LEACHING OF PRIMARY SULF DE ORES IN SULFURIC ACID SOLUTIONS AT ELEVATED

TEMPERATURES AND PRESSURES*

Abstract

This report covers laboratory experiments simulating in situ copper recovery from primary sulfide
ores in sulfuric acid systems pressurized with oxygen. Copper extraction and acid consumption
data are correlated with temperature in the range 50-150°C and quenched pH in the range 0.25-2.
Adverse changes in the chemi cal composition of the leach solution and the formation of gangue-
mineral alteration products and secondary sulfides can severely limit the extraction of copper
above 90°C. Keeping the pH below the self-buffered value of about 2 and the to^perature below
about 110°C produces the maximum extraction rates. Forcing the pH too low not only results in
excessively high acid consumption but may also limit the extraction of copper at temperatures
above 100°C.

INTRODUCTION

In situ leaching of primary sulfide ore provides a possible means of recovering copper from
deposits that are too deep to be mined economically by conventional methods. The first step in
such a process is the rubblization of a region of the ore deposit by the use of nuclear explosives orj
alternatively, by undercutting and caving assisted by conventional methods. The ore fragments are
then

*Work performed under the auspices of the U.S. Atomic Energy Commission.

leached in tne aqueous sulfuric acid system produced by introducing oxygen into the flooded ore
under hydrostatic pressure. The increase in solubility of oxygen under these conditions, together
with the elevation in temperature that results from the oxidation of the sulfide minerals, greatly
accelerates the rate of copper recovery fn m primary copper ores.

This concept has been under development at the Lawrence Livermore Laboratory. The objective of
our research program is to obtain experimental data and to develop a fundamental understanding
of the leaching process that will enable extrapolation of results from laboratory experiments to
field conditions. One important achievement in the LLL program was the development of a
leaching model by Braun, et al. that successfully describes the recovery of copper from ore
fragments of any given particle size distribution under limited conditions of temperature and pH.
The leaching mechanism involves mixed kinetics and includes a surface reaction within a moving
reaction zone plus pore diffusion of dissolved oxygen though the reacted portion of the ore
fragment to the reaction zone. The application of this model over a broad range in temperature
and pH is difficult because of changes in the chemistry of the leach solution, as well as
mineralogical and.physical changes within the ore. In order to broaden our knowledge of the
leaching process and to optimize the copper extraction rates, we have conducted leaching
experiments over a range of pH and temperatures. In this paper we will present some of the
experimental data and discuss some of the important factors that may limit the rate of copper
extraction.

EXPERIMENTAL

Leaching experiments were conducted in autoclaves of capacity 0.5-20 litre. The design of the
autoclaves has been described in detail by Braun, et al. The autoclaves are made of stainless steel
lined wi-h chemical lead or made entirely of titanium. These systems provide means for removing
liquid samples for chemical analysis during the course of the experiment without disturbing the
operating temperature or pressure.

Ore Type

The ore used in the experiments is a primary porphyry copper ore from the San Manuel Mine
(Magma Copper Company, Arizona), The chemical composition of the San Manuel ore appears in
Table I. The ore can be described as quartz-sericite, quartz-sericite-biotite, and altered quartz
monzonite rock types. Major gangue minerals include quartz, K-raica (predominantly sericite), and
biotite. Minor gangue minerals include K-spar (largely altered to sericite), plagiociase, hematite,
calcite, magnetite, kaolinite, montmorillonite, and hornblende. Sulfides include chalcopyrite (cp)
and pryite (py), with traces of covellite and bornite. The py/cp mole ratio is 2. The calcite/cp mole
ratio is 0.85.

The mineralogy and grain size are highly variable even within small samples, and gradations
between the rock types are common. The quartz-sericita rock type consists of fine to medium-
grained quartz (<1 mm) with pervasive and nonlocalized sericite and minor, fine-grained,
secondary biotite. The quartz-sericite-biotite rock consists of medium-grained (<1 mm) biotite,
quartz, and altered feldspar in a fine-grained matrix of sericite, quartz, and minor magnetite. The
altered quartz monzonite consists of coarse-grained quartz (<7 mm), highly altered feldspar
porphyblasts (<5 Mm), and localized concentrations of sericite and quarts pseudomorphing
original feldspar porphyblasts. Minor amounts of calcite (approximately 1 wt%) occur as small
veins and disseminated coarse grains.

Procedure

The autoclave is hand loaded with a weighed amount of ore sized by screening. A measured
amount of deiouized water, adjusted to the desired pH (25°C) by additions of concentrated H 2SO4,
is added to cover the ore. The system is pressurized with oxygen and heated to the desired
temperature. Flow of oxygen is then begun and the pressure of the system is adjusted to 2.8 MPa
(400 psig). Convective stirring induced by the rise of oxygen bubbles circulates the solution within
the vessel. After a period of one day, concentrated H2SO4 and water are added io adjust the pH
and reach the final liquid level. The pH and liquid level are cointained for the duration of the
experiment by weekly additions of H2SO4, and water if necessary to compensate for acid
consumed by the ore and water lost by evaporation and sampling.

The addition of sulfuric acid was completely omited in some of the experiments. In these
experiments at uncontrolled pH, only deionized water was added and the pH allowed to achieve
an equilibrium (self-buffered) value.

Samples of the leach solution are withdrawn from the autoclave every two weeks for copper
analysis. The small amount of copper contained in the samples not returned to the autoclave is
taken into account in the data reduction. The copper fraction extracted during the experiment is
based on the assay of 0.7 wt% Cu for the composite sample of San Manuel ore. Variations in the
grade of the ore used in the experiments are believed to be small. Several separate assays on
composite samples show variations to be within 0.01 wt%.

The leached ore fragments are studied by x-ray diffraction and optical microscopy to determine
changes in ore mineralogy produced during the experiment.

EXPERIMENTAL RESULTS

Table II lists the pertinent experimental conditions for each of the experiments to be discussed.
These experiments concerned the effects of Ph and temperature on the rate of copper recovery.
Therefore we attempted to keep other contributing factors (e.g., mineralogy, physical pronerties
of the ore) as constant as possible.

Experiments with Self-Buffered Ph

Braun et al. described a series of experiments that demonstrate San Manuel ore can be
successfully leached at 90°C and 2.8 MPa oxygen with deionized water as the starting solution. The
copper recovery for one of these experiments is shown in Fig. 1. For the experiments at 90°C, the
pH of the system decreased from an initial value near neutral to a steady-state value near 2 (Fig.
2). The buffering effect observed in these experiments is an important feature of the leaching
system and represents a steady-state balance of H+ -dependent reactions involving both the
sulfide and gangue minerals. These reactions will be discussed in detail in detail in a later section.

Below 90°C the rate of sulfuric acid production by sulfide oxidation is insufficient in the early
stages of leaching to balance the acid-consuming reactions with calcite in the ore. This is
demonstrated by Exp, 139 which had a starting solution of deionized water at 70°C. This
experiment was terminated after 13 days because the ph never dropped below 5.5 and
consequently no copper was recovered. Hence at 70°C acid must be added to the starting solution.
two experiments were conducted to determine the amount of acid required to start the leaching
reactions at 70°C for ore particles 16 ±3 mm in diam. In exp. 154 an initial acid addition of 15 meq
H+ /kg ore (1,5 lb H2SO4/ton ore) had little or no effect on copper Ieaching rate for the first 25 days.
Thereafter the leaching rate began to accelerate rapidly, however, reaching an extraction of 35 wt
% of the copper at 100 days. This indicates that with very little acid addition satisfactory leaching
rates can be obtained if an initial induction period, during hhich the rate of acid consumption by
carbonate minerals exceeds the rate of acid production from sulfide oxidation, is acceptable,

In exp. 163 the onset of leaching-rate acceleration at 70°C was decreased to 14 days by the
gradual addition of 50 meq H+/kg ore (5 lb H2SO4 /ton ore) to maintain a pH of 2.

Above 90°C the higher rate of acid production from sulfide oxidation makes acid additions
unnecessary, The requirement of acid addition at a given leaching temperature will, of course, be
modified for other ores containing different proportions of carbonate and pyrite.

Cooper extraction rate vs pH

A number of experiments were conducted at different pH values at 70°, 90°, and 110°C to
determine the effects of pH at the temperature of interest.

Figures 3 and 4 show that the copper extraction rates at 90° and 70°, respectively, are increased by
maintaining the pH below the self-buffered value of approximately 2, The effect of maintaining a
pH of 1 by additions of acid (Exp. 108, Fig. 3) is to accelerate the initial lnaching rate; however, the
latter stage of leaching approaches that. of the self-buffered system (exp. 89) . Therefore, there
appears to be no significant long-terra advantage in maintaining the pH at 1 rather than allowing
the system to be self-buffered at 90°C. The rapid initial leaching rate at pH1 may be due to the
increased oxidation of chalcopyrite by Fe +3, or to the increase in porosity of the rock because of
less precipitation of alteration products. Ferric iron is most abundant in the early stages of
leaching and gradually decreases in concentration with time because of the precipitation of
jarosite.

At 70°C, the copper recovry at a self-buffered pH (Exp. 163, Fig. 4) appears to be limited to
approximately 50%. This is surprising considering the excellent recovery at 90°C and self-buffered
pH. Considering the experimental results at 90°C (Fig. 3), the copper extraction curves at 70° and
pH 1 (Exp. 158), and at selfbuffered pH (exp, 163), were expected to be similar during the latter
stage of leaching. Experiment 163 is still in progress and the potential problems that could
produce the apparent leaching curve shown in Figure 4 have not been evaluated.

At 110°C (Fig. 5), a similar increase in the extraction rate was observed at pH values from 1.5 to
0.75. However, at 110°C and pH 0.5-0.25, the copper extraction reached a value of approximately
55-60%. This is in sharp contrast to the very rapid and complete recovery at 70°C and pH 0.25.

These results suggest that the most rapid recovery of copper can be accomplished at lower tempe
ature (<90°C) with the pH maintained below the self-buffered value. Temperatura greater tan 90°C
with very low pH (e.g., 0,5) área obviously undesirable.

Even though the increased rate of copper recovery that can often be accomplished by additions of
acid is desirable, the amount of acid consumed fay the gangue minerals ray be economically
unacceptable. For example, in the extreme case the acid consumption data presented in Fig, 6
show that at pH 0.25, 3.5 equiv. H+/kg ore was consumed in the first 50 days. This is equivalent to
0.175 kg of concentrated H2SO4 per kg ore (350 lb H2SO4 /ton ore). In contrast, the self-buffered
experiments require no acid additions.

The rate at which acid is consumed by the ore is largely a function of the, H+ activity (pH). Figure 7
shows the good agreenent between the H+ activity and the amount of acid consumed by the ore at
110°C.

Copper Extraction Rate vs Temperature

A set of experiments were conducted to determine the effect of temperature on the rate of
copper recovery. The experiments were conducted at pH 1 and near 2 (self-buffered value). These
pH values were chosen because the previous set of experiments show that very low pH may be
undesirable because of the high acid consumption by the gangue minerals.

Figure 8 shows the copper extraction rates at the self-buffered value near 2. During tne first few
days of the experiment, the higher-temperature experiments had a greater rate of recovery than
the 90°C experiment. However, Exp. 75 at 130°C showed an abrupt loss of copper at 18 days. This
is believed due to the formation of copper-bearing alteration products and secondary copper
sulfides as will be discussed in a later section.

The experiments at pH 1 are shown in Fig. 9. The highest rate of recovery was at 110°C whereas
the lowest rate of recovery was at 50°C.
These experiments show that temperatures above 110°C and below 70°C give unsatisfactory
recovery rates at pH 2-1.

Copper Extraction Rate at High Temperature and Low pH

Three experiments were conducted at 130° to 150°C and low p!l and each had poor copper
recoveries. The copper extraction data are shown in Fig. 10. Experiments 159 and 161 show
extraction curves similar to that of Exp. 75 (Fig. 8) . The abrupt loss of cupper shown by these
curves is believed to be due to the formation of copper-bearing phases produced during the
alteration Df the gangue minerals.

Summary of Experimental Results

The experimental data presented show that:

1. At 70°-90°C, successful leaching can be accomplished at approximately pll 2-0.25.

2. At 110°C, successful leaching can be accomplished only if the pH is between approximately
2 and 0.75.
3. At temperatures above 110° C, leaching is unsuccessful at any pH.
4. Highest extraction rates are at low pH and 70°-90°C.
5. Low pH may make the gangue minerals consume unacceptable amounts of acid.

MINERALOGY AND CHEMISTRY OF THE LEACHING SYSTEM

Before uiscussing the potential problems that may limit the extraction of copper, the chemical
reactions occurring during leaching will be considered. Tha acid-consuning reactions are:
(Less important acid-consuming reactions with minor gangue minerals, plagioclase, chlorite, etc.,
will not be considered.)

The acid-generating reactions are

For the self-buffered system where no acid was added, the pH of the leach solution (measured at
2S°C) decreased from an initial value near neutril to a steadystate value near 2. This buffering
effect represents a steady-state balance of the H+ -dependent reactions shown above as well as
ionic equilibria:
Where

The chemical composition of the leacii solution for a self-buffered experiment at 90°C is shown in
Pig. 11. The high sulfate concentration is an important feature of the systew. Most of the metal
ions produced during leaching are combined with sulfate as meral-sulfat e complexes or
precipitates.

The calcium concentration shown in Fig. 11 steadily decreases due t o the precipitation of
anhydrite, The Na, K, Fe, and Al concentrations ar e limited by the precipitation of jarosite-alunite
minerals [KFe 3 (SO4 )2 (0H)6, - KAl3(SO4)2 (OH)6 with Na substitution for K and limited substitution
between Al and Fe].

Jarosite and alunit e are the most common alteratio n products observed in the leached samples.
They occur as fine-grained material throughout the rock and as replacements of K-feldspar and K-
mica. The replacement reactions for K-feldspar and K-mica to jarosite are

With the exception of calcite, biotjte is attacked most readily by the acidleach solutions. The initial
step of Mofit e alteration is to a vermiculite-type material and jarosite. The more advanced stage
of alteration is to a mixture of jarosite and "brown clay." The brown clay is believed to be a
mixture of µm-size amorphous SiO2, jarosite, and montmorillonite.
The SiO2 concentration of the system ir at saturation with respect to amorphous silica. The amount
of amorphous silica observed in the ieached ore depends on the degree of alteration of the micas
and feldspars.

Secondary sulfides, digenite (Cu9S5) and covellite (CuS) are formed in all the leaching experiments.
Part of the copper released during oxidation of chalcopyrite within the reaction zone diffuses into
the ore and combines with unreacted chalcopyrite to form digenite and covellite. At pH values
above approximately 0.75, digenite is more abundant than covellite and below 0.75, covellite
predominates as the secondary sulfide.

DISCUSSION

The experimetal results show that temperature and pH have a pronounced effect on the copper
extraction rates. These effects are predictable in part from the leaching model developed by
Braun, et al. Their model is based unon diffusional rate constant for oxygen and a. chemical rate
constant, both of which are temperature dependent. Although the leaching model does not
directly account for variations in pH, several factors that are included in the model are affected by
pH. For example, the porosity of the ore would be greater at a low pH because of less precipitation
of iron salts.

The unpredictable loss of copper and the limited extractions observed in some of the experiments
at high temperatures and low pH may be due to the formation of copper-bearing alteration phases
or to the salting out or dissolved oxygen in the leach solution.

Copper Bearing Alteration Phases

The abrupt loss of copper from solution during leaching at high temperature (>110°C) may be due
to the incorporation of copper into gangue mineral alteration products. Petrographically, the only
unique feature of the leached ore in these experiments is the rather advanced state of gangue
mineral alteration. There is an appreciable amount of jarosite-alunite, montmorillonite, and a glass
phase (=60% SiO2).

The formation of montmorillonite is of considerable interest, both because it is capable of trapping

copper and because it is a common alteration product in the leached ore. The copper can be
adsorbed on the clay mineral surfaces, occur as an interlayer cation, or occur in a lattice position
substituting for Al [medmontite ) . The activity product constant for Mg-montmorillonite from Eq.
(10):
At a constant pH, the significant decrease in log K with an increase in temperature shows that the
formation of montmorillonite is favored at high temperatures. It is also possible that copper may
be incorporated in the glass phase or in the hydrated iron oxides that are produced. Detailed
petrographic and electron microprobe analyses are currently in progress on the alteration
products in order to locate the phases that may be acting as copper traps.

Salting Put of Dissolved Oxygen

The reaction for oxygen in equilibrium with the leach solution may be written as

where fO2(g) is the fugacity of oxygen gas and a02(aq) i s the activity is the oxygen in the aqueous
phase. The aO2(aq) is expressed by
Where yO2 is the activity coeddicient and mO2 is the molality of O2 at some finite electrolyte
concentration. Combining Eqs. (16) and (17) shows that the molality of O2(aq)

at a fixed temperature and fugacity of O2(g) depends on the value of y. By convention, the activity
coefficient of a gas dissolved in pure water is taken to be unity. However, y changes in value from
unity at zero ionic strength (pure water) to a number greater than 1 with increasing ionic strength.
Therefure, at fixed temperature and fO2, increasig ionic stregth leads to decrease in the
concentration of dissolved oxygen (salting out).

The lack of thermodynamic data on the chemical species in the leach solution at the temperatures
of interest prevents calculation of either the ionic strength or activity coefficients. However, the
minimum decrease in the oxygen concentration can be estimated from the sulfate analysis of the
leach solution and the data of Bruhn, et al., given in Fig. 12. Table III gives the chemical analyses
for the series of experiments at 110°C. The decrease in oxygen concentration can be significant for
experiments with high concentration of dissolved solids. For example, the estimated minimum
decrease in oxygen concentration in Exps. 120 and 121 would be 30% or greater as suggested from
the high sulfate concentrations. The high cation concentrations, particularly the trivalent Al +3 , may
contribute as much to the ionic strength of the solution as does sulfate. Therefore the actual
decrease in oxygen concentration may be significantly greater than 30%. Experiments 120 and 121
had the greatest concentration of Al+3 and SO4 -2 and they also had poor copper recoveries. The
limited copper extractions observed from these experiments may be due to salting out of oxygen
at the high ionic strengths.

Formation of Secondary Sulfides

Covellite and digenite have been observed in all the leached ore; they can be found in experiments
that have leached well and those that have leached unsatisfactorily.

Figure 13 shows chalcopyrite that has been partially replaced by digenite and covellite. During
leaching, covellite and digenite are cnntinually being formed in the interior of the ore fragments as
copper diffuses into the rock and reacts with the chalcopyrite. As the oxidizing zone advances into
the ore fragment, chalcopyrite, as well as the previously formed secondary covellite and digenite
are oxidized and release copper to the solution. As long as the rate of oxidation in the reaction
zone proceeds faster than the loss of copper to form digenite and covellite, the experiment will
have a favorable leaching rate. If the rate of oxidation decreases, perhaps due to the salting out of
dissolved oxygen, the rate of secondary sulfide formation may result in a decreased rate of copper
extraction or a loss of copper from solution. Therefore, the formation of secondary sulfides is not
believed to be the principal cause of poor extraction rates but rather reflects some other problem
in the leaching system.

Experiments are in progress to determine the factors that control the rate of formation of digenite
and covellite. Preliminary results show that the rate at which copper is lost from solution to form
secondary sulfides is strongly dependent on the pH. At pH 1.5 copper is lost from solution
approximately five times as fast as at pH 0.25.

SUMMARY AND CONCLUSIONS

The experimental and theoretical work conducted at the Lawrence Livermore Laboratory have
demonstrated that chalcopyrite ore can be successfully leached in sulfuric acid systems
pressurized with oxygen. We believe that this technology can be used successfully for the in situ
recovery of copper, and that this method is particularly attractive for ore deposits that cannot be
economically mined by conventional methods.

The results of this study show that pH and temperature have a pronounced effect on the rate and
ultimate recovery of copper. The pH and temperature of the leaching system determine, to a large
extent, the relative stability of gangue and alteration minerals, the rate of decomposition of the
ore, physical properties of the ore, and the chemistry of the leach solution. These factors in turn
are important in determining the rate at which copper can be leached from the ore.

The results of our study suggest that the most desirable conditions for leaching are near 90°C and
at a self-buffered pH near 2. At these conditions, the copper extraction rates are good and the
potential problems that can limit the recovery of copper are minimized. There is also the economic
benefit in that no acid additions are required. This is true for San Manuel ore; however, some acid
additions pay be required for ores with significantly different carbonate content or
chalcopyrite/pyrite ratios. Maintaining the pH of the system below the self-buffered value
generally has the effect of increasing the copper extraction rates, particularly in the earlystages of
leaching. This is true for leaching at temperatures less than 110°C. The enhanced extraction rate at
low pH may be due to:

(a) greater porosity of the ore because of less precipitation of alreration products.

(b) increased oxidation of chalcopyrite by Fe+3.

(c) and less replacement of chaicopyrite by digenite and covellite. The benefits obtained by
maintaining a low pH may in most cases be offset by the increased cost of adding sulfuric acid.
Another undesirable effect that low pH may have is to lead to high concentrations of dissolved
salts, giving an enhanced salting out of dissolved oxygen from the leach solution. This effect is
apparently nost serious when low pH is combined with high teraperatures, producing a rapid
decomposition of the gangue minerals.

The most desirable temperatures for leaching are near 90°C. At lower temperatures, dissolution
rates of the sulfides are smaller, and acid additions may be necessary. This is particularly true
during the early stages of leaching when acid consunption by the carbonate minerals is the
greatest. At higher temperatures, the rate of decomposition of the gangue minerals and the
production jf alteration products are greatly accelerated which may lead to poor copper
recoveries.

The experimental results reported in this study have increased our knowledge of the leaching
process, particularly in determining the Dpttnun as well .-is the undesirable conditions for leaching
San Manuel ore. Wc expect sintlnr leaching characteristics for chalcopyritc ores with mineralogies
similar to the San Manuel ore. However, leaching experiments should be conducted on a variety of
ore types to determine what variations In leaching characteristics can he expected for variations in
ore mineralogy. Particular attention should be given to ore that has abundant Fe-Mg silicates,
particularly biotite and hornblende which are readily attacked by acid solutions. An effort should
be made to determine the extent to which copper is trapped in clay minerals such as
montmorillonite.

ACKNOWLEDGMENTS

We gratefully acknowledge the technical contributions of A.E. Lewis and other LLL personnel and
the dedicated efforts of W. Beiriger and L. Marsh in the experimental work.