being separated
by (D) into floats (7) or sinks (8). The media drains through (5) and is either
recirculated via (6) or sent to further treatment, since in this particular case
it is an economic mineral. With a copious supply of autogenous media no
regenerative treatment is needed, though this can be achieved simply if
desired, since the grain size is coarse enough for slimed material to be screened
away. There is no problem of dilution of media by wet feed. and the process
tolerates unscreened run-of-mine feed.
DMS in eye/ones. In the Chance and Stripa processes an ascensionary
hydraulic pressure is imposed in place of a quiet bath. In the cyclone used
for OMS, though there is a minor movement of ore roughly parallel to the
long axis. the main driving force is centrifugal. Consideration of the for-
mula for centrifugal force (mv)p) shows a complicated work-pattern at any
cross-section in the sorting zone. The intensity of g-acceleration on an
individual particle in an evenly swirling system is a function of its radial
distance from the axis. This is modified in practice by the fact that the
particle is entrained in an environment which is itself being subjected to
sorting action. The swirling system increases in viscosity and S.G. outward.
but the swirl decelerates toward the periphery because of drag against the
containing walls of the cyclone, thus introducing a stirring element due to
shear.
Since the feed in a typical cyclone treatment is smaller than 10 mesh
(say minus 2 mm.) its area-volume ratio is much higher than that of coarse
material fed to a quiet bath. At first sight the viscous restraint exerted on
each particle might appear a major factor in dictating its response. Such
braking force is. however. being applied in a spinning system in which both
Mineral Processing- Dense Media Separation 289
SECTION OF TROUGH
VIEW OF UNIT
brake and particle are moving at roughly equal radial speed. Hence, the
viscosity of the medium only affects the situation while the particle moves
inward or outward along its radius, and is therefore a minor influencing
factor.
To simplify the approach to our study of the complicated dynamics of
cyclone separation we can start by assuming the media to be a heavy
liquid. This defers the extra problem of centrifugal distortion of the S.G.
of the bath from its lightest central core to its heaviest outer zone. Consider
next the distribution diagram in Fig. 132.
290 Mineral Processing- Dense Media Separation
~L~
Feed@ pF
I~ pL
F~
pF ~
r-CYdO"~l
T
He» L@
pH pL
Fluctuating p
H~
pH
Centrifugal pumps are less satisfactory for feeding in the ore pulp than are
those which deliver a constant volume at a steady rate, though they are in
wider use. The centrifugal pump is easily upset in delivery rate by wear in
any part of the system, since a change in the pressure head against which it
works varies the slippage of pulp passing through its impeller system. When
the feed rate is steady the vortex finder accepts delivery of the float fraction
without altering the pressure in the cyclone. If this pressure rises, the apex
discharge flow is increased and float product is being mishandled and wrongly
delivered. Where feasible, a constant-head tank between pump and cyclone
could be used. The volumes of solids rather than their weights in the pulp
are the key consideration in reaching a constant rate of properly blended
feed.
Smooth running is often aided by instrumental control based either on
pressure-sensitive devices which control the apex area or on radio-active
gauges which measure the pulp consistency and tonnage rate at a suitable
point. One such device measures changes in the vacuum in the vortex zone
and then controls the annular area at the apex by varying the hydraulic
pressure on a rubber ring. Another arrangement, best suited to large-scale
treatment where close sorting is not vital, is the "umbrella discharge", 12 in
which only a selected core of apex underflow is removed, the peripheral por-
tion being re-circulated.
Tests in a 6" cyclone working in the 2-4 to 3·5 range of density at a IS'
pressure head 13 confirm the importance of using ferro-silicon of a size appro-
priate to the required parting density of the bath. The finer grade used in
these tests contained 50% minus 20IL and only 2 % plus 43IL, the minerals
tested being in the minus Jt" plus 28 mesh range. Somewhat larger sizes
were successfully concentrated in a'12" cyclone, the test ores including such
minerals as limestone, silica, magnetite, phosphate rock, iron ore and chro-
mite. Commercial plants accepting feed up to I" in size are made, but in our
present state of knowledge ores exceeding some 10 mesh (2 mm.) are better
handled in quiet baths, provided the required degree of liberation can be
reached in these larger sizes. An exception might be a case where the bath's
parting density was just insufficient in a quiet bath. Here, the centrifugal
effect on a DM circuit charged with ferro-silicon could simulate a moderate
extra bath density.
This, together with the regulated return of cleaned media adjusted to the
correct bath density, is the critical control factor in DMS processes. The
usual sequence of operations is:
(a) Drainage of media from ore leaving the bath.
(b) Washing of ore products to remove balance of adherent medium.
(e) Collection of foul media, and cleaning it by a suitable process.
Since this is a mineral-dressing operation the methods are appli-
cations of those described later under:-
(I) Flotation. (For galena, and for barytes fouled with coal.)
(2) Magnetic Separation. (For ferromagnetic media.)
Mineral Processing-Dense Media Separation 293
Bath Density
The effect of varying the specific gravity of the dense media during a test
for amenability of ore to DMS is shown in Table 23. The purpose of the test
was to find whether a low-grade ore could be made worth treating. At a
separating density of 2·7 just under 54 /~ of the feed was retained as an en-
riched sink, while by stepping up the cone density to 2·8 the enriched sink
was reduced to 7 %, a float of 93 % of material too poor for treatment being
discarded.
If the cost of D MS treatment was of the order of 1s. per ton and of further
treatment to recover saleable cassiterite and wolframite 5s. per ton, the cost
would be, for 100 tons:
(a) Treating all the feed £30
(b) DMS at 2·7 S.G. £18t
(c) DMS at 2·8 S.G. £ 7.
The feed for this test was -~ 1" + 10 mesh, and the ore minerals included
cassiterite, wolframite, quartz, felspar, muscovite, and biotite.
TABLE 23
References
I. Bird, B. M., Mitchell, D. R., and Smith, F. E. (1943). "Coal Preparation",
A.I.M.M.E.
2. O'Connell, W. L. (1963). Trans. S.M.£. 226.
3a. Roe, L. A., and Tveter, E. C. (1963). Ibid.
3b. Baniel, A. M., et at. (1963). Ibid.
3c. Baniel, A. M., and Mitzmager. A. (1960). I.M.P.e. (London), I.M.M.
4. Pearson, A. Trans. I.M.M. (Lond.), 48.
5. Rodes, F., and Cremer, J. (1960). World Mining, March.
6. Geith, G. (1958). Trans. I.M.P.e. (Stockholm), Almqvist & Wiksell.
7. (1955). "Das Stripa-Schwimm-Sink-Ver:fahren" Erzmetall, 4.
8. American Cyanamid Co. Ore Dressing Notes, No. 14.
9. Moiset, P., and Dartois, R. (1960). I.M.P.e. (Lond.), I.M.M.
10. Anon. (1961). The Mining Magazine, May.
11. Cohen, E., and Isherwood, R. J. (1960). I.M.P.e. (Lond.), I.M.M.
12. Dreissen, H. H., and Fontein, F. J. (1963). Trans. S.M.E., March
13. Davies, D. S., Dreissen, H. H., and Oliver, R. H. (1963). I.M.P.e. (Cannes),
Pergomon.
i4. Nesbitt, A. c., and Weavind, E. G. (1960). I.M.P.e. (London), I.M.M.
15. Oss, D. G., and Erickson, S. E. (1962). S.M.E., May.
CHAPTER 13
Introductory
The laws governing the motion of a particle immersed in a fluid show its
behaviour to be determined by its mass and its surface friction. The gravi-
tational force necessary to initiate and to sustain its movement is modified
by the specific gravity of the circumambient medium, as was noted in
Chapter 12.
The net force acting on the particle is
f = v (Ll' - Lln)
(13.1)
g
where v is its volume and Ll' its density, Lln is the density of the fluid and g
the acceleration due to gravity. If Ll' >' Lln the particle gravitates downward,
if Ll' < Lln it tends to float. If the densities are almost equal it may lack
sufficient energy to move relative to its fluid surroundings. The greater the
excess of Ll' over Lln, and the larger the value of v (the bigger the particle)
the more readily and speedily it will fall. The potential energy of the im-
mersed particle is converted to kinetic energy, and used to overcome the
retardation (viscous, eddying and colliding) which opposes the attempts of
the particle to move. In commercial ore treatment by gravity methods it is
essential that the particles move briskly through the separating appliance.
When they settle so slowly that the flow of the pulp past them is laminar, or
viscous, the terminal velocity (in accordance with Stokes' law) varies as the
square of the diameter d. With fast and turbulent flow (Newtonian settle-
ment) it varies as vd.
This rules out the application of most gravity-based processes to solids
settling according to Stokes' law, because surface friction dominates their
reaction to the fluid medium. They therefore move too sluggishly to be
treated at the speed required if a reasonable tonnage is to be efficiently handled.
For gravity to act as the selecting force, there must be a marked difference in
settling rate between the heavy and the light particles in the feed. The closer
the specific gravity of the minerals bein~ separated, the coarser must be the
minimum size treated. For example, 14 150-mesh particle of quartz (S.G.
2'7) has a mean diameter of about 0'104 mm., and a falling rate in still water
of about 0·7 mm./sec. A quartz particle of 65 mesh (0·208 mm. diameter)
falls about 26 mm./sec. The tremendous increase in the falling rate due to a
doubling of the diameter might make it possible to use treatments based on
gravity for the coarser-sized material while ruling it out for the finer mesh.
296 Mineral Processing-Separation in Vertical Currents
Finally there is the hindered settling effect. The particle under considera-
tion is one of millions, each moving individually according to its physical
nature, but subject to the constraining influence of the mass movements in the
separating machine. The total number of particles for a given solid-liquid
ratio varies inversely as the average particle size. Consider a single cube of
quartz, I cm. long and weighing 2·7 gm., dropped into a vessel containing
I ml. of water. If this cube is reduced to smaller cubes by successive halving
of length of side, the changes in settling conditions take place in accordance
with Table 24.
TABLE 24
This suggests that with each halving of mesh size there is an eightfold
increase in opportunity for collision and a similar decrease in the kinetic
energy of the individual particle, together with a doubling of its surface
area/size ratio, although there has been no change in the solid-liquid ratio.
Only approximate prediction of the settling characteristics of particles
moving under constraint is possible, but empirically the retardation of fall is
of the order of 50 %-65 % of the free-settling rate.
Applied Forces
At first glance the use of gravity in mineral separation may appear some-
what elementary. In practice it can be extremely complicated. Process
control must always be kept as simple as possible, but a number of mutually
contending forces are at work. Though this chapter is concerned with the
use of vertical currents in mineral processing, streaming action is also occur-
ring. Study can usefully begin with the vertical component of separation,
exploited in classification and jigging.
Consider a mixture of grains of galena (S.G. 7'5) and quartz (S.G. 2'7)
which have been limited in size-range by screening through 14 mesh and on
20 mesh. Particles of similar shape and having a diameter of 1·1 mm.
have terminal velocities of 260 mm./sec. and 96 mm./sec. when falling freely
through water. If, instead, they are placed in a column of water rising at a
speed of 150 mm./sec., the particle of galena appears to fall 110 mm./sec.
while the quartz appears to rise 54 mm./sec. Because the size variable has
been eliminated by the use of screens, the settling variable can thus be used to
make these particles pursue opposed paths. (Fig. 133).
Mineral Processing-Separation in Vertical Currents 297
The more closely the feed to a rising current has been ranged by sizing
upon screens, the greater will be its accuracy of response to the combined
influence of the rising water and its own mass. The main conditions which
determine whether an ore can be treated by gravity methods can be repre-
sented as a concentration criterion (Taggart). This is a ratio
(L1h - L1f)
(L1L - L1f)
where L1h is the S.G. of the heavy mineral, L1r that of the fluid medium, and
L1L that of the light mineral. When the ratio exceeds 2'5, effective concentra-
- -
~
iiiiiiii
Water Rising
Fig. 133. Separation in Hydraulic Current
tion is possible down to the size of fine sands. With a ratio of 1'75, com-
mercial separation is possible down to 100 mesh. At \·5 the limit of fineness
is around 10 mesh, and at \·25 only gravel sizes can respond satisfactorily.
There is a big difference in sensitivity between the forms of gravity separation
now to be considered and the quiet ones of DMS in which accurate splitting
298 Mineral Prncessing--Separatioll ill Vertical Currents
The fifteen factors thus summarised do not complete the possible list, but
they indicate the complexity of gravity separation and explain the need for a
largely empiric approach to the treatment of each ore.
Elutriation of a single particle stages a conflict between its mass and the
upward urge exerted upon its area and cross-section by rising water. In
jigging this force is pulsed and new conditions enter. The hydraulic current's
rhythm takes the form of a sine curve (Fig. 136). The inertia of the particle
is overcome at some point on the upward part of this curve, when it starts to
rise and to acquire kinetic energy by transfer from the water. If light and
small, movement begins while the hydraulic urge is building up; if large and
heavy, as the water is decelerating toward zero from the top of the pulsion
stroke. Thus, as long as the water rises faster than the particle the latter
continues to receive energy. When the situation is reversed the particle
uses any excess energy it has stored in continuing its rise, and then starts its
fall, tracing a sine curve of smaller vertical amplitude which lags behind that
of the pUlsing water. The point can be reached for a sufficiently weighty
particle where it is still descending while the water is rising on its next pulsion.
The point at which its reversal occurs, and the direction and strength of water
flow at that instant varies according to the particle's mass. One might start
to fall during the dying away of pulsion while another might still be carried
by its stored energy over to the suction stroke.
When consideration moves from the single particle to the loaded jig-box a
complex situation is revealed. There is graduated teetering, with the top of
the box well expanded and the bottom rather close-packed. As stratification
proceeds the lightest particles in the upper strata are the first to loosen and
rise, and therefore travel furthest. Energy transfer from hydraulic water to
load is probably fairly even for all particles during rise, but as the teeter
collapses on the return half of the cycle loss by collision and friction is heaviest
300 Mineral Processing-Separation in Vertical Currents
in the lower strata. The hydraulic force consumed by the system must
equate the back pressure of the teetering load and overcome frictional
resistance. During pulsion the course of the water is turbulent, because it
bursts through the channels which offer least resistance as the packing of the
load changes. This causes momentary increased flow at such points. The
general picture must therefore include local disturbance in the teetering load.
At a late stage in the pulsion stroke the overall condition in the jig-box is one
of a system being pressed upon from below by water and from above by the
over-riding weight of the suspended solids. Since this storage-pressure is
released during the return half of the cycle, pulsion and suction are not
strictly equal, the suction stroke being reinforced by the downward move-
ment of the collapsing load.
In continuous jigging, with the load working across the jig-box so as to
introduce a streaming effect, extra water is usually added under the screen
(Fig. 137, water service) in order to increase the loosening action during pul-
sion and/or reduce it during the suction half of the cycle. New feed is con-
tinuously added to the top of the load on one side, and two products are
continuously withdrawn. These are the heavy fraction, either removed
through the screen or via bottom gates as shown in Fig. 137, and the light
fraction which works its way across the box. As this travels, it is repeatedly
dilated and resettled by the pulsing stream of hydraulic water.
Two jigging cycles are illustrated in Fig. 134, one at the first stage of
treatment and one when stratification is complete. In the pulsion stroke (a)
water has been forced up through the screen. The mixture of closely sized
l'
CL
particles (galena and quartz, for example) is dilated and teetering. Next the
pulsion decreases, dies out and the suction stroke commences. The heaviest
particles are the first to be affected by the reduced hyd:aulic lift, and they
therefore fall sooner and faster. As the teeter collapses it forms the bed (b)
with the galena beginning to concentrate at the bottom and the quartz at the
Mineral Processing-Separation in Vertical Currents 301
top. This stratification is helped by the fact that the galena particles have a
greater kinetic energy than the quartz and so are able to thrust aside the
lighter grains while the bed is still closing. (y) and (z) show the position
after a number of cycles, when the particles have arranged themselves in
strata determined by their mass and shape.
Instead of equal-sized and equal-shaped particles of galena and quartz,
consider next a mixture of various shapes and sizes, some being completely
liberated and some locked middlings, such as would be found in a typical
feed of crushed ore which had been screened to a size-range such as, say
-6 + 14 mesh.
Ignore for the moment the new feed and consider only a cycle of pulsion
and resettlement of the coarse grains. As water is forced up through the
bottom of the screen, part of its kinetic energy is transformed into an equiva-
lent static pressure. The back pressure at any depth in the dilated bed is
that of the overlying particles. If now the current is stopped (or reversed),
the bed packs tight, the action beginning at the bottom where the back pressure
is highest, and where any change in flow rate or direction is first felt. Thus, the
bottom particles interlock and present a barrier to those above, regardless of
density. During the change from dilation to interlock, the first heavy par-
ticles to be thus barred are the largest, and the last the smallest. This
explains an apparent anomaly in jigging-that small heavy particles move
downward more readily than large ones when the feed is long-ranged. These
small particles next pack the voids between the larger ones and thus increase
the bed's resistance to penetration. This packing effect must be con-
sidered when determining the size range which is to be used when screening
the feed. The greater the dilation of the bed the easier is the movement of
particles in jigging. This dilation is adjusted by controlling both the quan-
tity of hydraulic water used and the force applied to it.
Solid particles press down upon supporting media proportionally to their
immersed weights. Since the immersed weight of the particle is proportional
to its mass (volume x S.G.) this leads to the following effects:
(a) Particles of the same S.G. but of different sizes fall at rates varying
as size (or mesh).
(b) Particles of the same size but of different S.G. fall at rates varying as
specific gravity.
(c) Modifying (a) and (b) above, rate of fall is inversely as surface pre-
sented-equi-axial particles falling fast and plate-like particles
slowly for a given volume.
Jigs are run at speeds varying from 57 to 330 cycles/minute. With stream-
ing action added to the expanding and contracting teeter the movement
becomes one of plastic flow. With impulses arriving from two to five times
each second while the bed is being pushed toward the discharge side by
incoming feed, individual particles tend to dance sideways as they stratify.
The effects of k and I in the above list of forces now appear. The stirring
effect of pulsion helps the particles to drift with the streaming water and to be
pushed along by the new feed working its way through and across. At any
vertical cross-section the load density has fallen steadily from feed toward
discharge. Since the hydraulic lift is applied equably over the whole of the
302 Mineral Processing-Separation ill Vertical Currents
Uni-Directional Separators
In Chapter 12 two methods which combined the use of hydraulic water with
a heavy medium were discussed-the Chance and Stripa processes. Com-
parative newcomers to separation by more or less direct classifying methods
are the Lavodune and Lavoflux systems.~ The former resembles, broadly,
a straight-sided hydraulic classifier working at a tilt of 45° . Feed ranging
between 0·5 and 20 min. in size is treated, in a pulp diluted below 50 g./I. in
tubes 2 to 3 m. long, each of which handles a ton hourly per square decimetre
of its cross-section. The closer the size-range the more efficient is -the
action. Over 4,400 gallons of water per ton of 6re are kept in motion at a
water-gauge pressure of from 6 to \3 feet. This consumes from 0·2 to 0·8
kWh./t. of ore. As the fe~d moves upslope the lightest particles are swept
up and out of the tube while the heavier material forms a rotating dune from
which entrained light particles are freed as it performs an oval roll. During this
rotation part of the load breaks away periodically and falls to a separate
discharge. The Lavoflux, described in the same Paper, treats 0·1 to 0·5 mm.
feed and forms a fluidized dense medium instead of a dune.
Ore-box
Hutch
-.. . . .
.•' _ 't''' I" _ " !. . " :.,..
upward through the loaded box. The stroke direction is now reversed,
usually more gently in order to avoid making too compact a bed.
Hand-operated jigs include the Willoughby, the Cornish Kieve, and a
Chinese method of jigging on hand-held screens. Since the jig motions can
be duplicated with mechanical precision, hand jigging is not much used today.
The stratifying effect of the up-and-down movement of the ore box in a static
pool of water is like that produced when a vertical column of water is pulsed
upwards and downwards through a stationary ore box.
the whole of the material more or less mobile. Many jigs are driven by
eccentrics or linking motions which accelerate the stroke in accordance with
the harmonics of an angular motion. Thus the jigging cycle is seen to be far
more complex than a uni-directional teetering column.
The fluid normally used in jigging is water. Air has had limited applica-
tion in the pneumatic jig, used in desert countries. Attempts have been
made to jig in dense-media fluids, but no industrial process of importance
has been reported apart from the jigging element in the Stripa DMS machine.
The water used can be set in motion in several ways. Gentle oscillation
between the limbs of a U-shaped box is used in the Baum coal jig, the pulsing
force being applied by low-pressure air introduced through valves at inter-
vals, so as to sustain the oscillating action as the water rises and then gravi-
tates back through the material resting on the screen. For heavy minerals a
loosely fitting piston or a flexible diaphragm is used to impart more rapid
pulsion strokes, followed by positively powered return (suction) strokes.
Abrupt short strokes are possible because water used in this way exerts a
plastic thrust rather than a streaming action, and the particles with the least
mass and inertia are most readily set in movement. Before a particle thus
urged into motion has time to settle in response to the ensuing suction stroke,
the next thrust is delivered. The jig bed is thus kept in partial teeter, only
the upper layers being fairly free to move. By this relatively violent method
of applying force the work is done with a moderate quantity of water in motion.
The Denver, Bendelari and Pan-American jigs use compression in the water
chamber instead. An alternative is to move the jig box up and down in a tank
of water (Hancock and Halkyn Jigs).
Kirchberg and HentszcheP have derived this formula for resistance of a bed
of spherical grains to loosening (LJp)
V 2m h m
LJp=t/;-YM- - g./cm 2 ., (13.2)
2g dill V4 R
VR = voids in bed.
yF = specific weight of solid material.
Such a bed resists loosening while the general flow rate of the jigging fluid
(Vm) lies between Reynolds Nos. 20 and 3000. The maximum rate of flow such
a bed can resist without opening is:
306 Mineral Processing-Separation in Vertical Currents
-----------
(YF-YM) (1- VR) 2gdm V4 R
Vm = y' cm./sec. (13.3)
!fYM
Consider first the case of coal jigging, where there is a comparatively small
density difference between mineral and jigging water (say 1·5: 1'0). The
gentle pulsion stroke needed to dilate such a bed permits greater refinement
in the work than is practicable in jigging minerals with a density ratio of 3: I
or higher. The jigging cycle for coal may be represented as a sine wave
(Fig. 136) in which water velocity is plotted against time. 6 Starting from
rest, the water commences to rise through the closely packed bed of coal with
increasing velocity. Dilation of this bed begins with the loosening of the
uppermost stratum. (In jigging coal it is possible to apply water so as to toss
the whole bed solidly upward and then to allow it to open out upward and
downward, but this is not feasible with heavy minerals.) As the water
column attains a rising velocity of 100 mm./sec., the upper part of the coal
bed opens and begins to teeter. Velocity continues to increase, and layer
after layer opens until the whole bed is dilated and teetering. Then, as the
pulsion stroke passes its peak, the velocity of the water lessens and the suction
stroke begins. At this point each particle in the loosened bed has some
degree of mobility, and has acquired both kinetic and potential energy in
accordance with its movement and vertical displacement in the teeter bed.
These it dissipates as the suction stroke develops. The particles fall down-
ward in the falling column of-water and repack into a tight bed. First there
is momentary free settling, then hindered settling with the teetering mass
rapidly consolidating to the packing density of the solid phase, less voids.
Mineral Processing-Separation in Vertical Carrents 307
The bottom of the bed is the first to pack tight, then the middle shuts down
upon it, and last the top closes. Jigging is thus seen to be hindered settling
carried out under oscillating conditions during which the material passes
from fully fluid conditions (S.G. 1) to those of a bedded deposit at each
cycle. In coal jigging the pulsion-suction cycle is usually about 57/minute.
In treating heavy minerals, where more violent action is needed, speeds up to
280 cycles/minute are employed.
If a continuous upward current were used, stratification of the jig bed would
result, with the smallest and lightest particles uppermost, the largest and
heaviest ones at the bottom and a mixture oflargish light particles and smallish
heavy ones in the middle of the teeter bed.
If the material had been close-sized, this would result in clean separate
strata (heavy below and light above). Such close sizing would not be eco-
nomic in commercial practice.
When, instead of as a continuous rising current, the water is applied in a
succession of up-and-down pulses, a new factor appears. As the lifting current
becomes gentler, all the particles cease to rise, hover, and then begin to fall.
The heaviest are the first to change direction. and the lightest the last. Soon
the falling movement is accelerated as the upward pulsion stroke of the
hydraulic water dies out and downward suction commences. Hindered
settlement becomes rapidly more severe and the bed closes tightly. As the
spaces between the particles shrink, the bigger particles are arrested in their
fall. The smaller ones can still burrow down while this is happening, and are
the last to be stopped as the bed packs tight.
Since the heavy particles have the greater falling rate, stratification is effected
with the largest and lightest particles above and the smallest and heaviest
below, the remainder being disposed intermediately.
This accomplished, the rest of the work of the jigging appliance consists of
the orderly reception of new feed and its dispatch as separate stratified pro-
ducts. A list of requirements for efficient jigging of coal has been suggested
by Bird. 7
Two methods of withdrawing the heavy product are practised. In the
original "German" (on-the-screen) jigging the supporting screen is smaller
than the mesh of feed. Concentrate collects above the screen and is with-
drawn via side ports adjusted so as to maintain a bed of desired thickness
(Fig. 137, concentrate discharge). A later development, more widely used, is
"English" (through-the-screen) jigging, in which the screen mesh is larger than
that of the feed, so that the ore reaching the screen falls through to a receiving
chamber, the hutch. An advantage of the English method is that withdrawal
can be more closely controlled by use of an oversized layer of selected material
(broken ore, steel shot, etc.) called the bedding or ragging. This bed is
maintained in a suitably thick layer, and its component particles are chosen
for shape and specific gravity.
The controllable factors in through-the-screen jigging include:
(a) Amplitude of jigging cycle.
(b) Strokes per minute.
(c) Mechanical unbalancing of stroke-producing device.
(d) Hydraulic unbalaJ:lcing over selected part of stroke.
308 Mineral Processing-Separation in Vertical Currents
The two main types used in industry are "moving-screen" and "fixed-
screen" jigs. In the former the screen box is moved upward and downward
in a tank of water. In the latter, the screen box is stationary and the jigging
fluid is pulsed upward and downward through it.
Water is normally the pulsing fluid. When it is desired to aid selectivity
by applying the dilating impulse more gently, the density of the water is some-
times increased slightly by permitting it to carry solids in suspension. Little
commercial development of this possibility is known, as the method creates
problems which offset its advantages. In the treatment of dredged tin
gravels de-sliming before jigging has proved its worth. 8 Extra power would be
needed to handle and circulate dense media. The need for removing drag-
out material would add to working costs.
Pneumatic jigging has a definite field of application. In arid country, and
in cases where ·it is desirable to avoid wetting the feed, air is a good pulsing
fluid within somewhat severe limitations of particle size. It has the further
Mineral Processing-Separation in Vertical Currents 309
Fixed-screen Jigs
The oldest type of jig in use is the Harz, which takes its name from the
Harz Mountains in Germany, where it was developed for treating the lead-
zinc ores which have been mined in that region for several hundred years.
In the Harz jig (Fig. 137) an ore box, 18"-24" wide, 24"-48" long, and 6"-8"
deep, occupies about half the plan area of a compartment in a water-filled
tank, called the "hutch". In the other half of this compartment a loosely
fitting plunger moves verti:.:ally, so as to pulse water upward and downward
with a force proportional to:
(a) The displacement caused by the variable eccentric actuating the
plunger.
(b) The speed at which the plunger is reciprocated.
(c) The leakage past the plunger.
The body of water thus set in motion reverses its direction of flow after
passing the centre board, and alternately dilates and closes the bed of material
resting on the screen which forms the bottom of the jig box.
One reason for using a loosely fitting plunger is to reduce "water hammer"
in the reciprocating system. A second is to enable the operator to differ-
entiate between the strength of the pulsion and the suction stroke by applying
"hydraulic water" at a controlled rate. In Fig. 137 it will be noted that a water
service is provided above the plunger. All water introduced at this point
either finds its way out with the tailings or to the concentrate discharge after
passing through the screen. If no hydraulic water is used, only the water
entering the hutch with incoming feed is available. Since this arrives above
the screen, the pulsion and suction strokes are equal and opposite, save for
water lost through leakage from the hutch and periodic withdrawal of con-
centrates at the spigot of the hutch-work discharge. When hydraulic water is
added, the pulsion stroke is fortified and the suction stroke weakened, pro-
portionately to the amount so added. Thus, hydraulic water is employed to
aid in controlling the tightness of the closed bed and the differentiated speed
and violence with which it is opened and closed.
Several successive compartments-usually four-are placed in series in the
hutch. The sized feed is introduced gently at the head of the first compart-
ment, care being taken not to upset the stratifying work by plunging it in
violently. The feed should be distributed evenly over the full width of the box.
In the first compartment the controls are manipulated so as to produce as the
bottom layer a high-grade concentrate. The lighter fraction of the feed
stratifies upward and passes over the far end of the compartment into the
second section, and so on down to the final discharge. Concentrating
conditions are made progressively less discriminating from head to discharge
so that a lower-grade concentrate or "middling" is removed from the later
310 Mineral Processing-Separation in Vertical Currents
OGI.
C<>w(.wlnll.
DIICAarg,/
boxes, the top discharge from the last box being the jig tailing.
The concentrate can be withdrawn from the Harz jig in one of two ways.
If the feed particles are larger than the aperture of the supporting screen, they
cannot find their way down into the hutch. Jigging "over the screen"
originated in Germany, and is sometimes called German jigging. It is
characteristically used with fairly coarse feed and moderate suction, and the
grade of concentrate produced is partiy regulated by the thickness of the
bottom stratum, which is determined by the rate of withdrawal through the
undergate of the concentrate discharge.
A later modification, the English method, was introduced when it was
found possible to treat finer material. The feed is finer than the apertures
in the supporting screen, so the concentrate is mineral which has succeeded in
working its way down through the screen during the jigging cycle, while the
tailings consist of particles unable to descend under the prevailing operating
conditions. The hutchwork discharge is delivered to a receiving vessel in
such a way that concentrates can be removed when the hutch valve is closed
periodically, without upsetting the water-level in the jig-box.
A number of plunger mechanisms and product take-offs have been used.
The Baum jig is described below.
The amount of water used in jigging varies considerably from one mill to
another. If a round figure of 5; I per compartment be taken as a reasonable
average, twenty tons of circulating water are used per ton of ore treated in a
four-compartment plunger jig. This water is mostly reclaimable, but must be
kept in hydraulic action by applied power and transported through a conserv-
ing system. Instead of plungers. many modern jigs use flexible diaphragms
to produce the stroke. These tend to be more economical of water and allow
higher running speeds than are practicable with loosely fitting plungers.
Mineral Processing- Separation in Vertical Currents 311
In the Bendelari jig (Fig. 138) the diaphragm (A) is reciprocated from
below and hydraulic water is admitted at H. A modified form of this jig is
widely used for concentrating the values from alluvial gravels in dredging.
Jigging
The jig is widely used in coal cleaning. Reversing the conditions normal
with most minerals, coal is the abundant and light fraction. There is also a
marked shape factor, since the impurities are tabular. The range of size fed
to the jig determines the efficiency of separation. The closer the sizing, the
higher this becomes. If a jig is set to operate at a density of, say 1'5, this
means that it should produce washed coal with the same ash content as that
from a heavy-liquid float at a specific gravity of 1'5. Part of the sample, and
of the product, would have floated at a lower density. Of the several types of
coal jig, that in greatest use is the Baum. Various developments from the
original jig have been made. Where there is a marked difference in the breaking
characteristic of high-ash and low-ash coal it is possible, by splitting the feed
at two sizes to separate jigs, to obtain better results by operating each size at
its most appropriate density-equivalent. Normal modern practice avoids such
a complication and treats all the - 6" feed in one circuit.
The Baumjig (Figs. 140, 141) has a stroke rate of 55-57 strokes/minute, the
water being pulsed by means of compressed air. The bed recloses by unas-
sisted gravitation, no suction stroke being used. This closing can be modified
by the use of hydraulic water, or by slowing down the release of the pUlsing
air during the return stroke of the jigging cycle. The jig box has a V-shaped
cross-section and each of its two compartments forms a complete jig. The
screen of the first compartment slopes backward toward the feed end. The
feed is introduced gently, perhaps below a very shallow baffle to ensure wet-
ting, but not with sufficient disturbing influence to upset the stratification
taking place in the box. The weirs between the boxes in the older types have
been abolished for the same reason, so the feed proceeds with, if anything, a
slight climb through the jig. Air pressure is I t to 2t lb. at least. Most
of the dirt is dropped in the first compartment. Part of it falls through the
screen to a spiral conveyor which moves it out. With one type of control the
oversize dirt builds up on the screen until it has lifted a controlling float, which
responds to change in bed density at the interzone between dirt and coal. This
actuates a switch that starts the elevator motor and dirt is withdrawn through
the refuse gate until the falling density causes the float to switch off the motor.
In a variation, float motion controls secondary pulsion under the discharge
gate. Thus the thickness of bed is automatically controlled. The float
setting correlates the jig discharge products with the required average density
Mineral Processing-Separation in Vertical Currents 313
perforated plate. the greater the overall resistance to the water being pulsed
upward. Taking the tine of least resistance. water surges more strongly
through the tube and raises the float at each pulsation. Above the shale
gates (A and 8. Fig. 142) is an air chamber C which traps air when the release-
valve 0 is closed by the float. This air-cushion prevents water pulsation
near the end of the plate and shale discharge stops. When 0 opens the air-
pressure is released. pulsation occurs and discharge takes place.
To check performance the shiftsman is given a supply of heavy liquid.
He judges by the amount of floating coal shown in a sample whether the
setting is right. The second jig box delivers a coal middling which is either
suitable for use as a low-grade fuel or can be given a liberation crushing and
returned to circuit.
The Remer jig (Fig. 14:;) has a single long box. Impulses are produced by
dual eccentrics which give differential acceleration to the three hutches. This
Mineral Processing- Separation in Vertical Currents 315
VALVE
OPEN
Fig. 142A. Automatic Shale Dis- Fig. 142B. The Same Operating with
charge Control Operating with Thin Thick Shale Bed
Shale Bed (Automatic Coal Cleaning
Co. Ltd.)
Practical Application
Since jigging exploits the difference in density between particles of similar
size (and to a minor extent differences in shape), the process is mostly suitable
for ores which are adequately liberated at a fairly coarse m.o.g. It is theo-
Mineral Processing-Separation in Vertical Currents 317
Operation
Usually the size range of the feed is limited by screening ahead of the jig,
the spread of this range depending chiefly on the differences between the
specific gravities of the particles which are to be separated. When this
difference is small, the feed should be close sized and the suction stroke should
be gentle, as separation of equal-sized particles can best be achieved by
stratification according to density.
Sometimes-though this is not good practice-the feed to the jig is hy-
draulically classified, so that it lies in a range of particles which settle more or
less equally. Jigging action to be effective must now assist the small, heavy
particles to burrow downward through the interstices in the bed as it closes
after the dilating stroke. In this case the jig should be run with strong
suction, thus aiding penetration of the small sizes.
When jigging "through the screen", the apertures must not become blinded.
If the discharge of concentrate from the hutch is continuous. water in excess
of that required to modify the stroke cycle is needed to compensate that lost
with the discharge. It is added together with that used to modify the suction-
pulsion cycle.
Feed is usually sized by screening in an appropriate series of fractions from
-2" downward. Heavy-mineral jigs are run at speeds varying from 100
r.p.m. on coarse feed, with a stroke amplitude as high as 2", down to 300 r.p.m.
with at" stroke on fine feed. The relationship between speed and "throw"
must take account of the working stresses at the point where the jigging
impulse is given to the water column. As was noted earlier in this chapter,
this relationship does not by itself ensure correct hydraulic conditions in the
jig box. Coal jigs, such as the Baum. work at 57 cycles/minute. Here
gravity alone supplies the return half of the cycle and the speed is chosen to
suit the harmonics of the oscillating column of water.
The capacity of the jig is proportional to the transporting effect of the
horizontal current from feed to discharge end. as modified by the restricted
mobility of the bed. Two factors determine the maximum efficient rate of
feed. The first is the time needed for the finest desired particle of concentrate
to settle down and be withdrawn from below during its progress from feed to
discharge end of the box. The horizontal speed is determined by:
(a) The volume of feed (ore + water/minute).
(b) The cross-section normal to flow.
(c) The nett volume of jig-water (excess of pulsion over suction)/minute.
The second factor is the smoothness of working conditions in the ore box.
Mineral Processing-Separation in Vertical Currents 319
If the feed is run in so rapidly as to churn up the bed, it takes longer to secure
stratification, and separation suffers. If the feed is such that heavy use of
hydraulic water, or violence of stroke is needed to procure adequate dilation,
then there is risk of local breakthrough of excessive quantities of water.
This, by churning up the bed, upsets stratification and leads to loss. The
balance between total mass of material in the jig box, masses of individual
particles and blend of particle sizes must be such as will lead to even diffusion
of the pulsed water through the interstices of the jig bed. There must neither
be danger of local weakness leading to a breakthrough nor insufficient
dilation to achieve stratification. To aid transporting rate, a slight drop from
one compartment to the next is better than the use of streaming water.
Some operators, particularly those working on light material such as coal,
hold that the turbulence set up as ore falls over to the next box in a series is
undesirable, and prefer to run the compartments either level or at a slight
upward climb. The capacity of an ore box is proportional to the screen
area, and particularly to efficient delivery of feed over its whole width.
In addition to variation of the length and rate of the strokes, and the
addition of hydraulic water, there is another important variable. This is the
bedding or "ragging" layer, which rests immediately on the jig screen. Under
working conditions the contents of the ore box sort themselves into three
fairly distinct layers (Fig. 145). At the bottom is the separating layer,
which should be maintained at a grain size and density suited to the severity
with which it is to oppose penetration by particles from the feed. In jigging
through the screen the concentrate must fall through this bed, which may be
autogenous, but is usually composed of particles of a material chosen for its
specific gravity, of a size larger than the screen openings. Various materials
are used, among which are steel punchings, shot, galena, and other heavy
liquid ratio. This ratio changes from a maximum when the bed is c1ose-
packed at the end of the suction stroke, and a minimum when it is fully
dilated shortly after the peak of the pulsion stroke. The route taken by an
ore particle (also in up-and-down motion) is relative to the changing bed
density and to the residual kinetic energy of the particle after it ceases to
receive hydraulic thrust. In jigging "through the screen" heavy mineral
from the feed must pass through this bed. It is drawn down during the
suction stroke and in uue course falls to the hutch, since it is undersized to
the screen. At the same time, lighter mineral seeking to penetrate the
separating layer lacks the necessary dynamic thrust, and is rejected upward.
Thus, the thicker and heavier the separating layer the cleaner will be the
concentrate. The main factors determining whether the feed particle is re-
jected, held in the bed, or passed down through it, are:
(a) Bed density (solid plus liquid).
(b) Size, shape, and size-range of bed particles.
(c) Depth and uniformity of bed.
(d) Size, shape, and density of feed particles.
(e) Balance between pulsion and suction strokes.
(f) Bed dilation.
Too slow a stroke, by failing to dilate the bed, makes·it almost impossible
for particles to work through to the screen. Too sharp a stroke may upset
the stratification and evenness of the bed.
Above this bed is the roughing layer (Fig. 145). This also has sorting work
to do, though of a less severe nature. The layer is composed of partly
liberated particles of ore, which have stratified downward from the passing
feed, in accordance with the forces already discussed. The roughing layer
varies its composition somewhat in accordance with changes in the feed,
the setting of the jig, and the rate of withdrawal of concentrate which it allows
to fall through. It rejects any coarse gangue presented to its upper surface,
and if the jig is making a high-grade concentrate it either retains or rejects
(upward) any middlings. When the jig is set to make a "concentrate" of
middlings, as would be normal practice in the final compartments, the
adjustments are such as to allow these particles to drop through. The
roughing layer must be thick enough to press down on the separating layer
with moderate force, thus helping to resist any "boiling" through of water on
the pulsion stroke. The feed to the compartment must not arrive with suffi-
cient kinetic energy to scour the roughing layer.
Uppermost is the transporting layer (Fig. 145). Here the feed is received
and spread across the ore box, and from this layer of light and fairly mobile
gangue particles the heavier material in the feed drops down to the roughing
layer. Much of the gangue streams across to the discharge end without leav-
ing the roughing layer. Thus, part of the original feed is withdrawn below
from the system as concentrate, and part, after rejection from the roughing
layer, emerges and is lifted out at the discharge end.
In a series of four ore compartments there is progressive diminution of the
weight of entering feed and, consequently, a slowing down of the rate of
horizontal movement of the solids, proportional to the weight leaving as
concentrate. Against this, some increase in the total water probably occurs
Mineral Processing-Separation in Vertical Currents 321
Working Control
This accords with the conditions analysed in the preceding section. Changes
can be made in:
(0) The density of the oversize material in the separating (ragging) bed.
(b) The thickness of the separating bed.
(c) The addition of hydraulic water.
(d) The rate of withdrawing concentrates (German jigging).
(e) The feed rate.
(f) Speed and length of jig stroke.
When concentrate is drawn from above the screen, the slower the rate
the more severe is the selective action and the higher is the grade of the pro-
duct. Provided the final tailing is not too high in mineral value, the draw-off
from successive gates can be adjusted on some such lines as this (Fig. 146)
reed I 2 3 4 Tailing
- Hea~ Pruduct
Hign·graae Con,. Low· grade Cone. Good Midds. Poor Midds. (Gate) Discharge
Ta iling
Oischcugt
References
I. Kirchberg, H. and Hentzschel, W. (1958) I.M.P.e. (Stockholm), Almqvist & Wiksel.
2. Robinson, H. Y. Coal Cleaning Research, 9, Bull.l.
3. Michell, F. B. and Swarnapradip, P. (1963). I.M.P.e. (Cannes), Pergamon.
4. Batzer, D. J. (1962). Trans.I.M.M. (Lond.), 72.
5. Condolion, E., Hoffnung, G., and Moreau, C. (1963). I.M.P.e. (Cannes).
Pergamon.
6. Bird, B. M., Mitchell, D. B., and Smith, F. E. (1943). "Coal Preparation".
Seeley-Mudd Series, A.I.M.M.E.
7. Bird, B. M. (1943). Ihid.
8. Chaston, L. R. M. (1960). I.M.P.e. (Lond.), I.M.M.
9. Weavind, R. G., and McLachan, D. F. C. (1961). 1.S. All.M.M., Jan.
CHAPTER 14
Streaming Flow
The term "film sizing" has been defined as "Reverse classification; sorting
of mineral particles on such flattish surfaces as sluices and shaking tables in
accordance with the sizes of the particles moved by a flowing film of water,
which exercises transporting force proportional to the cross-section exposed
to flOW".1 The thrust exerted by the streaming water can be used to size
homogeneous sand, to wash particles of a light mineral away from those of a
heavier species, or to separate particles of similar mass but different shape.
In ore treatment all three effects are produced together unless pre-treatment
has modified one or more of the operating factors-particle size, density
and shape. This streaming force is manipulated in various ways and
in several types of machine. The appliances used to separate heavy minerals
from relatively light ores depend mainly upon the work done by water as it
encounters ore particles while flowing down an inclined plane. This plane
may be smooth, rough, or riffled; gently or steeply inclined; stationary or
mobile. Instead of a plane, concave or convex surfaces may be used, or
compound planes. Flow may be quiet or turbulent, steady or intermittent,
and elements of vertical or centrifugal motion can be introduced. This sec-
tion discusses the forces that act upon particles moved by a layer of water as it
flows down an inclined plane.
The kinetic energy carried by this moving water provides the dynamic
force. Constraints, modified· cross-sections, obstructions, changes of
direction, eddies and turbulent zones can be introduced in order to check,
accelerate or produce a centrifugal effect. Essentially, however, the separat-
ing effect on particles of ore introduced into a streaming current results from
their reaction to two successive sets of circumstances. The first is their
lateral displacement or drift, which is determined by the time taken by each
particle to fall from the point of entry to the floor of the Sluice (the channel
through which the water into which it is fed is streaming). The second is the
resistance offered by each particle to further lateral displacement after it has
reached this floor. The mode of presentation of kinetic energy to the resist-
ing particle is different from that at work in a vertical current. Efficient
use of flowing streams as a means of applying separating force to ore particles
therefore requires a different preparation of the feed.
Consider first a simple sheet of water flowing gently down a smooth plane.
The rate of flow and depth of the layer depend on:
(a) Volume of water available.
(b) Width of plane.
(c) Speed of water in direction of flow at moment of arriving on plane.
Mineral Processing-Separation in Streaming Currents 325
A short distance up from the plane the current of water runs with a swift-
ness depending chiefly on the slope (i.e. the gravitational pull). Near the
top of the layer the water is slightly retarded by friction with the layer of air
above it. This may set up wavelets or interfering currents.
Lightest
(sma:llst)
Fig. 149. Particle Drift in Flowing Water
At the moment of reaching the plane the biggest high-density sphere, having
fallen fastest and consequently been least affected by the current, lies nearest
downstream to its point of entry. The smallest low-density sphere has drifted
furthest downstream. The others are to some extent overlapped.
It would obviously be possible to base a separating process on this dis-
placing effect, with a clean coarse concentrate upstream, a clean light gangue
downstream, and an overlapped mixture between which could be separated
by screening. The practical difficulties, however, outweigh the advantages.
On arrival at the surface of the supporting plane, the spheres become
subject to a new system of forces, which may be represented as in Fig. 150.
The pull of gravity anchors the sphere to the plane if this is horizontal,
but aids rolling if it is inclined. The flowing water presses against the cross-
section of the sphere and encourages it to move downstream. The differential
rate at which the laminae of water are flowing causes pressure to be lowest
next the plane and highest at the top of the sphere, thus exerting a slight
overturning moment which strengthens the tendency for rolling downward to
commence.
Mineral Processing-Separation in Streaming Currents 327
flowing
Water
g
Fig. 150. Thrust 0/ Flowing Water on Particle
Smallest
( heavy)
'--'. ~~
--.
~%~
. • -FlI'at.er '- -
(lignt)
Application
It is now clear that three physical characteristics of the mineral particle play
a dominant part in determining its behaviour in film sizing or film separation.
These are
(a) Shape.
(b) Volume.
(c) Specific Gravity .
In the previous section it was convenient to simplify the picture by consider-
ing a sphere. In practice, the shape determines the cross-section presented to
the push of the stream and the ease with which rotting can be induced. If not
packed together, particles settle in their most stable positions, each on its
flattest available base. A single particle exposes a minimum cross-section to
the stream and offers maximum resistance to overturning movements. Since
it has achieved a maximum of frictional hold it is to that extent reluctant to
skid. If it lies close to the plane, the passing water rises over it and presses it
328 Mineral Processing-Separation in Streaming Currents
gravels to a sluicing height from which they can run gently down to a tailings
dump, the values being trapped along the sluice behind riffles. The Chinese
"palong", used in Malaya, may be fed thus or from a centrifugal sand pump
which elevates the tin sands (karang) from an opencut or lorn bong.
MINERAL BEARING
ALLUVIUM
In Fig. 154 one method of concentrating cassiterite from rich tin gravels is
shown. First the barren overburden is removed-perhaps by hydraulic
monitors and jet pumps if abundant pressure water can be cheaply led to the
workings. The mineralised gravel is next washed down to a sump cut in the
barren bed rock, and from here is pumped to one of the two or more long
parallel sluices. As the sands stream down this sluice they are kept from
packing too tightly by hand raking against the current. From time to time
riffles are set across the sluice, and existing ones are heightened. Thus the
light and valueless sands are washed down and run to waste while the heavier
ones accumulate and are trapped behind the riffles. When a full load of low-
grade concentrate has been thus produced, feed from the opencast (lombong)
is switched to a parallel sluice, and clean-up commences. Wash water is
streamed down, the sands being raked vigorously against the flow and the
riffles removed cautiously working upward from the discharge end of the
sluice. When the concentrates have been sufficiently upgraded in this way
they are removed for final cleaning by more delicate methods, and the sluice
is again ready to receive run-of-mine feed.
The simple but laborious operations of the lone prospector or "fossicker"
still produce a substantial yield of tin, gold, and alluvial diamond. His tools
include the gold pan, with such variants as the batea and calabash. The
cradle rocker (Fig. 158) allows a quantity of gravel to be washed, if necessary
with some recirculation of water. Feed is placed on a punched screen at the
bottom of a small feed box. The mineral-bearing sand is washed through this,
large pebbles being held up and discarded. Clay which might trap valuable
mineral is either rejected, or puddled and disintegrated by hand. Water is
either piped in or caught below the rocker and recirculated by the use of a
Mineral Processing- Separation in Streaming Currents 331
dipper. Particles of gold may be trapped upon canvas screens below the
hopper, or they may be washed through into the sluice box. This is carried
upon rockers, so that it can be oscillated from side to side, thus giving some
fluidity to the sand as it gravitates downward. The gold stratifies and is
arrested by the retaining riffle at the bottom of the sluice, while the im-
poverished sand climbs over and is discarded.
Another method uses a stationary sluice, recirculating the water when
necessary by means of a hand-operated diaphragm pump. The bed of mater-
ial is stirred by means of rakes, rabbles, or hoes to prevent it from consoli-
332 Mineral Processing-Separation in Streaming Currents
Fig. 157. Malayan Sluice (palong). Left compartment is being cleaned up;
middle and right are being loosened by raking to aid settlement of cassiterite
between riffles running across sluices (Tin Industry Board, Malaya)
Mineral Processing-Separation in Str2aming Currents 333
dating too tightly to permit stratification. The sand bed must be kept
sufficiently open to trap values if concentration is to result.
The gold strake is used in milling to entrap coarse particles of gold at an
early stage in the treatment of ore.
The simplest form consists of a stationary inclined plane, sloping downward
with just sufficient inclination for a film of pulp to flow gently along. The
light mineral particles roll down, while the heavier ones stop on a removable
cloth lying on the surface. The pulp is distributed evenly over the strake at
entry, and slope, feed size, solid-liquid ratio and feed rate are adjusted to give
smooth flow with no channel formation. At one time rough blanketing was
used as the catching fabric. The modern straking system uses strips of
corduroy cloth laid with its ribs across the direction of flow. The nap of the
pile points upward, so as to entangle the gold falling to it. Corduroy is sold in
bolts 75 yards long and 28" or 36" wide. Three pieces of cloth are laid on a
typical strake, the upper sheets overlapping the lower ones. The head cloth
may be held down by means of a strip of flat-sectioned iron. Two strakes are
set in parallel so that pulp can run over one while concentrate is being removed
from the other. If several strakes are set in series, the head strake catches
most concentrate, and the uppermost cloth needs removal and washing more
frequently than the others. Pulp is fed on at a size below 10 mesh, and the
table slope is adjusted till the largest gangue particles just manage to progress
over the corduroy ribbing. The solid-liquid ratio of the pulp may be from
20% up to 50%. The tonnage fed varies from 0·2 ton/sq. ft. of cloth to 2·0
ton/sq. ft. per day, 0·3 being a normal figure. Concentration ratio is 1 :2000
or less.
The cloths are washed regularly, at intervals of two hours or more. Where
this work is done by hand, the pulp is diverted to the parallel strake, and the
cloth is folded on itself, refolded, and carried to the concentrates tub. This is
a vessel filled with water, with a metal grid locked in place below water-level.
The cloth is opened out face downward on this grid and shaken so that the
concentrate falls through. When the values have been scrubbed and sluiced
334 Mineral Processing-Separation in Streaming Currents
away, the cloth is refolded and returned to the strake. Bottom cloths may
be left in place for four or eight hours. Old cloths contain an appreciable
amount of gold. They are burned, and the ash is sent to the cyanide plant.
Under the rough-and-ready conditions of elementary sluicing, little
attention is given to limitation of size as a variable. A coarse screen is
commonly used at the feed end to remove stones, the undersize from the
screen passing on to the sluice.
Mechanised Sluices
With the decline of the Cornish mining industry the use of devices which
incorporate mechanised movements has lessened, but a brief description of the
buddle is given since it illustrates certain important principles in pulp hand-
ling where gravitational force is used. In the convex round budd Ie (Fig. 159)
the pulp is distributed radially from the centre head which is usually of
concrete and about 5' in diameter, and drops some 12" to lY' on to an annular
cement bed, the overall diameter being about 15'. As the pulp spreads the
carrying film becomes thinner and moves more quietly, so that the particles in
it, though of subsieve size, are arrested and begin to build up into a layer of
mineral, the heaviest nearest the centre and the lightest out towards the
periphery, or even overflowing. Channelling is avoided by the light drag of
brushes which rotate slowly over the settling material. The word "buddle"
is believed to be derived from the building up of the ore bed as flow proceeds.
Periodically, feed is diverted to a duplicate buddle and the bed is sampled, the
products either being retained for further upgrading or rejected.
336 Mineral Processing-Separation in Streaming Currents
In the concave budd Ie (Fig. 160), feed is peripheral and material not settling
immediately is flushed away as the pulp-stream accelerates towards the centre.
A modern improvement on the concave budd Ie uses the same peripheral feed
and brushing gear. Instead of allowing build-up and periodic removal it is.
enriched pulp is delivered to the next machine in the series. The principle is
adapted from the Cornish round frame, a circular convex table made of wood,
which is rotated slowly and from which tailings and concentrates are removed
by separate channels.
One drawback to the use of sluices for treating fine sands is the speed with
which tightly packed beds are formed. These beds should not be smoothed by
raking, as the force used would interfere with the gentle separating action
needed. As with the sluicing of coarser sands, efficient trapping of the heavy
particles calls for even flow of pulp over loosely settled sands, with no channel-
ling to introduce scour or reduce the working width of the separating plane.
338 Mineral Processing-Separation ill Streaming Currents
For the first minute or two after flow of pulp has commenced a correctly
sloped plane forms a level bed of good trapping consistency. Mechanised
sluicing systems have been developed which exploit this, by feeding for a short
period and then flushing the settled material into a special launder for further
upgrading. In the old Cornish tin-dressing plants a large number of "ragging
frames" were used to catch "slime tin" in this way. These were automatic in
action. The feed carne gently on to a sloping wood table for a few minutes,
and at about the time when the bed was packing too tightly, water-actuated
balance weights tipped the table sharply and flushed its load into a concentrate
launder. Considerable space was needed in order to feed these tables at the
gentle velocities required for separation of particles, most of which settled in
accordance with Stokes' law. At a Canadian mill the rag-frame principle has
been rediscovered and developed in a space-saving form, for the recovery of
cassiterite from a lead-zinc tailing. This Buckman tilting table has since
been used successfully for other heavy minerals. Feed is distributed over
five hinged decks (Fig. 161) and after a short time is switched to a parallel unit.
The loaded decks are tilted sharply backward and washed with spray water
which removes the concentrate. The cycle is then automatically repeated.
Embossed rubber matting is favoured as the mineral-catching surface on these
tables. The vital point for success is a timing sequence in which the period of
sluicing is ended before the layer of settled material has packed down hard or
become channelled. If under this operating condition only a low ratio of
enrichment is advisable, the concentrate can be given further upgrading by
running two or more groups of tables in series.
Details of the operation of a test installation have been described by
Chaston. 2
Strakes and open tables are open to pilferage when used to catch gold.
In some African plants their place has been taken by the Johnson concen-
trator (Fig. 162). This is a 12' steel cylinder of 3' diameter, lined circum-
In le ! E:ld V.ew
shell rises, this mineral is washed off by a spray into an axial launder.
In coal cleaning by means of "trough washers" a stream of water carries the
raw coal down an incline. The lightest fraction stratifies at the top, the dirt
below and the middling forms an intermediate layer. Mobility of the sluiced
stream increases from the bottom and sides inward and the coal, being least
impeded, travels fastest. Shape factor aids the differentiation since shale is
tabular and slides quietly without projecting into the stream. Four types of
separator exploit these possibilities:
In the Rheolaveur the feed runs down a steep incline and acquires kinetic
energy as it commences to stratify. The trough then flattens abruptly to a
gentle slope, and in the stirring and dissipation of kinetic energy stratification
is completed. The dirt stratum is removed through apertures in the bottom
of the trough, guarded by a rising current of water (Fig. 163). Middlings are
recycled, and washed coal runs on and out.
Conical
Orifice
Fig. 163. Rheo/aveur Washer
These sluices are very simple. Light sheet metal, rubber coated against
wear, is adequate for construction. They are cheap to buy and to run, and
take but little space. Pullar 4 in reporting on their growing use in Australia,
mentions operating difficulties which stem from fluctuations in pulp density
or grade of feed. He also describes a type of sluice in which a transverse slot
of adjustable width with a movable "tongue" is used for splitting. An
element of flotation has been introduced in the treatment of flake graphite
and of rock phosphate. Suitable flotation reagents are added to the feed
and air is blown in at the head of the sluice. The splitter then works in a
water layer between the sunk fraction and the floating mineral/s.
Fan
Lon 9 I t u d i n a I !. e c tion
[
Splitter!.
II
PI an
Sand f rom p i t
~
Trash screen _ Root!.. nodules etc.
L----Tail M ld .-----i
Hou SuI'PQltr
IIH?l1or~.11
- 10··0·'---- - - - -"'1
,I
,,
I
I
I
I
I
I
I
I
I
I I
I I
rAILINC IMIOOLINC: CONCENTRATE
Vanners
Thus far separation above quiet planes has been considered. In the
vanner an endless belt with flanged edges moves slowly up a slightly inclined
plane, and over pulleys at the ends (Fig. 169). Feed is distributed gently in a
thin layer across the belt, and wash water runs down from above. As it
travels, the whole belt system is given a rapid shaking from side to side. This
keeps the feed spread and prevents channelling, and it helps the heavy mineral
to stratify down through the light gangue. The latter is washed down-slope
by wash water adjusted to a suitable velocity for the purpose, while the
heavier concentrate continues to cling to the rubber surface of the belt, until
it is carried over and washed into the concentrate hopper of the vanner.
Various oscillatory movements have been incorporated in the vanner, of
which those in the Frue and the Isbell are the most widely used. A few
vanners survive to this day in tin concentrators. They treat slimes too fine to
respond to ordinary tabling methods, and produce an impure concentrate
with a low recovery (SO~-;; or less). In the treatment of "tin slimes", as finely
ground cassiterite is usually called, the vanner lingers where low through-put,
moderate working costs, and ample space are to be had .
Mineral Processing-Separation in Streaming Currents 345
The feed to the vanner is first passed through settling cones or hydraulic
classifiers. It is therefore not an equal-sized but an equal-settled material,
comprising relatively coarse light gangue and relatively fine heavy concen-
trate. This discriminating sorting effect, produced by proper use of classifica-
tion of the feed, is important in the efficient running of shaking-bed sluicing
devices. Equal settling is valuable in simple sluicing, but cannot be used
with full effect unless the bed is sufficiently disturbed to permit the fine
Wash
Wate, Fe~
),~
~~o_o __o__o_o--,q
I Concentrate I Tailing
Fig. 169. Frue Vanner (after Truscott)
particles to burrow down, thus forcing the coarse ones up to where they meet
the full weight of the sluice water.
The difference between feed preparation for jigging and tabling is now
clear. Before jigging the feed is screened, the size variable is removed and the
density variable developed. Before tabling, which to an important degree
includes simple sluicing, the feed is hydraulically sorted into equal settling
fractions, the mass variable is removed and the size variable developed.
Stratification on the bed completes this phase of preparation for separatory
action.
The Frue vanner has an adjustable amplitude of side shake. This is varied
from I" when treating very fine cassiterite (say minus 270 mesh) up to 2" for
sand (at -30 mesh), at 200 shakes/minute.
Shaking Tables
The shaking table is widely used for the gravity separation of sands too
fine to treat by jigging. The physical principles utilised in tabling must be
understood if preparation of feed and application of control are to be efficient.
Consider a number of spheres rolling down a slightly tilted plane under the
urging influence of a flowing film of water. Some of the spheres (shaded) in
Fig. 170 represent heavy mineral and others (white) light gangue. The
largest sphere travels fastest and the smallest one slowest, under the com-
bined influence of streaming action and gravitational pull. Of two spheres
having the same density, the larger moves faster. Of two having the same
346 Mineral Processing-Separation in Streaming Currents
diameter, if the slope is relatively gentle and the hydraulic urge relatively
strong, the lighter sphere travels faster. If during the otherwise free down-
ward travel of these spheres the whole plane is moved sideways, then the
horizontal displacement of the spheres varies in accordance with the length
F••d 80x
__--~ Direction 01
Impulu
I
Middling Mlntrnl
of time they take to roll down. This is represented in Fig. 170, which shows
that the largest light sphere has undergone the least horizontal displacement
because it traveiled fastest, whilst the smallest heavy one has been carried
furthest to one side. From this it is seen that if a suitable displacing move-
ment can be applied to a plane, the feed can be spread into bands according
to the size and density of its constituent particles. If these bands are col-
lected into separate vessels as they leave this deck, the feed will have been
segregated into three main products:
break the grip between particle and deck. The deck accelerates, and in
doing so imparts kinetic energy to the material on it. Then the deck motion
is abruptly reversed so that it is snatched away from under the particles
resting immediately above it. These continue to skid sideways (across the
flow) until their kinetic energy has been exhausted. It is therefore essential
to provide a differential side-shake which builds up gently and then breaks
contact between deck and load. This is provided by the shaking mechanism
or head motion of the table. The slower the particle travels downstream, the
further it slides sideways under the influence of the shaking motion.
Thus far discussion has been limited to a series of individual particles fed to
the deck from one starting-point. If, instead, a layer several particles deep
is fed from a starting-line, it becomes possible to handle a greatly increased
load on the deck. The operating conditions have now changed. In the
cross-section through such a layer (Fig. 171), as seen normal to the direction
of shake, the mixed feed first stratifies itself under the disturbing influence of
the shaking action. The smallest and heaviest particles reach the deck,
the largest and lightest stay uppermost, with a mixture of large heavy and
small light grains between. This arrangement exposes the large, light
particles to the maximum sluicing force of the film of water as it streams
down the table. a force that can be controlled in intensity by varying the
volume of water used and the slope of the deck. It is thus possible to exert
some degree of skimming action to accelerate the downward movement of
the uppermost layer without disturbing those below.
The particles next to the deck are pressed to it by the material above,
and therefore can grip it with greater firmness than would be given by their
own unaided weight. They thus are able to cling during fast sideways ac-
celeration, and are only freed and set skidding by the sudden reverse action.
~ Direction of Progress
Final Stratifying
Fig. 171. Progressive Stratification along Riffle
The overlying particles have only a precarious hold. This aids the discrimi-
nating action of each stroke. The bottom particle travels furthest, breaks
free at stroke reversal and is the first to skid. Those above it sway backward
and forward and consequently receive less lateral movement. This accen-
tuates the separating action by giving the bottom (heavy mineral) particles
the maximum horizontal displacement per stroke and the upper (light gangue)
grains the least. This aids· the sorting discrimination. If the feed has been
properly prepared by hydraulic classification, ensuring that all the grains
have similar settling characteristics through vertical currents, film sizing can
now take advantage of the variation in cross-section between the heavy and
348 Mineral Processing-Separation in Streaming Currents
light particles in each stratum, sweeping down the lighter and leaving the
heavier untouched. The particles thus segregated are then removed in
separately discharged fractions, called bands, at the far end of the table's
deck.
It would not be possible to form and maintain an evenly distributed
thick bed of the kind called for by the foregoing considerations if a smooth
plane deck were used. Riffles are therefore employed to provide protected
pockets in which stratification can take place. They are usually straight and
parallel with the direction of shake, but may be curved or slanted. The
deck, instead of being plane, may be formed to provide pools in which the
feed can stratify (see Fig. 152). The riffles must
Thus (a) rules out as bad practice the use of "stopping" riffles set high above
the rest, sometimes used to arrest and spread entering feed. If all riffles are
not of similar initial height the stratifying action and transfer between them
is upset. Smooth delivery is best achieved with a feed box integral with the
moving deck, and aligned with the vibrator. It should let the feed down
gently to the head riffles. Items (b), (c), and (d) are arguments against the
use of curved riffles, which increase wall friction and upset stratifying action.
A badly maintained mechanical action and deck coupling may mislead the
engineer into re-designing his riffle plan, just as an incorrect stance may cause
the unwary golfer to modify his swing instead of standing correctly.
In the standard Wilfley table (Fig. 172) the riffles run parallel with the
long axis, and are tapered from a maximum height on the feed side (nearest
the shaking mechanism) till they die out near the opposite side, part of which
Oi,',ibutlng BOK
6/1.
. c=:J "
5/1.
: c=:J
is left smooth. Where the riffles stand high, a certain amount of eddying
movement occurs, aiding the stratification and jigging action in the riffle
troughs. As the load of material is jerked across the table, the uppermost
Mineral Processing-Separation in Streaming Currents 349
Applied Principles
The principles which underlie separating action, together with the methods
by which the combined energy of the sluicing water and the mechanical shaking
can best be exploited, may now be summarised.
(a) Feed must be in a size range chosen with regard to the particle size,
density and shape of each mineral being separated.
350 Mineral Processing-Separation in Streaminl? Currents
Mechanical Details
Shaking action on the table is controlled through an adjustable head motion,
which converts the rotary drive of the motor into a differentiated horizontal
stroke. In the Wilfley table (Fig. 173), one toggle (B) is seated against a fixed
mounting (C) which is bolted to the foundations. The other toggle (A)
bears on a yoke (D) which is connected to the table deck. The toggle sys-
tem is held together by the spring (E), and driven by the eccentric and pitman
(P). At the beginning of the forward stroke the toggles are at their flattest
and the spring in maximum compression. As the pitman rises, the toggles
steepen out. At the top of its stroke they are at their most acute angle and the
table has reached its maximum speed. The pitman now descends, flattening
the toggle angle and abruptly reversing the direction of the table. The abrupt
reversal occurs at the end of the forward thrust of the deck, and at this moment
the particles resting upon this deck skid forward. The decompression
of the spring aids the forward stroke to some extent. This spring is adjusted
until its increased compression just quietens the knocking sound of too loose
a set-up. For coarse feeds the Wilfley table is run at 240 full strokes/minute and
a stroke length < 1". For finer sands speed can be raised (by changing pul-
leys) to 300 r.p.m., but the stroke must be shortened. This adjustment is
made by means of the handwheel (G) which moves the sliding piece (F)
(Fig. 173). The toggle (B) in its seat (C) can thus be raised or lowered,
shortening the stroke (deck movement) from the full I" down to a minimum
of f'. In operation the table uses t h.p. but may have a starting torque of
twice this amount.
Various types of head motion are found on James, Deister, and other
tables. The Fraser & Chalmers table is moved by an electro-magnetic
motion (Sherwen).
Various materials of construction are available for the deck and its riffles.
A linoleum deck with hardwood riffles secured to it by copper tacks, is most
commonly used. Rubber decks with rubber riffles have a good working
life and freedom from leakage, but must be protected from full sunlight.
The rubber is bonded to the softwood underdeck with adhesive. On slimes
tables, brass riffles are used for the shallow section, above linoleum. Some
use of aluminium decking with metal riffles has been made. Where a large
tabling area is required, double-decked tables have been successfully employed,
suitable strengthening of the structure and increase of driving power being
provided.
Triple-deck tables are also manufactured. Lightness of the deck,
which is shaken 15,000 or more times hourly, has a direct bearing on working
life. In laboratory tests, the honeycomb aluminium sheet used in aeroplane
construction is excellent, though costly.
Some of the many riffling plans are shown in Fig. 175. The variations in
feed preparation from one mill to another are such that sound judgment
regarding the relative merits of the systems is difficult. Broadly, a riffle
plan in which the feed box moves with the deck and lets down the pulp smoothly
on to the top row of riffles is best. These top riffles should be high enough
to check any tendency of the newly arriving feed to scour its way down the
table without having time to settle and to begin its stratification in the sluicing
valleys between the riffles. Some operators check this tendency of the pulp
to rush down-slope by making every other riffle (or fourth one) at the feed
end twice the height of those between as in Fig. 175 (d). The desired effect
can usually be obtained more easily by careful control of the classifying and
Mineral Processing- Separation in Streaming Currents 353
feeding arrangements, thus ensuring that all the riffles are equally loaded
instead of overloading part of the deck at the expense of the low riffles.
Again, some operators consider that in place of the standard riffle (a)
which runs parallel to the shaking motion of the deck, efficiency is gained by
forcing the pulp to climb slightly on its way across. If the true function of the
sluicing space between the riffles is carefully helped by good feed preparation,
and by proper adjustment of deck tilt and wash-water the author considers
that deviations from straight-through sluicing have little to justify the com-
plication, wear and loss of sorting capacity they introduce. If, for any
over-riding reason it is not possible to provide a steady rate of feed at a
consistent solid-liquid ratio, of correctly classified sands, then deviations from
the standard riffle plan (Fig. 175 (a)) acquire some rather doubtful justification.
On the whole, however, an expert application of basic principles is better
than such compensating action.
The table deck is supported on wood carried on steel framing and a running
adjustment is provided for altering the tilt. In some tables a truss mounting
is used instead of rocker action. With this trussing, the whole deck rises
very slightly during the forward stroke and this may have a slight tossing
effect upon the load of sand. It is important for good efficiency that a table
be set so that its trusses are accurately aligned with the vibrator, that they
move in a vertical are, and above all that they do not reach a perpendicular
354 Mineral Processillg--Separation ill Streaming Currents
Table Controls
These are of three kinds. The first are:
(a) The type of riffling used;
(b) the material of which it and the deck surface are made;
(c) the acceleration and deceleration during one revolution of the drive;
(d) the mode of presentation of the feed.
These are major points of design, and can only be varied by reconstructIon.
They should therefore be decided by tests made on the material which is to be
treated, before the table is installed.
Mineral Processing-Separation in Streaming Currents 355
Next come running speed and stroke amplitude. Speed depends on motor
and pulley system. Stroke is adjusted by varying the toggle spread, or in the
Sherwen, by altering the amplitude of the solenoid vibration. With most
tables, stroke can be varied while running. It is desirable that these adjust-
ments should be settled and fixed as early as possible in a working plant, as
constant "fiddling" with the controls leads to inefficiency and working trouble.
If a small alteration is needed, it is easily made at any time during the working
life of the table.
Third there are the operating controls which must be made by the shiftsman
as the need arises. These are:
(a) The tilt across the table.
(b) The solid-liquid ratio in the feed distributing box.
(c) The wash water.
(d) The position of the product cutters.
The interplay of these controls will be understood if the operation of a
typical shaking table in a plant is considered.
Assume that the table is operating at half-tilt (for a Wilfley, which can be
adjusted from flatness to I "/foot, this is Y' per foot run from feed to discharge
side) and that the bands are spread as shown in Fig. 176.
Wash AWater Feed Zone
____
A~ ~
+--
- ' - - - - --; Oirectlon ur
__________--1 movement
- - - ----;along deck
is the stroke used. Somewhere below the -150 to 200 mesh sizes decks
with a different catchment plan are used. A typical slimes table as these are
called (Figs. 177/178) has a series of planes rather than riffles, on its linoleum-
covered deck. Plane A (Fig. 177) is the major settling section, some ten feet
along and four feet down the table. Here the finest particles settle and stratify.
There is no check of flow till the area B-B is reached, the settled rough con-
centrate being pushed toward the discharge from here. Gangue overflows to
plane C, where middlings are held and sent progressively to D, E, and F
for a repetition of the original action. The concentrate and/or middlings
produced on the slimes table may require further treatment on buddies or in
kieves before shipment.
One possible variation which should be mentioned must be guarded against.
If water used in treatment is caught and recirculated, it may become con-
taminated with colloidal iron and aluminium hydrates. These could, by
coating the deck and the particles with a tenuous layer of slippery slime,
upset the frictional grip of deck on particles, and also the stratification betweeen
riffles.
Slimes Treatment
Here the important particle' size is that lying between true Newtonian
and Stokesian movement, in the zone covered by Allen's equation
where p is the resistance of the fluid to the motion of the solid, K a constant
modified by shape and velocity of the particle relative to the fluid, r the radius
of the equivalent sphere, n a coefficient of velocity v, p the fluid density and
jA the kinematic velocity. Practice varies considerably as regards rate of
feed and range of size fed to a given type of concentrating appliance. Table
25 summarises a cross-section representative of reported practice.
TABLE 25
Average Ib.lft. 2 Particle Types of
per hour of feed size-range Concentrate
(a) Newtonian 60 20-200 mesh Shaking tables
(b) Allen 30 150-400 mesh Slime tables, vanners,
buddIes, strakes
(c) Stokesian down to 7 270- 30jA As (b)
Sampling Check
Although this is concerned with tabling control, it was not discussed in that
section, since it requires laboratory tests not possible in shift running. Where
a major tabling operation justifies the cost of sampling check, it becomes
important to have detailed information of the rate of change in concentration
along the discharge edges. Samples may be collected for a timed period in a
trough divided every few inches into compartments, a small head sample
being taken at the same time. Scrutiny, size analysis and assay then present
a picture of the value-change along the discharge area, and of the quantity
represented at the prevailing discharge-rate. It is not uncommon for the
most serious loss of readily recoverable value to be disclosed as coming from
feed running straight across the table from entry to the tailing discharge nearly
opposite, because it has not been properly arrested and bedded into the riffles.
Poor classification of the feed is also disclosed, when it is seen that material
which cannot be handled at the setting appropriate to the main-size range of
the feed is present. Various further improvements become possible once the
technical supervisor has reliable facts to work from. Cutters can be positioned
better, and the amount of middling best retained for re-circulation brought
under control.
Tabling-General
The maximum particle in the feed must not be thicker than the riffling or
more than one-third of a riffle in width. If tables are used in series, the first
group handling a given grade is set to produce a low-grade concentrate
containing all recoverable values, and a stripped gangue for discard. This is
called roughing and the low-grade concentrate is then passed to a smaller
group of tables for finishing, or cleaning, the tailings from these tables being
given further treatment, possibly by returning them to the rougher feed, as in
Fig. 179.
This is called a "rougher-cleaner" operation. Roughing tables have for
Mineral Processing--Separatiol1 in Streaming Currents 359
their primary task the rejection of barren tailing and operation can, at need, be
simplified to two-product work (a rejected gangue and retained low-grade
values) . This last will then go on to cleaning tables to be separated into two
bands-a final concentrate and a middling requiring further treatment.
Tables worked thus require less skilled attendance than where band-wander
in a three-band separation must be checked, since most of the gangue is
removed at the start, and operation such as that shown in Fig. 179 is steadied
by a wide-spread return middlings band, particularly if re-grinding is per-
formed between cleaners and classified re-distribution (not shown in diagram.)
With single-pass treatment any change in ore texture or richness calls for
Finished
Concentrate
Tables set for roughing work are fed more heavily than when a complete
operation is carried out in one pass since the operator has only to make sure
that one product (usually the tailing) is finished efficiently. More water is
used, and a greater tilt and stroke length, thus moving the sand vigorously
along and across the deck.
A further development usuiiIly found in tabling such ores as cassiterite is the
grouping of the film-sizing appliances. each group being fed with ore which
has been graded (probably through a series of hydraulic classifiers), in some
such manner as in Fig. 180.
Each group of tables in such an arrangement can either be operated in
parallel (one-pass treatment) or series (rougher-cleaner).
A further refinement may be practised with the coarse tabling group.
In order to avoid excessive production of "slime tin" (a cassiterite so finely
comminuted that much of it is lost with the tailings). the feed may be ground
and tabled in stages. Sliming loss occurs when fine particles settle very
slowly and are washed over the table without being trapped. It is therefore
sometimes advantageous to remove liberated concentrate early in the treat-
ment, using some such form of stage concentration as shown in Fig. 181.
Mineral Processing-Separation in Streaming Currents 361
1I Hydnmlk Wale,
Overflow to
rl------ Settling Cone - - - - - - - - ,
(+) (1)
t
Slimes Tables
t
Vanners
or Buddies
Fig. 180. Classification before Tabling
The coarser the feed to a table, the longer and slower should be the stroke.
One operator can take care of between ten and one hundred tables. Tables
used for roughing should have more riffling than those used for finishing the
concentration. They are required to hold the feed in teetering columns
between their riffles while the motion causes the particles to be jigged and
stratified. The deck of the table must be watertight.
Riffles are made of hardwood such as oak, of soft pine, or of plastics, e.g.
r
rubber. The maximum particle size usually treated by tabling is and sands
as fine as 300 mesh are handled. Frue vanners are given even finer feeds of
tin-bearing pulp, there being no simple alternative to gravity concentration
for separating this mineral.
The general tendency in modern mills where tables play an important part
is to "rough" a run-of-mill ore pulp which has been released from the grind-
ing circuit at a controlled upper size limit. The tailings from this roughing
are discarded, and the concentrates are classified and then given a cleaning
treatment.
Beside their use on cassiterite, tables are employed to treat free-milling gold
ore, auriferous pyrite, barytes, fluorspar, foundry dross, coal "smalls", etc.
A single miniature table is sometimes used to receive part of the tailings from
some other concentrating process (e.g. froth flotation) to aid the shiftsman by
showing him what minerals are leaving the circuit. Because tabling depends
on a good gravity difference between the minerals being separated, care must be
taken to avoid any unnecessary grinding of values which are sufficiently
362 Mineral Processing-Separation in Streaming Currents
1
______ Screen _ _ _ _ _--,
1
~
I (20 mesh)
.---J
(+) ( )
J
Hydraulic Classifiers
Coarse
1 Medium
1 1 Fine
T.bJ.s T.bJ.s T.blea
r Middling Tails
Tables make a middling product which, as has been noted, may contain
true tails and true concentrates. If this middling is recirculated, it grows in
volume, and the time arrives when it crowds over to join one of the other
products. Simple recirc.Ilation is not, therefore, the best way to handle a true
middling. If it consists of incompletely unlocked particles, part gangue and
part concentrate, these can only be freed by further grinding.
If the middling is partly composed of some mineral or minerals of inter-
mediate density, it should either be cleaned up and marketed, stockpiled, or
sent to waste. This involves separate retreatment of the middling, not simple
recirculation to the feeding head of the table producing the middling.
It is rarely feasible to make a high ratio of concentration in a single pass
without paying the penalty either in the form of dirty concentrate or of tailing
losses.
The ratio of concentration R is expressed in the simplest forms as weight of
feed F divided by weight of concentrate C. In cases where it is not convenient
to weigh feed and concentrate, a formula based on the assay grades of feed
Mineral Processing-Separation ill Streaming Currents 363
Dry Tabling
However carefully water used in gravity concentration is conserved, any
quantity up to fifteen times the weight of material undergoing treatment may
be in circulation. This water becomes foul with use, and if slimy, upsets
frictional grip of the sand to the deck.
Methods of separation based on gravity use least water in dense-medium
separation and most in the treatment of very fine slimes. They can only be
used where a plentiful supply of unfrozen water is available during the
operating season.
Sometimes the product would be injured by wetting, or may contain soluble
salts which preclude water treatment. Again, the cost of drying a wet
concentrate may be prohibitive. In such circumstances the possibility of dry
jigging, "dry-blowing", and dry tabling can be considered. In the pneumatic
jig air takes the place of water as the pulsating medium used to dilate the bed
of ore and promote stratification. Various forms of jig have been developed
for desert use.
364 Mineral Processing-Separation in Streaming Currents
Pneumatic tables use a throwing motion to move the feed along a flattish
riffled deck, and blow air continuously up through a porous bed. The general
principles of separation are similar to those applied in wet tabling. Use is
limited, and the recorded data of performance are somewhat confusing. 5 A
few coal plants use pneumatic methods, which have the advantage of pro-
ducing a dry product. The upgrading of asbestos by pneumatic tabling is
another important commercial application.
Hudson 6 has observed that whereas in wet tabling the particle size increases
and the mineral density decreases from the top of the concentrates band toward
the tailing (in the absence of marked shape differences), on an air table both
particle size and mineral density decrease from the top down, the coarsest
particles in the middlings band having the lowest density. Thus, air tabling is
similar in effect to hydraulic classification (the air blown through the porous
deck exerting an effect similar to that of the rising water in the classifier),
while wet tabling does not develop such an effect. This has, of course, been
developed by pre-classification of the feed in the case of the wet table. Logi-
cally, in air tabling a similar pre-treatment in reverse is indicated (sluicing
before tabling). In experimental work Hudson did this by sizing the feed
through a launder. The dry sands stratified during descent to the table and
fanned out as they fell from the launder, the coarser particles being thrown
further horizontally during fall than the finer ones.
In dry blowing, a horizontal air current is used to displace particles falling
vertically, each of which is displaced laterally in accordance with its rate of
drop.
A recent development of the Lavodune principles uses air instead of water
as the transporting and dune-forming agency.
Feed
~
,coarse rUSherS
(+) , Sereen(16#)----+(-)=:J
~
Series of rommelS j
Med. Fine
U.S. U.S.
! !
r- r-
Cone.
Jig
Cone.
Jig
Cone. _...J...--4---I----II...--~I...--+---I
Mid.
Tail_-------J.--------'--------l
I Hyd:aulie
Clasrfiers j
Coarse Med. Fine
Product Product Product
~ ! !
iTar
Shaking
Cone. Mid.
e
rTat
Shaking
Cone. Mid.
e
Shaking
rTable
Cone.
Tails _-----...L....-------I.-----...J
=
Fig. 182. Jigs followed by Tables
366 Mineral Processing-Separation in Streaming Currents
Fig. 182 gives the flow-sheet for the combined treatment of a sulphide
copper ore using jigs and tables. The main minerals are chalcocite, bornite,
and a siliceous gangue. For a number of years students at the Royal School
of Mines have worked parcels of this ore through the old pilot plant, and have
obtained excellent recovery of high-grade products, mostly in the jigging
section. This flow-sheet exemplifies the value of stage crushing with stage
concentration for material breaking free at a coarse enough mesh to permit
jigging.
Fig. 183 shows the series up-grading arrangement of groups of Humphreys
spirals.
Fig. 184 illustrates a Cornish sand treatment plant, in which rehandling
must be unusually thorough because an important part of the value resides
in the slimes, and there is no accepted alternative to gentle gravity separation
for its treatment.
Feed
1--(+)""
1
Screen ~ (-) I
Grind
j
Feed Tank
t
(T) ... t
Roug h·109 Splrals
. .(M;d.)]
I
Scavenging
1
(Cone.)
Spirals
1 1
Cleaning Spirals ~ (Tail)
Tail
--
1
(Cone.)
t.
Dewatenng > Water
--
t
Products
Feed
(a beach sand)
1
Reject _ ( + ) - Screen (20#)-- (-)
= ~
Conc ....- - - - Roughing Table --,-_- Tail
1
(Mid.)
!
Ball Mill
1
HYd .. UI] CI...ifi",
(Slimes)
1
(Mid.)
Roughing
BuddIes - - - - - - - f - + T a i l
I -
(Tail) ...4----Cleaning
BuddIes
1
Conc.-..------ Slimes Tables ---(Tail)
I
Ball Mill
References
1. Pryor, E. J. (1963). Dictionary 01 Mineral Technology, Mining Publications.
2. Chaston, I. R. M. (1961/62). Trans.I.M.M. (Lond.), 71.
3. Stewart, A. L. (1961). The Mining Magazine, Sept.
4. Pullar, S. S. (1963). Proc. Aust. I.M.M., March.
5. Knapp, E. A. (1953). Recent Developments in Mineral Dressing, I.M.M.
6. Hudson, S. B. (1962). Aust. I.M.M., Dec.
7. Knapp, E. A., and Sweet, C. T. (1953). Trans. I.M.M. (Lond.), I.M.M.
8. Conta del Fa, M., and Ferrara, G. F. (1963). I.M.P.e. (Cannes), Pergamon.
CHAPTER 15
Introductory
During the past few decades a radical change in the approach to inorganic
chemistry has occurred. This is due to an increasing knowledge of the
valences, structure, and bond strength of crystalline compounds. The forces
which integrate and largely determine shape, behaviour, and the energy used
in producing a change of compounded structure have been worked out for
many substances by precise mathematical methods. Physical and inorganic
chemistry overlap.
The work of such pioneers as Sidgwick 1 and Pauling 2 brought new concepts
of structural order to the chemical speculations of earlier workers who had not
the aid of radiography and spectrography which are today in common use in
research. The nature of a chemical bond can now be classified on considera-
tions of valence, ionic, covalent, and other forms of coupling, strength,
molecular shape, and crystal packing. The thermodynamics of reaction,
synthesis, and dissociation, the rate of crystal growth or degeneration, and
the chemical reactions possible in a given system are governed by forces
assessable in mathematical terms.
The segregation of valuable minerals into ore bodies during geological
aeons does not finish with the production of neat and formal end-products.
A metal will balance its hydrate, silicate, carbonate, and sulphide radicals in
various ways, and intruding ions of other elements have time, opportunity, and
sufficient directing energy to distort the lattice structures of the almost in-
soluble compounds that go to make up an ore complex. A general name for a
given mineral species has not, therefore, the same identifying significance as
would apply in the case of a fairly pure chemical compound. Classification
by name is not a guarantee of identical behaviour of, for example, galena
drawn from different deposits. Chemical analysis is helpful in the study of a
mineral, but is only a starting-point in the investigation of its nature. The
distribution of the elements of which it is mainly composed gives some guide
to its bond strength, but this is modified by any non-conforming ions in the
lattice and by the regularity or random dispersion of their distribution.
The particles produced by comminution have been severed from a rock
formation which for millions of years has been reducing its potential and free
energy to a mmmlUm. Grinding has given these particles a new stimulus
and disturbed their structural bonds. Concentrating processes can be used
to exploit the energy diffe'rences between the various mineral species thus
liberated. At the fundamental level, valency theory shows where bonds are
weak or strong, and this suggests promising lines of attack where changes of
state must be induced as a prelude to concentration or rejection. Among the
370 Mineral Processing-Physics and Chemistry in Ore Treatment
oxygen or any other gas molecules in the air to a reacting interphase. Since it
conveys all its molecules, the possibility of undesired ones could exist in
special cases (e.g. an atmosphere heavily charged with sulphur dioxide from
an adjacent smelter). Second, when air is blown through a pulp under suit-
able conditions it acts as a general stirring and transporting agent and per-
haps by picking up particles into the walls of individual bubbles, as a selective
transporting agent for mineral specifically attracted into the air-liquid
interface.
The liquid of dominant interest in mineral dressing is water. In the
research and testing laboratory this may be distilled or deionised water, to
which controlled additions of reagents are made. In the plant it is the
natural water of the local catchment area, mine water, river water, or piped
water from a distance. Normally it is used as received, but for some pro-
cesses it must be partially "softened" or otherwise prepared. Physically, a
liquid can be defined as a fluid which flows readily in such a way as to seek the
lowest available level and there moulds its underside to the shape of the con-
taining vessel and its surface to the mean curvature of the terrestrial globe.
A liquid differs from a gas because its molecules are bound together more
firmly by forces of adhesion, so that its particles settle in a closed vessel at
normal temperatures and do not distribute themselves evenly through the
container. If heat were applied to water in such a closed vessel, the liquid
molecules could be stimulated to speeds at which their group-bonds were
destroyed, after which the entire contents of the vessel would become gaseous.
Hence, water is the liquid form of a substance which can exist in more than
one state or phase, according to the temperature and pressure to which it is
subjected.
Since water is not the only liquid used in flotation, a fuller definition of the
liquid state is needed. Liquids occur in three forms. Associated liquids of
polar type have polar molecules, which seek to form oriented groups in accord-
ance with their polarity.
A polar molecule is one which has an electric moment, its positive and
negative charges being disposed at opposite ends. Its polar bond is the
electrostatic union of atoms due to the transfer of one or more electrons, as
with sodium chloride
..
(I) Na. - - . e!: (outer electron shells represented by dots).
(2) Na: <;:"1: (el octet completed by transfer of one electron).
(3) Na-:-Cl- (resulting electrostatic union and disposition of charge).
A polar molecule is chemically active and either soluble in water or hydro-
philic. A polar group is a portion of a molecule which has polar character-
istics.
In a non-polar molecule the electro-magnetic union is produced by the
sharing of electrons, as with carbon dioxide:
(4) : ~ : __ . ¢ . __ " ~ ..
(5) .. q: e: q:
(6) o=e=o (a covalent bond).
Mineral Processing~Physics and Chemistry in Ore Treatment 373
'~II*
"j'i'"1'ii:4:4:
~!:'Y-:I.:l.'J:l:
"",;-,:,',,',:.,:
11111'1 111111 -;:':-;"-;"""4:':
;: ",:4:4'1:,:
HUG,.. AbTwll WCIUr'
Nonpolar or 8emipolar or
bennne
Mr'7I'cU L, ."""""w.
aloohol
L.
Polar or
GI'oeicUd L,
alyool
carbon tetrachloride aJycerol liq, ammonia
hydrocarbon oila methylal water
ether
acetone
chloroform
Fig. 186. Normal, Semi-Polar and Associatt.'d Liquids (after Hackh)
Mineral Processing-Physics and Chemistry in Ore Treatment 375
The dividing line between the liquid and solid state is not always clearly
drawn. As the energy of the molecules in a substance becomes reduced
through loss of heat, they move more and more slowly. The "temperature"
of a molecule indicates the kinetic energy due to its thermal agitation. As this
falls the molecule becomes less able to force its way through the surrounding
molecules, and tends increasingly to become bound into a rigid structure or
lattice. At some stage of cooling, for all short term purposes, the substance
holds a definite shape and is no longer able to spread itself quickly at the lowest
available level. The molecules may become strongly locked in the regular
pattern of a crystal lattice, or there may be a fifer-range of several degrees
between the truly liquid and solid states, as with most commercial fatty
acids. Pitch, though solid over a period of hours, is flowing plastically, as
may be seen by leaving a lump of it for a year and then noting how it has
spread.
The atoms of a substance may be bound ionically along one or more
planes, producing considerable strength in the planes so bound, with weakness
planes when the mode of crystallisation is such that no ionic forces unite them
in other directions (cf. graphite, mica, and asbestos). Such symmetrical
binding can be found both in pure crystal lattices and in solid solutions,
where more than one element or compound has crystallised in an ionic
pattern.
In addition to heat and pressure changes, which can make the same com-
pound occur as gas, liquid, or solid-e.g. steam, water, and ice-solids can be
built up from discrete molecular dimensions to colloidal aggregates and thence,
via micro-crystals, to the macroscopic state. Sometimes single molecules
associate into a colloidal concentration; sometimes they grow into large
molecules (molecular colloids). Colloids can be considered as a stage in the
growth or, if produced by detrition, the breakdown of a substance.
The structure of a substance may be determined partly by its rate of growth
or solidification from the discrete molecular state of sub-colloidal disperse
phase. If barium sulphate is precipitated from a solution of barium chloride
by the addition of sulphuric acid, a flocculent precipitate is produced when the
reacting substances are fairly concentrated. The newly formed molecules
rush together so swiftly that they entangle water in the growing aggregates
and the true micro-crystalline form of the compound is not built up. If
extremely weak solutions are used, the molecules precipitate down upon suit-
able solid nuclei slowly enough to become arranged in a crystalline pattern
ionically bound, and no water is trapped into the precipitate.
A particle may be homogeneous, in which case the atoms of which it is
composed are bound together in a regular order. The pattern thus formed in
a crystal is termed the space lattice, and a straight line drawn through two
adjacent lattice points can be prolonged to pass through successive points
equally spaced. Parallel lines do the same and the crystal structure is refer-
enced to its x, y, and z distribution of such points with regard to a rectangular
tri-axial set of co-ordinates. The order will be repeated throughout such a
particle if it is a complete crystal, or a fragmentary portion of such a crystal.
It can be composed of an intergrowth of true crystals, in which case there
will be abrupt changes in lattice co-ordination at the boundaries between the
376 Mineral Processing-Physics and Chemistry in Ore Treutment
TABLE 26
SOME TWO-PHASED GAS-LIQUID-SOLID SYSTEMS
I. GAS-LIQUID SYSTEMS Mists and fogs a
Foams or froths b
II. GAS-SOLID SYSTEMS Aerosols c
Occlusions or deep sorptions d
Pumices e
III. LIQUID-LIQUID SYSTEMS Phases, Newtonian-normal f
Emulsions, Newtonian-normal g
Phases, normal-normal h
Emulsions, Newtonian-Newtonian i
IV. LIQl;ID-SOLID SYSTEMS Unstable dispersions j
Peptised dispersions k
Flocculations I
Solvating systems m
Slurries n
Gels o
Thixotropic gels p
Moist solids q
V. SOLID-SOLID SYSTEMS Two-component particles r
Multi-component particles s
Solid solutions t
Low dielectric systems u
High dielectric systems v
Zoned systems w
B c ....._0 __
Emulsions
A true emulsion is a dispersion of one liquid in another (usually oil and
water) under conditions in which they are practically immiscible, forming a
fairly stable suspension with little or no solubility of the one in the other.
If the dispersed droplets are of oil the emulsion is oil-in-water or O/W. If oil
forms the continuous phase it is WiO. This two-component system is
thermodynamically unstable and reverts to fully separated phases either
Mineral Processing-Physics and Chemistry in Ore Treatment 381
Floccules
In an emulsion the two-phase system is dominantly liquid-liquid, though
it was observed that a solid interphase could part these liquids. In typical
flocculation the two-phase system is liquid-solid. The liquid of main interest
in mineral processing is aqueous and the solid particles are too small to settle
down with appreciable speed from an unagitated system.
The surface potential of a particle tends to repel it from others similarly
charged. Opposing this, its surface tension promotes cohesion, whereby the
overall solid area of the cohering particles is reduced, and with this the total
surface energy. If repulsive force is dominant these minute particles remain
dispersed (peptised). If on close approach the mutual attraction offered as
reduced tension can overcome this electrical energy barrier they may floccu-
late. Floccules may consist largely of finely divided solids, or of particles
loosely clustered in an aggregate predominantly aqueous-perhaps micellar
in form. Forces disposing the particles to flocculate include magnetism
(used in the dense-media process to manipulate ferro-silicon through its
cleaning process); ionic forces of attraction; and secondary en tropic forces.
Each particle has an unmeasurable but considerable surface tension. This can
be partially satisfied, as has already been observed, by reactions between it and
the aqueous phase of the pulp. Further reduction of surface tension can also
be achieved by adherence to other particles. If a number of particles can
bind together so as to reduce their total free surface, the conditions for
flocculation are propitious. If, however, each particle has an ionised surface
and the bulk of the ions are of the same electric sign in each case, then a
repulsive force will be exerted between approaching particles. Should this
exceed the attraction offered by a reduction of total surface energy, the
particles remain dispersed in the pulp. The most favourable conditions for
flocculation therefore require electric neutrality between particles of similar
composition. This neutral state is called the iso-electric point, or the zero-
potential. It denotes the pH below which a given substance is acid, and
above which it is alkaline. At or near its iso-electric point a colloid coagulates
382 Mineral Processing-Physics and Chemistry in Ore Treatment
while a pulp tends to flocculate. The iso-electric point is also the pH of the
liquid phase at which there is minimum movement into it of ions from the
solid phase, or minimum solubility. Movement of a species from the solid
into solution as an ion increases the residual electrical charge at or near the
solid surface by the same amount, but of opposite sign. Electrostatic tension
is thus set up between ionizing particles and the system remains dispersed.
The adherence of particles of a given species as floccules must not be
confused with the phenomenon of slime-coating. In this an extremely fine
layer (e.g. of lime) adheres to the surface ofa relatively large particle of sulphide
and impedes true surface-reaction.!5 A pulp being prepared for flotation
may need dispersive conditioning to clean the particle surfaces, whereas one
containing the end-products of treatment may require flocculation to settle
and remove solids. In the first case the presence of two mineral species
which have been conditioned to develop strong differences in surface excita-
tion may produce ionic coupling; in the second the particle population in the
pulp is in the same general state of excitation and a tendency to flocculate may
be achieved near the iso-electric point. Modification of pH, to the extent
that it increases surface charge, is a dispersant, but so far as it neutralises
charge is a flocculating agent. The principal dispersing agents used are
weak-acid salts of the alkali metals, such as sodium silicate, carbonate,
sulphide, cyanide, and tri-phosphate. The most used flocculating agent is
lime, others including sulphuric acid, sodium aluminate, and ammonium
chloride. Trivalent ions appear particularly effective in inducing floccula-
tion.
If the floc is considered as a particle the effect of flocculation is to increase
particle size and facilitate sedimentation and filtration. The main stabilising
factor in a sol is its electrical charge, and flocculation is aided by the addition
of an electrolyte capable of discharging the component particles. Subsidiary
forces which influence flocculation include the physical state of the dispersion
medium, the temperature, population density (affecting opportunities for
contact between particles), particle shape and the viscosity of the system.
The Schulze-Hardy rule is that the ion of an electrolyte which causes a sol
to coagulate is of opposite electrical charge to that on the colloid particles,
and that its flocculating power increases greatly with an increase in ionic
valency.
The concentration of particles in the pulp has an important effect on rate of
settlement due to flocculation since collision must precede adhesion, and the
denser the population the higher will be the rate of collision. Once floccules
have formed, they act as a membrane when they drift downward, trapping
particles and filtering the pulp.
Before particles can flocculate from a pulp, they must come into contact.
When particles are already flocculated, some force is needed for their disper-
sion. This is provided by thermal motion. It can be observed with small
particles in a pulp as a vibratory effect, due to their ceaseless bombardment
by moving molecules of water. The energy transferred in each such collision
by a moving molecule or ion is about 6 x 1014 erg at room temperature.
The higher the pulp temperature and the lower its viscosity, the stronger
is this stirring effect. If the particle surfaces are nearly electrically neutral,
Mineral Processing-Physics and Chemistry in Ore Treatment 383
the movement will favour flocculation, as will also be the case if a flocculating
agent having bonding counter-ions has been dissolved in the water. If the
surfaces have similar electric charge but are still in contact, this Brownian
vibration may suffice to shake them apart. Since a very small gap suffices to
reduce the attractive forces across the interphase between surfaces, they will
then tend to remain dispersed. The thermal agitation is impartial, but will
aid any change which decreases the total surface energy in the system. Brown-
ian movement does not directly affect rate of settlement. A "stable" suspen-
sion of colloidal particles in water falls in accordance with Stokes Law unless
it is peptised by electrical charge. Molecular bombardment per se cannot
oppose gravitation, since it is directed at random on all sides of each particle.
The rate of fall may, however, be extremely slow.
Froths
The froths of technical concern in mineral processing are those produced in
the flotation process. A specialised texture is aimed at, in which an unstable
liquid-air foam increases its stability by becoming loaded partly by particles
of select mineral species and partly by the addition of a chemical reagent
(frother). The mineralised froth is a three-phase system and as such is more
closely considered in connexion with the flotation process in Chapter 17. A
foam may be defined as a dominantly gaseous mixture of gas and liquid, in
the form of more-or-less stable bubbles. The liquid phase may be colloidal.
Here again a foam is possibly three-phased in character. Such a system may
be stabilised by soaps which emulsify to form a binding network which en-
closes water in each bubble, or by capillary-active substances.
The work required to produce a foam 9 is the foam surface multiplied by the
surface tension. This foam is thermodynamically unstable since its collapse
is accompanied by a decrease in the total free energy. When a heteropolar
surfactant is dissolved in water its molecules preferentially concentrate at an
air-water interface with their hydrocarbon groups oriented toward the air
phase and their hydrophilic polar groups immersed in the aqueous phase3 .
The surface tension of the liquid decreases toward a minimum value as the
concentration of these molecules rises from zero to the saturation limit in
accordance with the Gibbs adsorption equation
1 oy
r=-- --- (15.3)
RT ologea
where y is the surface tension (dynes/cm.)
R is the gas constant
T is the absolute temperature
a the activity, related to concentration of solution.
The lowering of surface tension on addition of pine oil (a widely used flotation
frother) to pure water is from 72·3 dynes/cm. at N.T.P. to 63 dynes/cm. for
an addition of 0·2 g./1. and 58 for 0·4 g./1. Commercial pine oil contains an
erratic quantity of borneol and fenchyl alcohol (15-20%) and some 50%
of a-terpineol, and has a maximum solubility of 1·28 g./1. at 25°C.
384 Mineral Processing-Physics and Chemistry in Ore Treatment
For all pure liquids. including water. the life of an emerging bubble is about
0·01 second. Addition of any soluble substance increases this duration. A
highly stable bubble (e.g. one produced from water containing saponin) is not
usually able to pick up mineral particles.
Sutherland and Wark 18 observe "a correlation between the size of bubble
in the froth and the concentration of the frother. In very dilute solutions the
most stable bubbles are large, but as the concentration of the frother is
increased the average size of the stable bubble diminishes." They also dis-
cuss the bearing on froth stability of such factors as surface viscosity. liquid-
phase viscosity, rate of adsorption of frother and surface tension gradient.
Another factor which affects stability is the size range of the bubbles in a
blanket of froth emerging continuously from an aerated body of water.
Other things being equal. when these bubbles are sheared into the water by a
mechanical propeller a wide range of sizes is formed. but when it enters through
a porous septum a more closely sized and unstable foam is produced. (See
also Chapter 17.
Micelles
A micelle is an aggregation of organic surface-active ions or molecules so
clustered that their hydrophilic ends are oriented outward and their hydro-
carbon chains tightly packed inside the cluster. The size and shape of the
micelles is determined by the equilibrium between the attractive forces among
the hydrocarbon chains and the repulsive ones among the electrically charged
heads of the ions. When the concentration of surfactant ions in a solution
exceeds a certain point (the critical micelle concentration or CMC) the
independent ions aggregate in these complex groups. Three structural
shapes have been postulated.
From the specialised conditions in the transition zone which surrounds the
individual particle consideration now turns to the overall forces operating in
the pulp. These may be in part thermal. mechanical. magnetic and chemical
but include an important amount of electrical activity. Such localised fac-
tors as sorption. surface activity and ion exchange are dealt with later. This
section is concerned with generally accepted definitions of the forces which
react selectively on the mineral species in the pulp.
Electrochemistry has been defined 10 as the branch of physical chemistry
which studies the relationship between chemical transformations and energy in
reactions which involve electrical energy originating outside the affected sys-
tem. Although it is convenient at this point to consider pulp in a vessel
insulated from random external influences. the normal working conditions
under which ore is processed are fluent and unprotected from such forces.
Although process control in the mill must be applied in the context of these
conditions. causal forces are considered in isolation in this chapter.
Electrochemistry uses electrolysis and other electrical stimuli to promote
chemical changes. They act when a current flows through an electrolyte,
when potential difference is set up between electrodes, and when ionizing
Mineral Processing-Physics and Chemistry in Ore Treatment 385
fL=
[ ~]~€
1
(15.5)
Electro-osmosis is an effect arising from the partial separation of mobile
from fixed charges in an electrical double layer. Akin to it is the electro-
osmotic effect produced by the flow of liquid through a relatively fixed porous
solid mass or aggregate under the influence of an externally applied electric
field. Counter-ions surrounding the charged solid surfaces are attracted to
the area of applied charge of opposite sign while the relatively immobile solid
remains in place.
Three-phase Systems
TABLE 27
THREE-PHASE SYSTEMS
Type Example
\a) Mostly gas Froth
(b) Mostly liquid Floes
(e) Mostly solid Sediments
(d) Much gas and liquid, little solid Bottom of a froth column
(e) Much liquid and solid, little gas Thickener underflow
(f) Much gas and solid, little liquid Moist filtercake
Zeta Potential
partially seized layer of ions separated from the solid surface by a thin film of
water molecules. held to it condenser-wise by electrostatic force. Gouy. in
1910. added the concept of thermal motion which showed this layer as more
mobile. diffused or non-uniform. A moderately rigid part of the electrical
double layer is thus pictured as held in the liquid close to the solid surface
while the balance of the ions. and those of counter-charge (counter-ions or
gegen-ions) are diffused outward until their attenuation is that of the ions in
the characteristic liquid phase.
The gradient between the potential-determining ions. shown as anchored
circles at the interface in Fig. 187. and the normal concentration in the liquid
of fully hydrated ions and counter-ions from D outward (these hydrated ions
not being shown in the diagram) is the transition zone and the electric potential
r
difference across it is called the zeta-potential. At a very low temperature
the anchored charge B is embedded in the liquid phase. its ions being rigidly
adsorbed to the solid. As the temperature rises the system becomes increasing-
ly mobile and electrostatic coupling is more strongly opposed by molecular
movement until the characteristic situation is reached-the diffuse double
layer. The Gouy concept. which pictures a rigid inner layer of adsorbed or
structural ions diffusing for a few milli-microns outward from this interface. is
generally accepted in our present understanding. The presentation here
given is over-simplified. since electric charge on the surface of the solid can
arise from other causes than straight ionization-for example. from prefer-
ential ad- and de-sorption at selected sites on the surface. or the co-existence
of oppositely polarised surface areas. The double layer can be either ionic
or dipolar. The first of these can be several Angstrom units deep and the
second only a few molecular diameters. In the kinetics of the anchored
"wall". plane of shear or Stern layer (Fig. 187B) dispersion due to Brownian
motion is at work. Electro-kinetic effects also active include the streaming
potential (that of the plane of shear between the wall and the fully mobile
phase D).
The defects of many hypotheses concerned with the diffuse double-layer
theory arise from the concept of dimensionless counter-ions and hence from
preoccupation with electrical charges rather than mass-containing species. It
is not possible to follow this special branch of electro-kinetics further in a
textbook concerned with the practice of mineral concentration. but it must be
remembered that the surface of the particle is the doorway to its interior in
chemical extraction. and the selecting factor in flotation. As such. the surface
must be reckoned with if process control is to be scientific. In connexion with
traffic to and from the surface the significance of the iso-electric point must be
remembered-the point of minimum solubility. minimum electrical attack
and transfer. minimum ionization. It is represented by a critical pH value for
each species in an ore pulp.
(ideally) isolated system is constant, while its entropy tends toward a maxi-
mum. Entropy can also be thought of as "the measure of the capacity of a
system to undergo spontaneous change". 20 Clausius formulated the funda-
mental principles of thermodynamics thus:
(a) The total amount of energy in the universe is constant.
(b) The total entropy (S) is increasing, therefore entropy shows the
direction taken by a spontaneous natural process, which is irre-
versible.
For a stable or ideal crystal system 21
S = JdQrev
(l5.7)
o T
where T is the absolute temperature, rev indicates the reversibility of reaction,
the arrows in each case indicating two-way traffic tending toward equilibrium.
Metastability is a steady but incompletely satisfied state of equilibrium, which
will undergo further change on the introduction of something which aids
reaction tending to improve or complete stability.
Changes in the system may be self-initiated or may start with additions
from outside. The rate of change is expressed in terms of first, second and
third order reactions:
-d(A)
k1 (A) (15.11)
dt
- d(A) - d(B)
k~ (A). (B). (15.12)
dt dt
,' __ :
___
;,: K shell (2-)
" ....
/
,/ /"'"
,-e-..... ......,"
/ • ,_ -"' 'e\ \ Positive
/ " /' ", \ I \ nuclear
, cha rges
: I I I
I e ' : (protons)
I \ I I I
',\, ., / / I
-,' .... --
.,
'- ' ..
'e.., . . : '- -e---
--- ......
"
Its eleven electrons spin round the nucleus to which they are held by the
attraction exerted by the eleven positively charged protons at the centre,
round which they whirl in balanced orbits. The one M electron is unbalanced.
The chlorine atom has seventeen protons, balanced by seventeen electrons
(Fig. 189). ]n this case, one unbalanced electron on the outermost shell leaves
the atom dynamically unstable. The lone M electron (valence electron) of the
sodium atom is transferred to the "hole" in the chlorine's M shell to com-
plete the latter's octet. There is now one unneutralised proton on the
sodium, which gives this atom positive charge, making it positively ionised.
Similarly, there is now one excess negative charge on the chlorine atom, due
to its capture of an electron from the sodium, making it negatively ionised.
The two ions couple electrovalently to form
sodium chloride. Each ionised element has a perfect octet on its outermost
shell of electrons. The work done in removing this electron from the sodium
atom is 5·12 electron volts-the ionising potential. To form an ion by the
removal of two electrons in MgH) requires 22·58 e.v.; three (in AI + -t- +) need
53·0 and four (in Snt T + +) 101·9 e.v. Because the monovalent atoms can be
ionised at the cost of less work than those with more valence electrons, the
"alkali metals" are more active than the "alkali earths".
The valencies available at the active mineral surfaces offer four levels of
reaction to the attacking chemicals and, consequently, four possibilities for
differential surface modification. It is therefore of fundamental interest to
study the valencies offered by a newly cleaved surface in relation to the ionising
potentials required to attract ions from that surface into the liquid phase.
Even with the clues thus afforded the research worker, many further considera-
tions enter, and the overall picture is obscure. Electrical reaction proceeds
at electrical speed, and there is no established evidence that new surfaces are
immune from such rates of change. The author's hypothesis as to events
during comminution and cleavage is very tentative, but it may be found
Mineral Proce ssing-Physics and Chemistry in Ore Treatment 395
helpful. When mechanical impact and abrasion have created a new separate
particle it proceeds at electrical speed to neutralise the majority of its charged
lattice points from immediately available sources. Counter-charges are
available in the pulp and in its own substrate. and there may be some lattice
structure rearrangement. After an extremely short period of intense activity
on these lines, equilibrisation slows down until it reaches rates which are
measurable by laboratory methods. Much of the residual unbalanced
energy at this later stage is the result of lattice distortion, diffusion between
liquid and solid phases and chemical readjustment. The picture ends with
quiet relaxation, restoration of distorted lattices, and penetration of ions from
the aqueous phase where zones of high strain have been developed inside the
particle by comminution and chemical action. With moderately good con-
ductors such as the metal sulphides which are readily floatable, electro-valence
is an important factor in the general surface reactions during which the
sulphur lattice points appear to undergo a measure of hydration as a necessary
conditioning step. The nature and extent of this is neither fully clear nor
fully agreed. Electro-valence is defined by Hackh 16 as "a fundamental type
of atomic linkage, corresponding with ionised or polar linkages ... (and) due
to the transfer as distinct from sharing of electrons". Enough free energy
must be made available in the pulp to initiate and sustain each desired change.
The energising forces are supplied through hydrogen-ion concentration (by
pH control); anion concentration (hydroxyl. carbonate, silicate, cyanide, etc.);
available oxygen (either in the mill water or entrained); and surface-active
chemicals. A simple example of ionic structure is provided by crystals of
sodium chloride.
/ / /.
/ / /
/ V /-
V / ~
/ V
/ V V / --
o =CI
• =Na
Fig. 190. Sodium Chloride. The Unit Cell
The linit cell of NaCI has the form shown in Fig. 190. It can grow (repeat
itself) in all directions, provided sodium and chlorine ions arrive slowly
enough from a pure solution of these ions in water. If they reach the nuclear
lattice too rapidly, they have not sufficient time to find their ordered place in
the lattice. and a new oriented growth commences. For this reason, large
crystals are rare, whereas crystallites. which are intergrown systems of minute
crystals of similar composition. are common.
The second type of bond is the covalent. in which atoms share two electrons
396 Mineral Processing-Physics and Chemistry in Ore Treatment
in each valence bond. An example of this is the gas methane, CH' 4 The
hydrogen atom consists of a single proton and electron (Fig. 191a) and the
carbon atom of six protons and six electrons (Fig. 19Ib). These bond cova-
lently as shown in Fig. 192. The closed K shell of the carbon is omitted, since
only the valence electrons react to form an octet. Since no electrons are
transferred in covalent bonding, no ions are formed. Some of the most-used
organic reagents in flotation derive their aerophilic quality from covalent
bonding.
In the third type (the co-ordinate bond), valence electrons are again shared,
but all are contributed by one partner to the linkage. Examples are the sul-
phateand phosphate ions (S04) - - and (P0 4)-- -. Here, the group is ionised.
The four oxygen atoms require eight electrons to complete their octets, and in
addition to the six or five captured respectively from the sulphur and phos-
phorus atoms have acquired two and three respectively from external
atoms, thus acquiring a corresponding ionic charge. The fourth type of
bond, "metallic linkage", rarely affects flotation physics.
(a) (b)
Fig. 191. Atomic Hydrogen (a) and Carbon (b)
H
••
H• .H
• c •
••
H
Fig. 192. The Covalent Bond (Methane)
H:O:H
The latter forms are polar and are attracted to oppositely charged lattice
points. When an electrovalently bonded mineral particle is freshly severed,
its surface monolayer is charged wherever the ionic lattice has been ruptured.
It immediately begins to attract from the surrounding fluid the most suitable
available substances for reducing its electric potential. This it does by
sorption, hydration, oxidation (de-electronation), or exchange adsorption.
Because of the great variety of reactions possible in a flotation pulp, the
general word "climate", already used in this chapter, is often used. It refers
loosely to the physico-chemical state of the flowing stream which is under-
going commercial treatment. This climate is partly influenced and con-
trolled by means of reagents which vectorise the forces at work in desired
directions. These are concerned to accentuate differences between aerophilic
and hydrophilic mineral surfaces as a prelude to their separation by flotation.
A particle may be homogeneous, in which case the atoms of which it is
composed are bound together in a regular order. The pattern thus formed in
a crystal is termed the space lattice, and a straight line drawn through two
adjacent lattice points can be prolonged to pass through successive points
equally spaced. Parallel lines do the same and the crystal structure is refer-
enced to its x, y, and z distribution of such points with regard to a rectangular
tri-axial set of co-ordinates. The order will be repeated throughout such a
particle if it is a complete crystal, or a fragmentary portion of such a crystal.
It can be composed of an intergrowth of true crystals, in which case there will
be abrupt changes in lattice co-ordination at the boundaries between the
individual crystals forming the particle. A particle composed of minute
crystals, or crystallites, is usually tougher and harder than one derived from a
large single crystal. At the boundary of the particle its crystal lattice is
interrupted so that the atoms at the surface are only partially linked to those
in the body or substrate. The surface atoms are therefore in a less stable con-
dition than are the atoms deep inside the substrate, and are more sensitive to
the "climate" surrounding the particle. The electrical disturbances created by
the disruptive force which creates new surface produce polarisation of the
surface ions of an ionic compound. At the surface newly created during
comminution a reticulated and presumably uneven pattern of cations and
anions is developed, and their electrical disturbance affects both the substance
and the interphase. Considerations rising from this affect the consideration of
comminution as a process which, though in itself mechanical, has important
electro-chemical consequences. Throughout Nature all systems tend to
398 Mineral Processing-Physics and Chemistry in Ore Treatment
Ions and pH
The ionic theory, modified from that propounded in 1887 by Arrhenius, has
four points of interest in flotation:
I. When salts, acids, and bases dissolve in water they form positive and
negative ions. and the number of unit charges on each ion is determined by the
electrons it gains or loses when departing from atomic neutrality.
2. Ions are either produced by dissociation of the solid electrolyte,
NaCl-) Na + + Cl-, (15.15)
or, by reaction with water,
(15.16)
3. When an electric current passes through ion-containing water the
positively charged cations move toward the cathode (negative electrode) and
the anions toward the anode (positive electrode).
4. When an electrolytic compound dissolves in water, any acid properties
are due to hydrogen ions (protons attached to water molecules [H 2 0. H]+), basic
properties to hydroxyl ions (OH-), and neutral properties to the ions of the
salt concerned. save when these ions have reacted with ions of water (hydro-
lysed).
Acids in their pure state are unlike salts and bases, in that they are not
dissociated. They react with water to ionise:
HCl + H 20 -) H3 0+ + CI- (15.17)
H2 SO! + H 20 -) H30+ + HS04- (15.18)
H3 P04 + H20 -) H 30+ + H 2P0 4- . (15.19)
Ammonia acts as a base through reaction with hydrogen ion:
(15.20)
The neutral atom can acquire electrical charge. If it loses electrons
(usually from its outermost shell) it manifests one unit of positive charge for
each electron thus lost. If it captures electrons, it acquires one unit of nega-
tive charge for each such capture. Although changes of this kind are most
TARLE 32
PHYSICAL AND STRUCTURAL PROPERTIES ASSOCIATED WITH THE FOUR INTERATOMIC BoNDS
(after R. C. Evans)
readily made in the outermost shell where the electrostatic bond between
electron and nuclear proton is weakest, the second shell may also playa part
In Ionisation. Many atoms can acquire more than one charge. Where an
atom or group can have two valencies, the lower is by old custom called "-ous"
and the higher "-ic" e.g. ferrous chloride (FeCI 2 ) and ferric chloride (FeCi a).
If the ionisation of the atom is negative, the attraction should be directly
between a lattice point and a shell-electron. The hydroxonium ion of dis-
sociated water is usually called the hydrogen ion, H + , but is actually hydrated
as(H. HOH)+. If the H+ breaks away this reverts to H 20and H+ , a stripped
single proton. This might be expected to have, in addition to its known high
mobility and ease of transfer in the aqueous phase, considerable reactivity and
ability to penetrate an attracting crystal lattice. It would thus act as a de-
electronating (oxidising) agent.
The ionised lattice points at the surface of a particle suspended in water
attract oppositely charged mobile ions (counter-ions) (see Fig. 187). The
picture there given assumes a smooth sur:ace plane, but it is improbable that
this exists in fact. In the process of comminution the severance of structural
bonds is believed to leave a pitted irregular area (in terms of molecular
dimensions), and one in which the energised lattice sites are not necessarily
regularly disposed. These matters are conjectural, but it is convenient to
picture a Helmholtz mono-layer of ions strongly attracted to the counter-
charges on the majority of the lattice sites.
The potential effective charge in electrokinetic action (the ~-potential) is the
work needed to bring unit charge from infinity (in this case phase D) to the
surface of shear B. Hence ~ -potential can be modified either by changes in the
constitution of the electrolyte, or in that of the strongly adsorbed com-
pounds.
When a liquid moves relative to a solid, as when an electrolyte percolates
through a porous aggregate or a capillary tube, loosely adherent ions in the
diffused layer C are constantly removed, and a measurable potential (the
streaming potential) occurs between the ends of the system, in accordance with
the equation
417'1] KsEs
~~c _ _ _
(15.21)
PD
where E, is the streaming potential;
P is the pressure forcing the liquid through the capillaries;
K, is the specific conductivity of the liquid in the capillaries;
D is the dielectric constant of the liquid;
1] is the viscosity.
~-potential may also be derived from the electrophoretic behaviour. The
velocity with which a fine suspended particle moves is governed by its surface
charge and the applied electro-motive force. An approximate equation is
417'1]
~=--u, (15.22)
D
402 Mineral Processing-Physics and Chemistry in Ore Treatment
(15.23)
y being the dipole moment (y = Se). where S is the thickness of the double
layer and e the charge per cm. 2 of the double layer.
This "difference in potential" or potential gradient also provides the moving
force which causes electro-osmosis (diffusion through a membrane. and capil-
lary rise).
If changes are to be induced at the surface of a particle by ion-anchorage or
reaction with agents which have been added to the aqueous phase of a flotation
pulp there must be sufficient attractive force at that surface to overcome the
inertia of the system. It must act across the interphase. and as its attraction
varies as the square root of distance. the molar concentration of ions is
important. If a lattice ion captures a counter-ion from this zone. the surface
reaction at this point ceases to be polar. since the electrical tensions have been
neutralised. If the captured ion carries a non-polar group. this forms a
coating above the lattice point and the particle will react to flotation in accord-
ance with the dominant coating thus formed at the surface. In an electro-
valent compound such as a typical metal sulphide. there are positive and
negative ions in the newly sheared surface lattice. and there can be traffic in
both directions. It is possible to influence this traffic by adding desired ions
to the water. Balance is reached when the traffic from the mineral surface to
the water phase equals that in the opposite direction.
A major instrument of control of the ionisation in the pulp. and particu-
larly at the particle surfaces. is the hydrogen-ion concentration. pH is the
descriptive symbol. It denotes the logarithm of the reciprocal of the H-ion
concentration.
Even pure water can conduct some electricity. because it is slightly ionised:
2H 2 0 ~ (H 20. H)+ + (OH)-. (\5.24)
For any given temperature of water this is a constant:
Kw = [HsO+][OH-]. ( 15.25)
With pH defining the negative log H-ion concentration and pOH that of the
hydroxyl-ion concentration:
pH + pOH = pKw. (15.26)
At 25° C. the ion product for water is 10- 14 •
(H+) x (OH-) = 10-14 •
pKw is then 14, The pH at this temperature is therefore pKw-pOH. A pH
of 7 indicates neutrality. Lower values denote logarithmic increase in acidity.
Mineral Processing-Physics and Chemistry in Ore Treatment 403
and higher ones in alkalinity. Changes from the near-neutrality of pure water
occur when compounds are added which dissolve to produce H-ions. As these
increase, the hydroxyls (OH-) are correspondingly decreased.
An electrolytic compound becomes ionised on solution in water. If the
compound consists of a strong acid and a weak base or vice versa it ionises
fairly completely as regards the strong partner, but the relation between the
ionised weak partner and the water molecules is only partial. It resists
change in the pH of the solution such as would normally occur on the addition
of a strong base or acid, and the solution is said to be "buffered". A salt
such as aluminium chloride produces acidic buffering because the hydroxyls
are neutralised by hydrolisation.
This difference can be seen if water stained by the mixture of aniline dyes
called a "universal indicator" receives first a drop of dilute sodium hydroxide
(causing it to show the phenol-phthalein violet of pH 10+) and then small
additions of dilute sulphuric acid, which will produce a sudden transition to
the red of methyl orange at pH 4 -. If a salt such as sodium acetate (strong
base and weak acid) be now added and the experiment repeated, much greater
addition of acid becomes necessary to effect the colour change, which pro-
ceeds gently through the range of colours of the mixture in use.
The surfaces of the particles in a pulp exercise buffering effect, and the
opacity of such a pulp renders the use of dyes inconvenient for the measure-
ment of pH. The usual method of measurement is electrometric. Glass and
calomel electrodes are immersed in the pulp, and the potential difference
between them is read direct on a galvanometer calibrated in pH units.
The "common ion effect" is the change in concentration of an ion in a
saturated solution, through the addition of a different electrolyte which
yields an ion in common with solid substance present in excess. The ion
product remains constant, but with increase of one ion that of the other
diminishes correspondingly. Since the solution was already saturated
precipitation now occurs. The effect is to reverse the ionizing process.
substance is immersed in water, some ionises into solution, however low the
solubility. This, for a given temperature, may be expressed as a constant
Ks = (B+) (A-), (15.29)
but quantitatively it may be reduced by the presence of common ions or
increased by the "salt effect" of other electrolytes. Salting out, in general
terms, is the phenomenon in which a substance is precipitated from its
solution by the addition of soluble salts.
The mineral particles involved in the chemistry of flotation have very low
solubilities and their reactions take place either at the discontinuity lattice in a
monomolecular surface layer or with decreasing force for a very short
molecular distance inward. In electron-diffraction study of the boundary
lubrication of metals zones have been shown to exist, starting from the sub-
strate and proceeding outward I'ia increasing metallic oxidation anywhere up to
a thousand molecules deep to the lubricated interphase. No comparable
picture of the depth into the ore particle through which change is taking place
has yet emerged. When the wide variety of mineral lattice structures is
considered, together with the lattice defects and distortions possible in any
specific structure, it is surprising that research has led to so much agreement
in the whole field of mineral surface-physics, rather than that there is still much
that is controversial.
Surface chemistry is concerned with the specialised interface or interphase
in which the substrate of each reacting phase is unbalanced ionically. This
condition is induced by surface activity in which a surface-active agent (or
surfactant) modifies chemical, electrical or physical characteristics at the sur-
face of a solid. or changes the surface tension of either solids or liquids.
Usually the heteropolar molecules of the chosen surfactant are attracted to a
specific type of surface in the ore pulp so that only one species of mineral is
affected. Change of state at the surface of the attracting particle follows.
This may include either sorption or chemical change resulting in the formation
of a surface film differing in composition from the substrate. Sorption is a
general term for reactions at or near the surface. It includes the following:
Adsorption. The ability of the adsorbent (mineral surface in this case) to
hold adsorbates (gases, liquids. or solutes) adhesively at its surface. without
penetration. Desorption is the reverse action.
Absorption. Deep physical sorption by capillary action or superficial
migration of the sorbate.
Chemisorption. Deep absorption with some degree of chemical change or
lattice modification at the surface of the sorbing substance by means of
valence bonds.
Persorption. Permeation into a porous solid.
In connexion with surface change, the term "orientation" is concerned with
the position assumed by an adhering molecule, and the nature of the force
concerned. Hydrolysis and electrification are among the possible forces at
work on a zone (area of special interest on the solid surface) where molecules
are oriented.
One group of the surface-active molecule may form a polar monolayer
electrostatically attached to a counter-charged lattice point at the solid surface
Mineral Processing-Physics and Chemistry in Ore Treatment 405
while the rest of the molecule faces outward and thus presents a new type of
surface to the ambient phase at that point. If enough of these attachments
occur the lyophilic-lyophobic balance of attraction between the surface of the
particle and the ambient water and air decides whether it will remain wetted
in an aqueous phase or cling to the air-water interface of a coursing bubble
passing through the pulp. Other relationships modified by surfactants
include surface tension, emulsification and micellar grouping.
When the effect is to reduce the interfacial tension (surface tension) a solute
is said to be positively adsorbed. Such a solute's molecules contain both
hydrophilic groups (which adsorb to the water phase) and hydrophobic
groups (which are water-repellent). The three electro-chemical types of
molecule are the un-ionised, the anionic and the cationic. An amphoteric
surfactant is one which ionises in aqueous solution to become either anionic
or cationic in response to the pH of the solution. The molecules of an ampho-
lytic surfactant such as sodium lauryl sarcosinate contain both cationic and
anionic groups.
Surfactants 7 adsorb to surfaces to produce (a) lowering of one or more
interfacial boundary tensions in the system; (b) stabilisation of one or more
interfaces by forming adsorbed layers which oppose further change. Except
for solubilisation by some surfactants, action is either (a), (b) or a mixture of
the two.
Specific adsorption is sometimes suggested as a better term than chemi-
sorption. Change in the aerophilic quality of the mineral surface does not
become apparent in a two-phase solid-liquid system strictly isolated, so methods
such as contact-angle measurement described in Chapter 17 do not fully give
satisfactory research information regarding this type of sorption.
Where adsorption is due to positive attraction or linkage between polar
groups of opposite charge the surfactant is not necessarily amphipathic.
Moilliet 7 suggests that is should be called a specific surface-active agent, and
that it may be macro-molecular. Amphipathic adsorption is usually predict-
able, since the medium tends to expel one part of the molecule or ion and to
retain the other. Specific adsorption is not thus predictable since it varies
with the quality of the external attraction.
Surface energy can be stated in terms of free energy per unit area or surface
tension. It is provided by those molecules in the outermost liquid layer of a
liquid/air interface which are not compensated toward the substrate by the
mutual attraction of surrounding molecules. Any change in the total free
energy residing in the interfacial area depends on expansion or contraction
of this area. The effect of a surfactant in producing this change can often be
analysed thermodynamically.
In a liquid-liquid or liquid-solid interphase the hydrophilic groups of the
surfactant may be bound to the solid discontinuity lattice and the hydro-
phobic groups oriented outward. Here there may be some chemisorption,
as with the xanthates widely used in the flotation process to render a preferred
mineral's surface aerophilic.
Taggart's 23 solubility theory for flotation postulated metathesis between the
surfactant collector agent and the mineral surface, and has proved to be an
over-simplification. It failed to take into account crystal lattice structure,
406 Mineral Processing-Physics and Chemistry ill Ore Treatmellt
surface electric properties and physical adsorption. but was a good practical
guide.
An arbitrary definition of chemisorption regards it as an adsorption process
in which the adsorbate attaches to the adsorbing surface with a molar free
energy of \0 k.cal. or more. 8 Research has shown that with a few exceptions
uncontaminated and unoxidised sulphide mineral surfaces cannot be attracted
into a flotation froth. No fully agreed reason has yet been advanced but the
evidence points to the need for oxygen-attack to produce threshold sulpha-
tisation of the particle. which must be arrested far short of saturation. Plaksin 24
summarising part of a research programme. observes:
It is not necessary that the whole surface of the mineral should be affected.
Zones may develop on part of the surface. They may have ~pecial character-
istics such as an attracted layer of oriented molecules from the surrounding
fluid. A zone thus differs from a phase, the latter being homogeneous. while
the former has random disposition of its molecules.
Adsorption takes place at the solid boundary of the solid-liquid interphase.
some constituents of the fluid phase being attracted and held at the solid
surface. The attracting force may be ionic or entropic. In the latter case
the arriving molecules give a lower surface tension than that of the surface to
which they adsorb. resulting in an overall lowering of the potential energy of
the particle surface. If such a substance raises the surface tension, then it will
be repelled from the surface. since entropy tends to a maximum and is irre-
versible. Among positive (adsorbing) additives used in flotation are fatty
acids~ short chain alcohols. amines. phenols, and terpenes. Adsorption can
be caused by thermal or radiation energy. and can accompany hydrolysis or
other solvating actions changing electrical surface-potential. It can result
from residual valences when atoms at the surface of the adsorbent are only
partly saturated at their inner ends and have unsaturated valency points at the
outer ends capable of holding impinging counter-ions. An important charac-
teristic of chemical adsorption is the formation of a reaction product which
has the least solubility of all reaction products available for interfacial reaction.
In a saturated solution of an electrovalent compound having limited solu-
bility, the solubility product concentration is that of the ionic concentrations
and is constant at a given temperature. For a substance BA which yields
ions B+ and A- the equilibrium constant K is
(B+)(A-)
k cc= - - - (\ 5.30)
BA
and the solubility product K.BA. A molecule. atom. or ion giving the most
Mineral Processing-Physics and Chemistry in Ore Treatment 407
(a) The distance apart of the centres of charge across which attraction
must act.
(h) The spacing of the changed lattice points at the mineral surface.
(c) The size of attracted ions and their ability to pack themselves above
the lattice points.
Broadly. all adsorption is accompanied by reduction in free energy.
Physical (often called Van der WaaP5) adsorption requires only weak ad-
hesional forces between adsorbate and adsorbent while with chemisorption
the activation energy can be high. The change in interfacial tension on
charging an interface is sometimes called the electro-capillary effect. Ad-
hesion usually refers to an interface between dissimilar substances and should
not be confused with cohesion. which applies to the internal pressure due to
the force of attraction between molecules or particles in a mass and varies
inversely as temperature.
Changes in surface activity are accompanied by variations in surface ten-
sion. These can be measured for liquid/gas systems but for solids can at
present only be estimated indirectly. For a liquid it is expressed in dynes/cm.
as the force with which the surface on one side of a line I cm. long pulls that
on the other side of the line. If stated in ergs instead of dynes it is the work
required at constant temperature and liquid composition to increase the
surface area by one cm 2 of free energy per unit area measured in ergs/cm- 2 .
Further consideration is given to surface tension in relation to flotation in
Chapter 17.
A special type of sorption. indeterminate between ad- and chemisorption,
is that of the amalgams. These are so-called alloys of mercury with various
elemental metals. Though there is true chemical reaction in some cases
(e.g. sodium amalgam), the chemistry of gold, silver and copper amalgams is
more obscure. Some stoichiometric action takes place. Gold is slightly
soluble in mercury (0'06 'Yr. at 20° C.). The subject receives special reference
under Gold in Chapter 23.
Ion Exchange
tive ion sizes, valences and concentrations of the ions displaced by this cap-
ture. A liquid ion exchanger is a compound dissolved in a liquid immiscible
with the one carrying the ions which are to be exchanged. If this liquid is
aqueous. the immiscible phase may be kerosene. trochloroethylene. chloro-
form, etc .. a high flashpoint kerosene being preferred industrially. The
exchanging ions dissolved in the organic liquid contain hydrophobic groups
which hold them in this phase, the rest of the molecule being available for
capture. The compounds used include aliphatic amines. dialkyl phosphates
and fatty acids.
Solid exchangers are solid electrolytes with inert matrices (R) to which are
attached special polar groups (G) balanced by oppositely charged ions (X)
which are available for exchange. The formula is R(GX)n. If G is negative
X is a caticn and the material is a cation exchanger. 25 Inorganic exchangers
include zeolite. glauconite, natural montmorillonite and synthetic sodium
aluminosilicates. Some coals can develop IX qualities on sui phonation and
oxidation. The most important industrial IX materials are based on syn-
thetic resins such as phenol formaldehyde and styrene-divinylbenzene.
The polar groups in cation exchangers are sui phonic acid (-SOaH. strong).
carboxylic acid (-COOH, weak), sulphonium (-S.R 2 0H) and phosphonium
(PRaOH). The reaction is
R(GX)n -+- nM± -,. R(GM)n + nX± (15.31 )
The rate controlling step in IX is diffusion in the resin particle and across the
film of solution at its surface. and this is aided by increase in active area.
temperature rise. concentration of ions and their relative smallness. The
diffusion of counter-ions appears to be the critical factor in rate determination.
Any such ions which leave the LX media are replaced by an equivalent amount
of other counter-ions in order to maintain electro-neutrality. This step can
be subdivided into two rate determining steps-diffusion of counter ions in the
media, and inter-diffusion of these ions in the adherent films or diffusion layers.
and both may be at work simultaneously.
If a cationic resin is placed in an electrolytic solution there is cross-migration.
Cations go into the ambient phase and anions into the resin. This sets up a
positive charge in the solution and a negative one in the resin. and builds up a
potential difference called the Donnan potential. which pulls cations back
into the now negatively charged resin and I'ice versa. The counter-ion
concentration remains higher and the co-ion lower in the resin than in the
solution. With an anionic resin the effect is similar but of reversed sign. The
co-ion has but little effect on exchange kinetics.
For IX to succeed the exchanging ion species must be soluble in the selected
solvent. They must be at least partly dissociated and mobile, and the solid
must not be soluble or destructible in the solvent. The solvent should have
a high dielectric constant. That of water is E 81.
Resins are built from monomeric units forming linear chainlike polymers.
cross-linked so as to form flexible networks. The iogenic groups are carried
on the chains. These coiled networks unfold to permit limited entry of
solvent molecules and then swell till a balance is struck between the intruders
and the elasticity of the resin. Thus the resins are typically gels with an
}\4illeral Proce.\.~illg~Physics alld Chemistry ill Ore Treatment 409
References
I. Sidgwick, N.V. (1946). The Electronic Theory of Valency, Oxford University
Press.
2. Pauling, L. (1945). The Nature of the Chemical Bond, Cornell University
Press.
3. Adam, N. K. (1941). The Physics and Chemistry of Surfaces, Oxford Univer-
sity Press.
4. Gucker, F. T., and Meldrum, W. B. (1950). Physical Chemistry, American
Book Co.
5. Harned, H. S., and Owen, 8. B. (1950). The Physical Chemistry of Electro-
lytic Solutions, Reinhold.
6. Wells, A. F. (1950). Structural Inorganic Chemistry, Oxford University Press.
7. Moilliet, J. L., Collie, 8., and Black, W. (1961). Surface Activity, Spon.
S. Fuerstenau, et al. (1962) 50th Anniversary Vol. Froth Flotation, A.I.M.M.E.
9. Osipov, L. I. Surface Chemistry, Theoretical and Industrial Applications.
10. Milazzo, G. (1963). Electro-Chemistry, Elsevier.
II. Weye, W. A. (1956). Trans. N.Y. Acad. Science, 12.
12. Evans, R. C (1946). Crystal Chemistry, Cambridge University Press.
13. Gortner. R. A. (1937). Selected Topics in Colloid Chemistry, Oxford Univer-
sity Press.
14. McCarty, M. F., and Olsen, R. S. (1955). Eng. Min. J., Nov.
15. Del Guidice, G. R. M. (1934). Trans. A.I.M.E., 112.
16. Hackh (1949). Chemical Dictionary, Churchill.
17. Pryor. E. J. (1963). Dictionary of Mineral Technology, Mining Publications.
IS. Sutherland. K. L., and Wark, I. W. (1955). Principles of Flotation, Aust.
I.M.M.
19. Miall, L. M. (1961). A New Dictionary of Chemistry.
20. Thewles et al. Dictionary of Physics, Pergamon.
21. Fast, J. D. Entropy, Philips Technical Library.
22. Dash. W. C, and Tweet, A. C (1961). Observing Dislocations in Crystals.
Scientific American, Oct.
23. Taggart, A. F., Taylor, I. C, and Ince, CR. (1930). Trans. A.I.M.E., 87.
24. Plaksin. I. N. (1959). Trans. A.I.M.E., 214.
25. Heftmann, E. ( 1961 ). Chromatography, Pergamon.
20. Kakovsky, I. A. (1963). Study of Kinetics and Mechanism of Certain Hydro·
metallurgical Processes. I.M.P.C (Cannes)
CHAPTER 16
CHEMICAL EXTRACTION
Introductory
In the first edition of this book this chapter, then entitled "Gold Cyanida-
tion", confined itself to the hydro-metallurgical process of cyanidation. With
the steady extension of chemical treatment of ores and their concentrations
the term "hydro-metallurgy" is being replaced by "chemical extraction",
"chemical attack", and "chemical treatment". A change has therefore been
made and elementary discussion of additional processes is now included.
It could be argued that the extraction of specific elements or compounds
from their ores by dissolution is the province of the chemical rather than the
mineral technologist, and is therefore out of place in a textbook on mineral
processing. The counter-arguments are sufficiently cogent to outweigh this
objection. They are economic as regards the handling and treatment of a
moving tonnage of material, and technical in their use of the combined skills
of chemist and mineral engineer. Broadly, the same principles are applied in
Ore_ _ _ _ _ _ _N_e_w-,;r"I'
1-1 Sai"'" 1,,,,yd,dI soro"
Leaching ~ - - - ..J I
1
Solid/Liquid Separation - - -,
:
I
I I
I I
I
I
I
Tailings
. ~ . I
PreCipitatIOn I
orVal"" .---~
Concentrate
Fig. 193. The Basic Leaching Flow-sheet. Solids and pulps shown ill line,
liquids brokell
Mil/eral Pr(}('essiI/K~Chemical EXTractiol/ 411
Preparation has two main purposes in view. First, the ore must be ground
to such a fineness that all the economically recoverable gold will be sufficiently
exposed at the surface of a particle to ensure its dissolution during the optimum
time of chemical reaction. Second, the other compounds in the ore which are
liable to interfere with efficient recovery should, if possible, be removed or
rendered harmless.
The solvation of gold by cyanide takes several hours, speed being a function
412 Milleral Proc/'ssillg--Chemical Exlraclioll
of the total surface exposed to attack. Large particles of gold would need a
very lengthy period of exposure. This would in turn require an inordinate
volume of pulp-holding space if a big tonnage was treated daily. Fortunately
heavy and large particles of gold can readily be trapped so the holding capacity
need only be adequate to deal with particles small enough to escape from
gravity concentrators. Contact is rarely less than for eight hours and may exceed
twenty-four. During this period of solvation a large tonnage of gold ore.
together with a much greater weight of solution. must be kept moving and
thoroughly agitated. This is done in order to extract a few pennyweights of
gold from each ton of crushed ore. Pulp handling is therefore important
in the design and operation of the plant. in which the minimum amount of
power-consuming movement must be aimed at. consistent with efficient
extraction.
The weak cyanide solution circulates through the mill. becoming richer in
gold. Periodically the dissolved value is removed by precipitation. During
this circulation a little may be lost through spillage (though it is avoided as
much as possible). and there is some evaporation. After a while the cyanide
becomes "foul" with base-metal salts liable to lower the efficiency of recovery.
It must then either be regenerated or discarded. Since new water is con-
stantly coming from the grinding section. a balance must be struck between
entering water. circuit loss. and discarded solution. The foul solution is still
sufficiently poisonous to constitute a health hazard if allowed to run indis-
criminately into the district drainage area. and the working arrangements of the
plant must allow for safe disposal of waste liquor which cannot economically
be regenerated and returned to use.
In the third stage of the extraction process the solids and liquids must be
separated. It is customary to speak in cyanide plants of "sands" (relatively
coarse beds of particles through which water can percolate) and "slimes"
(true mineral particles which. when settled. would form a bed too closely
packed to allow such easy passage of water). The reason for this distinction
will be discussed later. The "slimes" are not a primary slime. such as a
clayey or talcose material produced by the weathering down of alkali-earth
minerals. Where such clays are mined with the ore. and are more or less
non-auriferous (or "barren"). trouble in the separating section can be avoided
if these primary slimes are removed before fine grinding is started. Some
gold flow-sheets therefore incorporate a washing stage for this purpose. This
has the further advantage of removing any soluble salts liable to interfere with
cyanidation. "Cyanicides" and other substances which jeopardise or upset
the gold-cyanide reaction include completely and partly oxidised sulphides of
antimony. zinc. iron. and copper; mineral acids produced by this oxidation;
some organic matter; graphite and charcoal (which can precipitate the dis-
solved gold prematurely). and chromium (which interferes with reprecipita-
tion of the gold at the appropriate stage of the process).
Electro-chemical Concepts
takes a positive charge to form the aurous ion Au + , according to the formula
for a metal /'vi of valence 11
(16.1 )
(.) being one farad (- -96,500 coulombs). The atomic structure of gold
(atomic number 79) shows the inner five electron shells to be filled (2, 8, 18,
32, and 18). This leaves one valence electron, the removal of which creates
the aurous ion. I n the electro-chemical series, with hydrogen = 0, the electrode
potential of the aurous ions on a gold anode is 1·36 -i- • Gold is therefore highly
resistant to attack by acids. Aurous ions can only exist in small concentra-
tions, and their salts readily form complex ions or covalent (auric) compounds,
or revert to metallic gold.
The trivalent form of gold (auric) has not at present been proved to take the
Au T + - ionised form, despite the weakness of the outermost shell of 18 elec-
trons. Two electrons from this shell readily accept covalent bonding. Thus
aurous chloride hydrolyses back to the non-ionised forms:
3AuCI = 2Au + AuCI 3 (! 6.2)
Chemistry of Cyanidation
This bromine salt is easily destroyed by excess alkali, iron, or pyrites, and
although it was favoured for treatment of telluride ores, its use has virtually
ceased in current practice. Improved grinding methods have led to better
exposure of the gold, and where necessary a selected fraction carrying the
troublesome mineral can be removed from the main stream of pulp by flotation,
and given special treatment.
The rate of solution of gold from a pure specimen under ideal laboratory
conditions was found by McLaurin 4 to be at its maximum with 0·25 % KCN,
a figure more than twice that obtained by Barsky, Swainson and Hedley 3 with
NaCN. In plant practice the supply of oxygen is limited and a further factor
must be considered-the rate at which dissolved gold moves away from the
surface of the particle and is replaced by fresh cyanide solution. The
"common ion effect" mentioned in Chapter 15 tends to hold solution ions
in super-normal concentration close to similar ions still partly bound in their
solid lattice. While it is true that this common ion attraction ceases with the
completion of the chemical reaction that locks the gold into a non-ionised
cyanide complex, there remains an envelope of partly spent enriched cyanide
around the auriferous particle. This enveloping action retards further
solvation unless overcome by such rate-determining factors as diffusion and
brisk agitation of the ore pulp. Since the effect is influenced by viscosity,
the rate of solution varies with temperature. In practice artificial heating is
not used. Wet grinding introduces an appreciable amount of heat into the
pulp. In sub-arctic conditions pulps must be kept from freezing.
The precipitation of gold from solution is effected by mixing the pregnant
solution, after removing as much oxygen as possible, with zinc in excess of its
gold content. The exact nature of the chemical reaction has not been agreed.
McFarren favours the reaction:
KAu(CN)2 + 2KCN + Zn +- H 20
= K 2 Zn(CN)4 -L Au -;.- KOH +H (16.11)
416 Mineral Processing- Chemical Extraction
d ~ ----(-)----- (1
Th'IC k-ener POSSI'bl y) (1)
. ." I
.S
::;,
-,
0,
00,
Q) , +
~:
all
:>, , •
1
o : AgItators (2)
~,
al I
II
d : ..
~: Solid-liquid
P:: ,:--~- ----(-)----Separation (3)_(+) Solids
L ___________________ ~
to Dump
,j,
Precipitation (4) ----- - - 7 - - - - - - - - J
1
Gold slimes to
Bullion Plant
~ l !
~ ~---------0-----------Thickener (and possibly filters) (la)
~
1 ~Solution
AgitT" (2a)
Switch:::..?" to Waste
,,
+,
,,
('0
~o I S
I'd to Dump - ~
Solid·liquid
. (3 --)-
0- P .: .
-- reclpltatlon (4a)
"eparatloll a
Fig. 195. Grinding in Water
418 Mineral Processing-Chemical Extraction
surface and the problem of balancing the removal of foul cyanide solution
against newly entering water is simplified. In ores in which the gold dissolves
quickly there is a good case for grinding in cyanide, but where dissolution
proceeds slowly, the extra time thus gained is a minor consideration. Gold in
characteristically sulphidic Canadian ores dissolves much faster than that in
typical Rand bankets. This may be due to plutonic as against alluvial
genesis. Some Rand plants now mill their ore in cyanide solution.
Consider two forms of a simple flow-sheet in which all the ore has been
finely ground and sent from the grinding section for cyanidation by agitation
(the method is discussed later). Let the grinding sections operate respectively
in cyanide and in water. The flow-sheets then are as shown in Figs. 194,
195.
In the cyanide grind, the only uncontrolled water entering the plant is the
moisture in the head feed. Some water is lost by evaporation, and a certain
amount of cyanide solution must be withdrawn when it has become too "foul"
with base-metal salts to be efficient as a gold solvent. This make-up water can
all be introduced at a point where it does useful work, and the quantity thus
introduced is under the control of the operators. This point is reached after
the pregnant cyanide has been displaced in stage (3) of the flow-sheet, but a
wetting film of auriferous cyanide still envelopes each particle of the solids.
If the water is used to wash or scrub these particles, the loss of dissolved gold
thus carried over to the dump is reduced. Item (I) of the flow-sheet is shown
as doubtful, since it might be better to agitate the effluent from the grinding
circuit without preliminary thickening.
Consider next the alternative flow-sheet (Fig. 195). Item Ia (thickener and
possibly filters) is no longer optional, as it is essential to remove most of the
water before introducing cyanide. If a thickener is used for this purpose, and
an underflow carrying 50 ~~ of solids is produced, each ton of ground ore
entering carries with it one ton of new water. This water must be accom-
modated in the solution-holding system, and must therefore displace an
equivalent quantity of cyanide solution, whether or no the latter is foul. The
operator has thus lost a useful manipulative control. If to the inexpensive
operation of thickening it is decided to add filtration, the new water is much
less-say 15 % entering on the filtered solids-but an extra element of com-
plication and expense has been introduced. This thickened pulp (or repulped
filter cake) must now be held in suspension in a collecting tank (lb, the stock
tank) and thinned down with barren solution. Since the barren solution
leaving 4b will now be diluted by fresh water, it must be adjusted in strength
before use, or alternatively fresh solid cyanide must be added at lb. Part of
this addition would be needed in the first of these flow-sheets, since cyanide
oxidises in use and must be replenished, but part of the addition now under
discussion is needed to replace loss incurred at 3a. When the solids are
separated at this point, any fresh water used to displace clinging auriferous
cyanide adds to the amount of barren cyanide which must be withdrawn and
run to waste, so the use of such water must be restricted.
In mineral dressing, the process should always be developed to suit the
specific ore, and to achieve desired ends with the greatest simplicity and econo-
my of effort. A strong case exists for milling in cyanide, but there are excep-
Mineral Processing-Chemical Extraction 419
Sand Leaching
Crushed ore from milling (together with auriferous tailing from gravity
or flotation treatment) should be brought to its optimum technical condition
before cyanidation. Cyanicides are removed, neutralised, or reduced to
tolerable concentrations. Comminution liberates the values to the point where
the greatest profitable recovery can be made, and the cost of further grinding
would not be compensated by increased recovery of bullion. Since grinding
is usually the most expensive item in treatment, the head value of the ore and
the mode of occurrence of its gold will, in each case, determine the amount of
grinding which should be done. At one extreme, ores such as the Mountain
Copper Co.'s are treated after crushing to -f', while at the other, the
auriferous sulphides of a high-grade ore may be ground down to extremely
fine slime before maximum profitable recovery is achieved.
When the crushed ore is sufficiently coarse, cyanide-bearing solution can be
made to percolate through moderately thick beds, with sufficient speed and
searching power to dissolve and remove the bulk of the exposed gold. The
rate of percolation determines the number of days a bedded tankful of such
crushed material must be treated, and hence the holding capacity (as capital
investment) of a leaching plant for a given throughput tonnage. The por-
osity of the bedded material determines the rate of percolation. If an un-
classified mass of sands were bedded, the associated slimes and fine sands
would obstruct the interstices between the large grains and thus interfere with
the percolation rate. This could in part be compensated by working with a
thinner bed, by using vacuum to pull the leach liquors down through the bed,
and by stirring these liquors into the upper layers of the bed. In the earlier
days of cyanidation the feed was classified into two types of product-sands
coarse enough to be leached and slimes too fine for percolation. With the
improvement and cheapening of grinding methods it has usually been found
better to do away with leaching of such ores and to reduce all the feed to a
fine state of grind-called on the Rand "all-sliming". Static leach treatment
continues to have value in a variety of special applications, notably with low-
grade ores, dump retreatment and low-capital projects where simple home-
made devices aid development finance.
The general scheme of leaching treatment can be pictured as in Fig. 196.
A certain amount of fine material can be tolerated in the sands classified to
the leaching section without retarding percolation too seriously, but it is
usually found undesirable to allow bad separation for another reason. This is
that gold-bearing solution tends to cling film-wise to the surface of ore particles,
and as the particle size becomes smaller not only does the total surface per
420 Mineral Processillg-Chemical Extraerioll
Assay
Sample
Make·up
,. - - - - - - - Solutions
unit volume increase exponentially, but more violent displacing action must
be used to remove these films than can easily be applied to leached sands.
At Golden Cycle 6 better classification of - 20 mesh material at a 200-mesh
split raised percolation rate from about I" to 10"/hour.
The coarse fraction can be produced as an underflow from hydraulic classi-
fiers or sand cones, or as the drag-out from rake, spiral, or bowl classifiers. It
can be fed into steel collecting vats (25' to 50' in diameter and 6' to 9' deep),
either through a hose directed so as to equalise the bedding, by turbo-distribu-
tors such as the "Butters and Mein" (Fig, 197) or by mechanical distribution,
Sand settles, while slime overflows with the excess water through vents in the
sides of the tank, which are plugged as the load rises, This gives some
further classification, the overflowing slimes being led to the slimes section
of the plant. (When calculating yield, correction must be made for gold thus
transferred), Sand can be leached direct in the collecting tank, but in a large
operation it is transferred ria bottom discharge openings to leaching vats,
These are large round tanks made of wood, concrete, or steel, up to 55' in
diameter and 15' in depth, the largest holding 850 tons, One ton of Rand
Mineral Processing-Chemical Extraction 421
barren sand occupies 23 cu. ft. as collected, and 21t cu. ft. when settled. If
crushed ore is leached without preliminary classification, the practicable
bedded depth rarely exceeds 3'.
The leaching tank has a false bottom through which solutions are dis-
charged (and occasionally introduced) and through which vacuum can be
applied. This bottom consists of an open wood framework of slats or per-
forated board. It is covered by twill, matting, or duck fabric, on which
wooden boards are laid to protect the textile from damage when workmen are
shovelling exhausted sands to the central bottom discharge. Fittings on appli-
ances which handle cyanide solution are of iron, not of copper alloys. Leach
rate varies from !" an hour up, 7" being a fair speed with vacuum aid applied
to the false bottom. The drained moist sands, still carrying up to 15 %
moisture, are levelled by hoes. Care must be taken that distribution of sand
is reasonably uniform or during the application of successive leaching washes
channels may form in the bed, through which the solutions will run without
making the searching contact with all the grains which is necessary to disso-
lution. The other essentials are the presence of oxygen and of sufficient lime
to prevent destruction of the cyanide. Some air is brought in with new
solution, and some is drawn in during drainage. Air can also be introduced
by top-raking, use of compressed air blowers, or by applying suction from
below. A more thorough method is to transfer the whole charge from one
tank to a second half-way through the leach, the movement breaking up
channels and ensuring good exposure to air. Leaching tanks are sometimes
built in pairs, one above the other, to facilitate this transfer and economise
space. "Strong" cyanide (0·05-0'1 ~;';NaCN) is used for leaching, and is
run on till the sand charge is covered, the solid-liquid ratio then being from
70:30 to 75 :25. If the liquor is permitted to percolate continuously, a
cover of up to 12" being maintained above the sand, there is less risk of
displacing entrapped air from the tank. This strong solution is sent for
precipitation. From time to time the charge is drained, and when dissolution
is nearing its end, weaker cyanide is employed (0·02 or 0·03 ~'-;;NaCN) to
complete the reaction and displace pregnant liquor. This solution is not
usually sent for precipitation but becomes stock for making up strong solution.
Final washing is with water or waste liquor from the slimes plant. The
treatment cycle takes from four to seven days. Tailing is then removed by
shovel, sluice, or mechanical plough and transported to the dump.
The double handling of sands from one leaching tank to another, or when
wetting with cyanide, begins in the collector tank. It aids aeration and coun-
teracts loss due to the tendency of the sands near the bottom of the bed to
become waterlogged and unaerated. The canvas bottom, which is held to the
tank sides by a ring of caulking rope, must be kept free from clogging due to
precipitated lime salts. A typical leaching sand is 50% plus 100 mesh in
grain.
Leaching has been successfully applied to clayey ores and the small-scale
retreatment of old dumps. The clayey material is worked by ploughing into
balls, sometimes consolidated with a little lime, and varying from ±n to 3" in
size. Leaching solution is run gently on to, and off from, this material either
in vats or in shallow trays CGnstructed from vertical boardings on an impervious
422 Mineral Processing-Chemical Extraction
clay flooring. In the case of reworking old dumps unusual care in needed to
avoid adding more than a bare minimum of protective alkali, if trouble due to
fouling of the solution by oxidised salts from the ore is to be minimised.
Slimes Agitation
less technically-minded age. It would be helpful if the use of the old term
could be confined to colloidal clayey material associated with the true ore as
the result of natural weathering, infiltration, or non-selective severance, and
not applied loosely to a variety of minerals, including the finer mesh sizes of
true ore produced during grinding. Primary slimes are usually loosely
adherent to the ore. and can be removed by washing, without substantial loss
of head value. When this operation can be justified, a nuisance is removed
and subsequent treatment facilitated. Primary slimes are apt to coat surfaces
which should remain open to chemical attack, to sorb expensive reagents
wastefully, and to cause trouble in the thickeners and filters by their slow rate
of settlement and ability to choke the interstices of separating media. The
word "slimes", as used on the Rand, may apply to pulps containing a certain
amount of such primary colloidal material, but is intended to describe ore too
finely ground to be effectively infiltrated by gently applied leach liquors. The
dominant characteristic of slimes treatment is its application to particles (say
minus \00 mesh) too small to be permeable by a percolating leach-liquor when
settled into a solid mass, but small enough to remain suspended in cyanide
solution under conditions of mild stirring and agitation.
Treatment may be generalised as in Fig. 198.
jCla,,;fio.
!
T I
I
To Leach
or Regrind
r Thickening Water--J
Slurry +-
1
- - - - - - -,
Barren
Cyanide
Agitation Solution
' kenmg-
1..------ Th 1C .
!
- - - - -"Pregs
I " I
~----i
t to I
Sand Filtration ____ • __ ~ __ ~:;p~ta~;""~,,.J
Discharge- and Washing
Gold
Slimes
Fig. 198. Slimes Agitation Flow-sheet
is being transferred into agitators. This method has largely been replaced by
one in which thickeners receive pulp from the grinding circuit and send a
thickened slurry to the agitators. Where space is limited, or winter conditions
sevet~, tray thickeners are sometimes used. For agitation the pulp is thinned
to a water-solid ratio of about 1·2: 1, with sufficienc dissolved cyanide to give a
strength (in terms of KCN) of 0·01 ';{, and protective lime 0·0005 % CaO.
426 Mineral Processing-Chemical Extraction
ADJUndL(
Willi
Separation
After optimum solvation of the gold has been reached (or in some processes
concurrently with this solvation), pregnant cyanide must be ~jsplaced.
The first wash with cyanide solution. applied to the new feed, dissolves the
bulk of the gold. This solution is sent direct to the precipitation plant.
However separation has been effected. the solid fraction is now wet with
pregnant solution which films each particle and is held by capillary action in
the spaces between particles. This cyanide must be removed by washing
with barren cyanide. low-value cyanide. or water. Cyanide solution in
various stages of gold-enrichment is available for this work. but the amount of
new wash water introduced must not exceed the foul cyanide liquor discarded.
since solution storage facilities are limited.
The various washings produce cyanide solution which is either (a) rich
~nough to be sent to the precipitation section. (b) mildly auriferous and ready
to pick up more gold before precipitation. or (e) foul enough to need regenera-
tion or to be run to waste, In addition. barren solution is returning from the
precIpitation section, The two main methods of effecting primary separation
are CCD. (counter-current decantation) and filtration. A combination of
the two is also used. CC 0, utilises the method in which upgrading or
downgrading is applied by stages (Fig. 202).Any numberofretreatments desired
are used in line. the down-graded product from each appliance providing the
head feed of the next succeeding appliance. and being joined there by the
upgraded product from the next appliance lower down the line.
428 Mineral Processin!{-Chemic:al Extraction
Cone.---+-----! I-----ll---- Ta iI
Tail
Cone.
Cone. Tail
Fig. 202. lOlillter-clirrellI Decalltatioll (CCD.)
cyanide. gold. or water justified the cost. this slurry could be filtered to recover
most of the liquor. A desirable working detail is that the slurry entering each
lower thickening stage should be turbulently mixed with the overflow cyanide
.----Water
Th,ck Pulp
to Waste
Precipitation
Fig. 203. CCD. ill Cyallidatioll
coming upline from the thickener below. This turbulence entrains necessary
oxygen and also helps to displace pregnant cyanide from the particles.
CCD. is not used on the Rand: where intermittent decantation (similar in
principle) was practised for many years until it was displaced by continuous
Milleral Processillg-Chemical EXlraclioll 429
1
Pulp
Overflow to .1
,. . ~
PreCipitatIOn
- - - - - - Thickener
I
- - - - - - + ) Slurry
Cake to Dump
Fig. 204. Rellloml of Solulioll /rOIll Pulp
In some plants which grind the ore in cyanide it is found that the moisture
leaving the process (on the discarded filter cake) balances the new intake of
water and provides a convenient discarding channel for foul mill solutions.
Some gold is lost in the discarded filter cake. Re-pulping and re-washing are
stopped when further recovery ceases to pay for the cost of extra treatment.
The influences which bear on efficiency of solvation are considered in Chapter
15 under Three-phase Systems and shown in (15.6) as a rate-determining
equation.
The position at this stage is that cyanide solutions of various strengths and
degrees of contamination, aeration, and gold content are circulating through
the plant or held up in tank storage. The richest of these solutions is sent
continuously to the precipitation section to be stripped of its gold. The
operational sequence is of the nature shown in Fig. 205.
Separating Section
I
I
I
{-
Filtrate
I
oj..
Clarification
1
I
I
t .
DeoxygenatlOn
I
-1
I
1
1
..
jr------F;lt"'t;on - - - - - - - - -
As the filtrate leaves the separating section it is more or less cloudy, owing
to the presence of material too fine to be trapped on ordinary cloth filters. If
zinc was introduced into solution in this condition the process of precipitation
would be interfered with, partly by films of slime settling on the zinc and
partly by adulteration of the gold precipitate with such slimes. This undesired
slime would flocculate and choke the filters used to retain gold slime. It is
Mineral Processing-Chemical Extraction 431
therefore standard practice to clarify the pregnant filtrate from the separating
section, the aim being to turn it into a limpid liquid. The oldest method is to
pass the liquor through a tank containing a bed of sand. Colloids are arrested
as a film at the top of the bed, which can be skimmed off periodically. A
certain amount of de-oxygenation accompanies this clarification. At one
time this was aided by the addition of iron filings to the sand, to take up
oxygen by rusting. This practice has been abandoned since it introduces
undesirable chemical reactions between the oxidising iron and the free cyanide.
Small thickeners have been used for clarification, the rake mechanism
being operated once daily and lowered a fraction of an inch, so as to plough
a thin skimming of foul sand to the central well for discharge. In a variation
of this the thickener works upon a bed of proprietary filtering material, which
is similarly shaved down.
The method of precipitation worked out by Merrill and Crowe, which
embraces clarification, is in widespread use today and is described below.
A short description of precipitation in boxes filled with zinc shavings is
desirable, as this gives some insight into the problems involved and provides
a practical method of working the small, undercapitalised type of venture.
Zinc shavings are swarfed on a lathe or spun from molten zinc just before use
(to minimise surface oxidation). They are dipped into 10% lead nitrate (or
acetate) solution, which precipitates on them a darkening coat of lead. Thus
an electro-couple is formed which has a stronger precipitating action on the
gold-bearing cyanide than has pure zinc. Lead nitrate may also be used in the
agitation section to precipitate soluble sulphides, and any excess arriving at
the zinc will help to maintain the PbZn coupling.
The zinc shavings (or extruded zinc wire) are packed into steel boxes, so
made that the pregnant solution will flow upward through each box, over a
weir, and down to the bottom of the next box in the series through which it is
to rise. Shavings are contained in wire baskets, through which precipitates
can fall to the bottom of the box for collection, via a drainage plug at one
downsloped corner. In use, the zinc in the head box of the series (five to
twelve boxes) disintegrates. New zinc is added to the final box, and is worked
up toward the head as replacement is made. Minus 20-mesh shavings are
periodically removed from the head boxes and sent to bullion treatment.
With too weak cyanide, a white precipitate forms on the shavings of zinc.
This masks the surface and inhibits precipitation of gold by preventing contact.
By raising the cyanide strength at entry this precipitation is largely avoided.
If the solution has been de-oxygenated before presentation to the zinc such
trouble is minimised.
Care must be taken lest extremely fine gold, as slimes, becomes detached
from the zinc and is flushed away. Sometimes the final box is left
empty or is packed with coke in order to trap such slimes, which can be
periodically reclaimed. No attempt should be made to dress the zinc boxes
while cyanide is flowing. Up to 10 tons of solution can be precipitated daily
for each cubic foot of carefully packed filiform zinc (weight about 5 lb.).
During clean-up, all gold slime must be carefully collected from the boxes
and transferred to the vessel in which cleaning up and/or acid treatment are
to Le performed.
432 Mineral Processing-Chemical Extraction
....
"-
S
:::
i .g
rii ! .~
"~
I...
0 ~"
'x"
(:l
~
~....
....
~'"
-.a
0
N
A
--L.!i __ __ __ _
:
i
434 Mineral Processing-Chemical Extraction
copper is present, it may cause trouble by plating out on the zinc. This is
not likely to occur with zinc dust, though it can happen with shavings where
the total available zinc surface is much more limited. If the quantity of zinc
present is insufficient for complete precipitation of the metal cyanides, reaction
is preferentially directed to the gold. This tendency can be made use of by
"starving" the zinc dust into the "pregs" in two successive additions, with
filtration of slimes after each. The first slime will be gold-rich, and the
second copper-rich.
Research has been applied for several years to the possible use of ion-
exchange methods of stripping the gold from the pregnant solutions, and
processes incorporating this method are working in some mills. The main
advantages which have been claimed are the possible continuous solution and
precipitation of gold in one operation of a resin-in-pulp (R.I.P.) type; the
reduced need for clarification or even filtration of the pulp before precipita-
tion; and the possibility of recovering other dissolved metals on the resin,
thus making a by-product or, at least, regenerating foul cyanide. Little has
yet been published.
l
A 3. Hydrocyanic acid.
4. Protective alkali.
5. Precipitation efficiency (gold in "barrens").
These (with the possible exception of No.3) are controlled by routine
checks made at frequent intervals during the day.
6. Degree of fouling of circulating solutions by zinc, copper, iron
B { as ferrocyanide and sulphocyanide.
~ 7. Available oxygen.
These are periodically checked.
C {8. Quality of zinc dust.
9. Available CaO in lime.
These affect new supplies.
Hydrogen cyanide (hydrocyanic or prussic acid) does not dissolve gold and
is therefore not "available", this term referring to that portion of the cyanide
salt in a suitable condition ("free") for making such an attack. Not all of the
"total cyanide" present is thus free, since substantial portions may have taken
up copper, zinc, etc., and thus become unable to dissolve gold. Since the
oxidation of various minerals in the ore releases acidic products into the
solution, a slightly alkaline condition is maintained by adding calcium oxide.
(Too high a content of lime slows down the cyanide process.) The solution
is checked periodically for loss of alkalinity and an appropriate correcting
addition is made. This lime combines with free hydrogen cyanide. restoring
the availability. Finally. routine checks are made by the shiftsmen as to
efficiency of precipitation. If insufficient fresh zinc surface is present, the
gold will not be adequately removed from the "pregs", and loss may result.
Simple rapid tests have been devised for the routine control of the foregoing
conditions, and they are made by the mill operators. These tests are described
below.
It is also necessary to decide when a cyanide solution should be discarded,
either because it is too foul for further reclamation or because newly arriving
water must be accommodated in the system. Control tests are made in the
mill laboratory. This also applies to the available oxygen, though a colori-
metric test can be made in the plant. Without adequate oxygen, as Elsner's
Equation shows, cyanide does not dissolve gold.
Tests of new supplies are also made in the laboratory. The criteria of zinc
are:
(a) Metal content and purity.
(b) Mesh size (available surface).
(c) Hydrogen emission (surface condition).
Lime, which may be kilned locally, is tested for its content of CaO in order
that the quantity to be added to a given bulk of liquor can be correctly
calculated.
Mineral Processing-Chemical Extraction 437
Solution Control
At various points in the solution circuit it is essential to know how much
cyanide is present in "available" form (capable of reacting with the gold in the
ore). Since this cyanide must be slightly alkaline, the "protective alkali"
438 Mineral Processing-Chemical Extraction
must also be maintained. Not all the cyanide present is available. Efficiency
of precipitation is also tested by a rapid method which shows the approximate
amount of residual gold in the "barren solution". The final routine test is for
efficiency of oxygenation, since the ability of the solution to dissolve gold
depends on adequate content of air. These tests are made at regular inter-
vals during each shift. Periodically the amount of zinc, copper, ferro-cyanide,
and thio-cyanate in solution may be checked in the mill laboratory. The
tests are made to determine the available CN- ion, and the results must be
translated in terms of the compound used to provide these ions. Tests are
reported in terms of potassium cyanide, but this is not used commercially,
the chemicals most employed being sodium cyanide and calcium cyanide.
Free Cyanide
The reaction between silver nitrate and potassium (or sodium) cyanide is:
AgNO a + 2KCN = KCN . AgCN + KNO a (16.15)
KCN . AgCN + AgNO a = 2AgCN + KNO a. (16.16)
The latter (AgCN) cyanide is insoluble, so that when this stage of reaction
is reached a permanent white precipitate appears. This end-point is made yet
more visible if one or two drops of a 10 % solution of potassium iodide are
added before commencing the titration. If 13·046 gm. of AgNO a are dissolved
in distilled water and made up to one litre, then each c.c. of this solution,
titrated into 100 C.c. of the cyanide solution (to the point when a precipitate
forms), is equivalent to 0·01 % of free KCN. To convert this to pounds per
short ton, the volume in c.c. of AgNO a used to produce a precipitate is
multiplied by 20.
Instead of titrating 100 c.c., the test may be run on 10 C.c. of solution.
Each c.c. of silver nitrate used now represents 0·1 % of free KCN or 2·0 lb./
short ton of ore.
Total Cyanide
If 5 C.c. of 10% sodium hydroxide is now added to the sample which has been
titrated, all the free cyanide, hydrocyanic acid, and zinc double cyanide is
converted to sodium cyanide. Further titration with silver nitrate until an
insoluble precipitate begins to form shows the total cyanide figure. This,
however, takes no account of any sodium ferricyanide or ferrocyanide present.
A simple semi-automatic method for monitoring the actual dissolving
strength of cyanide solution is reported10. Gold leaf is attached to a glass
slide mounted on a slowly revolving wheel. At each revolution the slide
passes through a light beam, and a photo-electric cell shows the rate of
dissolution. The method takes care of several factors, and includes warning
of inadequate oxygenation.
Protective Alkali
The 100-c.c. sample used to find the free cyanide can be used for this
determination if no alkali is added after completing the first test. It is titrated
with deci-normal oxalic acid (6'3 gm./litre of C2 H2 0 4 .2H 20), after adding 0·5
C.c. of a 10 % solution of potassium ferrocyanide and a drop of phenolph-
thalein indicator. Titration is continued until the purple colour disappears.
Mineral Processing-Chemical Extraction 439
ore. This yields a comparatively rich concentrate which can stand the cost
of intensely fine grinding or other specialised treatment. The second is the
steady improvement in roasting methods, which can today be applied in such
a way as to ensure almost complete destruction of the sulphides and their
associated trouble-making compounds at a controlled temperature. "Dead-
roasting" of the whole of the ore to expel all tellurium, sulphur, and/or
graphite is rarely practised, since it is expensive and difficult to control within
desirable temperature limits.
One difficulty with antimonial ores is that although protective alkali is
needed to protect the cyanide, antimony sulphide (stibnite) is soluble in alkalis
and the compound thus formed decomposes cyanide to thio-antimonite.
One possible reaction for either arsenic or antimony is
2 Sb2Ss + 6 Ca(OH)2 = Cas (SbOs)2 + Cas (SbSs)2 + 6 H20 (16.17)
Cas(Sb2Ss)2 + 6 NaCN + 302 = 6 NaCNS + Ca(SbOs)2 (16.18)
Pre-leaching of such ores with 2 % to 4 % of caustic alkali, oxidation
roasting, and prolonged weathering are among .the methods which have been
used to mitigate this trouble.
Where carbon is present in such a form that it absorbs the gold from its
solution in cyanide, it can sometimes be removed by pre-flotation. Special
carbon-depressing reagents have been developed for removing such undesired
material from concentrates before cyanidation. Graphite is sometimes
present in such a form that it will rise as a scum on the classifier if a little
paraffin or diesel oil is fed in with the ore, and this scum can then be removed
from circuit. One mill takes advantage of the fact that the graphite in its ore
preferentially sorbs mineral oils and causes such particles to become coated
with an oily envelope, after which they can no longer sorb gold from the
cyanide solution.
Since oxygen is essential if the cyanide is to dissolve gold, any mineral
present in the ore which is oxygen-avid may reduce the efficiency of extraction.
Pyrrhotite is the most active of these compounds. In some forms it is com-
paratively innocuous, while in others it is extremely unstable, taking up
oxygen so eagerly as to make the task of providing sufficient excess air diffi-
cult, and adding to this the further undesirable effect of reacting with the
cyanide to form thio-salts incapable of dissolving gold.
Among possible reactions between pyrrhotite and cyanide are
Fe5S6 + NaCN = NaCNS + 5FeS (16.19)
FeS + 202 = FeS04 (16.20)
FeS04 + 6NaCN = Na4 Fe(CN)6 T Na 2S0 4 (16.21)
in which series both cyanide and oxygen are consumed. In alkaline cyanide
the reaction may be
2FeS + 2NaCN + 3H20 + 30 = 2NaCNS + 2Fe(OH)s
(16.22)
and in neutral solutions
FeS + 7NaCN + H20 + ° = NaCNS -+- Na4Fe(CN)6 + 2NaOH
(16.23)
Mil1eral Processil1g-Chemical Extractiol1 441
Pyrrhotite may be regularly distributed through the ore body, or may vary
in its concentration and/or instability from point to point along the lode. Pre-
aeration at low alkalinity, followed by cyanidation, is used at Sub-Nigel to
combat this reaction. 12
If it is not convenient to remove pyrrhotite before treating the balance of the
ore, or to segregate ore containing excessive amounts until it can weather
down, then care must be taken to ensure ample aeration during agitation. It
has been suggested that apart from trouble for which pyrrhotite is directly
responsible, it is also able to act "as a sort of fulminator, and not improbably
as a catalyst ... (and) ... quite small quantities (can) introduce serious
amounts of cyanicides of many different kinds .... "
The gold-treatment plants in the North Ontario region commonly reinforce
the aeration provided by the standard agitator of the Dorr type by introducing
auxiliary pressure air deep into the tanks. In addition, the cyanide solution
in which the ore is milled is kept down to a low strength so that milling heat
and freshly sheared sulphide-mineral surfaces shall have as little opportunity
as possible to decompose the cyanide.
The effect of (a) and (b) is to justify more costly methods of concentration
than could be economically applied to run-of-mill ore. (c) means that sul-
phuric acid will be generated during aeration, and that the alkalinity of a
flotation pulp must be watched very closely if the cyanide is not to be des-
troyed. If such a pulp is aerated in the presence of lime, hydroxyl ions react
with the surface of the particles of pyrite and sulphide ions are liberated into the
pulp. Here they react with oxygen from the aeration, and thiosulphate and
sulphite ions are formed, also sulphur, all of these in time oxidising to form
sulphate ions.
If cyanide is also present, the above reaction may be accompanied by the
formation of a thio-salt, which is useless for the dissolution of gold. Pre-
aeration in the absence of cyanide is therefore preferable, if possible followed
by decanting away of the water in which the sulphur salts are now dissolved.
For some ores pre-aeration without added lime has been recommended.
The best procedure is arrived at by means of empirical experiments in the
ore-testing laboratory. The whole problem is frequently aggravated by the
fact that the flotation concentrate must be reground to a fine slime in order to
expose most of the locked-up gold to the cyanide solution. This greatly
increases the new sulphide surface and, consequently, the reaction with
oxygen. This reaction is not comparable in completeness with the roasting
treatment described below, in which the oxygen penetrates the fairly
solid lattice of the sulphide and destroys it. Broadly the difference is one of
degree, as between singeing and burning to ash. Aeration of freshly ground
pulp has a relatively gentle oxidising effect on the newly developed sulphide
surfaces.
To sum up, the efficient cyanidation of flotation pulps may need adequate
regrind, followed by aeration sufficient to stabilise the disturbed surfaces.
possibly removal of the fouled water and, finally, cyanidation with copious
aeration and vigilant watching of the protective alkalinity of the pulp. A
secondary benefit, obtained when all these matters are controlled, is that the
pregnant solution is not foul, and consequently precipitation is more efficient.
These controls are more easy to apply to batches ·of concentrate than con-
tinuously, particularly when the chemical reactions in the pulp are intricate.
The operator can deal more specifically with the minor variations in a batch
than with obscure changes in a continuous flow-line. In a small plant hand-
ling, say, ten tons daily, the flotation concentrate would be ground, collected
into an agitating tank, aerated in water which would then be decanted, and
finally cyanided in the same tank, each stage being carried to a satisfactory
completion. One such plant uses five home-made Pachuca agitators to
share the work, each holding up to five tons of dry solids. The flow-sheet is
shown in Fig. 207.
A tank is charged for 12 hours, the solids being settled and the overflow
water returned to the mill. The thickened pulp is then aerated without lime
for 18 hours and allowed to settle. Clear foul water is syphoned to waste
and 12 hours of agitation in weak cyanide, with lime, begins. The "pregs"
Mineral Processing-Chemical Extraction 443
I
with sulphuric acid.
Ball Mill
1
Corduroy Strakes
Pachucas
Fig. 207. Batch Cyanidation of Flotation Concentrate
Roasting
Since excessive consumption of cyanide arises from reaction with unstable
sulphides in the ore, roasting methods are sometimes used to deal with these
troubles. The change of state produced by pyrometallurgy is sometimes an
444 Mineral Processing-Chemical Extraction
were open, a fume nuisance would arise and this heat would be lost. The
concentrate (a filter cake) entering the furnace is very sticky, and the rabbles
must bring back enough dry, powdery charge in their circular motion to
coat this putty-like material with a film of dust. The filter cake now shreds up
into dust-coated lumps as the rabbles stir it. As the material is worked
forward into the hotter section of the furnace the trapped water turns to
steam and disintegrates the lumps. Thus the cost of an independent drying
section is avoided and the dust set up is kept sufficiently low to avoid the
need for costly trapping arrangements. In the roasting zone the charge has a
flocculated appearance. Progress is regulated by the gradient and by drops
between rabbles. These regulate the amount of backward travel. The
calcine is drawn from the cool discharge end by a raking mechanism and
agitated in water before being adjusted for alkalinity with lime. The main
object of roasting at Lake Shore is not to eliminate sulphur but to open the
material to cyanide attack by rendering it porous. The modifications used
at Lake Shore reduce flue dust at feed and discharge ends to a minor nuisance
of no great economic importance, the weekly clean-up of flue dust being some
0·4 % of the feed, carrying only an ounce or so of gold per ton. It is trapped
ahead of the stack.
446 Mineral Processing-Chemical Extraction
Carbon
Carbon, whether present as graphite in the natural ore body or added as
activated charcoal, has the power of sorbing gold from its solution in cyanide.
This can be a serious nuisance when treating graphitic ores, since gold-bearing
graphite then goes to waste. From time to time charcoal has been used
instead of zinc as the precipitating agent and it is thus employed in current
practice to a limited extent.
When such carbon occurs in the ore it can be removed by roasting, as at
Ashanti; floated out in the grinding circuit by the use of a little paraffin or
diesel oil; depressed from the ore by special flotation agents; or oiled over by
lubricating oil added to the grinding circuit and thus insulated from cyanide.
Fortunately, such graphite is not always a menace to plant efficiency. Several
Mineral Processing-Chemical Extraction 447
theories to account for the sorbing action have been advanced, none of which
will be discussed here.
Gold is deliberately sorbed from its cyanide solution in some plants where
special circumstances would lead to difficulties with the use of zinc as a pre-
cipitator. A method is described by Edmonds 18 and other applications are
described in the literature of the subject. The advantages c1aimed 5 are that
better recovery is made from foul solutions; that cyanide loss is lower; and
that refining is cheaper. In recent years methods have been developed for
removing the charcoal from the pulp. One is flotation, in which cyanidation,
aeration, and continuous removal of charcoal is used, thus permitting con-
tinuous dissolution and precipitation of the gold. Another is the use of
charcoal which has been rendered magnetic and can be removed expeditiously
by magnetic methods. In a third method a wire cage containing activated
charcoal is immersed in the tank containing pregnant cyanide. As gold
dissolves into the cyanide it is taken up by the charcoal, which is removed as
requisite.
Lewis and Metzner 19 have described the activation of carbon derived from
coconut, hardwoods, etc. The process "selectively removes the hydrogen
or hydrogen-rich fractions from a carbonaceous raw material in such a
manner as to produce an open porous residue". Charcoal can be exposed
to CO 2 , steam or a mixture of these gases at between 700 c. 0 and 1000° C.
Successive use of CS 2 and H is still better, using fluidised beds.
The gold-bearing charcoal can be sent to a smelter, or calcined, its ash
being re-cyanided. 20
Chlorination
The use of chlorine to extract gold from its ores pre-dates cyanidation. The
high cost of chlorine in the nineteenth century made it uneconomic (1881) but
this objection no longer exists. Gold is dissolved fairly quickly in the presence
of nascent chlorine, particularly when chlorinated brine is used,21 and some
of the difficulties presented by cyanicides are avoided. The method is not
used today.
Losses
In theory the amount of gold present in the tonnage of ore entering the
plant should all be accounted for in the equation for metallurgical balance.
ticles in tailing dumps. These inhibiting films are difficult to detect and, in
most cases, almost impossible to remove by commercial methods, particu-
larly when they are associated with partially oxidised materials.
Pressure Leaching
In most chemical reactions the temperature and the intimacy of contact
between the reacting phases are rate determining factors. Batchwise and
continuous systems have been developed which process superheated pulp
450 Mineral Processing-Chemical Extraction
Bacterial J"eaching
Biological research has shown that selected strains of bacteria can be made
to accelerate acid leaching. The micro-organisms found to render copper
minerals more readily leachable 26 are Ferrobacillus ferrooxidans, Thiobacillus
concretiverous, and Thiobacillus ferrooxidans, all of which occur in the effluent
waters of some mines in the southwestern United States. Research has
shown that some species of bacteria can feed on ore minerals in a highly
acidic environment. Apparently the sulphide is not attacked directly, but
ferrous iron dissolving as the result of chemical oxidation is converted by their
action to acid ferric sulphate, the process chemistry being
2 FeS2 + 702 + 2H 20 = 2 FeS04 + 2 H2S04' (16.27)
4 FeS04 + 2 H 2S04 + O2 + bacteria = 2 Fe2(S04)3 + 2 H 20
(16.28)
7 Fe2(S04)3 + FeS2 + 8 H20 = 15 FeS0 4 + 8 H 2S04 (16.29)
or, by reaction with any copper sulphide present
(16.30)
The resulting ferrous sulphate is then re-oxidised by further bacterial action
to ferric sulphate, together with the elemental sulphur, thus
(16.31)
This work confirms earlier research 27 in which the nutrient and environ-
mental requirements of bacteria used to accelerate the leaching of metal
sulphides were studied.
species it seems probable that industrial use will in due course extend to the
extraction of gold, silver, copper, nickel and other metals from their pregnant
solutions. 30 An outline of the methods now in commercial use is therefore
given at this point, together with some considerations affecting choice.
To the extent that mineral particles adsorb ions from the aqueous phase of
their solid/liquid system such particles are solid ion exchangers, a reaction
relied on in preparing (conditioning) an ore pulp for flotation. The present
discussion is not concerned with operations where low-solubility species
remain substantially unchanged throughout their concentration, but with
ion exchangers which act as intermediaries in the transfer of ionised com-
pounds, and which are not constituents of the original ore.
The solid ion exchangers used in mineral processing are resins chosen for
their stability (resistance to swelling), toughness when exposed to abrasion or
breakage, loading capacity and ionic sign. "Loading capacity" is definable
as the number ofiogenic groups per specific volume of weight or ion exchanger
and apparent or effective capacity as the number of exchangeable counter-ions
in that volume. The cation exchangers include sulphonated coals and sul-
phonated, carboxylic and phosphoric resins. Anion exchangers include a
range from weakly to strongly basic resin matrices.
The operating cycle in IX commences with the loading of the exchange
sites by ions drawn from the ambient solution, which must in its turn be cap-
able of receiving the ions then displaced from these resin sites. Exchange
continues till the Donnan equilibrium is reached in the semi-permeable resin
bead, at which point the work required for further adsorption exceeds the
residual electrostatic force available to procure more penetration. Choice
between competing ions for a capturing resin site is determined by the relative
sizes 2.:1d valences of these ions, and by the relative concentration (availability)
of the attracted species. The chemical nature of the ionized groups affects
resin capacity. A weak acid such as carboxylic (-COO-) is only ionised at a
high pH, and at a low pH combines with hydrogen to form the -COOH
group. A strongly acidic group such as -S02 remains ionised at a low pH.
Several methods of loading the ion exchangers are in industrial use. In the
oldest, the pregnant liquid ("royals") is separated from the sands in the pulp
after leaching is complete by decantation or filtration. This may be followed
by further clarification so as to reduce the entrainment of adulterating solids
in the end product and to avoid masking the surface of the resin beads with
slimes. The pregnant aqueous liquor with its load of B-type ions now per-
colates downward through a vessel such as the IX column shown in Fig. 209. 31
This column is packed with resin beads loaded with ions of type A. As the
solution enters it deposits its B-ions in a narrow zone near the point of entry,
and takes up the equivalent number of A-ions from the resin. As the
solution continues to percolate downward the resin becomes progressively
loaded and the zone of exchange travels through the column until the resin
bed approaches saturation. This stage is termed the "break-through point",
and B-ions now begin to appear in the effluent, though the lowest zone is not
yet fully loaded. To avoid risk of loss, a series of three IX columns is usually
kept on stream. The first is switched out when break-through point is reached.
New pregnant solution is then led to the second column, the original third
452 Mineral Processing-Chemical Extraction
Sem i-pregnant to
nex column
1
Resin
High value
barrel'
Barren •
Re turn tank
Back wash In
Pr egnant storag e
Fig. 209.
becomes the second, and a freshly prepared column is switched in behind it.
This arrangement allows three in-line columns to be continuously loading
while a fourth one is being regenerated (unloaded and re-activated).
When loading of the first column is stopped, its resin is not usually satu-
rated. Some channelling of the bed may also have occurred, leading to
incomplete contact between beads and liquor, and if clarification was incom-
plete some surfaces may have been masked by slime. The next stage is
usually a stirring of the bed by rising water sufficient to scrub the beads
without flushing them out. At the same time, deposited slimes are detached
and flushed out. This operation, called "backwashing", which also removes
any residual undrained pregnant solution, is sometimes omitted.
The next stage is elution, in which the captured ions are displaced from their
sites when a suitable chemical solution, the eluant, is flushed through the
column. On completion of this operation the resin sites are restored to their
original state and the beads are said to be regenerated. The now pregnant
eluate is run off either to storage or to precipitating treatment and the column
is given a further backwash to remove its residual eluant. It is then ready to
go on stream once more.
The chemical constitution of backwash water (which may be acidified) and
of the various eluants used (re-cycling, new and special) are specific to the
problem of displacing captured ions from the resin and re-activating its
sites. During their working lifetime, the beads become progressively
loaded or "poisoned" by undesired ions which have been picked up from the
Mineral Processing-Chemical Extraction 453
pregnant solution and which are too firmly held to respond to the eluant
used for normal regeneration. This build-up reduces the loading capacity and
is dealt with for most poisons by a periodic special regeneration. Some of
these shorten the working life of the beads. The poisons are of two broad
types, chemical and physical. Chemical poisons are ions which adsorb to
resin sites too firmly to be removed by standard elution. Physical poisons
(polymerising colloidal silica and organic, oily or fungoid and bacterial
accretions) may inhibit the diffusion of pregnant solution into the beads.
When this goes too far for efficient operation, the resin bed is replaced.
The final stage of IX, recovery in solid form of the values from the eluting
solution, is performed by chemical methods which yield the desired precipi-
tate. This is settled out, filtered and dried for despatch.
One variation on the use of static beds is in use in some American plants.
The pregnant liquor is thoroughly clarified, as no slimes can be tolerated.
It then is fed to adsorption. For this there are three groups of three columns,
two of which are on stream while the third is used for elution. The tenth
(backwash) column receives saturated resin from anyone of the six columns
of stream as it becomes saturated. It can also return resin to any of these
columns. In addition, in one plant, there are two spare columns used for
periodic regeneration of poisoned resin. In operation the pregnant solution
is split between two parallel lines for adsorption. When the first column is
saturated it is cut off and water-flushed. Its resin load is then pumped to the
backwash column. Stripped resin is then pumped into the empty column
from the first fully eluted column. Within sixteen minutes the regenerated
and re-loaded column is back on stream, this time as the third in line of
adsorption. The loaded resin is thoroughly washed in the backwash column
and then transferred to the eluting series of columns to be stripped. Among
the operating advantages claimed for this method are avoidance of flow of
pregnant liquor, better resin loading, reduced bulk of eluant and a higher
concentration ratio of "pregs" to eluate.
An entirely different approach is made in the resin-in-pulp method of IX
(R.I.P.). In this, the coarser sand particles are removed by repeated classi-
fication before the pulp, diluted to some 10 % solids, is allowed to make con-
tact with the resin. It then flows through a series of rectangular cells, called
"banks" in each of which a wire basket loaded with beads is jigged up and
down with sufficient force to dilate and contract its load. With fourteen
cells in line seven are adsorbing ions from the pregnant pulp, five are under-
going elution, one is receiving preliminary backwash and one post-eluting
backwash.
Solvent Extraction
This term has, unfortunately, more than one technical use in mineral
processing. Broadly, it applies to any process wherein chemicals in a liquid
phase selectively dissolve a designated mineral from its ore. In current
practice the usual meaning is "liquid-liquid extraction", a form of ion
exchange. One definition reads:
454 Mineral Processing-Chemical Extraction
Aqueous pregnant
liquor (feed) ~ r Barren organic
solvent ~------,
Mixing Treatment
t
Aqueous Extract~Separatin9· Treatment _ _ Organic
(raft inate) Extract
I
New Chemical : l J
11 CStrlPPln9 Treatment
Aqueous SIUrrYj
Aqueous
Chemical Solution ~
. 1 Concentrates
the relative densities of the liquids, the pH of the aqueous phase, temperature,
viscosity, absence of colloidal silica or other slimes and surfactants, and inter-
facial tension. The organic phase fll"st concentrate the values into a small
bulk while leaving unwanted impurities in the aqueous phase. The chemical
selectivity involved depends on the specific formation of extractable complex
ions, e.g. U0 2(S04)2', and on the insolubility of the reacting agent and the
reaction product in the aqueous phase. Extraction of uranium by tertiary
amine sulphate proceeds thus. 33
Ion-pairs such as that in this equation are neutral complexes which react
chemically as single units. Only oxonium systems are at present used in
mineral processing. Stripping is performed by aqueous extraction with some
pH modification. Selectivity between the phases may be reinforced by the
use of salting-out electrolytes in the aqueous feed. These are salts which
contain the anion of the extractable species and a non-extractable cation.
The organic solvent, or diluent, is not usually a reactant but rather a vehicle
for the reacting chemical used to transfer the reaction product from the
aqueous phase to the organic one in which it is more strongly soluble. In
ion-pair extraction the whole organic phase may be employed in this way.
In addition to the reactions thus far considered a "modifier" is often used to
produce synergism. Synergism has been defined 33 as the "co-operative
effect of two or more extractants that exceeds the sum of the individual
effects". This phenomenon, though widely exploited, is not yet fully clarified
by research.
As in the case of solid IX resins, there is a maximum loading of the extracting
solvent· above which its reacting chemical Is cannot take up further ions.
The proportioning of pregnant liquor and solvent during the mixing stage
must be sufficient to avoid such saturation.
The term "non-aqueous leaching" has been tentatively suggested34 to
describe techniques in which direct attack by an immiscible liquid is made on
solid mineral particles either dry or in aqueous pulp. This does not neces-
sarily involve ion-pair exchange since the sought value is characteristically
insoluble in water. This method of extraction is at present at the research
stage, and poses such economic problems as reagent cost and specially resistant
materials of construction of an industrial plant. Among proposed tech-
niques are direct solvation of sulphur by dimethyl disulphide. Other extract-
ants under test include liquid ammonia, liquid sulphur dioxide and liquid
chlorine.
References
1. Lord Fleck. (1962). Nat.ure, July
2. Thompson, P. F. "Dissolution of Gold in Cyanide Solutions." Trans. Electro-
chern. Soc., 91, p. 222.
3. Barsky, G., Swainson, S. J., and Hedley, N. "Dissolution of Gold and Silver
in Cyanide Solutions." Trans. Amer. lnst. Min. Metall. Engrs., 122, p. 660.
456 Mineral Processing-Chemical Extraction
References-continued
4. McLaurin, R. C "Dissolution of Gold in a Solution of Potassium Cyanide."
i. Chem. Soc., 63.
5. Rose, T. K., and Newman, W.A.C The Metallurgy of Gold, Griffin & Co.
6. Dorr, J. V. N., and Bosqui, F. L. (1950). Cyanidation and Concentration of
Gold and Silver ores, McGraw-Hili.
7. Clemes, A. "Modern Metallurgical Practice on the Witwatersrand." i. Chem.
Soc. S. Africa, 46.
8. Roberts, E. J. (1960). "Countercurrent Decantation, When and Why."
Trans. A.l.M.M.E., 217.
9. Goldblatt, E. (1956). "Recovery of Cyanide from Waste Cyanide Solutions
by Ion Exchange." Indust. Eng. Chem., 48.
10. Eicholz, G. C, and Josling, C A. (1963). Dept. of Mines Tech. Bull., Canada,
TB 43, March.
II. Hedley, N., and Tabachnick, H. (1958). "Chemistry of Cyanidation." Min.
Dress Notes, American Cyanamid Co.
12. King, A., Clemes, A., and Cross, H. E. "Treatment of Gold Ore Containing
Pyrrhotite at the Sub Nigel Ltd." Trans. I.M.M. (London), 56.
13. Bitzor, E. C, and Nines, C B. "Some Milling Problems at the Raub Australian
Plant." Eng. Min. i., 141.
14. Lake Shore Staff. "Milling Investigations into the Ore as occurring at the Lake
Shore Mines." Trans. Can. I.M.M., 39.
15. Matthews, O. "FluoSolids Roasting of Arsenopyrite Concentrates of Cochenour
Willans." Trans. Can. I.M.M., 52.
16. Parker, O. J. (1957). Proc. Aust.I.M.M., Part I, June.
17. Tait, R. J. C. (1961). Can. Min. & Met. Bull., April.
18. Edmonds, H. R. "Appl. of Charcoal to the Precipitation of Gold from its
Solution in Cyanide." Trans. I.M.M. (London), 27.
19. Lewis, and Metzner. (1954). "Activation of Carbon." Indust. Eng. Chem.,
May.
20. von Bernewitz, ·M. W. "Charcoal as a Gold Precipitant in Conjunction with
Flotation." Eng. Min. i., 141.
21. Putnam, G. L. "Chlorine as a Solvent in Gold Hydrometallurgy:' Eng. Min.
i., 145.
22. Bell, H. D. "Chromium in Cyanide Solutions." i. Chem. Soc. S. Afr., 36.
23. Barsky, G., Swainson, S. J., and Hedley, N. "Chemistry of Cyanidation."
Am. Cyanamid Co. Tech., Paper 21.
24. Head, R. E. "Physical Characteristics of Gold Lost in Tailings:' Trans.
A.I.M.M'£', 134.
25. Welch, A. J. E. (1957). "Extraction and Refining of the Rarer Metals."
Symposium I.M.M. (London), March.
26. Sutton, J. A., and Corrick, J. D. (1963). Min. Eng. (A.I.M.M.E.), June.
27. Razzell, W. E. (1962). Can. Min. & Met. Bull., March.
28. Mitchell, J. S. (1956). Min. Eng., Nov.
29. Forward, F. A., and Vizsolvi, A. (1963). Ethylene C/ycol Leach Process, etc.
6th I.M.P.e. Congress, Pergamon.
30. Everest, D. A., and Wells, R. A. (1963). Undeveloped Potential Uses of Ion-
Exchange in Hydrometallurgy. 6th I.M.P.e., Pergamon.
31. Ayres, D. E. R., and Westwood, R. J. Uranium in S. Africa 1946-56. Sym-
posium, Vol. 2.
32. Perry, J. H. (1950). Chemical Engineers' Handbook. 3rd ed., McGraw-Hill.
33. Bridges, D. W., and Rosenbaum, J. B. (1962). Metallurgical Applications of
Solvent Extraction., U.S. Bureau of Mines, I.C8139.
34. Lewis, C J., and Drobuck, J. L. (1963). Min. Eng. (A.I.M.M'£'), Nov.
CHAPTER 17
History
I
Reagents
for adjusting Dispersants
pH
1 t
Conditioning
l ~
Conditioning
Activators
Depressants
j! Collector(s)
Conditioning
II
Flotation
Frother(s)
1
Concentrate
t
Tailing
(Mineralised Froth) (Impoverished
Pulp)
was missed. The first collector not based on oil (x-naphthylamine) was
patented in 1917 by Corliss. In 1922 the use of cyanide to depress sphalerite
during the flotation of galena was patented by Sheridan and Griswold, and in
the next year or two commercial flotation of these sulphides in distinct
successive operations (differential flotation) was achieved by chance. Up to
that time only bulk flotation of the sulphides was possible, and separation of
the various metals in the float was made by expensive smelting methods. One
account of the discovery of a differential method begins with the laboratory
separation of lead and zinc sulphides, attributed by the workers to a change in
reagent dosage. A plant was built as a result, and failed to reproduce these
results. Close checking showed that the laboratory cell used a brass impeller
to agitate its pulp. Similar impellers were fitted in the plant and the process
became a success. The reasons are today fully understood and are discussed
later. Other milestones in progress are the Sulman and Edser patent of
fatty-acid soaps in 1924; Keller's discovery of xanthates in 1925; and Whit-
worth's development in 1926 of organic di-thiophosphates as collectors.
Those early days of groping are vivid memories of many mill men still
interested in the subject, but flotation is no longer the sport of empiricism. It
has become a science applied industrially to processes as far apart as the
preparation of special flour for invalid diet and the waterproofing of sparking-
plug terminals. The early years, and the machines and struggles have
been placed on record by Rickard, Pryor, Diamond, Hines, Crabtree,
Vincent and others.
Alternatively, it can be washed over the sides of a conveyor belt adapted for
this variation of the flotation process.
Surface Changes
Most mineral surfaces, when freshly cleaved, are readily wetted and prefer
water to air. After a period of exposure, however, some minerals sorb
~62 Mineral Processing-Principles of Froth Flotation
The surface energy of the phases in contact in the flotation system air-water-
mineral determines the relative attraction between the various species of
particle in the pulp toward either the air or the water phase, or in other words
their degree of floatability. For a liquid the surface tension y can be measured,
but for solids its value can only be indirectly assessed. For a liquid it is the
force which the surface on one side of a line 1 cm. long exerts on the other
side of this line, expressed in dynes/cm. Surface energy is the product of y
and area, expressed in ergs.
Surface energy is also definable as the work required at constant temperature
~64 ."vfineral Processing-Principles of Froth Flotation
and composition of the liquid to increase the surface area by 1 cm 2 • This repre-
sents the work required to bring additional molecules to the surface against the
attraction of surrounding molecules in the liquid body. Since the kinetic
energy of these molecules rises with temperature increase while inter-attrac-
The pressure within a curved liquid surface is higher than the external
pressure. A soap bubble of radius r with surface free energy 4 7Tr2y under-
goes change when r is decreased by dr to 87Try'r. The tendency to shrink
is balanced by the pressure difference LIP across the film. For an N-bubble
The angle across the water phase is called the contact angle 0 (Fig. 212). This
angle is the resultant of three tensions (T). The contact angle (0) is normally
expressed in Young's equation as
yS = ySL + yL cos 0 (17.1)
where y is the free energy per unit area and Sand L are the solid and liquid.
This was originally stated by Young in words rather than symbols. 9 Since y'S
is more or less unmeasurable the equation as stated above is of limited
practical use. It can also be written in terms of work (W)
W SL = yS + yL - ySL (17.2)
By combining 17·2 with Dupre's equation the expression is obtained
W SL = yL (1 ...l- cos 0) (17.3)
which shows the equlibrium for a liquid at rest on a solid. The righthand
components of Eq. 17.3 are measurable and can therefore be used to calculate
the force of liquid adhesion to a solid surface. A fuller discussion of contact
angle is to be found in research literature, (e.g. by Leja and Poling. 10)
The incidence of hysteresis on the measurement of 0 has been studied by
Gaudin et 01.11 This hysteresis is definable as the difference between the
maximum and minimum values of 0 under stated conditions of measurement.
It expresses the resistance to movement of the wetting perimeter and
depends on the direction of spread-gas displacing liquid at the solid surface
or vice versa. Its magnitude depends partly on the rate of movement of the
3-phase perimeter and is up to 30 % higher with static than with dynamic
friction. It is also affected by roughness of the solid surface and the extent
to which it has been modified by adsorption.
Contact-angle measurement, despite certain defects, not least of which is
the difficulty of obtaining suitably large plane surfaces of mineral, provides a
powerful investigating technique. For the given surface condition of the
specimen 0 is the index of surface energy. If (} is nil, the mineral will not float
under test conditions. Even a contact angle of a few degrees indicates that
there is some floatability. 0 is not a characteristic of the mineral phase in a
conditioned system, but of an aerophilic organic group in the molecules of the
adsorbed chemical, which is called the collector agent. FloataLility is
determined by the surface coating, not by the substrate. The contact angle of
all minerals conditioned by (i.e. carrying a monolayer of) potassium ethyl
xanthate (a collector described later) is approximately 60 0 , regardless of
whether these minerals are copper, lead, zinc, etc.
The contact angle method of establishing the conditions under which a
given surface becomes aerophilic is open to certain criticisms and experimen-
tal difficulties. Chief among these are the problems of obtaining a representa-
tive crystal of the required mineral of a sufficiently large size (over 0·5 cm. 2 in
plane area) which is also truly representative. Next comes the question
whether the finally prepared surface is still characteristic after its intensive
polishing with abrasives under water. If the mechanical action used distorts
the surface lattice or even produces a Rei/by layer, the evidence as to induction
and adhesion of air is suspect. A further disadvantage is the static nature of
466 Mineral Processing-Principles of Froth Flotation
the test system produced, which differs vitally from the dynamic one in which
the mineral particle must adhere to a coursing bubble (one rising toward the
surface of the test fluid). Finally, the need to preserve complete cleanliness
makes progress tedious if a new surface is prepared many times when such
variables as pH and reagent concentration are being studied.
Two newer techniques have been developed, each of which overcomes most
of these experimental difficulties while preserving the essentials of basic
control. The first of these is the bubble pick-up method. In this small
quantities of the mineral which is to be tested are crushed, sized to the
required mesh range, and cleaned by shaking in distilled water. The first
reference to its use is in a paper by Cooke and Digre. 12 An air bubble was
- Water - -
-(w)- -
- - ---
pressed against particles and then lifted. The quantity of grains adhering
was noted and it was found that correlation with contact angle tests on similar
material was good, while the pick-up method was more sensitive to changes in
reagent concentration than was measurement of 8. Improvements made by
Sun and Troxell 13 start with the sizing, cleaning and placing in 200 ml. of
distilled water of 0·5 g. of material. After adjustment of test chemicals and
allowing the determined conditioning period for reaction to occur, the par-
ticles are swirled to the centre of the beaker, pressed on by an air bubble held
in the concave end of a glass rod. This is then raised gently, moved to a clear
part of the beaker and tapped. The particles which drop are then examined.
Those so strongly aerophilic as to remain attached are usually visible. With
the bubble pick-up technique it is possible to relate changes of reagent and
pH to all stages between non-attachment and strong attachment of particle
to air-water interphase, and to make a series of tests fairly rapidly on a small
number of particles.
In the Hallimond tube, a modified form of which is shown in Fig. 213, the
particles are held on a porous surface of sintered glass at the bottom of the
tube containing distilled water and the testing chemicals. A stream of air
bubbles is blown upward through the sinter, and any particles adhering to
these coursing bubbles are floated. On emerging, each bubble explodes and
its load of particles slides down to the receiving pocket from which they are
retrieved as required. This apparatus marks an important further advance.
Mineral Processing-Principles of Froth Flotation 467
Levitation of particles is fully dynamic, and the conditions are to some extent
similar to those in a commercial flotation cell. A small sample is treated, and
the floating and non-floating fractions can be separately weighed and ex-
amined.
The only factor not brought into the test work by the Hallimond cell is the
influence of the frothing agent and its associated froth column. This can for
many purposes be left out of the tests until optimum conditions for levitation
have been established. Frothing can be studied under the same rigorous
conditions of cleanliness in a simple apparatus devised by the author 14 . A
sample of five grams or more can be treated, and given controlled flotation
with any desired degree of thoroughness in removing a true mineralised froth.
The wettability of small samples of powdered ore can also be assessed by the
SI NTERED SUPPORT
OF POROUS GLASS
high enough () to float without aid from collector agents. The greater the differ-
ence between the natural or induced contact angles of the minerals in the ore,
the more controllable is their selective separation by flotation. Among those
examining surface activity in its fundamental aspects are Moilliett l5 , Adam,16
Sutherland,s and Gaudin. 5
Activation Theory
In flotation technology activation is the development on the surface of a
designated mineral of selective reaction, usually attraction into the air-water
interface. This renders the activated particle more eager to be drawn into a
coursing bubble, rise and persist as part of a selectively mineralised froth.
To induce such changes reagents called activators are used. Other chemi-
cals, used to work to the opposite purpose are called wetting agents or depres-
sants. Both types are surface-active agents or surfactants, capable of influ-
encing conditions in the air-water-solid triphase. The reagents used in the
conditioning of mineral surfaces are for the most part charged with inorganic
ions (+ or -) chosen for their ability to increase the difference between the
adsorptive powers of the various species in the pulp with respect to the
collector agent used to gather the preferred particles into a mineralised froth.
In studies of the attachment of collector agents to minerals Plaksin 10 found
the tenacity and density of fixation of xanthate to depend on the structure of
the crystal lattice and the cleavage features developed during grinding. Sul-
phide minerals have heterogeneous surface properties which attract non-
uniform distribution of xanthate and give irregular response to the stimulation
coming from the aqueous phase. The electro-chemical potential of a species
may vary widely (e.g. between 0·2 and 0·75 volts for galena). Heterogeneity
in a continuous surface can lead to preferential "spotty" attachment to air,
preferential adsorption of a surfactant or localised patches of oxidation.
The same sulphide surface has been found to have both anodic and cathodic
areas. In the case of non-sulphide minerals (e.g. fluorite, calcite) the crystal
structure influences both the action of gases from the pulp and surface inter-
action with reagents. Even a pure mineral compound, prepared syntheti-
cally, exhibits a variety of responses to surface modification. Ore minerals
treated industrially are not pure and the study of their reaction to treatment
must be partly empiric. Plaksin's work has been summarised by Rogers 3 as
showing (a) that in the absence of oxygen the fresh surface of a sulphide
mineral is wettable to some extent; (b) oxygen is adsorbed from water in
preference to other co-existing gases; (c) adsorption occurs in stages; (d) first
oxygen is adsorbed, next xanthate is fixed. With non-sulphides such as
metal oxides, silicates, carbonates, sulphates and halides much weaker
adsorptive forces are at work than the chemi-sorptive bonds between sur-
factant and metal sulphide. Electrostatic and Van der Waals forces provide
the weaker and less selective adsorbing mechanisms for carboxylic collectors.
Metal sulphides are the most (;ommonly floated ore minerals. They are
only slightly soluble in water, but most of them oxidise superficially in moist
air or oxygenated water. Some are lustrous, some dull; some are good con-
Mineral Processing-Principles of Froth Flotation 469
ductors and others poor; some are tough, some brittle, some soft. Despite
this lack of common physical qualities, all float readily with a suitable col-
lector. The reason is not found in their common factor, low solubility.
When sufficiently oxidised to exhibit relatively high solubility, the metal ions
then generated precipitate the collector, thus wasting this expensive chemical
and diverting it from the work of surface modification. High solubility
would therefore be a most undesirable quality. Indeed, when it is encountered
(through partial formation of oxy-acids of sulphur at the surface by oxygen
attack) it must be neutralised as a preliminary conditioning stage. In most
minerals the metal-sulphide lattice is strongly bound together by internal
ionic forces. It is then immune from attack across the interphase by sol-
vating counter-ions at normal (near neutral) pH values in the pulp. A few
unstable sulphides (notably pyrrhotite and marcasite) readily take up external
oxygen, thus producing chemi-sorbed surface layers of sulphite and thio-
sulphite. This raises the acidity (reduces the pH) of the pulp to some extent,
and robs it of oxygen. The role of dissolved oxygen in activation of a sul-
phide surface is controversial.
The argument which requires a chemical for flotation holds that collector-
attachment is a stoichiometric reaction across the interphase. Wark, Cox,
Sutherland, and others 8 postulate an ion-exchange process as the attracting
mechanism. The chemical hypothesis requires incipient oxidation, and there
is considerable test evidence for the necessity of oxygen in the pulp as a pre-
requisite of collector-coating. If oxidation is considered as de-electronation,
with the removal of one or more electrons from the oxidising lattice point, a
mechanism is provided for attracting a negative ion. The purpose of this
attraction is considered later. There is at present insufficient evidence for
Taggart's claim of chemical oxidation, but much that suggests the need for
conditions which confer positive charge on a good percentage of discontinuity
lattice points at the appropriate stage of conditioning. This suggests that the
function of the hydrogen ion in the pulp is to promote an attracting potential
at the surface of each specific mineral within a certain pH range or below a
specific pH value. Activation is facilitated in accordance with the inverse
solubilities of the metal sulphides. Those of mercury, silver and copper
effectively replace zinc at the surface of sphalerite.
Study of the flotation of corundum by Modi and Fuerstenau 17 shows that it
can be regulated by pH control of the mineral surface charge, suggesting a
close connexion between ion association, electrokinetic behaviourandfloatabil-
ity. The adsorption of collector molecules is at its maximum at the iso-
electric point of a mineral, and therefore its floatability with that collector.
The electrical character of the surface is determined by two potentials, electro-
chemical and electrokinetic. With excessive transfer into solution of one
type of ion (positive or negative) the electrical equilibrium is disturbed and the
surface becomes charged. Ions of opposite charge now find it harder to
escape from the crystal lattice into the liquid phase. Ions already in solution
are drawn to the vicinity of the surface to balance its charge and an electrical
double layer forms. Its inner component is due to the charged surface and is
spread on, not into, the mineral while outer counter-ions are more mobile,
being stirred by thermal vibration in the liquid. Counter-ions next the
470 Mineral Processing-Principles of Froth Flotation
surface move with the mineral while those further out diffuse away, setting
up the zeta potential m. If a change in the magnitude of , is accompanied
by a change in its sign, then the ions of reagents at work in the pulp are poten-
tial-determining and can penetrate to the inner layer of the zone of shear and
thus influence adsorption to the mineral surface. Xanthates have been
shown 29 to affect the electro-chemical potentials of conducting minerals.
The structure and stability of hydrated surface layers depends on the nature
and interplay of the phases in contact. The process of hydration affects
solubility, attachment of particles to bubbles and coagulation. There is a
tendency in Russian research to analyse flotation in terms of surface hydration.
Wettability and change in floatability due to adsorption of heteropolar
substances can be investigated by contact angle tests. Attraction of dipoles
of water to surface lattice points is due to free atomic or molecular force
and surface hydration increases with the adsorption of hydrated (aquated)
ions. If during comminution strong electrostatic forces are set up the mineral
surface tends to be strongly hydrated because of ruptured bonds, and un-
compensated atomic and ionic charges. Where such charges are small the
hydration is slight (e.g. with graphite, sulphur, molybdenite or talc). Such
minerals have weakly adherent bonds across the cleavage slip planes and are
characterised by native floatability, air clinging to natural cleavage planes.
Gas adsorption considerably lowers the hydration of a mineral surface.
Multivalent cations appear to be hydrated by from six to eight molecules
of water, but anions attract a much weaker hydrating sheath. If dissolution
of a mineral constituent is to occur the energy of hydration must exceed the
lattice energy. In hydrated sheathing, the water molecules immediately in
contact with the true ion appear strongly oriented, the effect de~reasing
outward. Polar molecules become considerably hydrated and non-polar
ones only slightly. This hydration is exothermic.
Physical adsorption due to Van der Waals forces of some 3 to 4 calories/
mole is unstable, not particularly selective and is easily reversed. Multi-
molecular sorbed layers are possible. In chemisorption, owing to inter-
atomic action, the heat of adsorption is from 20 to 50 times higher, stable
attachment greater and de-sorbing harder to accomplish. Chemisorption
produces lower solubility of the resulting compounds at the crystal disconti-
nuity lattice. It is aided when the sorbed ions have ionic radii similar
to those in the lattice. Physical adsorption can overlie chemisorption.
It is easy to confuse electro-adsorption with the first stage of chemical
action. Many physical chemists consider that adsorption must in any case
precede full chemical action. While the possibility of chemical action in
surface modification is clearly established as an essential energising activity
for many minerals, ion exchange at a surface monolayer is a leading factor in
the work of collecting and modifying flotation agents.
The chemicals added in small and controlled quantities to the pulp before
or during flotation are, broadly, of three types--collectors, modifiers and
Mineral Processing-Principles of Froth Flotation 471
frothing agents. The division between these may be blurred, in which case
control of selectivity suffers.
Collectors, sometimes called promoters, are surfactants which preferentially
adsorb to specific types of particle surface, and by so doing render part of the
surface sufficiently hydrophobic to cling tenaciously in any available air-water
interface until finally removed as part of a mineralised froth when the pulp is
aerated.
Modifiers (regulators, activators, depressants and dispersants) are reagents
which intensify the selectivity of collector action. This they achieve in
various ways discussed later, the general effect being either to increase or to
decrease the hydrophobicity of a specific surface.
Frothers are reagents which tend to stabilise coursing bubbles as these
emerge from the pulp during its aeration, and which facilitate the retention of
specific minerals in a blanketing layer of mineralised bubbles. This layer is
removed continuously from above the pulp, bearing with it the desired con-
centrated mineral/so
Collectors act either by forming distinctive monolayers on part of the
mineral surface or by forming transition phases at the surface of no specific
composition or distinguishable nature. Modifiers act by increasing or
diminishing the attraction for collector ions through the electrical double
layer which surrounds each particle.
Collectors
A collector must sorb to the selected mineral surface so as to cover it
partially with a film which is water-repellent. Success in achieving this
specific sorption is essential if the particles thus modified are to be removed
from the pulp as a mineralised froth. It is commonly necessary to develop
the latent differences between the various mineral species before the collector
is added, in order to help this culminating action. This is done in a series of
conditioning steps, in which some of the minerals are caused to lose their
attraction to the collector, while in others it is increased. The delicate reac-
tions by which this differentiation is accomplished are caused by the addition
of small quantities of surface-active chemicals, sometimes in quantities
amounting to only a few grams per ton of ore. Technically, the mill operator
is more likely to maintain efficiency with "starvation" quantities than with
over-doses of these expensive chemicals, and economically he is concerned
to keep down costs by using them thriftily.
In the early days of flotation, oils were the only collectors used, and little or
no regard was had to their composition. Operations were in consequence
erratic and inconsistent. Since oil is also a somewhat inefficient frothing
agent, it was not possible to maintain separate control of concentration of
collector and texture of froth, though this is important in plant practice.
Oils are still used in flotation, and usually contain refinery sludges which re-
main after distillation and "cracking" with sulphuric acid. Such residues are
cheap and contain alkyl sui phonic and sulphuric acids. Operators using these
oils draw them from one source, thus assuring a fairly consistent composition.
472 Mineral Processing-Principles of Froth Flotation
Xanthates
These are the salts of xanthic acid, a sulphydril compound, which in theory
would be formed by the reaction
R . OH + CS 2 = R . OCSH
II (17.4)
S
ROC-SNa
II
S
drate. The R group in the case of sodium ethyl xanthate is C2H50, and the
structural formula of the salt is
H H
I I
H-C-C-O-C-S-Na
I I II
H H S
Non polar group: Polar group
'-----y---~ ~ '----y--~
The xanthates are prepared by reacting an aliphatic alcohol (or in the manu-
facture of rayon, cellulose) with carbon disulphide and a caustic alkali.
C 2H 50H -L NaOH + CS 2- -
C 2H 5 - OC - SNa -L H 20 (17'5)
II
S
If this somewhat unstable compound oxidises, it becomes dixanthogen
Thiocarbonates
These compounds are derived from the carbonate, and have collecting
power for sulphide minerals. Their structural formulae are
Thiocarbamates
In the xanthates the carbon atom connecting the hydrocarbon group to
the polar group is oxygen. In the thiocarbamates the connecting link is
nitrogen. The general formula for potassium diethyl dithiocarbamate is
Dithiophosphates
The general formula of these compounds is:
'"
RO
Here the connecting atom is phosphorus. They are marketed under the
trade name of "aerofloats". They are weaker collectors than the xanthates,
and are sometimes used to float the first mineral of a series, in order to avoid
collecting the less eager sulphides during this first stage of treatment. Again,
xanthate may be used after aerofloat, to bring up the weaker particles of the
mineral sought. This is one form of "scavenging" treatment. The aerofloats
are reaction products of phosphorus penta-sulphide with phenol, alcohol, a
mercaptan, thio-alcohol, amines, and nitriles. A compound with phenol and
alcohol is the aerofloat mostly used, and in the concentrator at Bingham,
Utah, where the aerofloats were first developed, the reagent is made in batches
by a mill hand as required. It is soluble in water and forms almost insoluble
salts with heavy metals. It is more sensitive to depressing action than xan-
thate, and therefore has special value in differential work, particularly when
iron sulphide must be kept down. Aerofloat 15 is a black reaction product
of cresol with 15 ~~ of P2SS, the excess cresol acting as a frother. A-25
carries 25 ~~ of P2SS, and A-31 is similar with an added 6 ~~ of thio-carbanilid.
A-239 is a di-amyl dithiophosphoric acid, neutralised with ammonia and
thinned with 10% of ethyl or iso-propyl alcohol. The ammonium ion may
account for the success of this reagent in floating copper in the presence of
hydrated iron. Aerofloat 241 is a water-soluble form of A-25, used as a
fast-acting collector for the sulphides of silver, copper, lead, and zinc in alkaline
circuit, and as a non-selective collector and frother for sulphides in acid
circuit. A-242, a water-soluble form of A-31, is similar to A-24l but has
Mineral Processing~Principles of Froth Flotation 477
Thiocarbanilide, etc.
The structural formula for this thiourea derivative is
It is only slightly soluble in water but disperses well, and where used is usually
added with the feed to the ball mill. It has selective qualities which have
some application in the flotation of copper and lead from complex sulphides.
The "minerec" reagents are derived from xanthates by oxidation, and have
some application as collectors for such reducing minerals as sulphides. A
third group of non-ionising reagents is the hydro-carbon oils, which confer
aerofilic quality by smearing rather than ionisation at the mineral surface.
Some' of these are marketed under the name "Nujol". Another proprietary
name "Flotagen" is given to mercaptobenzthiazole, the formula being
CO-
N
SH
S
It bases the American Cyanamid 400 series (also 125) and is used in alkaline
circuit for oxidised lead ores. Non-polar oils are bonded by Van der Waals
forces, by thio-compounds inherent or developed in the supplier's oil refinery,
or possibly by long-chain dipolar molecules such as higher alcohols of low
hydrating strength. The commercially supplied oily collectors vary both in
composition and contained surfactants, but tend to be consistent when coming
from some special source.
A completely saturated mineral oil or plant glyceride is not a collector per se.
It depends for its polar groups on impurities, additives such as sulphated or
sulphonated compounds picked up during commercial production, or on
unsaturated organic fractions. Where oil refinery residues are used as
collectors it is usually found desirable to rely on one source and its named by-
product. Hydroxyls, carboxyls, methoxy, sulphur or nitrogen groups may be
active. "Reconstructed", oils into which elemental sUlphur has been incor-
porated by heating to some 250 0 C. or by reaction with concentrated sulphuric
acid at about 50 0 C. (perhaps after the addition of alkali) have some use.
Carbon disulphide or other sulphur additives have been employed.
478 Mineral Processing-Principles of Froth Flotation
11
dithiophosphate, etc.
.r------- I
as collectors
Flotation of
sulphides
Sulphidos
~ Fatty acids,
Flotation of fatty acid soaps
~ non-sulphides or amines as collectors
Non-sulphides
The basis for collector action appears to be the abstraction of the carboxy-
late ion (CO. OH)- by the non-sulphide mineral. This reaction proceeds
slowly, particularly with saturated fatty acids. Gaudin and Toumesac19
found that 24 hours were needed for an approach to equilibrium in the
sorption by barite of laurate. The carboxylic fatty acids are good collectors
for polar minerals which contain alkali-earth metals (Ca, Ba, Mg, Sr),
carbonates and non-ferrous sulphates. Adsorption by oxides and silicates
is weaker. Process water containing a substantial trace of the above alkali-
earth ions must be softened before carboxyls can be used. Sulphonated fatty
acids are less susceptible to direct precipitation.
The saturated fatty acids are homologues of the basic structure CnH2n + 1 •
COOH or CnH2n02. They include, in ascending order of carbon chain
The soap is produced by substituting a sodium ion for the hydrogen of the
carboxyl -C02Na instead ofC02H.
The basic formula of the unsaturated fatty acid mainly used in flotation is
C nH 2n_ sCOOH. The formula of the widely used reagent, oleic acid, is CH3
(CH2)7CH = CH(CH2)7COOH and of its sodium soap or oleate C17H33
COONa.
Writing of the carboxylic collectors Taggart 20 says: "Fatty acids and soaps
are collectors for all minerals which, in water, free an earth-or heavy-
metal ion, or onto the surface of which such an ion can in any way be plated."
The "plating" or re-surfacing ions available in most pulps include Pb, FeN,
Ba, Ca, AI. By their indiscriminate reaction with such common gangue
minerals as quartz and calcite they lower the grade of non-sulphide concen-
trates unless they can be brought under control. " ... it is for this reason that
they have been supplanted for heavy-metal collection by the sulphydrate
collectors, which do not form insoluble salts with the earth metals. But
since no such substitute has yet been found for the earth-metal minerals, the
fatty compounds constitute the principal collectors for many of these. The
carbon content must be Cs or greater, preferably greater than C12 , on account
of the relatively high solubility of the earth-metal soaps of the lower acids,
and it should not be greater than CIS or C 20 , because of the low solubility of
the higher alkali-metal soaps, in which form the collectors are usually dis-
persed .... " The random activating effect of Ca ions is commonly reduced
by softening the mill water used in fatty-acid flotation. Close pH control is
helpful, particularly where the more specific action of oleic acid (as compared
to oleate) is desired.
A cheap source of fatty acid is tall oil or talloel (Am. Cyanamid 708).
That used in flotation is usually a distillate consisting mainly of linolenic acid
480 Mineral Processing~Principles of Froth Flotation
(an impure oleic acid), rosin acids, with some saturated fatty acids (palmitic
and stearic). The composition of tall oil varies considerably. Its collecting
properties for magnetite have been studied by Kivalo and Lehmusvaara. 21
The experimental flotation of hematite ores with an emubion of tall oil and
fuel oil has been reported by Kihlstedt 21 in connexion with progress in the
separation of apatite from hematite. The resins in tall oil act as frothers, and
this action may become excessive in alkaline pulp, particularly on slimy ores.
The tannin compounds in crude sulphite liquor from the wood-processing
industry (the main source of tall oil) are depressants. The reagent used in the
commercial flotation of phosphate rock is prepared by dissolving tall oil in
fairly strong caustic alkali (Cyanamid 708 of the 700 series), whereas the
Kihlstedt reagent, named UMIX, is an emulsification of tall oil, No.2 fuel
oil and a water-soluble alkyl-aryl sulphonate, precisely prepared. This
latter reagent, like one developed by the author and his colleagues on the
basis of reaction between oleic acid and infer alia mahogany soap22 is indiffer-
ent to slime and is not affected by ordinarily hard water.
The general formula of the alkyl sulphates is CH3(CH2),;-S03Na and of
alkyl-aryl sulphonates CH 3(CH 2);CaH; S03Na. Most mineral and vegetable
oils contain oleic acid or an associated compound. Among the commercially
obtainable water-soluble collectors based on fatty acids are the following;
o
R ~S-ONa
11
o
(c) Alkyl sui phonates with amide interlink (R has 15 to 18 CH)
R- CON C2 H4 S03 Na
I
CH 3
Mineral Processing-Principles of Froth Flotation 481
R COOCHS0 3 Na
(g) Alkyl-aryl sulphonates with R of up to 12 C.
R -c::::>- S03Na
An important series of reagents for the flotation of iron oxides, mica,
chlorite, chromite and garnet, ilmenite and talc is marketed by American
Cyanamid in its 800 series. These are anionic suI phonation products of
complex structure, containing such ingredients as mahogany soap and "green-
acids". They are water-soluble or water-dispersible and work best in an acid
pulp (pH 2 to 4).
Oleic acid and its sodium or potassium soaps are used in the flotation of
oxides, halides, silicates, and oxygen-salt minerals, e.g. fluorite, apatite, and
garnet. The relation between the adsorbing mineral and the fatty-acid
collector is not one of direct exchange reaction or simple monolayer formation.
It appears to be influenced by the action of cations such as those of calcium,
barium or iron at or near the surface of minerals such as silicates.
The adsorption of oleate ions on fluorite 18 is influenced by temperature,
being 193 g/ton at 16° C. and 240 at 30° C. With temperature decrease soap
and fatty acid molecules associate to form micelles, and activation suffers.
Micelles also form when the collector concentration exceeds 250 mg. II., but
this tendency is opposed if the oleate is emulsified with a very little pine oil.
Excessive use of fatty acid reduces its collecting power in other ways. It
can be adsorbed into bubbles to a sufficient extent to impede the picking up of
particles, or even to stabilise a more or less unmineralised froth. It can also
compete with the molecules of collector at the mineral surface by masking it
with hydrated micelles.
A mines
These are cationic collectors for non-sulphide minerals, the prefix "amino" or
the suffix "amine" denoting a - NH2 group. They are derived from ammonia
by replacing its hydrogen with an organic radical. The nitrogen core can be
Primary amine
Secondary amine
482 Mineral Processing-Principles of Froth Flotation
Tertiary amine
Quarternary amine
The solubility of amines (e.g. lauryl) is poor in an acidic pulp but increases
sharply above pH 8 to a maximum ionization at pH 9.
The mechanism of collection has been investigated by several workers using
contact-angle and bubble pick-up techniques.
Taggart and Arbiter 23 worked with dodecylammonium chloride on the
minerals wollastonite (CaOSi0 3), calcite, barite and fluorite. Their con-
clusion was that metathesis occurred between metal and collecting cation.
(17.12)
Modifying Agents
The chemical and physical additives discussed in this section are variously
described in flotation literature as regulators, conditioners or more specifically
as depressants, activators, wetting agents, flocculants and dispersants. The
practical requirement from the operator's viewpoint is the ability of the addi-
tive to sharpen selectivity as between the mineral species floated and that
(or those) left in the pulp. Though a slight over-simplification, this can be
considered in terms of a conflict between wetting and drying-out forces.
Surface hydration favours retention in the pulp and air-avidity attachment
of the particle to a coursing bubble followed by seizure in a mineralised
froth. Reaction which increases the hydrated area on the particle favours
wetting. Thermodynamically the spread of one fluid on another, or
the replacement on a solid surface of gas by liquid (or vice versa)
is due to a nett decrease in the total interfacial free energy of the
system. Gravity, mechanical force and chemical change also affect
spreading. The regulating chemicals used in flotation stimulate the desired
surface activity of specific minerals in the pulp. Modifying agents therefore
either reinforce or make possible the adsorption of collectors in the case of
floating species, or aid hydration if they are active depressants. A reagent is
not necessarily "active" in the sense of being measured into the pulp.
In the previous section collector agents were considered in connexion with a
clean and attracting surface. In the laboratory such conditions are readily
attained, but in plant practice this is not the case. Mill water, the dirt and oil
contributed during mining, transport, and, comminution, and the attempts of
all the newly sheared particles to pick up from these additives anything which
484 Mineral Processing-Principles of Froth Flotation
will reduce their surface potential, lowers the free energy of the surface lattices
of the various minerals in the ore. This gives them some degree of superficial
similarity, since they draw their contaminants from a common environment-
the pulp. Even a surface which has escaped random sorption may require
development of its ionic potential before it is adequately reactive. Condition-
ing agents are used to bring out latent aerophilic or hydrophobic qualities at
appropriate stages of the treatment.
The first modifying agency to affect the particle is mechanical. Commi-
nution detaches it from the parent mass, gives it form, surface, and energy due
to lattice deformation. The next is physical. Classification determines the
weight and mass of the particle on release from grinding. The third agency,
simultaneously at work, is surface modification of the new surfaces developed
by comminution. This change proceeds continuously from the birth of the
new surface to the time when it is either in chemical balance with its surround-
ings or removed from them. These surroundings contribute oxygen from the
air, ions from mill water and other minerals in the pulp, and surface-affecting
reagents introduced during treatment. Study of the preceding section
"Collectors" shows that, despite the thousands available, choice is usually
narrowed down to a few xanthates, dithiophosphates and fatty acids. Before
specific collector-action can be assured, there must therefore be equally
specific preparation of the surfaces to which that collector is to be attracted.
This is achieved by the use of modifying agents. One group of these is used
to increase the attraction of the collector toward the surface to be floated:
another group increases the wettability of the surfaces which are to remain
unfloated. A third group removes interfering ions from the pulp water. A
fourth disperses masking slimes from particle surfaces. A fifth plates, or
re-surfaces, selected minerals. The main groups of modifying agents are:
Most of these have dual roles-e.g. a wetting agent (b) does its work by surface
modification in which hydroxyls or other "wetting" ions increase their
Mineral Processing-Principles of Froth Flotation 485
TABLE 33
(After I. W. Wark. "Principles of Flotation", (1938): Australasian Institute of
Mining and Metallurgy.)
Mineral Max. flot. pH Formula
Chalcocite < 14·0 ClI,S
Covellite 13·2 CuS
Bornite IH ClI"Fe3"
Chalcopyrite 11·8 ClI,S. Fe,S"
Pyrite 10·5 FeS,
486 Mineral Processing-Principles of Froth Flotation
the balance between the two concentrations has a profound effect on the
heterogeneous state referred to in this book as the "pulp climate". If the
hydrogen and hydroxyl ions be pictured as a restless tide lapping at the shores
of the'surface lattices in the pulp, and if these surfaces be considered as
having a greater or less electrolytic attraction for these ions, according to the
Mineral Processing-Principles of Froth Flotation 487
TABLE 34
Critical pH Value
Collector and
Concentration
Sphalerite Galena Pyrite Chalcopyrite
(17.14)
r.1
...
~
Eo<
7r)()
- , ::-
~
~
h-
600 ~ ~ ~
r.l ~ r-
t
~
Il. ~
tt J l-
~
~
.0
~ 500 I
~ I I
0
Eo<
0 4«J L J
r.1
~ 1 I
I
~
0 I
0 JOO
rz.
0
II /
Z 200
0
V 1
....
Eo<
-<~ / / 1
/ V II
-
Eo< 100
Z
I'zl
0
/ V /
z .-' ,/
V V
8 0
2 J 4 5 6 1 8 9 10 /I
pH VALUE
Fig. 215. Critical pH, below which mineral floats, using sodium diethyl dithio-
phosphate collector. (After Sutherland and Wark, ibid., p. 119)
~ 200 ~I,-
ofo4 180 ,~ I 1 ~
.
~-
I'zl
:3 160 ~I I ;~~ r- I l _
~
t:
il- r--
~
8~ 140
~ -r- ; ~-
rz.f04 J ~
J
~
-
'I
o~ 120
;1
~
8~ 100
il ~ ~ -
E::~
-<~bD. 80
\ ,~'l ~ 'J/
fo4'" 60 1
~,
~+ ~I T
Z
~ 40 i 0/ <S
~ IJ
Z
o
o
20 ~ ./ / /
» i'- V VI-""
6 1 8 10 II
pH VALUE
Fig. 216. Critical pH-collector relationship in floating sphalerite
(After Sutherland and Wark, ibid., p. 119)
490 Mineral Processing-Principles of Froth Flotation
Process Water
The chief chemical constituent in the pulp is the mill water, together with
its content of salts and ions. If a pulp is being conditioned at 33 % solids (by
weight), the water-solid ratio is 2 to I. If a reagent is added at a concentra-
tion of 0'06Ib/ton of ore, this is 0·02 lb/ton of pulp, a little under 10 g. This
Mineral Processing-Principles of Froth Flotation 491
reagent, like the others added before and after it during a series of condition-
ing operations, must react with a large acreage of mineral surface. In a
big plant many tons of pulp flow through the conditioning section every
minute of the day, so timing and manipulation must be smooth and simple.
The flotation process depends for its efficiency 01. accurate treatment during
conditioning, so steady throughput rate in the grinding circuit is necessary
even if no reagent is added there. It is, however, customary to bring the pulp
to its correct pH value in the closed grinding circuit. If a series of minerals
is to be floated, the pH is adjusted as part of the reconditioning process after
the first mineral has been floated.
Mill water is usually drawn from the local rivers and lakes. The mineral
content may vary between summer and winter in temperate climates or
between the dry and wet season in the tropics. While rainy conditions pre-
vail, the water is less likely to carry calcium, magnesium, carbonate, sulphate,
and other ions than when it is being drawn from deep springs. In a sub-
arctic winter water drawn from below the ice of lakes is de-oxygenated by
algae, and has been found by some operators to produce too brittle a froth
for good flotation unless it receives preliminary re-aeration. This can also
apply to water recirculated from tropical tailings ponds, if vegetable growths
have stripped the water of its air. If the water is pumped from underground,
or even if the ore carries a substantial percentage of moisture from that source,
it may carry metal ions or vegetable acids derived from the rotting timber of
old stopes. These additives are fairly harmless as a rule. If the ore tends to
oxidise and produce metal ions, these may precipitate some of the small
quantity of collector with unfortunate effects on recovery, which depends in
the last resort on all of the reagent being on the desired surface. Unstable
sulphides-some pyrrhotites and marcasites, notably-may de-oxygenate the
pulp, which requires oxygen for the effective sorption of xanthate to the
minerals. The chemical role of the air carried in mill pulps at all stages in
flotation is not entirely clear, but evidence has accumulated as to its im-
portance. Unless threshold oxidation of a sulphide takes place, surface
excitation may be insufficient for the collector reagent to sorb. If oxidation
proceeds too far, direct chemical reaction between metal ions and collector
may occur in the liquid phase, with consequent poor flotation of the in-
adequately conditioned particle. The careless use of oil underground can so
contaminate the ore and foul the mill water as to upset process control in the
mill.
Mill water may depress or activate the pulp constituents if over-charged
with various ions. Those most liable to build up are Cl-, S04-, HCOs-,
C03~, Na+, K+, Ca++, Mg++, H+, and Fe++. When the intake of fresh
water must be restricted the c0ndition of recirculating process water must be
watched and if necessary corrected by appropriate treatment. Salts which
cause water hardness are particularly liable to react with fatty-acid reagents
and form insoluble complexes.
Mill floor washings are liable to carry oil or grease, and must not be
allowed to re-enter the circuit or to foul pumps or sumps connected with
returned water. Frothing agents, which are not appreciably adsorbed by
concentrates, overflow with these mineralised froths. If the water containing
492 Mineral Processing-Principles of Froth Flotation
Conditioning
The collector action is partly controlled by the pH value, but further con-
ditions may need to be satisfied before the collector is added, to heighten the
difference between aerophilic and hydrophilic mineral surfaces. This is
obvious in the case of differential flotation (Fig. 217) where the desired
products are floated successively. When a series of values are floated, a
further conditioning is necessary after each flotation.
Pulp
t
Condition to float
,..------{~ Reagents
Min~ral A
Float
(Mineral A) .c------
!
Flotation --Tail
I
i{R,"g'n~
y'--------'
Condition to float Mineral B
Float ~
Flotation - - Tail
(Mineral B)
I
etc.
There are three possible explanations for the potency of cyanide. It may
sorb to the mineral surface with greater tenacity than that of hydroxyls. It
may remove Fe ions from the surface by complexing them, as it does those of
copper and zinc. Copper is present on many sulphide surfaces, and as it has
strong attraction for xanthates, it must be "complexed" to cupro- or cupri-
cyanide if it is not to render floatable the mineral carrying it. The effect of
cyanide on floatability (contact-angle criterion) is shown in Table 36 and
Fig. 218. This graph shows the demarcation between clinging (0) and non-
adhesion (x) in contact-angle tests of bornite, using potassium ethyl xanthate
as 35 mg/1. as collector. There is a critical concentration of CN- above
which no flotation can occur.
Sulphides.-Soluble sulphides are often used as activators in the condition-
ing of oxidised sulphide surfaces. They can also act as depressants if added
in excess. If S ~ ions are adsorbed to a metal-sulphide surface they increase
the negative charge and also the adsorption of xanthate ions. Since this is a
depressing action, true metal sulphides are floated after straight collector-
conditioning before the residual pulp is further conditioned with a soluble
sulphide. Hydrogen sulphide, like hydrocyanic acid, is a weak acid. The
ionisation of its sodium salt Na2S is largely governed by the pH of the pulp,
as is the case with other weak acids. The ionisation product is
(17.18)
In this form the sulphide ion competes with the collector for sorption. If it
dissociates further, to
(17.19)
160
I~O
120
•
'"
.~
.g
o • •
;1:60
U
PQ '\
\
.
•
Z
40 0\
1\
20
'\'
t-!- II
o \.
• o to
pH VALUE
" 12 13 14
added S ~ ions is determined by its depth, and therefore by the rate of diffusion
into the surface during conditioning. This film is easily displaced by pulp
agitation. Pre-sulphidization of some minerals (e.g. malachite) reduces the
xanthate requirement considerably. Excessive residual sodium sulphide
after formation of the surface film has ceased depresses flotation by xanthate.
Floatability is, however, recovered after re-aeration. The time needed de-
pends on the rate of oxidation of hydrosulphide and sulphide ions in the pulp.
The period of agitation with air present is therefore important in sulphidising
treatment. Oxidation of a surface monolayer of S ~ or HS - takes only a
fraction of a second . Sodium sulphide is therefore added in stages along the
line of flotation cells, to suit the kinetics of the ore's reaction. Excess
sodium sulphide desorbs the collector if the mineral surface is not masked by
fine air bubbles. Soluble sulphides may also precipitate heavy metals and
cause them to activate particles indiscriminately and so lower the concentrate
grade.
Sill/ace oxidants.- Some sulphide minerals show a marked deterioration in
floatability when their freshly broken surface is given opportunity to oxidise
beyond the threshold degree which facilitates adsorption of xanthate. Several
sulphoxy compounds can be present in the zone of shear surrounding a
Mineral Processing-Principles of Froth Flotation 497
particle, apart from those intentionally added to the pulp. There are six
oxides of sulphur (27)-SO, S02, S20a, S03, S207 and S04, and twelve
oxy-acids. 5 Complex thio-compounds are formed by a number of metals.
The factors which affect the degree of oxidation include pulp temperature,
time, aeration, pH, extent of barrier films at the particle surface, and presence
of oxidising agents. The systems formed in surface layers are complex and
vary not only between species of sulphide, but also between minerals of the
same species derived from different geological formations.
Depressants sometimes used include lime, alkali sulphites, permanganates,
arsenates, chromates, ferricyanide, tartrate, acid-phosphates, citrates, oxalates,
and tungstate. They are shown in descending order of effectiveness as
depressant for pyrite in the presence of potassium ethyl xanthate. The
general action of all the above depressants is to neutralise positive lattice ions.
Differential Depression.-When more than one mineral has been recovered
simultaneously by bulk flotation it is often necessary to re-condition the pro-
duct and to re-float one or more constituent species from the mixture. To
achieve this the mechanism of depression may require selective desorption of
collector, with destruction of the original collector followed by specific re-
activation, or by specific weakening of aerophilic quality.
Specific surface oxidation or destruction of a xanthate film has been aided
by the use of permanganate, dichromate or ferri-cyanide. These oxidise the
xanthate to di-xanthogen. The alcohol group of the collector has been
preferentially vapourised by controlled heating prior to re-floating, as in the
separation of a molybdenite float from a copper sulphide sunk fraction. In
the removal of entrained silica from a phosphate-silica float, the anionic
fatty-acid collector is dissolved with caustic soda and the silica is then floated
with a cationic collector. Sphalerite which has been activated with copper so
as to float with galena can be depressed by re-floating the mixed concentrate
after conditioning with cyanide to dissolve the copper.
Sodium silicate. This is much used as a dispersant, for removing slimes
from sulphide surfaces to which they have sorbed. Up to 0'5 lb/ton i~ used
for dispersing quartz. The reagent sold as "sodium silicate" can vary widely
in its silica-sodium ratio, with corresponding changes in its value as a dis-
persant and gangue depressant (Sollengerger and Greenwalt). It also tends
to embrittle flotation froths. The salts are the meta-silicate (Na2SiOa),
disilicate (Na2SbO.j) and orthosilicate (Na4Si04). It may also be
present as hydrated silica. The commercially supplied composition is
mNa20.nSi02 and the n/m ratio or modulus mostly used varies between 2·2
and 3·0. With a lower modulus the pulp tends to alkalise and silicates are
depressed. Above 3·0 a coarsely dispersed hydrated silica is produced.
The dissociation of sodium silicate is complex, and preparation of its
solution for use must be standardised if the content of colloidal silica is not
to vary unpredictably, with consequent random deposition of silica gel.
Ions, molecules and micelles of silicic acid are strongly hydrated and if
adsorbed may have depressant effects. Sodium silicate yields ions which are
potential-determining for silicates and aluminosilicates, which are preferenti-
ally depressed. 29 A small addition activates the flotation of hematite by
sharply decreasing the negative zeta potential, but a further concentration
498 Mineral Processing-Principles of Froth Flotation
raises it again, and different dissociation products of the silicate then become
adsorbed.
Sodium metaphosphate has limited use in calcite dispersion. In the sense
that a dispersant restores the true surface to the mineral being activated it is
an activator, but sodium silicate is regarded by most workers as a depressant.
I t is attracted by common-ion effect to silicate lattices, where it acts as a wetting
agent.
One difficulty in the depression of slimes is the great surface area of these,
when colloidal and clayey. They sorb expensive reagents, contaminate froths
and stabilise them undesirably, and lead to difficulties in the settling of pulps
after treatment. They are rarely troublesome in the treatment of unaltered
ores but may be difficult with deposits which have been exposed to weather,
such as leached and shallow ore bodies and dumps. The methods of handling
them are preventative rather than chemical. If such slimes can be removed
before fine-grinding without undue loss of values, the problem is simplified.
Grinding should be kept to its minimum and the collector agent may be
found to act best if added in the grinding circuit, so that it can sorb before the
particles become slime-coated.
Apart from depressant action and the precipitation of fatty-acid collectors
as insoluble soaps, lime is used in sulphide flotation, inter alia to depress
pyrite. It appears to form a mixed surface film with ferrous and ferric
hydroxides and insoluble lime salts, thus reducing xanthate adsorption.
When it is desired to float pyrite despite interfering lime, ammonium ions can
be added to reduce or prevent adsorption of calcium.
Starch is one of several water-soluble polymers which act as depressants
when adsorbed. Some of the polar groups of its molecule orient toward
the mineral while other remain free and thus increase the hydrophilic quality
of the surface. Starch, and such other high molecular-weight colloids as
albumen, caseinate, gelatine, saponin and quebracho have shown depressing
effects in the flotation of several metal sulphides. The general formula of
starch is (C6HI005)n, with traces of phosphorus and silica. On warming to
dissolve it the giant molecules break down into polysaccharides or dextrins.
With non-sulphide minerals the depressant action of starch is due to hydrated
films on the minerals.
Quebracho is a wood extract containing tannic acid. It is an excellent
depressant for calcium salts in the flotation of fluorite by fatty-acid collectors,
but its concentration must be closely controlled if it is not to react with the
calcium in the CaF2. The chemistry of its reaction is obscure.
Regarded as hydrogen-ion phenomena, activation and depression are
relative to the pH. There are two special aspects of activation not thus dealt
with. In common-ion activation, an ion in solution is attracted to discon-
tinuity lattices in which similar ions are bound, provided excess binding force
is present at the surface of the particle. Thus, if the soluble sulphate of a
metal is added to a pulp which contains particles of the same metal in sulphide
form, the particle surface tends to become more basic, metal ions. being
drawn out of solution and captured at the surface by unsaturated lattice
points. This depressing action can be used to keep down sphalerite when
galena is floated with a xanthate or dithio-phosphate, by adding zinc sulphate
Mineral Processing-Principles of Froth Flotation 499
to the pulp. The zinc cation is attracted into the hydrated sheath which
surrounds the particle and aids depressing action by increasing the population
of accompanying hydroxy Is. Tests 29 show that zinc hydroxide begins to
precipitate on a sphalerite surface at a pH of 5·2 and that at pH 7·4 flotation
has fallen from 35-! % to IO %, even a copper-activated sphalerite being
affected. The zinc-coated surface can be reactivated by using ammonia
to dissolve the hydroxide, or sodium silicate to peptise it. Addition of
common ions sometimes increases the insolubility of the sorbing collector,
and in some cases allows economy in the use of this reagent. There is,
however, danger of direct chemical reaction between sulphate and collector,
resulting in precipitation of the latter.
Activation by Re-surfacing.-The modification of a mineral surface by
adsorption of metal ions has been referred to already. As widely practised
in the industrial flotation of sphalerite it is achieved at the surface of a weakly
floatable sulphide by building on to it a more actively floated metal. The
reagent most used for the work is copper sulphate, CuS04.5H20, and the
mineral to which it is most applied is sphalerite. Zinc sulphide floats weakly
or not at all with the xanthates. It is usually associated with galena and
copper in ores, and traces of the latter are "killed" during the first stage of a
differential float when the lead is being taken, by the uses of sodium cyanide.
After the galena has been floated, the pulp is brought to a pH of 9 or more by
adding lime. This lime is a depressant at such a pH for pyrite, which if
present might become activated in the next conditioning stage. A suitable
quantity of copper sulphate is now added, as a saturated solution. Most of
it, after reacting with residual cyanide, precipitates out as hydrate or carbon-
ate. These salts act as a reservoir and feed copper ions to the pulp to
replace those which now plate to the zinc lattice-ions thus:
If for any reason deactivation of the changed zinc particles is now needed, the
reaction can be driven from right to left by the addition of solvating CN - ions.
In the short reacting period used in a plant, the copper layer is kept as tenuous
as possible, as it is the zinc beneath, not the minute trace of surfacing copper,
which is wanted.
The case may arise of copper ionising into the pulp during grinding of an
ore containing zinc or other attracting sulphides. This would make differ-
ential flotation impossible and would also lead to heavy consumption of
xanthate by direct precipitation. Here, the addition of sodium or calcium
sulphide to the dry are entering the grinding section can sometimes induce
precipitation of the copper as an insoluble sulphide and thus prevent such
general activation. Several metal ions can act in this way. If they then
accept a collector, this may cause the mineral to float. Copper is the most
positive in action of these, and in the fractions of a pound used per ton
of ore is cheap.
Any metal in the electromotive series can displace from solution those
below it. The series is K, Na, Li, Ba, Sr, Ca, Mg, AI, Mn, Zn, Cr, Cd, Fe,
Co, Ni, Sn, Pb, H (hydrogen), Cu, As, Bi, Sb, Hg, Ag, Pd, Pt, Au. The
TABLE 37
RE-SURFACING AGENTS
Cationic:
Cu++ CUS04 0·\-2·0 Activator, Zn, Fe, Co, Ni sulphides Xanthate
Pb++ Acetate 0·\-2'0 Activator, stibnite Xanthate
Pb++ Acetate 0·\-0'3 Activator, halite Fatty acid
Ca++- CaO or (OHjt 0,5-10,0 Depressant for pyrite; activator for silica Xanthate/Fatty acid
Zn l +- ZnS04 0,2-2'0 Depressant for sphalerite Xanthate
Anionic:
0-- Air Depressant, pyrrhotite Xanthate
S03-- Na2S03 0,5-2,0 Depressant, sphalerite Xanthate
S-- Na2S 0·5-20·0 Activator, Pb, Cu oxides; depressant for
sulphides when excessive Xanthate
CN- . NaCN 0·5-\·0 Depressant, Cu, Zn, Fe sulphide Xanthate
Si02 Na2Si03 0,5-2,0 Depressant, gangue slime; activator for
silicates Xanthate/Cationic
COa-- Na2CO:l } 0-5-5'0 Activator, Pb, Fe sulphides Xanthate
\·0-10·0 Depressant, gangue Fatty acid
Organic colloids:
Dextrin 0'\-\'0 Depressants for gangue slimes, especially f Xanthate
Starch 1 carbonaceous slimes l Fatty acid
Lignin 0,2-5,0
sulphonate I
Mineral Processing-Principles of Froth Flotation 501
Four methods of aeration are possible, of which two are widely used. In
the so-called "mechanical" cell air is sheared into the pulp by a submerged
impeller which is also receiving air from an external source. In the "pneu-
Mineral Processing-Principles of Froth Flotation 503
matic" cell air is blown in through diffusing mats or nozzles. These two
methods may be mixed. In the third (little used) method a vacuum is set up
and air precipitates from the aqueous phase on to the most hydrophobic
particles with sufficient lifting power to buoy them to the surface. Some
measure of precipitation is also thought to occur in the low-pressure zone
swept out behind a fast-spinning impeller. This zone of cavitation increases
toward its periphery. The fourth mode of aeration, rarely seen today, is
that used in "cascade" flotation, where pulp falling turbulently over a weir
entrains air.
Practical details of flotation machines are discussed in Chapter 18, but
since most of those in use today incorporate mechanised aeration, a brief
discussion of this vital link in the oxygenation, levitation and selective re-
moval of particles from the pulp is appropriate at this point. A zone of low
pressure is generated near the hub of the impeller. If this is in direct com-
munication with the atmosphere or is fed by a controlled supply of low-
pressure air this air blends with the pulp before leaving at the periphery of the
spinning sytem. Air and pulp are discharged tangentially at differing pres-
sures, and along their line of departure a cavitated low-pressure void is
maintained behind the tips ofthe impeller blades. Thus, air is compressed into
the pulp during its accelerating flow along the front of the blade, and re-
precipitated to some extent where pulp is sucked back or slips back over the
top. As the issuing stream quietens and changes to a rising movement some
air precipitates preferentially on to aerophilic particles. This aids their in-
corporation into the larger (coursing) bubbles as they thrust toward the
surface. It has been calculated 29 that in a pulp of S.G. 2,7 which contains
30 % solids and is rising one metre from the impeller-swept zone to surface
there are some 3 ml of precipitated air per litre. Much of this pulp is in
circulation in the vertical plane. When a cell is started up the oxygen content
soon shows a threefold increase.
When a frothing agent (discussed later) is added to the pulp the surface
tension is lowered, power requirement for driving the impeller drops some
10 ~~ and both vacuum behind the blades and re-precipitation of gas are
increased. Bubbles precipitated from an aqueous pulp are below 0·2 mm in
diameter, and on addition of pine oil this is halved under running conditions
in which the average coursing bubble is some 0·9 mm in diameter. The size
range of bubbles generated in "mechanical" cells varies widely in accordance
with settings, dimensions and speeds, but tends to lie between 0·05 and \·5 mm
diameter, the bulk being around I mm. Bubbles diffused from pneumatic
sparging arrangements are much larger and more even in size, averaging
between 2·5 and 3 mm diameter.
Despite contortions, vibrations and collisions bubbles do not noticeably
coalesce while rising. Each has a hydrated enveloping barrier surrounded by
a slip stream, and an air-water interface toughened by surfactant molecules.
These fend off collisions and discourage size growth in the body of the pulp.
The situation changes abruptly on emergence. Bubbles between 2·5 and
6 mm rising freely through water (i.e. unobstructed by mineral particles and
unwarped by turbulent currents) steady down to the shape of a flattened
oblate spheroid. Bubbles below this size change their shape rapidly,
504 Mineral Processing-Principles of Froth Flotation
periodicity being about 1(1 ,OOOth of a second during ascent. Bubbles larger
than 6 mm have widely irregular shapes, indented or bellying out.
The peripheral speed of the impeller and the geometry of its enclosing
shrouding, is critically related to the intake and blending into the pulp of the
indrawn air. The balance between air and pulp passing through the impeller
is important, and too high an intake of pulp can prevent aeration. For a
given pulp density and depth of impeller immersion there is an optimum
point of balance between radial intake of pulp (as regards distance along
blade from hub) and its peripheral discharge for a given speed.
The Bubb/e.-In order to simplify a discussion involving these, and other
factors, it is helpful to begin with a system air-distilled water. At 4°C water
achieves its tightest molecular packing. It expands below this temperature
to the ice lattice structure at O°c. Above 4°C it expands with the increase
of molecular disturbance due to rising kinetic energy. At ordinary working
temperatures water contains lightly associated polar molecules and fugitive
lattice formations. The position of a molecule in water is considered to be
constrained by these light binding forces, so that it is always in some degree of
electrical association with the molecules surrounding it. The lower the
temperature and the higher the ion-dissociation, the greater, therefore, is its
binding. This is shown by the change in viscosity, which is 18 millipoises at
freezing point, to at 20°C, and 4 at 70 e at normal barometric pressure. At
0
boiling point the kinetic energy of the molecule transcends the restraint of the
adhesive forces, and the system converts to the vapour phase. In Fig. 219
two molecules are depicted. That lying well within the water is in balance
with the attractive forces on all sides of it (subject to slight unbalance due to
its movement in the system), but the molecule at the surface is only half
balanced. It, like the other surface molecules, is being drawn inward to the
body of the liquid, and the force of this attraction is related to surface tension
(y). At 20 o e, in air, y is 72·8 dynes/em. In order to minimise this potential
energy the water reduces its area as much as possible. Since the sphere con-
tains the maximum volume for the minimum area, a drop of water takes a
spherical shape. Larger volumes conform to the overriding pull of the earth's
,
Air
. ,
- '''-'' - -
.- . --
--- ,•••
, ...., -
~
-
-=-Water ~
Fig. 219
Mineral Processing-Principles of Froth Flotation 505
of froth the bubbles tend to pack with planar interfaces, curving where they
conform to adjoining bubbles.
If the optimum concentration of surfactant frother molecules is exceeded
the stability of the froth decreases, since a hydrated layer between these
molecules no longer exists. Highly stable froths are formed by evaporation
of water charged with strongly soluble alkaline salts.
The Three-phase System.-The whole system-water, frother, and aero-
philic particle-can now be considered. The newly born N-bubble is mo-
mentarily in rapid vibration 26 with maximum surface tension for the system.
Pine oil molecules and mineral particles compete for positions in the inter-
phase. The "values" have been selectively conditioned with a collector
agent. The gangue minerals also have some aerophilic qualities, due to their
sorption of lubricants and any other available hydrocarbons during their
journey from mine to conditioner. All the particles in the pulp may have
some degree of floatability as they arrive at the aerating stage of treatment.
There is, moreover, a great excess of air-water interphase over packing bodies,
so that all can be accommodated. This point is brought out by Taggart
who shows the rise in assay value of the desired mineral to occur almost
entirely in the froth column and not in the pulp. Particles which are com-
pletely wetted are unlikely to rise by air transport. They may, however, be
carried over into the froth by agitation. Hundreds of mineral-bearing
N-bubbles emerge from the pulp for everyone finally removed as a concen-
trate-loaded bubble from the top of the froth. The great majority of N-
bubbles fail to become bubbles. They burst, releasing air at the surface of
the froth, and particles which start to slip back into the pulp. Of those which
do develop into bubbles, only a minority stabilise sufficiently to be separated
with their cargo of concentrate. The great excess of N-bubbles over bubbles
finally needed to transport the concentrate out of the system is essential to
commercial flotation. Oversized particles which are heavy enough to over-
come the surface tension force holding them in the interphase may drop out
during this period of violent adjustment. Particles which are lightly held may
be shaken or torn away by the collisions and turbulences in the agitated pulp.
A further important feature of the change from N to F bubble state is that the
clinging grains of mineral now lose the aid of the pulp density, as they move
from water-solid system of pulp to one which is largely gaseous-the froth
column.
If a single N-bubble with its interfacial load is now considered, the chances
of its survival in bubble form are remote. In Fig. 220 the N-bubble a is
A IR
ob
o
d
W'" I l ~
_I,.
I • ....
.- ~. ,.,1, .111 f'-!/'
~
~
/,..,Ctwplt
1I,1f.,..,
., .... ~, .,
• ." •
v,; ~
u,'~
~ /, .. I• .,1/ ....' ...,_ ,,'. ...
_ I..
/.7~
Ollfl/lofll
It
Ii
iIT.
0
o. - - {~ 1,1 Of'.- -- --
o.
:X
;~
G 5 ~ ~ m ~
.
~ ~ ~ T,tIC;., /,.",
(B) Otwplt .'.'NIPtI.,M.
".. ctn copper
.,
p.l, od
f,. /Il /,.,., tI'If"'~' ./Ifrl.
P"'OfI'J$II
(e) {OJ
1m "0. I•• ".IfII-" CtlII,., c:tll
OIl
Taggart observes:
"All grains are falling relative to the bubble walls, streaming through the
inter-bubble spaces. Both dark and light grains (Fig. 222) fall upon and
slide along the upper surfaces of bubbles, but below the horizontal bubble
equators only dark particles cling. Groups of dark particles collect at the
lower poles of the bubbles, forming a pendant tip from which particles appear
to string out and fall away. Cross observation of the column shows by
colour that the concentration of dark particles in the inter-bubble spaces
increases from bottom to top of the column. This is readily confirmed by
sampling (Fig. 221). With all particles falling and the bubbles rising this
change in concentration can only occur by reason of the fact that the average
rising rate of the bubbles lies between the average falling rates of the light
and dark minerals."
The picture obtainable by direct observation is one of a sudden change in
ratio of gas to liquid from the predominantly liquid pulp to the predominantly
gaseous froth. An equally sharp change is from the coursing bubble and
free-swimming particle in the pulp to the packed N-bubbles at the bottom of
the froth with the spaces between them full of particles in pulp, pressed on all
sides by the air-pulp interphase. Here, if anywhere, the test conditions which
apply to bubble pick-up attachment must be valid. The need for several
types of mineral particle, present in abundance, can now be appreciated.
Somewhere between the one-walled N-bubble at the bottom of the froth-
column and the fully two-walled F-bubble at the top the unwanted particles
nlust be squeezed out of the interphase and sluiced down into the pulp.
There must be an excess of fully aerophilic particles at the top in order to
ensure this. This is why a concentrate should not be removed prematurely
(for example, in a closed grinding circuit). The valuable particles must help
Mineral Processing-Principles of Froth Flotation 509
rely on visual inspection of the froth to guide their controls, and association
between a pH upset and a "sticky" or dirty froth is one element of control.
A further activity probably proceeding in a froth column which contains
incompletely passivated particles is renewed surface activity between the
oxygen in the strongly aerated froth and the mildly excited mineral surfaces.
Enough has been written to show that the cleaning action in the froth is
complicated, and very important for good recovery. It should be studied
and experimented with. The froth column can be standardised for height
and condition to some extent as the result of specific tests on the mill pulp
as it is produced in the local mill water.
The mechanism by which bubbles increase in size upward through the
froth is mainly entropic. Although their surface tension has been appreci-
ably lowered by armouring, it is still considerable. When several tiny bubbles
merge, the total air volume remains the same but the total bubble surface
shrinks considerably. Stated mathematically, n bubbles, each of the same
radius and having a combined area A, will when merged to form one new
spherical bubble of volume equal to their combined volume, and area AI,
have shrunk in area thus:
(17.23)
Frothing Agents
o
CH3
/""
C-OH
CH3 CH3
and that of commercial cresylic acid (a mixture of phenol homologues,
cresols and xylenols).
These carry the hydroxyl which links the low-solubility HC group to the water
in the air-liquid interphase of the bubble. As the flotation process developed
precision of control, practice moved away from the confused use of these
compounds in the dual role of collector and frother. The modern method
exploits specific chemical and independently controlled frothing action in
separate reactions. Consideration of the possible significance of the alcohol
group led to successful use of amyl and hexyl alcohols to improve selectivity.
This in turn suggested that these reagents owed part of their value to their
greater solubility. Today the largely empiric position has been reached that
relative insolubility of the frother is by no means a .Iille qua 11011, and that
selectivity may actually be improved if completely water-soluble frothers are
used. Here research, though active, is not abreast of empirical application.
For sulphide flotation a series of water-soluble polypropylene glycol
methyl esters-the Dowfroths (Nos. 200 and 250) is now well established.
The formula is:
CH:l-(O-C:)H6)X-OH
and the number refers to the molecular weight. They produce a lively froth
which breaks down well on leaving the flotation cell. The American Cyana-
mid water-soluble frother is Aerofroth 65. Others, based on the higher
alcohols, are also marketed. Froths can be modified by the use of aerosols
such as the latter Company's Aerosol aT. These disperse gangue in non-
sulphide flotation.
The solubilitil of some frothers is
Methyl amyl alcohol 17g/lat20C
Capryl alcohol 1·28 gil at 25C
Pine oil 2'5 gil at 25C
a-terpineol I ·98 gil at 15-20 C
Cresylic acid \'66 gil at 20T
(Ajier Booth & Freybergerl
Mineral Processing-Principles of Froth Flotation 513
Particle Size
the froth. To the extent that specific surface is proportional to the transfer
of ions into the pulp from the mineral the climate of flotation may be affected.
Since hydration occurs during wet grinding and is opposed by surface
aeration there is room for research on the relative effects of dry grinding
where slime-formation is a problem. Stable hydration of small particles is
lessened in the presence of air and collector. 29
Research has shown a relationship between collector concentration and
maximum floatable size. A higher concentration of collector is therefore
necessary in floating a long-ranged feed than is required if grinding reduces
all desired constituents below the size readily floatable. As regards slimes
and the problems they introduce, care in the grinding circuit should be con-
cerned to keep these to a minimum. In running a plant, as in other humanly
controlled activities, it is easier to keep out of trouble than to get out.
Mineralization of Bubbles
A particle which had its surface fully covered by collector would find it
difficult, if not impossible, to cling to an air-water interface completely packed
with frother molecules. The incompleteness of adsorption of the contacting
phases is an important factor in attachment. Mineralizing contact may result
from collision between bubble and particle, from precipitation of gas from
water or through a mixed process in which precipitation initiates contact.
Tests made by Whelan and Brown on coal showed that attachment on collision
with a coursing bubble increases with the density of the particle concerned,
which by its greater mass is better able to break through the slipstream or
energy barrier surrounding the bubble. Taggart attaches importance to the
precipitation of air behind the impeller and therefore to the need for dissolved
gas in the process water. Other research 7 supports this by showing how a
minute bubble, such as might result from precipitation, initiates bubble
attachment or the build-up of a buoyant system. Field reports confirm the
need for good aeration of process water if a sustaining froth is to be produced.
The available gas also influences the kinetics of attachment.
Formation of the coursing bubble occurs in various ways in the flotation
cell. It may be diffused by blown-in air, in which case it starts in a state of
intensely rapid vibration, changing shape and· area hundreds of times per
second. It may be sheared into the pulp after induction via the low-pressure
body of air near the hub of the impeller of a mechanical cell and further
sheared by being milled with the solids in the turbulent zone swept by the tips
of the impeller blades. Precipitated bubbles can be pressed against particles
which settle below the impeller and are being drawn back toward the hub for
re-circulation. Formation, coalescence and attachment of bubbles to solids
are elements of flotation kinetics.
In agglomeration-flotation the deslimed and activated particles are tacked
together by minute air bubbles and form a flexible envelope (glomerule)
round an aqueous centre. This relatively light glomerule is removed in the
tailings band of a shaking table or by such other means as treatment on a
specially manipulated conveyor belt.
Mineral Processing-Principles of Froth Flotation 515
In froth flotation the energy barrier between the hydrated mineral surface
and the bubble must be surmounted by force, either due to the adsorptive
attraction of the collector or to collision resulting from mechanised energy.
The acceptance of a particle by an air bubble can be complete in 0·005 second.
The point of impact between bubble and particle is inter-related with
slippage along the surface. Whelan and Brown 38 found that the percentage
attachment decreased as the point of contact moved away from the vertical
bubble axis when coursing bubbles hit drifting particles.
Where the magnitudes are reversed, Dzieniewicz and Pryor found that a
small bubble attaches more readily to a polished surface than a large one 7 and,
once attached, improved the chance of attachment of larger ones which arrived
later. Two stages at least operate-first, saturation of the pulp with small
bubbles, which may be below visible size and second, attachment of the
particle thus "primed" to a larger bubble. Without the existence of con-
ditions in which the precipitation of gas from pulp can take place, flotation is
reduced and may even cease.
In both mechanical and pneumatic cells air precipitation takes place in the
body of the pulp. The special froth quality of the deep pneumatic cell (the
Britannia) suggest that bubbles may emerge on the particles with pressure
relief as the pulp streams upward. The amount and overall effect of precipi-
tation, horizon by rising horizon, is qualified by the initial aeration and
agitation.
Particle shape plays a part in mineralization. A sharp edge, beside being
the most likely point for a bubble to form as it emerges from water, is better
able to cleave the energy barrier round the coursing bubble. Roughness of
the surface of a hydrophilic particle aids wetting by trapping water. Con-
versely, a surface which shields a pocket of air helps contact with a bubble.
The greater the hydration of the mineral surface, the harder it is for gas
molecules in the surrounding aqueous phase to diffuse to the solid and initiate
bubble formation. Where a surface is incompletely wetted, molecules of air
should find it easier to displace water from the oriented sheath round the
particle than to force apart water molecules in the fully aqueous phase. This
aids coalescence and bubble formation at the surface rather than in the liquid
phase.
Some freshly cleaved sulphide minerals hydrate so strongly as to have a
natural zero contact angle and in this state to be indifferent to xanthate activa-
tion. Only when such a surface has had time to become oxygenated does
some degree of hydrophobicity appear. Research with various gases
dissolved in the pulp29 shows oxygen to be the most actively adsorbed. This
adsorption occurs in three stages (a) reversible adsorption (b) active attach-
ment and (c) chemisorption with superficial oxidation. The initial surface
layer of molecular oxygen dissociates to form a monatomic layer of atoms
coupled to charged metal points in the lattice. Electron-transfer ensues and
oxidation begins. An oxidic film then starts to form and at an early stage in
this reaction the maximum adsorption of competing collector ions can be
achieved. Excess oxidation reduces the adsorption of collector. It has
been suggested that this is a sort of "softening-up" process in which the oxy-
gen penetrates the surface layers and weakens the bonds between the ions.
516 Mineral Processing-Principles O/' Froth F/oIGtioll
This would facilitate ionization, ion mobility and either ion exchange with
collectors or ion migration with solubility increase.
Theorising in this field of surface physics is necessarily tentative in our
present state of knowledge, but the inferences drawn from research and con-
firmed by empiric industrial experience broadly confirm the suggestions in
this section as to the vital interplay between aeration of the pulp and selective
anchorage of the particle first in the coursing bubble and later in the overflow-
ing mineralized froth.
Ion Flotation
rate of air arrival should only just exceed that of foam collapse, and turbulence
must be avoided.
If the potentialities of ion flotation suggested by current research can be
translated into terms of process economics the handling of leach solutions,
mineralised wastes, process water etc. may be affected, and the still visionary
possibility of "mining" sea water for its dissolved mineral wealth brought a
step nearer.
Flotation Kinetics
This subject has been defined by Arbiter and Harris 3 as "the study of the
variation in amount of froth overflow product with flotation time, and the
quantitative identification of all rate controlling variables". Operating
variables, together with those of the pulp constituents, run into hundreds and
their interaction extends the possible permutations into almost astronomical
figures. Approach to this study is therefore empiric and confined to the
control of selected key factors in a severely simplified system, which the author
calls "the pulp climate".
Tomlinson and Fleming40 concluded from studies mainly concerned with
the behaviour of apatite and hematite that under precise laboratory control
of size, mineral purity and aeration the rate of flotation for a single mineral
of given size is first order and overall rate for a mixture of its sizes depends
on the weight proportions and the rates of flotation of the individual sizes.
With inhibited conditions flotation rates of fine sizes vary with the square of
the particle radius for readily responsive materials. In intermediate or
inhibited flotation dependence on particle size is always less than in free
flotation. In flotation of a wide range of mixed minerals the concentrate
grade increases with particle size.
During conditioning several activities can proceed simultaneously. Chi
and Young 41 include the ionization of reagents, diffusion, adsorption, chemical
surface reaction, desorption or diffusion of reaction products, sliming and
flocculation. These authors and others 17 agree that recovery of floatable
material from a pulp follows first-order law. This is defined by the equation
-d(A
~= ki(A) (17.24)
- dNm
- - - = klNm (\ 7.25)
dt
where t is flotation time, Nm the number of particles in the cell at the instant
t, kl a rate constant applicable as a criterion of floatability.
Mineral Processing-Principles of Froth Flotation 519
References
I. Rickard, T. A. (1916). The Flotation Process, Mining & Scientific Press.
2. (1961). Quarterly of Colorado School of Mines Vol. 56, NO.3 (Vols. I and 2).
3. (162). Froth Flotation, 50th Ann. Volume, A.1.M.M.E.
4. del Guidice, G. R. M. (1934). Trans. A.I.M.M.E.
5. Gaudin, A. M. (1957). Flotation, McGraw-Hili.
6. Mellgren, 0., and Subba Rao, M. G. (1963). Trans. I.M.M. (London), 72.
7. Dzieniewicz, J., and Pryor, E. J. (1950). Trans. I.M.M. (London), 59.
8. Sutherland, K. L., and Wark, 1. W. (1955). Principles ofFlotation, Aust. I.M.M.
9. Young, A. (1805). Phil. Trans. Roy. Soc., 84.
10. Leja,J., and Poling, G. W. (1960). Int. Min. Proc. Congress, 1.M.M. (London).
II. Gaudin, et al. (1963). 6th Int. Min. Proc. Congress (Cannes), Pergamon.
12. Cooke, S. R. B., and Digre, M. Trans. A.I.M.M.E., 184.
13. Sun, S. C, and Troxell, R. C Trans. A.I.M.M.E., 196.
14. Pryor, E. J., and Liou, K. B. (1948). Trans. I.M.M. (London), Oct.
15. Moilliett, J. L., Collie, B., and Black, W. (1961). Surface Activity, Spon.
16. Adam, N. K. (1941). The Physics and Chemistry of Surfaces, O.U.
\7. Modi, H. J., and Furstenau, D. W. (1960). Trans. A.I.M.M.E., 217.
18. Eigeles, M. A. (1950). Metallurgizdat.
19. Gaudin, A. M., and Tournesac. (1954). First World Congress on Detergence,
Paris.
20. Taggart, A. F. (1945). Handbook of Mineral Dressing, Wiley.
21. Kivalo, P., and Lehmusvaara, E. (1957). Int. Min. Proc. Congress, Stockholm.
22. Br. Patent 708475; U.S. Patent 2698088.
23. Taggart, A. F., and Arbiter, N. (1946). Trans. A.I.M.M.E., 169.
24. Bruyn, P. L. de. (1955). Trans. A.I.M.M.E., 202.
25. Hines, P. R. (1959). Trans. A.I.M.M.E., 214.
26. Tucker, et al. Trans. A.I.M.M.E., 183.
27. Sidgwick, N. V. (1950). The Chemical Elements and Their Compounds, O.U.P.
28. Sollengerger, C, and Greenwatt, R. B. (1957). Trans. I.M.M. (London), 65.
29. Klassen, V. I., and Mokrousov, V. A. Introduction to the Theory of Flotation,
Butterworth.
30. Derjaguin, B. V., and Dukhin, S. S. Trans. I.M.M. (London), 70.
3 \. U.S. Patent 2,990,(58.
32. Green, E. W., and Duke, J. B. (1962). Trans. S.M.E., A.1.M.M.E., Dec.
33. Sebba, F. (1959). Nature, Oct., 184.
34. Sebba, F. (1963). Royal School of Mines Jnl.
35. Sebba, F. (1962). Ion Flotation, Elsevier.
36. Haeck. (1964). Chemical Dictionary, Churchill.
37. Taggart, A. F. (1951). Elements of Ore Dressings, Wiley.
38. Whelan, P. F., and Brown, D. J. (1956). Trans. I.M.M. (London), 65.
39. Plaksin, I. N. et al. (1957/58). Trans. I.M.M. (London), 67.
40. Tomlinson, H. S., and Fleming, M. G. (1963). 6th 1.M.P.e. (Cannes),
Pergamon.
41. Chi, J. W. H., and Young, E. F. (1962/63). Trans. I.M.M. (London), 72.
42. Imaizumi, T., and Inoue, T. (1963). 6th I.M.P.e. (Cannes), Pergamon.
CHAPTER 18
FLOTATION PRACTICE
Introductory
Though a book cannot replace the need for practical experience it can give
its reader some understanding of the principles which are utilised in good
technical control. The presentation of these in Chapter 17 was inevitably
somewhat repetitive and tentative. In plant operation, however, a clear-cut
procedure must be agreed upon and adhered to, despite gaps in our funda-
mental knowledge and our conflicting hypotheses. The alert operator will
keep in touch with research in his field and be prepared to adapt the milling
technique whenever new discoveries, confirmed by local tests, justify change.
He is concerned not only with technical efficiency but also with the pressures
of management which take into account the overall economics of operation.
Working conditions in an industrial plant must therefore be flexible enough to
be adaptable to change in the business world. In what follows only the
technical aspect of operation can be considered.
The main components affecting steady control (without which efficiency
cannot be maintained) are
(I) Ore (texture, crystal interlock, variation in composition).
(2) Process water (seasonal variation, progressive fouling if re-circulated,
waste disposal).
(3) Reagents (nature, reaction rate, correct dosage, effect).
(4) Machines (max. and min. capacity, adjustment, reliability).
(5) Pulp (particle size range, solid-liquid ratio, density change along
circuit, pH stability, slime effects).
(6) Flow rates (aeration effects, operating sequences, time-temperature
changes, steady flow, changes of ore grade).
(7) Product grades (concentrates, tailings, re-circulating middlings, partly
processed stockpiles).
(8) Re-circulating build-ups (classifier returns, middlings re-treatment,
return water).
(9) Adjustment lag (sampling accuracy and assay speed, monitoring, auto-
matic control, surge suppression.
The interdependence of these items is not always realised. Anyone factor
in a chemical process which wanders from its norm of application changes
the working conditions for which all the other factors depending on it have
been fixed, and erratic running flllIsf follow. Though such factors as change
in composition and texture of the ore are to some extent unforeseeable, they
must be allowed for. If loss of control of one item is bad, multiple lack of
control is a great deal worse. Systematic checking, a good standard of
Mineral Processing-Flotation Practice 521
chemical cleanliness, and care on the part of the mill manager to interest the
workmen in methods necessary for a good standard of technical efficiency are
essential to success.
The Ore
It is important to observe any changes in the tenor of the incoming ore,
preferably well ahead of treatment. Variation in the crystal structures and
interlocks may have an important effect on liberation and optimum mesh-of-
grind. Changes in the proportion of associated minerals may be important.
As the mine deepens and develops, new minerals may enter the feed and
affect the concentration process. A patina or other indication of oxidation
on a particle too small to have received much attrition grinding, points to the
arrival of oxidised mineral which may either need special treatment or
exclusion. The most feebly floated particles rise in the scavenger froth.
They should be studied until the operator is thoroughly familiar with their
normal appearance, shapes, sizes, and rate of production. Any marked
change in the scavenger float will then give a clue to more subtle changes in
the ore from which they have come.
Primary metal sulphides are usually stable and react with good predicta-
bility in the flotation circuit. Secondary metal sulphides are less reliable.
They may occur as coatings on- primaries, or as penetrations along cleavages
in the altered ore in which they occur, as when covellite coats a pyrite crystal.
Since flotation is selective to a surface, and not to a core, a completely or
partially coated particle of alien mineral may float as though it were a true
concentrate. When geological conditions have favoured migration and
redeposition of the valuable ore-constituent, the problems of grinding and of
producing a clean concentrate are complex. Oxidised ores, according to the
extent of penetration and lattice "decay", are softened, mixed, and prone to
send metal ions into solution. Increased consumption of reagent, over-
grinding and the production of slimes are possible troubles under such
circumstances.
Oxidised or altered ore may arise from geological changes, or from slowness
in sending severed mineral from the mine to the mill. Hold-ups in stopes,
ore passes, and ore bins must be minimised when oxidation-prone sulphides
are being worked, particularly when unstable pyrite is present. Stope
washing, periodical clean-up of trucks, excessive production of "fines" when
blasting, and contamination of ore in transit by casual oil and foul water may
lead to poor recovery. Sometimes ore can be drawn from the various
stopes in such proportions as to produce a fairly steady blend. Sometimes
markedly different drafts of ore can be kept separate and treated according to
a plan. An ideal arrangement is for the mill to receive part of each stope
sample during valuation and to keep its own stope plan. These samples are
examined and tested, and the most suitable grinding and conditioning treat-
ment can be worked out. Then, provided good liaison is maintained be-
tween mine and mill, the right handling is assured at the right time. The
more prevision is exercised, the better the mill will do its work. Blending
522 Min?ral Proc?ssing-Flotation Practice
of mill feed from separate bins becomes increasingly desirable where the
stopes are widely separated vertically or horizontally, or where one mill
serves several mines.
When the feed is of high grade it is easy to produce a highly mineralised
froth and top-grade concentrate. At such times the possibility of insufficient
collector agent in the scavenging section must be guarded against. When
feed is low-grade it is harder to stabilise a highly loaded froth, and one of the
final cleaning cells may be switched to a lower-grade section if cells and launders
have been flexibly coupled.
Pulp Preparation
The most important factor in preparing the flotation feed is correct wet
grinding. Mineral particles as coarse as 28 mesh have been seen to float,
but this is abnormal, and even a 48-mesh sulphide needs a thick pulp to help
it to remain in the froth. Laboratory control of the milling operation in-
cludes assays and micro-observation of each screened fraction in a repre-
sentative tailings sample, made in order to find the amount of loss and the
reason for its occurrence. The coarsest particles are least likely to be
adequately liberated. If grinding all the ore to a finer mesh improves re-
covery sufficiently to yield a profit after the extra work is paid for, it should
be done. It may be that the trouble lies at the other end of the scale and that
the losses are due to over-grinding. This kind of loss is not clearly brought
out by tests of screen products and may pass unnoticed in the plant. The
sizing method called beaker decantation l is easy to use and reveals the trouble.
Tests are made on a composite sample cut from the routine daily tailing
samples, in proportion to the tonnage represented by each of these samples
over the period concerned. This may be a normal monthly check, or may be
for a shorter period when adjustments to the grinding circuit are proceeding.
Each fraction is assayed for its content of valuable metal. Results are worked
up as in the following table:
TABLE 38
In this instance the heaviest section of the losses is in the - 10,.,. band of the
sizing band, while nearly half the total loss lies in the - 30,.,. zone. This
would not be clearly revealed by routine screening down to 200 mesh, which
would only show 26·2 % of the loss above that mesh, and would fail to bring
out the significance of the much heavier loss in the fraction that had been
ground extremely small. The remedy is better adjustment and control of
ball mill and classifier. It may be necessary to "scalp" the closed grinding
circuit and lift out heavy-metal sulphides which are retained there beyond
optimum grind. This can be done by adding flotation agents to the mill
discharge and placing a unit flotation cell between this and the classifier.
The maximum floatable size of a fully liberated particle depends mainly on
its density, aerophilic attraction, and shape. The "toughness" or degree
of permanence of the froth is a modifying factor. The passing N-bubbles do
not affect the support given to the particle by pulp density, but this support
virtually disappears in the froth, which consists mainly of air. Fall is re-
sisted by the toughness of the froth and the strength with which the mineral
surface is attracted into the air phase.
Coal has nearly the density of the supporting pulp and there is no difficulty
in bringing 10-14 mesh material up to the frothing zone, though with the loss
of pulp support from thence on it may need aid from the froth texture when a
good deal of such coarse-meshed feed is being treated. For metal sulphides
the upper limit is 48-65 mesh and for metallic gold 100-150 mesh. Metallic
gold has a high density (19'2) and poor collection quality, so needs a tough
froth if it is not to slip back into the pulp. Mineral oils are sometimes fed
to the grinding circuit for this purpose, their passage through the mill en-
suring emulsification. At the other end of the scale, selectivity diminishes
when treating particles smaller than about 5,.,. and good recovery and grade of
concentration become more difficult to obtain.
The gangue particle size is limited only by considerations of mesh-of-grind
for liberation and smooth transport through the plant. The larger the particle,
the greater its tendency to settle and cause trouble but if machines, pipes, and
launders have been properly designed and installed, the risk is low. Large
particles need more vigorous agitation, with use of more power. They may
cause excessive abrasion of the flotation machine.
The reasons for the difficulty often experienced in the selective frothing of
very fine particles-say minus 5fL-are complex, varying from ore to ore,
and not fully understood. In the author's field experience it has been found
that, provided the commercial value of the mineral raised justifies meticulous
circuit control, particles down to 2,.,. can usually be floated at a good recovery-
rate. The difficulties centre mainly on the disproportion between the mass
and the surface of an extremely small particle. which causes it to behave as
though its surface tension is unusually high. As a result it reacts prematurely,
becoming oxidised, flocculated, or slimed before it reaches the conditioning
section. It is then difficult to coat with collector. The physics of attachment
to an N-bubble are also more delicate, since a particle of low mass is repelled
from the slipstream which surrounds a fast-rising N-bubble. It must there-
fore be offered small, slow-moving bubbles to sorb into, or achieve its attach-
ment in the mixing zone round the impeller. When a small particle (say
524 Mineral Processing-Flotation Practice
- 4/L) reaches the froth column, it tends to overflow with the concentrate
regardless of its composition. The down-drag of gravity hardly affects it
when opposed by the prevailing upward drift of the bubbles. Its dispropor-
tionately high ratio of surface to volume (as compared with the average
particle treated in the pulp) has given it undue selectivity of the energy-
neutral ising forces available during treatment, and thereby reduced its specific
reaction. The relative scarcity of such small particles in the tailings can only
be inferred in our present state of knowledge. Two indicators are observable
in most plants. First, a miniature shaking table can often remove further
value from the tailings band it is set to monitor, which would not be so easily
done if the feed to it had not reclassified itself to some extent. Second, the
last scavenger froth is white and foaming in many cases, which would not be
the case if the fraction of very small particles was similar to that at the head
end of the section. The conclusion drawn may at first sight seem to contra-
dict the sort of evidence given in Table 38, but this indicates the need for
research rather than objection, since the changes of size analysis, cell by cell,
for feed and discharge (the latter of both frothed minerals and cell tailings)
have not yet received the close attention than can best be given by a combined
effort between the research laboratory and the operating plant.
One method of combining the economy of coarse grinding with the re-
covery and grade best ensured by fine grinding is a stage treatment, when the
type of ore permits (Fig. 223).
Grind
Concentrate ~<----
1
Rough Float -----.~Tail
1
Regrind -----.1
jRefloat - - - - - ;..~(Tail)
Concentrate
Conditioning
Conditioning must provide the continuously passing pulp with the opti-
mum reaction time required, the appropriate reagents being added at each
stage in a series of operations. The efficiency of the conditioning depends on:
(a) Thorough mixing and dispersion of each reagent through the pulp.
(b) Repeated contacts between molecules of reagent and all the particles
concerned, at each conditioning stage.
(c) Development time for such contacts to produce the desired reactions,
if necessary in successive stages.
The work to be done includes:
(a) the dispersal of slime coatings.
(b) the correction of H-ion concentration.
(c) any needed depression or activation, including resurfacing.
(d) possibly final adjustment of pH.
(e) selective coating by collector agent.
(f) addition of the frother.
(g) adequate stabilisation of correctly loaded F-bubbles at the top of the
froth column.
Since the last additive (f) is not reacting chemically it needs dispersion only,
not time. A certain amount of entrainment of air by splash, cascade, and
vortex occurs during conditioning. If a frother was present a mineralised
froth would be formed and would be a nuisance during the conditioning
stage. Good operating control requires the accurate use of small quantities
of reagents and their thorough distribution through the pulp. If the mill
water is returning from dewatering end-products it may carry some frother,
since this reagent is not permanently removed by reaction with any of the
minerals. This might lead to the formation of froths or scums of the most
air-avid minerals, which could thus withdraw an undue share of collector
agent. It is therefore desirable that the mechanism of conditioning provides
for the beating back of such froths into the body of the pulp. A particle
which has once floated will float again whenever equally favourable oppor-
tunity is given, provided its aerophilic surface has not been destroyed, so
there need be no fear in "drowning" such material.
The work of the reagent may be hampered if it is not mixed continuously
into an equable stream or body of pulp. If the addition stage calls for
0·05 lb of reagent per ton of ore, harm will be done if all that 0·05 is surged
in with a hundredweight of ore, or if the pulp itself is surging so that it
alternates between double and half its correct rate of flow. Unless all the
particles have a chance, a small proportion of them may take up the whole
of the ration. Reagents which are readily soluble in water disperse quickly
into the pulp, but relatively insoluble oils must be thoroughly mixed, or a
Mineral Processing- Flotation Practice 527
in the flow-line can be used for conditioning. The ball mill is a good agitator
and emulsifier. Occasionally a collector can be added there, to ensure its
presence at the moment new surface is exposed, provided no edge-adherent
scum of rich mineral then appears in the classifier. Removal of a paraffin-
floated layer of graphite from the classifier has been made ahead of cyan ida-
tion.
The first stage of pH control may be an addition of lime in the fine-ore bin.
This has the advantage of reducing deterioration of sulphides. Final pH
adjustment would then be possible by feeding milk oflime (slurried Ca(OHh)
to the classifier. Another favoured addition point is the surge tank between
the classifier and the pumps which deliver to the flotation cells. Good
mixing and agitation are assured, but the dwelling time may be erratic.
Conditioning in the grinding circuit is, of cou'rse, only possible before taking
the first float of a series, unless there is to be further grinding between floats.
Conditioning tanks (Fig. 224) are cylindrical, and up to 16 in. diameter by
16 in. high, though a dimension of 10 in. would be usual with a moderately
coarse grind. Types such as the Denver and the Knapp & Bates are not
sanded up in the event of power failure, and the latter's slatted weir facilitates
return of prematurely floated mineral to the mixing zone. Capacity is
adjustable by provision of discharge ports at different heights. Some con-
ditioners have mild pumping ability, but in the event of breakdown this may
lead to flooding of the cells from which they normally draw their pulp.
The criteria for scale-up from laboratory conditioning requirements to
those in a full-sized operation have been mathematically considered by Chi
and Young 2 .
Reagents
Reagents are added to the pulp either as solids, immiscible liquids, emul-
sions, or solutions in water. Feeding arrangements are described later.
There is nothing static in the supply trade, new reagents being constantly
developed. Samples are usually sent by manufacturers on request, and are
tested in the mill laboratory. There are no fixed rules as to type and qualities,
though of the thousands which have been marketed relatively few are in
universal use. The tendency is for each plant to find the specific combination,
sequence, concentration, and staging time most suited to its own ore. and for
this rearrangement to be peculiar to the plant concerned. Since the amount
of reagent adsorbed is in ratio to the surface area entering reaction. the assay
value of the reacting mineral and the fineness of the grind are the main
determinants of optimum quantity added per ton of ore.
When the "fines" in the feed are excessive, and particularly if they derive
from barren earths and heavily oxidised ore, a dry-looking froth may appear.
Frothing agent should then be reduced. Little else can be done to help
matters. The risk of high tailing loss and poor concentrate grade, with
heavy consumption of reagent, is great when such froths appear, and the only
fully satisfactory remedy is selective mining or removal of troublesome
material during crushing. De-sliming as the first conditioning operation
may be practicable.
Mineral Processing-Flotation Practice 529
_ .. .. I----rootA ~
FIG. 226_ Feeding Lime Slurry
Mineral Processing-Flotation Practice 533
;=:
[LE\lATED STORAGE
~ rANK
V
/"
VALVE
-TO REAGENT
FEEDERS
01
II
"II
OVERFLOW
PIPE
_c::::=~~~
TO WATER
SUPPLY
The most widely used reagent feeder (Fig. 229) has small buckets mounted
at varying horizontal distances on a vertical revolving disc. A variable gear
allows changes in rate of rotation, and a retractable tipping device governs
the number of buckets tipped at each revolution. In a variation, buckets are
added or removed from the disc as required. The tipping bar is raised or
lowered to govern the amount spilt at each contact. Large and small buckets
are available and for oils dipping nails can be substituted. The rotating
buckets keep the solution stirred. Corrosive liquids are handled in fibre or
alloy buckets. The arrangement is simple, cheap, sturdy, and precise.
Evaporation may cause buckets to clog, so periodic cleaning is desirable.
A number of discs can be driven from one shaft, each dispensing a different
536 Mineral Processing-Flotation Practice
Flotation Machines
The design of the cell in which the concentrate is floated must facilitate as
many as possible of the following duties:
Usually cells are arranged in series, each cell in a "bank" receiving the
tailings from the one preceding it. The height of the froth column is de-
538 Mineral Processing-Flotation Pructice
termined for each cell by adjusting the height of the tailings overflow weir.
Since the overflow lip for the froth is fixed, the difference between this and the
weir level sets the height of the frothing column, subject to it being possible
for the froth to reach that height. This is largely determined by the availa-
bility of aerophilic mineral particles. The new feed enters the rougher
section of a typical bank. The froth column is kept high at the head of the
section and lessened from cell to cell by progressive raising of weir height.
The scavenger cells have but little mineral with which to stabilise a froth,
so their weirs are raised till, by the end of the scavenger section, the pulp is
almost spilling over to the froth launder. This end setting is part of the
policy of "pulling the cells hard" so that every particle showing a tendency
to float is removed. In the cleaner section the pulp is thinned and the weirs
kept low, so as to maintain a thick layer of froth and obtain the maximum
possible cleaning action.
The alternative to individual tanks is a series of long troughs, each con-
taining an appropriate number of aerating and agitating devices. A single
weir controls each trough. Current practice tends toward elimination of
individual square cells, each with its separate impeller, feed entry and tailings
weir, in favour of long tanks with unmechanised froth overflow. This trend
is marked in roughing and scavenging circuits where large tonnages are
handled. The Denver "free-flow" cell (Fig. 230) illustrates this simplifica-
tion.
Of the numerous cells manufactured, a few have been selected for descrip-
tion . Between them they incorporate the most important features of all c~lIs.
Developments in cell design are continuous, and the reader desiring detailed
information concerning a particular type is better served by the manufacturers'
bulletins than by a somewhat static text-book.
The history of cell design and development has been described by the
author elsewhere~. The hydrodynamics of flotation cells are still inadequately
understood despite much attention at research leveP. Satisfactory adherence
of the particle to the bubble depends not only on chemical preparation, but
also on the geometry and dynamics of the cell. So many variables are per-
force left uncontrolled in industrial flotation that a scientific basis for com-
parison of performance between various types of cell, with varied settings and
working conditions, is apt to be misleading. Operating skill is by no means
a negligible factor and this varies between shiftsmen.
Cells of type a include the Denver, Fagergren, Humboldt, Massco and
Knapp and Bates, and several other proprietary makes which, while in-
corporating special variations, conform in the main to a few general principles.
Tanks are square, round , or long (troughs), in the latter case having multiple
impeller units.
The standard Denver cell (Fig. 231) has a square tank. The illustration
shows this cell with a bonnet above the stand pipe (A) through which air can
540 Mineral Processing-Flotalion Practice
be blown down to the vortex near the hub of the impeller (G). At this hub
there is a partial vacuum which allows air to be drawn down to the impeller
blades (B) where it mixes with the pulp entering from the previous cell in line
~'ia the feed pipe. This pulp has fallen over an adjustable weir (Oil and, in
due course the tailings from the cell will pass under a baffle and over (02) to
the next cell in line. Sand which is too coarse to rise to this discharge weir
is by-passed through the cell by means of the sand-relief ports (E). Weirs
are raised and lowered by various mechanisms, operated in this instance from
the rods (F). When the stand pipe is arranged to receive pressure air the
cell is said to be supercharged. When the pipe is open the cell relies solely on
air drawn in, which varies with speed and condition of the impeller. A
cleaning rod can be used to clear the sand-relief ports (E). At the side of the
cell is a pipe wh.ich can be connected to the middlings return system. Two of
the stationary baffles (H) are shown. These receive pulp flowing from the tips
of the impeller-blades and past the stationary diffuser (I). At (1) is a re-
circulating port, called the "key stone". In some machines this can be
closed by a sliding plate. thus controlling re-circulation in the body of the cell.
In the author's experience better control is maintained when the keystone
aperture is permanently closed, and recirculation is confined to pulp drawn to
the zone between impeller tips and stationary baffles. A cross-section through
a standard Denver cell is shown in Fig. 232. Here the stand pipe is open and
the middlings pipe is shown connected to a return launder. A crowding
baffle (K) is depicted. The purpose of this is to push the rising bubbles
forward toward the skimming paddle (L) which removes a froth layer from
the cell. These baffles can readily be improvised where it is desired to hasten
removal of the risen material. They reduce the accumulation at the back of
the cell of mineral which fails to move over to the skimming zone. and which
may become over-oxidised by prolonged exposure, fall back into the pulp
and then be lost owing to its changed surface condition. The pocket or
spitzkasten, (M) below the skimming paddle is used in older Denver machines
to provide a quiet zone where the froth crowds out unwanted material for
return to the pulp. It is not much used in new installations, as better results
can be obtained by use of a rougher-cleaner-scavenger arrangement of cells.
in connexion with which a lively froth is removed as soon as possible for any
required re-cycling. By taking advantage of the ease with which a froth
from any section can be returned counter-current to join the feed with an
earlier cell in the line, upgrading can be more accurately controlled than by
the use of spitzkastens. Near the bottom of the cell is set a horizontal im-
peller which acts as a centrifugal pump. The impeller is driven by a vertical
shaft from an overhead motor at a peripheral speed not less than 1350 ft/min
and not usually more than 1800 ft/min, though up to 2000 is occasionally
used. There is an important connexion between peripheral speed and
efficient aeration. Several factors influence choice of speed. The feed to
the cell is introduced near the centre of the impeller. This feed may have
come from the conditioning section, from the previous cell in the bank. or
it may be a middling product overflowing as froth and being returned from
cells further down-line. In addition, any desired amount of pulp already
inside the cell can be drawn down to the vortex at the centre of the impeller
Mil1eral Processil1g-FlOfUliol1 PracTice 541
and centrifuged outward. Stationary baffles check the swirl of the outflowing
pulp. Some pulp may creep back along the bottom of the cell toward the
central vortex. All of the pulp in the swept area is caught up on the impeller
blades and accelerated to its terminal velocity (the peripheral speed of the
impeller less slip) before being abruptly arrested and turned over from
horizontal to vertical flow. The amount of slip is partly governed by the
clearance between the impeller and the stationary hood above it, and to a
less extent by the shape and number of blades and their deformation with
wear. The other slip-factor is the ratio of pulp to air being pumped through
the impeller. If all available pulp passages are open, there is less chance for
air to enter. The feed pipes are larger than is necessary for the pulp they
carry (unless the cell is being grossly overfed). If the internal pulp-port is
shut. the rest of the feed openings (new and return) and the standpipe sur-
rounding the impeller shaft are available for drawing in air. The adjustments
therefore centre on balancing the aeration by
In practice the gate adjusting (b) is usually kept shut, and the makers' speed-
setting is used. For a float requiring extra aeration, impeller speed may be
increased above 1800 ft/min peripheral, but wear increases as the sixth power
of velocity and power consumption with it, so this is an expensive alteration.
542 Mineral Processing-Flotation Practice
coverings of which have been worn to (or beyond) discard point. Poor
flotation would be obtained with impellers in this condition.
In the event of power failure sand settles above the hood, so the cell can
be restarted immediately power is restored. The height of the tailings weir
is adjusted by slats or by a slide. The vertical difference between the tailings
overflow and the bottom of the mechanical skimming paddles used to remove
froth is the height of the frothing column. The greater this is made, the more
severe is the cleansing and up-grading effect. Although the froth normally
rises over the whole horizontal area of the cell, it can only be removed at the
skimming side. This leads to the accumulation of ageing mineral in places
not much affected by skimming and also, in the case of difficulty floated
material, to the falling back of particles into the pulp. Where copious froth
is needed to get the float out before the weakly aerophilic minerals drop
back, special launders can be had which gather the froth more effectively.
In the flotation of an abundant concentrate from a minor gangue in such
cases as treatment of coal or phosphate, scrapers can bring out the froth
from the front half of the cell. At the tailings end provision is made for the
release of accumulated oversize too heavy to flow. This should periodically be
discharged from the circuit before it begins to impede smooth working. One
shiftsman controls up to 75 cells. He uses a vanning plaque, or inspects a
miniature shaking table working on a cut from the bank tailings, to provide a
visual check on performance.
Variations on the general Denver design include the Humboldt (Fig. 235)
and the M.S.S.A. Fahrenwald, in which pulp recirculates through apertures
in the hood. The size of the cell is determined by the economics of main-
taining a given volume of pulp in suspension and of transporting it through
the cell by means of the impeller. a centrifugal pump not necessarily being
544 Mil/era! Processil/f{-- F/otaTioll PracTice
The Fagergren can therefore float reluctant and slightly aerophilic particles
of the sort which slip back from most froths. The older Fagergrens have
single assemblies in square-section cells, but the trend is toward a long
rectangular trough, divided into sections each containing a rotor and stator.
The feed enters below the first partition. Tails go over partitions from one
compartment to the next and the froth level is adjusted at the end tailings weir,
which has a sandgate below for removing accumulated coarse material.
The cell lacks one great advantage possessed by the Denver, positive pull
from other cells to the impeller without a special pumping system. Any
moving of pulp from one point to another therefore requires an independent
pump, unless gravity flow is possible.
Cells of type b are typified by the M.S. Sub-A (Mineral Separations Sub-
Aeration). Here pulp enters below the hub of the impeller, which is an
546 Mil/eral Processil/g- -Flotatiol/ Practice
agitating device only . Air is blown in at low pressure by some such device
as a Rootes blower and is beaten up with the pulp. Froth is skimmed off
by paddles. and tailing overflows a slat weir and goes to the next following
cell. Again there is no positive suction for "shunting" pulp inside the bank
of cells.
In the Agitair cell. a long trough divided into several compartments. the
pulp moves from feed end to discharge. I n each compartment an impeller
(Fig. 137) rotates inside a balfle system (Fig. 238-9) . Air is blown in at from
~ p.s .i. to 11 p.s.i. through a hollow shaft (Fig. 240) and sheared into bubbles
Mineral Processing-Flotation Practice 547
as it enters the pulp. The volume of entering air is controlled for each
compartment separately. Impeller speed can be varied from 900 ft/min
(peripheral) to 1750. Adjustment of speed and air entry permit operational
control of the froth column, so that it flows counter-current to the pulp.
The assembly is shown in Fig. 240. These cells are modest users of power
c:
and produce a copious froth. This has won them preference in mills handling
ores of poor floatability, which require a big volume of froth to help their
mildly aerophilic particles to overflow. They have, indeed, found consider-
548 Mineral Processing-Flotation Practice
able application in flow-sheets once served by cells of type a. They lack the
ability to effect positive displacement of slime which is a valuable feature of
the violent impeller action in cells of type a. Against this must be set the fact
that improvements in desliming during the conditioning period have today
reduced dependence on this action.
Type c cells (cascades) use air entrained during the fall of splashing pulp
into the cell to effect flotation. Their use is largely confined to simple home-
made devices at the tailings end of a process where finished pulp is running
down to waste and can be cascaded.
Type d, the completely pneumatic, is one of the original forms of flotation
cell. In the Southwestern cell (Fig. 241) air is blown down vertical pipes
from a distributing header, and agitates the pulp in a long trough. The
froth rises to a quiet outside zone whence it overflows while the pulp, which
has risen to a top baffle under the pumping influence of the air, drops back
and is lifted again a little lower down the trough, thus slowly moving in
spirals from feed to discharge. These cells have a more voluminous and fragile
froth than those of type h. Maintenance is low, and the only power cost
is for low-pressure air. If internal recycling is needed, an air-lift pump must
be used. In the Britannia cell, a much tighter froth is produced, resembling
that of cells of type a. This is achieved by making the cell several times as
deep as the ordinarily shallow pneumatic cell (8 ft to \0 ft). The back
pressure of the extra hydrostatic head causes smaller bubbles to be generated.
An advantage is that for the same floor space several times the cell capacity is
available. Another is that with the larger volume of pulp passing through the
number of collector-coated particles on the bubbles is much higher. This
aids the cleansing effect in the froth column by increasing the competition
between particles.
A newcomer in the field of coal flotation is the cyclo-ceIJli. Agitation is
effected in submerged vortex chambers in which a high-speed jet of water is
formed into a hollow cone through which air is sheared into the cell in tiny
bubbles.
easily and cheaply maintained, thrifty in power consumption, and easily fitted
into the circuit. On the criteria of power and reagent consumption, product
grade and overall costs, pneumatic cells probably are best, but the tendency
in older plants is for them to be replaced by mechanically agitated machines.
The pneumatic has not the transporting power of the mechanical cell, nor
can it handle such a coarsely ground pulp. If sliming conditions are present,
the positive agitation of an impeller has a scouring effect not possible with
simple pneumatic agitation. For floating a single mineral from a simple ore,
particularly where vigorous frothing is called for, pneumatic cells have their
uses, especially when the plant must be inexpensive. Usually with expansion
of treatment rate, the flotation section is changed to use motorised cells.
When differential flotation is practised, the mechanical cell is dominant.
With a complex ore requiring running changes in the number of cells rough-
ing, cleaning, and scavenging, a machine which has positive pumping action
(such as the Denver) is often favoured. It is only a minute's work to remove
or insert a plug or two in a double launder, and the section's capacity can
thus be trimmed to suit a change in the ore-a convenience in a medium-
sized plant treating 500-1500 tons/day. A simple multiple-launder system
is shown diagrammatically in Fig. 242. Nine cells are served, and new feed
o 0 0 a
Fig. 242. Flexible Arrangement of Rougher-Cleaner-Scavenger Launders
can be brought to either No.3 or No.4, while the roughing section can
contain up to five cells, ending at No.7, or be shortened at either end. If the
spigot at 7 is removed and the shutter on its right dropped, the froth from
No.7 cell falls with the scavenger froth from 8 and 9 to the middlings return
pipe at either cell 3 or 4, according to whether the spigot and shutter for cell
4 are on or off. Thus, a lean ore which fails to produce sufficient crowding
of the froth when the cleaning section is working on three cells can be
strengthened by diverting new feed to No.3 and the froth from No.3 to the
rougher concentrate launder. With larger operations, the maximum capacity
for each section is usually built in and pulp-shunting is not necessary. The
technical criterion is the metallurgical balance-the grade of concentrate as
well as the percentage of value recovered being important. Factors in main-
taining full running time and smooth working are simplicity of the machines
Mil/eral Processil/!f---F/ofafiol/ Pracfic<' 551
and the ability of local labour to handle them. The corners of cells trap
ground-up wood and other organic materials, which decompose and release
depressants. Regular cleansing often improves recovery. In some plants a
monorail is installed above the bank, by means of which a unit cell can be
taken out of line for repair and a spare one can be dropped in. A circuit
using substantial quantities of lime may be troubled by precipitated scale,
which builds up and clogs pipes and other passages unless these are regularly
serviced.
The size of the cell selected depends on the total flotation time required
and the space and handling arrangements in the mill. In terms of capacity a
cell has four dimensions-length, breadth, height, and dwelling time (rate of
throughput). Part of the time is transferable to the conditioning section,
and part can be varied by changing the number of cells in line. This latter
element of capacity is based on the aeration requirements. The individual
cell's length and width depend on the agitating and searching power generated
by the impeller, which at /I revolutions/minute is a function of the square of
its diameter. Power consumption in unaerated pulp is a linear function of
this diameter, but in practice this dimension is limited, owing to the high
rate of wear should peripheral speed exceed some 1800 ft/min. The greater
the swept area, the more slowly must the impeller turn. The cell's overall plan
area can best beexpanded by using troughs instead of square cells. Height can
increase volume of pulp held, but the lifting must be done by the impeller
normal to its plane of rotation. Since this is done across the baffles required
to check surging, this is an inefficient mechanical design in terms of pumping
action. Power tests favour the mechanical cell which works in the shallow-
est volume of pulp, and as most of the particles in that pulp are barren, the
less power used to move them around, the better. Long troughs are suspect
for similar reasons, save where they handle large tonnages. Most of the
separating work is done in the first few feet, the balance of the distance travelled
by the non-floated particle between entry and discharge being an insurance
against bypassing of the impellers by values. Such bypassing can be more
accurately controlled, as can the other operating conditions, in a bank of
individual square cells than in a long tank. Here are the elements of com-
promise in design for a specific case of plant design.
Finally. there is the probability that new reagents may be developed which
will shorten the flotation time. When this happens, provided the mine does
not increase the tonnage sent for treatment, some of the flotation capacity is
no longer needed. A single cell is more flexible than a trough, whether it is
being tacked on to an under-aerated bank or removed to reduce floating time.
Agglomeration
position relatively large and light, so it works out to the top of a horizontally
moving band of feed on the shaking table. It then rolls down over the riffles
and is removed nearly opposite the feed entry.
Conditioning is usually carried out in a thick slurry of deslimed feed, fairly
closely classified as is customary in table gravity working. Conditions are
much the same as for flotation, pIllS the use of thickish oil to help to smear the
solids, instead of the frothers used in cells. The slurry is aerated by being
cascaded and tumbled in the air as it falls to the separating surface. The
heavy-metal sulphides or other floatable minerals now form rounded, flexible
glomerules which roll across the table to the gangue withdrawal launder,
while the relatively light gangue works over to t'"le heavy concentrate discharge
end. Desliming is important as the slimes would collect in the oil interfaces
and interfere with conditioning. The feed size range can be as high as
.~ in. and as low as 200 mesh, but the process is usually applied between the
limits of 6 mesh and 65 mesh, over a three-to-one range between maximum
and minimum particle in the feed. The use of a thick pulp saves reagent,
and aids in the smearing of oil. Dilution is performed at delivery to the table,
and cascade action is helpful in dragging in more air. The glomerules may
burst against sharp riffle-edges but will then stream as a skin-float down to the
tailings launder. The Holman-Michell flotation table is illustrated in
Fig. 243. Air is blown into the pipes which cross the table deck, at a light
pressure. In the operation shown, sulphides are being agglomerated and
removed from a rough concentrate of cassiterite.
When separating a concentrate on conveyor belts, as practised in phosphate
Minera! Prvcessinf{---F/otatioll Practice 553
treatment, the thick conditioned pulp is delivered to the middle of the belt
and is turned and ploughed as the belt passes under stationary fingers.
Sprays are directed on to the pulp and agglomerated phosphate aggregates
wash over the edge of the belt while the gangue remains wetted and travels to
the discharge end.
Middlings
Feed
1
Concentrate .......,;------ Flotation ------7"> Tailings
1
(Middlings)
t
Regrind
Special
Concentrate' ....- -
1
Flotation ..
----~ Tailings
' - - - - - - - - - - - M iddlings
1
Fif{. 244. F!o\\'-slleet-·-Specia! Re-freafmenf of'Middlings
554 Mineral Processin/?-F/ofafion Practice
The mill laboratory should be used not only in connexion with routine
checks but also to try new reagents, treatment rates, and other variables.
When tests have proved that an improvement can be made, the necessary
changes must be initiated by the mill superintendent and closely watched by
trained men in the early stages. The shiftsmen cannot be expected to
understand and control a circuit alteration until they have acquired experience
of it. If more than one factor is to be changed, the variations must be
introduced gradually, giving the plant a day or more to settle down after each
cautious alteration. Here lies the greatest danger since the day shift, during
which changes are usually initiated, hands over to workers who have had no
chance to get used to the new look of things. Alterations in a flow-sheet
have a knack of leading to, or being accompanied by, sudden choke-ups in
the mill pumps and piping, as most experienced engineers have found.
Changes should therefore be so arranged that the responsible engineer is on
hand for the following twenty-four hours or longer, while the shiftsmen
are being trained to the new conditions. Do not alter more than one thing
at a time, unless this is quite unavoidable. If the change leads to the re-
circulation of a different type of middling, some days will elapse before the
cumulative effect of this recycling can be assessed. It is essential that time
be given for this to happen before proceeding with any further change, if
confusion and misleading inferences are to be avoided. The difficulties of
carrying through a change successfully are greater than would theoretically
appear, but the pleasure of success is well worth the effort entailed.
The best concentrate is made in the first cell of a given section if condition-
ing has been correctly done. From this point on the froth mineralisation
diminishes and the grade of float drops. There are several reasons for this.
First, the head cell of a series contains the richest pulp, from which to draw
true concentrate into its froth. Second, the later cells, which have less true
concentrate, are less able to squeeze contaminated gangue back into the pulp
from their lower-grade froth. Third, the height of the froth-column di-
minishes in the later cells, since there are fewer floatable particles to stabilise
it. Fourth, the extremely fine particles entrained in the froth are removed
near the head of the section, and the bubbles are more fragile after their
departure. Thus for a given series of cells (whether set to rough, clean or
scavenge) the falling-off in grade of skimmed froth is accompanied by a
decrease in fully liberated value, and an increase in the middling and con-
taminated gangue which is then able to take its place. If a bank of cells is
divided into rougher, cleaner, and scavenger sections, the correct length of
each section varies with the assay grade of feed. It is sometimes helpful to
have a flexible system of launders and returns so that these sections can be
adjusted in accordance with changes in grade of feed. The height of the
frething head determines the amount of sorting work done on the floated
particles. To measure this, a floating gauge can be made and marked in
accordance with the depth of submergence in similar pulp ahead of the cells.
The frothing head should be diminished slowly from head to tail, the last
scavenger cell or so showing a barely mineralised fugitive froth. The weirs
of these final cells are set high so that the pulp almost brims over into the
froth launder.
556 Mineral Processing-Flotation Practice
In the conditioning stage, the higher the percentage of solids the less
agitation needed to maintain all the particles in suspension, and the more
concentrated will be the reagents in the pulp. This last point is not so critical
as it sounds, since the aqueous phase of a 15 % pulp is 941- % by volume as
against 87·5 % for 30% solids. Pulp density is usually dictated by that of the
overflow from the grinding circuit. In the flotation cell, the higher the pulp
density the easier it is for a particle to reach the frothing zone, but against
this must be set the need for good mobility of the solids. The practical upper
limit is 35 %-40 % solids, in the case of a slime-free feed. The presence of
primary slimes and colloidal material might reduce the permissible percentage
of solids, possibly even to 15 %. In the scavenging circuit a high density aids
the "reluctant" particles which are being forced to float. Most of these are
incompletely liberated, surface-contaminated by Iyophiles, largish, or of
awkward shape for holding in a bubble, so they need help. With a thicker
pulp there is more collision between bubble and particle and a longer dwelling
time in the cell for a given weight of solids. Before the pulp reaches the
scavenging section, most of the concentrate has been withdrawn, and thus
it may have become unduly thin. At Rammelsberg it has been found worth
while to switch pulp reduced below a critical operating density, thicken it,
and then return it to the line of cells. Other advantages of working with thick
pulp are the increased capacity thus conferred on the system and the better
carriage of particles in the transporting stream.
Just as thickening aids in forcing up the last reluctant particles which
would otherwise be lost as tailings, so does thinning of the pulp aid cleaning
of the rough concentrate. This leaves the roughing section as a froth which
breaks down to about a 2: I liquid-solid ratio as it flows toward the cleaning
section. Any desired dilution can be made by spraying this rough concentrate
en rOllte. Cleaning is usually practised on pulp containing about half the
percentage of solids used in roughing. A limiting factor in cleaning is the
dilution produced by the tailings from the cleaner cells, which are nearly
always sent to join the rougher feed, the whole bank of cells being arranged as
in Fig. 245.
This system can be extended as required in either direction, the number of
component cells in each block being suited to the processing time required.
The pulp is usually conditioned and floated at the prevailing mill temperature.
Sometimes the movement of copper on to sphalerite is speeded by warming
the conditioner, and where fatty acids are used in the flotation of non-
metallic earths a temperature of 30 C may be justifiable to ensure that the
oleic acid used is above its freezing point.
Fundamental improvements in process control have been made possible
by such technical aids as continuous fluorescence analysis from samples
drawn from appropriate points along the flow-line. These give rapid indica-
tion of change and. if coupled to correcting devices. make rapid corrections of
drift or irregular running. Other monitoring and controlling arrangements
deal with such matters as changes in pulp density, height of frothing blankets,
reagent availability and building up of re-circulating sands or pulps. These
are discussed in Chapter 21.
The largest number of cells is included in the rougher section and the smallest
Mineral Processing-Flotation Practice 557
in the final cleaning, the scavengers usually being run together. Though the
diagram shows the cells in separate blocks they would, if square units, be
arranged in an unbroken line. The chief advantage in having two scavenging
stages is that the first section can be "pulled" moderately hard, while the
final "save-all" cells can be pulled very hard, their froth being withdrawn
for special retreatment if unsuitable for return to the main circuit. Where a
high-purity product (e.g. acid-grade fluorspar at 97·5 ~~ + CaF2) is required,
several stages of recleaning may be employed. It is good practice to return
Feed
I
I
Roughing Cells ~
I Ig
(Froth) (Tail)
(T.;!) (Fcothl
1
Concentrate
1
Tail
a froth from a lower-grade section to a point in the preceding part of the line
of cells where the entering pulp has about the same assay value. Once a
particle has responded to flotation, it will almost certainly float again if
returned to the same cell, provided its surface has not deteriorated during
recirculation. An exception to this general probability would be a particle
of oxidised ore which had floated under the fugitive influence of conditioning
with a soluble sulphide. Re-surfacing would be needed, perhaps by a small
addition of the sulphide reagent to the scavenger float on its way back to the
preceding series of cells.
The rougher froth usually overflows at between 35 % and 45 % solids. It
may need the aid of water in the launder to break down the froth, to help it
to run freely, and to dilute it into the cleaning cells. In the preceding section
it was observed that a small plant could not normally justify a special re-
treatment of scavenger froth, which must build up and overflow till its con-
tained middlings either report as concentrate or tails. This leads to cyclic
surging, the circulating load building up until some circumstance initiates a
release. No amount of reagent or circuit manipulation can do much, since
only further liberation can upgrade such particles. The concentrate may be
deliberately downgraded by letting the higher-grade middlings work through.
558 Mineral Processing-Flotarion Practice
t ~Sample a
Regulating Ore Feeder
I
PHalkali~
Activating and
Pllmp
collecting reagents!
j
Conditioner/s
r
t
Distributor (if parallel banks of cells used)
Frother~ j
r
t Sample c,_ _ _ _--,
? Water t ? Water
returned --(C) - - Cells - (T) I pumped back
to circuit t to circuit
Pump i
+- Thickener ~ ~ I
Underflow Tailings Pond
V
rI-
Filter
Sample d
Cake to _ _ _ _ --'t
t
Drier
~
Shipment
depression of one from the bulk float. Method (c) is indirect, and is used in
such cases as the flotation of auriferous pyrite as a prelude to special treat-
ment of the fraction thus selected, or for upgrading a fraction of an uranium-
bearing ore by floating the minerals most closely associated with it. Method
(d) is an alternative to (b) used in such cases as the final upgrading of a dirty
phosphate float by first destroying its activation and then using amines to
remove the silica. Fig. 246 shows a general flow-sheet for single-product
flotation. Possible addition points for reagents are indicated, together with
those at which control samples may be taken.
The conventional reagent plan for floating the single minerals named away
from their gangues is shown in Table 39. The reagents prefixed Rare
Cyanamid reagents. In the column "Frothers", higher alcohols are often
used to modify the frothing agent named. Since the minerals of any named
type vary in their response to reagents, this Table can only indicate the best
starting point for an investigation. Some typical industrial treatments of
ores, most of which include flotation, are outlined in Chapter 23 under the
name of the principal element in the concentrate.
Broadly, the collectors used to float metal sulphides are thiophosphoric
(the aerofloats) and sulphydric (the xanthates). When such minerals are
partially (or in some cases completely) oxidised soluble sulphide may be used
ahead of, or concurrently with, the collector to restore for a short time an
attracting sulphide-metal surface. With the commonly floated non-sulphides
(excepting some copper, lead, and zinc minerals which accept resulphidisation)
the surfaces of the particles are relatively inert. It is correspondingly more
difficult to find a reagent combination which is strongly specific to one only
of the minerals present. Fatty acids and their salts are the most used anionic
collectors. Despite their higher cost, cationic amines are also in commercial
use. The consuming industries often demand a high standard of purity in
the concentrate they purchase. To attain this special care must be given not
only to specific collection, but also to specific depression and the lively froth-
column and multiple recleaning in which upgrading is achieved. The main
groups of non-sulphide minerals treated by flotation are:
I. Oxidised heavy-metal minerals.
'") Non-silicates of the alkalis and alkali earths.
3. Silicate minerals.
4. Inert minerals.
5. Carbonates.
The usual approach aims at depression of the unwanted minerals in order
to obtain a selective mineralisation of the froth. Depression achieves two
objects. It makes the non-floating surfaces more wettable, and also reduces
consumption of collectors by limiting the surface area which will attract them.
A vital controlling factor is the ionisation maintained in the pulp liquid.
The most ditncult operations are those which must separate minerals having
closely similar surface properties, such as fluorite from calcite: apatite from
calcite: fluor.ite from scheelite: beryl from felspar. The collectors used are
strongly reactive and their selectivity is poor. The crux of control, therefore.
lies in depression.
TABLF 39
GENERAL GUID~ TO REAGENT PRACTICF
.....
'" OIl
-~ ~
o
..... c-
til C .::- .... ro I ....
0.0) .;; c::':'" ~.~ ..... '".....
c:'"til '" o
* 0) " ~ 5h';; E 2 ~ '"
c0) OIl '..... '" 0) ~
0)
'"
'"<; :~'E ~v ~~'B ~ ~ ] ..c
E ..... '" 0. tU~C:: .~
'"
c c..c 0) • 6. '"E
0) 0) (3 o 0)
>0. .~::c o:;~ ~ u .....
i:iJ :::E ;':0 --<rJl Cl ~rJl"O til Cl --< U I..L. ~
Alunite K 9-4 2·7 Na 2 SiO" Excess Oleic acid Cresylic Associated minerals
K.AI:.(SO~k AI 19·6 Na2SiO" R.708 acid include Kaolin.
(OH)u Higher Pyrite
Alcohols
Antimony Stibnite Sb 71·7 4·7 Na2CO~ Excess CUSO, Xanthates Pine Oil Associated minerals
Sh 2 S" S 28·3 Na2SiO" Alkali (Ethyl sec. include Quartz,
NaCN Butyl) Barite, Galena,
Gold.
May be floated
for bullion content
or depressed before
cyanidation
Arsenic ARSENOPYRIH As 40 o'() NazCO" NaCN CUSO, Xanthates Pine Oil Associated minerals
FeAsS Fe 34·3 Na2SiO" Lime (Ethyl to include sulphides of
S 19·7 amyl) copper, iron, lead,
zinc, antimony.
May he auri-
ferous
Barium BARYTfS Ba 58·8 4·5 Na 2 SiO" AIC'" Pb(NO:ilz Oleic acid Pine oil Associations--
BaSO, NaZCO:l Citric FeCI2 and fatty Cresylic Iron, manganese,
Acid Aerosol acid soaps. acid galena, sphalerite,
Higher- chalcopyrite
alcohol
sulphates
Berylliulll Bervl Be 14·1 2·7 Washing, H2S0, Pb(NO:1l2 Aillines Associations-·
Be:,AI 2 Sil; Si 06·8 HF Oleic and Granite, schist.
0" AI 19·1 fatty acid mica, clay, slate
Salts. R.825
Boron BORA' B 11-4 1·7 Aniline, Starch, BaCl z Fatty Acids Aniline
Na2B,O~ Dextrin Pb(NO,,12 Xylidine
IOH20 Quebracho Pyridine
Calcium Calcite Ca<40 Na2COa Quebracho Oleic acid Pine Oil
CaCOa Na2SiOa Sulphonated
K2Cr207 oleic Fatty
acid salts
Calcium Dolomite Mg 13-2 2-9 Quebracho Fatty acids Pine Oil
CaMg(COal2 and salts
Carbon COAL C<95 1-4 Na2SiOa Tannin Fuel Oil Pine Oil
Quebracho Paraffin, Cresylic
Cresyls acid
Carbon Diamond C< 100 3-5 Natural Associations-
(boart grease, pet- Shale, clay, quartz,
CARBONADO) roleum jelly, iron, Fe+Mg
on Tables Silicates
Carbon GRAPHITE C<99 2-0 Na2COa Starch Paraffin Pine Oil Associations-
Na2SiOa 600 Series Fuel oil Long- Spinel, shale, mica,
Chain silica, schist
alcohol
Carbon Shale ? 600 Series Fuel oil Pine Oil Such shale may im-
Cyanamid Paraffin Higher pede cyanidation
alcohols unless removed
Cerium MONAZITE 5-\ Na2C03 Strong Oleic acid Pine oil
(CeLaYt)P04 Na2SiO'l Acids Fatty-acid
Formula salts
doubtful
TABLE 39-colltillued
Chromium CHROMIH Cr 46·5 4·5 NaOH Phospho- R.8oo Series Pine Oil Associations-
FeOCrzOa Na2CO:1 molybic amI Fatty acids Fuel Oil Serpentine, black
Fluosilicate p. tungstic Alkylo- sands
acids amines
sodium silicate, and flotation with a froth made fragile by use of such additives
as iso-butyl-carbinol. Cerussite floats fairly well with a heavy dosage of
amyl xanthate (up to severallb/ton), part of which can be omitted if sodium
sulphide can be made to film the surface with lead sulphide. The copper ores
malachite and axurite respond to mild sulphidising treatment with xanthate
and gas oil, or alternatively can be floated with fatty acids of the oleic type,
partially sulphonated, and preferably well emulsified by grinding with hot
water, sodium bicarbonate, and sodium silicate immediately before addition
to the pulp. Again, gas oil helps with collection and frothing. Where
siliceous gangue must be depressed, a moderately high pH aids by wetting the
calcium sorbed to the surface. If a sodium oleate soap is the collector, any
surface which carries calcium may sorb an oleate group and float, so pH up to
9t-IO, produced with soda-ash, is used to minimise this. Full depression
after this calcium surface-activation requires a pH of II.
Earths
In the flotation of earths low-solubility soaps are supposed to form be-
tween the Ca +, Ba +, etc., cations and the oleic anion in the discontinuity zone.
Whatever the reason, the froth is heavy and tenacious, being so strongly
flocculated in many cases as to earn the description "frozen". The author
finds that modification of the oleic acid by mild sulphonation removes this
qifficulty in the case of fluorspar flotation!, giving a lively, self-sorting froth.
pH control gives some differentiation between the metal "earths". Tannic
acid is a depressant for calcium, and the use of sodium silicate aids the
selective wetting of the associated siliceous gangue.
Silicates
Silicates are somewhat similar in behaviour to the above minerals. The
pH needs close control, and depression of gangue by a wetting agent is helpful.
If cationic reagents are used, the water must be soft and the pulp fairly free of
slime. Among the minerals floated by fatty acids or fatty acid amines are
feldspar, kyanite, garnet, and mica.
Non-ionizing Minera/s
The "inert" minerals include coal, graphite, and native sulphur. All these
preferentially sorb hydrocarbon oils and float readily without a frother, or
with a little pine oil. Sulphur acidifies its pulp, which usually must be
neutralised with lime to prevent corrosion of metal surfaces. The main
flotation difficulty is to disperse and wet down the gangue.
References
1. Pryor, Blyth and Eldridge. (1952). Recelll Dl'e/opmellls ill Milleral Dressing.
I.M.M., London.
2. Chi, W. H., and Ypung, E. F. (1962). TrailS. I.M.M. (Lolldon), 72. p. 169.
3. Ihid. (1962). Millillg World, March.
4. Pryor, E. J. (1961). "Flotation's Early Years" (Qrly. (I( Colorado Sch. o(
Mines), July.
5. Arbiter, N., and Steininger, J. (1963). I.M.P.e. (Calllle's), Pergamon.
6. Heal & Patterson Inc., Pittsburgh. Trade Bulletin.
7. Ejgeles, M. A. (1957). I.M.P.e. (Stockholm), Almqvist & Wiksell.
8. Pryor, E. J. (1952). Recellr Del'elopfl1ellfs ill Milleral Dre.lsill/:, I.M.M ..
London.
CHAPTER 19
from the poles normal to the surface. Maximum attraction therefore occurs
along the sharp edge of a wedge-shaped pole, while there is little or no
tractive tendency immediately adjacent to a flat polar surface. The attractive
effect is inversely proportional to the square of the distance between the
attracting pole and the attracted particle. Magnetic alloys of high gauss are
replacing electromagnets in a number of applications. They are described
not only according to composition but also in terms of their coercivity (He),
remanence (B,) and maximum gauss (BH max). Coercivity is the "magnetic
574 Mineral Processing-Magnetic and Electrical Separation
TABLE 40
Alloy· Br He BHmax
(gauss) (oersted) (megagauss-oersteds)
Applications
Magnetic Applicances
I
Dry
I
Detectors Wet
I
I
I
Concentrators
I
Protectors
I
D.M.S.
I
Concentration
Regeneration (belts, drums)
Belts
I
Discs Drums Static Pulley
(Belts, drums)
Feed
\
Revolving
Non-magnetic
Drum
Magnetic: Non-magnetic
Fig. 250. Arrangement of Wetherill Type Separator with Coils in Upper and
Lower Magnets for Fields of High Intensity
578 Mineral Processing-Magnetic and Electrical Separation
+- - ---
['--__ ~~ +
-- )-1'10
+ 100SCHAl<:ES SnONGEST
___ ___
MAGNETIC HAnklA\.
has the advantage that it is more likely to remove iron pinned down by a
piece of rock than is the suspended magnet. It may have Alnico permanent
magnets or be energised by D.C. In the latter case it must be inspected
regularly to ensure that current is flowing, that cooling air-gaps are open, and
that the pulley is functioning efficiently. To carry light pieces of iron clear of
the underside of the pulley a light strip of steel belt lacing can be tacked in at
intervals on the belt.
Special problems arise with non-magnetic steel. An excavator's digging
tooth would smash something if it ran through to the crushers. Several
sensitive methods of dealing with this type of problem have been worked out.
Small inserts of magnetic steel can be drilled into all parts liable to break off
in the mine and find their way up with the ore. These can sound an alarm
and shut off the belt motor, if necessary marking the spot by dropping white-
wash or chalk, since a loaded belt travels a little way before stopping. If
there are two or more such search points, the risk of the attendant's failing to
find all the signalling tramp metal at the first stop is removed. Mine-de-
tector types of coil are used in some arrangements. Extremely sensitive
arrangements can be used, but it is not necessarily desirable to stop the
transport of ore for a tintack.
580 Mineral Processing-Magnetic and Electrical Separation
Pole p'Pee
,..------'"------,
Brtdge _ __
bar
Cod
,I'
1st magne tIC
proClvcf
,,
2nd magnt>f,c
produc f Non- magneflc
producf
Fig. 254 shows arrangement for weakly magnetic sands (8-200 mesh) in
which the magnetic particle is deflected when falling through air. Up to
seven laminated rolls, with progressively smaller air-gaps, are used.
Mineral Processing-Magnetic and Electrical Separation 581
Much of the concentration required today is concerned with sands and fine
powders. In order to shake out trapped gangue particles by breaking up
magnetized aggregates the modern separator develops a rotating field. This
is produced by permanent magnets set with a pole distance of only a few
centimetres. The design recommended by Runolinna 18 has one or more
rotating drums, whose speed is so balanced as to use centrifugal force to
discharge the more weakly attached and feebly magnetic particles. The
magnets used in these drums can be closely packed. Cavanagh and Williams l9
have recorded studies made by the Ontario Research Foundation on sands
varying from 6~ down to - 325~. The basic developments in the work were
guided by these considerations:
(I) For dry work a single layer is ideal.
(2) Economically, high operating speeds are needed.
(3) The shape of the magnetic field can be developed to be circular, spiral,
or helical.
(4) No mechanical feeding and discharge devices should be used.
(5) Gravity should aid separation.
(6) An opposing air stream must replace the water-wash used in wet
magnetic concentration.
(I) above can be achieved by using small permanent magnets mounted on a
drum which rotates at a different speed inside an independent stainless steel
drum. The magnetic particles are thus constrained in movement. Four
types of machines have been developed from a "rougher" working on
- 6~ + I OO~ material down to one handling - 325~ dusts. The essential
difference is between the appropriate dynamics for easily settled material and
that easily airborne. In a specific case the relationship between average
particle size, magnet design, drum speeds, and differentials is considered in
setting the machine. Two types of rougher are used. In one the feed
travels on a belt at 400 ft/min and passes over the double drum, of which the
inner (magnet) drum is rotating in a reverse direction at 400 ft/min. The
non-magnetics are thrown clear as the belt turns over the drum, while the
magnetite clings as it is accelerated by the influence of the field flux, until
it is separately discharged as the belt leaves the drum.
In the other "rougher" type the magnet drum is mounted in a second unit
above the feed belt, and travels in the opposite direction. Gaps and flux
intensity are so adjusted that the magnetic particles are picked up more than
once and dropped, before finally adhering to the upper belt. As they climb
they are again accelerated by the differential effect. The upper belt moves at
twice the speed of the feed belt and the magnet drum rotates in the opposite
direction, as before. An air blast is directed upon the climbing particles to
aid separation.
- 60 + 200~ material is treated in a different way. Dust is fed from a
hopper on to a rotating stainless steel drum which forms its bottom. The
inner magnetic drum rotates inside this drum, but at a different speed. Its
magnetic particles thus aid with the feeding by stirring the material as they
feel the influence of the reversing flux. As feed leaves the hopper it is strati-
fied with non-magnetic gangue on top. This material leaves first, magnetic
582 Mineral Processing-Magnetic and Electrical Separation
middlings later, and concentrate on the rising side of the drum. Detachment
is produced by mounting the drums eccentrically, so that the gap widens
sufficiently for centrifugal force to overcome magnetic attraction in the con-
centrate discharge zone. Air elutriation aids the separation. With this
appliance magnetic flocculation may occur if the drums run too slowly, the
proportion of magnetic material is high, or the population in the air stream
too dense. This can be overcome by an arrangement of magnets which
induces a helical field on the drum surface. Usually the magnets are mounted
herring-bone fashion on the inner drum. Magnetic particles fed toward
the centre of the drum migrate outward and are discharged at the sides, while
the gangue falls off centrally. Several drums are normally used in series to
produce a good concentrate grade. The machine makes a finished grade at
- l50~ from ore containing not more than 30 % of - 325~ material.
Very fine dust must be treated in an airborne stream. A device called the
magnetic precipitator has been developed. The entering air travels in a
helical path, dust being pressed against the inner rotating wall of a cylinder,
while an outer system of stationary or rotating magnets induces a helical
magnetic field. The magnetic particles move upward against a draught of air
and at the top enter a weaker magnetic field and are thrown off, while tailings
are blown down to a central discharge. All these machines have proved
successful in pilot-scale runs, an Aerofall mill doing the dry grinding. Runo-
linna 20 has described development work in Finland by Professor Laurila on a
drum type of machine. High field strength improves separation (in the tests
900 gauss 2 mm from the drum surface (Fig. 255) at a flux density gradient
of 50 gauss 2/mm was used). The drum is run at a speed somewhat below
"critical", which is defined as the point where centrifugal force equals mag-
netic attractive force. A peripheral speed of 4·75 metres/second gives good
practical results.
In an alternative arrangement 21 , the Mortsell separator, the feed falls over
three drums in series. The two upper ones rotate slowly, and centrifugal
force is low. The top drum picks up most of the magnetic particles and drops
its tailings to the second one for scavenging treatment and discard. The
concentrates from both drums are delivered to the third, which is smaller and
rotates at a regulated speed. This is so adjusted as to remove by centrifugal
force any middlings which would lower the grade of the retained concentrate.
A pick-up type of drum separator (Fig. 256) is designed for use where a
A Foed
• - NOD-.IIt.clod product
C - nllddlln,.
D - "'.,nolle produd
• C D
large proportion of magnetic matter is contained in the feed and a high grade
concentrate is required. The magnetic material is lifted from the passing
stream and carried forward proportionally to its tenacity, to separate
middlings and fully magnetic drop-off points. Where the magnetic per-
meability of the feed is low, or its percentage small, the feed comes directly
on to the drum, from which non-magnetics are thrown while seized particles
are carried down till they leave the field and fall clear. This is a roughing
treatment.
High-intensity separation in which D.C. energised magnets are used has
been discussed by Palasvirta 22 . He notes that the rebounding of tailings
particles from the pole face back to the induced roll (see Fig. 254) increases
with rotor speed thereby reducing the grade of the magnetic product. This
can be suppressed by choke feeding, when the stream of particles blankets the
rebounding gangue. Within the sorting capacity limits of the machine feed
rate then becomes an operating control with a fast-run rotor.
584 Mineral Processing-Magnetic and Electrical Separation
fed. Magnets (3) above the belt lift and convey the flocculated ferro-mag-
netics (6) forward and out while the barrens (I) run on to tailing. The
Linney develops the same idea differently. When these machines are used
for dense-media work, the residual magnetism must be destroyed before
returning the ferro-silicon to the separating bath. This deflocculation is done
by passing the material through an A.C. coil.
In northern Europe drum-type wet separators are much used for concen-
trating finely ground iron ore. The "pick-up" principle is used (Fig. 258),
magnetic particles being lifted from the pulp while the gangue streams through.
Hydraulic currents keep the pulp in suspension, and the picked-up fraction
passes through a series of Nand S fields before it is removed by sprays.
The need for efficient methods of cleaning foul ferro-magnetic media from
D.M.S. baths stimulated the development of machines such as the Crockett.
Later, these were largely replaced by drum-type machines containing electro-
magnets, which were more compact and efficient. With the improved
permanently magnetic alloys now available, these have in turn given way to
drum separators of this type. In a review of current practice. Bronkala 23
gives as reasons for this change weight reduction, elimination of the need for
Mineral Processing-Magnetic and Electricul Separation 585
D.C. energising current, greater reliability and much lower cost of mainten-
ance. The problem of insulation of the electrical wiring also disappears.
The drum separator has from four to seven poles alternating north and
south, at the back of a non-magnetic drum which dips into a receiving tank
and transports the magnetic fraction of the feed to the discharge point. Feed
slurry is spread the width of the receiving tank as it leaves the feed box.
Fig. 259 shows a two-drum system in which the tails from the first drum are
scavenged in the second. With the most recently developed ceramic mag-
nets the second stage can often be omitted. A spigot usually delivers part
of the tailings discharge, thus avoiding retention of particles too coarse to
overflow. The rest of the tailings leave by weir overflow, thus ensuring a
steady height through which final washing of the magnetic concentrates is
performed.
586 Mineral Processing-Magnetic and Electrical Separation
Conctntratt
discharge
current is cut off and a short surge of high-pressure water enters via valves 6
and 3, the funnel being at the same time in position 14. The cycle is then
repeated.
Tests on some 34 minerals were made and good concentration of such
moderately magnetic minerals as hematite, ilmenite, garnet and biotite were
made, down to sizes as small as one micron. Efficient separation depends on
adequate liberation, distinctive discrimination in magnetic quality of the
species treated and a well-dispersed pulp. The operating economics thus far
reported are attractive, less than 3kWh/ton of feed being needed. This
machine should fill a special gap in the work of purifying high-grade clays,
talc, glass sands and ceramics where the presence of iron is undesirable.
Another cyclic arrangement the Ferro Filter, has been used at Kipushi in
the Congo to remove renierite (a germanium mineral) from a copper-sulphide
flotation concentrate. The filter, a horizontal disc, works in a 135-second
cycle of three stages-loading, flushing and discharge after cutting off the
electro-magnetic flux. The disc is fed from above and discharges into a
three-compartment tub.
gap. Air round the wire becomes ionized and is attracted toward the
grounded body, where it discharges its ions. If the voltage of the wire is
sufficiently high the ionized corona is visible as a luminous discharge.
Mineral particles which enter the electrified field are bombarded with gaseous
ions from which they acquire negative charge.
Two types of charge are possible 26 . If the particle is falling freely gaseous
anions tend to attach electrostatically to positively charged points in its
surface lattice. It then acquires overall negative charge as part of this
positively charged discontinuity lattice becomes neutralised and/or the
particle as a whole becomes a charged condenser. After partial saturation
the nett negative charge cannot increase, since further arriving gas-ions are
repelled from its now negative field.
If instead of falling freely between the corona-producing wire electrode
and the grounded surface the particle travels on the latter surface as it moves
past the corona discharge, it receives the same ion bombardment as in the
case above, but a new condition is added. If the particle is a good conductor
or if its surface is conductive, the arriving anions pass from it to ground.
If it is a good insulator, this leakage will be slower, less or even absent. A
conducting particle cannot, therefore, retain all or much of the charge in-
duced by ion bombardment but a non-conductor can. Build-up of retained
charge to the final strength (the steady-state) is practically instantaneous and
the polarity of this charge is opposite to that of the plate on which the particle
rests.
If the grounded surface is moving (e.g. as a drum or roll in rotation) the
particle is carried out of the ion-bombarding field into one where its electro-
static charge attenuates to its steady-state strength. A non-conductor
retains full charge, an intermediate one suffers some loss, but a conductor's
charge first decays to neutrality and then reverses its polarity until it has
acquired the charge density of the roll. Thus, the non-conductor remains
electrostatically coupled to the roll while the conductor either loses its grip
or is repelled.
With a two-component mixture of particles of which some are negatively
and some positively charged, separation can be effected by exploiting the
attraction of each species toward an oppositely charged electrode, and its
repulsion from one of like charge. Figure 261 shows separating conditions
for (a) free fall, (b) fall along an inclined plate, (c) fall between rotating rolls
instead of plates and (d) the additional use of an ionizing electrode. In
(b) the particles fall on the positive electrode. If non-conducting but initially
charged they take a steeper path on leaving the plate while the conductors
acquire positive charge and are repelled into a flatter trajectory. The roll-
type separator (c) aids removal of strongly adherent particles. In (d) the
ionizing electrode is either a highly charged wire or one with a series of needle
points (the spray electrode). It can either take the place of the negatively
charged roll or reinforce its action.
The complexities of the surface physics of conditioned minerals discussed in
earlier chapters are also encountered in H.T.S., since the ability of a particle
to acquire or retain charge depends in part on its crystalline nature and in
part on the state of its discontinuity lattice, particularly with respect to ad-
590 Mineral Processing-Magnetic and Electrical Separation
sorbed moisture. It has been suggested 27 that the treatment of minerals with
organic reagents can stimulate their electrical response by giving them differ-
ing affinities for adsorbed water, thus varying their surface conductivity.
A hydrophobic surface has a sharply decreased conductivity, and it is possible
that the conditioning of minerals before H.T.S. may have some analogy to
that preceding the flotation process. Pre-treatment, however, does not
markedly affect the electrical response of minerals which have high con-
ductivity.
Three types of particulate reaction are used in practice, and are defined by
Hudson 33 :
(a) Conductance charging, in which particles passed over a metal surface
charged by induction are themselves charged by conduction from it.
(b) Convective charging, in which particles are "sprayed" with a discharge
from a sharp edge, points or fine wire.
(c) Contact-potential or frictional charging in which particles acquire a
charge while passing over a surface and maintaining frequent intimate
contact with it.
On completion of the charging process the particles pass through a high-
intensity electric field and are deflected in accordance with the sign and magni-
tude of their charge. In view of the non-static nature of the convective
process, in which there is some current flow, and of the differential time-
element in discharge rates of various types of particle, the term "electrostatic"
is not entirely satisfactory. For the (b) type defined above the term "high-
tension" separation is perhaps preferable. Modern methods of rectifying
alternating current to yield steady D.C. at high voltage have broadened the
application of these methods, which are used in the benefication of various
minerals not readily treated by the flotation process.
H.T.S. is sometimes applicable to material too coarse for froth flotation,
or where dryness is important. It is also useful where small quantities of
Mineral Processing-Magnetic and Electrical Separation 591
The usual procedure is to re-pass either the residual fraction at lower potential
of ionizing electrode or the deflected fraction with a higher potential, or to
re-treat both fractions as shown in Fig. 263.
Material is heated while passing in a single-layer stream from the feed hop-
per to the grounded separating roll. The needle-point electrode charges the
passing stream. Good conductors discharge rapidly to the grounded roll and
thus become electrically inert. As they descend further on the rotating roll
they are strongly attracted toward the gas tube electrode and are deflected
592 Mineral Processing-Magnetic and Electrical Separation
. -DOCI4ot'OCI
tuell'COI
... I""IIfl)
•
_,
.4d/IJ.J/ob/e
..Jpli/ler.
In a specific case the maximum particle size which can be treated depends
on shape and S.G. Heavy sulphides up to 6 mesh may be amenable, and up
to I in. coke. The lower end of the size scale is about 325 mesh, smaller
particles no longer flowing with sufficient freedom through the separating
field.
In research on the basic principles which underlie mineral behaviour in
corona-discharge roll-type separators the Warren Spring Laboratory has
developed two novel laboratory appliances. One, christened the insiloscope,
has stainless steel electrodes working in a humidity-controlled atmosphere
in a brass box. Measurement of rate of charge leakage from mineral grains
has found this to vary only with respect to surface conductivity with changes
of humidity. The second instrument is a variation on the high-tension
separator in which a flat disc spinning at 80 r.p.m. is used instead of the
conventional roll. Mineral grains are fed radially so as to pass under a
stationary line filament. As they lose charge they are thrown off centri-
fugally into one of the compartments surrounding the disc.
are the least likely particles to be gripped, and therefore acquire the least
kinetic energy from the rotor. As they are in any case being electrically
repelled they are the earliest to be thrown outward and the flatness of their
trajectory can be augmented by attraction toward the positive electrode
system (Fig. 261d). Between these are the swept-off non-conductors falling
from the separating zone are two types of middlings-gravitational, and
ionically charged 34 • Ideally four kinds of product can be gathered from
separate trajectories. Furthest away from the grounded rotor and following
the flattest path are the electrically repelled good conductors. Next come
middlings which have mainly responded to gravity and air-classification, but
which have some electrical charge. Third are the ionically charged middlings
detached mainly by centrifugal force but mildly charged when leaving the
rotor. Last are the non-conductors too firmly pinned to the surface of the
roll to be removed by centrifugal force alone.
Conductance, charge etc. are relative to each mineral in the feed and to the
purity and surface cleanliness of each of these species. Centrifugal speed
and charge intensity are variables affecting the sorting, as is the positioning of
the charging electrodes. The charging electrode has two parts, one ionizing
by its corona field and the other not. The angular and radial positions of
these electrodes relative to each other and to the feed roll affects the response
of the passing particles. Roll speed and electrode voltage must be correlated.
The electrode system should be high and fairly close to the feed roll, with the
ionizing portion closer than, but not beneath, the non-ionizing electr"de.
Titanium-bearing beach sands which contain relatively heavy minerals such
as ilmenite, rutile, zircon and monazite together with lighter ones can be pre-
concentrated by gravity treatment so as to separate the bulk of the silica,
garnet, epidote etc. The heavy-mineral concentrate thus produced can
be treated on the lines indicated in Fig. 265, 266. 36 Recovery of values by
H.T.S. from such concentrates is usually high. 34 The method is particularly
interesting where high-grade products must be made and the densities of the
mixed minerals lie too close for effective gravitational separation to be used.
Good response has been obtained from chromite-bearing sands, tungsten
mixed minerals, microlite and tantalite, and cassiterite and columbite. Ad-
vanced laboratory tests on potash-bearing ores have been described 37 •
Pilot-scale work on coarse pebble phosphate 38 and feldspars 39 has bepn en-
couraging.
One important industrial application is the concentration of diamond from
associated gravity concentrates produced by D.M.S., jigging etc. The
diamond is distinguished from the accompanying minerals by its very
low conductivity40. In earlier appliances trouble was experienced when
the gravels harboured such metallic objects as wire and nails, which
gave trouble by short-circuiting the pin-point electrodes used. This
has been remedied by shielding these points with a strip of phenolic
laminated material (Celeron). This has the effect of intensifying deflection
of the gravels. The strip consists of layers of materials with different insulat-
ing qualities which cause it to act as a condenser. This leads to crowding of
the emerging field lines and thus increases the intensity of ion bombardment
in the air gap. The commercial separator replaces the pin-pointed electrode
596 Mineral Processing-Magnetic and Electrical Separation
Beach Deposit
!~----SPirals
!
Tails t
Rotary Drier
r-- H· h . ts . I
Ig -TensIon eparatlon
t
(Zircon, Monazite) (llmenite-
r l
R~~
GmV;" SeP"a.;on !
Magnetic
Se"''''';°1
r
Quartz-
Epidote
Tails Innren;«
Calcining Kiln
Recleaning
High-tension
Separation
----,_~------..J
l
RutIle
~
Magnetic Separation
-
z;ln ReeleLn.
~
Monazite
Fig. 265. TreClfmellf of Beach Deposit (after Carpenter and Griffith, Ameri-
can Institute of Mining and Metallurgical Engineers, Feb. 1957)
Mineral Processing-Magnetic and Electrical Separation 597
Dredge
1
Screen;n. and CI.,,;ficadon . ·~:l
J
Gravel
r--JigS-1 Slimes to
t Waste
Tails
Rotary Drier
I
t
1,'-----'
High Tension Separation
~ and Screening
t
f
Quartz Magnetic
--- Magnetic Separation
separation ~
Ilmenite 1
H igh-Tension
Separation ~
t Tails
Calcining Kilr. =
To Head_ High Tensiot Separation
Feed *
Air Table -_---------'
t
Magnetic Separation - _ - - - - - - '
~
Zircon
Fig. 266. Treatment oj."-"arine Alluvial Deposit (after Carpenter and Griffith)
598 Milleral Processillg-Magnetic and Electrical Separatioll
References
I. Goldman, J. E. Maglletism's Role ill the Nell' PhysiCS, Ford Motor Co.
2. Willis Jackson. (1956). "The Ferrites". Proc. Illst. EI. Eng., Oct., 100B.
3. Hogan, C. L. (1960). Scieflli{ic American, June.
4. Pryor, E. J. (1963). Dictionary (!f Milleral Technology. Mining Publications.
5. Edgar Allen, Ltd. Maglletic Materials, Pub. 153.
6. B.S. 3100, Brit. Standards Inst.
7. Herzog, M. E., and Backer, M. L. (1963). I.M.P.e. (Calllles). Pergamon.
8. Eketrop, S. (1957). I.M.P.e. (Stockholm), Almqvist & Wiksell.
Ibid. (1960). I.M.P.e. (LOlldoll), I.M.M.
9. Dahlem, D. H., and Sollenberger, C. L. (1963). I.M.P.e. (Calllles), Pergamon.
10. Boucraw, M., Koskas. R., and Michard. J. Ibid.
I \. Leon, J. B., and Worner. H. K. Ibid.
12. Mayer, K. Ibid.
13. Gagyi-Palffy, H., Palfi, G., and Halasz. H. Ibid.
14. Hencl, V. Ibid.
15. de Vaney, F. D. (1960). I.M.P.e. (LOlldoll), I.M.M.
16. Roe, L. A. (1958). TrailS. A.I.M.E. (I.M.M.), 211.
17. Goltz, A., and Neumann, K. (1955). Erzallfhereitllllg-salllagell ill West
Deutschlalld, Springer-Verlag.
18. Runolinna, U. (1961). Acta Polytechllica, ScolldillOl'io (Chem. Illd. Mer.), 16
(303).
19. Cavanagh, P. E., and Williams, E. W. (1957). TrailS. Call. I.M.M., Sept.
20. Runolinna, U. (1957). I.M.P.e. (Stockholm), Almqvist & Wiksell.
21. Kiblstedt, P. G., and Skold, B. (1960). I.M.P.e. (LOlldoll), I.M.M.
22. Palasvirta, O. E. (1959). TrailS. S.M.E., 214.
23. Bronkala, W. J. (1959). Ellg. alld Mill. J., 164.
24. Jones, G. H. (1960). I.M.P.e. (LOlldoll), I.M.M.
25. Wyman, R. A., Stone. W. J. 0 .. and Hartmann. F. H. (1962). Dept. of Mines,
Ottawa, T.B.36, June.
26. Barthelemy, R. E., and Mora, R. G. (1960). I.M.P.e. (Lolldoll), I.M.M.
/l4il/eral Processing-Magnetic and Electrical Separation 599
ReFerences-continued
27. Kakovsky, I. A., and Revnivtzer, V. I. (1960). I.M.P.e. (London), I.M.M.
28. Fraas, F. (1962). Electrostatic Separation of Granular Material, U.S. Dept. of
Interior B. of Mines Bull.. 603.
29. Linari-Linholm, A. A. (1950). J. Chern. Met. and Min. S. A/i'ica, Oct.
30. Fraas, F., and Ralston. O.C (]948). A.I.M'£', T.P. 2408, July.
31. Hatfield, H. S. (1924). Trans. I.M.M. (London), 33.
32. Holman. B. W., and Shepherd, St. J. R. C (1924). Trans. I.M.M. (London),
33.
33. Hudson, S. H. (1953). Recent Dereloprnents in Mineral Dressing, I.M.M.
(London).
34. Barthelemy, R. L, and Mora, R. G. (1960). I.M.P.e. (London), I.M.M.
35. Nilkuha, C, and Hudson, S. H. (1962). Aust. I.M.M., Dec.
36. Carpenter, J. H., and Griffith, R. F. (1957). A.I.M.E., New Orleans Meeting,
Feb.
37. Le Baron, I. M., and Knopf, W. C (1958). Trans. A.l.M.£., I.M.e. Cpn.,
Oct., Chicago.
38. Oberg. F. N., and Northcott, E. (1958). I.M.C Cpn. Chicago.
39. Northcott, E., and Le Baron, I. M. (1958). I.M.e. Cpn., Chicago.
40. Linari-Linholm, A. A. (1962/3). J. S. Ai, I.M.M., 63.
41. Plaksin, I. N., and Olofinsky, N. F. (]959/60). Trans. I.M.M .. 69.
CHAPTER 20
These stages are not always followed. New methods are often worked out
before the research scientist becomes aware of the problem and of the
empiric solution which has been arrived at in the plant. Hence, in practice
methods are at times used which are either harmful, needless, or only re-
quired because the whole problem has not been correctly analysed. In
mineral processing, as elsewhere, it is sometimes possible to maintain good
practice despite ignorance of basic principles, but the cost is apt to be ex-
cessive. Improvisation, occasional helplessness and frustration are the price
paid for lack offundamental knowledge. Logical activity, a keen and abiding
interest, and accurate control are among the pleasures which result from
correct application of scientific principles.
The orderly development of an ore body from original discovery to-and
through-its processing treatment usually proceeds along the following lines
(Table 41):
TABLE 41
TABLE 41--continued
Stage Place, method
(d) Tonnages opened out Central lab., batch tests
(e) Alternative treatments known Central lab., miniature pilot tests
(f) Flow-sheet decided on Full pilot-scale tests on mill site
(g) Concentrator built and operating Mill lab., routine tests
(h) Variations from original treatment
plan Mill lab., aided by central lab. at need
under the microscope, and check on the thoroughness with which it has been
wetted is then possible. Porous minerals may give misleading indications.
The particle shape resulting from comminution is a possible factor in treat-
ment since it affects both specific surface and rubbing contact. Cleavage,
liable to be associated with structural weakness or obdurate deep sorptive
contamination, can be observed under the microscope.
The refractive index aids in the identification of transparent minerals, which
merge with the ambient liquid when immersed in one having the same index
value (Table 422.)
TABLE 42
Refractive Refractive
Liquid Index Mineral Index
Ethyl alcohol. . \·36 Fluorite .. 10434
Chloroform \·44 Orthoclase 1·523
Glycerine 1·47 Gypsum .. \·524
Castor oil \ 048 Quartz .. 1·547
Xylol.. \'49 Muscovite 1'582
Benzene \·50 Beryl \·582
Cedar oil \·5\ Calcite .. \·60\
Clove oil, or monochlorbenzol 1·53 Topaz 1·622
Fennel oil \·54 Tremolite 1·622
Nitrobenzene, bitter almond oil, or anise oil \·55 Dolomite \·626
Monobromobenzene \·56 Aragonite 1·633
Bromoform .. \·59 Apatite .. 1·633
Cassia oil \·60 Barite 1·640
Monoidobenzol \·62 Diopside .. \·685
a-Monochlornaphthalene \·64 Kyanite .. \·723
a-Monobromnaphthalene \·66 Epidote .. \·750
Cadmium borotungstate \·70 Corundum \·765
condition of the freshly severed ore and to its ageing during transport and
storage. Two weighed samples are finely ground in distilled water. One is
immediately filtered and the other after standing for a defined period, pre-
ferably after settlement and exposure of the moist solids to air before re-
mixing with the decanted water and filtering. Both filtrates are evaporated
to dryness, weighed and analysed. Original compounds which might in-
fluence the pulp climate, and those which may develop as the result of ageing
before treatment, are thus discovered, and the final flow-sheet can be designed
to handle any problem thus exposed.
Liberation mesh gives a clue to optimum mesh of grind (m.o.g.), which is
further considered below. In this connexion the microscope yields valuable
information. It can reveal the mode of interlock of the ore minerals and the
type of association in particles of middlings, thus indicating the efficiency of
liberation. When it is possible to locate particles of valuable mineral which
have reported with the tailings, microscopic inspection may suggest, or even
expose, the reason for their failure to respond to treatment and thus open the
way to a remedy. Surface contamination, inadequate liberation, and partial
oxidation are usually revealed by thorough micro-inspection.
In order to examine a specimen of ore a properly polished surface must be
prepared. Harsh pressure and the use of high-speed abrading grind lead to
deformation, formation of a Beilby layer, a matted surface in which undue
removal of the softer constituents has occurred, or to deep scratches. Halti-
mond 4 recommends the use of a slow-running cool surface such as can be
mounted on a horizontal turntable, run at 50 to 100 r.p.m. With this a series
of sized Quick-cutting abrasives is used, thorough removal of each from both
lap and specimen preceding the use of a finer size. The specimen can be
trimmed to preliminary shape by a rock-slitting saw (a thin disc armed with
diamond-dust cutting edge and rotated at a high speed). Quite good work
can be done by hand trimming to about a I in. cube and producing the surface
by light hand pressure of the specimen on a piece of wetted plate glass sprinkled
with abrasive. Those in common use include jeweller's rouge, silicon carbide,
magnesia alumina and diamond dust. Aluminium foil, paper-backed and not
surface-varnished, makes an excellent medium-holding surface if trimmed and
attached to a small turntable. To examine aggregates of mineral sands or
crushed products the sample can be set in lucite and then polished in the same
way. A bibliography of microscopy as applied to mineral processing accom-
panies a paper by Amstutz'. His classification of intergrowth patterns is
shown in Fig. 267.
Opaque specimens are studied in the metallurgical microscope by means of
a vertical reflector, used either with monochromatic or polarized light.
Another technique now coming into use employs the x-ray micro-analyser
(Fig. 268)6. Electrons are focussed on the specimen and the resultant x-rays
are viewed by reflection. A method of identification which mixes electro-
chemistry with microscopy is chromography7. In one developmentS the
smooth surface of the specimen is pressed against cartridge paper soaked in an
appropriate attacking reagent while current is passed for about a minute.
The ions then transferred from the minerals to the paper are rendered visible in
characteristic differentiated colours by development and the distribution of
Mineral Processing-Testing and Research 607
()
Ty pe 1(0) Type I (b) Type 1(e)
Type 3(0)
Fig. 267.
Type 3 (b) •Type3(e)
- electron beam
1 micron diameter
electron spot on
specimen
specimen (5)
Amenability Tests
Bulk Samples
The field preparation of a large sample designed for exhaustive test work is
described in Chapter 21. It may be five or more tons in weight, and repres-
ents a large and partially proved body of similar ore. The testing laboratory
may be called on to do one or more of three things: (a) it may investigate in
detail the response of the ore to various methods of treatment, reporting in
terms of percentage recovery and indicated process cost; (b) and arising out
of (a) it may suggest the basis for a flow-sheet which will guide mill design,
full-scale equipment and commercial processing, (c) it may compare current
mill practice in the case of an already working mine with possibilities for
improvement revealed by the tests.
Work on the sample will usually take some months to complete, starting
with batch tests on small quantities (a kilogramme or less) and ending with
small-scale continuous runs, called pilot tests. The first step is to cut a
reliable head sample which will be truly representative of the bulk. This is
subjected to a "complete analysis" and must be sufficiently trustworthy to be
considered the representative feed in all further tests on the raw material.
Some hundreds of small batch samples will be tested during the work, and
part of the ore may be needed for dense-media tests and pilot runs on freshly
ground feed. The whole sample is therefore gently but thoroughly dried.
This should arrest any further ageing or acidic deterioration that may have
started in transit. The pile is then mixed and an adequate large sample is
withdrawn after coning and quartering. This is sealed up, while still dry and
warm, to preserve it until needed for D.M.S. or pilot-scale testing. The
610 Mineral Processing-Testing and Research
rest of the ore is,screened until it passes the coarsest mesh consistent with the
need for thorough blending and with due consideration of the liberation
mesh of the largest mineral grains observed in preliminary inspection. This
would not include barren gangue, but should be in line with the requirements
of good mixing discussed under sampling in Chapter 21. Probably the
material is now below coarse sand size. It is now given a final mixing and
sealed in small lots (say 7 lb. packets) in polythene bags, care being taken to
check that it is dry. If there appears to be a danger of oxidation, some of the
packets may be given the further protection of a small addition of dry calcium
hydroxide to maintain alkalinity, but the fact must be recorded and the weight
noted.
The final crushing size is chosen with regard to treatment in preparing
these packets. If jigging is contemplated as part of the flow-sheet the ore is
not brought down to sand size, for example. Crushing machines must avoid
local overheating of the ore, particularly at the later stages, where a metal
sulphide could easily be semi-melted in an appliance such as a laboratory
disc mill.
Comminution
If the testing laboratory is suitably equipped the coarse crushing tests can
be made there during the reduction of the bulk sample considered above.
Alternatively, a large sample may be sent to the manufacturer's more special-
ised testing plant. The test sample should be screened through mesh a and
on mesh b, to accord with the usual by-passing of undersize round the crusher.
The sizes are selected in accordance with the desired gape and set-for
example 3 in. and i in.-and between 1-2 cwt. of ore are needed. The
product is in due course returned to the bulk sample and mixed with it.
The crusher setting and throw are accurately measured, and its power
consumption is checked while idling after warming up. The sample is
weighed and fed through the crusher fast enough to maintain a full load, the
time taken and the average power readings being noted. If the crusher is of
the cone type, its setting can be checked by passing a short length of lead
piping (larger than the set) through and measuring the thickness on emergence.
From the figures obtained for various feed and discharge settings and nett
power consumed per ton, the coarse crushing behaviour of the ore is
observable.
Grindability tests are made in free-path crushing appliances. Their
general lines have been described by Bond et al ll . l~. From these the "stand-
ard work index" (W ) can he calculated.l~ This is the kWh/ton applied
612 Mineral Processing- Testing and Research
Nominal ball size Average ball size Number of balls Weight (grammes)
\'50" \·45" 43 8730
J'25" \·17" 67 7197
\·00" 1·02" 10 705
0·75" 0·76" 71 2058
0'50" 0·61" 94 1441
285 20131
examination of the returned grains, coupled with sizing analysis and check
with a suitable heavy liquid, will throw light on such a situation and suggest
useful economies in the final flow-sheet. These arise partly from increased
capacity and partly from reduced overgrinding of value while a lighter
tough gangue is being reduced in size. The final specification for overflow
product in the full-scale plant may include some such instruction as "so°;';
minus I00 mesh and not more than 7 %plus 65 mesh", based on the feasibility
of controlling classification so as to overflow up to 7 % of barren gangue at
the latter size.
Where facilities exist useful information as to the balance between grinding
and classification can be obtained in a miniature continuous operation. The
requirements are an accurate delivery of new feed, a small ball mill with
the necessary facilities for sampling, speed variation and rapid opening up,
and either a classifier or preferably a hydrocyclone closing the circuit. The
relation between feed and overflow size, feed rate and circulating load, crop
load and power consumption can be checked in such a system.
Gravity Separation
where Sh and S, are the specific gravities of fractions which are to be separated
and Sm is that of the separating fluid in which they are borne and through
which they must move to separate exits. Where water is the medium, as in
jigging and tabling, Sm is 1·0, but when an element of dense-media density
enters (as in D.M.S. and the use of spirals) the contributory effects of vis-
cosity and particle size appear. In tabling the shape factor modifies the
separating effect. Broadly, when C> 2·5, all sizes down to fine sands can
be treated. Thus cassiterite (S.G. circa 7·0) is readily separated from silica
(S.G. 2·7), the C value being of the order of 3·5. The C concept takes no
account of the time factor which enters increasingly as the particle size
diminishes. In the case of cassiterite, where direct commercial flotation
does not take over at any stage. treatment of the subsieve sizes becomes
gentler and less efficient down to 30!,- or less, when even a widespread settling
area fails to hal)dle a limited tonnage economically. Thus, in test work for
gravity concentration strong emphasis should be placed on feed preparation
Mineral Processill!{-- Testing alld Research 615
liquid, and the average voidage is therefore a significant factor in its composi-
tion. If a jig bed were composed of material of s.g. 4·0 and voidage such
that the solid-liquid ratio was either 60-40 or 70-30, the weight of a unit
volume of bed would vary by 10% and its mobility and discriminating effect
of passage would vary correspondingly. Tests should therefore take into
account the shape as well as the size and weight of the bedding material. At
the same time, the temptation to control this and any other factor beyond
the condition which could be easily maintained in plant practice must be
resisted. Wear of the bedding material must also be reckoned with. The
variables in a batch jig test, using a jig of the Denver type, include:
(I) Mesh of supporting screen.
(2) Mesh range of feed.
(3) Mesh of bedding material.
(4) Shape of bedding material.
(5) Thickness of bed.
(6) Density of bed, including void water.
(7) Rate of stroke.
(8) Length of jigging stroke.
(9) Rate of feed.
(10) Setting of hydraulic water valve.
(II) Percentage water in feed.
In view of the considerable number of variables involved-and this is
usually the case in good work, whatever concentration process is being
applied-a severely limited number of the possible permutations must be
explored. Fortunately, most of the above items will be found not to have
critical importance in a specific case, but 2, 3, 4, 5, 6 and particularly 10
need close attention. If possible, a rapid means of assessment of results
should be worked out before starting, as it is important that the effects of
each alteration be studied before proceeding with the next variation to be
explored.
Tabling tests can begin with individual treatment of each size from about
50 mesh down to, say, 20IL on a Haultain superpanner. If material can be
concentrated by gravity methods (tabling. vanning, and spiralling included),
it can be concentrated on a superpanner. If this instrument fails, no com-
mercial method will succeed on fine sizes. The superpanner is a miniature
shaking table into which are built most, if not all, of the motions used in
panning. A sample of a few grams can be tested, or some hundreds of
grams can be worked in successive additions and particle removals. There
is, in fact. no great virtue in using more than a light charge when testing for a
gravity process, but bigger quantities are treated when search is being made
for particles of heavy minerals only present in very small amounts. The
advantage to the microscopist of being able to examine a sample of, say,
half a gram which should contain all the heavier minerals from a total of
half a kilogram are obvious. Any desired texture of decking can be used on
the superpanner, linoleum and stainless steel being sufficient for most work.
Adjustments are for side and end shake. slope, bump, stroke length and speed,
and sluicing water. A variation on the superpanner has been described 1 ". It
Mineral Processing-Testing and Research 617
for each type, and the separating devices are controlled accordingly. To
check the results obtained with the small sample, a bulk sample may be run
through a washery test, the separating devices being adjusted to conform to
the laboratory findings. These tests have the further value of showing what
amount of particle disintegration must be expected in full-scale treatment of
the coal, and what problems are likely to arise in connection with the clarifica-
tion, re-use and change of the large quantities of water used. Some coals
are found to be very simple in their constitution and their breaking liberation,
so that a clean-cut density separation will drop the shale and lift the coal.
Others have a wide middling band, which may need further liberation and
separation. Laboratory batch and pilot tests can thus show the ash ;:ontent
for a given range and density. They can also be used to show how much re-
crushing of various middling sizes is needed to liberate them to treatment point.
From the ash analysis of each weighed fraction floated in the heavy-liquid
tests five graphs can be constructed. Cumulative yield percentage at each
S.G. of the parting liquid is plotted against the mean percentage ash of the
total floated at that S.G. to give the cumulative float curve. The cumulative
yield percentage of sinks at each S.G. is plotted against the mean percentage
ash to give the cumulative sink curve. The characteristic ash curve is
obtained by plotting percentage ash of successive fractions against cumulative
percentage yield of floats at the mid-point between the S.G. limits of each
fraction. The S.G. curve plots the upper value of the S.G. of successive
floated fractions against the cumulative percentage weight. This shows the
percentage of clean coal yielded by perfect separation at any S.G. The
Tromp, or distribution-partition curve relates the percentage floating in a
heavy liquid to the S.G. of that liquid.
Gold Amalgamation
The use of mercury to entrap gold goes very far back into history, mention
being made of its use by Vitruvius (13 B.C.). Its use became important from
the sixteenth century onward, and with sluicing and panning was virtually
the only gold-catching process in use before the arrival of cyanidation.
In mineral dressing, amalgamation is the process of separating gold and
silver from their associated minerals by binding them into a mixture with
mercury. Many other metals can be amalgamated, the list including zinc,
tin, copper, cadmium, lead, bismuth, and sodium. The word "amalgam"
is used to cover a variety of states, from those in which the solid metal is
wetted down into the mercury in a putty-like mixture to those in which the
metal dissolves in the quicksilver.
Gold is slightly soluble in mercury (0·06% at 20 c C.), and two compounds
(AU19Hg4 and AUI9Hga) have been isolated. When nitric acid is used to
dissolve away the mercury from amalgam, the gold particles are recovered
apparently in their original form. The wetting down of gold into mercury is not
alloying, but a phenomenon of moderately deep sorption, involving a limited
degree of interpenetration of the two elements (solid gold and liquid mercury).
In all wetting phenomena the surface tensions of the substances involved
Mineral Processing-Testing and Research 619
influence the nature of the reaction. All naturally occurring systems strive
to arrange themselves in a state of maximum order, at a minimum potential
energy. When water stands in a burette, a meniscus is formed with an up-
turned rim because the surface tension between glass and air is higher than
that of water and air, so that water is drawn up against the force of gravity
until the glass/water/air system has arranged itself at the maximum possible
stability. The surface tension of water is about 73 dynes/cm. That of
mercury is 375, and its meniscus turns down at the rim, showing that the
glass/air tension is sufficiently below that of mercury/air for repulsion to
take place. Gold is readily wetted by mercury because its surface tension is
still higher, and it is therefore absorbed into the quicksilver. There is also a
third factor at work-the dense-media effect. The density of quartz is
2'7, of mercury \3'5, and of gold about 19·2. With mercury as the separating
bath, quartz floats and gold sinks. Gravitational force acts to drown the gold
in the mercury, apart from any lowering of surface tension which might
result from sorption. Indeed, gravity may well be the most important force at
work since extremely small particles of gold cannot be amalgamated. Surface
contamination, however, may be partly responsible for this failure. If a
gold particle has reduced its surface tension sufficiently-e.g. by becoming
coated with an oily film-attraction into the body of the mercury may be
replaced by repulsion for reasons similar to those which affected the meniscus
in the mercury/glass/air system.
One vitally important condition for efficient amalgamation is that the
surface of both gold and mercury must be clean. "Clean" in this connection
implies purity at the surface, of a type which is studied more closely in the
chapters dealing with the flotation process. There must be no substantial
lowering of surface tension through the presence of impurities adsorbed to
either surface. Another is that the mercury must offer an adequate receiving
surfac.e to the particle of gold. If mercury is divided into minute droplets-
"f1oured"-it cannot open its surface to gold, nor do these droplets readily
reunite. A serious loss of mercury occurs when it is handled so roughly that
it becomes floured.
Mercury can be contaminated or "sickened" by several substances liable
to be present in an ore pulp. Fatty-acid oils and their salts (greases) finding
their way to the mill from oil leakages are a hazard with mechanised mining.
Such oils have a low surface tension-30 dynes/em or less-and if they make
contact with mercury they coat its surface. This "sick" mercury no longer
attempts to minimise its total surface tension by gathering itself into a sphere
but rolls sluggishly as a tear-shaped globule. The attractive force across the
interface is now that of the contaminating oil, and amalgamation does not
result from contact of this sick mercury with clean gold. A further difficulty
is that an oil film attracts graphite, talc, calcium earths, and metal sulphides,
which tend to form an impermeable film even if the mercury below has not
been sickened. Mercury is blackened by sulphur and by some sulphides,
particularly those of antimony, arsenic, and bismuth. This tendency is
minimised if the pulp is alkaline.
Amalgamation can be applied to run-of-mine ore or (more usually) to a
gravity concentrate. provided the metal is free-milling.
620 Mineral Processing-Testing and Research
Chemical Extraction
Flotation
Among the objects of these tests are:
(a) Research on unusual ores.
(b) Control of mill operations.
(c) Testing of new reagents.
(d) Trial of new methods.
(e) Improvement of concentrate grades.
(f) Check on new developments in the ore body.
(g) Amenability tests on drill cores and other samples.
(h) Development of a flotation flow-sheet.
Quite useful rough tests can be made with no better apparatus than a
stoppered test-tube, cylinder, or separating flask in which the pulp and
reagents can be shaken and from which a froth can be removed. The last
stage of grinding for any such test must be under water, using a pestle and
mortar if no better apparatus is available,
Scientific testing for flotation is based on these main principles:
(I) N-bubbles do not adhere strongly to most clean mineral surfaces, but
to suitable absorbed films at such surfaces.
(2) Particles borne up in a flotation froth are either aerophilic owing to a
hydrocarbon film, or are temporarily entrained.
(3) In the latter case, they can be got back into the pulp by simple cleaning
of a correctly textured froth.
(4) If mineral particles should have become froth-borne after conditioning
treatment and are not, there is a physical reason for the failure.
(5) This reason is probably:
(a) Inadequate surface liberation.
(b) Incorrect conditioning.
(c) Slime masking.
(d) Too small N-bubbles.
(e) Too large particles.
(f) Too violent aeration.
(g) Short-circuit through the aerating zone of the flotation cell.
(6) An adherent aerophilic film on a mineral surface lattice is developed
by conditioning with suitably ionised collector.
(7) The conditions under which such film is developed can be made
specific for floatable minerals by methods broadly described in previous
chapters.
(8) Conditioning treatment directs suitable collector agents on to specific
mineral surfaces, after the latter have been prepared for surface
activation.
(9) Most of such useful surface reactions are ionic. They lead to partial
coverage of the mineral's discontinuity lattices with a layer of polarised
collector.
(10) In addition to specific collection, the possibility of increasing the
wetability of minerals not required in a froth may exist. This calls
for the use of depressing agents.
Mineral Processing-Testing and Research 623
The facts listed above, logically employed to meet the case of a specific ore,
make it possible to apply scientific method to the problem of testing a
mineral for the purpose of floating it. Trial-and-error methods are not
appropriate in the light of modern knowledge of surface physics.
A common criterion of good technology is reproducibility of results when
a method which has proved successful is repeated. This though good as far
as it goes, would leave flotation test-work in an unsatisfactory and uninspired
state. If a scientific approach is made and the tests are all reasoned out and
recorded, whether or no they have fulfilled expectation, the work ceases to be
a "cook-book" routine and is set on a proper scientific basis.
Observation, analysis, and step-by-step variations of the conditions under
which the ore is subjected to flotation require skill, experience, and patience.
The list of factors affecting efficiency of concentration given below, though
long, is far from complete, and each important factor (for a given ore) must
be varied in isolation during the tests, the remainder being held unchanged
as far as possible. Assays are required on test products and take time to
perform, even if a short method can be relied on. The ultimate success of
the commercial operation reflects the care with which these tests are made,
and the cost of making them is a mere fraction of the money which will
accrue from improved efficiency of extraction if they have been correctly
performed and accurately established in the plant. The factors include:
ing on research into cell aeration, part of which has been published~:ul,
cells were designed which embodied some of these requirements.
Reagents can conveniently be added from a dropping tube which has been
calibrated. With a charge of one kilogram of solids, one millilitre of a
5 % solution of reagent is equivalent to 0·1 Ib./short ton. The tube for
introducing 100 o~ frothing agents can also be calibrated, though there is
some inaccuracy due to change of viscosity (this varies inversely as the lab-
oratory temperature, and specifically with different sources of pine oil).
Pine oil, like other reagents of low solubility, can be thinned down for
sensitive addition by emulsifying, say, five drops with 50 ml of water in an
agate mortar or by solution in ethanol.
Flotation cells can be cleaned between tests by scrubbing with a solution
of soda and then running them with clean sand and water. Tests should be
made in distilled water to begin with, mill water only being used after a
satisfactory treatment scheme has been evolved. A laboratory rod mill
loaded with - IO-mesh ore roughly simulates a standard mill receiving
-! in. feed. Aeration of the cell can be checked by filling a graduated
cylinder with water, inverting it under water in the cell, and timing the rate
of fall of the water-level in the cylinder. This should be about the same in all
parts of the cell at a cross-section a little below the surface.
When hand skimming, a rougher froth should be scraped somewhat deeper
than a cleaner froth. A final scavenger froth should be removed as com-
pletely as possible, even if some pulp is thereby included. To aid this
scavenging "pull", the pulp in the laboratory cell must be diluted till it nearly
brims over. [n the specially designed cell mentioned above this is avoided
by use of a tilting basin, and loss of pulp density and particle support is
thereby avoided. In a laboratory cell designed by the author the froth-
removal is performed by suction through an array of small tubes supported
on the housing of the impeller shaft. This array can be slid so as to maintain
the desired height of froth blanket. Vacuum for operating the suction can
be tapped from the centre of the zone swept by the cell's impeller, using a
receiving vessel to trap the removed froth. Concentrate grade can be
improved by restricting aeration, diluting the pulp, or using a weak frother,
apart from manipulating conditioning treatment. Recovery can be increased
by reversing these methods and by stiffening the froth.
When non-sulphide ores are being tested, particularly in the presence of
slimes, collectors containing fatty acids or amines should be added in
successive "starvation" quantities with skimming between additions. Tests
in which oxidised metal surfaces are treated by sulphidisation are best
performed after straight sulphide flotation has been completed. Skimming
should be completed quickly as the superficial sulphidisation is fleeting and
aeration of the pulp aids reversion to the oxide.
Table agglomeration tests on - IO-mesh + 100-mesh mineral can be per-
formed in a large watch-glass. The feed is deslimed, stirred with reagent and
frother into a thick slurry, and shaken vigorously in a test-tube to introduce
air. The slurry is then poured into the watch-glass using sufficient pulp to
ensure quiet peripheral overflow. If glomerules form they will overflow into
a surrounding \esse!, the tailing remaining in the watch-glass.
626 Mineral Processing- Testing and Research
The induction time for adhesion of air bubbles to mineral particles is one
determinant of flotation time, tenacity of mineral retention in the froth
blanket and general efficiency of levitation. Using a specially developed
technique Eigeles et a1 25 . 26 find induction time to be affected by such time
factors as conditioning period and age of air bubbles. In advanced test
work the optimum period for adsorption of reagents. also the pulp climate
and aeration during this period should be examined. Ionic activation and
depression introduced with the mill water must also be checked at a suitable
stage, to ensure that laboratory tests conducted in distilled water are repro-
ducible on the plant scale with local water. Bicarbonates. sulphates,
chlorides, calcium and magnesium ions and their possible build-up in re-
circulating mill water should be observed in this connexion. 27
Electro-Magnetism
Here tests are made to show the relative reponse of the ore's minerals to
the influence of fields of magnetic or electrical force. The chief matters
investigated may be summarised thus:
( I) Gel/eral Tests
(a) Particle size effect
(b) Particle density effect
(c) Degree of liberation
De-Watering
Tests may be called for in connexion with the following main operations:
(a) Control of pulp density.
(b) Thickening before filtration.
(e) Filtration.
(d) Flocculation.
(e) Counter-current decantation.
(f) Replacement of fouled water.
(g) Maximum extraction of pregnant solution.
surface and provision is made for use of vacuum or pressure, with appropriate
equipment for trapping filtrate, measuring flow rate, and for recording the
volume and rate at which air is drawn through during the drying test. The
slurry is adjusted as to flocculation, solid/liquid ratio and temperature and
kept mixed. The test leaf is gently immersed, face down, just below the
surface of the sample under recorded conditions of vacuum.
When batch tests have completed the closely controlled examination of the
ore constituents, and locked cyclic tests have shown as much as possible of
their inter-dependence, development of a new flow-sheet is carried on in a
continuously operating system. Here the optimum working conditions are
more clearly displayed and a compromise is reached between the technically
ideal treatment and one economically viable. Pilot tests in miniature mark
the first stage of scale-up in which this working adjustment begins. They are
best made in the central laboratory on two or three tons of the original bulk
sample. By starting on a small scale and with a feed rate well below 100
Ib/hour close scanning of the flowline is possible. The problem of re-
circulating products can be met by the use of holding tanks if full continuity
cannot be achieved on this small scale. Each variation of the flow-line
suggested by batch tests should be examined. Well-planned tests will then
confirm the general lines of the process and disclose potential difficulties not
shown in batch testing. When this work has been digested, together with
any further batch tests thus shown to be desirable, pilot testing should be
continued on a larger scale on location. This move away from the highly
specialised central facilities should not be made until the emphasis of the
test-work has shifted from basic research to that mainly concerned with
process economics and plant design. From now on tests are made on fresh
ore and use the proposed mill water. The data recorded during the tests
will deal not only with such things as particle sizes, recovery distribution and
grade, but will also record reagent consumption, power required and any
other matters of economic importance. A start may now be made with the
training of technical workers later to be employed in the concentrator. If
the ore changes from horizon to horizon the effects of blending and of treat-
ment of varying types can be compared. Change in grindability, liberation
mesh and their effect on products are now scrutinised in their bearing on
circulating load, plant capacity at bottle-neck points and optimum rate of
feed. This last is of direct concern to the mine's rate of ore extraction.
These continuous tests must show whether there is a difficult build-up of
re-circulated material anywhere, whether in near-liberation mesh tough
grains of ore or in the return treatment circuits (middlings and scavenger).
These matters affect the provision of re-grinding facilities and the decision
on optimum grade of concentrate and percentage recovery. If the concen-
trates are being sold, large samples may be called for testing by the proposed
market and the specifications drawn as the result of their tests may require
alterations in the treatment plan.
630 Mineral Processing-Testing and Research
Even a very large pilot plant cannot provide complete answers to questions
which depend mainly on changes in the ore as the mine develops. Too large
a plant is costly, and can defeat its own object by making the findings of a
trial run complex and obscure. Possible effects on flotation of seasonal change
in the water supply should be assessed where possible. Tests should be run
long enough to ensure that anything liable to upset the process by building
up to unmanageable proportions is disclosed, so that it can be traced back to
its cause and properly dealt with in the final operating plan and engineering
design. Surging and an erratic flow-rate are wasteful and should be dealt
with once for all in the pilot plant, where they can be exposed by suitable
instrumentation and measurement of flow-rates.
An automated pilot-scale flotation plant has been described,34 with a feed
rate between 33 and 66 lb/hour. It handles a minus lO-mesh feed and
includes automatic sampling and pulp level control and has proved reliable
in the testing of a variety of ores. One rule-of-thumb suggestion regarding
scale-up35 is that pilot tests should be performed on at least fifty times the
scale of bench and unit tests, and that the full-scale plant should have a
capacity of not more than fifty times that of the pilot plant. A final test-
production scale of between 10 and 25 tons per day is often found satisfactory.
One variation is to test only a novel unit operation on the pilot scale and to
rely on experience for the more common sections of the flow-sheet. Where
mechanised sorting or dense-media separation are proposed, portable plants
are available in many mining centres for such specialised unit tests, which
should be made before the overall plant capacity is estimated.
As the pilot plant is used intermittently special care is needed with regard
to cleanliness and the avoidance of corrosion. Plastic piping is non-corrosive
and easily re-coupled when testing flow-line variations or flushing down
between tests. It should be installed self-draining with no horizontal runs or
crests and bends where solids can lodge. Reliable flow-control valves are
needed, though in the final plant only on-off ones are usual. Accessibility
and safety in the lay-out are important, with good lighting plus plug points
for inspection lamps. Among the matters recorded during the tests are
temperatures, changes ofpH along the flow-Hne and pressure fluctuations. The
sequence of sampling along the flow-line should be planned to allow for time-
lags in the course of the pulp down-line, if the bearing of one sample on those
following is to be interpreted correctly. Periods between sampling must
be sufficiently small to expose fluctuations during treatment, but must not
overwhelm the laboratory which handles them. Two types of analysis are
required. Control assays feel the pulse of the current operation, while
special and detailed ones track down the chemical and physical causes of
variations in operating efficiency.
A second type of pilot testing is that designed to improve an existing
operation. It may be feasible to retain the original pilot plant on care-and-
maintenance for this purpose, or to design a large mill so that one treatment
line can at need be used for pilot tests. In chemical engineering the technique
called "evolutionary operation" is well established. It has been defined36
as a method of process operation which has a "built-in" procedure to increase
productivity. To put a designed experiment right into a main production
Mineral Processing-Testing and Research 631
of the mill affects the capital locked up in stores, suitable protection against
climatic conditions, and such matters as tailings disposal. Ample room for
expansion should be reserved in site-planning. When the other buildings
come into being expansion is restricted, should a big mine develop. The
character of the process may have to be modified to suit the labour force.
A few highly skilled men can work a difficult process, but too many under-
takings are wedded to a "cheap native labour" policy which is incapable of
running a complex ore-dressing operation smoothly.
Mineral-dressing costs fall into two divisions, capital and operating. The
former must be charged out as depreciation over the working life of the mine.
plus an obsolescence reserve, in case some new process is invented which
must replace part of the flow-sheet. If the mill is near a smelter, high main-
tenance costs may be needed to offset the corrosive effect of the sulphur-
acid fumes. Flotation plants are more costly to install than cyanide and
gravity mills.
Working costs include wages, stores, power, and overheads. Breakdown
allocations of cost should be made monthly, so that the working cost of each
operation in the mill is kept in sight. These include crushing, conveying.
grinding, pumping, classifying. lighting, heating, concentrating, samp-
ling, thickening, filtering, tailings disposal, and shipping, with detailed
analysis of the components of each cost. These figures disclose the insidious
entry of wasteful practices and stress the importance of full loading of the
plant, since labour is one of the main items of costs and is the same whether
work proceeds at full or part capacity. The same applies to power and
overheads. The mill should therefore never be so run as to risk breakdown
leading to serious stoppage.
References
I. Pryor, E. J. (1963). Dictionary of Mineral Technology, Mining Publications.
2. Winchell, A. N. (1951). Elements of Optical Mineralogy, Chapman and Hall.
3. Hosking, K. F. G. (1960). Mining Magazine, May.
4. Hallimond, A. F. (1963). Mining Magazine, April.
5. Amstutz, G. C. (1961). 50th Anniv. of Froth Flotatioll, Vol 2, July, Colorado
Sch. of Mines.
6. Allenden, D., and Mulvey, T. (1962). A.E.1. Eng., May/June.
7. Williams, D., and Nakhla, F. M. (1950/1). Trans. I.M.M. (LOlldoll), 60.
8. Hosking, K. F. G. (1963). Mining Magazine, July.
9. Pryor, E. J., Blyth, H. N., and Eldridge, A. (1953). Recent Developmellts ill
Mineral Dressing, I.M.M. (London).
10. (1963). B.S. 3406 Part 4, Br. Standards Inst.
II. Maxson, W. L., Cadena, F., and Bond, F. C. (1934). TrailS. A.I.M.M.E., 112.
12. Bond, F. C. (1949). Trans. A.I.M.M.E.. 183.
13. Bond, F. C. (1952). Min. Eng., 4.
14. Warren Spring Laboratory. (1962). Min. Proc. Information Note, No.3.
15. Muller, L. D., and Pownall, J. H. (1962). Trans. I.M.M. (LOlldoll), 71.
16. (1942). Rev. B.S. 1017. Sampling of Coal and Coke, Br. Standards Instn.
17. Kakovsky, I. A. (1963). I.M.P.e. (Cannes), Pergamon.
18. Parker, O. J. (1957). AI/st. I.M.M .. Part I. June.
19. Miller, R. P., Napier, E., Wells, R. A., Audsley, A., Daborn, G. R. (1963).
Trans. I.M.M. (London), 72.
20. Hancher, c..W. (1959). Eng. and Min. J., March.
21. Maltby, P. D. R. (1959/60). Trans. I.M.M. (London), 69.
Mineral Processing- Testing and Research 633
References-continued
22. Everest, D. A., and Wells, R. A. (1963). I.M.P.C. (Cannes), Pergamon.
23. Pryor, E. J., and Liou, K. (1950). Trans. I.M.M. (London), 58.
24. Pryor, E. l, and Dzieniewicz, J. (1950). Trans. I.M.M. (London), 59.
25. Eigeles, M. A., and Volova, M. L. (1960). I.M.P.e. (London), I.M.M.
(London).
26. Eigeles, M. A., and Volvenkova, V. S. (1963). I.M.P.e. (Cannes), Pergamon.
27. Agey, W. W., Salisbury, H. B., and Placek, P. L. U.S. Bureau of Mines
Research Inst. 6189.
28. Vijaykar, S. V., and Majumdar, K. K. (1962). Mining Magazine, July.
29. Cross, H. E. (1963). Jnl. S. Af I.M.M., Feb.
30. (1959). Cyanamid Flocculants, American Cyanamid Co.
31. Dorr, J. V. N., and Bosqui, F. L. (1950). Cyanidation and Concentration of
Gold and Silver Ores, McGraw-Hili.
32. Dahlstrom, D. A. (1959). Trans. Can. I.M.M., Feb.
33. Gieseke, E. W. (1962). Trans. Am. S.M.£., Dec.
34. Raffinot, P., and Formanek, V. (1963). I.M.P.e. (Cannes), Pergamon.
35. MacDonald, R. D., and Stevens, F. M. (1959). Min. Eng., May.
36. Box, G. E. P. (1958). 43rd Ann. Meeting Tech. Assn. of Paper Industry, New
York.
CHAPTER 21
Purposes in Sampling
The first and most obvious reason for sampling is to acquire information
about the ore entering the plant for treatment. The second is to inspect its
condition at selected points during its progress through the plant so that
comparison can be made between the optimum requirements for efficient
treatment and those actually existing, should these not coincide. The third
is to disclose recovery and losses, and to learn how to improve the former and
reduce the latter. Just as in military circles reconnaissance before the fight,
information during the fight, and a conference on "the lessons of the scheme"
are essential to intelligent control, so, translated to the civil field, does the
mill superintendent look ahead, keep his finger on the pulse of his charge, and
hold "post-mortems" on such failures as high tailing loss and poor concen-
trate. "What's gone is past", but knowledge of the reason for failure is the
first step toward preventing its repetition.
No readily visible financial profit results from sampling, which is a rather
dull and costly routine activity. Without its methodical and careful fact-
finding, however, the losses during treatment would be unassessed, unchal-
lenged, and largely unrecognised, and would be out of proportion to the samp-
ling cost. The quantity finally tested is a small fraction of the tonnage from
which it has been reduced. Chances of error in a single sample exist. Care-
ful use of a sound method, coupled with the taking of an adequate number of
samples, keeps the overall error within tolerable working limits. It some-
times happens that the least dependable labour is assigned to the collection
and reduction of samples, and that an unnecessary amount of dull manual
work is called for in their crushing, grinding, and bulk reduction. It is true
that a good shiftsman will get more out of his section than a bad one. It is
equally true that a bad sampler will so distort the contents of the fraction of
the original sample he finally takes to the assayer that the whole purpose of
the work is weakened and the returns rendered deceptive. There is a
tendency to rely on returns without checking the validity of basic data from
which they are compiled. Mill return sheets can degenerate unless an
efficient routine is maintained by sample-runners who understand the import-
ance of their work.
In a Symposium on "Statistical Methods in the Chemical Industry" a
Paper l compares a variety of appliances and methods used in automatic
sampling. It defines sampling as " ... the operation of removing a part,
convenient in quantity for analysis, from a whole which is of much greater ...
in such a way that the proportion and distribution of the quality to be tested
are the same in both the sample and the whole". Sampling is a statistical
Mineral Processing-Sampling and Controls 635
the true assay grade in a normal way. To use the nomogram the liberation
mesh of the value, its assay grade when pure, the density of value and gangue,
and the maximum size of the material being sampled must be known.
Sampling and process control are intimately associated. Here the domin-
ant need is the most rapid possible approximate information on any essential
quality in the material at the point sampled, and the rate at which it is passing
that point. Accuracy of control and advance in permissible complexity
of the flow-sheet have today led to a high efficiency and smoothness of
operation. Research has now made possible continuous check on some
vitally important processing factors, and the use of automation helps to
maintain smooth overall operation.
Sampling Solids
Occasionally the mill engineer is asked to prepare a representative sample
of ore either for a trial run, for dispatch to a distant ore-testing station, or in
connection with the valuation of a property. This is one of the less satis-
factory assignments, particularly if precautions against surreptitious tamper-
ing must be added to the natural problem. A proper proportion of freshly
severed ore should be cut from each exposed rock face that can be safely
reached. If the face has been exposed long enough to become chemically
altered a new surface should be exposed by blasting. All the material, large
and small pieces alike, should be gathered. If the face represents one-tenth
of the presumed ore body, and a ten-ton sample is wanted, then a ton should
be taken from a fresh, clean surface. If the ore is being shipped, it may well
be sealed into clean, empty oil-drums, to minimise oxidation en route. If the
ore varies from exposure to exposure, it may be wise to keep the sample from
each such point separate, but usually the whole amount is mixed as well as
possible before dispatch. Picked hand specimens which exhibit the con-
stituent minerals can usefully be sent separately, to aid the distant consulting
engineer with the microscopic work and contact-angle investigation. A few
large crystals of the valuable mineral are most helpful for this purpose, if they
can be found.
If the large heap of ore gathered is to be sampled on the spot. it must be
reduced to a manageable quantity. The distribution, grain size and frequency.
and the assay grade of the valuable minerals influence the accuracy with which
size reduction can be performed. The most difficult type of problem is that
of a very small quantity of value segregated in a large volume of gangue,
such as that presented by metallic gold in quartz. Each pennyweight is
1/600,OOOth of a ton of ore by weight and 1/4,OOO,OOOth by volume. The
quartz is brittle and the gold malleable so that crushing which will reduce the
former to sand will only distort particles of the latter. The sample finally
treated by the assayer is from one to four assay-tons. If the gold existed as a
single particle weighing one pennyweight and the ton of rock were reduced in
bulk to one assay-ton (29'1667 g in the short-ton system) the odds against any
gold being found are over 30,000 to I, while if it were in the portion finally
tested the assay would indicate richness. The example chosen, though
Mineral Processing-Sampling and Controls 637
tonnage received. In the sample reduction room a small jaw crusher and a
disc grinding mill would handle the comminution. Reduction would prob-
ably use riffiing followed by either rolling on glazed fabric or the use of
miniature riffies. In addition to the sample sent to assay, a reference sample
might be kept, and a quantity accumulated toward a monthly composite
sample.
Where crushing machines are used for sampling, a pile of barren rock or
sand should be kept, and some should be run through to clean each machine
after a sample has been crushed. A little of the new sample should then be
run through and discarded before crushing the rest.
A description of advanced design in sample reduction has been given in
connexion with the iron ore mined at Kiruna in Sweden 4 • Here the ore
varies in its phosphorus content between 0·0 17 % and over 2 %, and control
of this element in the finished product is essential. About 1,000 samples
must be assayed daily. Following statistical studies it was decided to take a
weight of I kilogram as representative of a carload or of the ore from one
round of shots during mining. For some grades of ore five samples are
combined, so the plant can handle either I kg or 5 kg lots. Twenty process-
ing lines are in use, through one of which each sample gravitates from a
drum drier to a gyratory crusher and automatic splitter and thence to a
mortar-type of grinder and a sample box. The entering sample is poured
into a drum, its tag number dropping into a holder below to which the final
sample will come. The sequence is automatic and between samples the
drum is cleaned by a strong air blast; 90 % of the sample is rejected after
passing through the miniature gyratory crusher, and the rest of the work
takes some three minutes.
Mineral Processing-Sampling and Controls 639
Sampling Pulps
Process control depends partly on testing of samples cut from the passing
stream of pulp. The information obtainable from these samples is discussed
later. Provided the load of solids is not distorted by irregular surges of
tramp oversize, abrupt changes in solid-liquid ratio, or uneven rate of passage,
a pulp sample can be made with good accuracy. If any of the above condi-
tions are present, they must be dealt with in any case in order to restore
operating efficiency, and it would be a waste of time to test a sample repre-
sentative of such poor operation.
The best place to take the sample is at a launder overflow where the pulp
stream is falling quietly and with a shallow flattish cross-section. Such
points exist where classifier overflow launders deliver to pump sumps or
conditioners. They can be contrived (head room permitting) by interrupting
the run of a launder so that it delivers to a second section. The sample
cut can be made either by a periodic diversion of the whole pulp stream, or by
means of a cutter (a rectangular bottomless box with its length normal to the
flow). The cutter traverses the stream at a steady rate, delivering the material
it gathers to a sample bucket by means of a flexible hose. The cutter is
moved by a reversible switch. (Fig. 5, p. IS). Several types are made. The
sample buckets are changed periodically by the sample runner and can be kept
locked between collections if desired. In this connection, the author's exper-
ience is that if shiftsmen want to tamper with samples nothing can prevent it,
but that proper relationship between management and labour should remove
this temptation. When assay returns show that a shift has had a bad run,
there is probably a reason for the trouble which can best be discovered by
co-operative investigation. Hectoring only confuses the shiftsman and tempts
him to take covert measures to ensure that future samples shall be perfect,
regardless of their validity.
Sample cutters which take part of the stream all the time are dangerous,
since there is usually segregation of the solids in various parts of the cross-
section. The amount gathered at each cut, and the period between cuts,
must be decided in each specific case. Some operators favour the use of a
master-switch which sets all the automatic samplers going at once. The
general purpose of the sampling is to provide information which will heIp
with the preparation of a metallurgical balance sheet, which accounts for the
units of value fed into the mill and those leaving it as concentrate and tailings.
Specific information is also gained at each sampling point which checks the
state of the pulp and its fitness for transfer to the next stage of treatment.
This will control technical performance and give data for possible improve-
ment. The general assay figures will show, subject to a fairly consistent
amount of value temporarily locked up in the various treatment stages, what
amount of saleable product has been made for the period and, when con-
joined with the cost accountants' figures, what the cost of production has been.
640 Mineral Processing-Sampling and Controls
Several types of machine have been developed for the purpose of doing this
while the ore is in transit. One widely used appliance is the Merrick weight-
ometer (Fig. 270).5 Troughing idlers are, mounted on a table which hangs
from scale beams that register the weight of the passing load . This weight
is integrated with the speed of the conveyor belt by the disc A3, which is
642 Mineral Processing-Sampling and Controls
driven from the underside of the returning conveyor belt through A5 and
the crowned rollers B3. The disc shaft is carried in ball bearings L2 and is
free to tilt under the pull of the weighing beam, transferred through G3.
The travel of the endless belt J5 is in fixed ratio to that of the conveyor belt,
and it is pressed against the disc rollers by guide rollers (not shown). While
two rollers diametrically opposite are rotating the disc does not turn. This
is zero position for an empty belt. When a load on the belt causes the disc to
tilt, B3's are skewed across J5 and the disc rotates in proportion to that tilt.
This rotation, recorded on the counter W2, shows the weight conveyed.
Remote counters can be added, for purposes of distant control. To obviate
belt slip and inaccuracy there should be a gravity take-up on the return side.
The "Adequate" Conveyor weigher picks up the belt on a single troughing
idler, and can readily be installed on existing systems without structural
alteration. The underside of the weighing component provides a compen-
sating return idler when sticky material is handled. Any weightometer must
be regularly calibrated, and this can be done by passing special weights over
it or by anchoring a length of heavy chain so that it lies on the empty moving
belt. If the entering ore changes seasonally from dry to moist, a layer of
fines may adhere to the belt and pass over repeatedly, thus showing a delivery
in excess of what the mill really receives. Recorded weight must therefore
be corrected when necessary by making suitable allowance for variation in
the weight of the conveyor belt, due to the clinging to it of an adhering layer of
ore or to any other cause. This is done by adjusting the check nuts (Fig. 270).
Various methods of rate-of-feed recording are possible on the discharge
side of the ore bin. The rate of travel of a feeder apron, the weight of ore
per foot run, and a record of the running time give a rough check. Less
satisfactory methods are the counting of trucks delivered, or use of the
figure for ore sent from the mine. The Humboldt feeder is a duplex auto-
matic hopper, one hopper filling while the other delivers its charge into the
ball mill. Each delivery is automatically weighed and recorded.
The Hardinge constant-weight feeder (Fig. 271) automatically adjusts the
then 9
p.Sp S8 D+I
Sp=p. Sp+ I (21.2)
S8 p+Ss(1-p) D+ I
S8
Ss(Sp-l) I
p = Sp(S.~-I) = D+ I (21.3 )
p. Sp Sp
S.. = I-Sp(l-p) 1- D(Sp-l)
(21.4)
I-p = Ss-Sp 1-
D= (21.5)
p Ss(Sp-l) p
Sp-I p
(21.6)
q = Ss-I Ss-p(Ss-I)'
Classifier Returns
Hand samples are collected periodically, and a sizing analysis is made.
This checks the efficiency with which the oversize is being returned and the
undersize released from the closed circuit. It is also possible to make a check
on the quantity of circulating load.
Classifier Overjfow
This extremely important sample should be cut automatically. It is
assayed for head value, which should check roughly with that of the solid
head sample unless a concentrate is being withdrawn from the closed grinding
circuit. Sizing analysis is also made on this fraction, to check efficiency of
grinding. Since efficient concentration depends on optimum liberation this
checking-point is essential. In addition to this sample, which is taken for
laboratory use and under laboratory control, the shiftsman usually takes dip
samples regularly at the weir and tests them for density and pH value. His
running controls are used to adjust water and alkali, and so maintain the
correct mesh-of-grind and alkalinity of the pulp.
Pulp density can be roughly assessed by the use of a hydrometer, though
the solids may settle so rapidly as to upset the reading. The usual method is
to weigh a known volume of pulp. Balances are made which give a d:rect
reading of specific gravity.
When concentration of metal sulphides is practised in the closed grinding
circuit, it is usually in order to avoid retention and overgrinding of a heavy
mineral. The concentrate thus made joins the classifier overflow. In this
case, provided they mix ahead of the automatic sample cutter, no special
Mineral Processing-Sampling and Controls 645
Flotation Products
Concentrates
In a flow-line along which several successive products are removed,
samples of each concentrate, rougher froth, and scavenger froth are needed.
Pulp flowing through intermediate stages may also require sampling. The
mos important of these is each final concentrate, which is sampled periodically
during the shift. The" cuts" thus taken over a period are mixed and assayed
for grade of end-product. This assay may include total sulphur, the sought
metal, and any other metals. It must be sufficiently comprehensive to provide
a basis for valuation if the product is being sold. Usually a simple assay
for metal grade is sufficient, a little concentrate being retained from each shift
assay to form a composite sample which is mixed and given the necessary
comprehensive assay at monthly intervals. At that time, a portion may be
sized, each screen and subsieve size being assayed for grade. The informa-
tion given by these last assays as to tonnage and grade of each size-fraction is
used to assess the work being done in the grinding circuit and the degree of
liberation of each value.
The assays of the partially cleaned concentrates are only made periodically
on accumulated bulk samples, except when alterations in the cell circuits are
being planned. The important day-by-day value of these samples is their
sizing analysis and microscopic appearance. These give direct information
which can be used to maintain efficiency despite changes in ore mineralisation,
or to improve any weak detail of the treatment.
Tailings
These are assayed from each shift's cumulative sample for grade. The
pH may be checked as each cut is taken, to provide a cross-check on con-
ditioning. The density of the pulp at overflow from the cells may also be
checked by the shiftsman, who should log his figures here, as elsewhere in the
plant when running check is made. This density check is only needed when
over-dilution of the pulp might lower flotation efficiency. This could arise
where a heavy draft of concentrate was made, or where tails from a series of
cleaning sections diluted the new feed immoderately.
A suitable weight of each sample should be retained and mixed to form a
special weekly or monthly sample. This is fractionated by screening and
646 Mineral Processing-Sampling and Controls
Cyanidation Products
Samples of pulp are collected by cutting through the whole stream at an
appropriate overflow point. Solutions may be sampled continuously by
diversion of a steady drip or by periodic draw-off. Examination of the pulp
samples is mainly concerned with losses in the tailings. Gold can be lost by
incomplete washing of the discard, allowing pregnant cyanide to run to waste.
This is checked by washing filter-cake and examining the filtrate. Another
cause is coating of the gold by iron or aluminium oxide, preventing solution.
A third is incomplete liberation, particularly of the gold carried in auriferous
pyrite. A fourth is too short a time in the agitators. This is usually due to
the failure of the gravity and amalgamation sections to trap and hold coarse
metallics. If graphite is present in the ore, its effect on the operation must be
watched.
The testing of samples of cyanide solution is concerned with the oxygen
content, freedom from fouling salts, strength of the "available" cyanide,
adequacy of its protective alkali, and the efficiency of precipitation. The
quantities of pulp, pregnant cyanide, stock solution, etc., held up in the
plant must be known if the day's recovery is to be correctly assessed. This
involves recording of flow rates.
I>ense-~ediumm Fluids
These are usually regenerated continuously. Sampling methods may be
used to control any or all of the following:
Mineral Processing-Sampling and Controls 647
Mill Control
Routine sampling should give all the information needed for technical
control and economic appraisal. In the foregoing sampling list discussion
was limited to single-mineral concentration, but checking would have to
cover the needs of a more complex operation by the addition of any further
control points needed. A metallurgical balance sheet can look complicated,
but when all the information from assays has been collated, its basic form
reads:
change in the ore mineralisation not taken into account in the original test-
work, and the place where preliminary investigation could best be initiated
would be the mill laboratory. The monthly tailings sample would be a good
starting point, and analytical methods might be worked out which would
show whether the lost value was of the true mineral that should have been
recovered, or whether there had been some change in composition. The mill
laboratory would try to produce isolated samples of the suspected materials
for x-ray, spectrographic, o~ other specialised tests only possible in a central
laboratory. Special care is needed in these not uncommon circumstances,
to ensure that the probable nature of the change has been discovered, and that
the minerals suspected are the ones calling for tests beyond the resources of
the mill laboratory. Such clarification as can be performed on the site aid
greatly in the work of a distant specialising organisation and may make the
difference between success and failure.
-I
,/ Detector
Fig. 272.
reasonably constant and an allowance is made for particle size. In the Simcar
apparatus a radio-isotope (thulium 170) is used as the radio-active exciting
source. The sample, at minus 1 in., weighs 30 lb and is collected in a steel
box with a perspex window through which radiation is beamed. Moisture
can be controlled by flooding the box with water and up to thirty samples can
be analysed hourly, ash content varying between 2 % and 30 % with a maxi-
mum error of 2 %.
A more elaborate system for continuous control and sampling has been
put into profitable use by Dutch State Mines 15 • Samples of coal are cut at
half-minute intervals, dried and ground in a hammer mill. They then pass
through an x-ray beam and the reflected rays are analysed. The operation per
sample takes some three minutes and its findings can be used as part of the
automatic control in the washery.
The use of a radio-active source in connexion with coal was mentioned
above. A Paper 16 on the use of isotopes describes their experimental em-
ployment in the rapid analysis of lead and tungsten ores.
650 Mineral Processing-Sampling and Controls
Automatic Control
Witj;l the growing complexity of ore treatment, the importance of maintain-
ing optimum conditions at key process stages has led to an increased use of
automatic control. This is still in its infancy, but most mills have devices at
critically important points which either regulate a process or give automatic
warning when something is going wrong. Against this, it must be recognised
that many mills have also put expensive equipment of this kind on their
junk heap. The reasons are simple. While labour is untrained, a good auto-
matic device, fresh from the manufacturer, is invaluable. As the shiftsmen
gain experience and the automatic appliance becomes clogged by pulp and
sand, the reliability of the man comes to exceed that of the appliance. Thus
many excellent devices fall by the way, though several remain. It is therefore
important not only to have the right appliance but also to be able to keep it in
going order, and this is made easier if it is simple and reasonably robust.
If the whole chain of operations is divided into sections, and the minimum
number of control factors essential to the efficient working of eath section is
known, then planning for automatic control can usefully begin. No ore
should leave a section for the next in line until it is in the correct condition to
serve as feed to that section. If any suitable automatic control can be used
to check part of the condition at the point of transfer, it is worth considera-
tion. Again, if any of the control factors in a given section can be made auto-
matic, the effect on the men in charge of that section should be studied, to see
whether the lessening of their duties will increase efficiency. There are human
problems as well as technical ones in these matters.
The elements of an automatic control system include:
These operations are performed in the order given above. There may be a
time-lag network coupled with stage correction, to compensate feed-back or
to prevent over-correction. If, for instance, pH is being measured in a
conditioner fifteen minutes beyond the point where the addition of lime is
controlled and the rate of addition of lime is automatically adjusted with each
measurement, then there must be a time-lag between successive measurements
of at least fifteen minutes so that the changed pH of the pulp will be measured
at the next measurement.
Automatic devices can monitor a detail in a process and can either give
appropriate warning of variation, keep a record, or take steps to correct a
drift from normal working. In the last case, the correction can be complete
at each resetting, or can be partial. The latter is to be preferred as a restora-
tion to normal, if made abruptly, leads to over-correction, whereas step-by-
step partial a·djustment, with a suitable time-lag between steps, brings the
condition to normal more satisfactorily and quickly.
Mineral Processing-Sampling and Controls 651
smoothly. The digital control would only recognise two states plus or
minus, and would not work continuously to re-balance the flowline. The
operational-digital technique of computer control combines the two methods
of operational analog control and programmed digital control. The process
controller gives feedback control based on input signals, coordination and
adjusting signals.
Feeding and Blending
Control of the rate and quality of process feed is highly important to smooth
and efficient running. In addition to the weightometers described earlier
some newer machines use electronic sensing devices to improve accuracy.
In one blending arrangement material A is continuously weighed and sampled
as it passes from its feed belt to the blending belt. The weight is signalled to
the sampling device and to a memory unit. The latter sends a delayed signal
to the feeder which delivers material B on to the blending belt and feed rate
is adjusted in accordance with the sampling information 18 .
In cement practice one operator19 stock·piles four heaps of raw material
in accordance with the chemical content. This, involving quarry control and
complex mathematical resolution to determine the lowest-cost blend for a
specific kiln feed, is controlled by linear programming with a computer.
Six ingredients must be proportioned within close limits. In another in-
stance the automation of the delivery belt tripper has been described20 •
Kennecott's blending trippers traverse the system continuously, the auto-
matic discharge limits being adjusted by hand so that each type of ore is
routed to its correct storage. In the filling system servo-devices are used to
sense the delivery position and the height of ore. When a bin is running low
the tripper leaves its continuous traverse and deals with it. It also signals the
main control centre if a bin is overfilled, shut down or wrongly working,
sounding an alarm if there is trouble with choked discharge.
Comminution and classification. In the crushi.ng plant automatic control
can deal with such matters as sequential starting and stopping, temperature
rise in bearings and loss of oil pressure. Transfer points can be scanned
either by closed-circuit television or with a photo-electric cell. This nor-
mally receives a ray of light, which can be arranged so as to be cut off when
ore piles up or conversely to emit a signal when the interruption of a properly
loaded conveyor belt is stopped by its running empty. Such cells monitor a
number of plant operations elsewhere, checking turbidity in bearing oil or in
clarified solutions, and loading levels in ore bins, steady-head tanks etc.
Automatic control in wet grinding can use several types of signal. These
include changes in:
(a) Grinding noise level in the ball mill.
(b) Power draft to mill.
(c) Circulating load in mill-classifier system.
(d) Pump loading (power draft).
(e) Pulp density (mill discharge or classifier overflow).
(f) Temperature rise through ball mill.
(g) Feed water to mill or classifier pool.
Mineral Processing-Sampling and Controls 653
Gravity separation, In DMS the first step away from periodic hand-check
and manual adjustment was continuous recording of the density of the bath
media, This is now used in various systems to control changes in the use of
diluting water and the rate of return of cleaned ferro-silicon from the densi-
fier 27 , Control of the water gives an immediate response while trimming of
the densifier rakes to increase or decrease the circulating load of solids makes
its adjustment more slowly, In one system four zones in the separating bath
are monitored by density gauges. Another (Fig, 273) has pneumatic controls
worked by signals originating with the returning dense media, This is
monitored both for S.G, and for pressure, thus affording a control level for
height of bath, In a system developed for the concentration of diamonds 28
bath density is controlled by the differential pressures between two bubble
pipes, and the viscosity is continuously monitored,
Improved de-sliming of a finely ground pulp in a hydro-separator was
found by a magnetite producer to depend mainly on the hydraulic current
introduced by four jets~l, The volume of water thus admitted is controlled
by means of a magnetic coil which senses the interface between a heavier
654 Mineral Processing~Sampling and Controls
References
1. Cook, P. E. (1959). Stevens lnst. of Technology Symposium. N.J., Jan.
2. Cy, P. ( ). "Sampling Nomogram". Minerals et Metaux, Paris.
3. Pryor, E. J. (1963). Dictionary of Mineral Technology, Mining Publications.
4. Eriksson, S., and Sundstrom, Y. (1961). World Mining, Aug.
656 Mineral Processing-Sampling and Controls
References-continued
5. Hudson, W. G. (1954). Conveyors and Related Equipment, Wiley.
6. Anon. (1961). "Automatic Grinding Control System". Can. Min. Jnl., May
7. Williamson, J. E. (1960). S. AI. I.M.M., Feb.
8. Weiss, N. (1960). World Mining, June.
9. Taggart, A. F. (1945). Handbook of Mineral Dressing, Wiley.
10. Bogert, J. R. (1961). Mining World, March.
11. Wood, R. E. (1959). Min. Eng. (A.I.M.M.E.), 214.
12. Lucy, W., Fulmer, T. G., and Holderreed, F. L. (1963). I.M.P.e. (Cannes),
Pergamon.
13. Carson, R. (1963). Rhod. Ch. of Mines Jnl., Sept.
14. Anon. (1964). Simon Eng. Review.
15. Balkestein, J. G., and Baerts, J. W. R. (1962). Trans. S.M.E., 223, Dec.
16. Kalmakov, A A., Polkine, S. I., Khan, G. A, and Smimov, V. V. (1963).
I.M.P.C. (Cannes), Pergamon.
17. Drevdahl, E. R. (1961). World Mining, June.
IS. Electro-Weighers, Trade Publication (Birmingham) Ltd.
19. Nalle, P. B. and Weeks L. W. (1960). Min. Eng., A.I.M.M.E., Sept.
20. West, H. H. (1963). S.M.E., A.I.M.M.E., March.
21. Conroy, K. D., and Kachel, G. C. (1963). Can. Min. and Met. Bull., Aug.
22. Amber, G. H., and P. S. Anatomy of Automation, Prentice-Hall.
23. Williamson, J. E. (1960). S. AI., I.M.M., Feb.
24. Weiss, N. (1960). World Mining, June.
25. Roe, L. A (1961). Min. Eng. (A.I.M.M.E.), Feb.
26. Anon. (1961). Automatic Grinding Control System. Can. Min. In!. May
27. Oss, D. G., and Erickson, S. E. (1962). Trans. A.I.M.E., May.
2S. Nesbitt, A. C, and Weavind, R. G. (1960). I.M.P.e. (London), I.M.M.
(London).
29. Kelly, F. J., and Stevens, C. S. (1964). Can. Min. J., Jan.
30. Piedboeuf, C., and Suys, F. .(1963). I.M.P.e. (Cannes), Pergamon.
31. Lawver, J. E., and Barbarowicz, W. (1962). Froth Flotation, 50th Ann. Vol.,
A.I.M.E.
32. (1961). Amer. Soc. Mech. Eng. A.S.A.C-S5, N.Y.
33. Bredberg, J. H., and Carroll, J. T. (1963). Eng. and Min. J., 164, Oct.
34. Anon. (1963). "Exper. at the Parc Mine". Mining Magazine, Jan.
35. Carson, R. (1963). RhodeSian Chamber of Mines Jnl., Sept.
36. Runnolinna, U., and Heikkinen, M. (1962). Mining World, Aug.
37. Anon. (1963). "The Parc Mine Project". Mine and Quarry Eng., Jan.
3S. Blumkine, G. V., et al. (1963). I.M.P.e. (Cannes), Pergamon.
39. Lyfield, A. J. (1963). I.M.P.e. (Cannes), Pergamon.
40. Hegarty, A. (1960). Mining Magazine, Aug.
41. Anon. (1962). Japan's Most Modem Cement Plant, etc. Int. Systems Con-
trol Ltd., Wembley, April.
42. Mokkan, A. H. Jnl. S. AI. I.M.M.
CHAPTER 22
Introductory
The movement of ore from the stopes to the mill is not usually the concern
of mineral processing. but the manner in which this task is performed can affect
subsequent treatment. If the blasting is done badly. too much fine material
is produced in the wrong place. and in the wrong way. Leaking lubricants.
if they fall on ore. become collector and frothing agents. Long delay between
stope-face and ore bin initiates undesirable chemical changes. The tempo of
transport. together with fore-knowledge of the ftotative qualities of ore from
different stopes. is sometimes of direct importance in process control. At the
ore bins dry transport usually ends. and the solids are moved in water from
thence to final discharge. In the older mill designs. great importance was
attached to gravity fall. Today belt conveyors handle dry solids and pumps
move pulps smoothly. cheaply. and efficiently. and it is no longer important
to choose a hill-slope down which flow can proceed by gravity.
Dry ore can be moved through chutes. provided slope and cross-section are
adequate. and awkward turns are avoided. Clean solids slide readily on a
15°-25° slope faced with bright steel, and on 40°-45° wooden surfaces. For
most ores a working slope of between 45° and 55° is used. Too steep a slope
is bad. Power is used in elevation of the feed, and fast-sliding rock is hard to
control. At the delivery end some cheaply replaceable restraining device may
be used, to brake the descending ore and deliver it gently. Iron sheeting. a
curtain of old drill shanks, or heavy chain are suitable.
Bucket elevators are used for the elevation of dry or pulped ore. They
consist of an endless chain or belt (Fig. 274) carrying the buckets vertically
or at a steep angle. On reaching the top pulley they should turn at sufficient
speed to fling their contents out with a measure of centrifugal force. Feed is
directed into a gathering boot or to buckets rising from a boot. Wet ore can
be drained through perforations in the buckets. A rod can be so suspended
as to rap the discharging bucket and so help to shake out packing sludge.
658 Mineral Processing-Unit Processes and Machines
r
~~ ~\\ ~~
11ft ]11t 1
] [
]H[ ]p 1
]
&dt A
Fig. 274.
8 c
Four Types of Bucket Elevator (after W. G. Hudson)
o
Belt conveyors handle the bulk of the dry ore moved through the modern
plant. A system including a tripper is shown in Fig. 275. They include the
belt, the carrying and return idlers (B and E), the take-up (F), the tripper
and shuttle belt/s, if used, the drive (C), the belt cleaner (near D).
The standard "rubber" belt has a cotton foundation which must be strong
enough to withstand the driving pull and the loading strains. The "weight"
of duck in ounces, refers to a single thickness or "ply" 36" x 42". Plies are
bonded together with rubber, and the pull needed to separate two plies in a
strip one inch thick, is termed the "friction". It is between 12 lb. and 24 lb.
and chosen for required duty and flexibility. The plies must not separate
under the worst conditions (i.e. when rounding the end pulleys). This
carcass (plies) is protected above and below by rubber, of specified tensile
strength (800-4000 p.s.i.). Covers vary in thickness from Dr" to -1". For
severe service, a "breaker strip" of open mesh fabric may bind cover to carcass
and confine any tear arising in use. Cord belts have additional longitudinal
cords embedded in rubber. Steel wire cables are also available for use under
Mineral Processing-Unit Processes and Machines 659
660 Mineral Processing-Unit Processes and Machines
feed end, and the shock of falling rock is avoided partly by delivering between
idlers, rather than directly above them. Rubber-covered idlers are often
used at such points. Transfer is kept to a minimum, to avoid this wear, and
extra troughing idlers are used at the loading end.
Return idlers are flat and more widely spaced. If abrasive material is
allowed to build up on the surface, undue belt wear will result, as there is
appreciable cling due to tension, even though the belt is not loaded. This
applies even more to the surface of the troughing idlers, which support the
underside of the fully loaded and tensioned belt. One essential condition
assisted by the idlers is the "training" of the belt. Misaligned structure, build-up
on idlers, bad loading, wind pressure, unequal stretch of the belt, insufficiently
heavy take-up, may cause the belt to climb off centre for part of its run.
In old installations guide idlers, which bore on the edge of the belt, were
used to prevent serious trouble, but they are obsolete. Troughing idlers with
tilting blocks or self-aligning arrangements (training idlers) are used instead,
combined with improved design. All idlers should run freely, and be checked
regularly to ensure that they are doing their work properly, and not damaging
the belt (for example, by being seized, or by having sharp stones stuck to them
which would punish the belt each time they turned). The troughing pulleys
are spaced at 3'-5' intervals and the return idlers at about 10'.
By using wefted nylon it has become practicable to increase the troughing
angle of the idlers to 45°, thereby raising the load-carrying capacity some
25 % for a comparable width of belt. The more severe flexing stress along
the "hinge lines" is met by superior elasticity and the need for fewer plies
than with cotton duck. The deep trough reduces the major causes of wear-
abrasion at the loading point and slipping on upgrades where a sagging belt
rises over an idler. Better training and centring are also claimed. The belt
speed must be kept above 400 ft./min. in order to ensure a good trajectory as
the ore discharges from the belt. Idlers as steep as 53° have been used in one
coking plant, where as a result spillage has virtually disappeared.
The take-up tensioning device (Fig. 277) acts on the return side. It should
be kept at the lowest practicable level so as to minimise belt stress. Auto-
mated adjustment is possible.
The take-up adjusts the belt for stretch or shrinkage, prevents undue sag
between idlers, and prevents slip at the drive-pulley. Provision for 11%
take-up is wise. Screw take-ups at the tail end have some use in short runs.
Gravity operated arrangements (Fig. 277) which adjust the tension continu-
ously to varying conditions, are usual in mill installations, together with a
screw take-up. With a belt inclined at 13° or more, the return side provides
sufficient weight for most, if not all, the take-up.
The tripper is used to distribute the discharge over a required length of ore
bin or dump. Instead of discharging over a drive pulley at the head end,
the belt is led back over a pair of bend pulleys. The load falls to a diverting
chute (Fig. 275). The tripper may be fixed or movable. In the latter case,
the bend (or snub) pulleys are on a travelling frame movable by hand, by
power from the main system, or by independent motor. It can be auto-
matically self-reversed.
The drive is at, or immediately following, the head end, and is supplied by
662 Mineral Processing-Unit Processes and Machines
HEAD DRIVE 7
~----------~
VERTICAL MINIMUM
-"":';"1
I I
GRAVITY BELT
T.U. ',:'- TENSION
COUNTERWEIGHT
n)f / a RIGGING
_ HEAOORIVE7
o ~~00------~
HORIZONTA~ ~ 0
GRAVln MINIMUM
1:U BELT
• • TENSION
one or two motors (the latter for heavy-duty work). With dual drive, the
first motor is synchronous and the second slip-ring, with automatically timed
acceleration and distribution of torque. Electrical interlock is usual in modern
systems. The final unit must be well under way before the next in front can
be started, and so through to the head feed. At stopping, the head machine
stops first. If a magnetic head pulley is in use, the belt cleaner must be set
well back, since magnetic material carries on round for a few feet.
The effective driving power depends on the difference between carrying
tension (T1) and return side tension (T 2)'
Effective belt HP = (T1 - T 2)S
33,000
where S is the speed (ft./min.) and T is expressed in lb./foot. Transmission
of driving grip depends on the area wrapped round the driving pulley (the
angle of wrap) and the coefficient of friction between this pulley and the under-
side of the belt, Tension on the loaded side is produced by weight of belt,
load, inclination and mechanical resistance of the system.
Heavy-duty conveyors with high lifts and/or long flights are widely used.
In the "Cable Belt Conveyor" two wire ropes at the edges of the belting take
the driving tension. In the Horstermann conveyor, used in less heavily
stressed work, the loaded belt is carried on 4' square steel plates attached to a
driving chain. In America the rigid structure is increasingly replaced by a
system in which the idling pulleys are supported at 10' to 20' intervals on
ropes in a light steel framework. Another radical departure from conven-
tional design is the German "Serpentix" conveyor. This breaks away from
the limiting requirement of straight-line movement from feed to discharge and
can even turn spirally on a radius of less than 10' while climbing, or wriggle
snakewise through obstacles. The carrying belt is in short sections each form-
ing a trough normal to direction of travel, and vulcanised to underlying steel
pans which are bolted to a driving chain. Side discharge is used at the
Mineral Processing-Unit Processes and Machines 663
delivery end after which the troughs are righted, and thus are available for a
return load. Another German development is the sandwich conveyor.3 In
this, a weighted belt over-runs the ore being lifted at an abnormally high
angle.
The maximum recommended belt speed for granular material coarser than
i" is 800 ft./min. and for heavy, large or abrasive ore 400'. Light and fine
sands should not be carried faster than 250 ft./min., and if discharge ploughs
are in use the speed should be still lower. Troughing rolls may be spaced at
from 3t ft. centres upward and return idlers from 8 ft. The Tl/T2 tension
ratio for a plain steel pulley with an angle of contact of 180° should be 2·19
and when snubbed to 240° 2·85, the coefficient of friction being 0·25. With a
rubber-lagged driving pulley having a coefficient of 0·35 the figures become
3·0 and 4·33 respectively. With tandem drive and bare steel pulleys at 360°
total contact the ratio becomes 4·8 and at 500° contact angle 8·86, Lagged
tandem pulleys would increase the safe working ratios to 9 and 21·2 respec-
tively. Corresponding belt tension Tl would lie between 1·85 and 1·13.
Since the cost of a belt may equal that of the drive and framework, care in
its maintenance is essential. Choice of the right belt is important. Starting
and loading strains and stresses must be controlled. Tension should not be
in excess of that required to avoid slip and sag. The belt must not bear on
anything except its properly maintained drums and idlers. Joints must be
square. Good design and maintenance at loading and discharge points are
necessary. Belts can convey their load at any slope up to 15°, or if the ore is
dry and not slippery, up to 20° or exceptionally 25°. On the Rand a certain
amount of washing of the banket ore is performed by spraying water onto the
rising belts. This assists the men picking off waste rock by displaying the
true surface, but does not, of course, remove any gold-rich slime adhering to
the underside of such waste. When the belt delivers downhill, the system
must be suitably braked. When it elevates its load, provision is necessary
to prevent run-back in the event of power failure. The tensile stress set up by
a load too heavy for a single belt to handle can be reduced by using a series of
belts each transferring to the next. Since the chief wear is at loading points,
such transfer is kept to the minimum.
Several methods are used to minimise the shock of loading. In one com-
mon arrangement (Fig. 278) the fine fraction of the feed falls to the belt first
and cushions the larger pieces which arrive later. The feed must not fall
directly above a roller, unless this is well padded with soft rubber. A good
arrangement is to bring the feed to a short unsupported length between two
such rollers. Skirting-boards, used to prevent side spill, must not be so
arranged that a stone could lodge between them and the belt, an occurrence
which could cut and ruin the belt in a short time. The surface of the belt
should be systematically inspected and all cuts, cracks, and damage sufficient
to let water into the fabric should be dealt with expeditiously. Portable
vuIcanising units can repair long cuts quickly. If acidic water is allowed to
penetrate to the textile" or mildew begins, the belt life is shortened. On
some conveyors idlers are used to bear against the edges of the belt if it should
ride up on its troughing idlers. If these seize, the edge of the belt saws into
them, and its protective rubber coating is removed.
664 Mineral Processing-Unit Processes and Machines
If such idlers must be used, they must be kept well lubricated and turning.
Another cause of wear is the build-up of spilt ore on the return track, so that
the belt surface rubs against it. Ore can also get on the inside of the belt and
weaken it by pressing up as the loaded fabric goes over the troughing pulleys.
The usual working speed of a belt is of the order of 600 ft./min., and
this is only suitable if the ore is free from dust. Each single section of a belt
series must go straight from feed to discharge point, as bends in a normal belt
line are not possible. The installation must be such that failure at any point
will automatically stop delivery in front of that point. A choked discharge
chute at the end of a series of belts would lead to chaos if the preceding sections
did not stop at once. Interlock with devices being fed by the belts is also
important for the same reason. It should not be possible to shut down any
machine on the working line without arresting the feed at the same time, and
similarly, motor failure should lead to the automatic tripping of all preceding
machines.
Hydraulic Transport
This may include delivery of ore to the mill, handling of pulp through grind-
ing and treatment, and/or product delivery such as that to tailings ponds or
underground backfill points.
The pulp varies in corrosiveness, abrasiveness, and particle size of its solid
content, and the channels of flow must be able to deal with these changes.
At some points flushing water may be needed, and at others it must be with-
drawn. At points of severe abrasive wear duplicate channels can simplify the
task of maintenance. Buried piping should be accessible, if necessary by
manholes at convenient distances apart for the use of drain-clearing rods and
"go-devils". Pulp transport is the arterial and venous system feeding and
Mineral Processing-Unit Processes and Machines 665
relieving the organs of the plant, and if it fails the mill must stop. All the
pulp-the solid, the liquid, and the deceptively small quantity of sticky,
corrosive or oversize material which is most prone to settle and form the
nucleus of blockage in a pipe-must move continuously.
Pulp may be made to flow through open launders. The correct slope is
established by tests which disclose the settling characteristics of the material
concerned-pulp and launder surface. The latter may be wood, metal, or
rubber.
In plants of any size, the pulp is moved through piping, which should be
kept travelling in as straight a line as possible to prevent abrasio n at bends.
The use of oversize pipe is dangerous whenever slow motion might give the
solids a chance to settle and choke the pipe. A self-draining arrangement
with a straight up-slope followed by a straight down-slope is less liable to
give trouble than a straight discharge from a pump, unless the shiftsman can
be relied on always to flush water through before shutting down the pumping
line. Pipes carrying pulps containing lime may become scaled, and periodic
treatment is helpful. Pipes of small bore can be scraped through if a rope
can be passed, while larger ones can be worked through with rods or go-
devils. Pipes should be accessible and identified as to function, to aid the
work of the maintenance gang. Materials of construction include wood stave
(for large diameter piping carrying tailings), iron and steel (which may be
rubber lined for handling corrosive pulp), plastics, e.g. Saran, concrete and
asbestos cement. The convenience of access provided by laundering can be
achieved by cutting slots along the top of fairly flat runs of piping at defined
points. The launder section may be rectangular (usual), Y, or trough-
shaped. Solids move by flowing in suspension when small and near the
density of the pulp. Larger particles roll if equi-axial or slide if
tabular, and there is some jumping due to eddies in the pulp stream.
When pulp consistency changes, or rate of flow is low, solids may
settle on the bottom of the launder and form a layer, which only moves when
it has built up sufficiently. If the launder is too shallow for this "bursting
head" to form, it will choke and overflow. The slope required varies in
direct ratio to particle size, particle density, and percentage of solids. Fines are
best moved in a deep flow, and coarse particles in a shallow one, down which
they can slide. The water should more than cover the largest particle.
The principles which govern hydraulic transportation of solids involve a
multiplicity of physical characteristics and design factors. Settling mixtures, as
distinguished from non-settling, have been arbitrarily defined as those in which
settlement of the particles in a stationary pulp is faster than 0·002 to 0·005
ft./sec. 4 There are four modes of flow at a given velocity of slurry movement.
In increasing rate these are (a) a stationary bed with saltation above it;
(b) a sliding bed with suspension; (c) heterogeneous flow; and (d) pseudo-
viscous flow. Condition (a) exists when a deep settled bed has formed 5 but
where the velocity is slightly above that at which the channel would be blocked
by deposited material. In condition (b) the bed thins and begins to slide.
When the channel is a pipeline any partial blockage due to settlement
restricts the cross-sectional area of free flow, and thus increases the pressure
ahead of this settlement. The factors involved in pipeline design and
666 Mineral Processing-Unit Processes and Machines
installation 6 include:
(I) Solid-liquid ratio
(2) Average S.G. of pulp.
(3) S.G. of solid constituent/so
(4) Size analysis and particle shape.
(5) Viscosity of fluid.
Among other factors are materials of pipeline construction, abrasive or
corrosive effects of the pulp, pump characteristics, required location, distan-
ces, grading of line, steadiness of operation and maximum load. Here the
Durand-Condolios sizing c1assiffication 7 is helpful. These authors have
developed a relationship between pipeline diameter and S.G. for "Limit
Deposit Velocity (LDV)" definable as the absolute minimum rate of flow to be
allowed for in designing a hydraulic transporting installation. Their equa-
tion is
where D is the diameter of the pipeline, r1 the mass per unit volume of solids,
r that of water and K a constant depending on the concentration and sizing
of the solid particles. For a specified pulp density and mineral content
LDV = Kl vi D (22.2)
and for variation in S.G. of solids
LOV = K2 vS.G. Solids -I (22.3)
Design factors for a slurry pipeline must allow for change in frictional
resistance where the walls of the piping are expected to roughen with use.
For a long line, boosting stations may be needed, their distance apart depend-
ing on the maximum permissible pressure to be developed at each pump
discharge. A typical phosphate slurry moved at 600 tons hourly might use
7,000 gal./min. of water at 15 ft./sec. and 40% solids. ]n Florida nearly 15
million tons of phosphate rock are pumped annually8 and automatic control
of series pumping is practised. 9
The capacity for type (a) (mill head storage) depends on the hoisting cycle
and days worked underground, and on the nature of the mill treatment. The
more complex the flow-sheet and the more costly the plant used, the stronger
becomes the argument for continuous milling. Not only must twice the
plant be provided if the mill is only to be run half the day, but recovery is
erratic during start-up and shut-down. An extremely simple operation,
particularly one closely linked with the entry of raw ore, can easily be started
and stopped, but such work as differential flotation tied to stage grinding
should be run continuously once started. Dense-media plants usually run in
phase with the delivery of ore to the secondary crushing plant, though there
is an increasing tendency to work the crushing plant most of the twenty-four
hours. If the mine is shut over the week-end, the mill must either have fine-
ore storage to tide it over or be shut down for the same period. A plant
requiring frequent overhaul might be run with a weekly shut-down, but a
better arrangement would be to shut down one section at a time for main-
tenance, provided the tonnage treated justified the laying out of treatment in
parallel.
A further condition to be studied in arranging ore storage is the effect of
exposure. Unstable sulphides must be treated with minimum delay. Wet
ore cannot be left exposed to extreme cold, or it will freeze and be difficult to
move. Ore in transit represents a tie-up of working capital which cannot be
released until the concentration is finished and the product shipped. The
extra handling involved adds to the total cost.
Where development ore must be stockpiled until the mill has been con-
structed, there is little choice as to storage arrangements. Again, political
considerations may call for special accumulation. A notable example was
the formation of special dumps of open-cast coal in Britain during World War
II, when the interruption of normal transport by damage through enemy
action had to be provided against. Here, the added problem of spontaneous
combustion had also to be met, by making the dumps of such dimensions that
internal heat did not rise to danger point.
The angle of repose of broken ore varies between 37° and 45°, and the
sliding angle, once movement has commenced, from 30° to 45°. An extra
5° to \00 must be allowed for, to ensure smooth withdrawal. To some extent
this depends on the surface condition of the rock, a clean pile of close-ranging
sizes being easy to control, while one of random sizes with deteriorated sur-
faces and half-consolidated fines can be difficult to start moving and to
keep running smoothly. A good large-scale method of coping with seasonal
surges, due to shortage of underground labour at sowing and harvesting
times followed by plenty at mid-season (a common condition where "cheap
native labour" is recruited from tribal sources), is shown in Fig. 279. The
undersize needed to keep the plant running is screened off and treated, while
oversize is formed in a long dump, above a tunnel equipped with a belt
conveyor. Such ore, provided it is non-caking, slides easily and drawing rate
is regulated by a curtain of chains. The dump cal" be formed by tripping
from an overhead conveyor belt.
When mixed ore is blended, it can be reclaimed by an arrangement such as
that shown in Fig. 280. The dump is formed by adding successive thin
668 Mineral Processing-Unit Processes and Machines
Ground Ground
Chain Control lever for
holding ore at lifting chair'
angle of repose
TravellIng
Hopper
Conveyor to Plant
Fig. 279. Stockpiling of Coarse Ore
..
~~ i
IHH~
I;·' m
~
Mineral Processing-Unit Processes and Machines 671
Since the water in the mill is used for a multiplicity of jobs, it is essential to
conserve adequate supplies which must allow for unfavourable rainfalls.
Water is a transporting agent, a separating medium in gravity work, the main
classifying control reagent, and may be the determining factor in selecting a
method of treatment.
Pulp Feed
l
Pumping
Pumps are used in ore dressing plants for a variety of purposes, to meet
which special designs and modifications have been developed. The main
types and duties are:
Miscellaneous
Frenier (ribbon-impeller).
Air lifts.
Hydraulic injection.
Pulsometers.
pumps are used, and the practical maximum lift is about fifty feet. If more is
needed, a second pump is put in series with the first. The efficiency, in terms
of volume swept, is between 50 ~~ and 75 ~{, when the pump is working at its
full load, and drops when it is only partly loaded (drawing air). The pump
casing must be protected against wear by means of renewable liners. These.
and the impellers. are cast of abrasion-resisting alloys, or are faced with
Mineral Processing~Unit Processes and Machines 677
rubber, the latter being preferred for finer sands and the former for coarser
grit and gravel. The chief costs are for power and replacement of worn
parts. There are two main modifications of design~glandless pumps, and
those with a water-seal. The purpose, in each case, is to protect the shaft
on which the impeller is mounted from being damaged by abrasive material
from the pulp. This is done in the latter case by the introduction of pressure
water to a lantern ring, so that it maintains a small positive flow through the
stuffing box into the casing. Pumps of the self-protecting type are more
widely used, since they are simpler to install and less liable to shaft damage
such as would occur in the event of gland-water failure.
The glandless pump has for some purposes the advantage of not intro-
ducing water into the pulp.
A second differentiation is between pumps which are designed to lift by
suction to the impeller and those which are meant to work on feed coming in
under slight pressure. In theory all pumps can lift the feed a short distance,
but in practice this is not the case. The Wilfley pump should be installed to
draw down from a surge tank so that there is at least a 4' head. The Vacseal,
if given a priming arrangement, can lift a moderate distance by suction.
The pumping system must be in balance with the load handled by the plant.
It must be able to deal with the maximum overload liable to enter the flow-
line and yet be able to work at reduced capacity when for any reason the flow
is restricted. Several methods of achieving this are possible. One scheme
for a small operation employing only one pump is shown in Fig. 282. The
ball valve controls recirculation of enough pulp to maintain a reserve and an
inlet head. This arrangement is wasteful of power, and sets up undue wear.
Where tonnage justifies it, the more elaborate system in Fig. 285 is good. A
two-pump scheme for ensuring completely steady delivery to a classifier or a
cyclone is shown in Fig. 286. Coupling units are marketed,19.20 which adjust
the pump speed to maintain a balance between feed and discharge rate when
the inflow varies. The present range of "]npower" couplings lies between
121 h.p. and 400 h.p. When the running speed of the pump is varied without
other change, the capacity is directly in proportion, the delivery head pro-
portional to the speed squared, and the power consumption to the speed
cubed. Within the limits of variation of impeller diameter (usually an effect
of wear rather than of design) the capacity varies directly as the diameter, the
head as its square, and power used as its cube. If a pump runs badly, the
impeller may be too much to one or the other side of the casing.
Periodically the seal must be renewed. After placing the seal, the packing
should be left loose and tightened down gently when the pump is working, so
that it beds down gradually. If properly packed, the shaft will turn easily
by hand. The maximum desirable rate of flow is 8 ft./sec. on the inlet side
and 12 ft.!sec. on the discharge side of a centrifugal pump. Sharp bends
are to be avoided, and change of section of piping should be made gently
through tapering reducing joints. The piping should be self-draining on
shut-down, and needs a reasonable velocity offlow through it to ensure against
blockage. A rate of 5 ft./sec. is a good minimum for ordinary pulp, or 4
ft./sec. for fine material in reasonably dense pulp. The power consumption
is proportional to the density of the pulp.
678 Mineral Processing-Unit Processes and Machines
LIOUID L!VEL
RECEIVER INOICATOR
SPEEO MOlO"
some 6 h.p. per ton-mile, sand averaging 0·4 mm. about 2·2 h.p. and raw
cement slurry at 40 - 50p. size only 0·12 h. p.IS Vertical lift is simpler than
horizontal transport where there may be settlement on route, in which case
the velocity must be twice the settling rate of the largest particles in free fall.
Transport interruption arises either from jamming or instability of flow.
The conduit must be at least three times the diameter of the largest particles
and where there is a pump in the flow line the maximum particle diameter
Dirty water
--
pumps
Electrically controlled
pneumatically operated
plug valves
ReI ief valve
loc k
hoppers
Diaphragm
pump
Slurry _
hopper
mJ Slurry
c:J Clear water
mZ:l Dirty wat r Bypass for
flushing Slurry delivery plpelln
pipeline
must not exceed one half of the smallest pumping cross section. If the
delivery grade rises at all steeply means of emptying the load must be pro-
vided, operable swiftly in the case of a sudden shut-down.
The main maintenance costs arise from pump and pipeline wear and
vary from below 0·03d. per ton-mile for slurries to over 3d. for gravels.
The diaphragm pump is used with most thickeners to remove the underflow.
It consists of a flexible diaphragm with a flap valve, which is reciprocated
vertically by an eccentric. Pulp is drawn up into the space swept by the
680 Mineral Processing-Unit Processes and Machines
diaphragm through a lower flap valve and overflows on the pressure stroke.
The throw of the eccentric is usually adjustable without stopping the pump, so
that the rate of withdrawal of settled slurry can be regulated. Since this
slurry is of high density, the piping from the thickener is of small size, and the
pulsing action set up by the pumping action helps to keep the material in
suspension. In one arrangement the density of the thickener underflow is
monitored by a gamma ray gauge. The resulting signals actuate an automatic
control on the diaphragm pump which varies the stroke amplitude. This
controls the rate of slurry withdrawal at the desired density.
Air-lifts are handy and easily improvised. They can be used to transfer
pulp from section to section of a flotation bank where the cells do not provide
suction, and for such jobs as collecting spillage and returning it to circuit.
An air-lift has the advantage, sometimes useful in cyanidation or in oxygen-
ating flotation water, of re-aerating the pulp. The inlet of the piping must be
at least one-third of the elevated distance below the pulp.
The Mono pump consists essentially of a rubber stator shaped as a double
internal helix, inside which a single helical rotor runs slowly with a slight
eccentric motion. It is self priming and delivers to a height of 100' or more.
In the Lock Hopper pump (Fig. 287) clean water at a high pressure is used
to evacuate slurry from a filled hopper into the delivery system. Operation is
continuous, one of two hoppers in parallel being filled while the other empties,
the control valves being automatically worked. The Westfield installation
in the figure delivers 28 tons of dry solids hourly some 3,600 feet at an 80'
lift, with a flow speed of 8·2 ft./sec.
Process Water
Among the factors which affect the use of water in ore treatment are:
(a) Availability of fresh supplies.
(b) Seasonal variation in quantity and quality.
(c) Disposal of radio-active, poisonous, tainted or acidic effluents.
(d) Softening or de-salting before use.
(e) Storage, conservation and re-circulation.
(f) Effect of build-up of dissolved salts when re-circulating.
Fatty acids are particularly sensitive to calcium ions, and water used for
flotation may require pre-treatment to reduce hardness. Where rainfall is
seasonal the concentration of such dissolved salts varies according to whether
run-off or deep spring water is being drawn to the plant. Aeration affects
both the effect of oxygen on the chemical action of flotation reagents and the
stability of the froth, so water drawn from beneath a wintry ice cover or
reclaimed from tailings dams in which algae have stripped the oxygen may
need some re-aeration before use.
When newly arriving water carries deleterious ions the re-circulated water
in a flotation circuit may hav~ been improved during treatment by loss of ions
to the solid products. Again, it mav have built up fouling ions to a dangerous
level by dissolution from the new ore. It is sometimes found advisable to
Mineral Processing~Unit Processes and Machines 681
Washery Water
veri sed fuel. Such a fuel can be used even with an ash content exceeding
40 %, provided the flues of the furnaces are able to deal with the resulting
fly-ash. In typical de-dusting systems (Fig. 288) the raw coal drops through
j~
:al
: ~ '. :'
"::':",:, ,
~"" ". ,.,:-.. ..,
F===r===~' ; :; ~ .
'----.llr---.l • • .
I :.~ l
I·' .
.J '. ..,Jo1
Adjusfing"':"!"···.. I
doors I I
8/rlley
deduster
....--I
!I
"'.
i
e
II:
ii
I eia
I:I!. . 1
;~III:
.....::s
1 ~
'"'
.!:::
U
...
....III
~
\
!:
-
Z
0:1 ~
too",
... Z .", ~
>0.
, t
-....
~% ~
... '"
CZ ...
III:lC
sJ ~; iat ~
f ... 0\
~~ 00
I
N
~i ~
:I
>-
III:
III:
~
,l~
§
1
allow solids to be handled, while still keeping th!! material sufficiently soft to
be worked. Flocculation is usually necessary to hasten the settlement of clay.
Starch, lime, and various proprietary compounds based on high-valency salts
are used. The three stages are:
(I) Addition of flocculating reagents.
(2) Mixing till these are dispersed.
(3) Gentle movement to help collision of dispersed particles without
breaking up formed floccules.
684 Mineral Processing~Unit Processes and Machines
persed. Zinc chloride and calcium chloride are among the proprietary addi-
tives.
Two main techniques of clarification are possible. In one the whole of the
water in the thickener is dosed. This gives a good settlement and a coarse
slurry, but much underflow must be removed and filtered. In the other a
moderately thickened natural sediment is underflowed and treated. The
thickener underflow is filtered, while clear water is drawn from the top of the
tank to the cleaning plant. An alternative to vacuum filtration is the running
of the slurry to a rectangular concrete tank, with a sand layer in the bottom
communicating to rows of well-points (draining pumping pipes). One tank
is filled while others are drained down and dug out.
Flocculation
Filtration
fine that it can pass the porous barrier is part of the filtrate. The pressure
drop may be due to simple gravity, to pressure on the pulp being fed, to a
vacuum on the filtrate side, or to the effect of centrifugal force. Filtration is
used to remove pregnant cyanide after leaching, and to clarify cloudy solutions.
If the viscosity in a pore is insufficient to neutralise the pressure differential
between its intake and discharge, a liquid can flow through it. Poiseuille's
formula for dynamic viscosity, v, as determined by use of the capillary tube, is
v = 71pr 4 t/8lV
in poises, where p is the pressure difference, r the radius of the tube, I its length,
V the volume of liquid delivered and t the time. Experimental verification
has been made of its implications, but the practical difficulty in evaluating
filter performance is that the porosity decreases rapidly as mineral particles
arrive and modify the overall structure of the system-septum plus arrested
layer. A further decrease in porosity is due to precipitation of calcium salts.
As the particles are arrested, the larger ones bridge the orifices. Smaller
particles next lodge in the interstices, so that after a very short run the filtrate
no longer contains material such as passed through initially. The effect of
this increase in I and the decrease in r is to decrease the dynamic viscosity
considerably. In practice, the rate of filtration and the ability of the barrier
to arrest solids depends on several main factors:
I. Available filtering area (direct).
2. Pressure difference across cake and membrane (direct).
3. Thickness of cake (inverse).
4. Pulp temperature (direct).
5. Specific surface of solids (inverse).
6. Size-range of particles (inverse).
7. Degree of flocculation (direct).
In gravity filtration the main function of the septum is to act as a support for
the bed of material. In one sense, leaching tanks used in cyanidation are
filters. Their thick beds are made possible by exclusion of most of the fine
sand which would clog the interstices between coarse particles and prevent
drainage. When the object is clarification, a thinner sand bed is used, the
fine material being allowed to obstruct the interstices until rate of filtration
and degree of solid arrest are adequately adjusted. Control can be maintained
by such means as a skimming cut to remove some of the uppermost sand.
Thickeners can be adapted for this purpose.
Pressure filters are robust vessels which contain porous filtering sections.
Feed is charged in and air pressure is turned on. In the developed form,
plate-and-frame filters, they are intermittent in action. The septum is stret-
ched over a frame which has channels for pulp feed, wash liquor, and filtrate
discharge. A number of these frames are assembled in a press, and' pulp is
pumped in. When pressure reaches a determined intensity, delivery of pulp
through its force pump is stopped. The solid cake now filling each frame is
washed by liquid under pressure, which can be discharged separately. The
press is next opened and the cake is removed. These devices are much used
in the chemical industry, but have declined in importance in milling.
Mineral Processing-Unit Processes and Machines 687
Vacuum filters are of two types, intermittent and continuous. The most
common intermittent type found in mineral dressing is used in clarifying
pregnant cyanide or in catching precipitated gold slimes. A number of
frames, each covered by a canvas envelope, are suspended in the tank which
receives the feed. Piping incorporated in the frame connects it with a com-
mon "leader" served by a vacuum pump.
Intermediate between this and the truly continuous filter is the "Genter".
71,1~·(~me .. __~'-~"""---i"'.
f/f~
Aulr:JmDllC
Cllr wrlvt \.
'1'mt con".cli1'lt} chmp ~
\, r,ube.(rpmtmanlfokl ~
F1IIf'Clfp· ;
!flttiqrlQrs •
~-~~
~ ~+.::~ I
H..t==~ t~======::::::(~~~ - 1r
- --- ----
4"HUdd'$cha"';lrne'
~:
~
~'
VACWU FltUA.c ~
~
§
~'
A [ttlV£I' I
'tAIl. ",PC
~
:::;.'
rlLT'AAT[
DISOWIG[ '0 ~
'il
OR. V~ UUU p ~ ~~
,." .. r_., ~
1:>
CON",,, ·l."'" tt..c""" 'ON
:::
...... 11( YI.It~ T t:>..
.II"~ ."
c..-lA' .... ...
............... tiM!!
II01'ru.1..
.", OVU n._TI'.-u.'
~ ~JII' "(NIIII(II
l
::;'
~
The sludge which has collected on the outside of each sock now slides down
to the bottom of the tank and is raked to a centre discharge. Low-pressure
air can be used to aid the displacement of the mud. This mud is intermediate
in consistency between the flowing slurry of a thickener discharge and the
solid cake of a drum filter.
Continuous vacuum filtration is today the standard practice in most mills.
The main types of machine are the drum and the disc filter. These have
several variants designed to meet special needs. Development research is
also active in the creation of new types. The drum filter is normally fed with
slurry underflowing from a thickener. If a dilute pulp, or an acidic one in a
highly dispersed state (such as that from an acid uranium leach), is fed direct
to the filter its solids are too mobile, and clog the system before a substantial
cake has formed. A similar feed if first flocculated and thickened produces a
wetter cake, but the rate of filtration is increased a hundredfold.
The general layout of a single-filtrate filtering system is shown in Fig. 292.
The cake detached by the scraper knife usually falls to a conveyor belt (not
shown). The cross-section of a drum filter is shown in Fig. 293. The machine
consists of the drum, of metal construction. On its circumference are the
filter plates, of wood or metal. These are grooved to permit flow of filtrate
drawn through the cloth. a membrane wrapped round the drum and bound on
by wire, so as to rest on the plates. The drum rotates at between two and
twelve minutes per revolution, in accordance with the length of gathering
time of cake required. This gathering is performed while the surface of the
drum is immersed in pulp in the trough. The trough, of metal structure, re-
ceives pulp and keeps it from settling by means of the reciprocating arms of
the agitator. Height of slurry is controlled between maximum and minimum
level by manual or automatic methods. Vacuum during submergence is
produced by connexion between the plate and the vacuum pump through a
valve head and distributing ports. Each port serves its own plate. As a
plate leaves the pick-up zone (Fig. 293) it passes progressively through the
drainage, washing, drying, and "blow-back" zones, finally reaching the
scraper-knife, where the cake and cloth, under the influence of compressed air,
bellies outward.
This assists the removal of the cake. The valve-ports can be con-
nected in various ways through the valve head in accordance with the
desired series of operations. Vacuum, first filtrate, second filtrate (if required),
and blow-back air are controlled by these settings. The scraper knife bears
lightly on the spirally wound wire which holds the cloth on the drum, and
aids the work of detachment mainly performed by the compressed air, which is
either blown steadily or pulsed so as to cause the cloth to flutter. The grooves
in the filter plates should be diagonal, to facilitate drainage of filtrate and
minimise the possibility of a wet blow-back. The scraper is usually edged
with rubber or belting to save wear on the cloth. In the Feinc system a
string discharge is used, cake being discharged at the sharp bend. To
minimise the formation of cracks which would lower the vacuum, Sullivan
Consolidated Mines has fast-rotating rubber strips beating on the emerging
cake. Deutsch Barit uses a perforated steel cylinder to bear against the cake
and squeeze out some residual moisture. Vacuum is normally applied from
690 Mineral Processing-Unit Processes and Machines
the entry of the segment into the pulp to a point on the rising side where
sufficient extraction of residual filtrate has been performed. The later this
is left once the drum has emerged, the more vacuum capacity must be pro-
vided.
Variations on the standard drum include a downward converging two-drum
WASH ."""YB
system for fast-settling material which could not conveniently be held in sus-
pension in the tank, and one in which feed first enters a small tank which
distributes its underflow on to the top of the drum (Fig. 294).
In another variation the pulp floods down the rising side of the drum, thus
sealing cracks in the cake. Where pregnant solution must be displaced before
discarding the filter cake (e.g. in the cyanide process), the vacuum must be
strongly maintained in most of the arc through which the cake moves after
emerging from the slurry tank, so that abundant wash water can be used to
displace and recover solution which would otherwise be lost. Blow-back
must be set well clear of this washing. Cracks in the cake would endanger
the efficiency of washing, and might overload the vacuum system. Regular
attention to the vacuum pumps should include check on leakage of air past the
valve seatings and correct tensioning of inlet valvesprings. Leaks in the
Mineral Processing-Unit Processes and Machines 691
Fig. 294. Section through Endflow Top Feed Filter (International Combustion)
Drying
Minerals nearly always carry inherent or acquired moisture. This IS
undesirable at times for several reasons:
~a) It increases transport cost.
(b) Water may be a nuisance or hazard in ensuing treatment of a mineral
product.
(c) Water introduces or increases humidity, corrosion and/or uncon-
trolled chemical action.
(d) Heat of reaction due to water may lead to a fire risk.
(e) Water may lead to difficulties in handling ore-e.g. frozen stockpiles
or unstable cargoes.
Against these possibilities is the probability that dusting loss in transit is
reduced when finely ground concentrates are exposed to wind en route or in
exposed stockpiles.
Moisture may be deeply held in cavities or cracks, carried as a surface film,
adsorbed or held by capillary attraction, or trapped by poor drainage in a
badly designed bunker. As distinct from good drainage, drying is definable as
removal of liquid (usually water) by evaporation. This may be effected by
exposure to sun and air or accelerated by heating. In the latter case convection
of hot air over and through the body of moist material is necessary, together
with removal of the resultant humid air. Alternatively, radiant heat may be
used. For efficiency there must be adequate contact between the drying
agency and the wet surface. Air may be pre-warmed or may gather its heat
from flues or plates over which the ore is passing. In rotary kilns a long
cylindrical shell slopes gently downward from feed to discharge. Hot gases
ascend counter to the stream of ore being tumbled as the shell slowly rotates.
This arrangement gives exposure of the mineral, while the upflow of hot air
from its entry, dry, to its moisture-laden discharge can be controlled by gates
or fans. A similar arrangement is used in roasting or in magnetic roasting, a
hot reducing gas being used in the latter case, to promote the chemical
reaction.
In the Lowden drier reciprocating rakes move the ore over heated plates
which constitute the roof of a furnace. Stationary kilns of various types
receive the wet feed at one end (above or horizontally) and either let it gravitate
or be rabbled to a discharge gate, moving counter to hot gases regulated to
produce the desired drying or pyro-chemical effect. Roasting action was
briefly described in Chapter 16 in connexion with the FluoSolids and Edwards
driers, used in recovery of gold from auriferous sulphides. Further discussion
Mineral Processing-Unit Processes and Machines 695
is presented below. One horizontal stationary drier (the Hazemag) uses screw
paddle action to work up to 100 tons/hour of asbestos through a drying pro-
cess. The machine is compact. The Raymond flash drier (Fig. 297) is
used for coal, lime or other finely divided material. The wet feed is mixed
with dry material in the mixer B and circulated in a turbulent stream of
hot air round the drying circuit.
fT
H- E.houst gas fon
F· Cyclon
G - Dry d l Idl'r
A-Wl'l
8-
Roasting
Sulphidic concentrates are frequently roasted either to produce SOz as a
stage in sulphuric acid manufacture or in connexion with further treatment
(e.g. copper sulphate leaching). Provided a sufficient percentage of sulphur
is present the process is exothermic, the driving force being oxidation of the
contained sulphur to SOz with, at lower temperatures, some S03. Control
may be used to avoid undue dilution of the sulphur-bearing exit gases. One of
the oldest arrangements, the multiple-hearth roaster, has been developed into
the flash roaster in which finely ground dry material is blown downward
through rising hot air. In fluidised beds the feed is introduced into a chamber
in which hot and partly roasted solids are teetering in a rising column of air.
When this rising air balances the weight of the bed there is generalised
stability. If the air current is increased beyond this expansion, mixing and
agitation increase as in elutriation. The interplay between the size and den-
696 Mineral Processing-Unit Processes and Machines
sity of the mass of particles, the size range of the load, together with the
velocity and reaction characteristics of the gas can be held in control by
choice of a suitable shape of container and by regulation of reaction rate,
temperature and gas velocity. Schytil82 has described two Dorr systems,
the BASF and the INCO. Cooling control in the fluidised roaster is achieved
in various ways. Water may be directly injected; the solid-liquid ratio of the
newly-entering pulp may be varied; waste-heat boiler elements can be built
into the bed; cold-roasting gas can be partly de-oxygenated and re-cycled or
cold re-cycled material from previous roasting can be added with the feed.
In the Dorr FluoSolids plant used typically to treat flotation slurries at minus
150 mesh low gas velocities are used. Where coarser grains are to be roasted
cooling is controlled by cold re-cycled gas injected into the upper part of the
bed. Schytil describes research on temperature effects in which a dispersion
of pyrite in sand was used. The threshold roasting temperature is about
400 0 C., and results in copious sulphatisation. As a grain reacts the reaction
product migrates into its substrate and this diffusion governs the situation:
Pyrite converting to Fe208 is reduced 37 %in volume and pores are formed.
When sulphate is formed, however, the mol volume increases by 160% so
that the pores are closed and reaction stops prematurely owing to the for-
mation of a non-porous sulphate skin. The rate-determining step is control-
led by the speed of this diffusion. At 5000 C. retardation ceases and roasting
proceeds with the availability of oxygen. At still higher temperatures reaction
stops before complete oxidation, due to distillation of sulphur from the pyrite
particles under the influence of heat at a rate exceeding the availability of
oxygen for reaction.
Tait3S has described research which led to improved recovery of a refrac-
tory gold ore carrying pyrite, arsenopyrite, stibnite, sphalerite, pyrrhotite and
galena. In order to produce a good porous structure for subsequent cyani-
dation the arsenic eliminating temperature requires strict control of available
oxygen in roasting. After this, further roasting is less critical as regards
temperature save that with stibnite present the extraction is badly interfered
with unless the control continues. At 4800 C. gold extraction from the roasted
calcine exceeds 90 % even when 2 % of antimony is present. During oxi-
dation the temperature must not rise above this point until the charge has
been visibly dead-roasted. Arsenic must be eliminated at minus 496 0 C.
before this further oxidation at a somewhat lower temperature, starvation of
oxygen being used in the first stage. These research findings were successfully
transferred to plant scale in a two-stage fluid roaster.
Magnetic roasting of iron ores in a reducing atmosphere (Fe20a to FC;a04
to FeO to Fe, or intermediately) was referred to in Chapter 19.
Tailings Disposal
At completion of treatment the tailings may be in the form of filter cake or.
more usually, a pulp. A filter eake could be handled by any convenient
form of solid transport, but a pulp must be gravitated or pumped to the
impounding dam. One general pulp-handling scheme, used with various
Mineral Processing-Unit Processes and Machines 697
dam traffic. Feed lines from mill to dam, if gravitational, should be graded
gently so as to minimise abrasion. Slope adjustment can be made by the use
of drop boxes.
In one method of build-up tailings at around 46 % solids are delivered to
cyclones slanted inward 15° along the berm. The coarser and thicker under-
flow falls nearest the inside of the wall and the overflowing slimes are laundered
well into the impounding area. When the berm is sufficiently consolidated a
dragline picks up this underflow material. Each dam is rested for from 4 to
10 weeks between fillings and after rising from 30' to 35' a new dragline crest
road is built. The use of mobile cyclones has been described,36 a telescopic
arrangement of the feedpipe allowing up to 37' of travel without having to shut
off delivery.
The stability of slimes dams has been discussed by Donaldson,37 who lists
three causes of failure:
(a) Surface erosion by rain and wind.
(b) Seepage through the toe end.
(c) Shear failure due to low strength of incorporated slimes or of
foundation soil.
Vegetative cover helps to prevent erosion and is easy to establish on flat sur-
faces. On berms Chenik 38 suggests planting suitable grasses into holes filled
with soil, or the use of 30" lengths of reed inserted 16" into the sand. Grass
must be given a start by the use of a fertiliser and perhaps helped until a natu-
ral humus has been formed. Creeper grasses can be used initially, giving way
to tufted ones when a growing compost has formed. Marram and couch
grass have been successful, and have also combated the nuisance of dust blown
from the dam.
As an alternative to impoundment, mill tailings may be returned under-
ground for hydraulic stowing in exhausted stopes or as backfill. The material
must be sufficiently deslimed (e.g. by removal of minus 15/L particles with
cycloning) to ensure that it will drain and consolidate stably in situ.
Soil mechanics has been defined as "that discipline of engineering sciences
which studies theoretically and practically soils, by means of which and upon
which engineers build their structures".39 These include the foundations of
heavy dynamic loads such as reciprocating crushers as well as the silts and
slimes liable to enter the walls of tailings dams. The moisture content of a
soil is related to the surface available for wetting and specific surface is pro-
bably more important than solid-liquid ratio when considering the effect of
this moisture. The colloidal fraction of the pulp sent to a settling pond need
not be great. If it segregates and finds its way through crevices it can influence
the creation of a slipping plane by virtue of its concentration locally rather
than its size percentage. Tailings may carry chemicals which are probably
flocculative in tendency and therefore strengthening. If the tendency is
reversed-for instance, through subsequent oxidation of contained sulphides
---entrained slimes would be a source of weakness. Compressive action as
the dam rose could produce a sh~ar zone, and streaming potentials might
aid flow of water and further liquidation. If the retaining walls are reasonably
coarse there ShOlild be no risk. It is, however, well to choose a site where
Mineral Processing-Unit Processes and Machines 699
failure would not endanger life or property. Streaming potentials are more
important where the soil structure must carry heavy engineering loads.
Thixotropic clays and thermo-osmotic migration of water in the subsoil,
particularly of varved clays and silts subject to permafrost conditions, call for
experienced judgement.
Maintenance
This section can only touch briefly on the routine plant activity required to
ensure smoothness and continuity of operation, without which good effciency
can neither be reached nor maintained. Among the elements of a well-
conceived routine maintenance schedule are:
to entry of water or chemicals, and the surface physics of the metals in con-
tact. It takes three basic forms. True lubrication requires the interposition
of a laminar, viscous or fluid film between the moving surfaces which is thick
enough to prevent high spots on the two surfaces from making contact. Part-
ing lubrication is that in which this layer is held under sufficient pressure to
fend off the moving surfaces from each other. Special lubricants are those in
which the oil adsorbs to one or both surfaces and maintains a wetting film
which neither ruptures under load nor deteriorates at working temperatures.
Wrong loading conditions in heavily stressed components lead to metal~to
metal contact, shown by scuffed, ridged, galled or scored gears. Incipient
welding occurs if the lubricant fails to maintain a barrier and to remove the
intense local heat generated during sliding contact. Pitting is due to surface
fatigue of the affected metal and is not appreciably ameliorated by lubrication.
Abrasive wear, due to the entrainment of mineral dust, shows as a smooth
matte surface. Here a more viscous oil may help things, but the best remedy
is to suppress the dust at source.
Filed records are helpful in building up information as to the probable
working life of parts of a machine not readily seen without shutting down.
If it is known that a submerged impeller has an average life of so many hours
inspection can await approach to that limit. Clean-up of odd corners often
improves metallurgical recovery by removing depressive wood pulp from the
corners of such places as flotation cells. Safety check should be made as
though one is a stranger to the plant, since habit leads the workers to ignore
danger points. Spot checks can be made outside routine checking, and deal
closely with quality control. Quiet deterioration in performance is apt to go
unregarded and an examination in which localised laboratory sampling aids
inspection is sometimes illuminating.
References
I. A Handbook on Belt Conveyor Design, General Electric Co. Ltd.
2. Grierson, A. (1963). Trans. I.M.M. (London), 73, Dec.
3. Rasper, E. H. and P. (1960). Eng. and Min. J., Nov.
4. Govier, G. W., and Charles, M. E. (1961). Eng. Jnl., 44, No.8.
5. Ellis, H. S., and Round, G. E. (1963). Bull. Can. I.M.M., Oct.
6. Roberts, C. L. (1959). Proc. Aust. I.M.M., Dec.
7. Durand, R., and Condolios, E. (1952). Proc. Nat. Coal Board.
8. Bowen, F. B. (1963). Bull. Can.I.M.M., Oct.
9. Hardy, H. S., and Canaris, S. A. (1961). Min. Eng. (A.I.M.M.E.), March.
10. (1960). G.E.C. Journal, Vol. 27, 3.
11. Larsen, E. P. (1962). Min. Eng. (A.I.M.M.E.), Jan.
12. Jensen, V. J. E., and MacDonald, J. V. (1962). Western Miner and Oil Re'iel1',
June.
13. Mclellan, R. S. (1961). Min. Eng. (A.I.M.M.E.), March.
14. Couche, R. A. (1961). Aust. I.M.M., June.
15. Western Machinery Co., San Francisco, U.S.A., Trade Bulletin.
16. Stephenson, W. B. Trans. A.I.M.M.E., 187.
17. Denver Equipment Co. (1958). Bulletin, Jan-Feb.
18. Condolios. E.. Couratin and Pariset. (1961). Eng. Jnl. Canada, June.
19. Townend, D. S. (1957). Petroleum Times, Sept.
20. Anon. (1964). Mining Journal, 3rd Jan., p. 8.
21. Michaelson, Ensign and Hubbard. (1960). Trans. A.I.M.E., July.
22. Mussey, O. D. U.S. Geological Survey.
Mineral Processing-Unit Processes and Machines 701
References-continued
23. Rapier, P. M. (1960). Eng. and Min. In/., Dec.
24. Garman, C. N. (1962). Trans. A.I.M.E., 223, June.
25. Lynn, R. D., and Arlin, Z. E. (1962). Trans. S.M.£., Sept.
26. McCarty, M. R., and Olson, R. S. "Separan 2610", Dow Chemical Co.
27. Amer. Cyanamid Co. (1959). Cyanamid Flocculants.
28. Nalco Chern. Co., Chicago, Trade Bulletin.
29. Gieseke, E. W. (1962). Trans. Am. S.M.E., Dec.
30. Cornell, Emmett and Dahlstrom. (1957). New Orleans A.I.M.E., Feb.
31. Ritter, J. (1962). New Scientist, Jan.
32. Schytel, F. (1959). Metallgesellschaft A. G., No.1.
33. Tait, R. J. C. (1961). Can. Min. and Met. Bull., April.
34. Windolph, F. (1961). Mining Eng. (S.M.E.), Nov.
35. Given, E. V. (1959). Mining Eng. (A.I.M.E.), July.
36. Holiday, J., and Wilks, R. (1959). Eng. and Min. J., Oct.
37. Donaldson, G. W. (1960). Jnt. S. AI I.M.M., Oct.
38. Chenik, D. (1961). Jnt. S. AI, I.M.M., 65, Nov.
39. Jumikis, A. R. (1962). Soil Mechanics, van Nostrand.
CHAPTER 23
Aluminium
Bauxite
The formula is AlzO a. 2H zO, but this clayey or lateritic mineral includes
among its impurities iron oxide, phosphate and titania, together with silica.
Preliminary treatment often consists of a wash-scrubbing which leads to the
discard of a substantial amount of alumina-rich tailing. This can be treated
after a comparatively coarse grind and de-sliming with an 800 series collector
in acid (sulphuric) circuit, with fuel oil to aid froth stabilisation. Where the
concentrate thus produced carries iron or titanium minerals, tabling or high-
intensity magnetic treatment has been used for fjnal cleaning.
At Weipa, N. Australia, minus 3" material is wet-screened at 250 ton/hour.
Plus i" and minus 10 mesh sizes are rejected. The intermediate size, some
I 56 tons/hour, is sent without further treatment to the refinery. This screening
removes the bulk of the silica. The Jamaican bauxite treated by Reynolds
Metal Co. consists mainly of gibbsite and boehmite. It is low in silica but
high in iron and is treated by a modified Bayer process. l The essential
reaction is
Al z0 3. 3 HzO -' 2 NaOH = 2 NaAIO z -'-- 4 H 20 + red mud.
In simplified form the f10wsheet is shown in Fig. 299. Ore assaying 40 to
59 % AlzOa with 20 to 24 ~~ Fe203 and from 2 to 3 ~~ each of Si0 2 and TiO z
is fed at minus 2" to rod mills where it is ground with spent liquor. The
discharge slurry, at minus 20 mesh and 45~';; solids, is digested under pressure
at 200° C. without mechanical agitation. After half an hour pressure is
released by stages and the hot slurry is settled in four-tray thickeners which
discharge red mud to be washed by countercurrent decantation and discarded.
The thickener overflow is cooled in heat exchangers to 65° C. and sent te
Pachuca tanks, where it is precipitated with "seeded" alumina from storage.
Mineral Processing-Selected Ore Treatments 703
The solution is agitated by compressed air for 65 hours during which alumina
precipitates on the "seeds". These are then settled, washed and filtered
before kilning to yield 99 % Al 20 3, anhydrous. The aluminous liquor
remaining after precipitation passes via heat exchange to evaporating units and
is then topped up with sodium hydroxide and re-cycled. Some 50% of the
alumina is thus re-cycled, the balance leaving as precipitate.
t ~~-----
Wet grind (rod mills)
Steam
1----+ Digestion
~------'I---
!
Washing, filtration
!
Precipitation - - - - - - - - - ,
~
(recirculating liquor
to evaporators)
Calcining
Cryolite (Na3AIF6)
Activated by copper sulphate, collected by fatty acid salts or oleic acid,
aided by ortholuidine. Higher alcohols for frothers.
t
-325 mesh tailing -E<--- De-sliming
j SulPhuric acid
1 tr--- l
Fuel oil
Mahogany soap
Frother
ferrO- First flotation
II
E
{
Tail m~gneslans.
{ Hydrofluoric acid
mIca Amine acetate
t
Feldspar
t
Silica
carrying 60-70% feldspar, 20-30% quartz, 5% mica and some slate, garnet
and iron to obtain a 50% recovery, much of the loss being in a -200$ fraction
which is run to waste. After starvation conditioning with amine, fuel oil
pine oil, and sulphuric acid, the mica is floated off at 30-40 % solids. The
tailing is thickened to 70-80 % solids in a cyclone, further conditioned with a
very little hydrofluoric acid, R.825, and pine oil, and garnet and iron are
floated. Finally the feldspar is floated after depression of silica with further
hydrofluoric acid and the use of more amine, fuel oil and pine oil.
Kyanite (AlzSiOs)
Treatment may start by scrubbing with a solution of sodium hydroxide
followed by desliming. The collectors are oleic acid and fatty-acid salts, and
perhaps R.825, aided by aerosol. Pine oil and higher alcohols are used as
frothers, with sulphuric acid for cleaning. Pilot-plant tests are reported by
Browning. l3 Laboratory and pilot-scale work is also reported by R. A.
Wyman.152 The flow-sheet of a small American producer (Fig. 301) shows
several stages of flotation.3
Comminution
and Screening
I
< - n - - - p o t . Ethyl Xanthate
Flotation ---------,~
~
C
Pyrite, etc.
H 2S04 (pH3)
Sulphonated -----,1
1
Classification - - ( -) ~
{-
-150 ~ waste
Petroleum t
t
Roughing Floating ~
Tail
Cleaning Float
t
Drying
~
Magnetic Separation --------.t
I
Kyanite
Magnetic Waste
The U.S. specification for kyanite and sillimanite (AI 20 J • Si0 2) calls for at
least 59?1o Al 2 0 3 and less than 39 % Si0 2 , with under 0·75 % Fe;03.
Antimony, Arsenic
Antimony sulphide (stibnite) and arsenopyrite are among the metal sul-
phides liable to contain sufficient gold to justify flotation as a stage in its
recovery. The response of these minerals to a xanthate collector is much
improved by pre-activation with copper sulphate in a pulp rendered alkaline
with soda-ash. The specific dosage, reaction time, pH, etc., must be worked
out in each case. If the concentrate is to be roasted and treated for its gold,
it may be necessary to depress the antimony minerals or to remove them from
the roasted product by alkaline leach before cyanidation.
At Turhal, in Turkey, a readily slimed soft stibnite with some mispickel
and pyrite in a quartz gangue is first hand-picked to remove barren rock.
At the same time small lump stibnite assaying up to 55 % Sb is selected. The
ore is next impact-crushed through 8 mm., screened into three fractions and
jigged and tabled. 4 Middlings and slimes are floated after activation with
copper sulphate, using locally produced reagents.
A standard plan calls for a large circulating load in the grinding section in
order to minimise overgrind, since stibnite is very friable. Pulp is adjusted
with lime to a pH around 7·6 and activated either with copper sulphate or
lead acetate, up to 6-!- Ibs. of the latter being required in the conditioning
stage. If arsenopyrite or pyrite are present they are depressed in the flotation
cleaning stage by the use of 0·04 lb/ton of sodium cyanide. Where talc is
present it is also depressed in this stage, using yellow dextrin at 0·04 lb./ton.
Sodium isopropyl xanthate, also at about O·04lb./ton is the collector. Smelter
prices are based on a 65 % content of antimony and a grade below 50 % is
rarely profitable.
Asbestos
The objectives in milling asbestos are to free the potential fibre from the
slabby gangue in which it is sandwiched; to maintain the maximum length of
fibre in the end product; and to remove dust and grit at each stage of release
and progressive opening out so that as little as possible is trapped in the final
asbestos. Up to lO tons of air are drawn through the passing material for
each ton of fibre finally produced and special problems of dust catching and
disposal arise. Work may start with impact crushing, or the use of Aerofall
mills, Symons cones, impactors, ball mills or hammer mills as feed size
decreases and fluffiness increases. After rough drying, the broken ore is
repeatedly screened and re-crushed. It stratifies during passage along the
screens, the undersize being waste or a middlings of gangue and unopened
short fibre worth further treatment. The screen discharge goes on to finer
comminution and fiberisation while freed fibre is lifted by air elutriation,
using aspirators placed over the screens and trapping the lifted fibre in cyclones.
Mineral Processing-Selected Ore Treatments 707
Barite
Treatment aims at producing barium sulphate suited to several main
requirements, in the paint, paper, pharmaceutical and other industries.
Where colour is important it must meet rigid specifications, while lower
grade concentrates can tolerate discoloration. These latter are used to
blend drilling muds in oil wells and must have a density of at least
4'25, be 98 % minus 200 mesh, and have a viscosity of less than 60
centipoise. The ores carry iron carbonates, quartz, iron oxides, calcite
and sometimes fluorite. A generalised flow-sheet is given in Fig. 302,
in which gravity concentration at the coarser liberation sizes is followed by
froth flotation. In this sodium silicate is used as a dispersant in a circuit
brought to a pH of between 8 and IO with soda-ash and/or hydroxide. The
collectors for the iron minerals may include R 824 or R 825, and tall oil or
sodium oleate, with pine oil as frother.
At Rammelsberg the tailings from sulphide flotation (lead, zinc and pyrite)
are first cycloned at a dividing point of 40fllO to remove floating pyrite and
thicken the feed to between 45 and 40 % solids. It is roughed with Resanol
(a mixture of fatty acids) which has been fluidised with tall oil or oleic acid,
pH being adjusted to 6·5.
708 Mineral Processing-Selected Ore Treatments
Ore (-1~6")
r-+t" -- Scr ~
n-- -t" j
Jigs , Tails ~ Tables
t
Barite
1
Barite
Comminution
to -65 Mesh
I Float--- Flotation
~
Drying,
1
Tailings
Bagging
(BaS0 4 )
If the product is to be sold as drilling mud, heating during the drying stage
may be utilised to remove reagents undesirable for that purpose.
Beryllium
The pure mineral beryl (3 BeO . Al 20 3 • 6 Si0 2) contains 14 % BeO, and
the market calls for 10 % BeO minimum. The pegmatites in which this
mineral is found contain as major gangue minerals quartz, albite and musco-
vite, while associated "heavy minerals" include cassiterite, celumbite and tan-
talite, rarely in economic quantities. Spodumene, a lithium mica, may occur
in viable quantity. Hand-picking at the supply source is increasingly giving
place to flotation. After floating off muscovite the pulp is conditioned at
pH 5 with a tallow amine acetate and its beryl and feldspar are bulk-floated.
using an alcohol frother. 9 These minerals are separated by conditioning
with calcium hypochlorite followed by flotation of the beryl with a petro:eum
sulphonate.
Chemical attack at the research level has been described l l in connexion
with the upgrading of rough concentrates. For successful flotation pilot-
plant studies12 show that the basic requirements are (a) recognition of the
effect of the heavy associated minerals on product grade; (b) determination
of the flotation characteristics of these minerals and of beryl both in the
natural and the acid-washed state and (c) selective removal of heavy minerals
Mineral Processing-Selected Ore Treatments 709
Calcite
Limestone is the general term for rocks rich in calcium carbonate. Treat-
ment may be directed to upgrading this either by comminution and production
of an enriched slime fraction by classification, or by flotation of the calcite (and
if present dolomite) away from silica and sericite. Sodium silicate is then
used as a dispersant and oleic acid as collector agent, in a pulp made alkaline
with soda-ash. Sulphonated fatty acids may also be used. A secondary
flotation process is used when calcite must be reduced or removed before
leaching an uranium ore with acid.
Carbon
This is dealt with under the headings "Coal"; "Diamond"; and
" Graphite".
Chromium
The industrial mineral chromite (FeO. Cr203) is marketed either in terms of
its chromium content or of its Cr/Fe ratio where metallurgical grade is con-
cerned. Chromite brick is assessed on the refractory character of the raw
material. Beneficiation by gravity methods, which include DMS, tabling
and the use of spirals, is concerned to raise the Cr 203 content well above 40 %.
Iron may be partly replaced in the crystal lattice by other elements.
The gangue minerals are chiefly serpentine and olivine. At the I.M.D.C.
Congress at Goslar, 1955, L. Wenz reported studies on a Turkish ore con-
taining 30% Cr203, associated with oxides and sulphides. When ground to
-100ft the serpentine dominated the fines, chromite the medium particles
and olivine the coarse. A strongly alkaline pulp (pH + 10) was used, with
sodium oleate and oleic acid (llb./ton) to float the serpentine. The pulp
was then rendered acid (pH 3) and chromite was floated with t lb./ton
of a fatty-amine collector. Olivine, which only floats in a neutral solution
under these conditions, remained in the final tailing, the concentrate grade
being 45 %Cr03.
Mixed pyro-metallurgical, leaching and electro-chemical methods are used
to raise the Cr/Fe ratio to the degree required for ferro-chrome. In one
approach applied to a table concentrate14 the chromite is ground to minus
52 mesh and mixed with coke ground to minus 200 mesh before reduction
roasting at 1,250° C. The product is acid leached in 10 % boiling sulphuric
acid to yield a residual ferro-chrome of 11/1 ratio.
710 Mineral Processing-Selected Ore Treatments
Clay
Coal
For most purposes the calorific value of the coal is the quality required. The
ash content of this "packaged power" is worse than useless. It must be paid
for as coal, the freight charge of transport must be met, and the residual ash
712 Mineral Processing-Selected Ore Treatments
Cleaning Treatment
The purpose of this is to remove "free" dirt. Coal varies widely through
TABLE 43
(R('yi::;ion of H156)
Coals with ash of 0\,('1' 10,/,:, Il1U"t 1)(' riMIHJ(1 Lefore analysis for classi-
fication to gin' a maximulll ~'il'ld of coal with ash of 10% or less.
I
IVolatile Mulier
---,-
(~'r!!'ljwKii1f1 ('fl/tr
I
GrOl/p(,) CI".s Sub·(,hl •.'
I (d.w.mI) aeriMal Description
TYIlI'
U'u
100'
lOlt I ender 9'1
{Tml t• I' ,:·1
A
.\uthrarites
I .\
lO~t G'1-\I-(J
}
202 1~H--l;;'11 ll-O
203 L5'1-17'() 1~-n4 ('oking ,teum c""ls
204 17'1-IH'5 HI-UH
2()6 P·l-HHl A--ll for V.M. lIeat·altere([ low· volatile
9·1-lfi·n st<'am roal.
A-lJ for V.M.
15·1-19·fi
300 19·6-3~·O A-t:9 amI owr Me(lium·Yolatile coal.
301 19·6-32·0 U4 and over
301<, 19'6-27'5l (;4 and over
Prime coking coals
30111 ~7'6-32'0 I
305 19'6-3~'O 0-03 (Mainly) heat-altered
2(l6 19'6-3~'0 A-F medium· volatile coals
400 to Over 32·0 A-G9 and (lver High· volatile coals
900:-
400 OYer 32'0 0\1 and oyer
401 3N-36'0 Yery strongly caking coals
402 Over 36·0 09 and oyer
RUN'Of'MINE COAL
CONVfYQM
FROl'l WAGON ~ __ • __________ !Y!~~S___ ..
-----rIPPwis-----... ---------i. TO WAGONS
PIIII1AI\Y SCREE~ 10 'yO 8~
OVERSIZE
PICKING BELT i-oW
(OR INSPECTION BEl. T ). LAAGE S>!ALE ta'
=---....
8I\EAIlfR t:.o" OR HI
-~-.:;..;..;..;:-.---. t- ______________ .AL~'T'"
2 scmN 101VAGOHS
6'-l OR &-2" -t-ff
3 PRODOCT FLOAT
t. SINK SEPARATOR
CI.EAH
COAL flIES
fLOTATION
CLASSIFYING SCPHNS PLANT
,
1
I
I
I
Yi..o
•
I
HAND PlClED 6"-3" 3~z" z"-t" 1'~Yz' SHALE
LAP.GE IF
RfQulp.eo t~· 12M/M-O
\~-------~v~----------~
CLEAN
Dry-Cleaning Methods
Part of the value of coal cleaning is lost if as the result of removing dirt a
substantial amount of residual water must be shipped. At coarse sizes
washed coal drains readily to a fairly low moisture, but with the smaller sizes
capillary action in the storage bunkers and on the dewatering screens becomes
a problem. Washed "slack" at -in carries over 10% of water, and smaller
sizes are much more retentive, even when centrifugal force or vacuum filtration
is applied. A 10 % charge of water on ten tons of moist coal entering a
coke oven requires nearly 2! million B.T.U.'s for its evaporation (over 27%
of the heat applied to the charge). This would take more than a quarter of a
32-hour coking cycle, and if this feed moisture were reduced to 3 %, six hours
would be saved. Slurry, even after forced drainage, carries up to 25 % of
residual water and presents problems which stilI elude satisfactory solution.
No dry cleaning method yet evolved gives a complete solution of the com-
mercial problem. Those used include:
(1) Stationary devices worked with continuous or interrupted air streams.
(2) Dry reciprocating tables with air classification via porous beds.
(3) Appliances exploiting frictional and shape differences between coal
and dirt.
(4) Methods using centrifugal force.
The machines include air jigs. pneumatic tables ,and other oscillating
devices. With the increased use of water for suppressing dust in British
coal mines dry cleaning methods have suffered a set-back.
716 Mineral Processing-Selected Ore Treatments
Flotation
Compared with hard-rock treatment, coal flotation (as practised) is ex-
tremely simple. It is applicable to a fairly coarse mesh-of-grind, considera-
tions of liberation permitting, and gives a docile float at minus 48 mesh.
Current practice is extending froth flotation to coarser sizes, up to 2 mm.
being reported. 17 In 1960 four million tons of minus O' 5 mm. concentrate were
produced in Britain. In that year 92 % of the United Kingdom coal output
was mechanically mined and in many collieries the minus 12 mesh fraction
was around 30%, and was relatively high in ash. In the U.S.A.IS an average
top size of 28 mesh is floated, alcohol possibly aided by kerosene and/or fuel
oil being the reagent. Germany floats a minus 3 mesh feed, most of which is
below 150 mesh, and brings its ash content down from 20 % to below 8 % in
the product.
The fine coal thus produced is either dried for use as a powdered fuel, or
blended with coarser coals. Low rank brown coal and lignite is hydrophilic,
but when the carbon is above 80 % the chemical strut.ture changes and by
about 89 % carbon maximum hydrophobic qualityjs reached.
The flotation reagents most used are locally produced by-products rather
than those of specialised manufacture. They include cresyls, pine oil,
kerosene, creosote, fuel oil, gas oil and aliphatic alcohols. Among the
reagent combinations favoured in current British practice are creosote and
"cresylic acid" in 2:1 ratio, and medium fuel oil and "cresylic acid" in 5:1
ratio, the latter being used for lower-rank coals. These so-called cresylic
acids are often mixed with phenolic compounds distilled well above the nor-
mal boiling range for commercial cresylic acid. The bulk of the feed floats,
so withdrawal arrangements for the froth must be able to handle a large volume.
The value of the concentrate is far lower than in ore treatment, so high per-
centage recovery is less important than a good grade. Coal, being light, can
be worked in a dilute pulp. This aids the process by reducing trouble with
slimes from associated clay.
Owing to the high residual water in the final product, the direct use of
floated concentrate presents difficulties. It is increasingly mixed as a newly
produced filter cake with washed smalls, in which condition it is easy to move
through bunkers to automatically fired boilers, provided the blending has
been properly done. Owing to this moisture limitation, flotation was once
looked on rather as a purification process for colliery wash-water and sink-
float liquids than as a direct beneficiating process applicable to fines.
The ash content of coal in washery water rises as the rmrticle size drops. In a
typical test the + 14-mesh material has an ash content of 121 % and the - 200-
mesh 43* %. Flotation aims to reduce this. In a pilot test on feed carrying
the ash percentages just quoted, the washery slurry was thickened in a hydro-
separator with rising water at 12 ft./hour. This gave an overflow containing
much of the - 200 mesh material, and a reasonably thickened underflow
from an entering slurry which carried 90 <95 % water. Roughing flotation
of the underflow with a light fuel oil as collector and pine oil, the pulp con-
taining 17 < 20 % solids, gave a good coal recovery with 12i % ash at negligible
cost. Despite. centrifuging, the moisture of the product exceeded 25 %.
Provided the total output of flotation fines does not exceed some 10 % of the
Mineral Processing-Selected Ore Treatments 717
available "smalls" suitable for blending, its use in this way is economically
interesting.
In Britain the laws affecting discharge of washery water into rivers are
stringent. Effluents from flotation plants must not contain more than 2 p.p.m.
of phenols or 100 p.p.m. of suspended solids Flotation tailings are not now
impounded, but flocculated and filtered.
Apart from the importance of flotation in cleaning contaminated dense
media and in upgrading the high-ash coals in washery slurries, the possi-
bilities of flotation in the direct treatment of low-rank coal is being studied.
In Europe flotation is a useful stage in the further upgrading of already
high-purity coals, as part of the treatment to remove fixed ash before chemical
methods are used. The end-product is used in the production of carbon elec-
trodes.
Dewatering
Residual moisture is of five kinds:
runs the bunker and breaks the slurry into small lumps for delivery to mechani-
cal stokers. In another a layer of slurry is deposited on one of fuel passing
under the bunker on its conveyor belt. Here control of blend is made by
setting a guillotine gate and any required blending and delivery rate can be
obtained.
Dewatering of flotation concentrates has been cheapened and accelerated 19
by the use of f10cculants before filtration. Two main types of centrifuge are
also in use-basket and solid-bowl. The former are of three sub-types 20_
those which have scrapers to effect discharge, those with vibrating baskets to
move the solids and those with "pushers" or "peelers" to remove these. The
solid bowl types used in Britain are the Bird and Dynacone centrifuges.
The free liquid associated with a mass of small particles is either interstitial
or held as a surface film. Voidage depends mainly on particle shape and
size, and surface liquid on surface area. Displacement by unaided drainage
can only use small forces to oppose the capillary hold, but a centrifuge of the
type used in industrial plants develops up to seve(al hundred times the gravi-
tational effect. Experience with Durham coals indicates that minus feed r'
to centrifuges and filters should be split at about t mm. (32 mesh). The
coarser fraction, after centrifuging had a moisture of 4-:\- %. In the solid
bowl type of machine a cylindrical bowl, or one shaped like a truncated cone,
revolves on a horizontal axis. Its settled contents are moved from the wall
by a screw conveyor moving at a slightly different speed. Feed enters cen-
trally and solids and liquids are discharged from opposite ends. Effective
classification, in terms of residual moisture and of solids left in the liquid
effluent, is related to the depth of liquid being centrifuged (the "pond" depth.)
The shallower this is, the longer is the "beach" over which solids must move
to the discharge and the lower the discharge moisture. Solid bowls are
used in the minus 1" size range and for de-watering flotation froth and tailings.
Automated thermal drying by use of a rotary kiln in one Canadian operation
uses a drum 6' in diameter and 40' long, turned at 8 r.p.m. 21 Heating is by a
mixture of oil and coal dust, the latter being regulated by the temperature of
the exhaust gases and the oil acting mainly as a pilot flame. The system
described has a throughput of 120 tons/hour of minus t" coal fed at between
10% and 15 % moisture and discharged at 4 % moisture. About ~ gallons of
oil and up to 40 lb of coal dust are consumed per ton dried.
Grading
The oldest method of grading made use of convaiescent mining labour fit
for light surface duty. The coal was displayed on picking belts and worked
over for sizes, dirt, and pyritics. Below a handsorting size screening was
used to produce the various sizes (variously named in different coalfields)
such as "treble nuts" (3"-Ir'), "double nuts" (Jr'-I"), or "peas", beans,
pearls, nutty slack, and fine slack (duff or dart). A quiet revolution in
practice has been under way since the nationalisation of Britain's coal in-
dustry. New coal-preparation plants have better feed arrangements and
more mechanisation, and a measure of standardisation of screen sizes has
replaced the traditional confusion.
The nomenclature used by the National Coal Board is based on that set out
Mineral Processing-Selected Ore Treatments 719
Cobalt
Most of the world's supply is a by-product from the concentration of
copper or nickel ores. At Chibuluma in Zambia 22 a 3 % concentrate
containing some 36 % of iron, 44 % sulphur and 10 % silica is upgraded to
about 20% each of cobalt and copper by mixed smelting and leaching. This
starts with production of a low-grade matte in which re-cycled slag containing
some cobalt forms part of the charge. The product assays 4t %each Co and
Cu and is crushed, ground and calcined. Soluble sulphates are thus produced
in fluidised roasting, and these are dissolved and upgraded first by reaction in
a cobalt-copper-iron sulphate solution which removes a barren liquor and
second by further leaching which upgrades the insoluble portion to shipping
assay, the liquor becoming the leach solution for the next batch of fresh matte.
The pressure leaching of a concentrate assaying ITt % Co, 24 % As, 20 %
Fe with some Ni and Cu has been described by J. S. Mitchel1. 24
After cleaning to an As/Fe ratio of 1·2: I it is pulped to 26 % solids and
oxidised in a carbon steel autoclave cladded with stainless steel, with an
additional lead lining and acid-resistant brick. The pulp is wei red through
six successive compartments of the 40' long vessel. Each has its own agitator
and aeration, and high pressure pumps force the feed in. Reaction is exo-
thermic at an operating pressure of 500 p.s.i. and 190 0 C. On discharge,
90 % of the cobalt is in solution as sulphate. The residual acid in the dis-
charge is nearly neutralised with lime. Pregnant solution is filtered off
and stripped of its gypsum and iron salts by precipitation methods. Copper
is next removed by precipitation as a cement. Finally, metallic cobalt is
produced by batch treatment in vertical autoclaves which hold 450 gallons.
Anhydrous ammonia is added, the autoclave is heated to 190 0 C. and hydro-
gen is blown in, the pressure being 750-800 p.s.i. The cobalt amine now
seeds down as metal.
The use of Flotagen (mercaptobenzothiazole plus sodium carbonate) is
reported as a specific collector for cobaltite, smaltite, and erythrite from a
diabase gangue. Diphenyl thiocarbazid (C6H5NHNHhC = S has been
usec\ for the selective flotation of cobalt and nickel from copper and iron
720 Mineral Processing-Selected Ore Treatments
Copper
The copper-bearing ores include so wide a range that industrial treatment
uses nearly every processing technique, sometimes in highly specialised ways.
Copper minerals range from straight sulphides via copper-iron sulphides to
oxides, carbonates, silicates and chlorides. Often associated minerals must
be recovered separately, including gold, silver, pyrites, cobalt, molybdenum,
germanium, lead and zinc. The practical examples outlined in the following
sub-sections illustrate the use of chemical extraction (leaching) and its modi-
fications in the L.P.F. process (leach-precipitation-float); straight flotation of
typical copper ores; differential flotation of mixed ores; and pyro-metallurgy
jn the segregation process.
Leaching
The chemical reactions used in leaching the more common ores are:
Azurite: CU3(OHh. (C0 3h + 3 H 2S04 - 3 CuS04 + 2 C02
+4HzO
Brochanite: CuiOH)6. S04 + 3 H2S04 - 4 CUS04 + 6 H20
Chalcocite: CU2S + Fe2(S04h - CuS + CUS04 + 2 FeS04
(Covellite)
Chalcopyrite: CuFeS2 + 2 Fe2(S04h + 2 H20 + 302 - -
CUS04 + 5 H20 + 2 H2S04
Chrysocolla: Cu . Si03 . 2H20 + H2S04 - - CUS04 + Si02 + 3H20
Copper, native: Cu + Fe.2(S04h - CUS04 + 2 FeS04
Covellite: CuS -I- Fe2(S04h - CUS04 + 2 FeS04 + S
Cuprite: CU20 + H2S04 + Fe2(S04h - 2 CUS04 + 2 FeS04
+ H20
Malachite: CU2(OH}zC03 + 2 H2S04 - 2 CUS04 + C02 + 3 H20
Pyrite: 4 FeS2 + 1502 - 2 Fe2(S04h + 2 H2S04
Tenorite: CuO + H2S04 - CUS04 + H20
Heap roasting and leaching have been practised in Spain and Germany
since the sixteenth century. Dump leaching is currently an important
source of copper.
Pit leaching has been practised for many years, but its methods have today
been mechanised and accelerated. Weed 26 has described the Cananea
Mineral Processing-Selected Ore Treatments 721
method used since 1946. The material includes pit dumps (0'2% or so Cu)
and caved stopes. The chalcocite this carries is readily leached in weak
sulphuric acid carrying ferric iron. Water is led in, percolated, and pumped
up to a precipitation plant built of wooden grids in wood cells. Shredded
iron from old cans catches the contained copper. The pregnant solution
carried 3· 3 g./I. of Cu, 3 g./l, of ferrous, and 7·4 of ferric iron and has a pH of
2·3. It emerges from the precipitation cells with 0·25 g./I. of Cu and a pH of
2·9.
At Bingham, Utah, dump leaching, which produced 16,678 tons in 1962, is
planned for expansion to 72,000 tons. 27 Several factors have led to this
revival of an ancient technique. Leaching is largely self-sustaining, provided
handling arrangements permit adequate exposure of the copper mineral to air
and water. The role of bacteria in chemical attack is now understood and
exploited. Huge tonnages can be moved by mechanised handling. Protective
materials minimise corrosion of exposed installations. The process is cheap
and requires but little manpower. For success the dump must not contain
much acid-consuming limestone and must be underlain by an impervious
stratum. Natural oxidation may contribute to the supply of sulphuric
acid and the iron sulphate essential to leaching action. Any pyrite present
must make contact with the leach liquors if it is to help the replenishment of
acid and ferrous sulphate, but must not build up excessively re-circulating
liquor or the retrieval of the copper from the pregnant solution will run into
difficulties.
Treatment suited to dump leaching is also applicable to caved copper-
bearing rock or ore in situ, as has been proved at Ray and Miami, in the
American southwest mines. Here the ore is already broken or is pervious
and the leaching liquor is sprayed over the deposit, whence it percolates
through and drains to old workings. It is then pumped to surface and strip-
ped of its copper. At Kennecott research has shown the conditions under
which bacterial activity thrives include humidity, warmth and darkness.
These bacilli oxidise pyrite to sulphuric acid and ferrous sulphate and convert
ferrous to ferric iron. The acid takes up ferric sulphate which during attack
on the sulphide copper is reduced to ferrous, thus forming food for bacterial
sustenance. Leaching solutions are controlled to a pH of 2'3, at which hy-
drolysis of the copper is avoided.
Stripping of the pregnant liquor is by reaction with iron. Crumpled cans
from which lacquer or tin has been removed are favoured, the reaction
being
Fe + CUS04 = Cu + FeS04 This is followed by
Fe + H 2S04 = H2 + FeS04 and finally
Fe -!- Fe2(S04) = 3FeS04'
Scrap iron and sponge iron are also used. Stoicphiometrically, one unit of
iron precipitates nearly 1·14 of copper but in practice the Fe/Cu ratio is
usually nearer 3 to 1. The resulting low-grade cement copper assays from
50 % to 90 % Cu, and recovery is high.
Acid strength is usually below 7 %and where necessary is maintained by the
"autoxidation process".28 In this ferrous sulphate catalyses the leach liquor
722 Mineral Processing-Selected Ore Treatments
in the presence of sulphur dioxide gas and oxygen and produces the needed
strength of sulphuric acid and ferric sulphate. To do this barren solution from
copper precipitation is diluted to some 5 % strength of ferrous sulphate and
reacted in aerating cells with gas produced by burning elemental sulphur.
At Bagdad, in Arizona, a leaching plant produces some 20 tons of cement
copper daily from low-grade oxidic stockpiles and oxide ore overburden.
Each of two stockpiles, in adjacent canyons, has its surface divided into
100' square pools separated by embankments and delivering to a common
overflow. Acidic water is sent from the precipitation plant to the leaching
area and run on to one of two settling ponds. From the pond pregnant
liquor containing 1 g./I. of copper passes to ten 2-compartment stripping
cells in series. Each has a steel grid on which rests a load of de-tinned cans,
the depth to this grid being 13' and the cross-section 10' by 91'. Cement
copper falls through and the ratio of iron to copper is 1·6 to 1.29 N'Changa,
on the copperbelt of Zambia, has the world's largest plant engaged in the
electro-winning of copper. In the concentrator three products are made-
a sulphide which is smelted, an oxide float containing from 11 % to 14 %
acid-soluble copper and a low-grade oxide with about 2 % copper. These
are leached with spent electrolyte from the "tank house" where cathodes are
made electro-chemically. The higher grade oxides are treated in agitators
and the lower ones in pachucas. Pregnant liquor is filtered off and sent to the
tank house. 30
A useful recovery of copper from old mine workings, and from drainage
areas in copper-bearing rocks, can be made by simple precipitation on scrap
iron if the iron in the pregnant water has been reduced to the ferrous state.
In one area the mine water is left in contact with pyrrhotite for this purpose,31
the reaction being
Fe 7SS + 32 H 20 + 31 Fe2(S04h -69 FeS04 + 32 H2S04
and the effect that pyrrhotite replaces a substantial part of the scrap iron.
The leaching of atacamite. a copper oxychloride with the formula CuCl 2
3Cu(OH}z presents special problems. The method developed at Mantos
Blancos, in Chile32 .33 , uses percolation leaching with sulphuric acid followed
by precipitation of the copper with sulphurous acid as cuprous chloride. The
presence of chlorine precludes electro-precipitation and no local supplies of
scrap iron are available. Precipitation is effected in spray towers through
which gaseous S02 moves counter-current to the pregnant solution, reaction
being:
(a) 2 CuCh -1- S02 + 2 H20 - 2 CuCI + H2S04 + 2 HCI
(b) 2 CUS04 -1- 2 NaCI + S02 -i- 2 H20 _2 CuCi +
2 H2S04 -;- Na2S04
At Bagdad Copper Cpn., a 5 ton per day pilot plant is working out a novel
method. Sulphide concentrates from copper flotation are roasted by the
Fluo-Solids method. The calcine is then leached and the copper is recovered
by electro-deposition. Sulphuric acid generated during this work is thus
available for .heap leaching oxidic copper ore too poor to warrant normal
treatment. Part of the expected saving in this process will be the cutting of
Mineral Processing-Selected Ore Treatments 723
Leach-precipitation~float(L.P.F.)
A typical flow-sheet for this process is that of the Hayden plant (Fig. 304).
I
(- ) - - - Grind and - - (-'-)
Classify I
f ~-tr----~
Leach < - - (Slimes) - - Leach
(pH 2·1) and Classify
1
Precipitation <_----,
1
(Sands) Lime
I _I Raconite
I /(M",,,;,,,
"
I
Magnets r Tail
(Float)
I Sponge
--C(c---- Flotation Iron Roast
I -""---- Reduce
t to iron---)~SO,
Tail ~-
Sponge Contact
Copper plant
t
H 2S04 -+-,
In other variations of this scheme sulphuric acid is pumped over old dumps,
ore piles, etc., ill situ. The resulting pregnant solution is stripped of its
copper. reconstitllted and recirculated.
724 Mineral Processing-Selected Ore Treatments
The use of sulphuric acid in the L.P.F. process has the further effect of
cleaning partially oxidised sulphide surfaces during treatment with sulphuric
acid, following which flotation with xanthate may be improved.
At Inspiration, Arizona, the old process called for leaching followed by
electrolysis. With impoverishment of the mine at depth together with an
increase in the proportion of sulphidic copper the old acid ferric sulphate
process which took nine days has been dropped. Instead, leach contact
with straight acid for four days now deals with the oxidic copper, leaving
clean sulphides to be recovered by flotation. Ray (Kennecott) has
developed a L.P.F. flow-sheet for recovering oxidic copper. The values
include chrysocolla, cuprite, malachite, tenurite, and native copper.
The carbonate, silicate, and sulphate are the main objects of attack. Copper
is got into solution by standard leaching methods. The residue is reground,
given froth flotation to recover the sulphide copper, and then refloated for its
pyrite. This pyrite is calcined, producing SOb which after reaction to H 2S04
is used for leaching. The ash is reduced with coal to yield finely divided iron.
This iron is circulated through the precipitation plant and is then removed by
flotation as a copper-iron froth, any iron not caught at that stage being picked
up, if needed, by magnetic separation of the tailing.
L.P.F. as a method of treating mixed sulphide-oxidic copper ores was
developed from tests started in Arizona in 1929. "Oxidic" includes copper
oxides, carbonates and silicate. The original full-scale operation (1934-
1943) and a revived form (1957-1959) have been reviewed by Bean 34 .35 . In
the finalised process the sulphide minerals were first floated and the tailings,
in which chrysocolla was the dominant copper value, were thickened and
leached with sulphuric acid in air-lift agitators. Copper was then precipi-
tated during passage of the "pregs" over de-tinned and shredded cans agitated
in wooden drums, and floated. At Hayden it was found that there should be
an excess of minus 35 mesh powdered iron in the secondary (acid) stage of
flotation. Any unconsumed iron was removed by magnetic separation
applied to the scavenger tailings. At Rosita, Nicaragua, highly refractory
ores are treated by L.P.F.36 The copper is precipitated on shredded iron,
screened to pass 6 mesh and floated with Minerec A and Aerofloat 25.
Three types of precipitate must be allowed for-a readily floated fluffy cement
copper, a reluctant hard and granular cement, and copper-plated iron particles
giving a low-grade but payable concentrate.
Flotation, sulphide copper
The copper-containing ores normally treated by flotation may be divided
into two classes with respect to reagent response. The sulphides, with which
are included native copper and such ores as selenides, tellurides, arsenides,
and antimonides, respond readily to xanthate collectors, notably potassium
ethyl xanthate. Selectivity is improved by care in grinding, by avoidance of
undue oxidation through delay in treatment after severance, and by judicious
use of polysulphides where such oxidation cannot be avoided. The main
sulphides are chalcocite (CU2S), covellite (CuS), chalcopyrite (CuFeS2)' and
bornite (CuiFeS4)' Minor sulphides include chalmosite (CuFe2S3) and
cubanite (CuFe2S4)' Broadly, the greater the copper-iron ratio the higher is
Mineral Processing-Selected Ore Treatments 725
the upper limit of pH and the lower the risk of depression through the use of
lime, which is used not only to control pH but also to depress pyrite. This
depression may be increased by pre-aeration to decrease the attraction of
xanthate to the altered iron surface, and by the use of sodium cyanide in very
small quantities. If copper is only one of the minerals to be floated, pre-
aeration must be considered in relation to its possible effects on the later
work. Copper, being the most readily floatable of the major sulphides barring
molybdenite, is taken first in a differential treatment. The use of cyanide
might be precluded in the case of an ore carrying appreciable gold which
could best be recovered from a floated product.
At Noranda copper occurs mainly as chalcopyrite with associated pyrite-
pyrrhotite, all carrying finely disseminated gold. The pyrrhotite is unstable
and care is essential in the transport of the severed ore, since delay produces
oxidation which has an adverse effect on subsequent treatment. Fine secon-
dary grinding is necessary in order to free gold before rejecting pyrrhotite
which, by reason of its instability, is a stong cyanicide. It is also necessary to
float substantially all the copper before cyanidation of the auriferous
pyrite. The flow-sheet (Fig. 305) of the heavily pyrrhotitic section of the
Ore pulp
1
(Float) ...-,,------ Flotation (copper sulphide
and gold)
I Flotition
Mixed (pyrites)l
-<Float) (
(Float)
~_R_e_tnd alnd float j
(Tail)
ore body shows in condensed form the method used. A fuller description
is given by Ames. 3s At Adak (Sweden)39 ores from two adjacent mines are
treated, the principal minerals being copper sulphide and pyrrhotite. The
pulp pH is under automatic control, lime being the alkali used. Potassium
amyl xanthate is the collector. The flow-sheet (Fig. 306) is unusual in the
number of re-cleaning stages it incorporates.
726 Mineral Processing-Selected Ore Treatments
t
~Iassification -------.t
(--) i
1 (mid)
(I·)
Ball mill
n
Roughing
t.----------Flotation
Tail 1
leaning
Flotation
1
(first stage)
(Tails)
i<-_____ Re-c1eaning
(four stages)
t
Final Copper
Concentrate
The change from the pure sulphide CUzS which floats at pH up to 14, with
increasing structural iron down to chalcopyrite (maximum tolerable pH 11.8),
shows the coincidence of structural change and electro-chemical response.
If it were necessary to float one copper-bearing sulphide away from others.
this could be utilised. Normally. the purpose of the work is to float as much
of the copper as possible for a specified grade of concentrate. and a suitable
pH for this purpose is maintained. Pyrite is usually present in the ore. and a
lime-controlled pH above II is used to aid its depression. Iron sulphide
floats with increasing reluctance as the pH rises above 6·5 and lime has a
depressing effect. When copper minerals are treated, there is some ionisation
during grind and conditioning. and this results in a certain amount of activa-
tion of any non-copper sulphides present. as well as of selenides. tellurides, ar-
senides, or antimonides which may I::e in the ore. The soluble copper causing
this can be complexed by the addition of a little sodium cyanide, but this may
depress bornite and chalcopyrite which carry enough structural iron to be
easily depressed by a conditioning treatment intended to keep down iron.
Bornite is also prone to oxidise. and the pre-flotation processes (mining.
transport, storage, and grinding) must deal with this. The reagent plan for
the copper sulphides is dispersion of masking slimes (if needed) by the use of a
specified sodium silicate (the trade terminology is loose as regards this reagent);
pH control by calcium hydroxide; possible use of cyanide to de-activate
Mineral Processing-Selected Ore Treatments 727
copper-surfaced iron or other alien sulphides; and the most suitable xanthate
(found by test). Usually potassium ethyl xanthate is preferred, in quantities
varying from O·OI-lb./ton up to 0·1. The pulp is conditioned for pH and
dispersion in the grinding circuit, and requires only a few minutes with the
xanthate, at normal milling temperatures and from 20% to 30% solids. As
with all floats, the higher the temperature' (and therefore the lower the pulp
viscosity) the shorter can be the final conditioning period. Also, the coarser
the grind of the ore being floated, the higher should be its percentage of pulp
solids in the first roughing stage. The frother is usually pine oil, added
immediately ahead of the first cell. pH drop through the cell series is rarely
serious enough to require attention, but occasionally a make-up addition of
pine oil or xanthate in the scavenger section may be economically justified.
Native copper floats strongly with a xanthate.
A special case is the reflotation of copper sulphide from a copper-nickel
matte 40 as practised at Sudbury, N. Ontario. The pH is held between to and
12'4, a high OH- concentration being needed to depress nickel sulphide.
Xanthate may be used, but the preferred collector is diphenyl guanidine
(0'5 Ib./ton) or di-artho tolyl guanidine, added during grinding. Pine oil may
be added as frother, and lime is used to control alkalinity. From three to six
rougher-cleaner stages are used with intermediate regrind and some 7%+
325-mesh at completion of separation. The guanidine provides sufficient
frothing action as a rule.
The chief non-sulphide ores to be floated are azurite (2CuC0 3Cu(OHh),
cuprite (Cu20), and malachite (CuC0 3 Cu(OHh). No satisfactory method of
floating chrysocolla (CuSi0 32 . H 20) has yet been developed. The formulae
given should be accepted with reserve. X-ray and chemical analysis of
malachite'! proved that in the case of a typical ore no clear-cut formulation
was justified. The general treatment plan for ores found in the Congo and
Rhodesian copper belt commences with a xanthate float for sulphide copper.
The tailings are conditioned with collectors based on fatty acids derived mainly
from palm kernels and cottonseed oil. Sodium sulphide is added in the
conditioners and along the flotation line. The prevailing mill temperature
is subtropical, which aids the rate of reaction. Research 41 has shown that
given close control of the hydroxyl ion balance against the carbonate ion,
straight xanthate flotation is possible.
In the old N'Changa treatment, as described by Talbot,42 an important
consideration affected flow-sheet design. Coal must be brought over a
congested railway system at considerable transport cost whereas electricity,
now being developed from hydro-electric resources, was temporarily in
short supply. Thus, both from cost and availability considerations, it
was decided to leach as much concentrate as possible and to smelt only the
residue. The flow-sheet was, therefore, developed as in Fig. 307. The main
copper values in the ore are sulphides, native copper, azurite. and malachite.
At Kolwezi in the Belgian Congo copper and cobalt are bulk floated. The
upper mining levels carry malachite, chrysocolla, heteroganite (CoO. 2 CO Z0 3
. 6H 20) and cobaltiferous wad, which are floated with hydrolysed palm oil,
dispersion being aided by sodium silicate and carbonate. A mill-head of
5 to 6 % copper and O' 2 ~~ cobalt yields a 26 % copper concentrate with up to
728 Mineral Processing-Selected Ore Treatments
Ore
!
Sulphides
!
(Float)
Float
!
(Tail)
t
Mixed Float l
1
(Float) Tail
1
Leach
Plant
!
l
Pregs. to
Sewndary Treatment
Sulphide
Float
Smelter
t
Tail
In the subsequent zinc flotation live steam is used to keep the cleaning section
of conditioning up to 30 e C Flotation tails are cyanided to recover the rest
of the gold. 43
Separate circuits are used for the sulphide and talcose ores (Fig. 308).
Talc is roughed out, cleaned twice, and discarded, the tail from this float
joining the feed to the secondary sulphide float. Copper is floated in two
grind-and-float stages, the concentrates being combined and recleaned. Zinc
is taken from the copper flotation tail, cyanidation being a final scavenging
operation used to recover residual gold and silver after this operation.
Special care is needed in the talc flotation to limit the loss of valuable sul-
phides. Some of the floatable talc mineral is allowed to escape to the sul-
phidecircuit. These "insolubles" are depressed by adding sulphurous acid and
dextrin at the head of the copper cleaner circuit.
1
Comminution and
.1.
CommmutlOn an d
Classification Cla"'i]",lion
1
r--Talc Float~
Primary Copper Float
I (f)
(f) '''' (1) j
1
Talc ""~ 1 (f)
Regrind
I (t)
~1
I ~ec. Cl Float
r Zinc Float!
L
(~) (f.2)
(I) (f) j
j ) T,rti.,y Co Floal----..l
to Cyanide Zinc Copper
Plant -
In the talc circuit zinc is depressed with 0·03 Ib./ton of lime at the ball mill.
together with 0·034 of zinc sulphate. The only flotation agent is pine oil
(0,007). The copper reagents are lime (I·n), soda-ash ('55), zinc sulphate
730 Mineral Processing-Selected Ore Treatments
('63), Aerofloat 25 ('041), Minerec T-27 (0'076) in the primary and secondary
stages. Sodium ethyl xanthate ('022), dextrins 615-620 ('41), and sulphur
dioxide (S = 0'25) are used in the late stages. The zinc is floated with lime
as iron depressant (2'57), copper sulphate (Cu = '15), minerec 748 ('03), and
Aerofloat 25 ('056) for roughing, with further copper (0'033) and ethyl xanthate
('029) in the scavenge-float.
A substantial amount of molybdenite is bulk-floated with sulphide copper,
notably in America's western porphyrites. A typical concentrate assays
30% Cu and from 0·5 to 0'9% MO.44 Five main methods are used to separate
these minerals from the float. Steam or dry-heating may be lIsed to partially
remove or destroy the collector agent, which adheres preferentially to molybd-
denite. Temperatures from 90 0 to 140 0 C. are used over periods from I to 4
hours (prior to re-flotation of the MoS 2 away from the copper using only a
frothing agent). In a second method the molybdenite is depressed by organic
colloids such as starch and glue. One plant applies this treatment to a float
originally made with a dithiophosphate collector and re-floats the copper with
dithiocarbonates after depression with dextrin. A third method uses
sodium ferrocyanide in a weakly alkaline pulp to depress the copper, with
fuel oil and an alcohol frother to lift the molybdenite. A fourth treatment
depresses the copper with sodium hypochlorite in an alkaline pulp, in com-
bination with the third of these methods. In a fifth method, copper minerals
are depressed by use of phosphorus, arsenic or antimony salts. The products
from these various methods of differentiation assay from 3 % to 30 % MoS z
and to bring them to shipping grade repeated cleaning, re-grinding, roasting,
leaching and the use of special reagents may be needed.
A summary of Russian practice in copper-molybdenum flotation has been
made by Crabtree. 45
The Lynn Lake deposits mined by Sherritt Gordon in North Manitoba
contain copper as chalcopyrite, pyrrhotite, pentlandite and some cobalt, zinc
and gold. Talc and actinolite are troublesome gangue minerals dealt with by
organic depressants. Ore is ground to 50 % minus 100 mesh and separated
into magnetic and non-magnetic fractions. The former is re-ground to
90?;'; minus 200 mesh, re-magnetised and floated to yield a 3 % nickel concen-
trate. Non-magnetics are ground to 50% minus 200 mesh and the copper
plus the rest of the nickel is bulk-floated. The concentrate is then acidified
to pH 5·6 and re-floated in a 16-cell cleaning bank. Float from the first four
cells, is alkalised with lime, and the nickel is depressed with cyanide, the
resultant tailing joining the magnetic concentrate and bringing it up to an
assay value between 10 and 12 % Ni, while the float assays 31 ?;.; Cu. The
balance of the cleaner section is scavenged before discard. 46
Research and large-scale pilot tests show good possibilities for silicates
and oxidic ores not readily amenable to straight leaching and/or flotation.
Sutulov48 estimates Chile's reserves of "difficult" mixed copper ores to be some
100 million tons assaying from It% to 5% Cu. The general pattern of
treatment starts with crushing to minus 10 mesh. The product is thoroughly
mixed with common salt of minus 20 mesh size and minus 40 mesh coke. Up
to 2 % salt and 1 %coke may be needed. Excess of these leads to embrittlement
of the froth at the flotation stage and too much salt may react unfavourably
during pre-flotation grinding even when the pulp is raised to a pH of 10·5
with lime.
Kilning is performed in a rotary calcining furnace indirectly heated. the
main factor here being the furnace temperature. In tests in a 175-ton/day
pilot operation on southwestern American oxide-sulphide copper ores the
reaction temperature aimed at was 760 0 C. reached 22' along a 48' kiln 54"
in internal diameter and having a surface temperature of about 8700 C The
copper is volatilised from its minerals as cuprous chloride and then reduced
to metal flakes up to i" by t" which are deposited on the coke particles. 49
These flakes separate on agitation. The optimum temperature quoted in the
Chilean tests is 750 0 C. volatilisation leading to loss beginning at about 800°
C. but here pre-heating and slower calcining times were used. Ninety per
cent recoveries were obtained save where there was calcium in the gangue
when yield dropped below 86°,{,. Specular hematite made its ores refractory.
recovery falling below 70 %. The collector agent found best with the Chilean
ores was Z-200 with a cresylic acid frother. American work used amyl
xanthate and methyl-isobutyl-carbinol at a pH of 11·5.
Water
Figures published by the U.S. Geological Survey show that about 100.000
gallons of water are used per ton of copper produced from U.S. domestic
ores of which 70 % is consumed in mining and milling. Of this 46 % is saline
with 1.000 p.p.m. of dissolved solids and 54 % is fresh. In arid country
considerable economy is practised by recirculation of mill water. coupled with
precautions against excessive evaporation.
Diamond
The dominant facts in diamond concentration are that the concentration
ratio is of the order of 20 millions to I. that the stones must be recovered
intact and undamaged and that. despite the security risks attendant on the
handling of such values a high percentage recovery is required. Three main
types of deposit are exploited-kimberlite ("blue ground"). marine terraces
and alluvial gravels. the latter including offshore deposits. Kimberlite, the
ore of volcanic pipes and fissures. has a blue-grey matrix which consists
mostly of serpentine. with such subsidiary minerals as calcite. diopside.
enstatite. ilmenite. phlogopite. pyrite. pyrope. and-rarely-a diamond. 50
Marine terraces and alluvial gravels need only screening before treatment, but
kimberlite must be disintegrated. At Kimberley the stages of concentration
732 Mineral Processing-Selected Ore Treatments
today start with washing pans worked at a 32/1 ratio, followed by DMS at a
6/1 ratio. The sinking fraction is then re-concentrated on greasy surfaces at a
50,000/1 ratio and hand sorting at 2/1 completes the work. At Premier
Mines on the Rand, grease vibrating tables take over from a 25/ I concentra-
tion by DMS and work at a ratio of 80,000/1.
Despite experiments with alternative methods, Kimberley retains pan
washing for its clayey blue ground. 51 This clay acts as a natural dense media
in a slurry blended to a density of 1.26 to 1.28. Mine ore at minus I ±" is fed
to primary concentrating pans where 80% of the diamonds are held. The
overflow from the inner concentric weir of the pan is screened through r
and oversize is crushed to minus i" and returned. Undersize is given a
secondary panning, and thence goes to 6-mesh screens from which oversize
is discarded. The underflow is cycloned to retain minus 6 pills 28 mesh
and the cyclone overflow recirculates as "puddle" through the plant.
A washing pan has two concentric walls at 7' and 3' radii, the outer of which
is 20· high and the inner 12". A sliding bottom door facilitates clean-up. A
vertical shaft at the axis has ten stirring arms ·furnished with small triangular
teeth which thrust the settled material outward against the inward flow of the
slurry. A single circular tooth at the end of the arm moves settled material
to a discharge port in the outer wall. Arms revolve at 8 r.p.m. The settling
minerals are diamond, ilmenite, garnet and diopside. The rakes keep the
bottom slurry in semi-suspension at a S.G. between 1·5 and 1·6 with a 50%
slip past the stirring teeth. The concentrate taken at this stage should be
0·25 % but for safety's sake a settled cut of 2 % is removed. This is cleaned up
by grease tabling after DMS. Since 1958 the plus 10 mesh fraction has been
sent to DMS and the mil/us 10 pills 28 and mil/us 28 mesh sizes to separate grease
table operations. The grease used is a refined petrolatum having four
grades of penetration point, the one selected having the best hardness of
arresting surface for the grain size of the diamondiferous feed. Diamonds
caught on the grease are scraped off periodically by hand. tables being in
pairs so that one is being fed while the other is dressed. About an equal
quantity of adherent gangue is removed at the same time. The grease is melted
away with boiling water and the diamonds are dried and sieved on IO-mesh.
Undersize is wet-milled, de-slimed and again degreased in chromic acid, and
finally concentrated either by flotation or electrostatic separation. From a
daily input of 13,500 tons of kimberlite the final concentrate of diamond
weighs 2,800 carats (1,25 Ibs.). Until 1948 the grease tables used at Kimberley
had a side-shake imparted by a vibrator but this was then superseded by
high-frequency vibrations of the table deck.
Although diamonds from freshly mined kimberlite respond to grease
tabling, those from alluvial and marine deposits do not. The surface is
restored to its water-repelling state by gentle scrubbing of the DMS product
and conditioned with a fatty acid if grease tabling is to follow. For electro-
static separation, one mine in S.W. Africa 52 sends the DMS concentrate to
screens after drying at 130 0 C. Dust and oversize are removed and the minus
6 plus 1·9 mm. fraction is separated at 22,000 volts.
Checking of separating densities is performed either by use of a variant of
the coal washery sink-float curve or by feeding radio-activated isotopes
Mineral Processing-Selected Ore Treatments 733
through the recovery plant and tracing their progress with Geiger-Muller
counters. The S.G. of diamond is 3·5 but 90% of the associated mineral is
2,8. 53 Control of the viscosity of the separating media is important.
Fluorine
The commercially floated ore, fluorite, occurs in association with silica,
calcite, barite and sometimes such sulphides as lead, zinc or iron. These
last are floated or removed by gravity concentration before dealing with the
fluorite. The market for acid-grade CaF2 demands a high degree of purity,
and the treatment usually calls for thorough recleaning to remove the last of
the calcite and silica. The specification for acid-grade fluorite requires at
least 97'5 ~~ CaF 2, and less than 1'5% Si02, with iron (as Fe203) below 0'5%.
The conventional reagent plan control" pH by the use of soda-ash or sodium
hydroxide. The depressants for gangue mmerals, notably calcite, are que-
bracho (a tannin extract from the bark of a tree of that name), dextrin and
sodium silicate. The collector agent is oleic acid or one of its modifications.
In some plants a mildly acid pulp is preferred. The frother is pine oil. Hard
water increases the difficulty of obtaining a freely break in g froth, without
which it is not possible to depress the last of the calcite. A further cause of
trouble is the freezing range of oleic acid, varying with commercial reagents
between 21 0 C. and 28° C. Unless the pulp is warmed well above this tem-
perature, a "frozen" froth tends to form, and selecting action in the froth
column is impaired. Reagents have been developed which modify the oleic
structure without loss of collecting power. With these mildly sulphonated
reagents, ordinary temperatures and untreated milJ water can be used. The
cool pulp also aids the specific depressing of calcite, and in a working plant
control of the depressing quebracho reagent can be closely made by maintain-
ing the frosty blue-white froth of high grade fluorite until the scavenging
section, when a faint pinkness warns that excess is being approached. Potas-
sium dichromate may be used to depress any sulphides not previously re-
moved by sulphydril collectors. The all-important factor for high-grade
concentration is a "lively" froth, in which the particles have not become
stuck together by frozen oleic acid or calcium oleate. Sodium silicate is
used as a dispersant in fluorite flotation.
Germanium
A method of recovery of renierite (Cu, Ge, Fe, Zn, GaMS,As)4 from its
flotation with copper sulphide has been developed at Kipushi.ii4 It occurs
finely disseminated in both copper and zinc sulphides in this deposit in a
concentration of 220 p.p.m., which become 600 p.p.m. in the copper roughing
float. This is passed through a magnetic filtering arrangement-the Ferro
filter-on which a fraction assaying 0'919~~ Ge is held. The germanium is
later abstracted by volatisation as fume during smelting in an electric furnace,
the final product assaying from 4 % to 9 %.
734 Mineral Processing-Selected Ore Treatments
Gold
General
In its ores gold is usually present as the metal, alloyed with metallic silver
and perhaps copper. Its high specific gravity (19·3) causes particles of gold,
even when of subsieve size, to settle readily from pulps in which the chief
gangue mineral is silica. Gold is malleable, and a grinding treatment which
breaks the gangue minerals may only flatten the metal without substantially
reducing the size of its liberated particles. This differential grinding effect,
by developing size differences, may assist gravity separation once the pulp is
clear of the grinding section. Against this, the weight and malleability of
gold particles causes them to be retained in a closed grinding circuit even after
they have been adequately liberated.
In addition to "native" gold, the element may occur in forms of associations
unfavourable to direct recovery by gravity methods. Metal sulphides such
as pyrite, pyrrhotite, stibnite, arsenopyrite, and galena frequently contain
inclusions of gold, perhaps as nodules only a few microns in diameter. Such
sulphides are called "auriferous". It is not always practicable to grind these
sulphides to the fineness required to liberate this finely disseminated gold,
nor would it always be possible to trap it efficiently once it had been freed.
The usual practice in such cases is to concentrate the gold-bearing sulphides
at a relatively coarse m.o.g., regrind them, and then extract the gold by chemi-
cal attack.
A third class of gold ore has its values combined in the form of telluride or
sulpho-telluride. These compounds are not malleable. They slime readily
and their gold content is extracted by chemical methods.
The process selected depends, therefore, on whether the gold can be freed
from its gangue at a sufficiently coarse mesh, or whether it is carried in a
heavy sulphide which can be similarly freed. If the ore contains other
valuable minerals, it may also be necessary to provide for their recovery. A
notable example is the flow-sheet used in a number of mills on the Rand,
where half the gold is recovered from strakes (a form of sluice) and the
balance by the chemical process of cyanidation, after which the tailings are
re-treated to recover uranium.
Another important factor must also be considered in commercial treatment.
Whereas in separating base-metal values the concentrate forms a significant
percentage of the ore, only a few pennyweights of gold per ton exist in the
average run-of-mine ore. Thus a large quantity must be handled in order to
extract a very small amount of extremely valuable concentrate. If all the
gold is carried in a small percentage of the ore, as is the case with auriferous
pyrite, considerable handling economies may be made possible by using froth-
flotation to separate this pyrite from the barren gangue at an early stage. If
part or all of the gold is disseminated through a siliceous gangue, all the ore
must be treated.
Gravity separation of gold is practised on strakes, shaking tables, and in
sluices and jigs. Amalgamation with mercury is used in connexion with this
work. Froth-flotation can be employed to remove gold and sulphide
minerals from a finely ground pulp.
Mineral Processing-Selected Ore Treatments 735
Ore
1
I ·I
DIrect
Gravity Concent.rat.ion
~ Chemical Attack
(I) Amalgamation -
1
(6) Cyanidation
Mixed Methods ~
~(8)+(6)
(6)+(5)
(5)_ (8)-- (6)
(6)- (5)-- (8)-- (6)
Modes of Occurrence
The process selected for separating any mineral from its associated gangue
must take into account its condition after comminution. Gold, silver, and
copper occur in nature in the free metallic state and also in chemical com-
bination. The modes of occurrence of gold are shown in Fig. 310.
(I) Much of what follows applies both to silver and gold. Usually the
native metal is a gold-silver alloy. Other elements and minerals may be
present.
Free gold occurs in alluvial and eluvial deposits as the result of the action of
weather upon the original ore, during long geological periods. Perhaps, as
with the Alaska beaches, a mere trace of the precious metal existed in the
original rocks, from which frost and tidal action have removed the lighter
sands, leaving enriched patches or beds of pay-dirt. Under these circum-
stances only minerals which are heavy, tough, and chemically inert remain
when the bulk of the gangue is weathered down and washed away. Con-
centration is therefore fairly simple, requiring only the use of gravity treatment
to exploit differences between the specific gravity of the gold (19'0+) and the
sands (2' 7 + ). If the surface of the gold particle is bright and clean it can also
be trapped by clean mercury, with which it combines physically to form an
amalgam. Amalgamable gold is said to be "free milling". It is possible to
make a gravity concentrate with no use of mercury, and to apply smelting
methods direct to this concentrate in order to produce bullion. When the
gold is not free milling, this is sometimes done. If the gold occurs in an
auriferous sulphide (e.g. galena) and the latter is worth recovering. the gold
'1-------
may be produced as a smelted by-product.
Gold - - - , - - - - - - ,
Combined (4)
r----'
Free (I) (tellurides)
(2)
I n siliceous gangue.
Clean Surface Tarnished made free b)
(free milling- Surface crushing
can be (may be too (then can be
amalgamated) badly coated amalgamated)
to amalgamate)
During the geological ages of exposure to wind, rain. sea action and chemi-
cal attack from salts of associated minerals. the particle of gold may have
acquired a patina or coating of some kind, such as oxidised iron. It retains
the high density of clean gold but can only be amalgamated if the tarnish is
Mineral Processing-Selected Ore Treatments 737
thin and discontinuous, or if scouring treatment has removed this film. Such
particles are sometimes called "refractory", though the term should be
reserved for occurrences of types (3) and (4) in Fig. 310. These films pro-
bably originate with the solution of iron from sulphides in the original ore bod y
from which the alluvium has been derived, and its redeposition as a hydrated
iron oxide around the gold. Particles of alluvial gold occur in many forms,
varying from large (and rare) nuggets down to specks only a few microns in
size. The shapes include threads. wires, plates. foils, and spongy gold, and
specific gravity varies between 12 and 19 according to the associated alloying
elements.
Much of the world's lode gold occurs in a simple siliceous gangue, from
which it must be partially or completely unlocked by crushing and wet grind-
ing before separation. If at the end of the grinding process the gold is free
fron gangue, clean and bright, it can readily be coated by mercury (amalga-
mated). Because of its simplicity and efficiency, amalgamation was the main
process used for concentrating lode gold before the cyanide process came into
use early in this century. If during comminution fragments of quartz are
driven into the malleable gold, amalgamation is more difficult, and losses may
occur. If the associated minerals reduce the chemical or physical cleanliness
of the mercury, the amalgamation process may be unsuitable for direct appli-
cation to run-of-mill ore. If the gold is so fine in grain that its complete
liberation entails very fine grinding, amalgamation may again be an unsuit-
able method, partly because of the cost of such fine grinding and partly because
very fine gold does not readily amalgamate. There are therefore several cir-
cumstances in which "straight" amalgamation is not the most efficient process.
The truly "refractory" or non-amalgamable ores are of two types. In (3)
above (Fig. 310) the gold exists as metal which during the formation of the
mineral deposit has precipitated from its solution in molten metal-sulphides,
and is now disseminated through these sulphides as minute specks. Ordinary
grinding of the ore might liberate these sulphides and might partially expose
the gold locked inside them. More frequently it is necessary to separate
auriferous sulphides from the siliceous gangue and subject these concentrates
to intensive grinding before the contained gold can be exposed. Ores of type
(4) are completely unamalgamable, as regards their tellurides and selenides.
Direct chemical attack upon the gold is the only known method for its recovery.
Concentration before Amalgamation
It was stated above that gold could exist in the ore either in a "free-milling"
or a refractory state. Both types can co-exist. It is, however, essential to
ensure clean surface contact (clean gold and clean mercury) if good amalgama-
tion is to be assured. Even if the gold can be superficially cleaned and scoured
by grinding, any associated minerals liable to upset the purity of the mercury
still remain. The selected treatment must avoid such contamination. Three
approaches are possible:
(a) Direct amalgamation of all the ore.
(b) Amalgamation of a selected concentrate, after bringing deleterious
material under control.
(c) Use of methods not involving amalgamation.
738 Mineral Processing-Selected Ore Treatments
Case (a)
Grinding 1000 tons at 2s. per ton £100
Bulk amalgamation at Is. per ton 50
£150
Case (b)
Grinding 1000 tons at 2s. per ton £100
Concentration to 0·1 % at 6d. per t o n . 25
Grinding and amalgamation at 40s. of the ton of corcentrate 2
£127
Process (b) shows a dennite saving. Suppose further that it is found that
this intensive grinding of all the auriferous concentrate improves recovery,
which with process (a) is40% and with process (b) 50%. The real comparison
Mineral Processing-Selected Ore Treatments 739
is between 4000 dwt. recovered for £150 (= 9d./dwt.) and 5000 dwt. recovered
for £127 (= 6d./dwt.). Iffurther savings resulted by lowering the grade fed to
the cyanide plant, the advantages in this case of preconcentration would be
even more striking. Broadly, an operation of good magnitude and continuity
will usually benefit by some measure of pr,econcentration.
Preconcentration Methods
Whether or no preconcentration is intentionally practised, it will occur in the
grinding circuit. An appreciable amount of metal may be locked up tem-
porarily in such places as the mill feed box, the liner joints, launders, and
beds of mechanical classifiers. This represents at best a tie-up of values and
at worst an opportunity for theft. Metallic gold should be removed from the
grinding circuit as soon as possible after liberation. The chief preconcentrating
devices used in the grinding circuit are:
(a) Jigs.
(b) Hydraulic traps.
(c) Shaking tables.
(d) Blanket, corduroy or rubber strakes.
(e) Mechanised strakes.
(f) Froth-flotation cells.
Strakes
The gold strake or corduroy table was described in Chapter 14. Strake
concentrates are high in value. and strict precautions are taken in handling
and transferring them. Spare cloths are not left lying around. Despite this,
the open-strake system, beside being monotonous, unpleasant, and laborious
to operate, is a security risk. There is the further objection to its use, that
trapping of values from the flowing stream is only efficient during the early
stages of deposition after clean up. As the sand beds down tightly, it becomes
harder to trap the desired particles, channels begin to form, and concentrate
is scoured away. Against these defects must be set the fact that a simple
740 Mineral Processing-Selected Ore Treatments
extract a very small amount of gold-rich heavy sand. Riffled sluices and jigs
are the chief appliances used in this work.
Amalgamation Practice
The methods used to bring the liberated gold particles into contact with
mercury are:
Some operators consider that a little sodium amalgamated into the mercury
improves recovery. The pool (b) has been made one terminal of an electric
circuit. Pressure has been used to force the gold into the mercury.
From 1860 to 1925 much of the world's gold was concentrated by surface-
contact methods. Plates were hung in the mortar boxes of stamp batteries,
placed so as to receive the issuing pulp, and set in launders. Mortar plates
were found unsatisfactory, as the splashing pulp scoured away both mercury
and amalgam. The addition of mercury inside the mortar box was also dis-
continued owing to excessive flouring. Plates, where still used, may be of
copper, silver-plated copper, or muntz metal. They are usually t" thick,
4'-5' wide, and 4'-12' long, the size varying with tonnage and fineness of
grind. If the liberated particles of gold are very small, the pulp must be
dilute and must flow gently in order to ensure seizure by the mercury. The
area of plate per ton per 24 hours varies between 0·2 sq. ft. and 3 sq. ft.,
chiefly in accordance with the thoroughness used to search the pulp. The
plates are inclined between It" and 3" per foot run, 2" being common. They
are mounted in such a way as to protect them from the jarring vibration of the
stamps, which would cause the mercury to pack hard and to be less receptive to
gold. A new plate may receive its first dressing with mercury into which a
little silver has been amalgamated. When correctly prepared or "dressed"
it is clean and bright, with its adhering film of mercury soft and not quite
fluid enough to start "weeping" down-slope. All discoloration is removed
from the plate metal, stains being rubbed with a solution of ammonium chlor-
ide, sodium hydroxide, or cyanide. Mercury is then sprinkled on to the oil-
free, clean plate. A "bottle" made from a short length of iron pipe, plugged
at one end and having a piece of cloth tied over the other, forms a good shaker.
Riffles can be scribed across the soft finished surface with a whisk broom to aid
arrest of gold. Pulp is allowed to flow at 10 %-20 % solids, just fast enough
to ripple gently and to move evenly. As gold is trapped, the amalgam
hardens and must be freshened and kept soft by the addition of more mercury,
care being taken neither to allow hard spots to develop nor on the other hand
an excessive fluidity. A mercury trap may be used at the tailings end of the
plate to catch any detached amalgam (Fig. 312).
742 Mineral Processing-Selected Ore Treatments
From time to time the plates are dressed. Pulp flow is stopped and the
plates are washed clean. The amalgam is then worked upward, new mercury
being added if necessary to soften it. A rubber or hardwood scraper is used
to remove amalgam. Stains are dealt with, new mercury is put on, and work
restarted. The time interval between dressings depends on the richness of the
ore and on the rate at which the plates begin to show staining or "sickening"
of the mercury.
Immersion contact, in which either an aqueous pulp or the dry ore is forced
through a bath of mercury. has been developed in several ways, but its use has
not become widespread. Gold is more readily wetted into a soft layer of
amalgam than into a pool of mercury, and contact between the two metals is
more thorough when the particles roll over' plates than when a separating
film of air or dust obstructs the union. An ingenious use of pressure amal-
gamation was at one time made on alluvial black sands. 55 A cast-iron bowl
flaring in diameter from 17" at the bottom to 23i" at the top rotated at 150
Mineral Processing-Selected Ore Treatments 743
r.p.m., under which conditions a t" thickness of mercury built up on its walls,
forming a layer of 8+ sq. ft., weighing 100 lb. The speed was adjusted so that
even rusty and greasy gold stayed in the machine while the associated heavy
black sands worked up and out.
The most common method of amalgamation in use today depends on grind-
ing contact. It is applied intensively to concentrate produced by gravity
methods. Its use in an amalgamation barrel is described in the next section.
The older methods, in which mercury was ground together with the whole of
the ore, sometimes led to serious working losses. These are largely avoidable
when the concentrates are gathered into a small bulk and treated for removal
or neutralisation offouling substances before any mercury is added. The Patio
Process has been described by Rickard. 56 Pan amalgamation in the clean-up
pan (Fig. 313) is usually applied to a concentrate.
r- SIt - - - -
Toward the end mercury was added and during the last few minutes water
was flushed through. Further batches of ore were added for some days, after
which the amalgam and any sand not displaced during flushing was retrieved
and panned. The track was then repacked with clay between the stones and
the work re-started. 57
Handling Mercury
The quantities used, and the intermittent way in which they are handled,
make it unlikely that the workers in an amalgamation plant will be exposed to
serious health hazards. The subject is covered in various official publications
which deal with the large-scale use of mercury. 59 Chronic poisoning develops
by insidious stages. Poisoning can result from exposure to mercury vapour
in the atmosphere or from mercury spillage finding its way into the worker's
system. Mercury should be stored in tight flasks and not left where food is
eaten or stored. If spilt, it should be covered with water and immediately
picked up. It should never be removed by blowing with pressure air since
this might cause small globules to become airborne. Workers should wash
their hands thoroughly after handling any mercury-containing materials.
They should rinse their mouths before eating or drinking, and should wear
special clothing in the work sheds. This clothing should be washed weekly.
Process Control
Process control in the ore-treatment plant has two aspects, financial and
technical. In the case of gold both are complicated by the high value of the
concentrate and the heavy density of gold, which hangs up in any quiet, rough
place in the flow-line. Ideally, a known tonnage of ore would be sampled and
assayed with good accuracy, and the various products would be similarly
checked in order to produce a metallurgical balance sheet. This might be
depicted as in Fig. 315.
Mineral Processing-Selected Ore Treatments 747
Feed (I)
Straking (2)
Conc. (3) - - - - - - Amalgamation - - - - - - Tails (4)
etc.
Suppose the feed (I) to be WOO tons assaying 5 dwt., then concentrate (3)
might be I ton assaying 2000 dwt. The equation for metallurgical balance
(Ff = Cc -l- Tt) would read:
1000 tons at 5 dwt. I ton at 2000 dwt. + 999 tons at 3 dwt.
= 5000 units = 2000 units = 3000 units
(I) (3) (4)
In the rough and tumble of milling it is not possible to obtain a neat equation
of this kind. Consider a developed flow-sheet (Fig. 314) for a plant using two
stages of wet grinding, each with gravity extraction of its "metallics", rusty
gold, and auriferous pyrite. Items (3) and (8) may consist of jigs, flotation
cells, shaking tables, strakes, etc. The concentrates they produce are closely
guarded and handled under conditions which provide a high degree of security.
In order to bring this flow-sheet into metallurgical balance we must be able to
equate the total pennyweights (or other units) of gold at certain key points,
thus,
Owts. (I) = Owts. (6(-»+Owts. (15)
The difficulties in obtaining a balanced equation of this sort are almost
insuperable. Since accurate knowledge is the foundation of good control,
this unfortunate fact is a challenge to ingenuity, not an inevitable disability.
Although perfection in this important matter is not yet possible, a great deal
can be done to obtain valuable information at a reasonable cost. If the f1ow-
sheet is now considered in detail a good idea of some of the practical difficul-
ties confronting the mill superintendent emerges, and the flesh of day-by-day
operation begins to clothe the skeleton of basic principles thus far discussed.
At the same time the matters requiring technical control are noted.
New Feed.-For accuracy, the following information is needed:
(a) Tonnage (dry).
(b) Assay value.
(c) Size analysis.
(a) and (b) show the total value delivered by the mine, for which the mill
accepts responsibility. (a) can be known with fair precision, if the feeding
arrangements into the scoop box (2) are properly calibrated and the moisture
of the ore is checked. Alternatively, the tonnage drawn daily can be arrived
at by checking the weightometer reading at a fixed hour, and compensating
for tonnage rise or fall in the ore bin. With item (b) the difficulty begins.
Gold (S.G. < 19·3) is distributed through gangue (S.G. >2·65). The gold is
748 Mineral Processing-Selected Ore Treatments
!
Primary Mill ( 2 ) - - - - - - - -.....
!
Cone. - - - - Concentrating Device (3) - - - - Tail
Cp Tp
I
~
Pump (4)
~
Ef"==Classifier (5) +
+
Tail
~Classifier (6)~(To n~xt
~ sl'ctlOn)
l
Conc. ----Concentrating Device (8) ---J--Tail
~ Th
I
• I
Amalgam
I
..
Amalgamation Press (12)
!
Mercury Retort (13)
!
Bullion Furnace (14)
.!
Bulhon Safe (15)
Fig. 316. The Amalgamation Flow-sheet
Cp Tp Primary }
Cs Ts Secondary Concentrate or Tail
Ct Tt Table
Mineral Processing-Selected Ore Treatments 749
finely divided and locked into pieces of ore varying in size from, say, t" down
to a few microns. If the ore is "spotty"-i.e. if it has its gold irregularly
distributed (like a very few plums in a very large cake)-a big sample must be
taken in order to achieve some approximation of accuracy. If a 10% sample
were used on a 1000 tons/day feed, its reduction to a bulk suitable for the
assay office would be too costly. Yet, as mathematical analysis shows, only a
large sample, ground fairly fine, can be made to yield a representative assay
sample of a few pounds weight. Some psychological resistances to accuracy
complicate an already formidable technical problem. The mill is judged by
the percentage recovery, and that judgment ultimately resides with a partially
informed and distant Board of Directors. It is tempting to give the mill the
benefit of every doubt, real or imaginary. The stope assay plans, since they
suffer from most of the disabilities which affect the mill head sample, are
sometimes adjusted by a "milling factor". Again, moisture figures might be
too low after a head sample had dried between collection and test. A wrong
assay return can arise from difficulty in reducing the size of the original
sample. However much care is taken, the head value as calculated from
sample is unusually open to error in the treatment of gold ores. Better
reliance can be placed on the figures obtained toward the end of the flow-
line, and the head assay usually provides only a somewhat unreliable secon-
dary check. As technical control improves in the mill, this state of affairs
will gradually disappear. Its existence will be indignantly denied by some
metallurgists, but the author has been at pains to verify the facts. In the case
of finely disseminated and evenly distributed gold values such as are charac-
teristic on the Rand, the head sample is closely representative of the grade. It
then forms a reliable integer in the equation of metallurgical balance. Here,
as in some other respects, the Rand "bankets" provide a model of milling
docility rarely met with in the ordinary lodes.
Primary Mill (2).--This provides numerous traps in which metallic gold
can lodge. Control is based on the concept of continuous flow accompanied
by continuous separation into a poorer and a richer fraction, thus:
Entering Pulp = Richer Product + Poorer Product,
all in terms of rate of flow and pulp consistencies. Should any constituent of
the ore become static, the rate of flow ceases to be representative. If, for
example, a serious amount of gold hung back in the mill and was released
during occasional surges, any pulp sample would be unrepresentative. If
cut during such a surge it would be too rich, and if between surges too poor.
Over a long term the control data would be reasonably accurate, provided the
trapped gold were all recovered and brought safely to the bullion safe.
Since the feed box to the mill and the joints between the liners provide
lodging points at which gold is retained, safety precautions must be taken to
prevent theft when the mill is stopped, or is being relined.
Concentrating Del'ice (3) and (8).-This is designed to be appropriate to the
technical condition of the desired product. If the gold is sufficiently coarse,
in association with a 4- 60 mesh gangue, Denver jigs or hydraulic traps pro-
vide a compact and thief-proof method for trapping out liberated gold. If
below 60 mesh, shaking tables or flotation cells can be used, the latter being
750 Mineral Processing-Selected Ore Treatments
should occur in this section. The metallurgical balance can be struck with
good precision:
Bullion (15) = Ct - Tails (II).
The above general picture shows the strength and weakness of control on
this very vulnerable part of the gold mill. Good management, maximum
mechanisation, and the employment of a few well-paid shiftsmen rather than
a large number of "cheap" labourers are the best safeguard against loss,
provided a good basic flow-sheet is being operated according to a strict
technical routine. Samples at key points show the assay value, degree of
liberation, and solid-liquid ratio.
In many plants the gold is milled not in water but in cyanide solution. In
this case the problem of economic and technical control in the grinding
section becomes far more complicated, since gold-bearing solutions are
withdrawn.
Flotation
Where possible, gold is removed by gravity, amalgamation and/or simple
cyanidation. Theonlyflotation in such cases might bethatof associated graph-
ite which otherwise could upset amalgamation or reprecipitate values from the
pregnant solution, with ensuing loss. Such graphite is also dealt with by
ignition (Ashanti) and by surface closure and passivation, using oil. One or
two mills use activated charcoal to precipitate the gold contemporaneously
with its dissolution by the cyanide, and this auriferous carbon can be recovered
by flotation.
The high density of gold (19,2) makes it hard to hold in a flotation froth,
though this is by no means impossible. A further difficulty is that even an
extremely rich gold ore contains less than 0·0002 % of metal, which is quite
inadequate for stabilising a froth. For gold occurring in a siliceous gangue
flotation would therefore be unsuitable. If, however, part or all of that gold
is locked in an auriferous sulphide, sulpharsenide, telluride, etc., a very differ-
ent picture is presented. Such gold frequently exists as inclusions only a
few microns in size, and could not justify the cost of the extremely fine grinding
which would be needed to expose it. If all the gold-bearing sulphide can be
floated, the exposure of gold in a regrinding treatment is relatively cheap.
Alternatively, the float can be sweet-roasted (by the FluoSolids process or in
roasting furnaces), the residue being cyanided.
In gold flotation the end-product is so valuable that no risk of loss is taken.
Powerful collectors are used, so that every particle tending to float is captured.
The bubbles may be stiffened by the use of mineral oil to reinforce the frothing
agent, if the natural armouring of mineral particles is insufficient. Mechanical
cells are used, and the bubbles are mostly small, providing a stiff mat through
which a floated gold particle finds it hard to slip back to the pulp. As a
result of so severe a float, the cyanicides which might otherwise run through
to the cyanide plant are lifted in the float, which can contain a complex
mixture of sulphides. If this is now ground intensively to expose the gold in
the auriferous particles, it becomes oxygen-avid, slimy and difficult to thicken
or filter. Hence. the cyanidation of a flotation concentrate is more difficult
752 Mineral Processing-Selected Ore Treatments
and costly than is that of a straight pulp. Far more intense aeration is re-
quired, and the addition of protective alkali must be closely watched, as the
sulphides will not only denude the pulp of its oxygen but will release sulphuric
acid as a result. Such a pulp can be pre-aerated in water to reduce these
troubles, after which it may be run to join the flotation tailings proceeding to
cyanidation (Fig. 317).
During pre-aeration ferrous salts in the water are largely oxidised to the
ferric state. If an appreciable amount of dissolved (ferrous) iron remains it
may be necessary to discard the water used. If sulphides oxidise to sulphites
and sulphates and are left in the pulp, they react with cyanide to form thio-
cyanate. Unoxidised iron may also produce ferrocyanide, again wasting
reagent. Soluble sulphides can be precipitated with litharge or lead acetate.
The pulp should be aerated at a cyanide strength of between 0·02 %and 0·05 %.
Grinding
t
Conditioning
r t
(C) .....- - - - Flotation -----+> (T)
Regrinding --------,1
Pre-aeration
I
To Cyanidation
Sulphur dioxide also surface-cleans the particles and sulphatises the hy-.
drated iron in the pulp.
Among the reagents used for the first two types of ore are Aerofloat 208,
with a cresylic frother such as Aerofloat 15 or 25. Xanthates (butyl or
amyl) are also employed, either alone or with the Aerofloats. Where a little
sulphidisation may help, Cyanamid 404 with Aerofloat 242 may be used.]
Froth may be stiffened by the addition of lubricating oil in the grinding circuit,
but a more controllable procedure is to use cresyl ahead of flotation, perhaps
with pine oil or one of the proprietary frothers based on the higher alcohols.
Pulps are kept thick (30% + solids).
The reagent plan for the third type of ore must conform to the requirement of
the associated metal sulphide being floated. ]f free gold co-exists, the pH
control should be made with soda-ash, not with lime. The periodical techni-
cal liter;>ture is rich in descriptions of gold-milling practice.
Working Precautions
The cyanidation of gold (and silver) was considered at the basic level in
Chapter 16. Selected flow-sheets in which the process is used, alone or in
combination with other methods, are given below.
Mercury, cyanide and prussic acid gas are deadly poisons if absorbed into
the system in more than minute quantities. Mercury is most dangerous
when present in the atmosphere as the result of leakage during the distillation
of amalgam. Good ventilation, meticulous care in sealing the retort, hood-
ing of furnaces, extraction of fume from danger points, and care in handling
liquid mercury are the best safeguards. Men working with mercury should
wash and change their clothing before eating, and food should not be allowed
in the working places. This applies also to cigarettes or any other thing which
might serve as a vehicle for the intake of condensed poisonous fume. ]n
addition, periodical medical examination is desirable.
Soluble lead salts-nitrate and acetate-are normally used in the precipi-
tation plant. These are both quick and cumulative poisons. They can be
picked up on the skin or from clothing, if lead solutions are handled carelessly.
They are unlikely to be inhaled, however, as the lead is not atomised during
its use. Cleanliness and periodic medical examination are desirable, as with
mercury.
Cyanide can be dangerous in several ways. "Cyanide rash" affects some
people seriously, and such individuals should not be employed where cyanide
is used. For normal workers the chief danger with the very dilute solutions
used is chafing of the flesh by wet clothing. Where this carelessness has
caused skin irritation, the remedy is a I %solution of lime chloride and boracic
acid, allowed to dry on the affected part. If accidental contact with cyanide
has been made, the affected part should immediately be washed and moistened
with either a 2t %solution of Epsom salts or of 2t %acetic acid.
Prussic acid gas is deadly in its effects, and takes less than a minute to kill, if
breathed in any quantity. Here, even more emphatically than with the other
poison-hazards in the gold mill, prevention is better than cure. Caution in
working round piping which may have accumulated a pocket of gas, liberal
ventilation, and extraction of fume are important. Swift remedial action is
Mineral Processing-Selected Ore Treatments 755
vital if gas has been inhaled. The symptoms are throat irritation, breathing
trouble, watering eyes, heaviness, headache, dizziness, an irregular pulse,
pallor, and lastly unconsciousness leading to death. These develop in this
succession in a matter of minutes.
If breathing has stopped, artificial respiration should be started and con-
tinued till the doctor's arrival. Oxygen may be administered, and a capsule
of amyl nitrite may be crushed and held close to the patient's nose while
j
Waste
1
Crushing,
Screening
Slimes Washing
Waste Sorting
l
(2)lI
Sized Grinding
Pebbles
Fine (3) (Slimes)
Ore Bins
Mine water
r- - - - --- - - - - --
1
~Grinding,
: Classification (4)
1
Storages ~ - - - - - - - - - - Thickening (5)
~
Agitation (CN) (6) +--"',
_.:
- -,
r·
,~---
I
1 I
:
L----------+Thickening (7) :
Filtration ~ - - - - - -." ,--- ----/I ~,
: Barrens
l '.
oj. Storage
TaIling Clarification (8) ----~, :-----~:
__
D_aw_ Precipitation
Bullion
Production
Crushing to - 1"
1
1 Grinding, + - - - -,
Amalg. Bpi,.'
L : :
Bullion: Filtration' "_:J
~ I
Precipitation, - - - - - - - -', " -I - - - - - ~
etc. -. 1
1
I
I
Grinding
.1 .
j
Agltatlon
I motn,
(Tails) (Conc.)
I
1l
Cyclones I Roastingl
1
(+) (-) 1.
1 1 Arsenic
Underground
(Back Fill)
Dam
-
Fume
--- 1
Cyanidation
Residues to
Stockpile
Crushing to - in
t
Ball Mills
t
Classification,
1 t
~ pebbleMilling,~
and Jigs
t
Jig. Thickening
Conc.
Pre-aeration
1
-----,Jl
De-watering
(Filtration)
Cyanide
Cyanidation
(Pachucas)
Hollinger starts its 97% recovery with grinding in cyanide to 60% -200
mesh. Here 45 % of the gold is extracted. Classifier overflow goes to 90
double-deck shaking tables which remove one sixth of the feed as a pyritic
concentrate carrying four fifths of the residual gold. This is ground to 80 %
- 325 mesh, thickened to 55 % solids and agitated for 24 to 36 hours in Dorr
Mineral Processing-Selected Ore Treatments 759
machines, the tailings then joining the thickened table tailings for 16 hours in
Pachucas. Residual solids are split between filtration and C.C.D. (Fig. 321).
Kerr-Addison mines two types of ore. One is a quartzose stock-work in
the carbonate zone and yields a 99 %gold recovery. The other has auriferous
quartz and pyrite in fractured lavas and gives below 95 %. In order to avoid
trouble in the precipitation section, due to nickel, copper and arsenic, half a
ton of foul cyanide is discarded for each ton of ore treated. Nickel must be
kept below 0·2 lb./ton. Feed to the roasting section is at 77% solids (Fig.
322).
Grinding and +- - - - - - - - - - ""I
Classification
1 '
I
I I
(crre
Shak;ng Tables
(Tail)
Classifier,
Ball Mill Thickening
TlllCkLg,
Dorr Agitation
1
Pachucas
"Balen"
Solution
1 :
I
• I
~
I
DecantatIon - - - - - - - - - - -, I ~ - - - - ~,
Filtration' '-:-/ I
1
Repulping +- - - - - - - - - - - -',
! I
1
,~- - -l
'!':
I - -
1
FI'ltratlon
' - - - - - - - - - - - -, ,I r I
,,:_1 - - - - ~
I
,
Tails "tt
PreClpl atIon,- - - J'
-- etc.
Fig, 321. The Hollinger Flow-sheet 1957
760 Mineral Processing-Selected Ore Treatments
Crushing
1
r - - ... Grinding------.
,-----___lj __ T1U.J.ning
:II i
I
I
j
Preis to
r-
I
I Desliming
(eydon,,)
Slimes
ll"'"------Flotation
Tails
to Mine cone. (Sulphides)
Ro.Lng
(FluoSolid)
~calcine
Fig. 322. Kerr-Addison Flow-sheet 1958
Mineral Processing-Selected Ore Treatments 761
Comminution
!
l,---------- Jigs - - - - - - - - - - ,
(Cone.) (tail)
To Refinery
I
~
Classification (+ ) _ _ _ _ _ _ _....J
t
Lime,
Cyanide ----,1 I
Flotation
Regrind,
--------..1
Tail
Cyclone Classn.
t'------ t -----------,1
Cyanidation
Tail to Refinery
Dam
Mclntyre's gold occurs coarse and fine in quartz and as auriferous pyrite.
The flow-sheet (Fig. 323) uses a sequence jig-flotation-cyanidation.
At Lake Shore 20% of the gold is associated with tellurium and requires
very fine grinding. This is done after flotation of the first cyanide tail (Fig.
324) an operation in which similar tails from two adjacent mines are treated.
762 Mineral Processing-Selected Ore Treatments
Ore (-f)
commin1ion and
l
,. - - - - - - - - - - - - - -+
Classification
Barren
Solution
1
Rake
Agitation
I
~ - - - - - - -Tray Thickening
P'~gs.
: Agitation
Precip1tation,
etc.
j
I
I<- - - - - - -
I
_1 - -- - - -- -- -Filtration
Cyanide
r - - - - - Tails from
1 adjacent Mills
Flotation------,
RoJing
. jjLiffi'
~- - - - - - - - - - - - - - - - - - ... Cyanidation
1
Tail-------'
t
Comminution
I
Na2 COa
1 . . - - - - - - R.301
CuS0 4
and Classification
Flotation
! "..---- R.301, Pine oil
Tail to
1
Float to Roasting
Cyanidation, re-grind, tabling,
etc. cyanidation, etc.
Graphite
This form of carbon occurs both in coarsely crystalline form and finely
divided, with no clear dividing line. The usual treatment starts with a pulp
of about pH 8, made alkaline with soda-ash, and aided as regards dispersion
and selectivity by sodium silicate. The collecting agents for this readily
floated mineral are paraffin or diesel oil, perhaps aided by pine tar oil. The
frother is pine oil. Pine oil alone may suffice with finely ground graphite.
When clean coarse flake can be liberated by primary grinding in a rod mill it is
removed by screening. If the product is insufficiently clean, it can be up-
graded by some such treatment as is shown in Fig. 326.
In the Kaiserberg mines in Styria two types of graphite schists occur. One
yields microcrystalline scales, the other soft and plastic carbon. Associated
minerals include mica, sericite, chlorite, quartz, plagioclase, hornblende, tremo-
lite, limonite, and small fractions of other undesired minerals. Hand-sorting
and air classification are used to produce a 66 % + carbon product. This
contains too many abrasives and other impurities for special uses and is
pulped, ground finely and floated at pH 8 to upgrade it to 90 % --t carbon, the
main impurity being a mica which does not impair its lubricating quality.
Coarse flake commands the best market price, so care is taken to avoid
overgrinding. Impact crushing may be followed by screening, with tabling
of the screen undersize to produce a coarse flake, tailings and a middling
which goes, together with the screen oversize, to rod milling. The mill
discharge can then be classified, undersize being floated and oversize tabled.
764 Mineral Processing-Selected Ore Treatments
Feed
1
Rod Mill
(screen
1
(+) I (tail) - - Unit Flotation j
~ t
.. - - - Classifier - ( -)
(+) .-- I Cell
Coarse
GfaPI1lie
1 Conditioning and~
Flotation t
Tail Fine Graphite
(for shipment
or gravity upgrading)
Iron
The main concentrating processes are based on the degree of natural ferro-
magnetism in the iron minerals, or on the extent to which this can be developed
by reduction roasting. Gravity concentration is also extensively practised
and there is large-scale use of the flotation process both for pyritic ores and
for those of low ferro-magnetism. Sulphide iron may be floated in connexion
with its use in producing sulphuric acid, or for recovery of associated gold.
The machines used were described in Chapter 19.
Reduction roasting at research, pilot and operating level was reported in
six Papers in the 1963 I.M.P. Conference. Roast-reduction of hematitic
quartz was shown as a topochemical process in which reduction proceeds
evenly inward from the exposed surface. 63 Concurrently the quartz matrix
is micro-cracked, thus reinforcing the thermal stress in the heated mill in
which the ore is being autogenously ground at some 600° c., and the iron,
now substantially reduced to magnetite, is liberated. Non-magnetic Swedish
taconites often contain iron too finely divided for economic liberation by
Mineral Processing-Selected Ore Treatments 765
which rough out a middling and discard fifty tons of tailing assaying 7-!- %
Fe and 2 % P, the phosphorus content of the entering ore being 1 %. The
middling goes to three-turn spirals for cleaning, its tailing being returned to
the rougher spirals. Here 30 tons per hour assaying 63 % Fe are recovered.
Spirals are worked at a pulp density of 25 % to 30% and splitters are set by
calibrating gauges. One man operates the whole spiral section in which
each machine has a capacity of 1-!- tons per hour and stands in an area of one
square yard. In Canada's Carol concentrator spirals are used on an ore
which is liberated at 14 mesh. Aerofall mills work in closed circuit with 14
mesh screens71 and the undersize is classified into a coarse product, about a
splitting point of 150 mesh. Each split fraction goes to rougher-cleaner-
recleaner spirals. An unusual de-sliming concentration is practised with the
Moose Mountain ores of Capreol, Ontario, which must be ground substan-
tially through 325 mesh for liberation before producing a 66-!- % Fe concentrate.
After a first-stage magnetic concentration of a rod-mill discharge the con-
centrate is finely ground and given a second magnetic treatment. The con-
centrate from this is re-ground, magnetically flocculated and sent to hydro-
separators from which the dominantly siliceous non-settling fraction over-
flows. The operation depends on automated sensing of the separating line
between the silica with its density of 2·8 and the zone of flocculated magnetics
at S.G. 5'2.72
In one Nevada flow-sheet the minus t" magnetics are recovered by dry
treatment and the minus 48-mesh screened undersizes by wet magnets. 73
Treatment of a quartz-siderite ore assaying 20 % to 28 % Fe in Hungary by
roasting in a reducing atmosphere using natural gas, followed by dry mag-
netic separation has been used in conjunction with recovery of copper and
barite from the ore. 74
Iron ore flotation was pioneered in Michigan in 1954. At Humboldt
where a cherty hematite and magnetite is treated. Minus 65 mesh classifier
overflow is deslimed and thickened by cyclones, with rejection of some 3-!- %
of the mill feed and 2-!- % of the iron. The pulp is conditioned at 70 % solids
with 0·005 lb./ton of Aerosol, 1t lb. of tall oil and then floated. Concentrates
are filtered, stockpiled and drawn to the re-grind section for reduction to at
least 75 % minus 325 mesh, a size essential for formation of a robust "green"
ball suitable for pelletising. The need for integration of flotation treatment
with pellet production causes the latter to dominate procedure. At Cleveland-
Cliffs a rougher concentrate produced by flotation with 1'25 lb./ton of a tall
oil fatty acid in Fagergren cells is made, at an average grade of 61·7 % Fe. 76
This is re-ground to at least 80 %minus 325 mesh, to satisfy the requirements
for pelletising. It then passes, without further reagent addition, through a
series of four conditioning tanks in which it is progressively heated above 98°
C. The fatty acids present migrate selectively to the hematite. The effluent
is diluted to between 32 % and 34 % solids and re-floated, iron recovery being
nearly 98 % and grade 67 % Fe.
partially oxidised sulphide, the sulphate (anglesite, PbS04 ), and the carbonate
(Cerussite, PbC0 3).
Galena usually occurs in association with sphalerite and other sulphides,
from which it is differentially floated. Some complex treatments are dis-
cussed later. The mineral floats readily with aerofloat or xanthate, in a pulp
made alkaline with sodium carbonate. It may be depressed by lime or by a
pH exceeding 10-4. Potassium dichromate forms a non-reactive coating of
lead chromate, and is sometimes used to depress lead from a bulk float. If
Aerofloat 25 or 31 is used, little or no pine oil is needed as a frother. Where
galena has become tarnished the sulphidising action of R.404 (Cyanamid)
may aid collection. With a calcite gangue the adherence of slimed gangue
may occur and cause loss unless a dispersant, sodium silicate, is used. If
oxidised lead is to be floated, sodium sulphide is commonly used to produce a
surface attractive to xanthate collectors. As soluble sulphides are strong
depressants for clean lead and silver sulphides, theses minerals must be removed
in an earlier flotation operation or they will probably be lost. Since excess
sodium sulphide is in any case a depressant, starvation quantities should be
added cell by cell in the second flotation treatment. The surfacing effect
produced by sulphide addition is transient, and the froth should be removed
speedily to avoid reversion of the mineral surface to its oxidic state. The
use of copper sulphate to control and stabilise the newly sulphidised mineral
particles is favoured by some operators. This is done by two conditioning
stages, starting with sulphidising treatment and followed by the use of the
copper salt.
In controlling the flotation of lead carbonate, a quick vanning test on the
tailings can be made by adding a few drops of sodium sulphide to any whitish
heavy mineral, which will turn brown if it consists of insufficiently sulphidised
lead.
With most oxidised lead ores it is good practice to take gravity concentra-
tion as far as possible and thus to reduce the production of slimed values.
Sulphidising agents are used immediately before flotation, with sodium bi-
carbonate. In one Moroccan mine where Hancock jigs pre-concentrate the
flotation feed of a cerussite-limestone ore, considerable improvement was
made by cutting down conditioning time and amount of reagents (hydro-
sulphide and amyl xanthate) and intensifying aeration during flotation.77 A
special impeller was developed, with rotor bars but no stator or deflecting
baffles. Since excessive reagent consumption occurs in hot weather condition-
ing time is cut down at such periods. Aeration in the conditioner is sometimes
helpful, the reason not being yet known. Alkaline earths are liable to be
harmful since they tend to form coatings of insoluble carbonate on cerussite. 78
By using relatively cheap barium sulphide instead of sodium sulphide, and
adding sulphuric acid in the conditioner deep enough to prevent escape of the
H 2S evolved one Sardinina operation has improved its metallurgy, the pH
being controlled between 6·8 and 7·2.
Lithium
Lithium is mainly extracted from spodumene, which has a theoretical lithia
768 Mineral Processing-Selected Ore Treatments
Manganese
Concentration methods include hand sorting with cobbing (in South
Africa), DMS, jigging and tabling for manganese minerals sufficiently liberated
at a coarse size for marketing. Where flotation is used the concentrates must
reach the specifications required for use in furnaces-48 % Mn, low iron and
phosphorus and below II ~~ silica plus alumina after pelletisation. Three
main flotation types of ore are those with a high calcite gangue, those inter-
mediate in calcite and silica, and those with a siliceous gangue. With the first
two, the calcite is floated with from I to 4 Ib./ton of an oleic collector, the
pyrolusite or manganite being depressed with t-3 lb. of dextrin. With the
third type, if non-clayey or slimed, the thickened pulp is conditioned at
about pH 7 with S02, tall oil is used as collector and the manganese is then
floated after some dilution.
An unusual agglomeration-float of wad has been described 88 . Wad is a
complex of manganese oxides which may include psilomelane, hollandite,
coronadite, and pyrolusite. This slimes during grinding and part of the
flotation problem is concerned with recovery of the slimed manganese. At
Three Kids the main gangue minerals are quartz, kaolinite, montmorillonite,
calcite, gypsum, and barite. The specific surface of this material can reach
58 m. 2/g., and reagent consumption is correspondingly high. The flow-sheet
(Fig. 327) requires intense mechanical stirring during conditioning, to break
Comminution
1 jS0 2
Tall-oil soap
Petroleum oil
Intense mechanical Petroleum sulphonate
conditioning
1
Roughing and 4 stages
of cleaning
Mercury
The main mineral treated is cinnabar, from which the mercury is usually
extracted by direct distillation. One Californian producer that has worked
a 12 Ib./ton deposit intermittently since 187089 feeds minus I-V ore to rotary
furnaces down-sloped f' per foot, and turning I r.p.m. Heating is by fuel
oil, and the mercury-laden gases leaving at between 260 0 C. and 315 0 C. are
drawn by suction fans to cyclones and condensers. Condensates are collected
daily and batch-treated by rabbling with quicklime, when 80 ~~ of the mercury
drains away. The rest is recovered by batch distillation. Mercury from
both operations is next filtered through a bed of quicklime and sold as "prime
virgin" mercury, 99·9 % or more pure.
Pre-concentration by gravity methods or by flotation may be used with a
complex or lowgrade ore before retorting. Slimes and clays must be dis-
persed prior to effective flotation, but sodium sili!:ate and sodium carbonate
have proved unsuitable dispersing agents, as they exercise a depressing effect.
An organic chemical based on lignin (called Palcotin), has been successfully
used in amounts varying from 0·04 to 0·1 Ib./ton. pH is not critical, but
should be held between pH 6 and pH 8. The cinnabar readily accepts surface
activation by ions of lead or copper. A secondary butyl xanthate (Reagent
301) and Frother 65 are used to complete the work.
Mica
Molybdenum
In a number of plants molybdenite is recovered as a by-product bulk-
floated with sulphide copper, in which from 0'3% to 2 % may occur. In a
typical operation in Arizona 90 rougher concentrates are reground through
200 mesh and cleaned. The clean concentrate is thickened and steamed,
ferro-cyanide being added to depress the copper. Three stages of flotation
are then given, following which the roughed molybdenite froth is reground
and recleaned.
The major molybdenite operation is that of Climax, Colorado, where the
main selling product is this dominant mineral. Advantage is taken of the
strong floatability of MoS 2 in a flow-sheet (Fig. 328) which discards most of
r--
I
Comminution
Primary Grind
I
Primary Float
~d
~ondary Grind
I ~ S:~ondary Float - - - - -
Mid 1 Tail
I T'::r~I~:Fl-
Mid
Mid
I • Final Float
1
Molybdenite
the gangue at a coarse grind. The ore body is a silicified altered granite,
with its molybdenite finely disseminated in stringers and thin veins. Minor
quantities of wolframite, monazite, and cassiterite are recovered from the
flotation tailings by gravity methods. Because of the flaky nature of molybde-
nite it is possible to float particles consisting mainly of quartz gangue, using
oil only as the collector, aided in its emulsification by a sulphated mono-
Mineral Processing-Selected Ore Treatments 771
glyceride, and a little pine oil. Cyanide and soda-ash are used in the cleaning
stages at pH 8'3, to keep down traces of pyrite and chalcopyrite. Primary
grinding at 43 % + 1OO~ is followed by a 25 to 1 concentration in the roughing
section, yielding a 9 % MoS 2 concentrate. This is thickened and given three
further stages of grinding in pebble mills, with staged flotation. Ninety-eight
and a half per cent recovery of high-grade molybdenite is maintained. Tail-
ings are sent to a by-product plant where tin and tungsten minerals are
recovered by treatment in which gravity, flotation and magnetic separation are
used.
More usually molybdenite is bulk-floated with copper sulphides, which latter
are then depressed. The method used at Morenci has been described by
Papin. 9a Thickened bulk concentrates are dewatered, repulped with fresh
water, and brought to pH 7'5 with sulphuric acid. The thiophosphate collect-
ing agent is selectively removed from the copper mineral by sodium ferro-
cyanide, the effects of which are short-lived. Flotation of the molybdenite is
performed after brief conditioning, and the rough concentrate thus produced
is given further grinding and cleaning, a little sodium cyanide and sulphide
being used in this work to aid further additions of ferro-cyanide. In other
mills soluble starch has been used as a depressant for molybdenite collected
with xanthates, but this did not work with the thiophosphate collector used
by Morenci. At Bingham, Utah, steam distillation removes decomposed
xanthate selectively from copper, the molybdenite being refloated after this
treatment. The general pattern (Fig. 329) is:
1
Destruction of Copper
Collector
Smearing ----,..
Collector Molybdenite Flotation ------,1
(saturated petroleum) I of
~ (Tail)
(Float) Cu Sulphides
1
Molybdenite
Other methods used include depression of the copper with sodium hypo-
chlorite in an alkaline pulp together with sodium ferrocyanide, fuel oil and an
alcohol frother being used to float the molybdenite. Copper minerals can
also be depressed by phosphorus, arsenic or antimony salts. The float from
such differentiations assl;lYs from 3 % to 30% MoS 2, and is brought up to
772 Mineral Processing-Selected Ore Treatments
Nickel
The main source of the world's nIckel is the copper-nickel sulphide ores
typified by those mined at Sudbury, Ontario. The principal sulphide is
pentlandite (NiFeS2), which is usually associated with chalcopyrite and iron
sulphides. International Nickel (Inco) and Sherritt Gordon follow a
Scavenger
Float
------,1 1 (Tail)
1
I
Ni to Smelter
Tails
I
Matte Grinding,
(Float)
-------,l~_-----JI
Screening, and
1
Nickel to
Flotation
Smelter
Screen oversize
(iron, precious
metals to further
treatment)
bulk CuNi float by differentiation, while other Canadian producers leave the
l"tter problem to the smelters.
The Inco ore carries pyrrhotite and ferro magnesian silicates. Its pentlan-
dite is intergrown with the pyrrhotite, but the associated chalcopyrite is more
readily liberated by grinding. This facilitates differentiation of the bulk
float (Fig. 330). Lime controls the pH in both bulk and separating floats, a
further addition depressing the pentlandite from the copper sulphides in the
latter operation. A unique feature of the Inco operation is the use of
flotation to separate cuprous sulphide (floated) from nickel sulphide with
removal of iron, by grinding the nickel-copper matte produced by smelting
the high-nickel concentrate originally depressed from the bulk float. Slow
cooling of the matte is essential so that crystals of the sulphides can grow to a
suitable size before grinding. At Outokumpu Oy (Finland) a somewhat
similar ore is similarly treated, save that lime is not added in the roughing
stage and dextrin is added as a depressant for pentlandite when differentiating
the bulk float.
The froth in the primary plant at Creighton is judged by fluorescent lighting,
daylight being excluded. Winter pulp temperature is 9° C.
The Sherritt Gordon operation,81 mentioned earlier in this chapter, is com-
plicated by the presence of large amounts of talc which cannot be pre-floated
since it would take with it 10% of the nickel and half the copper in the ore.
A depressant has been developed which, when used after conditioning with
amyl xanthate and a phosphate dispersant, has met the need. Bulk flotation
produces a concentrate carrying 70% of the mill-head as free pentlandite and
85 %of the copper as chalcopyrite. These are separated in a further flotation
stage by depressing the nickel with lime and cyanide. The balance of the
mill-head nickel is too intimately associated with pyrrhotite to be liberated
economically by grinding. This is upgraded in a second float to 5 % Ni and
I % Cu in a circuit made acid with S02 (pH 6). This concentrate joins the
high-grade nickel product, the combined assay being 13·5 % Ni and 2 % Cu.
Niobium (columbium)
One mill which started production in the Oka (Quebec) district in 1961
removes ferro-magnetic minerals before floating successively the pyrite,
apatite, calcite, mica and silicates, leaving pyrochlore as a tailing. Test-
work resulting from which a small concentrator is to be built for treating
pyrochlore and perovskite to produce a 45% to 50% concentrate of Nb20 5
has been described. 95 The process uses mixed amines and di-amines together
with a sulphonated wetting agent to produce a low-grade concentrate.
Reagents are mixed hot and used as an emulsion in warm water. Ammonium
bicarbonate depresses some gangue minerals and aeration is starved, to keep
down apatite. In cleaning, a stronger addition of the bicarbonate is made
(3%-4%) to depress diopside and mica still further and produce a 10% to
20% grade of concentrate. This is now conditioned with calcium hypo-
chlorite to remove collector, after which the pyrite is floated and the residual
pulp is tabled to yield a 50% concentrate of Nb 20 5 •
774 Mineral Processing-Selected Ore Treatments
Phosphorus
An important branch of non-sulphide flotation is concerned with the
flotation (froth and agglomeration) of the apatites and phosphatic deposits.
The general formula Ca3(P04h of true apatite may be modified by inter-
growth with carbonate, and the term "phosphate rock" embraces chlorapatite
(CaS(P04)3CI), fluorapatite (Cas(P04hF), and hydroxylapatite (CaS(P04)3
(OH». The Florida rocks, called collophanes, lie roughly in a thirty-mile
circle under a sand cover. They are composed of clay slimes, silica and phos-
phate pebble in equal proportions. "Pebble" ranges from 1n down to 400
mesh and varies from well-defined hard particles to soft weathered rock.
The clay in this matrix aids pumping of the quarried material to a distant
washery. Pebble supplies are low and fine sands are the main source of con-
centrates, produced by flotation. The deposits are stripped, delivered to
sluices feeding a pumping and pipeline systein and, at the delivery end,
screened. Oversize up to 1-!- n is hammer milled to break down pebbles and
mud balls. All products then go to screens, hydrosizers and/or log washers.
A pebble product (minus -!-" plus 14 mesh) is recovered by screening and the
undersize is de-slimed, minus 150 mesh material being rejected. Oversize is
classified into plus 20 mesh pebble, minus 20 plus 35 mesh feed for agglomera-
tion tabling and minus 35 mesh feed to froth flotation.
Agglomeration feed is again de-slimed and conditioned at 70% to 75%
solids in rotary drums with fuel oil, fatty acid and sodium hydroxide. It is
then concentrated by tabling, spirals or on specially adapted conveyor belts,
silica being discarded. The minus 35 mesh feed is conditioned at 65 % to
70% solids, diluted to 25 % and then floated to produce a 68 % + grade of
bone phosphate of lime (BPL). Cells must have good resistance to abrasion,
such as neoprene linings, good sand relief and copious overflow. To reach
a 72 % + BPL, the Crago method of reverse flotation is used. The froth
product is de-oiled by the use of sulphuric acid and the residual silica
is then floated by use of an amine collector. Ideally, the head feed is
thickened to 70 % solids and agitated with acid, and thoroughly rinsed to
remove previous reagents. Fig. 331 gives a general flow-sheet compiled from
several sources.
Sun, Snow, and Purcell98 report that 18-C unsaturated fatty acids are the
best collectors for Florida phosphate in the descending order linoleic, oleic, and
linolenic acid. A liquid hydrocarbon such as fuel oil is necessary, more
being used when less fatty acid is required. The ions of Ca, Mg, AI, or Fe if in
the pulp at concentrations exceeding 17 p.p.m. seriously restrict flotation.
Best research results were obtained at 20° C. and a pH of 8·9.
Hughes 99 in an earlier description of the 30 m. tons/annum industry from
which some ten million tons of concentrate are made in Florida, states that
after washing and removal of pebble phosphates the treatment head assays
minus 5 % P 20 S• The -14 + 35# material is agglomerated and the -35 +
50# is floated at pH 9 to 9t. Tall oil, caustic alkali, and fuel oil are the re-
agents, the agglomeration feed· being conditioned as a thick pulp in which the
phosphates form glomerules with sufficient entrapped air to cause them to
ride up and over to the light-fraction discharge on the concentrating appliances.
Mineral Processing-Selected Ore Treatments 775
The standard treatment applied to the fine sands roughs out a phosphate-rich
float. This is de-oiled by scrubbing with sulphuric acid, rinsed and floated at
pH 7·4 with an amine and paraffin, the floating silica being discarded. The
Coronet Phosphate Co. treats its coarse sands on 42" wide belt conveyors.
The thick pulp is ploughed toward the centre by bars and baffles as it travels
beneath them, and the glomerules which work upward are flushed over the
sides by water jets.
Israeli phosphate treatment in the Negev is handicapped by lack of water.
A 28t %concentrate ofP 20 s was made at first by dry methods. These exploit
differences in hardness and fracture between the phosphatic oolites, the
cementing calcite and the free chert and limestone. Large lumps of the last
named are rejected and the balance is crushed through 4". It is autogenously
dry-ground to minus It". Oolites are liberated at about 35 mesh and a fatty
acid collector can be used in a de-slimed pulp to float off calcite and leave a
32 % phosphate tail. 97
Potassium
Soluble potassium salts can be extracted by means of "solution mining"
where the beds lie between impermeable covers. One such operation, at
Kalium in Saskatchewan, draws the resulting mixture of saturated halite
(NaCl) and sylvite (KCl) from underground to ponds where solar evaporation
produces differential crystallisation. Contamination is minimised by the use
of monel piping, iron being undesirable in the product. The physical chemis-
try of this fractional crystallisation can be studied by means of phase dia-
grams. tOO At Bonneville the Utah brines and bitterns contain ions of chlorine,
sulphate, sodium, potassium, and magnesium. Solar evaporation is aided
by dry, hot summers and low rainfall. Ninety per cent of the sodium
chloride is crystallised out in a general pond area and the residual bittern is
then given further evaporation in a separate pond, where sylvinite (mixed
sodium and potassium chlorides) crystallises to yield fertiliser grade potash
(96% KCl).
Potash salts can be floated away from those of sodium in a saturated brine.
This process is used extensively in the treatment of mined sylvinite ores. After
de-sliming to remove clay, which interferes seriously with the use of an
amine collector, a cationic float is made. Alternatively the ore may be
conditioned with lead chloride to activate the potassium and floated with a
7 to 9 carbon-chain aliphatic acid, such as a mixture of caprylic soap. tOt
Choice between cationic and anionic reagent plans is largely governed by
sensitivity to clay dispersed in the feed and by the size of particle which will
float. Soaps are not clay-sensitive and do not require careful scrubbing
together with cooked starches, natural gums or synthetic polyglycol or acryla-
mid. Activating lead in soap flotation is added during the wet grinding, up
to 4 lb./ton of chloride or sulphate being used. Alcohol, cresylic acid and
cooked starch are among the reagents used to control froth volume.
776 Mineral Processing-Selected Ore Treatments
(+)~(---
~
Screen (from i" to I-!") --~) (-)
t
Hammer Mill
t
Screen or
I (+)~(----- Hydrosizer_ ( - )
'------<E('-----' I set at 1 mm.
1
)~~---
or Cyclones
set 150 mesh
I
(- ) ....."'~--- Primary Classifiers
at35 m"h 1
) (+)
L(-)
r rl
conditiontg +- Secondary ----)- ( +)
and Phosphate Classifiers I
1
Flotation ~
Comminution
t
(-, ) ~ 20-mesh screen ~ (- )
t
150 mesh -<f-- Air Classifier
t
Attrition Mills
~
(slimes) ";'(--- Wet Cyclone
t
(floats) ""<---- Flotation
-j,
Drier
~
Magnetic Separator
150 ~
tails I
(30% of ~
orig. feed) Flint Sand
Among the minerals which are removed may be iron oxides and stained
particles, which respond readily to Cyanamid 400 series reagents, notably 425;
garnet, chromium and zirconium minerals, feldspar and mica. Not more
than 0·03 % of iron can be tolerated in a good optical sand and it may be
found profitable to use a warm sulphuric acid leach to dissolve this, when
present as Fe z0 3 .
Foundry sands are reclaimed by a thorough scrub with water to remove
bonding materials, clays, slimes and fine particles which are classified out and
discarded.
Si1ver
The sulphides include argentite (AgzS); polybasite (Ag 2S . CUS)9 (Sb 2S3 .
As ZS 3); proustite (3Ag 2S . As ZS 3); pyrargyrite (3Ag zS . Sb 2S3); and stephanite
(5Ag zS . Sb 2S3). Flotation characteristics have been studied at length by
Leaver, Woolf, and Towne.'o4
Their general finding is that these sulphides, and also tetrahedrite (4Cu2S .
Sb zS3), in which part of the copper may be replaced by silver, float well in a
natural circuit with ethyl xanthate or Aerofloat 15. Depressants include
sodium sulphide. sodium hydrate, lime. and starch. Slimes containing iron
oxide or talc interfere with grade and recovery. If free gold is present.
Aerofloat 208 may be helpful. Copper sulphate may help in activation if
zinc is associated with the silver sulphides. Where selective depression of
silver away from base-metal sulphides is needed, this is best performed after
bulk flotation, using lime. cyanide. or zinc sulphate. Oxidised silver ores are
aided by sulphidising treatment.
Following success in pre-concentration of gold and silver ahead of cyanida-
tion on a small scale. mines in the Pachuca district of Mexico now operate a
1,000 ton flotation plant which makes separate concentrates of lead. zinc and
iron. tails being cyanided. Gold and silver contained in these concentrates
are paid for by the smelters and cyanidation is improved by removal of cyan i-
cides during flotation. The galena concentrate carries as much refractory
gold and silver as possible. together with the copper in the ore. The sphalerite
carries bullion values which can be recovered by cyanidation. Iron minerals
are also floated, the amenable ones going to cyanidation while those carrying
refractory bullion metal are sent to smelting. I05 A pilot plant has shown the
feasibility of extraction of copper and silver from a low-grade Peruvian ore by
the segregation process. I06
Sulphur
similar methods used for sulphide removal), the main requirement is a grade
of pyrite which will burn strongly; high recovery is a secondary consideration,
economy in the use of small quantities of collector, frother, and perhaps acid
being of prime importance.
Separative treatment to recover elemental sulphur cemented between the
grains of the host rock must take precautions against corrosive attack. Fine
crushing in rolls may be followed by pebble milling in a silica-lined mill,
using scuffing in open circuit to free the sulphur. Flotation at apH of2 to 3
in acid-resistant cells follows, perhaps using only ar. alcohol frother as reagent.
Recovery of sulphur from low-grade deposits by simple solvation has been
proposed,lo7 using dimethyl disulphide as the leaching agent. It might take
over where grade was too poor for use of the Frasch process, in which
superheated water is pumped down to liquefy and bring to surface the mineral.
In Italy kilning is used for this liquefaction, part of the sulphur acting as fuel.
Talc
Talc is mostly processed by dry grinding and air classification. The specifi-
cation for the paint industry requires more than 981- % to be minus 325 mesh.
Where exceptional purity is needed, flotation is possible, in which a wetting
agent (e.g. alkyl sodium sulphonate) can be used.
Tin
Gravity treatment of cassiterite-bearing sands and gravels dominates
industrial practice. Where hard-rock mining is used, or a roughed gravity
concentrate must be upgraded after further grinding, emphasis is on the
avoidance of over-grinding since slimed cassiterite is difficult to recover and
tailings losses may be high. A beach-sand flow-sheet was given in an earlier
chapter (Fig. 184). Detailed descriptions of current methods in which
jigs, tables, spirals and tilting frames are integrated with close screening and
classification have been madel08.109.
Among chemical methods which have shown laboratory promise is vola-
tization of tin as its tetrachloride, a process which has had limited use in
plant practice. Another proposal is based on the equations
Sn + CuCI2 = Cu + SnCh
the cassiterite being reduced by carbon monoxide at 850°C before leaching Uo •
Tintanium
The minerals mainly concentrated for titanium are rutile (S.G. 4'2, non-
magnetic) and ilmenite (S.G. 4'5-5'0, weakly magnetic). Where these
occur as beach sands they are separable by mixed H.T. separation and
magnetic treatment, as shown in Fig. 266. Ilmenite ores are treated by
780 Mineral Processing-Selected Ore Treatments
Tungsten
The principal ores treated are wolfram (S.O. 7-1 to 7'9) which may be
feebly magnetic, and scheelite (S.O. 5'9-6'1). In gravity treatment, care
must be taken to concentrate the value at the coarsest possible grind, with
avoidance of over-grinding. Jigs, tables, rag frames and spirals are used.
Using flotation, scheelite (CaW04) is the only tungsten ore to have received
important commercial treatment. It floats readily at a pH of lOt in softened
water, using a carboxyl collector, sodium silicate as dispersant and tannin as a
depressant for associated calcite, fluorite, dolomite, and apatite. The
formidable difficulty is in obtaining sufficient separation from the gangue
minerals to give an economic grade of concentrate, as the scheelite itself is
readily depressed. Spraying of the froth with Aerosol OT has been recom-
mended.
Roasting and/or leaching of tungsten ores and concentrates involves
complex treatment to remove impurities which include tin, arsenic, phos-
phorus, manganese, bismuth, molybdenum, copper, antimony, zinc, lead and
iron 113 . A shorter route for scheelite starts with a sulphuric acid leach
followed by re-Ieach of the resulting tungstic acid with ethylene glycol 114.
Uranium
Twenty-four types of uranium deposit have been classified by Everhart115
in accordance with associated rocks, minerals of value and type of deposit.
The more important uranium minerals, as listed by Patching116 are:
TABLE 44
MORE IMPORTANT URANIUM BEARING MINERALS
TABLE 44-continued
Mineral Composition S.G. Hardness
Meta-Torbernite CuO.2UOaP.Os8H.0 3·68 2
Parsonsite 2PbO.UOa.P.Os.H 20 6·23 ?
Pyrochlore Na, Ca, Nb.06 F 4·2-4·9 5-5'5
Saleeite MgO.2UOa.P20s·8H.0 W3·3 2-3
Thucholite Hydrocarbon with U, Th, etc. }'5-2'0 3·5-4
Torbernite. . CuO.2UOa·P.Os·12H 20 3·2-3·5 2-2,5
Uraninite pitchblende xU0 2 ·yU03 6,5-10,6 5-6
Uranocircite BaO.2U03.P.Os·8H.0 3·5-4·0 2
Uranophane CaO.2U0 3 .2SiO•. 6H.0 3·8-4·9 2-3
Uranothorite ThO z.SiO•. U0 3CaO, etc. 4·1-4·4 4'5-5
Of these, the most important economically are the oxides pitch-blende
and uraninite. Veins of uranium minerals are rarely of mining width, and
often are networks. This leads to dilution in mining by waste host rock.
The more usual deposits are disseminated among other minerals. Physical
pre-treatment aims at reducing the bulk which must be subjected to chemical
attack: at removal of interfering ore constituents; and at recovery of associ-
ated values which can be separated after appropriate comminution. With the
low head value of typical uranium ores (as little as 0·\ ~;,;) any roughing step
which will remove a substantial fraction without undue loss is useful, since
the physical methods employed in mineral dressing are cheaper than leach
treatment. Interferants in acid leaching include acid-consuming carbonates
(more than about 4 %) chlorite, earthy hematite, apatite, and fluorides.
Those inimical to carbonate leaching include sulphides (above 0·5 % S.).
Instead of a discard of gangue, selective sorting has limited use, the notable
appliance being the Lapointe picker 1l7 which exploits the radio-active
mineral's ability to signal its position on a miniature conveyor. DMS is
used at Radium Hill to separate the heavy sulphides associated with the
uranium from an impoverished gangue. The possibilities of froth flotation
have received much research attention. Indirect upgrading by removal of
gangue and direct flotation of values is an economic process in some plants.
One operation, described by Read, Carman, and Gough 1l8, treats old tailings
and current cyanide residues to obtain 40% recovery of U308 and 90% of
the sulphur in a concentrate weighing 6 % of the head feed.
The dominant chemical treatments are the acid leach and the carbonate
leach. In the former a fairly thick pulp (60% solids) is agitated with sul-
phuric acid and oxidising agents such as manganese dioxide and sodium
chlorate (to maintain the uranium in its reactive state as U308)' The preg-
nant liquor is then removed and stripped. Where excessive carbonate in the
ore precludes this method alkaline leach with sodium carbonate and bicarbon-
ate is used. Oxidation is obtained either by pre-aeration or the use of
permanganate, and hot pulps are sometimes needed (70°C or more). Car-
bonate leaching does not entail the costly materials of construction required
to prevent corrosion in an acid leach plant, and treatment of the "pregs" is
simpler, but both processes have operating difficulties.
Recovery of uranium from pregnant solutions can be performed by
chemical precipitation, static-bed ion exchange, fluid-bed ion exchange or
by the use of organic solvents. The first of these methods, based on the use
782 Mineral Processing-Selected Ore Treatments
(A) Sites. Exchange positions on surface of bead where ions are captured
by resin.
(S) Solvent extraction. Selective transfer of metal salts from aqueous
solution or pulp to immiscible organic liquid.
Pre-concentration
Of the numerous oxidic uranium minerals, those commercially exploited
include autunite, carnotite, and pitchblende or uraninite (XU02.yUOa,
etc.). All are of variable composition. Thucolite is mixed with hydrocar-
bons. Tobernite is a hydrated compound with copper phosphate. Urano-
phane is a compound silicate with calcium. Carnotite is a hydrated com-
pound with vanadate and potassium. The ores treated commercially
contain only minute quantities of these uranium-bearing compounds and
include quartz, schists, shales, and dolomite. The problem of upgrading
prior to leach treatment has engaged much laboratory research, and limited
industrial application of its findings has followed. Three lines of attack
have been opened up. First the uranium mineral might be concentrated.
Second, associated minerals might be floated, taking with them the uranium.
Third, gangue minerals (particularly those liable to interfere with subsequent
chemical treatment) might be removed, leaving an enriched uranium-bearing
tail.
In pilot-scale work Fitzgerald and Kelsalll~o report on a micaceous feld-
spathic sandstone impregnated with carbonates and carrying uraninite.
Levinl~l has described testwork on thucolite in Witwatersrand ores and
residues. 20 to 30% of the uranium floated with a frother only. Xanthate
brought this figure to between 30 and 70 With some ores oleic acid was a
good collector. Altogether 80 % of the mill-head reported in a 20% con-
centrate, though pre-cyanidation to recover gold reduced flotation efficiency
unless counteracted by an acid wash. A valuable review of the research
aspects.of uranium concentration is given by Patching ll6 . Lord and Light 122
describe work on uraninite and three other uranium-bearing minerals which
gave high laboratory recovery and upgrading. Read, Carman, and Gough liH
describe the operation of two flotation plants. One treats cyanidation
tailings on the Rand to give a concentrate containing 40% of the U30a in
6 % of the feed weight. The other treats the residue from the uranium acid-
leach plant. The reagents used are xanthate, Aerofloat 25, and Aerofroth 70.
Butler and Morris Uj have developed a series of organic collectors which
give up to 15: I ratios of concentration from heads assaying 0·1 to 0'5~~
U:J0 H under laboratory conditions. In research on low-grade Saskatchewan
ores Tinker et a/l~c) have used mercaptans under experimental conditions
with promising results.
Laboratory tests on a Canadian ore were made in two stages 12fi , in which
acid-consuming gangue minerals were first floated, and flotation of the uran-
ium minerals followed. Working on ore ground in Faraday Uranium's
mill circuit fro111 91 j., to 93/\, of the uranium was recovered in less than half
the weight of the head feed. More than a third of the ac.id consumed in
straight leach-extraction was saved. Sodium oleate was used in stage I,
and a proprietary emulsion (Acintol FA-I) in stage 2. Other research on
784 Mineral Processing-Selected Ore Treatments
most of the lime, and some iron and aluminium go into solution when the
material is pulped and agitated while 70 %of the residual solids are discarded
as a coarse tailing. This is possible because the uranium is interstitial in the
grains of sandstone.
Secondary uranium minerals dissolve readily in dilute H2S04, but the
tetravalent oxides require oxidising conditions. Where suitable pyritic
mineral occurs in the ore in sufficient concentration, it is the source of the
sulphate. Otherwise, an oxidising agent such as pyrolusite is necessary.
A description of Rand practice has been given by McLean and Prentice 130 .
The pregnant liquor from acid leaching carries concentrations of other ions
much in excess of the uranium content. Post-war research first worked on
straight precipitation methods, but these gave way to the now widely used
ion exchange (IX) processes.
Metal sorption on cation-exchange substances (natural zeolites and
synthetic resins) has long been used in the removal of Ca and Mg from "hard"
water. This form of activated resin is, however, useless because it is non-
selective. Uranium recovery is based on anion-exchange resins. The
metal sorbs as an anionic complex. When uranyl sulphate solution flows
through a bed of strong base anion-exchange resin the uranium is absorbed,
while substantially all other metal ions go through. From time to time load-
ing is stopped, and the uranium is displaced by an acidified salt solution. In
a pure pregnant sulphuric solution the uranium is present as the uranyl
(U02H) ion or an undissociated sulphate (U02S04). If this is passed over
such base IX-resins as De-Acidite FF or Amberlite IRA-400 (chloride form)
one U02 group and three S04 groups are sorbed and four Cl- ions are
displaced from the resin. With a sulphate-form resin the sorption equation
is
U02 2 +S04 2 + 2R= S04~
R ~ U02(S04)a
The sulphate ion does not compete for resin sites, but bisulphate does.
Resin can be loaded at pH 2 to 7-lb/ft 3 , but if the pH falls to 0·5 this figure
drops to 2-lb/ft 3 , sites not occupied by the uranyl in sulphate anion being
used by bisulphates. Operation in acid leaching is concerned to give the
highest pH compatible with effective extraction, together with stable solutions.
This requirement is met at a pH between \·8 and 2. If the build-up of ferric
sulphate in the circulating leach liquors is excessive a certain amount sorbs
to the resin. This ceases if the pH is decreased toward 1·4.
At the point where uranium begins to "break through", the resin in the
loading column is far from saturated and the uranium is of low purity.
Plants are so designed that two or more columns are in series, and that
break through on the second column begins when the first column is satur-
ated. Column I is now taken off line, column 2 receives the entering preg-
nant solution and column 3 replaces column 2 as the save-all. During the
last part of the loading to saturation some iron is crowded off the resin.
The loaded resin is next eluted with an acidified mineral salt, the most-used
being molar ammonium nitrate containing 0·2 molar nitric acid. The
residual ferric ion is the first to be displaced and can be separately discharged
786 Mineral Processing-Selected Ore Treatments
further stages of agitation, heat and additional acid being used. The dis-
charge is thickened, "pregs" being returned to the first stage of agitation while
tails are washed and discharged. Unclarified solution is fed direct to one of
four 8 ft x 15 ft columns each equipped with a bottom distributing system of
piping covered by coarse sand. The base anion resin resting on this sand
occupies 270 ft3. Three columns are on stream during loading, with a
pressure drop of 15lb between bottom and top and an estimated flow rate of
5 to 6 gal/min/ft 2 of cross-section. A column is considered loaded when its
liquor has the same uranium content at both entry and discharge. The
second column in line should by then be showing a break-through value of
up to 10% of the IX feed. The third column gives a virtually barren effluent.
After loading, the column is taken off stream and given a water-rinse to
displace most of the remaining "pregs". This liquor is sent to the next
column. Vigorous back-washing is now applied at 200-300 gal/min to
loosen the bed and flush out slimes. Any resin carried out by this treatment
is trapped on a l00~ screen. The column is next eluted with ammonium
nitrate in HN03, the eluate going to precipitation. The cycle is completed
by acid rinse with H2S04 to regenerate the Amberlite 400 anion resin, which
has a uranyl nitrate capacity of 4·2 Ib/ft3. The concentration ratio in the
IX step is 20: 1 from a feed assaying 0·5 to 0·6 g/l. The eluate goes to three
pachucas in series, lime slurry being added to the first two and gaseous am-
monia to the third. The lime slurry is precipitated as gypsum and takes with
it most of the iron as ferric hydrate. The product from the third pachuca is
thickened, the underflow, which assays 1-2 % uranium, being filtered and
returned to the head of the leaching circuit. The filtrate is fed to a series of
three pachucas where the pH is adjusted to a final 6·8-7·0 with gaseous
ammonia. The overflow is thickened, heated to 90°C, re-thickened and
filtered. The reagents consumed per ton of feed are 115 tb H2S04, 3·3 lb
HN03, 6-!- lb CaO, 0·28 lb Separan (a Dow flocculant) and 2·8 lb anhydrous
ammonia. The power is 35 Kwh/ton. The plant is well "instrumentated"
and most controls are air powered.
An alternative to fixed-bed and R.I.P. methods in the stripping of pregnant
liquor is the moving-bed I35 . One system has ten columns, six of which are
grouped in parallel lines of three for adsorption. Three are used for elution
and the tenth back-washes uranium-saturated resin beads. Piping is so
arranged that resin in any adsorption column can be transferred to backwash,
from backwash to any elution column and from this back to .any adsorption
column.
For resin recovery the use of flotation has proved interesting in laboratory
research l36 • The R.I.P. process cannot tolerate more than some 10% of
solids in the pulp, but direct recovery of loaded resin by flotation would not
be subject to this limitation, and the problem of resin abrasion by coarser
particles might be simpler. Cation-exchange resins are readily floated by
amine-type collectors and anion exchange ones by all anionic collectors.
Tests showed resin flotation to be clean, rapid and nearly complete while
loading capacity and elution were not upset by bead conditioning for flotation.
Bead size in the minus 16 plus 400 mesh range is not critical. Regeneration of
cobalt-poisoned resins in which boiling with either thiocyanate or nitrite is
788 Mineral Processing-Selected Ore Treatments
FLOW SHEET
URANIUM SOLVENT EXTRACTION
EXTRACTION
RAFFINATE TO TAILS
(4 STAGES)
BARREN SOLVENT
FIR
STRIPPING
SODA ASH
(2 STAGES)
FILTRATION BY-PRODUCT
ACIDIFICATION
NH 3 -------I PRECIPITATION
FILTRATION
measurement of the e.mJ., which is held at 400 millivolts in the first two tanks.
Leaching time is about 16 hours.
Bicroft142 , in Canada, started acid leaching in 1956, using conventional
fixed-bed columns. The main operating difficulty has been with poly-
thionate poisoning of the resin. The pegmatitic ore yields a low-grade leach
liquor at a pH of between 1·8 and 2'1, carrying 0-4 g/l. of UaOs, ferrous and
ferric iron, fluorine, chlorine, thoria, alumina and over 15 p.p.m. of poly-
thionates. Following backwash with warm water the IX columns are eluted
with sodium chloride solution acidified with sulphuric acid. Eight bed
volumes are sent to precipitation and nine to re-cycling eluate. A "bed
volume" is that of the liquor which corresponds to the space occupied in the
column by the resin and its voids. The eluate is stripped by precipitation
with magnesia, the reaction being U02S04 + Mg(OH)2 = U02(OHh + MgS04.
The sulphate building up by this reaction is held near 120 g/l. by periodic
discharge of neutral eluate. About a quarter of the 0·9 g/l. of fluorine in the
pregnant liquor is adsorbed and then eluted with the uranium, from 5 % to
10 % later precipitating during magnesia reaction while the rest re-cycles with
the eluate. Every few days some neutral eluate is withdrawn and agitated
with slaked lime at pH 9·1 to precipitate this as calcium fluoride, at the same
time as sulphate build-up is being controlled.
The IRA-4oo resin used at Bicroft shows a plant loading of 3·6 lb/fta of
UaOs. When fouled by polythionates this falls to 2 Ib/ft a, at which point
resin regeneration is necessary. Break through from the first column when
the resin is fouled by slime or polythionates starts at some 65 %of saturation.
The general formula of polythionates is Sn06 ~ with n between 2 and 6. They
come from any sulphur source, such as sulphide in the ore. One preventative
is to maintain the free leaching acidity above 5 g/l., when sulphates instead of
polythionates are formed but at Bicroft this would lead to excessive con-
sumption of acid, which is held in a free range between 1 and 3 gil. Three
control variables are of practical use-temperature, retention time and point
of addition of sodium chlorate as an oxidising agent. Poisoning is seasonal,
being worst in spring and autumn. Resin regeneration is done after these
seasonal build-ups of poison by taking a set of four IX columns out of service
for cleansing. After elution the column receives a weak NaOH/NaCI wash,
the resin being stirred with an air lance to prevent channeling. It is then
flushed with the same solution, washed with water, then with normal sul-
phuric acid and finally with water before being returned to service. The
alkali decomposes the polythionates (2S506 + 6(OH)-= 5S20a + 2H20), the
sodium chloride speeding this reaction. Fluoride is removed at the same
time (ThF4+4NaOH=Th(OH)4+4NaF), the thorium hydroxide dissolving
during re-acidification of the resin.
The moving-bed system of IX following acid leach is used by Consolidated
Dennison in a 6,000 ton/day plant 14a. Maintenance of anti-acid protection
for structures, machines, tanks and piping are an appreciable item of operat-
ing cost. The ore minerals include brannerite, uraninite, monazite, sericite,
pyrite and rutile. A long leaching period in strong acid at raised temperatures
is necessary unless a good addition of chemical oxidants is made. Leach is
in three parallel lines, each headed by two pachucas and followed by six
Mineral Processing-Selected Ore Treatments 791
mechanical agitators with wood-stave tanks and stainless steel gear. Live
steam and strong sulphuric acid are introduced in the pachucas and steam
can also be blown in further down-line. Sodium chlorate (2'72 lb/ton
milled) is added at the second pachuca and average working temperature
along the line is 57°C. The acid strength falls from 85 g/l. to 70 g/l. along
the leaching line, and retention time is 50 hours. Agitator maintenance
arising from corrosion accounts for a third of the milling cost. The filtered
pregnant solution is clarified in leaf filters and carries from 9 to 19 g/l. of
U aOs. Resin, after loading, is hydraulically transferred from the adsorption
columns to the backwash column and thence to elution. The advantages
claimed include lower installation cost, higher resin loading, lower con-
sumption of nitrate (41b/ton of ore), better concentration with lower cost of
precipitation, and better shake-up and. slime removal. Against these,
attrition of the resin may be somewhat higher and loss of barren solution a
little larger. Amberlite IRA-400 is used and loading is of the order of
4'61b/c ft at pH 2·2. Some 400 tons of eluate, assaying 19 g/l., are produced
daily and precipitated in two stages. In the first the pH is adjusted with lime
slurry to 2, and sulphate is removed as gypsum. Magnesia is then slurried in
to bring the pH to 3·2 for precipitation of iron a1'\d other impurities in a series
of three pachucas and a time of five hours. The purified eluate next goes to
three mechanical agitators in series, and magnesia slurry raises the pH to 5
and at discharge to 6'5. The emerging slurry is thickened for removal of
uranium, overflow returning to eluate tankage.
A general description of solvent extraction techniques is given in a trade
bulletinl44 and an interesting account of R.I.P. leaching coupled with solvent
extraction in another from the same source14S • The cleaning of process water
used in the carbonate leach plant at Homestake has also been described146•
Zinc
The conventional flotation of sphalerite, usually following that of associ-
ated galena, was considered under "Lead". The surface is activated by
copper sulphate in a pulp brought to a pH of 8-9 with lime, and a straight
xanthate float follows. More complicated is the problem of the oxidised zinc
ores. These include smithsonite (ZnCOa), hydrozincite (ZnCOaZn(OH)2),
hemimorphite (Zn2SiOa(OH)2), and willemite (ZN2Si04). Research work
and practical mill circuits are described by Rey et aI147.149. With a mixed
ore of lead and zinc sulphides and oxides, the treatment starts with sulphide
flotation. Tails are then conditioned, sulphidised, and given collector con-
ditioning. If amines are used, the slimes must first be removed or neutralised
to avoid heavy reagent consumption and contamination of concentrates.
They can be neutralised with soda-ash, sodium silicate, and such polyphos-
phates as calgon. Organic colloids such as starch and carboxymethyl
cellulose are also helpful where desliming in cyclones is avoidable. As slimes
carry zinc, the question of their handling is economic, and depends partly
on the zinc market price. Unlike its action with sulphide minerals, sodium
sulphide in reasonable excess has no depressing effect on oxide zinc minerals.
792 Mineral Processing-Selected Ore Treatments
Only primary amines have collecting values. They are preferably aliphatic,
in the form of soluble acetates or hydrochlorides. Frothers include pine oil,
Dowfroth and alcohols. The amine is added with the frother, immediately
after the sulphide.
Plant-scale flotation of oxidised zinc ores following sulphidization at 50°C
is practised at Ammi 149 • Activation with copper sulphate and flotation with
amyl xanthate completes the reagent plan. The method does not suit
strongly ferruginous ores. An alternative treatment for the finer fraction of
the zinc oxidic minerals uses lauryl amine emulsified with pine oil and a
mineral oil with sodium sulphide and a pH of 12_
Some Generalities
This description of typical industrial treatments is necessarily incomplete
and has omitted many important ores and elements. Practice improves
continually, as new processes are developed and fresh demands are made by
the consuming industries served by mineral processing. Some basic principles
underlie the planning of most, if not unfortunately all flow-sheets. A sum-
mary of the more important is offered in conclusion.
(1) Avoid over-grinding.
(2) Avoid surging and irregular rate of flow at each stage, using auto-
matic aids wisely.
(3) Watch the state of return water and the effect of build-up of detri-
mental constituents.
(4) Take products, whether values or tails, out of the main circuit
coarse and early.
(5) Make good use of roughing followed by cleaning.
(6) Watch the ore and its products for by-product values.
(7) Stockpile any middlings not at present worth treating but of potential
value.
(8) Maintain routines, including mill cleanliness and safety practises.
(9) Operate planned maintenance schedules and avoid the need for
emergency action.
(10) Break down the cost per ton of each unit operation in terms of
labour, material, power and equipment and graph monthly changes.
(11) Make regular analyses of values in each size of tailing, and use the
information thus given to improve grinding and treatments in the
zones of maximum loss of value.
Mixed ores
The sulphides of lead and zinc, together with lesser amounts of those of
other metals, commonly occur in economic quantities and are separately re-
covered. Where liberation mesh permits, galena can be recovered by jigging
Mineral Processing-Selected Qre Treatments 793
and tabling before finer grinding is used to free the species involved at a mesh
suitable for differential flotation.
The normal procedure in differential flotation is to activate and float the
minerals successively. An alternative commonly practised is to make a bulk
float of two or more minerals, and then to differentiate them by depressing
one in a secondary flotation. An ore containing payable quantities of copper,
lead, zinc, and iron sulphides, together with, say, barytes, might be treated as
in Fig. 334.
This example of bulk flotation followed by successive differentiation is
taken from pre-war practice at Rammelsberg Mine in the Harz Mountains.
Bulk flotation may be a good starting-point when soluble metal salts in the
pulp have activated more than one of the sulphides, so that it is easier to lift
two together. One sulphide in the mixed concentrate is then deactivated as
shown in Fig. 334, where lead is depressed from a bulk copper-lead float.
Another case in which bulk flotation might be chosen would be that of an ore
in which the valuable sulphides were coarsely aggregated with respect to the
gangue but so closely interlocked with one another as to require fine grinding
Feed
t
(C) - - - B u l k Pb, Cu Flotation - - - ( T )
!
Condition to
~
Condition to
depress Pb activate Zn
!
Flotation
1i
Flotation
t t ~
Cu Pb Zn (T)
Float Sink Float
t
~
(C) Condition Tails (T)
~.------ and Float - - - - - - - ,
Fe
Float
Condition Tail
1
Barytes
C and Float
t
Tails
Float to Waste
before separation. Bulk flotation might be used to reject the gangue and thus
reduce the tonnage sent to fine grinding. The general principle in differential
flotation is to use a collector in starvation quantities, after selectively con-
ditioning the pulp to promote activation of the first mineral to be raised. As
a certain amount of activation of the second mineral usually occurs under
such conditions, a lively and fragile froth should be maintained. Various
frothers derived from alcohol have been developed to aid in obtaining the
desired texture.
Lead-Zinc Differentiation
The main factors which influence the behaviour are given by Rey 79 as:
1. Abundance and nature of iron sulphides.
2. Degree of oxidation.
3. Basic or acid nature of the gangue.
4. Presence and nature of copper minerals.
Basic gangue contains sufficient available calcium to neutralise acid and to
precipitate as hydroxide any ferrous sulphate derived from oxidation. In
aerated water the reaction then follows:
2FeS04 + 2CaCOa+ 3H20 +102= 2Fe(OH)3 + 2CaS04 + 2C02
The sulphates of lead and zinc react similarly, nearly to completion. The
gangue tends to control surface contamination of the particles through its
influence on the solubility products. If sulphides are oxidised in an acid pulp,
they remain bright and highly floatable, but become contaminated by basic
salts when oxidised at about pH 7. From this Rey proposes a classification
into four main ore types, arranged in increasing order of difficulty (as regards
differential flotation):
(A) Unoxidised ores.
(1) Galena-sphalerite or galena marmatite.
(2) The same, plus increased amounts of iron sulphides (pyrite,
marcasite, pyrrhotite). This class can be subdivided into ores
with low, medium or high content of iron sulphides.
(B) Oxidised ores-acid gangue.
(1) Weakly acid ores.
(2) Strongly acid ores (galena-sphalerite-anglesite).
(C) Oxidised ores-basic gangue (galena-sphalerite-cerussite, possibly
smithsonite).
(1) Without iron sulphides.
(2) With iron sulphides.
(D) Ores containing copper minerals of secondary origin.
A generalised reagent plan for the successive flotation of galena and sphalerite
follows these lines. The pulp is conditioned first for activating the galena.
pH is controlled usually by soda-ash, but sometimes by lime, to a figure
between neutral and pH 10, and a dispersant such as sodium silicate may
also be employed. If there are traces of copper in the ore, from 0·1 lb/ton to
0'5 lb/ton of sodium cyanide may be needed to prevent random activation of
the sphalerite .during the collector-coating of the galena. To aid in de-
pression of the zinc, a solution of zinc sulphate may also be used. The de-
TABLE 45
UNOXIDISED ORES: USUAL REAGENT FORMULAS AND FLOTATION RESULTS
(After Rey)
Proportion of Alkali Depressant Other Reagents
Iron Sulphides (g./ton) (g./ton) (g./ton) Grade Concentrates
None CaO 0-200 NaCN 0-20 Zn Circuit Zn 52/62
CUS04 250-500 Pb 0·5/0·1
Pb Circuit Na2COa 0 Pb Circuit
None Pb 75/80
An 0·5/3
Low
Pb Circuit Na2C03 0-100 NaCN 0-100\ Pb 60/78
ZnS04 0-2001 Zn ·7/5
Zn Circuit CaO 0-750 CUS04 300-600 Zn 62/60
Pb ·65/-3
Medium
Pb Circuit Na2C03 0-1000 NaCN 30-100 Pb 60/78
orCaO 0-500 ZnS04 0-350 Zn 2·5/7
Zn Circuit CaO 500-2500 CUS04 300-600 Zn 52/60
Pb ·65/-3
High
Pb Circuit CaO 200-500 NaCN 150-300 I Na2S03 or Pb 55/70
(in mill) (in mill) i NaHSOa 0-500 Zn 3/8
ZnS04 0-500
Zn Circuit CaO 500-2500 CUS04 400-800 Zn 48/57
Pb 1/3
* Concentrate grade varies with marmatitic or sphaleric composition.
796 Mineral Processing-Selected Ore Treatments
TABLE 46
After McQuiston
Major copper Method of Approximate Lead:
Mine Mineralisation Separation Copper Mineral Ratio
Buchans . Chalcopyrite SO. 2:1
Idarado. . Chalcopyrite Cyanide I: I
San Francisco. Chalcopyrite SO. 5 :1
Casapalca Diversified SO. 4:1
Mahr . Chalcopyrite Cyanide 2:1
Tsumeb . Tennantite Cyanide 2:1
NaCN 0'5\
Dowfroths ·04
Z.l1 ·07 Comminution and
(lbs./ton) Classification
(6%+65# and 54% -200#)
1
CaO
CUS04
Dowfroth
Z.l1
·5
·13
I
~o·06 ___!
.-----Lead Flotation
r
(Tail)
Zinc Flotation
1
Cone. 721% Pb
2% Zn
1 and 931% recovery
Cone. 491% Zn
11% Pb
Tail and 931% recovery
·25% Pb
·32% Zn
This flow-sheet (Fig. 335) has been simplified in block form from one given
by Walton s1 . It illustrates the typical sequence in a simple separation, with
low reagent consumption and high specific recoveries.
calls for fine grinding in controlled stages, and a special "de-zincing" opera-
tion in which the lead recleaner concentrate is treated at 40°C with sodium
dichromate before removing the bulk of its residual zinc as a float. A block
flow-sheet (Fig. 336) shows the increased complexity dictated by the nature
of the ore.
Mine Ore
t
Dense-media Separation -------.1
Primary Comminution t
Classification to - 20# Tail
lr-------- t
Lead, Coarse Flotation
Pb Conc. 1
Sec. and Tertiary Comm.,
Classification to
84%-200#
t
Float to
I Lead Flotation - - - ) (Tail)l
at 50% Solids
Initial
Zinc Flotation
1 1 TiilS to
Dezincing Treatment of Tin Plant, etc.
and Reflotation Zinc
-l- Cleaning
(Tail) t
Zinc Conc.
The final lead concentrate assays 65 % Pb and 3 % Zn. The final zinc
averages 45 % Zn and 9 % Pb. The reagent plan is summarised in Table 48.
TABLE 47
SULLIVAN FLOTATION REAGENTS
lb./ton Lead Zinc
(original feed) Flotation Flotation Desliming Total
Cyanide ·063 ·063
Lime ·77 ·39 ·39 1·55
Z.ll Xanthate ·14 ·07 ·01 ·22
Dowfroth ·006 ·008 ·014
Dichromate ·063 ·063
Cresylic Acid ·008 ·008
Copper Sulphate ·58 ·29 ·87
800 Mineral Processing-Selected Ore Treatments
1
N82S~
(6 kg.fton) ~
I
Bulk Flotation----'l
Agitation at
T_ail
50% solids
1
Classification
Thickening
Agitation with KCN
Comminution
1
Float
Lead Sulphide
Float
r
lr---Lead Flotation---""'l
ZinC Flotationl
Zinc Sulphide
rPyrite Flotationl
Float
Pyrite Tail
Ammeberg, Sweden
Here the ore sulphides are sphalerite, galena and pyrrhotite with quartz,
felspars, mica, calcite, garnet, pyroxene, and hornblende. Crystallisation
varies, but a liberation limit of .1601-' is worked to, liberation of sphalerite
from galena being still finer, at 501-'. The mill receives, in addition to its local
ore, the sink product from the dense-media plant at the associated mine at
Mineral Processing-Selected Ore Treatments 801
1
, - - - - - - - - Classifier - - - - - - . \
----Il
(+) (-)
Tsumeb, S. W. Africa
A complex ore is treated in this plant. Work started with the milling of
20-year-old dumps of heavily oxidised Cu-Pb-Zn minerals on the evidence of
802 Mineral Processing-Selected Ore Treatments
batch testwork. Much was learned as the mine developed and the f1ow-
sheet was adapted to deal with newly exposed problems. The deposit contains
forty-one valuable minerals of copper, lead, zinc, vanadium, and germanium.
In the upper levels the copper is oxidised. In the middle zone it becomes
chalcocitic and at depth changes to tennantite. In a Paper describing this
complex but outstanding operation by Ratledge, Ong, and Boyce B3, stress is
laid on the importance of adding the xanthate collector in the sulphide
section to the grinding mills rather than in a later conditioning stage. Tests
confirmed that with this ore sulphides which have been finely ground do not
respond so well as comparatively granular material. Right through the ore
body the zinc mineral has been activated by soluble metal salts, and strong
depressing action is necessary to inhibit its premature flotation.
Primary Comminution
and Classification
.-----(C) ~
1
.. - - - CuPb Flotation - - (T)l
I Secondary
t Comminution
... - - -
~ (C)+<---Sec. CuPb Flot. (T)l
1
r--------Tertiary
Comminution
Germanium
Conc.
The sulphide flow-sheet (Fig. 339) was developed after varying the experi-
mental grind. It uses three grind-and-float stages, the third milling stage
closing with the secondary flotation section. The CuPb concentrate contains
0·08 % of germanium, mainly as germanite, but with some magnetic renierite.
The latter can·be checked by placing a sample on a glass slide and moving it
Mineral Processing-Selected Ore Treatments 803
over a magnet. The sphalerite is fluorescent and plant control makes use of
this and of pilot tables which monitor the flotation tails. In the oxide ore
flotation circuits lead sulphide, oxidised lead minerals, slimed lead concen-
trates, copper sulphide, copper non-sulphides, germanium and zinc products
are recovered separately. Preliminary grinding is to 50 % minus 325 mesh and
a bulk sulphide roughing float follows 84 • Magnetic germanium (renierite)
is recovered from this float during passage through a series of Franz ferro-
magnetic filters, but recovery is low as the germanite also contained does not
respond to magnetic arrest.
Reagents used in lb/ton of sulphide ore treated in July 1954 were frother
(3 cresylic acid to I Powell frother) 0·11 ; collector (50/50 sodium ethyl xan-
thate and sodium isopropyl xanthate) 0·26; sodium cyanide 0·74; zinc
sulphate 1·75; soda-ash 0·76; lime 2·25; and copper sulphate 0·47. The
1954 CuPb concentrate grade was 11·35 % Cu and 50·88 % Pb, with 86 % and
95± % recovery. The zinc concentrate assayed 57 ·16 % Zn, 1-13 % Cu, and
2-47% Pb, with a zinc recovery of 62·35 %. Germanium is floated from the
CuPb concentrate after ten minutes. conditioning at pH 5·2 with 1-46 Ib/ton
of starch and 6 Ib (in three stages) of sulphurous acid. A quick roughing
float follows, and this product is cleaned after alkalising with 1·46 Ib/ton of
limetopH 10<1O-!-.
Ore
I1mmLon
I ---~~(T)
1 1
Bulk ROUgtng Float
(T)
I
l-BUlk Cleaning Float Oxide
Roughing
j Float ~
(T)
~
Dichromate-61 O~(T)
Roughing Float
. 1
T 1 T~1
Oxide
L I_ _ _ _ --~Dlchromate~610 I Cleaning
Cleaning Float and Recleaning
Float
1
I
. ~
ZInC conc. CuPb conc.
CuPb CuPb
Bulk Sepn. Zn Pyrite Cu Zn Pyrite Cu Pyrite As-Pyrite
The reagents used in the "high-oxide" section (Fig. 340) to July 1954 were
frother 0·06 lb/ton; ZA+343 0'46; Z.6 0·53; RA04 0'31; soda-ash 6'29;
sulphidizer 11'55; sodium silicate 0·83. The dichromate - 610 circuit was
temporarily out of action, but 1953 consumption had been 3·61 dichromate
and 0·17 R.610.
This flow-sheet utilises the fact that the sphalerite is already highly activated
and good recovery of the sulphidic values is possible in a bulk float. Copper
and lead are then depressed from this with a combination ofR.61O and sodium
dichromate, which allows the zinc to float from the depressed CuPb minerals.
The bulk float tailings are conditioned with sodium sulphide and then floated.
Fresh water is important to success, as cyanides and zinc sulphate, which are
present in the mill return water, have deleterious effects.
Boliden, Sweden
This central mill, built in 1951, treats three types of ore. Those from
Renstrom and Langsele yield copper, lead, zinc, and iron sulphides while the
Boliden ore body is milled for its copper, pyrite, and arsenopyrite. Part of
the auriferous arsenopyrite from Boliden is smelted without pre-treatment in
the mill. Renstrom ore is floated in three stages. First comes bulk CuFeS2
and PbS with precious metals, followed by galena depression with sodium
dichromate. Next is flotation of Zn(Fe)S. Finally the FeS2 is floated for its
sulphur (49% in the end product). The flow-sheet for the Langsele ore
yields copper, zinc, and pyrite in three stages of flotation. That for Boliden
produces successively copper, pyrite, and arsenopyrite. The reagent plan is
shown in Table 48.
Sherritt Gordon Mines, Ltd., Canada
At Lynn Lake, 150 miles north of Flin Flon, a nickel-copper-cobalt ore is
treated. The values lie in massive sulphides, disseminated sulphides, and
stringers. The troublesome gangue minerals are talc and actinolite. Talc
is depressed before floating the sulphides by the use of tetra-sodium pyro-
phosphate (0'4 lb/ton) with the amyl xanthate used in flotation (0'05) for a
five-minute conditioning period. A locally patented reagent called Guartec
is added (0'3) immediately ahead of flotation. The first flotation product is
a high-grade nickel copper concentrate containing 70 % of the nickel, mainly
as free pentialldite, and 85 % of the copper as chalcopyrite. These are
differentiated by depression of the nickel with lime and cyanide. The nickel
unfloated in the first bulk operation is recovered as a low-grade concentrate
after further flotation in a circuit made acid with S02. It is too intimately
associated with pyrrhotite for separation by grinding.
Differentiation of Non-Sulphides
A research Paper by Ejgeles 154 make the points that selectivity of non-
sulphides is achieved by depression rather than activation. The pulp
"climate" is important in this respect. Research has shown dextrin and
sodium silicate to be of value in the separation of fluorite from calcite;
dichromate and dextrin for fluorite from barite; starch and lignin sulphonate
for diaspore from pyrophyllite. All this work was done with softened water.
806 Mineral Processing-Selected Ore Treatments
References
1. (1959). Eng. and Min. J., May.
2. Scott, T. R. (1963). Min. and Chem. Eng. Review, May.
3. Denver Equip. Co. Bull., M.4-B.68.
4. Finn, W. K. (1963). World Mining, Jan.
5. Gartshore, J. L. (1962). Bull. Can. I.M.M., June.
6. (1963). World Mining, Feb.
7. Wiser, J. P. (1962). Can. I.M.M. Bull., June.
8. Wilson, J. N. (1960). I.M.P.e. (London), I.M.M. London.
9. Thorn, e. (1962). Froth Flotation, 50th Anniv. Vol., A.I.M.E.
10. Clement, M., and Klossel, E. (1963). I.M.P.e. (Cannes), Pergamon.
II. Everest, D. A., and Napier, E. (1963). I.M.P.e. (Cannes), Pergamon.
12. Moir, D. N., Collins, D. N., Curwen, H. e., and Manser, R. M. (1963).
I.M.P.C. (Cannes), Pergamon.
13. Browning, J. S. (1961). Min. Eng. (A.l.M.M.E.), July.
14. Misra, R. N., and Bhatnagar, P. P. (1963). Jnl. Nat. Met. Lab. India., Nov.
15. Phillips, W. M. (1963). Trans. S.M.E. (A.l.M.M.E.), June.
16. Greene, E. W., and Duke, J. B. (1962). Trans. S.M.E. (A.l.M.M.E.), Dec.
17. Jones, W. I. (1956). Royallnst. Chem., 3,
18. Geer, M. R., and Yancey, H. F. (1962). U.S. Dept. Int. Bur. of Mines.
I.e. 8093.
19. Gieseke, E. W. (1962). Trans. Amer. S.M.E., Dec.
20. Mason, G. A., and Armson, L. (1963). Steel and Coal, Feb.
21. Russell, J. A., and Laffin, J. J. (1961). Can. Min. and Met. Bull., Oct.
22. (1963). Horizon, Oct.
23. Formanek, V., and Lauvernier, J. (1963). I.M.P.e. (Cannes), Pergamon.
24. Mitchell, J. S. (1956). Mining Eng., Nov.
25. (1958). U.S. Bur. of Mines IC 5388, Feb.
26. Weed, R. e. (1956). Trans. A.l.M.M.E., July.
27. Argall, In., G. O. (1964). World Mining, Jan.
28. Denver Trefoil. (1962). Spring-Summer.
29. (1961). Eng. and Min. J., June.
30. (1963). Chamber of Mines Jnl., Rhodesia, Dec.
31. Jacobi, J. S. (1963). Min. Eng. (A.I.M.M.E.), Sept.
32. Markel, W. (1960). World Mining, May.
33. Knobler, R. R., and Joseph, W. (1962). Min. Eng. (S.M.E.), Jan.
34. Bean, J. J. (1960). Trans. A.l.M.E., Dec.
35. Bean, J. J. (1961). Qly. Colo. School of Mines, Vol. 56.
36. (1963). Bull. Can. I.M.M., Aug.
37. Last, Stevens and Eaton. (1957). Trans. A.l.M.E., Feb.
38. Ames. (1957). 6th Commonwealth Min. and Met. Congress.
39. Anon. (1957). Swedish Min. Dressing Plants, I.M.D.e., Stockholm, Almqvist
& Wiksell.
40. British Patent Spec. 602028.
41. Pryor, E. J., and Lowe, G. M. Trans. I.M.M. (London), 65, 67.
42. Talbot, H. L. (1946/7). Trans. I. M. M. (London), 56.
43. Coulter, R. F. (1962). Can. Min. and Met. Bull., June.
44. Hernlund, R. W. (1961). 50th Anniv. Froth Flotation (Colo. Sch. of Mines),
Vol. I.
45. Crabtree, E. H. (1963). Eng. and Min. J., Dec.
46. Deco Trefoil. (1963). May-July.
47. (1960). Eng. and Min. J., Nov.
48. Sutulov, A. (1962). World Mining, Aug.
49. Freeman, Rampacek and Evans. (1961). A.l.M.E., Oct.
50. Adamson, R. J. (1959). Jnl. S. Al I. M. M. , Aug.
51. Colvin, E., and Simpson, H. S. (1960). Jnl. S. Al I. M. M. , May.
52. Linari-Linholm, A. A. (1961). Jnl. S. AII.M.M., Feb.
53. Nesbitt, A. e., and Weavind, R. G. (1960). I.M.P.e. (London), I.M.M.
London.
Mineral Processing-Selected are Treatments 807
References-continued
54. Steygers, L. (1960). World Mining, Jan.
55. Mackenzie, D. C. Mining Magazine, 56.
56. Rickard, T. A. Trans. Can. I.M.M. 39.
57. Lydon, P. A. (1959). Mineral Information Service, California, Nov.
58. Woodcock, J. T. (1962). C.S.l.R.O. (Australia). "Ore Dressing lnves.
Report 628", March.
59. Duschak, L. H., and Schutte, C. N. Bull. U.S. Bur. Min. 222.
60. Mokkan, A. H. (1958). Jnl. S. Af I.M.M., Feb.
61. Cross, H. E., and Bragman, C. F. (1962). Jnl. S. Af I.M.M., Oct.
62. (1957). The Milling of Canadian Ores. 6th Commonwealth Min. and Met.
Congress.
63. Lean, J. B., and Warner, H. K. (1963). I.M.P.e. (Cannes), Pergamon.
64. Dahlem, D. H., and Sollenberger, C. L. (1963). I.M.P.e. (Cannes), Pergamon.
Press.
65. Boucraut, M., Koskas, R., and Michard, J. (1963), I.M.P.e. (Cannes),
Pergamon Press.
66. Meyer, K. (1963). I.M.P.e. (Cannes), Pergamon.
67. Hencl, V. (1963). I.M.P.e. (Cannes), Pergamon.
68. Frommer, D. W. (1964). Min. Eng. (S.M.E.), April.
69. Herzog, M. E., and Backer, M. L. (1963). I.M.P.e. (Cannes), Pergamon.
70. Syoberg, S. E. (1961). Mining World, Jan.
71. Giambra, S. P. (1963). Can. Min. and Met. Bull., May.
72. (1963). Bull. e.I.M.M., Aug.
73. (1963). Eng. and Min. Jnl., Dec.
74. Gagyi-Palfy, A., Palfy, G., and Halasz, H. (1963). I.M.P.e. (Cannes),
Pergamon.
75. Barkhahn, R. W. (1961). Mining World, Nov.
76. Argall, G. O. (1957). World Mining, March.
77. Rey, M., de Merre, P., Mancuso, R., and Formanek, V. (1961). 50th Anniv.
Froth Flotation. Colo. Sch. of Mines Qly., July.
78. Rey, Chataignon, and Formanek. Am. lnst. Min. and Met. Eng., 187.
79. Rey, M. (1957). I.M.P.e. (Stockholm), Almqvist & Wiksell.
80. McQuiston, F. W. (1957). I.M.P.e. (Stockholm), Almqvist & Wiksell.
81. (1957). The Milling of Canadian Ores. 6th Commonwealth Min. and Met.
Congress.
82. Anon. (1957). Swedish Mineral Dressing Mills, Almqvist & Wiksell.
83. Ratledge, J. P., Ong, J. N., and Boyce, J. H. (1955). Trans. A.I.M.M'£', 202.
84. Beall, J. V. (1962). Trans. A.l.M.M'£', Dec.
85. Bear, I. J. (1958). Chem. Eng. and Min. Review, Feb.
86. Dresher, W. H. (1963). S.M.£., Feb.
87. (1962). e.S.I.R.O. Report 717, June.
88. Yates, E. H. (1957). Min. Eng. A.I.M.E., Dec.
89. (1961). Abbott Mercury Mill, Denver Bull. M.4-B.\08.
90. Curtis, C. H. (1961). Min. Eng. A.l.M.E., Nov.
91. Hernlund, R. W. (1961). 50th Anniv. Froth Flotation. Colo. School of
Mines Qly., July.
92. (1964). World Mining, May.
93. Papin, J. E. Trans. A.l.M.E., 202.
94. Gaudin, A. M. (1957). Flotation, McGraw-Hili.
95. Noblitt, H. L. (1963). I.M.P.e. (Cannes), Pergamon.
96. Houston, W. M., and La Venue, W. A. (1962). Min. Eng. S.M.£., Nov.
97. Hoffman, I., and Mariacher, B. C. (1961). Min. Eng. S.M. Eng., May.
98. Sun, S. c., Snow, R. E., and Purcell, V. I. (1957). Trans. A.I.M.E.. 208.
99. Hughes, C. V. O. (1956). Min. Eng., Jan.
100. Hadzeriga, P. (1964). Trans. S.M.£., June.
101. Clark, H. N., and Reiter, J. S. (1961). 50th Anniv. Froth Flotation. Colo.
Sch. of Mines Qly., Vol. 56, July.
102. Jenkinson, D. W. (1960). A.I.M.M.E., July.
808 Mineral Processing-Selected Ore Treatments
References-continued
103. Hanes, F. E., and Wyman, R. A. (1962). Can. I.M.M., July.
104. (1939/42). U.S. Bur. of Mines Inv. 3436 and 3661.
105. Bryan, R. B. 50th Anniv. Froth Flotation. Colo. Sch. of Mines Qly., Vol. 50.
106. PoIlandt, E., and Pease, M.E. (1959/60). Trans. I.M.M. (London), 69.
107. Huillet, F. D., and Lewis, C. J. (1961). World Mining, JUly.
108. Dalton-Brown, H. (1960/61). Trans. I.M.M. (London), 70.
109. Chaston, I. R. M. (1961/2). Trans. I.M.M. (London), 71.
110. Burdon, R. G. (1964). Mining Magazine, May.
111. Eidsmo, 0., and MeIlgren, O. (1960). I.M.P.e. (London), I.M.M. London.
112. Runolinna, U., Rinne, R., and Kurronen, S. (1960). I.M.P.e. (London),
I.M.M. London.
113. Li, K. C. (1961). S.M.E. (A.l.M.E.), Preprint 61.B7.
114. Forward, F. A., and Vizsolyi, A. (1963). I.M.P.e. (Cannes), Pergamon.
115. Everhart, D. L. (1954). Eng. Min. i., 155.
116. Patching, S. W. F. (1956). Symp. I.M.M. (London), Paper 7.
117. Kaufran, L. A. Trans. Can. I.M.M., 53.
II 8. Read, Carman and Gough. (1957). inl. S. Af I.M.M., Feb.
119. Pryor, E. J. (1963). Dictionary of Mineral Technology, Mining Publications.
120. Fitzgerald and Kelsall. (1956). "Extraction and Refining of the Rarer
Minerals". Symposium, I.M.M. (London), March.
121. Levin, J. (1957). I.M.P.e. (Stockholm), Paper VI.
122. Lord and Light. (1956). Can. Min. and Met. Bull., Feb.
123. Arden, T. V. (1957). "Symposium. Extraction and Refining of Rare
Minerals",I.M.M.
124. Butler, J. N., and Morris, R. J. (1956). Trans. A.l.M.M.E., Oct.
125. Tinker, et al. (1957). Precambrian, 30, Aug.
126. Honeywell, W. R., and Harrison, V. F. (1963). Can. Min. and Met. Bull.,
Aug.
127. Somnay, J. Y., and Light, D. E. (1963). Amer. S.M.£., March.
128. Behan, E.L. (1956). Can. Min. and Met. Bull., April.
129. (1955). Mining World, July.
130. McLean and Prentice. (1955). Geneva Conf on Peaceful Use of Atomic
Energy.
131. McCarty, M. F., and Olsen, R. S. (1956). Eng. Min. i .. , Nov.
132. (1957). World Mining, Jan.
133. ArgaIl, G. O. (1957). Mining World, Dec.
134. Hargrove, D. (1958). Eng. Min. i., March.
135. Maltby, P. D. R. (1959/60). Trans. I.M.M. (London), 69.
136. Bhappu, R. B. 50th Anniv. Froth Flotation. Colo. Sch. of Mines Qly.,
Vol. 56.
137. Nugent, Hoogervorst and Wheeler. (1960). S. Af I.M.M., Feb.
138. Miller, R. P., Napier, E., WeIls, R. A., Audsley, A., and Daborn, G. R. (1963).
Bull. I.M.M. (London), 72.
139. Hannay, R. L. (1956). Can. Min. inl., June.
140. Black and Koslov. (1958). Inst. Chem. Eng., Chicago.
141. (1963). Can. Min. inl., July.
142. Edwards, I. c., and Smith, D. C. (1962). Bull. Can. I.M.M., Oct.
143. (1962). Trans. Can. I.M.M., March.
144. Hazen, W. C. (1963). Denver Bull. N. T4-B.32, Oct.
145. Seeton, F. A. (1963). Denver Bull. No. M4. B.l20, Nov.-Dec.
146. Garman, C. N. (1962). Trans. A.l.M.E., 223, June.
147. Rey et al. (1954). Trans. A.l.M.M.E., 199, April.
148. Rey, M., and Formanek, V. (1960). I.M.P.e. (London), Paper 18.
149. Billi, M., and Quai, V. (1963). I.M.P.e. (Cannes), Paper 43 J.
150. Lewis, J. L. (1961). 50th Anniv. of Froth Flotation. Colo. School of Mines.
151. Bearce, W. E. (1961). Ibid.
152. Wyman, R. A. (1958). Trans. A.I.M.M.E., 21 I.
153. American Cyanamid Co.
154. Ejgeles, M. A. (1957). T.M.P.e. (Stockholm), Almqvist & Wiksell.
APPENDIX A
GLOSSARY
Belt Conveyor.-An endless belt on which ore is placed and moved from point
to point in the transport system.
Benejiciation.-Treatment of crude ore in order to improve its quality for
some required purpose.
Blocking (0/ Crusher).-Obstruction of crushing zone by clayey material or
rock which refuses to break down and pass to discharge.
Break.-Optimum mesh-of-grind (q. v.).
Bubble.-The dictionary definition of "bubble" is modified in this book to
define two physical states:
(a) N-Bubble. A bubble immersed in water or pulp, with surface
tension at or approaching that of water.
(b) F-Bubble (usually written "bubble" in text). An emerged free
bubble of surface tension substantially below that of water and
having an area twice that of an N-bubble of equal diameter.
Bubble Pipe.-Tube inserted in pulp at regulated depth, through which com-
pressed air is gently bubbled. The air-pressure indicates the pulp density
and provides a means of control.
Bulk Flotation.-The intentional raising as a mineralised froth of more than
one mineral in one operation.
r: [. :' ::j
progressively enriched as it moves in the opposite direction.
Country Rock.-The valueless rock surrounding a lode.
Coursing Bubble.-One rising freely through cell during froth flotation.
C'U;::~::~~:m::~::l~::n::s:::n~::;
where n = r.p.m., and d the mill diameter in feet.
Crop Load.-The mixture of crushing bodies, ore particles, and water being
tumbled in the ball mill.
Crystallography.-Science of the shapes, colour, chemical, electrical, physical,
and optical properties of crystals.
Customs Mill.-A plant receiving ore for treatment from more than one mine.
Cyanicide.-Any substance present in a pulp which attacks or destroys the
cyanide salt being used to dissolve precious metals.
Cyanide.-Usually refers to cyanide solution in circulation in a mill treating
gold or silver ores. The "stock" or "solution" is of two main types,
"barren" from which all possible value has been extracted, and "pregs"
or "pregnant", which is charged with gold or silver and awaits their
removal.
Cyclone.-Hydro-cyclone. A classifying (or concentrating) separator into
which pulp is fed, so as to take a circular path. Coarser and heavier
fractions of solids report at apex of long cone while finer particles over-
flow from central vortex.
Lode.-An ore body occurring as a steep vein or layer filling a rock fissure.
Long-ranged.-Of crushed or ground ore, size distribution covering a wide
range of meshes.
Lyophilic.-Having the property of attracting liquids.
Magnetic Separation.-Use of permanent or electro-magnets to remove rela-
tively strong ferro-magnetic particles from para- and dia-magnetic ores.
Mesh.-The aperture framed by the wires of a sieve opening. With square-
woven sieves the number of wires per linear inch defines the mesh-e.g.•
30 mesh means that there are thirty wires per inch-run.
When crushed ore is fed on to two sieves, the lower being the finer, the
coarse sieve defines the limiting mesh and the fine one the retaining mesh
of the sand held between the screens. It can be defined as "-12 + 120~",
the + 12 for this particular combination being called oversize and the
-120~ undersize.
The terms "long-ranged" or "short-ranged" are sometimes used to
indicate roughly the cut-off limits of the sand sizes in an aggregate of
particles.
Mesh-of-grind.-Written m.o.g. It states optimum particle size in grinding,
in terms of the percentage of the material passing (or alternatively retained
on) a given screen-mesh.
Metallurgical Balance Sheet.-Statistical presentation of an operation in terms
of the general equation:
Units of value in feed = Units in concentrate + units in tailings
+ units retained in circuit.
An equation is needed for each valuable mineral undergoing extraction.
Micelle.-(Part}. A colloidal aggregate of molecules present in aqueous
solutions of many soaps and dyestuffs.
Middlings.-In two-component ore, particles incompletely liberated by com-
minution into concentrate or gangue. In complex ores, in addition to
incomplete liberation, there may be muItiphased particles of middling,
or intermediate species which react too feebly to treatment to report as
concentrate or tailing.
Mill (Milling ).-Also concentrator, plant, reduction works. The place where
run-of-mine ore is received and treated, and from which end-products are
dispatched.
Mill Head.-Ore accepted for treatment in concentrator, after any preliminary
rejection such as waste removal.
Mineral Dressing.-Treatment of natural ores or partly processed products
derived from such ores in order to segregate or up-grade some or all of
their valuable constituents, and/or to remove those not desired by the
industrial user. Mineral-dressing processes are applied to industrial
waste to retrieve useful by-products.
Mineral Species.-Native element or inorganic substance corresponding
generally with ideal mineral having a precise formula and known lattice
structure.
Muck-shifting.-Term used for extensive earth-moving operations.
Near-mesh.-Near-sized; grains close in cross-section to a specified screening
mesh, which tend to blind apertures and slow down sizing.
Nine-Point Samp/e.-At final stage of sample preparation, manipulation of
small quantity of finely ground mineral. Material is rolled on glazed
cloth or paper, flattened into a disc, and divided into eight equal segments.
Final sample is drawn equally from each segment and from centre.
Mineral Processing-Glossary 815
Nip, Angle 0[-The angle of nip, or wedge angle, is that below which a piece
of ore can be seized between two crushing surfaces without slipping.
Opencast.-A working place open to the sky, and from which ore is extracted.
Open Circuit.-A flow-line in which the solid particles pass from one appliance
to the next without being screened, classified, or otherwise checked for
quality, no fraction being returned for retreatment.
Optimisation.-Co-ordination of various processing factors, controls and
specifications to provide best overall conditions for technical and/or
economic operation.
Optimum .-The use of this word in mineral dressing is associated with the idea
of quality control. As an example, a pulp might be required to be de-
livered from one section to the next at an optimum pH, mesh size, and
percentage of solid plus or minus 10% to ensure smooth working of the
receiving section. The + or - tolerance must be acceptable to the
treatment which follows.
Ore.-A naturally occurring complex of minerals from which any fraction of
commercial value can be extracted and used. "Raw" and "run-of-
mine" ore is ore as it comes from the working place.
Overburden.-Valueless earth or rock overlying the ore body.
Over-grinding.-Reduction of size of particles of a given mineral below that
aimed at in comminution.
Oversize.-See Mesh.
Ragging.-See Bedding.
816 Mineral Processing-Glossary
dumps, and in ore deposits which have been exposed to climatic action
and include clay, alumina, hydrated iron, near-colloidal common earths,
and weathered feldspars.
"Secondary" slimes are very finely ground minerals from the true ore.
Sluicing.-Separation of minerals in a flowing stream of water.
Solution.-See Cyanide.
Solvation.-A combination between solute and solvent (Hackh).
Sorption.-Reaction at or immediately adjacent to, the surface of a solid.
AdS, monomolecular; AbS, deeper physical penetration; ChemiS,
sorption with chemical action and change.
Sorting.-Removal by hand (hand picking) of selected pieces of rock. Term
also applied to classification (q. v.) of finely ground pulps.
Specific Population.-Number of particles in unit volume of pulp.
Specific Surface.-Total area of particles in unit-weight of ground material,
usually reported in cm. 2/gm.
Specimens.-In mineral dressing, unusually rich pieces of ore or characteristic
constituents thereof in coarsely crystalline form-not representative
samples.
Spills, Spillage.-Ore, pulp, circulating liquor inadvertently discharged from
flow-line and requiring appropriate means of recovery or removal.
Stage.-Stage-grinding is successive comminution. Stage-concentration is
stage-grinding repeated on the concentrate produced by treatment
between grinding stages. Stage-addition in flotation refers to deliberate
use of insufficient reagent in the early part of the treatment, in order to
increase selectivity of conditioning, followed by further addition at a
later point in the process.
Starvation.-As used in flotation, the deliberate inadequate addition of a
reagent in order to restrict its effect. In comminution, avoidance of
crowding in the machine by restricting rate of feed.
Stope.-Underground working place formed as ore is extracted.
Stope Assay Plans.-Plan showing assay value of exposures of ore in a stope,
together with any other data desired,.
Subsieve Sizing.-Size analysis of particles too small for efficient grading by
use of screens, usually minus 200 mesh.
Substrate.-The true lattice of a crystal, as distinct from its discontinuity
lattice, or surface.
Superpanner.-Mechanism invented by Professor H. T. Haultain which-simu-
lates rocking, bumping, and sluicing action used in panning, and gives
precise information as to possibility of gravity treatment of sands. Used
in rapid assays and as a research aid.
Surge (tanks, bins).-Receptacles capable of receiving and redispensing small
tonnages, thus steadying any fluctuations in a flow-line.
Sweet Roasting.-Removal of sulphur from metal sulphides by ignition.
Tailing.-See Gangue.
Teeter.-The dance of a bed of particles in a column of rising water, or teeter
column.
Theta (8).-The symbol for contact-angle in flotation testing.
Thickener.-Large circular tank in which solids settle slowly and form a slurry
which is continuously removed from below while fairly clear water over-
flows.
Thixotropy.-Property of certain mineral suspensions in water (e.g., bentonite)
of remaining fluid while agitated, but gelling when quiescent.
818 Mineral Processing-Glossary
BEAKER DECANTATION
METHOD OF PROCEDURE
No special apparatus is required. The equipment consists of two 2-litre
beakers, five I-litre beakers, a glass rod covered with rubber tubing, a stop-
watch and, if available, a mechanical stirrer.
(I) Using the mechanical stirrer, 100 g. of the material to be sized is dis-
persed in a litre of water in one of the 2-litre beakers for a suitable period, with
the aid of a deftocculant. The stirrer is removed when dispersion appears
complete.
(2) Water is added until the surface of the suspension is about 5 cm. below
the lip of the beaker.
(3) The height from the surface of the pulp to the level of the settled sand is
measured (hi cm.) and the time for the smallest particle (d microns in diameter)
required in the coarsest fraction to settle this distance is calculated according
to Stokes's law (x cm./sec., giving a settling time of!!2 sec.). The pulp is
x
stirred until all particles are in suspension. The maximum possible vertical
motion is imparted to the pulp with the minimum of circular motion in a
horizontal plane. When the suspension is uniform, the outside of the beaker is
sharply tapped and the suspended particles allowed to settle for the calculated
time multi pled by 1·05 (i.e. ~ x 1·05 sec.). This increase in settling time over
x
that calculated ensures that all particles at and above the required size have
settled after allowing for hindered settling. The beaker is vigorously tapped
with the rod for 15 sec. before the end of the settling time. This tapping
forms the settled material into a compact cake and ensures that particles built
up as a thin layer in the curve between the sides and bottom of the beaker join
the main cake.
(4) At the end of hi x 1·05 sec, the supernatant pulp IS
. poured ' kl y mto
qUlC .
x
the second 2-litre beaker (B). The settled particles remain as a compact cake
on the bottom of beaker A. This cake now contains all the particles of size d
and above, but the fraction is stiII contaminated with residual "fines".
(5) Five I-litre beakers (C, D, E, F, and G) containing water and dispersant
are prepared and the solution from the first of these beakers, C is poured on t.o
the cake in A. The depth from the liquid surface to the sand surface IS
820 Mineral Processing-Beaker Decantation
measured (h2 cm.) and the time for a particle of size a to settle this distance is
calculated h2 sec.. The pulp is brought into uniform suspension by stirring
x
as before, the beaker tapped, and the particles allowed to settle for ~sec.,
x
the correction for hindered settling not being required in this thinner pulp;
15 sec. before this time expires the beaker is tapped vigorously. At~sec.
x
the supernatant pulp containing only those particles finer than d is poured
back into the beaker C.
The settled material in A now contains the required fraction with a reduced
proportion of that finer material which has settled in the first stage.
(6) Repeating the process carried out with beaker C by using the other four
beakers D, E, F, and G, the material less than d contained in the cake is suc-
cessively reduced.
If, on adding more wash water and stirring, all the particles settle within the
calculated time, then theoretically no particles less than d will be found in the
cake, and conversely, if, when the final wash water is added, particles are still
in suspension at the end of the calculated time, then some particles finer than d
will be found in the settled material. Observation indicates that five washes are
sufficient.
(7) After the final wash from the fifth beaker (G), the cake is removed from
beaker A, dried, and weighed.
(8) Beaker B now contains most of the particles smaller in size than d, the
remainder being in the wash solutions. The height from pulp surface to that
of the settled sand is measured and the time for a particle :2 to settle this
cate head samples are set out in the Table together with the nominal micron
sizes calculated from Stokes's law. The total time for one complete frac-
tionation was 5t hours. Losses were due to the difficulty in collecting all the
fines and were therefore added to the minus 9-micron fraction.
TEST 1 TEST 2
Size in
Cut Microns Cum. per Cum. per
Number (calc.) Wt. WI.
per cent. cent. Fines cent. Fines
x ·52 per cent. x ·52
I +59 32·75 35·1 31·20 35·8
2 -59+45 15·55 26·9 16·70 27·1
3 -45+ 32 14·75 19·2 15·00 19·3
4 -32+21 11·20 IN 11·85 13·1
5 -21 + 15 7-60 9-4 7·55 9·2
6 -15+ 9 5·85 6-4 5·25 6·5
- 9 12·30 IN5
100·00 100·00
Although a total of five washes was given, it was noticed that when treating
material finer than 32 microns the final wash still contained particles in sus-
pension.
APPENDIX C
ABBREVIATIONS
A barrel, 744
gold, 416, 736
Absorption, see Sorption health hazards, 746
Acceptance into mill, 14 pan, 743
Acid cure, 784 patio process, 743
Activation, see also Flotation and Modi- plates, 741
fiers practice, 741
common-ion, 498 product treatment, 746
copper-ions in, 499 sick mercury, 619
by re-surfacing, 499 tests, 618
and solubility, 469 Amberlite, see Resin
sulphides, 495 Amenability, see Glossary, 264, 609
theory, 468 Amex, see Uranium
Adsorption, see Sorption Amine, see Reagents
Aeration, in cyanide process, 414, 426, Andreason Pipette, 165
441 Angle of nip, 23
in flotation, 514 and rolls, 65
methods, 502 Angle of repose, 55
Aerofall, see Mill Anion, see Ions
Aerophil, 393, 472 Antimony ore treatment, stibnite, 706
Aerophobe,393,472 Arrastres, 743
Aerosol, see Frothers Arsenic ore treatment, arsenopyrite, 706
Agitator, Agitation, see also Gold Asbestos, milling of, 128, 706
Brown, 426 Assay, complete, 603, 609
Dorr,427 test control, 630
gold "slimes", 422 Automation, see Instrumentation
machines, 426
Pachuca, 426
of pulp, 672 B
as reaction control, 392,415
of slimes for cyanide, 672 Bacteria, see Leaching
Wallace, 427 Balls, see Crushing bodies
Agglomeration, see also Flotation Barium, barite treatment, 707
on conveyor-belt, 552 Barren solution, see Glossary
and de-sliming, 552 Base mill, 109, 613
general, 460, 551, 514 Batch tests, see Tests
glomerules, 514, 551 Baum, see Jig
tables, 552 Beaker decantation, see Appendix B
Air, 372, see also Aeration, gas Bedding, see Glossary
lance, see Glossary Beilby flow, 465, 606
sizing, 228 Belt conveyor, see Conveyor
Allen, equation, 170, 357 Beryllium, concentration of ores, 708
All-sliming, see Glossary B.E.T., see Adsorption
Aluminium ore treatment Bins,
bauxite, 702 blending from, 92
Bayer process, 702 ore, 91, 669
corundum, 703 segregation in, 91, 669
cryolite, 704 surge, 54
feldspars, 704 Blake crusher, see Crusher
kyanite, 705 Blyth elutriator, 163
sillimanite, 706 Bond,
Amalgamation, 258, see also Gold co-ordinate, 396
amalgam, 407 covalent, 372, 395
826 Mineral Processing-Index