279Powder Technology, 14 (1976) 279 - 285e Elsevier Sequoia S.A., Lausanne - Printed in the Netherlands

Mineralogical Heterogeneity of Ore Particles and Its Effects on Their Interfacial Claracteristics.

R. D. KULKARNI and P. SOMASUNDARANHenry Krumb School of Mines, Columbia Unive",ity, New York, N. Y; 10027 (US.A.)(Received October 10, 1975)

tl'"t

\i

SUMMARY

:;~ A significant proportion of the U.S.hematitic ores contain silica and otherimpurities in a very finely dispersed form.The mineralogical heterogeneity of theparticle surface and sub-surface regions ofsuch nattlral hematite and the effect ofdesliming treatments, are examined in thiswork using scanning electron microscopy,energy dispersive analysis of X-rays and Augerspectroscopy techniques. The implications ofthe observed mineralogical heterogeneity oninterfacial properties such as zeta potentialand adsorption, and on the validity of priorinterpretations of results obtained for suchproperties are discussed. Specifically, thereasons for a lower isoelectric point of naturalhematites than that of the synthetic hematite,and for the adsorption of anionic collectors onnatural hematite above their isoelectric point,are explained on the basis of the concentrationof impurities, as in this case silica and clay, inthe surface region.

"'~~

INTRODUCTION

knowledge of Ute nature of such mineralogicalheterogeneity on the surface of particles, andUtat of the manner in which Utey affect thesurface properties is essential for accurateelucidation of mechanisms involved in Uteinterfacial processes.

In tJ1e past two decades, tJ1e interfacialprocesses, such as flotation and flocculation,have been correlated witJ1 Ute zeta potentialof minerals. Thus, mechanisms of adsorptionof surfactants have been studied by measuringtJ1e variation in zeta potential of mineralsupon adsorption. While significant advancehas been made along Utis line, particularly forsynthetic minerals, confusion prevails fornatural minerals witJ1 regard to the actualmagnitude of tJ1e zeta potential of the mineralparticles and the location of Uteir isoelectricpoints. Widely differing values have oftenbeen reported by different authors for thesame mineral. For example, isoelectric pointsranging from 2 to 9 for various naturalhematites [1, 2] and from 8 to 9.1 forsynUtetic hematite [2] have been reported.Such differences can arise from tJ1e naturalvariations in Ute crystal structure [3J or thoseproduced during preparation [4].

In the case of natural minerals tJ1ere canalso be differences owing to the mineralogicalor chemical heterogeneity of Ute particleseiUter as a result of incomplete liberation, orbecause of the type of pretreatment, such asacid leaching, often used for cleaning theminerals. In the case of low grade iron ores,silica and clays are disseminated in a matrixof iron oxide; various ratios of silica tohematite can therefore result at the surfaceupon subjecting it to different treatments.WiUt such ore particles, Ute measured zetapotential and isoelectric point will lie be-tween the values for the individual mineralspresent in Ute particle surface region. The

~

Selective adsorption of reagents on one ormore minerals, in preference to others thatare present in an ore, has been the basis forflotation and flocculation for beneficiation ofores. The degree of separation obtained inUtese processes depends essentially upon thedegree of liberation of minerals during grind-ing. Minerals are not, however, usually fullyliberated; as a result the ore particles remainheterogeneous. Under such conditions, a

At

.Paper presented at the 7th Annual Fine ParticleSociety Conference, Philadelphia, Auguat, 1975.

i,~~,:,,",-;:.,: :~4;'

280

TABLE 1

Description of the mineral samples used and thetreatments received by them

(-No. Sample Treatment received

1

2

lh-in. hematite

lh-in. hematite

Unwashed

Washed with distilledwater several times

Washed with distilledwater and treatedwith 3 x 10-6 mole{lpotassium oleatesolution at pH 4

Des1imed

Obtained duringpreparation ofsample 4

Ultrasonicallytreated

3 ~-in. hematite

~":.~:F

4

5

325/400 mesh hematite

SlimesFig. 1. Typical scanning electron micrograpb of an un-wasbed natural bematite sample (No.1).

6 Ih-in. hematite

(.

:"¥~~ Fig. 2. Typical scanning electron micrograph of thewashed natural hematite sample (No.2).

MATERIALS AND METHODS

Hand-picked pieces of massive naturalMinnesota hematite sample (obtained fromWard's Natural Sciences Establishment) wereused for this study. Chemical and X-rayanalyses of the bulk sample showed it toconsist of 94 wt. % hematite and the restessentially quartz. Samples for tests wereprepared by roll-crushing and sieving. Thevarious size fractions used for the differentstudies are as follows: the fractions +4 mesh,325/400 mesh and the slimes for the spectro-scopic studies; the 35/65 mesh fraction forelectrokinetic studies; and the 325/400 meshfraction for adsorption and titration studies.

presence of a second mineral on the surfaceneed not, however, affect the response of theoriginal mineral to properties such as adsorp-tion and flotation. Thus, adsorption of ananionic surfactant, for example oleate oralkylsulfate, on the siliceous hematite canhave the same pH dependence as that for purehematite (except for a difference in the totalavailable surface area, and hence the mag.nitude of surfactant adsorbed in the completepH range). However, the pH dependence ofthe zeta potential of it will be different fromthat of pure hematite or silica. Clearly, acorrelation of results of electrokinetic studieswith other interfacial properties for determin.ing their mechanisms can, in such cases, yieldmisleading results. Such correlations areusually made using mineral samples hand-picked for impurity content. The purpose ofthe present study was to examine thechemical composition in the surface and sub.surface region of such a selected mineralsample, and its relation with the interfacialproperties. A natural hematite sample wasexamu1ed using scanning electron microscopy,low.angle X-ray analysis and Auger spectro-scopy techniques before and after variouspretreatments. The results obtained arecorrelated with the electrokinetic, titrationand surfactant adsorption properties of thenatural minerals.

'1-

281

TABLE 2

Probe analysis of various hematite samples showing the effects of washing, desliming, and ultrasonic treatment:oxygen is excluded for calculation of the atomic percentages

Elements Sample 1 Sample 2Unwashed, at. % Washed, at. %

Sample 5Slime, at. %

Sample 4Deslimed, aL %

Sample 6Ultrasonically treated.at. %

FeSiAINaMg

88.78.41.9

0.9

79.410.3

1.94.13.4

81.910.64.40.51.7

75.110.57.71.5..9

88.57.72.00.70.9

Spectroscopic studies were conducted usingunwashed, wllShed and ultrasonically treatedsamples, and slimes which resulted from thewashing of 325/400 mesh hematite sample(See Table 1). Zeta potential and isoelectricpoint* were detennined using streamingpotential [5] and microelectrophoresis [6]techniques, and point of zero charge* wasdetennined by titrating the mineral suspen-sions with acid or alkali. Adsorption capacitieswere detennined using radioactive oleate [7] .

RESULTS AND DISCUSSION

The scanning electron micrograph of atypical region on an unwashed hematitesample (No.1) is given in Fig. 1. The analysisshowed an iron content of 88 at. %** for thesurface and sub-surface region (1 Jl.m thick),which is lower than the bulk concentration of96%**. Auger analysis of the surface regionalone gave lower values in the range of 50 -85%**. Washing of the sample removed mostof the slimes on its surface (see Fig. 2). Acomparison of tJie results from the probeanalysis of the unwashed and washed samplesgiven in Table 2 shows a decrease in ironcontent upon washing, suggesting removal ofhematite during that operation.

The heterogeneity of the surface wasstudied by analyzing different spots on the

*While isoelectric point describes that condition ofthe system at which the potential at the plane ofshear, i.e. the zeta potential as obtained from electro-kinetic measurements, is zero, the point of zero chargerefers to that at which the net surface charge is zero,and can be different from the isoelectric point [8].

**Oxygen and carbon are excluded for the calcula-tion of the atomic percentages.

same hematite particle, for example tJIe spotsE, G and H shown in Fig. 3. Probe analyses ofE and G spots were similar and indicated tJIemto consist mostly of AI and SL Spot H, on tJIeother hand, is composed mainly of Fe. Sig-~cant variation in chemical compositionexists between various spots of suffi-cient magnitude to consider it as tJIe result ofmineralogical heterogeneity of the surfaceregion. The effects of such variation of thesurface region will be examined later.

The distribution of Fe and Si on the sur-face of a washed hematite sample, shown inFig. 4, suggests hematite and silica or clay tobe intimately associated with each other.Liberation of the minerals in such samples forsubsequent separation would obviouslyrequire imer grinding than is usually employ-edin their processing.

Slimes produced during mineral processingoperations often respond poorly to mostseparation treatments, and yet they are thesource of a huge loss of mineral values.Differences between surface regions of des-limed hematite samples and those of the slimederived therefrom were therefore studied. Inaddition to the presence of particles of varioussizes and shapes, the surface and sub-surface ofslimes were found to analyze higher in Mg, AIand Na than those of the deslimed sample(Table 2). This is possibly due to the presenceof a large amount of clays in the slime frac-tion. Auger analysis of the surface of thesesamples also showed the presence of largeamounts of Al and Mg (see Table 3).

The effect of ultrasonic treatment on thehematite samples was to make the surfacemore uniform in appearance and pitted, andto cause a significant reduction in Si, Na andMg, possibly due to removal of clay (see Fig. 5and Tables 2 and 3). Whereas simple washing

282

c

(a) (a)

(AI Si Fe

(b) (b)

Fig. 4. (a) Fe distribution, (b) Si distribution in atypical washed natural hematite sample (No.2).

removed hematite slime only, ultrasonic treat-ment apparently removed clays also, changingthe Fe content of the surface to approximatelythat of the unwashed samples. These resultsclearly suggest that even simple physical treat-ments such as washing or ultrasonic vibrationcan cause marked alterations in the surfacechemical composition, which in turn can beexpected to influence such surface propertiesas zeta potential.

The most important among the aboveeffects is that of the surface chemical ormineralogical heterogeneity on the interfacialproperties. The isoelectric point as determinedby streaming potential and electrophoreticmeasurements of the natural hematite used inthe present study was at about pH 3.0. Thetitration experimen~, however, yielded apoint of zero charge of 7.1. Also, a maximum

AI Fe FeSi(c)

Fig. 3. (a) Scanning el~tron micrograph of a hematitesample on which spot analysis was conducted. (b)EDAX analysis of spot G shown in Fig. 3(a). Analysisof spot E was similar to that obtained for spot G. (c)EDAX analysis of spot H shown in Fig. 3(a). )-

, '-::""',;' c'

,~ f~;'::i'~,""'~' .~:..'",Z';:~~;

283

TABLE 3Auger analysis of various hematite samples

Sample 6lntrasonicallytreated, at. %

Sample 2Washed, at.%

Sample 5Slime, at. %

Sample 3Washed and oleicacid treated, at. %

Sample 4Deslimed, at. %...

c~S~

&1 Excluding Including Excluding Including Excluding Including Excluding Including Excluding Including0 and C 0 and C 0 and C 0 and C 0 and C 0 and C 0 and C 0 and C 0 and C 0 and C

,27.14.23.20.30.10.3

43.420.8

FeSiAIMgCaTi0C

73.218.06.81.50.5

22.05.42.10.50.2

37.232.3

34.657.3

5.61.70.8

50.824.313.9

3.06.81.2

24.511.76.71.43.30.6

50.6

40.823.326.6

7.12.2

16.99.7

11.12.90.9

54.81.7

76.911.9

9.20.850.30.85

11.118.4

1.80.60.3

19.548.3

!

)~~

+30

'~~~:::~I I II I I I I

ct.:' I RSo.4~

IEP -- 0 -.I . 1.0 . 10 M.2!! / .(,. a 7.6. 10'.

. 1.0. 10-'0 2.'.'0-10 5.0kl0.1+ 1.0.10-4x 1.0.10-'

>+20E

i+'OIz\&I ,... 004--«... -10\&IN

\,

"-20 .-301~~ -

Fig. 5. Typical scanning electron micrograph of anultrasonically treated hematite sample (No.6).

~ x-x--;:-x ~ ~

I I I. I I I I2 3 4 5 6 7 8 9 10

pH

: Fig. 6. Zeta potential of natural hematite in sodium. dodecylsulfate solutions (data of Shergold and

Mellgren [11,12]).

i

hematite can be higher than that of the sameamount of silica [10]. The difference inappearance between silica regions, E and G,and hematite region, H, in Fig. 3 might alsobe noted in this regard. Nature of adsorptionof a surfactant such as oleate that adsorbs onhematite but not c;>n quartz will be the ~efor natural hematite containing quartz and forpure hematite. It follows that the electro-kinetic processes and adsorption processes canrespond differently to the existence ofmineralogical heterogeneity on the surface ofa particle. Such differences in responsereported in the literature, which have oftenbeen the source of misleading conclusionswith regard to the mechanisms involved, cannow be explained on the basis of t1le observa-

in the floatability of hematite using oleate,which in many earlier studies has been foundto occur at its point of zero charge [9] , wasobtained here at pH 8 [7] . The abovediscrepancY between isoelectric point andpoint of zero charge can be explained on thebasis of the mineralogical heterogeneityidentified for the surface regions. The presenceof silica on the particle surface in significantproportions has evidently caused a shift in theisoelectric point of the sample from that ofpure hematite towards that of silica. However,values obtained for the point of zero chargefrom results of adsorption and flotation areexpected to be much closer to that of purehematite, since the magnitude of theseprocesses js controlled mainly by the surfacearea and surface charge density of minerals,and since effective surface charge density ofL

284

,(

1000 2000 3000 4000

AGING TIME, MINUTES

Fig. 7. Zeta potential of natural hematite as a function of aging time with intermittent addition of 0.1 M KOHsolution.

The complex electrokinetic behaviorexhibited by natural hematite as a function oftime can also be explained on the basis of themineralogical heterogeneity. The change inzeta potential of hematite as a function oftime with intemrittent addition of alkali isshown in Fig. 7. Initially, a negative zetapotential was obtained for hematite at pH3.5 - 3.6. Addition of potassium hydroxidesolution changed the zeta potential to apositive value, With aging, however, the zetapotential reverted to a negative value. Thisbehavior repeated i~lf several times until apH of 5 was reached, above which add~tion ofthe alkali produced the usually expectedincrease in negative zeta potential. Generationof positive potential due to the addition ofalkali is unusual, and is possibly due to thedissolution of silica from the surface by theadded alkali and! or precipitation of hematitefrom solution upon the instantaneousincrease in pH. Either phenomenon will resultin a hematite-rich surface that is positivelyCharged. During subsequent aging, a decreasein pH occurs and this can be expected tocause both reprecipitation of dissolved silicaand redissolution of iron hydroxide. Bothreactions will again cause the particles to bemore siliceous on the surface and hence morenegatively charged.

tions made in this study. An interestingexample is the observation by Shergold andMellgren [11] that anionic dodecylsulfateadsorbs above the isoelectric point of theirnatural hematite sample when its particles arenegatively charged. Their plot of zetapotential of hematite as a function of pH inthe presence of dodecylsulfate is reproducedin Fig. 6 [11, 12] . The isoelectric point ofthis hematite is at pH 4.7. The increase innegative charge upon the addition of dodecyl-sulfate, even above the isoelectric point,suggests adsorption of the surfactant in thatpH range also.

Similar adsorption of oleate on negativelycharged alumina has been reported by Guptaand Smith [13]. The existence of such anadsorption of negatively charged collectors onnegatively charged minerals has often beeninterpreted to suggest chemisorption, which isnot of coulombic origin. In the light of thepresent discussion on the effects ofmineralogical heterogeneity on isoelectricpoint, it is evident that such an interpretationis unnecessary. Though the isoelectric pointof the natural hematite (which usually containssilica in it) used by Shergold and Mellgren [11]was 4.7, the hematite regions on the surfacecan be considered to be positively chargedeven in the pH region 4.7 - 8 so as to favouradsorption of sulfonate on it.

""... '~':;

~~~j

285

CONCLUSION ACKNOWLEDGEMENTS

Support of the American Iron and SteelInstitute Grant (No. 23-274) is acknowledged

REFERENCES

)

1 A. S. Joy and D. Watson, Occurrence andproperties or hematite slimes, Proc. Vlth Int.Mineral Processing Congress, Pergamon Press,Oxrord, 1965, pp. 355.369.

2 G. E. Parks, The isoelectric points or solid oxides,solid hydroxides, and aqueous hydroxo complexsystems, Chem. Rev., 65 (1965) 177 - 198.

3 T. W. Healy, A. P. Herring and D. W. Fuerstenau,The errect or crystal structure on the surfaceproperties or a series or manganese dioxides,J. Colloid Interface Sci., 21 (1966) 435 . 444.

4 I. J. Lin and P. Somasundaran, Alterations inproperties or samples during their preparation bygrinding, Powder Technol., 6 (1972) 171 . 179.

5 P. Somasundaran and R. D. Kulkarni, A newstreaming potential apparatus and study of temp-perature effects using it, J. Colloid Interface Sci.,45 (1973) 591 - 600.

6 P. Somasundaran and D. W. Fuerstenau.Mechanisms or alkyl sulronate adsorption at thealumina-water interface, J. Phys. Chem., 70(1966) 90.

7 R. D. Kulkarni and P. Somasundaran, Oleateadsorption at hematite/solution interface and itsrole in notation, AIME Meeting, New York.Feb. 1975.

8 P. Somasundaran, Zeta potential or apatite inaqueous solutions and its change during~uilibration, J. Colloid Interrace Sci.. 27 (1968)660.

9 A. S. Peck, L. H. Raby and M. E. Wadsworth, Aninfrared study of the flotation or hematite witholeic acid and sodium oleate, Trans. AIME, 235(1966) 301; R. W. M. Lai and D. W. Fuerstenau,A model ror the surface charge or oxides andflotation response. 104th AIME Annual Meeting,New York, Feb. 1975.

10 A. Breeuwsma, Adsorption of ions on hematite(Q-Fe20a), A colloid chemistry study, Thesis.Agricultural Univ., Wageningen. 1973, p. 65.

11 H. L. Shergold and 0. Mellgren. Concentration orminerals at the oil-water interface: hematite-isoocatane-water system in the presence orsodium dodecylsulfate, Trans. !MM, C123 - C132.

12 D. W. Fuerstenau, Adsorption of surractants atthe solid-water inter races. in M. L. Hair (Ed.),The Chemistry or Biosurfaces, 1, Marcel Dekker,New York. 1971, pp. 143 - 176.

13 S. C. Gupta and R. W. Smith, Amphoteric collector-mineral charge interaction and flotation, inP. Somasundaran and R. B. Grieves (Eds. ),Advances in Interfacial Phenomena or Particulate/Solution/Gas systems, AIChE Symp. Ser., Vol. 71,no. 150, New York. 1975, pp. 94.99.

(1) The mineralogical heterogeneity of thesurface and sub-surface regions of hematiteparticles and slime was examined in this study,and the effect of the heterogeneity on theirinterfacial properties is discussed. Using low-angle X-rays and Auger spectroscopic techni-ques, hematite particles were found to bericher in silica and clay in the surface regionthan in the bulk, the proportion of mineralsin the surface region being dependent uponthe pretreatment that the particles received.

(2) Samples of natural minerals such ashematite exhibit complex interfacial behaviordifferent from that of the pure or synthetichematite. For example, negatively chargedalkylsulfate adsorbs on natural hematite inmeasurable amounts even above its isoelectricpoint. Also, the electrokinetic behavior ofnatural hematite and its response to aging incertain pH ranges upon addition of alkali isunusual. Addition of alkali below pH 5 causedan instantaneous change of zeta potential to apositive value, which upon aging reverted to anegative value.

(3) The complex behavior of natural oreparticles can be explained on the basis of themineralogical heterogeneity of the surface andthe concentration of impurities in the surfaceregion. While the electrokinetic properties,such as isoelectric point, will be an average ofthose for the individual minerals present onthe surface, adsorption properties, being theresult of processes at a molecular level, can becharacteristic of an individual mineral, if otherminerals are inert towards the adsorbents.Adsorption, for example, of alkylsulfate onnatural hematite can be considered as electro-static in nature, even above its isoelectricpoint, if the effect of the concentration ofsilica or clay on the surface is recognized.Alternative mechanisms, proposed in the past,involving for example chemisorption, are un-warranted.

(4) The study emphasizes the need formonitoring of surface composition whilestudying interfacial properties of naturalminerals, and for the recognition of theeffect of the variation in chemical andmineralogical composition on these properties.

(,~~~~