Wills Mineral Processing Technology Elsevier

Esta é uma pré-visualização de arquivo. Entre para ver o arquivo original

Mineral Processing Technology 
An Introduction to the Practical Aspects of Ore Treatment and Mineral 
Recovery, by Barry A. Wills, Tim Napier-Munn 
 
 
 
 
· ISBN: 0750644508 
· Publisher: Elsevier Science & Technology Books 
· Pub. Date: October 2006 
 
Preface to 7th Edition 
Although mining is a conservative industry, economic drivers continue to encourage innovation and 
technological change. In mineral processing, equipment vendors, researchers and the operations them- 
selves work to develop technologies that are more efficient, of lower cost and more sustainable than 
their predecessors. The results are apparent in new equipment and new operating practice. Any textbook 
needs to reflect these changes, and Barry Wills' classic is no exception. 
It is nearly 30 years since Mineral Processing Technology was first published, and it has become the 
most widely used English-language textbook of its kind. The sixth edition appeared in 1997 and Barry 
and his publishers felt that it was again time to bring the text up to date. They approached the Julius 
Kruttschnitt Mineral Research Centre at the University of Queensland to take on the challenging task. 
My colleagues and I agreed to do so with some trepidation. The book's well-deserved reputation and 
utility were at stake, and the magnitude of the task was clear. Revising someone else's text is not an 
easy thing to do successfully, and there was a real danger of throwing the baby out with the bath water. 
The value of Mineral Processing Technology lies in its clear exposition of the principles and practice 
of mineral processing, with examples taken from practice. It has found favour with students of mineral 
processing, those trained in other disciplines who have converted to mineral processing, and as a reference 
to current equipment and practice. It was important that its appeal to these different communities be 
preserved and if possible enhanced. We therefore adopted the following guidelines in revising the book. 
The 7th edition is indeed a revision, not a complete re-write. This decision was based on the view 
that "if it ain't broke, don't fix it". Each diagram, flowsheet, reference or passage of text was considered 
as follows. If it reflected current knowledge and practice, it was left unchanged (or modestly updated 
where necessary). If it had been entirely superseded, it was removed unless some useful principle or 
piece of history was being illustrated. Where the introduction of new knowledge or practice was thought 
to be important to preserve the book's currency, this was done. As a consequence, some chapters remain 
relatively unscathed whereas others have experienced substantial changes. 
A particular problem arose with the extensive references to particular machines, concentrators and flow- 
sheets. Where the point being illustrated remained valid, these were generally retained in the interest of 
minimising changes to the structure of the book. Where they were clearly out of date in a misleading sense 
and/or where alternative developments had attained the status of current practice, new material was added. 
It is perhaps a measure of Barry Wills' original achievement that it has taken more than a dozen people 
to prepare this latest edition. I would like to acknowledge my gratitude to my colleagues at the JKMRC 
and elsewhere, listed below, for subscribing their knowledge, experience and valuable time to this good 
cause; doing so has not been easy. Each chapter was handled by a particular individual with expertise in 
the topic (several individuals in the case of the larger chapters). I must also thank the editorial staff at 
Elsevier, especially Miranda Turner and Helen Eaton, for their support and patience, and Barry Wills for 
his encouragement of the enterprise. My job was to contribute some of the chapters, to restrain some of 
the more idiosyncratic stylistic extravagancies, and to help make the whole thing happen. To misquote 
the great comic genius Spike Milligan: the last time I edited a book I swore I would never do another 
one. This is it. 
Tim Napier-Munn 
December 2005 
Contributors 
Chapters 1, 4, 9, 11, 14 
Chapters 2 and 16 
Chapter 3 
Chapters 5 and 6 
Chapter 7 
Chapter 8 
Chapters 10, 13 and 15 
Chapter 12 
Appendix III 
Prof. Tim Napier-Munn (JKMRC) 
Dr Glen Corder (JKTech) 
Dr Rob Morrison (JKMRC) and Dr Michael Dunglison (JKTech) 
Dr Toni Kojovic (JKMRC) 
Dr Frank Shi (JKMRC) 
Marko Hilden (JKMRC) and Dean David (GRD Minproc, formerly with 
JKTech) 
Dr Peter Holtham (JKMRC) 
Dr Dan Alexander (JKTech), Dr Emmy Manlapig (JKMRC), Dr Dee 
Bradshaw (Dept. Chemical Engineering, University of Cape Town) and 
Dr Greg Harbort (JKTech) 
Dr Michael Dunglison (JKTech) 
Acknowledgements 
Secretarial assistance 
Other acknowledgements 
Vynette Holliday and Libby Hill (JKMRC) 
Prof. J-P Franzidis (JKMRC) and Evie Franzidis for their work on an 
earlier incarnation of this project. 
Dr Andrew Thornton and Bob Yench of Mipac for help with aspects 
of process control. 
The Julius Kruttschnitt Mineral Research Centre, The University of 
Queensland, for administrative support. The logos of the University 
and the JKMRC are published by permission of The University of 
Queensland, the Director, JKMRC. 
 
Table of Contents 
 
1 Introduction 1 
2 Ore handling 30 
3 Metallurgical accounting, control and simulation 39 
4 Particle size analysis 90 
5 Comminution 108 
6 Crushers 118 
7 Grinding mills 146 
8 Industrial screening 186 
9 Classification 203 
10 Gravity concentration 225 
11 Dense medium separation (DMS) 246 
12 Froth flotation 267 
13 Magnetic and electrical separation 353 
14 Ore sorting 373 
15 Dewatering 378 
16 Tailings disposal 400 
App. 1 Metallic ore minerals 409 
App. 2 Common non-metallic ores 421 
App. 3 Excel spreadsheets for formulae in chapter 3 
 
Introduction 
Minerals and ores 
Minerals 
The forms in which metals are found in the crust 
of the earth and as sea-bed deposits depend on 
their reactivity with their environment, particularly 
with oxygen, sulphur, and carbon dioxide. Gold and 
platinum metals are found principally in the native 
or metallic form. Silver, copper, and mercury are 
found native as well as in the form of sulphides, 
carbonates, and chlorides. The more reactive metals 
are always in compound form, such as the oxides 
and sulphides of iron and the oxides and silicates of 
aluminium and beryllium. The naturally occurring 
compounds are known as minerals, most of which 
have been given names according to their composi- 
tion (e.g. galena- lead sulphide, PbS; sphalerite - 
zinc sulphide, ZnS; cassiterite- tin oxide, SnO2). 
Minerals by definition are natural inorganic 
substances possessing definite chemical compo- 
sitions and atomic structures. Some flexibility, 
however, is allowed in this definition. Many 
minerals exhibit isomorphism, where substitution 
of atoms within the crystal structure by similar 
atoms takes place without affecting the atomic 
structure. The mineral olivine, for example, has 
the chemical composition (Mg, Fe)2 SiO4, but the 
ratio of Mg atoms to Fe atoms varies in different 
olivines. The total number of Mg and Fe atoms 
in all olivines, however, has the same ratio to 
that of the Si and O atoms. Minerals can also 
exhibit polymorphism, different minerals having 
the same chemical composition, but markedly 
different physical properties due to a difference in 
crystal structure. Thus, the two minerals graphite 
and diamond have exactly the same composi- 
tion, being composed entirely of carbon atoms, 
but have widely different properties due to the 
arrangement of the carbon atoms within the crystal 
lattice. The term "mineral" is often used in a 
much more extended
sense to include anything 
of economic value which is extracted from the 
earth. Thus, coal, chalk, clay, and granite do not 
come within the definition of a mineral, although 
details of their production are usually included in 
national figures for mineral production. Such mate- 
rials are, in fact, rocks, which are not homoge- 
neous in chemical and physical composition, as 
are minerals, but generally consist of a variety 
of minerals and form large parts of the earth's 
crust. For instance, granite, which is one of the 
most abundant igneous rocks, i.e. a rock formed by 
cooling of molten material, or magma, within the 
earth's crust, is composed of three main mineral 
constituents, feldspar, quartz, and mica. These 
three homogeneous mineral components occur in 
varying proportions in different parts of the same 
granite mass. 
Coals are not minerals in the geological sense, 
but a group of bedded rocks formed by the accu- 
mulation of vegetable matter. Most coal-seams 
were formed over 300 million years ago by 
the decomposition of vegetable matter from the 
dense tropical forests which covered certain areas 
of the earth. During the early formation of the 
coal-seams, the rotting vegetation formed thick 
beds of peat, an unconsolidated product of the 
decomposition of vegetation, found in marshes 
and bogs. This later became overlain with shales, 
sandstones, mud, and silt, and under the action 
of the increasing pressure and temperature and 
time, the peat-beds became altered, or metamor- 
phosed, to produce the sedimentary rock known 
as coal. The degree of alteration is known as 
the rank of the coal, the lowest ranks (lignite or 
brown coal) showing little alteration, while the 
highest rank (anthracite) is almost pure graphite 
(carbon). 
2 Wills' Mineral Processing Technology 
Metallic ore processing 
Metals 
The enormous growth of industrialisation from the 
eighteenth century onward led to dramatic increases 
in the annual output of most mineral commodi- 
ties, particularly metals. Copper output grew by a 
factor of 27 in the twentieth century alone, and 
aluminium by an astonishing factor of 3800 in the 
same period. Figure 1.1 shows the world produc- 
tion of aluminium, copper and zinc for the period 
1900-2002 (data from USGS, 2005). 
All these metals suffered to a greater or 
lesser extent when the Organisation of Petroleum 
Exporting Countries (OPEC) quadrupled the price 
of oil in 1973-74, ending the great postwar indus- 
trial boom. The situation worsened in 1979-81, 
when the Iranian revolution and then the Iran-Iraq 
war forced the price of oil up from $13 to nearly 
$40 a barrel, plunging the world into another and 
deeper recession, while early in 1986 a glut in the 
world's oil supply cut the price from $26 a barrel 
in December 1985 to below $15 in 1986. Iraq's 
invasion of Kuwait in 1990 pushed the price up 
again, from $16 in July to a peak of $42 in October, 
although by then 20% of the world's energy was 
being provided by natural gas. 
In 1999, overproduction and the Asian economic 
crisis depressed oil prices to as low as $10 a barrel 
from where it has climbed steadily to a record 
figure of over $60 a barrel in 2005, driven largely 
by demand especially from the emerging Asian 
economies, particularly China. 
These large fluctuations in oil prices have had 
a significant impact on metalliferous ore mining, 
due to their, influence both on the world economy 
and thus the demand for metals, and directly on the 
energy costs of mining and processing. It has been 
estimated that the energy cost in copper produc- 
tion is about 35% of the selling price of the metal 
(Dahlstrom, 1986). 
The price of metals is governed mainly by 
supply and demand. Supply includes both newly 
mined and recycled metal, and recycling is now 
a significant component of the lifecycle of some 
metals- about 60% of lead supply comes from 
recycled sources. There have been many prophets 
of doom over the years pessimistically predicting 
the imminent exhaustion of mineral supplies, the 
most extreme perhaps being the notorious "Limits 
to Growth" report to the Club of Rome in 1972, 
which forecast that gold would run out in 1981, zinc 
in 1990, and oil by 1992 (Meadows et al., 1972). 
In fact major advances in productivity and tech- 
nology throughout the twentieth century greatly 
increased both the resource and the supply of 
newly mined metals, through geological discovery 
and reductions in the cost of production. This 
actually drove down metal prices in real terms, 
which reduced the profitability of mining compa- 
nies and had a damaging effect on economies 
heavily dependent on mining, particularly those 
in Africa and South America. This in turn drove 
further improvements in productivity and tech- 
nology. Clearly mineral resources are finite, but 
supply and demand will generally balance in such 
Figure 1.1 World production of aluminium, copper and zinc for the period 1900-2002 
a way that if supplies decline or demand increases, 
the price will increase, which will motivate the 
search for new deposits, or technology to render 
marginal deposits economic, or even substitution 
by other materials. 
Interestingly gold is an exception, its price 
having not changed much in real terms since 
the sixteenth century, due mainly to its use as 
a monetary instrument and a store of wealth 
(Humphreys, 1999). 
Estimates of the crustal abundances of metals are 
given in Table 1.1 (Taylor, 1964), together with the 
actual amounts of some of the most useful metals, 
to a depth of 3.5 km (Chi-Lung, 1970). 
The abundance of metals in the oceans is related 
to some extent to the crustal abundances, since 
they have come from the weathering of the crustal 
rocks, but superimposed upon this are the effects 
of acid rain-waters on mineral leaching processes; 
thus the metal availability from sea-water shown in 
Table 1.2 (Chi-Lung, 1970) does not follow 
precisely that of the crustal abundance. The sea- 
bed may become a viable source of minerals in the 
future. Manganese nodules have been known since 
the beginning of the nineteenth century (Mukherjee 
et al., 2004), and recently mineral-rich hydrothermal 
vents have been discovered and plans are being made 
to mine them (Scott, 2001). 
It can be seen from Table 1.1 that eight elements 
account for over 99% of the earth's crust; 74.6% is 
Introduction 3 
Table 1.2 Abundance of metal in the oceans 
Element Abundance Element 
in sea-water 
(tonnes) 
Abundance 
in sea-water 
(tonnes) 
Magnesium 
Silicon 
Aluminium 
Iron 
Molybdenum 
Zinc 
Tin 
Uranium 
Copper 
Nickel 
1015-1016 Vanadium } 
1012-1013 Titanium 
Cobalt } 
101~ Silver 
Tungsten 
Chromium / 
Gold 
Zirconium 
109-10 l~ Platinum 
109_1010 
1012_1013 
<10 8 
silicon and oxygen, and only three of the industri- 
ally important metals (aluminium, iron, and magne- 
sium) are present in amounts above 2%. All the 
other useful metals occur in amounts below 0.1%; 
copper, for example, which is the most important 
non-ferrous metal, occurring only to the extent of 
0.0055%. It is interesting to note that the so-called 
common metals, zinc and lead, are less plentiful 
than the rare-earth metals (cerium, thorium, etc.). 
It is immediately apparent that if the minerals 
containing the important metals were uniformly 
distributed throughout the earth, they would be 
so thinly dispersed that their economic extraction 
would be impossible. However, the occurrence of 
minerals in nature is regulated by the geological 
Table 1.1 Abundance of metal in the oceans 
Element Abundance (%) Amount in 
3.5km of 
crust (tonnes) 
Element Abundance (%) Amount in 
3.5 km of crust 
(tonnes) 
(Oxygen) 46.4 
Silicon 28.2 
Aluminium 8.2 
Iron 5.6 
Calcium 4.1
Sodium 2.4 
Magnesium 2.3 
Potassium 2.1 
Titanium 0.57 
Manganese 0.095 
Barium 0.043 
Strontium 0.038 
Rare earths 0.023 
Zirconium 0.017 
Vanadium 0.014 1014-1015 
Chromium 0.010 
1016-1018 Nickel 0.0075 
Zinc 0.0070 
Copper 0.0055 1013-1014 
Cobalt 0.0025 
1016-1018 Lead 0.0013 
Uranium 0.00027 
Tin 0.00020 
1015-1016 Tungsten 0.00015 1011-1013 
Mercury 8 • 10 -6 
Silver 7 • 10 -6 
Gold <5 x 10 -6 / 
1014--1016 Platinum metals <5 • 10 -6 [ <1011 
4 Wills' Mineral Processing Technology 
conditions throughout the life of the mineral. A 
particular mineral may be found mainly in asso- 
ciation with one rock type, e.g. cassiterite mainly 
associates with granite rocks, or may be found asso- 
ciated with both igneous and sedimentary rocks 
(i.e. those produced by the deposition of material 
arising from the mechanical and chemical weath- 
ering of earlier rocks by water, ice, and chemical 
decay). Thus, when granite is weathered, cassiterite 
may be transported and re-deposited as an alluvial 
deposit. 
Due to the action of these many natural agencies, 
mineral deposits are frequently found in sufficient 
concentrations to enable the metals to be profitably 
recovered. It is these concentrating agencies and the 
development of demand as a result of research and 
discovery which convert a mineral deposit into an 
ore. Most ores are mixtures of extractable minerals 
and extraneous rocky material described as gangue. 
They are frequently classed according to the nature 
of the valuable mineral. Thus, in native ores the 
metal is present in the elementary form; sulphide 
ores contain the metal as sulphides, and in oxidised 
ores the valuable mineral may be present as oxide, 
sulphate, silicate, carbonate, or some hydrated form 
of these. Complex ores are those containing prof- 
itable amounts of more than one valuable mineral. 
Metallic minerals are often found in certain associ- 
ations within which they may occur as mixtures of a 
wide range of particle sizes or as single-phase solid 
solutions or compounds. Galena and sphalerite, for 
example, associate themselves commonly, as do 
copper sulphide minerals and sphalerite to a lesser 
extent. Pyrite (FeS2) is very often associated with 
these minerals. 
Ores are also classified by the nature of their 
gangues, such as calcareous or basic (lime rich) 
and siliceous or acidic (silica rich). An ore can be 
described as an accumulation of mineral in suffi- 
cient quantity so as to be capable of economic 
extraction. The minimum metal content (grade) 
required for a deposit to qualify as an ore varies 
from metal to metal. Many non-ferrous ores 
contain, as mined, as little as 1% metal, and often 
much less. 
Gold may be recovered profitably in ores 
containing only 1 part per million (ppm) of the 
metal, whereas iron ores containing less than 
about 45% metal are regarded as of low grade. 
Every tonne of material in the deposit has a 
certain contained value which is dependent on the 
metal content and current price of the contained 
metal. For instance, at a copper price of s 
and a molybdenum price of s a deposit 
containing 1% copper and 0.015% molybdenum 
has a contained value of more than s The 
deposit will be economic to work, and can be clas- 
sified as an ore deposit if: 
Contained value per tonne > (total processing costs 
+ losses + other costs) per tonne 
A major cost is mining, and this can vary enor- 
mously, from only a few pence per tonne of 
ore to well over s High-tonnage operations 
are cheaper in terms of operating costs but have 
higher initial capital costs. These capital costs are 
paid off over a number of years, so that high- 
tonnage operations can only be justified for the 
treatment of deposits large enough to allow this. 
Small ore bodies are worked on a smaller scale, to 
reduce overall capital costs, but capital and oper- 
ating costs per tonne are correspondingly higher 
(Ottley, 1991). 
Alluvial mining is the cheapest method and, if 
on a large scale, can be used to mine ores of very 
low contained value due to low grade or low metal 
price, or both. For instance, in S.E. Asia, tin ores 
containing as little as 0.01% Sn are mined by allu- 
vial methods. These ores had a contained value 
of less than s but very low processing costs 
allowed them to be economically worked. 
High-tonnage open-pit and underground block- 
caving methods are also used to treat ores of low 
contained value, such as low-grade copper ores. 
Where the ore must be mined selectively, however, 
as is the case with underground vein-type deposits, 
mining methods become very expensive, and can 
only be justified on ores of high contained value. 
An underground selective mining cost of s 
would obviously be hopelessly uneconomic on a 
tin ore of alluvial grade, but may be economic on a 
hard-rock ore containing 1.5% tin, with a contained 
value of around s 
In order to produce metals, the ore minerals must 
be broken down by the action of heat (pyromet- 
allurgy), solvents (hydrometallurgy) or electricity 
(electrometallurgy), either alone or in combination, 
the most common method being the pyrometallur- 
gical process of smelting. These chemical methods 
Introduction 5 
consume vast quantities of energy. Treatment of 1 t 
of copper ore, for instance, consumes in the region 
of 1500-2000 kWh of electrical energy, which at 
a cost of say 5 p/kW h is around s well above 
the contained value of all current copper ores. 
Smelters are often remote from the mine site, 
being centred in areas where energy is relatively 
cheap, and where access to roads, rail or sea-links is 
available for shipment of fuel and supplies to, and 
products from, the smelter. The cost of transporta- 
tion of mined ore to remote smelters could in many 
cases be greater than the contained value of the ore. 
Mineral processing is usually carried out at the 
mine site, the plant being referred to as a mill or 
concentrator. The essential purpose is to reduce the 
bulk of the ore which must be transported to and 
processed by the smelter, by using relatively cheap, 
low-energy physical methods to separate the valu- 
able minerals from the waste (gangue) minerals. 
This enrichment process considerably increases the 
contained value of the ore to allow economic trans- 
portation and smelting. 
Compared with chemical methods, the phys- 
ical methods used in mineral processing consume 
relatively small amounts of energy. For instance, 
to upgrade a copper ore from 1 to 25% metal 
would use in the region of 20-50kWht -1. The 
corresponding reduction in weight of around 25:1 
proportionally lowers transport costs and reduces 
smelter energy consumption to around 60-80 kW h 
in relation to the weight of mined ore. It is impor- 
tant to realise that, although the physical methods 
are relatively low energy users, the reduction in 
bulk lowers smelter energy consumption to the 
order of that used in mineral processing, and it is 
significant that as ore grades decline, the energy 
used in mineral processing becomes an important 
factor in deciding whether the deposit is viable to 
work or not. 
Mineral processing reduces not only smelter 
energy costs but also smelter metal losses, due to 
the production of less metal-bearing slag. Although 
technically possible, the smelting of extremely low- 
grade ores, apart from being economically unjusti- 
fiable, would be very difficult due to the need to 
produce high-grade metal products free from dele- 
terious impurities. These impurities are found in the 
gangue minerals and it is the purpose of mineral 
processing to reject them into the discard (tailings), 
as smelters often impose penalties according to 
their level. For instance, it is necessary to remove 
arsenopyrite from tin concentrates, as it is difficult 
to remove
the contained arsenic in smelting and the 
process produces a low-quality tin metal. 
Against the economic advantages of mineral 
processing, the losses occurred during milling and 
the cost of milling operations must be charged. The 
latter can vary over a wide range, depending on 
the method of treatment used, and especially on 
the scale of the operation. As with mining, large- 
scale operations have higher capital but lower oper- 
ating costs (particularly labour and energy) than 
small-scale operations. As labour costs per tonne 
are most affected by the size of the operation, so, 
as capacity increases, the energy costs per tonne 
become proportionally more significant, and these 
can be more than 25% of the total milling costs in 
a 10,000 t d-1 concentrator. 
Losses to tailings are one of the most impor- 
tant factors in deciding whether a deposit is viable 
or not. Losses will depend very much on the ore 
mineralogy and dissemination, and on the tech- 
nology available to achieve efficient concentration. 
Thus, the development of froth flotation allowed 
the exploitation of vast low-grade copper deposits 
which were previously uneconomic to treat. Simi- 
larly, the introduction of solvent extraction enabled 
Nchanga Consolidated Copper Mines in Zambia 
to treat 9 Mt/yr of flotation tailings, to produce 
80,000 t of finished copper from what was previ- 
ously regarded as waste (Anon., 1979). 
In many cases not only is it necessary to sepa- 
rate valuable from gangue minerals, but it is 
also required to separate valuable minerals from 
each other. For instance, porphyry copper ores 
are an important source of molybdenum and the 
minerals of these metals must be separated for 
separate smelting. Similarly, complex sulphide ores 
containing economic amounts of copper, lead and 
zinc usually require separate concentrates of the 
minerals of each of these metals. The provision 
of clean concentrates, with little or no contamina- 
tion with associated metals, is not always econom- 
ically feasible, and this leads to another source of 
loss other than direct tailing loss. A metal which 
reports to the "wrong" concentrate may be difficult, 
or economically impossible, to recover, and never 
achieves its potential valuation. Lead, for example, 
is essentially irrecoverable in copper concentrates 
and is often penalized as an impurity by the copper 
6 Wills' Mineral Processing Technology 
smelter. The treatment of such polymetallic base 
metal ores, therefore, presents one of the great chal- 
lenges to the mineral processor. 
Mineral processing operations are often a 
compromise between improvements in metallur- 
gical efficiency and milling costs. This is particu- 
larly true with ores of low contained value, where 
low milling costs are essential and cheap unit 
processes are necessary, particularly in the early 
stages, where the volume of material treated is rela- 
tively high. With such low-value ores, improve- 
ments in metallurgical efficiency by the use of 
more expensive methods or reagents cannot always 
be justified. Conversely high metallurgical effi- 
ciency is usually of most importance with ores of 
high contained value and expensive high-efficiency 
processes can often be justified on these ores. 
Apart from processing costs and losses, other 
costs which must be taken into account are indirect 
costs such as ancillary services- power supply, 
water, roads, tailings disposal- which will depend 
much on the size and location of the deposit, 
as well as taxes, royalty payments, investment 
requirements, research and development, medical 
and safety costs, etc. 
Non-metallic ores 
Ores of economic value can be classed as metallic 
or non-metallic, according to the use of the mineral. 
Certain minerals may be mined and processed for 
more than one purpose. In one category the mineral 
may be a metal ore, i.e. when it is used to prepare 
the metal, as when bauxite (hydrated aluminium 
oxide) is used to make aluminium. The alternative 
is for the compound to be classified as a non- 
metallic ore, i.e. when bauxite or natural aluminium 
oxide is used to make material for refractory bricks 
or abrasives. 
Many non-metallic ore minerals associate with 
metallic ore minerals (Appendix II) and are mined 
and processed together, e.g. galena, the main source 
of lead, sometimes associates with fluorite (CaF2) 
and barytes (BaSO4), both important non-metallic 
minerals. 
Diamond ores have the lowest grade of all mined 
ores. The richest mine in terms of diamond content 
(Argyle in Western Australia) enjoyed grades as 
high as 2 ppm in its early life. The lowest grade 
deposits mined in Africa have been as low as 
0.01 ppm. Diamond deposits are mined mainly for 
their gem quality stones which have the highest 
value, with the low-value industrial quality stones 
being essentially a by-product; most industrial 
diamond is now produced synthetically. 
Tailings retreatment 
Mill tailings which still contain valuable compo- 
nents constitute a potential future resource. New 
or improved technologies can allow the value 
contained in tailings, which was lost in earlier 
processing, to be recovered, or commodities 
considered waste in the past can become valuable 
in a new economic order. Reducing or eliminating 
tailings dumps or dams by retreating them also 
reduces the environmental impact of the waste. 
The cost of tailings retreatment is sometimes 
lower than that of processing the original ore, 
because much of the expense has already been met, 
particularly in mining and comminution. There are 
many tailings retreatment plants in a variety of 
applications around the world. The East Rand Gold 
and Uranium Company (ERGO) closed its oper- 
ations in 2005 after 28 years of retreating over 
870 Mt of the iconic gold dumps of Johannesburg, 
significantly modifying the skyline of the Golden 
City and producing 250t of gold in the process. 
Also in 2005 underground mining in Kimberley 
closed, leaving a tailings dump retreatment opera- 
tion as the only source of diamond production in the 
Diamond City. Some platinum producers in South 
Africa now operate tailings retreatment plants for 
the recovery of platinum group metals (PGMs), and 
also chromite as a by-product from the chrome-rich 
UG2 Reef. 
Although these products, particularly gold, tend 
to dominate the list of tailings retreatment oper- 
ations because of the value of the product, there 
are others, both operating and being considered as 
potential major sources of particular commodities. 
For example coal has been recovered from tailings 
in Australia (Clark, 1997), uranium is recovered 
from copper tailings by the Uranium Corporation 
of India, and copper has been recovered from 
the Bwana Mkubwa tailings in Zambia, using 
solvent extraction and electrowinning. The Kolwezi 
Tailings project in the DRC which proposes to 
recover oxide copper and cobalt from the tailings 
of 50 years of copper mining is expected to be the 
largest source of cobalt in the world. 
Introduction 7 
The re-processing of industrial scrap and 
domestic waste is also a growing economic activity, 
especially in Europe. It is essentially a branch 
of mineral processing with a different feedstock, 
though operation is generally dry rather than wet 
(Hoberg, 1993; Furuyama et al., 2003). 
Mineral processing methods 
"As-mined" or "run-of-mine" ore consists of valu- 
able minerals and gangue. Mineral processing, 
sometimes called ore dressing, mineral dressing or 
milling, follows mining and prepares the ore for 
extraction of the valuable metal in the case of 
metallic ores, and produces a commercial end 
product of products such as iron ore and coal. Apart 
from regulating the size of the ore, it is a process 
of physically separating the grains of valuable
minerals from the gangue minerals, to produce an 
enriched portion, or concentrate, containing most 
of the valuable minerals, and a discard, or tailing, 
containing predominantly the gangue minerals. The 
importance of mineral processing is today taken 
for granted, but it is interesting to reflect that less 
than a century ago, ore concentration was often a 
fairly crude operation, involving relatively simple 
gravity and hand-sorting techniques performed by 
the mining engineers. The twentieth century saw 
the development of mineral processing as a serious 
and important professional discipline in its own 
right, and without physical separation, the concen- 
tration of many ores, and particularly the metallif- 
erous ores, would be hopelessly uneconomic (Wills 
and Atkinson, 1991). 
It has been predicted, however, that the impor- 
tance of mineral processing of metallic ores 
may decline as the physical processes utilised 
are replaced by the hydro and pyrometallurgical 
routes used by the extractive metallurgist (Gilchrist, 
1989), because higher recoveries are obtained by 
some chemical methods. This may certainly apply 
when the useful mineral is very finely dissemi- 
nated in the ore and adequate liberation from the 
gangue is not possible, in which case a combina- 
tion of chemical and mineral processing techniques 
may be advantageous, as is the case with some 
highly complex ores containing economic amounts 
of copper, lead, zinc and precious metals (Gray, 
1984; Barbery, 1986). Also new technologies such 
as direct reduction may allow direct smelting of 
some ores. However, in the majority of cases the 
energy consumed in direct smelting or leaching 
of low-grade ores would be so enormous as to 
make the cost prohibitive. Compared with these 
processes, mineral processing methods are inexpen- 
sive, and their use is readily justified on economic 
grounds. 
If the ore contains worthwhile amounts of more 
than one valuable mineral, it is usually the object 
of mineral processing to separate them; similarly 
if undesirable minerals, which may interfere with 
subsequent refining processes, are present, it may 
be necessary to remove these minerals at the sepa- 
ration stage. 
There are two fundamental operations in mineral 
processing: namely the release, or liberation, of 
the valuable minerals from their waste gangue 
minerals, and separation of these values from 
the gangue, this latter process being known as 
concentration. 
Liberation of the valuable minerals from the 
gangue is accomplished by comminution, which 
involves crushing, and, if necessary, grinding, to 
such a particle size that the product is a mixture 
of relatively clean particles of mineral and gangue. 
Grinding is often the greatest energy consumer, 
accounting for up to 50% of a concentrator' s energy 
consumption. As it is this process which achieves 
liberation of values from gangue, it is also the 
process which is essential for efficient separation 
of the minerals, and it is often said to be the key 
to good mineral processing. In order to produce 
clean concentrates with little contamination with 
gangue minerals, it is necessary to grind the ore 
finely enough to liberate the associated metals. 
Fine grinding, however, increases energy costs, and 
can lead to the production of very fine untreat- 
able "slime" particles which may be lost into the 
tailings. Grinding therefore becomes a compromise 
between clean (high-grade) concentrates, operating 
costs and losses of fine minerals. If the ore is 
low grade, and the minerals have very small grain 
size and are disseminated through the rock, then 
grinding energy costs and fines losses can be high, 
unless the nature of the minerals is such that a 
pronounced difference in some property between 
the minerals and the gangue is available. 
An intimate knowledge of the mineralogical 
assembly of the ore is essential if efficient 
processing is to be carried out. A knowledge not 
8 Wills' Mineral Processing Technology 
only of the nature of the valuable and gangue 
minerals but also of the ore "texture" is required. 
The texture refers to the size, dissemination, 
association and shape of the minerals within the 
ore. The processing of minerals should always be 
considered in the context of the mineralogy of 
the ore in order to predict grinding and concen- 
tration requirements, feasible concentrate grades 
and potential difficulties of separation (Hausen, 
1991; Guerney et al., 2003; Baum et al., 2004). 
Microscopic analysis of concentrate and tailings 
products can also yield much valuable informa- 
tion regarding the efficiency of the liberation 
and concentration processes (see Figures 1.2a-i 
for examples). It is particularly useful in trou- 
bleshooting problems which arise from inadequate 
liberation. Conventional optical microscopes can 
be used for the examination of thin and polished 
sections of mineral samples, and in mineral sands 
applications the simple binocular microscope is a 
practical tool. However, it is becoming increas- 
ingly common to utilise the new technologies of 
automated mineral analysis using scanning elec- 
tron microscopy, such as the Mineral Liberation 
Analyser (MLA) (Gu, 2003) and the QEMSCAN 
(Gottlieb et al., 2000). 
The most important physical methods which are 
used to concentrate ores are: 
(1) Separation based on optical and other proper- 
ties. This is often called sorting, which used 
Figure 1.2a Chromite ore. Relatively coarse grain size, and compact morphology of chromite (C) grains 
makes liberation from olivine (O) gangue fairly straightforward 
Figure 1.2b North American porphyry copper ore. Chalcopyrite (C) precipitated along fractures in quartz. 
Liberation of chalcopyrite is fairly difficult due to "chain-like" distribution. Fracture is, however, likely to occur 
preferentially along the sealed fractures, producing particles with a surface coating of chalcopyrite, which can 
be effectively recovered into a low-grade concentrate by froth flotation 
Introduction 9 
Figure 1.2c Mixed sulphide ore, Wheal Jane, Cornwall. Chalcopyrite (C) and sphalerite (S), much of which is 
extremely finely disseminated in tourmaline (T), making a high degree of liberation impracticable 
Figure 1.2d Hilton lead-zinc ore body, Australia. Galena (G) and sphalerite (S) intergrown. Separate "clean" 
concentrates of lead and zinc will be difficult to produce, and contamination of concentrates with other metal is 
likely 
Figure 1.2e Copper-zinc ore. Grain of sphalerite with many minute inclusions of chalcopyrite (C) along 
cleavage planes. Fracturing during comminution takes place preferentially along the low coherence cleavage 
planes, producing a veneer of chalcopyrite on the sphalerite surface, making depression of the latter difficult in 
flotation 
10 Wills' Mineral Processing Technology 
Figure 1.2f Lead-zinc ore. Fine grained native silver in vein networks and inclusions in carbonate host rock. 
Rejection of this material by heavy medium separation could lead to high silver loss 
Figure 1.2g Flotation tailings, Palabora Copper Mine, South Africa. Finely disseminated grains of 
chalcopyrite enclosed in a grain of gangue, and irrecoverable by flotation. Maximum grain size of chalcopyrite 
is about 20 microns, so attempts to liberate by further grinding would be impracticable 
Figure 1.2h Gravity circuit tailings, tin concentrator. Cassiterite (light grey) locked with gangue (darker grey), 
mainly quartz. The composite particle is very fine (less than 20 I~m), and has reported to tailings, rather than 
middlings, due to the inefficiency of gravity separation at this size. Loss of such particles to tailings is a major 
cause of poor recovery in gravity concentration. In this case, the composite tailings particles could be 
recovered by froth
flotation into a low-grade concentrate 
Introduction 11 
Figure 1.2i Tin concentrate, assaying about 60% tin. Although there is some limited locking of the cassiterite 
(light grey) with gangue (darker grey), the main contaminant is arsenopyrite (white), which, being a heavy 
mineral (S.G. 6), has partitioned with the cassiterite (S.G. 7) into the gravity concentrate. The arsenopyrite 
particles are essentially liberated, and can easily be removed by froth flotation, thereby increasing the tin grade 
of the concentrate and avoiding smelter penalties due to high arsenic levels 
(2) 
(3) 
(4) 
to be done by hand but is now mostly accom- 
plished by machine (see Chapter 14). 
Separation based on differences in density 
between the minerals. Gravity concentra- 
tion, a technology with its roots in antiq- 
uity, is based on the differential movement of 
mineral particles in water due to their different 
hydraulic properties. The method has recently 
enjoyed a new lease of life with the develop- 
ment of a range of enhanced gravity concen- 
trating devices. In dense medium separation 
particles sink or float in a dense liquid or 
(more usually) an artificial dense suspension; 
it is widely used in coal beneficiation, iron 
ore and diamond processing, and in the pre- 
concentration of metalliferous ores. 
Separation utilising the different surface 
properties of the minerals. Froth flotation, 
which is one of the most important methods 
of concentration, is effected by the attach- 
ment of the mineral particles to air bubbles 
within the agitated pulp. By adjusting the 
"climate" of the pulp by various reagents, 
it is possible to make the valuable minerals 
air-avid (aerophilic) and the gangue minerals 
water-avid (aerophobic). This results in sepa- 
ration by transfer of the valuable minerals to 
the air bubbles which form the froth floating 
on the surface of the pulp. 
Separation dependent on magnetic properties. 
Low intensity magnetic separators can be used 
to concentrate ferromagnetic minerals such 
(5) 
as magnetite (Fe304), while high-intensity 
separators are used to separate paramagnetic 
minerals from their gangue. Magnetic sepa- 
ration is an important process in the bene- 
ficiation of iron ores, and finds application 
in the treatment of paramagnetic non-ferrous 
minerals. It is used to remove paramagnetic 
wolframite ((Fe, Mn) WO4) and hematite 
(Fe203) from tin ores, and has found consid- 
erable application in the processing of non- 
metallic minerals, such as those found in 
mineral sand deposits. 
Separation dependent on electrical conduc- 
tivity properties. High-tension separation can 
be used to separate conducting minerals 
from non-conducting minerals. This method 
is interesting, since theoretically it represents 
the "universal" concentrating method; almost 
all minerals show some difference in conduc- 
tivity and it should be possible to separate 
almost any two by this process. However, 
the method has fairly limited application, and 
its greatest use is in separating some of the 
minerals found in heavy sands from beach or 
stream placers. Minerals must be completely 
dry and the humidity of the surrounding air 
must be regulated, since most of the electron 
movement in dielectrics takes place on the 
surface and a film of moisture can change 
the behaviour completely. The biggest disad- 
vantage of the method is that the capacity of 
economically sized units is low. 
12 Wills' Mineral Processing Technology 
In many cases, a combination of two or more of fine magnetite particles onto the surfaces of non- 
of the above techniques is necessary to concen- magnetic minerals in the slurry. 
trate an ore economically. Gravity separation, for Some refractory copper ores containing 
instance, is often used to reject a major portion sulphide and oxidised minerals have been pre- 
of the gangue, as it is a relatively cheap process, treated hydrometallurgically to enhance flotation 
It may not, however, have the selectivity or effi- performance. In the Leach-Precipitation-Flotation 
ciency to produce the final clean concentrate, process, developed in the years 1929-34 by the 
Gravity concentrates therefore often need further Miami Copper Co., USA, the oxidised minerals 
upgrading by more expensive techniques, such as are dissolved in sulphuric acid, after which the 
froth flotation, copper in solution is precipitated as cement copper 
Ores which are very difficult to treat (refractory), by the addition of metallic iron. The cement 
due to fine dissemination of the minerals, complex copper and acid-insoluble sulphide minerals are 
mineralogy, or both, respond very poorly to the then recovered by froth flotation. This process, with 
above methods, several variations, has been used at a number of 
A classic example is the huge zinc-lead-silver American copper concentrators, but a more widely 
deposit at McArthur River, in Australia. Discov- used method of enhancing the flotation perfor- 
ered in 1955, it is one of the world's largest mance of oxidised ores is to allow the surface 
zinc-lead deposits comprising measured, indicated to react with sodium sulphide. This "sulphidisa- 
and inferred resources totalling 124 Mt with up tion" process modifies the flotation response of 
to 13% Zn, 6% Pb and 60g/t Ag (in 2003). For the mineral causing it to behave, in effect, as a 
35 years it resisted attempts to find an economic pseudo-sulphide. Such chemical conditioning of 
processing route due to the very fine grained mineral surfaces is widely used in froth flotation 
texture of the ore. However, the development of (see Chapter 12); sphalerite, for example, can be 
the proprietary IsaMill fine grinding technology made to respond in a similar way to chalcopy- 
(Pease, 2005) by the mine's owners Mount Isa rite, by allowing the surface to react with copper 
Mines (now Xstrata), together with an appro- sulphate. 
pilate flotation circuit, allowed the ore to be Recent developments in biotechnology are 
successfully processed and the mine was finally currently being exploited in hydrometallurgical 
opened in 1995. The concentrator makes a bulk operations, particularly in the bacterial oxidation 
lead-zinc concentrate with a very fine product of sulphide gold ores and concentrates (Brierley 
size of 80% smaller than 7 p~m. There are many and Brierley, 2001; Hansford and Vargas, 2001). 
stages of flotation cleaning to achieve the neces- The bacterium Acidithiobacillus ferrooxidans is 
sary product grades with sufficient rejection of mainly used to enhance the rate of oxida- 
silica. McArthur River is a good example of how tion, by breaking down the sulphide lattice and 
developments in technology can render previously thus liberating the occluded gold for subse- 
uneconomic ore deposits viable. Process evolution quent removal by cyanide leaching (Lazer et al., 
for McArthur River continues, with the propri- 1986). There is good evidence to suggest 
etary Albion atmospheric leaching process being that certain microorganisms could be used to 
considered for the direct treatment of concentrates enhance the performance of conventional mineral 
(Anon., 2002). processing techniques (Smith et al., 1991). It has 
Chemical methods, such as pyrometallurgy or been established that some bacteria will act as 
hydrometallurgy, can be used to alter mineralogy, pyrite depressants in coal flotation, and prelim- 
allowing the low cost mineral processing methods inary work has shown that certain organisms 
to be applied to refractory ores (Iwasaki and can aid flotation in other ways, with potential 
Prasad, 1989). For instance, non-magnetic iron profound changes to future industrial froth flotation 
oxides can be roasted in a weakly reducing atmo- practice. 
sphere to produce ferromagnetic magnetite. It has Extremely fine mineral dissemination leads to 
also been suggested
(Parsonage, 1988) that the high energy costs in comminution and high losses 
magnetic response could be increased without to tailings due to the generation of difficult- 
chemically altering the minerals, by the adsorption to-treat fine particles. Much research effort has 
Introduction 13 
been directed at minimizing fines losses in recent 
years, either by developing methods of enhancing 
mineral liberation, thus minimizing the amount 
of comminution needed, or by increasing the 
efficiency of conventional physical separation 
processes, by the use of innovative machines or 
by optimising the performance of existing ones. 
Several methods have been researched and devel- 
oped to attempt to increase the apparent size of 
fine particles, by causing them to come together 
and agglomerate. Selective flocculation of certain 
minerals in suspension, followed by separation 
of the aggregates from the dispersion, has been 
successfully achieved on a variety of ore-types at 
laboratory scale, but plant application is limited 
(see Chapter 15). 
Ultra-fine particles in a suspension can be 
agglomerated under high shear conditions if the 
particle surfaces are hydrophobic (water-repellent). 
A shear field, caused by vigorous agitation, of suffi- 
cient magnitude to overcome the energy barrier 
separating the particles is necessary to bring them 
together for hydrophobic association. Although the 
phenomenon of shear flocculation is well known 
it has not, as yet, been exploited commercially 
(Bilgen and Wills, 1991). 
Selective agglomeration of fine particles by oil 
is a promising method, and has been devel- 
oped to a commercial scale for the treatment 
of fine coal (Capes, 1989; Huettenhain, 1991). 
In the oil agglomeration process an immiscible 
liquid (e.g. a hydrocarbon) is added to the 
suspension. On agitation, the oil is distributed 
over oleophilic/hydrophobic surfaces and particle 
impact allows inter-particle liquid bridges to 
form, causing agglomeration. The oleophilicity of 
specific minerals can be controlled, for example, 
by adding froth flotation reagents. As yet, the oil 
agglomeration process has not been used to treat 
ultra-fine minerals outside the laboratory (House 
and Veal, 1989). 
Run-of-mine ore 
i' I Comminution 
t , 
, t 
Product 
handling 
, , , 
Figure 1.3 Simple block flowsheet 
rejection. The next block, "separation", groups 
the various treatments incident to production 
of concentrate and tailing. The third, "product 
handling", covers the disposal of the products. 
The simple line flowsheet (Figure 1.4) is for 
most purposes sufficient, and can include details of 
machines, settings, rates, etc. 
(+) 
(+) 
Ore 
I 
l 
Crushers 
Screens 
(-) 
Gdr ding 
Classification 
(-) 
j Separation 1 
Concentrate Tailing 
Figure 1.4 Line flowsheet. (+)indicates oversized 
material returned for further treatment and (-) 
undersized material, which is allowed to proceed to 
the next stage 
The f lowsheet 
The flowsheet shows diagrammatically the 
sequence of operations in the plant. In its simplest 
form it can be presented as a block diagram 
in which all operations of similar character are 
grouped (Figure 1.3). In this case comminu- 
tion deals with all crushing, grinding and initial 
Milling costs 
It has been shown that the balance between milling 
costs and metal losses is crucial, particularly with 
low-grade ores, and because of this, most mills keep 
detailed accounts of operating and maintenance 
costs, broken down into various sub-divisions, such 
14 Wills' Mineral Processing Technology 
as labour, supplies, energy, etc. for the various areas 
of the plant. This type of analysis is very useful in 
identifying high-cost areas where improvements in 
performances would be most beneficial. It is impos- 
sible to give typical operating costs for milling 
operations, as these vary enormously from mine 
to mine, and particularly from country to country, 
depending on local costs of energy, labour, water, 
supplies, etc., but Table 1.3 is a simplified example 
of such a breakdown of costs for a 100,000t/d 
copper concentrator. Note the dominance of 
grinding, due mainly to power requirements. 
Table 1.3 Costs per metric tonne milled for a 
100,000 t/d copper concentrator 
Item Cost- US$ Percent cost 
per tonne 
Crushing 0.088 2.8 
Grinding 1.482 47.0 
Flotation 0.510 16.2 
Thickening 0.111 3.5 
Filtration 0.089 2.8 
Tailings 0.161 5.1 
Reagents 0.016 0.5 
Pipeline 0.045 1.4 
Water 0.252 8.0 
Laboratory 0.048 1.5 
Maintenance support 0.026 0.8 
Management support 0.052 1.6 
Administration 0.020 0.6 
Other expenses 0.254 8.1 
Total 3.154 100 
Efficiency of mineral processing 
operations 
L iberat ion 
One of the major objectives of comminution is 
the liberation, or release, of the valuable minerals 
from the associated gangue minerals at the coarsest 
possible particle size. If such an aim is achieved, 
then not only is energy saved by the reduction of 
the amount of fines produced, but any subsequent 
separation stages become easier and cheaper to 
operate. If high-grade solid products are required, 
then good liberation is essential; however, for 
subsequent hydrometallurgical processes, such as 
leaching, it may only be necessary to expose the 
required mineral. 
In practice, complete liberation is seldom 
achieved, even if the ore is ground down to the 
grain size of the desired mineral particles. This 
is illustrated by Figure 1.5, which shows a lump 
of ore which has been reduced to a number of 
cubes of identical volume and of a size below 
that of the grains of mineral observed in the 
original ore sample. It can be seen that each 
particle produced containing mineral also contains 
a portion of gangue; complete liberation has not 
been attained; the bulk of the major mineral- the 
gangue - has, however, been liberated from the 
minor mineral- the value. 
Figure 1.5 "Locking" of mineral and gangue 
The particles of "locked" mineral and gangue 
are known as middlings, and further liberation 
from this fraction can only be achieved by further 
comminution. 
The "degree of liberation" refers to the 
percentage of the mineral occurring as free parti- 
cles in the ore in relation to the total content. 
This can be high if there are weak boundaries 
between mineral and gangue particles, which is 
often the case with ores composed mainly of 
rock-forming minerals, particularly sedimentary 
minerals. Usually, however, the adhesion between 
mineral and gangue is strong and, during comminu- 
tion, the various constituents are cleft across. This 
produces much middlings and a low degree of liber- 
ation. New approaches to increasing the degree of 
liberation involve directing the breaking stresses 
at the mineral crystal boundaries, so that the rock 
can be broken without breaking the mineral grains 
(Wills and Atkinson, 1993). 
Many researchers have tried to quantify degree 
of liberation with a view to predicting the behaviour 
Introduction 15 
of particles in a separation process (Barbery, 1991). 
The first attempt at the development of a model for 
the calculation of liberation was made by Gaudin 
(1939); King (1982) developed an exact expres- 
sion for the fraction of particles of a certain size 
that contained less than a prescribed fraction of 
any particular mineral. These models, however, 
suffered from many unrealistic assumptions that 
must be made with respect to the grain structure 
of the minerals in the ore, in particular that libera- 
tion is preferential, and in 1988 Austin and Luckie 
concluded that "there is no adequate model of 
liberation of binary systems suitable for incorpora- 
tion into a mill model". For this reason liberation 
models have not found much practical application.
However, some fresh approaches by Gay, allowing 
multi-mineral systems to be modelled (not just 
binary systems) free of the assumptions of pref- 
erential breakage, have recently demonstrated that 
there may yet be a useful role for such models 
(Gay, 2004a,b). The quantification of liberation is 
now routinely possible using the dedicated scan- 
ning electron microscope MLA and QEMSCAN 
systems mentioned earlier, and concentrators are 
increasingly using such systems to monitor the 
degree of liberation in their processes. 
It should also be noted that a high degree of 
liberation is not necessary in certain processes, 
and, indeed, may be undesirable. For instance, it 
is possible to achieve a high recovery of values 
by gravity and magnetic separation even though 
the valuable minerals are completely enclosed by 
gangue, and hence the degree of liberation of the 
values is zero. As long as a pronounced density 
or magnetic susceptibility difference is apparent 
between the locked particles and the free gangue 
particles, the separation is possible. A high degree 
of liberation may only be possible by intensive fine 
grinding, which may reduce the particles to such a 
fine size that separation becomes very inefficient. 
On the other hand, froth flotation requires as much 
of the valuable mineral surface as possible to be 
exposed, whereas in a chemical leaching process, a 
portion of the surface must be exposed to provide 
a channel to the bulk of the mineral. 
In practice, ores are ground to an optimum grind 
size, determined by laboratory and pilot scale test- 
work, to produce an economic degree of libera- 
tion. The concentration process is then designed 
to produce a concentrate consisting predominantly 
of valuable mineral, with an accepted degree of 
locking with the gangue minerals, and a middlings 
fraction, which may require further grinding to 
promote optimum release of the minerals. The 
tailings should be mainly composed of gangue 
minerals. 
Figure 1.6 is a cross-section through a typical 
ore particle, and illustrates effectively the libera- 
tion dilemma often facing the mineral processor. 
Regions A represent valuable mineral, and region 
AA is rich in valuable mineral but is highly 
intergrown with the gangue mineral. Comminu- 
tion produces a range of fragments, ranging from 
fully liberated mineral and gangue particles, to 
those illustrated. Particles of type 1 are rich in 
mineral, and are classed as concentrate as they 
have an acceptable degree of locking with the 
gangue, which limits the concentrate grade. Parti- 
cles of type 4 would likewise be classed as tail- 
ings, the small amount of mineral present reducing 
the recovery of mineral into the concentrate. Parti- 
cles of types 2 and 3, however, would probably 
be classed as middlings, although the degree of 
regrinding needed to promote economic liberation 
of mineral from particle 3 would be greater than in 
particle 2. 
Figure 1.6 Cross-sections of ore particles 
During the grinding of a low-grade ore the bulk 
of the gangue minerals is often liberated at a 
relatively coarse size (see Figure 1.5). In certain 
circumstances it may be economic to grind to a 
size much coarser than the optimum in order to 
produce in the subsequent concentration process a 
large middlings fraction and a tailings which can 
be discarded at a coarse grain size. The middlings 
fraction can then be reground to produce a feed to 
the final concentration process (Figure 1.7). 
This method discards most of the coarse gangue 
early in the process, thus considerably reducing 
16 Wills' Mineral Processing Technology 
I 
Middlings 
Se~ila~ dti~d n 
9' 
Concentrate 
Feed 
1 Primary grind 
Pre-con!entration I 
1 
Tailings 
Middlings Taili~ngs 
Figure 1.7 Flowsheet for process utilising two-stage 
separation 
grinding costs, as needless comminution of liber- 
ated gangue is avoided. It is often used on minerals 
which can easily be separated from the free gangue, 
even though they are themselves locked to some 
extent with gangue. It is the basis of the dense 
medium process of preconcentration (Chapter 11). 
Concentration 
The object of mineral processing, regardless of the 
methods used, is always the same, i.e. to separate 
the minerals into two or more products with the 
values in the concentrates, the gangue in the tail- 
ings, and the "locked" particles in the middlings. 
Such separations are, of course, never perfect, 
so that much of the middlings produced are, in 
fact, misplaced particles, i.e. those particles which 
ideally should have reported to the concentrate or 
the tailings. This is often particularly serious when 
treating ultra-fine particles, where the efficiency of 
separation is usually low. In such cases, fine liber- 
ated valuable mineral particles often report in the 
middlings and tailings. The technology for treating 
fine-sized minerals is, as yet, poorly developed, 
and, in some cases, very large amounts of fines are 
discarded. For instance, it is common practice to 
remove material less than 10txm in size from tin 
concentrator feeds and direct this material to the 
tailings, and, in the early 1970s, 50% of the tin 
mined in Bolivia, 30% of the phosphate mined in 
Florida, and 20% of the world's tungsten were lost 
as fines. Significant amounts of copper, uranium, 
fluorspar, bauxite, zinc, and iron were also simi- 
larly lost (Somasundaran, 1986). 
Figure 1.8 shows the general size range applica- 
bility of unit concentration processes (Mills, 1978). 
It is evident that most mineral processing techniques 
fail in the ultra-fine size range. Gravity concentra- 
tion techniques, especially, become unacceptably 
inefficient. Flotation, one of the most important of the 
concentrating techniques, is now practised success- 
fully below 10 txm but not below 1 p~m 
It should be pointed out that the process is also 
limited by the mineralogical nature of the ore. 
For example, in an ore containing native copper it 
is theoretically possible to produce a concentrate 
Figure 1.8 Effective range of application of conventional mineral processing techniques 
Introduction 17 
containing 100% Cu, but, if the ore mineral was 
chalcopyrite (CuFeS2), the best concentrate would 
contain only 34.5% Cu. 
The recovery, in the case of the concentration 
of a metallic ore, is the percentage of the total 
metal contained in the ore that is recovered from 
the concentrate; a recovery of 90% means that 90% 
of the metal in the ore is recovered in the concen- 
trate and 10% is lost in the tailings. The recovery, 
when dealing with non-metallic ores, refers to the 
percentage of the total mineral contained in the ore 
that is recovered into the concentrate, i.e. recovery 
is usually expressed in terms of the valuable end 
product. 
The ratio of concentration is the ratio of the 
weight of the feed (or heads) to the weight of the 
concentrates. It is a measure of the efficiency of 
the concentration process, and it is closely related 
to the grade or assay of the concentrate; the value 
of the ratio of concentration will generally increase 
with the grade of concentrate. 
The grade, or assay, usually refers to the content 
of the marketable end product in the material. Thus, 
in metallic ores, the per cent metal is often quoted, 
although in the case of very low-grade ores, such 
as gold, metal content may be expressed as parts 
per million (ppm), or its equivalent grams per tonne 
(gt-1). Some metals are sold in oxide form, and 
hence the grade may be quoted in terms of the 
marketable oxide content, e.g. %WO 3, %U308, etc. 
In non-metallic operations, grade usually refers to 
the mineral content, e.g. %CaF 2 in fluorite ores; 
diamond ores are usually graded in carats per 100 
tonnes (t), where 1 carat is 0.2 g.
Coal is graded 
according to its ash content, i.e. the amount of 
incombustible mineral present within the coal. Most 
coal burned in power stations ("steaming coal") 
has an ash content between 15 and 20%, whereas 
"coking coal" used in steel making generally has an 
ash content of less than 10% together with appro- 
priate coking properties. 
The enrichment ratio is the ratio of the grade of 
the concentrate to the grade of the feed, and again 
is related to the efficiency of the process. 
Ratio of concentration and recovery are essen- 
tially independent of each other, and in order to 
evaluate a given operation it is necessary to know 
both. For example, it is possible to obtain a very 
high grade of concentrate and ratio of concentration 
by simply picking a few lumps of pure galena from 
a lead ore, but the recovery would be very low. On 
the other hand, a concentrating process might show 
a recovery of 99% of the metal, but it might also 
put 60% of the gangue minerals in the concentrate. 
It is, of course, possible to obtain 100% recovery 
by not concentrating the ore at all. 
There is an approximately inverse relationship 
between the recovery and grade of concentrate in 
all concentrating processes. If an attempt is made 
to attain a very high-grade concentrate, the tailings 
assays are higher and the recovery is low. If high 
recovery of metal is aimed for, there will be more 
gangue in the concentrate and the grade of concen- 
trate and ratio of concentration will both decrease. 
It is impossible to give figures for representative 
values of recoveries and ratios of concentration. A 
concentration ratio of 2 to 1 might be satisfactory 
for certain high-grade non-metallic ores, but a ratio 
of 50 to 1 might be considered too low for a low- 
grade copper ore; ratios of concentration of several 
million to one are common with diamond ores. The 
aim of milling operations is to maintain the values 
of ratio of concentration and recovery as high as 
possible, all factors being considered. 
Since concentrate grade and recovery are metal- 
lurgical factors, the metallurgical efficiency of any 
concentration operation can be expressed by a curve 
showing the recovery attainable for any value of 
concentrate grade. Figure 1.9 is a typical recovery- 
grade curve showing the characteristic inverse 
relationship between recovery and concentrate 
grade. Mineral processes generally move along 
a recovery-grade curve, with a trade-off between 
grade and recovery. The mineral processor's chal- 
lenge is to move the whole curve to a higher point 
so that both grade and recovery are maximised. 
0 r 
n" 
Grade of concentrate 
Figure 1.9 Typical recovery-grade curve 
18 Wills' Mineral Processing Technology 
Concentrate grade and recovery, used simulta- 
neously, are the most widely accepted measures 
of assessing metallurgical (not economic) perfor- 
mance. However, there is a problem in quanti- 
tatively assessing the technical performance of a 
concentration process whenever the results of two 
similar test runs are compared. If both the grade 
and recovery are greater for one case than the 
other, then the choice of process is simple, but if 
the results of one test show a higher grade but a 
lower recovery than the other, then the choice is 
no longer obvious. There have been many attempts 
to combine recovery and concentrate grade into a 
single index defining the metallurgical efficiency of 
the separation. These have been reviewed by Schulz 
(1970), who proposed the following definition: 
Separation efficiency (S.E.) - Rm - Rg (1.1) 
where Rm- % recovery of the valuable mineral, 
Rg- % recovery of the gangue into the 
concentrate. 
Suppose the feed material, assaying f% metal, 
separates into a concentrate assaying c% metal, 
and a tailing assaying t% metal, and that C is the 
fraction of the total feed weight that reports to the 
concentrate, then: 
100Cc 
Rm -- ~ (1.2) 
f 
i.e. recovery of valuable mineral to the concentrate 
is equal to metal recovery, assuming that all the 
valuable metal is contained in the same mineral. 
The gangue content of the concentrate- 100- 
(lOOc/m)%, where m is the percentage metal 
content of the valuable mineral, 
100(m - c) 
i.e. gangue content -- 
m 
Therefore, Rg = C x gangue content of concen- 
trate/gangue content of feed: 
100C(m - c) 
(m- f ) 
Therefore, Rm - Rg = 100Ccf { lOOC(m-c) f ) 
lOOCm(c- f) 
(m- - f ) f 
(1.3) 
Example 1.1 
A tin concentrator treats a feed containing 1% 
tin, and three possible combinations of concentrate 
grade and recovery are: 
High grade 
Medium grade 
Low grade 
63% tin at 62% recovery 
42% tin at 72% recovery 
21% tin at 78% recovery 
Determine which of these combinations of grade 
and recovery produce the highest separation effi- 
ciency. 
Solution 
Assuming that the tin is totally contained in 
the mineral cassiterite (SnO2), which, when pure, 
contains 78.6% tin, and since mineral recovery 
(Equation 1.2) is 100 • C x concentrate grade/feed 
grade, for the high-grade concentrate: 
100 
62-Cx63x- - , and soC=9.841x10 -3 
1 
Therefore, 
0.984 x 78.6 x (63- 1) 
(S.E.) (Equation 1.3)- 
(78.6- 1) • 1 
=61.8% 
Similarly, for the medium-grade concentrate, from 
Equation 1.2: 
Therefore, 
72= 100 • C x 42/1 
C = 1.714 • 10 -2, 
and S.E. (Equation 1.3) -- 71.2%. 
For the low-grade concentrate, from Equa- 
tion 1.2: 
78 - 100 x C x 21/1 
Therefore, C - 3.714 x 10 -2, and S.E. (Equa- 
tion 1.3) - 75.2%. 
Therefore, the highest separation efficiency is 
achieved by the production of a low-grade 
(21% tin) concentrate at high (78%) recovery. 
Although the value of separation efficiency can 
be useful in comparing the performance of different 
operating conditions on selectivity, it takes no 
account of economic factors, and, as will become 
apparent, a high value of separation efficiency does 
not necessarily lead to the most economic return. 
Since the purpose of mineral processing is to 
increase the economic value of the ore, the impor- 
tance of the recovery-grade relationship is in deter- 
mining the most economic combination of recovery 
and grade which will produce the greatest financial 
return per tonne of ore treated in the plant. This will 
depend primarily on the current price of the valu- 
able product, transportation costs to the smelter, 
refinery, or other further treatment plant, and the 
cost of such further treatment, the latter being very 
dependent on the grade of concentrate supplied. A 
high grade concentrate will incur lower smelting 
costs, but the lower recovery means lower returns 
of final product. A low grade concentrate may 
achieve greater recovery of the values, but incur 
greater smelting and transportation costs due to the 
included gangue minerals. Also of importance are 
impurities in the concentrate which may be penal- 
ized by the smelter, although precious metals may 
produce a bonus. 
The net return from the smelter (NSR) can 
be calculated for any recovery-grade combina- 
tion from: 
NSR = Payment for contained metal 
- (Smelter charges + Transport costs) 
This is summarised in Figure 1.10, which shows 
that the highest value of NSR is produced at an 
optimum concentrate grade. It is essential that the 
mill achieves a concentrate grade which is as close 
s "••••Valuation 
~NSR 
Target 
~._ ~ m e n t 
' Transport 
Concentrate grade 
Figure 1.10 Variation of payments and charges with 
concentrate grade 
Introduction 19 
as possible to this target grade. Although the effect 
of moving slightly away from the optimum may 
only be of the order of a few pence per tonnes 
treated, this can amount to very large financial 
losses, particularly on high-capacity plants treating
thousands of tonnes per day. Changes in metal 
price, smelter terms, etc. obviously affect the NSR- 
concentrate grade curve, and the value of the 
optimum concentrate grade. For instance, if the 
metal price increases, then the optimum grade will 
be lower, allowing higher recoveries to be attained 
(Figure 1.11). 
High 
P ~ s ~ ~ ~ Low 
Concentrate grade 
Figure 1.11 Effect of metal price on NSR-grade 
relationship 
It is evident that the terms agreed between the 
concentrator and smelter are of paramount impor- 
tance in the economics of mining and milling oper- 
ations. Such smelter contracts are usually fairly 
complex. Concentrates are sold under contract 
to "custom smelters" at prices based on quota- 
tions on metal markets such as the London Metal 
Exchange (LME). The smelter, having processed 
the concentrates, disposes of the finished metal to 
the consumers. The proportion of the "free market" 
price of the metal received by the mine is deter- 
mined by the terms of the contract negotiated 
between mine and smelter, and these terms can 
vary considerably. Table 1.4 summarises a typical 
low-grade smelter contract for the purchase of tin 
concentrates. As is usual in many contracts, one 
assay unit is deducted from the concentrate assay in 
assessing the value of the concentrates, and arsenic 
present in the concentrate is penalised. The concen- 
trate assay is of prime importance in determining 
the valuation, and the value of the assay is usually 
agreed on the result of independent sampling and 
assaying performed by the mine and smelter. The 
20 Wills' Mineral Processing Technology 
Table 1.4 Simplified tin smelter contract 
Material 
Tin concentrates, assaying no less than 15% Sn, to be free from deleterious impurities not stated, 
and to contain sufficient moisture as to evolve no dust when unloaded at our works. 
Quantity 
Total production of concentrates. 
Valuation 
Tin, less 1 unit per dry tonne of concentrates, at the lowest of the official London Metal 
Exchange prices. 
Pricing 
On the 7th market day after completion of arrival of each sampling lot into our works. 
Treatment charge 
s per dry tonne of concentrates. 
Moisture 
s per tonne of moisture. 
Penalties 
Arsenic s per unit per tonne. 
Lot charge 
s per lot sampled of less than 17 tonnes. 
Delivery 
Free to our works in regular quantities, loose on a tipping lorry or in any other manner acceptable 
to both parties. 
assays are compared, and if the difference is no 
more than an agreed value, the mean of the two 
results may be taken as the agreed assay. In the 
case of a greater difference, an "umpire" sample is 
assayed at an independent laboratory. This umpire 
assay may be used as the agreed assay, or the mean 
of this assay and that of the party which is nearer 
to the umpire assay may be chosen. 
The use of smelter contracts, and the impor- 
tance of the by-products and changing metal prices, 
can be seen by briefly examining the economics 
of processing two base metals - tin and copper - 
whose fortunes have fluctuated over the years for 
markedly different reasons. 
Economics of tin processing 
Tin constitutes an interesting case study in the 
vagaries of commodity prices and how they impact 
on the mineral industry and its technologies. 
Almost a half of the world's supply of tin in the 
mid-nineteenth century was mined in south-west 
England, but by the end of the 1870s Britain's 
premium position was lost, with the emergence 
of Malaysia as the leading producer and the 
discovery of rich deposits in Australia. By the end 
of the century, only nine mines of any conse- 
quence remained in Britain, where 300 had flour- 
ished 30 years earlier. From alluvial or secondary 
deposits, principally from South-East Asia, comes 
80% of mined tin. 
Unlike copper, zinc and lead, production of tin 
has not risen dramatically over the years and has 
rarely exceeded 250,000 t/yr. 
The real price of tin spent most of the first 
half of the twentieth century in a relatively narrow 
band between US$10 and US$15/t (19985), with 
some excursions (Figure 1.12; USGS, 2005). From 
1956 its price was regulated by a series of 
international agreements between producers and 
consumers under the auspices of the International 
Tin Council (ITC), which mirrored the highly 
successful policy of De Beers in controlling the 
gem diamond trade. Price stability was sought 
through selling from the ITC's huge stockpiles 
when the price rose and buying into the stockpile 
when the price fell. 
From the mid-1970s, however, the price of 
tin was driven artificially higher at a time of 
world recession, expanding production and falling 
consumption, the latter due mainly to the increasing 
use of aluminium, rather than tin-plated steel, 
cans. Although the ITC imposed restrictions on 
Introduction 21 
Figure 1.12 Tin prices 1900-2002 
the amount of tin that could be produced by its 
member countries, the reason for the inflating tin 
price was that the price of tin was fixed by the 
Malaysian dollar, while the buffer stock manager's 
dealings on the LME were financed in sterling. The 
Malaysian dollar was tied to the American dollar, 
which strengthened markedly between 1982 and 
1984, having the effect of increasing the price of 
tin in London simply because of the exchange rate. 
However, the American dollar began to weaken 
in early 1985, taking the Malaysian dollar with it, 
and effectively reducing the LME tin price from 
its historic peak. In October 1985, the buffer stock 
manager announced that the ITC could no longer 
finance the purchase of tin to prop up the price, as 
it had run out of funds, owing millions of pounds to 
the LME traders. This announcement caused near 
panic, the tin price fell to s and the LME 
halted all further dealings. In 1986 many of the 
world's tin mines were forced to close down due 
to the depressed tin price, and prices continued 
to fall in subsequent years. The following discus- 
sion therefore relates to tin processing prior to the 
collapse, including prices and costs. The same prin- 
ciples can be applied to the prices and costs of any 
particular period including the present day. 
It is fairly easy to produce concentrates 
containing over 70% tin (i.e. over 90% cassiterite) 
from alluvial ores, such as those worked in South- 
East Asia. Such concentrates present little problem 
in smelting and hence treatment charges are rela- 
tively low. Production of high-grade concentrates 
also incurs relatively low freight charges, which is 
important if the smelter is remote. 
For these reasons it has been traditional in the 
past for hard-rock, lode tin concentrators to produce 
high-grade concentrates, but high tin prices and 
the development of profitable low-grade smelting 
processes changed the policy of many mines 
towards the production of lower-grade concen- 
trates. The advantage of this is that the recovery of 
tin into the concentrate is increased, thus increasing 
smelter payments. However, the treatment of low- 
grade concentrates produces much greater prob- 
lems for the smelter, and hence the treatment 
charges at "low-grade smelters" are normally much 
higher than those at the high-grade smelters. Freight 
charges are also correspondingly higher. 
Suppose that a tin concentrator treats a feed 
containing 1% tin, and that three possible combi- 
nations of concentrate grade and recovery are (as 
in Example 1.1): 
High grade 
Medium grade 
Low grade 
63% tin at 62% recovery 
42% tin at 72% recovery 
21% tin at 78% recovery 
Assuming that the concentrates are free of 
arsenic, and that the cost of transportation to 
the smelter is s of dry concentrate, then the 
return on each tonne of ore treated can be simply 
22 Wills' Mineral Processing Technology