- Email: [email protected]
Reactivity comparison of Australian chalcopyrite concentrates in acidified ferric solution
2 (1976/1977) 219-233 219 © Elsevier Scientific Publishing Company, Amsterdam -- Printed in The Netherlands
Hydrometallurgy,
R E A C T I V I T Y C O M P A R I S O N O F A U S T R A L I A N C H A L C O P Y R I T E CONC E N T R A T E S IN A C I D I F I E D F E R R I C SOLUTION*
tt.G. LINGE C S I R O Division o f Mineral Chemistry, P.O. B o x 124, Port Melbourne, Victoria 3 2 0 7 (A us tralia )
(Received August 2nd, 1976)
ABSTRACT Linge, H.G., 1977. Reactivity comparison of Australian chalcopyrite concentrates in acidified ferric solution. Hydrometallurgy, 2: 219--233. The leaching of chalcopyrite from several Australian chalcopyrite concentrates by the reaction CuFeS: + 4 Fe(III) * Cu(II) + 5 Fe(II) + 2 SO obeyed parabolic kinetics in acidified nitrate solution between 25 and 40 ° C: The chalcopyrite reactivity was dependent on the mineral composition of the concentrate : the presence of pyrite accelerated the reaction markedly, but sphalerite and bismuthinite slowed it slightly. Galvanic interaction between the minerals cannot account for this change: instead, the associated minerals must influence the rate determining diffusion of the lattice elements within the chalcopyrite crystal.
INTRODUCTION The leaching o f c h a l c o p y r i t e b y ferric ions in dilute acid s o l u t i o n is of i m p o r t a n c e in the n e w h y d r o m e t a l l u r g i c a l processes being d e v e l o p e d as alternatives t o existing smelting t e c h n o l o g y and for r e c o v e r y o f c o p p e r f r o m heaps and d u m p s o f processed waste a n d o t h e r l o w grade ore ( W o o d c o c k , 1 9 6 7 ; D u t r i z a c and M a c D o n a l d , 1974a). In the practical o p e r a t i o n o f these n e w processes, o n e o f the m o s t i m p o r t a n t factors is the variable mineral c o m p o s i t i o n o f the feed material. However, n o t h o r o u g h investigation has y e t been d o n e o f the e f f e c t o f this variability on the rate o f c h a l c o p y r i t e dissolution f r o m industrial c o n c e n t r a t e s . T h e r e is a g r o w i n g b o d y o f i n f o r m a t i o n t o suggest t h a t changes in the mineral c o m p o s i t i o n o f c o n c e n t r a t e s can m o d i f y the intrinsic reactivities o f the mineral c o n s t i t u e n t s present. H o w e v e r , the m e c h a n i s m s o f mineralogical i n t e r a c t i o n are n o t well u n d e r s t o o d (e.g. Prosser, 1970). I t has been p r o p o s e d (e.g. G o t t s c h a l k a n d Buehler, 1 9 1 2 ; Peters a n d Majima, 1 9 6 8 ; W a d s w o r t h , *Presented in part at the 5th National Convention of the Royal Australian Chemical Institute, Canberra, May 1974.
220 1972) that a reactivity dependence on mineral composition should arise in the aqueous oxidation of concentrates containing different metal sulphides, due to complex galvanic coupling of the electrochemical oxidation and reduction steps occurring at tile surfaces of each sulphide present, cf. aqueous corrosion of different metals in electrical contact. The sparse experimental results available {e.g. Gottschalk and Buehler, 1912; Ichikuni, 1960, 1962; Majima, 1969; Dutrizac et al., 1971) for the leaching of mixed sulphide ores seem to disagree. The magnitude of the effect cannot be calculated, and the validity of the galvanic coupling theory is not proven. In the present work we have examined the dissolution kinetics of several chalcopyrite concentrates from different Australian mines, to cover the likely practical reactivity range. The specific rate of chalcopyrite leaching was measured by potentiometric titration of the ferric ions consumed in the reaction, as previously described in detail (Linge, 1976), and the result correlated with the mineral composition of each concentrate. Whilst this work was in progress, a similar study, using sintered discs of synthetic chalcopyrite mixed with different metal sulphides, was published by Dutrizac and MacDonald (1973), who concluded that the galvanic coupling mechanism explained the significant chalcopyrite-reactivity changes observed. The present work establishes such changes, also, for the dissolution rate of chalcopyrite from concentrates; but the measured kinetics of the reaction show conclusively that another mechanism must be responsible for this reactivity change. EXPERIMENTAL Materials
The concentrates used in this study contained chalcopyrite in association with at least one other mineral sulphide. This is shown in Table 1, which lists the source, composition, and some properties of each material used. The mineral composition was determined by AMDEL (The Australian Mineral Development Laboratories, Adelaide, South Australia) from point counting in reflected light and electron probe analysis of composite mineral grains. The formulae shown in the table summarize the molar ratio of the minerals present in each concentrate at concentrations greater than 10% of the chalcopyrite content. In each sample the chalcopyrite was more than 90% liberated. The non-opaque minerals (designated collectively as NOP) were primarily quartz and smaller amounts of complex metal silicates and carbonates. Each concentrate composition was confirmed by appropriate chemical analysis for the major elements (using atomic absorption for metal values and the Leco (Laboratory Equipment Corporation, St. Joseph, Michigan 49085, U.S.A.) combustion technique for the total sulphur present). Unscreened samples (0.5--5 g) of concentrate, consisting of mineral frag-
Peko, Northern Territory
Moline, Northern Territory
Peko, Northern Territory
Mt Carrington, New South Wales
Juno, Northern Territory
Orlando, Northern Territory
2
3
4
5
6
Mine
1
Concentrate No.
Chalcopyrite concentrates
TABLE 1
(CuFeS2 )1.00 (FeS2)0.72 (CuS)o.16 (Fe304)0.13 (NOP)2.2
(CuFeS2)t .00 (Bi2S3)o.32 (FeS2)0.11 (Fe3 04 )0 .19 (NOP)0.7
(CuFeS2 )1.00 (FeS2)0.91 (ZnS)0.46 (NOP)o.s
(CuFeS2 )1.00 (ZnS)o . ~ ( F e S 2 )0.14 (CuS)o .1 (NOP)o .5
(CuFeS2)1 .oo (FeS2)0.34 (FeAsS--FeS2)o .14 (NOP)o.2
(CuFeS2)1.oo (FeS2)o.17 (Fe304)0.1 (NOP)o.1
Concentrate composition
1.3
1.6
1.0
1.1
0.8
1.2
Specific surface area (m2/g)
0.29
0.44
0.44
0.59
0.68
0.77
Mole area fraction as chalcopyrite (~) (calc.)
t-O t~
222 ments ranging from 1 to 100 t~m in size, were used for the individual dissolution experiments. Each sample was c o m p o u n d e d from three smaller lots, separately scooped from representative 50--100 g batches of the material, split by riffling from the kg samples that had been obtained from each mine. The specific surface area of each concentrate was determined by argon adsorption at --195°C in a conventional BET apparatus using the value 13.8 A2 for the adsorbate area (De Boer, 1970). All chemicals used were of A.R. quality. A good grade of demineralized water (specific conductance at 20°C about 0.5 t~mho cm--l), prepared as required in a commercial deionizer (Elgastat Type Bl14) was used throughout. Ferrous nitrate solution, used in making up the leachant, was prepared from fresh ferrous sulphate and barium nitrate solutions for each experiment. The concentrates were used in their "as received" condition except for a routine wash in acetone and ethyl ether to remove any organic surface contaminants (e.g. flotation agents) which, if present, would interfere with the potentiometric titration procedure used.
Dissolution experiments The titration m e t h o d and the apparatus used have been previously described in detail (Linge, 1976). Samples of concentrate were dissolved as slurries in a stirred reactor (capacity 250 ml) at constant temperature in deoxygenated ferric-ferrous nitrate mixtures (total [ Fe(III) ] / t o t a l [ Fe(II) ] = 0.1; total [Fe] ca. 10 - 3 M; solution oxidation potential E h ca. 680 mV SHE) in a 0.01 M HNO3 and 0.5 M NaNO3 buffer. The oxidant consumption in the mineral dissolution reactions was compensated by titrating oxygen-free ferric nitrate into the slurry, at a constant solution oxidation potential. The time variation of the ferric nitrate was used as a continuous in situ measure of the concentrate oxidation rate. An experiment was commenced by dispensing a known weight of sample from a plastic bucket, suspended inside the reactor, into the solution. Measurement of the ferric nitrate consumed was then begun. Presaturated oxygen-free nitrogen was passed continuously over the cell solution to prevent ingress of the atmosphere during a run. Stirring of the solution was normally at 1000 r.p.m., which suspended the sample uniformly in the solution. In some experiments, the stirring rate was doubled. Although chalcopyrite was always the principal copper bearing constituent in each concentrate, covellite (CuS) was also present in a few samples. This mineral was n o t expected to contribute significantly to the measured copper dissolution rates, since the published results of Dutrizac and MacDonald (1974b) show that covellite dissolves far more slowly than chalcopyrite in acidic ferric sulphate at moderate solution temperatures (e.g. up to 50 ° C). The solution copper concentration was therefore used as a specific measure of the a m o u n t of chalcopyrite which had dissolved from each sample. Small (e.g. 1--2 ml) filtered solution samples were taken from the vessel
223
at regular intervals during the reaction using a sintered-glass suction device, and the metal solution-content was analysed by atomic absorption spectroscopy. RESULTS AND DISCUSSION
Oxidation rates and stoichiometries
Figures l(a) and l(b) show plots of the time variation of the ferric ion consumption (U mol) for two of the concentrates (Nos 1 and 5) at early (e.g. to 5 h) and later (e.g. to 1 week) stages of the leaching reaction. The rate of oxidation (A U/At) always decreased with time, varying rapidly by about one to two orders of magnitude in the first hour of reaction, but much more slowly later on in the experiment. The shape of the plot for concentrate No. 1 was representative of all the samples (except No. 5). For No. 5 concentrate the rate of oxidation remained constant at intermediate stages of the dissolution. This is evident in Fig. l(b). Runs 10, 12 and 14, 15 illustrate the good reproducibility of the experiments. The U--t characteristics agree to better than + 5% (typical of all the concentrates), and this proves that the sampling procedure gave representative sub-batches.
/ 25
NoI
20
/No
I
}<
D
/
,~A
c
~ioI , /
08
/ 5f _ /
0G _
~......~---------"--~
2 39, 25"C
E ~04 % No5
02 •
-----"---~-o
J , ,
5'0
'
•
os'/Aos
Not
o.5~,25"c (o) i
,oo
J
200 0 Time /mm
2000
4000
G000
8000
10,000
Fig. 1. Ferrie ion uptake vs. time for eoneentrates Nos 1 and 5, showing the effect of changes in temperature, sample weight, and solution stirring, and illustratingreprodueibility. (a) At early stage (e.g. to 5 h), (b) at later stage (e.g. to I week), of the reaction. For explanation o f A, B, C, D s e e t e x t . • Run 20; o Run 10; • Run 12; • Run 14; ~ R u n 15; ~ R u n 39 (2000 rpm); x R u n 30.
224
Each concentrate oxidized more rapidly when sample weight and temperature were increased; cf. runs 20, 15 and 30 (Figure 1). The reaction rate differed for each sample. The results for concentrate 1 were representative of the highest values measured, but these showed no dependence on solution agitation. The stirring rate was doubled from 1000 rpm for experiment 39 (Fig. 1) and at point A later on in run 14, with reversal at B and a similar cycle at C and D. This did n o t change the rate of the reaction. This implies that the limiting mass transfer rate of ferric ion in the h y d r o d y n a m i c boundary layer was never reached in the reaction vessel at any stage in the dissolution. The reactivity of the crystal surface, therefore, exerted its m a x i m u m control on the measured rates of oxidation. The dependence of the a m o u n t of copper dissolved from a sample (Cu d) on the equivalent oxidant consumption was examined in detail so that a specific measurement for the rate of chalcopyrite dissolution could be obtained from the values forA U/At. No more than 5--10% of the total copper available for the reaction was dissolved in an individual experiment. The a m o u n t of chalcopyrite available for leaching remained, therefore, effectively unchanged during a run. The shape of the plots of Cu d against U were similar for all concentrates, and Figs. 2(a) and 2(b) show typical results. There were always two successive reaction stages, each being characterized by a line of constant slope. The initial slope (ACud/A U) was about an order of magnitude greater than t h a t for the subsequent reaction, and the point of changeover from the initial to the final value was proportional to the weight (and hence the surface area) of the concentrate used.
15I
/z~z~
10
/
o
05
°'
0
I
2
3 fO3U/ m0L
//
•
1
2
3
Fig. 2. E x a m p l e s o f r e l a t i o n s h i p o f c o p p e r release to o x i d a n t c o n s u m p t i o n , a n d i n f l u e n c e of changes in s a m p l e w e i g h t o n t h e p o i n t o f c h a n g e o f t h e c o p p e r - s t o i c h i o m e t r y . All results a t 250C, E h ~ 6 8 0 m V SHE, a n d p H 2 in n i t r a t e solution. (a) C o n c e n t r a t e No. 1;
(b) No. 3.
225 The experimental values of ACUd/A U for the final reaction stage varied from 0.04 to 0.17. These data are consistent with the expected dissolution reaction CuFeS2 + 4 Fe(III) ÷ Cu(II) + 5 Fe(II) + 2 So
(1)
governing the long-term oxidation of chalcopyrite in acidic ferric nitrate. The present values are lower than the calculated result (Linge, 1976, equation (2) } ACud/A U= 0.22 for equation (1), because of the simultaneous oxidant consumption by the other minerals present in each concentrate; but, as expected, this deviation was least for the concentrates of higher chalcopyrite grade (e.g. Fig. 2, slope for No. 1 = 0.17; for No. 3 = 0.08). The enhanced copper dissolution at the start of the reaction was far more pronounced for the concentrates than for the pure chalcopyrite mineral (Linge, 1976). The congruent dissolution (reaction ( 1 ) ) of pure chalcopyrite was preceded by the formation of a metal deficient "chalcopyrite-like" phase. In this reaction, about 50% more copper was dissolved per mole of oxidant consumed than in the subsequent reaction (1). The ten-fold difference netween the initial and final values of ACUd/A U for the concentrates is not consistent with this initial reaction. Instead, these larger values imply that the change in the oxidation state of copper during the initial stage of concentrate dissolution was only marginal (e.g. less than 1/10th of the change in reaction (1), i.e. less than +0.2). This indicates that preoxidized copper mineral was present at the concentrate surface at the start of the leaching reaction. This would have formed by oxidation during the production of each concentrate, to an extent dependent on the intrinsic reactivity of chalcopyrite in the original ore deposit; and it would be unaffected by the pretreatment we applied before the dissolution tests, but would dissolve in acid present in the leaching mixture. Acid pre-leaching of a few concentrates confirmed this reaction. The initial rates of copper dissolution, therefore, did not relate to the chalcopyrite reaction rate. Consequently, these values could n o t be used in the chalcopyrite -reactivity assessment. The initial oxidation of concentrates preleached in acid was n o t separately examined. The solution concentrations of the principal lattice constituents in each of the concentrates were also analysed, but a closed analysis of reaction stoichiometries and mass balances was n o t attempted. Iron and zinc always dissolved freely; bismuth (in No. 5) reacted similarly, except at 25 ° C, when a bismuth precipitate formed at intermediate reaction time. This reaction correlated with the additional linear segment found in the U--t plots (Fig. 1) for concentrate No. 5. The nature of this kinetic change was n o t investigated in detail, since there was no interference to reaction (1) at later times, when the rates of copper dissolution were a usable measure of the reactivity of chalcopyrite in each sample. Both elemental sulphur (shown by carbon disulphide leaching of reaction residues) and sulphate ions (shown by BaSO4 precipitation from the solution) also formed in the reaction.
226
1"5
No4
10 o/O/°~°/
m/= ~
~
o
/o /
~
'
So
I
~
0 '5
No Z _Xx.XX~X---''~
~x ~ ~x x ~ × ~
0
No 3
50
100
Fig. 3. Parabolic leaching kinetics of chalcopyrite for each concentrate (numbered on curves). All results at 25 ° C and 1000 rpm for 2.3 g of sample.
Chalcopyrite reactivities The copper dissolution kinetics for each concentrate were obtained from the measured time-dependence of the ferric ion consumption, converted to an equivalent measure of the a m o u n t of copper dissolved by interpolation of the appropriate plot of Cu d against U. These data obeyed the parabolic leaching kinetics previously observed for pure chalcopyrite when equation (1) applied. This is shown in Fig. 3 which gives a representative plot of Cu d against t 1/2 for each concentrate for the same experimental conditions. The slope of the linear segment was different for each concentrate, but in each case, was proportional to the weight (and hence the surface area) of sample used, with an activation energy (Ea) similar to that for pure chalcopyrite (14 kcal mol -~) (Linge, 1976) in the same temperature range. Figure 4 shows this for one material. Table 2 summarizes all the data in the form of the slope (S A) per unit area of concentrate and the derived values for E a. The uniform fit of these results by the same rate law and activation energy shows that the mechanism of reaction (1) is the same for each sample. The rate of the reaction, however, is dependent on the mineral environment of chalcopyrite. The S A values are influenced by both the a m o u n t and the intrinsic reactivity of the chalcopyrite in the sample. The proportionality of the observed
227 15
3 , 35°C
/
Io
~"
S
-g E
/
2
t_)
J
/oj /
%
/
o-5
xx~X~
Ix 0'1[~ 0
x~x
loci, 25 C
~x"-'-"
"~x""'x
j~X -''~ X~X~x~ ~"
~ I
n I
~ L
0.59' 25% I
50
I
I
I
I
I
I00
T,r~e & / r, io& Fig. 4. Effect of changes in temperature and sample weight on parabolic leaching kinetics, for concentrate No. 6.
dissolution rates on sample weight proves that these two factors can be separated by an equation of the form S O = SA/¢ (2) where ¢ is the proportion of the total concentrate surface occupied by chalcopyrite, and S O is the rate constant for reaction (1) per unit area of chalcopyrite. Equation (2) compensates for the surface area dilution due to the physical presence of the other minerals in the sample. S O measures the intrinsic chalcopyrite reactivity; it should have the same constant value, independent of the chalcopyrite source, unless specific interactions with the mineral environment separately modify the rate of chalcopyrite oxidation. ¢ must reflect the average effect of the complex distribution of chalcopyrite surface patches. There is no well-defined procedure for computing this from the mineralogical properties of the sample. We have obtained ¢ (Table 1) from the molar area fraction of chalcopyrite, assuming that the average mineral composition at the concentrate surface was the same as in the bulk when reaction (1) applied. Thus dp = n V2m/3/ E n V m 2/3
228 TABLE 2 Parabolic dissolution parameters for chalcopyrite leaching Concentrate No.
1011 SA (mol cm-2 s -~/~)
Ea (kcal tool ~)
1 2 3 4 5 6
1.13 1.18 0.61 3.92 0.53 2.28
14 12 11 14 9 12
where V m is the molar volume and n the number of moles of each mineral appearing in the formula for each concentrate (as given in Table 1). Implicit in the calculation is the assumption that the crystal shape factor is the same for all minerals present. However, the approximation is n o t considered to have had significant effect on the dissolution results: even the more drastic assumption t h a t each mineral has the same molar area (obtained by equating ¢ to the mole fraction n/En of chalcopyrite in the concentrate ) depressed each value by less than 10%, w i t h o u t any effect on the ranking of the dissolution response of each concentrate. The values calculated for S O are listed in Table 3. The uncertainty in these data is estimated to be less than 10%, therefore the changes indicated in Table 3 are significant. The reactivities of the samples presently investigated differ by up to almost an order of magnitude, indicating that mineral composition exerts an influence on the intrinsic oxidizability of chalcopyrite in a ferric-ion leach. The previous result (Linge, 1976) for a pure chalcopyrite specimen is also shown in Table 3 for comparison, and lies within the range of the present data. This gives support to the procedure used for evaluating ¢ . Comparison of the mineral c o n t e n t of each sample with the results for S o in Table 3 shows t h a t the pyrite-rich concentrates (Nos 4 and 6) are the most reactive (e.g. about 10 times as reactive as pure chalcopyrite). The samples that contain smaller quantities of pyrite but major amounts of another sulphide mineral (Nos 3 and 5) are the least reactive (e.g. up to 25% less reactive than pure chalcopyrite), whilst the materials containing smaller quantities of pyrite with, or without, a similar a m o u n t of another sulphide (Nos 1 and 2) have intermediate reactivity (e.g. up to 25% more reactive than pure chalcopyrite). This reactivity sequence bears a resemblance to the results of Dutrizac and MacDonald (1973) for the rate of leaching of synthetic chalcopyrite from sintered discs of different sulphide mixtures in acidified ferric sulphate solution. The published plots of Cu d against t have a similar shape to the present
229 TABLE 3 Chalcopyrite reactivities for the parabolic (S o ) and the initial (TO) dissolution response Concentrate No.
1012 S O (mol cm--2 s ~2)
109 T (mol 0m--2)
1 2 3 4 5 6 pure
15 17 10 89 12 79 14a
26 22 9 64 7 42
aChalcopyrite reactivity of a pure sample for the same experimental conditions (Linge, 1976). results, b u t the dissolution response was evaluated f r o m the initial slope of a q u a d r a t i c e q u a t i o n in t, which was fitted t o the data for each disc b y a least-square analysis. This s h o w e d t h a t pyrite, as well as m o l y b d e n i t e , enhanced the dissolution rate. Galena and i m p u r e (high iron) sphalerite r e t a r d e d the r e a c t i o n , whilst the p r e s e n c e o f p u r e r f o r m s o f sphalerite caused erratic behaviour, raising as well as lowering the leaching rate. T h e p r e s e n t results c o n f i r m the beneficial e f f e c t o f p y r i t e f o r the leaching o f c h a l c o p y r i t e concentrates, b u t the sphalerite in o u r sample, and also b i s m u t h i n i t e , like galena, slightly r e t a r d e d the leaching reaction. T h e values o b t a i n e d f o r S O were f o u n d to c o r r e l a t e with the e x t e n t o f the initial c o p p e r dissolution which p r e c e d e d r e a c t i o n (1). This is s h o w n b y the values o f T O in Table 3. T O is the a m o u n t o f c o p p e r dissolved t o the instant of change in t h e c o p p e r s t o i c h i o m e t r y (ACUd/A U), expressed per real unit area o f c h a l c o p y r i t e in each sample. T h e values for ¢ used in this calculation were the same as in e q u a t i o n (2) above. It is seen t h a t S O and T O give a similar r e a c t i v i t y sequence. This f u r t h e r c o n f i r m s t h a t the e x t e n t o f p r i o r oxidation o f each c o n c e n t r a t e is related to the r e a c t i v i t y o f the c h a l c o p y r i t e f r o m which the o x i d i z e d c o p p e r mineral was originally f o r m e d . Log T O is linearly related to S O. A l t h o u g h this relation c o u l d n o t be t h e o r e t i c a l l y explained, it d e m o n s t r a t e s t h a t the initial dissolution response o f each c o n c e n t r a t e can serve as a q u i c k c o m p a r a t i v e i n d i c a t o r o f the c h a l c o p y r i t e reactivity in an acid-ferric leach. This c o u l d be useful as a rapid and practical r e a c t i v i t y test for industrial application. Mechanism o f mineral in teraction D u t r i z a c and M a c D o n a l d ( 1 9 7 3 ) suggest t h a t the mineral i n t e r a c t i o n is a
230 consequence of the electrochemical nature of reaction (1). Chalcopyrite dissolves anodically, CuFeS2 -- 4 e + Cu(II) + Fe(II) + 2 S O
(3)
This part-reaction is depolarized by the reduction of ferric ions at suitable cathodic sites, Fe(III) + e ÷ Fe(II)
(4)
The electron transfer between the anodic and cathodic reaction sites is completed through the mineral phase. This is feasible for sulphide mixtures in physical contact because they generally have good electronic conduction. According to this explanation the secondary sulphide present in the mixture changes the chalcopyrite dissolution rate by influencing the rate of ferric ion reduction. A single sulphide mineral in acidified ferric solution has an electrode potential less noble than the equilibrium value for the Fe(III)/ Fe(II) redox couple, symptomatic of the rate of mineral leaching by reaction (4). The relative values for this potential determine the mineral influence. Electrical contact to a sulphide with a higher potential raises the chalcopyrite potential, and this raises the dissolution rate. Conversely, coupling to a site of lower potential diminishes the dissolution rate. Values for sulphide electrode potentials in acidic ferric solutions have n o t been published. Such measurements are available for dilute acid in contact with the atmosphere. These data, which presumably indicate the relative extent of corrosion of the mineral by oxygen gas dissolved in small amounts in the solution, have instead been used in the argument. On this scale, chalcopyrite has the highest electrode potential, except for that of pyrite. The theory therefore predicts an accelerated chalcopyrite dissolution in the presence of pyrite, and a reduced rate for all other mixtures, apparently in accord with most of Dutrizac's (Dutrizac and MacDonald, 1973) and also with the present results. This electrochemical explanation implies that the charge transfer across the mineral/solution interface limits the rate of the reaction. Any influence on the passage of current can then change the dissolution rate. This mechanism predicts (e.g. Habashi, 1966; Wadsworth, 1972) linear dissolution kinetics (i.e. Cu d ~ t) for reaction (1), and it is therefore not consistent with the parabolic kinetics experimentally observed. This parabolic form of rate law matches previous results (Dutrizac et al., 1969; Baur et al., 1972; Wadsworth, 1972; Dutrizac and MacDonald, 1974a; Linge, 1976) for the leaching of pure chalcopyrite by ferric ions in both sulphate and nitrate solutions at low pH. The only satisfactory conclusion is that another step must control reaction (1). This implies that any change in the rates of charge flow across the mineral/ solution interface cannot explain the observed dependence of S O on mineral composition. The mineral interaction must therefore arise by a different mechanism. There are two rate-controlling steps consistent with the observed parabolic
231
kinetics. These are (a) oxidant diffusion in the solution-filled pores in the sulphur layer which is formed during the reaction, and (b) solid-state diffusional processes within the chalcopyrite crystal involving the lattice elements. The calculated pore diffusion rates for the present experimental conditions are about four orders of magnitude larger than the measured dissolution rates, and, therefore, this mechanism cannot control the leaching reaction (Linge, 1976). Calculations based on the solid-state diffusion mechanism are in much better agreement with the results. This is shown below. Extrapolation of recent tracer measurements (Chen and Harvey, 1975) of the 64 Cu and s9 Fe self-diffusion coefficients in chalcopyrite at temperatures above 100°C by an Arrhenius plot shows that the cation diffusion coefficients are in the range 1 0 - 1 6 --10 -~s cm 2 s-1 at room temperature. Chen and Harvey suggest that the metal atoms move from occupied tetrahedral lattice positions into adjacent octahedral interstices as positive ions, leaving their covalently shared electrons behind, and that this process continues via unoccupied tetrahedral and octahedral interstices within the chalcopyrite lattice. There is no information on the corresponding rates of sulphur atom (or ion) diffusion, but they are probably lower than the cation diffusion flux since the sulphur sublattice in chalcopyrite is close-packed (Craig and Scott, 1974). The m a x i m u m chalcopyrite concentration difference between the surface and crystal bulk is of the order of Vm-1, where Vm is the molar volume. The m a x i m u m flux (]') of the lattice elements capable of sustaining the reaction at the crystal surface is, therefore (Jost, 1960), j = 0.5 A V m ~ ( D / ~ t ) 1/2
(5)
where A is the chalcopyrite surface area and D the diffusion coefficient. For rate control by such a step, I jl = A(Co d ) / A t . Experimentally (Figs. 3 and 4), Cu d = S o A t 1/2, with S O as defined in equation (2). Hence: S O = ( D / n ) 1/2 Vm~
(6)
This predicts a m a x i m u m value (using D ~ 10 -15 cm 2 s-1 and Vm ~ 50 cm 3 ) of about 4 × 10 -1° for S 0. This upper limit for the dissolution rate satisfies the measured results in Table 3, which are about an order of magnitude lower. The nature of the diffusing species controlling the rate cannot be deduced from this result, but the value suggests t h a t lattice elements diffusing at a lower rate than the Cu and Fe cations present in the structure might be responsible. This is in accord with the idea that the diffusion of sulphur atoms, in complex association with the holes injected during the reaction, is the rate controlling step (Linge, 1976). Since solid-state diffusion is the best interpretation for the parabolic dissolution kinetics of chalcopyrite, the mineral interaction must be couched in terms of a mechanism influencing this particular rate-determining step. The variation of S O with concentrate composition implies (eqn. ( 6 ) ) that the minerals associated with chalcopyrite in the concentrates change the dis-
232
solution rate by altering the diffusivity or concentration (or both) of the rate-controlling species within the chalcopyrite lattice. The detailed operation of this interaction cannot be established from the present results, since the identification of the rate-controlling species is still uncertain. However, the galvanic interaction previously described (Dutrizac and MacDonald, 1973) cannot account for the observed change. ACKNOWLEDGEMENTS
The author is particularly grateful to Dr. G.M. Lukaszewski for the gift of all samples used; for commissioning the Modal analyses from AMDEL; and for his constructive advice throughout this work. He is indebted to Ralph Tyler for the BET measurements and to Phil Strode for the chemical analyses.
REFERENCES Baur, J.P., Gibbs, H.L. and Wadsworth, M.E., 1972. Initial stage sulphuric acid leaching kinetics of chalcopyrite using radiochemical techniques. Paper 72-B-96, AIME Annual Meeting, San Francisco, February 1972. De Boer, J.H., 1970. In: D.H. Everett and R.H. Ottewill (Editors), Surface Area Determination. Butterworth, London, p. 10. Chen, J.H. and Harvey, W.W., 1975. Cation self-diffusion in chalcopyrite and pyrite. Metallurgical Transactions (B), 6B: 331--339. Craig, J.R. and Scott, S.D., 1974. In: P.H. Ribbe (Editor), Sulfide Mineralogy. Mineralogical Society of America (Short Course Notes, November 1974, Vol. 1, Chapter 5). Dutrizac, J.E. and MacDonald, R.J.C., 1973. The effect of some impurities on the rate of chalcopyrite dissolution. Canadian Metallurgical Quarterly, 12(4): 409--420. Dutrizac, J.E. and MacDonald, R.J.C., 1974a. Ferric ion as a leaching medium. Minerals Science and Engineering, 6(2): 59--100. Dutrizac, J.E. and MacDonald, R.J.C., 1974b. The kinetics of dissolution of covellite in acidic ferric sulphate solutions. Canadian Metallurgical Quarterly, 13(3): 423--433. Dutrizac, J.E., MacDonald, R.J.C. and Ingraham, T.R., 1969. The kinetics of the dissolution of synthetic chalcopyrite in aqueous acidic ferric sulphate solution. Transactions of The Metallurgical Society of AIME, 245: 955--959. Dutrizac, J.E., MacDonald, R.J.C., and Ingraham, T.R., 1971. Effect of pyrite, chalcopyrite and digenite on the rate of bornite dissolution in acidic ferric sulphate solutions. Canadian Metallurgical Quarterly, 10(1 ): 3--7. Gottschalk, V.I-1. and Buehler, H.A., 1912. Oxidation of sulphides. Economic Geology, 8: 15--34. Habashi, F., 1966. The mechanism of oxidation of sulfide ores in nature, Economic Geology, 61: 587--591. Ichikuni, M., 1960. The dissolution of sulphide minerals in various media. III. Factors intervening in the dissolution of chalcopyrite. Bulletin of the Chemical Society of Japan, 33: 1159--1162. Ichikuni, M., 1962. The action of ferric ions on chalcopyrite. Bulletin of the Chemical Society of Japan, 35: 1765--1768. Jost, W., 1960. Diffusion in Solids, Liquids, Gases. 3rd edn. Academic Press, New York, Chapter 1. Linge, H.G., 1976. A study of chalcopyrite dissolution in acidic ferric nitrate by potentiometric titration. Hydrometallurgy, 2(1): 51--64.
233
Majima, H., 1969. How oxidation affects selective flotation of complex sulfide ores. Canadian Metallurgical Quarterly, 8(3): 269--273. Peters, E. and Majima, H., 1968. Electrochemical reactions of pyrite in acid perchlorate solutions. Canadian Metallurgical Quarterly, 7(3): 1 1 - 1 1 7 . Prosser, A.P., 1970. Influence of mineralogical factors on the rates of chemical reactions of minerals. In: M.J. Jones (Editor), Mineral Processing and Extractive Metallurgy. The Institution of Mining and Metallurgy, London, pp. 59--80. Wadsworth, M.E., 1972. Advances in the leaching of sulphide minerals. Minerals Science and Engineering, 4(4): 36--47. Woodcock, J.T., 1967. Copper waste dump leaching. Proceedings of the Australasian Institute of Mining and Metallurgy, No. 224: 47--66.
Hydrometallurgy,
R E A C T I V I T Y C O M P A R I S O N O F A U S T R A L I A N C H A L C O P Y R I T E CONC E N T R A T E S IN A C I D I F I E D F E R R I C SOLUTION*
tt.G. LINGE C S I R O Division o f Mineral Chemistry, P.O. B o x 124, Port Melbourne, Victoria 3 2 0 7 (A us tralia )
(Received August 2nd, 1976)
ABSTRACT Linge, H.G., 1977. Reactivity comparison of Australian chalcopyrite concentrates in acidified ferric solution. Hydrometallurgy, 2: 219--233. The leaching of chalcopyrite from several Australian chalcopyrite concentrates by the reaction CuFeS: + 4 Fe(III) * Cu(II) + 5 Fe(II) + 2 SO obeyed parabolic kinetics in acidified nitrate solution between 25 and 40 ° C: The chalcopyrite reactivity was dependent on the mineral composition of the concentrate : the presence of pyrite accelerated the reaction markedly, but sphalerite and bismuthinite slowed it slightly. Galvanic interaction between the minerals cannot account for this change: instead, the associated minerals must influence the rate determining diffusion of the lattice elements within the chalcopyrite crystal.
INTRODUCTION The leaching o f c h a l c o p y r i t e b y ferric ions in dilute acid s o l u t i o n is of i m p o r t a n c e in the n e w h y d r o m e t a l l u r g i c a l processes being d e v e l o p e d as alternatives t o existing smelting t e c h n o l o g y and for r e c o v e r y o f c o p p e r f r o m heaps and d u m p s o f processed waste a n d o t h e r l o w grade ore ( W o o d c o c k , 1 9 6 7 ; D u t r i z a c and M a c D o n a l d , 1974a). In the practical o p e r a t i o n o f these n e w processes, o n e o f the m o s t i m p o r t a n t factors is the variable mineral c o m p o s i t i o n o f the feed material. However, n o t h o r o u g h investigation has y e t been d o n e o f the e f f e c t o f this variability on the rate o f c h a l c o p y r i t e dissolution f r o m industrial c o n c e n t r a t e s . T h e r e is a g r o w i n g b o d y o f i n f o r m a t i o n t o suggest t h a t changes in the mineral c o m p o s i t i o n o f c o n c e n t r a t e s can m o d i f y the intrinsic reactivities o f the mineral c o n s t i t u e n t s present. H o w e v e r , the m e c h a n i s m s o f mineralogical i n t e r a c t i o n are n o t well u n d e r s t o o d (e.g. Prosser, 1970). I t has been p r o p o s e d (e.g. G o t t s c h a l k a n d Buehler, 1 9 1 2 ; Peters a n d Majima, 1 9 6 8 ; W a d s w o r t h , *Presented in part at the 5th National Convention of the Royal Australian Chemical Institute, Canberra, May 1974.
220 1972) that a reactivity dependence on mineral composition should arise in the aqueous oxidation of concentrates containing different metal sulphides, due to complex galvanic coupling of the electrochemical oxidation and reduction steps occurring at tile surfaces of each sulphide present, cf. aqueous corrosion of different metals in electrical contact. The sparse experimental results available {e.g. Gottschalk and Buehler, 1912; Ichikuni, 1960, 1962; Majima, 1969; Dutrizac et al., 1971) for the leaching of mixed sulphide ores seem to disagree. The magnitude of the effect cannot be calculated, and the validity of the galvanic coupling theory is not proven. In the present work we have examined the dissolution kinetics of several chalcopyrite concentrates from different Australian mines, to cover the likely practical reactivity range. The specific rate of chalcopyrite leaching was measured by potentiometric titration of the ferric ions consumed in the reaction, as previously described in detail (Linge, 1976), and the result correlated with the mineral composition of each concentrate. Whilst this work was in progress, a similar study, using sintered discs of synthetic chalcopyrite mixed with different metal sulphides, was published by Dutrizac and MacDonald (1973), who concluded that the galvanic coupling mechanism explained the significant chalcopyrite-reactivity changes observed. The present work establishes such changes, also, for the dissolution rate of chalcopyrite from concentrates; but the measured kinetics of the reaction show conclusively that another mechanism must be responsible for this reactivity change. EXPERIMENTAL Materials
The concentrates used in this study contained chalcopyrite in association with at least one other mineral sulphide. This is shown in Table 1, which lists the source, composition, and some properties of each material used. The mineral composition was determined by AMDEL (The Australian Mineral Development Laboratories, Adelaide, South Australia) from point counting in reflected light and electron probe analysis of composite mineral grains. The formulae shown in the table summarize the molar ratio of the minerals present in each concentrate at concentrations greater than 10% of the chalcopyrite content. In each sample the chalcopyrite was more than 90% liberated. The non-opaque minerals (designated collectively as NOP) were primarily quartz and smaller amounts of complex metal silicates and carbonates. Each concentrate composition was confirmed by appropriate chemical analysis for the major elements (using atomic absorption for metal values and the Leco (Laboratory Equipment Corporation, St. Joseph, Michigan 49085, U.S.A.) combustion technique for the total sulphur present). Unscreened samples (0.5--5 g) of concentrate, consisting of mineral frag-
Peko, Northern Territory
Moline, Northern Territory
Peko, Northern Territory
Mt Carrington, New South Wales
Juno, Northern Territory
Orlando, Northern Territory
2
3
4
5
6
Mine
1
Concentrate No.
Chalcopyrite concentrates
TABLE 1
(CuFeS2 )1.00 (FeS2)0.72 (CuS)o.16 (Fe304)0.13 (NOP)2.2
(CuFeS2)t .00 (Bi2S3)o.32 (FeS2)0.11 (Fe3 04 )0 .19 (NOP)0.7
(CuFeS2 )1.00 (FeS2)0.91 (ZnS)0.46 (NOP)o.s
(CuFeS2 )1.00 (ZnS)o . ~ ( F e S 2 )0.14 (CuS)o .1 (NOP)o .5
(CuFeS2)1 .oo (FeS2)0.34 (FeAsS--FeS2)o .14 (NOP)o.2
(CuFeS2)1.oo (FeS2)o.17 (Fe304)0.1 (NOP)o.1
Concentrate composition
1.3
1.6
1.0
1.1
0.8
1.2
Specific surface area (m2/g)
0.29
0.44
0.44
0.59
0.68
0.77
Mole area fraction as chalcopyrite (~) (calc.)
t-O t~
222 ments ranging from 1 to 100 t~m in size, were used for the individual dissolution experiments. Each sample was c o m p o u n d e d from three smaller lots, separately scooped from representative 50--100 g batches of the material, split by riffling from the kg samples that had been obtained from each mine. The specific surface area of each concentrate was determined by argon adsorption at --195°C in a conventional BET apparatus using the value 13.8 A2 for the adsorbate area (De Boer, 1970). All chemicals used were of A.R. quality. A good grade of demineralized water (specific conductance at 20°C about 0.5 t~mho cm--l), prepared as required in a commercial deionizer (Elgastat Type Bl14) was used throughout. Ferrous nitrate solution, used in making up the leachant, was prepared from fresh ferrous sulphate and barium nitrate solutions for each experiment. The concentrates were used in their "as received" condition except for a routine wash in acetone and ethyl ether to remove any organic surface contaminants (e.g. flotation agents) which, if present, would interfere with the potentiometric titration procedure used.
Dissolution experiments The titration m e t h o d and the apparatus used have been previously described in detail (Linge, 1976). Samples of concentrate were dissolved as slurries in a stirred reactor (capacity 250 ml) at constant temperature in deoxygenated ferric-ferrous nitrate mixtures (total [ Fe(III) ] / t o t a l [ Fe(II) ] = 0.1; total [Fe] ca. 10 - 3 M; solution oxidation potential E h ca. 680 mV SHE) in a 0.01 M HNO3 and 0.5 M NaNO3 buffer. The oxidant consumption in the mineral dissolution reactions was compensated by titrating oxygen-free ferric nitrate into the slurry, at a constant solution oxidation potential. The time variation of the ferric nitrate was used as a continuous in situ measure of the concentrate oxidation rate. An experiment was commenced by dispensing a known weight of sample from a plastic bucket, suspended inside the reactor, into the solution. Measurement of the ferric nitrate consumed was then begun. Presaturated oxygen-free nitrogen was passed continuously over the cell solution to prevent ingress of the atmosphere during a run. Stirring of the solution was normally at 1000 r.p.m., which suspended the sample uniformly in the solution. In some experiments, the stirring rate was doubled. Although chalcopyrite was always the principal copper bearing constituent in each concentrate, covellite (CuS) was also present in a few samples. This mineral was n o t expected to contribute significantly to the measured copper dissolution rates, since the published results of Dutrizac and MacDonald (1974b) show that covellite dissolves far more slowly than chalcopyrite in acidic ferric sulphate at moderate solution temperatures (e.g. up to 50 ° C). The solution copper concentration was therefore used as a specific measure of the a m o u n t of chalcopyrite which had dissolved from each sample. Small (e.g. 1--2 ml) filtered solution samples were taken from the vessel
223
at regular intervals during the reaction using a sintered-glass suction device, and the metal solution-content was analysed by atomic absorption spectroscopy. RESULTS AND DISCUSSION
Oxidation rates and stoichiometries
Figures l(a) and l(b) show plots of the time variation of the ferric ion consumption (U mol) for two of the concentrates (Nos 1 and 5) at early (e.g. to 5 h) and later (e.g. to 1 week) stages of the leaching reaction. The rate of oxidation (A U/At) always decreased with time, varying rapidly by about one to two orders of magnitude in the first hour of reaction, but much more slowly later on in the experiment. The shape of the plot for concentrate No. 1 was representative of all the samples (except No. 5). For No. 5 concentrate the rate of oxidation remained constant at intermediate stages of the dissolution. This is evident in Fig. l(b). Runs 10, 12 and 14, 15 illustrate the good reproducibility of the experiments. The U--t characteristics agree to better than + 5% (typical of all the concentrates), and this proves that the sampling procedure gave representative sub-batches.
/ 25
NoI
20
/No
I
}<
D
/
,~A
c
~ioI , /
08
/ 5f _ /
0G _
~......~---------"--~
2 39, 25"C
E ~04 % No5
02 •
-----"---~-o
J , ,
5'0
'
•
os'/Aos
Not
o.5~,25"c (o) i
,oo
J
200 0 Time /mm
2000
4000
G000
8000
10,000
Fig. 1. Ferrie ion uptake vs. time for eoneentrates Nos 1 and 5, showing the effect of changes in temperature, sample weight, and solution stirring, and illustratingreprodueibility. (a) At early stage (e.g. to 5 h), (b) at later stage (e.g. to I week), of the reaction. For explanation o f A, B, C, D s e e t e x t . • Run 20; o Run 10; • Run 12; • Run 14; ~ R u n 15; ~ R u n 39 (2000 rpm); x R u n 30.
224
Each concentrate oxidized more rapidly when sample weight and temperature were increased; cf. runs 20, 15 and 30 (Figure 1). The reaction rate differed for each sample. The results for concentrate 1 were representative of the highest values measured, but these showed no dependence on solution agitation. The stirring rate was doubled from 1000 rpm for experiment 39 (Fig. 1) and at point A later on in run 14, with reversal at B and a similar cycle at C and D. This did n o t change the rate of the reaction. This implies that the limiting mass transfer rate of ferric ion in the h y d r o d y n a m i c boundary layer was never reached in the reaction vessel at any stage in the dissolution. The reactivity of the crystal surface, therefore, exerted its m a x i m u m control on the measured rates of oxidation. The dependence of the a m o u n t of copper dissolved from a sample (Cu d) on the equivalent oxidant consumption was examined in detail so that a specific measurement for the rate of chalcopyrite dissolution could be obtained from the values forA U/At. No more than 5--10% of the total copper available for the reaction was dissolved in an individual experiment. The a m o u n t of chalcopyrite available for leaching remained, therefore, effectively unchanged during a run. The shape of the plots of Cu d against U were similar for all concentrates, and Figs. 2(a) and 2(b) show typical results. There were always two successive reaction stages, each being characterized by a line of constant slope. The initial slope (ACud/A U) was about an order of magnitude greater than t h a t for the subsequent reaction, and the point of changeover from the initial to the final value was proportional to the weight (and hence the surface area) of the concentrate used.
15I
/z~z~
10
/
o
05
°'
0
I
2
3 fO3U/ m0L
//
•
1
2
3
Fig. 2. E x a m p l e s o f r e l a t i o n s h i p o f c o p p e r release to o x i d a n t c o n s u m p t i o n , a n d i n f l u e n c e of changes in s a m p l e w e i g h t o n t h e p o i n t o f c h a n g e o f t h e c o p p e r - s t o i c h i o m e t r y . All results a t 250C, E h ~ 6 8 0 m V SHE, a n d p H 2 in n i t r a t e solution. (a) C o n c e n t r a t e No. 1;
(b) No. 3.
225 The experimental values of ACUd/A U for the final reaction stage varied from 0.04 to 0.17. These data are consistent with the expected dissolution reaction CuFeS2 + 4 Fe(III) ÷ Cu(II) + 5 Fe(II) + 2 So
(1)
governing the long-term oxidation of chalcopyrite in acidic ferric nitrate. The present values are lower than the calculated result (Linge, 1976, equation (2) } ACud/A U= 0.22 for equation (1), because of the simultaneous oxidant consumption by the other minerals present in each concentrate; but, as expected, this deviation was least for the concentrates of higher chalcopyrite grade (e.g. Fig. 2, slope for No. 1 = 0.17; for No. 3 = 0.08). The enhanced copper dissolution at the start of the reaction was far more pronounced for the concentrates than for the pure chalcopyrite mineral (Linge, 1976). The congruent dissolution (reaction ( 1 ) ) of pure chalcopyrite was preceded by the formation of a metal deficient "chalcopyrite-like" phase. In this reaction, about 50% more copper was dissolved per mole of oxidant consumed than in the subsequent reaction (1). The ten-fold difference netween the initial and final values of ACUd/A U for the concentrates is not consistent with this initial reaction. Instead, these larger values imply that the change in the oxidation state of copper during the initial stage of concentrate dissolution was only marginal (e.g. less than 1/10th of the change in reaction (1), i.e. less than +0.2). This indicates that preoxidized copper mineral was present at the concentrate surface at the start of the leaching reaction. This would have formed by oxidation during the production of each concentrate, to an extent dependent on the intrinsic reactivity of chalcopyrite in the original ore deposit; and it would be unaffected by the pretreatment we applied before the dissolution tests, but would dissolve in acid present in the leaching mixture. Acid pre-leaching of a few concentrates confirmed this reaction. The initial rates of copper dissolution, therefore, did not relate to the chalcopyrite reaction rate. Consequently, these values could n o t be used in the chalcopyrite -reactivity assessment. The initial oxidation of concentrates preleached in acid was n o t separately examined. The solution concentrations of the principal lattice constituents in each of the concentrates were also analysed, but a closed analysis of reaction stoichiometries and mass balances was n o t attempted. Iron and zinc always dissolved freely; bismuth (in No. 5) reacted similarly, except at 25 ° C, when a bismuth precipitate formed at intermediate reaction time. This reaction correlated with the additional linear segment found in the U--t plots (Fig. 1) for concentrate No. 5. The nature of this kinetic change was n o t investigated in detail, since there was no interference to reaction (1) at later times, when the rates of copper dissolution were a usable measure of the reactivity of chalcopyrite in each sample. Both elemental sulphur (shown by carbon disulphide leaching of reaction residues) and sulphate ions (shown by BaSO4 precipitation from the solution) also formed in the reaction.
226
1"5
No4
10 o/O/°~°/
m/= ~
~
o
/o /
~
'
So
I
~
0 '5
No Z _Xx.XX~X---''~
~x ~ ~x x ~ × ~
0
No 3
50
100
Fig. 3. Parabolic leaching kinetics of chalcopyrite for each concentrate (numbered on curves). All results at 25 ° C and 1000 rpm for 2.3 g of sample.
Chalcopyrite reactivities The copper dissolution kinetics for each concentrate were obtained from the measured time-dependence of the ferric ion consumption, converted to an equivalent measure of the a m o u n t of copper dissolved by interpolation of the appropriate plot of Cu d against U. These data obeyed the parabolic leaching kinetics previously observed for pure chalcopyrite when equation (1) applied. This is shown in Fig. 3 which gives a representative plot of Cu d against t 1/2 for each concentrate for the same experimental conditions. The slope of the linear segment was different for each concentrate, but in each case, was proportional to the weight (and hence the surface area) of sample used, with an activation energy (Ea) similar to that for pure chalcopyrite (14 kcal mol -~) (Linge, 1976) in the same temperature range. Figure 4 shows this for one material. Table 2 summarizes all the data in the form of the slope (S A) per unit area of concentrate and the derived values for E a. The uniform fit of these results by the same rate law and activation energy shows that the mechanism of reaction (1) is the same for each sample. The rate of the reaction, however, is dependent on the mineral environment of chalcopyrite. The S A values are influenced by both the a m o u n t and the intrinsic reactivity of the chalcopyrite in the sample. The proportionality of the observed
227 15
3 , 35°C
/
Io
~"
S
-g E
/
2
t_)
J
/oj /
%
/
o-5
xx~X~
Ix 0'1[~ 0
x~x
loci, 25 C
~x"-'-"
"~x""'x
j~X -''~ X~X~x~ ~"
~ I
n I
~ L
0.59' 25% I
50
I
I
I
I
I
I00
T,r~e & / r, io& Fig. 4. Effect of changes in temperature and sample weight on parabolic leaching kinetics, for concentrate No. 6.
dissolution rates on sample weight proves that these two factors can be separated by an equation of the form S O = SA/¢ (2) where ¢ is the proportion of the total concentrate surface occupied by chalcopyrite, and S O is the rate constant for reaction (1) per unit area of chalcopyrite. Equation (2) compensates for the surface area dilution due to the physical presence of the other minerals in the sample. S O measures the intrinsic chalcopyrite reactivity; it should have the same constant value, independent of the chalcopyrite source, unless specific interactions with the mineral environment separately modify the rate of chalcopyrite oxidation. ¢ must reflect the average effect of the complex distribution of chalcopyrite surface patches. There is no well-defined procedure for computing this from the mineralogical properties of the sample. We have obtained ¢ (Table 1) from the molar area fraction of chalcopyrite, assuming that the average mineral composition at the concentrate surface was the same as in the bulk when reaction (1) applied. Thus dp = n V2m/3/ E n V m 2/3
228 TABLE 2 Parabolic dissolution parameters for chalcopyrite leaching Concentrate No.
1011 SA (mol cm-2 s -~/~)
Ea (kcal tool ~)
1 2 3 4 5 6
1.13 1.18 0.61 3.92 0.53 2.28
14 12 11 14 9 12
where V m is the molar volume and n the number of moles of each mineral appearing in the formula for each concentrate (as given in Table 1). Implicit in the calculation is the assumption that the crystal shape factor is the same for all minerals present. However, the approximation is n o t considered to have had significant effect on the dissolution results: even the more drastic assumption t h a t each mineral has the same molar area (obtained by equating ¢ to the mole fraction n/En of chalcopyrite in the concentrate ) depressed each value by less than 10%, w i t h o u t any effect on the ranking of the dissolution response of each concentrate. The values calculated for S O are listed in Table 3. The uncertainty in these data is estimated to be less than 10%, therefore the changes indicated in Table 3 are significant. The reactivities of the samples presently investigated differ by up to almost an order of magnitude, indicating that mineral composition exerts an influence on the intrinsic oxidizability of chalcopyrite in a ferric-ion leach. The previous result (Linge, 1976) for a pure chalcopyrite specimen is also shown in Table 3 for comparison, and lies within the range of the present data. This gives support to the procedure used for evaluating ¢ . Comparison of the mineral c o n t e n t of each sample with the results for S o in Table 3 shows t h a t the pyrite-rich concentrates (Nos 4 and 6) are the most reactive (e.g. about 10 times as reactive as pure chalcopyrite). The samples that contain smaller quantities of pyrite but major amounts of another sulphide mineral (Nos 3 and 5) are the least reactive (e.g. up to 25% less reactive than pure chalcopyrite), whilst the materials containing smaller quantities of pyrite with, or without, a similar a m o u n t of another sulphide (Nos 1 and 2) have intermediate reactivity (e.g. up to 25% more reactive than pure chalcopyrite). This reactivity sequence bears a resemblance to the results of Dutrizac and MacDonald (1973) for the rate of leaching of synthetic chalcopyrite from sintered discs of different sulphide mixtures in acidified ferric sulphate solution. The published plots of Cu d against t have a similar shape to the present
229 TABLE 3 Chalcopyrite reactivities for the parabolic (S o ) and the initial (TO) dissolution response Concentrate No.
1012 S O (mol cm--2 s ~2)
109 T (mol 0m--2)
1 2 3 4 5 6 pure
15 17 10 89 12 79 14a
26 22 9 64 7 42
aChalcopyrite reactivity of a pure sample for the same experimental conditions (Linge, 1976). results, b u t the dissolution response was evaluated f r o m the initial slope of a q u a d r a t i c e q u a t i o n in t, which was fitted t o the data for each disc b y a least-square analysis. This s h o w e d t h a t pyrite, as well as m o l y b d e n i t e , enhanced the dissolution rate. Galena and i m p u r e (high iron) sphalerite r e t a r d e d the r e a c t i o n , whilst the p r e s e n c e o f p u r e r f o r m s o f sphalerite caused erratic behaviour, raising as well as lowering the leaching rate. T h e p r e s e n t results c o n f i r m the beneficial e f f e c t o f p y r i t e f o r the leaching o f c h a l c o p y r i t e concentrates, b u t the sphalerite in o u r sample, and also b i s m u t h i n i t e , like galena, slightly r e t a r d e d the leaching reaction. T h e values o b t a i n e d f o r S O were f o u n d to c o r r e l a t e with the e x t e n t o f the initial c o p p e r dissolution which p r e c e d e d r e a c t i o n (1). This is s h o w n b y the values o f T O in Table 3. T O is the a m o u n t o f c o p p e r dissolved t o the instant of change in t h e c o p p e r s t o i c h i o m e t r y (ACUd/A U), expressed per real unit area o f c h a l c o p y r i t e in each sample. T h e values for ¢ used in this calculation were the same as in e q u a t i o n (2) above. It is seen t h a t S O and T O give a similar r e a c t i v i t y sequence. This f u r t h e r c o n f i r m s t h a t the e x t e n t o f p r i o r oxidation o f each c o n c e n t r a t e is related to the r e a c t i v i t y o f the c h a l c o p y r i t e f r o m which the o x i d i z e d c o p p e r mineral was originally f o r m e d . Log T O is linearly related to S O. A l t h o u g h this relation c o u l d n o t be t h e o r e t i c a l l y explained, it d e m o n s t r a t e s t h a t the initial dissolution response o f each c o n c e n t r a t e can serve as a q u i c k c o m p a r a t i v e i n d i c a t o r o f the c h a l c o p y r i t e reactivity in an acid-ferric leach. This c o u l d be useful as a rapid and practical r e a c t i v i t y test for industrial application. Mechanism o f mineral in teraction D u t r i z a c and M a c D o n a l d ( 1 9 7 3 ) suggest t h a t the mineral i n t e r a c t i o n is a
230 consequence of the electrochemical nature of reaction (1). Chalcopyrite dissolves anodically, CuFeS2 -- 4 e + Cu(II) + Fe(II) + 2 S O
(3)
This part-reaction is depolarized by the reduction of ferric ions at suitable cathodic sites, Fe(III) + e ÷ Fe(II)
(4)
The electron transfer between the anodic and cathodic reaction sites is completed through the mineral phase. This is feasible for sulphide mixtures in physical contact because they generally have good electronic conduction. According to this explanation the secondary sulphide present in the mixture changes the chalcopyrite dissolution rate by influencing the rate of ferric ion reduction. A single sulphide mineral in acidified ferric solution has an electrode potential less noble than the equilibrium value for the Fe(III)/ Fe(II) redox couple, symptomatic of the rate of mineral leaching by reaction (4). The relative values for this potential determine the mineral influence. Electrical contact to a sulphide with a higher potential raises the chalcopyrite potential, and this raises the dissolution rate. Conversely, coupling to a site of lower potential diminishes the dissolution rate. Values for sulphide electrode potentials in acidic ferric solutions have n o t been published. Such measurements are available for dilute acid in contact with the atmosphere. These data, which presumably indicate the relative extent of corrosion of the mineral by oxygen gas dissolved in small amounts in the solution, have instead been used in the argument. On this scale, chalcopyrite has the highest electrode potential, except for that of pyrite. The theory therefore predicts an accelerated chalcopyrite dissolution in the presence of pyrite, and a reduced rate for all other mixtures, apparently in accord with most of Dutrizac's (Dutrizac and MacDonald, 1973) and also with the present results. This electrochemical explanation implies that the charge transfer across the mineral/solution interface limits the rate of the reaction. Any influence on the passage of current can then change the dissolution rate. This mechanism predicts (e.g. Habashi, 1966; Wadsworth, 1972) linear dissolution kinetics (i.e. Cu d ~ t) for reaction (1), and it is therefore not consistent with the parabolic kinetics experimentally observed. This parabolic form of rate law matches previous results (Dutrizac et al., 1969; Baur et al., 1972; Wadsworth, 1972; Dutrizac and MacDonald, 1974a; Linge, 1976) for the leaching of pure chalcopyrite by ferric ions in both sulphate and nitrate solutions at low pH. The only satisfactory conclusion is that another step must control reaction (1). This implies that any change in the rates of charge flow across the mineral/ solution interface cannot explain the observed dependence of S O on mineral composition. The mineral interaction must therefore arise by a different mechanism. There are two rate-controlling steps consistent with the observed parabolic
231
kinetics. These are (a) oxidant diffusion in the solution-filled pores in the sulphur layer which is formed during the reaction, and (b) solid-state diffusional processes within the chalcopyrite crystal involving the lattice elements. The calculated pore diffusion rates for the present experimental conditions are about four orders of magnitude larger than the measured dissolution rates, and, therefore, this mechanism cannot control the leaching reaction (Linge, 1976). Calculations based on the solid-state diffusion mechanism are in much better agreement with the results. This is shown below. Extrapolation of recent tracer measurements (Chen and Harvey, 1975) of the 64 Cu and s9 Fe self-diffusion coefficients in chalcopyrite at temperatures above 100°C by an Arrhenius plot shows that the cation diffusion coefficients are in the range 1 0 - 1 6 --10 -~s cm 2 s-1 at room temperature. Chen and Harvey suggest that the metal atoms move from occupied tetrahedral lattice positions into adjacent octahedral interstices as positive ions, leaving their covalently shared electrons behind, and that this process continues via unoccupied tetrahedral and octahedral interstices within the chalcopyrite lattice. There is no information on the corresponding rates of sulphur atom (or ion) diffusion, but they are probably lower than the cation diffusion flux since the sulphur sublattice in chalcopyrite is close-packed (Craig and Scott, 1974). The m a x i m u m chalcopyrite concentration difference between the surface and crystal bulk is of the order of Vm-1, where Vm is the molar volume. The m a x i m u m flux (]') of the lattice elements capable of sustaining the reaction at the crystal surface is, therefore (Jost, 1960), j = 0.5 A V m ~ ( D / ~ t ) 1/2
(5)
where A is the chalcopyrite surface area and D the diffusion coefficient. For rate control by such a step, I jl = A(Co d ) / A t . Experimentally (Figs. 3 and 4), Cu d = S o A t 1/2, with S O as defined in equation (2). Hence: S O = ( D / n ) 1/2 Vm~
(6)
This predicts a m a x i m u m value (using D ~ 10 -15 cm 2 s-1 and Vm ~ 50 cm 3 ) of about 4 × 10 -1° for S 0. This upper limit for the dissolution rate satisfies the measured results in Table 3, which are about an order of magnitude lower. The nature of the diffusing species controlling the rate cannot be deduced from this result, but the value suggests t h a t lattice elements diffusing at a lower rate than the Cu and Fe cations present in the structure might be responsible. This is in accord with the idea that the diffusion of sulphur atoms, in complex association with the holes injected during the reaction, is the rate controlling step (Linge, 1976). Since solid-state diffusion is the best interpretation for the parabolic dissolution kinetics of chalcopyrite, the mineral interaction must be couched in terms of a mechanism influencing this particular rate-determining step. The variation of S O with concentrate composition implies (eqn. ( 6 ) ) that the minerals associated with chalcopyrite in the concentrates change the dis-
232
solution rate by altering the diffusivity or concentration (or both) of the rate-controlling species within the chalcopyrite lattice. The detailed operation of this interaction cannot be established from the present results, since the identification of the rate-controlling species is still uncertain. However, the galvanic interaction previously described (Dutrizac and MacDonald, 1973) cannot account for the observed change. ACKNOWLEDGEMENTS
The author is particularly grateful to Dr. G.M. Lukaszewski for the gift of all samples used; for commissioning the Modal analyses from AMDEL; and for his constructive advice throughout this work. He is indebted to Ralph Tyler for the BET measurements and to Phil Strode for the chemical analyses.
REFERENCES Baur, J.P., Gibbs, H.L. and Wadsworth, M.E., 1972. Initial stage sulphuric acid leaching kinetics of chalcopyrite using radiochemical techniques. Paper 72-B-96, AIME Annual Meeting, San Francisco, February 1972. De Boer, J.H., 1970. In: D.H. Everett and R.H. Ottewill (Editors), Surface Area Determination. Butterworth, London, p. 10. Chen, J.H. and Harvey, W.W., 1975. Cation self-diffusion in chalcopyrite and pyrite. Metallurgical Transactions (B), 6B: 331--339. Craig, J.R. and Scott, S.D., 1974. In: P.H. Ribbe (Editor), Sulfide Mineralogy. Mineralogical Society of America (Short Course Notes, November 1974, Vol. 1, Chapter 5). Dutrizac, J.E. and MacDonald, R.J.C., 1973. The effect of some impurities on the rate of chalcopyrite dissolution. Canadian Metallurgical Quarterly, 12(4): 409--420. Dutrizac, J.E. and MacDonald, R.J.C., 1974a. Ferric ion as a leaching medium. Minerals Science and Engineering, 6(2): 59--100. Dutrizac, J.E. and MacDonald, R.J.C., 1974b. The kinetics of dissolution of covellite in acidic ferric sulphate solutions. Canadian Metallurgical Quarterly, 13(3): 423--433. Dutrizac, J.E., MacDonald, R.J.C. and Ingraham, T.R., 1969. The kinetics of the dissolution of synthetic chalcopyrite in aqueous acidic ferric sulphate solution. Transactions of The Metallurgical Society of AIME, 245: 955--959. Dutrizac, J.E., MacDonald, R.J.C., and Ingraham, T.R., 1971. Effect of pyrite, chalcopyrite and digenite on the rate of bornite dissolution in acidic ferric sulphate solutions. Canadian Metallurgical Quarterly, 10(1 ): 3--7. Gottschalk, V.I-1. and Buehler, H.A., 1912. Oxidation of sulphides. Economic Geology, 8: 15--34. Habashi, F., 1966. The mechanism of oxidation of sulfide ores in nature, Economic Geology, 61: 587--591. Ichikuni, M., 1960. The dissolution of sulphide minerals in various media. III. Factors intervening in the dissolution of chalcopyrite. Bulletin of the Chemical Society of Japan, 33: 1159--1162. Ichikuni, M., 1962. The action of ferric ions on chalcopyrite. Bulletin of the Chemical Society of Japan, 35: 1765--1768. Jost, W., 1960. Diffusion in Solids, Liquids, Gases. 3rd edn. Academic Press, New York, Chapter 1. Linge, H.G., 1976. A study of chalcopyrite dissolution in acidic ferric nitrate by potentiometric titration. Hydrometallurgy, 2(1): 51--64.
233
Majima, H., 1969. How oxidation affects selective flotation of complex sulfide ores. Canadian Metallurgical Quarterly, 8(3): 269--273. Peters, E. and Majima, H., 1968. Electrochemical reactions of pyrite in acid perchlorate solutions. Canadian Metallurgical Quarterly, 7(3): 1 1 - 1 1 7 . Prosser, A.P., 1970. Influence of mineralogical factors on the rates of chemical reactions of minerals. In: M.J. Jones (Editor), Mineral Processing and Extractive Metallurgy. The Institution of Mining and Metallurgy, London, pp. 59--80. Wadsworth, M.E., 1972. Advances in the leaching of sulphide minerals. Minerals Science and Engineering, 4(4): 36--47. Woodcock, J.T., 1967. Copper waste dump leaching. Proceedings of the Australasian Institute of Mining and Metallurgy, No. 224: 47--66.