thebehaviourofminerals insulphideflotation processes ... location i main sulphide i gangue sulphide...

The behaviour of minerals in sulphide flotationprocesses, with reference to simulation andcontrol.

SYNOPSIS

by A. J. LYNCH*, Ph.D.,N. W. JOHNSON*, Ph.D.,D. J. McKEE*, Ph.D., andG. C. THORNE*, B.Sc. (Hons.) (Visitors)

In sulphide flotation processes, there are two mechanisms by which particles are transferred from the flotationpulp to the concentrate: adhesion to air bubbles and natural flotation, and entrainment in the froth. The entrainmentmechanism is particularly important for particles of non-sulphide gangue, which do not absorb collector. The im-portance of controlling the recovery rate of water was illustrated in plant testwork in which selectivity betweenvaluable sulphides and sulphide gangue, and between valuable sulphides and non-sulphide gangue, was altered bymanipulation of variables affecting the water recovery rate.

The two mechanisms should be recognized in the development of a control system. Most of the control schemesin operating plants have concentrated on reagent additions, which affect the natural flotation mechanism. Althoughcontrol of the chemical environment is clearly of paramount importance in flotation, plant tests have shown thatcontrol of the variables affecting the entrainment mechanism is also important.

SAMEVATTING

In die sulfiedflottasieprosesse is daar twee meganismes waardeur partikels uit die flottasiepulp na die konsentraatoorgedra word: adhesie aan lugborrels en natuurlike flottasie, en saamsleping in die skuim. Die saamsleepmeganismeis van besondere belang vir partikels van nie-sulfiedaarsteen wat nie die versamelaar adsorbeer nie. Die belang-rikheid van beheer oor die herwinningstempo van die water is geHlustreer in aanlegtoetswerk waar die selektiwi-teit tussen waardevolle sulfiede en sulfjedaarsteen en tussen waardevolle sulfiede en nie-sulfiedaarsteen gewysig isdeur die manipulasie van die veranderlikes wat die waterherwinningstempo be"lnvloed.

Die twee meganismes moet in die ontwikkeling van 'n kontrolestelsel erken word. Die meeste kontrolestelsels inbedryfsaanlegginge het hulle toegespits op die byvoeging van reagense wat die natuurlike flottasiemeganismebe"lnvloed. Hoewel beheer oor die chemiese omgewing vanselfsprekend van die allergrootste belang in flottasie is,het aanlegtoetse getoon dat beheer oor die veranderlikes wat die meesleepmeganisme bernvloed, ook belangrik is.

INTRODUCTION

In the concentration of mineralsulphides by flotation, advantage istaken of differences in the surfacenature of the mineral species toseparate the collector-coated val-uable sulphides from the morehydrophilic sulphide and non-sul-phide gangues. The metallurgicalefficiency is governed by the abilityof the process to produce sufficientdifference in the recovery rates ofthe minerals to obtain a highrecovery of the valuable sulphide atan acceptable grade. Selectivity isaltered by a change in the operatingvariables, e.g., reagent additions,pulp level, air addition.

In recent years, there has been amovement in industry towards auto-matic control, the incentive beinghigher metallurgical efficiency. Thisimproved efficiency stems fromgreater circuit stability and reagentsavings, and from closer operationto the process limits. Many attempts

*Julius Kruttschnitt Mineral ResearchCentre, University of Queensland, Aus-tralia.

have been made to describe the pro-cesses mathematically so that com-puter simulation techniques can beused to develop circuit optimizationand control procedures. Types ofmodel have been described by Wood-burn1 and Jowett et al.2. The im-portance of the recovery rate ofwater on the process has not gen-erally been recognized in simulationmodels, and this has resulted inpoor agreement between the ob-served and predicted plant per-formance.

In the development of the models,much emphasis has been placed ontheoretical and laboratory work, butlittle systematic data collection hasbeen done in operating plants. Con-sequently, some of the complexitiesin the flotation process, such as theeffect of entrainment, which mayexert marked influences on theresults of the process, have notbeen recognized. In this regard, it ispertinent to repeat the comments ofWoodburnl, that 'the formulation of[flotation] models can only be im-

proved if a clear understanding ofthe role that the model is destined to

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY

fulfil is possessed by the modeller',and that 'this normally requires aclose interaction between operationalpractice and modelling theory'.

In research work on flotationmodelling at the Julius KruttschnittMineral Research Centre (JKMRC),the differential flotation rates ofvaluable sulphides and gangue min-erals in sulphide ores have been in-vestigated. Testwork has been con-ducted in several concentrators treat-ing a variety of copper and lead-zincsulphide ores. The plant work hasbeen supplemented where necessarywith particular laboratory and pilot-scale tests. The data obtained, to-gether with such other plant data asare available, have been used in aninvestigation of the pattern ofbehaviour of the valuables andgangue minerals with a view todeveloping simulation and controltechniques. These patterns of be-haviour, as observed in operatingplants and as affected by changes insome variables with control possi-bilities, are discussed in this paper.

Bubble attachment to the col-lector-coated sulphide surface has

APRIL 1974 349

Company Location

I

Main sulphide

I

Gangue sulphide Other gangue Referenceminerals minerals

Philex Mining Corporation Luzon, Phillipines Chalcopyrite Negligible Quartz 93% Johnson I"(0,5 % Cu) Magnetite 6%

Mount Isa Mines Limited Queensland, Chalcopyrite Pyrite6%

Quartz 75% McKee"Australia (3,0 % Cu) Pyrrhotite Dolomite 10%

Mount Lyell Mining and Tasmania, Chalcopyrite Pyrite 10% Quartz 87% McKee"Railway Company Limited Australia (1,0% Cu)

0I

Peko Mines Limited Northern Territory, Chalcopyrite Pyrite 11% Magnetite 50%Australia (3,0 % Cu) Quartz 1...30% Johnson I"

Chlorite J 0

Bougainville Copper Pty Bougainville Chaleopyrite Negligible Quartz 95% J ohnson I"Limited Island, P.N.G. (0,7 % Cu) Magnetite 3%,

New Broken Hill N.S.W., Australia Galena, Marmatite Negligible Quartz Thorne12Consolidated Limited (8% Pb 13% Zn) Rhodonite

North Broken Hill Limited N.S.W., Australia Galena, Marmatite Negligible Quartz, Calcite Thorne12(13% Pb, 10% Zn) Rhodonite

TABLE IDESCRIPTION OF SULPHIDE ORES THAT WERE EXAMINED BY THE JKMRC

TABLE IISOURCES OF OTHER PUBLISHED DATA ON SULPHIDE ORES

Company--

Location

Nchanga Consolidated Copper Mines Limited

Main minerals Reference

Bull"

Zambia Chalcopyrite, malachite, quartz, silicates

Mount Isa Mines Limited . . . . . Queensland, Australia

Bull"

Kelsall13

Chalcopyrite, pyrite, quartz, dolomite

The Zinc Corporation Limited. . . . . . . . N.S.W., Australia Galena, marmatite

long been recognized as the principalmechanism for the recovery ofmineral sulphides by flotation.Considerable research (for example,Gaudin3, Glembotskii et al.4, Bull5,Woodcock and Jones6, 7) has beendone on the importance of variablesthat affect this mechanism. Theentrainment mechanism, in whichpulp is entrained between bubble-mineral aggregates and carried overinto the concentrate launder, wasdiscussed by JowettS in 1966, and ithas been shown in more recent workboth in the laboratory and in con-tinuous plants (Johnson et al.9) tobe responsible for the recovery offree non-sulphide gangue. This mech-anism is more selective with regardto particle size and specific gravity,but air bubble attachment ismore selective with regard to min-eral type since this determines theextent to which the particle surfacecan be made hydrophobic.

In this paper, the observed flo-tation behaviour of the mineraltypes is described, and an expla-nation for the behaviour is given interms of flotation mechanisms. Sim-

350 APRIL 1974

ulation models are also discussed.Various flotation control schemes arediscussed with reference to theobserved mechanisms.

The sulphide ores studied in theJKMRC test programme are listedin Table I.

The sources of other publishedplant data on sulphide flotationprocesses are given in Table H.

SUMMARY OF THEFLOTATION BEHAVIOUR

OF SULPHIDE ORESIN PLANT CIRCUITS

The recovery-time relationshipsfor sulphide minerals, non-sulphidegangue, and water are shown inFig. 1 for many of the rougher/scavenger banks discussed in thispaper. The high flowrates of con-centrate encountered in plant samp-ling usually make direct measure-ment of flowrates impracticable.Consequently, the recoveries of thecomponents were based on the flow-rates calculated by the two-productformula applied to assays and pulpdensities.

The various minerals in the differ-

ent ores showed a common patternof behaviour. The behaviour of thethree main groups (i.e., sulphidevaluable, sulphide gangue, and non-sulphide gangue) is discussed inmore detail in the next section. Therecovery-time curves for each sizefraction of minerals in the MountIsa rougher bank are shown in Fig. 2,and the consistency of behaviourfrom size to size will be noted. Fromthe data available on industrialsulphide flotation processes, the fol-lowing comments are made.(1) For each ore, there is a sig-

nificant difference between therates of flotation of the valuablesulphides, gangue sulphides, andnon-sulphide gangue.

(2) For each there is a similaritybetween the recovery rates ofnon-sulphide gangue and water.

(3) Even with the sulphide minerals,there is a dependence of re-covery rate on the recovery rateof water.

FLOTATION RATES ANDRECOVERY MECHANISMS

The recovery of a particular


>-a::w>0uwa::

Mt. Iso Chalcopyrite Nth. B H Galena1111-11 -2=32-3 3

1 ~23-

90 ,1 ~l;:::;:- ~6- Peke Chalcopyrite

/2

"><J6C(---Phile~~~;pyrite

80

/

, ,/lOON

BHe Galena

703 57

5S~ Mt. Lyell Chalcopyrite1 2 5

56

IY50 ,'tt/ Mt. Iso Pyrite

I 511

30

I

II/"

Peke Pyrite

20 M /'

Nth. B H Ma~matite

11 9----"

7 9------10 ~ ~o N B H C Marmatite

0---908

,o[~~~ Mt. Lyell Pyrite0 7~ 8

40

mineral is the combined result of theselective bubble-attachment andwater-entrainment recovery mech-anisms. The relative importance ofeither mechanism depends on thenature of the mineral (e.g., type,fineness of grind), the chemicalflotation environment (e.g., col-lector, modifier, salt concentrations),and the physical variables that affectthe availability of air in the pulpand the recovery rate of water(e.g., frother, air, pulp level). Thisis illustrated, for example, in thelead flotation circuit of New BrokenHill Consolidated Limited (NBHC).Galena is selectively floated free ofmarmatite and non-sulphide gangue.The recovery-size curves after sixrougher machines are shown inFig. 3.

For galena, because the mineralhas a high specific gravity and be-cause most of the particles have ahydrophobic surface, selective at-tachment is more important. Theshape of the curve is typical forvaluable sulphides (Kelsall and Stew-arP5, Morris16). Maximum recoveryis obtained in an intermediate sizerange between about 15 and 70 /Lm,with lower recovery in the finerand coarser fractions. For mar-matite, which is a less dense sul-phide than galena and activated toonly a minor extent, entrainmenthas a greater effect. In Fig. 3, theshape of the curve describing therecovery of the coarser sizes ischaracteristic of selective attach-ment, whereas the increase in re-covery in the fines is evidence ofentrainment. For the non-sulphidegangue, entrainment is still moreimportant, except for the smallpercentage in composites withfloatable sulphide.

A quantitative relationship hasbeen developed between th ~ re-covery rates of water and free non-sulphide gangue, and this will bediscussed later. These equations canbe extended to describe the influenceof the water recovery rate on therecovery rate of the sulphides.

FLOTATION MODELS

Non-sulphide GangueSome non-sulphide particles ad-

sorb sufficient collector to make themhydrophobic, and they float bybubble attachment. However, themajority of non-sulphide particlesdo not become hydrophobic, and

APRIL 1974 351

Fig. I-Behaviour ofsulphlde and gangue minerals and water In several chalcopy-rite and galena flotation processes

.JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY

100

60

~

WATER

>-a::w>0Uwa::

# Mt. Iso

r Peke

cP ./20 d /I /

r Mt. /. NBHC Nth. B H10 C Leel! . --=.L:.-:::8=Q-t:.-t:.-t:.

\ 8~8;:::::;- Phi lex 0. 0 0----,8+ +-+0+- - - --.0 -

16 NON-SULPHIDE GANGUE

.Peke/Mt. Iso /'

c .

8!:t '/Nth BH

4 I Lyell / _t:.-t:.-t:.c . t:.--t:.~ .1~~: O_8_N!-~~o-

__~~lexob+-, I I I I

4

NOMINAL

30

~

12>-a::w>0uwa::

~

8 12 16 20 24

RETENTION TIME (min.)

Symbol. Size Fraction.~(microns)

z 100 .- +147z .« . x- +104 -147

~80. 0-+74 -104.w + 0- +53 -74a::: x . . . . . .w 60 l1-+43 -53I- 0

0 ia:::<>~ v- +37 -43>-0...

40 ~..-+28 -370 0 xU 0 + x . -+25 -37--1 0 x« a + + x x x x

~0x x:r: 20 a a + + + to tou a a Dc Dc

~<fJt0~0 ~iv2i <P- ~<>II ~<JJ0~

100 ~. . .x . .Rv ~x x x . . . . .

0 x x0 0

x x x x x80 ~~0 0

~l1 0

Z . 0 v 0 00

Z <:8° l1 0 0 0 0 0

60 + v l1 0« . ~0 'g ~l1 0

W . i+ E l1 0a::: c 0 c c40 ~VG l1. l1 l1 l1 l1

W . <>+V0 vG '\P V0 V0

I- ~-+20 -28 . . <>+<>+a:::20 <> -+18 -25 . <>+ <>+ <>+ <>+>- .

0... . . . .~-+14-20 . . .. . ~..0 . - +12 -18

100 t io ~. .8 xC .~~&0R<

... V l1 . . . .

~~~vv f r xo Xo Xo Xcw . Xo~°

iv v & & H &z . ~itJO &.0 vz 90 . ~8

v v v Vif . 0« ~.- ~° 8 8 8~w .-~w . . ~.-.- .- ..a::: .~~w .~~~a::: ~*w . . .~.-+10-14 . . . . .3: + - -12 .

80 .w . . .w . .::J w~.- -10z

« w~0- Overall Result. w

w~w ww- Water. w w

701 2 3 4 5 6 7 8 9 10 11 12

e6MPARTME~r NUMBER

352 APRIL 1974

Fig. 2-Relationshlps between percentage of chalcopyrlte, pyrite, and gangueremaining and compartment number for various mineral size-fractions In theMount Isa plant rougher section. The relationship between percentage of water

remaining and compartment number Is also shown. .


1O0r

<1:z

~ 80<1:~

CL0

>-et:W

6 60uWet:

;;'<

4010

w

'=~Ler:<1:L

CL0

6

>-et:'JJ>0uWet:

;;'< 4

25

w 3:::>~z<1:~

CL0

~ 2w>0uWet:

;;'<

.- ........ "."'" "'/ .

../

./.

8

. Characteristic.of I

~ntralnment. I

I

:Characteristic of

:

air bubble attachment..~i

'",.'\

.\ .

\.

4

.~.

~. "'.""

.\.01 502 10

SIZE

203 5

PARTICLE IN MICROMETRES.

Fig. 3-The relationship between recovery after six rougher compartments andparticle size for galena. marmatlte. and non-sulphide gangue. NBHC lead rougher

bank


2nn

APRIL 1974 353

in this discussion the term non-sulphide gangue will refer to theseparticles. Most published modelsfor gangue recovery have used semi-empirical models that are identicalto the published semi-empiricalmodels for valuables. A low valuefor the first-order rate constant wasassumed. Jowett8 proposed a differ-ent type of model for gangue, whichassumed that mechanical entrap-ment is the major mechanism for therecovery of gangue.

This hypothesis was investigatedin JKMRC pilot-plant tests. Puresilica slurry was continuously floatedwith frother only at different pulpdensities and at varying waterrecovery rates. The relationship be-tween the recovery rates of waterand silica is shown in Fig. 4. Thisfigure gives strong support to the

importance of the water-entrain-ment mechanism for non-sulphidegangue. The relationship betweenthe recovery rates of water and non-sulphide gangue for some operatingplants is shown in Fig. 5.

There is also evidence that hy-draulic classification occurs in theentrained pulp in the froth column,and this is illustrated in Fig. 6,which shows what may be regardedas the classification function forparticles of non-sulphide gangue indifferent ores. It is probable that theparticles in region A are pre-dominantly free and those in regionB are predominantly composite.The small particles enter the con-centrate after being classified in therising column of entrained water inthe froth, and their behaviour canbe described by a classificationmatrix that can be written

(Recovery rate of free gangue" /Recovery rate of water)C

OFt-oncentrate

(Mass of free gangue*/Unit mass of water)Pulp

wr-«a::>-a::w>0Uwa::

Pulp Density

42,0 % solid~35,5 % solids24,0 % solids17.0 % solids

+

/ /~

+ 0

!+~c

0 8/8

0 0/ /.

"+tf 0/ /~Ol 0 ./0 / 0 ,/'+0 .

/0 /+00 /0 /'

/0' /0 /-/tP' gO .-.

+0/ ,.-/0 rrO .........

R6 "'iII""00

+0

0

.c0- 100E-..9

,00

50

«u.--Jtf)

WATER

500RECOVERY

1000RATE (g/min)

Fig. 4-Relationship between the recovery rates of silica and water at varying pulpdensities in a pilot-scale machine

354 APRIL .1974

orOFt = (RRFGt/ RR W)/(MFGU WPt)

. . . . . . . . (1)for the ith size fraction.This can be rewritten and summedover n size-fractions to give theequation

nRRFG=RRW. LOFt.

i= 1MFGUWPt. .. (2)

*io the ith size-fraction.In equation (2), the quantity

(MFGUWPt) is determined by thepulp density and sizing of thegangue in the compartments of arougher/scavenger section. For eachcompartment, other variables (e.g.,pulp level, air-addition rate, andfrother addition) determine thewater recovery rate (RRW) in equa-tion (2). From the collected resultsit has been noted that values in theclassification matrix (OFt) for aparticular type of free gangue areconstant over a normal range ofwater recovery rates. Typical valuesare listed in Table Ill. The valuesin the classification matrix effect-ively express the efficiency of thewater in transporting the varioussize-fractions of free non-sulphidegangue from the pulp, through thefroth column, and into the concen-trate.

Equation (2) can be modified togive an important alternative form:

RRFGt=RRW . OFt. MFGUWPt.(3)

-RRW . OFt. MFGtMW AT . . . (4)

=kt . MFGt. (5)

Equation (5) is simply the trad-itional first-order rate equation forthe flotation process. However, if arougher/scavenger bank is con-sidered, the value of the rate constantfor the ith size-fraction of ganguecan be evaluated for each compart-ment from the quantities (RRW),(MW AT), and (OFt). It is importantto note that the recovery rate ofwater is related to the physicalvariables in all operations and maychange down a bank. This is thereason why a conventional rateequation to describe gangue be-haviour must be treated with care.

This discussion has been confined


-420 -300 -210 -150 -75 -44 -33 -23 -16 -11Size-fraction (microns)+420 +300 +210 +150 +75 +44 +33 +23 +16 +11

,04 ,11 ,24 ,44 ,83Va]ue in classification 0 0 0 0 0 ,03matrix (CFi)

Rougher First Second Thirdscavenger scavenger scavenger

--Survey 10-12-70

77,5 84,] 86,2 86,9ObservedPredicted 78,9 83,3 85,5 86,9

Survey 12-]2-7084,3 87,.5 88,6 89,0Observed

Predicted 83,0 86,9 88,8 90,0

Survey 17-12-7083,4 87,2 88,8 89,5Observed

Predicted 82,7 86,5 88,5 89,6

Survey 22-12- 7080,8 86,1 87,9 88,3Observed

Predicted 81,5 85,4 87,5 88,8

TABLE IIITYPICAL VALUES IN THE CLASSIFICATION MATRIX }'OR SILICEOUS GANGUE

to particles that do not adsorbcollector in sufficient quantities tomake them partially floatable. Someparticles of non-sulphide gangue dofloat by adhesion to air bubbles, andthe classification function is in-adequate to describe their behaviour.These should really be treated in thesame manner as the sulphide part-icles, but the problem of identifyingthese particles and describing themquantitatively has not yet beensolved.

The importance of entrainment isimplicitly recognized by flotationoperators who vary pulp depths an.dair additions to cells as part of theirnormal operating routine. It is im-portant that it should be explicitlyrecognized in simulation models andcontrol systems, and, for this pur-pose, data on the flotation rate ofwater as well as the mineral speciesmust be collected. It seems reason-able to suggest that this mechanismis just as important with oxideflotation as with sulphide flotation,but data for testing the hypothesisare not available.

Sulphide Valuables

The flotation of valuable sul-phides is essentially a rate process.Wood burn and Lovedayi7, Harrisand Chakravartii8, and others haveproposed semi-empirical modelsusing a continuous distribution offlotation properties. Thp, equationthat describes the process for aparticular mineral in a single batchcompartment is

00-kt

M =M 0 J e f (k,O) dk .. (6)0

and in a continuous compartment is

.00 00-kt

M =M 0J J e E(t) f (k,O) dt dk.

0 0. . . . . . . . . . . . . (7)

The problem in using these equa-tions, which are first-order andrecognize the existence within avaluable component of species havingdifferent flotation properties, is todetermine the form of the continuousdistribution of rate constants.

Other authors (for exampleImaizumi and InoueI9, Kelsalli3)

TABLE IVCOMPARISON OF OBSERVED AND PREDICTED VALUES OF PERCENTAGE RECOVERY FOR

VALUABLES IN THE PHILEX ROUGHER/SCAVENGER SECTION

High rate constant(min-I)

Low rat~ constant

I

Proportion of(mm-I) mineral with

Ilow rate constant

Survey 10-12.70 . .12-]2-7017.12-7022-12-70 . .

I

1,331,331,331,33

0,0250,0250,0250,025

0,2200,]850,1800,200


have used a discrete approximationto the form of the continuous dis-tribution:

FCR=</J-ks~(l-</J)e-kft.

This approximation can be il-lustrated by reference to Fig. 2, inwhich it will be noted that each size-fraction of chalcopyrite can be re-garded as containing fast- andslow-floating fractions, the majordifference between the size-fractionsbeing the proportion of the fractionthat is slow floating. Kelsall de-scribed this behaviour for a copperore in 1961, and it has since beenfound that the same comment canbe made about the valuable mineralsin many ores. Thus, the three para-meters that describe the behaviourof the valuable mineral on a size-by-size basis are as follows:(1) the flotation rate constant of

the fast-floating fraction in eachfraction (kf),

(2) the flotation rate constant ofthe slow-floating fraction in eachfraction (ks), and

(3) the proportion of the slow-floating fraction «/J).

The approximation that each sul-phide mineral exists in the orewholly as fast- and slow-floatingcomponents can be demonstratedwith respect to the Philex ore. Foursurveys were carried out on therougher/scavenger banks and, foreach survey, the rate constants ofthe fast- and slow-floating fractions,and the proportion of slow-floating,were calculated from the followingequation:

Recovery rate of chalcopyritefrom a compartment=kfMf+ksMs.

The results, and a comparison be-tween the observed and calculatedresults for each survey, are given inTable IV.

The value of using this simpleapproximation in simulating ~hebehaviour of the valuable sulphideminerals is evident.

Further results have shown thatthe recovery rate of water also in-

APRIL 1974 .

Mt. Iso rougher section.Mt. Lye!! rougher section.Philex rougher-scavenger

section.Peko rougher-scavenger

section.Nth. B H rougher

section.N B H C rougher

section.

10 15 20 25

RECOVERY OF WATER

D

14 .0

12w=><..9Z«<..9

x

10 1:1.

LL0

8 +

>-0:::W>0uW0:::

6

4 1:1.

1:1.4:x

1:1.

0 1:1.++

o~+~0 2

00 5%

x

DD

D

DD

x D

D

D

D

D

X D

30

Fig. 5-Relationship between recovery rates of non-sulphide gangue and water inseveral chalcopyrite and galena flotation processes

fluences the recovery rate of sulphideparticles, and that the recovery ofeach sulphide mineral should bedescribed by the equation

Recovery rate of mineral from acompartment (RRV)

= (k 'f+ kfw)Mf+ (k' s+ks w)M s.

. . . (8)The rate constants k'f and k's are themore basic rate constants, depend-ent on the chemical environmentand obtained at a minimum waterrecovery rate. The rate constantskfw and ksw describe the con-tribution of the water-entrainmentmechanism, which depends on therecovery rate of water controlled bythe physical operation of the cell,i.e., froth depth, air addition, andfrother addition.

The effect of varying waterrecovery rate on the recovery of avaluable by the entrainment mech-

;856 APRIL1974

anism can be described by a modi-fied form of equation (4):

Recovery rate of fast floatingvaluable of size i by entrain-ment

RRW- MW AT

. CFi . Xi . Mf,

(9)

where Xi=proportion of fast-floating valuable inthe ith size fraction.

Over 11size fractions,Recovery rate of fast-floating

valuable by entrainment

RRWn

=Mf' MW AT.

i~lCFi. Xi

=Mf.kfw. (10)

By substitution of the values ofkfw and a similarly derived ksw inequation (8),

RRW nRRV =(k'f+ MWAT

.i~l

CFi .X;}Mf+(k~+ :;~T

n

.L CF'i . Yi)Ms,i= 1

. . . (ll)where Yi=proportion of slow-

floating valuable in theith size fraction.

This is the proposed model for therecovery of sulphide minerals. Itincorporates the recovery rate ofwater, which has been observed toaffect the rate of recovery of val-uables. Experimental studies of theeffect of entrainment on the rates offlotation are to be carried out. Theelements of the classification matriceswill be determined in a series ofsteady-state pilot-scale tests.

It is clear that the method ofrepresenting a comminuted sulphidemineral in terms of fast- and slow-floating fractions is a considerablesimplification of the real case. De-spite this, the approximation gives amodel that is simple and that ac-curately describes the behaviour ofthe valuable sulphides in severaldifferent ores, and it appears to be ofgeneral use.

Sulphide GangueThe behaviour of gangue sul-

phides has been observed to besimilar to that of valuable sulphides,but the rates of flotation are lower.The rate of recovery of water has alarger proportional influence on therate of recovery of gangue sulphidesthan of valuable sulphides. This canbe explained by the greater im-portance of the water-entrainmentrecovery mechanism for gangue sul-phides.

SummarySome of the features common to

sulphide flotation processes are asfollows.(a) The valuable sulphides can be

regarded as containing fast-floating and slow-floating frac-tions, and the total rate of flo-tation is a function of thebehaviour of these fractions andthe rate at which water is re-covered.


0..

:J0..

'w I

:)0::<..9Wz~«3<..9

LLLL00

tf)(f)tf)tf)««LL

I I

1,0

0,9\

0,8 fJ

0,7 \

~0,6 0\ A§0,5 tu x

I Ol. 0\0::

~

~ 0,3\~

~ 02 -8-00,1

\:

00

I

W:J<..9Z«<..9

LLLLO

°(f)(f)(f)(f)««LLI

Alternative control objectives canbe constant concentrate grade atmaximum mineral recovery, maxi-mum grade at constant recovery,maximum throughput at some re-quired grade and recovery, optimi-zation of reagent usage, economicoptimization of circuit performance,or some combination of these. Theprimary function of the controlscheme is to detect a disturbance andto alter the control variables tomaintain the control objectives.Reagent additions are frequently

><0 used as control variables, and thelevel of control varies from manualadjustment to digital control such

240 as is used at the Ecstall concen-trator (Amsden et al.2O).

Disturbances to the process mayinclude changes in feed grade or oretype. If the disturbance is a changeonly in feed grade, a change inreagent addition may be sufficient toallow for the altered mass of sul-phide in the pulp. Changes may alsobe necessary in the physical vari-ables if the correct flowrates ofconcentrates are to be maintained,particularly where there is a heavy

8 J KM R C pilot plant. Silica

pulp. 33,5% solids.

Philex Mining Corporation.

Rougher section.

Mount Iso Mines Limited.

Rougher section.

NBHC Limited.Rougher section.

x

0

0

0

~

Bo_~XJ

80

40 80 120 160 200

NON-SULPHIDE GANGUEPARTICLE SIZE (microns)

Fig. 6--Relationship between the classification function, which describes hydraulicclassification in the entrained pulp in the froth column, and particle size for gangue

in several sulphide rougher sections

(b) The gangue sulphides behavein a similar manner to thevaluable sulphides, except thatthe flotation rate constants aremarkedly reduced.

(c) The composite particles appearto behave as slow-floating sul-phide particles, although the

100

slow-floating fractions may alsocontain some liberated particles,particularly in the very finefractions.

4 CELLTAILING

(d) The free particles of non-sulphide, siliceous gangue enterthe concentrate by an entrain-ment mechanism, and a hy-draulic classification mech-anism operates while they areretained in the froth. Theirrecovery rates can be related tothe recovery rate of water.

CONTROL OF SULPHIDEFLOTATION PROCESSES

12 CE LL

TAILING

BO

---= = =:. -::. -= - Valuable: ""'ph.de

FEED

~

~GO

J~ 40

:

'"

~,.

-'

COLLECTOR ADDITION

4 CELLCONCENTRATE

Fig. 7- The typical responses of valu-able and gangue sulphide recoveries to

collector addition

.

0

X R F Analysis (%Cu, 0;0

Mag netic Flowmeter.

Fe, %Solids.)


Fig. 8--lnstrumentation on the test rougher bank at Mount Isa

APRIL 1974 357

rW +-.0 22<l:0::(9

w 20>

~-.J18::)

2:::)u

16 .90 91

\.. --.J92 93 94 95

a/a RECOVERYw

~300 .0::-/c

ffi ~~OO

L"/+\

> 0) \

o..Y '!

[cl ---100 ~+0:: ....0::W

00 ~4 8 12 16~CELL NUMBER

Compartment No.

~ 12~ .0 6

~ i116 x 5 - 8offi + 1COt-« ~12

/x +

~ ~ 8 IQ x )

~q.- . +

~ 4 ~ /x ,/5? ~

x +LL 0 -

0 2 4 6 8 10 12 14 16 1820% COPPER

Eu

Fig. 9 - Relationship between copperassays and height above the froth-pulpinterface for concentrate from various

points in Mount Isa rougher bank

flaw. Assays .of the feed. and banktailing may be sufficient in this casefar carrective actian. If, an the .otherhand, the disturbance is a change inare type, it is likely that a differentchemical enviranment is necessaryta give best flatatian perfarmance. Inthe semi-empirical madel, the pro-partian .of slaw-flaating mineral (cp)must be altered. A feed and tailingassay is nat sufficient ta allawcarrective actian. It has been faundin plant tests (Lynch21) that thefirst few cells in a bank suitablyequipped with sensing instrumentscan be used as an an-line detectar .ofchanges in are flaatability, i.e., aretype.

Importance of Chemical and PhysicalControl Variables

Ta effectively cantral a sulphideflatatian pracess, bath the physicaland chemical variables shauld beincarparated in the cantral scheme.The chemical variables are mast use-ful far altering the rate .of recavery.of particles by the selective-attach-ment mechanism. Mast flatatiancantral schemes use reagents as thecantral variables; far example, Ecs-tall,~NBHC (Presgrave22), Lake Du-fault (Smith and Lewis23 andLewis24), and Tennessee Capper'sLandan mill (Faulkner25). Cantrol.of the chemical enviranment is anessential part .of any cantral scheme.

The typical respanse .of sulphidemineral recavery ta a change incallectar cancentratian is shawn inFig. 7. In many cantral schemes, anattempt is made ta minimize reagentcansumptian and maximize can-centrate grade by .operating as claseas passible ta the edge .of the plateau

358 APRIL 1974

regian an the curve far valuablesulphides. A scheme far the manitar-ing .of circuit feed and .of cancentrateand tailing assays, and far thesystematic variatian .of reagents in asearch far the .optimum grade-recavery paint has been imple-mented successfully at the Ecstallcancentratar20, This was a majaradvance in the evalutian .of cantrolschemes.

In .other plant testwark, thecantrol .of the physical variables hasbeen faund ta be necessary taminimize the water recavery rate,which affects the recaveries .of sul-phide and nan-sulphide gangues bythe entrainment mechanism(Tharne12 and Lynch21). It wasshawn that it is passible ta vary thegrade-recavery curve far a particularmineral by manipulatian .of thevariables that cantral .only the frathbehaviaur, withaut alteratian .of the

chemical enviranment. Since thiswark canfirmed the impartance .ofthe entrainment mechanism, it willbe discussed here in mare detail.

Experimental Systems

The testwark was canducted inthe Na. 1 capper cancentratar atMaunt Isa during the periadDecember 1972 ta April 1973, andalsa in the lead-zinc cancentratar atNBHC at Broken Hill during theperiad February ta May 1973. Thechalcapyrite flatatian circuit in theNa. 1 cancentratar at Maunt Isaincluded six .open-circuit Fagergrenbanks in parallel. One bank was setup far the test pragramme with theinstrumentatian shawn in Fig 8. AHewlett Packard 2100A digital cam-puter was used far data callectianand pracessing an-line. It was natpassible ta vary the air additian tathe cells, but the frath behaviaur

Fig. IQ-The influence of change in water recovery rate on the copper grade-recovery curve, Mount Isa rougher bank


NB HC Lead rougher. t:,.First six compartments. /

Froth Depth (cm.) /+

" 10,8

,~/!

+ 12,1 0/ /° 14,0

;~:z ;)+

0

/0

//° .

l/O kx/0 ~~~

~y ~~ ~

:::::-----

~/

1 2 3 4

CELL NUMBER.

.8

7

6

er::w 5f-

~

.x

4LL0

>- 3er::w>8 2wer::

0 1~

00 5 6

Fig. 11- The effect of change in froth depth on water recovery from the N BHCrougher bank

was changed by alteration of frothdepth.

In the galena flotation circuit atNBHC, the test bank was an open-flow rougher bank of Denver 21machines, which treated about 35per cent of the total rougher feed.The only measuring instrument onthe bank was a magnetic flowmeteron the feed line. In this case also,froth behaviour was changed byalteration of the froth depth.Steady-state Results

The relationship between heightabove the froth-pulp interface andcopper assay for froths from variouspoints in the Mount Isa rougherbank is shown in Fig. 9. This showsqualitatively the importance of thefroth as a concentrating region inthe cell.

The influence of change in waterrecovery rate, obtained by changein froth height, on the grade-recovery curve for the Mount Isarougher bank is shown in Fig. 10.

The improvement in the grade-recovery curve with decreased waterrecovery is the result of an in-creased drainage from the frothof particles that arrived there byentrainment. The effect of change infroth depth on the water recoveryrate in the NBHC rougher bank isshown in Fig. 11, and on the select-ivity of separation of galena frommarmatite in Fig. 12. The improve-ment in selectivity with increasingfroth depth is due to the increasedpreferential drainage from the frothof that fraction of the ganguesulphide (in this case marmatite)which reached the froth by entrain-ment.

On-line tests

On-line tests were carried out atMount Isa in which the froth depthwas varied and the effect on theconcentrate grade and recovery wasobserved continuously. A typicalresult is shown in Fig. 13. It is clear


that the water recovery can becontrolled to optimize grade.

CONCLUSIONS

The observed behaviour ofminerals in sulphide flotation pro-cesses has been described. Modelsthat recognize the importance ofthe water-entrainment mechanismhave been discussed. Plant testworkhas shown that it is possible tooptimize the grade-recovery curve bycareful control of the water recoveryrate, and that control of the waterrecovery rate should be an integralpart of a flotation control system.

ACKNOWLEDGEMENTSThis work was supported

financially by Mount Isa MinesLimited and the Australian MineralIndustries Research Association.The assistance given by the com-panies mentioned in the paperwhile the work was being carried outis gratefully acknowledged.

OFi

NOMENCLA TURE

the ith element of theclassification matrix de-scribing the hydraulic classi-fication of mineral particlesduring their passagethrough the froth columnrecovery rate of watertotal recovery rate of freeganguemass of water in a com-partmentmass of free gangue in theith size-fraction in a com-partmentrate constant for the ithsize-fraction of ganguemass of mineral in a batchcompartment, or flowrateof a mineral in the tailingfrom a continuous com-partment (mass/time)residence time distributionof solids (in tailings exit)mass of mineral initially ina batch compartment orflowrate of a mineral in thefeed to a continuous com-partment (mass/time)first-order rate constant fora mineralfraction of component un-recovered after time tmass of sulphide mineralin the pulp of a compart-

RRWRRFG

MW AT

MFGi

ki

M

E(t)

Mo

k

FOR

Mf,Ms

APRIL 1974 359

ment with high and lowrate constant respectively

The other symbols used in this paperare defined in the text.

REFERENCES

1. WOODBURN, E. T. Mathematicalmodelling of flotation processes.Miner. Sci. Engng, vo!. 2, no. 2.1970. pp. 3-17.

2. JOWETT, A., PICKLES, G., andWILLIAMS, P. H. Applications ofmathematical models of flotation,Automatic control systems in mineralprocessing plants (Symposium Aust-ralas. Inst. Min. Metall., Brisbane).1971. pp. 197-209.

3. GAUDIN, A. M. Flotation. New York,McGraw-Hill, 1957. Chapter 10.

4. GLEMBOTSKII,V. A., KLASSEN, V. 1.,and PLAKSIN, 1. N. Flotation. NewYork (Primary Sources). Translatedby R. E. Hammond. 1963. Chaptersn-v.

5. BuLL, W. R. The development of amethod of digital computer simu-lation of the flotation process bymeans of a mathematical model.Ph.D. thesis, University of Queens-land, 1966.

6. WOODCOCK,J. T., and JONES, M. H.Chemical environment in Australianlead.zinc flotation plant pulps. 1. pH,redox potentials, and oxygen con-centrations. Proc. Australas. Inst.Min. Metall., no. 235. 1970. pp. 45-60.

7. WOODCOCK,J. T., and JONES, M. H.Chemical environment in Australianlead-zinc flotation plant 'pulps. n.Collector residuals, metals in solution,and other parameters. Proc. Australas.Inst. Min. Metall., no. 235. 1970.pp. 61-76.

8. JOWETT, A. Gangue mineral con-tamination of froth. Br. Chem.Engng., vol. 2, no. 5. 1966. pp. 330-333.

9. JOHNSON, N. W., McKEE, D. J., andLYNCH, A. J. Flotation rates of non-sulphide minerals in chalcopyriteflotation processes. A.l.M.E AnnualMeeting, California, February 20-24,1972.

10. JOHNSON, N. W. The flotation be-haviour of some chalcopyrite ores.Ph.D. thesis, University of Queens-land, 1972.

11. McKEE, D. J. Studies in the controlof chalcopyrite flotation processes.Ph.D. thesis, University of Queens-land, 1972.

12. THoRNE, G. C. The behaviour of thegalena flotation circuits at NorthBroken Hill Limited and New BrokenHill Consolidated Limited. St. Lucia(Australia), Julius Kruttschnitt Min-eral Research Centre, Internal Re-ports No. 26 and 27. 1973.

13. KELSALL, D. F. Application of prob-ability in assessment of flotationsystems. Bull. Inst. Min. Metall.,vol. 70. 1961. pp. 191-204.

14. BuLL, W. R. The rates of flotation ofmineral particles in sulphide ores.

360 APRIL 1974

5 NHBCFirst

Lead rougher.sixteen compartments

t:,

/0

Froth Depth (cm) 11

ce~l No 1 9 16 ix

22,9 17,8 12.120,3 16,5 6,3 Ix

60 flQ

1-/X

I ! I

20

RECOVERY

WI-I-<!::La:<!::L

'-+

3LL0

0

f:j.

>-a:w>0uwa:

2

~

00

'.!fo

!~O GO

GALENA.

80

OFFig. 12- The effect of change in froth depth on the selectivity curve describing

galenajmarmatite recovery in the NBHC lead rougher bank.

Proc. Australas. Inst. Min. Metall.,no. 220. 1967. pp. 69-78.

15. KELSALL, D. F., and S"l'EWART,P. S. B. A critical review of appli-cations of models of grinding andflotation. Automatic control systems inmineral processing plants (SymposiumAustralas. Inst. Min. Metall., Bris-bane). 1971. pp. 213-232.

16. MORRIS, T. M. Measurement andevaluation of the rate of flotation as afunction of particle size. Trans. Soc.Min. Engrs AIME, no. 187. 1952.pp. 91-95.

17. WOODBURN, E. T., and LoVEDAY,B. K. Effect of variable residencetime on the performance of a flo-tation system. J. S. Afr. Inst. Min.Metall., vo!. 65. 1965. pp. 612-628.

18. HARRIs, C. C., and CHAKRAVARTI, A.Semi-batch froth flotation kinetics:species distribution analysis. Trans.AIME, 1970. vol. 247. pp. 162-172.

19. IMAIZUMI, T., and INouE, T. Kineticconsiderations of froth flotation. Min-eral processing, ed. A. Roberts, 6thInt. Min. Proc. Congress, Cannes,1965. pp. 581-583.

20. AMSDEN, M. P., CHAPMAN, C., andREADING, M. G. Computer control of

flotation at the Ecstall concentrator.10th Int. Symp. on Application ofComputer Methods in the MineralIndustry, Johannesburg, 1972.

21. LYNCH, A. J. The operation andcontrol of a chalcopyrite rougherflotation bank. St. Lucia (Australia),Julius Kruttschnitt Mineral ResearchCentre, Technical Report, Jan.-Jun.1973. pp. 1-35.

22. PRESGRAVE, D. K. Development ofthe control system installed in TheNew Broken Hill Consolidated Limit-ed concentrator, Australia. Proc. 9thComm. Min. Met. Congr., ed. M. J.Jones, London, Instn. Min. Metall.,1969. pp. 717-732.

23. SMITH, N. W., and LEWIS, C. L. Com-puter control experiments at LakeDufault. Can. Inst. Min. Bull.,vo!. 62, no. 682. 1969. pp. 109-115.

24. LEWIS, C. L. Application of a com-puter to a flotation process. Can.Inst. Min. Bull., vo!. 64, no. 705.1971. pp. 47-50.

25. FAULKNER, B. P. Computer controlimproves metallurgy at TennesseeCopper's flotation plant. Min. Engng,N.Y., no. 18. 1966. pp. 53-57.


l- -- ---

c :- ~=!Froth FrothDepth Depth Froth Depth 8.9 cm

8.9 cm 3,8 cm3 -------------. -...." .. .". . .......y". .. Off-line ~.. ... ... I

% CuIN FEED

2% Cu--

-0,3

IN 12 CEllTAilS

- -" .% Cu

0,6

IN 4 CErlTAilS. 0,5

... .. .."" .."" "" "" ", "... ..

-.------. .. . - - ----. . .... ...'... . .e. ... .. ...:------.---.---.-.-.._-. . .. ... ......1 .. .. ..C>j

- ---0,425

.. .....

% Cu IN1- 4 CEllROUGHER

CDNC..

.~ e. 0 0...-. "' - .. ...-I.. . ' ... .... . . ..I

.. ... ... .

-"O;"""~I---- - - - - -- - - - -- - - ----

20 - -~',-- - --. .... ......15 - - -

100

A122 4 6

I~~

8

HOURS.

10 14

TIME

Fig. 13-Results from the No. 5 copper rougher bank, Mount Isa. to demonstratethe effect of change in froth depth on the process.

Book reviewRamsey, Robert H. Men and

mines of Newmont. A fifty yearhistory. New York, Octagon Books,1973. $11.00.

It is interesting to consider howone man can influence a develop-ment by courage, luck, and perse-verance. This is largely the story ofhow Colonel WaIter Boyce Thomp-son and his later associates, datingfrom the year 1921, founded amining and finance empire; a pre-vious company started in 1916apparently gave him the experienceto go on to bigger things despiteearlier set-backs. Hence, he en-visaged that Newmont, with an

initial capital of $8 million, wouldbe a trading vehicle for his manypromotions. It has since (1971) be-eome a eompany with a market valueof $950 million spread over 24million ordinary shares.

The other chief executives wholargely helped towards this achieve-ment were Charles F. Ayer (a lawyer),Fred Searls (a metallurgist), PlatoMalozemoff (a financier), and MarcusD. Banghart (a mining engineer, whospent some time in Southern Africa).

Newmont's largest project is per-haps San-Manuel-Magma CopperMine in Arizona. It also has accessto many others: inter alia, Kenne-


cott and Texas Gulf Sulphur inCanada; O'Okiep, Tsumeb, and Pala-bora in South Africa; elsewhere, toD.S.A. companies like Homestake,Hanna, etc., through the associationof one-time Newmont personnel(e.g., John Gustafson). This indi-cates a spirit of co-operation on aglobal scale that has paid handsomedividends.

All are mentioned in a skilfullywoven tale of success and failures bythe author, who was a MetallurgicalConsultant to Newmont and is nowa member of the staff in New York.

J.T.M.

APRIL 1974 361

NIM ReportsThe following reports are avail-

able free of charge from the Nat-ional Institute for Metallurgy, Priv-ate Bag 7, Auckland Park 2006.

Report No. 1567

Estimation of the parameters in thedistributed-constant flotation model.

A complete system is developedfor the estimation, from batch orcontinuous data, of the parametersin the distributed-constant flotationmodel. Various hypotheses concern-ing the model are tested, and adescription is given of a method bywhich a particular slurry can beadequately and significantly charac-terized from available data.

The model can predict total re-covery as a function of residencetime, but is not able to predict theparticle-size distribution of the con-centrate. The incorporation of asimple model for froth behaviour

results in it marked improvement inthis respect.

Report No. 1591The inductive measurement of con-

ductivity in slags.An investigation of the use of the

inductive-loop method for the de-termination of conductivity in slagsresulted in the design of severalinstruments. The development ofthe different systems is described,and the advantages and disad-vantages of the various types ofinstruments are discussed.

Report No. 1609An interlaboratory analysis of two

ferrochromium slags.This report describes the analyses,

by ten different laboratories, of twoslags, one associated with the pro-duction of high-carbon ferrochrom-ium, and the other with the pro-duction of ferrochromium-silicide.The results, their statistical treat-

ment, and the methods of analysisused are given. Preferred values areassigned to the two samples.

Report No. 1611The X-ray-fluorescence determi-

nation of percentage concentrations ofniobium and tantalum in ores andminerals.

This report describes the investi-gational work in the development ofa procedure for the determination oftantalum and niobium in ores wherethe concentration of either elementexceeds 0,5 per cent. The samplesare decomposed by the use of a fluxthat incorporates two internal stand-ards, and the technique obviates theneed for the determination of back-grounds and corrects for matrixeffects. Calibration is achieved bymeans of synthetic standards. Theprecision of the method compareswell with chemical methods, and thetime required is much less.

Company affiliatesThe following members have beenadmitted to the Institute as Com-pany Affiliates.

AE & Cl Limited.AfroxjDowson and Dobson Limited.Amalgamated Collieries of S.A. Limit-

ed.Apex Mines Limited.Associated Manganese Mines of S.A.

Limited.Blackwood Hodge (S.A.) Limited.Blyvooruitzicht G.M. Co. Ltd.Boart & Hard Metal Prod ucts S.A.

Limited.Bracken Mines Limited.Buffelsfontein G.M. Co. Limited.Cape Asbestos South Africa (Pty) Ltd.Compair SA (Pty) Limited.Consolidated Murchison (Tvl) Gold-

fields & Development Co. Limited.Doornfontein G.M. Co. Limited.Durban Roodepoort Deep Limited.East Driefontein G.M. Co. Limited.East Rand Prop. Mines Limited.Free State SaaiplaasG.M. Co. Limited.Fraser & Chalmers S.A. (Pty) Limited.Gardner-Denver Co. Africa (Pty) Ltd.Goldfields of SA Limited.The Grootvlei (Pty) Mines Limited.

Harmony Gold Mining Co. Limited.Hartebeesfontein G.M. Co. Limited.Hewitt-Robins-Denver (Pty) Limited.Highveld Steel and Vanadium Corpo-

ration Limited.Hudemmco (Pty) Limited.Impala Platinum Limited.Ingersoll Rand Co. SA (Pty) Ltd.James Sydney & Company (Pty)

Limited.Kinross Mines Limited.Kloof Gold Mining Co. Limited.Lennings Holdings Limited.Leslie G.M. Limited.Libanon G.M. Co. Limited.Lonrho S.A. Limited.Loraine Gold Mines Limited.Marievale Consolidated Mines Limit-

ed.Matte Smelters (Pty) Limited.Northern Lime Co. Limited.O'okiep Copper Company Limited.Palabora Mining Co. Limited.Placer Development S.A. (Pty) Ltd.President Stern G.M. Co. Limited.Pretoria Portland Cement Co. limit-

ed.Prieska Copper Mines (Pty) Limited.Rand Mines Limited.

Rooiberg Minerals Development Co.Limited.

Rustenburg Platinum Mines Limited(Union Section).

Rustenburg Platinum Mines Limited(Rustenburg Section).

St. Helena Gold Mines Limited.Shaft Sinkers (Pty) Limited.S.A. Land Exploration Co. Limited.Stilfontein G.M. Co. Limited.The Griqualand Exploration and Fi-

nance Co. Limited.The Messina (Transvaal) Develop-

meltt Co. Limited.The Steel Engineering Co. Ltd.Trans-Natal Coal Corporation Limit-

ed.Tvl Cons. Land & Exploration Co.Tsumeb Corp. Limited.Union Corporation Limited.Vaal Reefs Exploration & Mining Co.

Limited.Venterspost G.M. Co. Limited.Vergenoeg Mining Co. (Pty) LimiteVlakfontein G.M. Co. Limited. d.Welkom Gold Mining Co. Limited.West Driefontein G.M. Co. Limited.Western Deep Levels Limited.Western Holdings Limited.Winkelhaak Mines Limited.

362 APRIL 1974 JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY

thebehaviourofminerals insulphideflotation processes ... location i main sulphide i gangue sulphide...

Documents