hydrometallurgy

ELSEVIER Hydrometallurgy37 ( 1995 ) 1-21

Bacterial leaching of complex sulfide ore samples

in bench-scale column reactors
L a s s e A h o n e n a, Olli H. T u o v i n e n b
=GeologicalSurvey of Finland, FIN-02150 Espoo, Finland
bDepartment of Microbiology, Ohio State University, 484 West 12th Avenue,
Columbus, OH 43210-1292, USA
Received 17 August 1993; revised version accepted 4 February 1994

Abstract

Several variables were examined in column bioleaching of a complex sulfide ore mate-
rial which contained chalcopyrite, pentlandite, pyrite, pyrrhotite and sphalerite as the main
sulfide minerals. Samples were used with varying proportions of pyrrhotite, pyrite, quartz-
ite (low acid consumption) and skarn (high acid consumption). The experiments were
carried out using bench-scale column leaching reactors which were inoculated with acido-
philic, Fe- and S-oxidizing bacteria, initially derived from the source mine water. Leaching
rates in sterile controls were negligible. In inoculated columns new solid phases (covellite,
jarosite, Fe (III) oxide and elemental sulfur) were formed. Acid consumption was highest
under low pH and low redox potential conditions. The solubility of ferric iron was con-
trolled by jarosite and an Fe (III) hydroxide (initially amorphous). The leaching rates of
Co (from pyrite and pentlandite), Cu (chalcopyrite), and Zn (sphalerite) showed a ten-
dency to increase with dissolved ferric iron concentration. The leaching of Ni (from pyr-
rhotite and pentlandite) did not correlate with the concentration of ferric iron in solution.
Microscopic counts of bacteria in solution, deemed insufficient to represent total bacterial
counts, showed a tendency to be higher at the lower pH and intermediate redox potential
ranges. Trickle-leaching conditions yielded higher acid production and redox-potential
values compared with flood leaching. The leaching rates of Co, Cu, Ni and Zn each re-
sponded differently to redox potential and pH regimes. The accelerating effect of a de-
creasing particle size on the metal leaching rates was amplified by low pH values.

1. Introduction

Industrial scale dump- and heap-leaching processes are especially suited for
low-grade porphyry copper ores, in which primary copper minerals have been

0304-386X/95/$09.50 © 1995 Elsevier ScienceB.V. All rights reserved

SSDI 0304-386X ( 94 ) 00011-Q
2 L. Ahonen, O.H. Tuovinen/Hydrometallurgy 37 (1995) 1-21

partially altered to secondary sulfide and oxide minerals, which are readily ame-
nable to oxidative and non-oxidative acid leaching. Acidophilic thiobacilli play
an important role in these processes [ 1 ]. Other metals are dissolved from asso-
ciated minerals in varying concentrations, but the current commercial heap-
leaching processes are designed for Cu recovery. In the present report, the results
are summarized from an extensive, laboratory-scale study dealing with the bac-
terial leaching of a complex Cu-Co-Zn ore.
The ore material was obtained from the Keretti mine, Outokumpu, eastern
Finland. The geological setting and mineralogy of the deposit have been de-
scribed by Peltola [ 2 ] and Koistinen [ 3 ]. Mining was carried out under ground.
Upon the termination of mining activity, ore reserves remained which were not
exploitable by conventional excavation techniques. An underground, in situ,
leaching process was therefore considered as one of the possible alternative tech-
niques. Several microbiological, chemical and physical factors influencing the
leaching were included in the experimental plan. In the biological leaching pro-
cess, the proper contact of ore material with the leach liquor is essential for fluxes
of reactants and products, such as bacteria, dissolved gases (02 and CO2), solu-
bilized metals and sulfur species. In bench-scale studies, various percolator and
column designs have been developed for studying the kinetics and optimization
of biological leaching processes. Rate-controlling steps in copper heap-leaching
systems have not been elucidated.
Ferric iron is an important oxidizing agent in the bacterial leaching of sulfide
minerals. Soluble iron species are the main determinants of redox potential, with
active iron-oxidizing bacteria ( Thiobacillusferrooxidans, Leptospirillumferroox-
idans) contributing to high Fe3÷/Fe 2+ ratios. Precipitation of ferric iron in the
leaching system may suppress the metal solubilization by preventing the contact
between the leaching agent and the mineral. The solubility of iron is defined by
the solution redox potential and pH. Thus, the optimization of these parameters
may greatly improve the metal recovery.
The bacterial leaching process requires acidic conditions, the acidity often being
autochthonously produced by the oxidation of pyrite and hydrolysis of ferric iron.
The acid may, however, be neutralized in various acid-consuming reactions; for
example, involving the leaching of carbonate minerals. Some silicate minerals
also react in acid solution.
The solid/liquid contact area, which in low-porosity rocks is mainly a function
of particle size of the material, is a major factor in determining the kinetics of the
leaching reactions. Typical to leaching processes is their selectivity with respect
to certain minerals or groups of minerals. The leaching effect is also able to pen-
etrate into microfissures and micropores of the rock mass, thus reducing the need
for energy consuming crushing and grinding operations. The optimization of a
leaching process thus requires quantitative information on the effect of grain size,
as well as on the propagation of the leaching effect inside the mineral grains.
As in all biochemical and chemical processes, the rates of leaching reactions
are also temperature dependent. An experiment was also included in this study
program to evaluate the effect of temperature on the bacterial leaching in column
 L. Ahonen, O.H. Tuovinen / Hydrometallurgy 37 (1995) 1-21 3

reactors; the results have been presented elsewhere [ 4 ]. The characterization of

the temperature effects of bioleaching on suboptimal temperatures was consid-
ered to be particularly important. An additional set of column leaching experi-
ments was thus carded out to obtain more quantitative information on the tem-
perature effects; the results have been reported [ 5 ].
In the leaching of complex sulfide ore, electrochemical interactions between
sulfide minerals may play important role. Attention was thus also paid to the
relative proportion of the different sulfide minerals, their mode of occurrence
and textural relationships. The effect of silver addition was also tested in this set
of long-term bacterial leaching experiments, as previously reported [6 ]. All these
factors have a considerable interaction in active leaching processes and their rel-
ative effects should be taken into consideration in optimization studies.

2. Materials and methods

In the course of the study several samples from the ore, of different mixtures of
the sulfide ore and gangue, were tested. Copper was present almost exclusively as
chalcopyrite (CuFeS2), zinc as sphalerite (ZnS), while cobalt was distributed in
pentlandite [ (Co,Ni,Fe)9S8] and pyrite (FeS2). Nickel was associated with pen-
tlandite and with pyrrhotite (Fel_xS). The average pyrite to pyrrhotite ratio in
the ore is about 1 : 1, but spatial variation in their relative abundance in the ore
occurs. Both pyrite-dominated and pyrrhotite-dominated samples were used in
the experiments. In the leaching tests the sulfide ore samples were mixed with
gangue material. In most experiments, the gangue material was carbonate and
graphite-bearing siliceous metasediment, hereafter referred to as quartzite. In ad-
dition to quartzite, which is the immediate host rock of the ore, more effective
acid-consuming material was also tested in the experiments. This material was a
mixture of diopside- and tremolite-skarn, serpentinite and black schist, which
also are typical in the mineralized rock association. The rock material was crushed
and sieved in different size fractions. A summary of the experimental conditions
and sample components in different experiments is presented in Table 1. The
partial elemental composition of the ore samples is presented in Table 2.
The experiments were carded out with glass column reactors which, unless oth-
erwise specified, contained 800 g of ore sample and 800 ml of leach solution (Ta-
ble 1 ). The columns were 50 cm high with an internal diameter of 9 cm. The ore
was placed on a fritted glass base which was raised from the bottom of the col-
umn. The leach solution was passed through the ore sample by gravity and recir-
culated through a side loop either by airlift or with a peristaltic pump. In two
leaching experiments the column system ( 100 × 11 cm) comprised 12 kg ore ma-
terial and 31 of mineral salts solution.
The bacteria used in the column leaching experiments were derived by enrich-
ment of mine water samples, using sulfide ore material from the mine site as the
substrate. The leach solution contained 3.0 m M (NH4)2SO4, 2.3 m M K2HPO4
and 1.6 m M MgSO4-7H20 in dilute sulfuric acid, with an initial pH of 2.5. Two
4 L. Ahonen, O.H. Tuovinen/Hydrometallurgy 37(1995) 1-21

~o~

-~

N8

.*~ .~.
0

l l l l l l l l ~ ~ ~ ~
~ ~ l l l l ~ I I I I I I I I I

.~ .~ °~

~-oo-- ~ o~.~ 0

o ~ _ ~

0 0 0 ~ ~'~ ~'~
"~
::I ~ ~ ~ ~ :~ c~ ~ o o ~ ~ ~ ~ c~ O - - ~ e ~ c , ~ ~ ~ c~

•==.~ •o~.,.o ~~. o.~.~o

~ ,, ~:._-~ ~ ~~~ .~-~ ~-,~o ~.~o ~ g _ ~ _ ~ - ~ ~ ~
ul
 L. Ahonen, O.H. Tuovinen / Hydrometallurgy 37 (1995) 1-21

Table 2
Partial elemental analysis of ore samples used in column leaching experiments

Ore Composition (%)

sample
Cu Zn Ni Co Fe S

A 0.29+0.04 0.32+0.03 0.16+0.02 0.056+0.003 11.6+0.3 8.5

B 0.23+0.05 0.31+0.12 0.17+0.01 0.058+0.008 12.2+1.4 7.4
C 0.59+0.26 0.08+0.01 0.11 +0.04 0.077+0.020 9.6_+ 1.3 10.7
D 0.52+0.28 0.08+0.02 0.12+0.04 0.057+0.015 8.2+ 1.5 7.6
E 0.68+0.31 0.12+0.06 0.12+0.02 0.087+0.037 10.3+2.6 7.8
F 0.042_+0.13 0.04+0.01 0.11 +0.01 0.049+0.013 7.8+ 1.3 6.5
G 0.50+0.15 0.18+0.10 0.10+0.02 0.056+0.014 8.8+1.3 6.5
H 1.2+1.2 1.16+0.15 0.08+_0.02 0.079+0.026 11.5+1.8 7.3
I 0.64+_0.28 0.07_+0.02 0.10+0.02 0.079+0.029 8.3+2.2 10.1
J 6.15+0.25 0.43+_0.02 0.08+_0 0.61+_0 38.1+_2.1 36.0
K 1.0 0.50 0.30 0.10 10.5 9.1

The maximum deviation in results is given if three or more parallel analyses were available.

columns were initially set up for enrichment and operated for about 300 d. The
cultures thus derived were used for inoculation in subsequent experiments. Un-
less otherwise stated, the column leaching experiments were carried out at
20_+ 2 oC. Sterile controls were included in three experimental sets.
During the experiments, samples (30 ml) were taken every 2 weeks for chem-
ical analysis. Sample volumes were replaced with equivalent aliquots of mineral
salts solution. Solution pH and redox-potential values were measured 2-3 times
a week. Redox potential was measured with a Pt electrode using standard calomel
electrode (SCE) as a reference. Loss of solution due to evaporation was esti-
mated volumetrically and compensated for by adding sterile distilled water. The
initial acid consumption was satisfied with sulfuric acid. Where necessary, so-
dium hydroxide was used to neutralize excessive acid production.
Dissolved metals were analyzed by atomic absorption spectrometry (AAS).
The elemental composition of ore samples is based on averaged results obtained
by AAS and by X-ray fluorescence spectrometry (XRF). Upon termination of
the experiments, the solid residues were air dried and samples were taken for
chemical analysis, X-ray diffraction (XRD) and ore microscopy. In some exper-
iments, individual samples were taken for mineralogical examination during the
active phase of leaching. The solid residues were dissolved in HC1-HNO3 mix-
ture and analyzed directly by AAS or indirectly by XRF. Total sulfur in solid
residues was analyzed by combustion and IR detection (LECO IR 32H), SOby
iodometric titration and sulfate indirectly by AAS determination of Ba after pre-
cipitation of BaSO4. The chemical analyses were carried out by the analytical
laboratory of Outokumpu Oy.
Leaching rates were calculated from the slopes of the metal dissolution curves
for each data point. In order to reduce the effect of short-term changes and ana-
6 L. Ahonen, O.H. Tuovinen/Hydrometallurgy 37 (1995) 1-21

lytical errors on rate calculations, the metal dissolution curves were linearized
with five data points, using the sum of least squares.
The concentration of dissolved Fe (III) was estimated on the basis of redox
potential measurements, applying the Nernst equation. In these estimations, the
speciation of the soluble complexes of ferric and ferrous iron was not systemati-
cally considered; only the species FeSO~- was accounted for. Concentrations were
taken to equal activities:
RT, [FeSO + ]
E=Eo - - ~ l n [Fe2+ ] [SO24_ ]
The thermodynamic data used in this work were obtained from the MINTEQ
data base and from Brookins [ 7 ].

3. Results and discussion

3.1. General trends

A summary of metal dissolution rates in various experiments is given in Fig. 1.

The leaching data are primarily based on experiments carded out at 20 + 2°C,
with the exception of one column leaching experiment each at 4 + i °C and
10 + 1 °C. The particle size encompassed three sizes in the range of 1.68-20 mm.
Most of the data were obtained using a particle size fraction with a 1.68-5 m m
diameter.
Long-term column leaching experiments lasted for about 2 yr, by which time
more or less complete leaching of zinc was achieved. The yields of leaching
amounted to about 50% for Ni and Co and < 25% for Cu.
Leaching rates gradually decreased in the long-term leaching experiments (Fig.

~m

! ll !-
x m
1~ 6 8'o 1do 240 ~ ' ~0 4'0 sb 6 2'0 4'0 6'0
Cu °/oo per month Zn °/oo per month Ni °/oo per month Co °/oo per month
Fig. 1. Leaching rates of copper, zinc, nickel and cobalt in column leaching experiments, expressed as
permil solubilization of each metal content per month. A total of 30 experiments were carried out, of
which the first two were used only for culture enrichment purposes. The mean (.~) of leaching rates
in experiments 3-30 is also given.
 L. Ahonen, O.H. Tuovinen / Hydrornetallurgy 37 (1995) 1-21 7

2). Because changes in leaching rates were evident even at low metal recoveries,
this phenomenon cannot be entirely explained by the reduced amount of mineral
material available. The decreased rates may be at least partially attributed to the
formation of a product layer on the mineral surface which creates a diffusion
barrier to the interfacial fluxes of reactants and products.

A
Ni,
Co"
• = Cu o..o--o-'°~" v''"''i
Cu 600-
= = Zn f 1 3 % Cu
800 500"
700 400. 57°,°00 - Zn

$ 6O0 300" -1200

L~500 200. Zn -1000
400 100. -800
300 0 600
200 -400
100 -200
0 i-- ~ | , i ! | , , , i !
0 50 100 150 200 250 300 350 400 450 500 550
Days

B
100 /

m.c: 8°t -"'.A

80 t .iZn J ~ ,n,..-
3~

i l i- i i i i i i i i i

600 [ mV t, ~

500 -

3oo ]
2OO
3.0 )
soo .~' ~'
2.5]
=~ 20[ 400 § --c
1.5]
1.0" I,
0 5~) 100 150 2()0 250 3(~0 350 41)0 4~50 5(~0 5Ei0

Days

Fig. 2. Results of column leaching experiment no. 14. (a) Solubilization of copper, zinc, nickel and
cobalt. (b) Changes in leaching rates, redox potential, pH and acid consumption.
8 L. Ahonen, O.H. Tuovinen/Hydrometallurgy 37 (1995) 1-21

The metal solubilization was simultaneously affected by several independent

factors. Very low leaching rates were attributed to acid-consuming gangue mate-
rial (experiments 17-19), low temperature (experiments 11 and 12), or a lack
of bacterial catalysis (chemical controls, experiments 9, 21 and 23 ) (Fig. 1 ).
Nickel leaching rate was relatively high in the beginning of the experiments,
decreasing strongly during the first half year (Fig. 2B ). The maximum of the Ni
leaching rate in the beginning of the experiments was usually accompanied by a
minor increase in cobalt leaching rate. The main phase of cobalt leaching usually
started later on. Solution analysis often showed zinc depletion in the beginning
of the experiments, but subsequently the metal was readily solubilized. In many
experiments soluble copper was virtually absent until pH adjustment to pH < 2.5
started up the solubilization.

3.2. Mineralogical observations

Microscopic examination of the partially leached ore material revealed the

common existence of Fe (III) oxide or oxyhydroxide coatings on mineral surfaces
(Fig. 3). In experiments conducted at higher pH values (pH ~ 3 ), Fe (III) pre-
cipitation in the particle interstices partially consolidated the material, suppress-
ing its fluid permeability. Fine grained, loose precipitates were also frequently

Fig. 3. Schematic diagram of mineralogical observations of leach residues. A compact Fe (III) oxide
layer (Fe-ox) is illustrated as a layer on pyrrhotite (Fel_~) surface. Pyrrhotite surfaces were char-
acteristically covered by elemental sulfur (S °) layer. Pentlandite inclusions [(Co,Ni,Fe)gSs] are shown
relatively intact to indicate preferential dissolution of pyrrhotitc. Pyrite (FeS2) grains contained fis-
sures with chalcopyrite (CuFeS2) seams, both being relatively recalcitrant to bioleaching under these
conditions. Shown also are covellite (CuS) precipitation on Fe~ _xS and residual sphalerite (ZnS) in
a fracture reaching inside Fe~_ xS.
 L. Ahonen, O.H. Tuovinen / Hydrometallurgy 37 (1995) 1-21 9

observed on the surface of the leached material. At higher pH values the color of
the precipitate was typically brown. At experiments conducted at pH values be-
low pH 2.5, bright yellow precipitates were formed. The brown precipitate ap-
peared to be amorphous in XRD analysis. The yellow precipitate was well crys-
tallized and was identified by XRD as jarosite (general formula
XFe3 (804)2(OH)6, X = K +, Na ÷, H3 O+, NH~- ). Sodium jarosite was the main
component in yellow precipitates. The major XRD reflections displayed spread-
ing, which was attributed to the presence of potassium and hydronium jarosite.
The chemical composition of some of the jarosite precipitates has been reported
previously [ 6 ].
Precipitation of a secondary, Cu sulfide, covellite, on the surface of pyrrhotite
was observed by microscopy (Fig. 3 ) and visual inspection. The presence of cov-
eUite was particularly pronounced when, in spite of the regular addition of sul-
furic acid, the solution pH remained relatively high (pH 2.5-4), owing to the
presence of alkaline gangue minerals. In some experiments the release of copper
into solution was completely suppressed. Chemical analysis of leach solution
samples suggested the precipitation of zinc in the course of the leaching but sec-
ondary zinc precipitates were not identified.
The solubility of iron monosulfide in an acid solution can be calculated from
the following equation, which predicts that ferrous iron and hydrogen sulfide
concentrations could exceed 1 m M concentration at pH 4.5 5:
FeS + 2H ÷ ~ F e 2÷ + H2 S logK= 3.1
Although thermodynamic properties of natural pyrrhotite may differ from the
values used in this calculation, it can be concluded that non-oxidative dissolution
of pyrrhotite, accompanied by the evolution of H2S, takes place in acid solutions
[81.
The solubility of cupric sulfide is several orders of magnitude lower than that
of iron sulfide and copper will precipitate if even minor amounts of hydrogen
sulfide are present:
C u S + 2 H + ~ C u 2 + + H 2 S l o g K = - 16.0
The solubility of ZnS (log K = - 4.6) is several orders of magnitude lower than
that of iron monosulfide. The exchange reaction between FeS and Zn 2+ is also
thermodynamically favorable:
FeS + Zn 2+ ~ Z n S + Fe 2+ logK= 7.7
Covellite was observed to precipitate even at high redox potentials, if the so-
lution pH was higher than about 2.5, and ferric iron was thus practically absent.
In the presence of ferric iron, covellite formation seemed to be suppressed. This
indicates that, in the absence of ferric iron, copper sulfide precipitation on the
pyrrhotite surface is a faster process than the oxidation of soluble sulfide. Ferric
iron is known to be an effective oxidizing agent for sulfide minerals but it may
also oxidize soluble sulfide.
Elemental sulfur (S °) was observed in the chemical analysis of leach residues.
10 L. Ahonen, 0.11. Tuovinen / Hydrometallurgy 37 (1995) 1-21

Elemental sulfur may be lost during preparation of polished sections for micro-
scopic determinations. Partially leached pyrrhotite grains were frequently sur-
rounded by dark rims (Fig. 3 ), containing unidentified powder material or casts
of removed substance. The same texture was observed in shake flask experiments
with pure pyrrhotite, with elemental sulfur as the main component in leach resi-
due [9 ]. The formation of elemental sulfur is attributed to partial oxidation of
pyrrhotite:
4Fel_xS+ ( 3 - 3 × ) 0 2 + ( 1 2 - 12× ) H + ~ ( 4 - 4 × ) F e 3+
+4S°+ ( 6 - 6 × ) H 2 0
The leaching of pyrrhotite seemed to be controlled by crystallographic direc-
tions; preferentially advancing parallel to the direction of pentlandite lamellae.
Comb-like textures were often observed. Pentlandite was observed to protrude
on the partially leached pyrrhotite surface, indicating its relative resistance in
comparison to pyrrhotite (Fig. 3). No significant differences in pyrrhotite leach-
ing textures between the bacterial and solely chemical leaching were observed.
Leaching of pyrite typically propagated along microfissures and grain bounda-
ries. Solid residues were not found for pyrite leaching. Textures indicating pyrite
leaching were not observed in the chemical controls.
In complex sulfide ore material, galvanic interactions between sulfide minerals
may take place. Microscopic examinations [ 10 ] showed that in chemical con-
trois chalcopyrite was preferentially leached with respect to pyrite when the two
minerals were in contact. In bacterial leaching experiments, microscopic obser-
vations on polished specimens indicated that pyrite was more corroded than
chalcopyrite. The leaching also seemed to propagate along the grain boundaries
between pyrite and chalcopyrite grains.

3.3. Acid consumption

Active bacterial leaching necessitated periodic pH control, acid or base was

added if the pH value deviated from the preferred pH range. Sulfuric acid was
added during the pre-leaching phase and, even though acid production took place
in the long-term experiments, the net process was acid-consuming in all tests. The
acid demand varied depending on the experimental conditions; being highest un-
der low pH and low redox potential conditions (Fig. 4). The acid addition aver-
aged 43 g (0.44 mol) H2SO4/kg ore [91 g (0.93 tool) H2SO4/g Cu leached],
including the initial pH adjustments. If estimated only for the pH and redox po-
tential regime, where copper leaching was fastest (pH < 2.5; redox potential > 460
mV), the acid demand averaged 81 g H2SO4/g Cu leached.
Carbonate minerals are readily solubilized in acid solutions. The leaching of
carbonate-rich gangue material, for example skarn, repeatedly increased the pH
to values which were prohibitively high (pH > 4 ) to the acid and bacterial dis-
solution of metals. Many silicate minerals are also partially solubilized in long-
term acid leaching processes. The solubility of silica is low, and secondary sili-
 L. Ahonen, O.H. Tuovinen / Hydrometallurgy 37 (1995) 1-21 11

mV

..= 18 pH
=E
o o 2.9
=o_~ 12
N~o'~ 3.3

o !
Fig. 4. Acid consumption as a function of pH and redox potential in column leaching experiments.

ceous precipitates may form. The dissolution and formation of solid phases of
silicates have yet to be investigated for bacterial leaching processes. The leaching
rates of silicates (Me2 SIP4) are much lower than those of carbonates but their
relative contribution to acid consumption is high, as is generically shown by the
following equation:
Me2SiO4+ 4 H + ~ 2Me 2+ + SiO2 + 2H2 O
During the pre-leaching phase the smell of hydrogen sulfide was sometimes
observed in the columns, indicating non-oxidative leaching of pyrrhotite. An in-
crease in acid consumption coincided with an increase in redox potential and the
concurrent leaching of nickel, suggesting partial oxidation of pyrrhotite.

3.4. Iron

A significant amount of iron was released from the sulfide minerals during the
leaching process. It is evident that a major part of the dissolved iron was subse-
quently hydrolyzed and precipitated as ferric oxyhydroxide (used collectively for
oxides, hydroxides and oxyhydroxides of ferric iron) but jarosites were also
formed:
F e O O H + H 2 0 ~ F e 3+ + 3OH-
The solubility products of ferric oxyhydroxides reported vary from 10 -36.6 to
10 -43"7, depending on the nature of crystalline phase, on the degree of crystaUin-
ity and on the crystal size [ 11 ]. Poorly crystallized ferrihydrite and lepidocrocite
are common products in the rapid hydrolysis of ferric iron. During the aging pro-
cess these will be recrystallized and partially dehydrated, slowly converting to
goethite (ot-FeOOH; log K = - 4 3 ) . The transformation to goethite is, however,
12 L. Ahonen, O.H. Tuovmen / Hydrometallurgy 37 (1995) 1-21

a slow process in acid solutions [ 12 ] and kinetic factors seem to favor the for-
mation ofjarosite. The solubility of Na-jarosite is defined by the following equa-
tion and the corresponding solubility product value was inferred from data pub-
lished by Chapman et al. [ 13 ]:
NaFe3(SO4)2(OH)6~ N a + +3Fe 3+ +2SO42- + 6 O H - logK= - 8 9
A compilation of the dissolved iron concentration values as a function of mea-
sured pH in different experiments is presented in Fig. 5a. The sulfate concentra-
tions in leaching solutions were not systematically analyzed. The total sulfate
added for pH adjustment averaged 0.5 tool/1. It was also estimated that complete
oxidation of sulfur in ore material would produce additional 0.5 mol sulfate/1.
The concentration of pure sulfuric acid at pH 2 may be taken as representing the
lowest limit of total soluble sulfate.
The respective ferric iron concentrations calculated from the total iron and
redox potential data are shown in Fig. 5b. The solubilities of ferric iron precipi-
tates in acid solutions are strongly affected by sulfate, which forms stable com-
plexes with soluble ferric iron. Thermodynamic calculations based on these data
suggested that the solubility of ferric iron was controlled by jarosite precipitation
at pH < 2.5 [ 8 ]. The concentration of dissolved ferric iron at pH > 2.5 was neg-
ligible, presumably because of precipitation of Fe (III) oxyhydroxide or schwert-
mannite, a poorly crystalline Fe (III) oxyhydroxysulfate [ 14,15 ].
Soluble iron may act as an effective leaching agent whereas iron precipitates
may suppress the dissolution processes. Fig. 6 shows the relationship between the
amount of ferric iron in solution and the leaching rates of various metals. The
average oxidation rates of Cu, Co and Zn increased when soluble iron was present
in the leach solution at low concentrations; whereas the leaching of Ni did not
exhibit any apparent correlation with the concentration of soluble iron.
The effect of iron precipitation on metal leaching rates was evaluated by com-
paring the concentration of total dissolved iron in solution with the estimated
total amount of iron released from the ore material. The estimation was based on
the known Co/Fe and Ni/Fe ratios in the minerals, and on the assumption that
Co and Ni would not be incorporated into the precipitates. There appeared to be
no relationship between the percentage fraction of iron in solution and the rate
of leaching of either Cu, Ni, Co or Zn.
Attempts were also made to precipitate ferric iron by forced aeration of column
effluents at pH > 2.5. The increase in turbidity as a result of the forced aeration
was negligible, suggesting that bacterial oxidation of iron was sufficient to main-
tain the iron in the oxidized form. In some experiments leach solutions were par-
tially replaced periodically but no discernible effect could be seen on the leaching
rates.

3.5. Bacteria

Bacterial leaching rates were deafly higher than the chemical leaching rates for
the metals which were present in sulfide minerals in the ore samples. Chalcopyr-
 L. Ahonen, O.H. Tuovinen / Hydrometallurgy 3 7 (I 995) 1-21 13

A
12-
• O •

10 .:

8 ......
041 • • •

-~ 6
-:.; ee •co e=•

• .9...:. • .
•~ J • .N~'~ ",,_ .
'=1 . ',.J'.,,;.;~l~l- .. . •
• t •"1 • e" o • •0 •

• O~ • • ee • • • •
0 - " ~"P - " J ~ " "" " " "

15 2.0 2.5 3.0 35

pH

•
#
e °
B
10

8
._

6
v
tD
LL
......-::..:...

•. . : . : . . .

1.5 2.0 2.5 3.0 3.5

pH
Fig. 5. Scatter diagram of ( a ) total dissolved iron concentration and (b) ferric iron concentration as
a function of pH in columnleachingexperiments.

ite leaching was negligible without bacteria, as also was the solubilization of co-
balt. Total iron reached higher concentrations with bacteria and the redox poten-
tial values were also higher, compared with the chemical controls. In the chemical
controls, because of the lack of bacterial ferrous iron oxidation, redox potentials
 14 L. Ahonen, O.H. Tuovinen / Hydrometallurgy 37 (1995) 1-21

1ooi 200 % .

8oi =o
= .
,-~ 160 - °•
". •• °
E .• . q --•,.;"
6oi Io ".'.• "" ~ 120-
.:-'.:•••
8 4oi Iii
% : ~.,to° ° • ~ 80-
• • • •oe ° • •
• . . ' . ' . ' . -=..t :
•

• • ",'~ll,,q, °~t, ,.o •

2oi • : --"~ • 40- ° z°°Oo °° '~o .~v,

• : .t.. :
•
: .'."
." . .: 1"~'." ". " " ; . " r ' t . ' ~
.". °~. " '*-.. Oo • . . , . ~ .. , . .
n

..... " C . ~_..'L~.:;.. :'" " ~ o

10 100 10;0 100'00 10 1O0 1000 10000

200 100

160 80

E E
~ 120 60
E E
~ 80 40
• " ....," • , 8 .'.'i.~.: .
• "" • • "" ">": 0 • ,t;;',';~'"
40 20
• ,,~ttlb • ,* ,, ",,/',-"-.'9'~ : •
- - . . ... ' -. -.,,'.,t:
. . . . .

0
10 1O0 1000 10000 10 1O0 1000 10000

Fe(lll) mg/liter

Fig. 6. Leaching rates o f copper, zinc, nickel a n d cobalt as a f u n c t i o n o f ferric iron c o n c e n t r a t i o n in

c o l u m n leaching experiments.

remained low. Acid consumption was greatly reduced in the bacterial leach
columns.
In some experiments attempts were made to enumerate bacteria by micro-
scopic counts of effluent samples during column operation. It was a general trend
with these results that higher bacterial counts were obtained at the lower pH range
and at intermediate redox potential values (Fig. 7 ). This general tendency is in
keeping with the dynamics ofbioleaching whereby active bacteria account for the
formation of oxidized chemical species in solution; the oxidation being coupled
to growth, and thus to cell density. The bacterial counts are based only on free-
swimming bacteria and no attempt was made to determine bacterial densities
either qualitatively or quantitatively on mineral surfaces. At present, there is not
sufficient information to evaluate the relative role of cell attachment in contrib-
uting to bacterial counts in active bioleaching systems.
3.6. Ore~liquid~air contact
In initial experiments, the daily column operation was based on either: ( 1 ) 2 h
circulation followed by 22 h draining; or (2) continuous circulation of the leach
 L. Ahonen, 0.11, Tuovinen / Hydrometallurgy 37 (1995) 1-21 15

mV
300 340 380 420 460 500 ,54.0 ~ . 7

E 600, .9

rn 400 3.3

$ 200
.~ 8o0

rn 1-- 400

i°°
Fig. 7. Bacterial counts as a function o f p H and redox potential in column leaching experiments.

solution while the ore was completely immersed and in full contact with the leach
solution (flood leaching). The continuous circulation regime was found to be
more effective.
Column design was subsequently modified to approximate trickle leaching
conditions by raising the level of the fritted base in the column. This modification
necessitated the insertion of a port below the base for pressure equilibration in
the effluent holding zone. Larger columns, approximating trickle leaching con-
ditions, were also set up to accommodate 12 kg ore samples and 3 1 of leach
solution.
A direct comparison with the two column designs, trickle leaching versus flood-
ing (immersion), is difficult because of a number of other parameters that varied
at the same time. If the comparison is based on the same ore material and ap-
proximately similar pH profiles under flood and trickle leaching conditions, the
trickle leaching yielded approximately three times faster rates of copper and co-
balt solubilization (experiments 20 and 26 at pH 2.5; experiments 22 and 27 at
pH 2 ). More or less comparable leaching rates were obtained for nickel and zinc
dissolution with the two techniques. Compared with trickle leaching, acid con-
sumption was higher under flood leaching conditions. Trickle leaching consis-
tently involved higher redox potential values (difference ~ 100 mV) and the sys-
tem became acid-producing. The flood leaching, on the other hand, displayed
prolonged acid demand; evidence for the dissolution of acid-consuming min-
erals. Trickle leaching facilitates oxidative processes because of improved oxygen
transfer conditions, whereas a flood leaching system gives better contact between
mineral surface and water, thereby facilitating the leaching of pyrrhotite and non-
oxidative dissolution of carbonate minerals.
16 L. Ahonen, O.H. Tuovinen / Hydrometallurgy 37 (1995) 1-21

3. 7. Redox potential and p H

The redox potential and pH values during leaching experiments varied for many
reasons. Prior to inoculation, the columns were pre-leached with sulfuric acid
(pH ~ 2 ) in order to neutralize the alkaline components of the ore material. Bac-
terial oxidation of ferrous iron to the ferric form usually took place during the
first weeks of the experiment, increasing the redox potential from initial values
of ~ 300-400 mV to > 500 mV. The increase in redox potential was usually ac-
companied by an increase in acid consumption and in metal solubilization. While
the redox conditions were mainly controlled by bacterial activity and oxygen sup-
ply, the solution pH was usually maintained at predetermined levels.
Fig. 8 illustrates the general effects of pH and redox potential on the leaching
of various metals. The average rates of solubilization of Co and Cu were highest
at high redox potential and low pH values, while the leaching of Ni displayed
practically no pH or redox potential dependence. The average copper leaching
rate increased rapidly at pH < 2.5 and at redox potential values > 500 mV. Rea-
sonable Zn leaching rates were obtained at lower redox potential values ( ~ 400-
500 mV) and they were also less dependent on solution pH. Compared with the
other metals, the rate of cobalt leaching displayed a more apparent dependence
on the change in pH and redox potential.
The general trends observed here can be attributed to the electrochemical na-
ture of the respective sulfide minerals. The sulfide minerals studied here are ther-
modynamically unstable in oxidative conditions. A generalized half-cell reaction
describing the complete oxidation and leaching of a sulfide mineral (MeS) can
be written as:
MeS + 4H20 ~ Me 2+ + SO42- + 8e- + 8H +

Assuming activities of [ Cu ] = 0.01, [ Zn ] = 0.0 l, [ Fe ] = 0.1, [ SO 2- ] = 0.1 and

pH = 2, the following thermodynamic equilibrium potentials (SCE) were calcu-
lated from the respective equilibrium constant (K) values: chalcopyrite = 273 mV;
pyrite = 267 mV; sphalerite = 232 mV; and pyrrhotite = 167 mV. The solution re-
dox potential values measured in the experiments clearly exceeded the calculated
equilibrium potentials.
An updated compilation of measured sulfide mineral single electrode poten-
tials is given by Rossi [ 1 ]. Values from 500 to 600 mV (standard hydrogen scale,
corresponding to about 250-350 mV SCE) have been measured for pyrite, with
slightly lower values for chalcopyrite. This is in agreement with the observations
that galvanic coupling between pyrite and chalcopyrite promotes chalcopyrite ox-
idation [ 16,17 ]. In the present experiments, galvanic effects between pyrite and
chalcopyrite were clearly demonstrated with chemical controls, in which redox
potentials were typically low. A major source of Co in this ore was pyrite, which
has the highest electrode potential in the electrochemical series of the common
sulfide minerals [ 1,18 ]. In bacterially catalyzed leaching, when redox potential
and ferric iron concentrations were high, rapid pyrite leaching appeared to pre-
 L. Ahonen, O.H. Tuovinen / Hydrometallurgy 3 7 (1995) 1-21 17

~ ~ o ~ ~ ~ o

E
8
.=.
..,=,

0
Ca,

qluotu~l/~u uz

e~
0e~

0
 18 L. Ahonen, O.H. Tuovinen / Hydrometallurgy 3 7 (1995) 1-21

Particle size (ram) Particle size (ram)

20 10 5 1.68 20 10 5 1.68

ao ~ : 7 pH 12o ~l~

_ so IIV////~.llll]~~ao _ ~ 3
40 ~0 J 100

Fig. 9. Leaching rates of copper, zinc, nickel and cobalt as a function of particle size fraction in column
leaching experiments.

dominate. Pyrrhotite leaching did not show any dependence on the redox poten-
tial, suggesting that the rate-determining step is not a redox reaction.
Pyrite and chalcopyrite have higher electrode potentials and are more recalci-
trant than sphalerite, pyrrhotite and pentlandite. Pyrrhotite (the main source of
Ni in this ore) has a low electrode potential and is more readily dissolved than
CuFeS2 and FeS2, as is apparent from the broad pH-redox potential range where
nickel dissolution took place.
The electrochemical control of leaching in these column experiments is ob-
scured by several factors. For example, the minerals may mask each other in ore
particles, in which case sequential chemical leaching may override galvanic ef-
fects. Bacterial distribution and indiscriminate oxidative attack of metal sulfides
 L. Ahonen, O.H. Tuovinen / Hydrometallurgy 37 (1995) 1-21 19

may also obscure effects attributable to the leaching of minerals in an electro-

chemical series.

3.9. Particle size

Fig. 9 shows the effect of particle size on metal leaching rates, summarized
from column leaching experiments 14-16 and 17-19. The leaching rates for all
metals increased with a smaller particle size fraction and this effect was amplified
at decreasing pH values. This trend is attributed to the increased solubility of
Fe 3+ and higher redox potential at low pH values which, together with a geomet-
ric increase in the available surface area, cause the positive rate effect. Copper
and nickel leaching rates approximately doubled when the particle diameter was
decreased from 5-10 m m to 1.68-5 mm. Concurrent zinc dissolution displayed
a five- to seven-fold increase in the respective rate and that of Co was accelerated
over ten-fold.

4. Concluding remarks

The results demonstrate that there are several interacting parameters in the
bacterial leaching of sulfide minerals. Low leaching rates were typical with acid-
consuming gangue material, at low temperatures, or in the absence of bacterial
catalysis. Hydrogen sulfide formation was evident; suggesting non-oxidative
leaching, which was attributed to the presence ofpyrrhotite. Owing to the toxicity
of H2S, the non-oxidative leaching in the initial stages may be detrimental to
bacterial activity. Active leaching of pyrrhotite, most important in the initial
phase, was a net acid-consuming reaction. When pyrite oxidation was the major
oxidative reaction, the leaching of the ore material became a net acid-producing
phase. The length of the acid-consuming and acid-producing phases during the
leaching was controlled by the relative amount of sulfide minerals, the tempera-
ture and the presence ofgangue minerals.
Ferric iron is a key component in sulfide mineral leaching. It is important to
ensure the bacterial regeneration of ferric iron, otherwise the leaching becomes
non-oxidative and occurs at a low redox potential level. With little or no bacterial
activity, ferrous iron predominates as the redox species and oxidative leaching is
almost negligible. Of the various sulfide minerals present in the ore, pyrrhotite
leaching was the least sensitive to low redox potential. However, for all sulfide
minerals present in this ore, the data emphasize that the fastest leaching rates are
achieved under oxidized conditions (high redox potential) and at low pH values,
where a considerable portion of ferric iron remains in solution; that is, under
conditions where bacterial activity is responsible for oxidative leaching of sulfide
minerals. The relationships between leaching rates and pH, redox potential and
ferric iron in solution, which vary for each discreet mineral, suggest that there is
no uniform mechanism of mineral dissolution in biological leaching. The appar-
ent lack of a universal mechanism to explain the biological leaching of sulfide
20 L. Ahonen, 0.1t. Tuovinen / Hydrometallurgy 37 (1995) 1-21

minerals in this work warrants fundamental studies of biological and electro-

chemical mechanisms to understand, optimize and develop models for biological
leaching processes.

Acknowledgements

The experimental work was carried out in the Department of Microbiology,

University of Helsinki. We are grateful to Ms. Paula Hiltunen for her skillful par-
ticipation in the planning and experimental phases of the work, Ms. Sinikka Pah-
kala for technical assistance and Dr. Martti Lehtinen (Geological Museum, Uni-
versity of Helsinki) for XRD analysis of leach residues. We appreciate the
analytical services and the collaboration with several experts from Outokumpu
Oy during the experimental phase of the project. The work was funded by the
Ministry of Trade and Industry (Finland). Additional financial support was re-
ceived from Outokumpu Research Oy and the Nordisk Industrifond.

References

[ 1 ] Rossi, G., Biohydrometallurgy. McGraw-Hill, Hamburg (1990).

[ 2 ] Peltola, E., Origin of Precambrian copper sulfides of the Outokumpu district, Finland. Econ.
Geol., 73 (1978): 461-477.
[ 3 ]Koistinen, T.J., Structural evolution of an early Proterozoic strata-bound Cu-Co-Zn deposit,
Outokumpu, Finland. Trans. R. Soc. Edinburgh Earth Sci., 72 ( 1981 ): 115-158.
[4 ]Ahonen, L. and Tuovinen, O.H., Temperature effects on the microbiologicalleaching of complex
sulfide ore material in percolators containing chalcopyrite, sphalerite, pentlandite, and pyrrho-
rite as main minerals. Biotechnol. Lett., 11 ( 1989): 331-336.
[5 ]Ahonen, L. and Tuovinen, O.H., Bacterial oxidation of sulfide minerals in column leaching ex-
periments at suboptimal temperatures. Appl. Environ. Microbiol., 58 ( 1992): 600-606.
[ 6 ] Ahonen, L. and Tuovinen, O.H., Silver catalysis of the bacterial leaching of chalcopyrite-contain-
ing ore material in column reactors. Miner. Eng., 3 (1990): 437-445.
[ 7 ] Brookins, D.G., Eh-pH Diagrams for Geochemistry. Springer, Berlin ( t 988).
[ 8 ] Ahonen, L. and Tuovinen, O.H., Solid-phase alteration and iron transformation in column bio-
leaching of a complex sulfide ore. In: C.N. Alpers and D.W. Blowes (Editors), Environmental
Geochemistry of Sulfide Oxidation. Am. Chem. Soc., Washington, D.C. (1994), pp. 79-89.
[9]Ahonen, L., Hiltunen, P. and Tuovinen, O.H., The role of pyrrhotite and pyrite in the bacterial
leaching of chalcopyrite. In: R.W. Lawrence, R.M.R. Branion and H.G. Ebner (Editors), Fun-
damental and Applied Biohydrometallurgy. Elsevier, Amsterdam (1986), pp. 13-22.
[ 10 ] Ahonen, L. and Tuovinen, O.H., Alterations in surfaces and textures of minerals during the bac-
terial leaching of a complex sulfide ore. Gcomicrobiol. J., 10 ( 1993 ): 207-217.
[ 11 ] Macalady, D.L., Langmuir, D., Grundl, T. and Elzerman, A., Use of model-generated Fe3+ ion
activities to compute Eh and ferric oxyhydroxide solubilities in anaerobic systems. In: D.C. Mel-
chior and R.L. Bassett (Editors), Chemical Modeling of Aqueous Systems II. Am. Chem. Soc.,
Washington, D.C. (1990), pp. 350-367.
[ 12 ] Schwertmann, U. and Murad, E., Effect of pH on the formation of goethite and hematite from
ferrihydrite. Clays Clay Miner., 31 ( 1983): 277-284.
[ 13 ] Chapman, B.M., Jones, D.R. and Jung, R.F., Processes controlling metal ion attenuation in acid
mine drainage streams. Geochim. Cosmochim. Acta, 47 ( 1983): 1957-1973.
 L. Ahonen, O.H. Tuovinen / Hydrometallurgy 37 (1995) 1-21 21

[ 14]Bhatti, T.M., Bigham, J.M., Carlson, L. and Tuovinen, O.H., Mineral products of pyrrhotite
oxidation by Thiobacillusferrooxidans. Appl. Environ. Microbiol., 59 ( 1993): 1984-1990.
[ 15 ] Bigham, J.M., Carlson, L. and Murad, E., Schwertmannite, a new iron oxyhydroxysulfate from
Pyhiisalmi, Finland, and other localities. Miner. Mag. (1994), in press.
[ 16 ] Berry, V.K., Murr, L.E. and Hiskey, J.B., Galvanic interactions between chalcopyrite and pyrite
during bacterial leaching of low grade wastes. Hydrometallurgy, 3 (1978): 309-326.
[ 17 ] Mehta, A.P. and Murr, L.E., Fundamental studies of the contribution of galvanic interaction to
acid-bacterial leaching of mixed metal sulfides. Hydrometallurgy, 9 ( 1983): 235-256.
[ 18 ] Natarajan, K.A., Electrochemical aspects of bioleaching of base-metal sulfides. In: H.L. Ehrlich
and C.L. Briedey (Editors), Microbial Mineral Recovery. McGraw-Hill, New York (1990), pp.
79-106.