fundamental studies of the flotation process:

8/10/2019 Fundamental Studies of the Flotation Process:

1/14

Fundam ental Studies of the Flotation Process:The Work of the National Institute for M etallurgy

by N. P. Finkelstein,* M .Se. Ph.D (Fellow) and V. M . Lovell,* M .Se. Ph.D (Visitor)

SYNOPSIST he co nc epts in vo lv ed in flo ta tio n re sea rch a re d isc us se d a nd a im s a re p re se nte d. T hre e d iffe re nt m eth od.

o f in ve stig atio n c an be u sed , a ll o f w hich a re a pplied a t t he N atio na l Ins titu te fo r M eta llu rg y N .I.M .), a nd th es ea re th e e mp iric al, tria l a nd e rro r m eth od , th e c he mic al e ng in ee rin g m eth od , a nd th e p hy sic oc he mic al a pp ro ac hsT his paper is devoted to the last of these.

B ec au se o f th e c om ple xity o f th e p hy sic oc he mic al in te ra ctio ns in vo lv ed , th e w ork h as , to a la rg e e xte nt, b ee nc on ce ntra te d o n on ly o ne a sp ec t, viz. th e in te ra ctio n b etw ee n th e m in era l s urfa ce a nd re ag en ts in so lu tio n,w hich w as considered of m ajor im portance in the understanding of the process. The aspects on w hich w orkhas been done are the concentration of uranium , the reaction betw een salt m inerals and fatty acids, m ixedx an th ate -c atio nic c olle cto r s ys te ms , a nd th e re ac tio n b etw ee n s ulp hid e m in era ls a nd th io l re ag en ts . S ele cte dproblem s from the first tw o of these are discussed in som e detail and deal w ith the agglom eration flotationof uranium and the chem ical control of the flotation process for phoscorite. A detailed discussion is alsop re sen te d o f th e p ro gre ss m ad e in th e u nd ersta nd in g o f th e in te ra ctio n b etw ee n su lp hid e m in era ls a nd th io lre ag en ts , w hic h is o f m ore g en era l a pp lic atio n.

F in ally , a n in dic atio n is g ive n o f the fu tu re d ire ctio n o f p hysic oc he mic al re se arch a t the N .I.M .

SINOPSISD ie b eg rip a ang aa nd e flo ta sie n av orsin g w ord b esp re ek e n d ie d oe ls te llin gs d aa rva n u ite en ge sit. D rie v er-

sk ille nd e o nde rso ek m eto de s, alm al ta ns d eu r d ie N as ion ale In stitu ut va n M eta llu rgie N .I.M .) a an ge we nd ,k an v an g eb ru ik g em aa k w ord ; n l. d ie e mp irie se p ro be er e n fo ute er m eto de , d ie c he mie se in ge nie urie se m eto de ,asook n fis ie s-c he mie se b en ad erin g. D ie la as te m eto de w ord in h ie rd ie v erh an de lin g b eh an de l.

A sg ev olg v an d ie in ge wik ke ld he id v an d ie fis ie s-c he mie se w is se lw erk in g w ord h ie rd ie v erh an de lin g g ro te n-d ee ls to t e en a sp ek b ep erk ; n l. d ie w is se lw erk in g tu ss en d ie m in era al o pp erv la kte e n d ie re ag en se in o plo ss in gw at as baie belangrik beskou w ord om nvo lle b egrip va n d ie p ro se s te ve rkry. N ade rh and sa l an de rfa se s va nd ie p ros es o ok b es tud ee r w ord. D ie a sp ekte w aa ro p w erk a lre ed s g ed oe n is , is d ie ko nse ntra sie v an u ra an , d iere ak sie tu ss en g ou t m in era le e n v ets ure , g em en gd e x an th ate -k atio nie se v ers am elin gs s ys te em , e n d ie re ak sietu ss en s ulfie de m in era le e n th io l re ag en se . G ek os e p ro ble me a an ga an de d ie e ers te tw ee a sp ek te , w at g ro ot-lik s h an de l o or o ore en ho pin g flo ta sie v an u ra an e n d ie c he mie se k on tro le v an d ie flo ta sie p ro se s v an fo sk orie t,w ord in b es on de r b eh an de l. n B es pre kin g w ord o ok g ev oer in d ie fy nste b eso nd erh ed e o or die v oo ru itga ng o pd ie g eb ie d v an d ie w is se lw erk in g tu ss en s ulfie de m in era le e n th io l re ag en se w at v an m ee r a lle da ag se b ela ngis .

Ten slotte w ord n rig sn oe r g eg ee vir v erd ere fisie s-c he mie se n av orsin g de ur die N .I.M .

INTRODUCTION

Flotation is acknowledged to be the most importantprocess in use at the present time for the concentrationof minerals. Flotation is used in the processing of thebulk of the world s non-ferrous metals, as well as of agrowing proportion of its iron, non-metallic minerals,and coal. However, until the middle of the 1950 s, theprocess played a less prominent part on the SouthAfrican mineral-extraction scene than it did in othermining countries because hydrometallurgy was applieddirect to the ores in the extraction and concentration ofthe two principal products of the industry and becausesulphide minerals were relatively less im portant. Duringthe past decade and a half, however, several very largeflotation plants, as w ell as num bers of sm aller operations,have been established in the country. The increase inproduction of flotation concentrates is clearly illustratedby the figures in Table 1. A look into the future leaves nodoubt that the importance of flotation to the SouthAfrican mineral-extraction industry will continue togrow.

Before 1960 no long-term research and developmentdevoted to the improvement and understanding of theflotation process were undertaken in South Africa. W ith

*T he N atio nal Institute fo r M etallurgy

the increase in application of the process, the need wasfelt for such an effort to be made, and work in thisdirection was initiated at the Government M etal-lurgical Laboratory, the forerunner of the NationalInstitute for M etallurgy (N.1.M .). The progress of thisprogramme has built up facilities and a body of know -how that should be of use to the South African industryin evaluating the significance of new developments andnew ideas in this specialised field.

The present paper is intended to make the w ork moregenerally known to metallurgists in the country. Itdeals, first, w ith the general concept and planning of this

long-term research, and goes on to describe one aspect-the study of the physicochemical basis of flotation-and to discuss the progress that has been made onseveral of its com ponent projects.

AIM S AND DIRECTIONS IN FLOTATIONRESEARCH

Research and development on the flotation process canbe conducted with varying aims. Short-term work isgenerally directed towards the application of know nprocesses, based on available machines and reagents, to aparticular ore. Long-term investigations could be aimedat the follow ing :

(1) The development of generalized methods of cir-cuit design that w ill enable the performance ofexisting processes in particular situations, or the

JO U RN A L O F T HE S OU TH A FR IC AN IN ST IT UT EO F M IN IN G A ND M ETA LLU RG Y9 6 F EB RU ARY1972


2/14

development of generalised methods for the con-tro l o f flo tatio n p roces ses.

2) The extension of the scope of the process tocom binations of m inerals that are not readilyseparated to be optim ized, to minerals that are notflotable with existing reagents, and to the coarseand fine sizes of particles that are less am enableto the process. W ithin this aim is included thedevelopment of new reagents, new machines,and new tech niq ues.

3) A m ore-com plete understanding of the mech-anism of the different processes taking placeduring flotation. U nderstanding of the physical andmechanical aspects could be of im portance infurthering aim 1) above; the understanding ofthechem ical aspects would assist in realizing aim 2),and is essential to the solution of the importantproblem associated with the fact that nom inallyidentical m inerals can show significantly differentbeh av iou r in flo tatio n.

Three essentially different m ethods can be used in theinvestigation of flotation. Firstly, the empirical trial-and-error method in which previous experience is appliedto new m aterial by way of systematic process testing.For all its unwieldiness, this is the only technique bywhich processes can be developed and proved at thepresent tim e. S econdly, th e c he mic al- en gin ee rin g m eth od .This is a synthetic approach that concentrates on them ixing and transport aspects, and is designed to meetthe first and, to a lesser extent, the second of the aimsset out above. It seeks to characterize flotation as a rateprocess and express characteristics of a plant as a ma-thematical model that relates all the m ost importantfactors governing the operation of the process. A satis-factory model would indicate how the process should beoperated for greatest efficiency. Thirdly, th e p hy sic o-c he m ic al a pp ro ac h. This analytical approach is directedtowards the realization of both the second and thirdaims set out above. It is based on the prem ise that theprim ary, if not the m ost important, factor governing theefficiency and selectivity is the proper conditioning ofthe solid-solution and air-solution interfaces so thatcollisions between them are fruitful and selective. Ittherefore seeks to understand the nature and the m ech-anism of the chem ical and physical reactions takingplace at and betw een the interfaces.

At the N.I.M . all three approaches are applied to thestudy of flotation. The Ore Dressing Division carries outem pirical developm ent in relation to specific ores, theUniversity Research Group in the Departm ent ofChemical Engineering at the University of Natal con-cerns itself with an engineering approach based onm athem atical m odelsl, and the M ineral and ProcessChem istry D ivision studies the physicochem ical aspectsof f lo ta tion .

PHYSICOCHEM ICAL RESEARCH W ORK

Fig. 1 presents a break-down of the flotation process inthe form of a flow diagram, which shows the m ost im -portant stages that take place and the interaction takeover between them . The process is m anifestly a com plexone. The heterogeneous m ixture of solid particles that

make up a typical flotation feed is concentrated andseparated through interaction with chemical reagents,air, and energy. Important interactions take place atthree interfaces-the solid-solution, solid-air, and air-solution interfaces-and in the solution phase.

Before a detailed consideration of some of the variousresearch projects that have been undertaken and areunder way in the M ineral and Process Chem istryDivision, it is necessary to explain the principles thathave governed the general direction of the work.

At the outset it was accepted that it would not bepossible to work imm ediately on all stages of the process c, d, e, and g of Fig. 1) in which important chem icalinteractions take place. It would be preferable to directa concerted effort at one of the stages and to acquire onthat one aspect a sound foundation of knowledge andexperience on which a wider understanding could bebuilt. M ilitating against a more general attack on thesubject was not only the considerable complexity of thephysicochem ical interactions taking place, but alsothe uncertainties and contradictions in the existingknowledge of the subject - a circumstance that made itdifficult to choose a secure point of knowledge from whichto m ount new research program mes.

It was decided that all studies should be conducted onas rigid and quantative a basis as rigidly as possible. Atthe time the work was started, there existed a certaingeneral framework of understanding that explained,probably correctly, how hydrophobicity developed fromthe oriented absorption of collector molecules at themineral surface. However, this general understandingdid not explain the processes taking place in sufficientdetail to be of practical assistance in process design andcontrol, or in the prediction of the behaviour of flotationsystems. It was considered that further progress wouldbe possible only on the basis of a very m uch more quan-titative and exact knowledge of the reactions takingplace. The decision to adopt this type of very thoroughapproach carried with it the acceptance that the pro-gress of the work m ight necessarily be slow and that, in-itially at least, the field of endeavour would have to besevere ly res tr ic ted .

The aspect of the process chosen for initial study wasthe interaction between the mineral surface and reagentsin solution see Fig. stage d)), and the fields in whichwork has been carried out are shown in Fig. 2. Thefollow ing sections describe a few of the specific problem sthat have been undertaken and the conclusions that havebeen reached will be outlined and placed in context withthe general state of knowledge of the particular subject.

THE INTERACTION BETW EEN SULPHIDEM IN ERALS AND THIOL REAGENTS14

To place the work that has been done at N.I.M . in itscontext, it is necessary to sum marize the general state ofknowledge of the interactions that take place betweenthiol reagents and sulphide m inerals.

By the early 1950 s, a clear picture ofthe events takingplace when thiol reagents interact with sulphide m ineralsurfaces appeared to be em erging. It was known thatsulphide surfaces abstracted thiols from aqueoussolution and that in the process anions such as sulphates,

JOURN L O F T HE SO UTH FRIC N INSTITUTE O F M INING ND M ET LLURG Y FE RU RY 1972 97


3/14

(a ) Comminutiono f minera ls

(c ) R ea cti on s o f m in era ls urfa ce w ith w ate r

(d )R ea ctio n o f m in era l

sur face withr eagent s i n solut ion

(f) Coll is ion of part ic lesw ith bubb le s

(g ) Int er ac ti on betweenpart ic le sur face

and bubb le s

Part ic le bubbleaggregate

Froth

Free part ic les

Separation

I nt er ac ti on b etwe enfro th a nd ta ilin gs

FIG UR E 1 - Schem atic flow diagram of flotatio n process

9 8 F EB RU ARY 1 97 2

(b ) Generationo f a ir

(e )R eac tio n o f airsurf ace wi th

re ag en ts in w ate r

F ree bubb le s

Tailings

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF M INING AND M ETALLURGY


4/14

I C O N C E NT R ATI O N O F U R A N IU M I

I IElectrokinetic

study Agglomeration Selectiveof Witwatersrand ores flotation flocculationRef: Ref: 3 Ref: 2

I R E AC T IO N B ET W E EN SA LT M IN E RA LS A N D FAT T Y A C ID S I

I ISurface distribution Adsorption Chemica l con trol

o f o leat es studies of f lo ta tion

R ef: 4 5 6 7 Ref: 4of phoscori te

Ref: 8

I R E A CT IO N B E T W EE NS U LP H ID E M IN E R AL S A N D T H IO L S 1;

I

I T I 1

R ole of oxygenProducts

Homogeneous ActivationR ef: 1 4 1 5 1 6 re ac tio ns o f t hio ls and depression

17 18 22 R ef: 19 20 Ref: 9 16 21 22

~

Galena wi thMetal sulp Xanthate Catalyzed

xanthatehides with stability oxidationxanthate of xanthates

I 1

Metal Sulphur a ssulphide with a product

thiols

M IX ED XA NTH ATE C ATIO NIC C OL LE CT OR S YS TEM S

Adduct formationRef: 9 10 11 12

Flotationproperties

Ref: 3

FIGURE 2 F ield s of p hysic oc he mic al flo tation stu dies

JO UR NAL O F TH E SO UTH AFR ICAN INSTITUTE O F M ININ G AN D M ETALLUR GY FEBRUARY 972


5/14

hydroxides, or carbonates appeared in solution inequivalent quantities. The product of the reaction wasconsidered to be the metal-thiolate, and it was tacitlyassumed that this adsorbed regularly, layer by layer.There was some evidence that the first layer was morestrongly held than the subsequent ones. The reactionwas generally regarded as one that rapidly reached a

reversible state of equilibrium. This conclusion carriesthe important implication that the state of the surface isuniquely related to the state of the solution- con-centrations of different io ns, pH , tem perature, etc.-w ithwhich it is in contact. There was some controversyw hether clean sulphidemineral surfaces w ere floatable,but it was accepted that the surfaces encountered inpractice would, to a greater or lesser extent, be oxidizedand that this would make them hydrophilic. It wasthought that hydrophobic properties were conferred onsuch surfaces by the adsorbed metal xanthates that wereknown to be so oriented that the hydrocarbon portionextended away from the solid and into solution.

During the mid 1950 s facts began to accumulate thatwere not in harmony with this picture. The doubtscrystallized around the work of the Russian school, w hichpublished a series of imported papers at about this time.These showed that the presence of oxygen was essentialif there was to be reaction between sulphide minerals andthiol reagents, and if the minerals were to be renderedflotable. , They also showed that the disposition of theadsorbed reagents at the surface, far from being regular,consisted of a number of isolated multilayer patches. Itwas suggested that the adsorption of reagent was con-trolled by the semiconductor properties of the surface,which were favourably influenced by interaction withoxygen.

During the past ten to fifteen years, a number ofinvestigations have been undertaken in which sophisti-cated modern techniques have been applied to the studyof the chemistry of these interactions. These investi-gations have concentrated on the following problems.

(i) How does oxygen participate in the interactionbetween thiol reagents and sulphide minerals?

(ii) W hat are the products ofreaction?(iii) W hat makes the surface flotable after reaction

w ith th io ls?The work programme of N.I.M . has been directed

primarily at the first and second of these problems.

THE ROLE OF OXYGEN IN THE INTERACTIONBETW EEN XANTHATE AND GALENA15, 16, 17, 18, 22

One of the mechanisms that have been proposed toaccount for the participation of oxygen in the inter-action between xanthates and sulphide minerals involvesthe oxidation of xanthate by oxygen in solution to formdixanthogen, which then adsorbs and reacts at thesurface:

2X--I -t 02 -1 -H2O - --+X2-1 -20H-

I P bS -I- X 2 ---+ I PbS - S urfa ce P ro du ctsA series of simple experiments showed that this

mechanism could be eliminated from consideration. Therate of decomposition of xanthates in pure, alkaline

solutions was found to be the same in the presence andin the absence of dissolved oxygen, and no dixanthogencould be detected among the products of decom position.Reaction w ith oxygen and the formation of dixanthogenwas found to be catalysed by the presence of ions ofseveral transition metals. (See Table HI). However, therate of the catalysed formation of dixanthogen was toolow for it to be of importance in flotation, where allreaction must take place within a few minutes.

It is clear, therefore, that oxygen must play its role bydirect action at the m ineral surface. There are essentiallytwo ways by which this interaction might take place.The first of these involves the oxidation of the surface toform insoluble products that subsequently undergo anexchange with xanthate ions from solution, for example,

I P bS -\-2 02 ---+I PbS04IP bS 04 -1 -2 X- ---+IPbX2.

It has been convincingly demonstrated29, 30thatxanthate can react with heavily oxidized surfaces in

this way, but this does not preclude the second directmechanism - the simultaneous interaction of xanthateand oxygen at the surface - which may be w rittensch em atica lly as fo llo ws:

I P bS -I-O -I-n X- ---+IPb S Surface -I- Solution2 product(s) product(s)The N.I.M . group has devoted a major effort to the

elucidation of this mechanism and has concentrated ongalena and ethyl xanthate as a model system. Thisinvestigation is one of considerable experim entaldifficulty, which stems from two sources: firstly, therequirement that the sulphide surfaces used should befree of oxidation initially and, secondly, the need for theconcentration of oxygen in the reaction system to becontrolled and measured, and for the system to beisolated from contact with the abundant oxygen of thesurrounding atmosphere. Thus, the progress of theinvestigation is conveniently followed through thedevelop ment of experim ental techn iques.

The first stage of the investigation was the develop-ment of a means by which oxidation products could beremoved from the surface of a sample of ground galenaand by which the surface could subsequently be con-tacted with solutions containing either no oxygen orsome oxygen16. The apparatus consisted essentially of acolumn through which solutions flowed upwards, main-taining a bed of ground galena in fluidized condition.Contact with sodium sulphide solution at pH 8, followedby rinsing with water, was found to effectively free thesurface of oxidation: the resulting clean surface didnot react w ith an oxygen-free xanthate solution. Asexpected, when oxygen was introduced to the xanthatesolution, reaction between xanthate and the surface, asindicated by a decrease in concentration of xanthate insolution, was observed. Q uite unexpected, however, w asthe observation that this reaction did not reach com-pletion after a time, as would be expected if it w aslim ited by the build-up of a layer of adsorbed product on

the surface. On the contrary, reaction persisted for aslong - and experiments extended over several days-as oxygen and xanthate were kept in contact with them in era l su rfa ce .

2 F EB RU ARY 1 97 2 JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF M INING AND M ETALLURGY


6/14

The finding that the abstraction of xanthate by galenais a continuing process that does not reach completion orequilibrium is of the greatest importance and hasnecessarily influenced the course of subsequent work. Itimplies that the reaction is not reversible and that theex ten t o f re actio n is no longer a primary parameter thatcan be related simply to such variables as the concen-

trations of xanthate and of oxygen, pH, and tempera-ture. The system can effectively be studied only withreference to the k in etics o f rea ctio n. Therefore, the nextstage of the work emphasized the determination ofrate s o f reac tio n.

The second stage of development was carried out withan apparatus17 shown in Fig. 3 that was similar to thatused previously, in that solutions having passed througha fluid bed of mineral were analysed for their xanthatecontent and then discarded. However, provision wasmade for the concentration of oxygen in the enteringsolution to be precisely controlled. It was found17 thattwo distinct reactions between xanthate, oxygen, and

galena could be distinguished. The first of these resultsin the adsorption of xanthate, and the second gives riseto soluble products. The adsorption reaction predomi-nates in the early stages of contact between oxygen-containing xanthate solutions and clean galena, andfollow s an exponential rate equation:

dqfdt=f3 expo -aq ,where q is the quantity of xanthate adsorbed anda an df3 are constants. At any time, both the rate and the ex-tent of adsorption vary directly with the square root ofthe partial pressure of oxygen in solution. The secondtype of reaction becomes predominant once the rate ofadsorption has been sufficiently reduced by the accum u-

lation of adsorption p rod ucts. T his reaction is responsiblefor the continuing abstraction of xanthate from solution,which had been noted in the first stages of experimen-tation. A fter several hours of contact between the surfaceand the reagent, this reaction proceeds at a constantrate, which also varies with the square root of the partialpressure of oxygen in solution. Although it w as not foundpossible to identify the chemical changes taking place, itwas noted that the reaction was associated with theappearance of a water-soluble product that was tenta-tively identified as the monothiocarbonate ROCOS-.It can be seen that this substance can be obtained fromthe equivalent xanthate, ROCS;, by replacement of

one of the sulphur atoms with oxygen.The design of the second apparatus did not allow

meaningful observations to be made within the first20 minutes of contact between the solutions and themineral. This meant that the results do not apply to thereactions that are of primary importance in flotation.Accordingly, attention w as focused on the developmentof an apparatus that would allow this early period to bestudied, and would also allow the consumption ofoxygen during reaction to be monitored. After a pro-longed period of unsuccessful innovation, the apparatusshown in Fig. 4 was developed31. In use, suitably pre-treated galena is dropped into the upward-flow column,which contains a solution of the reagents that are re-quired for the particular experiment. The column is thensealed in closed circuit with a spectrophotometer cell for

determining the concentration of xanthate, a Clarkelectrode for monitoring the concentration of dissolvedoxygen, and a completely sealed glass centrifugal pump.This system is capable of yielding meaningful resultsw ithin 2 minutes of the commencement of reaction andhas proved itself to be extremely productive. Theresults are very much more reproducible than those

obtained by the previous techniques.W hereas numbers of reliable observations have beenmade by this technique, the work of interpretation isstill in progress. At this time, however, a number ofinteresting findin gs can be m entioned .

1. The two stages of reaction, which are againclearly revealed, are characterized by differentvalues of the stoichiometric ratio - number ofmolecules of xanthate reacted to number ofmolecules of oxygen reacted. The ratio has a valueof two in the first stage of reaction and, dependingon the conditions, varying values of less than onein the second stage.

2. Clear confirmation was obtained that the final,continuing stage of the abstraction of xanthate isdu e to the fo rm ation o f m onothiocarbonate.

3. Under certain conditions of pH and xanthateconcentration, adsorbed xanthate is converted tomonothiocarbonate M TC-) at the surface, whichthen desorbs. It is thought that this effect mighttake place by reaction between adsorbed leadxanthate and lead hydroxide:

Pb OHh+PbX2+20H + 2MTC- +2PbS+~O.4. The new experimental technique allow s the

quantity of monothiocarbonate formed at anytime to be accurately determined and, therefore,

the amount of xanthate adsorbed at the surface tobe estimated. Thus it has been shown that theadsorption of xanthate can be reversed if suitableadjustments are made to the concentration or pHof the solution with which it is in contact. W hereasit is not yet certain how quantitative and com-plete this reversibility is, it seems highly probablethat former indications of the adsorption ofxanthate as irreversible derived from confusionbetween adsorption and the side reaction by whichxanthate is converted to monothiocarbonate-and the latter certainly could not be reversedunder the conditions obtaining in flotation.

5. The total rate of consumption of oxygen bygalena is decreased by the presence of xanthate.

PRODUCTS OF REACTION BETW EEN THIOLSAND M ETAL SULPHIDES19, 20

The investigation of the nature of the products ofreaction was prompted by the controversy that hasgrown up in the literature between those authors whobelieve the product of reaction between xanthates andsulphide minerals to be the metal xanthate, those whobelieve it to be dixanthogen, and those who believe itto be a mixture of the two. In the case of galena, themost convincing piece of evidence in favour of dixan-thogen was provided by the work of Abramov32, whoused infrared spectroscopy to identify the products ex-tracted from the surface of xanthated galena. W hen

JO URN L O F THE SO UTH FRIC N INSTITUTE O F M INING ND M ET LLURG Y FE RU RY 1972 2


7/14

members of the N.I.M. team repeated this work, theywere unable to identify dixanthogen in extracts fromxanthate- treated galena surfaces but, instead, found onlylead xanthate to be present. Further investigation of thedetailed procedure used by Abramov revealed that oneof its stages - the evaporation of the organic extractfrom the m ineral surface - had the effect of convertingany lead xanthate present to dixanthogen and showedAbramov s finding to be an experimental artifact. Itw as therefore confidently concluded that m etal xanthatewas a major product of the reaction between galena andxanthate solutions. Since dixanthogen is more solublethan lead xanthate in the solvents used, and since thereis no reason to believe that it is more strongly held atthe surface than is the lead xanthate, it was also con-cluded that dixanthogen is not a major product. How-ever, it could not be concluded that lead xanthate wasthe only product of reaction because there was no as-surance that all the products formed at the surface wereextracted by the methods used.

The investigation was extended to the determ inationof the products of reaction between som e tw elve differentsulphide minerals and several different xanthate homo-logues. The purpose of this work was to establishwhether the different minerals would form the sametype of product (metal xanthate or dixanthogen) and tosee whether any observed differences could be ex-plained. The results, which are summarized in Table IV,show that, where minerals react to form an extractableproduct, this product tends to be either metal xanthateor dixanthogen. The formation of mixtures of the two is ararity. W hich of the two products is formed correlateswith the potential that the mineral assumes in a solution

of the xanthate. Those minerals that take up a potentialmore positive than the equilibrium potential for thereaction

X2+2e-+2X-form dixanthogen; those that assume a more negativepotential either form the metal xanthate or do not react.

The reactions between the same suite of minerals anddifferent types of thiol flotation reagents are currentlybeing studied. The reagents are m ercaptobenzthiazole,and alkyl dithiophosphates, thiocarbamates, and thio-nocarbamates. Results are being obtained that are inharmony with the behaviour observed in reaction withxanthates. Again it is found that, where reaction takesplace, it is either the metal thiolate or the dithiolate thatis formed; and, again, the formation of the dithiolatetakes place on those minerals that assume a potential in asolution of the reagent that is higher than the equi-librium potential for the appropriate reduction reaction.

Although this work has provided an overall view of thebehaviour of a number of related reagents with a wholerange of metal sulphide minerals, and has thrown somelight on a controversial subject, it by no means provides acomplete pr an adequate understanding of the productsthat are formed at the mineral surface. There is evidencethat some minerals, at least, form a tightly-bound thiol-

containing product that is not amenable to extractionby common organic solvents and hence cannot be de-tected by; the techniques used at the N.I.M . For galenaand xanthate, this product is thought to be a species in

2 2 F EB RU ARY 1 972

which a surface lead atom is attached to one xanthate,instead of to the normal two xanthates. Furthermore,there is reason to suspect that elementary sulphur is al-ways formed as a primary product of the reaction be-tween oxygen and sulphide surfaces in the presence ofthiol reagents and that this sulphur may play an im-portant part in rendering the sulphide mineral surfacesfloatable.

THE CAUSE OF FLOATIBILITY

Both the metal xanthate and dixanthogen have beensuggested as the entities that confer hydrophobicproperties on sulphide mineral surfaces after they havebeen treated with xanthate. The finding that sulphideminerals react with xanthates to produce, in general,either xanthate or dixanthogen makes it extremelyunlikely that either of these can be the sole cause ofhydrophobicity at sulphide surfaces after treatm ent withxanthates. Indeed, it is now not even certain that eitherof them need necessarily be the primary hydrophobic

entity. In an attempt to explain their observations onthe role of oxygen, the workers at N.I.M. have been ledto postulate reactions in which elemental sulphur isalways one of the products - whether the reaction isadsorption or the formation of monothiocarbonate. Thequestion now arises whether sulphur can be either solelyor partly responsible for hydrophobicity. W hereas thesuggestion is by no means a new one and some objectionsto it have been proposed, it is thought to warrant re-examination.

AGGLOM ERATION FLOTATION OF URANIUM -BEARING ORES2. 3

For some time, interest has been shown by the miningindustry in the development of processes for the con-centration of uranium from minerals present in W it-watersrand ores. That none of the processes thus farproposed has been adopted might be ascribed to theirfailure to recover the uranium values that are associatedwith the finer particle sizes. M uch of the uranium occursin the slime fractions as is shown in Table n, whichgives the analyses of the various size fractions from asample of ore from W estern Deep Levels, and whichshows that 23 per cent of the uranium is associated withthe fraction smaller than 12/Lm. The uranium in the oresis usually associated with the phyllosilicate fraction(which lim its the extent of beneficiation that is possible),and the concentration problem therefore resolves itselfinto the separation of the phyllosilicates from the quartzgangue.

It was decided to try a new approach that would takeinto account the special requirements of the slimefractions. This approach was based on the principle thatthe fine particles m ust be induced to associate selectivelyinto larger aggregates as a precursor to eventual sepa-ration. Two separation procedures can be considered,viz., settling and flotation. W hen settling is used, theoverall process is called selective coagulation or

flocculation , and the agglomerated particles settlefaster than those that remain dispersed. W hen flotationis used, the process becom es agglom eration flotation andhere agglomeration is induced by the presence of finely

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND M ETALLURGY


8/14

9 5 9 5 5 96

Pyrite 2 7 75 3 5 7 2 9 2 5 D 8

Fluorspar:

Acid - - -Ceramic - - -

Copper* 3 2 73 4 8 89 3 6 3 3 3

Apatite 26 213 64 310 1 34 4 69

Antimony 13 6 00 23200 20200

Platinumt 5 2 2 5

TotalI

1 80 8 61 613695 841 010

1965 1970

428 29 4 607 851

4835 71 1994828 5257

201 66 6 464 673

501 06 8 9 29 6 71

20600 28 800

50000 100 000

1 211 291 2207451

80*

xxX X

X

90

70

~

~0:

60

50

0

300 20 0 60 80 100 120S I LI C ON , p .p . m. 0

16 0 180 200

Fig

TABLE IPRODUCTION OF FLOTATION CONCENTRATES IN THE REPUBLIC OF SOUTH AFRICA. M ETRIC TONS

*Calculated from data for metal by assuming 30 per cent contained in flotation concentrate.t A pproxim ate figures.

JO URN L O F THE SO UTH FRIC N INSTITUTE O F M INING ND M ET LLURG Y FE R U RY 9 7 2 3


9/14

0 F F /R E C O V E R Y. F F I G R A D EA 5 T / R E C O VE R Y. G R A D E

. .. . . A 0

A 0~ 0 1 .0

'T- .: - .:

~ff i~ is

< > UIt >

W

0 3 2 7 0 0 41.6 23.46 62 2 37.0 2080 85

< 41.6 >32.3 21.82 323 18.0 1380 85

23.1 16.03 326 13.0 1212 89

15.7 10.54 308 8.0 832 88

< 15.7 > 12.0 4.97 205 3.0 419 87

< 12.0 23.18 349I

21.0 717 97I

0+ 90 0+90

GIF F / R. F F /G

2 ' Is o

1 1 9 1 '0p H V A L U E

.\

3 6

1

8 0 ~- .: . : .W 0 -C 0 U0 WIL 3 2 7 0

0

.~ . 0'0

< :>/

2 8 + 6 0

2 (.5 0

2 6 8U N I1 0L , p .p .m .

1 0 1 2

F ig 2

T A B L E IIEFFECT O F PA RTICLE SIZE O N THE O IL-A GG LO MERATIO N FLO TATIO N O F A SA MPLE OF O RE FRO M W ESTERN D EEP LEV ELS

2 0 4 F E B R U AR Y1972 J O U R N A L O F T H E S O U T H A F R IC A N IN S T IT U T E O F M IN IN G A N D M E


10/14

E L U T R T O R

FO R

P RE PARATION

O F S AM PL E

SPECTROPNOTOMETER

CELL

REACTION

COLUMN

CLARK TYPE

O X YG EN E LE CT RO DE

GLASS PUMP

Fig 3

Nz N2 102 N2

XANTHATE

SOLUTION

R ECO RD ER I

A BS O RB ANCE

VS

WAVELENGTH

~

R E C O R O E R n

A B S O R B A N C E

VS

TIME

~

G LA SS T UB IN G

TEFLO N TUBING 5mm BO RE

Fig 4

divided oil droplets. The resulting agglomerates arestrong enough to survive the severe mechanical stressesof flotation and massive enough to attach to air bubbles.

The success of both the above approaches depends onthe existence of reagents that produce aggregation ofonly one of the mineral fractions present, i.e., eitherphyllosilicates or quartz. It was considered that the

choice of suitable reagents could be simplified if morewere known about the structure of the interface betweenthe mineral surface and its aqueous environment- inparticular the sign and magnitude of the charge at theinterface. A ccordingly, electrokinetic investigations on auranium -rich fraction and a low -uranium (mainly quartz)fraction from a typical W itwatersrand ore were carriedout. The aims of this investigation were as follows:

(a) To see whether exploitable differences in the zetapotentials of the two fractions could be induced byvariation in the electrolyte concentrations. Forexample, if conditions could be found under whichthe potential of one species was negative and the

potential of the other positive, then a negativelycharged reagent might be expected to adsorbselectively to the positive surfaces and eitherselectively coagulate them or render them selec-t ively f loatable .

(b) To detect whether chemisorption of reagents toone or other, or both, of the two types of mineralta kes p lace.

It was found that most of the range of electrolytes andother reagents studied brought about very similarchanges in the potentials of the uranium-rich and theuranium-poor fractions. W ith sodium oleate, however,the decrease in zeta potential was significantly greaterfor the uranium-rich fraction. Since this might indicate agreater affinity of oleate for the uranium-containingminerals than for the quartz, attempts were made to usethis reagent in laboratory tests on selective coagulationan d ag glo meratio n flo tatio n.

The agglomeration flotation tests gave promisingresults. The method involved the addition to the pulp ofan emulsion comprising a non-polar hydrocarbon(kerosene, 25-30 lb/ton), oleic acid, and a non-ionicem ulsifying agent. A ddition of sulphuric acid in am ountsof 0.4 to 0.8 lb/ton was also necessary, the usual effectbeing an increase in grade and a decrease in recovery withincreasing additions. The amounts of oleic acid added inthe tests were high (varying between ca 3 and 7 lb/ton),but no attempt was made to optimize the process andadditions as low as 1.5 lb/ton may well prove to beadequate. The emulsifying agent used was an alkylaryl polyether alcohol, Triton X.lOO, and was addedin amounts of ca 0.75 lb/ton.

An important difference from the techniques of con-ventional flotation was the requirement of long con-ditioning times at high energy inputs and of high pulpdensities in the conditioning stage. The use of highconditioning energies in agglomerate flotation is es-sential and may arise from the need for the removal ofcollector that becomes loosely attached to the gangueduring the early stages of conditioning23. The manner inwhich high conditioning energies achieve this removal isto unnderstood, but it may be brought about by the

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF M INING AND METALLURGY FEBRUARY1972 5


11/14

attrition of the more weakly held collector at the surfaceof the gangue m inerals. In the tests, conditioning carriedout at impeller speeds of 1 800 rev/min and times of 30minutes was found to give good results. Pulp densities inthe conditioning stage varied from 30 to 50 per centsolids, although the pulp was diluted to 20 per centsolids before aeration and collection.

Recoveries of high-grade ore (400 p.p.m . ofU3Os) wereusually between 85 and 92 per cent, with grades ofapproximately 1200 p.p.m . of U3Os in concentratescontaining about 30 to 35 per cent by mass of the feed.Low-grade ores (ca lOO p.p.m. of U3Os) gave recoveriesof about 75 to 80 per cent from concentrates containing25 to 35 per cent of the feed, with grades varying be-tween 200 and 400 p.p.m . of U3Os.

Of considerable im portance is the fact that significantrecovery of uranium from the fines was achieved. Thisis shown in Table II, where the response of differentsize fractions to the flotation process is shown.

Laboratory-scale tests on selective coagulation of theuranium-containing species were not successful inbringing about differential coagulation2. H ow ever, testsusing high energies and pulp densities in the conditioningstage comparable with those used in agglomerationflotation did achieve some concentration, although thiswas not of sufficient magnitude to generate optimismabout the future of the process in the treatment ofW itwatersrand conglom erate ore.

FLOTATION OF PHOSCORITE ORESs

In conjunction with the research department of thePhosphate Development Corporation, a search was made

for correlations between the concentration of flotationreagent in the solution phase of phoscorite pulp and themetallurgical performance of the flotation. This searchwas prompted by the possibility that the compositionof the solution phase of the flotation pulp might governthe concentrations of the collector adsorbed at thedifferent m ineral surfaces, on w hich differential flotationof the minerals ultimately depends. Hence, the solutionconcentrations might provide an important factor forthe control and optimization of the float. Such corre-lations for the flotation of galena by xanthate have beendemonstrated by Bushell and used in the control of alarge concentrator24. The initiation of the work was

prompted by the availability, as a result of more funda-mental studies of the Mineral and Process ChemistryDivision, of suitable methods for the determination ofthe concentrations of most of the principal reagents usedin the flotation of phoscorite.

In the flotation of phoscorite, four reagents are addedto the pulp in the conditioning stage, viz., caustic soda,sodium silicate (Na20.2SiO2), Unitol DSR (fatty acid,collector) and Berol EM U (polyglycol ether, frother anddepressant). In this work, the levels of concentration insolution of three of these, viz., caustic soda (pH),sodium silicate, and Unitol DSR, were determined.The concentration of the remaining additive, Berol

EMU, was not measured, mainly because no analyticalprocedure was readily available. The concentrations ofcalcium and magnesium were also determ ined, althoughprevious work at Foskor25 suggested that correlations

with these ions would be unlikely.The investigation was carried out on the I-ton per

hour pilot plant at Foskor, and the procedure wasbriefly as follows. Sam ples of pulp were withdrawn fromthe plant at various points in the circuit, as shown in theschematic diagram of the pilot plant in Fig. 5, and atregular intervals of one hour to 90 minutes. After the

pH value had been measured, the samples were centri-fuged and the supernatant solutions were analyzed forUnitol DSR, silicate, calcium, and magnesium. Thephosphate content of the solids was also determined.The levels of concentration of the various species insolution were established when the pilot plant wasoperating under standard conditions, Le., conditionsof reagent addition that were required for maximumrecovery at merchant-grade (36.5 per cent P2O5) Aseries of runs was then performed in which the additionof the three relevant reagents (caustic soda, sodiumsilicate, and Unitol DSR) was varied. This was toestablish the effect of different concentrations on gradea nd re co ve ry.

W hen the plant was running under standard con-ditions , a pronounced variation in concentration of thereagents was obtained, although the metallurgy of theplant was normal. Similar behaviour has been observedby two other groups of workers. W oodcock andJones26, 27, in m easuring up to 19 chemical param eters inflotation pulps in six Australian lead-zinc plants, ob-served that vastly different chemical environmentsexisted in the six plants, with a great variability at manypoints in each circuit. Although this finding suggeststhat chemical control of these plants is unlikely to beeffective, this deduction cannot be confirmed becauseW oodcock and Jones do not quote metallurgical per-formance data. However, correlations between some ofthe variables measured, viz., pH, redox potential, andoxygen concentration in solution, were obtained.Plaksin et al28 made measurements of the residualxanthate concentrations at various points in a flotationplant processing copper-lead-zinc-pyrite ore. T hey, too,found that the residual concentration of xanthate variedwithin wide limits, although the performance of theplant remained satisfactory. Decrease in the xanthatedosage rate likewise did not adversely affect plant per-formance, although it lowered the residual xanthate

concentration.Although this result was disappointing from a controlpoint of view, some interesting figures concerning theuptake of the reagents during the conditioning stageswere obta ined .

Comparison of the concentration of Unitol DSR insolution in the flotation feed (8.1 p.p.m .) with the theo-retical initial concentration (105 p.p.m .) indicated that alarge proportion had apparently been removed fromsolution in the 2-minute period after its addition.However, from a previous study of the absorption ofsodium oleate on pure apatite4, it was shown that theamount of Unitol DSR that could be absorbed on the

apatite fraction of the phoscorite was only ca 4 per centof the amount that appeared to have been removedfrom solution. A hypothesis that could be put forwardto explain this anomaly is that there was not time for

2 6 F EB RU ARY 1972 JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF M INING AND M ETALLURGY


12/14

co A DD ITIO N OF NDU AL

WATERto tH E F RO TH L AU ND ER S

UNnOL S B E RO L E N U

S A NP U NG P D ON T S

F r- F LO TA O N F EE D

-ROUGNER CONCENTRATES C -S C AV EN G E R C O H C EH T R ATE

- SC AV EN GE R . .U H GSC C - C lE A NE R C O , EN T RAT Eet -ClEANER TA GSRCC -R E C LE A N E R C O , E NT R AT EReT -REClEAHER TA HGSI W - IN OU ST A L WATE R

XC - F I RS T C O H O IT I O RAN

- B OREREAGENT ADDITIOH

ST

FlHALtA INGS

CONCENTRA

Fig 6

saponification to be completed before sampling took

place and that the droplets of unsaponified materialw ere separated from the solution during centrifugation.The corresponding decrease in soluble silicate con-

centration (from 150 to 108 p.p.m. of silicon) could beexplained by its precipitation as calcium and m agnesiumsilicate, especially since the levels of these two ions insolution decreased sharply when sodium silicate wasadded to the circuit.

Variation in the addition of caustic soda to the planthad the expected effect on pH, but these changes hadlittle effect on the flotation performance (Fig. 6). Thelevels of Unitol DSR and silicate in solution increasedwith increase in pH, whereas magnesium decreased,

persumably because of formation of the hydroxide.However, the results did suggest that, within the range

TABLE IIIRATES OF DECOM POSITION OF SOLUTIONS OF PO TASSIUM ETH YL

XANTHA TE AT pH 8

AtmosphereH alf- life o f

decompositionh

M etal ion added

N.

O.

52 0

52 0

o. Zn + 52 0

Mn + 50 0

45 0a +

AP + 43 5

Pb + 43 5

Nio+ 37 3

Hg + 35 2

33 5U +

Sn + 31 0

31 0O2+

Fe3+ 23 6

52 0. Any of the above

ReT

of values tested, the concentration of caustic soda is nota )o ntro llin g v aria ble .

The effect of silicate concentration on grade and re-covery was more pronounced and is shown in Fig. 6.These results are unusual in that variations in gradeand recovery are parallel.

The effect of Unitol DSR concentration on perfor-mance is also shown in Fig. 6. No real effect on eithergrade or recovery is apparent beyond a solution concen-tration of ca 8 p.p.m., and this surprising observationmay indicate that the species of Unitol that is importantin flotation is not measured in solution. This would bethe case if the unsaponified material-whose existencewas suggested earlier-was to govern the hydrophobicityof the particles. The failure of different additions ofUnitol to affect the pH value of the pulp provides someconfirm ation of this m echanism .

More detailed analysis of the results by multilinear,stepwise regression analysis did not produce anymeaningful correlations with flotation performance,although silicate was shown to be a significant variable.That silicate concentration is not an important variablefor control purposes, however, is apparent in Fig. 7,w here recovery is plotted against silicate concentrationin the feed. The figure includes the results of the stand-ard runs as well as those of the runs in which the ad-dition rates of both silicate and the other reagents werevaried. The best computed straight line is shown.However, if the distribution of points is such that alarge number fall within a narrow range of recoveriesand concentrations (as occurs here), such a line is mis-leading because the least squares treatment gives undue

weight to the few data that lie outside these ranges.These extreme conditions, however, are artificial andwould not be likely in practice. Therefore, without amore complete understanding of the part played bysilicate, this variable could not be used as a control ofth e p ro ce ss .

It was established, therefore, that solution-concen-tration variables cannot provide a useful basis for thecontrol of the Foskor process. Such a conclusion, how-ever, raises some important considerations. It suggeststhat the interaction between calcite and apatite and thefatty acid collector, the magnitude of which should beindicated by the residual concentrations in solution, isnot an important factor in this process. Alternatively, ifthe essential relationship between adsorption of col-lector and floatability is accepted, it suggests that thisadsorption is not relately uniquely to the concentrationof reagent in solution. From this particular study,neither of these alternatives can be invalidated, al-though there are strong indications that the amount ofUnitol DSR at the conditioned apatite surfaces was notrelated to the concentration in solution because of thepresence on the m ineral surface of unsaponified m aterial.However, further work is necessary to substantiate thispicture.

PROSPECT

In the future, the work of the N.I.M. group on thep hy sic oc he mic al a sp ec ts o f flo ta tio n w ill b e b ro ad en ed to

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF M INING AND M ETALLURGY FEBRUARY1972 7


13/14

TABLE IV

PRODUCTS EXTRACTED FROM SULPHIDE SURFACES

M X=m etal xanthate X.=dixanthogen N PI n o p ositiv e id en tific atio n

Mineral Normal Alkyl Xanthates

M ethyl Ethyl Propyl Butyl Amyl Hexyl

Orpiment NPI NPI NPI NPI NPI MX

Realgar NPI NPI NPI NPI NPI MX

Sphalerite NPI NPI NPI NPI NPI MX

Stibnite NPI NPI NPI NPI NPI MX

Antimonite NPI MX MX MX M X MX

Cinnabar NPI NPI NPI NPI NPI MX

Galena M X M X M X MX M X MX

Bornite NPI NPI M X MX M X MX

Chalcocite NPI NPI M X MX M X MX

Covellite X. X. X.+MX X.+M X X.+MX X.+MX

Chalcopyrite X. X. X. X. X. X.

Pyrite X. X. X. X. X. X.

Pyrrhotite NPI X. X. X. X. X.

Arsenopyrite X. X. X. X . X. X.

Alabandite NPI X. X. X . X. X.

Molybdenite X.+? X.+? X.+? X.+? X.+? X.+?

cover aspects of the process other than the interactionbetween reagents and mineral surfaces. Although theknowledge gained of this aspect confirms the group sinitial belief that it is one of the keys to the understand-ing and the more systematic development and operationof flotation processes, it is certainly not the only one.

Least understood and, according to present thinkingat N.I.M., most vital of these key areas is the link be-tween the chemical reactions that take place at asurface and the flotation properties that result. To lendthis opinion point, reference has only to be made to thestudy described here of the products of reaction betweenvarious sulphide minerals and thiol reagents. Although

the products of reaction are now known, it is not possibleto gauge the significance for flotation of the observeddifferences. To provide this link will be one of the goalsof the expanded programme of the group. Being plannedat present is an investigation of the interaction of airbubbles with hydrophobic surfaces- the ultimate aimof which is to provide an explanation of the way inwhich the attachment and detachment of bubblesdepend on the chemical composition of the surface. Thiswork will naturally include studies of the mechanism bywhich non-polar additives operate in processes such asa gg lo me ra tio n flo ta tio n.

ACKNOWLEDGEMENTSThis paper is published by permission ofthe Director of

the N ational Institute for M etallurgy.

REFERENCES1. KING, R. P., J. S. Air. Inst. Min. Metall. To b e pub lished.2. M ACKENZIE, J. M . W . and LOVELL, V. M., Double layer

studies of W itwatersrand ores with particular reference tothe concentration or uranium by selective flocculation.Johannesburg, National Institute for M etallurgy. N.I.M.R eport N o. 1 41 3, 1 97 2

3. M ACKENZIE, J. M . W ., The agglomeration of flotation ofuranium-bearing minerals from W itwatersrand conglome-rate ores. Johannesburg, N ational Institute for M etallurgy,R epo rt N o. 1386, 6th N ovem ber, 1971.

4. LOVELL, V. M. and FINKELSTEIN, N. P., The study ofmineral surface properties by the use of gas adsorptionmeasurements. Johannesburg, National Institute for M etal-lurgy, Report No. 1 40 2., 1 97 2

5. LOVELL, V. M . and FINKELSTEIN, N. P., The influence of

preadsorbed oleate ions on the adsorption of water vapourby insoluble ionic crystals. Johannesburg, National In-s titu te f or M e ta llu rg y, R eport N o. 5 04 , 1 96 9.

6. HALL, P. G., LOVELL, V. M . and FINKELSTEIN, N. P.,Adsorption of water vapour on ionic solids containing pre-adsorbed sodium oleate. Part I. Calcium fluoride.Trans.F ara d. S oc . 66, 152 0-152 9, 19 70.

7. Idem., Part n. Calcium carbonate. Trans. Farad. Soc. 66 ,2 62 9- 26 35 , 1 97 0.

8. LOVELL, V. M., et 1101.,The flotation of phoscorite ores.P ossible correlation s b etw een chem ical v ariab les in solu tionand metallurgical performance. Johannesburg, NationalIn stitu te fo r M eta llu rg y, Report No. 1 20 4, 1 97 1.

9. F IN KE LS TE IN , N . P., Influence of alky l-trim ethyl am mon iu mhalides on the stability of potassium ethyl xanthate ina qu eo us s olu tio n. J. A ppl. O hem . L ond. 19, 73-76 , 196 9.

10. GOOLD, L. A. and FINKELSTEIN, N. P., An infrared in-vestigation of the potassium alkyl xanthate and alkyltrimethyl ammonium bromide mixed-collector system.Johannesburg, National Institute for M etallurgy, ReporNo. 4 98 , 1 96 9.

11. FINKELSTEIN, N. P., The bathochromic shift of the 381 nm

208 FEBRUARY 1972 JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY


14/14

absorption peak in aqueous solutions of potassium ethylxanthate containing dodecyltrimethyl am monium brom ide.Jo hann esb urg N ation al In stitu te fo r M etallu rg y. Repor~No. 437 1968.

12. FIN KE LST EIN N . P. and FA ZA KE RLY G . V. The detection ofthe formation of a bulk adduct between potassium ethylxanthate and dodecyl trimethyl ammonium bromide inaqueous solution. Johannesburg National Institute forMetallurgy. R epo rt N o. 340 1968.

13. M AHN E E. J. L ovELL V. M . and FINKELSTEIN N. P.S mall-scale flota tion tech niqu es. Joh an nesb urg N ation alI ns titu te f or M eta llu rg y. Report N o. 1026 1970.

1 4. G RA N VIL LE A . A LL IS ON S . A . a nd F IN KE LS TE IN N . P. Areview of reactions in the flotation system galena-xanthate-o xyg en . Jo han nesbu rg N atio nal In stitu te for M etallurgyR eport N o. 1296 3 rd September 1971 .

15. G AU DIN A . M . and FIN KE LST EIN N . P. I nteractions in thesystem galena-potassium ethyl xanthate-oxygen. NatureLond. 207 389-391 1965.

16. LEVIN J. and FIN KELSTEIN N. P. An apparatus for use inthe study of interactions between sulphide mineralss ulp hy dr yl r ea ge nts an d o xy ge n. J oh an ne sb urg N atio na lI ns titu te f or M e ta llu rg y. Report N o. 4 5 2 8th J un e 1 96 6.

17. FINKELSTEIN N. P. Quantitative aspects of the role ofoxygen in the interaction between xanthate and galena.S cp n. S ci. 5 227-256 1970.

18. GRANVILLE A. A stirred-vessel batch apparatus for the

study of galena-xanthate-oxygen interactions. Johannes-burg N ational Institute for M etallurgy. Report No. 93231st Augus t 1970 .

19. ALLISON S. A . and FINKELSTEIN N. P. A study of theproducts of reaction between galena and aqueous xanthatesolutions. Trans. Inst. M in. M etall. Sect C 80 C.235 -2391971.

2 0. A LL lS ON S . A . G OO LD L . A . Nrco L M . J. and G RA NV lL LEA. The products of reaction between sulphide m inerals andaqueous xanthate solution. To be published.

21. FIN KEL STElN N . P. K inetic and therm odynam ic aspects ofthe interaction between potassium ethyl xanthate andoxygen in aqueous solution. Trans. Inst. M in. M etall. 76

C.51.C.59 1967.22. ALLlSON S. A. and FINKELSTEIN N. P. S tudies of the

stability of xanthates and dixanthogen in aqueous solutionand the chemistry of iron xanthates and cobalt xanthates.Joh an nesb urg N ation al In stitute for M eta llurgy . ReportNo. 1125 1971.

2 3. R U NN O LI NN A U . R I N NE R . a nd K u RR A N EN S . A gg lo m er -ation flotation of ilm enite at Otanmaki. Intern. M in. Proc.Congo 1960. London Institution of M ining and M etallurgy447-475 1960.

24. BusHELL C. H. G . CHA LM ERS H . J. and GO WAN S W. K.Residual xanthate m onitoring and control.C on tin uo us P rocessing and Process Control. Proc. of Symp. PhiladelphiaDec. 5-8 1966. Ed. T. R. Ingraham. New York. GordonB re ac h. 4 1- 53 1 96 8.

25 . R ET IE F I. M . Io nic po pu latio n o f flota tion p ulp s. Jo han ne s-b urg P ho sp ha te D ev elo pm en t C or po ra tio n. R eport No. 53Jan. 1970.

26 . W O OD CO CK J. T. an d JO NE S M . H . C hem ical env iro nm en tin Australian lead-zinc flotation plant pulps: I.pH redoxp ote ntia ls a nd o xy ge n c on ce ntr atio ns . P roc . A ustra las. Inst.M in. M eta ll. 235 45 1970.

27 . W O OD CO CK J. T. e t a l. 11. C ollecto r re sid uals m etals insolution and other param eters. 61Ibid.

28. PLAKSIN I. N. OKOLOVlTCH A. M . and Popov R. L.Residual concentration of xanthate in the industrial con-c en tra te p la nts. Akad. Nauk S.S.S.R. Izv. Otdelenie T ekhn.

Nauk. N o. 1 204-208 1963. N IM Translation 246.29. M ELLGREN 0. Heat of adsorption and surface reactions of

potassium ethyl xanthate on galena.Trans. Am. Inst. Min.Engrs. 235 46-60 1966.

30. M ELLGREN O. and RAO S. R. Heat of adsorption andsurface reactions of potassium diethyldithiocarbamate ongalena. Trans. Inst. M in. M etall. 77 C.65-C.71 1968.

31. H AR RIS P. J. and FIN KE LST EIN N . P. JohannesburgN ational Institute for M etallurgy. To be published.

32 . A BR AM OV A . A . M eth od fo r th e q ua ntitative de term ina tionof the sorption form of solvents on mineral surfaces.SovietM in in g S cie nc e. 4 384-389 1968.

JO URN L OF THE SO UTH FRIC N INSTITUTE O F M IN ING ND M ET LLURG Y FE RU RY 9 72 2 9

fundamental studies of the flotation process:

Documents