Acid Pressure Oxidation of Arsenopyrite

Canadian Metallurgical Quarterly, Vol. 29, No.1, pp. 13-20, 1990Printed in Great Britain

0008--4433/90$3.00+ .00Canadian Institute of Mining and Metallurgy

Pergamon Press pIc

ACID PRESSURE OXIDATION OF ARSENOPYRITE:PART II, REACTION KINETICS

V. G. PAPANGELAKIS and G. P. DEMOPOULOSDepartment of Mining and Metallurgical Engineering, McGill University, Montreal,

Quebec H3A 2A7, Canada

(Received 9 February 1989; and in revised/arm 28 September 1989)

Abstract-The kinetics of pressure oxidation of narrow-sized arsenopyrite fractions (triclinic and mono-clinic) was experimentally determined in the temperature range 120-180C and pressure range of 2-20atm 02' A surface reaction control shrinking core model was found adequately to describe the kineticresults. The partial formation of molten elemental sulphur and the massive precipitation of scorodite wasfound not to interfere with the progress of the heterogeneous oxidation process. The dependence of theapparent rate constant on temperature (EA = 66-72 kJ mol-I), oxygen pressure (first order), and particlesize (ks vs 11ro linear) were in agreement with the surface reaction control model. The true rate determiningstep was postulated to be electrochemical in nature involving the first electron transfer during reductionof the surface-adsorbed oxygen. Finally the following intrinsic rate equation was developed to representthe pressure oxidation of arsenopyrite in a batch reactor:

dN (-8672) .--=49.527exp -- Po molmm-Icm-2Sdt T 2

INTRODUCTION

Acid pressure oxidation is currently considered as the mosteffective process for the treatment of refractory gold con-centrates [1]. Arsenopyrite and pyrite are the most commonhost minerals of refractory gold. From a process optimizationstandpoint it is important to know the behaviour of each ofthese minerals during acid pressure oxidation. Although theaqueous oxidation of pyrite by molecular oxygen has beeninvestigated in the past by various researchers [2, 3] the sameis not true for arsenopyrite. As a part of an on-going researchprogram focussing on the chemistry, kinetics and modelling ofthe pressure oxidation process the system FeAsS-H2S04-02has been studied. The subject-matter of the present paper, isthe kinetics of pressure oxidation of arsenopyrite. The reactionchemistry of the system is described in Part I [4]. Other papersreport on the mathematical modelling of the batch [5] andcontinuous oxidation [6] of arsenopyrite.

Pressure oxidation of arsenopyrite has been found [4] to bestoichiometric and to follow two parallel competing paths:

4FeAsS(s) + 1302(aq) +6H20(l) --+ 4H3As04(aq)

+4Fe2+ (aq) +SO~- (aq) (1)

4FeAsS(s) + 702(aq) +8H+ (aq) +2H20(l) --+ 4H3As04(aq)

+4Fe2+ (aq) +4S(l) (2)

The SOforming reaction occurs to a significant degree (approx.20% yield) only at temperatures below 160C and in the pres-ence of high sulphuric acid concentrations. Arsenic reports insolution directly as H3AsOiaq) while ferrous ion is furtheroxidized to yield ferric ion

2Fe2+ (aq) +102(aq) +2H+ (aq) --+ 2Fe3+ (aq) +H20(I).

arsenate which precipitates as crystalline scorodite

Fe3+ (aq) +H3As04(aq) +2H20(l)

--+ FeAs04 2H20(s) + 3H+ (aq). (4)

For comparison reasons the kinetics of pyrite oxidation isbriefly reviewed here in view of their common occurrence inrefractory gold concentrates. The pressure oxidation of pyritehas been found [7-10] to be chemically controlled exhibitingpredominantly first order dependence [7, 8] at the lower O2pressure region (

14 V. G.PAPANGELAKIS and G. P. DEMOPOULOS: ACID PRESSURE OXIDATION OF ARSENOPYRITE: PART II

mineral specimens originated from Gold Hill, Utah (monoclinicvariety) and Mexico (triclinic variety). No detectable sulphide/arsenide impurity phases were present in either specimen.Quartz was the only impurity phase identified (120/0in mono-clinic and 1.30/0in triclinic). The massive mineral specimensafter crushing, grinding and wet sieving were split into threenarrow-sized fractions, namely -147 104 /lm, -74+ 53 /lmand - 44+ 37 /lm. In Fig. 1 typical ground mineral particlemorphologies are depicted. Some quartz-totally liberated-particles are seen to be present in the monoclinic specimen(Fig. la).All pressure oxidation tests were conducted in a 300 ml Parr

mini autoclave with the internal parts made of titanium. Typ-ically for each experiment 0.2 g of arsenopyrite was chargedinto the autoclave along with 220 ml of sulphuric acid solution.Temperature and pressure were kept constant throughout eachexperiment. Periodically, aliquots of leach liquor were with-drawn from the pressure reactor and analyzed for As(tot) andoccasionally for Fe(tot) by AA spectrophotometry. For thecalculation of conversion or equivalently "fraction extracted"(R), correction factors were applied to account for the volumeand mass losses due to sampling. More details on the exper-imental procedure have been given previously in Part I [4].

Fig. 1. Scanning electron micrographs of the two arsenopyrite speci-mens: (a) monoclinic, (b) triclinic.

RESULTS

As it was established in Part I of this work [4],during pressureoxidation of arsenopyrite the highly insoluble scorodite com-pound forms. It was decided thus at the early stage of the kineticinvestigation to establish a slurry density that would not beassociated with precipitation. In that way the oxidation kineticsof FeAsS could be studied by monitoring simply the dissolvedarsenic and/or iron concentrations and thus greatly facilitatingthe overall investigation. Furthermore, by using a low slurrydensity, possible O2 mass transfer problems were effectivelyprevented thus securing the unambiguous measurement of theintrinsic reaction kinetics. From a number of tests, wherethe amount of arsenopyrite mineral leached was varied, whilethe other variables were kept constant at 130C, 10atm O2 and0.5 N H2S04, a slurry density of 0.1% solids was found to giveprecipitation-free conditions. Thus, throughout this series ofexperiments the slurry density was kept constant at 0.1% solids,that is 0.2 g FeAsS/0.21 solution.Having thus established the slurry density for the kinetic

investigations, the following variables were then studied: stir-ring speed, acid concentration, temperature, oxygen pressure,and particle size.Stirring speeds in the reactor were varied from 530 to 930

rev./min. Lower speeds than 530. rev./min were not appliedsince this was found to be the minimum speed to give adequatesolids suspension, as it was determined by carrying out a seriesof agitation tests in glass liner that allowed visual observation.Results from these experiments expressed in terms of "fractionextracted", are shown in Fig. 2. Over this range of stirringspeeds, no significant effect on the oxidation rate was found.Based on these data, the stirring speed of 730 rev./min waschosen for all subsequent tests.Before the effects of the other variables are presented, it is

important to clarify that the "fraction extracted" used to illus-trate the results in this paper is based principally on As measure-ment. In several instances "fraction extracted" based on Femeasurement was determined as well for comparison p~rposes.However, As was chosen as the main element to monitor theoxidation kinetics of FeAsS as a precaution against possibleinterference in the iron determination from corrosion of thesampling valve made of stainless steel.All the results reported here were conducted at a constant

0.5 N H2S04 concentration. The acid generated by the oxi-

1.001lJI- 1300 Cu

10 atm O20:: 0.5 N H2SO4I- gx -200+270 mesh1lJ0.5 ~

z g ft0 ~ 930 rpmI-

t\ c 730u ~ 0 o 5300::ll. ~

10 30 60 120TIME (min)

Fig. 2. The effect of stirring speed on arsenopyrite conversion. (Exper-imental conditions: Boac; 1013 kPa O2; -74+ 53 .urn.)

V. G. PAPANGELAKIS and G. P. DEMOPOULOS: ACID PRESSURE OXIDATION OF ARSENOPYRITE: PART II 15

1.0r-r--~"----,---r-~-"'-----r-----,oWI-U

DISCUSSION

16 V. G. PAPANGELAKIS and G. P. DEMOPOULOS: ACID PRESSURE OXIDATION OF ARSENOPYRITE: PART II

Shrinking core models

The object of the present kinetic study was primarily todevelop a rate equation useful for reactor design and processmodelling. To this end, the obtained experimental data areanalysed in this section with the aid of shrinking core modelswhich have been previously established [14, IS]. Moreover amechanistic interpretation of the kinetic results is presented.

Disregarding diffusion control through the external bound-ary layer (see Fig. 2), then the pressure oxidation of arsen-opyrite is expected to be controlled by either the surface reactionor the diffusion process through the product layer or com-bination of the two steps (i.e. mixed control). In the case ofspherical (or more precisely equiaxed) particles the followingshrinking core models are available for single particles andconstant solution concentration [IS]

1-(1-R)1/3 = kst

for surface reaction control,

l-1R-(1-R)2/3 = kDtfor product layer diffusion control and

for mixed control, where R is the "fraction extracted" at timet and ks, kD and km are apparent rate constants. These constantsare given in equations 8, 9 and 10.

bkM[02Yks=---- pro

k = bkM[02]m pro

Moreover

where

b the arsenopyrite stoichiometric coefficient, equal to ~ sincereaction 1 dominates in the present system.M molecular weight of FeAsSp density of FeAsSro initial particle radiusk intrinsic rate constantDc effective diffusivity of 02(aq) through the product layer[02] bulk concentration of dissolved oxygena reaction order in terms of dissolved oxygen concentration.

The concentration of dissolved oxygen is considered to beproportional to its partial pressure according to Henry's law[16] i.e. [02] = Kt; I POry.The shrinking core "inodel equations can be applied to the

present heterogeneous reaction system provided that some basicassumptions are met. These are: (a) the particle retains itsequiaxed shape, (b) the product layer is porous, (c) the con-

centration in the bulk of the solution remains constant, (d) thetelnperature is uniform throughout the heterogeneous reactionzone. In addition, the mixed-control model equation 7 is appli-cable provided the reaction is of first order and the pseudo-steady state assumption holds [15, 17]. In the present system theproduct layer may consist of molten SOand scorodite particlesadhering on the unreacted mineral surfaces. The pressure waskept constant. Finally despite the high exothermic nature of theoxidation reaction I (~H~9S = 1430 kJ molj;-c.Ls) [4] assump-tion (d) is believed to be met in view of the very dilute slurryemployed.

(5)

Initial reaction order

Before the shrinking core model equations are applied to theexperimental kinetic data the order of the initial reaction ratewith respect to the partial pressure of oxygen will be established.The pressure instead of the dissolved oxygen concentration isused here which implies that Henry's constant is incorporatedinto the intrinsic rate constant.

The rate of arsenopyrite oxidation expressed as moles ofoxygen disappearing per unit time, per particle is

dN- ~ = 4nr2k' P/dt o~

(6)

(11)

(7) where

(8)

No, number of moles of oxygen at time tr nidius of unreacted core at time tk' apparent intrinsic rate constant (= Khll k)P o~ partial pressure of O2

From the stoichiometry of reaction I

(9)dl\To~

dt( 12)

13dlV4 dt

(10)where

N number of moles of FeAsS present at time t.

On the other hand, arsenopyrite fraction extracted, R, is givenby the following equation

lVO-1VR=--

I\T 0( 13)

where

No initial number of moles of arsenopyrite.

By differentiating equation 13 and combining with eqs. II and12, equation 14 is obtained

dR 16 ,dt = 131\To m-k' Pb~.

A plot of In dRjdt at t = 0 versus In P o~ would result in astraight line with slope a. The derivatives of R are taken att = 0 since no elemental sulphur has yet been formed to interferewith the oxidation reaction. Second order polynomialregression was performed to fit the experimental data of Fig. 5.By taking the logarithms of the initial slopes (SJ of theregression polynomials at t = 0 and plotting them against thelogarithms of the respective oxygen pressures the graph of Fig.

(14)


2 10 20

o /-1

(j)

c n = l06:!: 0.06-2

0.5 2.0 3.01.0lnP02

Fig. 8. Determination of the reaction order with respect to oxygenpressure by considering initial kinetics.

8 was obtained. The resulting straight line (correlation coef.0.997) has a slope of 1 indicating that the reaction is of firstorder.

Surface reaction control

Analysis of the arsenopyrite pressure oxidation results (Figs2-7) indicated that all were in very good correlation with equa-tion 5. Typically the agreement obtained is illustrated in Fig. 9,which represents the results of Fig. 4. The linearity of the"1- (1- R) 1/3 vs time" plots suggests that the reaction con-trolling step of the pressure oxidation of arsenopyrite is thechemical reaction taking place at the surface of the mineral. Theother two model equations 6 and 7 failed to yield satisfactorycorrelation.From these findings, it appears that the minor liquid sulphur

formation (10-20% yield [4]) especially at the lower tem-perature range investigated, does not interfere with the reactionkinetics of arsenopyrite oxidation. In other words, liquid sul-phur does not form a protective impervious layer around thearsenopyrite particles. This is in contrast to the pyrite pressureoxidation system, where liquid sulphur has been found essen-tially to stop the reaction below the 1000/0oxidation level [9].The totally different behaviour of the two minerals can beexplained by comparing their respective Pilling-Bedwarth

n

~0.40:

17~/'60/ lwe ./ ./. 140C

. /./ / / /. 130C/.. .~or- 0.2 / ./ _

, / /. ./~. __ 1200C

~.. ----. --.----./ . -----. .--.

:::::::::::::::--.--0.0 --=::. __ ...&.--...J.----JI.....---'-_..L..-.----L -'-- --'10

0.6

30 12060TIME (min)

Fig. 9.Plots of the shrinkingcore--ehemicalreactioncontrolmodelatdifferent temperatures. (Data from Fig. 4.)

ratios [18],z, defined as the volume of insoluble product formedper unit volume of solid reactant. Thus, the z values are

2 x 16.3 cm3 g-atomsolz = 24 3 1-1 = 1.36cm mo FeS2

for pyrite and

16.3 cm3 g-atomso1z = . 3 I = 0.62

26.5 cm molpeAss

for arsenopyrite. The molar volume values were calculated fromWeast [19]. For sulphur, the value refers to crystalline mono-clinic Sf3 (m.p. 119C). At temperatures above 119C the molarvolume would increase but it is not expected to significantlyaffect the z values. It can be clearly seen that z is greater thanunity in the case of FeS2 (assuming that all sulphidic sulphuroxidizes to SO)and this can result in a sulphur layer envelopingcompletely the reacting pyrite particle. However, in the case ofFeAsS (z < 1) the produced sulphur does not suffice to form asimilar protective layer thus allowing the reaction to proceedto completion without any hindrance.Apart from the elemental sulphur, of concern to the present

reaction system is also the possible effect of precipitation offerric arsenate could have on the progress of the oxidationreaction. Application of the shrinking core-reaction controlmodel to the experiments represented by Figs 2 and 5 publishedin Part I [4], in which precipitation had taken place, gaveexcellent agreement with the experimental data. The linear plotobtained is given in Fig. 10. In the same Figure, data from aprecipitation-free (0.1% solids) experiment is plotted for thepurpose of comparison. Thus, it can be concluded that ferricarsenate precipitation does not interfere with the overall oxi-dation kinetics of arsenopyrite. This observation supports thefundamental assumption made in the early stages of the presentinvestigation to monitor the kinetics of the oxidation reactionunder precipitation-free conditions.From the slopes of the "1- (1- R) 1/3 vs time" plots the

ks values were derived and the Arrhenius plot (Fig. 11) wasconstructed for both the monoclinic (Fig. 4) and triclinic (Fig.7) arsenopyrite oxidation test results. Since the ks values aredirectly proportional to the intrinsic rate constant (eqn 8) andthe data represent experimental runs of the same particle size

0.4

0:0.2I

00.1 % dsYO.5-to 2.5

60 120 240TIME (min)

Fig. 10. Plot of the shrinking core-ehemical reaction control modelfor differentslurry densities.(Experimentalconditions: Boac; 1013

kPa O2; 0.5 N H2S04; -74+53 }lm.)


180TEMPERATURE (Oe)160 140 120

EA=72.15.8 kJ/mol

EA=65.54.9 kJ/mol

6.0 monoclinic

triclinic

6.8 ."2.2 23 2A

1/ T (K-~103)Fig. 11.Arrhenius plots for monoclinic and triclinic arsenopyrite. (Data

from Figs 4 and 7.)

2.5

(-74+ 53 flm), the activation energy calculated is directlyrelated to the intrinsic kinetics. The Arrhenius plot revealed anactivation energy of 72.1 kJ mol- I (17.2 kcal mol-I) with acorrelation coef. 0.988 for the tests involving the monoclinicarsenopyrite specimen and 66.5 kJ mol-I (15.9 kcal mol-I)with a correlation coefficient of 0.998 for the tests involvingthe triclinic mineral variety. This range of activation energysupports the view that the pressure oxidation of arsenopyriteis controlled by the rate of the chemical reaction at the surfaceof the particles.Comparison of the activation energy determined for the pres-

sure oxidation of arsenopyrite with those reported for pyriteoxidation (50-55 kJ mol- I), shows stronger temperaturedependence for the former mineral.The pre-exponential factor of the intrinsic rate constant was

also determined with the aid of the Arrhenius plot (monoclinicspecimen). Finally the explicit expression of the intrinsic rateconstant for the pressure oxidation of FeAsS (monoclinic)becomes

(-8672) .k'=49.527exp --T- molmln-Icm-2atm-I

Overall oxygen pressure dependence

The linear plots of Fig. 12 add further support to the findingthat the oxidation process is chemically controlled. Here theresults of Fig. 5, representing experiments run at various partialoxygen pressures, are plotted in terms of the chemical reactioncontrol model (eqn 5). The straight lines obtained indicate thatthe reaction mechanism does not change within the oxygenpressure range of 2-20 atm (202.6-2026.4 kPa).The consistency of the first order dependence on pressure

which was determined using initial rates (Fig. 8) was verifiedwith the shrinking core model results of Fig. 12. Thus by plot-ting the logarithm of ks versus the logarithm of oxygen pres-sure (Fig. 13) a linear plot with slope essentially 1 (correlationcoefficient 0.988) was obtained. This plot confirms that pressure

20 atm/'/. ,/ 10.. ~./ .~.

/_ ...-/.~ .-- 5

/. ~. .--g--- . .,..,.,-__ .-- ..- 21/ ...,..,.,_. _._.-

0.0~I,-::-~-

0.4

~0.2

10 30 50 120TI ME (min)

Fig. 12. Plots of the shrinking core-ehemical reaction control modelat different oxygen pressures. (Data from Fig. 5.)

oxidation of arsenopyrite is first order with respect to oxygenpressure for the whole pressure range studied (2-20 atm). Firstorder dependence on pressure has also been reported for pyriteoxidation at temperatures and pressures in the range of 25-160C and 0.2-10 atm, respectively [7, 8, 20]. However, forhigher pressures a 0.5 order dependence has been reported forpyrite [9]. No tendency towards fractional order was detectedin the present arsenopyrite oxidation system as it is testified bythe plots of Figs 8 and 13.

Particle size and acid effects

Considering equations 5 and 8 and for a given temperatureand oxygen pressure, plots of 1'0[1- (1- R) 113] vs time wouldbe expected to give a common straight line for the differentparticle sizes investigated. Such a "normalization" plot is shownin Fig. 14, where 1'0 is the geometric average radius of theparticles corresponding to each size fraction. As it is seen,correlation of the three size fractions is very good.The results of Fig. 3 were similarly analyzed with the aid of

the shrinking core model and the dependence of the apparentrate constant (ks) on H+ concentration was found to be 0.3(correlation coefficient 0.995). The pertinent data are given inTable 1. For the estimation of free H+ concentration, H2S04was taken as being 500/0dissociated, since its second dissociationconstant is only 8 x 10-3 at 130C [21]. Bailey and Peters [9]

(15) Po (atm), 5 202 10

-6

V)oX

c.-7

/// n= 0.90! 0.03

-805 1.0 2.0 3.0

1nPo,

Fig. 13. Determination of the reaction order with respect to oxygenpressure by considering overall kinetics.


-147+104E 0-74+53:::L. 0-44 + 37......'0

~m

30 12060TIME (min)

Fig. 14. Normalization plot of the shrinking core model. (Data fromFig. 6.)

also found that increased H 2S0 4 concentrations have a ben-eficial effect on pyrite pressure leaching.

Electrochemical mechanism

The pressure oxidation of pyrite (rest potential 0.62 V [22])has been well established to be electrochemically controlled [9].Similarly the pressure leaching of chalcopyrite (rest potential0.5 V [23]) has been interpreted as an electrochemically con-trolled process [24]. Arsenopyrite is also a semiconducting min-eral (rest potential = 0.58 [25]) varying from p-type to n-type[26]. The behaviour of arsenopyrite as an electrocatalyst foroxygen reduction has been shown to be similar to that of othersulphide minerals such as pyrite [27]. It is therefore a reasonableassumption that its oxidation mechanism is of electrochemicalnature in a similar fashion with that of pyrite.

The surface of arsenopyrite is envisaged to consist of anodicand cathodic sites on which the following two principal reac-tions* are believed to occur

FeAsS(s) +8H20(l) ~ Fe2+ (aq) +H3As04(aq)

+ SO~- (aq) + 13H+ (aq) + 13e- (16)

02(aq) +4H+ (aq) +4e- ~ 2H20(l). (17)

The cathodic reduction of dissolved molecular oxygen at themineral surface according to reaction 17 is considered the keyprocess affecting the overall reaction kinetics. Among the sev-eral reaction paths proposed previously to account for theelectro catalytic reduction of O2 [28, 29] the following is pos-tulated as representing the present system

1. Chemisorption of O2

02(aq) + 2FeAsS(s) ~ 2FeAsS(0)ads

2. First electron-transfer

2(FeAsS(0)ads + H+ +e- ~ FeAsS(OH)ads)

3. Second electron-transfer

2(FeAsS(OH)ads+H+ +e~ ~FeAsS+H20) (20)

*Only the sulphate-forming anodic reaction, which dominates athigh temperatures, is considered here; the possible influence of theFe 3+ /Fe2+ couple on the overall mechanism via the alteration of themixed potential on the mineral surface is neglected.

Table 1. The dependence of rate con-stant ks on H+ concentration at 130aC

(In ks = 0.3 In [H+])

[H2SO4] [H+] ks(N) (M) min-1

0.1 0.05 1.40 x 10-3

0.25 0.125 1.94 x 10-3

0.5 0.25 2.31xl0-3

The first electron transfer step (reaction 19) is taken as therate determining step. This is in agreement with the positiveeffect of H+ concentration on the oxidation kinetics (see Table1). Alternative mechanisms call for negative or no effect at allas it has been shown in the case of low pressure (< 1 atm) pyritesystems [20, 30]. On the other hand in high pressure leachingsystems involving pyrite [9] and chalcopyrite [24] a similar effectwith the present study has been found. The rate of reaction 19with the aid of the Butler-Volmer equation (high field approxi-mation) [29] can be expressed as follows

. = -4Fk 8[H+] -acFI1eIe 2 exp fYlT . (21)

Similarly for the anodic oxidation of FeAsS (reaction 16) it canbe shown that

(22)

where

ie' iacathodic and anodic current densitiesk2 and ka forward rate constants for cathodic (19) and anodic(16) reactions.8 fraction of arsenopyrite surface sites covered by dissolvedoxygen (surface coverage by intermediates is neglected).[H+] concentration of H+ae, aa transfer coefficients for the forward cathodic and anodicreactions respectivelyl1e, l1a cathodic and anodic potentialsF, f!lt, and T Faraday constant, gas constant and temperaturerespectively.

At the mixed potentiall1m, l1m = l1e = l1a, the followingequation applies (since reactions 16 and 17 do not carry thesame number of electrons)

4ia = -13ie

which yields

(18)(23)

(19) Applying equation 23 to equation 22 gives

. _ (k28[H+])aa/aa+acIa - 13Fka k .

a(24)

Assuming aa = ac, the last equation becomes

(25)

Equation 25 suggests the reaction rate to have half-order depen-


dency on the hydrogen ion concentration and on the surfaceconcentration of molecular oxygen (8). The experimentallydetermined orders were found to be 0.3 for [H+] and 0.9 forp02' Different adsorption isothermes (Langmuir, Temkin,Freundlich) are available to describe the relationship between8 and the applied partial oxygen pressure [31, 32]. A Freundlichtype isotherm of the form 8 = kjP6, which reduces to first orderbehaviour (eqn 25) may be assum-ed to apply. Given that noinformation on the exact form of this isotherm is available forthe present high temperature three phase system and that thecomparison with experiment ignores the contribution from theminor reaction path 2 and that a number of simplifying assump-tions are inherent to the electrochemical kinetic theory no fur-ther analysis is attempted. However the proposed mechanismoffers satisfactory explanation as to the effect of H+ on theoverall oxidation kinetics of arsenopyrite and the measuredactivation energies in the range of 66-72 kJ mol- I furthersupport the postulated activation control.

CONCLUSIONS

The pressure oxidation kinetics of arsenopyrite has beenstudied in the temperature range 120-180C and 2-20 atm O2pressure. The oxidation kinetics was found to follow a shrinkingcore model with the surface chemical reaction, as the rate con-trolling step. Neither the liquid elemental sulphur that formsto a certain degree nor the precipitation of scorodite whichoccurs with high slurry densities appear to interfere with theprogress of the heterogeneous oxidation process. The activationenergy was determined to be in the range of 66 (triclinic arsen-opyrite) to 72 (monoclinic arsenopyrite) kJ mol-I for the wholetemperature range investigated. The reaction order with respectto oxygen partial pressure was found to be 1 by consideringboth initial and overall kinetics. The acid concentration wasalso found to have a beneficial effect on the oxidation kinetics.Finally normalization plots of the shrinking core model in theform of ro[1- (1-R) 1/3] vs time resulted in one straight linefor various particle size fractions, thus further supporting theconclusion that the oxidation process is chemically controlled.

Having determined the activation energy and reaction orderin terms of oxygen pressure the following intrinsic rate equationwas developed, based on the surface reaction control model, torepresent the aqueous pressure oxidation of arsenopyrite in abatch reactor (sulphate-forming reaction path 1).

dN (-8672) .- S dt = 49.527 exp --T- P02mol mIn-I cm-2

where N is the number of moles of arsenopyrite, t is the time,S the total surface area of FeAsS particles, and P02 the partialoxygen pressure in the reactor.

Finally an electrochemical reaction mechanism has beenadvanced involving chemisorption of oxygen on arsenopyritesurface followed by two single-electron transfer steps. The firstelectron transfer is postulated as being the rate determiningstep.

Acknowledgement-This work was supported by an NSERC OperatingGrant. A McGill University scholarship awarded to one ofus (V.G.P.)is also gratefully acknowledged.

REFERENCES

1. G. P. Demopoulos and V. G. Papangelakis, Proc. Int. Symp. onGold 1\1etalllirgy (edited by R. S. SaIter, D. M. \Vyslowzil and G.\V. McDonald), p. 341, Pergamon Press, Oxford (1987).

2. R. T. Lowson, Chem. Rev. 82,461 (1982).3. J. B. Hiskey and W. J. Schlitt, Imelfacing Technologies ill Sollition

A1ining (edited by W. J. Schlitt and J. B. Hiskey), p. 55, SME-AIME, Littleton, Co. (1982).

4. V. G. Papangelakis and G. P. Demopoulos, Part I, Can. metall. Q.29, I (1990).

5. V. G. Papangelakis, D. Berk and G. P. Demopoulos, Hydro-metallurgical Reactor Design and Kinetics (edited by R. G. Bautista,R. J. Wesley and G. W. Warren), p. 204, Tl\1S-AIME, \Varrendale,PA (1986).

6. V. G. Papangelakis, D. Berk and G. P. Demopoulos, J\1etal. Trans.B (1990) in press.

7. D. R. McKay and J. Halpern, Trans. T,US-AlJ\1E6, 301 (1958).8. J. Gerlach, H. Hahne and F. Pawlek. Er=metallI9, 66 (1966).9. L. K. Bailey and E. Peters, Can. metall. Q. 15, 333 (1976).ro. I. H. Warren, Aust. J. appl. Sci. 7, 346 (1956).11. R. J. Cornelius and J. T. \Voodcock, Proc. Aust. Inst. l\1in. J\1etal.

185,65 (1958).12. C. T. Mathews and R. G. Robins, Aust. Chem. Engng 15, 19(1974).13. W. E. King and J. A. Lewis, Ind. Engng Chem. Proc. Des. Del'. 19,

719 (1980).14. O. Levenspiel, Chemical Reaction Engineering, p. 357, John \Viley,

(1972).15. J. M. Smith, Chemical Engineering Kinetics, 3rd edn, p. 642,

McGraw-Hill, New York (1981).16. H. A. Pray, C. E. Schweickert and B. H. l\1innich. Ind. Engng

Chem.44, 1146 (1952).17. P. R. Taylor, M. de Matos and G. P. Martins, Aletall. Trans. B

148,49 (1983).18. N. B. Pilling and R. E. Bedworth, J. Inst. l\1etals 29,529 (1923).19. R. C. \Veast, Handbook of Chemistry alld Physics, 66th edn, p. B-

148, B-198, B-201, CRC Press, Cleveland, OH (1985).20. T. Biegler, D. A. J. Rand and R. \Voods, Electroanalyt. Chem.

Intelf Electl'oeil. 60, 151 (1975).21. W. L. Marshall and E. V. Jones, J. Pill'S. Chem. 70,4028 (1966).22. E. Peters and H. Majima, Can. metall: Q. 7, 111 (1968).23. G. W. Warren, M. E. \Vadsworth and S. M. El-Raghy, Hydro-

metallurgy: Research, Derelopment alld Plant Practice (edited byK. Osseo-Asare and J. D. Miller), p. 261,TMS-AIME, \Varrendale,PA (1983).

24. H. K. Lin, H. Y. Sohn and M. E. \Vadsworth, HydrometallurgicalReactor Design and Kinetics (edited by R. G. Bautista, R. J. \Vesleyand G. W. Warren), p. 149, TMS-AIME, \Varrendale. PA (1986).

25. M. J. V. Beattie and G. W. Poling, Int. J. of l\1in. Processing 20,87 (1987).

26. R. T. Shuey, Semiconducting Ore A1inerals p. 205, Elsevier, Amster-dam (1975).

27. T. Biegler, D. A. J. Rand and R. Woods, Trends in Electrochemistry(edited by J. O'M. Bockris, D. A. J. Rand and \VeIch). p. 291,Plenum Press, New York (1977).

28. A. Damianovic, Modern Aspects of Electrochemistry (edited by J.O'M. Bockris and B. E. Conway), No.5. p. 370, Plenum Press,New York (1969).

29. J. O'M. Bockris and A. K. N. Reddy, 1Uodern Electroc!lemistry. p.991 and 1251, Plenum Press, New York (1970).

30. T. Biegler and D. A. Swift, Electrocllim. Acta 24, 415 (1979).31. J. J. Carberry, Chemical and Catalytic Reaction Engineering, p.

378, McGraw-Hill, New York (1976).32. E. Gileadi and B. E. Conway, Modern Aspects of Electrochemistry

(edited by J. O'M. Bockris and B. E. Conway), No.3. p. 347,Plenum Press, New York (1964).

Acid Pressure Oxidation of Arsenopyrite

Documents

Transcript of Acid Pressure Oxidation of Arsenopyrite