scopy for Chemical Analysis (ESCA) was used to peiform the

suiface analyses.

S. Acar and P. Somasundaran

Abstract-Results of selective flocculation of naturalores containin,i\' two or more minerals and of synthetic mineralmixtures do not usually agree with what is expected on thebasis of the results from single mineral tests (Acar andSomasundaran. 1985). This is mainly due to the ionic speciesreleased by the minerals. which act differently under differentpH values. The objective of this investigation was to study 1 )the surface chemical composition of selected sulfide miner-als-natural chalcopyrite and pentlandite and synthetic cov-ellite. millerite. and pyrrhotite-and 2) the changes broughtabout by the dissolved mineral species. Electron Spectro-

Introduction

S. Ace" member SME, is currently a metallurgist with NewmontMetallurgical Services, Salt Lake City, UT. He was with the HenryKrumb School of Mines, Columbia University, New York, NY when thestudy reported in this paper was conducted. P. Somasundaran,member SME, is chairman at the Henry Krumb School of Mines.M&MP paper 89-649. Manuscript Aug. 17, 1989. Discussion of thispaper must be submitted, in duplicate, prior to Aug. 30, 1990.

Electron Spectroscopy for Chemical Analysis (ESCA) is asurface analysis technique that provides a depth of analysisranging from 100 to 500 A. The method involves irradiation ofthe solid sample in high vacuum with monoenergetic soft X-rays (Mg Ka or AI Ka X-rays) and sorting the emittedelectrons by energy (Perkin Elmer). The spectrum obtained isa plot of the number of emitted electrons per energy intervalversus their binding energy. Each element (with the excep-tions of hydrogen and helium) has a characteristic set ofelectron binding energies that can be used to identify theelement. Moreover, ESCA can often differentiate betweenchemical environments or valence states of any given ele-ment.

ESCA has been extensively used in the minerals and metals

MINERALS & METAlLUAGK:Al PROCE~94 MAY 1990

the state of oxidation of the sulfide mineral surface (Cliffordand Miller, 1974; Majima, 1970; Plaskin, 1958; Yamaski andUsui, 1965). ESCA has been used to characterize the surfacesof galena, sphalerite, chalcopyrite, chalcocite, covellite, bor-nite, pyrite, pyrrhotite, molybdenite, and millerite (Clifford,Purdy and Miller, 1975). Surface analyses were made aftercomminution in air and comminution in a nitrogen atmos-phere, and after adsorbtion of xanthate or dithiophosphatecollector. Analyses were also made after treating with modi-fiers such as N~Cr2O7' S02' starch, NaCN, and CuS04.

Most of the sulfide samples ground in air showed signs ofelemental sulfur or sulfate formation. The adsorption of KAXand DTP was observed but with difficulty by following thesimultaneous increase in the C, S, and P molar amounts. It waspossible also to follow the action of CuS04, N~Cr2O7' S02'and starch but not NaCN in a Pb-Zn-Cu flotation separation.

In the present study, ESCA was used to investigate theeffect of dissolved species from natural and synthetic sulfideminerals on the surface chemical composition in selectiveflocculation systems.

Experimental materials and methods

Materials

industry. Manocha and Park (1977) used ESCA to character-ize the nature of chemical species on galena (PbS) surfacesexposed to air and aqueous environments and concluded thatlead sulfate is the major product of oxidation of PbS in air andaqueous environments. The presence of small amounts ofelemental sulfur in the early stages of oxidation suggested thatthe initial oxidation step differs from the later major oxidation

processes.ESCA has also been used to determine the bonding nature

of adsorbed cations.Koppelman and Dillard (1978) used X-ray photoelectron

spectroscopy (XPS) to study the bonding nature of metal ionsadsorbed from aqueous solutions by the marine clay mineralskaolinite, illite, and chlorite. Binding energies for the latticeelements Si, AI, 0, Mg, K, and Ca were found to be in goodagreement with the published values. However, significantdifferences in bonding energies (1.geV) for the Fe 2P3/) levelfor iron in chlorite and nontronite were observed. lOts wasattributed to the two different oxidation states of iron, Fe++ inchlorite and Fe+++ in nontronite. Fe+++ adsorbed onto kaolinitehad an Fe 2P3/2 level binding energy 1.2 e V lower than latticeFe+++. This lowering of binding energy was attributed to thenegative potential of the electrical double layer at the mineralsurface when the cation is adsorbed.

Flotation responses of sulfide minerals and the amount ofcollector adsorption have been shown to be dependent upon Minerals: Natural chalcopyrite was purchased from Ward's

MAY 1QQO 9SMINERAlS & METAllURGK:AL PROCESSING

Corp. as minus 2.36-mm (minus 8-mesh) lumps. It was alsoground in an agate mortar under nitrogen atmosphere andstored in a Teflon bottle. The surface area of this sample wasfound to be 0.75 m2/g (250 sq ft per oz).

Inorganic reagents: Fisher certified CuCl,2H2O,NiClr6H20, and FeClr4H20 were used in the experiments asthe source of the desired metallic cations. Fisher certified HCland NaOH (IN, O.lN, and O.OIN) were used to adjust the pH.Amend Drug and Chemical Co. reagent grade NaCl, which is99.96% pure, was used to make 3.10-2 kmol/m3 salt solutionfor all the experiments.

Water: Triple-distilled water with specific conductivity ofabout 10-6 mmhos/cm was collected from a quartz still.

Methods

Experimental procedure for the surface compositionalstudies began with pretreatment of the mineral surfaces byconditioning them in various inorganic solutions and thenanalyzing the treated and untreated surfaces using ESCA.

1.5-g (0.05-oz) mineral samples were conditioned in Tef-Ion bottles using a wrist action shaker in 50 ml (1.7 fl oz) ofthe required solution at the desired pH for three hours. Themineral samples were rinsed several times with triple-distilledwater and separated from the solution by filtration followed

Natural Sciences Establishment Inc. After crushing and con-centrating using an electrostatic separator, the chalcopyritewas ground in a stainless steel ball mill and wet screened atminus 20 ~m (625 mesh). It was dried and stored in a Teflonbottle under nitrogen atmosphere to prevent further oxidation.Surface area of the sample a.., measured by the BET nitrogengas adsorption method was 1.22 m2fg (406 sq ft per oz).

Flotation concentrate of pentlandite was obtained fromInco Limited and was cleaned to remove the flotation reagent.The cleaned concentrate was screened to minus 20 ~m andleached with dilute HCI. Leached material was repeatedlywashed with distilled and triple distilled water until natural pHof triple distilled water was obtained. The sample was thenfreeze dried and stored in a Teflon bottle under nitrogenatmosphere. The surface are of this sample was determined tobe 1.43 m2fg (476 sq ft per oz).

Synthetic covellite of surface area 2.74 m2fg (912 sq ft peroz) purchased from Morton Thiokollnc., Alfa Products wasused as received.

Synthetic millerite, also purchased from Morton ThiokolInc., was ground under nitrogen atmosphere in an agatemortar and stored in a Teflon bottle. The surface area of thesample was determined to be 0.75 m2fg (250 sq ft per oz).Synthetic millerite was stated to be 99.9% pure, and its X-raydiffraction pattern matched ASTM pattern number 2-1280 fornickel-2-sulfide.

Synthetic pyrrhotite was obtained from Fisher Scientific

U)20a:..-u~-J~

LL.0

a:~CD~::)

.2

(/)z0(r.

Ul1J-Jl1J

LL0

(r.l1Jm~::>z

606.6806.61006.6606.61006.6 806.6BINDING ENERGY, eVBINDING ENERGY. eV

Fig. 3-XPS spectrum of millerite (a, left) and millerite conditioned in 10-2 kmoVrn' FeClr4H20 (b. right).

MINERALS & ueT~ ~96 MAY 1990

by several washes with triple-distilled water. After drying,solids were removed from the filter paper and kept in glassvials for surface analysis.

XPS spectra were obtained using the Perkin Elmer Physi-cal Electronics Division ESCNAUGER system model num-ber 548 equipped with a magnesium anode. The samples wereloaded by first spreading the powders on a double-stick scotchtape attached to a molybdenum mask. The electrons leavingthe sample were detected by an electron spectrometer accord-ing to their kinetic energy. Identification of chemical stateswas made from the positions and separations of the peaks, aswell as from certain spectral contours.

Resull~ and discussion

XPS spectrum of natural pentlandite conditioned in 3.10-2kmol/m3 NaCI solution at the natural pH of the mineral isshown in Fig. la. As can be seen from the figure, the mostintense Ni peak [(Ni(2p )] is located at 855.0 eV along withNi(2p I) peak at 877.0 e Vbinding energies. The spectrum alsoexhibits Fe(2p]) peak at 714.0 e V and Fe (auger) peak at 608.0eV. .

synthetic sulfide mineral systems were investigated, and theresults are discussed below.

The XPS spectrum obtained for millerite is given in Fig. 3a.The binding energy of Ni(2pJ peak has been reported to be inthe range from 853.6 to 855.2 eV for Ni, from 852.9 to 857.4eV for Ni++, and approximately 861.0 eV for Ni++++ com-pounds (Clifford, Purdy, and Miller, 1975). In the presentinvestigation, the Ni(2P3) spectrum exhibited the most intensepeak at 857.0 e V and had a broad shoulder or a separate peakat 860.0 eV. However, when millerite sample was condi-tioned in 10-2 kmoVm3 FeCI2.4H2O, the X-ray photoelectronspectrum also contained Fe(2P3) peak at 712.0 eV of bindingenergy (Fig. 3b).

This suggests that the ferrous ions are adsorbed on themillerite surface as indicated by the conclusion from theelectrokinetic results that ferrous ions are adsorbed on themillerite surface electrostatically (Acar, 1985).

The spectrum obtained for millerite conditioned in 10-2kmol/m3 CuClr2H20 is shown in Fig. 4. It was reportedearlier that surface conversion of millerite to covellite isthermodynamically possible (Acar, 1985). In fact, CU(2P3)peak at 932.6 e V suggests such reactions are taking place onthe millerite surface.

As shown in Fig. 5a, the synthetic covellite spectrumdisplays a most intense copper peak at 934.1 eV for Cu(2p,).A secondary peak is located at 956.2 eV of binding energy.Conditioning of the covel lite particles in 10-2 kmol/m3FeClr4H20 results in another peak at the binding energy of

On the other hand, when pentlandite is conditioned inchalcopyrite supernatant, Cu(2p ) and Cu(2p) peaks are seenat the binding energies of933.6eV and 953.0eV, respectively(Fig. I b). Comparison of the two spectra shows that pentlan-dite surface can have significant amounts of copper specieswhen it is conditioned in chalcopyrite supernatant.

The spectrum obtained for chalcopyrite conditioned in3.10-2 kmol/m3 NaCI solution is illustrated in Fig. 2a. Thisspectrum exhibits copper peaks at 954.0 e V and 934.0 e V ofbinding energies for Cu(2p ) and Cu(2p]). Electron bindingenergies ofCu(2P3) for five Jifferent chalcopyrite samples arereported to be in the range of 932. 16 - 934.61 eV. The S(2p )peak is located at 163.6 eV, which corresponds to sulfide(Clifford, Purdy, and Miller, 1975).

While the chalcopyrite spectrum showed no indication ofa nickel peak, the spectrum of chalcopyrite conditioned inpentlandite supernatant exhibited a small Ni(2P:l) peak at854.6 e V electron binding energy (Fig. 2b). This most proba-bly resulted from the adsorption of nickelous ions at thisrelatively low pH (natural pH of 4.2) on the chalcopyritesurface. The indication of the adsorption of the dissolvedspecies of chalcopyrite and pentlandite was seen from theresults of zeta potential measurements of the minerals in thepresence and in the absence of supernatant of each other(Acar, 1955).

Based on the results obtained from surface chemical analy-sis of the natural minerals, a more systematic study employingselected synthetic sulfide minerals was initiated. Chemicalequilibria calculations reported earlier for the sulfide mineralsystems showed covellite to be more stable, followed bymillerite and pyrrhotite, and that surface conversion of theleast stable mineral to a more stable mineral can occur undercertain conditions (Acar, 1985).

In fact, comparison of supernatant composition of pynbotitewith that of pyrrhotite in CuCI., solution showed that themineral dissolution was increased significantly by CuCI."which itself was completely depleted, suggesting the ex-change of iron on the pyrrhotite with copper to form covelliteon the surface. Similar results were also obtained for themillerite/CuCI, and pyrrhotite/NiCI, systems, showing sig-nificant surface conversions. The surface conversion and/oradsorption of dissolved species in the selective flocculation of~

MAY'-. 97MINERALS & METAlLURGICAl PROCESSN;

CuS

CI)z0~~(J\&I-J\&I

...0

~\&Im~~z

(I)z0~I-(,).....J...

I&.0

~...m~:;)z

~

~

electrokinetic experiments suggest thatthe nickel species adsorb electrostaticallyat this acidic pH range (Acar, 1985).

The ESCA spectrum of syntheticpyrrhotite is given in Fig. 7a. The mostintense Fe photoelectron line (2p,) is lo-cated at the binding energy of 710.6 e V ,which is in agreement with the publishedreports (Perkin Elmer). Auger peaks of Feare also very clearly seen at 553.0 an~607.6 e V of binding energies.

Pyrrhotite is the most unstable of thesulfide minerals chosen for this study(Acar, 1985). Therefore, conditioning ofpyrrhotite particles in copper or nickelsolutions should be expected to result insignificant surface conversions. The re-sults of the ESCA spectrum of FeS in thepresence of 10-2 kmol/m CuCI2 (Fig. 7b)and 10-2 kmol/m3 NiCI2 (Fig. 8) supportsuch surface conversions of pyrrhotite tocovellite or to millerite, respectively.Spectra show intense copper and nickelpeaks at the binding energies of 934 e Vand 857 e V along with an Fe peak. at about714 e V, respectively.

966.6 916.6

1006.6 606.6 ~1006.6 806.6

BINDING ENERGY. .VFig. 5-XPS spectrum of covellite (a. left) and covellite conditioned in 1-0-' kmoVm3 FeCI.,4H.O(b. left).

714.6 e V , which is the Fe(2PJ) peak (Fig. 5b). Presence of thisiron peak on the covellite- contacted with FeCI2 solution Conclusionssuggests that ferrous ions, i.e. Fe++, the only major species atthe natural pH of 5.1, do adsorb on the negatively chargedcovellite. This was also seen from the electrokinetic data

reported previously (Acar, 1985).The spectrum of covellite treated with 10-2 kmol/m3

NiCI.,.6H.,O is given in Fig. 6. Although the resolution ofNi(2p~) peak is not good, this figure suggests that nickel ionscan adSorb on the covellite surface when contacted with NiCl2solution. This is also attributed to the adsorption of positivelycharged nickel species on the negatively charged mineralsurface. Correlation of this result with that obtained from

806.6

BINDING ENERGY. .V

ESCA studies have proven to be useful in investigating thesurface composition of natural and synthetic sulfide mineralsafter various treatments. Thus, comparison of spectra ob-tained for millerite in the absence and in the presence of CuCI.,showed copper to be present on the millerite surface whentreated with CuCI.,. Similar results were also obtained forpyrrhotite conditioned in CuCI., or NiCI., solutions. Since themineral particles were rinsed Several tImes with triple-dis-tilled water prior to surface analysis, these results show thatdissolved mineral species can strongly interact with the min-

Fig. 6.-XPS spectrum of oovellite conditionedin 10-2 kmoVm3 NiClr6H20.

MINERAlS & METAlLURGICAl P~SSING~ MAY 1~

eral surfaces and alter their surface properties. It is to be notedthat presence of such species can cause non-selectivity inseparation processes using flocculation or flotation of mineralmixtures (Acar and Somasundaran, 1985; Acar, 1985).

Ackno""edgment

The authors acknowledge the National Science Founda-tion for financial support of this work.

References

kat, S" 1985, "tnIeractions ~ Polymer, ~xWIg ~ and DissdYed Species in Selective ~ sy.-ns," DESc. 0iSSeI1aIkIn. Cdunna Univel8ty, SdIOOI

of E~ring and ApPi8d Science, ~ YOfk, May

Acar, S., and ~, P., 1985, "Effect 01 Dissolved Miner. Species on ~Fk)cculation 01 ~ " MkwaIs & ~ ~ November, pp. 231-325.

A-.inson, S.J., en.de, C.R., andRobefls, M.W., 1973, "ESCAs..-solC~onMetals: CaIbon Mono~ on Molybdenum and Tungslen Films," JoumaI of E/ec1Ion Spec-

troscopy and Related ~ Vol. 2. pp. 1 05- 1 09.

Brion,D., Hayer,J.andPredai,J.J., 1980, ~byESCAof&.faC8~of FWIe PyYh DuIWlg the FIOIation Process," F... Paltides PIocessbIg, ed. P. Sornasundaran,

Syn.,osium ProceectWIgS, SUE AmuaI MeeWIg. pp.544-557.

CIIfOfd, R.K.. Purdy, K.LandWler,JD, 1975, ~of~Mineral~

.. FrOItI FIotaIion SysIemS Using EI~ ~ lor ChenDI Analysis," AJCIIE

S~, pp. 138.147.

FnISI, D.C- leeder, W.R.. Tapping, R.L and w , B.. 1977, "An XPS SIudy of die

0xid8Ii0n of Pyrite ~ Pyrites.. CoeI,"Fu8;~, pp. 277-280.

~an, M.H., ~ DiUard, J.G.."An X-Ray :':-~ ~ (XPS) SblJyofCabal Adsorbed on 1118 Clay Mi c ," Journal of CoIa;d and kJtett8OB s.:I8nC:I8. Vol.

66. SepIenmer, RI. 345-351.

KoppeI~ M.H ~ DIIan1. J.C.. "An ESCA SbIdy of SOIbed M8IaI kIns on Clay""'"

MariIIe Chemistry in /he C-' ~ ACS sympoa.n, Vol 1, RI. 1 ~1.

Majima, H.. ~ MeIsRutgIcs/ QJaIWly, Vol. 8. RI. 269.

Manocha, A.S.. and PaIk, R.L, 1977, "FIoIaIKxI ReI8I8d ESCA S1ucies on PbS ~ "

~of&JttaCBSdencs, Vol 1, RI. 129-141.

P~ EJmw. Pt!ysk:aI Ei8droIj(:s Division, ~ of Chemical Analysis.

Ptakm, I.. 1959, "I.-radion of Mnerals wtfI Gases and Reagents In ~. Mining

~ Vol. 11, Mardt, RI. 319-324.

Ywnaski, T., and UU. S.. 1985, ~ared ~ Sl.des 01 Xan1haIe AdsoIbed on

ZW.c Sulfide," T~ SME-AIME, Vol. 232. M8d1. RI ~.

UAYI990 99MINERAlS & METAlLURGICAl ~~