Mineral Biotechnology Microbial

Mineral Biotechnology:
Microbial Aspects of Mineral Beneficiation,

Metal Extraction, and Environmental Control
Edited by S.K. Kawatra and K.A. Natarajan
Published by the
Society for Mining, Metallurgy, and Exploration, Inc.

8307 Shaffer Parkway
Littleton, CO 80127
© 2013 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Society for Mining, Metallurgy, and Exploration, Inc. (SME)
8307 Shaffer Parkway
Littleton, Colorado, USA 80127
(303) 973-9550 / (800) 763-3132
www.smenet.org
SME advances the worldwide minerals community through information exchange and
professional development. With members in 50 countries, SME is the world’s largest
professional association of mineral professionals.
Copyright © 2001 Society for Mining, Metallurgy, and Exploration, Inc.

Electronic edition published 2013.
All Rights Reserved. Printed in the United States of America
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in

any form or by any means, electronic, mechanical, photocopying, recording, or otherwise,
without the prior written permission of the publisher.
Disclaimer
The papers contained in this volume are published as supplied by individual authors. Any
statement or views presented here are those of individual authors and are not necessarily those
of the Society for Mining, Metallurgy, and Exploration, Inc. The mention of trade names for
commercial products does not imply the approval or endorsement of SME.
On the Cover
Inset: Photomicrograph of Thiobacillus ferrooxidans. Together with Leptospirillum ferrooxidans
and Thiobacillus thiooxidans, these mesophillic bacteria are the main constituents of a mixed
bacterial culture used in the BIOX• process. Top: The Fairview BIOX• plant in South Africa was
the first commercial biooxidation plant for refractory gold concentrate in the world. It was
commissioned in 1986 and is currently treating 55 metric tons per day of concentrate. Bottom:
The Harbour Lights BIOX• plant near Leonora, in Western Australia was commissioned in
1992 and treated 40 metric tons per day of a refractory pyrite/arsenopyrite concentrate.
ISBN 0-87335-201-7
ISBN-13: 978-0-87335-201-7
Ebook: 978-0-87335-378-6
Library of Congress Cataloging-in-Publication Data

Mineral biotechnology : microbial aspects of mineral beneficiation, metal extraction, and
environmental control / edited by S.K. Kawatra and K.A. Natarajan.
p. cm.
Includes bibliographical references and index.
ISBN 978-0-87335-201-7 -- ISBN 978-0-87335-378-6 (Ebook)
1. Bacterial leaching. I. Kawatra, S.K. II. Natarajan, K.A.
TN688.3.B33 M57 2001

622'.7--dc21
2001018684
Contents
FOREWORD v
SECTION 1 BIOBENEFICIATION 1
Depression of Pyrite Flotation by Yeast and Bacteria 3
S.K. Kawatra and T.C. Eisele
Calcium Removal from Bauxite Using Paenibacillus polymyxa 13
J.M. Modak, S.S. Vasan, and K.A. Natarajan
Desulfurization of Coal by Microbial Flotation in a
Semicontinuous System 27
T. Nagaoka, N. Ohmura, and H. Saiki
Biobeneficiation of Mineral Raw Materials 37
S.N. Groudev
Role of Corundum-Adapted Strains of Bacillus polymyxa in the
Separation of Hematite and Alumina 55
Namita Deo and K.A. Natarajan
Role of a Heterotrophic Paenibacillus polymyxa Bacteria in the
Bioflotation of Some Sulfide Minerals 67
P.K. Sharma and K. Hanumantha Rao
SECTION 2 BIOLEACHING 83
Commercialization of Bioleaching for Base-Metal Extraction 85
P.C. Miller, M.K. Rhodes, R. Winby, A. Pinches,
and P.J. van Staden
Microbiological Leaching of Uranium Ores 101
O.H. Touvinen and T.M. Bhatti
Advances in the Application of the BIOX® Process for Refractory
Gold Ores 121
P.C. van Aswegen and H.J. Marais
SECTION 3 BIOREMEDIATION 135

Degradation of Metal Cyanide Complexes by Microorganisms 137
R. Fedel-Moen, S.R. Ragusa, R.W.L. Kimber, and B.D. Williams
Biochemical Removal of HAP Precursors from Coal—INEEL Slurry
Column Testing 153
K.S. Noah and G.J. Olson
Microorganisms, Biotechnology, and Acid Rock Drainage—Emphasis
on Passive-Biological Control and Treatment Methods 169
N. Kuyucak
iii
Sawdust-Supported Passive Bioremediation of Western United States
Acid Rock Drainage in Engineered Wetland Systems 189
D.N. Thompson, R.L. Sayer, and K.S. Noah
Biotechnologies for Remediation and Pollution Control in the
Mining Industry 207
L. Bernoth, I. Firth, P. McAllister, and S. Rhodes
SECTION 4 BIOMINERALIZATION 219

Utility of Bioreagents in Mineral Processing 221
P. Somasundaran, Namita Deo, and K.A. Natarajan
Effect of Mesophilic Microorganisms on the Electrochemical Behavior
of Galena 229
J.L. González-Chávez, F. González, A. Ballester, and M.L. Blázquez
Ocean Manganese Nodules: Biogenesis and Bioleaching
Possibilities 239
H.L. Ehrlich
Immobilization of Free Ionic Gold and L-Asparagine-Complexed Ionic
Gold by Sporosarcina ureae: The Importance of Organo-Gold
Complexes in Gold Mobility 253
G. Southam, W.S. Fyfe, and T.J. Beveridge
INDEX 261
iv
Foreword
This new millennium rightly belongs to biotechnology, and rapid progress in minerals pro-
cessing based on biological principles is just around the corner. Materials processing by
microorganisms is reported to have begun almost 3.5 billion years before human interven-
tion. Over geological periods of time, the tiny earthly microorganisms have evolved to pro-
vide process energy, use waste metal compounds, and produce enormous quantities of
valuable minerals. In addition, microorganisms can live in hostile environments.
The interactions among biotechnology, metals, and minerals are relevant to a number of
emerging interdisciplinary areas:
Biogenesis and biomineralization
Biomaterials processing and biomimetics
Ceramics and biomedical engineering
Biomineral beneficiation
Bioleaching
Biocorrosion, biofouling, and biodeterioration
Bioenvironmental control
Amazing processing functions are incorporated into a tiny microbial cell—accumulation of
metal ions, generation and synthesis of polymers and mineral composites, and synthesis of a
variety of biopolymers and catalysts. Sensors, regulators, and adaptive machinery associated
with microorganisms make them akin to a modern microprocessor-controlled biochemical
factory. With the advent of genetic engineering, it has become possible to modulate microor-
ganisms to make them perform a desired function at faster rates. For example, because metal
resistance in bacteria is known to be plasmid-mediated, it may become possible to develop
“super bugs” through plasmid transfer.
In nature, biological systems are replete with examples of organic and inorganic supramo-
lecular assemblies and superior architectural styles. Unique and exquisite biominerals such
as diatoms, coccoliths, seashells, and bones exhibit controlled processing with respect to
structure, size, shape, orientation, and texture. Biological routes are used to fabricate tough,
durable, and adaptive polymer–ceramic composites under natural conditions. There is a
strong interrelationship between biomineralization and materials chemistry based on biomi-
metics and molecular tectonics.
Microbes also participate in lithification, mineral formation, conversion, precipitation, and
transport in nature, as well as in mineral diagenesis and sedimentation. Natural metal–
microbe cycles are responsible for the various biochemical and geochemical reactions that
lead to the formation of mineral deposits. Among the essential elements required for living
organisms are carbon, oxygen, hydrogen, silicon, magnesium, phosphorus, calcium, iron, and
manganese—all common ingredients of biominerals. Calcium biominerals predominate. Many
elements such as copper, zinc, and lead are deposited on the external surfaces of bacteria as
sulfides. Biomineralization takes place in well-defined spatially delineated sites and involves
molecular construction of discrete, self-assembled, organic supramolecular systems. Intimate
association of inorganic and organic phases is biomineralization’s hallmark. Possibilities for
modifying crystal shapes by the interaction of soluble molecules with crystal faces do exist;
crystal formation can either be promoted or inhibited biologically.
More than 50 biominerals, which include various metal carbonates, phosphates, halides,
oxalates, oxides, and sulfides, have been recorded. Elemental selenium, tellurium, gold, sil-
ver, sulfur, and mercury are biogenic. Minerals are generated by most groups of organisms
such as bacteria, fungi, algae, and plants. The scale of biogenic production in comparison to
human usage is indeed staggering. For example, the biogenic generation rate for CaCO3 is
estimated to be about 5 × 1012 kg/yr and that of SiO2 is about 5 × 1011 kg/yr. Structurally
ordered materials in nature are produced by higher organisms such as mollusks. Structural
polymer–mineral composites are composed through a polymer framework. Extracellular
microstructures are assembled that resemble a brick wall with the mortar (polymer) laid
before the ordered bricks of calcium mineral, and elegant structures of silica and iron oxides
are biologically constructed.
Various biopolymers find applications in ceramic and mineral processing. Several types of
polyanionic proteins can modify or inhibit crystallization, and can alter the crystal morphol-
ogy as well. Large-scale commercial applications of such bioreagents include antisealants,
dispersants, and antifreezes. Ceramics processing with biogenic additives has thus become
possible. Biopolymers can be used to control plasticity of clays, and colloidal stability of fine
particles can be enhanced through biogenic additions. Dispersion or flocculation of ceramic
and mineral particles can be achieved through biological treatment.
Microorganisms that inhabit ore bodies and water systems can play a significant role in
causing environmental pollution, such as acid water generation. Acidic waters originating
from bacterial sulfidic mineral oxidation, especially from metalliferrous and coal mines, con-
stitute a major source of environmental pollution. Both active and abandoned mines remain
a source of this problem—termed acid mine drainage (AMD)—which leads to contamination
of groundwater tables, rivers, streams, and even seacoasts.
At the same time, appropriate use of various microorganisms can bring about environmen-
tal protection. In nature, the iron- and sulfur-oxidating group of bacteria called Thiobacillus
is associated with mineral sulfides such as arsenopyrite, pyrite, chalcopyrite, sphalerite,
galena, molybdenite, millerite, orpiment, and antimonite, all of which serve as energy
sources for the microbes. The abundance of iron and sulfur in natural sulfide mineralization
makes it easier for the Thiobacillus group of bacteria to colonize on them. Biooxidation of
pyrite and sulfur leads to the formation of sulfuric acid containing Fe+3, which subsequently
dissolves various toxic metal ions through its solvent action.
The microorganism Thiobacillus ferrooxidans, which is known to be effective in the leach-
ing of several minerals, was first isolated in the laboratory in 1947 from the AMD of bitumi-
nous coal mines. Throughout the world, bioleaching processes are increasingly used as
alternative and supplementary methods because high-grade ore reserves are being continu-
ously depleted, energy costs are increasing, and environmental preservation is an all-
encompassing issue.
Biohydrometallurgical extraction of metals from a wide variety of ores is being commer-
cially practiced all over the world—in Canada, the United States, Russia, South America, Aus-
tralia, and a few European countries. Large quantities of copper, uranium, and gold ores are
processed by microbial technology on an industrial scale, and the recovery of several other
metals is also possible using such methods.
In situ, dump, and heap leaching are the techniques practiced in microbial leaching. In
particular, heap leaching in the presence of bacteria is a very useful method for recovering
copper, uranium, and gold from its low-grade ores and tailings. The contribution of bioleach-
ing is estimated to be approximately 15%, 13%, and 25% of the total world production of
copper, uranium, and gold, respectively.
vi
The following features of microorganisms are significant in detoxification of liquid and
solid effluents from mining and mineral processing industries:
Removal of dissolved metal ions even at low part per million levels
Concentration of accumulated metals for recovery
Degradation of toxic organic chemicals from effluents to inert products
At least four major mechanisms have been identified for biological removal of metal ions
from liquid effluents—bioadsorption, bioaccumulation, precipitation, and volatilization. Bio-
adsorption is primarily an adsorption-type phenomenon that takes place through electro-
static attraction of metal cations to the negatively charged cell surfaces. The chemical
composition of the bacterial cell wall also plays a role in bioadsorption through metal bind-
ing to exopolysaccharides, proteins, and other functional groups. Bioaccumulation is the pro-
cess of metal uptake by living microorganisms, which are dependent on metabolic energy.
This requires specific transport systems and depends on metal tolerance of the organisms.
Inter- as well as intracellular accumulation can occur. In biodegradation, the microorganisms
transform the organic chemicals into innocuous forms, degrading them to carbon dioxide
and water. The microorganisms can also decompose the organic chemicals anaerobically.
In summary, microorganisms can play a beneficial role in all facets of minerals processing,
from mining to waste disposal and management. In this publication, the utility of mineral
biotechnology as an emerging area is illustrated with respect to biobeneficiation, bioleach-
ing, and bioremediation.
We would like to thank all the authors for contributing valuable technical papers, and also
extend our thanks to Curtis J. Tompkins for financial support and continuous encouragement.
S. Komar Kawatra
K.A. Natarajan
vii
SECTION 1
Biobeneficiation
Depression of Pyrite Flotation by Yeast and Bacteria 3
Calcium Removal from Bauxite Using Paenibacillus polymyxa 13
Desulfurization of Coal by Microbial Flotation in a Semicontinuous System 27
Biobeneficiation of Mineral Raw Materials 37
Role of Corundum-Adapted Strains of Bacillus polymyxa in the Separation of

Hematite and Alumina 55
Role of a Heterotrophic Paenibacillus polymyxa Bacteria in the Bioflotation of

Some Sulfide Minerals 67
Depression of Pyrite Flotation
by Yeast and Bacteria
S.K. Kawatra* and T.C. Eisele*
ABSTRACT
Microorganisms, such as the bacteria Thiobacillus ferrooxidans, have been reported to depress pyrite
flotation (Elzeky and Attia, 1987; Atkins, 1990). However, the dependence of the depression on the
type of organism or on suspension pH is still being determined. In this study, the relative effectiveness of
various microorganisms (chemolithotrophic bacteria, chemoorganotrophic bacteria, and yeast) over
the pH range of 2 to 12 was studied. Screening tests using microflotation showed that every microor-
ganism tested was capable of depressing naturally hydrophobic pyrite at acidic pH. Larger-scale experi-
ments with both mineral pyrite and coal pyrite using Thiobacillus ferrooxidans and Saccharomyces
cerevisiae as depressants, showed that these microorganisms are very effective depressants for mineral
pyrite at acid pH, but they are largely ineffective at neutral and alkaline pH, where the mineral pyrite
surface is not naturally hydrophobic. The flotation response of the coal pyrite was completely different
from the mineral pyrite. The coal pyrite was most floatable near neutral pH, with the floatability
decreasing in acidic or alkaline solutions. Depression of the coal pyrite by yeast was not selective
between the pyrite and the associated coal under the experimental conditions.
INTRODUCTION
Pyrite depression is of great interest in sulfide-mineral flotation, where pyrite is a common
gangue mineral. Because the floatability of pyrite is a function of the pH and a variety of
other chemical factors, pH control is a critical part of sulfide-flotation circuits. A wide variety
of pyrite-depressing chemicals have been studied and used in sulfide mineral processing,
with considerable success. However, there are certain applications where there are still no
satisfactory pyrite depressants, and, even when suitable depressants are available, there may
be a need for less-expensive reagents.
One condition where pyrite depression is very difficult is when the material being treated
has a high hydrocarbon content, with an extreme example being coal flotation (Yancey and
Taylor, 1935; Baker and Miller, 1971; Raleigh and Aplan, 1991). The main difficulty with
chemical pyrite depressants in coal flotation has been that their benefits are small to nonex-
istent in industrial-scale coal-flotation operations (Kawatra et al., 1991). Several investiga-
tors have, therefore, turned to the study of microorganisms, particularly the bacteria
Thiobacillus ferrooxidans, as pyrite depressants (Capes et al., 1973; Townsley and Atkins,
* Department of Mining and Materials Process Engineering, Michigan Technological University.
1986; Atkins et al., 1987; Elzeky and Attia, 1987; Stainthorpe, 1989; Atkins, 1990; Ohmura
et al., 1993). The basis for this is that certain microorganisms have been suggested to have a
specific affinity for pyrite surfaces (Elzeky and Attia, 1987), and, therefore, it is hoped that
they can be made to attach to the pyrite in large numbers with a higher degree of selectivity
than is possible with chemical depressants. If hydrophilic organisms are used, then once the
pyrite surface is completely covered by microorganisms, the pyrite recovery by froth flotation
would be depressed.
In this paper, the ability of several different microorganisms to depress pyrite flotation is
evaluated using two carbon-free mineral pyrites and a coal pyrite under a range of pH condi-
tions. The goal was to determine what microorganism types are most effective for this type of
work and to determine whether the effect varies significantly with changes in pH or with
pyrite type.
THEORETICAL DISCUSSION
There are three ways in which a pyrite particle can reach the froth product in flotation: flota-
tion, entrainment, and locking (or “induced floatability”). For flotation to occur, the pyrite
surfaces must become hydrophobic, so that they can attach directly to the air bubbles.
Entrainment occurs when small particles are suspended in the water making up the bubble
films and are not actually attached to the bubbles. Locking of pyrite particles to floatable par-
ticles will allow even large pyrite particles to be carried into the froth.
If pyrite is reaching the froth product by flotation, its recovery can be significantly affected
by chemical depressants, which work by altering the pyrite surface chemistry to make it less
hydrophobic. However, if the pyrite reaches the froth by entrainment or locking, then pyrite
depressants will have no effect. Both entrainment and locking are mechanical effects that
have little to do with the hydrophobicity of the particles involved, although locked particles
generally do not float as rapidly as completely hydrophobic particles. Entrainment can only
be reduced by reducing the amount of water carried into the froth or by increasing the set-
tling rate of the particles. The effects of locking can only be reduced if the ore is either
ground to a finer size or if some recovery is sacrificed to avoid floating all of the locked parti-
cles (Lynch et al., 1981; Kawatra et al., 1991).
The flotation of pyrite from both synthetic coal/pyrite mixtures and high-sulfur coals has
reportedly been depressed by the bacterium Thiobacillus ferrooxidans (Elzeky and Attia,
1987). This organism has been extensively studied for its ability to dissolve pyrite by using it
as an energy source. It was, therefore, originally investigated as a pyrite depressant because
it has a biological reason for having a specific affinity for pyrite surfaces. It has been reported
(Elzeky and Attia, 1987) that this bacterium specifically and selectively attaches to pyrite
within a matter of minutes. Early work by Elzeky and Attia (1987) was carried out by condi-
tioning coal/pyrite slurries with bacteria at pH 2, followed by flotation over the pH range of 7
to 9. The rationale was that T. ferrooxidans would be most active at a pH 2, while it is known
that coal is most floatable at near-neutral pH. Other investigators have generally used a pH
that was near 2 for both bacterial conditioning and for the flotation step (Townsley and
Atkins, 1986; Atkins et al., 1987; Stainthorpe, 1989; Atkins, 1990; Ohmura et al., 1993). It
has also been reported that organisms other than T. ferrooxidans are effective pyrite depres-
sants, as are a number of miscellaneous proteins, such as those found in whey (a byproduct
of cheese making) (Stainthorpe, 1989; Atkins, 1990).
The research described in this paper was intended to address the following questions,
which have not been completely answered by the studies reported in the literature: What
microorganisms are most effective for pyrite depression? What is the effect of pH on the
depression of pyrite flotation by microorganisms? The answers to these questions will deter-
mine whether microbial pyrite depression is potentially a useful industrial technique or sim-
ply a laboratory curiosity.
EXPERIMENTAL PROCEDURE
Materials
Three materials were used in the froth flotation studies: a high-purity mineral pyrite, a
lower-purity mineral pyrite (from a different source) and a coal pyrite.
The mineral pyrites were purchased from Wards Natural Science Establishment. The
high-purity pyrite was collected from a deposit in Rico, CO, and the low-purity pyrite was
collected rom a deposit in Custer, SD. Both pyrites were stage ground and screened to max-
imize the amount in the 105 × 74-µm size fraction. This size fraction was used in all experi-
ments with these materials, because it was fine enough to be floatable but coarse enough to
minimize entrainment effects. It was then stored at –20°C until needed. The Rico pyrite
sample in this size fraction was over 95% pyrite (by weight), and the Custer pyrite sample in
this size fraction was 52% pyrite (by weight), with the remainder being calcium carbonate
and silicates.
The coal pyrite was hand-collected as coarse nodules from the jig refuse belt at the Empire
Coal Mine, Gnadenhutten, OH, which processes a mixture of bituminous coals from the
Lower Kittanning (#5) seam, Middle Kittanning (#6) seam and Upper Freeport (#7) seam.
The hand-picked pyrite was then stage-ground, screened, and stored in the same manner as
the mineral pyrite. The 105 × 74-µm size fraction of the coal pyrite was 70% pyrite, with the
remainder being mainly coal and clay.
Organisms
Several species of microorganism were used in the microflotation screening tests. These were
selected to cover a wide range of microorganism types, including Pseudomonas fluorescens (a
gram-negative strict aerobe), Lactobacillus acidophilus (a gram-positive anaerobe),
Staphylococcus epidermis (a gram-positive aerobe), Klebsiella terrigena (a gram-negative
nitrogen-fixing aerobe), two strains of Thiobacillus ferrooxidans (a gramnegative lithotroph)
and Saccharomyces cerevisiae (a yeast). These organisms, with the exception of Thiobacillus
ferrooxidans, were taken from standard stock cultures maintained by the Environmental
Microbiology Laboratory, Department of Biological Sciences, Michigan Technological Uni-
versity. The Thiobacillus ferrooxidans culture was originally obtained from EG&G Idaho (cul-
ture DSM 83) and was subsequently adapted to grow on pyrite to create strain PA 1. The
Thiobacillus strains were grown in T&K nutrient medium with the following composition
(Tuovinen and Kelly, 1973):
Ammonium sulfate, (NH4)2 SO3 0.4 g
Potassium phosphate, K2HPO4 0.4 g
Magnesium sulfate, MgSO4·7H2O 0.4 g
Ferrous sulfate, FeSO4·7H2O 33.3 g
Distilled, sterilized water 1,000 mL
Sulfuric acid, H2SO4 to pH 2.0
The other organisms were grown in standard nutrient broth, which is a general-purpose cul-
ture medium that consists of 5 g/L peptone and 3 g/L beef extract in distilled water. This
medium was purchased premixed from DIFCO, Detroit, MI. All organisms were cultured on
an orbital flask shaker at 200 rpm and at 30°C. After growing to a sufficiently high cell den-
sity, the non-Thiobacillus microorganisms were all washed and rinsed twice by centrifuging
for 20 minutes at 15,000 G (gravities) and resuspending in distilled water at 4°C. Following
the second rinse, the cells were diluted to 2 × 108 cells/mL, based on cell counts using a
Petroff-Hauser counting chamber under a phase-contrast microscope at 400x magnification.
The Thiobacillus strains were treated similarly, except that they were washed in distilled
water with the pH adjusted to 2 with sulfuric acid. This was necessary because T. ferrooxidans
is an obligate acidophile, and the cells are killed by near-neutral pH. When the pH is raised to
near neutrality, the T. ferrooxidans plasma membrane dissolves (Brock, 1984), which allows
the cell to rupture, releasing its contents into solution (lysis).
For the larger-scale flotation tests with impure mineral-pyrite and coal-pyrite, only
Thiobacillus ferrooxidans culture DSM 83 and Saccharomyces cerevisiae were used. The
Thiobacillus were cultured in the T&K nutrient media described above using an orbital flask
shaker at 200 rpm and at 30°C. The cells were grown for approximately one week. The cell
density of the culture used for pyrite flotation tests was 3.4 × 108 cells/mL, as determined
with a Petroff-Hauser cell-counting chamber. These microorganisms were not washed, so
that the results would be more comparable to likely industrial practice. The Saccharomyces
cerevisiae used in these larger-scale experiments was a commercially available active dry
yeast, obtained from Red Star Inc., that contained no additives. Dosages of S. cerevisiae in the
flotation experiments were determined by weighing the dry yeast with a standard four-place
analytical balance.
MICROFLOTATION PROCEDURE
Each microflotation test used 5 g of 105 × 74-µm Rico pyrite. The pyrite was suspended in
80 mL of washed bacterial suspension with the pH adjusted to 2 with sulfuric acid and with
the sample mixed for 15 minutes. This gave a bacteria dosage of 3.2 × 109 cells/g. This corre-
sponds to an approximate cell dosage of 0.2 to 0.5 kg/t. Controls were suspended in distilled
water with the pH also adjusted to 2 with sulfuric acid and with the sample mixed for
15 minutes before flotation. Each charge was then mixed for an additional minute with
0.005 mL (0.86 kg/t) of a 50/50 mixture of #1 fuel oil (the 215° to 288°C petroleum distilla-
tion fraction) and #2 fuel oil (the 282° to 338°C petroleum distillation fraction). A dosage of
0.003 mL (0.60 kg/t) of methyl isobutyl carbinol (MIBC) was then added. The slurry was
then transferred to a glass microflotation cell with a working volume of 80 mL and an airflow
rate of 100 mL/min. The pyrite was then floated for 1 minute.
MINERAL PYRITE AND COAL PYRITE FLOTATION PROCEDURE

Each test used 150 g of either the Custer or the Empire pyrite sized to 105 × 74 µm. Each
charge was suspended in 1.9 L of distilled water (7.3% solids) in a Denver flotation machine
and treated as follows:
The pH was adjusted as desired with either sulfuric acid or sodium hydroxide solution,
and the charge was conditioned for 5 minutes to ensure that the pH had stabilized at
the target value.
The desired microorganisms were added, and the slurry was conditioned for
10 minutes.
The collector (#2 fuel oil, 3.0 kg/t) was added, and the slurry was conditioned for
2 minutes. A large dosage of collector was used to ensure that all particles that were
even slightly hydrophobic would float.
The frother (MIBC, 0.2 kg/t) was added, and the slurry was conditioned for 30 sec.
The pyrite was floated for 5 minutes.
ANALYSES AND CALCULATIONS

Because the Rico pyrite used was essentially pure, the pyrite recovery in the microflotation
tests was equal to the total weight recovery. Therefore, no composition analysis was needed.
Because of the very high pyrite levels in the impure mineral pyrite and coal pyrite samples,
a simple procedure for measuring the pyrite content could be used. For each sample, two 1-g
portions were weighed out. The first portion was leached by adding 50 mL of 4.8-N hydro-
chloric acid solution and boiling gently for 30 minutes. The second portion was leached by
adding 75 mL of 1.0-N nitric acid solution and boiling gently for 30 minutes. The residues
from each leach were then filtered, dried, and weighed to determine the percentage weight
loss.
Because pyrite does not dissolve in hydrochloric acid but dissolves completely in nitric
acid (Hurlbut and Klein, 1977), the pyrite content of the sample was calculated by
% Pyrite = (% wt. loss in HNO3) – (%wt. loss in HCI) (EQ 1)
The total weight-percent recovery to the froth product was calculated by
Froth Wt.
Recovery, %W = ------------------------ × 100 (EQ 2)
Feed Wt.
The recovery of pyrite was calculated by
( %Wt. in Froth × Froth Wt. )

%Recovery = ------------------------------------------------------------------------ × 100 (EQ 3)
( %Wt. in Feed × Froth Wt. )
RESULTS AND DISCUSSION
Microflotation Tests
From the microflotation results shown in Table 1, it is seen that all of the microorganisms
tested are capable of effectively depressing the flotation of pure Rico mineral pyrite at pH 2.
From this, it is concluded that an organism need not be a lithotroph to depress pyrite during
flotation at acid pH. In fact, the two strains of Thiobacillus are the least effective of all of the
microorganisms tested, and the pyrite-adapted strain is even less effective than the parent
strain. Of the other organisms, the yeast Saccharomyces cerevisiae was the most effective.
It should be noted that, of the organisms tested, only the T. ferrooxidans could survive at pH
2. All of the other organisms tested need a near-neutral pH to survive; they were all killed by
such acid conditions. Because cell death is often accompanied by lysis (Brock et al., 1984), the
dead cells release their internal contents into the solution. It is, therefore, likely that the pyrite
depression by these organisms is caused by released proteins and other complex molecules and
not by attachment of intact cells to the pyrite surface.
TABLE 1 Results of flotation of the pure Rico pyrite with various microorganisms at pH 2
Percentage of pyrite
Organisms Source reporting to froth
Pseudomonas fluorescens MTU stock culture 2.0 ± 0.25
Lactobacillus acidophilus MTU stock culture 2.4 ± 1.40
Staphylococcus epidermis MTU stock culture 2.3 ± 0.60
Klebsiella terrigena MTU stock culture 4.4 ± 0.40
Saccharomyces cerevisiae MTU stock culture 1.8 ± 0.02
Thiobacillus ferrooxidans EG&G Idaho, Culture DSM-83 9.4 ± 1.80
T. ferrooxidans strain PA-1 Pyrite-adapted strain derived from 39.9 ± 2.70
DSM-83
Distilled water controls — 76.5 ± 2.50
NOTES: Tests were run in triplicate and the results were averaged.
The microorganism dosage for each test was held constant at 3.2 × 109 cells/g of pyrite, and the solids
concentration was 5.88%.
The microorganisms for each test were centrifuged, washed, and resuspended in clean water to prevent side
effects from the growth medium.
FIGURE 1 Flotation of 150 × 200-mesh mineral pyrite over a wide pH range, with varying dosages
of both Thiobacillus ferrooxidans and Saccharomyces cerevisiae. The T. ferrooxidans dosage was
estimated from cell counts and is only approximate. A large dosage of microorganisms is required
to depress flotation at pH 2. At higher pH, the floatability of this pyrite is lost and little effect is
seen from microorganisms.
Mineral Pyrite Flotation

In the results shown in Figure 1, it is seen that the mineral pyrite is most floatable under
acidic conditions, but that its floatability becomes negligible at neutral and alkaline pH. This
effect is believed to be due to the changes in solubility of iron hydroxides on the pyrite sur-
face as the pH changes (Kawatra et al., 1991). These hydroxides form as a result of partial
oxidation of the pyrite, which appears to occur almost immediately and make the pyrite sur-
face hydrophilic and prevent its flotation (Baker and Miller, 1971).
The froth product at pH 2 contains virtually all of the pyrite, with the nonfloating cell prod-
uct being silicates and carbonates with less than 1% pyrite remaining. Because coal flotation is
normally carried out at near-neutral pH, pyrite with surface chemistry similar to the mineral
pyrite sample will not be a problem in coal flotation, because it will only float in acid solution.
To achieve complete pyrite depression at pH 2, it was necessary to add 6.6 kg of S. cerevi-
siae per metric ton of pyrite, which is a very high dosage level. At neutral pH, the effect of
S. cerevisiae was negligible, simply because pyrite was not being recovered by flotation in the
first place, with any pyrite in the froth being recovered by entrainment. Under alkaline condi-
tions, the S. cerevisiae actually caused a slight increase in the pyrite recovery into the froth.
This was because the microorganism acts as a mild foaming agent under these conditions.
Because the T. ferrooxidans was largely ineffective at the acid pH where it was expected to
work best, it was not used in the neutral and alkaline pH flotation experiments because it is
much more difficult to grow and handle than the S. cerevisiae.
COAL PYRITE FLOTATION

The coal pyrite was hand-collected as coarse nodules from the jig refuse belt at the Empire
Coal Mine, Gnadenhutten, OH. The behavior of the coal pyrite was completely different from
FIGURE 2 Flotation of 150 × 200-mesh Empire coal pyrite, with and without added yeast
(Saccharomyces cerevisiae). The yeast produces a general depression of all components and is
not selective toward pyrite, as is shown by the close similarity of the values for total weight
recovery and pyrite recovery.
the mineral pyrite, as shown in Figure 2. First, the coal-pyrite was most floatable at near-
neutral pH, with over 60% of the total weight floating at pH 6 in the absence of microorgan-
isms, but less than 25% floating at either pH 2 or pH 12. This behavior is very similar to that
reported for coal (Zimmerman, 1979), but very unlike that seen for the mineral pyrite. Sec-
ond, there was no pH level at which the pyrite recovery was much different from the coal
recovery. The recovery of pyrite was always very close to the total weight recovery, which
shows that the pyrite particles were being recovered at the same rate as the coal particles.
When S. cerevisiae was added to the coal pyrite, it did depress the pyrite recovery. However,
it also depressed the total weight recovery to the same degree, showing that it was a general
flotation depressant and not a highly selective one. It is believed that the coal pyrite was
floating because the pyrite was locked to hydrophobic coal particles and not because the
pyrite was itself hydrophobic. If the actual pyrite surfaces were contributing to the hydro-
phobicity, then the flotation of pyrite should have been considerably enhanced at pH 2, as
was seen for the mineral pyrite. Instead, the maximum flotation was at pH 6, as has been
reported for coal (Zimmerman, 1979).
CONCLUSIONS
From the results presented, the following conclusions were drawn:
In microflotation experiments at pH 2, every organism tested was found to depress the
flotation of mineral pyrite.
At pH 2, the T. ferrooxidans strains tested were less effective pyrite depressants than any
of the other organisms studied. This is believed to result from the other organisms dying
at the highly acidic pH levels, with the depression mainly being caused by complex
organomolecules such as proteins that the dying cells released into solution. Because
the T. ferrooxidans strains remain alive at pH 2, they would not be expected to release
these molecules in such large quantities. They, therefore, would not be as effective as
pyrite depressants at this pH.
In larger-scale flotation experiments with both mineral pyrite and coal pyrite, the two
types of pyrite were found to behave completely differently. The mineral pyrite was
most floatable under acidic conditions and practically unfloatable under neutral or
alkaline conditions, while the coal pyrite was most floatable at neutral pH. The behav-
ior of the coal pyrite was very similar to the reported flotation behavior of coal, suggest-
ing that the coal pyrite was floating because of physical locking to coal particles and not
because the pyrite itself was particularly hydrophobic.
In the larger-scale experiments, neither T. ferrooxidans nor S. cerevisiae affected mineral
pyrite flotation at neutral or alkaline pH. This is because the pyrite was not floatable
under these conditions. Under acidic conditions, both organisms were effective mineral
pyrite depressants, but only at high dosages. In coal pyrite flotation, S. cerevisiae was
found to be a depressant at neutral pH. However, the depression is unselective between
the pyrite and the associated coal, and it is, therefore, unlikely to be industrially useful.
ACKNOWLEDGMENTS
Support for this research was provided by the State of Michigan Research Excellence Fund,
Consumers Power Co., Detroit Edison Co., and the Ohio Coal Development Office. The bacte-
rial cultures were kindly provided by Dr. D.R. Lueking and Dr. S.T. Bagley of the Department
of Biology, Michigan Technological University. The authors would also like to thank Ms. J.F.
Bird and Dr. H. Johnson of the Ohio Coal Development Office and Dr. R.R. Klimpel of the
Dow Chemical Co., for their useful suggestions and critical discussion of this project.
REFERENCES
Atkins, A.S., 1990, “Developments in the biological suppression of pyritic sulfur in coal flotation,”
Bioprocessing and Biotreatment of Coal, D.L. Wise, ed., Marcel Dekker, NY, pp. 507–548.
Atkins, A.S., Bridgewood, E.W., Davis, A.J., and Pooley, F.D., 1987, “A study of the suppression of
pyritic sulfur in coal froth flotation by Thiobacillus ferrooxidans,” Coal Preparation, Vol. 5,
pp. 113.
Baker, A.F., and Miller, K.J., 1971, “Hydrolyzed Metal Ions as Pyrite Depressants in Coal Flotation:
A Laboratory Study,” US Bureau of Mines Report of Investigations, RI 7518.
Brock, T.D., Smith, D.W., and Madigan, M.T., 1984, Biology of Microorganisms, Prentice Hall,
Englewood Cliffs, NJ.
Capes, C.E., McIlhinney, A.E., Sirianni, A.F., and Puddington, I.E., 1973, “Bacterial oxidation in
upgrading pyritic coals,” CIM Bulletin, November, pp. 88–91.
Chander, S., and Aplan, F.F., 1990, “Surface and Electrochemical Studies in Coal Cleaning,” Final
Report, US Department of Energy, DOE/PC/80523-T11 (DE 9000–7603).
Eisele, T.C., 1992, “Coal Desulfurization by Bacterial Treatment and Column Flotation,” Ph.D.
Dissertation, Michigan Technological University.
Elzeky, M., and Attia, Y.A., 1987, “Coal slurry desulfurization by flotation using thiophilic bacteria
for pyrite depression,” Coal Preparation, Vol. 5, pp. 15–37.
Hurlbut, C.S., and Klein, C., 1977, Manual of Mineralogy, 19th Ed., John Wiley & Sons, NY
Kawatra, S.K., Eisele, T.C., and Johnson, H., 1991, “Recovery of liberated pyrite in coal flotation:
entrainment or hydrophobicity?,” Processing and Utilization of High-Sulfur Coals IV, Dugan,
Quigley, and Attia, eds., Elsevier, Amsterdam, pp. 255–277.
Kuenen, J.G., and Bos, P., 1989, “Habitats and ecological niches of chemolitho(auto)trophic
bacteria,” Autotrophic Bacteria, Shlegel and Bovien, eds., Science Tech Publishers, Madison,
WI, pp. 53–80.
10
Lynch, A.J., Johnson, N.W., Manlapig, E.V., and Thorne, C.G., 1981, Mineral and Coal Flotation
Circuits: Their Simulation and Control, Elsevier, Amsterdam.
Ohmura, N., Kitamura, K., and Saiki, H., 1993, “Mechanism of Microbial Flotation Using
Thiobacillus ferrooxidans for pyrite suppression,” Biotechnology and Bioengineering, Vol. 14,
pp. 671–676.
Raleigh, C.E., and Aplan, F.F., 1991, “Effect of feed particle size and reagents on coal-mineral
matter selectivity during the flotation of bituminous coals,” Minerals and Metallurgical
Processing, Vol. 8, No. 2, pp. 82–90.
Stainthorpe, A.C., 1989, “An investigation of the efficacy of biological additives for the
suppression of pyritic sulfur during simulated froth flotation of coal,” Biotechnology and
Bioengineering, Vol. 33, pp. 694–698.
Townsley, C.C., and Atkins, A.S., 1986, “Comparative coal fines desulphurization using the iron
oxidising bacterium Thiobacillus ferrooxidans and the yeast Saccharomyces cerevisiae during
simulated froth flotation,” Process Biochemistry, Vol. 21, No. 6, pp. 188–191.
Townsley, C.C., Atkins, A.S., and Davis, A.J., 1987, “Suppression of pyritic sulfur during flotation
tests using the bacterium Thiobacillus ferrooxidans,” Biotechnology and Bioengineering, Vol. 30,
p. 108.
Tuovinen, O.H., and Kelly, D.P., 1973, “Studies of the growth of Thiobacillus ferrooxidans,” Arch.
Mikrobiol., Vol. 88, pp. 285–298.
Yancey, H.F., and Taylor, J.A., 1935, “Froth Flotation of Coal—Sulfur and Ash Reduction,” US
Bureau of Mines Report of Investigations, RI 3263.
Zimmerman, R.E., 1979, “Froth flotation,” Coal Preparation, J.W. Leonard, ed., AIME, New York,
Chapter 10, Part 3.
11
Calcium Removal from Bauxite
Using Paenibacillus polymyxa
J.M. Modak,* S.S. Vasan,* and K.A. Natarajan†
ABSTRACT
The biological removal of calcium from a bauxite ore in the presence of the soil bacterium Paenibacillus
polymyxa is demonstrated. A column bioreactor that can be either operated in a fluidized-bed (FB)
mode for coarser particles or in a total-recycle-slurry (TRS) mode for finer particles was designed, fabri-
cated and tested for this purpose. For the efficient removal of calcium from the ore to the required levels,
leach solution should not be allowed to accumulate in the reactor. Therefore, the solution needs to be
drained out periodically. Cascade leaching was found to be more efficient. A direct correlation between
pH changes in the leach solution and calcium solubility could be made. More than 85% of the calcium
could be removed through cascade leaching.
INTRODUCTION
Low-grade bauxite ores that contain less than about 50% aluminum are used in the manufac-
ture of alumina-based abrasives and refractory materials. The bauxite ores need to be benefi-
ciated to remove undesirable mineral constituents before they can be used as raw materials
for the above purposes. Calcium and iron are the major bauxite impurities that affect its com-
mercial applications as abrasives and refractories. Abrasive applications require that the cal-
cium content of the bauxite ore should not exceed 0.5% (expressed as percent CaO), and
refractory applications require that Fe2O3 should be as low as 1% (Namita Deo et al., 1999).
Several physico-chemical processes, such as froth floatation, gravity separation, reduction
roasting and magnetic separation, are available for the beneficiation of bauxite. However,
these methods are energy and cost intensive, are less flexible, and pose environmental prob-
lems. A biotechnological solution to the problem promises to be economical, efficient, and
environmentally benign. The use of microorganisms in the extraction of copper, gold, and
uranium has been established and commercialized (Natarajan, 1998). However, biobenefici-
ation of nonsulfidic ores is not well understood and is not now exploited commercially.
Biobeneficiation differs from bioleaching in that it refers to the selective dissolution of unde-
sirable mineral component(s) from an ore by direct or indirect action of the microbes,
thereby, enriching the desirable mineral content of the ore.
* Department of Chemical Engineering, Indian Institute of Science, Bangalore, India.

† Department of Metallurgy, Indian Institute of Science, Bangalore, India.
13
Earlier laboratory investigations have shown that bacteria and fungi can remove iron
and silica from clays, sands, and bauxite ores (Groudev et al., 1983; Karavaiko et al., 1984;
Groudev et al., 1985; Karavaiko et al., 1989, Ogurtsova et al., 1990). Reports on silica
removal from bauxite using Bacillus circulans and Bacillus mucilaginosus (Groudeva and
Groudev, 1983; Karavaiko et al., 1989; Ogurtsova et al., 1990) showed that heterotrophic
bacteria, especially Bacillus species, can be used in biobeneficiation. Recently, the micro-
bial ecology of bauxite ore and water samples from Jamnagar mines (India) were reported
to be comprised of bacteria such as Bacillus polymyxa and Bacillus coagulans (Natarajan et
al., 1997), and preliminary investigations on calcium and iron removal from bauxite by
B. polymyxa were carried out (Anand et al., 1996). The results of their study showed signif-
icant potential for using biobeneficiation in the removal of calcium and iron from bauxite
using B. polymyxa. B. polymyxa is a N2-fixing chemoorganotroph and a facultative anaer-
obe. Various metabolites, comprised of organic acids such as formic, acetic, lactic and suc-
cinic, as well as ethanol, 2,3-Butanediol (2,3-Butyleneglycol) and acetoin, are produced by
B. polymyxa (Roberts, 1947; Groudev and Groudeva, 1986; Mankad and Nauman, 1992).
It also produces extracellular polysaccharides (ECP), especially levan that forms the capsule
of the organism (Murphy, 1952). These metabolic products can solubilize calcium and iron
from the ore (Charley et al., 1963) resulting in beneficiation. Furthermore, B. polymyxa has a
calcium-dependent metabolism; Ca is required for the production of enzymes such as amy-
lases and proteases (Gottschalk, 1989); for the synthesis of Ca-dipicolinate, an essential
component of its endospores; and also for the production of slime (or myxa) (Wilkinson,
1958). It has been shown that the growth of B. polymyxa is low in the absence of calcium,
that the bacterium can utilize calcite in the ore to meet its calcium requirements (Anand et
al., 1996; Natarajan et al., 1997) and, hence, bring about removal of calcium impurities
from the ore.
This paper discusses the design of a flexible reactor to carry out bioleaching or biobenefici-
ation in general and, in particular, as it concerns the optimization of operating conditions to
maximize calcium removal from bauxite using B. polymyxa for abrasive applications. The
emphasis is on the design of the process that can easily be adapted by the mining industry. As
the focus of this work is on abrasive applications, the removal of calcium is given priority
over iron removal. Besides, excessive iron removal is not very beneficial for abrasive applica-
tions, because iron is externally added to remove silica as ferro-silicon (Khanna, 1997). In
accordance with the recent change of appellation (Ash et al., 1993), Bacillus polymyxa is
referred to in this work as Paenibacillus polymyxa.
MATERIALS AND METHODS
Ore
Bauxite ore samples were obtained from Orient Abrasives’ Jamnagar mines in Gujarat, India.
The raw bauxite was crushed, ground and screened into the following two size fractions:
–4+5 mesh (4 to 4.8 mm) and –200+300 mesh (53 to 74 µm). The chemical compositions of
the two size fractions are given in Table 1.
TABLE 1 Chemical composition of bauxite ore used in biobeneficiation experiments

Mesh size Al2O3 CaO Fe2O3 SiO2 TiO2
–200+300 58.6% 2.80% 2.32% 3.0% 2.2%
–4+5 57.1% 2.96% 2.58% 3.5% 2.5%
14
Microorganism and Growth Medium
A pure strain of Paenibacillus polymyxa NCIM 2539 (National Collection of Industrial Micro-
organisms, Pune) was used in all of the experiments. Bromfield medium (Bromfield, 1954)
and its modification (Anand et al., 1996) are not economical for use in large-scale biobenefi-
ciation studies. The expensive carbon source in the modified medium, i.e., analytical-grade
sucrose, was replaced with a cheaper sucrose source, namely, cane sugar. The modified
Bromfield medium with cane sugar was used in all of the tests.
Growth of Microorganism
A 10% (v/v) of an active inoculum (from the late exponential phase) containing at least
109 cells/mL was added to a flask containing modified Bromfield medium at pH 7. (Cell den-
sities were determined by microscopic counting using a Petroff-Hauser Counter under a
phase-contrast microscope and were also determined by colony counting after plating.) The
flask was incubated on a rotary shaker operated at 240 rpm and 30°C. The growth of the
microorganism was monitored by measuring the dry weight and the pH of the growing cul-
ture. In the cascade leaching operation (described later), the fully-grown culture after
24 hours of growth was used for leaching experiments. In the case of uncascade-leaching
experiments, bauxite ore was added at the time of inoculation, so that bacterial growth and
calcium leaching occurred simultaneously. A bacterial culture containing 109 cells/mL was
used in the cascade-leaching operations.
The consumption of sucrose was estimated by the phenol-sulfuric acid method and plotted
against the dry weight of cells. Through a linear-regression analysis of the plot, yield of cells
per substrate consumed was estimated. The contribution of extracellular polysaccharides
(ECP) to the total sugar is reported to be about 0.24 g/g of the bacterium when 2% sucrose
was used. Using this relationship, the error introduced by considering ECP was not more
than 10% (Vasan, 1998).
Bioreactor Design
A column glass bioreactor (48-mm ID, 1.2-m long), which was designed to carry out the
leaching experiments, is shown in Figure 1. Stainless steel flanges with several ports (to be
used as inlets and outlets) were attached to the top and bottom of the glass column. An air-
vent tube (1-m) that was tall enough to avoid spillage of the circulating liquid was also pro-
vided at the top of the bioreactor. Due to the pressure gradient in the column, the pressure at
the top of the bioreactor was lower than the atmospheric pressure. As a result, air was contin-
uously sucked into the bioreactor through the air-vent tube. Thus, the aeration was achieved
without sparging the air at the bottom of the reactor, as is typically done in stirred-tank
bioreactors. This air-vent tube could be closed whenever aeration was not required. An auto-
clavable 15-L plastic container with an outlet tap connection was used as the storage tank.
The tap was used for taking samples. In addition, three ports were provided on the storage
tank: two were leaching-medium inlets from the top and bottom of the column bioreactor
and one was an outlet to the bottom of the bioreactor. A mono-block centripetal pump (rated
at 0.18 kW, 0.25 hp) was used for pumping the liquid. The whole set up was mounted on
slatted angle scaffolding.
The bioreactor was operated in two different modes: a fluidized-bed (FB) mode for
coarser particles (–4+5 mesh) and a total-recycle-slurry (TRS) mode for fine particles
(–200+300 mesh). The FB reactor was operated by introducing a wire mesh at the bottom to
hold the solid ore particles. A 30-mesh screen was used to hold particles in the –4+5-mesh
range. The fluidization was achieved by pumping leaching media at the bottom of the col-
umn. The liquid flow rates were controlled such that the transport disengaging height (TDH)
of the particles was less than the height of the reactor (1.2 m). It was not necessary to fluid-
ize the ore all the time. The ore particles were fluidized by pumping liquid at the bottom for a
desired period (called the flood cycle) and then shutting off the pump (called the drain
15
FIGURE 1 Schematic representation of column bioreactor used for biobeneficiation
cycle). This flood/drain cycle could be repeated by using a suitable timer. If the air vent were
open, the drain cycle would lead to the complete withdrawal of the culture into the storage
tank, resulting in aeration and stratification of the bed. If the air vent were shut off, the ore
would be submerged in a fixed volume of culture, although circulation by the pump could be
stopped. The bioreactor that operated in flood/drain-cycle mode was similar to that reported
by Andrews et al. (1994). In TRS experiments, the bauxite ore and leaching medium were
added to the storage tank, and the slurry was pumped from the storage tank to the bottom of
the reactor. The slurry coming out of the reactor at the top of the column was fed back to the
storage tank, thereby, making it a total-recycle operation. Thus, TRS operation is equivalent
to conventional stirred-tank slurry reactor operation. However, a uniform suspension of par-
ticles was achieved by the action of the circulating liquid, without the use of an agitator, and
no forced pumping of air is required for aeration.
Leaching of Bauxite Ore

Two types of leaching experiments were conducted: uncascade and cascade operations. In
uncascade operation, the leaching medium consisted of a modified Bromfield medium,
which is freshly inoculated and used for leaching of bauxite ore. In this operation, bacterial
growth and leaching occurred simultaneously, and the operation was allowed to continue for
four days. In cascade operation, bacterial culture was first grown for 24 to 30 hours in the
Bromfield medium in the absence of bauxite ore. The grown culture containing bacteria, and
their metabolite products, was used as the leaching medium. Furthermore, leaching was
stopped after 24 hours, and liquid medium was allowed to drain into the storage tank. Fresh
leaching medium (grown culture) was added to the storage tank, and the pumping was
restarted. The entire operation was repeated four times. In the case of the FB reactor, the
16
FIGURE 2 Growth of Paenibacillus polymyxa in modified Bromfield medium
bauxite ore remained in the bioreactor while, in the TRS, the bauxite ore drained back to the
storage tank along with the liquid medium. The ore was allowed to settle at the bottom of the
storage tank (for about 0.5 hour), and the leached liquor was decanted. The solid residues
were analyzed for the calcium remaining in the residue, so that the calcium removed from
the ore could be monitored.

Figure 2 shows a typical growth profile of Paenibacillus polymyxa inoculated in modified
Bromfield media containing 2% cane sugar without any bauxite added. After a lag phase of
about 45 minutes, cells start growing rapidly and enter a logarithmic growth phase. This
phase lasts for about 20 hours, beyond which cell growth is very slow (i.e., stationary
phase). The sucrose concentration decreases sharply in the first 20 hours, and, thereafter,
the sucrose consumption rate also slows down significantly. The cell growth is accompanied
by a decrease in the pH of the culture. After 24 hours of inoculation, the pH decreases from
7 to 2.2 and reaches a saturation value. Paenibacillus polymyxa is known to produce organic
acids such as formic, succinic, acetic, and lactic (Mankad and Nauman, 1992), which would
decrease the pH of the culture. The optimal pH for the growth of Paenibacillus polymyxa is
in the range of 4 to 7. When the pH of the culture decreases to below 4, the growth of the
bacteria is severely hindered, and, therefore, growth ceases beyond 20 hours, even though
there is sufficient carbon available for growth.
Initial leaching experiments were conducted on the uncascade operation with –4+5- and
–200+300-mesh particles. Four liters of modified Bromfield medium were prepared in the
storage tank and inoculated with a 10% (v/v) actively growing culture of P. polymyxa. The
experiments with –4+5-mesh particles were conducted with the reactor operated in FB mode.
Four hundred grams of –4+5-mesh bauxite ore was added to the reactor to yield an equiva-
lent of 10% pulp density. The reactor was operated in flood/drain cycle mode with a 1-hour
fluidization cycle and a 4-hour drain cycle. In the case of –200+300-mesh particles, the ore
was added to the storage tank, and the slurry was circulated in the reactor continuously.
17
FIGURE 3 Calcium removal with time (A) fluidized bed reactor for –4+5-mesh size particles (B)
total recycle slurry reactor for –200+300-mesh size particles
Figure 3 shows the calcium removed from the bauxite ore. In both of these experiments,
cell growth was observed (data not shown), although the cell densities were lower than that
obtained in the absence of bauxite ore. The percent calcium removal increases as time
progresses, but it reaches a saturation value of 20%. It is interesting to note that the calcium
removal from coarser particles (Figure 3A) is similar to that obtained from finer particles
(Figure 3B), even though the time required is longer. In the case of the TSR experiment, the
saturation level is attained in 24 to 48 hours, while 60 hours are required in FB experiments.
The kinetics of calcium solubilization can be expected to be faster for the finer particles
(TRS), as compared to coarser particles (FB). These results indicate that a considerable sav-
ing in grinding costs can be achieved if coarser particles are treated. However, the time
required with coarser particles will be higher. Furthermore, the processing of bauxite for the
manufacture of abrasive-grade material requires the particle size to be –200+300 mesh
(Khanna, 1997). Thus, grinding to finer sizes is an inevitable step, even if higher particle
sizes are used for bauxite biobeneficiation.
In view of these requirements, further experiments were conducted with –200+300-mesh
particles by operating the bioreactor in the TRS mode. The calcium content in the leached
residue was about 2.4%, which does not meet the requirements of the abrasive industries
(<0.5%), and a further improvement in calcium removal is required. One possible way in
which the improvement in calcium removal can be achieved is to replace the leaching
medium periodically in cascade operation. Anand et al. (1996) showed that the growth of
bacteria is delayed in the presence of bauxite ore. It can be expected that the leaching can
also be improved if the growth phase is separated from the calcium-solubilization (leach-
ing) phase. Therefore, the bacteria were first grown in the absence of bauxite ore and, then,
this grown culture, which contained the cells and bacterial metabolites, was used as a
18
FIGURE 4 Calcium removal and pH variations in column bioreactor for 5% pulp density bauxite
slurry
leaching medium in cascade experiments. The results of Figure 2 clearly show that the
growth of the bacteria ceases after about 20 to 24 hours. Therefore, the culture grown for
24 hours was used as the leaching medium in cascade experiments.
The results of a typical cascade biobeneficiation TRS experiment with 5% pulp density are
presented in Figure 4. Bauxite ore with a calcium content of 2.8% CaO was beneficiated with
24-hour-grown culture of P. polymyxa, and the target of 0.5% CaO in the ore is reached after
88 hours of operation, consisting of four cascades. In this experiment, the duration of each
cascade was maintained at 24 hours, except for the last cascade, which was stopped after
16 hours because no change in the leached liquor was observed.
It was observed in preliminary experiments that the addition of bauxite ore to grown cul-
ture increased the pH of the solution, almost intantaneously, from about 2 to 5. To examine
the implication of such pH jumps for calcium removal, the solid residues were collected
10 minutes after the start of each cascade. Ten minutes was chosen because this was the
shortest time in which a sample could be collected from the bioreactor. Figure 4 shows that
the biobeneficiation of calcium in each cascade is characterized by two distinct phases: Phase
I—the initial phase of rapid removal of calcium and Phase II—a gradual rate of calcium
removal. This trend is observed for calcium content as well as in the pH of the leach liquor.
For example, in the first cascade, the calcium content decreased from 2.8% to 2.24% (as
CaO) within 10 minutes of the addition of the culture (Figure 4). The calcium content further
decreased to 2.18% (as CaO) in the remaining duration of the cascade. The initial rapid
phase results from the action of a well-grown culture (≈pH 2) coming into contact with the
ore. The dissolution of calcium by the metabolite was almost instantaneous, resulting in the
rapid removal of calcium. As a result of calcium removal, the pH of the leach liquor rose from
2.1 to 5.45 (first cascade in Figure 4). Such an increase in alkalinity has two effects on bene-
ficiation. The ability of metabolite to dissolve calcium is reduced as the pH rises to near-neutral
levels. As a result, the dissolution of calcium slows down. However, higher pH values are favor-
able for bacterial growth. It should be noted that the grown culture used for leaching has about
19
FIGURE 5 Calcium removal in Phases I and II in four cascades
10-g/L sugar, which can support the growth of the bacteria. Thus, the growth of the bacteria
is likely to pick up as the pH increases (in spite of some growth inhibition due to the ore).
The growth of the bacterium, along with weathering effects, leads to more removal of cal-
cium over the next 24 hours, which results in a gradual decrease in the calcium content (Fig-
ure 4). This phase is accompanied by the gradual increase in pH from 5.45 to 6.22.
The observed reduction of CaO content from 2.8% to 0.5% in the bauxite ore at 5% pulp
density corresponds to the removal of 820 ppm of calcium. Figure 5 compares the calcium
removal (expressed as ppm calcium in leach liquor) in the initial rapid phase (Phase I) and
the slow phase (Phase II) in four different cascades. This rapid removal of calcium in Phase
I can be attributed to the action of acidic culture alone, and the time scale (10 minutes) is
too small for the contribution of the cells by direct interaction with the ore. Thus, Phase I
can be looked upon as the contribution of indirect mechanism of leaching by action of bac-
terial metabolites. It is interesting to note that the calcium removed in Phase I in three cas-
cades is approximately the same, namely, 180- to 200-ppm Ca. This points to the
saturation solubility of calcium in the 24-hour-grown culture. Phase II, on the other hand,
is more gradual (24 hours), which can be viewed as a combined contribution of the direct
mechanism involving bacteria and indirect leaching by bacterial metabolites. It can be seen
from Figure 5 that the gradual phase (Phase II) brings about good leaching, as in the case
of the second or third cascade. This may be due to the growth of bacteria and the activity of
the culture in Phase II. It is also possible that the cells attached to the ore from earlier cas-
cades have contributed to leaching through a direct attack mechanism, as the contact time
is large. Bacterial adhesion to mineral particles was observed under a scanning electron
microscope. Direct bacterial attack through enzymatic oxidation can take place.
Exopolysaccharides present around the cell wall facilitate bacterial adhesion and the sub-
sequent uptake of calcium and iron from the ore matrix. Because significant amounts of
loose slime, residual carbohydrates and organic acids are present in the liquid phase, cal-
cium can be solubilized through chelation and acidolysis (Anand et al., 1996).
Biodissolution of hematite, corundum (alumina) and calcite in the presence of Bacillus
polymyxa has been reported (Namita Deo and Natarajan, 1997), and it would be interesting
20
FIGURE 6 Calcium removal and pH variations in column bioreactor for 10% pulp density bauxite
slurry
to compare the dissolution of calcium and iron from bauxite with individual pure mineral
systems. The presence of bacterial cells promoted the dissolution of calcite, hematite, and
alumina at a much faster rate than that observed in the presence of bacterial metabolite
alone. Among the above three minerals, the biodissolution tendency was the highest for cal-
cite, followed by hematite. Alumina solubility was very limited. The direct role of bacterial
cells in enhancing the solubilization of calcite through attachment and direct enzymatic
attack could be clearly seen.
The results presented in Figures 4 and 5 indicate that calcium removal slows down in
the fourth cascade. The calcium content of the bauxite ore has decreased to 0.8% at the
start of the fourth cascade. The calcium content in the leach liquor at the start of the cas-
cade is also found to increase (data not shown). This is primarily due to the inability to
remove the leach liquor completely at the end of the cascade. In these experiments, simple
settling and decantation were used for solid-liquid separation at the end of the cascade. In
this process, only about 80% of the leached liquor could be separated. In other words,
every time a new cascade was started, 20% of the calcium-rich (300-ppm), alkaline (pH 6
to 7), old liquor was mixed with calcium-free acidic (pH 2 to 2.1) grown culture. Thus,
decreasing calcium content in the ore and increasing calcium content in the leaching
media contribute toward lowering of the driving force for calcium removal, and less leach-
ing was observed in the fourth cascade.
The results of the cascade experiments (four cascades) with bauxite ore at 10% pulp den-
sity are shown in Figure 6. Even at this pulp density, the calcium removal is characterized by
a rapid calcium removal in Phase I, followed by a gradual removal in Phase II. It is interesting
to note that the calcium removal in Phase I of each of the first three cascades correspond to
about 200-ppm calcium leached by the action of the grown culture. This experiment, there-
fore, supports the earlier conjecture that the 24-hour-grown culture has the saturation limit
for calcium solubilization. It should be pointed out that the contribution of Phase II in calcium
21
FIGURE 7 Correlation of pH increase and calcium dissolution in Phases I and II of bauxite leaching
removal is somewhat higher at a pulp density of 10% than it is at a pulp density of 5%.
Increasing the pulp density implies that the calcium to be removed is also higher, and four
cascades are not enough to remove the calcium to the desired level of 0.5% CaO. As higher
pulp densities are preferred in industrial-scale operations, biobeneficiation looks promising
to give good results with higher slurry concentrations as well, even though the number of
cascades required will increase as the pulp density is increased. Figures 4 and 6 also compare
the results of cascade experiments with respective experiments in which the bacterial culture
is not replenished every 24 hours (uncascaded experiments). It can be expected that, in the
uncascade experiments, calcium removal in the first 24 hours will be the same as in the cas-
caded mode. However, no further calcium removal takes place in the next 64 hours of opera-
tion, and the calcium removal saturates at around 20%. In fact, a small increase in the
calcium content of the bauxite ore, possibly due to reprecipitation, is observed (Figures 4 and
6) after 48 hours of treatment.
The results of Figures 4 and 6 show that Phases I and II observed in calcium removal is
apparent from the data of the pH of the culture media. It is, therefore, interesting to see
whether the pH of the leaching media can be correlated to the calcium leached. Such a corre-
lation can be useful in industry as an indirect estimation of the calcium removal. The incre-
mental alkalinity (∆ pH) and calcium leached (∆ Ca, ppm) were calculated for the 5% and
10% pulp density cascade experiments (data of Figures 4 and 6) and are plotted as shown in
Figure 7. A clear segregation of the data from two experiments corresponding to Phases I and
II is clearly seen in Figure 7. The data in Phases I and II can be correlated by a straight line
with correlation coefficients of 0.93 and 0.95, respectively. The slope of the line for Phase I is
higher than that for Phase II; that is, Phase I is characterized by a higher rate of increase in
pH compared to Phase II. At the start of Phase I, the grown culture is acidic, which is neutral-
ized by the instantaneous solubilization of calcium in the leach liquor. The contribution of
cells in this phase is negligible. On the other hand, the pH variations in Phase II are the
22
FIGURE 8 Calcium removal in pulse leaching experiment conducted at 20% pulp density
results of two opposing phenomena, namely, a pH increase due to calcium dissolution, which
is countered by the production of acidic metabolites by the bacteria. Thus, the increase in pH
is slower in Phase II than it is in Phase I. However, the increase in pH can be well-correlated
to calcium leached in both these phases and can be treated as an indirect estimation of cal-
cium leached.
An interesting observation can be made from the above-described cascade experiments:
the cumulative time required for Phase I leaching (40 minutes) is extremely short compared
to the time required for Phase II leaching (almost 87 hours). This observation has an impor-
tant implication for industrial application in reducing the time required for calcium removal
by carrying out a series of Phase I operations, that is, conducting cascade experiments with
very short duration. Such an experiment can also be viewed as a pulse-leaching experiment.
This concept was tested at 20% pulp density of the bauxite ore, as higher pulp densities are
preferable in industrial applications. A grown culture at pH 2.1 was contacted with the ore
for 10 minutes, after which the ore was separated from the leaching medium and fresh-
grown culture was added. The procedure was repeated several times. As pointed out above,
the duration of contact (10 minutes) was determined by the time required for drainage of liq-
uid in the storage tank and solid/liquid separation. The results of this experiment are shown
in Figure 8. The brief contact time between ore and grown culture rules out any possibility of
involvement of bacterial cells in calcium removal. It can be seen that the calcium content in
the bauxite ore is brought down from 2.8% to 0.5% in 13 cascades. The decrease in calcium
content is almost linear (correlation coefficient 0.98) with the number of cascades. This
implies that the calcium leached in each cascade is almost constant and corresponds to about
200 ppm Ca. The number of cascades required is higher than that that required in earlier
experiments with 5% and 10% pulp densities. The pulp density (20%) used in this experi-
ment is high, and the contribution of Phase II leaching is completely absent in the pulse-
leaching experiments. In spite of the increased number of cascades, the entire operation is
over in about 4 hours, a substantial reduction in the operating period. However, this saving
in the operating time is at the cost of an increase in the volume of grown culture required.
23
CONCLUSIONS
The following major conclusions were made based on the results of this study:
The soil bacterium Paenibacillus polymyxa was found to be efficient in the removal of
calcium from bauxite ore.
A column reactor that could be operated either in a fluidized-bed mode for coarser par-
ticles or in a total-recycle-slurry mode for finer particles was designed and tested for the
beneficiation of bauxite.
To obtain higher leaching efficiencies, it was found that it was essential to drain off
the leach liquor periodically from the reactor, replenishing it with fresh medium in a
cascade operation.
Changes in the pH of the leach medium correlated well with increasing calcium
dissolution. The presence of bacteria aids in the lowering of the pH, facilitating calcium
dissolution.
Calcium removed from the bauxite ore was found to be linear with respect to the num-
ber of cascades in pulse-leaching experiments. The calcium-removal efficiency was
found to be in excess of 85%. The developed bioprocess appears to be a simpler,
cheaper and environmentally benign alternative to existing acid-dissolution processes.
ACKNOWLEDGMENTS
The authors gratefully acknowledge help received from Orient Abrasives Ltd. (OAL), New
Delhi, for sponsoring this work. Special thanks are due to Mr. P.P. Khanna, Executive Direc-
tor of OAL, for supplying ore samples and valuable ideas from time to time.
REFERENCES
Anand, P., Modak, J.M., and Natarajan, K.A., 1996. “Biobeneficiation of bauxite using Bacillus
polymyxa: Calcium and iron removal,” Int. J. Mineral Processing, 48, pp. 51–60.
Andrews G.F, Noah K.S., Glenn A.W., and Stevens C.J., 1994. “Combined physical/microbial
beneficiation of coal using the flood/drain bioreactor,” Fuel Processing Technology, 40,
pp. 283–296.
Ash, C., Priest, F.G., and Collins, M.D., 1993. “Molecular identification of rRNA Group 3 bacilli
using a PCR probe test,” Antonie van Leeuwenhoek, 64, pp. 253–260.
Bromfield, S.M., 1954. “Reduction of ferric compounds by soil bacteria.” J. Gen. Microbiol., 11,
pp. 1–6.
Charley, P.J., Sarkar, B., Stitt, C.F., and Saltman, P., 1963, “Chelation of iron by sugars,” Biochem.
Et Biophy. Acta, 69, pp. 313.
Gottschalk, G., 1989. “Nutrition of bacteria,” Bacterial Metabolism, M.P. Starr, ed., Springer
Verlag, New York, pp. 1–10.
Groudev, S.N., Groudeva, V.I., Genchev, F.N., Petrov, E.C., and Mochev, O.J., 1983. “Removal of
iron from sands by means of microorganisms,” Progress in Biohydrometallurgy, Cagliari,
pp. 441–450.
Groudev, S.N., Genchev, F.N., Groudeva, V.I., Mochev, O.J., and Petrov, E.C., 1985. “Biological
removal of iron from quartz sands, kaolins and clay,” Proc. XV Int. Min. Process. Congress, Vol.
II, Floatation, Hydrometallurgy, Cannes, France, pp. 378–387.
Groudev, S.N., and Groudeva, V.I., 1986. “Biological leaching of aluminium from clays,”
Biotechnol. Bioeng. Symp. No.16, John Wiley & Sons, Inc., New York, pp. 91–99.
Groudeva, V.I., and Groudev S.N., 1983, “Bauxite dressing by means of Bacillus circulans,”
Travaux ICSOBA, 13, pp. 257–263.
24
Karavaiko, G.I., Belkanova, N.P., Eroshchev-Shak, V.A., and Avakyan, Z.A., 1984. “Role of
microorganisms and some physico-chemical factors of the medium in quartz destruction,”
Microbiology, 53, pp. 795–800.
Karavaiko, G.I., Avakyan, Z.A., Ogurtsova, L.V., and Safanova, O.F., 1989, “Microbiological
processing of bauxite,” Biohydrometallurgy 1989, J. Salley, R.G.L. McGready and L. Wichlacz,
eds., CANMET, Ottawa, pp. 93–102.
Khanna, P.P., 1997, personal communication, Executive Director, Orient Abrasives Ltd., New
Delhi.
Mankad, T., and Nauman, E.B., 1992. “Effect of oxygen on steady state product distribution in
Bacillus polymyxa fermentations,” Biotechnol. Bioeng., 40, pp. 413–426.
Murphy D., 1952. “Structure of levan produced by Bacillus polymyxa,” Can. J. Chem., 30, pp. 872–878.
Namita Deo, Vasan, S.S., Modak, J.M., and Natarajan, K.A., 1999, “Selective biodissolution of
calcium and iron from bauxite in the presence of Bacillus polymyxa,” Biohydrometallurgy and
the Environment Toward the Mining of the 21st Century, R. Amils and A. Ballester, eds.,
Elsevier, Amsterdam, pp. 463–472.
Namita Deo, and Natarajan, K.A., 1997, “Interaction of Bacillus polymyxa with some oxide
minerals with reference to mineral beneficiation and environmental control,” Minerals
Engineering, Vol. 10, pp. 1339–1354.
Natarajan, K.A, Modak, J.M., and Anand, P., 1997. “Some microbiological aspects of bauxite
mineralization and beneficiation,” Minerals and Metallurgical Processing, Vol. 14, No. 2,
pp. 47–53.
Natarajan, K.A., 1998. Microbes, Minerals and Environment. Geological Survey of India, Bangalore.
Ogurtsova, L.V., Karavaiko, G.I., Avakyan, Z.A., and Korenevskii, A.A., 1990. “Activity of various
microorganisms in extracting elements from bauxite,” Microbiology, 58, pp. 774–780.
Roberts, J.L., 1947, “Reduction of ferric hydroxide by strains of Bacillus polymyxa,” Soil Sci., 63,
pp. 135–140.
Vasan, S.S., 1998, Master of Science Thesis, Indian Institute of Science, Bangalore, pp. 32–34.
Wilkinson, J.F., 1958. “The extracellular polysaccharides of bacteria,” Bact. Rev., 22, pp. 46–73.
25
Desulfurization of Coal by
Microbial Flotation in a
Semicontinuous System
T. Nagaoka,* N. Ohmura,* and H. Saiki*
ABSTRACT
A microbiological flotation system that facilitates the removal of pyritic sulfur from coal is reported.
The chemoautotrophic bacterium Thiobacillus ferrooxidans is utilized as a pyrite-selective biological
surfactant that augments the flotation-mediated rejection of pyritic sulfur. The bacteria were pro-
duced by continuous cultivation in 50-L tanks, where it was possible to produce 3.8 × 1012 cells/day
when the dilution rate was optimal and the pH was 1.6, a condition that prevented ferric precipita-
tion. Prior to flotation, the bacteria were mixed with coal, which allowed cells to selectively adhere to
the pyrite. The adherence of bacteria made the pyrite surfaces more hydrophilic, which facilitated sep-
aration by increasing the tendency of pyrite to sink, while the hydrophobic coal floated. When a syn-
thetic coal/pyrite mixture (10.1% sulfur) was subjected to microbial flotation, the level of sulfur
rejection was stable during the entire period of operation (i.e., one hour). In the absence of bacteria,
flotation reduced the sulfur content of a coal/pyrite mixture to within the range of 6.7% to 7.4%, with
the combustible recoveries ranging from 93.9% to 99.0%. In the presence of bacteria, the sulfur con-
tent of cleaned coal was in the range of 1.3% to 1.4%, with the combustible recoveries ranging from
80.8% to 86.3%. Bacterial activity accounted for 57% of the total sulfur rejected through flotation.
The desulfurization capacity of the system was further confirmed by testing it with run-of-mine coal.
The pyritic sulfur content was reduced from 2.9% in the feed coal to around 1.2% in the froth prod-
uct, while 80% of the combustible recovery was retained. Microbial flotation removed 60% of the
pyritic sulfur from the feed. These findings demonstrated that microbial flotation could be utilized for
the removal of pyritic sulfur from coal in a semicontinuous system.
INTRODUCTION
Coal is receiving increasing attention as an energy resource. Coal combustion, however,
releases sulfur dioxide into the environment, and the desulfurization of coal prior to combus-
tion would be an important step toward reducing the release of sulfur dioxide. This would
allow for the use of high-sulfur coal to meet the energy demands of the future. Sulfur in coal
* Bio-Science Department, Abiko Laboratory, Central Research Institute of Electric Power Industry
(CRIEPI), Abiko-city, Chiba, Japan.
27
occurs in two forms: organic and inorganic. Organic sulfur exists as a part of the coal matrix
in forms such as thiophene, aromatic sulfide and thiol compounds. These forms are difficult
to remove. Inorganic sulfur in coal, by contrast, is mainly concentrated as pyrite particles,
which are separable from coal particles (Couch, 1991). Therefore, the removal of pyritic sul-
fur is possible for the cleaning of coal prior to combustion. To increase pyrite removal from
coal, it is necessary to liberate the pyrite particles by grinding. However, separation of the
finely ground pyrite from the coal is difficult with conventional physical methods such as a
jigs and heavy-medium cyclones. Column flotation has been investigated as a means of coal
desulfurization, although it is unclear how to regulate the pyrite floatation.
The first report on the use of biological materials for coal desulfurization was noted in
1973 (Capes et al., 1973). The addition of the iron-oxidizing bacterium Thiobacillus
ferrooxidans improved pyrite removal from coal in oil agglomeration. Biological conditioning
with T. ferrooxidans was then applied to froth flotation for coal to reduce pyrite flotation
(Attia and Elzeky, 1985; Dogan et al., 1985). Some researchers investigated a microbial con-
ditioning process with synthetic coal and several run-of-mine coals (Townsley and Atkins,
1986; Townsley et al., 1987; Atkins et al., 1987; Elzeky and Attia, 1987; Stainthorpe, 1989;
Atkins, 1990; Kawatra and Eisele, 1993; Eisele and Kawatra, 1994). There is growing con-
sensus about the efficacy of biological conditioning using the bacterium as a way of enhanc-
ing pyrite removal during flotation.
The role of biological conditioning for the removal of pyrite from coal in flotation has been
considered from the point of view of bacterial oxidation (Attia et al., 1990). It had been
thought that a few molecular layers of the pyrite surfaces were oxidized by bacteria and that
the oxidation might induce the suppression of pyrite floatability by changing its surface prop-
erty from hydrophobic to hydrophilic. However, this speculative mechanism of surface modi-
fication by bacteria has not been proved. In 1993, it was reported that the biomodification of
pyrite surfaces was caused not by surface oxidation but by bacterial adhesion to pyrite sur-
faces (Ohmura et al., 1993a). The adhesion of T. ferrooxidans occurred selectively onto pyrite
surfaces (Ohmura et al., 1993b) and induced the suppression of pyrite flotation (Ohmura et
al., 1993a). The bacterial adhesion enhanced the removal of pyrite from both synthetic and
Pittsburgh coals during flotation (Ohmura and Saiki, 1994; Ohmura and Saiki, 1996). The
bacterium acts as a pyrite-selective, biological surfactant suppressing pyrite floatabilities. In
addition, other microorganisms in addition to T. ferrooxidans were reported to adhere to coal
and associated minerals and, as a result, have an affect on the separation of minerals from
coal in flotation (Raichur et al., 1995).
Every report on microbial modification for pyrite removal from coal has described batch
tests on a laboratory scale. There is no additional knowledge available for establishing the
biomodification process for coal desulfurization. The effect of biomodification on pyrite
removal should be discussed in a continuous process. This report describes the application of
a bacterial function in adhesion to pyrite removal from coal in a semicontinuous system. The
system features continuous cultivation and the use of a flotation column cell instead of a con-
ventional froth-flotation cell to treat fine coal.
Coal Samples
A synthetic coal/pyrite mixture and a run-of-mine coal were used in the flotation experi-
ments. The mixture was made of mineral pyrite and run-of-mine Datong (China) coal.
Microorganism and Culture Medium

The iron-oxidizing bacterium Thiobacillus ferrooxidans ATCC 23270 was used in this
study. The cells were cultured in m9K medium that contained 3.0 g/L (NH4)2SO4,
28
0.5 g/L MgSO4·7H2O, 0.5 g/L KH2PO4, 0.01 g/L Ca(NO3)2 and 44.2 g/L FeSO4·7H2O in dis-
tilled water. The medium was a modified 9K medium (Silverman et al., 1959). In m9K
medium, the precipitation of the ferric compounds (e.g., jarosite) is less formative than in the
9K medium.
Continuous Cultivation
Cells were produced in a 50-L cultivation tank (Model MSJ-N, B.E. Marubishi Co., Ltd.,
Tokyo, Japan), and any part of the tank in contact with the culture medium was provided the
Teflon-rubber coatings to prevent corrosion.
The cultivation was carried out in m9K medium while being aerated (15 L/min) and agi-
tated (180 rpm) at 28°C. The pH of the medium was automatically adjusted to 1.6 by the
addition of 6N H2SO4 to avoid the precipitates mentioned above.
Fresh m9K medium (pH 1.5) was pumped into the cultivation tank, and the culture,
including the bacterial cells, was pumped out at the same rate and collected in a second tank.
Varying the dilution rate regulated the production capacity of the cultivation system. When
the cell density reached a stationary phase, dilution was initiated by pumping fresh medium
into the cultivation tank and pumping cultured medium out at the same rate.
The cell density was determined by a previously contracted calibration curve in which the
optical density at 610 nm was plotted as a function of the density of cells. Measurement of
the concentration of the total soluble iron and ferrous ion was carried out by the o-phenan-
throline method.
Flotation Column Apparatus

The flotation column used for the desulfurization of coal (Figure 1) had a total height of 1.5 m
and an internal diameter of 80 mm. The column was equipped with upper (10-mm radius,
70-mm long) and lower (50-mm radius, 140-mm long) micro air-bubble generators. The gen-
erators were of the double-pipe type constructed of an inner micro-pore pipe (mean pore size
of 1 µm) and an outer acrylic pipe. As the flotation liquor flowed through the inner pipe, bub-
bles were generated in the flotation column by shearing air, which was compressed in the
space between the inner and outer pipes (Figure 1 insert). The upper generator was set up to
generate a stable froth to optimize coal recovery.
Operation of the Semicontinuous Microbial Flotation System

The system was comprised of a cultivation tank and a flotation column, as shown in Figure 1.
To enable bacterial adhesion to pyrite, the culture medium from the second tank was mixed
with a coal slurry that contained methyl isobutyl carbinol (25 µL/L) as a frother. In the syn-
thetic coal, the collector (kerosene) was added at the concentration of 1% (v/w). After mix-
ing, the final concentration of the coal slurry was adjusted to 1.0% (by weight). The column
was filled with the flotation liquor, and air bubbles were generated. The flow rates of the
upper and lower air-bubble generators were 1.0 and 0.3 L/min (0.3 and 0.1 MPa), respec-
tively. The bacteria-containing coal slurry was fed into the flotation column at a rate of
315 mL/min through the upper air-bubble generator. After the coal slurry was fed to column,
the product was collected every 10 minutes for an hour. Clean coal was overflowed from the
top of the column and was collected for subsequent sulfur and ash analysis.
Analysis of Sulfur and Ash

The sulfur content in the coal was determined based on International Organization for Stan-
dardization (ISO) method 157 (Hard coal—Determination of sulfur). The total sulfur content
was measured using a sulfur analyzer (EMIA-512FA, Horiba Ltd., Kyoto, Japan). Ash was
determined by ISO method 1171 (Solid mineral fuels—Determination of ash).
29
FIGURE 1 Schematic diagram of a semicontinuous microbial column flotation system
The pyrite rejection (combustible recovery) was calculated by subtracting the pyrite (ash)
in the froth from that in the feed coal.
RESULTS
Continuous Cultivation of Thiobacillus ferrooxidans

The system utilized a 50-L continuous cultivation tank for producing bacteria. The dilution
rate was changed from 0.01 to 0.06 h–1, regulating the production capacity of the cells. At
dilution rates of 0.01 and 0.03 h–1, the cell density was maintained at about 1.6 to 2.8 ×
108 cells/mL (Figure 2). However, when the rate increased to 0.06 h–1, the cell density in the
culture medium declined, and the ferrous content in the medium increased due to the wash-
ing out of the cells. Therefore, the dilution rate for cell production was set to 0.03 h–1 to pre-
vent washout. At that condition, 3.8 × 1012 cells/day were produced. The doubling time of
T. ferrooxidans is thought to be about 8 hrs. The time corresponded to the dilution rate of as
much as 0.08 h–1. Nonetheless, it was found that the maximum dilution rate was less than
0.06 h–1. This meant that the doubling time under the conditions in this study was much
30
cell density in medium
pH
ferrous concentration
▼ total iron concentration
FIGURE 2 Continuous cultivation of iron-oxidizing bacteria. The dilution rate (D) was 0.01, 0.03,
and 0.06 h–1
lower than what was previously reported. One possible reason for the low growth rate is the
medium pH. The optimum pH for growth of this strain may be substantially higher than 1.6.
Increasing the medium pH, however, would induce the formation of insoluble ferric com-
pounds (e.g., jarosite). Such precipitation would make the separation efficiencies lower,
because nonselective bacterial adhesion of it onto both pyrite and coal particles increase the
tailing in column flotation. In this cultivation protocol, precipitation was avoided by adjust-
ing the pH; the total iron concentration was kept constant at about 160 mM.
Semicontinuous Microbial Column Flotation

Once bacteria were cultivated, the removal of pyritic sulfur from synthetic coal/pyrite mix-
ture and ROM coal was investigated. To enable bacterial adhesion to pyrite, the bacteria-
containing cultured medium was mixed with coal slurry that also contained flotation
reagents. Synthetic coal/pyrite mixture was mixed with the bacteria-containing cultured
medium to a final cell density of 7.4 × 109 cells/g of coal. The flotation was carried out for
one hour, and 1% (by weight) coal slurry was fed to the column at the rate of 315 mL/min.
When the flotation was carried out in the absence of bacteria, the sulfur content was
reduced, on average, from 10.1% in the feed coal to 7.1% in the product (Figure 3); combus-
tible recovery ranged from 93.9% to 99.0%. In the absence of bacteria, flotation removed
29% of the sulfur from the feed coal. In the presence of bacteria, the sulfur content of the
product was reduced to 1.4%, and the combustible recovery was 82.7%. In this case, 86% of
the added sulfur was removed by flotation, and 57% of the sulfur removal was attributable to
bacterial activity.
A semicontinuous column flotation was then carried out with ROM coal, Pittsburgh coal
(particle size 53 to 75 µm), under conditions identical to those described above. In the
absence of bacteria, the pyritic sulfur content of the product was 2.0% on average, which
means that 66% of the pyritic sulfur in the feed coal remained in the froth product
(Figure 4). When bacteria were mixed with the coal to a density of 7.7 × 109 cells/g of coal
prior to flotation, the pyritic sulfur content was reduced from 2.9% in the feed coal to 1.2%
31
combustible recovery without cells
combustible recovery with cells
total sulfur in concentrate without cells
▼ total sulfur in concentrate with cells
FIGURE 3 Semicontinuous desulfurization of synthetic coal/pyrite mixture by microbial column

flotation
in the product coal. This corresponds to a 60% rejection of pyrite from the feed coal; at the
same time, the combustible recovery was approximately 80%. Flotation also reduced the ash
content of the coal from 9.8% of feed coal to 6.5% of product coal.
DISCUSSION
The pyritic sulfur content of the product was maintained at around 1.2% during an entire
operation when Pittsburgh coal was processed (Figure 4). This value agreed well with the
0.98% on batch flotation with Pittsburgh coal in a previous study (Ohmura and Saiki, 1996).
Desulfurization efficiencies were approximately the same under both batch and continuous
conditions. However, 40% of pyritic sulfur in the feed coal still remained in the product coal
after microbial flotation. The following are thought to be possible reasons for the incomplete
pyrite removal:
A previous report showed that bacterial conditioning enhanced the removal of the liber-
ated pyrite particles but not that of the locked ones (Ohmura and Saiki, 1996). A part of
the locked pyrite in coal particles would not be rejected and remained in the product
coal. The locked pyrite may be the major form of pyrite in the product coal.
Another possibility for the incomplete rejection of pyrite may be that the froth still con-
tained fine liberated particles after flotation (Trahar and Warren, 1976; Kawatra and
Eisele, 1997). This meant that fine pyrite particles might entrain into the froth, even
though the particles are liberated. However, little information regarding the effect of
fine particle sizes on pyrite removal in microbial flotation was available.
Therefore, a series of microbial flotation experiments was carried out with three different
sizes of pyrite particles (53 to 75 µm, 26 to 38 µm and <26 µm), as shown in Figure 5. When
the particle size was 53 to 75 µm, the floatability of pyrite decreased with an increase in the
amount of added cells and finally reached 7.9%. In contrast, even with the excess addition of
bacteria, pyrite floatabilities with particle sizes of 26 to 38 µm and less than 26 µm were sup-
pressed only to 17.6% and 28.1%, respectively. It is more difficult to suppress pyrite float-
ability with fine pyrite particles than with coarse pyrite particles. These results suggest that
there are minimum sizes of pyrite particles that will sink during microbial flotation.
32
combustible recovery without cells
combustible recovery with cells
pyritic sulfur or ash in concentrate without cells
▼ pyritic sulfur or ash in concentrate with cells
FIGURE 4 Semicontinuous desulfurization and deashing of Pittsburgh coal by microbial column

flotation
To clarify how the fine particles could be removed, the actual sizes of the particles in the
froth and tailings were determined by a laser-scattering particle-size analyzer (LA-910W,
Horiba Ltd., Kyoto, Japan), as shown in Figure 6. When less than 26-µm pyrite was fed into the
flotation with bacteria, the mean size of froth and tailing was 6.5 and 15.1 µm, respectively.
It seemed to be difficult to suppress the flotation of particles that were <10 µm in size.
Those results suggest that fine pyrite particles in the feed coal remained in the product coal.
This might be one of the reasons why microbial flotation could not completely eliminate the
pyritic sulfur content of the product coal.
Microbial flotation removes pyrite particles in coal by changing the surface property
through bacterial adhesion. Previously, it was shown that T. ferrooxidans had an affinity for
pyrite adhesion (Ohmura et al., 1993a). The adhesion of the bacterium increased with an
increase in the pyritic sulfur content of the coal (Bagdigian et al., 1986). The same phenome-
non concerning adhesion to coal was reported with the other iron oxidizer (Vitaya and Toda,
1991). Recently, the mechanism of adhesion of the bacterium to pyrite was made more clear
by the discovery of a pyrite-binding protein (Ohmura and Blake, 1997; Sasaki et al., 1997).
From these reports, it was thought that the pyrite removal in both synthetic and ROM coal
was induced by selective adhesion of the bacterium to pyrite.
33
–26 µm
26–38 µm
53–75 µm
FIGURE 5 Effect of particle size on suppression of pyrite by bacteria. Data points are means
±S.D. of triplicate determinations. Floatability was calculated as described in the Materials and
Methods section.
Feed
♦ Froth
Tailing
FIGURE 6 Size distribution of froth and tailing after the microbial flotation with fine pyrite particles.
Fine pyrite particles (–26 µm) were fed into the flotation column with 4.2 × 108 cells.
However, a decrease in the combustible recovery was observed in the flotation with bacte-
ria when compared with that without bacteria (Figures 3 and 4). It was reported that the
excess addition of the bacterium induced not only the adhesion to pyrite but also that to coal
(Ohmura and Saiki, 1994; Ohmura and Saiki, 1996). The loss of coal recovery would be due
to a decrease of coal floatability by nonselective adhesion to coal particles due to the physical
interactions between coal particle and cell.
At present, the available information for microbial flotation is insufficient to incorporate the
process in an industrial situation. The present study provides information for a semicontinuous
system to enhance the removal of sulfur from coal by bacterial conditioning. There is now a
34
perspective of how a continuous-flow system would operate. It was shown that the flotation
column in the system could treat 0.19 kg/h of coal when the concentration and feed rate of
coal slurry was 1% (by weight) and 315 mL/min, respectively (Figure 4). If the system oper-
ated continuously, it would be possible to treat 4.5 kg of coal in a day. To treat this amount of
the coal, 3.4 × 1013 cells would be needed. The cell producing ability of the culture tank in the
present system was estimated as 3.8 × 1012 cells/day at a dilution rate of 0.03 h–1. This sug-
gested that culture volume should be increased from 50 to 500 L for continuous operation.
It seemed to be difficult to establish a 500-L culture tank with continuous operation.
Recently, electrochemical cultivation for T. ferrooxidans was reported as one of the methods
for producing a high cell yield (Matsumoto et al., 1999). With this method, a cell density of
1.0 × 1010 cells/mL could be maintained, as compared to about 1 to 2 × 108 cells/mL in a con-
tinuous culture. Electric cultivation may reduce the culture volume from 500 to 5 L, to provide
cell biomass to a continuous-flow system of microbial flotation.
CONCLUSION
The authors constructed experimental facilities for the continuous microbial desulfurization
of coal. The facilities are comprised of a cultivation tank (with a working volume of 50 L) and
a flotation column (with an inner radius of 80 mm and a height of 1.5 m). The removal of
pyritic sulfur from a synthetic coal/pyrite mixture and ROM coal was investigated as follows:
T. ferrooxidans was cultivated at the rate of 1.8 × 1011 cells/h for 20 days continuously,
without precipitating iron compounds.
The desulfurization of a synthetic coal/pyrite mixture was carried out in the microbial
column-flotation system. The cells of 7.4 × 109 cells/g of coal were added to the flota-
tion column, and the sulfur content in the froth was reduced from 10.1% in the feed to
1.4% in the product. The combustible recovery was about 85%.
For the ROM coal, the sulfur content in the froth was reduced from 2.9% in the feed to
1.2% in the product by the bacterial addition of 7.7 × 109 cells/g of coal. The combusti-
ble recovery was about 80%.
The authors demonstrated the operation of a semicontinuous desulfurization system using
microbial column flotation. To develop the method further, microbial column flotation can be
stepped up from its use in batch experiments to its use in a continuous system.
REFERENCES
Atkins, A.S., Bridgewood, E.W., Davis, A.J., and Pooley, F.D., 1987, “A study of the suppression of
pyritic sulfur in coal froth flotation Thiobacillus ferrooxidans,” Coal Preparation, Vol. 5, pp. 1–13.
Atkins, A.S., 1990, “Developments in the biological suppression of pyritic sulfur in coal flotation,”
Bioprocessing and Biotreatment of Coal, D.L. Wise, ed., New York, Marcel Dekker, pp. 1–13.
Attia, Y.A., and Elzeky, M.A., 1985, “Biosurface modification in the separation of pyrite from coal
by froth flotation,” Coal Science and Technology, Vol. 9, Processing and Utilization of High-
Sulfur Coal, pp. 673–682.
Attia, Y.A., 1990, “Feasibility of selective biomodification of pyrite floatability in coal desulfurization
by froth flotation,” Resources, Conservation and Recycling, Vol. 3, pp. 169–175.
Bagdigian, R.M., and Myerson, A.S., 1986, “The adsorption of Thiobacillus ferrooxidans on coal
surfaces,” Biotechnology and Bioengineering, Vol. 28, pp. 467–479.
Blake, R.C., Howard, G.T., and McGinness, S., 1994, “Enhanced yields of iron-oxidizing bacteria
by in situ electrochemical reduction of soluble iron in the growth medium,” Applied and
Environmental Microbiology, Vol. 60, pp. 2704–2710.
Capes, C.E., McIlhinney, A.E., Sirianni, A.F., and Puddington, I.E., 1973, “Bacterial oxidation in
upgrading pyritic coals,” CIM Bulletin for November, pp. 88–91.
35
Couch, G., 1991, “Advanced coal cleaning technology,” IEA Coal Research.
Dogan, M.Z., Ozbayglu, G., Hicyilmaz, C., Sarikaya, M., and Ozcengiz, G., 1985, “Bacterial
leaching versus bacterial conditioning and flotation in desulfurization of coal,” 15th Congr. Int.
Mineralurgie, Vol. 2, pp. 304–313.
Eisele, T.C., and Kawatra, S.K., 1994, “Use of fungi and bacteria for depression of mineral and coal
pyrite,” Reagents for Better Metallurgy, Chapter 11, SME, pp. 91–100.
Elzeky, M., and Attia, Y.A., 1987, “Coal slurry desulfurization by flotation using Thiophilic bacteria
for pyrite depression,” Coal Preparation, Vol. 5, pp. 15–37.
Kawatra, S.K., and Eisele, T.C., 1993, “Depression of pyrite flotation by microorganisms as a
function of pH,” Processing and Utilization of High-Sulfur Coals V, Parekh and Groppo, eds.,
Elsevier, Amsterdam, pp. 139–148.
Kawatra, S.K., and Eisele, T.C., 1997, “Pyrite recovery mechanisms in coal flotation,” International
Journal of Mineral Processing, Vol. 50, No. 3, pp. 187–201.
Matsumoto, N., Nakasono, S., Ohmura, N., and Saiki, H., 1999, “Extension of logarithmic growth
of Thiobacillus ferrooxidans by potential controlled electrochemical reduction of Fe(III),”
Biotechnology and Bioengineering, Vol. 64, No. 6, pp. 716–721.
Ohmura, N., Kitamura, K., and Saiki, H., 1993a, “Mechanism of microbial flotation using
Thiobacillus ferrooxidans for pyrite suppression,” Biotechnology and Bioengineering, Vol. 41, pp.
671–676.
Ohmura, N., Kitamura, K., and Saiki, H., 1993b, “Selective adhesion of Thiobacillus ferrooxidans to
pyrite,” Applied and Environmental Microbiology, Vol. 59, pp. 4044–4050.
Ohmura, N., and Saiki, H., 1994, “Desulfurization of coal by microbial column flotation,”
Biotechnology and Bioengineering, Vol. 44, pp. 125–131.
Ohmura, N., and Saiki, H., 1996, “Desulfurization of Pittsburgh coal by microbial column
flotation,” Applied Biochemistry and Biotechnology, Vol. 61, pp. 339–349.
Ohmura, N., and Blake II, R., 1997, “Aporusticyanin mediates the adhesion of Thiobacillus
ferrooxidans to pyrite,” Proc. of International Biohydrometallurgy Symposium, IBS97-
BIOMINE97, PB1.
Raichur, A.M., Misra, M., and Smith, R.W., 1995, “Differential adhesion of hydrophobic bacteria
onto coal and associated minerals,” Coal Preparation, Vol. 16, pp. 51–63.
Sasaki, K., Ohmura, N., and Saiki, H., 1997, “Purification and localization of the receptor for the
adherence of Thiobacillus ferrooxidans to pyrite: another role for rusticyanin,” Proceedings of
International Biohydrometallurgy Symposium, IBS97-BIOMINE97, PP6.
Silverman, M.P., and Lundgren, D.G., 1959, “Studies on the chemoautotrophic bacterium
Ferrobacillus ferrooxidans. I. An improved medium and a harvesting procedure for securing
high cell yields,” Journal of Bacteriology, Vol. 77, pp. 642–647.
Stainthorpe, A.C., 1989, “An investigation of efficacy of biological additives for the suppression of
pyritic sulfur during simulated froth flotation of coal,” Biotechnology and Bioengineering,
Vol. 33, pp. 694–698.
Townsley, C.C., and Atkins, A.S., 1986, “Comparative coal fines desulphurization using the iron
oxidizing bacterium Thiobacillus ferrooxidans and the yeast Saccharomyces cerevisiae during
simulated froth flotation,” Process Biochemistry, Vol. 21, pp. 188–191.
Townsley, C.C., Atkins, A.S., and Davis A.J., 1987, “Suppression of pyritic sulphur during flotation
tests using the bacterium Thiobacillus ferrooxidans,” Biotechnology and Bioengineering, Vol. 30,
pp. 1–8.
Trahar, W.J., and Warren, L.J., 1976, International Journal of Mineral Processing, Vol. 3, pp. 103–131.
Vitaya, V.B., and Toda, K., 1991, “Kinetics and mechanism of the adhesion of Sulfolobus
acidocaldarius on coal surfaces,” Biotechnology Progress, Vol. 7, pp. 427–433.
36
Biobeneficiation of Mineral
Raw Materials
S.N. Groudev*
ABSTRACT
A process for the combined chemico-biological removal of iron present as oxide minerals in different
mineral raw materials (quartz sands, kaolins, clays, etc.) was developed. In the process, mineral raw
materials are leached at about 90°C with lixiviant containing microbially produced oxalic and hydro-
chloric acids. The leaching is carried out in mechanically stirred acid-resistant reactors for periods of
from 1 to 6 hours, depending on the iron content and the forms of iron in the raw materials being
leached. The iron contents of some sands treated by this method were lowered from levels that were in
the range of 0.035% to 0.088% Fe2O3 to less than 0.012% Fe2O3, making them suitable for the prep-
aration of high-quality glass. The iron contents of different kaolins were lowered from levels that were
in the range of 0.65% to 1.49% Fe 2O3 to levels in the range of 0.44% to 0.75% Fe 2O3. As a result, the
whiteness was increased from values of 55% to 87% to values of 86% to 92%. The iron content of clay
was lowered from 6.25% Fe2O3 to 1.85% Fe2O3, and this increased the “fireproofness” of the clay
from 1,670°C to 1,750°C.
A similar process was used for the leaching of aluminum from aluminosilicates, mainly clays and
kaolins. However, in this case, the microbial fermentation fluid containing citric acid was acidified by
means of sulfuric or hydrochloric acid or by means of different mixtures of mineral acids. For enhanc-
ing aluminum solubilization, the aluminosilicates were heated before leaching at 600° to 650°C for 1
to 2 hours. Over 90% of the aluminum present in different clays and kaolins were leached within 3 to
6 hours.
“Silicate” bacteria related to the species Bacillus circulans were used to leach silicon from low-grade
bauxite ores containing aluminosilicates as impurities. The bacterial action was connected with the
degradation of the mineral structures (by means of microbial metabolites such as organic acids and
exopolysaccharides), as well as with the selective separation of the rich-in-aluminum fine fractions,
which were retained by the mucilaginous capsules of the bacteria. The solid residues after treatment
were characterized by higher values of alumina and silicon module (Al 2O3:SiO2 ratio), and they were
suitable for processing by means of the Bayer process for recovering aluminum.
The bending strength and other ceramic properties of kaolins were improved by contact with well-
developed cultures of “silicate” bacteria. The improvement was caused mainly by bacterial metabolites
(exopolysaccharides) that acted as resins during drying.
* Department of Engineering Geoecology, University of Mining and Geology, Durvenitza, Sofia, Bulgaria.
37
Pyritic sulfur and different metals (i.e., uranium, vanadium, molybdenum, aluminum, etc.) were
removed from shales by means of acidophilic chemolithotrophic bacteria, which were able to use the
shale pyrite as a source of energy for their growth. The desulfurization of the oil shales turns them into
rich-in-kerogen concentrates suitable for producing petroleum-like oil.
Conclusions concerning the prospects of applying the above biobeneficiation processes under
commercial-scale conditions are presented.
INTRODUCTION
Biotechnological methods in mineral processing have, until now, only been applied on a
commercial-scale to the leaching of sulfide-bearing ores and concentrates by means of acido-
philic chemolithotrophic bacteria. At the same time, a large variety of microbial processes
connected with the removal of impurities from different mineral raw materials are known,
but most of them have only been applied on laboratory and pilot scales. Different het-
erotrophic microorganisms carry out a considerable portion of these processes. In contrast to
the autotrophic bacteria, the heterotrophic bacteria and fungi require organic carbon for
growth and energy-generation and, with a few exceptions, they do not derive any benefit
from the degradation of the minerals. This makes the heterotrophs less amenable for com-
mercial operations than the autotrophs. However, some leaching processes connected with
the activity of different heterotrophic microorganisms seem promising for a real application.
The same is also valid for some processes carried out by chemolithotrophic bacteria, which
remove some impurities (mainly pyritic sulfur) from mineral raw materials such as shales
and coals.
Data about some biobeneficiation processes that seem promising for use in minerals pro-
cessing are presented in this paper.
MICROBIAL REMOVAL OF IRON FROM MINERAL RAW MATERIALS

The iron oxides contained as impurities in quartz sands, kaolins, and clays can greatly lower
the quality of these mineral raw materials. Various nonbiological methods are used indepen-
dently or in different combinations to remove the iron. However, some of these methods,
e.g., magnetic separation and flotation, are not fit for universal application, and their effi-
ciency greatly depends on the properties of the mineral raw material being treated. The
chemical methods that consist of leaching with mineral or organic acids and treatment with
reducers usually are suitable for achieving a higher degree of iron removal. However, they
are more expensive, the operating conditions are very difficult and the processes are environ-
mentally dangerous.
The leaching of iron from oxide minerals by means of microorganisms is well known. Solu-
bilization is connected with the formation of organic acids and other metabolites acting as
complexing agents (Berthelin and Kogblevi, 1974), as well as with enzymatic and nonenzy-
matic iron reduction (Bromfield, 1954; Ottow, 1969; Lundgren et al., 1983). These processes
greatly influence the biogeochemical cycles, not only of iron but also of carbon and many met-
als. Until recently, it was considered that much of the Fe(III) reduction in natural environ-
ments was the result of nonenzymatic processes. However, a large number of microorganisms
that can obtain energy for growth by the complete oxidation of organic compounds to carbon
dioxide with Fe(III) as the sole electron acceptor have now been described, and it is becoming
increasingly apparent that the metabolism of these microorganisms is responsible for most of
the Fe(III) reduction taking place in sedimentary environments (Lovley, 1991).
It must be noted that, in most cases, the biological leaching of iron from oxide minerals
takes place at a low rate. Comparative experiments with different heterotrophic bacteria and
fungi isolated from soils, freshwater pipeline precipitates and iron ore deposits revealed that
most bacterial species leached iron from oxide minerals more efficiently in cases where these
bacteria were grown in the presence of the minerals being leached (Table 1) (Groudev et al.,
38
FIGURE 1 Leaching of iron from goethite by means of (1) Aspergillus niger and (2) Bacillus polymyxa
TABLE 1 Iron solubilization from kaolins by means of different microorganisms

Iron solobulized at 14 days, %
Leaching by means of Leaching by means of spent
cultures grown in the nutrient media containing
Microorganisms presence of kaolin microbial metabolites
Aspergillus niger 23.0 33.4
Penicillium sp. 13.2 17.0
Pseudomonas sp. 9.9 13.2
Bacillus polymyxa 4.1 1.5
Bacillus circulars 3.3 0.8
NOTE: The experiments were carried out at 30°C with a kaolin containing 1.22% Fe2O3.
1985). However, the iron release by most fungi and by some bacteria was more efficient in
cases where these microorganisms were grown prior to leaching in nutrient media contain-
ing sucrose (as a carbon and energy source) in the absence of iron minerals and where the
leaching was carried out by means of the culture solutions, i.e., the spent nutrient media. In
other experiments, the culture solutions in which the different microorganisms had been
grown were sterilized by means of membrane filtration, and they were also used to leach iron
from oxide minerals. It was found that the leaching carried out by the most active strains of
bacteria and fungi, as well as by some other strains, was due to the action of their soluble
metabolic products, while the principal mechanism of the rest of the strains in solubilizing
iron was an enzymatic one. This finding provided for the possibility of carrying out the leach-
ing of iron minerals under conditions that were much more favorable for achieving maxi-
mum iron-release rates and extents than those needed for achieving an optimum microbial
growth. The best results were achieved by means of microorganisms producing organic acids
(mainly oxalic acid). The most active strains were related to the Aspergillus and Penicillium
genera. Some strains of Bacillus, Clostridium, and Pseudomonas were able to dissolve iron
anaerobically, but the rates of iron release were lower than those achieved by most strains
aerobically (Groudev and Groudeva, 1986a).
The susceptibility of the different iron oxide minerals toward microbial leaching was dif-
ferent. The iron was solubilized from limonite more slowly than from goethite but more rap-
idly than from hematite (Table 2).
39
TABLE 2 Iron solubilization from limonite, goethite, and hematite by means of different
microorganisms
Iron solubilized at 14 days, %
Microorganism Limonite Goethite Hematite
Aspergillus niger 5.1 8.6 1.4
Penicillium sp. 2.1 4.4 1.0
Mucor pyriformis 2.0 4.1 0.9
Pseudomonas sp. 1.9 3.7 0.8
Bacillus sp. 1.2 1.7 0.6
Bacillus polymyxa 1.0 1.7 0.5
Micrococcus lactilyticus 1.0 1.5 0.5
Bacillus circulars 0.8 1.2 0.3
Micrococcus sp. 0.7 1.3 0.3
Rhodopseudomonas sphaeroides 0.6 1.2 0.3
Enterobacter cloacae 0.5 1.0 0.2
Enterobacter aerogenes 0.5 0.9 0.1
Candida lypolitica 0.3 0.8 0.1
Escherichia coli 0.3 0.7 0.1
The rates of iron release from the minerals were not constant (Figure 1). The dissolved
iron concentrations usually fell, in some instances, after reaching a maximum, and this was
due to reoxidation and precipitation of some solubilized iron and/or due to adsorption of
some iron onto the residual mineral. However, the iron concentrations in cultures of some
fungi fell only slightly during the cultivation and leaching. This was connected with reaching
a low pH (below 3), which prevented the chemical reoxidation of iron as well as the forma-
tion of large amounts of secreted complexing agents, which prevented the precipitation of
the dissolved iron. No iron was solubilized in the absence of microorganisms.
It was also found that a mixture of oxalic and hydrochloric acids was very efficient for
leaching of iron oxide minerals. These acids exert a marked synergetic effect during the
leaching. The hydrochloric acid serves as a donor of protons (H+) for the degradation of the
iron oxide minerals, while the oxalic acid enhances the leaching by reducing the Fe(III) to
Fe(II) and by complexing the iron ions.
Based on these data, a combined chemico-biological method for the removal of iron from
mineral raw materials was developed (Groudev et al., 1985; Groudev, 1987, 1988b). The
flowsheet of the method is shown in Figure 2. It includes the following major steps: the prepa-
ration of leach solution, the leaching of iron from mineral raw materials containing iron
oxides as impurities and the removal of iron from the pregnant solution.
Strains of heterotrophic bacteria and fungi (the most suitable are some strains of the fun-
gus Aspergillus niger) capable of producing oxalic acid are grown in suitable nutrient media
containing carbohydrates (e.g., molasses) as sources of carbon and energy. After the culti-
vation, the fermentation fluid is separated from the fungal mycelium by filtration, diluted
with water and acidified to pH 0.5 using hydrochloric acid. The resulting fluid is used as a
leach solution.
Different mineral raw materials, i.e., quartz sands, kaolins, clays, kieselguhrs (diatoma-
ceous earths) and bauxites, containing iron oxides as impurities were treated by such solu-
tions. The concentration of oxalic acid in the solutions, i.e., the extent of dilution of the
fermentation fluid with water, depended on its intended use. For example, to remove iron
from sands, which normally contained less than 0.1% Fe2O3, solutions containing 2 to 3 g/L
40
FIGURE 2 Flowsheet of a process for the microbial removal of oxide iron from mineral raw materials
oxalic acid were found to be very efficient. However, for removal of iron from some rich-in-
iron clays, the concentration of oxalic acid was higher than 30 g/L.
The leaching was carried out in acid-resistant reactors with mechanical stirring at about
90°C and under batch or continuous-flow conditions. A countercurrent-leach unit consisting
of two to four reactors connected in a series was very suitable.
The iron contents of some quartz sands treated by this method were lowered from the
range of 0.035% to 0.088% Fe2O3 to less than 0.012 % Fe2O3 (Table 3) within 60 to 90 min-
utes, making them suitable for the preparation of high-quality glass. The iron contents of
different kaolins were lowered from the range of 0.65% to 1.49% Fe2O3 to the range of
0.44% to 0.75% Fe2O3 (Table 4) within 3 to 6 hours and, as a result of this, their whiteness
was increased from levels of 55% to 87% to levels of 86% to 92%. The calcium was also
removed during the leaching. Such kaolins can be used for the preparation of high-quality
chinaware, as well as in the paper industry. The iron content of clay was lowered from
41
TABLE 3 Removal of iron from sands by means of leach solution containing microbial metabolites
Fe2O3 content, %
Sand no. Before leaching After leaching Fe2O3 released, %
1 0.035 0.009 74.3
2 0.045 0.008 82.2
3 0.051 0.008 84.3
4 0.056 0.010 82.1
5 0.056 0.016 71.4
6 0.062 0.014 77.4
7 0.068 0.010 85.3
8 0.071 0.009 87.3
9 0.080 0.009 88.8
10 0.083 0.018 78.3
11 0.083 0.014 83.1
12 0.088 0.012 86.4
TABLE 4 Removal of iron and aluminum from different kaolins and a clay by means of different
microorganisms
Fe2O3 content, % Al2O3 content, %
Before After Fe2O3 Before After Al2O3
Mineral raw material leaching leaching released, % leaching leaching released, %
Kaolin no.
1 0.65 0.44 32.3 32.0 31.7 0.9
2 0.80 0.51 36.3 34.2 34.0 0.6
3 0.91 0.50 45.1 33.5 33.2 0.9
4 1.05 0.44 58.1 32.7 32.3 1.2
5 1.13 0.47 58.4 32.3 32.0 0.9
6 1.16 0.52 55.2 34.3 34.0 0.9
7 1.22 0.55 63.1 34.7 34.3 1.2
8 1.27 0.60 52.8 32.7 32.3 1.2
9 1.32 0.60 54.6 34.1 33.8 0.9
10 1.40 0.68 51.4 33.2 32.9 0.9
11 1.49 0.75 49.7 35.2 34.8 1.1
Clay 6.25 1.85 70.4 33.8 33.2 1.8
6.25% to 1.85% Fe2O3 (Table 4). The “fireproofness” of the clay was increased from 1,670°
to 1,750°C, making it suitable for preparing fireproof materials. Only small amounts of alu-
minum (less than 1%) were leached together with iron from the kaolins and the clay.
It must be noted that the final iron content after the leaching did not depend on the initial
iron content before leaching. Instead, it depended on the forms in which the iron was present
in the mineral raw materials. Only the iron not included in the crystal lattice of the minerals
is removed by this method (as well as by any other hydrometallurgical method).
After leaching, the dissolved iron can be extracted from the pregnant solutions as hydrox-
ide precipitates by increasing the pH. More attractive, however, is the removal of iron by sol-
vent extraction or ion exchange, which provides for the possibility of recycling the solutions
42
more efficiently. The above-mentioned chemico-biological method can be combined with
conventional methods such as flotation and magnetic separation.
MICROBIAL LEACHING OF ALUMINUM FROM MINERAL RAW MATERIALS

Different methods for the extraction of aluminum from mineral raw materials are known, but
the Bayer process is the only one that has been widely used on an industrial scale. However,
only high-quality bauxite ores can be treated by this method, and, despite some recent modi-
fications, it still presents several shortcomings. The geographical distribution on a world
scale of the high-quality bauxite ores is quite uneven, and the bauxite ore reserves deplete
rapidly. On the other hand, different aluminum-bearing nonbauxite raw materials, such as
clays and kaolins, have an almost ubiquitous distribution. Some of them, especially those rich
in alumina, can be regarded as aluminum ores. At the present time, the recovery of alumi-
num from such raw materials is not competitive to the treatment of bauxite by the Bayer pro-
cess. This is due not only to their lower alumina content in comparison with that of the
bauxites but also to the lack of efficient processing methods.
A wide range of microorganisms has been reported to be capable of leaching aluminum
from different minerals and rocks, generally through the production of organic acids
(Silverman and Munoz, 1971; Eckhard, 1978; Mehta et al., 1979; Groudev et al., 1982a;
Vachon et al., 1994; Gazso, 1997). The organic acid attack on minerals involves both H+ ion
attacks on the mineral matter and chelation of the dissolved metals, with the mechanisms
applying to both silicates and nonsilicates. The most active acids seem to be citric and oxalic,
though some other acids are also effective in some cases.
Chemolithotrophic microorganisms that can grow at the expense of energy liberated from
microbial degradation of aluminosilicates structures have not yet been isolated. Some chem-
olithotrophic bacteria (mainly thiobacilli and nitrifying bacteria) can degrade aluminosilicates
by means of mineral acids (such as sulfuric and nitric) produced as a result of the bacterial
oxidation of some sulfur or nitrogen inorganic compounds.
In most cases, the biological leaching of aluminum from aluminosilicates proceeds at slow
rates. However, some results obtained from leaching aluminum from clays (Groudev et al.,
1982a) suggest that such a process could be feasible on a large scale. Furthermore, because
the aluminum solubilization is due to the action of secreted soluble metabolites, it is possible
to carry out the leaching under conditions that are much more favorable for achieving both
maximum aluminum release rates and yields than are needed for achieving an optimum
microbial growth. It has been shown that mixtures consisting of some organic and mineral
acids are very suitable as leach solutions for aluminum extraction from aluminosilicates. On
this basis, a combined chemico-biological method for aluminum extraction was developed
(Groudev et al., 1982b; Groudev and Groudeva, 1986b; Groudev, 1988a). The flowsheet of
the method (Figure 3) is similar to that of the method for iron removal. It includes the fol-
lowing major steps: the preparation of leach solution, the leaching of aluminum from suit-
able aluminum-bearing mineral raw materials and the extraction of aluminum from the
pregnant solution.
Strains of heterotrophic bacteria and fungi (the most suitable are some strains of the fungus
Aspergillus niger) capable of producing organic acids (mainly citric acid) are grown in suitable
nutrient media containing carbohydrates as sources of carbon and energy. After the cultiva-
tion, the fermentation fluid is separated from the fungal mycelium by filtration, diluted with
water and acidified to a pH of 0.5 by either hydrochloric or sulfuric acid or by a mixture of
mineral acids. The resulting fluid is used as leach solution.
Different raw materials can be a target for aluminum extraction by the above-mentioned
leaching method. The aluminosilicates are the largest and most important group of aluminum-
bearing minerals. In principle, the susceptibility of aluminosilicates toward degradation
decreases with an increase in the amount of the Si-O-Si bonds. Furthermore, any isomorphous
substitution will, in general, decrease the stability of the relevant mineral structure. The clays
43
FIGURE 3 Flowsheet of a process for the microbial leaching of aluminum from mineral raw
materials
are suitable to be used as sources of aluminum because they are abundant and relatively
cheap, the costs for their comminution are low and some of them are characterized by a rela-
tively high aluminum content. The susceptibility of the different clays to biodegradation is
quite different (Groudev and Groudeva, 1986b). The dioctahedral clay minerals (kaolinite,
halloysite, illite, montmorillonite, etc.) are more stable than the trioctahedral clay minerals
(vermiculite, serpentine, chrysotile, genthite, etc.). Chlorites are intermediate in behavior.
The susceptibility is increased considerably by heating the clays at 600° to 650°C for 1 to 2
hours. The treatment caused amorphication of the raw material due to the separation of
water from the hydroxylic groups in the crystalline structures of the clay minerals. The heat
treatment not only enhances the aluminum leaching but also inhibits the iron leaching
44
(a)
(b)
FIGURE 4 Biological leaching of aluminum and iron from (a) untreated and (b) thermally
activated clay
(Figure 4). The latter effect is considered very important, as iron impedes the subsequent
extraction of aluminum from the pregnant solution.
The leaching is performed in acid-resistant reactors with mechanical stirring at a tempera-
ture of about 90°C and can be performed under batch or continuous-flow conditions. A
counter-current leach unit consisting of two to four reactors connected in series is very suit-
able. Under optimum conditions, as much as 90% to 95% of the aluminum was extracted
within 3 to 4 hours from some thermally pretreated clays (Groudev et al., 1982). Efficient
extraction can also be obtained from some other aluminum-bearing mineral raw materials,
such as anorthosite, alunite, and low-grade bauxites.
Various techniques have been used for recovering aluminum from the pregnant solutions
(Groudev, 1988a). The aluminum may be precipitated as aluminum hydroxide by auto-
claving these solutions and raising their pH. Al(OH)3 precipitation may also be achieved by
increasing the redox potential of the pregnant solution. The Al(OH)3 is then converted to alu-
mina (Al2O3) by means of calcining.
The aluminum may also be recovered as potassium aluminum sulfate (alum) by adding
potassium sulfate and sulfuric acid to the pregnant solution, heating the solution at 90°C for
15 minutes and then cooling.
The aluminum may also be precipitated as AlCl3·6H2O by saturating the pregnant solution
with hydrogen chloride. The precipitate is separated by filtration and dissolved in 20%
hydrochloric acid. This solution is saturated with hydrogen chloride, and the aluminum is
reprecipitated as purified crystals of AlCl3·6H2O.
45
TABLE 5 Leaching of aluminum from clay minerals by culture solutions produced by Aspergillus
niger (strain G-8) and acidified to pH 0.5 with sulfuric acid
Al2O3 solubilized at 4 hrs, %
Thermally
Mineral Untreated mineral activated mineral
Kaolinite 4.1 62.8
Halloysite 10.4 74.7
Illite 12.5 77.0
Montmorillonite 2.6 18.5
Vermiculite 34.7 95.6
Serpentine 17.3 84.2
Chrysotile 18.5 83.0
Genthite 24.2 91.4
The organic acids are separated from the aluminum during the treatment of the pregnant
solutions and can be used again as leaching agents. This enhances the commercial feasibility
of the combined chemico-biological method.
MICROBIAL REMOVAL OF SILICON FROM LOW-GRADE BAUXITES

It has been reported that different “silicate” bacteria are able to remove silicon from low-grade
bauxites containing aluminosilicates (mainly kaolinite) as impurities (Groudev et al., 1982a;
Polkin et al., 1982; Groudeva and Groudev, 1983; Yakhontova et al., 1987; Melnikova et al.,
1990; Malinovskaya et al., 1990; Gazso, 1997). The bauxites after treatment are characterized
by higher values of alumina and silicon module (i.e., the alumina to silica ratio) than those pro-
duced by the conventional methods (e.g., gravimetric separation and flotation) and are suit-
able for processing by means of the Bayer process for recovering aluminum.
The “silicate” bacteria are characterized by well-expressed requirements for silicon during
their growth. It has even been reported that the action of these bacteria on aluminosilicates is
of an enzymatic nature and that they are able to utilize energy released from the aluminosili-
cates biodegradation (Yakhontova et al., 1987). However, it is now assumed that the “sili-
cate” bacteria are typical heterotrophic microorganisms (Groudev, 1987; Ogurtsova et al.,
1991). Their taxonomic status is uncertain. A large number of such bacteria have been iso-
lated by different researchers and have been related to different species of the genus Bacillus.
It has been reported that the “silicate” bacteria are a “mucilaginous phase” during the growth
and of the well-known soil bacterium Bacillus circulans or merely are its ecotypes and variet-
ies (Tesic, 1972; Groudev, 1990). It is known, however, that the species Bacillus circulans is a
very heterogeneous group, and its phenotypic heterogeneity is not due to inherent variability
of genetically related strains but rather is due to inclusion of genetically unrelated organisms
in this taxon (Yakhontova et al., 1991).
Typical “silicate” bacteria are red-shaped cells (2.0 to 5.0 µm by 0.7 to 1.5 µm). They form
large slimy capsules (5 to 10 µm by 7 to 20 µm) during growth on synthetic, nitrogen-free or
high C:N-ratio nutrient media. The capsules consist of exopolysacaharide that contains galac-
tose, sucrose, mannose, and small amounts of amino sugars. The relative amounts of these
components vary in the different strains and partially depend on growth conditions.
Organic acids, mainly citric, gluconic and oxalic are produced from the carbohydrate sub-
strates during bacterial growth and are secreted into the culture media. The optimum growth
is at neutral or slightly acid pH. Some strains survive at pH<4. The optimum growth temper-
ature is 30° to 37°C. Some strains grow under anaerobic conditions.
46
TABLE 6 Bauxite treatment by means of “silicate” bacteria and their metabolites
Extraction, %
Way of treatment SiO2 Al2O3
Washed cell suspension without glucose and 3.29 0.09
with mineral salts
Washed cell suspension with 1% glucose 60.26 12.92
Washed cell suspension with 1% glucose and mineral salts 71.80 14.23
Washed cell suspension with 1% glucose, mineral salts, and 76.43 15.05
solution replacement
Culture solution without cells 1.35 1.94
Spent nutrient medium containing cells 36.87 10.07
Ashby’s sterile nutrient medium 0.03 0.01
NOTE: Leaching was performed in agitated flasks at 30°C for 10 days. When necessary, fresh solution replacement was car-
ried out after the fifth day of treatment.
TABLE 7 Bauxite d ressing by means of Bacillus circulars under continuous-flow leaching conditions
Component, % Extraction, %
Way of
Product Recovery % Al2O3 SiO2 Al2O3 SiO2 Ms* treatment
Bauxite ore 100 43.4 25.9 100 100 1.7 —
Concentrate 57.6 63.9 9.1 84.8 20.2 7.0 Biological
Solution 42.4 15.5 48.7 15.2 79.8 0.3
Concentrate 63.5 47.9 21.8 70.1 53.4 2.2 Flotation
Solution 36.5 35.6 33.1 29.9 46.6 1.1
*Ms is the silicon module, i.e., the Al2O3:SiO2 ratio.
The degradation of aluminosilicates by “silicate” bacteria greatly depends on the way of

treatment (Table 6). At present, the action of “silicate” bacteria on aluminosilicates is
connected with the degradation of the mineral structures and solubilization of their com-
ponents (Groudev and Groudeva, 1988), as well as with the selective separation of the
rich-in-kaolinite fine fractions from the bauxites (Karavaiko et al., 1989). The degrada-
tion is carried out by means of organic acids (mainly oxalic and citric acids) secreted by
the bacteria. The mucilaginous bacterial capsules, consisting of exopolysaccharides, then
bind a significant portion of the dissolved silicon and aluminum. These mucilaginous cap-
sules retain also the fine kaolinite particles.
It has been found that the different natural specimens of kaolinite differ considerably from
each other with respect to their amenability to both chemical and biological degradation. The
specimens with a well-arranged structure consisting of relatively large blocks of helicoidal lay-
ers with high concentrations of helicoidal dislocations and with high permeability are more
amenable to degradation than the other specimens (Yakhontova et al., 1991). The ability of
the different strains of “silicate” bacteria to remove silicon from bauxites is also different. The
best results have been achieved by the strains possessing larger mucilaginous capsules and a
higher acid formation ability (Groudev, 1990).
The bacterial treatment of low-grade bauxites has, as yet, only been applied on laboratory
and pilot scales under both batch and continuous-flow conditions using reactors with
mechanical stirring and enhanced aeration (Table 7) (Groudev and Groudeva, 1988). The
bauxites being treated are finely ground (usually to less than 100 µm) and the leach solu-
tions inoculated with “silicate” bacteria also contain soluble organic compounds (usually
47
sugars) as sources of carbon and energy. The pH of the mineral suspension is usually main-
tained in the range of 4 to 8, i.e., under conditions at which the bauxite minerals (gibbsite,
boehmite and diaspore) are relatively stable and refractory to leaching. The main problems,
which should arise in connection with the commercial-scale application of such treatment,
are the long residence times (usually 2 to 3 weeks) and the possible contamination of the
leach suspensions with undesirable heterotrophic microorganisms.
IMPROVEMENT OF THE CERAMIC PROPERTIES OF KAOLINS

In the past, in China and some other east Asian countries, clays and kaolins intended for
preparations of fine ceramics were matured for years in humid pits. Small amounts of plant
wastes were occasionally added to the surface of the clay material in the pits. This practice
improved the ceramic properties of the minerals. At present, it is clear that the improvement
in properties was due to alterations caused by microorganisms that were naturally associated
with the clay minerals. Little is known about the microbial communities in such ecosystems,
the distribution and the number of microorganisms in the clay or the biological processes
occurring during the maturation of the clay.
Information about the feasibility of improving ceramic properties of clays and kaolin by
microbial treatment is scarce (Vlasov et al., 1980; Kromer et al., 1988; Groudev, 1990;
Groudeva and Groudev, 1993). Heterotrophic bacteria and fungi, as well as the chem-
olithotrophic Thiobacillus thiooxidans, have been used in these studies. The microorganisms
have been used to treat kaolins in two ways: by means of microbial cultures growing on an
organic substrate (e.g., sucrose) in the presence of the kaolin being treated and by means of
microbial cultures grown prior to the treatment on sucrose in the absence of kaolin.
Strains of the “silicate” bacteria were the most efficient for improving the ceramic proper-
ties of the kaolins. Both methods of treatment exerted positive effects. However, the treat-
ment of the kaolins by means of preliminary grown cultures was usually more efficient than
the treatment by means of cultures growing in the presence of the kaolins (Table 8)
(Groudeva and Groudev, 1993).
The most important factor affecting the microbial action was the ratio between the
amount and density of the microbial culture being used, on the one hand, and the amount of
the kaolin being treated, on the other. The residence time and temperature during treatment
were not as important. Even contact for 1 to 4 hours of well-developed “silicate” bacteria cul-
ture (containing 108 cells/mL and 10 g/L of exopolysaccharides) resulted in improving the
bending strength and other ceramic properties of the kaolins. Short contact slightly changed
the chemical composition of the kaolins and did not affect their alumina content. Usually,
less than 1% of the kaolins were degraded after microbial treatment for 72 hours (Table 9).
Longer contacts increased the degree of kaolin degradation.
Different strains related to the same species often differed considerably from each other
with respect to their ability to improve the bending strength of the kaolins. On the other
hand, the amenability of the different kaolins toward microbial action was also different.
The positive effect achieved by the “silicate” bacteria and most of the other microorgan-
isms was caused mainly by the mucilaginous exopolysaccharides produced during their
growth (Table 10). They acted as resins during drying. The heteropolysaccharides usually
were more efficient than the homopolysaccharides in this respect. The sugars in the nutrient
media also acted as resins. In some cases, the organic acids produced by the bacteria reduced
the size of the kaolin particles. This increase in the contact surface has a positive effect on the
plasticity of the kaolin. Some colloidal substances formed as a result of the partial degrada-
tion of the kaolins being treated also caused a positive effect. These were mainly iron and
aluminum hydroxides and silicon compounds.
This treatment can be conducted under nonaseptic conditions, which makes it technically
feasible. The process is economically attractive on a commercial scale only if an inexpensive
waste biomass is used to treat the kaolins.
48
TABLE 8 Effect of different microorganisms on the bending strength of kaolins
Bending strength, kg/cm2
Type of Kaolin No. 3 Kaolin No. 5

Microorganisms exopolysaccharides I II I II
Aeromonas sp. Fructane 10.4 12.0 17.4 19.0
Alcaligenes sp. Curdlane 11.3 12.3 19.0 20.1
Bacillus circulars HPS 14.5 16.0 32.0 35.8
Bacillus megaterium HPS 11.9 11.7 19.0 20.4
Bacillus mesentericus HPS 11.4 12.5 20.8 23.3
Bacillus mucilaginosus HPS 14.0 15.4 27.0 31.4
Bacillus polymyxa HPS 12.0 13.6 24.4 28.6
Bacillus subtilis HPS 11.4 11.3 18.1 20.8
Brevibacterium sp. Dextrane 10.8 11.0 17.7 18.5
Methylomonas sp. Glucane 9.9 10.7 16.4 17.7
Pseudomonas aeroginosa Alginate 9.9 11.0 16.9 19.1
Xanthomonas campestris Xanthane 9.5 10.7 15.3 16.1
Sterile controls:
Distilled water 8.2 8.2 13.7 13.7
Solution of sucrose in distilled water:
2 g/L 9.1 9.1 15.0 15.0
5 g/L 9.9 9.9 16.1 16.1
20 g/L 10.7 10.7 17.7 17.7
Ashby modified nutrient medium 11. 0 11.0 18.0 18.0
Ashby modified nutrient medium 8.4 8.4 14.0 14.0
without sucrose
BACTERIAL LEACHING OF SHALES

Shales are fine-grained sedimentary rocks consisting mainly of clay minerals such as illite
and montmorillonite mixed with fine particles of quartz and mica. The shales are considered
important mineral resources in several countries. The oil shale contains considerable
amounts of organic material (kerogen) that can be processed to yield petroleum-like oil. Like
petroleum, shale oil can be refined into gasoline, fuel oil and many other products. The black
shale (containing relatively lower amounts of organic material than the oil shale) can be
used as a raw material for the preparation of building materials, a component for cement
production, a fertilizer and plant stimulant, and a raw material for the extraction of metals
(Maremäe, 1988).
Until now, the biobeneficiation of shales includes the following two processes:
The leaching of metals (uranium, vanadium, molybdenum, aluminum, etc.) by means of
acidophilic chemolithotrophic bacteria (Thiobacillus ferrooxidans, Thiobacillus thiooxidans
and Sulfolobus acidocaldarius) (Iskra et al., 1980; Tasa, 1998). These bacteria grow on
shales using the sulfide minerals (mainly pyrite) included in the shales as impurities.
The removal of pyritic sulfur also by means of acidophilic chemolithotrophs. It has been
found that the bacterial leaching is the most suitable method for shale desulfurization,
because it causes only negligible alterations in the fuel structure and characteristics
(Vrvic et al., 1988; Tasa and Lindström, 1996; Tasa, 1998). The clay minerals in the
shales are also solubilized during the leaching by a proton attack caused by the sulfuric
49
TABLE 9 Effect of microbial treatment with preliminary grown Bacillus circulars on the composition
and ceramic properties of kaolin
Variable Before treatment After treatment
Al2O3, % 32.7 32.4
SiO2, % 50.9 50.6
Fe2O3, % 1.13 1.02
TiO2, % 0.24 0.21
CaO, % 0.82 0.75
Loss on drying, % 0.34 1.25
Loss on burning, % 12.0 13.1
Bending strength on drying, kg/cm2 13.7 35.8
Plasticity, % 32.3 37.0
Drying shrinkage, % 3.4 3.7
Firing shrinkage at 960°C, % 4.4 4.6
Water saturation capacity at 960°C, % 28.4 30.2
Formation water, % 32.5 34.7
Whiteness on drying, % 76.5 78.5
Whiteness on firing at 1,350°C under 85.1 87.5
reduction atmosphere, %
Size fractions, %
>60 µm 0.10 0.02
10–60 µm 1.93 0.68
5–10 µm 12.25 8.15
1–5 µm 34.08 34.43
<1 µm 51.64 56.72
NOTE: Pulp density = 40%, residence time = 72 hours and temperature = 30°C.
TABLE 10 Treatment of kaolin using different agents obtained from a culture of Bacillus circulars
Bending strength,
Agent used for kaolin treatment kg/cm2
Washed cells without slimy capsules consisting of exopolysaccharides 14.0
Washed cells and sucrose (2%) 17.7
Washed cells and sucrose (2%) and mineral salts 18.1
Culture solution (spent nutrient medium) without cells and 14.5
exopolysaccharides
Culture solution containing viable cells and exopolysaccharides 35.6
Culture solution containing dead cells and exopolysaccharides 33.0
Culture solution containing no cells but containing exopolysaccharides 33.4
acid, which is generated as a result of the pyrite oxidation. The desulfurization of the oil
shales turns them into rich-in-kerogen concentrates, which are suitable for producing
petroleum-like oil.
OTHER BIOBENEFICIATION PROCESSES

Microbial Fe(III) reduction has been proposed as a mechanism for the extraction of Fe(III)
from ore (Arnold et al., 1988). Fe(III)-reducing microorganisms could potentially selectively
dissolve the Fe(III) in the ore, yielding the soluble iron form (Fe[II]), which would require
50
less reduction for conversion to elemental iron than would be required for Fe(III). A micro-
bial process for removing Fe(III) impurities from low-grade kaolins by means of such micro-
organisms has been developed (Hintz et al., 1977).
A mixed culture containing Enterobacter aerogenes and Leuconostoc mesenteroides was
used in a nutrient medium containing molasses, urea, and kaolin. Depending on the kaolins
utilized, the microorganisms removed from 27.5% to 43.9% of the initial Fe(III). The treat-
ment did not cause any unfavorable modifications in the chemical and physical characteris-
tics of the kaolins, and it increased the whiteness of the kaolins upon burning.
“Silicate” bacteria are used as biofertilizers in the bioremediation of contaminated soils
and dumps consisting of low-grade ores and coals (Groudev, 1997b). The bacteria degrade
aluminosilicates presented in the soils and dumps and solubilized potassium and some
microelements that are essential for the growth of the soil microflora and plant communities.
Furthermore, these bacteria stimulate the growth of the microbial and plant communities
also by secreting biologically active substances related to the giberelines and heteroauxines.
“Silicate” bacteria have been used under laboratory conditions to pretreat gem-bearing
rocks to facilitate the subsequent physical separation of the gems (Groudev, 1997b) and facil-
itate the precipitation of the red mud resulting from the Bayer process (Gazso, 1997).
“Silicate” bacteria have also been used to remove silicon and phosphorous from a manga-
nese ore containing 40.0% Mn, 18.8% SiO2 and 0.35% P (Karavaiko et al., 1988). The treat-
ment was carried out in reactors with mechanical stirring and intensive aeration. About 40%
of the silicon and 35% of the phosphorous were removed as a result of the bacterial action.
The P:Mn ratio was decreased from 0.09 to 0.004, making the ore to be related to the highest-
quality type.
More than 60% of the silica were removed within eight days from a magnesite ore by
means of a strain of Bacillus sp. (Mohanty and Mishra, 1991). Different sugars (glucose, fruc-
tose, sucrose, and commercial sugar) were used as carbon and energy sources in this study.
The treatment was carried out in a stainless steel fermentator with mechanical stirring and
enhanced aeration. It was found that the release of silica from the ore was independent of
particle size.
A culture of Hyphomicrobium sp. was able to remove about 30% of the phosphorous from
a manganese ore. The removal of phosphorous by means of Aspergillus niger was not so effi-
cient (Agate, 1985). Nitrifying bacteria have also been used for removal of phosphorous from
manganese ore (Khan et al., 1991). The phosphate content of the ore was reduced from
0.25% to about 0.16% within 53 days of treatment. The results obtained also revealed that
manganese of the ore did not leach out with phosphate resulting in beneficiation of the ore.
Different heterotrophic bacteria and fungi have been found to solubilize phosphate from
an Idaho rock phosphate ore (Rogers and Wolfram, 1991). More than 80% of the phosphate
were solubilized within nine days under optimum conditions with respect to nutrient supply
and pH. In all cases, the analysis of the solutions in which phosphate had been extracted
from the ore showed that organic acids, which have been implicated as the mechanism of sol-
ubilization, were present.
A very efficient removal of phosphorous from iron ores was achieved by means of different
acid-producing heterotrophic bacteria and fungi (Groudev, 1997b).
Organic acids-producing microorganisms have been reported to leach metals from differ-
ent mineral raw materials (Burgstaller and Schinner, 1993; Piga et al., 1993; Vachon et al.,
1994; Tzeferis et al., 1994; Tarasova et al., 1995; Tungkaviveshkul et al., 1995; Ehrlich et al.,
1995; Veglio et al., 1995). Similar processes are used under field conditions to treat soils
contaminated with heavy metals, arsenic, and radioactive elements (Groudev, 1997a).
Different acidophilic chemolithotrophic bacteria are able to remove efficiently pyritic sul-
fur from coals (Rossi, 1993; Andrews and Noah, 1995). The information about this process is
abundant. However, the bacterial desulfurization until now has been applied only under lab-
oratory and pilot-scale conditions in spite of some very promising data.
51
CONCLUSIONS
The microbial treatment of nonsulfide-mineral raw materials includes a large variety of pro-
cesses. However, with a few exceptions, these processes are carried out by means of secreted
microbial metabolites. Only the solubilization of some iron oxide minerals in some cases is
connected with an enzymatic reduction of Fe(III). The desulfurization of shales and coals is
connected with the bacterial oxidation of pyrite, which is carried out by acidophilic chem-
olithotrophic bacteria. In most cases, the secreted metabolites are organic compounds that
can be used by different heterotrophic microorganisms as sources of carbon and energy. This
is an essential barrier for these processes to be carried out under commercial-scale conditions
where it is not possible to prevent the contamination by undesirable microorganisms. For
that reason, the treatment of mineral raw materials by heterotrophic microorganisms (de
facto by their metabolites) is arranged mainly as a process separated from the microbial
growth, during which the relevant metabolites are produced. This provides for the possibility
of arranging the treatment under conditions that are more favorable for achieving maximum
rates and extents of mineral degradation (e.g., higher temperature and pulp density and the
addition of some chemical reagents such as mineral acids) than those needed to achieve opti-
mum microbial growth.
However, it must be noted that, in some cases, the microbial production of secreted
metabolites, e.g., organic acids and polysaccharides, is too expensive to allow for their use in
the mineral-processing industry. For that reason, it is necessary to search suitable waste
products from the pharmaceutical, food, or chemical industries that can directly be used for
treatment of minerals.
REFERENCES
Agate, A.D., 1985, In Biogeotechnology of Metals, G.I. Karavaiko and S.N. Groudev, eds., Centre for
International Projects GKNT, Moscow, pp. 377–395.
Andrews, G.F., and Noah, K.S., 1995, In Minerals Bioprocessing II, D.S. Holmes and R.W. Smith,
eds., The Minerals, Metals & Materials Society, Warrendale, PA, pp. 219–230.
Arnold, R.G., DiChristina, T.J., and Hoffman, M.R., 1988, Biotechnol. Bioeng., 32, pp. 1081–1096.
Berthelin, J., and Kogblevi, A., 1974, Rev. Ecol. Biol. Sol, 11(4), pp. 499–509.
Bromfield, S.M., 1954, J. Gen. Microbiol., 11, pp. 1–6.
Burgstaller, W., and Schinner, F., 1993, J. Biotechnol., 27, pp. 91–116.
Eckhard, F.E.W., 1978, In Proc. 4th Int. Biodeterioration Symp., Berlin, p. 107.
Ehrlich, H.L., Wickert, L.M., Noteboom, D., and Doucet, J., 1995, In Biohydrometallurgical
Processing, Vol. I, T. Vargas, C.A. Jerez, J.V. Wiertz, and H. Toledo, eds., Universidad de Chile,
Santiago, pp. 395–403.
Gazso, L., 1997, Personal communication.
Groudev, S.N., 1987, Acta Biotechnol., 7, pp. 299–306.
Groudev, S.N., 1988a, In Biogeotechnology of Metals. Manual, G.I. Karavaiko, G. Rossi, A.D. Agate,
S.N. Groudev, and Z.A. Avakyan, eds., Centre for Int. Projects GKNT, Moscow, pp. 314–318.
Groudev, S.N., 1988b, ibid., pp. 318–322.
Groudev, S.N., 1990, “Microbiological Transformations of Mineral Raw Materials,” Doctor of
Biological Sciences Thesis, University of Mining and Geology, Sofia.
Groudev, S.N., 1997a, In Proc. XX Int. Min. Proc. Congr., Aachen, September 21–26, Vol. 5,
pp. 729–736.
Groudev, S.N., 1997b, Unpublished data.
Groudev, S.N., and Groudeva, V.I., 1986a, Ind. Minerals, No. 222, pp. 81–84.
Groudev, S.N., and Groudeva, V.I., 1986b, Biotechnol. Bioeng. Symp., 16, pp. 91–99.
Groudev, S.N., and Groudeva, V.I., 1988, In Biohydrometallurgy, P.R. Norris and D.P. Kelly, eds.,
Antony Rowe Ltd., Chippenham, UK, pp. 397–406.
Groudev, S.N., and Groudeva, V.I., 1993, In Third European Ceramic Society Conference, Madrid,
September 12–17.
52
Groudev, S.N., Genchev, F.N., and Groudeva, V.I., 1982a, Travaux ICSOBA, 12(17), pp. 203–212.
Groudev, S.N., Genchev, F.N., Petrov, E.C., Groudeva, V.I., Mochev, D.J., Genchev, D., and
Rusenov, G., 1982b, Bulgarian Patent No. 38860.
Groudev, S.N., Groudeva, V.I., Genchev, F.N., Mochev, D.J., and Petrov, E.C., 1985, In Proc. XVth
Int. Min. Proc. Congr., Cannes, June 2–9, Vol. II, pp. 378–387.
Groudeva, V.I., and Groudev, S.N., 1983, Travaux ICSOBA, 13 (18), pp. 257–263.
Hintz, I., Kiss, S., Papacostea, P., Radulescu, D., and Dragan-Bularda, M., 1977, In Fourth
Symposium on Soil Biology, Cluj-Napoca, Romanian National Society for Soil Science,
Bucharest.
Iskra, A.A., Nosov, V.D., and Shatalov, V.V., 1980, In Proc. of the Int. Conf. on Use of
Microorganisms in Hydrometallurgy, Hungarian Academy of Sciences, Pecs., pp. 34–41.
Karavaiko, G.I., Avakyan., Ogurtsova, L.V., and Safonova, O.F., 1989, In Biohydrometallurgy 89,
J. Salley, R.G.L. McCready, and P.C. Wichlacz, eds., CANMET SP89-10, pp. 93–102.
Karavaiko, G.I., Rossi, G., Agate, A.D., Groudev, S.N., and Avakyan, Z.A., 1988, Biogeotechnology
of Metals. Manual, Centre for International Projects GKNT, Moscow.
Khan, S.S.U., Bansal, S.S.K., Narayanrao, S.V., Deshpande, W., and Chakraborty, T., 1991, In
National Symposium on Application of Geomicrobiology in India, Dhule, March 25–27,
pp. 80–92.
Kromer, H., Mortel, H., and Ditz, H., 1988, Science of Ceramics, 14, pp. 113–118.
Lovley, D.R., 1991, Microbiol. Rev., 55, pp. 259–287.
Lundgren, D.G., Boucheron, J.A., and Mahony, W.B., 1983, In Recent Progress in
Biohydrometallurgy, G. Rossi and A.E. Torma, eds., Associazione Mineraria Sarda, Iglesias,
pp. 55–69.
Malinovskaya, I.M., Kosenko, L.V., Votselko, S.K., and Podgorskii, V.S., 1990, Microbiologiya, 59,
p. 70.
Maremäe, E.J., 1988, Oil Shale, 5, pp. 407–417.
Mehta, A.P., Torma, A.E., and Murr, L.E., 1979, Biotechnol. Bioeng., 21, pp. 875–885.
Melnikova, E.O., Avakyan, Z.A., Karavaiko, G.I., and Krutzko, V.S., 1990, Microbiologiya, 59,
pp. 63–69.
Mohanty, B.K., and Mishra, A.K., 1991, In National Symposium on Applications of Geomicrobiology
in India, Dhule, March 25–27, pp. 117–125.
Nakamura, L.K., and Swezey, J., 1983, Int. J. Syst. Bacteriol., 33, pp. 46–52.
Ogurtsova, L.V., Avakyan, Z.A., and Karavaiko, G.I., 1991, Microbiologiya, 60, pp. 823–827.
Ottow, J.C.G., 1969, Zbl. Bakt., 6, pp. 600–615.
Pichinoty, F., Mandel, M., and Asselineau, J., 1983, Ann. Microbiol., Inst. Pasteur, 134 B,
pp. 353–356.
Piga, L., Pochetti, F., and Stoppa, L., 1993, J. Miner. Met. and Mater. Soc., 45(11), pp. 54–59.
Polkin, S.I., Adamov, E.V., and Panin, V.V., 1982, Technology of Bacterial Leaching of Non-Ferrous
and Rare Metals, Nedra, Moscow.
Rogers, R.D., and Wolfram, J.H., 1991, In Mineral Bioprocessing, R.W. Smith and M. Misra, eds.,
The Minerals, Metals, and Materials Society, Warrendale, PA, pp. 219–232.
Rossi, G., 1993, Fuel, 72, pp. 1581.
Silverman, M.P., and Munoz, E., 1971, Appl. Microbiol., 22, pp. 923–924.
Tarasova, I.I., Khavski, N.N., Kharullina, R.T., Karavaiko, G.I,. and Dudeney, A.W.L., 1995, In
Biohydrometallurgical Processing, Vol. I, T. Vargas, C.A. Jerez, J.V. Wiertz, and H. Toledo, eds.,
Universidad de Chile, Santiago, pp. 379–384,
Tasa, A., 1998, “Biological Leaching of Shales: Black Shale and Oil Shale,” Ph.D. Thesis, University
of Tartu, Tartu.
Tasa, A., and Lindström, E.B., 1996, Oil Shale, 13, pp. 133–143.
Tesic, Z., 1972, Rev. Ecol. Biol. Sol, 9, pp. 589–593.
Toro, L., Veglio, F., Terreri, M., Ercole, C., and Lepidi, A., 1993, FEMS Microbiol. Rev., 11,
pp. 103–108.
53
Tungkaviveshkul, T., Thizavetyan, P., and Tanticharaen, M., 1995, In Biohydrometallurgical
Tzeferis, P.G., Agatzini, S., and Nerantzis, E.T., 1994, Lett. Appl. Microbiol., 18, pp. 209–213.
Vachon, P., Tyagl, R.D., Auclair, J., and Wilkinson, K.J., 1994, Environ. Sci. Technol., 28, pp. 26–30.
Veglio, F., Beolchini, F., Ubaldini, S., Abruzzese, C., and Toro, L., 1995, In Biohydrometallurgical
Vlasov, A., Veinberg, S.N., and Skripnik, V.P., 1980, Steklo i Keramica, 8, pp. 14–16.
Vrvic, M.M., Dragutinovic, V., Vucetic, J., and Vitorovic, D., 1988, In Biohydrometallurgy, P.R.
Norris and D.P. Kelly, eds., Antony Rowe Ltd., Chippenham, UK, pp. 545–547.
Yakhontova, L.K., Andreev, P.I., Ivanova, M.J., and Nesterovich, L.G., 1987, Comp. rend. Acadamy
Sci. USSR, 296(1), pp. 203–206.
Yakhontova, L.K., Groudev, A.P., Krinari, G.A., Morozov, V.P., and Sidyakina, G., 1991, Comp.
rend Acad. Sci. USSR, 320(6), pp. 1459–1462.
54
Role of Corundum-Adapted
Strains of Bacillus polymyxa
in the Separation of Hematite
and Alumina
Namita Deo* and K.A. Natarajan*
ABSTRACT
Strains of Bacillus polymyxa, preadapted and grown in the presence of corundum, were found to be
capable of the efficient separation of hematite from alumina. Results of tests performed using binary
hematite-corundum and ternary hematite-quartz-corundum mixtures in the presence of cells and met-
abolic products separated from the adapted bacterial culture indicated that more than 99% of the
hematite could be efficiently separated through selective flocculation after desliming. It was found that
alumina-specific bioproteins and other nonproteinaceous compounds were secreted by bacterial cells
after adaptation to the mineral. The utility of this bioprocessing is demonstrated in the removal of
iron from bauxite ores through selective flocculation in the presence of the adapted bacteria.
INTRODUCTION
The influence of cells and metabolite (spent growth medium containing extra cellular met-
abolic products such as polysaccharides, proteins, enzymes and organic acids) of Bacillus
polymyxa (a neutrophilic, heterotrophic bacteria associated with mineral deposits) on the
settling and flotation behavior of different minerals was recently reported (Namita Deo
and Natarajan, 1997). It was established that quartz (silica) could be efficiently separated
from hematite, corundum, and calcite through biotreatment. Both selective flocculation
and flotation can be used to achieve the above separation after the appropriate pretreat-
ment with either bacterial cells or their metabolites.
It has also been shown that the bacterial proteins present in the metabolite are mainly
responsible for specific adsorption onto quartz and kaolinite, bringing about their selective
dispersion or flotation when present along with hematite, corundum, or calcite. However,
satisfactory separation of hematite and corundum from their mixtures could not be achieved
through such biotreatment procedures. Separation of alumina from iron oxides such as
hematite and magnetite is known to be very difficult due to similarities in their surface chem-
ical, crystallographic, and chemical properties.
* Department of Metallurgy, Indian Institute of Science, Bangalore, India.
55
Unadapted cells of Bacillus polymyxa grown in the presence of Bromfield medium exhibit
similar surface affinities toward both hematite and corundum, as shown by zeta potential
and FTIR spectroscopy studies (Namita Deo and Natarajan, 1997). Similar surface chemical
changes were brought about both on hematite and corundum due to prior bacterial interac-
tion and, as a result, both minerals behaved similarly under flocculation or flotation condi-
tions. The success of any biological process developed to separate hematite and corundum
(alumina) would then depend on whether one could achieve exclusive biospecificity for
either of the two minerals.
If a biological protocol could be developed to generate either hematite- or alumina-
specific biosurfactants, their separation could be achieved. Similarly, one could also attempt
to train the organism to adapt to a specific mineral (hematite or alumina, as in this case) so
that such specially adapted strains could exclusively interact (adhere) to the desired mineral
phase in a mixture and bring out selective separation through conferment of diverse surface
chemical changes. It is well established in bio-hydrometallurgy that adapted strains of
Thiobacillus ferrooxidans (Srihari et al., 1991) are more efficient than wild unadapted strains
in the bioleaching of sulfide minerals. The role of bacterial adaptation in enhancing mineral
adhesion and dissolution is well documented in the bio-hydrometallurgy literature (Preston
Devasia et al., 1993).
To investigate the role of mineral adapted cells of Bacillus polymyxa in mineral-beneficiation
processes, attempts were made in this study to develop specially adapted strains through serial
subculturing over prolonged periods of time in contact with different mineral samples. The
beneficial role of corundum- (alumina-) adapted cells of Bacillus polymyxa in the separation of
hematite from alumina is examined in this paper.
Bacterial Strain, Growth, and Adaptation Conditions

A pure strain of Bacillus polymyxa NCIM 2539 (National Collection of Industrial Microorgan-
isms, National Chemical Laboratory, Pune, India) was used in all the studies. A 10% (v/v) of
an active inoculum was added to Bromfield medium (Phalguni et al., 1996) and incubated at
30°C on a rotary shaker at 240 rpm. The bacterial growth pattern was studied by microscopic
counting using a Petroff-Hausser counter under a phase-contrast microscope and, also, by
colony counting after plating.
The pH changes during bacterial growth were monitored and enough cell mass was gener-
ated by continuous growth for 8 hours. The culture was filtered through Whatman No.1 filter
paper and centrifuged at 15,000 rpm for 15 minutes. The cell pellet was washed several
times and resuspended in distilled water. The purity of the bacterial strain was ensured. A
special medium containing 1% peptone, 1% starch, 0.005% neutral red dye, and 2% agar
was used for enrichment of the bacterial cultures and purification of the strain after interac-
tion with minerals.
TABLE 1 Composition of the control medium

Component Units Analysis
KH2PO4 g/L 0.5
MgSO4·7H2O g/L 0.20
(NH4)2SO4 g/L 1.00
Yeast extract g/L 0.15
Sucrose g/L 5.0
pH — 7.0
56
The above strain of Bacillus polymyxa was adapted to corundum (–38 µm) in Bromfield
medium at 5% pulp density. Before using for adaptation, the minerals were steam sterilized
to kill the endogenous organisms, if any. The subculturing was done every 24 hours for
approximately six months. Adaptation to the mineral was considered achieved when the
growth rate of the adapted strain was identical to that of the control. The composition of the
control medium is shown in Table 1.
The adaptation period extended up to six months. Sufficient cell biomass (6 × 109 cells/mL)
was generated by repeated growth in Bromfield medium in the presence of the mineral. The
culture was filtered through Whatman No. 42 filter paper and centrifuged at 15,000 rpm for
15 minutes. The pellet was washed several times with 10–3 M KNO3 solution and resuspended
in 10–3 M KNO3 solution. The cells harvested from the exponential growth phase were used in
all the studies. The wet weight of the biomass generated due to bacterial growth in the pres-
ence of the mineral was also determined as well as the cell number.
Mineral and Ore Samples

The handpicked pure mineral samples of hematite and corundum that were used in this
study were obtained from Alminrock Indscer Fabriks, Bangalore, India. The purities of the
minerals were checked by chemical analysis. The hematite, corundum and quartz (silica)
samples were found to have purities of 95%, 97%, and 98%, respectively. The samples were
subjected to dry grinding in a porcelain mill. The ground sample was then dry screened, and
the –38-µm size fraction was used in the adsorption, flocculation and electrokinetic studies.
In some tests, a bauxite ore sample that was composed of 85% Al2O3, 11% hematite and
4% silica was used. The –38-µm size fraction of the ground ore was used in selective floccula-
tion studies.
Electrokinetic Measurements
Electrokinetic measurements were made using a Zeta-meter 3.0. Zeta-potential changes as a
function of pH were determined for cells adapted to corundum and compared with those
obtained using unadapted cells.
Proteinase-K Treatment
The adapted strains (109 cells/mL) were suspended in 1 mL of Tris buffer (hydroxymethyl
aminomethane hydrochloride) of pH 7.0 and 25 µg of Proteinase-K (Sigma) per 100 µg of
protein was added. The suspension was incubated at 30°C for 1 hour. The cells were pelleted
in a microfuge. The pellet was washed three times with 10–3 M KNO3 and resuspended. This
was followed by adjustment of the solution to the required pH.
Protein and Polysaccharide Production by Adapted and Unadapted Strains

A 10% (v/v) of active adapted strain was added to Bromfield medium containing 5% corun-
dum and incubated at 30°C on a rotary shaker at 240 rpm. The protein and polysaccharide
production was monitored from cell-free supernatant at regular intervals of time by the
Bradford method (Bradford, 1976) and the phenol sulfuric acid method (Dubois et al., 1956).
Flocculation Studies Using Adapted Cells

To establish the effect of corundum-adapted cells on the settling rate of hematite and corun-
dum, flocculation studies were carried out in the presence of the above-adapted strains at
different pH and periods of time. The settling rates of the above two minerals in the presence
of the metabolite derived from the adapted strains were also studied.
57
For selective flocculation studies, 5 g of an artificial mixture of either 1:1 hematite and
corundum or 1:1:1 corundum, hematite, and quartz were used to determine the separation
efficiency in the presence of the above adapted strain and metabolites.
Selective Flocculation of Hematite-Corundum Mixture by Corundum-Adapted Cells from

Different Phases of Growth
To 100 mL of cell suspension (109 cells/mL, harvested from different phases of growth such
as lag, log, late log, and stationary phases), 5 g of the 1:1 mineral mixture was added, and
the pH was adjusted to 7.0. This was done to establish whether cells harvested from different
phases of growth would have varying influences on mineral separation. Then the cylinder
was inverted five times and set down. After 1 minute, 90 mL of the supernatant was decanted
out, and another 90 mL of water, which was preadjusted to the same pH value, was added to
the remaining solids. This desliming process was performed six times. The sink and float frac-
tions were analyzed each time.
Similar experiments were also carried out in the presence of bacterial metabolites gener-
ated from different phases of bacterial growth.
Iron, aluminum, and silica contents were analyzed using Atomic Absorption Spectroscopy,
as well as ICP.
Separation of Hematite from Corundum by Different Concentrations of Corundum-Adapted

Protein and Polysaccharide and with Protein After Proteinase-K Treatment
One hundred milliliters of corundum-adapted metabolite (from the stationary phase) was
precipitated by 65% (NH4)2SO4. The precipitate was collected by centrifugation, and the
supernatant was desalted and used for polysaccharide experiments. The protein precipitate
was dissolved in 20 mL of Tris-hydroxide buffer of pH 7.0. It was dialyzed against the same
buffer over 18 hours. The precipitate, which formed during dialysis, was removed by centrif-
ugation and discarded. Different concentrations of protein and polysaccharide were used for
selective flocculation studies to compare their efficiencies.
For selective flocculation studies, 0.5 g of the artificial 1:1 mixture of hematite and
corundum were placed in a 10 mL graduated cylinder and diluted to 10 mL by different con-
centrations of protein (4, 2, 1, 0.5, and 0.25 mL) and polysaccharide (10, 5, 2.5, 1.25, and
0.65 mL). The pH was adjusted to 7.0, and the cylinder was then inverted five times and set
down. After 1 minute, 9 mL of supernatant was decanted and another 9 mL of water, which
was preadjusted to the same pH value, was added to the remaining solids. This desliming
process was performed six times. The sink and float products were analyzed each time for
iron and aluminum.
The same series of experiment as above was also carried out with the protein solution after
interaction with Proteinase-K and with adapted metabolite after boiling for 15 minutes.
Deadaptation
The adapted strain was deadapted by repeated subculturing in Bromfield medium in the
absence of minerals. After each subculturing, the efficiency of the prior adapted strain was
ascertained by selective flocculation of hematite-corundum. The adapted strain was consid-
ered deadapted when it lost its ability to separate hematite and corundum from its mixture
or the growth rate of adapted strain in absence of minerals became equivalent to that of the
wild strain.
Protein and Polysaccharide Separation from Bacterial Metabolite and Characterization

Batch culture that had reached stationary phase was centrifuged (15,000 rpm for 20 minutes)
to separate cells from extracellular material, followed by filtration of the supernatant through
sterile 0.2-µm pore size filters. The protein in the cell-free supernatant fluid of the culture
58
FIGURE 1 Measured zeta potentials as a function of pH for unadapted and corundum-
adapted cells of Bacillus polymyxa
liquor was precipitated by the slow addition of ammonium sulfate to a saturation level of
65%. After 2 hours, the precipitate was collected by centrifugation, and the supernatant was
desalted and used for polysaccharide experiments after lyophilization.
The protein precipitate was dissolved in a minimal volume of 0.02 M Tris-hydrochloride
buffer at a pH 7.0. It was dialyzed against the same buffer over 18 hours. The precipitate that
formed during dialysis was removed by centrifugation and discarded. The supernatant fluid
was lyophilized and used for experiments.
Cytoplasmic Protein
Cell biomass was generated by repeated growth in Bromfield medium in the presence and
absence of corundum. Bacterial cells in late exponential growth phase were filtered through
Whatman No. 42 filter paper and centrifuged at 15,000 rpm for 15 minutes. The cell pellets
were washed several times with 10–3 M KNO3 and resuspended in 5 mL of 30 mM Tris-HCl
pH 8 buffer. The cell membranes were disrupted by sonication in ice (two cycles of 40 s,
16 mHz using a Soniprep 150). Cell debris was removed by centrifugation for 25 minutes at
25,000 rpm, and the supernatant was saved in ice.
Five µL of 2 X SDS-PAGE sample buffer was added to 5 µL of each of the above protein.
The sample was boiled for 5 minutes, and then centrifuged in a microcentrifuger for
1 minute. Then all the samples were loaded on a 15% SDS PAGE gel.
Gel Electrophoresis for Isolation and Characterization of Bacterial Proteins

Polyacrylamide gel electrophoresis (PAGE) in the presence of an anionic detergent, namely,
sodium dodecylsulfate (SDS), by which proteins can be characterized in terms of the molecu-
lar size of the constituent polypeptides, was used. In the electrophoresis apparatus, the sup-
porting medium that was impregnated with buffer solution was suspended between the two
buffer components. Electrophoresis was carried out in vertical position that allows longer
separations. The buffer compartments in the tank were separated physically from the elec-
trode compartment. This prevents the buffer compartments from becoming contaminated
with the products of electrolysis formed in the electrode compartments. The electrodes were
constructed of platinum. The procedures involved are detailed elsewhere (Deutscher, 1990).
59
Scanning Electron Microscopy
Bacterial attachment to different minerals was monitored using a scanning electron microscope
(JEOL-840). The mineral specimens were gold coated in vacuum before SEM examination.

Variation of measured zeta potentials as a function of pH for cells of Bacillus polymyxa before
and after adaptation to corundum is illustrated in Figure 1. The isoelectric point (IEP) of
unadapted cells was found to be at a pH of about 1.5 to 2, (estimated by extrapolation of curve
at zero zeta potential), whereas, after adaptation to corundum, the IEP shifted to a pH of about
3.7. The surface chemical behavior of unadapted bacterial cells is thus found to be significantly
different from that of corundum-adapted cells. Significant shifts in IEP and the measured zeta
potentials of the adapted cells in the positive direction are indicative of the presence of an
excess of positively charged functional groups (NH3+) on cell surfaces. A new surface compo-
nent is secreted by the bacterial cells during growth in the presence of corundum.
The settling rate of corundum was found to be significantly enhanced in the presence of
adapted cells. For example, within 1 minute of cell interaction, 99% settling of the corundum
could be observed with adapted cells compared to about 90% with unadapted cells at neutral
pH. For hematite, the settling rate decreased from 97% (with unadapted cells) to about 80%
(with adapted cells) at neutral pH during this period. Similar settling tests for corundum and
hematite were also carried out using the metabolites separated from fully-grown cultures of
adapted and unadapted cells. It was observed that the metabolites derived from adapted
cells increased the settling rate of corundum, while decreased that of hematite. For example,
while 99% of the corundum could be settled in 1 minute with metabolites from adapted cells,
only 80% to 90% of settling could be obtained with metabolites from unadapted cells. Simi-
larly, the settling rate of hematite decreased from 96% to 85% in 1 minute after interaction
with metabolites from adapted cells. Higher specific affinity of corundum-adapted cells, as
well as their metabolic products toward corundum could thus be observed.
Selective separation of corundum from hematite from 1:1 mixtures through their differen-
tial settling characteristics was then examined using 6 × 109 cells/mL of corundum-adapted
cells and their metabolites. Typical results are presented in Table 2. Percent iron removal (as
hematite) from the settled samples as a function of the number of desliming stages and the
solution pH is given. Selectivity of corundum-hematite separation was found to increase with
the number of stages of desliming. The highest separations of more than 99% were achieved
at a pH 7.0 in the presence of corundum-adapted cells. Solution pH is an important parame-
ter influencing the selective separation of hematite from corundum. Similar tests carried out
using a higher number of adapted cells (6 × 1010 cells/mL) indicated that percent iron
removal decreased significantly. For example, after six deslimings, the percent iron removal
from the mixture was only about 65% at pH 7.0 (in the presence of a higher cell population)
as compared to 99.2% in the presence of 6 × 109 cells/mL.
With unadapted cells, no significant separation of hematite and corundum could be
achieved. Metabolites at about pH 6.0 separated from exponential phase of growth of corun-
dum-adapted cells were also found to function efficiently in the separation of corundum
from hematite. It was also observed that cells harvested from the exponential phase of
growth were more efficient in such separation. For example, cells harvested from the lag
phase could remove only 40.6% of iron from the settled corundum, compared to the 99.8%
removal exhibited by cells harvested from the mid log phase of growth. Cells from the late
log phase of growth weren’t quite as efficient (96% iron removal). It is interesting to note
that only about 36% of the iron could be removed using death phase cells, implying that
active cell metabolism contribute significantly in bringing about selective separation. In case
of cells from log and initial stationary phases of growth, secreted metabolic products would
be readily and increasingly available, both at the cell wall and at the interface. During the lag
phase of growth, on the other hand, cell growth is limited, and so is its metabolic activity.
60
TABLE 2 Separation of corundum-hematite through selective bioflocculation using corundum-
adapted cells of Bacillus polymyxa and their metabolite (6 × 109 cells/mL, particle size = 38 µm,
pulp density = 5%)
Total % iron removal by cells
Total % iron removal by
pH
Number of metabolite
deslimings 4.0 7.0 9.0 12.0 at pH 5–6
1 31.4 50.0 40.3 33.3 52.0
2 50.0 58.3 50.0 41.7 59.0
3 58.3 66.7 58.3 50.0 68.0
4 63.3 75.2 61.7 53.3 79.0
5 70.1 83.3 66.7 56.7 89.0
6 80.1 99.2 66.7 58.3 99.9
TABLE 3 Separation of corundum-hematite-quartz mixture through selective flocculation using

corundum-adapted cells (6 × 109 cells/mL, particle size = 38 µm, pulp density = 5%)
Total % iron removal Total % silica removal
pH pH
Number of
deslimings 4.00 7.00 9.00 12.00 4.00 7.00 9.00 12.00
1 50.6 58.1 49.7 40.1 54.0 56.0 50.0 40.5
2 59.5 66.5 58.1 50.5 60.5 65.9 60.3 55.5
3 69.5 74.8 66.7 59.6 70.6 78.9 75.5 60.5
4 75.6 83.2 75.0 62.6 85.7 89.5 80.6 75.6
5 85.5 94.3 83.2 70.5 90.1 95.6 94.6 80.5
6 95.6 99.5 85.0 89.0 95.6 99.9 99.5 90.5
TABLE 4 Separation of corundum-hematite-quartz mixture through selective flocculation using

metabolite (pH 5–6) separated from corundum-adapted cells
Number of Total % Total %
deslimings iron removal silica removal
1 59.7 60.6
2 68.1 71.0
3 75.8 85.6
4 83.3 97.0
5 95.3 99.5
6 99.9 99.9
To understand the beneficial role of corundum-adapted cells in the selective separation of

hematite and quartz with respect to corundum, further settling (flocculation) tests were car-
ried out using 1:1:1 ternary mixtures containing hematite, corundum and quartz. Typical
results are presented in Table 3 as a function of solution pH and number of desliming stages.
While silica removal was found to be efficient at practically all the pH values (4.0, 7.0, 9.0,
and 12.0), iron removal from the mixture was the highest at pH 7.0, followed by pH 4.0. At
neutral pH, more than 99% of iron and silica could be removed from the alumina-rich settled
product, after six desliming stages. As observed with the binary mixtures of hematite and
corundum, there is an optimal cell population corresponding to highest separation.
61
TABLE 5 Separation of corundum-hematite mixture through selective flocculation using corundum-
adapted bacterial protein and protein-free metabolites (pH 7.0, –38 µm, pulp density = 5%)
Total % iron removal
Number of Bacterial protein Metabolite after
deslimings separated from metabolite number of protein separation
1 55.0 30.0
2 60.0 40.0
3 70.0 50.0
4 85.0 65.0
5 98.0 70.1
6 99.0 85.2
Too high or too low of a cell population was found to be detrimental to selectivity. For
example, at 6 × 1010 cells/mL, only about 70% to 75% of iron and silica could be removed
from the ternary mixture. Similarly, at a lower cell density of 6 × 108 cells/mL, only about
60% to 65% of iron and silica removal could be achieved. Metabolites separated from the log
phase of growth of corundum-adapted cells were also effective in the separation of hematite
and quartz from corundum from their 1:1:1 ternary mixtures as illustrated in Table 4. After
six desliming stages, 99.9% of both iron and silica could be separated from the alumina-rich
flocculated product.
To investigate the specific role of any additional proteins (or polysaccharides) present in the
metabolites of corundum-adapted cells, the proteins were separated by precipitation using
ammonium sulfate. Selective flocculation tests were carried out using the proteins separated
from the corundum-adapted cell metabolites. Typical results are presented in Table 5. Ninety-
nine percent of the iron could be removed from the alumina-rich flocculated product after six
desliming stages with bacterial proteins separated from the metabolite. The above observation
establishes that corundum-adapted cells, which exhibit enhanced specific affinity toward
corundum (alumina), generate some additional specific protein compounds. Further experi-
ments were carried out at different concentrations of proteins and polysaccharides and similar
results were obtained.
It was also observed that, if the metabolite collected from adapted cells was heated
(boiled) before use, it partially loses its ability to separate alumina and hematite. The pro-
teins separated from the metabolite were treated with Proteinase-K for about an hour, and
then they were used to test its effectiveness in separating hematite from corundum. As
expected, no separation could be observed after Proteinase-K treatment, which results in the
degeneration and disintegration of the generated bio-proteins that are responsible for the
selective flocculation.
It becomes further evident that adaptation of bacterial cells to corundum results in the gen-
eration of newer and specific bioproteins, which are responsible for the selective separation of
alumina through preferential flocculation of corundum in the mixture. Even after separation
of the bioproteins from the metabolite, there were other biosurfactants present in the protein-
free metabolite that could bring out at least partial selective separation of corundum and
hematite, as shown in Table 5. The exact nature and composition of nonproteinaceous com-
pounds present in the metabolites of adapted cells could not, however, be established. Further
detailed investigations are essential for the purpose.
To confirm whether the protein that is secreted during adaptation is either an extracellular
protein or a cytoplasmic protein synthesized during adaptation and secreted into solution, gel
electrophoresis was carried out with the protein precipitated from metabolite and the cytoplas-
mic protein extracted from the corundum-adapted bacterial cells by sonication. The gel electro-
phoresis results are shown in Figure 2. From the figure, it is clear that, on adaptation, the cells
secrete some specific type of proteins that were not present in the unadapted cells. It is also
62
FIGURE 2 SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel electrophoresis) of (A)
intracellular and (B) extracellular proteins of Bacillus polymyxa. Lane 1: standard molecular
weight; Lane 2: Bromfield medium grown; Lane 3: Corundum grown. (The arrows indicate 31
and 20 Kda proteins in Case A and 31 and 29 Kda proteins in Case B.)
clear that the new protein, which is secreted during adaptation, is a cytoplasmic protein that is
synthesized during adaptation to corundum. The molecular weight of the protein, which was
present in both cytoplasm and the extracellular metabolite, was found to be 31 kDa.
Compared to hematite, corundum exhibits higher surface affinity to bacterial proteins.
For example, after 45 minutes of equilibration, adsorption density of bacterial proteins on
corundum was found to be about 2 mg/m2 compared to 0.4 mg/m2 for hematite. Besides
corundum-specific bioproteins, nonproteinaceous polysaccharide species are also generated
that exhibit very high specific surface affinity toward corundum (and not hematite). The
presence of such polysaccharide species was ascertained by FTIR spectroscopy. Bacterial
cells during growth in the presence of corundum particles were found to adhere to mineral
surfaces and colonize as revealed by scanning electron microscopy. The cells separated from
the mineral particles possess specific affinity for corundum, as evidenced by their rapid
adsorption on the mineral surface.
To establish whether the changes brought about by adaptation were permanent genetic
changes or only temporary reorderings of the cell membrane, deadaptation experiments
were carried out in the presence of Bromfield medium without corundum. Simultaneously,
flocculation experiments were also carried out with the cells and metabolite from each trans-
fer to examine the hematite-alumina separation efficiency. Typical results are shown in Table
6. After the first transfer, the cells, as well as the metabolite, were able to separate more than
99% of the iron from the alumina in 1:1 mixtures after six deslimings. As the number of
transfers in the mineral-free medium increased, the percentage of iron removal decreased
and after the seventh transfer, both cells and their metabolite lost their ability for iron-
alumina separation. This observation confirms that the specific hematite-alumina-separation
ability exhibited by the adapted cells is a temporary trait. As the bacteria grow in the absence
of the above minerals, they gradually loose their acquired tendency (property) and revert
back to their natural conditions. After the seventh transfer in the corundum-free medium,
corundum-adapted cells behaved similarly to unadapted cells. Thus, to preserve the benefi-
cial acquired properties of the corundum-adapted cells, it is essential to grow and preserve
63
TABLE 6 Separation of corundum-hematite mixture through selective flocculation using cells
deadapted through serial subculturing in mineral-free medium (after six deslimings, pH = 7.0, pulp
density = 5%)
Number of subculture in
mineral-free medium Deadapted cells Metabolite
1 99.8 99.5
2 91.7 91.7
3 66.1 70.0
4 50.0 53.3
5 33.3 32.3
6 16.7 10.7
7 0.0 0.0
TABLE 7 Iron removal by selective bioflocculation of bauxite after interaction with corundum-
adapted cells and their metabolites (pH 7 to 9 for cell interaction and about pH 5 to 6 for the
metabolite, 5 minutes interaction, 5% pulp density)
Number of
deslimings Cells only Metabolite
1 23.0 15.3
2 39.8 25.7
3 50.7 38.6
4 68.9 55.9
5 85.3 75.6
6 96.0 88.1
them continuously in the presence of the mineral (namely, alumina). Under such conditions
only, the newer biosurfactants secreted by the bacteria could be preserved with full activity
and vigor. The newer surface and metabolic components generated by bacterial cells on con-
tinuous growth in the presence of alumina particles are specific to the mineral. However,
detailed structural investigations are called for in order to understand the biochemical and
chemical nature of such biopolymers.
The utility of such bacterial cells preadapted and grown in the presence of alumina parti-
cles was further tested in the beneficiation of a bauxite ore with respect to selective removal
of iron. Iron was present mostly as hematite in the bauxite ore samples used in this study.
Typical results are illustrated in Table 7. Iron oxide particles were dispersed while the alu-
mina rich flocculated product was recovered as sink.
CONCLUSIONS
The following major conclusions were made from this study:
Bacillus polymyxa could be adapted to different mineral substrates. Corundum-
(alumina-) adapted cells and metabolic products of the bacteria were found to be
capable of bringing about effective separation of alumina from hematite.
Corundum-adapted cells of Bacillus polymyxa were found to secrete special proteina-
ceous compounds that exhibit significant surface affinity toward only corundum, bring-
ing about its effective flocculation and subsequent separation in the presence of
hematite and quartz.
64
To preserve the beneficial acquired properties of the bacteria, it is essential to grow and
preserve them continuously in the presence of the mineral, namely, alumina. Resubcul-
turing the bacteria in the Bromfield medium in the absence of alumina led to a gradual
loss of their ability for hematite-alumina separation.
The metabolic products separated from corundum-adapted bacterial cells were also
found to be effective in hematite-alumina separation.
REFERENCES
Bradford, M.M., 1976, “A rapid and sensitive method for the quantitation of microgram quantities
of protein utilizing the principle of protein-dye binding,” Anal. Biochem, 72, pp. 248–254.
Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A., and Smith, F., 1956. “Colorimetric method for
the determination of sugars and related substances,” Ana. Chem., 28, pp. 350–356.
Deutscher, M.P, 1990, Methods in Enzymology, Academic Press, New York.
Namita Deo, and Natarajan, K.A., 1997, “Surface modification and biobeneficiation of some oxide
minerals using Bacillus polymyxa,” Min. Metall. Process, August, pp. 302–309.
Namita Deo, and Natarajan, K.A., 1998, “Interaction of Bacillus polymyxa with some oxide
minerals with reference to mineral beneficiation and environmental control,” Minerals
Engineering, 10, pp. 1339–1354.
Phalguni, A., Modak, J.M., and Natarajan, K.A., 1996, “Biobeneficiation of bauxite using Bacillus
polymyxa,” Int. J. Miner. Process, 48, pp. 51–60.
Preston Devasia, Natarajan, K.A., Sathyanarayana, D.N., and Ramananda Rao, G., 1993, “Surface
chemistry of Thiobacillus ferrooxidans relevant to adhesion on minerals,” App. Environ
Microbiol., 59, pp. 4051–4055.
Srihari, R., Kumar, R., Gandhi, K.S., and Natarajan, K.A., 1991, “Role of cell attachment in
leaching of chalcopyrite mineral by Thiobacillus ferrooxidans,” App. Microbiol. Biotech., 36, pp.
278–282.
65
Role of a Heterotrophic
Paenibacillus polymyxa
Bacteria in the Bioflotation of
Some Sulfide Minerals
P.K. Sharma* and K. Hanumantha Rao*
ABSTRACT
A pure strain of Paenibacillus polymyxa and mineral-adapted strains are used to bring about surface
chemical changes on pyrite and chalcopyrite and, thus, their flotation. Paenibacillus polymyxa was
adapted by repeated subculturing in the presence of pyrite and chalcopyrite. The surface chemical
changes of bacteria due to adaptation and of minerals after their interaction with bacterial cultures
are evaluated by electrokinetic and infrared spectroscopy and are discussed with reference to their flo-
tation responses. Interaction of bacterial cells, bacterial metabolites and whole bacterial cultures
affected the Hallimond flotation behavior of the sulfide minerals. The Xanthate flotation results show
that pyrite, but not chalcopyrite, is depressed when the tests are carried out after interaction with
chalcopyrite-adapted Paenibacillus polymyxa. This investigation demonstrated that the surface
chemical properties of bacteria can be manipulated successfully to achieve the desired effects in the flo-
tation process.
INTRODUCTION
Paenibacillus polymyxa is a Gram-positive, aerobic, hetertrophic bacterium associated with
oxide mineral deposits. It secretes exopolysaccharides, several proteins, enzymes and organic
acids such as acetic, formic, and oxalic acid. The old nomenclature of this bacterium was
Bacillus polymyxa, but it was recently renamed (Ash et al., 1993). The extracellular polysaccha-
ride (ECP) aids in the biological uptake of the metal ions necessary for metabolism and growth.
The structure of the bacterial cell wall plays a significant role in bacterial adhesion to min-
eral surfaces. Peptodoglycan, accounting for 40% to 90% of the dry weight of the cell wall, is
covalently linked to other macromolecular constituents, including several types of polysac-
charides and polyphosphate polymers that are known as teichoic acids (Stanier et al. 1993).
The adhesion of bacteria to the sulfide minerals results in the modification of surface prop-
erties to that of the bacterial cell surfaces (Devasia et al., 1993; Blake et al., 1994). The bac-
terium surface properties depend on the growth conditions (Amaro et al., 1990; Jerez et al.,
* Division of Mineral Processing, Luleå University of Technology, Luleå, Sweden.
67
1992; Devasia et al., 1993). Components on the bacterial surface play an important role in
adhesion (Marshal, 1976; Irvin, 1990).
Although the use of different microorganisms in ore leaching is well established, biological
methods of mineral beneficiation are not satisfactorily understood. As distinct from bioleach-
ing, biobeneficiation by definition refers to the selective removal of undesirable mineral con-
stituents from an ore through interaction with microorganisms, thus enriching it with respect
to the desired valuable mineral. Even though there have been some studies carried out on the
use of microorganisms in beneficiation, it has been hitherto done with chemolothotrophs and
mainly Thiobacillus ferrooxidans (Dogan et al., 1985; Lyalikova and Lyubavina, 1986;
Townsley et al., 1987; Yelloji Rao et al., 1991; Mishra and Chen, 1995). Only recently, have
heterotrophs been studied with respect to oxide mineral beneficiation (Phalguni et al., 1996;
Deo and Natarajan, 1997, 1998a, 1998b).
It is essential to understand the mechanisms and resulting consequences of microbe-
mineral interaction before the utility of microorganisms in the processing of minerals could
be established. Because mineral beneficiation processes such as flotation and flocculation
depend on the surface chemical properties of the minerals, any changes in the surface chem-
istry brought about by biotreatment would be significant. The role played by Paenibacillus
polymyxa in the beneficiation of iron ore minerals was examined recently by Deo et al.
(1997, 1998a).
In this paper, an attempt is made to discuss the surface chemical changes engineered by
Paenibacillus polymyxa on sulfide minerals and their flotation responses. The manner in
which surface chemical properties of bacterial cells changed due to adaptation is elucidated
by electrokinetic and FT-IR and FT-Raman spectroscopy studies. Finally, the surface chemi-
cal changes imparted by these bacterial cells on the minerals are analyzed by microflotation
and electrokinetic studies.
METHODS AND MATERIALS
Materials and Reagents

Pure crystalline mineral samples were purchased from Gregory, Bottley & Lloyd Ltd., Lon-
don. The chemical composition of the pyrite was: Fe = 27.9%, S = 53.5% and Cu = 0.04%,
and the chemical composition of the chalcopyrite was: Cu = 29.8%, Fe = 27.9% and S =
31.9%. The samples were crushed and finely wet ground in a ceramic ball mill. The ground
material was wet sieved, and the –106+38-µm fraction was collected. The –38-µm fraction
was microsieved using an ultrasonic bath to obtain –5-µm material. The three size fractions
from each mineral were dried in an oven for two days at 50°C. They were then stored at 4°C
in plastic bags. Table 1 lists the surface areas for the mineral size fractions and the test in
which they were used.
The collector, potassium isopropyl xanthate, was from Hoechst AG, Germany. Analytical-
grade 4-methyl-2-pentanol (MIBC frother) was obtained from Merck-Schuchardt. All other
reagents were procured from KEBO Lab, Sweden.
TABLE 1 Physical properties of the mineral samples

Surface area, m2/g
Size fractions, µm Pyrite Chalcopyrite Applications
–5 2.64 4.166 electrokinetic
measurements
–38+5 0.192 0.7543 adaptation
–106+38 0.831 0.362 flotation
68
Bacterial Strain and Adaptation
A pure strain of Paenibacillus polymyxa NCIM 2539 was used in all the studies. A sucrose-rich
Bromfield medium was used for subculturing the bacteria. A 10% active cell culture was
added to the medium, and it was then incubated in a rotary shaker at 150 rpm at 30°C. The
cell number increases during growth, and the pH drops due to the generation of organic
acids. Therefore, the growth was monitored by cell count and pH measurement.
Adaptation of Paenibacillus polymyxa was performed by repeated subculturing of the bac-
teria in Bromfield medium and in the presence of an increasing pulp density of the minerals.
The subculturing was initiated with only 0.1% pulp density, and the pulp density was suc-
cessfully increased during subsequent subculturings. At each new step, the inoculum from
the previous lower pulp density was introduced into the medium containing the higher pulp
density. When the growth curve in the presence of the mineral agreed with that of the
growth without the mineral, the bacteria was considered to be adapted to that particular
mineral and pulp density.
Cell Harvesting
The cells were harvested just at the beginning of the stationary phase of its growth. Cells
grown in pyrite and chalcopyrite were filtered through Whatman filter paper (No. OOR) to
remove the cells from the suspended solid materials. The liquid containing the cells was then
centrifuged at 15,000 rpm for 10 minutes, and the cell pellet was obtained. It was then
washed twice in deionized water to remove any trapped ions. The concentrated cells were
stored at 30°C for use in the interaction with minerals and zeta-potential measurements.
Some portion of the pellet was dried at room temperature so as to record the FT-IR and FT-
Raman spectra.
Cell Counting
A calibration curve was made between the cell counting using a Petroff-Hausser counting
chamber and an enumeration counting method (serial dilution plate count) (Aneja, 1996). A
concentrated solution of bacterial cells was prepared by dispersing the centrifuged cell pellet
in deionized water. The cell densities (cells mL–1) were determined using a Petroff-Hausser
counter. A series of cell density stock solutions, ranging from 1011 to 107 cells mL–1, were
made by diluting the concentrated solution.
One milliliter from each stock solution was serially diluted by a factor of 1/10 in test tubes
containing 9 mL of saline (0.85% NaCl). This was done eight times, so that the last one com-
prised a total dilution of 10–8. One-tenth of a milliliter from each of the 10–5, 10–6, 10–7, and
10–8 dilution test tubes was spread on sterilized petri dishes with solidified nutrient agar (8%
nutrient medium, 12% agar). These petri dishes were incubated at 37°C for one day, and,
after the growth of the cell colonies, the number of cells were counted. From the dilution fac-
tor, the original cell count was obtained in the stock solutions. Finally, a straight-line calibra-
tion curve between the Petroff-Hausser count and the enumeration count was obtained on a
log-log graph.
During the usual laboratory experiments, the cells were counted using a Petroff-Hausser
counter, as this method is quite convenient. The cell numbers were reported based on plate
count using the calibration curve.
Zeta-Potential Measurements
After 60 minutes of conditioning at a specified pH value, the zeta-potential measurements of
Paenibacillus polymyxa cells and sulfide minerals were made with a Laser Zee Meter (Pen
Kem, Model 501) equipped with a video system. All the measurements were made at con-
stant ionic strength (I = 0.001) using KNO3. The zeta potentials of the bacterial cells alone
were measured with a bacterial population of about 2 × 106 cells mL–1, and, for the mineral
69
samples, a concentration of 1 g L–1 was used. In the case of mineral interacted with bacteria,
the zeta potential was measured after interacting the mineral with a sufficient amount of
cells to achieve a monolayer.
FT-IR and FT-Raman Measurements

The infrared spectra of the samples were recorded using a Bruker Fourier transform spec-
trometer with a diffuse reflectance attachment. The radiation was measured with a MCT
nitrogen-cooled detector against a nonabsorbing KBr matrix, which was used as a reference.
The samples for diffuse reflectance were prepared by dispersing 53 mg of the sample in
387 mg of KBr. The typical measurement time while recording the spectra was about 5 min-
utes (100 scans) at a resolution of 4 cm–1.
The FT-Raman spectra were recorded using a Perkin-Elmer NIR, FT-Raman 1700X spec-
trometer. The samples were excited with 100, 300, and 600 mW of unpolarized intensity-
stabilized 1064 nm radiation from a Spectron SL 301 Series Nd:YAG laser. The spectra were
registered at 4 cm–1 resolution using InGaAs detector. About 200 to 250 scans were required
to get good noise to signal ratio. Because the quantity of bacterial cell samples was not
enough to record the spectra, it was diluted with KBr and the pure KBr spectrum was later
subtracted.
Flotation Tests
The single-mineral flotation tests were carried out in a Hallimond tube using 1 g of mineral
sample. The mineral samples were first conditioned with a predetermined bacterial cell con-
centration in 100 mL of water at a specified pH for 5 minutes. Then potassium isopropyl xan-
thate collector was added and the sample was conditioned further for 5 minutes. The entire
solution was transferred into a Hallimond tube and floated for 1 minute after the introduc-
tion of MIBC frother. An airflow rate of 13 L h–1 was applied during flotation. Tests were car-
ried out at the natural pH with 1 g of mineral in 100 mL of deionized water.
The influence of initial cell concentration on the xanthate flotation of sulfides was also
examined. Some tests were performed in the absence of collector but after bacterial cell con-
ditioning of the minerals. All the tests were conducted at the natural pH of the respective
minerals.
RESULTS AND DISCUSSIONS
Characterization of Paenibacillus polymyxa Cells
Calibration Curve. Petroff-Hausser chamber counting was used for determining the cell
densities during the usual experimentation, as this method is quite fast and convenient.
Because an overestimation of the actual cells was conceivable in this method of counting, a
calibration curve was constructed between the Petroff-Hausser counting and the enumera-
tion counting. The curve was accepted to be quite accurate.
The calibration curve (Figure 1) delineates the fact that an overestimation of approxi-
mately 100 times the actual cell count is present in the Petroff-Hausser counting method. A
straight-line fit was performed on the points obtained, and the following equation was
obtained with a standard deviation of 0.13178 and a correlation factor of 0.99458.
Log(Plate Count) = –1.03357 + 0.93303 Log(Petroff-Hausser Count) (EQ 1)
All the cell densities presented in the following results were measured by Petroff-Hausser
counting, but they were later converted to plate count (real count) using the above equation.
70
FIGURE 1 Calibration curve between Petroff-Hausser chamber counting and enumeration (serial
dilution spread plate) counting
Adaptation of Paenibacillus polymyxa. The pure culture of P. polymyxa was adapted to

pyrite and chalcopyrite by repeated subculturing. The subculturing was started with a 0.1%
(by weight) pulp density of the –38+5-µm mineral, and slowly the pulp density was
increased. Adaptation with pyrite is achieved until a pulp density of 1.8% and a growth rate
the same as unadapted P. polymyxa are attained, i.e., the stationary phase starts after about
8 hours. While, for chalcopyrite, the adaptation is achieved until a pulp density of 0.5% is
attained, but the stationary phase starts after 20 hours, i.e., the chalcopyrite adapted cells
are harvested after 20 hours of growth.
FT-IR and FT-Raman Studies. The diffuse reflectance FT-IR spectra of dried Paenibacillus
polymyxa cells grown in different conditions are shown in Figure 2. All three spectra are sim-
ilar, apart from the fact that the mineral-adapted P. polymyxa have more intense peaks then
the unadapted P. polymyxa.
The peak assignments are listed in Table 2. The characteristic peaks at 2957, 2923, 2852,
1455, 1378, and 1332 cm–1 are due to the alkyl groups present on the cell surface. While the
peaks at 3284, 3071, 1656, 1637, and 1544 cm–1 are assigned to the amide A,B,I (α helical
protein), I (β helical protein), and II vibrations, respectively. Therefore, the presence/promi-
nence of these peaks indicates the presence/prominence of proteins on the cell surfaces. The
peaks at 1235 and 1071 cm–1 are related to the polysaccharides present on the cell surface. It
is known that some pathogenic Bacillus species possess poly-glutamic acid capsules, which
are essential for virulence, and the FT-IR spectra from these contain bands near 1605 cm–1
and near 1407 cm–1, which are associated with a strong amide I band at 1653 cm–1, indicat-
ing the presence of α helical structure. Some nonpathogenic Bacilli are also known to possess
this polypeptide capsule (Naumann et al., 1996).
The chalcopyrite-adapted P. polymyxa contains higher amounts of proteins on its surface,
because the bands at 1656 and 1544 cm–1 are more intense than unadapted P. polymyxa. The
bands at 1235 and 1071 cm–1 are weaker when compared to the ones on unadapted cells, indi-
cating that the amount of polysaccharides are less on the chalcopyrite-adapted P. polymyxa.
The spectra of pyrite-adapted P. polymyxa shows stronger bands at 1656, 1544, 1235,
and 1071 cm–1, which indicates that pyrite-adapted cells have higher amounts of secretions
on its surface. These cell surfaces contain higher protein and higher polysaccharide contents
71
FIGURE 2 The diffuse reflectance FT-IR spectra of (1) pyrite-adapted, (2) chalcopyrite adapted,
and (3) unadapted Paenibacillus polymyxa cells
TABLE 2 Peak assignments for the FT-IR and FT-Raman spectra

Wave number, cm–1
FT-IR FT-Raman Assignments
3284 3277 NH stretching of proteins (Amide A)
3071 3065 Amide B vibration
2957 — (CH3) stretching
2923 2916 (CH2) stretching in fatty acids
— 2878 (CH3) stretching of methyl
2852 — (CH2) stretching in fatty acids
1656 1665 Amide I of protein of α-helical structure
1637 — Amide I of protein of β-pleated sheet structure
1544 — Amide II of protein
1455 1454 Asymmetric –CH deformation of C–CH3 group
1406 — Symmetric stretching of carboxylate group
1378 — Asymmetric –CH deformation (bending) for C–CH3 group
1331 1337 –CH deformation (bending) for –CH-group
1235 — (PO2–) stretching of phosphodiesters
1071 — C–O–C, C–O, C–O–P, and P–O–P vibrations of polysaccharides
Naumann et al., 1996, Bellamy, 1975, Twardowski and Anzenbacher, 1994.
compared to chalcopyrite-adapted cells and unadapted cells. The shoulder at 1637 cm–1
along with the band at 1656 cm–1 show that the cell surfaces have proteins with both α heli-
cal structure and β pleated sheet structure.
The band at 1406 cm–1 and a stronger band at 1656 cm–1 indicate the presence of a
polypeptide capsule, the function of which is not fully known, but it could play a role in the
formation of biofilms (Naumann et al., 1996). Although more proteins are present on the
72
FIGURE 3 FT-Raman spectra of (1) pyrite-adapted, (2) chalcopyrite-adapted, and
(3) unadapted Paenibacillus polymyxa cells
pyrite-adapted P. polymyxa, because of the presence of higher amounts of polysaccharides,

the cells are not hydrophobic.
The FT-Raman spectra shown in Figure 3 exhibit new bands at 2878 cm–1. These bands
represent CH3 stretching. The other bands at 3277, 3065, 2916, 1665, 1454, 1337 cm–1 are
the same as the ones present in FT-IR spectra. The spectra recorded at 100 and 300 mW
energies show some differences (Figure 3). At 300 mW, the band at 2916 cm–1 for unadapted
cells is much more pronounced then it is at 100 mW. An effort was made to also record the
spectra at 600 mW, but it was only possible for unadapted cells, as the other two samples
were burnt.
In summary, the FT-IR and FT-Raman spectra showed the presence of CH, CH2, CH3, NH,
NH2, NH3, COOH, CONH, C-O-C, C-O, and C-O-P groups on the surface of all the grown cells.
As revealed from the spectra, the intensity of all the absorbance peaks is higher for adapted
cells than for unadapted cells.
73
unadapted cells
pyrite-adapted cells
chalcopyrite-adapted cells
FIGURE 4 Zeta potentials of Paenibacillus polymyxa cells
The groups obtained are the ones present in any protein or polysaccharide molecule. Both
pyrite- and chalcopyrite-adapted cells have higher amounts of protein on their surfaces than
do the unadapted cells. Pyrite-adapted cells had higher amounts of proteins on their surfaces
than did chalcopyrite-adapted cells. The amount of polysaccharide was found to be higher on
pyrite-adapted cells than on chalcopyrite-adapted cells.
Hydrophobicity depends on the relative presence of proteins and polysaccharides on the
cell surface. However, even though the pyrite-adapted cells have more proteins on their sur-
faces, because of the higher amounts of polysaccharides, they are not as hydrophobic as the
chalcopyrite-adapted cells.
Zeta-Potential Studies. The surface of a bacterial cell is charged due to the presence of func-
tional groups such as carboxyl (–COOH), amino (–NH2) and hydroxyl (–OH) originating from the
cell wall components of lipopolysaccharides, lipoprotein, and bacterial surface proteins. The zeta-
potentials of unadapted and pyrite-adapted P. polymyxa are negative for the entire pH range, and
they exhibit an iso-electric point (iep) of about pH 2. However, the chalcopyrite-adapted
P. polymyxa cells have an iep of pH ∼4, above which they are negatively charged.
The above results indicate that the surface chemical properties of unadapted, pyrite-
adapted and chalcopyrite-adapted P. polymyxa are significantly different, and the net surface
charge of the bacterial cells is dependent on their growth history. The iep of a bacterium
reflects a balance between the anionic and cationic acid-base groups on the bacterial cell sur-
face and specific ion adsorption. An iep between pH 2.0 and 2.8 results from a cell surface
predominated by polysaccharide-associated carboxyl groups (Rijnaars et al., 1995). Bacteria
with an iep greater than or equal to pH 3.2 have cell walls prevalent with protein molecules
rather than polysaccharides (Nikaido et al., 1985; Deflaun et al., 1990; Bendinger et al.,
1993; Rijnaars et al., 1995).
The zeta-potential of chalcopyrite-grown P. polymyxa illustrates that the cell surfaces con-
tain higher protein content, which is in direct accordance with the FT-IR results. In the case
of pyrite-grown P. polymyxa, although the protein content on the cell surface is higher and
because polysaccharides are also present in higher quantities (Figure 2 and Table 2), the iep
is around pH 2.
74
pyrite mineral
× pyrite interacted with chalcopyrite-adapted cells
pyrite interacted with pyrite-adapted cells
pyrite interacted with unadapted cells
FIGURE 5 Zeta potentials of pyrite after interaction with different Paenibacillus polymyxa cells
pyrite mineral
× pyrite interacted with chalcopyrite-adapted cells
pyrite interacted with pyrite-adapted cells
pyrite interacted with unadapted cells
FIGURE 6 Zeta potentials of chalcopyrite after interaction with different Paenibacillus polymyxa
Mineral Interaction with Paenibacillus polymyxa
Zeta-Potential Studies. The zeta-potential behavior of pyrite and chalcopyrite before and
after interaction with different types of P. polymyxa are presented in Figures 5 and 6, respectively.
The iep of pyrite was observed to be less than pH 6.8, which is in agreement with previously
reported values (Girczys and Laskowski, 1972). However, the iep obtained for chalcopyrite was
less than pH 7.4, which is not in agreement with a previously reported value of less than pH 3
(McGlashan et al., 1969; Healy and Moignard, 1976).
75
The iep for both pyrite and chalcopyrite was shifted to less than pH 2 when they were
interacted with unadapted and pyrite-adapted P. polymyxa, and this happened because the
ieps of both of these cells are below pH 2. On the other hand, the iep of pyrite shifted from
pH ∼6.8 to ∼5.2 and the iep of chalcopyrite shifted from pH ∼7.4 to ∼3 when they were inter-
acted with chalcopyrite adapted P. polymyxa. The shift in the iep of minerals toward the iep
of the cells indicates specific adsorption of cells on the minerals. The shift on chalcopyrite
was observed to be much more pronounced than on pyrite, which can be due to higher affin-
ity of chalcopyrite-adapted cells toward chalcopyrite mineral.
The zeta-potentials of the minerals at and above pH 8 are not much influenced by the
presence of P. polymyxa cells. Because the minerals and cells are negatively charged at these
pH values, the adsorption of cells cannot be realized through electrostatic interaction, but
chemical interaction between the cell wall groups and surface metal ions is feasible. The
changes in the zeta-potentials at acidic pH values are obviously caused by the electrostatic
interaction between the oppositely charged cells and minerals.
Flotation Studies. Flotation tests with pyrite and chalcopyrite were performed after inter-
action with the bacterial cells alone, the metabolite alone and the whole culture. The natural
floatabilities of pyrite and chalcopyrite were found to be approximately 10% and 48%,
respectively. In the presence of 10–5 M potassium isopropyl xanthate, the recoveries for
pyrite and chalcopyrite were 62% and 74%, respectively. A lower collector concentration of
10–5 M was chosen, so that the effect of interaction with bacterial cells, metabolite, and
whole culture could be more pronounced.
Flotation after interaction with cells alone was performed as a function of cell density. The
flotation results after interaction with unadapted, pyrite-adapted, and chalcopyrite-adapted
P. polymyxa are presented in Figures 7, 8, and 9, respectively. The effect of the interaction of
unadapted cells was noticeable until a cell density of 107 cells mL–1, after which flotation
behavior of mineral with or without collector was the same, i.e., it increased drastically to
about 80% to 90% at a cell density of 108 cells mL–1, which was observed to happen due to
excessive frothing. The natural floatability of pyrite increased from 10% to 20% at a cell den-
sity of 106 to 107 cells mL–1, and, for chalcopyrite, it decreased from 48% to 25%. In the pres-
ence of collector, the recovery of chalcopyrite was nearly constant at 70% to 75%, but the
pyrite was depressed gradually with about 2 × 106 cells mL–1, and then it increased. There-
fore, a selective flotation of chalcopyrite from pyrite was possible at an unadapted cell den-
sity of 2 × 106 cells mL–1.
Interaction of minerals with pyrite-adapted P. polymyxa caused depression of both pyrite
and chalcopyrite, even in the presence of collector, as shown in Figure 8. The difference in
recovery of chalcopyrite and pyrite was about 15%, the highest recovery for chalcopyrite was
observed to be 50% at 4 × 106 cells mL–1. Even at higher cell densities, no excessive frothing
was observed, and, hence, the flotation recoveries were in the range of 27% to 40%. In the
absence of collector, the flotation behavior of both minerals was similar and the recoveries
were about 15%.
Flotation results after interactions with chalcopyrite-adapted cells are shown in Figure 9.
In the presence of collector, the recovery of chalcopyrite is constant at 65% up to a cell den-
sity of 3 × 106 cells mL–1, and pyrite is depressed to about 30%. Therefore, a selective flota-
tion of chalcopyrite from pyrite was possible in a large range of cell density, from 6 × 104 to
3 × 106 cells mL–1. This shows that, after interaction with chalcopyrite-adapted cells, the col-
lector adsorption on chalcopyrite is not hindered much, but the adsorption on pyrite is hin-
dered to a large extent.
Flotation tests were also performed after mineral interaction with metabolite and the
whole culture, and the results are presented in Figures 10 and 11, respectively. In general,
the flotation recoveries after interaction with metabolite are lower than the other two
cases, which means that metabolite is richer in polysaccharide content. The metabolite of
adapted P. polymyxa was more effective in depressing than was unadapted cell metabolite.
76
pyrite
pyrite + 0.01 mM xanthate
○ chalcopyrite
chalcopyrite + 0.01 mM xanthate
FIGURE 7 Flotation of minerals in the presence of unadapted Paenibacillus polymyxa cells
pyrite
○ chalcopyrite
FIGURE 8 Flotation of minerals in the presence of pyrite-adapted Paenibacillus polymyxa cells
77
pyrite
○ chalcopyrite
FIGURE 9 Flotation of minerals in the presence of chalcopyrite-adapted Paenibacillus polymyxa

cells
FIGURE 10 Flotation of mineral after interaction with metabolite from different types of
Paenibacillus polymyxa cells in the presence of 0.01-mM xanthate collector
78
FIGURE 11 Flotation of mineral after interaction with whole culture of different types of
Paenibacillus polymyxa cells in the presence of 0.01-mM xanthate collector
The flotation recoveries in the case of whole cultures is higher than metabolite interaction.
Again, the recoveries with unadapted cell culture were higher than with adapted cells. The
recoveries were at a minimum in the case of chalcopyrite-adapted cell culture.
CONCLUSIONS
The surface properties of Paenibacillus polymyxa depend on their growth history. The FT-IR
and FT-Raman studies of the cells revealed that the pyrite- and chalcopyrite-adapted cells
generate more proteins on the surface than the unadapted cells. Higher amounts of polysac-
charides are secreted in the case of pyrite adaptation than with chalcopyrite adaptation.
Accordingly, the surface potentials of the cells differed due to the protonation and deproto-
nation of the cell-wall surface groups having different amounts of protein and polysaccha-
rides on its surface. The iep of pyrite-adapted cells is less than pH 2 because of the presence
of higher polysaccharides. But, for chalcopyrite-adapted cells, the iep was pH 4. The
unadapted and chalcopyrite-adapted cells adsorption inhibited the xanthate flotation of
pyrite but not chalcopyrite.
Although the reason is not fully understood, the results showed that P. polymyxa cells
could be used successfully in the depression of pyrite during the flotation of valuable sulfide
minerals at natural pH levels.
ACKNOWLEDGMENTS
The authors are thankful to Prof. K.A. Natarajan, Indian Institute of Science, Bangalore, for his
invaluable suggestions and fruitful discussions. Financial support for the project, titled “Bio-
mineral and Bio-hydrometallurgical Processing,” was provided by the Swedish Foundation for
International Cooperation in Research and Higher Education (STINT).
79
REFERENCES
Amaro, A.M., Chamorro, D., Seeger, M., Arredendo, R., Peirano, I., and Jerez, C.A., 1991, “Effect
of external pH perturbations on in vivo protein synthesis by acidophilic bacterium Thiobacillus
ferrooxidans,” J. Bacteriol, Vol. 173, pp. 910–915.
Aneja, K.R., 1996, Experiments in Microbiology, Plant Pathology, Tissue Culture and Mushroom
Cultivation, Wishwa Prakashan, New Delhi, ISBN: 81-73280827, 35 pp.
Ash, C., Priest, F.G., and Collins, M.D., 1993, “Molecular identification of rRNA group of 3 Bacillii
using a PCR probe test,” Ant. Van. Leeuwenhoek, Vol. 64, pp. 253–260.
Bellamy, L.J., 1975, The Infrared spectra of complex molecules, Vol. I, 3rd Ed., Chapman & Hall,
London.
Bendinger, B., Rijnaarts, H.M., Altendork, K., and Zehnder, A.J.B., 1993, “Physiochemical cell
surface and adhesive properties of coryneform bacteria related to the presence and chain
length of mycolic acids,” Appl. Environ. Microbiol, Vol. 59, pp. 3973–3977.
Blake, R.C., Shute, E.A., and Howard, G.T., 1994, “Solubilisation of minerals by bacteria:
Electrophoretic mobility of Thiobacillus ferrooxidans in the presence of iron, pyrite and sulfur,”
Appl. Environ. Microbiol, Vol. 60, pp. 3349–3357.
Deflaun, M.F., Tanger, A.S., McAteer, A.L., Marshall, B., and Levy, S.B., 1990, “Development of an
adhesion assay and characterisation of an adhesion-deficient mutant of Pseudomonas
fluorescens,” Appl. Environ. Microbiol, Vol. 56, pp. 112–119.
Deo, N., and Natarajan K.A., 1997, “Interaction of Bacillus polymyxa with some oxide minerals
with reference to mineral beneficiation and environmental control,” Mineral Engg., Vol. 10,
No. 2, pp.1339–1354.
Deo, N., and Natarajan K.A., 1998a, “Biological removal of some flotation collector reagents from
aqueous solutions and mineral surfaces,” Minerals Engg., Vol. 11, No. 8, pp. 717–738.
Deo, N., and Natarajan, K.A., 1998b, “Studies on interaction of Paenibacillus polymyxa with iron
ore minerals in relation to beneficiation,” Int. J. Miner. Process, Vol. 55, pp. 41–60.
Devasia, P., Natarajan, K.A., Sathyanarayana, D.N., and Ramananda Rao, G., 1993, “Surface
chemistry of Thiobacillus ferrooxidans relevant to adhesion on mineral surfaces,” Appl. Environ.
Microbiol, Vol. 59, pp. 4051–4055.
Dogan, M.Z., Ozbayoglu, G., Hicyilmaz, C., Sarikaya M., and Ozcengiz, G., 1985, “Bacterial
leaching versus bacterial conditioning and flotation in desulfurisation of coal,” XVth Int. Miner.
Process. Cong., Cannes, pp. 304–313.
Girczys, J., and Laskowski, J., 1972, Transactions AIME, Vol. 81, pp. C118.
Healy, T.W., and Moignard, M.S., 1976, “A review of electrokinetic study of metal sulfides,”
Flotation, A.M. Gaudin Memorial Volume I, M.C. Fuerstenau, ed., American Institute of Mining,
Metallurgical and Petroleum Eng., 275 pp.
Irvin, R.T., 1990, “Hydrophobicity of proteins and bacterial fimbriae,” Microbial cell surface
hydrophobicity, R.J. Doyle and M. Rosenberg, eds., Am. Soc. Microbiology, Washington D.C.,
pp. 137–138.
Jerez, C.A., Seeger, M., and Amaro, A.M., 1992, “Phosphate starvation affects the synthesis of outer
membrane proteins in Thiobacillus ferrooxidans,” FEMS Microbiol Lett., Vol. 98, pp. 29–33.
Lyalikova, N.N., and Lyubavina, L.L., 1986, “On the possibility of using a culture of Thiobacillus
ferrooxidans to separate antimony and mercuric sulfides during flotation,” Fundamentals and
Biohydrometallurgy, R.W. Lawrence, R.M.R. Branion and H.B. Ebner, eds., Elsevier, New York,
pp. 403–406.
Marshal, K.C., 1976. Interfaces in Microbial Ecology, Harvard University Press, Cambridge,
Massachusetts, pp. 27–52.
McGlashan, D., Rovig, A., and Podonik, D., 1969, “Assessment of Interfacial reactions of
chalcopyrite,” Transactions AIME, Vol. 244, pp. 446.
Mishra, M., and Chen, S., 1995, “The effect of growth medium of Thiobacillus ferrooxidans on
pyrite and galena flotation,” Mineral Bioprocessing, D. Holmes and R.W. Smith, eds., Miner.
Met. Mater. Society.
80
Naumann, D., Schultz, C.P., and Helm, D., 1996, “What can infrared spectroscopy tell us about the
structure and composition of intact bacterial cells?,” Infrared spectroscopy of Biomolecules,
H.H. Mantsch and D. Chapman, eds., Wiley-Liss Inc., ISBN 0471021849, pp. 279–310.
Nikaido, H., and Varra M., 1985, “Molecular basis of bacterial outer membrane permeability,”
Microbiol Rev., 49, pp. 1–32.
Rijnaars, H.M., Norde, W., Lyklema, J., and Zehnder, A., 1995, “The isoelectric point of bacteria as
an indicator for the presence of cell surface polymers that inhibit adhesion,” Coll. Surf., B., 4,
pp. 191–197.
Stanier, Y.R., Ingraham, J.L., Wheelis, M.L., and Painter, P.R., 1993, General Microbiology, 5th
Ed., Macmillian, London.
Townsley, C.C., Atkins, A.S., and Davis, A.J., 1987, “Suppression of pyrite sulfur during flotation
by Thiobacillus ferrooxidans,” Biotech. Bioengg., Vol. 30, pp. 1–8.
Twardowski, J., and Anzenbacher, P., 1994. Raman and IR spectroscopy in biology and
biochemistry, Ellis Horwood, London and Polish Scientific Publishers, Warsaw.
Yelloji Rao, M.K., Natarajan, K.A., and Somasundaran, P., 1991, “Effect of bacterial conditioning of
sphalerite and galena with Thiobacillus ferrooxidans on their floatability,” Mineral Bioprocessing,
R.W. Smith and M. Mishra, eds., Miner. Met. Mater. Soc., pp. 105–120.
81
SECTION 2
Bioleaching
Commercialization of Bioleaching for Base-Metal Extraction 85
Microbiological Leaching of Uranium Ores 101
Advances in the Application of the BIOX® Process for Refractory

Gold Ores 121
83
Commercialization of Bioleaching
for Base-Metal Extraction
P.C. Miller,* M.K. Rhodes,* R. Winby,* A. Pinches,†
and P.J. van Staden†
ABSTRACT
This paper is a progress report on the commercialization of using bioleaching for base-metal concen-
trates. The paper focuses on bioleach processes for recovering copper from chalcopyrite and nickel/
cobalt from pentlandite/pyrrhotite. Data is discussed from pilot-plant trials in which an overall recov-
ery >95% was obtained for the bioleaching of copper from chalcopyrite. The pilot plant was operated
in closed circuit with solvent extraction and electrowinning circuits for final metal recovery. For the
bioleaching of nickel and cobalt from pentlandite/pyrrhotite, an overall recovery of 97% was
achieved. Precipitation routines were used to produce a final nickel/cobalt product. The pilot plants
were capable of treating a few kilograms of concentrate per day. Prominent features in the design of a
1-t/day (1.1-stpd) copper-bioleach demonstration plant for the treatment of chalcopyrite concentrate
are discussed. The plant, which is now under design, will be constructed and operated based on the
results obtained from the laboratory pilot-plant campaigns. Issues of scale-up for the demonstration
plant, together with integration into upstream and downstream processing, are also addressed. Inde-
pendently derived capital and operating costs are presented for a possible commercial plant. These cost
studies indicate the principle economic issues in considering the application of bioleaching to the
extraction of base metals. The benefits of bioleaching complex concentrates that are not amenable to
physical beneficiation and economic treatment by conventional smelting are highlighted. Issues, such
as the environmental aspects, that illustrate the benefits of bioleach technology are given.
INTRODUCTION
Bacterial oxidation of refractory gold concentrates is now accepted practice for the leaching of
pyrite and arsenopyrite to expose the contained gold for recovery by cyanidation (van Aswegen,
1993). All of the commercial bioleach plants, with the exception of one facility that used moder-
ate thermophiles at a temperature of 50°C, have used mixed mesophilic cultures that operate
within a temperature range of 35° to 42°C (Brierley and Brans, 1994; Miller, 1997).
The most significant current application of microorganisms in the field of base-metal
extraction is heap bioleaching of secondary copper sulfides, particularly in South America. In
* BacTech Metallurgical Solutions Ltd., Toronto, Ontario, Canada.

† Biotechnology Division, Mintek, Randburg, South Africa.
85
addition, copper is leached from low-grade chalcopyrite dumps very slowly over a number of
years. (Schnell, 1997). Numerous studies have been conducted since the 1970s on the use of
various strains of chemolithotrophic bacteria for the recovery of base metals from concen-
trates, and various authors, such as Kelly et al. (1979), have summarized this early work.
To date, no commercial facility has been operated for base-metal extraction. However,
there is now a plant under construction to treat pyrite-containing cobalt. It is accepted that
the basic technology is similar to that for treating pyrite for refractory gold concentrates
(Briggs and Millard, 1998).
There is increasing interest in the potential use of higher-temperature organisms for
bioleaching base metals. The increased interest results from reports of the advantages
offered by operating at a higher temperature, which includes the potential for an increase in
leach kinetics (Murr and Chakraborti, 1980). The technology for the only bioleach plant for
refractory gold pretreatment that used moderate thermophiles was supplied by BacTech,
and, in 1994, this plant commenced operations at Youanmi in Western Australia (Brierley
and Winby, 1996). The plant was highly successful and operated continuously without inter-
ruption for more than three years, until the operation was closed due to ore depletion. Since
then, BacTech, in conjunction with Mintek, has used its knowledge to investigate the applica-
tion of higher-temperature organisms to the leaching of base-metal sulfides. A program of
development with Peñoles in Mexico is now progressing toward the construction and opera-
tion of a 1-t/day (1.1-stpd) copper bioleaching demonstration plant designed to treat 5-t/day
(5.5 stpd) of chalcopyrite concentrate.
It is important to also give recognition to the other unit operations that are required in a
bioleaching process for treating concentrate. Figure 1 is a simplified flowsheet for the treat-
ment of chalcopyrite by bioleaching. Figure 2 shows a similar flowsheet for the potential treat-
ment of a nickel/cobalt concentrate. These figures are used throughout this paper to illustrate
the integration of bioleaching with the necessary upstream and downstream operations.
OVERALL REACTIONS AND PROCESS IMPLICATIONS

The chemical reactions that occur, together with the mechanisms and associated role of
organisms in bioleaching, have been well documented and will not be discussed in detail
here (Tuovinen et al., 1991).
The fundamental process implications of such reactions relate to the overall process acid
balance, which is obtained with consideration given to the downstream neutralization of acid
produced. The reactions given below are presented in their simplest forms—in which all of
the sulfur is converted to sulfate. These reactions indicate the most common reactions
responsible for acid production and consumption in sulfide bioleaching practice. They
exclude gangue reactions, such as neutralization of acid by carbonates and the generation of
acid by precipitation reactions of solubilized ferric iron in the bioleach, both of which con-
tribute to the process acid balance.
The principle reaction for chalcopyrite leaching, which is an acid consumer, is as follows
4CuFeS2 + 17O2 + 2H2SO4 → 4CuSO4 + 2Fe2 (SO4)3 + 2H2O (EQ 1)

A very important consideration in chalcopyrite bioleaching is the downstream acid pro-
duction from electrowinning, which can provide return acid to the bioleach operation. Twice
as much acid is produced in electrowinning the copper as is consumed in the bioleaching of
chalcopyrite.
4CuSO4 + 4H2O + 2e → 4Cu + 4H2SO4 + 2O2 (EQ 2)

Pyrite is present in most chalcopyrite concentrates, and some of the pyrite is oxidized in
the process to produce acid as follows
4FeS2 + 15O2 + 2H2O → 2Fe2 (SO4)3 + 2H2 SO4 (EQ 3)
86
FIGURE 1 Simplified flowsheet of an integrated circuit for the bioleaching of chalcopyrite
concentrate
For some copper concentrates, zinc can be present as sphalerite and will leach readily as follows
ZnS + 2O2 → ZnSO4 (EQ 4)

For complex nickel sulfides, the leaching of pentlandite occurs as follows
2Ni9S8 + 33O2 + 2H2SO4 → 18NiSO4 + 2H2O (EQ 5)

In many cases, valuable cobalt is associated with the pentlandite as small quantities within
the lattice and is leached with the nickel values.
Pyrrhotite, the most common iron sulfide associated with pentlandite, reacts readily by
consuming acid as follows
4FeS + 2H2SO4 + 9O2 → 2Fe2(SO4)3 + 2H2O (EQ 6)
87
FIGURE 2 Simplified flowsheet for the bioleaching of nickel concentrate
In practice, intermediary reactions occur in which elemental sulfur is produced. Although

it is not the aim to present a discussion here, the quantity of the elemental sulfur formed is
possibly dependent upon the chemical environment, residence time, pH, redox, and the type
of bacterial culture used.
When elemental sulfur is produced, lesser quantities of sulfate require downstream neu-
tralization, which may enhance the overall economics of bioleaching. However, if the
bioleach residue contains precious metals, the formation of elemental sulfur may have nega-
tive implications for precious-metal recovery.
This is because of increased reagent consumption in downstream cyanidation. It is, there-
fore, essential that an accurate knowledge of the sulfur moiety and speciation of the final
residue be known, due to its impact on the process acid balance and relevant downstream
operations for metal recovery from bioleach residues.
88
MATERIALS AND METHODS FOR BATCH AND CONTINUOUS TEST WORK
There are a number of sequential phases to be followed for process commercialization. To
put into context the results of laboratory test work conducted to date and the materials and
methods used, these are summarized as follows:
Testing of the technology on a small laboratory batch scale.
Continuous testing on small-scale pilot plants with a throughput of 1 to 10 kg/day (2.2
to 22 lb/day) of concentrate, with semicontinuous integration of upstream and down-
stream processing.
Operation of a continuous demonstration plant treating a few tons of concentrate per
day and producing approximately 1 t/day (1.1 stpd) of final metal values. The plant
should use commercially available equipment.
Design, construction and operation of a full-scale plant based on the demonstration-
plant data.
Proving the technology on a small laboratory-batch scale was performed by the authors
laboratories on a large variety of concentrates obtained from sources worldwide. This stage
of testing has the advantages of requiring relatively small sample sizes of a few kilograms,
which are often readily available, and these exercises can be conducted at relatively low cost
within a time period of a few months.
Following this work, a number of small-scale pilot-plant campaigns have also been oper-
ated, but on a smaller range of concentrates. These require an increased sample size of a few
hundred kilograms of concentrate and continuous campaigns that can operate for up to a
year in this initial pioneering work.
This paper focuses on two particular materials that have been tested on a pilot-plant scale
over the last two years. One of these materials was a chalcopyrite concentrate and the other
was a pentlandite/pyrrhotite concentrate. The characteristics of these materials and test
methods are described below.
Copper
The main chalcopyrite concentrate sample tested was from Mt. Lyell in Tasmania, Australia.
A summary of the analysis and mineralogy of this sample, Tables 1 and 2, respectively, reveal
the copper mineralogy to be pure chalcopyrite, with no other copper minerals present. This is
an important feature of this material for testing, as the objective of the test program was to
successfully demonstrate the bioleaching of copper from chalcopyrite.
The Mt. Lyell concentrate was considered to be one of the most difficult to bioleach suc-
cessfully. The presence of any other copper minerals, such as chalcocite or covellite (which
leach quite readily), would, therefore, have interfered in the demonstration of process viabil-
ity. A bioleach process developed using Mt. Lyell material would be readily applicable to
many other copper concentrates that may contain more readily leachable copper minerals, as
well as chalcopyrite.
TABLE 1 Average analysis of concentrates

Cu, % Ni, % Fe, % Co, % S2–, %
Copper conc. 25 — 30 — 33
Nickel conc. 5.5 7.5 50 0.45 28
89
TABLE 2 Mineralogy of concentrates
Mineralogical content, %
Mineral Nickel conc. Copper conc.
Metal content 7.5 Ni 5.0 Cu
Chalcopyrite 5 74
Pentlandite 12 —
Pyrite — 11
Pyrrhotite 80 —
Sphalerite — 2
Gangue 2 7
Nonsulfide iron — 6
Nickel and Cobalt

The concentrate sample tested was generated from ore samples obtained from the Radio Hill
mine in western Australia. Tables 1 and 2 also give the mineralogical and chemical character-
istics of this material. The material contained pentlandite and pyrrhotite with some flame
pentlandite (less than 2%) occurring within the pyrrhotite. Cobalt was also present in small
quantities. A feature of the mineralogy is the intergrown nature of the pyrrhotite with the
pentlandite, which historically gave difficulty in the production of a high-grade nickel con-
centrate of smelter quality due to dilution of the pentlandite by excessive quantities of pyr-
rhotite. Therefore, a major feature of the concentrate is the high quantity of iron present,
much of which is present as pyrrhotite and readily acid labile. Taylor et al. (1997) provided
further details on the Radio Hill material.
Batch Testing Methods

Although extensive batch test work was performed on both concentrates, the methods and
results of batch test work have been described previously and will only be summarized here
(King and Budden, 1996; Miller and Winby, 1997; Fewings and Wilathgamuwa, 1998). Agi-
tated aerated reactors with capacities ranging from 1.5 to 3 L were used. In these reactors,
air is delivered underneath an impeller using a single pipe sparger. Concentrate material was
suspended in a basal salts medium (King and Budden, 1996) at a pulp density of 5% to 25%
(w/w), and pH adjustments were made, as required, with known volumes of sulfuric acid.
Inoculation was performed with either moderate thermophiles or with extreme thermo-
philes. The temperature was maintained at 48°C for moderate thermophiles and at 70°C for
extreme thermophiles. Morphological descriptions of the moderate thermophile culture used
have been described previously and are comprised of three morphological types: long bacilli,
short oval cocci, and short bacilli (Williams, 1997).
Regrinding the concentrate to a p90 of 10 µm was necessary with the Mt. Lyell copper con-
centrate, while a number of particle sizes were tested for the nickel/cobalt concentrate,
including fine grinding to the above size.
Continuous Piloting Methods

The pilot plants used for both the nickel and copper bioleaching were previously described
(Fewings and Wilathgamuwa, 1998; Rhodes et al., 1998). As mentioned above, Figure 1
gives an overview of the bioleaching flowsheet for copper, while Figure 2 gives a similar illus-
tration for nickel/cobalt bioleaching. These serve to illustrate the downstream operations
employed for metal recovery operations and the recycles employed.
90
The bacteria used in the majority of pilot-plant campaigns were moderate thermophiles
maintained at 48°C, although a campaign using extreme thermophiles at 70°C was also con-
ducted. A pilot plant consists of a series of either three or four reactors constructed of 316L
stainless steel. The reactors are generally equipped with MX04 and HA630 Metquip impellers
for agitation. Air is sparged above the tip of the impeller, and facilities for carbon dioxide
addition to the air supply are available.
The primary reactors are at least twice the capacity of the secondary reactors. This is to
promote a more stable bacterial population and reduce system perturbations due to the
effects of fresh incoming concentrate feed. The copper pilot plant had a 150-L primary reac-
tor and three subsequent 50-L staged reactors. The nickel pilot plant had a 45-L primary
reactor and two subsequent 15-L staged reactors.
The temperatures of all reactors were maintained by a combination of electrical trace
heaters and water-cooling coils, as required. Both pH and dissolved oxygen probes were
installed. The redox was measured by hand-held units. Concentrate feed for the plants was
generally supplied on a conveyor at a rate of between 2 and 5 kg/day (4.4 and 11.0 lb/day)
and was added with the associated liquid nutrient feeds, pulp dilution water, and any recy-
cles by peristaltic pumps. Concentrate preparation by grinding was conducted on a daily
semicontinuous basis to ensure freshness of feed using a Metprotech vertical stirred mill
(Liddell, 1996). Laboratory pressure filters were used for downstream solid-liquid sepa-
ration, and, although this was a batch operation, it was not considered different to an
operation in which stock tanks are installed to act as a buffer between bioleaching and
downstream processing. Iron precipitation from the liquor was performed in a 100-L reactor,
equipped with heating and temperature control.
For the copper pilot plant, the solvent-extraction rig was supplied by Simplicity Engineer-
ing, Tucson, AZ, and it operated on a continuous basis using stored liquor from iron precipi-
tation. Acid raffinate was returned to the bioleach after stripping trace organic in a carbon
column. The electrowinning cell was comprised of four lead, tin, and calcium anodes and
three stainless steel cathodes.
For the nickel pilot plant, two-stage precipitation of iron followed by final nickel-cobalt
precipitation using lime was conducted batchwise only when the plant was at steady state.
The second-stage iron precipitate contaminated by nickel and cobalt was returned to the
bioleach to minimize metal losses. The “final effluent” was recycled back to the bioleach
plant as pulp dilution water.
Samples were taken from bioleach vessels and downstream unit operations, and liquor and
solid analyses were conducted as appropriate to confirm metal extractions on both pilot plants.
BATCH TEST WORK AND CONTINUOUS PILOTING

Batch Test Work

Results of batch test work on chalcopyrite copper and nickel/cobalt-bearing concentrates are
summarized in Table 3, and these results compliment data previously published (Neale et al.,
1996; Miller and Winby, 1997; Rhodes et al., 1998). This table illustrates the different
behavior in leaching the copper concentrate in comparison to the nickel/cobalt concentrate
with respect to iron solubilization and final oxidation level for maximum metal extraction.
The copper concentrate contains pyrite, much of which remains unoxidized while the chal-
copyrite is leached preferentially. This is likely to be due to galvanic interactions between the
pyrite and chalcopyrite minerals, as noted by other workers (Holmes, 1995). The bioleach
can, therefore, be terminated prematurely when maximum copper solubilization has been
achieved. This leads to lower soluble iron levels for downstream neutralization and a more
economical process if only a partial oxidation is required.
91
TABLE 3 Summary of batch results
Metal Iron Sulfide
Sample Bacteria type extraction extraction oxidation
Cu conc. moderate thermophiles 97% Cu 49% 80%
Cu conc. thermophiles (Sulfolobus) 97% Cu — —
Low-grade Cu conc. moderate thermophiles 98% Cu 47% 39%
Ni conc. moderate thermophiles 98% Ni and Co 99% 98%
FIGURE 3 Cumulative extraction of Cu into solution for pilot plant on Mt. Lyell chalcopyrite
copper concentrate
Because bioleaching costs are strongly influenced by the amount of sulfur to be oxidized,
it is important to examine on a case-by-case basis the degree of oxidation required to maxi-
mize metal extraction. However, in many cases, for base-metal bioleaching, a complete oxi-
dation will be considered necessary. For example, if precious-metal values are associated
with the pyrite and these are to be recovered from the bioleach residue, then complete oxida-
tion of the pyrite and, therefore, a full oxidation of the concentrate is required.
For the Mt. Lyell copper concentrate, much of the refractory precious-metal concentrate
was associated with the chalcopyrite, and high precious-metal recoveries are, therefore, pos-
sible by oxidizing only the chalcopyrite portion of the concentrate.
The high metal extractions achieved in all cases provided the incentive for proceeding to
the continuous piloting stage in the laboratory program.
Continuous-Test Results for Copper

Figure 3 illustrates the profile of copper extraction against plant residence time for the Mt.
Lyell pilot-plant operation using moderate thermophiles. The profile is generated by examin-
ing the extraction obtained in each of the reactors operating in series (labeled as R1 to R4) in
which an overall plant-residence time of six days was used.
It is evident that much of the extraction occurred within the primary reactor, in which an
extraction of 80.3% copper was obtained in the reactor residence time of three days. The
duty of subsequent reactors is to increase this extraction over the remaining residence time
92
FIGURE 4 Total copper extraction into solution using extreme thermophiles
to give an overall extraction of 96.4% copper within the overall residence time of six days.
These were the final results obtained after operating the plant for a one-year period and in
final closed circuit with recycle of acidic raffinate from solvent extraction to the bioleach. In
the circuit, there was no net acid requirement for the leach circuit. Accurate mass balances of
solids and solution analysis over the final three months of operation verified this high extrac-
tion to be correct and sustainable.
Mineralogy conducted on the final bioleach residue indicated that the unrecovered copper
values were present as relatively coarse free chalcopyrite, and, although a p90 of <10-µm
feed size to the bioleach was used, there still existed a coarse tail component in the feed. In a
commercial fine-grinding system, a much closer control on the size range could be achieved,
giving the potential to further improve copper extractions by a reduction in this coarse-sized
feed component.
Figure 4 shows the final copper extraction obtained vs. the duration of the pilot-plant cam-
paign in which the plant was operated using extreme thermophiles at 70°C. The profile illus-
trates how improvement in metal extraction can occur over a period of time. It is speculated
that the improvement is due to long-term adaptation of the microbial population to the par-
ticular sulfide substrate being leached. An average of 96% copper extraction was achieved,
which is similar to the final result achieved using moderate thermophiles.
The downstream iron precipitation routines conducted on bioleach liquor in the moderate
thermophile plant resulted in a 1% copper loss, which gave a final overall plant recovery of
95.4% copper. The solvent extraction plant worked well throughout the campaign, using LIX
622 as an extractant and Shellsol 2046 as a diluent. A recognized potential phenomenon of
“crud formation” was found to be minimal. This phenomenon in the context of bioleaching is
sometimes attributed to debris such as bacterial cellular matter carried over in the bioleach
liquor, giving interference in organic and aqueous metal species transfer and subsequent
phase separation.
An analysis of the cathode copper produced continuously from the electrolyte by electro-
winning is given in Table 4. The quality obtained was just below LME “A”-grade specifica-
tion. This result suggests that the production of “A”-grade copper on a commercial scale
should be readily achievable with the use of larger cells operated continuously over a long
term and with inclusion of a scavenging circuit. To indicate the potential for precious-metal
93
TABLE 4 Cathode analysis on electrolyte from integrated pilot plant
Other metals, ppm
Cathode Cu, % Pb Ag
ML1 99.971 130 —
ML 2 99.963 210 —
SA 1 99.9891 0.3 76
SA 2 99.9919 23 5
TABLE 5 Precious metals recovery from pilot-plant copper concentrate bioleach residues
Bacteria type Cyanidation addition, kg/t Gold, g/t Silver, g/t
Moderate thermophiles, 48°C 10 93 81
Sulfolobus, 70°C 10 87 N/A
recovery from the bioleach residue, nonoptimized cyanidation work was conducted, and the
results are illustrated in Table 5. It is evident that good recoveries were obtained from resi-
dues obtained with both types of thermophilic bacteria using an unoptimized NaCN quantity
of 10 kg/t (20 lb/st). There is confidence that optimization work in downstream precious-
metal recovery would result in improvement in gold recovery together with a reduction in
reagent consumption.
More detail on test programs and other work on chalcopyrite have been provided by van
Staden et al. (1996), Neale et al. (1996) and van Staden (1998).
Continuous Test Results for Nickel and Cobalt

Figure 5 illustrates the profile of nickel and cobalt extraction against plant residence time
from the pilot-plant operation using Radio Hill concentrate with moderate thermophiles.
Nickel and cobalt extractions were found to be almost equal, as anticipated by the mineral-
ogy. Therefore, for the purposes of this paper, they are represented by a single figure. The
profile has many similar features to that discussed above for copper bioleaching. The extrac-
tion profile is generated by examining the extraction obtained in each of the reactors operat-
ing in series (labeled as R1 to R3) with an overall plant-residence time of five days. It is
evident in the nickel/cobalt-leaching process that much of the extraction occurs within the
primary reactor, in which an extraction of 79% nickel/cobalt was obtained in the equivalent
of three days. This gave a similar result to that experienced for copper. The duty of subse-
quent reactors is to increase this extraction over the two remaining days of residence time to
give an overall extraction of 98% nickel and cobalt.
This campaign was also operated over a one-year period to show the stability of the sys-
tem. A number of different concentrate types were fed to the plant to demonstrate process
flexibility to the variety of feed types anticipated over the life of the project. A discussion of
these particular results was given by Fewings and Wilathgamuwa (1998). The results
showed that nickel and cobalt extraction remained at 98% and showed that the extraction
was not sacrificed as a result of a change in the feed type to the bioleach pilot plant.
A minimum of three precipitation steps are required after the bioleach of the nickel and
cobalt to reject ferric iron in two stages. For this particular project, this is followed by precip-
itation of nickel/cobalt values as a hydroxide. A number of saleable nickel-cobalt products
could be produced if required, including a nickel/cobalt sulfide or the final metals by sequen-
tial solvent extraction and electrowinning.
Bulk volumes of synthetic solutions were prepared for the initial precipitation trials based
on a multielement analysis of the bioleach liquor and confirmation test work carried out on
94
FIGURE 5 Cumulative extraction of nickel into solution for piot plant trial on Radio Hill concentrate
TABLE 6 Grades of Ni/Co hydroxide product using nonoptimized reagent addition (Taylor et al., 1997)
Final product grade
Reagent quantity, Product mass,
Reagent type kg/t of liquor kg/t of liquor Ni, % Co, %
Mg O 3.1 8.1 39 2.3
Lime 5.2 15.2 22.4 1.3
the real bioleach liquor after initial optimization trials. A two-stage iron precipitation was
used for total iron rejection, with the bulk of the iron being rejected in the first stage by a pH
adjustment to a value approaching three using limestone. In the second stage, a further pH
adjustment was conducted to remove residual ferric species from the liquor, resulting in a
relatively small mass of iron precipitate from this stage contaminated with nickel/cobalt.
This precipitate was returned to the bioleach to minimize metal loss.
Aeration was conducted during these neutralization steps to convert the minor amounts of fer-
rous iron to ferric, resulting in an overall iron removal approaching 99%. The co-precipitation of
nickel and cobalt from solution was less than 1% (Taylor et al., 1997), and an overall recovery of
97% nickel/cobalt was used in cost analysis assuming a 1% nonrecoverable metal loss in down-
stream practices. Both the use of magnesium oxide and lime were tested in the production of a
final nickel/cobalt hydroxide.
The final grades of precipitates produced, together with reagent consumptions, are given
in Table 6 and are cited from Taylor et al., 1997. Although the use of lime produced a lower-
grade product, Taylor et al. (1997) concluded from a preliminary cost analysis that the use of
lime was preferable for this particular project. The analysis accounted for the increased
transport cost of the lower-grade product and the increased reagent costs if magnesium
oxide were used as the neutralization reagent.
DEMONSTRATION PLANT
Because a demonstration plant is a major exercise with similar implementation stages to that
of a commercial project, proceeding to a demonstration plant requires considerable financial
95
commitment. Laboratory-based pilot plants have clearly demonstrated the ability of the pro-
cess to obtain high base-metal extractions from the bioleaching of chalcopyrite or pentlandite/
pyrrhotite on a continuous basis. This provided the incentive to test the process on a demon-
stration scale, and a demonstration-scale plant is now planned for operation in 2000. More
importantly, the work demonstrated the ability to produce high-purity, saleable products and
demonstrated the potential for precious-metal recovery if required. These results were inte-
grated with a number of prefeasibily studies (discussed below) in making the decision to pro-
ceed to the demonstration-scale operation.
Chalcopyrite is considered the most refractory of the base-metal containing minerals for
bioleaching, and it is possible that demonstration plants for other base metal minerals
would not be required if the feasibility of chalcopyrite bioleaching were demonstrated on a
large scale. Therefore, a commitment was undertaken to design, construct and operate a
1 t/day (1.1 stpd) bioleach facility treating approximately 5 t/day (5.5 stpd) of chalcopy-
rite concentrate.
The plant will use commercially available equipment wherever possible, and some of the
major issues to be addressed relate to successful scale up of operations outside of the
bioleach section. The main areas of focus will be in the upstream regrinding and downstream
dewatering areas, with both of these being challenged by the fine particle size required for
the bioleach operation.
The upstream regrind operation will be concerned with critical particle-size control to pre-
vent a coarse tail component in the feed, while the downstream dewatering will be concerned
with the handling of slurries containing fine leached solids. In the laboratory pilot plant, thick-
ening and filtration had to be carried out in batch mode. Continuous thickening and filtering of
products , therefore, has to be confirmed, with the influence of these unit operations on the
water balance being viewed as one of the more important areas for investigation.
Scale-up of oxygen utilization in bioleaching will be modeled on a template of data from a
commercial bacterial-oxidation plant treating refractory gold concentrates using BacTech/
Mintek technology. Preferences for the use of either extreme thermophiles or moderate ther-
mophiles on a commercial scale will be made according to issues relating to reaction rates,
heat balances, the influence of precipitation phenomenon, and the final residue sulfur speci-
ation obtained.
Although the commercial scale up of the bioleach tanks to 500 and 1,000 m3 (17,670 and
35,340 cu ft) is now established practice in refractory gold plants, the demonstration facility
also provides an opportunity to test new reactors that are now under development. These
new reactors may have lower capital and operating costs on a commercial scale. The larger
demonstration-scale plant allows accurate data to be derived for the performance of these
new reactor designs in direct comparison to conventional agitated air-sparged reactors.
An advantage of a larger-scale demonstration plant is also its amenability to more effec-
tive instrumentation than is possible with a pilot plant on a laboratory scale.
In particular, the measurement of pH, redox and dissolved oxygen levels with an ability to
accurately assess gas utilization and the use of logic controllers to ease optimization. Solvent
extraction and electrowinning are mature technologies. However, scale-up will continue to
examine areas such as crud formation, as the program of work advances to this larger scale.
With respect to electrowinning, treatment in a commercial cell will allow LME “A”-grade cop-
per to be produced. Conditions in the electrowinning cell would be the same as in a full-scale
plant, except for having fewer electrodes.
CAPITAL AND OPERATING COST ESTIMATES
Copper
Miller and Winby (1997) detailed a cost comparison for an operation in which bioleaching of
chalcopyrite was being considered as an alternative to toll smelting. The study concluded
that the economic merits of bioleaching must be examined on a case-by-case basis. In this
96
TABLE 7 Preliminary capital costs US$ (± 30%) for integrated bioleach/SX-EW plant that
produces 30,000 t/a (33,000 stpy) copper
Copper
Cost $, millions $/annual lb Cu
Regrinding 3 0.05
Bioleaching 15 0.22
SX/EW 10 0.15
Reagent area 2 0.03
Services 5 0.07
EPCM and contingency 11 0.16
Total 47 0.68
particular study, the cost of transport was highlighted as a major component for smelting,
while the power cost for bioleaching and on-site solvent-extraction-electrowinning were
major components for the bioleach route. Thus, a project with high transport costs, poor
smelter terms (due to concentrate impurities), and low on-site power cost should favor a
bioleaching route.
For the Mt. Lyell project, Rhodes et al. (1998) cited a potential economic benefit to consid-
ering bioleaching, in that a lower-grade concentrate could be accepted as a feed to bioleach-
ing than was tolerable by toll smelting. This would give an improved copper recovery in
flotation of 3% by decreasing the concentrate grade from 26% to 20% copper.
Table 7 gives a summary of the results from an independent engineer who estimated the
capital cost for the Mt. Lyell project based on producing 30,000 t/a (33,000 stpy) of copper
metal. It is evident that a substantial component of the capital expenditure would be for the
downstream solvent-extraction and electrowinning steps to produce copper metal on-site
and that the bioleaching operation represents about a third of the projected capital require-
ments. Table 8 illustrates the operating costs predicted for the Mt. Lyell study compared to
another copper project in which a different engineer conducted a costing exercise for a pro-
posed bioleach option. It can be seen that the predicted operating costs for both projects are
in reasonable agreement, with the differences being due to the fact that the undisclosed
project has higher power and reagent-transport costs and the fact the operation would be
located in a remote region.
The operating costs also suggest that the cost of power is an important determinant in the
viability of a bioleach project using solvent-extraction-electrowinning routines. For these par-
ticular projects, it is anticipated that operating costs would have been competitive with toll
smelting if bioleaching were to be adopted. The results suggest that for a mine amortizing
the capital over ten years, the total cost may be between 44 and 53 cents/kg (20 and
24 cents/lb) of copper produced. Importantly, the costs are little affected by the presence of
impurities such as arsenic or zinc, which are expensive to deal with in conventional smelting
operations. Thus, bioleaching is particularly attractive for the treatment of “dirty” concen-
trates that attract significant penalties when treated conventionally.
Nickel and Cobalt

Table 9 indicates the preliminary capital and operating cost estimate for cobalt and nickel
production based on a plant producing 4,000 t/a (4,400 stpy) of nickel in precipitate. Taylor
et al. (1997) discussed the original costing exercise conducted by an independent engineer.
However, this study was for a combined copper/nickel/cobalt concentrate. As indicated by
Fewings and Wilathgamuwa (1998), the copper values were subsequently viewed as being
processed as a separate concentrate. Therefore, Table 9 in this discussion only reflects the
97
TABLE 8 Preliminary operating costs* US$ (± 20%) for integrated plant producing 30,000 t/a
(33,000 stpy) copper
Copper
Mt. Lyell, Other,
Cost cents/lb Cu cents/lb Cu
Reagents 3 4
Power 5 8
Labor 3 3
Other 2 2
Total 13 17
* All costs have been determined in Australian dollars and converted to US dollars at a rate of A$ 1.00 = US$ 0.60. Signifi-
cant factors affecting the Mt. Lyell costs are power at 2US-cents/kWhr and local limestone available at US$8/ton.
TABLE 9 Preliminary capital* and operating costs† US$ (± 20%) for integrated plant producing
4,000 t/a (4,400 stpy) nickel in precipitate
Capital Operating
Cost $, million $/annual lb Ni $/lb Ni
Regrinding 0.5 0.04 0.02
Bioleaching 4 0.35 0.07
Iron removal and 3 0.25 0.29
Ni precipitation
Reagent area 1 0.07 —
Services 1 0.11 0.16
EPCM and contingency 2.5 0.20 —
Total 12 1.02 0.54
* All costs were determined in Australian dollars and converted to US dollars at a rate of A$1.00 = US$ 0.60.
† Unit costs/lb Ni allows for the cobalt credit as the final product in a nickel/cobalt precipitate.
nickel/cobalt component of the costs. The unit costs are quoted per pound of nickel, allowing
for the cobalt credit, as the final product is a nickel cobalt precipitate. Table 9 shows that a
large component of the cost is for iron removal and nickel cobalt precipitation and, similar to
the copper study, the bioleach section accounts for a third of the projected capital require-
ment. The high contribution of the downstream processing to the total operating cost is due
to iron removal. This is because this concentrate had a high pyrrhotite content, which gave
rise to large quantities of readily solubilized iron in the bioleach liquor.
CONCLUSIONS AND THE FUTURE ROLE OF BIOLEACHING IN BASE-METAL

CONCENTRATE TREATMENT
Results have been presented that indicate the technical viability on a pilot-plant scale for the
continuous bioleaching of chalcopyrite. A minimum overall recovery of 95% copper was
obtained from a fully integrated circuit. The ability to use bioleaching for nickel and cobalt
extraction has also been shown on a pilot-plant scale, giving an overall recovery of 97%, using
precipitation routines to produce the final product. These results provided the incentive to
advance to a tonnage demonstration plant treating a chalcopyrite concentrate at Peñoles in
Mexico. Commissioning of the plant is planned during the first half of 2000.
98
The current costing studies indicate that bioleaching would compete effectively with smelt-
ing. It is believed that a good portfolio of applications exists for the use of bioleaching in the
immediate future. Environmental considerations are also prevalent in many projects, and this
is likely to encourage adoption of bioleach processes, as the products are either benign precipi-
tates or the final metal. Bioleaching is making effective use of natural organisms under con-
trolled conditions by accelerating the degradation process of sulfides, which occur in nature.
With the exception of a small amount of oxygen being evolved in electrowinning, there are no
gases produced and waste precipitate products can be readily returned to the environment. The
environmental advantages of the technology may also result in a reduction in the time for per-
mitting approval, thereby, decreasing the lead-time for commercialization of a project. Other
advantages for considering bioleach technology were cited by Miller (1997).
The “niche” area for the first applications of concentrate bioleaching for base metals is
likely to be with concentrates that are either untreatable by smelting or that attract severe
penalty charges. As already mentioned, these are often termed “dirty” concentrates, and, in
the context of copper, they contain impurities such as zinc, arsenic or bismuth. In addition,
copper recovery may be sacrificed in concentrate production to minimize contamination by
these elements to ensure that the concentrate meets smelter specification. By contrast, in the
bioleaching of a copper/zinc concentrate, both metals report to the leach liquor and can be
recovered separately by sequential solvent-extraction-electrowinning operations.
The use of solvent extraction and electrowinning for the recovery of zinc was reported by
Steemson et al. (1997). The zinc content of a copper solvent extraction raffinate stream can be
increased by recycle through the bioleach and an increased zinc tenor bled off for metal recovery,
thus improving the economics of zinc recovery. It is, therefore, believed that treating a combined
copper/zinc concentrate with bioleaching would have a dual benefit by eliminating penalty pay-
ments otherwise incurred by smelting and also adding value by economic zinc recovery.
The ability to use bioleaching to leach complex base-metal concentrates in which it is diffi-
cult to obtain a grade of smelter quality has been well demonstrated here in the leaching of a
combined nickel/cobalt concentrate.
A further area for potential commercialization is the treatment of copper/arsenic concen-
trates containing copper sulfides and arsenopyrite or containing copper arsenic minerals
such as enargite. Arsenic-containing concentrates are more costly to smelt because of the
need to recover and dispose of the arsenic. In bioleaching, arsenic can be precipitated as an
environmentally stable iron arsenate in a similar manner to that now used in the bacterial
oxidation of refractory arsenopyrite gold-bearing concentrates.
In conclusion, it is believed that the increasing requirement for environmentally clean pro-
cess technology will be a major incentive for adopting bioleach technology. This paper pre-
sented results from long-term pilot plant studies and independent costing exercises, which
give support to the technical and economic process viability. The objective of the larger-scale
demonstration plant will be to confirm these positive findings and to advance bioleaching of
base-metal concentrates toward commercialization.
REFERENCES
Brierley, C.L., and Brans, R., 1994, “Selection of BacTech’s thermophilic biooxidation process for
Youanmi Mine,” Biomine 94: Applications of Biotechnology to the Minerals Industry,
International Conference and Workshop, Perth, Western Australia, Sept. 19–20, Australian
Mineral Foundation Inc., Glenside, Australia.
Brierley, C.L, and Winby, R., 1996, “BacTech’s thermophilic biooxidation plant at Youanmi Mine:
An update on performance and cost,” Randol 96, Squaw Creek, CA, April 21–24, 1996. Randol
International Ltd., Golden, CO.
Briggs, A.P., and Millard, M., 1998, “Engineering aspects of a bacterial oxidation circuit for cobalt
recovery at Kasese, Uganda,” ALTA 1998, Nickel/Cobalt Pressure Leaching & Hydrometallurgy
Forum, Perth, ALTA Metallurgical Services, Melbourne, Australia.
99
Fewings, J., and Wilathgamuwa, A., 1998, “Pilot plant experience on the bacterial leaching of Ni/Co
sulfide concentrates at BacTech (Australia) Ltd.,” ALTA 1998, Nickel/Cobalt Pressure Leaching &
Hydrometallurgy Forum, Perth, ALTA Metallurgical Services, Melbourne, Australia.
Holmes, P.J., and Crundwell, F.K., 1995, “Kinetic aspects of galvanic interactions between
minerals during dissolution,” International Symposium, Principles and Practices of Leaching,
Winnipeg, Canada, October.
King, A.J., and Budden, J.R., 1996, “The bacterial oxidation of nickel and cobalt polymetallic
concentrates,” ALTA 1996, Nickel/Cobalt Pressure Leaching & Hydrometallurgy Forum, Perth,
ALTA Metallurgical Services, Melbourne, Australia.
Kelly, D.P., Norris, P.R., and Brierley, C.L., 1979, “Microbial methods for the extraction and recovery
of metals,” In: Microbial Technology, Society for Microbiology Symposium 29, A. Bull, D. Ellwood,
and C. Ratledge, eds., Society for General Microbiology Ltd., UK, pp. 263–308.
Liddell, K., 1996, “Design of ultra-fine grinding mills and their potential application in copper
hydrometallurgy.” ALTA 1996, Copper Hydrometallurgy Forum, Brisbane, ALTA Metallurgical
Services, Melbourne, Australia.
Miller, P.C., and Winby, R., 1997, “The potential commercialisation of bioleaching for the
treatment of chalcopyrite ores and concentrates,” Preprint 97-094, presented at the SME
Annual Meeting, Denver, CO, SME.
Miller, P.C., 1997, “The design and operating practice of bacterial oxidation plant using moderate
thermophiles,” In: Biomining: Theory, Microbes and Industrial Processes, D.E. Rawlings, ed.,
Springer-Verlag and Landes Bioscience, Georgetown, pp. 81–100.
Murr, L.E., and Chakraborti, N., 1980, “Kinetics of chalcopyrite-bearing waste rock with
thermophilic and mesophilic bacteria,” Hydrometallurgy 1980, 5, pp. 337–354.
Neale, J.W., Pinches, A., Kruger, P., and Van Staden, P.J., 1996, “Copper bioleaching,” ALTA 1996,
Copper Hydrometallurgy Forum, Brisbane, ALTA Metallurgical Services, Melbourne, Australia.
Rhodes, M.K., Deeplaul, V., and van Staden, P.J., 1998, “Bacterial oxidation of Mt. Lyell
concentrates,” ALTA 1998, Copper Sulphides Symposium, Brisbane, ALTA Metallurgical
Services, Melbourne, Australia.
Schnell, H.A., 1997, “Bioleaching of Copper,” In Biomining: Theory, Microbes and Industrial
Processes, D.E. Rawlings, ed., Springer-Verlag and Landes Bioscience, Georgetown, pp. 21–43.
Steemson, M.L., Wong, F.S., and Goebel, B., 1997, “The integration of zinc bioleaching with
solvent extraction for the production of zinc metal from zinc concentrates,” International
Biohydrometallurgy Symposium, Metallurgy 1, IBS97 Biomine 97, Australian Mineral
Foundation Inc., Glenside, Australia.
Taylor, A., Fairley, H., and Winby, R., 1997, “Re-opening of the Radio Hill nickel project, Western
Australia using bacterial leaching of nickel sulphide concentrates,” In: ALTA 1997, Nickel/
Cobalt Pressure Leaching and Hydrometallurgy Forum, Perth, ALTA Metallurgical Services,
Melbourne, Australia.
Tuovinen, O.H., Kelley, B.C., and Groudev, S.N., 1991, “Mixed cultures in biological leaching
processes and mineral biology,” In: Mixed Cultures in Biotechnology, J.G. Zeikus and E.A.
Johnson, eds., McGraw-Hill Publishing Co., New York, pp. 373–427
van Aswegen, P.C., 1993, “Commissioning and operation of biooxidation plants for the treatment
of refractory gold ores,” Hydrometallurgy Fundamentals, Technology and Innovations,
Proceedings of Milton E Wadsworth (IV) International Symposium Hydrometallurgy, AIME, Salt
Lake City, UT, Aug. 1–5, 1993.
van Staden, P.J., 1997, “The simulation and rough economic evaluation of copper bioleach
options,” Extraction Metallurgy ’97 Conference, June 24–26, South African Institute of Mining
and Metallurgy.
van Staden, P.J., 1998, “The Mintek/BacTech copper bioleach process,” ALTA 1998, Copper
Sulphides Symposium, Brisbane, ALTA Metallurgical Services, Melbourne, Australia.
Williams, T.L., 1997, “Factors affecting bacterial population dynamics at the Youanmi bacterial
oxidation plant,” International Biohydrometallurgy Symposium, IBS-BIOMINE97, Metallurgy 4,
Australian Mineral Foundation Inc., Glenside, Australia.
100
Microbiological Leaching
of Uranium Ores
O.H. Touvinen* and T.M. Bhatti†
ABSTRACT
Microbiological leaching has been used as an alternative approach to conventional hydrometallurgi-
cal methods of uranium extraction. In the microbiological leaching process, iron-oxidizing bacteria
oxidize pyritic phases to ferric iron and sulfuric acid, and uranium is solubilized from the ore due to
sulfuric acid attack. If uranium in the ore material is in the reduced, tetravalent form (U IV), a redox
reaction is involved whereby uranium is oxidized to the hexavalent form (U VI) upon dissolution. In
acid-leaching systems, the primary oxidant is ferric iron, which is reduced to ferrous iron by its chem-
ical reaction with UIV. The ferrous iron thus formed is reoxidized to ferric iron by iron-oxidizing bacte-
ria such as Thiobacillus ferrooxidans and Leptospirillum ferrooxidans. Nutritional requirements
and responses to environmental extremes of acidophilic iron-oxidizing bacteria are appraised. The
S-entity in Fe-sulfides is oxidized to sulfate by bacteria similar to Thiobacillus ferrooxidans and Thio-
bacillus thiooxidans. Pyrite and marcasite oxidation is a sulfuric acid forming reaction. Heap, dump,
and in situ leach techniques are feasible as bacterial leaching systems.
INTRODUCTION
This paper is a review of the microbiological leaching of uranium ores. Uranium demand has
been on the decline, and the use of microbiologically mediated uranium recovery from
low-grade ores is now almost extinct because of the worldwide availability of high-grade ura-
nium deposits. The application of microbiological leaching processes to the beneficiation of
uranium ores dates back at least to the 1950s.
It was observed that the Urgeirica uranium ore-processing plant in Portugal was not
achieving the expected yield of uranium. The discrepancy, in terms of the loss of uranium
from the stockpile, was eventually attributed to substantial leaching caused by rainwater.
Subsequently, it was found that uranium solubilization from the ore was catalyzed by acido-
philic Fe- and S-oxidizing autotrophs (Cameron, 1963). These bacteria oxidized the pyrite
present in the ore to sulfuric acid and ferric sulfate, and these products solubilized uranium
from the ore.
* Department of Microbiology, Ohio State University, Columbus, OH.

† National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan.
101
The initial work on uranium bioleaching in the early 1950s was directed toward preventing
solubilization, but it soon became apparent that this process could be applied on a commercial
scale to extract uranium from low-grade ores. During 1952 and 1953, the plant at Urgeirica
started a uranium heap-leaching process on a commercial scale. This is one of the early mile-
stones in the microbiological leaching of uranium ores (Lowson, 1975). Subsequent bioleach-
ing tests gave promising results for other uranium ore samples from different geographical
regions (Audsley and Daborn, 1963; Miller et al., 1963).
Subsequently, microbiological leaching processes were applied for commercial-scale
extraction of uranium by heap, dump, and stope leaching of mine waste rocks and worked-
out stopes in uranium mines in the Elliot Lake Area, Ontario, Canada, in the early 1960s
(MacGregor, 1966; McCready and Gould, 1990). Leach solutions containing acidic ferric
sulfate were circulated through surface heaps and underground stopes. Additionally, some
mines recovered uranium by hosing down roofs, walls, and floors of mine stopes at intervals
of several months. Supporting structures (pillars, walls, and roofs) in underground mines
cannot be safely mined via conventional processing. Other left-behind ore materials and
waste piles were also treated with leach solutions. The resulting leach solutions were acidic
(pH 2 to 3) and contained dissolved uranium at concentrations that were economically
recoverable by ion-exchange techniques (MacGregor, 1966; Wadden and Gallant, 1985;
Marchbank, 1987). Some pregnant solutions from underground-stope bioleaching circuits
contained as much as 2.5 to 10 g/L U3O8.
Some mines experimented with stope leaching using bacteria-containing solutions that were
supplemented with fertilizer N and P to provide inorganic nutrients, but any beneficial effects
of using nutrients on uranium bioleaching were not established unequivocally. In some
instances, yttrium was also recovered as a byproduct from leach solutions, which further
improved the overall process economics. Thorium (232Th) is sometimes associated with ura-
nium minerals. Solubilization of thorium occurs during uranium leaching, and this may lead to
elevated thorium levels because this actinide is usually not recovered from leach solutions.
Radium (226Ra), a decay product of 238U, is a particularly problematic aspect in mine and tail-
ings management because of its decay to gaseous radon (222Rn), which poses a health hazard.
Laboratory, pilot, and commercial-scale bioleaching methods have been evaluated in many
uranium-producing countries of the world. The potential of this biotechnology was recog-
nized as particularly amenable to low-grade uranium ores containing <0.05% U3O8
(<0.042% U). Commercial-scale bacterial leaching of uranium has been practiced with
low-grade uranium ores (0.01% to 0.05% U3O8; 0.0085 to 0.042% U) that would otherwise
have been uneconomical to process by conventional hydrometallurgical methods.
The Agnew Lake Mine (Ontario, Canada) was considered to be the first operation where
uranium bioleaching process was applied to a virgin orebody. A generalized flowsheet of the
bioleaching operation at the Agnew Lake Mine is shown in Figure 1. The mine started pro-
duction in the late 1970s and employed surface heap-leaching and underground stope-
leaching circuits, but the production was discontinued and the mine site was closed within a
few years. A fractured orebody that failed to prevent the loss of leach solution contributed to
the mine closure (Anonymous, 1979). Uranium recovery by biological leaching processes has
been quiescent in the 1980s and 1990s, the result of declining demand. The Denison Mines
in the Elliot Lake area, Ontario, practiced underground stope bioleaching in the 1980s and
early 1990s. The mine was eventually closed in the adverse economic situation.
Closure of uranium mines poses serious major environmental issues because bacterial
action continues to dissolve iron from sulfide mineral surfaces. Mine flooding will eventu-
ally result in anaerobic conditions that will prevent oxidative leaching of sulfide minerals
(Glombitzal et al., 1995), but uranium (and other actinides) can also be solubilized under
reduced conditions in the absence of molecular O2, especially if it is already in the hexa-
valent form in the exposed mineral.
In general, microbiological leaching processes of uranium have been applied to ores that
contain accessory Fe-sulfides. Microbiological leaching applications for the solubilization
of uranium from sedimentary deposits such as sandstone have not been practiced on a
102
FIGURE 1 A generalized flowsheet of surface heap and underground stope bioleaching circuits for
uranium based on the Agnew Lake Mine operation (modified from Tuovinen et al., 1981). Only
uranium was recovered during the active life of the mine. The ore material contained uranium
(mainly as uranothorite) and thorium, as well as the rare earths. Pyrite and pyrrhotite were the main
Fe-sulfides that contributed to the biogenic formation of acidic ferric sulfate as the leach liquor.
commercial scale. Amenability studies have been reported for the bacterial leaching of
uranium-containing sandstone deposits (Tomizuka and Takahara, 1972; Bosecker and
Wirth, 1980; Mahmood, 1994; Bhatti and Malik, 1997a, 1997b), but economic feasibility
remains doubtful. Because of the lack of pyrite and other Fe-sulfides in sandstone deposits,
iron and sulfur have to be supplied extraneously to thiobacilli.
RECOVERY OF URANIUM FROM ORES

The choice of lixiviant for uranium leaching from ores depends on the valence of the ura-
nium, the nature and composition of the ore matrix, the solubility of contaminants, the mill
tailings management and other environmental considerations. The choice also depends on
103
the economical feasibility. Sulfuric acid (H2SO4) is the most widely used leaching agent for
uranium extraction. In the H2SO4-based leaching process, UVI is solubilized from mineral
matrix in dilute sulfuric acid and forms soluble uranyl sulfate complexes [H4[UO2(SO4)3]
with the following redox reaction:
UO3 + 3H2SO4 → H4[UO2(SO4)3] + H2O (EQ 1)
Tetravalent uranium (e.g., UO2) is insoluble in dilute sulfuric acid and must be oxidized to
uranyl ion for the dissolution to occur. Molecular oxygen (O2) is a relatively slow and ineffi-
cient oxidant for UIV in sulfuric acid leaching (Merritt, 1971; Haque and Ritcey, 1982). The
oxidation of uraninite is relatively slow in sulfuric acid, and ferric iron accelerates the oxida-
tive dissolution of UIV. Ferric iron (Fe3+) must by regenerated to maintain oxidizing condi-
tions during the leaching. Sodium chlorate (NaClO3) and pyrolusite (MnO2) have been used
as Fe2+-oxidizing agents in leach solution (Haque and Ritcey, l982).
Ferric iron is an oxidant via an electrochemical surface reaction, and a concentration of
about 1 to 2 g/L Fe3+ is usually adequate for the oxidation-reduction reaction (Mattus and
Torma, 1980).
Nitric acid (HNO3) and Caro’s acid (H2SO5) have also been used as oxidants for UIV leach-
ing processes (Haque and Ritcey, 1982; Chutinara et al., 1984). The oxidation of UIV occurs
at >400 mV redox potential values. Relevant leaching reactions can be summarized by
UO2 + 2Fe3+ → UO22+ + 2Fe2+ (EQ 2)
2Fe2+ + MnO2 + 4H+ → 2Fe3+ + Mn2+ + 2H2O (EQ 3)
6Fe2+ + ClO3– + 6H+ → 6Fe3+ + Cl– + 3H2O (EQ 4)

Hydrogen peroxide (H2O2) has also been used as a chemical oxidant for solubilization UIV
(Lawes, 1978; Eligwe and Torma, 1986). H2O2 forms an insoluble uranium peroxide com-
plex [UO4(H2O)2] as follows
UO2 + H2O2 + 2H+ → UO22+ + 2H2O (EQ 5)
UO22+ + H2O2 + 2H2O → UO4(H2O)2 + 2H+ (EQ 6)

A generalized flowchart of uranium recovery from pregnant leach liquors is presented in
Figure 2. The extraction has multiple steps that lead to the production of yellowcake. Ura-
nium can be extracted from sulfuric acid or bacterial leach solutions with ion-exchange resins
or with cationic or anionic organic solvents.
Both anionic and cationic complexes of uranium, such as UO22+, UO2(SO4)22–, and
UO2(SO4)34–, are present in the leach solutions. The anionic uranyl species UO2(SO4)34– is
readily adsorbed by some ion-exchange resins as a sulfate or bisulfate complex. The
ion-exchange reaction with resins such as Amberlite-IRA-400 is selective for uranium, and
the uranium loading capacity is pH-dependent. The resin has a higher affinity for
[UO2(SO4)4]2– than for [Fe3(SO4)2]2–. The resin is suitable for uranium recovery from acid-
leach liquors with high ratios of Fe3+/UO22+. Amberlite XE-270, a weak base anion
exchanger, adsorbs a low amount of Fe in proportion to uranium as compared to Amberlite
IRA-400 that has a loading ratio of about one Fe to five U3O8 (Gow et al., 1971). Other
ion-exchange resins for uranium extraction are Amberlite IRA-405 and IRA-427; Dowex 1,
11 and 21K; Duolite A-101D and A101D-U; Ionac A-580 and A-590; Nalcite; and Permutit SK
and SKB. Uranium speciation in the leach solution and in the ion-exchange reaction can be
summarized by
UO22+ +2SO42– → UO2(SO4)22– (EQ 7)
UO2(SO4)22– + SO42– → UO2(SO4)34– (EQ 8)
104
FIGURE 2 Generalized flowchart of uranium recovery by ion exchange, solvent extraction, and
precipitation as yellowcake. Various adjustments of the process (e.g., makeup of eluant, solvent
cleanup, and strip solution bleed) are omitted from this diagram.
4(R+)HSO4– + UO2(SO4)34– → (R+)4UO2(SO4)34– + 4HSO4– (EQ 9)

+
The complex [(R )4UO2(SO4)34–] can be eluted from ion-exchange resin with organic sol-
vents, followed by solvent stripping (Figure 2). Commonly used organic solvents are alkyl
phosphoric acids, tri-butyl phosphate, and Alamine-336 (Merritt, 1971; Naden et al., 1985).
Kerosene is generally used as a diluent in solvent-extraction systems. Solvent extraction has
been believed to be incompatible with bacterial leaching because of the toxicity of kerosene
and other solvent residues to thiobacilli. These findings are based on laboratory studies, but
it is not known whether it is a relevant consideration in uranium leaching. Information on
concentrations of residues of solvent-extraction chemicals in uranium bioleaching solutions
is scanty, and the half-lives and susceptibility of these chemicals to abiotic and biological
decomposition in acid, iron-containing solutions are not known.
Elution of ion-exchange resins is also possible with inorganic solutions such as sulfate,
nitrate, or chloride salt that give pure and highly concentrated uranium solution (strip solu-
tion) from which high-grade uranium oxide can be precipitated with a base (e.g., NH3, MgO
or NaOH). Uranium is formulated to sodium diuranate (Na2U2O7) or ammonium diuranate
[(NH4)2U2O7], which is commercially known as yellowcake and which has a grade equiva-
lent to about 88% to 91% U3O8 after thickening, washing and drying as follows
2UO2SO4 + 6NH4OH → (NH4)2U2O7 + 2(NH4)2SO4 + 3H2O (EQ 10)
FE- AND S-OXIDIZING MICROORGANISMS AND BIOLEACHING REACTIONS

Bench-scale studies concerned with the microbiological leaching of uranium ores have exten-
sively used enrichment or pure cultures of iron- and sulfur-oxidizing Thiobacillus ferrooxidans
or mixed cultures of T. ferrooxidans and Thiobacillus thiooxidans (a sulfur-oxidizer). In addi-
tion, although it was not recognized until the 1970s, Leptospirillum ferrooxidans, an acido-
philic iron oxidizer, is common in enrichment or mixed cultures derived from mineral
surfaces, precipitates, or mine waters. All three bacteria use CO2 as the sole source of carbon
and grow in the acidic pH range of 1 to 3, which is typically used in uranium leach solutions.
As a general rule, T. ferrooxidans is inhibited below pH 1 and above pH 5. T. thiooxidans and
L. ferrooxidans have been shown to grow at pH levels as low as 0.5, and it seems probable
that the pH limits for extreme acidophiles are even lower in the environmental situations
found in metal and coal mines.
T. ferrooxidans has been assumed to be the most important bacterium for microbiological
leaching of uranium ores. There is not enough information available to assess the relative
abundance of T. thiooxidans and L. ferrooxidans in uranium mine environments. Recent find-
ings (Rawlings et al., 1999) suggest that L. ferrooxidans are numerically more dominant than
105
T. ferrooxidans in many copper leach mines, but it is not known whether the same applies to
uranium mines. Several reasons, including superior tolerance to low pH and high Fe3+ concen-
trations, are believed to account for the dominance of L. ferrooxidans (Rawlings et al., 1999).
For many decades, the species name T. ferrooxidans was used to generically signify bacte-
rial communities that were active in uranium leaching operations. This generic use of
nomenclature stemmed from the common practice of employing ferrous sulfate or
pyrite-based media to confirm the presence of iron-oxidizing bacteria.
According to Berthelot et al. (1997), moderately acidophilic thiobacilli (e.g., T. thioparus,
T. neapolitanus, T. intermedius, and T. novellus) are found in ore samples and mine waters at
circumneutral pH values. The role of these and other thiobacilli in the bioleaching of uranium
remains unclear. These thiobacilli oxidize sulfooxyanions (thiosulfate, polythionates, and
sulfite) found in mine tailing ponds (Silver and Dinardo, 1981; Silver, 1985a, 1985b, 1987).
Consortia of acidophilic thiobacilli (T. ferrooxidans and T. thiooxidans) and leptospirilli
(L. ferrooxidans) are believed to be superior to pure cultures for the biological oxidation of
sulfide minerals. Microbial diversity is extensive in acid mine waters, including uranium
mine leach solutions (Schippers et al., 1995; Berthelot et al., 1997). While the moderately
acidophilic thiobacilli are most numerous, thermoacidophilic archaea, resembling Sulfolobus
and Acidianus spp., have also been isolated from uranium-mine waste heaps (Schippers et
al., 1995).
Numerous heterotrophic bacteria are present in leach mines, and it is believed that only a
fraction of the microbiological diversity has been characterized to date. The presence of het-
erotrophic bacteria and eukaryotes (yeasts and fungi) in acid leach mines is thought to be the
consequence of organic compounds derived from the source water used in the leach solution,
from uranium processing chemicals and from the runoff water from adjacent surface and
subsurface soils. Iron- and sulfur-oxidizing bacteria are ubiquitous in leach mines, and the
release of metabolites and the eventual decomposition of biomass are also sources of organic
carbon for heterotrophs.
Several chemical and biochemical reactions take place concurrently in mineral-leaching
processes. Some principal leaching reactions mediated by thiobacilli (oxidation of Fe and S
entities) and leptospirilli (oxidation of Fe entity) and related bacteria are
2FeS2 + 7.5 O2 + H2O → 2Fe3+ + 4SO42– + 2H+ (EQ 11)
4Fe2+ + O2 + 4H+ → 4Fe3+ +2H2O (EQ 12)
2So + 3O2 + 2H2O → 2SO42– + 4H+ (EQ 13)

The oxidative leaching of uraninite (UO2) involves a redox reaction with Fe3+ as an oxidant
and bacterial reoxidation of Fe2+ (Figures 3 and 4) (Tuovinen, 1986; Muñoz et al., 1995; Bhatti
et al., 1998). The redox reaction resulting in uranium oxidation is between reduced uranium
(e.g., UO2) as the electron donor and Fe3+ as the electron acceptor. Uranium oxidation with O2
as the electron acceptor (Figure 3) is relatively slow compared with the Fe3+-mediated oxida-
tion. Dissolved Fe3+ and Fe2+ usually originate from the bacterial oxidation of Fe-sulfides such
as pyrite, pyrrhotite, marcasite and mackinawite. Pyrite is by far the most abundant Fe-sulfide
present in the uranium ores. Continuous regeneration of ferric sulfate by bacterial oxidation of
Fe2+ and Fe-sulfides (Figure 4) maintains the high Fe3+/Fe2+ ratios and redox potential values
needed for uranium leaching (Torma 1986; Tuovinen, 1986).
The regeneration of Fe3+-containing leach solutions by T. ferrooxidans-mediated oxidation
has been proposed as a separate process as it could be controlled to maximize oxidation kinet-
ics via nutrient addition, aeration, and pH-adjustment. Livesey-Goldblatt et al. (1977) devel-
oped a pilot-plant unit for the continuous oxidation of recycled ferrous sulfate leach solution
from uranium ore-processing plants. This process, called Bacfox, was based on the rapid oxida-
tion of an acidified solution of ferrous sulfate to ferric sulfate. The influent solution was satu-
rated with air and passed over a biofilm of T. ferrooxidans adhering to a solid (corrugated)
surface. The rate of Fe2+ oxidation was reported at 7.5 g/m2/h.
106
FIGURE 3 Schematic presentation of uranium oxidation coupled with Fe 3+ reduction (two Fe3+
reduced per each UIV oxidized) or O2 reduction (0.5 O2 reduced per each UIV oxidized). Both types of
redox reactions are relevant in uranium bioleaching systems that involve acidophilic iron- and sulfur-
oxidizing bacteria, but Fe3+-mediated oxidation is kinetically much faster, sometimes even by orders
of magnitude, depending on experimental conditions.
Other fixed-film systems have since been reported for the biological conversion of ferrous
sulfate to ferric sulfate, which are capable of exceeding the productivity of the Bacfox pro-
cess. These include various packed-bed and fluidized-bed reactor designs to promote oxygen
mass transfer and high biomass density (Jensen and Webb, 1995). Activated carbon, sand,
glass, and resin beads, polyvinyl chloride and ore particles have been tested for solid matrix
in these designs. Sankaran et al. (1996) used a Bacfox-type process for the extraction of ura-
nium from an Indian schistose quartzite ore. The performance of these biological fixed-film
Fe-oxidation systems under extremes of pH, toxic metals, and other shock loads have not
been thoroughly assessed, but they are much more resilient than bacteria in suspended
growth mode. Fixed-film iron oxidation systems involve biofilm formation, but much of the
mass in the biofilm is due to jarosite precipitation that serves as a porous matrix for bacteria
(Karamanev, 1991).
Mathematical models have been presented to define bacterial-growth associated bioleach-
ing reactions and to predict sulfide-mineral leach plant performance. Recent modeling efforts
in the bioleaching area have been summarized and critiqued by Haddadin et al. (1995). Such
models can be complex, and sometimes they tend to ignore many interactive variables that
impact the chemical and microbiological reactions. These include the type and geometry of
the mineral, the particle size distribution and the reactive surface area, the residence-time dis-
tribution of the solids, the pulp density, the bacterial numbers, the pH and redox potential, the
nutrient availability, the ferrous and ferric iron concentration and iron solubility, the agita-
tion, and the aeration. Bacteria in bioleaching systems are free-swimming in solution and
attached on solids, and the proportionality between bacterial attachment and detachment var-
ies. Bioleaching systems have multiple substrates in solution phase (dissolved iron and sulfur
107
FIGURE 4 Schematic presentation of bacterial reoxidation of Fe 2+ that occurs concurrently with
uranium leaching. Bacterial oxidation of pyrite, if present, produces acidic ferric sulfate and net
sulfuric acid. Bacterial oxidation of pyrrhotite (Fe 1–xS) and other monosulfides does not result in
net acid formation.
TABLE 1 Examples of crystallized Fe(III)-precipitates found in oxidized zones in acid mines.

All these precipitates are formed via solution-phase transformations.
Fe(III)-mineral Nominal composition
K-jarosite KFe3(SO4)2(OH)6
Na-jarosite NaFe3(SO4)2(OH)6
H3O-jarosite H3OFe3(SO4)2(OH)6
NH4-jarosite NH4Fe3(SO4)2(OH)6
Schwertmannite Fe8O8(OH)6SO4
Ferrihydrite Fe5HO8·4H2O
Goethite α-FeOOH
compounds) and heterogeneous solid phases (e.g., elemental S, sulfide minerals, and oxide
minerals). In uranium bioleaching, the oxidation of UIV on mineral surfaces is mediated by
Fe3+, and the oxidation and solubilization of uranium and regeneration of ferric iron each has
different rate expressions. The surface area and mass of particles change with time because of
the leaching, which makes it difficult to model the kinetics of solution-phase, solid-phase, and
solution-solid interfacial reactions. The formation of new solid phases in bioleaching systems
(e.g., So and Fe(III) precipitates) further complicates model development. Derivation of
kinetic parameters and data fitting into first-order rate expressions and shrinking-core and
shrinking-particle models (and various modifications thereof) have been presented for several
sulfide-mineral bioleaching systems with limited success, perhaps because of the lack of
understanding of all aspects of the bioleaching process coupled with limited parameters.
The solubility of Fe3+ is greatly influenced by pH, concentration, temperature, and ionic
composition. The solid phase controlling Fe3+ solubility at pH 2 to 3 is jarosite (Table 1), and
108
it can be minimized below pH 1.7. The abiotic oxidation of Fe is relatively rapid and affected
by pH and O2 concentration at pH >3.5 (Ackman and Kleinmann, 1984). In high-sulfate acid
environments, the hydrolysis of Fe3+ produces various Fe(III)-hydroxysulfate complexes, of
which poorly crystalline schwertmannite (Table 1) is the predominant product at pH 3 to 3.5
(Bigham et al., 1996a). Thus, schwertmannite can be found in leach residues when the pH
values are too high to promote jarosite precipitation (Bhatti et al., 1993). In an increasing pH
scale, poorly crystalline ferrihydrite and well-crystalline goethite become increasingly abun-
dant in acid sulfate systems (Bigham et al., 1996b). Fe(III)-precipitates can adsorb dissolved
ions including UO22+.
Microcalorimetric experiments with powdered uranium dioxide (UO2) and T. ferrooxidans
suggested that bacteria could directly oxidize uranous (UIV) compounds (Soljanto and
Tuovinen, 1980). Heat evolution was not observed under abiotic conditions (UO2 with or
without heat-inactivated cells of T. ferrooxidans). Because the reaction mixtures contained Fe
as an impurity, the results could have represented a combination of direct and indirect oxida-
tion of UO2 by T. ferrooxidans and Fe3+. Further experiments involving uranous sulfate con-
firmed that the oxidation was coupled with CO2 fixation, O2 consumption and reduction of
cytochromes and rusticyanin (DiSpirito and Tuovinen, 1981, 1982a, 1982b). Ivarson (1980)
also demonstrated O2 uptake coupled with uranous sulfate oxidation by T. ferrooxidans. But
the role of trace amounts of iron in uranous sulfate stock solution as an electron carrier
between U4+ and T. ferrooxidans could not be ruled out.*
In uranium-leaching systems, iron is invariably present in solution and the rate of ura-
nium oxidation to soluble species is a function of the rate of ferric iron regeneration by
iron-oxidizing bacteria. The production of H2SO4 and Fe3+ from Fe-sulfides (pyrite, marca-
site) and iron-mediated oxidation are all important mechanisms by which T. ferrooxidans
accelerates the leaching of reduced forms of uranium from ores.
Many factors influence the rate of bioleaching, including morphology and electronic struc-
ture of the mineral surface, particle size and surface area, pH, redox potential, temperature,
catalysts, partial pressure of oxygen, relative humidity, and chemical and mineralogical com-
position of the ore. Climatic conditions (season, ambient temperature, and rainfall) also
affect the kinetics of bioleaching reactions. Highly siliceous or carbonaceous gangue minerals
(e.g., calcite and dolomite) consume acid, thereby changing conditions beyond the pH range
suitable for acidophilic bacteria.
Temperature is among the most important parameters influencing the diversity of the
microbial communities. Thiobacillus acidophilus and Acidiphilium spp., in association with
T. ferrooxidans, have been isolated from a bioleaching uranium mine in Ontario, Canada
(Berthelot et al., 1993). The same study also demonstrated the psychrotrophic nature of
many Fe-oxidizers isolated from underground uranium mines where the water temperature
varied between 13° and 18°C. In general, acidophilic heterotrophic bacteria in mine waters
constitute a diverse group of organisms (Kishimoto et al., 1995) that vary in their ability to
utilize different carbon sources. Heterotrophic iron-oxidizing acidophiles have also been iso-
lated from mine water samples (Johnson, 1995).
Several acidophilic iron-oxidizing thiobacilli have been isolated from the Agnew Lake
mine, which employed surface-heap and stope-leaching processes for uranium solubilization
(Tuovinen et al., 1981). They varied in their plasmid composition, but the functions of the
products of plasmid-borne genes remained obscure. Numbers of iron-oxidizing acidophiles
varied in the range of <1 to >106 cells per milliliter of leach solution, but their distribution did
not have an apparent trend with respect to the season or the sampling site in the mine.
A diverse group of acidophilic autotrophs (T. ferrooxidans) and heterotrophs (Acidiphilium
cryptum, A. versutus, and A. symbioticum) have been isolated from uranium mine tailings
(Berthelot et al., 1997), and the presence of Thiobacillus acidophilus was also noted in these
* Curiously, the direct oxidation of uranous compounds by T. ferrooxidans prompted Sankaran (1985)
to suggest that uranium-oxidizing acidophiles may rightly be called Uranobacillus uranooxidans. No
further suggestions to this effect have appeared in the literature.
109
samples. Neutrophilic S-oxidizing bacteria resembling Thiobacillus thioparus have been iso-
lated from sandstone uranium-mine drainage (pH 7.2 to 7.8) (Mahmood, 1994). A novel iso-
late, named Thiobacillus plumbophilus because of its ability to grow with galena (PbS) as sole
source of energy, has been isolated from a German uranium mine (Drobner et al., 1992).
T. ferrooxidans derives energy from the oxidation of reduced iron and sulfur compounds
and fixes CO2 via the reductive Calvin-Benson cycle. A number of growth media for nutri-
ent requirements of T. ferrooxidans have been described in the literature. This bacterium
synthesizes its cell materials from inorganic constituents, namely, CO2, NH4+, K+, Mg2+,
Ca2+, SO42–, and PO43–. Several minor and trace metals (e.g., Cu, Zn, Co, Fe, and Mo) are
also essential growth factors. Organic cofactor requirements are not provided for any of
the commonly used media.
Phosphate- or sulfate-limited growth conditions cause serious metabolic imbalance and
restricted growth. This is because both nutrients are essentially required by growing cells of
T. ferrooxidans (McCready et al., 1986; Tuovinen, 1990). High phosphate concentrations in
mineral salts lead to the formation of insoluble complexes with Fe(III) (Hoffmann et al., 1985)
that may effectively coprecipitate and scavenge soluble U from leach solutions. Phosphate star-
vation of T. ferrooxidans acts as a signal for the expression of several new proteins (Jerez et al.,
1992, Varela et al., 1998) that are thought to be involved in scavenging phosphate through a
high-affinity uptake system. A great deal is known about the molecular basis of phosphate sen-
sory and regulatory system and signaling of gene expression in Escherichia coli and related bac-
teria (Wanner, 1998). The similarity in the phosphate sensory and regulatory systems during
phosphate starvation between these bacteria and T. ferrooxidans is unclear.
T. ferrooxidans and other iron-oxidizers can reduce sulfate to meet with sulfur require-
ments for biosynthetic purposes. Sulfate is also involved in Fe2+ oxidation by T. ferrooxidans,
as it has been shown to stimulate rusticyanin oxidation and reduction activity (Blake and
Shute, 1994; Yamanaka and Fukumori, 1995). One aspect of the sulfate requirement in iron
oxidation is the formation of a sulfate-ferrous ion complex as the physiological electron donor
(Blake and Shute, 1987; Kai et al., 1992; Blake et al., 1993).
In bioleaching situations, it is to be expected that many inorganic nutrients (e.g., K+,
Mg2+, Ca2+, SO42– and PO43–) are solubilized from minerals in host rock. Minor and trace
nutrient requirements have not been quantitatively determined for pure cultures nor for
bioleaching processes.
Ammonium ion (NH4+) is the preferred source of nitrogen for T. ferrooxidans and is sup-
plied as (NH4)2SO4 salt in the growth medium. Residues of inorganic nitrogen may be present
from explosives used for blasting and fracturing the orebody (Torma, 1987). Runoff (espe-
cially of nitrate) from adjacent surface soils and rainwater are also natural sources of inor-
ganic nitrogen in mine water. Fertilizer-grade ammonium sulfate supplements have been used
at some mine sites to promote acidophilic Fe- and S-oxidizing bacteria. Depending on the pre-
vailing pH, ferric iron concentration, and ionic composition, excess ammonium may poten-
tially be stripped from solution due to the formation of poorly soluble ammonium jarosite.
In general, microbial assimilation of ammonia is routed via biosynthesis of glutamate, ala-
nine, and aspartate. Glutamate biosynthesis is the most common pathway of ammonium assim-
ilation, and glutamate is the precursor of many amino acids, purines, and pyrimidines. The
assimilation involves the reductive formation of L-glutamate from α−ketoglutarate and ammo-
nium acceptor, mediated by glutamate dehydrogenase. Below 1 mM ammonium concentra-
tions, glutamate dehydrogenase is nonfunctional, and glutamate biosynthesis is mediated by
glutamine synthetase and glutamate synthase via a different pathway. The two pathways of
ammonia assimilation are regulated by ammonium, glutamate, and glutamine concentrations,
usually at transcriptional level. In the case of T. ferrooxidans, the glutamine synthetase struc-
tural gene glnA and glutamate synthase small subunit gene gltD have been cloned and
sequenced (Barros et al., 1985; Rawlings et al., 1987; Deane and Rawlings, 1996). Ammonium
starvation involves a stringent response that is regulated through a multigene network Ntr. The
regulatory genes ntrBC are usually in linkage to glnA in the γ-subdivision of proteobacteria,
whereas this linkage between glnA and ntrBC does not exist in T. ferrooxidans (Kilkenny et al.,
110
1994; Rawlings and Kusano, 1994) that belongs to the β-subdivision. It is not known whether
these distinct linkages reflect differences in Ntr gene regulation by gene products NtrB and
NtrC in proteobacteria in γ− and β−subdivisions.
Nitrate also serves as a nitrogen source for some iron-oxidizing thiobacilli, but the reductive
pathway involving nitrate and nitrite reductases has not been elucidated at molecular biochem-
ical level in T. ferrooxidans or other acidophilic iron- or sulfur-oxidizers.
Some isolates of T. ferrooxidans and at least one isolate of L. ferrooxidans have been
reported to fix molecular nitrogen as the nitrogen source (Norris et al., 1995). The nitroge-
nase activity has been assayed with acetylene reduction technique in the presence of ferrous
iron as the energy source and electron donor. The diazotrophic nature of several Fe-oxidizing
bacterial isolates has been verified by genetic analysis of the nif operon (Pretorius et al.,
1986, 1987; Rawlings, 1988). The nitrogenase structural genes nifHDK have been
sequenced, and the nifH promoter does not seem be uniquely different from those in diaz-
otrophic heterotrophic bacteria.
The environmental regulation of N2 fixation in T. ferrooxidans is not understood, as the
nitrogenase complex itself is sensitive to molecular oxygen, which can be abundantly present
in mine waters. It is not known whether the diazatrophic property of T. ferrooxidans substan-
tially contributes to nitrogen balance in mine waters and enhances the competitiveness of the
bacterium in its natural environment.
Aeration with CO2-enriched air has been reported to enhance the growth of acidophilic
thiobacilli with Fe2+ and sulfur compounds and accelerate bioleaching rates of sulfide miner-
als (Bailey and Hansford, 1993; Jaworski and Ubanek, 1997). Apparent CO2 limitation
results from mass-transfer rates that slow down as the pulp density increases and from the
kinetic properties of the ribulose-1, 5-bisphosphate carboxylase enzyme responsible for the
initial step of CO2 fixation. Acid dissolution of carbonaceous gangue minerals contributes to
available CO2 balance in microsites, but dissolution products are often poorly soluble (e.g.,
gypsum, CaSO4·2H2O) and can create diffusion barriers.
Mass transfer of dissolved oxygen is also believed to be a rate-limiting step in bioleaching
processes. Oxygen demand in reactor leaching systems is a function of the reactive surface
area (particle size), pulp density, and Fe-sulfide content of the ore material (Bailey and
Hansford, 1993). In heap- and stope-leaching systems, oxygen demand is difficult to predict
and to satisfy because of heterogeneity of ore material and nonuniform rates of oxygen con-
sumption and mass transfer at different depths in ore piles. Oxygen deficiency is inevitable in
deep zones in sulfide ore piles during active bioleaching. Rest periods, which are practised in
some leach mines, may help in oxygen diffusion and redistribution, although this has not
been unequivocally proven with analytical data. Schippers et al. (1995) reported that oxygen
concentration was a limiting factor for T. ferrooxidans at depths below 1.5 to 2 m in uranium-
mine waste heaps. Great variability in mass-transfer rates is to be expected at different sites
because of variations in compaction and hydraulic permeability.
T. ferrooxidans can grow anaerobically using sulfur compounds as electron donors and fer-
ric iron as electron acceptor (Pronk and Johnson, 1993). This redox reaction can be coupled
at the level of hydrogen sulfide, sulfite, and elemental sulfur oxidation via a sulfur-ferric iron
oxidoreductase enzyme complex (Sugio et al., 1987, 1989, 1992). Pronk et al. (1991, 1992)
demonstrated that ferric iron reduction with elemental sulfur as the electron donor repre-
sents anaerobic respiration because it was accompanied by electron transport and proton
translocation activity.
Inorganic nutrient formulations have been used for biological leaching processes involving
sulfide mineral concentrates. These formulations are based on empirical approaches that
estimate inorganic nutrient requirements from the chemical composition of microbial biom-
ass. They typically contain excess levels of inorganic nutrients, which may promote
Fe(III)-precipitation. Sometimes, the precipitation may defeat nutrient provision, especially
with monovalent cations (e.g., NH4+ and K+), which precipitate as jarosites and, thus, are
effectively sequestered to a biologically unavailable form. Factors controlling ferric iron
111
solubility and jarosite formation are also important in maintaining a balance of nutrients
available for bacteria in leach solutions.
T. ferrooxidans is metabolically active within pH values of 1.5 and 5.0. Ferric iron solubil-
ity is also controlled by pH, and precipitation is pronounced above pH 2. Below pH 3, ferric
iron solubility is controlled by jarosite formation (Ahonen and Tuovinen, 1994, 1995). The
acidity of mine waters is the result of carbonate-bicarbonate alkalinity. Ferrous iron oxida-
tion is a proton-consuming reaction, whereas the formation of jarosites or schwertmannite
are acid-forming reactions. Furthermore, complete oxidation of monosulfides are acid-
consuming reactions, whereas the oxidation of disulfides (e.g., FeS2 and CuFeS2) are
acid-producing reactions. Thus, the acidity of a leach solution is the sum of several
dynamic microbiological and abiotic reactions.
Surface properties of ores and minerals play an important role in the rate of metal solu-
bilization. Sulfide content greatly determines the reactivity of exposed mineral surface
with leach liquor and bacteria. The presence of exposed sulfide may also be a determinant
for promoting bacterial attachment on ore particles, but experimental evidence for this is
still ambiguous. A decrease in grain size increases specific surface area, and higher metal-
leaching yields can be obtained with no change in the total mass of particles. A decrease in
grain size also exposes surfaces of gangue mineral and may lead to increased acid con-
sumption and solubilization of minor and trace elements. An increase in pulp density may
suppress bacteria due to mass-transfer limitation (O2 and CO2) and may have inhibitory
effects on acidophilic iron- and sulfur-oxidizing thiobacilli due to the dissolution of toxic
ions. It is not known whether the activity of bacteria is hampered because of elevated levels
of soluble uranium in leach solutions.
Temperature is an important environmental parameter that influences bacterial activity in
bioleaching operations. Rates of chemical and biological reactions are influenced by the pre-
vailing temperature conditions. In harsh winter climatic conditions, the temperature of the
leaching solution could be as low as just above 0°C, and summer temperatures can be as high
as 35°C. Berthelot et al. (1993, 1994, 1997) found numerous acidophilic iron-oxidizers and
heterotrophs in samples from uranium mines in Ontario, Canada. The temperatures of the
mine water samples were in the range of 13° to 18°C. Iron dissolution from a pyritic ore,
coupled with O2 consumption and 14CO2 incorporation into cell material, has been reported
for mixed bacterial cultures growing at 0°C (Langdahl and Ingvorsen, 1997). These cultures
were derived from samples of a weathered sulfide ore deposit in North Greenland, and they
could be classified as psychrotophs because the optimum temperature was near 21°C. In
general, T. ferrooxidans are active at temperatures up to 40° to 42°C (Niemelä et al., 1994).
Temperatures in the range of 45°C are prohibitive to T. ferrooxidans (Modak et al., 1996).
T. ferrooxidans responds to shifts to unfavorable temperatures by the synthesis of multiple
heat-shock proteins (Varela and Jerez, 1992; Hubert et al., 1995), which is a universal
response in microorganisms to adverse temperatures. Although the function of the heat-
shock proteins has not been elucidated specifically in T. ferrooxidans, it is known from stud-
ies with other bacteria that heat-shock proteins act as molecular chaperones that are
involved in the assembly and folding of other proteins (Yura et al., 1993; Dawes, 1999).
Some heat-shock proteins have proteolytic activity, which enables the cells to dispose of
heat-denatured proteins via enzymatic degradation.
It is not known whether moderately thermophilic acidophiles (e.g., Thiobacillus caldus) are
present in uranium leach mines; these bacteria have an optimum at about 45°C and grow at
temperatures up to about 52°C (Hallberg and Lindström, 1994). Other examples of moder-
ately thermophilic iron-oxidizers are Leptospirillum thermoferrooxidans (optimum at 45° to
50°C, maximum at about 60°C), described by Golovacheva et al. (1992); an unnamed isolate
TI-1 (optimum 48°C) described by Sugio et al. (1993, 1995), Acidimicrobium ferrooxidans
(optimum between 45° and 50°C), described by Clark and Norris (1996) and several
Sulfobacillus spp. (optimum between 45° and 50°C) described by Norris et al. (1996) and Nor-
ris (1997). Similarly, the role of thermophilic archaea (e.g., Sulfolobus spp.) in uranium mines
practicing bioleaching is unknown. Elevated temperatures are usually the result of intense
112
oxidation of pyrite phases. Thermoacidophilic iron- and sulfur-oxidizers are discussed in recent
reviews by Norris (1997), Johnson and Roberto (1997) and Norris and Johnson (1998).
Acidophilic iron-oxidizers have been found abundantly in zones at temperatures up to
45°C (and even higher) in waste rock dumps in the Rum Jungle area in Australia (Goodman
et al., 1981a, 1981b). Further identification of these bacteria was not reported, but it is plau-
sible that they represent the T. caldus-type or other related moderate thermophiles. Temper-
atures in underground mines fall in the general range of 5° to 15°C, depending on the
geological setting of ore-deposits (Murr and Brierley, 1978; Kelley and Tuovinen, 1988). Iron
oxidation occurs at temperatures as low as near 0°C, but the rate constant of the reaction is
kinetically not favorable (Kovalenko et al., 1982; Ferroni et al., 1986). For harsh tempera-
ture regions in Pakistan, Bhatti and Malik (1997b) reported a significant decrease in rates of
uranium bioleaching during hot summer months when the temperatures were in excess of
42° to 50°C.
Active T. ferrooxidans bacteria have been noted in uranium mine leach solutions that con-
tained as much as 12 g/L U3O8 (Duncan and Bruynesteyn, 1971). In pure culture, some iso-
lates of T. ferrooxidans can develop resistance to over 2 g/L U3O8 (>10 mM U) (Tuovinen and
Kelly, 1974; DiSpirito and Tuovinen, 1981). Uranium resistance T. ferrooxidans needs con-
tinuous selection, because repeated subculture of uranium-resistant T. ferrooxidans in the
absence of uranium can result in the loss of the resistance trait (Martin et al., 1983). The
genetic basis of uranium resistance remains unknown. Genes encoding uranium resistance
would be of great interest, not only for genetic manipulation, but for the fundamental under-
standing of how microorganisms cope with toxic actinides in their environment.
CONCLUDING REMARKS
T. ferrooxidans and other acidophiles display considerable diversity in phenotypic character-
istics, such as metal resistance, pH tolerance, temperature extremes, and substrate oxidation
kinetics, but the underlying fundamental genetic system is poorly known. Molecular genetic
studies with T. ferrooxidans genes have relied on cloning and expression in Escherichia coli,
because recombinant DNA technology for gene-transfer systems among iron-oxidizing thio-
bacilli is not well established. This is a serious limitation to using molecular genetic
approaches to improve desirable properties in these bacteria. Plasmids are common in these
acidophilic thiobacilli, and they vary in size. Their role in conferring phenotypic variation is
obscure because most T. ferrooxidans plasmids have remained cryptic in spite of extensive
efforts to find the function of plasmid-borne gene products. The molecular genetic character-
ization of Leptospirillum ferrooxidans and Thiobacillus thiooxidans is in very incipient stages.
To date, there is no evidence to suggest that iron-oxidizing bacteria in uranium mines
have unique resistance traits that are not found in similar acidophilic bacteria living in other
metal mines, coal mines or in tailings. Several T. ferrooxidans isolates also have insertion
sequences (Rawlings and Kusano, 1994, Chakravarty et al., 1997; Holmes et al., 1999;
Rawlings, 1999) that are capable of integrating themselves into plasmids or chromosomal
DNA independently of the host cell recombination. Because insertion sequences and plas-
mids are native genetic elements in T. ferrooxidans, they are likely to confer resistance factors
that help the bacteria survive under adverse environmental conditions. In the absence of
established gene-transfer technology, however, improvement of T. ferrooxidans and other
acidophilic cultures for bioleaching processes must continue to resort to traditional methods
of enrichment, selection, and adaptation.
REFERENCES
Ackman, T.E., and Kleinmann, R.L.P., 1984, “In-line aeration and treatment of acid mine
drainage,” US Bureau of Mines, Report of Investigation 8868.
113
Ahonen, L., and Tuovinen, O.H., 1994, “Solid-phase alteration and iron transformation in column
bioleaching of a complex sulfide ore,” Environmental Geochemistry of Sulfide Oxidation, C.N.
Alpers and D.W. Blowes, eds., American Chemical Society, Washington, D.C., pp. 79–89.
Ahonen, L., and Tuovinen, O.H., 1995, “Bacterial leaching of complex sulfide ore samples in
bench-scale column reactors,” Hydrometallurgy, Vol. 37, pp. 1–21.
Anonymous, 1979, “Solution mining of U3O8 at Agnew Lake is jeopardized by faulting in
orebody,” Journal of Mining Engineering, Vol. 180, p. 52.
Audsley, A., and Daborn, G.R., 1963, “Natural leaching of uranium ores. 2. A study of the
experimental variables. 3. Application to specific ores,” Transactions of the Institution of Mining
and Metallurgy, London, Vol. 72, pp. 235–254.
Bailey, A.D., and Hansford, G.S., 1993, “Factors affecting bio-oxidation of sulfide minerals at high
concentrations of solids,” Biotechnology and Bioengineering, Vol. 42, p. 1164–1174.
Barros, M.E., Rawlings, D.E., and Woods, D.R., 1985, “Cloning and expression of the Thiobacillus
ferrooxidans glutamine synthetase gene in Escherichia coli,” Journal of Bacteriology, Vol. 164,
pp. 1386–1389.
Berthelot, D., Leduc, L.G., and Ferroni, G.D., 1993, “Temperature studies of iron-oxidizing
autotrophs and acidophilic heterotrophs isolated from uranium mines,” Canadian Journal of
Microbiology, Vol. 39, pp. 384–388.
Berthelot, D., Leduc, L.G., and Ferroni, G.D., 1994, “The absence of psychrophilic Thiobacillus
ferrooxidans and acidophilic heterotrophic bacteria in cold, tailings effluents from a uranium
mine,” Canadian Journal of Microbiology, Vol. 40, pp. 60–63.
Berthelot, D., Leduc, L.G., and Ferroni, G.D., 1997, “Iron-oxidizing autotrophs and acidophilic
heterotrophs from uranium mine environments,” Geomicrobiology Journal, Vol. 14, pp. 317–324.
Bhatti, T.M., Bigham, J.M., Carlson, L., and Tuovinen, O.H., 1993, “Mineral products of pyrrhotite
oxidation by Thiobacillus ferrooxidans,” Applied and Environmental Microbiology, Vol. 59, pp.
1984–1990.
Bhatti, T.M., Vuorinen, A., Lehtinen, K.M., and Tuovinen, O.H., 1998, “Dissolution of uraninite in
acid solutions,” Journal of Chemical Technology and Biotechnology, Vol. 73, pp. 259–263.
Bhatti, T.M., and Malik, K.A., 1997a, “Pilot-scale processing of 125 tones low-grade sandstone
uranium ores by microbial heap leaching process,” Technical Report No. NIBGE-3/1997,
National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan.
Bhatti, T.M., and Malik, K.A., 1997b, “Acid and microbiological leaching of a sandstone uranium
ore,” Biotechnology Comes of Age; Conference Proceedings of the International
Biohydrometallurgy Symposium IBS’97, BIOMINE 97, Australian Mineral Foundation, Sydney,
NSW, Australia, pp. PM4.1–PM4.2.
Bigham, J.M., Schwertmann, U., Traina, S.J., Winland, R.I., and Wolfe, M., 1996a,
“Schwertmannite and the chemical modelling of iron in acid sulfate waters,” Geochimica et
Cosmochimica Acta, Vol. 60, pp. 2111–2121.
Bigham, J.M., Schwertmann, U., and Pfab, G., 1996b, “Influence of pH on mineral speciation in a
bioreactor simulating acid mine drainage,” Applied Geochemistry, Vol. 11, pp. 845–849.
Blake, R.C., II, and Shute, E.A., 1987, “Respiratory enzymes of Thiobacillus ferrooxidans. A kinetic
study of electron transfer between iron and rusticyanin in sulfate media,” Journal of Biological
Chemistry, Vol. 262, pp. 14983–14989.
Blake, R.C., II, Shute, E.A., Greenwood, M.M., Spencer, G.H., and Ingledew, W.J., 1993,
“Enzymes of aerobic respiration on iron,” FEMS Microbiology Reviews, Vol. 11, pp. 9–18.
Blake, R.C., II, and Shute, E.A., 1994, “Respiratory enzymes of Thiobacillus ferrooxidans. Kinetic
properties of an acid-stable iron:rusticyanin oxidoreductase,” Biochemistry, Vol. 33,
pp. 9920–9228.
Bosecker, K., and Wirth, G., 1980, “Bacterial leaching of carbonate bearing uranium ore,”
Biogeochemistry of Ancient and Modern Environments, P.A. Trudinger, M.R. Walter, and B.J.
Ralph, eds., Australian Academy of Sciences, Canberra, Australia, pp. 557–562.
Cameron, J., 1963, “Discussion on natural leaching of uranium ores,” Transactions of the
Institution of Mining and Metallurgy, London, Vol. 72, p. 52.
114
Chakravarty, L., Kittle, J.D., Jr., and Tuovinen, O.H., 1997, “Insertion sequence IST3091 of
Thiobacillus ferrooxidans,” Canadian Journal of Microbiology, Vol. 43, pp. 503–508.
Chutinara, D., Torma, A.E., and Singh, A.K., 1984, “Kinetics of sulfuric acid leaching of a uranium
ore with oxone,” Metall, Vol. 38, pp. 121–126.
Clark, D.A., and Norris, P.R, 1996, “Acidimicrobium ferrooxidans gen. nov., sp. nov.: mixed-culture
ferrous iron oxidation with Sulfobacillus species,” Microbiology, U.K., Vol. 142, pp. 785–780.
Dawes, I.W., 1999, “Stress response,” The metabolism and molecular physiology of Saccharomyces
cervisiae, J.R. Dickinson and M. Schweizer, eds., Taylor & Francis, London, U.K., pp. 277–326.
Deane, S.M., and Rawlings, D.E., 1996, “Cloning and sequencing of the gene for the Thiobacillus
ferrooxidans ATCC33020 glutamate synthase (GOGAT) small subunit and complementation of
an Escherichia coli gltD mutant,” Gene, Vol. 177, pp. 261–263.
DiSpirito, A.A., and Tuovinen, O.H., 1981, “Oxygen uptake coupled with uranous sulphate
oxidation by Thiobacillus ferrooxidans and Thiobacillus acidophilus,” Geomicrobiology Journal,
Vol. 2, pp. 275–291.
DiSpirito, A.A., and Tuovinen, O.H., 1982a, “Uranous ion oxidation and carbon dioxide fixation
by Thiobacillus ferrooxidans,” Archives of Microbiology, Vol. 133, pp. 28–32.
DiSpirito, A.A., and Tuovinen, O.H., 1982b, “Kinetics of uranous ion and ferrous iron oxidation by
Thiobacillus ferrooxidans,” Archives of Microbiology, Vol. 133, pp. 33–37.
Drobner, E., Huber, H., Rachel, R., and Stetter, K.O., 1992, “Thiobacillus plumbophilus sp. nov., a
novel galena and hydrogen oxidizer,” Archives of Microbiology, Vol. 157, pp. 213–217.
Duncan, D.W., and Bruynesteyn, A., 1971, “Enhancing bacterial activity in a uranium mine,”
Canadian Institute of Mining and Metallurgy Bulletin, Vol. 64, pp. 32–36.
Eligwe, C.A., and Torma, A.E., 1986, “Advances in the hydrogen peroxide leaching of uranium
ores,” Metall, Vol. 40, pp. 491–496.
Ferroni, G.D., Leduc, L.G., and Todd, M., 1986, “Isolation and temperature characteristics of
psychrotrophic strains of Thiobacillus ferrooxidans from the environment of a uranium ore,”
Journal of General and Applied Microbiology, Vol. 32, pp. 169–175.
Glombitzal, F., Eckardt, L., Hummel, A., Loffler, R., Schreyer, J., Nindel, K., and Zimmermann, U.,
1995, “Biological activity and water quality at flooding of the uranium mine ‘Konigstein’,”
Biohydrometallurgical Processing, Vol. II, C.A. Jerez, T. Vargas, H. Toledo, and J.V. Wiertz, eds.,
Universidad de Chile, Santiago, Chile, pp. 315–322.
Golovacheva, R S., Golyshina, O.V., Karavaiko, G.I., Dorofeev, A.G., Pivovarova, T.A., and
Chernykh, N.A., 1992, “A new iron-oxidizing bacterium, Leptospirillum thermoferrooxidans, sp.
nov.,” Mikrobiologiya, Vol. 61, pp. 1056–1065.
Goodman, A.E., Khalid, A.M., and Ralph, B.J., 1981a, “Microbial ecology of Rum Jungle. 1:
Environmental study of sulfidic overburden dumps, experimental heap-leach piles and tailings
dam area,” Australian Atomic Energy Commission Report No. AAECIE531, Lucas Heights,
NSW, Australia.
Goodman, A.E., Khalid, A.M., and Ralph, B.J., 1981b, “Microbial ecology of Rum Jungle. II:
Environmental study of two flooded open cuts and smaller associated water bodies,”
Australian Atomic Energy Commission Report No. AAEC/E57, Lucas Heights, NSW, Australia.
Gow, W.A., McCreedy, H.H., Ritcey, G.M., McNamara, V.M., Harrison, V.F., and Lucas, B.H.,
1971, “Bacteria-based processes for the treatment of low-grade uranium ore,” The Recovery of
Uranium, International Atomic Energy Agency, Vienna, Austria, pp. 195–211.
Haddadin, J., Dagot, C., and Fick, M., 1995, “Models of bacterial leaching,” Enzyme and Microbial
Technology, Vol. 17, pp. 290–305.
Hallberg, K.B., and Lindström, E.B., 1994, “Characterization of Thiobacillus caldus sp. nov., a
moderately thermophilic acidophile,” Microbiology, U.K., Vol. 104, pp. 3451–3456.
Haque, K., and Ritcey, G.M., 1982, “Comparison of oxidants for the leaching of uranium ores in
sulfuric acid,” Canadian Institute of Mining and Metallurgy Bulletin, Vol. 75, pp. 127–133.
Hoffmann, M.R., Hiltunen, P., and Tuovinen, O.H., 1985, “Inhibition of ferrous ion oxidation by
Thiobacillus ferrooxidans in the presence of oxyanions of sulfur and phosphorus,” Processing
and Utilization of High Sulfur Coals, Y.A. Attia, ed., Elsevier, Amsterdam, pp. 683–698.
115
Holmes, D.S., Jedlicki, E., Cabrejos, M.E., Bueno, S., Guacucano, M., Inostroza, C., Levican, G.,
Varela, P., and Garcia, E., 1999, “The use of insertion sequences to analyse gene function in
Thiobacillus ferrooxidans: a case study involving cytochrome c-type biogenesis proteins in iron
oxidation,” Biohydrometallurgy and the Environment Toward the Mining of the 21st Century,
Part B, Molecular Biology, Biosorption, Bioremediation, R. Amils and A. Ballester, eds., Elsevier,
Amsterdam, pp. 139–147.
Hubert, W.A., Leduc, L.G., and Ferroni, G.D., 1995, “Heat and cold shock responses in different
strains of Thiobacillus ferrooxidans,” Current Microbiology, Vol. 31, pp. 10–14.
Ivarson, K.C., 1980, “Enhancement of uranous-ion oxidation by Thiobacillus ferrooxidans,” Current
Jaworska, M., and Urbanek, A., 1997, “The influence of carbon dioxide concentration in liquid
medium on elemental sulphur oxidation by Thiobacillus thiooxidans,” Bioprocessing
Engineering, Vol. 16, pp. 361–365.
Jensen, A.B., and Webb, C., 1995, “Ferrous sulphate oxidation using Thiobacillus ferrooxidans:
a review,” Process Biochemistry, Vol. 30, pp. 225–236.
Jerez, C.A., Seeger, M., and Amaro, A.M., 1992, “Phosphate starvation affects the synthesis of
outer membrane proteins in Thiobacillus ferrooxidans,” FEMS Microbiology Letters, Vol. 98, pp.
29–34.
Johnson, D.B., 1995, “Mineral cycling by microorganisms: iron bacteria,” Microbial Diversity and
Ecosystem Function, D. Allsopp, R.R. Colwell, and D.L. Hawksmith, eds., CAB International,
Wallingford, U.K., pp. 137–159.
Johnson, D.B., and Roberto, F.F., 1997, “Heterotrophic acidophiles and their roles in the
bioleaching of sulfide minerals,” Biomining: Theory, Microbes, and Industrial Processes, D.E.
Rawlings, ed., Springer-Verlag, Heidelberg, pp. 259–279.
Kai, M., Yano, T., Tamegai, H., Fukumori, Y., and Yamanaka, T., 1992, “Thiobacillus ferrooxidans
cytochrome c oxidase: purification, and molecular and enzymatic features,” Journal of
Biochemistry, Tokyo, Vol. 112, pp. 816–821.
Karamanev, D.G., 1991, “Model of biofilm structure of Thiobacillus ferrooxidans,” Journal of
Biotechnology, Vol. 20, pp. 51–64.
Kelley, B.C., and Tuovinen, O.H., 1988, “Microbiological oxidations of minerals in mine tailings,”
Chemistry and Biology of Solid Waste: Degraded Materials and Mine Tailings, W. Salomons and
U. Förstner, eds., Springer-Verlag, Berlin, pp. 33–53.
Kilkenny, C.A., Berger, D.K., and Rawlings, D.E., 1994, “Isolation of Thiobacillus ferrooxidans
ntrBC genes using a T. ferrooxidans nifH-lacZ fusion,” Microbiology, U.K., Vol. 140,
pp. 2543–2453.
Kishimoto, N., Kosako, Y., Wakao, N., Tano, T., and Hiraishi, A., 1995, “Transfer of Acidiphilium
facilis and Acidiphilium aminolytica to the genus Acidocella gen. nov., and emendation of the
genus Acidiphilium,” Systematic and Applied Microbiology, Vol. 18, pp. 85–91.
Kovalenko, T.V., Karavaiko, G.I., and Piskunov, V.P., 1982, “Effect of Fe3+ ions in the oxidation of
ferrous iron by Thiobacillus ferrooxidans at various temperatures,” Mikrobiologiya, Vol. 50, pp.
231–236.
Langdahl, B.R., and Ingvorsen, K., 1997, “Temperature characteristics of bacterial iron
solubilisation and 14C assimilation in naturally exposed sulfide ore material at Citronen Fjord,
North Greenland (83°N),” FEMS Microbiology Ecology, Vol. 23, pp. 275–283.
Lawes, B.C., 1978, “The effect of sodium silicate on leaching of uranium ores with hydrogen
peroxide,” In-Situ, Vol. 2, pp. 75–79.
Livesey-Goldblatt, E.P., Tunley, T.H., and Nagy, I.F., 1977, “Pilot plant bacterial film oxidation
(BACFOX process) of recycled acidified uranium plant ferrous sulphate leach solution,”
Bacterial Leaching, W. Schwartz, ed., Verlag Chemie, Weinheim, pp. 175–190.
Lowson, R.T., 1975, “Bacterial leaching of uranium ores—a review,” Australian Atomic Energy
Commission’s Publication AAEC/ E356, Lucas Heights, NSW, Australia.
MacGregor, R.A., 1966, “Recovery of U3O8 by underground leaching,” CIM Bulletin, Vol. 59,
pp. 583–587.
116
MacGregor, R.A., 1969, “Uranium dividends from bacterial leaching,” Journal of Mining
Engineering, Vol. 21, p. 54.
Mahmood, T., 1994, “Bacterial heap leaching studies of low-grade ores from Siwalik sandstone
ore deposits, Sulaiman Range, Pakistan,” Ph.D. Thesis, University of the Punjab, Lahore,
Pakistan.
Marchbank, A., 1987, “Update on uranium leaching at Denison Mines,” Proceedings of the 4th
Annual General Meeting of BIOMINET, R.G.L. McCready, ed., CANMET Special Publication No.
SP87-10, Canadian Government Publication Centre, Ottawa, Ontario, pp. 3–18.
Martin, P.A.W., Dugan, P.R., and Tuovinen, O.H., 1983, “Uranium resistance of Thiobacillus
ferrooxidans,” European Journal of Applied Microbiology and Biotechnology, Vol. 18,
pp. 392–395.
Mattus, A.J., Jr., and Torma, A.E., 1980, “Uranium extraction from a low-grade ore by acidic
leachants,” Metall, Vol. 34, pp. 33–36.
McCready, R.G.L., Walden, D., and Marchbank, A., 1986, “Nutrient requirement for the in-place
leaching of uranium by Thiobacillus ferrooxidans,” Hydrometallurgy, Vol. 17, pp. 61–71.
McCready, R.G.L., and Gould, W.D., 1990, “Bioleaching of uranium,” Microbial Mineral Recovery,
H.L. Ehrlich, and C.L. Brierley, eds., McGraw-Hill, NY, pp. 107–125.
Merritt, R.C., 1971, The Extractive Metallurgy of Uranium, Colorado School of Mines Research
Institute, Golden, CO.
Miller, R.P., Napier, E., and Wells, R.A., 1963, “Natural leaching of uranium ores. 1. Preliminary
tests on Portuguese ores,” Transactions of the Institution of Mining and Metallurgy, London, Vol.
72, pp. 217–234.
Modak, J.M., Natarajan, K.A., and Mukhopadhyay, S., 1996, “Development of
temperature-tolerant strains of Thiobacillus ferrooxidans to improve bioleaching kinetics,”
Hydrometallurgy, Vol. 42, pp. 51–61.
Muñoz, J.A., González, F., Blázquez, M.I., and Ballester, A., 1995, “A study of a Spanish uranium
ore. Part I. A review of the bacterial leaching in the treatment of uranium ores,”
Hydrometallurgy, Vol. 38, pp. 39–57.
Murr, L.E., and Brierley, J.A., 1978, “The use of large scale facilities in studies of the role of
microorganisms in commercial leaching operations,” Metallurgical Applications of Bacterial
Leaching and Related Microbiological Phenomena, L.E. Murr, A.E. Torma, and J.A. Brierley, eds.,
Academic Press, NY, pp. 491–520.
Naden, D., Rowden, G.A., Lunt, D.J., and Wilson, D., 1985, Development and application of
solvent extraction, ion exchange (RIP) and liquid membrane processes for uranium recovery,”
Advances in Uranium Ore Processing and Recovery from Non-Conventional Resources,
IAEA-TC-491/4, International Atomic Energy Agency, Vienna, Austria, pp. 53–72.
Niemelä, S.I., Sivelä, C., Luoma, T., and Tuovinen, O.H., 1994, “Maximum temperature limits for
acidophilic, mesophilic bacteria in biological leaching systems,” Applied and Environmental
Norris, P.R., Clark, D.A., Owen, J.P., and Waterhouse, S., 1996, “Characteristics of Sulfobacillus
acidophilus sp. nov. and other moderately thermophilic mineral-sulphide-oxidizing bacteria,”
Microbiology, U.K., Vol. 142, pp. 775–783.
Norris, P.R., Murrell, J.C., and Hinson, D., 1995, “The potential for diazotrophy in iron- and
sulfur-oxidizing acidophilic bacteria,” Archives of Microbiology, Vol. 164, pp. 294–300.
Norris, P.R., 1997, “Thermophiles and bioleaching,” Biomining: Theory, Microbes, and Industrial
Processes, D.E. Rawlings, ed., Springer Verlag, Heidelberg, pp. 247–258.
Norris, P.R., and Johnson, D.B., 1998, “Acidophilic microorganisms,” Extremophiles: Life in
Extreme Environments, K. Horikoshi and W.D. Grant, eds., Wiley-Liss, NY, pp. 133–153.
Pourbaix, M., 1966, Atlas of Electrochemical Equilibria in Aqueous Solutions, Pergamon Press,
London.
Pretorius, I.M., Rawlings, D.E., and Woods, D.R., 1986, “Identification and cloning of Thiobacillus
ferrooxidans structural nif genes in Escherichia coli,” Gene, Vol. 45, pp. 59–65.
117
Pretorius, I.M., Rawlings, D.E., O’Neill, E.G., Jones, W.A., Kirby, R., and Woods, D.R., 1987,
“Nucleotide sequence of the gene encoding the nitrogenase iron protein of Thiobacillus
ferrooxidans,” Journal of Bacteriology, Vol. 169, pp. 367–370.
Pronk, J.T., Liem, K., de Bruyn, J.C., and Kuenen, J.G., 1991, “Energy transduction by anaerobic
ferric iron respiration in Thiobacillus ferrooxidans,” Applied and Environmental Microbiology,
Vol. 57, pp. 2063–2068.
Pronk, J.T., de Bruyn, J.C., Bos, P. , and Kuenen, J.G. , 1992, “Anaerobic growth of Thiobacillus
ferrooxidans,” Applied and Environmental Microbiology, Vol. 58, pp. 2227–2230.
Pronk, J.T., and Johnson, D.B., 1993, “Oxidation and reduction of iron by acidophilic bacteria,”
Geomicrobiology Journal, Vol. 10, pp. 153–171.
Rawlings, D.E., 1988, “Sequence and structural analysis of the α− and β−dinitrogenase subunits of
Thiobacillus ferrooxidans,” Gene, Vol. 69, pp. 337–343.
Rawlings, D.E., 1999, “The molecular genetics of mesophilic, acidophilic, chemolithotrophic, iron,
or sulfur-oxidizing microorganisms,” Biohydrometallurgy and the Environment Toward the
Mining of the 21st Century, Part B, Molecular Biology, Biosorption, Bioremediation, R. Amils and
A. Ballaster, eds., Elsevier, Amsterdam, pp. 3–20.
Rawlings, D.E., Jones, W.A., O’Neill, E.G., and Woods, D.R., 1987, “Nucleotide sequence of the
glutamine synthetase gene and its controlling region from the acidophilic autotroph
Thiobacillus ferrooxidans,” Gene, Vol. 53, pp. 211–217.
Rawlings, D.E., and Kusano, T., 1994, “Molecular genetics of Thiobacillus ferrooxidans,”
Microbiological Reviews, Vol. 58, pp. 39–55.
Rawlings, D.E., Tributsch, H., and Hansford, G.S., 1999, “Reasons why ‘Leptospirillum’-like species
rather than Thiobacillus ferrooxidans are the dominant iron-oxidizing bacteria in many
commercial processes for the biooxidation of pyrite and related ores,” Microbiology, U.K., Vol.
145, pp. 5–13.
Sankaran, R.N., 1985, “Biogenic uranium fractionation,” Transactions of the Indian Institute of
Metals, Vol. 38, pp. 76–77.
Sankaran, R.N., Yadava, R.S., Sen, D.B., Kulshrestha, S.C., Mohanty, K.B., and Singh, J., 1996,
“Leachable U and Au in an Indian schistose quartzite,” Hydrometallurgy, Vol. 43, pp. 387–389.
Schippers, A., Hallmann, R., Wentzein, S., and Sand, W., 1995, “Microbial diversity in uranium
mine waste heaps,” Applied and Environmental Microbiology, Vol. 61, pp. 2930–2935.
Silver, M., and Dinardo, O., 1981, “Factors affecting oxidation of thiosalts by thiobacilli,” Applied
and Environmental Microbiology, Vol. 41, pp. 1301–1309.
Silver, M., 1985a, “Water leaching characteristics of uranium tailings from Ontario and northern
Saskatchewan,” Hydrometallurgy, Vol. 14, pp. 189–217.
Silver, M., 1985b, “Parameters for the operation of bacterial thiosalt oxidation ponds,” Applied
and Environmental Microbiology, Vol. 50, pp. 663–669.
Silver, M., 1987, “Distribution of iron-oxidizing bacteria in the Nordic uranium tailings deposit,
Elliot Lake, Ontario, Canada,” Applied and Environmental Microbiology, Vol. 53, pp. 846–852.
Soljanto, P., and Tuovinen, O.H., 1980, “A micro-calorimetric study of U(IV) oxidation by
Thiobacillus ferrooxidans and ferric-iron,” Biogeochemistry of Ancient and Modern Environments,
P.A. Trudinger, M.R. Walter and B.J. Ralph, eds., Australian Academy of Science, Canberra,
Australia, pp. 469–475.
Sugio, T., Mizanashi, W., Magaki, K., and Tano, T., 1987, “Purification and some properties of
sulfur:ferric iron oxidoreductase from Thiobacillus ferrooxidans,” Journal of Bacteriology, Vol.
169, pp. 4916–4922.
Sugio, T., Katagiri, T., Inagaki, K., and Tano, T., 1989, “Actual substrate for elemental sulfur
oxidation by sulfur:ferric iron oxidoreductase purified from Thiobacillus ferrooxidans,”
Biochimica et Biophysica Acta, Vol. 973, pp. 250–256.
Sugio, T., White, K.J., Shute, E., Choate, D., and Blake, R.C., II, 1992, “Existence of a hydrogen
sulfide:ferric iron oxidoreductase in iron-oxidizing bacteria,” Applied and Environmental
Sugio, T., Takai, M., and Tano, T., 1993, “Isolation and properties of a moderately thermophilic iron-
oxidizing bacterium,” Bioscience, Biotechnology, and Biochemistry, Vol. 57, pp. 1660–1662.
118
Sugio, T., Kishimoto, K., Takai, M., Oda, K., and Tano, T., 1995, “Growth of moderately
thermophilic iron-oxidizing bacterium strain TI-1 in synthetic medium,” Journal of
Fermentation and Bioengineering, Vol. 79, pp. 290–293.
Tomizuka, N., and Takahara, Y., 1972, “Bacterial leaching of uranium from Ningyo-Toge ores,”
Fermentation Technology Today, Proceedings of the 4th International Fermentation Symposium,
G. Terui, ed., Society for Fermentation Technology, Osaka, Japan, pp. 513–520.
Torma, A.E., 1986, “Biotechnology as an emerging technology,” Biotechnology and Bioengineering
Symposium, Vol. 16, pp. 49–63.
Torma, A.E., 1987, “Impact of biotechnology on metal extractions,” Mineral Processing and
Extractive Metallurgy Reviews, Vol. 2, pp. 289–330.
Tuovinen, O.H., and Kelly, D.P., 1974, “Studies on the growth of Thiobacillus ferrooxidans. II.
Toxicity of uranium growing cultures and tolerance conferred by mutation, other metal cations
and EDTA,” Archives of Microbiology, Vol. 95, pp. 153–169.
Tuovinen, O.H., Martin, P.A.W., Dugan, P.R., and Silver, M., 1981, “The Agnew Lake uranium
mine leach liquors: Chemical examinations, bacterial enumeration and composition of plasmid
DNA of iron-oxidizing thiobacilli,” Proceedings of International Conference on Use of
Microorganisms in Hydrometallurgy, Hungarian Academy of Sciences, Pécs, pp. 59–69.
Tuovinen, O.H., 1986, “Acid leaching of uranium ore material with microbial catalysis,”
Biotechnology and Bioengineering Symposium, Vol. 16, pp. 65–72.
Tuovinen, O.H., 1990, “Biological fundamentals of mineral leaching processes,” Microbial Mineral
Recovery, H.L. Ehrlich and C.L. Brierley, eds., McGraw-Hill, NY, pp. 55–77.
Varela, P., Levican, G., Rivera, F., and Jerez, C.A., 1998. “An immunological strategy to monitor in
situ the phosphate starvation state in Thiobacillus ferrooxidans,” Applied and Environmental
Varela, P., and Jerez, C.A., 1992, “Identification of GroEL and DnaK homologues in Thiobacillus
ferrooxidans,” FEMS Microbiology Letters, Vol. 98, pp. 149–154.
Wadden, D.A., and Gallant, A., 1985, “The in-place leaching of uranium at Denison Mines,”
Canadian Metallurgical Quarterly, Vol. 24, pp. 127–134.
Wanner, B.L., 1998, “Phosphate signaling and the control of gene expression in Escherichia coli,” Metal
Ions in Gene Regulation, S. Silver and W. Walden, eds., Chapman & Hall, NY, pp. 104–128.
Yamanaka, T., and Fukumori, Y., 1995, “Molecular aspects of the electron transfer system which
participates in the oxidation of ferrous ion by Thiobacillus ferrooxidans,” FEMS Microbiology
Reviews, Vol. 17, pp. 401–413.
Yura, T., Nagai, H., and Mori, H., 1993, “Regulation of heat-shock response in bacteria,” Annual
Review of Microbiology, Vol. 47, pp. 321–350.
119
Advances in the Application of
the BIOX® Process for
Refractory Gold Ores
P.C. van Aswegen* and H.J. Marais*
ABSTRACT
The Biological Oxidation (BIOX®) Process for the treatment of refractory gold ores was commercialized
in 1988. During the last decade, the technology has developed into a mature, commercially viable pro-
cess, with five additional operations commissioned, including operations at Sao Bento in Brazil, Har-
bour Lights and Wiluna in Australia, Ashanti in Ghana, and Tamboraque in Peru. With ongoing
development, the efficiency and cost effectiveness of the process have both improved, making it an
attractive alternative to conventional refractory treatment processes such as roasting and pressure oxi-
dation. The bacterial culture is robust and can withstand the normal fluctuations experienced on an
operating plant. The plants are simple to operate, making it ideal for remote areas. The scale-up poten-
tial of the process has been proven, and it can be applied to large refractory ore bodies. The process is
environmentally friendly. Neutralization of plant effluents produces precipitates that comply with the
most stringent environmental regulations.
INTRODUCTION
The year 1998 marked the tenth anniversary of the commercialization of the BIOX® Process
for the pretreatment of refractory sulfide ores. What began in the 1970s as a small research
project at GENCOR Process Research (now Billiton Process Research) has culminated over
the last decade in a mature, commercially viable process. During this period, five new plants
using the process have been commissioned.
In 1988, when GENCOR decided to market the technology on a licensing basis, the 10-t/day
(11-st/day) demonstration plant at Fairview Mine in South Africa was the only operating com-
mercial BIOX® plant. At that stage, it was operating in parallel with an old Edwards roasting
plant. The plant was extended in 1991 to treat the total concentrate production of the mine,
which amounted to 35 t/day (39 st/day). The roasters were decommissioned and demolished,
eliminating the threat of environmental pollution.
In 1990/1991, a second plant was built at the Sao Bento Mine in Brazil. The original plant
consisted of a single reactor operating in series with the existing pressure-oxidation circuit. A
second reactor was commissioned in 1994 with a third commissioned in January 1998.
* GFL Mining Services Ltd., Parktown, South Africa.
121
Harbour Lights in Western Australia was the first licensed plant to use the process. This
40-t/day (44-st/day) operation was commissioned in 1992 and ran until 1994, when the
mine was closed down.
The next plant was installed at Wiluna Mine in Western Australia in 1993. The original
plant had a capacity of 115 t/day (127 st/day) of concentrate, but it has since been expanded
to 158 t/day (174 st/day) by the addition of three new reactors.
The process took a quantum leap with the commissioning of Ashanti’s Sansu plant, with a
design capacity of 720 t/day (794 st/day) of concentrate in 1994. The original plant con-
sisted of three modules, but a fourth module was added in 1995 to increase the capacity to
960 t/day (1,058 st/day).
The 60 t/day (66 st/day) Tamboraque plant in Peru was recently added to the list of com-
mercial applications of the process.
The robustness, simplicity of operation, environmental friendliness and cost effectiveness
of the technology has been demonstrated at all of these operations. The BIOX® Process has
been a technical and economic success and offers real advantages over conventional refrac-
tory processes, such as roasting and pressure oxidation. Ongoing development work on
bench and pilot scales, as well as on the operating plants, is aimed at improving the efficiency
and cost effectiveness of the process even further.
License agreements have been signed with Amantaytau Gold Fields in Uzbekistan; TVX
Hellas in Greece, for their Olympias project at Stratoni; and Perseverance Exploration, for
their Fosterville operation in Victoria, Australia.
When the gold interests of GENCOR and Gold Fields of South Africa were merged in Feb-
ruary 1998 to form the new Gold Fields Ltd., the BIOX® Process and its holding company,
Biomin Technologies S.A., were transferred to the new company.
T H E B I O X® C U L T U R E
The process utilizes a mixed population of Thiobacillus ferrooxidans, Thiobacillus thiooxidans
and Leptospirillum ferrooxidans to break down the sulfide mineral matrix, thereby, liberating
the occluded gold for subsequent cyanidation.
Thiobacilli consist of 0.3- to 0.6-µm-diam, 1- to 3.5-µm-long straight rods (Figure 1). Lep-
tospirillum has similar dimensions but occurs as a vibroid when young and as a highly motile
spiral when mature. The bacteria are believed to attach themselves to the metal sulfide sur-
faces in the ore where they cause accelerated oxidation of the sulfides. The composition of
the population is influenced by factors such as temperature and pH.
The leptospirillum numbers are enhanced by a low pH and by a high slurry temperature
(Lawson, et al., 1991). Shake-flask tests conducted on T. ferrooxidans bacteria showed that
their oxidative activity was inhibited at a pH of less than 1.6. Optimum oxidation was
achieved in the pH range of 2 to 3. Tests with T. thiooxidans showed very little growth
between pH 0.5 and 1.0.
Because L. ferrooxidans can only oxidize reduced iron compounds and T. thiooxidans can
only oxidize sulfur compounds, it is important to control the pH and temperature within nar-
row ranges to maintain the right balance of bacterial species to optimize the rate of oxida-
tion. The typical operating pH range in the BIOX® process is 1.2 to 1.8. Lime, limestone, and/
or sulfuric acid are used to control the pH in the reactors.
The BIOX® culture operates best at a temperature of 40°C. However, it is possible to run at
45°C in the primary stage and even at 50°C in the final secondary reactors, as was proven
during pilot-scale test work on Fairview concentrate (Miller, 1991). The oxidation reactions
of sulfide minerals are exothermic. Therefore, it is necessary to cool the process to maintain
the slurry temperature within the optimum range.
122
FIGURE 1 Photomicrograph of Thiobacillus ferrooxidans bacteria
The oxidation reactions of the main sulfide minerals usually present in refractory ores may
be summarized as follows:
2FeAsS + 7O2 + H2SO4 + 2H2O → 2H3AsO4 + Fe2(SO4)3 (EQ 1)
4FeS2 + 15O2 + 2H2O → 2Fe2(SO4)3 + 2H2SO4 (EQ 2)
4FeS + 9O2 + 2H2O → 2Fe2(SO4)3 + 2H2SO4 (EQ 3)
FeS + Fe2(SO4)3 → 3FeSO4 + So (EQ 4)
FeS2 + Fe2(SO4)3 → 3FeSO4 + 2So (EQ 5)
2FeS + 2H2SO4 + O2 → 2FeSO4 + 2So + 2H2O (EQ 6)
4FeSO4 + 2H2SO4 + O2 → 2Fe2(SO4)3 + 2H2O (EQ 7)
2So + 3O2 + 2H2O → 2H2SO4 (EQ 8)

The oxidation reactions indicate the high oxygen demand of sulfide oxidation. Large vol-
umes of air have to be injected and dispersed into the slurry. This is one of the main engi-
neering challenges in the design of a full-scale bioreactor, as will be described below.
The bacteria require sufficient carbon dioxide to promote cellular growth. Carbon dioxide
is obtained from the carbonate minerals in the ore and from the air added to the process. If
no carbonate minerals are present, limestone must be added to the feed, as in the case of the
Tamboraque plant. Alternatively, the air injected must be enriched with carbon dioxide.
The bacteria also require nutrients to sustain growth. Nitrogen, phosphorous, and potas-
sium are the essential elements and are added to the primary reactors as a solution of ammo-
nium sulfate and potassium and phosphate salts. The amount of nutrients required depends
123
on the composition of the ore or concentrate treated. Ammonium sulfate is included in the
nutrient suite because there is no definite proof that all the bacteria in the culture can fix gas-
eous nitrogen in air.
Certain substances are potentially toxic to the bacteria. These include:
thiocyanate and cyanide at very low concentrations;
bactericides, fungicides and descaling reagents that are normally used for water treatment;
oil, grease and degreasing compounds;
chloride concentrations above 7 g/L, which inhibit bacterial activity (there are indica-
tions that the chloride causes membrane damage to the bacterial cells) (Lawson et al.,
1995); and
arsenic at high concentrations (although the culture is tolerant to arsenic(V) concentra-
tions as high as 20 g/L, a high concentration of arsenic(III) can be toxic) (Dew, 1995).
T H E B I O X® P R O C E S S
The basic flowsheet for the treatment of refractory-sulfide flotation concentrate is shown in
Figure 2.
Feed Preparation
Thickened flotation concentrate is normally remilled to 80% –75 µm and 95% –150 µm. A
finer grind may enhance the rate of biooxidation, but could have a detrimental effect on
down-stream processes such as solid-liquid separation. The remilled concentrate then goes
into a surge tank before it is fed to a splitter ahead of the reactors.
The slurry density of the feed to the splitter is controlled automatically at 20% solids by
adding dilution water into the suction line of the feed pump. The solids content of the slurry
has a definite effect on the oxygen mass-transfer rate. For flotation concentrates with a sul-
fide sulfur content of 20% to 30%, the required oxygen mass-transfer rate limits the solids
content to 20%. For concentrates or ores with a lower sulfide concentration, higher solids
concentrations may be tolerated.
Nutrient solution is added to the feed slurry at the splitter. Most of the operating plants
use a premixed nutrient blend, which is received in powder form and which is made up in
process water to 10% to 15% (w/w) solution strength.
Biooxidation
The biooxidation section consists of a number of equally sized tanks, configured as primary
reactors (typically three) and operated in parallel, and a number of secondary reactors con-
nected in series. With this configuration, a longer residence time is achieved in the primary
tanks, allowing a stable population to be established and preventing washout of the bacteria.
Between 50% and 70% of the sulfide sulfur is oxidized in the primary reactors. The sulfide
oxidation duty of these tanks is generally higher than that of the secondary tanks.
The overall residence time in the biooxidation, which is mainly a function of the mineral-
ogy, typically varies from four to six days. For an ore in which the gold is locked mainly in
arsenopyrite, one can expect a shorter residence time to achieve the optimum gold liberation
than with an ore in which most of the gold is occluded in pyrite. This is because the oxidation
rate of arsenopyrite is faster than that of pyrite.
With the high oxygen requirement of the oxidation reactions, the supply of air to the reac-
tors represents the largest consumer of power in the process. Low-pressure air, supplied by
blowers, is injected into the reactors and is dispersed by mechanical agitators specially
designed to provide an acceptable oxygen-transfer rate and oxygen utilization. Radial flow
impellers, such as the Rushton, were traditionally used for high gas dispersion.
124
FIGURE 2 Typical BIOX® flowsheet
Significant advances have been made in the development of axial-flow impellers for gas
dispersion over the last decade. These impellers have been tested and found to be more
power efficient with respect to oxygen utilization and oxygen mass-transfer rate. The higher
pumping rate of these impellers also provides better solids suspension. With the exception of
the Sao Bento plant, all of the plants are using Lightnin, Model A315, axial-flow impellers for
air dispersion in the reactors.
The oxidation of sulfide minerals is exothermic. To maintain the slurry temperature
between 40° and 45°C, the tanks have to be cooled. This is effected by circulating cooling
water through stainless steel coils and removing the heat in an evaporative cooling tower.
The efficiency of evaporative cooling depends on the local climatic conditions, mainly the
wet bulb temperature.
The mineralogy of the material treated determines whether the process is acid generat-
ing or acid consuming. The oxidation of pyrite produces sulfuric acid, whereas acid is con-
sumed during the oxidation of arsenopyrite and pyrrhotite. A high carbonate content of the
feed material can also increase acid consumption. The pH in the reactors is maintained
125
between 1.2 and 1.8 by adding concentrated sulfuric acid, lime, or limestone. In the case of
acid-consuming ores, acidic biooxidation products or solution may be recycled to reduce
the consumption of fresh acid. The preceding flotation process can also be used to depress
acid consuming minerals, such as dolomite and calcite, to minimize operating costs.
For an acid-generating material, local limestone, when available, can reduce the cost of
neutralization. Wiluna Mines, for instance, uses a locally mined calcrete as a neutralizing
agent in the plant (Stephenson et al., 1997).
Counter-Current Decantation
The oxidized product is washed in a counter-current decantation circuit prior to cyanidation.
High-rate thickeners are generally used in this application. The circuit is designed to give an
iron concentration of less than 1 g/L in the underflow from the final thickener, which is
pumped to the cyanidation circuit. Acidic overflow from the first thickener, containing dis-
solved iron, arsenic, and sulfates, is treated in a neutralization circuit.
Neutralization
Extensive research and development work has gone into the optimization of the design of the
neutralization section of the BIOX® process in recent years (Broadhurst, 1993). This was to
ensure that the effluents comply with the most stringent international environmental regula-
tions with respect to stability and residual arsenic concentrations. The most efficient neutral-
ization flowsheet was found to be a two-stage process effected in a series of agitated tanks,
usually six or seven. In the first stage, the pH is raised to between 4 and 5 with limestone or
lime to form ferric arsenate and gypsum according to the reactions
2H3AsO4 + Fe2(SO4)3 + 3CaCO3 → 2FeAsO4 + 3CaSO4 +CO2 + H2O (EQ 9)

and
Fe2(SO4)3 + 3CaCO3 + 3H2O → 2Fe(OH)3 + 3CaSO4 + 3CO2 (EQ 10)

In the second stage, the pH is further raised to 7 by adding milk of lime
H2SO4 + CaO → CaSO4 + H2O (EQ 11)

Pilot-plant tests, as well as operational results from the commercial plants, have shown that a
sufficiently stable arsenic precipitate is formed, provided that the molar ratio of iron to
arsenic is higher than 3:1. The precipitates subjected to the standard, as well as modified
toxicity characteristic leaching procedure (TCLP) of the USEPA, proved to be stable over a
wide range of pH values (5 to 9). Test work also indicated that continuous dehydration/
rehydration of the precipitates, simulating seasonal weathering on a tailings dam, resulted in
an increase in arsenic stability (Batty et al., 1995).
Incorporated in the neutralization circuit is a recycle pump that transfers the partially neu-
tralized effluent from the third or fourth tank back to the head of the tank train. The recycle
ratio is normally between 200% and 500% of the fresh feed. Recycling is employed to
improve the settling characteristics of the precipitates by providing seed for growing larger
crystals. It also improves the neutralization efficiency and reduces limestone consumption.
Air is sparged into the tanks to ensure complete oxidation of ferrous iron. At most of the
plants, the neutralized effluent is mixed with flotation tailings before dewatering in a dedi-
cated thickener. The overflow from the dewatering thickener is then recycled back to the
milling and biooxidation sections as dilution water. The underflow is deposited on a tailings
dam. The recycled solution has no detrimental effect on flotation performance or bacterial
activity, as proven at Fairview. However, it does help to conserve water, and it can be vital in
arid regions.
126
FIGURE 3 The Fairview plant in South Africa
It is not possible to reuse tailings-dam return water in the reactors and upstream sections
of the plant. This water is normally contaminated with cyanide and thiocyanates, which are
highly toxic to the culture.
COMMERCIAL PLANTS
The commercial BIOX® plants mentioned in the introduction have been described in a num-
ber of papers. A brief summary of these operations is presented below.
Fairview, South Africa

Much of the innovative research on biooxidation of refractory gold ores conducted at
GENCOR Process Research in the late 1970s and early 1980s was driven by the need to
replace the outmoded Edward’s roasters at Fairview, which, at the time, were seriously con-
tributing to atmospheric and water pollution in the sensitive Barberton area in South Africa
(Dew et al., 1993).
In 1984, a pilot plant was commissioned to treat flotation concentrate from Fairview. The
pilot plant operated for two years, and its success led to the installation of a demonstration
plant at Fairview. The demonstration plant was commissioned in October 1986 and was
designed to treat 10 t/day (11 st/day) of flotation concentrate at a water to solids ratio of 8:1
and with an overall residence time of four days. With ongoing development, the water solids
ratio of the feed was decreased to 4:1, and the feed rate increased to 17 t/day (19 st/day).
After running the plant for four years in parallel with the roasters, so much confidence was
gained in the process that it was decided to expand the plant to treat the full 35 t/day (39 st/day)
concentrate production from the mine (Figure 3). These expansions were commissioned in 1991,
after which the roasters were decommissioned. Another reactor was added to the circuit in 1994
to increase the plant capacity to 40 t/day (44 st/day) of concentrate.
127
TABLE 1 Summary of Fairview plant performance
Yearly average
Operating parameter 1988 1990 1991 1995 1996 1997
Concentrate treated, t/min 263 350 712 906 754 865
Concentrate gold grade, g/t 99 109 127 151 127 116
Concentrate S2– grade, % 27.4 23.1 22.9 18.0 16.8 14.3
Gold recovery in cyanidation, % 93.0 92.5 93.4 93.9 96.6 97.1
Plant availability, % 99 98 98 98 99 99
FIGURE 4 The Sao Bento metallurgical flowsheet
The Fairview plant played a vital role in the ongoing development of the process. The
scale of operation lends itself perfectly to the testing of new equipment, design modifica-
tions, and process optimization. It has also been used as a training ground for operators of
new BIOX® plants.
The mine was sold to Avgold Ltd.’s Eastern Transvaal Consolidated Division (ETC) at the
beginning of 1998. ETC also manages the Sheba, New Consort, and Agnes Mines in the Bar-
berton area. Until recently, refractory gold concentrate from these mines was treated in a flu-
osolids roaster at New Consort. The acquisition of Fairview offered ETC the opportunity to
utilize the more efficient BIOX® process for the treatment of its concentrate. The plant was
expanded to 55 t/day (61 st/day) of concentrate, and the New Consort roaster was decom-
missioned. Associated with this purchase was the conclusion of a licensing agreement,
whereby Avgold obtained the right to use the BIOX® technology for the treatment of its con-
centrate. At the same time, Gold Fields was granted free access to the site for development
work and for the training of operators.
128
TABLE 2 Summary of plant performance at Sao Bento
Current
Concentrate feed rate, t/day 27 42 51 75 105 120 80
Residence time, days 3.8 2.4 1.9 1.4 1.0 0.9 3.0
2–
Concentrate feed S grade, % 22.1 23.7 24.0 22.3 18.2 17.6 20.0
S2– oxidation, % 81.4 79.4 73.6 72.5 71.9 68.8 75.0
S2– oxidation based on design, % 57.8 94.1 107.3 144.4 163.7 173.0 142.6
FIGURE 5 The Harbour Lights plant in Western Australia
The performance of the Fairview plant over the past ten years is summarized in Table 1.
The plant has been a technical and commercial success, proving the simplicity of the process
and the robustness of the bacteria.
Sao Bento, Brazil

The Sao Bento mine in Brazil was commissioned in 1986 and employs a pressure oxidation
circuit to treat refractory flotation concentrates that typically contain pyrite (16%), arse-
nopyrite (38%), and pyrrhotite (46%). The pressure oxidation plant was originally designed
to treat 240 t/day (265 st/day) of concentrate. In 1990/1991, after extensive pilot-scale test
work at GENCOR Process Research, the capacity of the oxidation circuit was expanded with
the installation of a single 580-m3 (20,000 cu ft) reactor. The unit was designed to treat
150 t/day (165 st/day) of concentrate with a sulfide sulfur grade of 18.7% and to achieve
30% sulfur oxidation (Slabbert et al., 1992). The semioxidized product is blended with the
feed to the autoclaves (see Figure 4).
Although the design concentrate feed rate of the unit has not been achieved, the plant per-
formed well with respect to the tons of sulfide sulfur oxidized per day (see Table 2). Sulfide
oxidation in excess of 70% is achieved in the relatively short residence time of 1.5 days. This
is substantially higher than what was predicted from pilot-scale test work.
129
A second reactor was commissioned in December 1994. The feed rate to the plant is controlled
at a point where the oxygen addition to the autoclaves is optimized. In early 1998, a concentrate
stock tank was converted to a BIOX® reactor to increase the plant capacity even further.
The Sao Bento operation has proven the viability of combining biooxidation with pressure
oxidation. It is a cost-effective method of increasing the capacity of an existing pressure-
oxidation plant.
Harbour Lights, Australia

The Harbour Lights Mine, near Leonora in Western Australia, mined a refractory orebody,
with the gold associated mainly with arsenopyrite and pyrite. Sulfide flotation was used to
produce a concentrate containing 80 g/t (2.34 oz/st) gold, 18% sulfur, and 8% arsenic. The
concentrate was cyanide leached, and the residue, still containing 40 g/t (1.17 oz/st) gold,
was stockpiled.
The technology was licensed to Harbour Lights for the treatment of stockpiled and fresh con-
centrate. A process-design package was provided to an Australian engineering company for the
construction of a 40-t/day (44-st/day) plant (Figure 5). Construction started in June 1991, and
commissioning commenced at the end of 1991 (Barter et al., 1992). Full production was
achieved by June 1992. A performance test run conducted in October 1992 demonstrated that
the guaranteed gold recovery of 92% could be achieved at the design throughput rate.
The stockpiled concentrate, which was previously leached with cyanide, contained thiocy-
anate, which is toxic to the bacteria. The concentrate had to be washed, remilled, and thick-
ened prior to biooxidation.
The successful operation of the Harbour Lights plant again confirmed the commercial via-
bility of the technology. BIOX® plants are easy to operate (Harbour Lights used one operator
per shift), and the bacterial strain can adapt to new environments within a relatively short
period. Operations at Harbour Lights were terminated in 1994 when the orebody was
depleted. The plant was decommissioned at the end of that year.
The capital cost of the 40 t/day (44 st/day) Harbour Lights plant was Aus$4.5 million
(1991 dollars), and the operating cost over the life of the project was Aus$260/t of concen-
trate treated.
Wiluna, Australia
The Wiluna Gold Mine is located in the Northeastern Gold Fields Region of Western Austra-
lia, some 600 km (370 miles) north of Kalgoorlie. During the 1930s, Wiluna was one of the
largest gold producers in Australia when it treated high-grade refractory ore in a flotation/
roaster plant. Operations ceased in 1947 with the depletion of the high-grade ore reserves
known at the time. Production recommenced in 1984 with the treatment of flotation tailings
from earlier operations and later from oxide ore.
By 1990, the remaining reserves were largely refractory sulfides, and a decision on a suit-
able treatment process was required. During an extensive metallurgical test program, various
process options, such as roasting, pressure oxidation, and biooxidation, were evaluated. The
BIOX® process was finally selected based on better gold recoveries, lower capital and operat-
ing costs, and shorter permitting and construction periods. The process was also selected for
environmental reasons (Stephenson et al., 1997).
The Wiluna plant was successfully commissioned in 1993. The plant was originally
designed to treat 115 t/day (127 st/day) of flotation concentrate with a sulfur grade of 24%,
which is equivalent to 27 t/day (30 st/day) of sulfur. The plant consisted of six reactors, each
with a live volume of 470 m3 (16,600 cu ft). The overall residence time in the biooxidation
section was five days. During a performance test run conducted in late 1993, an average sul-
fide sulfur oxidation of 96.5% was achieved, whereas 93.6% was guaranteed.
In 1996, the capacity of the plant was increased to 158 t/day (174 st/day) of concentrate,
equivalent to 35 t/day (39 st/day) of sulfur, by the addition of three new reactors.
130
FIGURE 6 The Sansu plant in Ghana
The capital cost of the original 115 t/day (127 st/day) plant was Aus$9 million (1993 dol-
lars). The operating cost is Aus$70/t of concentrate treated, with the cost of power making
up 50% of the total. To reduce power costs, a natural gas fired power station has recently
been constructed to replace the diesel generators. Gas is supplied via the Goldfields gas
transmission system. Wiluna also has the advantage of having a cheap source of neutralizing
agent in the form of calcrete, which is mined on site.
Ashanti, Ghana
A major breakthrough came for the process when Ashanti Goldfields Co. (AGC) selected the
process for its Sansu Sulfide Treatment Plant near Obuasi in Ghana. This was after AGC
examined six refractory treatment options, including roasting, pressure oxidation, and nitric
acid leaching (Nicholson et al., 1993). BIOX® was again selected based on the reduced capi-
tal and operating costs, technical risk, environmental impact, and simplicity of operations.
Continuous pilot test work was performed at GENCOR Process Research on bulk concen-
trate samples from the primary sulfide and semioxidized ore bodies to determine the design
parameters for a full-scale plant. The widely different mineralogical and biooxidation charac-
teristic of the two ore types demanded a plant design with sufficient flexibility to treat both
concentrates individually or as a blend.
The original Sansu plant had a design capacity of 720 t/day (794 st/day) of concentrate
and consisted of three identical modules, with six 900 m3 (32,000 cu ft) reactors per module
(see Figure 6). The plant was successfully commissioned in February 1994, two months
ahead of schedule. The satisfactory performance of the plant (Table 3) led to the installation
of a fourth module in 1995. This module treats concentrate from the Pampora Treatment
Plant, thereby, reducing the throughput of the less efficient Edward’s roasting plant.
The Sansu plant is by far the largest biooxidation plant in the world. This installation has dem-
onstrated the scale-up potential of the process. With the modular design applied at Sansu, it is
possible to apply the technology to large refractory gold deposits. The simplicity of the process
makes it ideal for remote areas. The capital cost of the 960 t/day (1,060 st/day) plant amounted
to US$25 million (1994 dollars) and the operating cost is now US$66/t of concentrate.
131
TABLE 3 Sansu plant performance
Operating parameter Mar 1994 May 1994 Jun 1994 Jun 1995 Jun 1997 Jun 1998
Concentrate feed rate, t/day 307 532 586 745 880 1,007
Ave. S2– content of feed, % 10.5 7.5 7.4 7.5 10.3 9.1
2–
Ave. S oxidation, % 94.8 94.6 97.6 92.5 92.1 91.0
Ave. Au dissolution in CIP, % 91.9 94.2 94.2 92.0 92.2 91.5
TABLE 4 Design criteria for the Amantaytau, Olympias, and Fosterville plants
Amantaytau Olympias Fosterville
Concentrate feed rate, t/day 1,100 772 120
Concentrate S2– grade, % 25 36.3 17.8
Concentrate As grade, % 1.2 13.8 8.1
Solids concentration in feed, % 20 20 20
Number of BIOX® modules 4 3 1
Number of reactors per module 6 4 6
Total residence time, days 4 4 5
Reactor live volume, m3 974 1,328 531
Overall S2– oxidation, % 95 85 98
Tamboraque, Peru
The Tamboraque mine, owned by Minera Lizandro Proaño, is located in the Viso-Aruri min-
ing district of Peru, some 90 km (56 miles) east of Lima near the town of San Mateo. The
plant site is at an elevation of 3,000 m (9,800 ft) and is adjacent to the Rio Rimac, a fast flow-
ing perennial river that provides Lima with most of its water. The mine can easily be accessed
from Lima via Peru’s central highway and central railway.
Until 1996, the mine treated 200 t/day (220 st/day) of ore, mined from a vein deposit
containing galena, sphalerite, pyrite, and arsenopyrite, to produce lead and zinc concen-
trates. The tailings from the concentrator, containing 3 to 4 g/t (0.09 to 0.12 oz/st) gold and
34 g/t (0.99 oz/st) silver locked in arsenopyrite, was stockpiled. The mine has embarked on
an expansion program involving the construction of a new 600 t/day (660 st/day) concentra-
tor and a 60 t/day (66 st/day) BIOX® plant for the treatment of stockpiled tailings and, later,
for the treatment of run-of-mine ore.
Proaño experimented with bacterial oxidation since the late 1980s and later constructed a
50-kg/day (110 lb/day) pilot plant at the TECSUP Institute in Lima for the biooxidation of
arsenopyrite concentrate from Tamboraque. The bacterial culture used in this test work was
isolated from mine water. In 1994, GENCOR Process Research was contracted to assist with
the optimization of the pilot plant in order to collect sufficient data for the design of a full-
scale plant. Pilot-plant results indicated that a sulfur oxidation of less than 85% was required
to yield a gold recovery of 92%.
A license agreement was concluded between BIOMIN and Proaño, and BIOMIN provided a
process-design package for a 60-t/day (66-st/day) plant, which was commissioned at the end
of 1998. An interesting feature of the plant is the very high arsenic content (26%) of the feed
material (Table 4). With 56% arsenopyrite in the feed, the process is acid consuming. Acid
mine drainage from old mine workings is used as dilution water to the plant. Not only does it
alleviate a pollution problem, but it also provides additional iron to the process, which assists in
the production of a stable ferric arsenic precipitate during neutralization. Due to the low car-
bonate content of the concentrate, limestone is added to the feed as a source of the carbon
dioxide required for bacterial growth. The capital and operating costs for this 60-t/day (66-st/
day) plant are estimated at US$3 million and US$71 per metric ton of concentrate respectively.
132
FUTURE PLANTS
To date, more than 150 concentrate and whole ore samples have been tested for amenability
to the BIOX® process. Over 15 integrated pilot-plant runs have been completed to obtain
information for the design of commercial plants. Several projects are at the feasibility stage
or are awaiting financial backing.
License agreements have already been signed with Amantaytau Gold Fields in Uzbekistan,
TVX Hellas in Greece, and Perseverance Exploration in Victoria, Australia.
Amantaytau Goldfields Project, Uzbekistan

The Amantaytau Goldfields comprise a number of gold deposits, mainly refractory, situated
in the Navoi region of Uzbekistan in Central Asia. A joint venture was established between
Lonrho Plc., the International Finance Corp. (IFC), the Uzbek State Committee of Geology,
and the Navoi Mining and Metallurgical Combinat for the development of these deposits
with combined reserves of more than 160 t (5 million oz) of gold.
After extensive metallurgical research involving laboratory and pilot-scale testing by
GENCOR Process Research in South Africa and on-site testing at the Ingichki Research Insti-
tute (Batty et al., 1995), the BIOX® process was selected as the preferred route for the treat-
ment of the refractory sulfide ore. A feasibility study was completed by Lonrho in 1995. From
geological and mining modeling, the operations were optimized to a mill throughput rate of
3.5 Mt/a (3.9 million stpy), with oxide ore being mined for the first three years and sulfide
ore for the remaining 14 years of the mine’s life.
GENCOR provided the process-design package for the oxide plant, as well as the sulfide
treatment plant, which included a 1,100-t/day (1,200-stpd) BIOX® plant. The main process
design criteria are summarized in Table 4.
However, in May 1997, Lonrho announced that it would not proceed with the develop-
ment of the mine. The low price of gold, increases in the cost of mining supplies, and local
factors were given as reasons for Lonrho’s decision.
Olympias Project, Greece

TVX Gold’s Kassandra mining complex is located on the Chalkidiki Peninsula in northeastern
Greece, 100 km (60 miles) east of Thessaloniki. It includes the Olympias mine and mill, the
Skouries copper-gold porphyry deposit, the Madem Lakkos and Mavres Petres mines, a base
metal milling facility at Stratoni, and a number of exploration facilities. The Olympias mine
began operations in 1976, producing silver, lead, and zinc. More than 130 t (4 million oz) of
gold were left in the flotation tailings, due to the refractory nature of the ore. TVX Gold,
through its wholly owned subsidiary, TVX Hellas, has recently completed a feasibility study
on the construction of a gold recovery plant to treat old and current tailings. A combined
BIOX®/pressure oxidation circuit was selected. Up to 85% of the sulfide sulfur will be oxi-
dized by biooxidation, after which the product will be pressure leached to complete the oxi-
dation process. The feasibility study was based on a process design package provided by
BIOMIN for a 772-t/day (850-stpd) plant. The main design criteria for this plant are shown
in Table 4. The sulfide sulfur oxidation duty of this plant will be more than double that of the
Sansu plant. The reactors will also be significantly larger.
Fosterville, Australia
Perseverance Exploration’s Fosterville Mine near Bendigo in Victoria, Australia, is an open-
pit, heap-leach operation for oxide ore. The sulfide ore reserves are now estimated at 3.54 Mt
(3.90 million st) at 3.27 g/t (0.095 oz/st) gold. The sulfide ore is highly refractory and direct
cyanidation yields recoveries lower than 2%. The gold is locked in pyrite and, to a lesser
degree, in arsenopyrite, and the ore contains preg-robbing carbonaceous minerals.
133
Test work has shown that it is possible to depress the organic carbon during flotation and
to obtain more than 93% gold recovery after biooxidation of the flotation concentrate. Perse-
verance completed a feasibility study based on a 120-t/day (132-stpd) plant (Table 4). This
project is also in the financing stage.
REFERENCES
Barter, J., Carter, A.J., Holder, N.H.M., Miller, D.M., and Van Aswegen, P.C., 1992, “Design and
commissioning of a 40 tonne per day flotation concentrate biooxidation treatment plant at the
Harbour Lights Mine,” Extractive Metallurgy of Gold and Base Metals Conference, Kalgoorlie,
Australia, October.
Batty, J.dK., Dew, D.W., Stallknecht, H., and Marais, H.J., 1995, “The Amantaytau Refractory
Gold Project,” Randol Gold Forum, Perth, Australia, March 1995.
Broadhurst, J.L., 1993, “The Nature and stability of arsenic residues from the BIOX® process,”
Biomine ’93 Conference, Adelaide, Australia, March 1993.
Dew, D.W., 1995, “Comparison of performance for continuous biooxidation of refractory gold ore
flotation concentrates,” in Biohydrometallurgical Processing, Volume 1, T. Vargas et al., ed.,
University of Chile Press, Santiago.
Dew, D.W., Miller D.M., and Van Aswegen, P.C., 1993, “GENMIN’s commercialization of the
bacterial oxidation process for the treatment of refractory gold concentrates,” Randol Gold
Forum, Beaver Creek, CO, Sept. 1993.
Lawson, E.N., 1991, “Environmental Influences on BIOX®,” Bacterial Oxidation Colloquium,
SAIMM, Johannesburg.
Lawson, E.N., Nicolas, C.S., and Pellat, H., 1995, “The toxic effect of chloride ions on Thiobacillus
ferrooxidans,” in: Biohydrometallurgical Processing, Vol. 1, T. Vargas et al., ed., University of
Chile Press, Santiago.
Miller, D.M., 1991, Effect of temperature on BIOX® operations,” Bacterial Oxidation Colloquium,
SAIMM, Johannesburg, South Africa, 1991.
Nicholson, H.M., Lunt, D.J., Ritchie, I.C., and Marais, H.J., 1993, “The design of the Sansu
Concentrator and BIOX® Facility, Biomine ’93 Conference, Adelaide, Australia, March.
Slabbert, W., Dew, D.W., Godfrey, M.W., Miller, D.M., and Van Aswegen, P.C., 1992,
“Commissioning of a BIOX® module at São Bento Mineração,” Randol Gold Forum, Vancouver,
BC, Canada, March.
Stephenson, D., and Kelson, R., 1997, “Wiluna BIOX® Plant—Expansion and new developments,”
Biomine ’97 Conference, Sydney, Australia, August.
van Aswegen, P.C., 1993, “Biooxidation of refractory gold ores—The GENMIN experience,”
Biomine ’93 Conference, Adelaide, Australia, March.
van Aswegen, P.C., 1998, “The BIOX® process ten years after commercialization,” Au ’98, 11th
International Gold Symposium, Rio de Janeiro, Sept.
134
SECTION 3
Bioremediation
Degradation of Metal Cyanide Complexes by Microorganisms 137
Biochemical Removal of HAP Precursors from Coal—INEEL Slurry

Column Testing 153
Microorganisms, Biotechnology, and Acid Rock Drainage—Emphasis on

Passive-Biological Control and Treatment Methods 169
Sawdust-Supported Passive Bioremediation of Western United States Acid

Rock Drainage in Engineered Wetland Systems 189
Biotechnologies for Remediation and Pollution Control in the

Mining Industry 207
135
Degradation of Metal Cyanide
Complexes by Microorganisms
R. Fedel-Moen,* S.R. Ragusa,* R.W.L. Kimber,* and B.D. Williams†
ABSTRACT
The use of cyanide to extract gold from ores is widespread. In the process of extracting gold, cyanide
may form metal cyanide complexes with a variety of other metals present in the ore. This study was
carried out to investigate the biodegradation of the Cu(I)-cyanide complex and tetracyanonickelate
under varying pH and nutrient additions. Detailed studies using shake-flask cultures and bioreactors
showed that naturally occurring bacteria residing on gold-leach-pad ore utilized the Cu(I)-cyanide
complex (CuCN) and tetracyano-nickelate, Na 2(NiCN4), as carbon and/or nitrogen sources. Ammo-
nium-nitrogen (NH4+-N) and/or cyanate (CNO–) accumulated in solution as the metal cyanide com-
plexes were degraded. The accumulation of cyanate as a product is a concern due to its toxicity.
Degradation rates were increased if a growth substrate (peptone, 0.1%) was added to the medium or
if bioreactors were primed with peptone before addition of the metal-cyanide complex. Tetracyano-
nickelate was degraded at pH 8 without added peptone, but it was not degraded at pH 10. The addi-
tion of peptone was required for degradation of the Cu(I)-cyanide complex at pH 8 and pH 10. At pH
8, the addition of peptone resulted in the accumulation of cyanate, whereas, at pH 10, cyanate was
not detected in the cultures. Biodegradation of metal-cyanide complexes was faster (approximately
1.5 to ten times faster) in bioreactors than in shake-flask cultures. The addition of peptone was not
required for degradation of metal-cyanide complexes in bioreactors. However, priming the bioreactor
with peptone resulted in higher metal-cyanide degradation rates over multiple cycles.
INTRODUCTION
Cyanidation of gold with sodium cyanide is a widely used and well-documented process for
the extraction of gold from ores (Scott and Ingles, 1981). However, many other metals
present in ores also react with the cyanide anion to form metal-cyanide complexes that
remain in the effluents (Scott and Ingles, 1981; Byerley et al., 1988). Two of the most com-
mon complexes found in gold-mine tailings are copper and nickel cyanides, both of which are
classified as moderately strong metal-cyanide complexes (Hoecker and Muir, 1987). Bacte-
rial degradation of the Cu(I)-cyanide complex has rarely been reported. Bacterial degrada-
tion of the nickel-cyanide (as tetracyanonickelate) complex has been reported by a number
* CSIRO Land and Water, Glen Osmond, Australia.

† Environmental Science and Management Department, University of Adelaide, Australia.
137
of authors (Rollinson et al., 1987; Silva-Avalos et al., 1990). Pseudomonas fluorescens strain
NCIMB11764 was shown to utilize cyanide as a source of nitrogen when glucose was sup-
plied as the carbon source (Rollinson et al., 1987). It was also shown that tetracyano-
nickelate degradation was associated with the induction of cyanide oxygenase enzyme
activity, which was not induced in the presence of ammonia (Rollinson et al., 1987).
In another study, Pseudomonas putida BCN3 (Silva-Avalos et al., 1990) degraded potas-
sium tetracyano-nickelate(II), K2(NiCN4), and used it as a sole nitrogen source in the pres-
ence of glucose. The study of Silva-Avalos et al. (1990) showed that the degradation of
K2(NiCN4) led to the formation of NiCN2, which only occurred when ammonia was absent.
Shpak et al. (1995) used a strain of Pseudomonas fluorescens to degrade a range of different
metal-cyanide complexes. They concluded that a possible mechanism for degradation of
metal-cyanide complexes may be a shift in the chemical equilibrium in the dissociation of
complex anions, which takes place at the expense of the utilization of cyanide ligands by the
degrading strain (Shpak et al., 1995).
Cyanide-degrading microorganisms have been used for the remediation of cyanide-
contaminated mine effluents. The Homestake Mine (USA) Biological Decyanidation Process
was designed to use various species of Pseudomonas bacteria (Mudder et al., 1985). The pro-
cess operates at low levels (0.5 to 5.0 mg/L) of weak-acid dissociable cyanide (WAD-CN) in
the presence of phosphoric acid. The main source of bacteria used in establishing such degra-
dation processes have come from either soil, sewage, or, in the case of the Homestake Mine,
the bacteria were isolated from process solutions. Previous degradation studies have usually
been carried out in the presence of a nutrient source, such as glucose.
Reactions between cyanide and metals occur readily but are dependent on the concentra-
tion of metal, level of cyanide, pH, redox potential and radiation energy (light). The cyanide
ion, CN–, is a ligand that allows many reactions to occur, and metal complexes with many
coordination groups can result (Dickerson et al., 1984).
Cuprous cyanide is formed when cyanide and Cu2+ ions are present together and results in
the precipitation of cupric cyanide CuCN2, which readily decomposes to CuCN and cyanogen
(C2N2). In the presence of excess free cyanide, the CuCN dissolves to form further cyanide
complexes (Ford-Smith, 1964; Sharpe, 1976).
2Cu2+(aq) + 4CN–(aq) → 2CuCN(s) + C2N2 (EQ 1)
Copper (I) cyanide complex (CuCN) is insoluble in water but is considered very stable.
The Cu(I)/Cu(II) equilibrium can easily be made to proceed in either direction depending on
the anions present in the aqueous solution.
The Cu(I)-cyanide complex (CuCN) is soluble in excess cyanide to give the ion Cu(CN)2–,
with the progressive formation of Cu(CN)32–and Cu(CN)43– (Cotton and Wilkinson, 1988)
and is soluble in low pH solutions (Sharpe, 1976). It is found that, at pH 8 to 14 and Eh of
+0.5 to –0.5 volts, the main copper ion in solution is Cu(CN)32–(Osseo-Asare et al., 1984).
Cu+ + 2CN– → (Cu(CN)2)– (EQ 2)
Cu(CN)2– + CN– → (Cu(CN)3)2– (EQ 3)
Cu(CN)32– + CN– → (Cu(CN)4)3– (EQ 4)

The chemical reactions in the gold tailings are not fully understood nor have they been
extensively studied. With the presence of cyanide, sodium, copper, other transition metals,
cyanate, ammonia, and thiocyanate, interactions between many are expected. Some interac-
tions are known, but many are not known.
The addition of stoichiometric quantities of sodium or potassium cyanide to a solution of a
nickel (II) salt will form sodium tetracyanonickelate (Ford-Smith, 1964; Sharpe, 1976).
4NaCN + NiSO4 → Na2Ni(CN)4 + Na2SO4 (EQ 5)
138
The complex Ni(CN)42– is destroyed by strong mineral acids to produce Ni(CN)2. While in a
solution of excess cyanide there may be the formation of the deep red Ni(CN)53– anion.
The tetracyanonickelate ion (Ni[CN]42–) is one of the most stable complexes of nickel, but
it is decomposed by acids with the formation of hydrated NiCN2.
For both nickel and copper, reactions with the cyanide ligand are numerous, producing
heavy-metal cyanides, alkali metal/heavy-metal cyanides and heavy-metal/heavy-metal cya-
nides. In addition, complexes that include ammonia or sulfur are formed. It is also highly
likely that compounds like cyanate and other byproducts present in the gold tailings material
interact with nickel cyanate, but they have not yet been studied.
In this paper, the authors report the degradation of the Cu(I)-cyanide complex and tetra-
cyanonickelate by naturally occurring bacteria present on gold ore material from leach pads.
Degradation was determined with and without the addition of an organic nutrient (pep-
tone). The concentrations of metal-cyanide complexes used in this study were reflective of
values frequently found in gold-mine tailing dams.

Leach-pad ore was obtained from a gold mine in Queensland, Australia. The mine used heap
leaching as a method of extracting gold from ores. The leach-pad ore was stored at room
temperature until required.
Media and Stock Solutions

Modified Davis mineral salts (DMS) medium (Davis et al., 1959) containing KH2PO4 (0.2 g/
L), Na2HPO4·2H2O (0.3 g/L), KCl (1.0 g/L), MgSO4·7H2O (0.2 g/L) and CaCl2·2H2O (0.01 g/
L) dissolved in distilled water. Where necessary, peptone was added as a growth substrate at
a concentration of 1 g/L, and, for the isolation of bacteria, Bacto agar (15.0 g/L) was added.
The pH of DMS was adjusted to pH 8 or pH 10 using NaOH. Tetracyanonickelate was pre-
pared by adding 119 mg (0.45 mM) nickel sulfate (Analar) and 98 mg (2 mM) sodium
cyanide (Merck) per liter of deionized water. Solutions were sterilized by filtration using a
0.45-mm pore size membrane filter and added to autoclaved DMS medium.
Degradation of Cu(I)-Cyanide and Tetracyanonickelate Complexes

Investigations into the microbial degradation of the Cu(I)-cyanide complex and tetracyano-
nickelate were performed using sterile and nonsterile ore. Degradation experiments were
conducted in both shake flasks and bioreactors.
Shake Flasks
Shake-flask cultures (250-mL Erlenmeyer flasks) contained DMS medium (100 mL; pH 8
and 10), with or without the addition of peptone (0.1%). After the addition of the Cu(I)-
cyanide complex or tetracyanonickelate (final concentration of 150 and 95 mg/L, respec-
tively), flasks were inoculated with ore (0.5% w/v). Controls consisted of the respective
media with sterile ore (autoclaved) added. Sterile ore was added to determine whether
abiotic adsorption or degradation reactions occurred between the metal-cyanide com-
plexes and the ore itself. Incubations were performed in duplicate for each set of culture
conditions. Flasks were shaken slowly (50 rpm) and incubated at 25°C. For each analysis,
6 mL (flasks) or 10 mL (bioreactor) of supernatant solution was filtered through a 0.45-µm
pore size filter. The filtrate was used for quantitative chemical analysis, and the pH was
measured using a Corning pH meter with a silver/silver chloride combination electrode.
Ammonium-nitrogen was determined by the Berthelot color reaction (Stainton et al.,
1977) using a LKB Biochrom Ultraspec 3 UV/visible spectrophotometer. The degradation
139
of cyanide complexes was monitored over time. Microbial numbers were determined by
inoculating agar plates containing Davis mineral salts plus metal cyanide and/or peptone.
Bioreactors
The degradation of cyanide complexes in bench-scale bioreactors was performed in BELLCO
Spinner Flasks containing 1 L of DMS medium. Bioreactors were inoculated with 10% (w/v)
leach-pad gold ore and were held at room temperature. For the primed bioreactor (Bioreac-
tor B), peptone (0.1%, w/v) was added to the DMS and heterotrophic bacteria were allowed
to grow for 20 days before the addition of a metal cyanide. All subsequent metal-cyanide
degradation cycles were carried out without further addition of peptone. The Cu(I)-cyanide
complex (140 to 150 mg/L) was added to each bioreactor and the degradation monitored.
Once degradation was complete, the ore was allowed to settle, the supernatant was
removed, and the medium was replaced with fresh DMS spiked with CuCN (150 mg/L). After
two degradation cycles with the Cu(I)-cyanide complex, the medium was replaced with fresh
DMS spiked with tetracyanonickelate (Na2NiCN4) at a level of 95 to 104 mg/L. Samples
(10 mL) were taken at set intervals over the incubation period and the pH, metal cyanide and
ammonium-nitrogen concentrations determined.
Determinations of Free Cyanide, Cyanate, and Weak-Acid Dissociable Cyanide

Samples were filtered through a 0.45-µm syringe tip filter and the pH, free cyanide, cyanate,
and weak-acid dissociable cyanide (WAD-CN) concentrations were determined. Free cyanide
was determined according to the procedure described in Standard Methods for the Examina-
tion of Water and Wastewater (Greenberg et al., 1992). Briefly, filtrates were titrated with sil-
ver nitrate (0.0031 M) after the addition of p-dimethylaminobenzylidene rhodanine. Total
cyanide and WAD cyanide were analyzed using an Alpkem Flow Solution 3000 Auto Ana-
lyzer (Perstorp Analytical Environmental) according to the procedure described in Standard
Methods for the Examination of Water and Wastewater (Greenberg et al., 1992). Cyanate
(CNO–) was converted to ammonia and analyzed using an ammonia ion selective electrode
supplied by Greenberg et al. (1992).
HPLC Determination of the Cu(I)-Cyanide and Tetracyanonickelate Complexes

For HPLC analysis, aliquots (0.5 mL) of copper cyanide-containing media were added to 0.5
mL of sodium cyanide (NaCN) solution (200 mg/L) to help stabilize the CuCN during analy-
sis. No addition of sodium cyanide solution was necessary for the tetracyanonickelate deter-
minations. Quantitative determinations of metal-cyanide complexes were performed using a
GBC HPLC equipped with a UV detector and a NOVA-PAK C18 3.9-mm × 150-mm column.
For the detection of the Cu(I)-cyanide complex, a wavelength of 234 nm was used to avoid
confusion from peptone absorbance at lower wavelengths. The mobile phase consisted of
acetonitrile (25% in water) and tetrabutyl ammonium sulfate (0.005 M) at a flow rate of 1.0
mL/min (Hilton and Haddad, 1986). Tetracyanonickelate was detected at 267 nm, using an
eluent of acetonitrile (18% in water) and tetrabutyl ammonium sulfate (0.005 M), at a flow-
rate of 1.0 mL/min. The Cu(I)-cyanide complex and tetracyano-nickelate concentrations
were calculated against calibration curves. As a qualitative check of metal-cyanide complexes
remaining in solution, filtrates were scanned in the range between 210 to 600 nm using a
UV/VIS spectrophotometer, which quickly gave an indication of the level of metal-cyanide
complex remaining in solution.
140
Treatment:
○ –P–B: minus peptone, minus bacteria
+P–B: plus peptone, minus bacteria
–P+B: minus peptone, plus bacteria
+P+B: plus peptone, plus bacteria
FIGURE 1 Biodegradation of Cu(I)-cyanide complex in shake-flask cultures at pH 8
TABLE 1 Cu(I)-cyanide complex degradation products (mg/L) for shake-flask cultures incubated
at pH 8 and pH 10 for 120 days
pH 8 pH 10
Treatment WAD-CN, mg/L CNO, mg/L WAD-CN, mg/L CNO, mg/L
–P –B 34.6 3.4 45.5 8.6
+P –B 34.2 1.2 26.9 10.4
–P +B 40.9 2.8 36.2 6.3
+P +B <0.1 33.0 <0.1 1.8
WAD-CN = weak acid dissociable cyanide

CNO = cyanate
Treatment: –P–B = minus peptone, minus bacteria
+P–B = plus peptone, minus bacteria
–P+B = minus peptone, plus bacteria
+P+B = plus peptone, plus bacteria
Cu(I)-Cyanide Complex Degradation in Shake-Flask Cultures

Chemical and microbiological analyses indicated that breakdown of the added complexes
was occurring. Biodegradation of the Cu(I)-cyanide complex was confirmed by the reduction
in WAD-CN concentration over the incubation period and the detection of cyanate (Table 1).
In no case was free cyanide detected in cultures.
At pH 8, degradation of the Cu(I)-cyanide complex did not occur in sterile controls or in inoc-
ulated cultures in the absence of peptone (Figure 1). After 120 days of incubation, WAD-CN
concentrations were greater than 30 mg/L, and cyanate concentrations were less than
3.4 mg/L, indicating little activity toward the cyanide complex (Table 1). However, in the pres-
ence of peptone, biodegradation of the Cu(I)-cyanide complex commenced after seven days
with a degradation rate of 0.62 mg/L/day. WAD-CN concentrations were below detection lim-
its (0.1 mg/L), and cyanate concentrations reached a maximum concentration of 33 mg/L after
120 days (Table 1). An increase in pH of approximately 1.0 pH unit was observed in cultures
exhibiting a reduction in the Cu(I)-cyanide complex concentration. The increases in pH were
141
Treatment:
+P+B: minus peptone, plus bacteria
FIGURE 2 Biodegradation of Cu(I)-cyanide complex in shake-flask cultures at pH 10
attributed to the formation of ammonia by the microorganisms from degradation of peptone

and/or metal cyanide. These results indicate that the indigenous ore bacteria actively oxidized
CuCN to cyanate (CNO–) if peptone was supplied. An enzyme, cyanide monoxygenase, is
known to convert cyanide to cyanate. Under ideal conditions, a second enzyme (cyanase) acts
on cyanate and converts cyanate to ammonia and carbon dioxide (Dubey and Holmes, 1995).
In the study reported in this paper, the cyanase enzyme was either inhibited by the pH 8 envi-
ronment or its formation was repressed, possibly by ammonia. The formation of cyanate is a
concern because cyanate is a toxic byproduct, although not as toxic as cyanide.
Although shake-flask experiments were also conducted at pH 10, interaction between the
CuCN medium and ore resulted in a reduction in the pH of the medium to approximately
pH 8.5 after one week. Sterile controls and inoculated cultures without peptone supplementa-
tion showed no degradation of the Cu(I)-cyanide complex. However, with peptone present,
biodegradation of the Cu(I)-cyanide complex did occur (Figure 2). The rate of degradation
was slightly faster at the higher pH 10 (0.74 mg/L/day). In this case, cyanate (CNO–) did not
accumulate (Table 1) in the medium, indicating that other end products, such as ammonia
(NH3) and carbon dioxide, formamide or formic acid, may have formed. Alternately, cyanate
may have been produced and then degraded to ammonia (NH3) and carbon dioxide (CO2) by
the cyanase enzyme. In this instance, it is not known whether cyanase was responsible for
degradation of cyanate or whether the cyanide was converted directly to ammonia and carbon
dioxide by an alternative enzyme pathway, which utilizes the cyanide dioxygenase enzyme.
Tetracyanonickelate Degradation in Shake-Flask Cultures

Shake-flask experiments were also prepared to evaluate the ability of the indigenous ore bac-
teria to degrade tetracyanonickelate. Similarly, degradation experiments were prepared at
pH 8 and 10.
At pH 8, tetracyanonickelate biodegradation occurred in the presence and absence of pep-
tone (Figure 3). The rate of cyanide degradation was, however, higher in the presence of
peptone, 2.35 mg/L/day compared to 0.92 mg/L/day without peptone. With the addition of
peptone, degradation of tetracyanonickelate was complete by day 20, whereas it took 55
days for complete degradation in its absence. This may be reflective of the added carbon
142
Treatment:
FIGURE 3 Biodegradation of tetracyanonickelate in shake-flask cultures at pH 8
Treatment:
FIGURE 4 Biodegradation of tetracyanonickelate in shake-flask cultures at pH 10
source (peptone) or increased bacterial number. Degradation of tetracyanonickelate did not

occur under sterile conditions (Figure 3).
No degradation of tetracyanonickelate was evident when peptone was omitted or under
sterile conditions at pH 10 (Figure 4). Degradation of tetracyanonickelate was only observed
when peptone was supplemented into inoculated cultures, with a degradation rate of
0.66 mg/L/day.
Because of tetracyanonickelate degradation, ammonium-nitrogen concentration increased
over the incubation period at pH 8 (Figure 5). In the presence of peptone, ammonia was
formed from peptone and by cyanide degradation with ammonium-nitrogen concentrations
reaching a maximum value of 18 mg/L within 41 days (Figure 5). In the absence of peptone,
ammonia concentrations increased slowly and then rapidly over the incubation period, and by
143
Treatment:
FIGURE 5 Ammonia-nitrogen formation during tetracyanonickelate biodegradation at pH 8
Treatment:
FIGURE 6 Ammonia-nitrogen formation during tetracyanonickelate biodegradation at pH 10
Day 80, the ammonia concentration had reached 12.1 mg/L (Figure 5). Ammonia concentra-
tions were highest after complete degradation of cyanide. WAD-CN concentrations were
reduced to below detection limits, however, the accumulation of cyanate varied among treat-
ments (Table 2). Cyanate (CNO–) was formed at pH 8 in the presence of peptone (Table 2) and
not formed in its absence. This finding indicates either that a different degradation pathway
operated in the absence of peptone or that the formation of end products such as carbon diox-
ide and ammonia had occurred. In the absence of peptone, either the cyanase enzyme was
active due to lower concentrations of ammonia, or an alternative pathway that utilized the cya-
nide dioxygenase enzyme was operational and formed ammonia and carbon dioxide directly.
144
TABLE 2 Tetracyanonickelate degradation products (mg/L) for shake-flask cultures incubated at
pH 8 and pH 10 for 77 days
pH 8 pH 10
Total CN, WAD-CN, CNO, Total CN, WAD-CN, CNO,
Treatment mg/L mg/L mg/L mg/L mg/L mg/L
–P–B 15.4 13.4 0.45 16.3 14.2 0.5
+P–B 13.5 13.2 7.2 15.2 13.9 1.5
–P+B <0.1 <0.1 0.45 13.8 13.8 <0.3
+P+B <0.02 <0.02 12.5 0.45 0.6 <0.3
WAD-CN = weak acid dissociable cyanide

CNO = cyanate
Treatment: –P–B = minus peptone, minus bacteria
+P–B = plus peptone, minus bacteria
–P+B = minus peptone, plus bacteria
+P+B = plus peptone, plus bacteria
In the pH 10 cultures where tetracyanonickelate degradation occurred, the accumulation

of ammonium-nitrogen was observed (Figure 6). The concentration of ammonia was lower
at pH 10 (5 mg/L at day 60) than it was at pH 8 (18 mg/L at day 41) This was due to the
slower degradation rate of peptone and of tetracyanonickelate at the higher pH. However, at
pH 10, ammonium-nitrogen concentrations reached a maximum concentration of 18 mg/L
after 80 days (Figure 6), which was similar to maximum concentrations observed at pH 8.
Although WAD-CN concentrations were reduced at pH 10 in the presence of peptone,
cyanate did not accumulate in the medium (Table 2). Presumably, degradation of tetra-
cyanonickelate at elevated pH results in the formation of different end products, compared
to those formed in the presence of peptone at pH 8. Possibly, the enzyme responsible for
cyanate degradation (i.e., cyanase) was not produced by bacteria at pH 8 due to the higher
ammonia concentrations or lower pH environments, causing its repression or inhibition.
Alternately, the degradation of tetracyanonickelate at pH 10 may be occurring by a pathway
where cyanate is not produced. This contrasts with the work of Silva-Avalos (1990), who
found that tetracyanonickelate is converted to nickel cyanide (NiCN2). The work reported
here shows that, at pH 8 in the presence of peptone, cyanate was formed. However, without
peptone, cyanate was not found. The authors have also shown that, under certain conditions,
tetracyanonickelate is degraded completely.
Cu(I)-Cyanide Complex Degradation in Bioreactors

The results described in the previous section indicated that the indigenous microorganisms
found on leach-pad ore were able to degrade the Cu(I)-cyanide complex in shake-flask cul-
tures if peptone was supplied. A bioreactor system was designed to determine whether
metal-cyanide biodegradation processes could be scaled up. In this system, feed batch cul-
tures were used. Bioreactor A was operated at pH 9 without added peptone, whereas Biore-
actor B, also at pH 9, was “primed” with peptone before the start of the experiment. In
Bioreactor B, heterotrophic bacteria were allowed to grow for 20 days in the presence of pep-
tone (priming) before the addition of the Cu(I)-cyanide complex. Once the bioreactor was
primed, peptone was not added again. The degradation of the Cu(I)-cyanide complex in the
bioreactors was followed over two degradation cycles.
Degradation of the Cu(I)-cyanide complex during the first cycle was observed in both
bioreactors (primed and unprimed), which was surprising given that the cyanide complex
was not degraded in the absence of peptone in shake-flask experiments. In Bioreactor A, deg-
radation proceeded after a lag period of seven days and was completed by Day 28, resulting
in a degradation rate of 1.6 mg/L/day (Figure 7). In Bioreactor B, which was primed, the lag
period before degradation commenced was reduced to three days, and degradation was com-
pleted by Day 10 at a degradation rate of 8.7 mg/L/day (Figure 7).
145
FIGURE 7 Degradation of Cu(I)-cyanide complex in Bioreactor A ( ) and Bioreactor B () over
28 days (unprimed and primed respectively) Cycle I
FIGURE 8 Degradation of Cu(I)-cyanide complex in Bioreactor A ( ) and Bioreactor B () over

33 days (unprimed and primed respectively), Cycle II
During the second degradation cycle in Bioreactor A, there was a longer lag period before
the Cu(I)-cyanide complex degradation commenced. Degradation proceeded after 17 days
and continued until all of the Cu(I)-cyanide complex had been degraded at Day 33
(Figure 8). The rate of cyanide degradation increased in the second cycle for Bioreactor A,
resulting in a degradation rate of 2.56 mg/L/day. The faster rate in the second cycle was
probably due to the adaptation of bacteria to the degradation of the Cu(I)-cyanide complex.
In the primed bioreactor (B), all of the Cu(I)-cyanide complex was degraded within 10 days.
Although the degradation rate for Bioreactor B, at 5.67 mg/L/day (Figure 8), remained
faster than Bioreactor A, there was a reduction in the rate of cyanide degradation when com-
pared to the first cycle. The slower degradation rate during the third cycle could be due to
less nutrients being available to bacteria. Measurements taken for ammonium-nitrogen dur-
ing the priming period (before the addition of the Cu(I) cyanide complex) showed that no
ammonia was present in Bioreactor A (Figure 9). The high level of ammonia (up to 80 mg/L)
found in Bioreactor B before the addition of the Cu(I)-cyanide complex was due to the degra-
dation by microorganisms of amino acids and peptides present in peptone (Figure 9).
The degradation of the Cu(I)-cyanide complex in Bioreactor A resulted in the concomitant
increase in ammonia concentration. After the addition of the CuCN (Figure 10), ammonium-
nitrogen slowly accumulated in the medium, resulting in a final concentration of approxi-
mately 17 mg/L. Any change in ammonia concentration due to cyanide degradation in
Bioreactor B was masked by the already elevated ammonia concentrations due to the degra-
dation of peptone in the priming phase (80 to 100 mg/L ammonia).
146
FIGURE 9 Ammonium-nitrogen in Bioreactors A and B over 20 days with ( ) and without ()
added peptone and before addition of Cu(I)-cyanide complex
FIGURE 10 Ammonium-nitrogen in Bioreactor A ( ) and Bioreator B () for 28 days after first
addition of Cu(I)-cyanide complex (Cycle I)
During the second cycle of cyanide degradation, ammonia was found to accumulate in
both bioreactors. In Bioreactor A, a gradual increase in ammonia concentration was
observed, resulting in a maximum concentration of 20 mg/L (Figure 11). Bioreactor B also
showed an increase in ammonium-nitrogen concentration of up to 37 mg/L over the first
10 days (i.e., when the Cu[I]-cyanide complex was being degraded), resulting in a final con-
centration similar to that of Bioreactor A (Figure 11). In each of the bioreactors, the pH
remained relatively constant throughout the experiment (pH 9.0), varying only by 0.5 pH
units. The pH dropped slightly at the start of the experiment and slowly increased as degra-
dation proceeded (data not shown).
Priming of the bioreactor with peptone resulted in higher Cu(I)-cyanide complex
degradation rates and better growth of bacteria within the bioreactor. The higher rate of
degradation seen in the primed reactor was probably due to the presence of a larger pop-
ulation of cyanide-degrading bacteria on the surface of the ore particles.
147
FIGURE 11 Ammonium-nitrogen in Bioreactor A ( ) and Bioreactor B () for 33 days after
second addition of Cu(I)-cyanide complex (Cycle II)
FIGURE 12 Degradation of tetracyanonickelate over 56 days in Bioreactor A ( ) and

Bioreactor B () during Cycle I
Tetracyanonickelate Degradation in Bioreactors

Once the degradation of the Cu(I)-cyanide complex was complete, the supernatant from the
bioreactors was drained, and the bioreactors were refilled with fresh DMS spiked with tetra-
cyanonickelate (95 mg/L).
In Bioreactor B, degradation of tetracyanonickelate started within two days and was
complete within 26 days, whereas, in Bioreactor A, degradation commenced after 20 days
incubation and was complete within 56 days. Bioreactor B (primed) showed a faster tetra-
cyanonickelate degradation rate (2.81 mg/L/day) compared to Bioreactor A (1.40 mg/L/day),
which had not been primed (Figure 12). Biodegradation rates within the bioreactors were
faster than those determined in the shake-flask experiments.
Tetracyanonickelate was degraded at faster rates in both bioreactors during the second
degradation cycle, which was due to the microorganisms becoming adapted to the tetra-
cyanonickelate. In Bioreactor A, tetracyanonickelate degradation was complete within
27 days, and, in Bioreactor B, tetracyanonickelate degradation was complete within 10 days.
148
FIGURE 13 Degradation of tetracyanonickelate over 27 days in Bioreactor A ( ) and
Bioreactor B () during Cycle II
The degradation rate in Bioreactor A was 2.94 mg/L/day, and, in Bioreactor B, it was
9.55 mg/L/day. Bioreactor B (primed) had faster degradation rates in both experiments,
and, although bioreactor A improved in the rate of degradation by a factor of 2.1 in the sec-
ond cycle, bioreactor B improved by a factor of 3.4 during the second cycle (Figure 13).
Throughout the test period, the pH in both bioreactors fluctuated by one pH unit, dropping
in the initial stages and gradually increasing as ammonia was produced (data not shown).
The ammonium-nitrogen produced during the breakdown of tetracyanonickelate (first
cycle), was greater in Bioreactor A, where it reached a maximum value of 27 mg/L/day com-
pared to 17 mg/L/day reached in Bioreactor B (Figure 14). The formation of ammonia dif-
fered slightly in each bioreactor in that, for Bioreactor A, the ammonia concentration
increased gradually throughout the incubation period, whereas, in Bioreactor B, it reached a
maximum after 33 days and then decreased.
In the second cycle, Bioreactor A had a maximum ammonia concentration of 29 mg/L at
20 days, with Bioreactor B reaching a value of 28 mg/L at day 10 (Figure 15). As with the
first cycle, ammonia increased gradually in Bioreactor A until it reached a maximum,
whereas, in Bioreactor B, it reached a peak when tetracyanonickelate degradation was com-
plete. The decrease in the amount of ammonium-nitrogen in Bioreactor B in both degrada-
tion cycles toward the end of the experiments could indicate a utilization of the ammonia for
ongoing metabolic processes. Theoretically, the level of ammonium ion that could have been
produced from the breakdown of cyanide was 27.9 mg/L. This suggests that the tetracyano-
nickelate in each of the bioreactors was first degraded to ammonia, which, in Bioreactor B,
was subsequently utilized by the bacteria.
Biodegradation of metal-cyanide complexes may provide an economic means of removing
potentially toxic compounds from tailings or contaminated mine water.
CONCLUSIONS
The biodegradation of both the Cu(I)-cyanide complex and tetracyanonickelate was demon-
strated in shake-flask experiments and in bioreactors. In shake-flask experiments, peptone was
required for the degradation of the metal-cyanide complexes at pH 8 and pH 10. The exception
being tetracyanonickelate, which was degraded at pH 8 without added peptone. Increasing the
pH markedly slowed the degradation rates of metal-cyanide complexes. The optimum pH for
degradation of metal-cyanide complexes was in the range of 7.5 to 8. In addition, peptone sup-
plementation greatly improved the rate of cyanide degradation. The addition of peptone led to
149
FIGURE 14 Ammonium-nitrogen from tetracyanonickelate degradation over 56 days in
Bioreactor A () and Bioreactor B () during Cycle I
FIGURE 15 Ammonium-nitrogen from tetracyanonickelate degradation over 27 days in

Bioreactors A () and Bioreactor B () during Cycle II
the formation of cyanate as an end product when cultures were incubated at pH 8. Differences
in the formation of end products suggested that either the major degradation pathways are dif-
ferent at pH 8, when compared to pH 10, or that ammonia inhibits cyanate degradation by the
enzyme cyanase at pH 8.
In the bioreactor system, degradation of the Cu(I)-cyanide complex occurred in the
absence of peptone. However, degradation rates were faster when the bioreactor had been
primed with peptone. Priming the bioreactor with peptone before the addition of metal-
cyanide complexes led to shorter lag times and faster degradation rates. A comparison of
cyanide degradation rates for shake-flask cultures and bioreactors illustrated that the
bioreactor system was more efficient.
150
ACKNOWLEDGMENTS
The authors thank K. Riley and R. Wood, C.S.I.R.O. Division of Energy and Technology, Lucas
Heights Research Laboratories, Menai N.S.W., Australia, for their analytical services.
REFERENCES
Byerley, J.J., Enns, K., Trang, C.V., and Le, V.T., 1988, “A treatment strategy for mixed cyanide
effluents-precipitation of copper and tetracyanonickelates,” Randol Gold Forum, pp. 331–339.
Cotton, F.A., and Wilkinson, G., 1988, “The elements of the first transition series,” Advanced
Inorganic Chemistry, 5th Ed., John Wiley and Sons, Inc., Canada.
Davis, J.B., Raymond, R.L., and Stanley, J.P., 1959, “Areal contrasts in the abundance of
hydrocarbon oxidizing microbes in soils,” Applied Microbiology, Vol. 7, p. 156.
Dickerson, R.E., Gray, H.B., Darensbourg, M.Y., and Darensbourg, D.J., 1984, Chemical Principles,
4th Ed., The Benjamin/Cummins Publishing Company Inc., Menlo Park, CA.
Dubey, S.K., and Holmes, D.S., 1995, “Biological cyanide destruction mediated by
microorganisms,” World J. Microbiol. Biotechnol., Vol. 11, pp. 257–265.
Ford-Smith, M.H., 1964, “The chemistry of complex cyanides, A literature survey,” Department of
Scientific and Industrial Research, National Chemical Laboratory, Her Majesty’s Stationary
Office, London, pp. 30–35.
Greenberg A.E., Clesceri L.S., and Eaton A.D., editors, 1992, Standard Methods for Examination of
Water and Wastewater, The American Public Health Association Publishers, Washington, DC.
Hilton, D.F., and Haddad, P.R., 1986, “determination of metal-cyano complexes by reversed-
phase ion-interaction high-performance liquid chromatography and its application to the
analysis of precious metals in the gold processing solutions,” Journal of Chromatography, Vol.
361, pp. 141–150.
Hoecker, W., and Muir, D., 1987, “Degradation of cyanide,” Research and Development in
Extractive Metallurgy, The AusIMM, Adelaide Branch, May, pp. 29–36.
Mudder, T.I., Fox, F., Whitlock, J., Fero, T., Smith, G., Waterland, R., and Vietl, J., 1985,
“Biological treatment of cyanidation waste waters: Design, start-up, and operation on a full
scale facility,” Homestake Mining Company Report.
Osseo-Asare, K., Xue, T., and Ciminelli, V.S.T., 1984, “Solution chemistry of cyanide leaching
systems,” Process Metals, Mining, Extraction and Processing, V. Kudryk, C.A. Corrigan and W.W.
Leang, eds., AIME TMS Conference Proceedings, Los Angeles, February, 1984.
Rollinson, G., Jones, R., Meadows, M.P., Harris, R.E., and Knowles, C.J., 1987, “The growth of a
cyanide-utilizing strain of Pseudomonas fluorescens in liquid culture on nickel cyanide as a
source of nitrogen,” FEMS Microbiology Letters, Vol. 40, pp. 199–205.
Scott, J.S., and Ingles, J.C., 1981, “Removal of cyanide from gold mill effluents,” Canadian Mineral
Processors, Thirteenth Annual Meeting, Ottawa, Ontario. January 20–22, pp. 380–418.
Sharpe, A.G., 1976, “The chemistry of cyano complexes of the transition metals,” Organometallic
Chemistry, A Series of Monographs, University Chemical Laboratory, Cambridge, England,
Academic Press, London, pp. 265–271.
Shpak, V.E., Podol’skaya, V.I., Ul’berg, Z.R., and Shpak, E.A., 1995, “Degradation of metal
cyanides complexes in microbe dispersions,” Colloid Journal, Vol. 57, No. 1, pp. 108–112.
Silva-Avalos, J., Richard, M.G., Nagappan, O., and Kunz, D.A., 1990, “Degradation of the metal-
cyano complex tetracyanonickelate (II) by cyanide-utilizing bacterial isolates,” Applied and
Environmental Microbiology, Vol. 56, No.12, pp. 3664–3670.
Stainton, M.P., Capel, M.J., and Armstrong, F.A.J., 1977, “The chemical analysis of fresh water.
Fisheries and environment,” Canada and Fisheries Marine Service, Misc. Spec. Pub., No. 25.
151
Biochemical Removal of HAP
Precursors from Coal—INEEL
Slurry Column Testing
K.S. Noah* and G.J. Olson†
ABSTRACT
A combined physical/biochemical process for the precombustion removal of 13 inorganic hazardous
air pollutant (HAP) precursors, i.e., Sb, As, Be, Cd, Cr, Cl, Co, F, Pb, Hg, Mn, Ni, and Se, from coal
were tested. Biochemical processes for removal of HAP precursors from coal potentially offer advan-
tages of deeper cleaning, more specificity, and less coal loss. The slurry column is a second-generation
process for the beneficiation of fine (60-mesh × 10-µm) coal by a combination of physical separation
of mineral matter and biooxidation of pyrite. Sixty-seven percent removal of pyrite from a 60-mesh
Pittsburgh #8 coal was achieved at a 35% (w/w) slurry concentration and a five-day reactor resi-
dence time. Ninety percent of the heating value of the feed coal was recovered. Among the HAP precur-
sors of most concern, over half of the Se, As, and Hg were removed from the feed coal. From 40% to
70% of most HAP precursors were removed from the feed coal. Hg in the feed coal was reduced from
0.12 to 0.054 µg/g in the product coal, while waste coal contained 0.24 µg/g.
INTRODUCTION
The US Clean Air Act and its amendments regulate the emission of S and N oxides from coal-
fired utility boilers. This legislation also emphasizes reducing public exposure to other so-
called hazardous air pollutant (HAP) precursors resulting from coal combustion. The HAP
precursors include 13 trace elements (Sb, As, Be, Cd, Cr, Cl, Co, F, Pb, Hg, Mn, Ni, and Se),
radionuclides, and polynuclear aromatic hydrocarbons (PAHs). There is great interest in the
development of processes that can efficiently remove HAP precursors from coal.
Most of the 13 trace elements identified as HAP precursors in coal are associated, in part,
with the mineral fraction (especially in the case of pyrite) as discrete sulfides, as carbonates,
and in clay minerals (Valkovic, 1983; Swain, 1985; Spears et al., 1991; Martinez-Tarazone et
al., 1992). Consequently, conventional physical coal cleaning can remove a portion of these
components from coal. However, physical cleaning is not very efficient in removing finely dis-
seminated mineral matter from coals, especially pyrite.
* Biotechnology Department, Idaho National Engineering and Environmental Laboratory, Idaho Falls, ID.
† Little Bear Laboratories, Inc., Golden, CO.
153
Pyrite-oxidizing bacteria offer the potential of a biochemical process for precombustion
leaching of certain trace elements from coal (Olson, 1994). The bacterial oxidation of pyrite
to ferric sulfate and sulfuric acid may solubilize trace elements occurring as sulfides or car-
bonates in coal. A biochemical approach potentially offers the advantages of deeper removal
of trace elements, more specificity, and less loss of coal than physical cleaning. Biochemical
processes operate under milder conditions and are generally less expensive than chemical
treatments.
HAP precursors such as Mn, Ni, Co, As, Pb, Cd, and Cr can be removed from coal by
bioleaching (Jilek and Beranova, 1982; McCready et al., 1985; Francis et al., 1989; Olson et
al., 1999). However, Hg, perhaps the most important of the HAP precursors, could not be
removed from US coals by bioleaching, and bioleaching removed only a portion of other HAP
precursors (Olson et al., 1999).
A slurry column reactor developed at INEEL for coal depyritization uses a combination of
physical separation of mineral matter and biooxidation of pyrite (Andrews et al., 1993b).
The goals of this project were to determine the effectiveness of the slurry column reactor in
removing HAP precursors and pyrite from Pittsburgh seam coal. This coal contains finely dis-
seminated pyrite and is difficult to clean by conventional processes.
SLURRY COLUMN REACTOR

The slurry column reactors were made of Plexiglas and have the shape of a tall, narrow-angled
wedge (Figure 1). The columns are 380-mm wide with a half-angle of 5° and a working vol-
ume of 34 L. The aeration tube, made of perforated silicon rubber, goes completely across the
reactor near the bottom, leaving a gap between it and the wall. The feed coal is slurried in a
microbial culture in the clean coal reactor (CCR), which also acts as a physical separator. Lib-
erated mineral particles, being denser and more hydrophilic than coal, settle out during the
biodegradation/separation operation (aeration step) and fall through this gap into the solids
collection region. The liberated mineral particles are then transferred to a second identical
reactor, the rougher/propagator (R/P) reactor. This acts as a second-stage separator (a
“rougher”) that divides the true mineral particles from the organic coal particles that are
transferred with them due to the imperfect separation in the CCR. The former again settle out
and are discarded as solid waste, while the latter are returned to the CCR with the coal slurry.
Many of the bacteria that grow on the large inclusions are returned to the CCR with the coal
slurry, providing an additional well-acclimated inoculum for the feed coal (hence “propaga-
tor”). When the air is turned off, the solids settle out and the supernatant is pumped out
through a manifold tube lowered onto the top of the settled bed. A screen was installed in one
wall of the solids collection region to study the possiblity of draining more water out of the
base of the settled bed, but this did not significantly improve dewatering. The other wall of the
reactor contains a door through which the treated coal can be discharged.
Unlike the aerated trough (Andrews et al., 1993a), these units were deliberately designed
as physical separators as well as bioreactors. Factors contributing to enhanced separation are
the height of the column and the details of the base of the reactor, including the aeration
tube, the tube-to-wall gap size, and the positioning of baffles. Previous experiments identi-
fied a range of air flows that keep the coal suspended while allowing the minerals to settle
and that were also sufficient to supply the oxygen and carbon dioxide needed by the microor-
ganisms (Andrews and Noah, 1997). In addition, details of the theory and design of the
slurry column reactors can be found in this paper.
Startup
Pittsburgh #8 coal was ground at Pittsburgh Energy Technology Center’s Coal Preparation
Division, and the ground coal was shipped in drums to INEEL. The coal contained 17.3% ash,
3.5% total sulfur, 1.3% pyritic sulfur, and approximately 2.0% moisture. The particle size
was approximately 30% +60 mesh and approximately 100% +170 mesh.
154
FIGURE 1 The slurry column reactor
Details of the process startup procedure have been described previously (Andrews et al.,
1992). Both the CCR and the R/P were filled with microbial medium containing 1.2 g/L Fe2+
as the energy source and were inoculated with a mixed culture of Thiobacillus spp. and
Leptospirillum spp. The medium was periodically recycled through an electrolytic cell to
regenerate the Fe2+, which allowed continued microbial growth. At the same time, cells were
grown on Pittsburgh #8 coal in shake flasks whose initial inoculum was a consortium of
mesophilic pyrite-oxidizing bacteria.
When the microbial cell concentration reached 1.0 × 108 cells/mL, coal was added to the
reactors. The resulting 35% (w/w) slurry was then inoculated with a consortium of bacteria
grown on coal. The reactors were then aerated for 41 days, allowing the microbial culture to
become properly adapted to the coal before any steady-state measurements were taken.
After this first 41-day run, five runs (charging the CCR with fresh coal) were performed to
test operating and sampling protocol for this specific experiment.
Operational Procedure
The process consists of two reactors, the CCR and the R/P, together with associated pumps
and holding tanks, operated in the sequencing batch mode (Andrews and Noah, 1997). For all
155
runs in this test, the working volume of each reactor was 34 L at a 35% (w/w) slurry (11.25 kg
dry coal). Feed coal was added to the CCR and aerated for approximately one hour. Then, the
air was shut off and followed by a strict three-hour waiting period. After the three-hour wait-
ing period, a predetermined amount of wastewater was pumped off from the top of the CCR
into a liming tank. The waiting period followed by wastewater pumping fixed the smallest
particle size remaining in the system, which affects physical separation and effectively
removed all of the ultra fines and clays from the process. The wastewater volume determined
the hydraulic residence of the batch. Thus, this volume was a variable that can be manipulated
to reach varying results in the experiment. The wastewater was treated by lime precipitation.
Our experience (Andrews et al., 1992) was that relatively small amounts of lime produce an
excellent, rapid-settling floc. This floc incorporates the bacteria and colloidal coal, leaving a
clear supernatant. The supernatant retains a sulfate concentration on the order of 1 g/L, pre-
sumably because it is saturated with gypsum (CaSO4·2H2O has a solubility of 2.41 g/L).
The majority of the runs were done using a five-day coal residence time, eight with 10 L of
wastewater and four with 5 L of wastewater, followed by six runs using a two-day coal resi-
dence time, five with 5 L of wastewater and one with 10 L of wastewater. During the course
of aeration, five settled solid drains, the region below the aeration tube, were taken from the
CCR and allowed to dry before being placed back in the R/P. Weights were taken of these on
a daily basis to preserve the integrity of the water balance. In addition, two settled solid
drains were taken from the R/P. These two drains constituted the waste coal.
Analytical Procedures
Seven samples were analyzed for percent total S, percent pyritic S, percent sulfate S, percent
organic S, moisture content and 13 HAP precursors (Hg, As, Cd, Cr, Se, Sb, Be, Co, Pb, Mn,
Ni, Cl, and F). Trace detection limits of the HAP precursors in the coal were approximately
0.5 µg/g, except for Cd, Mn, Be (0.2 µg/g), Hg (0.02 µg/g), and Pb (2.0 µg/g) (Olson et al.,
1999). The seven samples were as follows (see Figure 2):
Feed coal: added to the CCR;
New water 1: before the first wash water was acidified, a liquid sample was taken from
approximately 15 L (this sample was only taken if new water was added to the system
at this step, if no new water was added for the first wash, no sample was taken);
New water 2: 15 L of fresh dechlorinated tap water for the second wash;
Waste coal: the coal removed from the two solids drains of the R/P;
Product coal: the coal removed from the CCR during the wash step;
Wastewater solids: the solids that settled out after the wastewater had been limed; and
Wastewater liquid: the clear supernatant left on the top of the settled solids.
Daily samples were taken from both reactors and analyzed for sulfate sulfur, total iron,
and ferrous iron. Sulfates were analyzed on a spectrophotometer using a barium chloride
turbidimetric method (ASTM D–516–82), and total iron was measured by Atomic Absorption
Spectroscopy using diluted samples. Ferrous iron was analyzed by placing a 1-mL sample of
reactor liquid into a solution containing 4 mL of acid solution with 10 mL of nanopure water
and then titrated using a 0.01-N KMnO4 solution. The amount of Fe2+ was then calculated.
Eh (reference electrode silver-silver chloride), pH and air flow rate were monitored daily,
while cell counts were done occasionally to ensure that cell populations were being main-
tained. All solid samples were analyzed for ash content.
156
FIGURE 2 Slurry column overall mass balance flowsheet
Eh, pH, and Bacteria Counts

The oxidation/reduction potential (Eh) is a good indicator of the Fe3+/Fe2+ ratio in solution
for microbial depyritization systems. This ratio is decreased by the consumption of ferric ion
for the chemical oxidation of pyrite according to
FeS2 + 2Fe3+ → 3Fe2+ + 2S (EQ 1)
FeS2 + 14Fe3+ + 8H2O → 15Fe2+ + 2SO4 = + 16H+ (EQ 2)

It is increased by the oxidation of ferrous ion by the iron-oxidizing bacteria Thiobacillus fer-
rooxidans and Leptospirillum ferrooxidans
Fe2+ + H+ + 1⁄4 O2 → Fe3+ + 1⁄2 H2O (EQ 3)

As can be seen in Figure 3, the Eh varied over the residence time of the reactor. As expected,
at the beginning of a run, when the CCR was charged with fresh feed coal, a relatively low Eh
was observed due to the addition of the feed coal. Hence, fresh pyrite was added to the sys-
tem and ferric ions were consumed by the reactions represented by Eqs. (1) and (2). Interest-
ingly, it was also observed that there was a sharp decrease in the Eh of the R/P at the
beginning of the run. This was due to the addition of the pyrite-rich drainings from the CCR
added to the R/P at the beginning of the run. Eventually, the rate of microbial oxidation of
ferrous iron, as represented by Eq. (3), exceeded the rate of ferric ion consumption by pyrite
and an increase in Eh was observed. The concentration of Fe2+ decreased in the CCR over the
course of the run (data not shown), in agreement with the Eh data. The concentration of Fe2+
ions in the R/P remained low and fairly constant. The relatively large pyritic inclusions in the
R/P are more slowly oxidized over a longer period.
A major process control objective was to keep the pH within the tolerance of the microor-
ganisms and low enough to prevent the formation of precipitates (i.e., ferric hydroxysulfate
157
FIGURE 3 Eh of various runs
FIGURE 4 pH during various runs
precipitates). The pH was maintained at 1.5 to 2.5 by adding sulfuric acid to the reactors.
There would be theoretical advantages to using phosphoric acid instead, which provides an
essential microbial nutrient. However, ferrous phosphate and other phosphates have a very
low solubility and so would coat the pyrite surfaces and would complicate washing of the
product coal. The bacteria appear to obtain sufficient phosphate from the feed water and coal.
The variation of the pH in both reactors is shown in Figure 4. The sharp increase in the pH
of the CCR at the beginning of a run was due to the basic minerals in the feed coal. The addi-
tion of sulfuric acid at the beginning of the run caused the sharp decrease in pH between day
one and day two. Due to microbial oxidation of pyrite (Reaction [2]) and the sulfur formed
158
TABLE 1 The process water balance
Reactor Waste- Product
residence water Feed coal New wash coal water, Waste coal Waste- CCR drains Percent
time, days volume, L water, kg water, kg Total kg water, kg water, kg water, kg Total recovered
5 10 0.195 20.0 20.070 6.852 0.267 10.150 0.510 17.777 89.9
5 5 0.211 17.0 16.961 6.869 0.28 6.549 0.616 14.341 85.3
2 5 0.154 14.6 14.754 7.418 0.184 4.812 0.514 12.929 91.2
2 10 0.152 18.5 18.652 10.290 0.231 10.046 0.478 21.044 112.8
in Eq. (1) (by Thiobacillus ferrooxidans or Thiobacillus thiooxidans), the pH continued to

decrease slightly by
S + 3⁄2 O2 + H2O → 2H+ + SO42– (EQ 4)

The Eh and pH were excellent indirect measures of the activity of the microbial population.
A direct measure involved direct microscope counts of the total bacterial concentration in the
reactor liquids and in the wastewater. These measurements were difficult and imprecise
because of the presence of coal fines and because of the tendency of the bacteria to adhere to
solids. These measurements, therefore, show considerable daily variation in the R/P and CCR.
Part of the operating strategy was to recycle the cells to inoculate the next batch of feed coal. In
most cases, there were 2.6 E7 cells/mL in the wastewater, implying good cell recycle.
The Water Balance

The water balance is shown in Table 1. The greatest concern was the amount of water in the
product coal. Knowing the coal mass fraction and grind size of the coal, the amount of water
that theoretically should be associated with the product coal can be calculated. However, the
critical amount of water left in the coal was more than the calculated amount. Another con-
cern was that the measured amount of water fed into the system did not equal the water out-
let. There were several possibilities for this. First, it was assumed that the evaporation water
was equal to the water makeup. The aeration made it difficult to see the actual slurry level of
the reactor making the addition of makeup water imprecise. Second, the sample used to
determine percent water could have been wetter or drier than the rest of the coal on the air-
dried tray. As seen in Table 1, there was a significant amount of variation in the percent of
water recovered, from 85% to 113%, corresponding to +4 L to –6 L.
The Solids and Heating Values Balance

Table 2 represents the process solids balance. This balance was calculated before lime addition.
The average overall percent recovery for the five-day coal residence times was about 89%, and
the average for the two-day coal residence times was 91%. Longer residence times allowed for
more dissolution of the basic ash and pyrite by the acidic media. Another reason for not achiev-
ing 100% recovery was that some residual coal was left on the sides of the reactor when collect-
ing the coal. The fact that some mass was lost could contribute to the lack of balance in water,
ash, and HAPs, because the majority of these balances were mass dependent. Although not
measured, some of the coal organic matter may have been solubilized due to chemical and
microbial reactions (Andrews et al., 1993b).
More important than the recovery of the mass of the feed coal is the recovery of its heating
value. The heating value of the product coal (Table 3) averaged approximately 8% more
than that of the feed coal. This is what should happen when the ash or mineral content is
removed from the coal. The heating value recovery, defined as the product coal heating
value times the mass of product coal, divided by the feed coal heating value times the mass of
feed coal, was 88% for the five-day coal residence time runs, and 91% for the two-day coal
residence time runs.
159
TABLE 2 The process solids balance. The feed coal was 11.25 kg (dry weight).
Reactor Wastewater
residence time, volume, Product coal, Waste coal, Recovery,
days L % of total % of total %
5 10 81.0 8.1 89.1
5 5 81.2 8.2 89.4
2 5 83.3 7.8 91.1
2 10 82.2 7.8 90.0
TABLE 3 Heating value recovery. The feed coal was 12,043 Btu/lb.
Reactor Wastewater Product coal Heating
residence time, volume, heating value, value recovery,
days L Btu/lb %
5 10 13,082 87.48
5 5 13,188 87.98
2 5 12,867 90.04
2 10 12,947 92.06
TABLE 4 The process ash balance. The feed coal was 17.35% ash.
Reactor Wastewater Ash in
residence time, volume, Product coal, product, Waste coal, Ash in waste,
days L % ash % % ash %
5 10 10.98 51.6 43.24 20.3
5 5 10.46 55.7 43.71 23.3
2 5 12.10 53.9 50.80 20.6
2 10 12.27 48.3 60.65 22.7
TABLE 5 Sulfur measurements of solids. The feed coal was 3.63% total S and 1.39% pyritic S.
Reactor Wastewater Product coal Waste coal
residence time, volume,
days L Total S, % Pyritic S, % Total S, % Pyritic S, %
5 10 3.02 0.59 4.74 2.78
5 5 3.02 0.46 4.48 2.14
2 5 3.26 0.62 5.87 3.37
2 10 3.12 0.73 5.09 2.09
The Ash-Forming Minerals

The data on the ash content of the product coal and solid waste are summarized in Table 4. The
ash in the product coal and the solid waste do not add up to that in the feed coal because some
of the ash-forming minerals (including pyrite) were dissolved during the process. Typically,
20% to 23% of the ash in the feed coal appeared in the solid waste, 51% to 56% appeared in
the product coal and 20% to 30% was solubilized. Twenty percent physical removal was an
improvement over the trough design (Andrews et al., 1993b), where only 10% of the ash was
removed physically. Improving this variable was not an objective in these runs, so it could pre-
sumably be raised further in dedicated work.
160
The Sulfur Balance and Pyritic Sulfur Removal
The results of the sulfur analysis of the solid samples are shown in Table 5. The two-day coal
residence time had lower pyritic reduction because the coal spent less time in the reactor. The
increase in pyrite removal during the five-day coal residence time, 5-L wastewater runs was
more a function of the bacteria adapting to the coal than the amount of wastewater removed.
On average, the waste coal had twice the pyritic sulfur of the feed coal. This implies that
there was good physical separation in the reactors. However, the pyritic sulfur for the waste
coal does fluctuate. It was difficult to remove just the material below the aeration tube (assum-
ing that this material was high in pyrite) without also getting some of the organic fraction, (i.e.,
the slurry above the aeration tube) when draining the solid collection region. Improving the
draining procedure would produce higher pyritic sulfur numbers in the waste coal.
Pyrite reduction was somewhat over 50% based on pyritic sulfur measurements. The total
sulfur reduction in product coal averaged about 15%, but should have been approximately
22% based on pyritic sulfur reduction. This discrepancy is due to the inherent imprecision of
the ASTM method for coal sulfur speciation and because the biooxidation of coal pyrite does
not necessarily proceed stoichiometrically (Andrews, 1989; Schippers et al., 1996). Elemen-
tal sulfur (formed by biological oxidation of pyrite) is counted as total sulfur and also reports
as organic sulfur, not pyritic sulfur. The ASTM procedure for coal sulfur actually measures
pyritic iron, with the sulfur value being based on the stoichiometry of pyrite.
A complete sulfur balance is shown in Table 6 as grams of sulfur and usually shows a loss
of sulfur. The authors have shown that ASTM procedures, as were used here, do not remove
all the sulfate sulfur from coal that has been biooxidized (Andrews, 1989). Additional wash-
ings, stronger normality acid or longer wash times (or a combination) were needed to wash
all sulfate and precipitated sulfate from the coal. There also could have been some loss of sul-
fate with the mist in the air leaving the reactor. Although there were two stages of demisting,
baffles, and screens, there was, on average, about 1,500 mL of water loss from each reactor
daily. In addition, some of this apparent sulfur loss may be caused by the accumulation of
solids on the walls of the CCR. However, most of the sulfur loss must be caused by the partial
oxidation of pyrite, as discussed above. Some of the pyritic sulfur in the feed was converted
to elemental sulfur rather than sulfate, and it does not appear in any of the outflow measure-
ments. This hypothesis was supported by the pyritic and total sulfur measurements from the
product coal listed in Table 5. For example, there was 2.2% organic sulfur in the feed coal
(Table 5). For a coal residence time of five days and 5 L of wastewater, there was an average
of 0.46 ± 0.19% pyritic sulfur in the product coal, which implies that there was 2.66 ± 0.19%
total sulfur, less than the 3.02 ± 0.06% total sulfur shown in Table 5. This suggests that the
product coal contains 0.37 ± 0.25% elemental sulfur, corresponding to a sulfur flow of
between 11 and 57 g. This is in the range to close the sulfur balance.
The percent physical removal (Table 6) of pyritic sulfur was calculated from the grams of
pyritic sulfur appearing in the waste coal divided by the grams of pyritic sulfur in the feed
coal. For coal residence times of five and two days, the physical removal will be independent
of the residence time. With longer residence times, the percent physical removal may
decrease due to dissolution of the pyrite. On average, there was about 16% pyrite removal by
physical means. Again, by changing the number of drainings, the ratio of R/Ps to CCRs and
improving the solid collection, the percentage physical removal could be improved. The per-
cent removal of pyritic sulfur caused by microbial oxidation was calculated from the data in
Table 6 for the rate of appearance of sulfate sulfur minus the sulfate sulfur added, divided by
the amount of pyritic sulfur. As expected, the percent oxidation decreased with decreased
reactor residence time.
HAPs Balance
To verify the efficiency, with which the slurry column removed the 13 HAPs from the feed
coal, a material balance was performed on each of the 13 elements (Table 7). For clarity,
only the data for the five-day residence time, 10-L wastewater runs are shown. Detailed data
161
TABLE 6 The process sulfur balance
Steady-state
conditions Sulfur inflow, g Sulfur outflow, g Percent removal
Reactor Sulfuric Waste Product Waste Product
residence Waste- Feed coal Feed coal acid coal coal WWS coal coal WWS WWL Percent Percent
time, water pyritic sulfate sulfate pyritic pyritic pyritic sulfate sulfate sulfate sulfate physical microbial
days volume, L sulfur sulfur sulfur Total sulfur sulfur sulfur sulfur sulfur sulfur sulfur Total removal oxidation
5 10 156.47 33.11 29.58 219.16 25.31 52.80 1.80 7.65 29.32 52.15 3.26 166.89 16.18 16.41
5 5 145.13 29.25 29.29 203.67 20.04 41.53 1.26 9.36 30.79 41.13 1.33 145.44 14.20 17.47
2 5 156.70 34.85 44.08 235.63 29.14 59.54 1.29 9.61 42.78 43.86 0.35 186.37 18.73 11.96
2 10 168.75 49.50 30.74 248.99 18.34 67.47 1.87 9.30 30.50 75.85 — 203.32 10.87 20.98
TABLE 7 The process HAPs balance for the five-day coal residence time and 10-L wastewater runs. Values are averages of eight runs.
Feed coal Product coal Waste coal Wastewater solids and liquids Percent
HAP N+S+NW*, Total in, accounted
element µg/g mg mg lime, mg mg µg/g mg µg/g mg µg/g mg mg/L mg for
162
Hg 0.123 1.38 0.006 0.004 1.389 0.060 0.546 0.235 0.214 0.114 0.085 0.001 0.003 63
As 10.5 118 0.034 0 118.2 5.0 45.8 19.5 17.7 25.89 18.78 0.005 0.024 70
Cd 0.14 1.6 0.034 0.024 1.630 0.16 1.40 0.37 0.32 1.09 0.80 0.003 0.014 162
Cr 26.9 302 0.073 0.789 302.5 20.0 182.0 83.9 76.4 44.1 32.9 0.007 0.033 97
Se 1.73 19.4 0.058 0 19.5 0.8 7.0 3.4 3.1 2.45 1.8 0.005 0.025 58
Sb 1 11 0.113 0 11.3 1.0 9.1 1.0 0.9 0.9 0.62 0.01 0.05 94
Be 1.10 12.3 0.037 0.024 12.4 1.1 9.5 1.0 1.0 1.34 1.00 0.003 0.015 91
Co 5.50 61.7 0.290 0.024 62.0 3.1 28.0 6.6 6.0 23.4 17.4 0.01 0.047 83
Pb 10.9 122 0.591 0.485 123.3 7.5 67.9 20.7 18.8 24.3 18.1 0.08 0.34 87
Mn 295 3,309 19.6 1.1 3,329.6 63.1 577.0 629.0 569.0 1,267 935 4.0 19.9 65
Ni 13.6 152 0.620 0.061 152.8 6.7 60.9 25.3 23.0 50.6 37.4 0.013 0.065 79
Cl 291 3,269 838.125 0 4,106.8 353.0 3,201.0 241.0 218.0 393 296 54 260 98
F 103 1,160 7.7 0 1,167.6 79.9 726.0 129.0 117.0 0.114 0.085 3.1 15 74
* Nitric + sulfuric + new water

can be seen in Olson (1997). There was greater than 45% reduction in concentration of Hg,
As, Se, Co, and Ni for the feed coal. There was excellent removal of Mn from the feed coal,
approximately 80% reduction in concentration. Removal of Be, Cl, and F from the feed coal
was poor. Sb was below detection limits (1.0 µg/g and 0.01 mg/L) for all samples including
the waste coal. Unfortunately, no conclusions can be drawn for Sb. HAPs concentrations in
the waste coal was doubled or tripled, except for Sb, Be, Cl, F, and Co, which were about
equivalent in concentration to the feed coal. This means that the percent physical removal of
the other HAPs was equivalent.
The Cd content of the feed coal for the runs shown in Table 7 averaged only 0.14 µg/g.
This value was at least four times higher in the other runs (data not shown). This helps
explain why the output for the first set was 162% of the input. The data shown is not a true
representation for Cd. In general, Cd had fair reduction in the product coal of 20% to 35%,
similar to Pb and Cr. Because the concentration of Co in the waste coal approximated that of
the feed coal and there was approximately 45% reduction in concentration in the product
coal, the majority of Co that was removed became soluble and was, therefore, removed
chemically or microbiologically. Cr concentration was tripled in the waste coal along with lit-
tle reduction in the product coal. This implies that, overall, there was a greater physical
removal of Cr than microbial reduction.
There was little to no difference in HAPs reduction based on wastewater volume. How-
ever, in some cases, coal retention time made a difference. For example Se, Co, and Ni had
lower concentrations in the product coal at the five-day coal residence time.
Also shown in Table 7 is the percent recovery or closure on a mass balance of each HAP.
What is interesting is that the elements that had very poor removal, Be and Cl, had good
mass-balance closure. For some of the elements, the mass balance did not close. One expla-
nation may be that the elements accumulated in the R/P. To verify this, an overall mass bal-
ance on each element was performed (Table 8). All elements showed a loss, except for Cr,
which showed an overall gain of 68 mg. Next, the solids and liquids left in the R/P after the
last run, and the fines left over were analyzed for HAPs (Table 9). These “R/P-accumulation”
terms were on the same order of magnitude as the overall mass-balance accumulation terms,
except for Cr, Sb, and Mn. This suggests that, except for Cr, Sb, and Mn, elemental mass bal-
ances can be closed. Cr gain could be from dissolution of the steel screws. There were some
stainless steel screws used in the reactors to hold together screens and gaskets. Because Sb
was always below detection limits in the feed coal, product coal, and waste coal, one cannot
perform a mass balance. However, Table 9 shows that Sb accumulated in the R/P; there was
21 µg/g Sb in the R/P solids. Mn did accumulate in the R/P (Table 9) but does not account
for all the loss. As indicated in Tables 7 and 8, up to 35% of the Mn was solubilized into the
wastewater. Therefore, Mn might not be accounted for due to the water balance being off by
up to 30% some of the time.
Finally, using the overall mass balance, the performance of the slurry column can be eval-
uated (Table 10). A ranking of the elements removed from the feed coal was done based on
how much ended up in the product coal, the waste coal (physical removal) or the wastewater
(biological/chemical solubilization). The slurry column process reduced Mn the best. After
that, As, Ni, Se, and Co were within the same removal efficiency, followed by Hg, Cd, Pb, and
F, followed by Cr, Be, Cl, and Sb. Cr had a low removal efficiency, but was removed only by
physical means. Cr, Mn, Se, Hg, Pb, Cd, As, and Ni were removed with approximately the
same efficiencies physically. Mn and Co were reduced 25% by biological/chemical means,
followed by Ni, Cd, and As.
Although Hg did not go into solution by biotreatment, this work has shown that biotreated
coal can apparently enhance the subsequent physical removal of Hg. Approximately 50% of
the Hg was removed from the feed coal, winding up in the waste coal stream, or physical
removal process of the slurry column. Hg in the feed coal was 0.12 µg/g and was reduced to
0.054 µg/g in the product coal. The waste coal contained 0.24 µg/g Hg.
163
TABLE 8 Slurry column overall mass balance of HAP precursor elements
NW1* NW2 Accumulation

Total Feed new new Nitric Sulfuric Total Waste Product Balance term
HAP input, coal, water, water, acid, acid, Lime, output, coal, coal, WWS, (out/in), ins-outs,
element mg % % % % % % mg mg % % % mg
Hg 24 96.4 1.9 3.1 0 0 0.4 18 25.1 68.1 6.1 76.5 5.53
As 2,165 100 0.6 0 0 0 0 1,607 23.9 56.9 19.2 74.2 558
Cd 87 98.7 2.9 0.7 0 0 0.6 80 19.5 59.4 20 91.9 7.1
Cr 6,224 99.7 0.3 0 0 0 0.3 6,292 26 66.7 7.1 101.1 –68.2
Se 322 99.7 0.3 0.3 0 0 0 243 27.6 58.5 13.7 75.4 79.5
Sb 205 98.9 0.8 1.1 0 0 0 194 9.2 85.3 5.2 94.8 10.6
164
Be 218 99.4 0.6 0.4 0 0 0.3 198 8.4 84.3 6.9 90.8 20.1
Co 1,094 99.5 2.7 0.4 0 0 0.1 906 13.2 56.4 28.8 82.8 189
Pb 2,152 99.1 0.4 0.4 0 0 0.5 1,869 20.9 65.7 13.2 86.8 284
Mn 58,907 99.4 3.7 0.5 0 0 0 41,543 31.6 32.7 35.2 70.5 17,365
Ni 2,669 99.6 3.1 0.4 0 0 0.1 2,183 19.7 52.2 26.4 81.8 486
Cl 97,687 85.6 1.1 14.4 0 0 0 89,186 5.3 85.1 5.6 91.3 8,501
F 21,866 99.3 2 0.7 0 0 0 17,336 12.5 74.6 11.4 79.3 4,530
* New water 1 is an intermediate value and does not add to the overall balance.
TABLE 9 Overall accumulation of HAPS based on analysis of R/P and fines
R/P solids R/P water Fines
HAP Accumulation
element µg/g kg mg mg/L L mg µg/g kg mg total, mg
Hg 0.178 12.95 2.3051 0.005 5 0.025 0.206 0.664 0.137 2.47
As 14.4 12.95 186 2.86 5 14.3 25.1 0.664 16.7 217
Cd 0.8 12.95 10.4 0.167 5 0.835 1.36 0.664 0.903 12.1
Cr 44.6 12.95 578 1.2 5 6 47.5 0.664 31.5 615
Se 2.82 12.95 36.5 0.34 5 1.7 3.53 0.664 2.34 40.6
Sb 21 12.95 272 0.01 5 0.05 1 0.664 0.664 273
165
Be 1.02 12.95 13.2 0.062 5 0.31 0.94 0.664 0.624 14.1
Co 4.82 12.95 62.4 1.43 5 7.15 10.2 0.664 6.77 76.3
Pb 16.8 12.95 218 0.34 5 1.7 36.6 0.664 24.3 244
Mn 360 12.95 4,662 192 5 960 525 0.664 349 5,971
Ni 14.1 12.95 183 3.88 5 19.4 21.9 0.664 14.5 217
Cl 520 12.95 6,734 115 5 575 420 0.664 279 7,588
F 136 12.95 1,761 3.7 5 18.5 206 0.664 137 1,916
TABLE 10 Removal ranking. HAP elements are shown in order of rank from best to worst case.
Overall Overall
Overall physical bio/chem
percent in HAP removal,† HAP removal,§ HAP
product coal* element % element % element
23.1 Mn 26.4 Cr 25.1 Mn
42.2 As 22.3 Mn 25.1 Co
42.8 Ni 20.8 Se 22.9 Ni
44.1 Se 19.3 Hg 18.7 Cd
46.7 Co 18.2 Pb 14.3 As
52.3 Hg 18.1 Cd 11.1 Pb
54.9 Cd 17.7 As 10.5 Se
57.3 Pb 16.1 Ni 10.2 F
59.1 F 10.9 Co 8.8 Cl
67.7 Cr 9.9 F 7.0 Cr
76.7 Be 8.8 Sb 6.4 Be
77.7 Cl 7.7 Be 5.2 Sb
80.8 Sb 4.8 Cl 4.8 Hg
* [Total mg Product Coal / (total mg Feed Coal + Total mg NW2 + total mg Nitric Acid + Total mg Sulfuric Acid)]*100
† [Total mg Waste Coal / (Total mg Feed Coal + Total mg NW2 + Total mg Nitric Acid + Total mg Sulfuric Acid)]*100
§ [(Total mg WWS + Total mg WWL – Total mg Lime) / (Total mg Feed Coal + Total mg NW2 + Total mg Nitric Acid +
Total mg Sulfuric Acid + Total mg Lime)]*100
CONCLUSIONS
Overall, the slurry column reactor performed well for its first trial. The narrower, taller
design improved physical removal of pyrite over the trough design used in earlier work.
There was approximately 20% removal of ash and pyrite. There was also close to 90% Btu
recovery and good closure on the solids balance. The sulfur balance did not close. Items to be
addressed in future work are:
improving the washing of the product coal;
improving sulfur oxidizing activity to prevent the formation of elemental sulfur; and
performing an iron balance, which is easier to perform than a sulfur balance, and verify-
ing the sulfur balance and formation of elemental sulfur.
Overall, the five-day coal residence times had better pyrite and HAPs reduction. The two-
day retention performed well also, but a one-day residence time would be more feasible for
an actual process. The amount of wastewater purged from the reactor made little or no dif-
ference on pyrite or HAPs reduction.
Mn was removed best using this system. The overall ranking, based on product coal, was
Mn, Ni, Se, Co, and Hg. Physically, Cr, Mn, Se, Hg, and Pb were removed best. Microbial/
chemical solubilization reduced Mn, Co, Ni, Cd, and As. In the feed coal, most elements were
low in concentration to start with, and these elements were even lower in the product coal,
making analysis difficult. Between analysis error, water-balance error and accumulation
within the R/P, HAPs balances can be closed.
Items that require further testing are the product coal collection procedure, the waste coal
collection procedure to include the number of drainings and the number of CCRs to R/Ps.
Addressing these items would improve physical separation, Btu recovery, HAPs removal, and
mass balances of water, coal, and HAPs.
166
ACKNOWLEDGMENTS
This work was supported in part by the US Department of Energy, Assistant Secretary for
Fossil Energy, under DOE Idaho Field Office Contract DE-AC07–99IDO13727. The US
Department of Energy’s Federal Technology Center, under contract number DE-AC22–
95PC95155 with Little Bear Laboratories, Golden CO, sponsored this research. The FETC
COR was Dr. M. Nowak. Dr. D. Finseth of FETC assisted in obtaining Pittsburgh coal. The
contribution and expertise of Mr. J. Richards of Covenant Engineering Services is acknowl-
edged. Dr. R. Cherry of the INEEL is thanked for reviewing the paper.
REFERENCES
Andrews, G.F., 1989, “An examination of the kinetics of coal pyrite decomposition,” Biotechnology
in Materials and Metal Processing, B. Scheiner, F. Doyle, and S. Kawatra, eds., Society of Mining
Engineers, pp. 87–93.
Andrews, G.F., Dugan, P.R, McIlwain, M.E., and Stevens, C.J., 1992, “Microbial Desulfurization of
Coal,” INEL Report EG&G–2669.
Andrews, G.F., Stevens, C.J., Glenn, A., and Noah, K.S., 1993a, “Microbial coal depyritization;
why and how?” Biohydrometallurgical Technologies, Vol. I, A.E. Torma, J.E. Wey, and V.L.
Lakshaman, eds., pp. 381–391, Warrendale, PA, TMS.
Andrews, G.F., Stevens, C.J., Noah, K.S., and McIlwain, M.E., 1993b, “A Combined Physical/
Microbial Process for the Beneficiation of Coal,” INEL Report EG&G–2712.
Andrews, G.F., and Noah, K.S., 1997, “The slurry-column coal beneficiation process,” Fuel
Processing Technol., 52, pp. 247–266.
Francis, A.J., Dodge, C.J., Rose, A.W., and Ramirez, A.J., 1989, “Aerobic and anaerobic microbial
dissolution of toxic metals from coal wastes: mechanism of action,” Environ. Sci. Technol., 23,
p. 435.
Jilek, R. and Beranova, E., 1982, “Some experiences with bacterial leaching of brown coal,”
Proceedings International Conference on Use of Microorganisms in Hydrometallurgy, Pecs,
Hungary, p. 167.
Martinez-Tarazone, M.R., Spears, D.A., and Tascon, J.M.D., 1992, “Organic affinity of trace
elements in Asturian bituminous coals,” Fuel, 71, p. 909.
McCready, R.L., and Zentilli, M., 1985, “Beneficiation of coal by bacterial leaching,” Can. Metall.
Quarterly, 24, p. 135.
Olson, G.J., 1994, “Prospects of biodesulfurization of coal: mechanisms and related process
designs,” Fuel Process. Technol., 40, p. 103.
Olson, G.J., 1997, “Biochemical Removal Of HAP Precursors From Coal,” Final Report, Project DE-
AC22–95PC95155, submitted to US Dept. of Energy, Pittsburgh, PA, May 1997.
Olson, G.J., Tucker, L.R., and Clark, T.R., 1999, “Bioleaching of trace elements from U.S. coals by
pyrite-oxidizing bacteria,” Proceedings International Biohydrometallurgy Symposium, El
Escorial, Spain, June 20–23, pp. 483–491, In: Biohydrometallurgy and the Environment Toward
the Mining of the 21st Century, Part A, R. Amilis and A. Ballester, eds., Amsterdam, Elsevier.
Schippers, A., Jozan, P.G., and Sand, W., 1996, “Sulfur chemistry in bacterial leaching of pyrite,”
Applied and Environmental Microbiology, Vol. 62, No. 9, Sept., pp. 3424–3431.
Spears, D.A., 1991, “Pyrite in some U.K. coals,” Processing and Utilization of High Sulfur Coals IV,
P.R. Dugan, D.R. Quigley, and Y.A. Attia, eds., Amsterdam, Elsevier, pp. 85–93.
Swain, D.J., 1985, “Modern methods in bituminous coal analysis: trace elements,” CRC Crit. Rev.
Anal. Chem., 15, p. 315.
Valkovic, V., 1983, “Trace elements in coal: occurrence and distribution,” Trace Elements in Coal,
CRC Press, Boca Raton, FL.
167
Microorganisms, Biotechnology,
and Acid Rock Drainage—
Emphasis on Passive-Biological
Control and Treatment Methods
N. Kuyucak*
ABSTRACT
Acid rock drainage (ARD) resulting from the natural oxidation of sulfide mineral wastes has become a
major concern to the mining industry, both during the operational period and after mine closure. ARD
is generally characterized by high acidity and high concentrations of dissolved metals and sulfates. If
the generation of acid cannot be prevented or controlled, it must be treated to eliminate acidity and to
reduce the concentration of heavy metals and suspended solids before release to the environment.
Microorganisms play an important catalytic role in the generation of ARD. Several prevention, con-
trol, and treatment technologies that are based on microbiological activities have been developed. The
role of microorganisms and biotechnology in the generation of ARD and their role in the methods used
for prevention, control, and treatment are discussed in this paper. Recently developed passive and bio-
logical treatment processes that have been implemented at mine sites are emphasized in this paper.
Benefits, limitations, and design criteria for passive biological processes are also presented.
INTRODUCTION
Acid Rock Drainage

The management of waste materials, such as tailings and waste rock, from the mining of sul-
fidic metals, uranium, and coal is an environmental challenge. Acid generation occurs when
the sulfide minerals contained in the waste material, predominantly pyrite (FeS2) and pyrrho-
tite (FeS), are exposed to oxygen and water. The primary step is the oxidation of the sulfide
minerals and the generation of acid. Subsequently, the leaching of oxidized products occurs as
rainwater and snowmelt enters the waste pile or dump. If sufficient alkaline or buffering min-
erals (e.g., calcite) are not present to neutralize the acid, the resulting leach water becomes
acidic. This water is generally known as acid rock drainage (ARD) or acid mine drainage
(AMD). ARD is characterized by high acidity (low pH) and high concentrations of sulfate and
* Golder Associates Ltd., Ottawa, Ontario, Canada.
169
metals such as iron (Fe), manganese (Mn), aluminum (Al), zinc (Zn), copper (Cu), nickel
(Ni), lead (Pb), cadmium (Cd), and arsenic (As). The steps involved in the ARD generation
process, using pyrite as the example, can be represented by
FeS2(s) + 7⁄2 O2(g) + H2O → Fe2+ + 2SO42–(l) + 2H+ (EQ 1)
Fe2+ + 1⁄4 O2(g) + H+ → Fe3+ + 1⁄2 H2O (EQ 2)
FeS2(s) +14Fe3+ + 8H2O → 15 Fe2+ + 2SO42–(l) + 16H+ (EQ 3)
Fe3+ + 3H2O → Fe(OH)3(s) + 3H+ (EQ 4)

The overall sulfide to sulfate oxidation is summarized by
FeS2(s) + 15⁄4 O2(g) + 7⁄2 H2O → Fe(OH)3(s) + 2SO42–(aq.) + 4H+ (EQ 5)

As shown in Eqs. (1) and (2), the principle components required for the formation of ARD
are:
wastes containing reactive sulfides, including sulfide minerals (S2–or S22–), elemental
sulfur (So), and various sulfur intermediates (e.g., thiosalts, S2O32–);
molecular oxygen; and
water.
Other factors influencing the rate of acid generation include bacterial activity, temperature,
pH, the presence of alternate oxidants, e.g., ferric iron (Fe3+), and manganese (Mn3+ or
Mn4+) as illustrated by Eq. (3), and the presence of buffer or alkaline minerals (e.g., calcite
and silicates). The hydrolysis of ferric iron and precipitation of ferric hydroxide also produce
acid, as shown in Eq. (4). Chemical oxidation of sulfides by ferric iron or manganese and
hydrolysis reactions can take place under anoxic conditions. Equations (1) and (2) could
occur as a result of either an abiotic or microbially catalyzed chemical reaction. Microbial
oxidation reactions require the presence of oxygen. The role of microorganisms in the acid-
generation process is briefly discussed below.
In addition to waste rock and tailings, the sources of ARD from mining operations may
include underground mines, open pit mines, spoil piles, stock piles, and spent heap-leach
piles. The development of ARD is time dependent. At some sites, ARD may evolve over a
period of years after mine closure or may occur during the initial stages of mining operation.
If the formation of ARD cannot be prevented, it should be collected and treated. Otherwise, it
has the potential to contaminate groundwater and local watercourses, damaging the health
of plants, fish, and humans.
Chemical Composition of ARD

The chemical quality of ARD can vary widely because it is dictated by the physical, chemical,
mineralogical, and microbiological properties of each site. Physical site characteristics
include the size, density, and source (e.g., tailings and waste rock) of the waste material, as
well as the hydrological properties of the site. The type of mining operation (coal vs. metal or
uranium mining) may also have a large influence on ARD composition. Recently, however,
an industry government consortium on mine environment neutral drainage, MEND, in Can-
ada conducted a nation-wide survey and compiled the results in a report (MEND, 1994).
Studies conducted by Paine (1987) and Bhole (1994) also reported typical characteristics
and ranges for ARD. Reported characteristics and ranges of ARD for coal-mining operations
are summarized in Table 1. Usually, metal mining ARD contains a greater variety of metals,
such as copper, zinc, cadmium, arsenic, lead, nickel, etc., in elevated concentrations.
Dissolved salts other than sulfates (calcium, magnesium, etc.) may also be present in ARD.
The expected concentration range for total dissolved salts (TDS) is generally in the order of
100 mg/L to 30,000 mg/L. These levels can lead to a reduction in the quality of potable
170
TABLE 1 Reported characteristics of ARD from coal mine drainage (Paine, 1987)
Parameter Units Range
Flow rate L/s 57–157
pH — 1.4–6.2
Total acidity as CaCO3 mg/L 0–45,000
Fe mg/L 1–10,000
Al mg/L 1–2,000
Mn mg/L 1–50
SO4 mg/L 1–20,000
Total suspended solids (TSS) mg/L 5–3,000
surface and groundwater supplies, and, therefore, these compounds require removal/
reduction from the ARD before it reaches the natural environment or water resources.
Factors Affecting ARD Generation

The type of sulfur minerals, the presence of oxygen (air), and the quality and quantity of
alkaline minerals found in the waste are the primary factors that affect ARD generation. In
addition, temperature, pH, distribution of sulfide and alkaline materials in the waste pile,
and the surface area of sulfide minerals are important variables in controlling the rate of oxi-
dation and, hence, ARD generation. Because bacteria play a catalytic role, temperature and
pH appear to be the determining factors for microbial growth in mine-waste environments.
Differences in conditions found in tailings and waste rocks, such as particle size, surface area
of sulfide minerals, and the homogeneous distribution of sulfide and alkaline mineral content,
affect the rate of potential oxidation and neutralization process and, hence, the quality of ARD.
As the mean particle diameter in the waste-rock environment is typically greater than 200 mm,
tailing material is much finer, generally less than 0.2 mm (Broughton and Robertson, 1992;
Nicholson, 1994). The finer particle size has a larger surface area for sulfide minerals and
enhances the rate of oxidation. On the other hand, the finer and more homogeneous material
that facilitates the neutralization process within the tailing impoundment allows the alkaline
materials to be in close proximity to the acid generating sulfide minerals.
The oxidation of sulfide minerals is an exothermic process. Thus, a significant quantity of
heat is released, and the interior temperature encountered in mine waste piles can reach 80°C.
Microorganisms, mainly bacteria, as mentioned above, are indigenous to the ARD environ-
ment. These microorganisms play a role in the direct and indirect oxidation of sulfur minerals
(e.g., pyrite and pyrrhotite). Normally, sulfur-oxidizing bacteria are incorporated for extracting
metals from ores in metallurgical processes. Bioleaching is beyond the scope of this report and
will not be discussed. The concepts of ARD generation and bacterial oxidation of pyrite via
direct and indirect pathways are illustrated in Figure 1.
ROLE OF MICROORGANISMS AND BIOTECHNOLOGY
Microorganisms in ARD Production

It has been reported that a number of acidophilic or acid-tolerant bacteria, namely, Thiobacil-
lus, Leptospirillum, Sulfolobus, Sulfobacillus, and Metallogenium, are associated with the
mediation of acid generation from sulfide minerals at pH levels below 4. Most of these micro-
organisms are gram-negative, autotrophic mesophile and chemolithotrophs that exhibit high
tolerance to various metal ions and also to some anions such as arsenate.
The Thiobacillus species, except Thiobacillus ferrooxidans, are unable to oxidize iron, but
they can oxidize reduced sulfur compounds to sulfuric acid (H2SO4). T. ferrooxidans has been
171
FIGURE 1 Conceptual presentation of ARD generation (British Columbia Acid Mine Drainage Task
Force, 1989; Johnson, 1995)
the most documented microorganism having a role in the production of ARD. Thiobacillus
species are typically mesophiles with optimum growth observed between 25° and 35°C.
T. thiooxidans is the most acid-tolerant of the thiobacilli, with growth occurring over a pH
range of 0.5 to 4. T. denitrificans can use nitrate as a terminal electron acceptor for the oxida-
tion of reduced sulfur compounds at pH levels of 6.0 to 8.0 and can reduce nitrate (NO3–) or
nitrite (NO2–) to nitrogen gas (N2). T. novellus is a chemolithoheterotroph capable of growing
on inorganic and organic substrates.
Similarly to Thiobacillus bacteria, Leptospirillum bacteria are also gram-negative, meso-
philic, acidophilic, and obligate autotrophs. Although Leptospirillum ferrooxidans cannot oxi-
dize sulfur like its Thiobacillus counterparts, it is capable of oxidizing ferrous iron in acidic
environments at temperatures above 20°C. Sulfolobus sp. and Sulfobacillus sp. are thermo-
philic acidophiles capable of oxidizing both sulfur and iron. These bacteria can grow at pH
levels of 1 to 3 over a temperature range of 50° to 90°C.
Role of Biotechnology in ARD Prediction, Control/Prevention, and Treatment

Several research programs have been undertaken to understand the role of bacteria in the
generation of ARD. The results obtained from these investigations have helped in the devel-
opment of strategies to control ARD and methods to predict, prevent, and treat ARD. For
instance, a method where activities of T. ferrooxidans are enhanced was developed to predict
the potential for ARD generation from a given mine waste. The use of anionic surfactants has
been found to be useful in inhibiting the growth of T. ferrooxidans (US Environmental Protec-
tion Agency, 1995).
172
It has been revealed that creating anoxic conditions would not only inhibit the activities of
sulfur-oxidizing bacteria, but also would enhance activities of sulfide-reducing bacteria
(SRB), which could successfully be used in preventing and/or treating ARD. Several methods
to control, prevent, and treat ARD have been developed based on SRB. This paper discusses
the main biological processes that have been developed or are under development for ARD
prevention, control, and treatment.
ARD CONTROL/PREVENTION AND TREATMENT METHODS
Prevention/Control Methods
Mining industries face the challenge of preventing ARD occurrences in a cost-effective man-
ner. The most feasible method would be to control and prevent ARD at the source. The best
way to achieve this is to limit the oxidation of sulfide minerals and/or increase the portion of
neutralizing minerals in the mine waste. Several methods, including the use of soil (or dry)
and water (or subaqueous) covers, have been investigated to curtail oxidation of sulfidic
wastes (MEND, 1993, 1994). It has been demonstrated that soil covers, particularly multi-
layer methods, are effective. However, in many cases, they are not economically feasible
(Yanful and Nicholson, 1991).
Water covers have been shown to be an economical alternative to dry covers. Because of the
low oxygen-diffusion rate in water with respect to air, the oxidation of reactive wastes can be
minimized and water covers can be an effective long-term control method for acid generation.
However, the application of water covers is limited to site conditions. Site conditions with
respect to hydrology, topography, and the presence of a water source in the vicinity should be
suitable for the water-cover application (Dave, 1992; St-Germain and Kuyucak, 1998). In addi-
tion, although water covers can significantly reduce acid generation, a slow release of some
metals may still occur to the water column, resulting in an increase in some metal concentra-
tions, which may exceed regulated water standards. Thus, the water will require treatment
before its discharge to the environment.
The use of biological materials and the growth of aquatic plants in situ (“biologically sup-
ported water cover”) have been recommended to further improve the effectiveness of water
covers. The concept and the process will be discussed in detail below.
Prevention of ARD by Water Covers and Biological Methods

Because the effectiveness of a water cover is based on the solubility and diffusion rate of oxy-
gen in water, the rate of oxygen transport through water can be sufficiently slow so that acid
generation is insignificant. However, tests carried out at the Noranda Technology Centre in
Canada revealed that, although water covering unoxidized waste rock could reduce the acid
generation rate by more than 99.7%, concentrations of some metals, such as lead and zinc,
have increased to 4 and 6 mg/L, respectively (Aube et al., 1995). To further improve the effec-
tiveness of water covers, the feasibility of a layer of organic material on top of tailings, in addi-
tion to the water cover, was investigated (St-Germain et al., 1997; Beckett et al., 1998;
St-Germain and Kuyucak, 1998). The organic layer could be built up by growing aquatic
plants in situ in the water cover. Aquatic plants could consist of emergent, floating, and sub-
merged species. The investigations revealed that the benefit of the organic layer at the tailings
and water interface was two-fold:
Oxygen would be consumed in the organic material through bacterial activity, and, thus,
oxygen diffusion from the surface water to the tailings pore water would be prevented.
Upward metal fluxes from the tailings to the water cover would be prevented due to the
formation and precipitation of metal-sulfide complexes through the activity of sulfate
reducing bacteria (SRB). The organic material degrades to lower molecular organic com-
pounds during the consumption of oxygen by aerobic microorganisms, and anaerobic
173
FIGURE 2 The concept of biologically supported water cover
conditions at the interface are created. The presence of lower molecular organic com-
pounds and the anaerobic conditions enhance the growth of SRB, as explained below in
Eqs. (9) and (10). Additionally, metals form complexes with organic compounds, and
biosorption and bioaccumulation phenomena occur in the system. These processes fur-
ther help to retain metals at the tailings-water interface.
Furthermore, the presence of plants in the tailings impoundment, especially emergent plants
growing at the edges, improves the physical stability of the tailings through their roots. The
concept of a “biologically supported water cover” is illustrated in Figure 2.
CONVENTIONAL AND BIOLOGICAL METHODS TO MITIGATE ARD

If the occurrence of ARD cannot be prevented, the ARD must be treated by means of a chem-
ical and/or biological process to eliminate or minimize its impact on the environment. Acid-
ity is neutralized and metals are removed or reduced to acceptable levels to comply with
regulated water standards.
Heavy metals such as iron, zinc, copper, lead, cadmium, aluminum, and manganese,
which are found in acid mine waters, become insoluble and precipitate in solution at a cer-
tain pH level when they react with a chemical reagent to form a metal complex (Hedin et al.,
1994; Kuyucak, 1995). The type of chemical reagent, the pH of the water, oxidation/
reduction and hydrolysis reactions, the presence of biotic and abiotic catalysts, and the reten-
tion time of the water are important parameters that govern metal-removal processes.
Conventional and Chemical Methods

The neutralization of waters and the precipitation of metals with a neutralizing reagent (e.g.,
calcium or sodium hydroxide and calcium or sodium carbonate) in a treatment plant is a
common method because of its simplicity and relatively low cost. Neutralization/precipita-
tion processes are particularly feasible for treating large volumes of highly contaminated
waters. The precipitation of metals such as copper, zinc, cadmium, manganese, lead, and fer-
rous iron (Fe2+) require a pH greater than 9, which provides a low solubility for each of the
given metal ions. Ferric iron (Fe3+) and aluminum are hydrolyzed and precipitate out at pH
levels less than 5. In some cases, neutralization processes may be inefficient to reduce metals
to low levels as desired.
Lime neutralization/precipitation, referred to as the “chemical process,” is the most common
method used for treating ARD within the mining industry. The use of lime, as CaO or Ca(OH)2,
is often preferred over other alkaline reagents, particularly for treating ARD in large quantities,
due to its high reactivity and abundance (Kuyucak, 1998). In the lime-neutralization process,
metals and sulfate (SO4) are precipitated as metal hydroxide complexes and gypsum (CaSO4),
174
respectively. The precipitate, which is called “sludge,” is separated from the neutralized water
in ponds or through a mechanical solid/liquid separation device (e.g., clarifier/thickener) to
obtain clean water, and the sludge is stored in a controlled area. The mode of process applica-
tion may vary from simple to sophisticated, depending on the level of control and equipment
used in the process. The lime neutralization/precipitation process may pose some drawbacks. It
produces a large volume of sludge and may result in inefficient treatment, specifically for
removal of some metal ions to low levels.
The removal of metals to low levels can be achieved with the help of sulfide reagents (e.g.,
Na2S, H2S, and CaS) by forming sulfide compounds at pH levels less than 6. Metal concentra-
tions in the treated water are usually lower than those obtained with hydroxide precipitates.
Metal sulfide complexes offer some advantages over hydroxide precipitates. They are less
voluminous and are chemically more stable, being less susceptible to changes in pH as long
as they are stored under anaerobic conditions. However, due to the colloidal nature of metal-
sulfide precipitates, the treated effluent may require filtration to obtain low suspended solids
in the treated effluent. In addition, the sulfide-precipitation process is more expensive than
that of lime neutralization. Therefore, its application is limited to site-specific conditions.
The biologically generated sulfide-precipitation process has been investigated as an alterna-
tive treatment method.
Role of Microorganisms in ARD Mitigation

Microorganisms can be involved in ARD abatement, primarily through the reduction of met-
als and sulfates, as well as other alkalinity generating processes. The extent to which each
process may contribute to the neutralization of ARD depends upon the chemical composition
of ARD, the availability of necessary electron donors and acceptors, and the temperature and
pH conditions within the mine-waste environment.
Acidophilic heterotrophic bacteria present in the ARD environment may be involved in
ARD mitigation, as well as iron and manganese recycling. Some of these bacteria, such as
Acidiphilicum spp., play a passive role by metabolizing organic materials that are potentially
toxic to iron-oxidizing bacteria, thereby inhibiting biologically mediated iron oxidation reac-
tions (Johnson, 1995). Other species demonstrated the ability to reduce the iron, present
either as soluble or as solid-phase compounds, to ferrous iron.
Manganese and iron reduction may also contribute to the neutralization process. Micro-
organisms, including the heterotrophic bacteria Pseudomonas, Clostridium, and Desulfovibrio,
can directly reduce manganese and iron by using them as final electron acceptors under
anaerobic conditions. The ability to reduce ferric iron is widespread among acidophilic het-
erotrophic bacteria and has been reported for Thiobacillus ferrooxidans growing on elemental
sulfur. When ferric iron is reduced to ferrous iron, the removal of iron from ARD becomes
easier, because ferrous iron reacts with sulfide produced by sulfate reduction, and this, ulti-
mately, results in the removal of iron and promotes alkalinity generation. Sulfate reduction
leads to permanent alkalinity production when hydrogen sulfide gas (H2S) is released from
the mine-waste environment.
Other biologically mediated processes that can contribute to ARD neutralization by ulti-
mately consuming H+ ions include:
ammonification by various microorganisms and enzymes;
denitrification, where a number of bacterial species, such as Pseudomonas, Paracoccus,
Flavobacterium, Alcaligenes, and Bacillus spp, convert ammonia to nitrates under anaer-
obic conditions; and
methane generation by methanogenic bacteria.
Discussion regarding the processes involved in sulfate reduction follows.
175
PASSIVE-BIOLOGICAL ARD TREATMENT METHODS
These processes have, in some circumstances, proven to be feasible alternatives to conven-
tional lime-neutralization/precipitation and sulfide-precipitation methods (Hedin et al.,
1991; Blowes et al., 1995). Properly functioning passive treatment systems can produce com-
pliance level effluents with no additional costs other than the initial construction and limited
periodical maintenance. Even if supplementary chemical treatment is required to meet efflu-
ent limits at some sites, it still can be cost effective. They are particularly ideal for decommis-
sioned sites and for the treatment of seepage where the temperature, flow rate, and chemical
composition do not fluctuate and remain optimal all year round. Alkalinity required for
removing acidity and metals can be generated by the following two naturally occurring pro-
cesses: the dissolution of limestone or other carbonate rocks; and bacterial sulfate reduction,
which also generates sulfide to precipitate/remove metals.
The most common passive-treatment systems are anoxic limestone drains, constructed
anaerobic and aerobic wetlands and biosorption. The performance of individual systems is a
function of both quality and quantity of the raw mine drainage. Experience has demon-
strated that the influent flow rate, contaminant concentrations, pH and alkalinity (or acidity)
are all extremely important to system performance. Moreover, in addition to temperature,
the capacity of the biological treatment system is significantly affected by the changes in pH.
Anoxic Limestone Drain

Anoxic limestone drains (ALD) treat ARD having a net acidity by adding alkalinity to the
water. The system consists of a bed of crushed limestone (CaCO3), which is installed below
the ground surface and covered with a fabric filter and clay (or other materials) to promote
anoxic conditions (Brodie et al., 1992). ARD flow through the ALD is gravity driven, allowing
the acid component of the water to react with the limestone and release carbonic acid
(Brodie et al., 1991; Brant et al., 1995), as was presented in Eq. (1). The carbonic acid then
reacts with the limestone to produce bicarbonate (HCO3–) alkalinity and, consequently, the
pH level in the water is increased as follows
CaCO3 + 2H+ → Ca2+ + H2CO3 (EQ 6)
CaCO3 + H2CO3 → Ca2+ + 2HCO3– (EQ 7)

When the pH in the water increases to above 6.4, bicarbonate becomes the dominant dis-
solved carbonate species. At this stage, the acidity reacts with limestone, generating alkalin-
ity directly according to
CaCO3 + H+ → Ca2+ + HCO3– (EQ 8)

The degree of dissolution is enhanced in anaerobic environments because of a decrease in
the formation of ferric iron precipitate (which armors carbonate surfaces) and, because of an
increase in the partial pressure of dissolved CO2, which directly affects the solubility of car-
bonate compounds (Hedin and Nairn, 1992). Development of high CO2 partial pressures
increases the alkalinity beyond that possible under normal atmospheric conditions (Kilborn,
1996). The size of the limestone particles compromises the quantity of free flow and the
presence of sufficient surface area for dissolution. Faulkner and Skousen (1994) suggested
that the larger particle sizes (80 to 250 mm), along with the finer particles (20 to 40 mm)
should be used in an ALD to increase the hydraulic conductivity and reduce the potential for
plugging. Mine water with low dissolved oxygen (O2), ferric iron (Fe3+), and aluminum is
ideally suited for the treatment with ALD, based on the results reported by US Bureau of
Mines (USBM) for the treatment of ARD at coal mines in the United States.
176
The waters with sulfate concentrations greater than 2,000 mg/L should also be treated
with care to prevent CaSO4 (gypsum) formation and its potential precipitation in the system.
For instance, partial neutralization of ARD and/or operation of the system with high velocity
lowers the risk of plugging and system failure.
Microbiological Sulfate-Reduction Principles of Sulfate Reducer Bacterial Processes

A group of bacteria called sulfate reducers (SRB), such as Desulphovibrio spp, can convert
sulfate contained in ARD to sulfide and can generate bicarbonate (HCO3–) in the presence of
organic carbon (nutrient) sources using it as an electron donor under anoxic and reducing
conditions. Sulfate reduction first produces hydrogen sulfide (HS–), which attracts free
hydrogen ions (H+) and produces hydrogen sulfide (H2S). The hydrogen sulfide generated
forms insoluble metal complexes and results in the removal of metals. The bicarbonate
released results in an increase in alkalinity (i.e., pH). The reactions are explained by
2CH2O + SO42– → HS– + 2HCO3– + H+ (EQ 9)
Me + H2S → MeS (EQ 10)

where
CH2O represents the organic matter and
Me represents the heavy metal.
SRB are known to be natural soil bacteria and can be found in soils, sewage sludge, and
manure. They require low-molecular-weight organic carbon compounds (e.g., simple organic
acids), suitable concentrations of sulfate (>200 mg/L), a pH level greater than 4.5, and low Eh
levels (<–150 mV). SRB can function in the absence of oxidizing agents such as O2 and Fe3+.
Low-molecular-weight carbon compounds (e.g., lactic acid and acetate) used by SRB are com-
mon products of natural degradation (i.e., microbial fermentation) processes, which occur in
anoxic environments (Wetland, 1992; Bechard et al., 1994; Kuyucak and St-Germain, 1994a,
1994b; Eger et al., 1997). A variety of materials, depending on their cost and availability, can
be used as nutrients. They may include industrial wastes such as molasses, sewage sludge,
compost, manure, wood chips, wood or paper pulp, brewery waste, and hay. Natural organic
materials provide slow-release, long-term nutrient sources to SRB. The materials can be sup-
plemented with materials containing nitrogen and/or phosphorus to obtain the optimal nutri-
ent (i.e., substrate) composition required. Bacteria found in natural organic substrates other
than SRB utilize oxygen to degrade the material and, in turn, create anoxic conditions
required for the SRB.
The pH requirements are obtained by the alkalinity generated by microbial activity and
carbonate dissolution. Temperature significantly affects the rate of microbiological pro-
cesses. At low temperature levels (<10°C), the reaction rate decreases more than 50% of that
which could be obtained at 20°C. The average sulfate reduction (or H2S generation) rate has
been found to be 0.3 mol/m3 of the nutrient per day at temperature levels higher than 10°C
in the tests conducted where agricultural wastes were used as nutrients (Kuyucak and
St-Germain, 1994). However, because the average temperature found in the in-seepage and
in the groundwater is known to be above 10°C year-round and because it does not fluctuate
seasonally, temperature requirements do not pose serious limitations for treatment of these
water bodies.
The overall process results in an improvement in the water quality due to the precipitation
of metals as sulfides, with the H2S generated in organic substrate and neutralization of the
acidity due to the bicarbonate released during sulfate reduction. This process is particularly
good for the reduction of heavy metals such as cadmium, copper, lead, mercury, zinc, and
iron to low concentrations.
177
BIOLOGICAL AND SRB-BASED TREATMENT PROCESSES
The SRB can be used as a cost-effective alternative to the conventional lime-neutralization
process for treating ARD. The function of SRB can be used in either wetland system, i.e., in
the bioreactive walls which are placed at the leading edge of an underground acid plume or
in bioreactors operated under controlled conditions. The SRB-based passive processes are
able to treat seepage, small streams and acid water accumulated in open pits or lagoons at
mine sites.
Constructed Wetlands
Wetlands are composite systems where a variety of physical, chemical, microbiological, and
plant-mediated processes occur that can cause significant changes in the water chemistry.
Investigations on natural wetlands indicated that wetlands have a significant capacity for
removing trace metals and improving low pH values (MEND, 1990; Pett et al., 1990;
Wheeler et al., 1991). Alkalinity generation and improvements in metal and acidity removal
enhanced by the addition of organic amendments and nutrients to the natural systems and
directing the flow path through the system was observed.
The general removal mechanisms that were identified in both natural and constructed
wetlands include (Cohen, 1996):
the oxidation and precipitation of metal oxides,
the adsorption and complexation of metals by organic substrates,
sedimentation,
filtration of suspended and colloidal particles,
active plant uptake, and
microbial sulfate reduction followed by precipitation of metal sulfides.
Plant growth and decay in the wetland provides a constant source of organic substrate. The
organic matter provides ion exchange and adsorption sites while stimulating a consortium of
bacterial activities that promote a number of biotic and abiotic processes. Wetland vegetation
also provides attachment sites for bacterial growth and flow channels, thereby, increasing
contact with the microorganisms and nutrients and promoting good flow distribution within
the system.
In most wetlands where a free water surface is present, aerobic conditions exist through-
out the water column, and anaerobic conditions develop mainly below the sediment surface.
Metals are removed in the aerobic zones by oxidation, precipitation, adsorption, and com-
plexation reactions, while neutralization is primarily achieved in the anaerobic zones by SRB
activity and by the increase in alkalinity associated with other chemical and microbial reac-
tions, such as the dissolution of limestone beds.
In recent years, wetlands have been constructed for the treatment of ARD. Constructed
wetlands can be classified as aerobic, anaerobic, or combined wetlands. Sequential (or
staged) treatment using a combination of aerobic and anaerobic wetland cells was found
effective in generating alkalinity and attenuating ARD (Kepler and McCleary, 1994). An
example of combined passive treatment systems, including aerobic and anaerobic con-
structed wetland cells is illustrated in Figure 3.
Anaerobic Wetlands. Anaerobic wetlands are generally employed to treat drainage
waters that contain elevated concentrations of iron and aluminum, have a total acidity greater
than 300 mg/L as CaCO3, or have a pH of less than 4 (Hedin et al., 1994). Anaerobic wetlands
may also be referred to as compost wetlands or subsurface wetlands. Subsurface flow through
a 300- to 450-mm deep organic substrate under anaerobic conditions promotes chemical and
microbial processes, which generate alkalinity. The substrate usually consists of a low-cost,
high organic-content material, such as mushroom compost, sawdust, manure, leaf mulch, hay,
178
FIGURE 3 An example of combined passive treatment systems
or peat, which acts as a nutrient source for reducing bacteria. A neutralizing source such as
limestone is also generally included below the substrate layer (Kleinmann and Hedin, 1993).
Aerobic Wetlands (Marshes). Aerobic wetland cells are typically 150- to 450-mm-
deep water columns designed to maximize oxidation reactions that cause the precipitation of
metals, primarily iron and/or manganese, as oxides and hydroxides. The precipitates are
then incorporated into the wetland sediment. This type of wetland is most effective for drain-
age waters that are alkaline in nature (Hedin, 1996). As dissolved metals are oxidized, a net
increase in acidity is generally observed in these systems due to the release of H+ ions and/or
consumption of alkalinity associated with oxidation reactions. Thus, the use of an aerobic
wetland followed by an ALD system has been highly recommended if the quality of ARD is
suitable for the use of an ALD system.
Most aerobic wetlands (or marshes) contain plants growing in a clay or soil substrate. The
dense stands of reeds growing in these systems serve as a support for bacteria and algae and
act as a hydraulic barrier to reduce the flow of the water and enhance its residence time. In
aerobic wetlands, several oxidation reactions take place. As a result of these reactions, metals
(e.g., Fe and Mn) precipitate as oxides and hydroxides, which provide adsorption sites for
metal ions, and the organic material decomposes (Hedin et al., 1994). Metals can form com-
plexes with organic materials becoming immobilized and consequently are retained in the
system. The wetland also acts as a filter and enhances the settling of suspended solids.
An aerobic wetland system was recently adapted to a ditch treatment method at mine sites
located in the United States and in Quebec, Canada (Eger et al., 1997; Kuyucak, 1998). An
open ditch was constructed around waste rock piles, and the seepage originating from the
waste rock dumps was collected into it. The ditch was furnished with a layer of limestone
and peat obtained from a nearby bog area. Stone berms were also installed in the ditch to
control the flow of the ARD. The ARD remained in contact with atmospheric air, and the iron
species found in the ARD was oxidized to ferric iron and was precipitated in the ditch. A
small ALD system was installed at the outlet of the ditch to neutralize the acidity of the
water, which consisted predominantly of protons (H+) that resulted from the oxidation,
hydrolysis, and precipitation of iron. No plant transplantation or growth was involved in
these systems. The aerobic ditch treatment system is illustrated in Figure 4.
Sulfate Reduction and Bioreactor Systems

Bioreactor treatment systems are designed to optimize wetland treatment systems without
the presence of wetland plants. They are strictly dependent on bacterial activities. In bio-
chemical reactor applications where the process parameters such as pH, temperature, anaer-
obic, nutrients, and SRB are controlled, a short-chain organic acid or fatty acid is used as
substrate. For instance, Budelco Mining and Smelter in the Netherlands had evaluated the
effectiveness of several processes (e.g., SRB, ion exchange, and membrane) with laboratory
and pilot-scale tests. They found that the microbial sulfate reduction was the most cost-
effective method for the treatment of groundwater containing trace metals (e.g., Zn, Cd, and
Cu) and sulfate. At the site, a reduction in sulfate concentrations from 1,500 mg/L to less
than 500 mg/L was required to comply with the regulated standards. This could be achieved
179
FIGURE 4 Schematic presentation of the aerobic ditch treatment system
with the microbial sulfate reduction process. In the full-scale process operation, ethanol is
used as the nutrient for the bacteria, and 250 to 400 m3/hr is treated in up-flow anaerobic
blanket reactors. All the process parameters are controlled and adjusted to optimal levels.
Excess H2S, which is left after the precipitation and separation of trace metals from the sys-
tem, is converted to elemental sulfur with the help of photosynthetic sulfur oxidizing bacteria.
In passive processes, biodegradable substrates are used as nutrients, and the efficiencies of
these processes are dependent on a consortium of bacteria rather than SRB alone. Biodegrad-
able substrates are metabolized, and organic acids are produced that become available for
the growth of SRB. As in wetland systems, the microbial activity within the bioreactors can
be supplemented with inorganic chemical reactions such as pH neutralization via limestone
dissolution or other neutralizing reagents (Eger et al., 1997). The bioreactors rely on several
microbial reactions, which require different levels of oxygen to treat ARD. A cellulose-based
material must be supplied to the system on a periodic basis or when it has been consumed.
The substrate is degraded by cellulolytic bacteria, which generate free sugars and other
metabolites. Aerobic and facultative heterotrophs can then further metabolize these products
to provide substrates for the growth of fermentative anaerobes. Under anaerobic conditions,
the free sugars are fermented to short-chain organic acids or short-chain fatty acids, which
are suitable substrates to support the growth of the SRB. The SRBs then reduce sulfate to
hydrogen sulfide, which then reacts with dissolved metal ions forming low solubility metal
sulfide precipitates. The SRBs concurrently consume hydrogen ions and produce carbon
dioxide that generates the alkalinity, thereby, increasing pH levels. This is due to the reduced
concentration of hydrogen ions and the buffering effect of bicarbonates. Other biologically
mediated reactions such as ammonification, metal reduction, and methane production can
also contribute to the alkalinity generation depending upon the chemical composition of
ARD, as mentioned above.
Bioreactor applications that have been investigated for ARD mitigation have included open
pit, underground mines, biotrench, and a series of cell configurations, e.g., acid reduction
using microbiology ARUM process (Fyson et al., 1995). These processes involve the addition
of SRB inoculum along with the necessary substrate and nutrients to the acidic drainage. Due
to activities of a consortium of bacteria, anaerobic and reducing conditions are generated in
situ in the open pit, in trenches, or in cells. These, then, promote the growth of SRB, resulting
180
in alkalinity generation, acidity reduction, metal precipitation, and, hence, ARD mitigation
(Kuyucak and St-Germain, 1994a, 1994b; Bechard et al., 1995; Fyson et al., 1995).
In the biotrench and cell applications, the cells are physically separated by a clay wall, and
the drainage is allowed to flow slowly through the system. Various organic substrates, such
as wood shavings, straw, manure, and alfalfa, are used in these systems. A bag of organic
substrate was prepared and suspended in an abandoned mine shaft with cables for the miti-
gation of acid mine water in place (Canty, 1998). Except during the spring runoff, the pro-
cess is capable of producing a high-quality effluent.
Bioreactive Walls
The use of sulfate-reducing reactive walls has been proposed for those waters affected by sul-
fide mineral oxidation, which migrate downward to aquifers underlying the mine-waste
environment. These can be transported through the groundwater system, eventually dis-
charging into a surface-receiving stream (Blowes et al., 1995; Benner et al., 1998). A portion
of the existing aquifer is excavated, and the original material is replaced with an organic sub-
strate. The porosity of the wall is an important factor; the wall should be sufficiently perme-
able (e.g., 10–3 cm/s) to allow water to flow through it (Waybrant et al., 1995).
A field test was conducted in 1995. In this test, mine drainage flowing within an aquifer
was intercepted and treated using a 15-m-wide, 4-m-thick, 3.6-m-deep wall. The design rate
of flux through the wall was 288 m3/yr (Blowes et al., 1997). The results revealed that the
downstream sulfate concentrations had been reduced by 50%, as iron concentrations were
reduced by 95% (Benner et al., 1998). The pH was raised from 5.8 to 7.0, coinciding with an
increase in alkalinity from 0 to 50 mg/L as CaCO3.
Biosorption and Bioaccumulation Systems

Materials of biological origin that have the ability to remove metal ions from aquatic solu-
tions are called “biosorbent.” Sorption processes that play the role in removing metals are
referred to as “biosorption.” Usually, biosorption refers to the removal of metals by nonliving
biomass, and bioaccumulation refers to the removal of metals by living cells. The main sorp-
tion mechanisms involved include surface adsorption, ion exchange, complex formation,
and/or precipitation with either living or nonliving biomass. These processes exhibit poten-
tial as ARD abatement methods (Kuyucak, 1987).
Metals Removal via Living Cells (Bioaccumulation). Wetlands for ARD mitigation
were initially designed to imitate natural Sphagnum peat moss wetlands, which exhibit rela-
tively high adsorption capacities for heavy metals. Investigations conducted on a white cedar
bog in Minnesota revealed that the natural peat land removed 80% of the nickel and nearly
100% of the copper from a tailings drainage (Eger et al., 1980; Eger and Lapakko, 1988).
Removal by the peat accounted for more that 90% of the total metal reduction. In Canada,
Kalin et al. (1990) reported that acidic drainage, which contained elevated heavy-metal con-
centrations and low pH levels, emanating from a coal mine in Nova Scotia caused considerable
damage to a peat bog’s natural vegetation, with the exception of Typha and Sphagnum species.
It has been found that Sphagnum moss generally accumulates metals to toxic levels after sev-
eral months of exposure to mine drainage. However, Sphagnum moss is not readily available
and is difficult to transplant (Hedin and Nairn, 1993). A laboratory study indicated that the for-
mation of organically bound iron was the principle phenomenon for the removal of dissolved
iron from a coal-mine drainage (Tarleton et al., 1984). The high metal sorption capacity of peat
(living or nonliving) has been attributed to the presence of humic substances containing car-
boxyl, amino, quinone, hydroxyl, and other functional groups. These functional groups are
responsible for surface adsorption, ion exchange, chelation, and complex formation with met-
als in solutions.
The underwater meadow of macrophytic algae chara, Nitella flexilis, was evaluated for its
capability to treat mine drainage from a uranium mine (Smith and Kalin, 1989). The algae act
181
as a filter, which effectively transfers the contaminant radium (Ra 226) and, to a lesser extent,
uranium (U) from the water to the sediment during their continuous growth and decay.
The use of microbial mats immobilized on glass wool or wire mesh was examined in con-
structed ponds to reduce concentrations of dissolved metals such as zinc, manganese and
iron from mine drainage (Vatcharapijarn et al., 1994; Phillips et al., 1995). Microbial mats
are composed of heterotrophic and autotrophic microorganisms, mainly blue-green algae
Oscillatoria sp., green filamentous algae, and Chromatium sp.
It has been reported that the microorganisms are held together by slimy secretions and that
extracellular metal deposits could be responsible for minimizing toxic effects in the biological
system. Because these mats are both nitrogen fixing and photosynthetic, they are self-sufficient
and solar driven with few growth requirements. As well, microorganisms have been genetically
engineered to sequester metals such as gold, silver, platinum, cadmium, cobalt, copper, ura-
nium, and mercury, mainly for the purpose of metal extraction and recovery (Smith et al.,
1994; Ledin and Pedersen, 1996). However, in spite of several pilot studies conducted under
low flow rates of 2 to 5 L/min, these processes have not yet been tested on a large scale.
Metals Removal via Nonliving Cells (Biosorption). The drawbacks associated with
the use of living cells, such as metal toxicity, adverse climatic conditions and costs of nutrient
supply and culture maintenance, can be avoided with the use of nonliving biomass as biosor-
bents (Kuyucak, 1987; Kuyucak and Volesky, 1989; Kilborn, 1996).
Dried nonliving biomass such as marine algae (Ulva sp.), blue-green algae (Spirulina sp.),
yeast (Saccharomyces cerevisiae), common duck weed (Lemna sp.), and finely ground peat
(Sphagnum peat moss) have been immobilized by blending them into a high-density polysul-
fone dissolved in an organic solvent.
BIO-FIX beads are fabricated from the immobilized biomass and are commercially used to
reduce dissolved metal concentrations, including arsenic, cadmium, copper, silver, lead,
manganese, and zinc from the mine waste environment (Jeffers et al., 1989; Bennett et al.,
1991; Kilborn, 1996). BIO-FIX beads are generally enclosed in meshed polypropylene bags
and are placed in a trough directly in an ARD stream or in conventional equipment (e.g., a
tank). The beads are particularly effective in adsorbing metals from dilute ARD streams.
Biosorbents such as marine and fresh algae are also employed for the recovery of metals,
including silver, copper, cobalt, mercury, or cadmium from aqueous solutions (Kuyucak and
Volesky, 1989). An immobilized algae, mainly Chlorella vulgaris, which commercially is called
AlgaSORB, has been found effective for treating groundwater (Darnall et al., 1989). The lev-
els of metal ions can be reduced in concentration from low ppm levels to low ppb levels.
Another commercial biosorbent (MediaflexMC), consisting of Sphagnum peat moss and a
carbon material, has been reported to be effective in treating various types of wastewaters,
including ARD (Belanger et al., 1995). MediaflexMC was able to treat drainage from an aban-
doned copper mine in a basin with a capacity of 6,000 m3 using flow rates in the order of
250 m3/d. MediaflexMC acts as a biofilter having adsorptive and precipitation properties and
can remove Fe, Zn, Cu, Ni, Cr, and Cd.
Chitosan (a natural biopolymer from chitin) and calcium-alginate have been investigated
to remove cadmium, barium, uranium, and zinc (Kilborn, 1996). Forest products and cellu-
loytic materials such as bark, sawdust, and litters also show metal sorption capacity
(Kuyucak and St-Germain, 1993, 1994).
When the biosorbent in use becomes saturated, it is replaced with fresh material. Metals
can be stripped from the metal-laden biosorbent and can be recovered. The biosorbent can be
used several times or it can be disposed of into a safe environment.
POTENTIAL BENEFITS AND SUITABILITY OF PASSIVE TREATMENT

With the placement of the cover over the tailings, the quantity and quality of the seepage origi-
nating from the tailings pond will become manageable for passive in situ treatment processes.
182
Potential benefits include the following:
passive systems are a sustainable alternative to the expensive chemical methods;
passive systems form a potentially low maintenance option with low requirements for
energy and material input compared to chemical treatment plants;
passive systems can protect groundwater resources and remediate contaminated
plumes, and they can keep constituents immobilized in the system;
passive systems incorporate a range of physical, chemical, and biological processes that
can reduce metal concentrations to very low levels similar to the chemical sulfide pre-
cipitation process;
passive systems eliminate sludge collection and disposal concerns; and
passive systems use natural materials rather than chemical reagents, and they are aes-
thetically attractive with consequent “green” appeal (i.e., wetland can eventually
become a valuable wildlife refuge).
Examples to Passive Treatment Field Applications

The technology is relatively new and has been undergoing continuous development.
Although there are several field-scale applications, information related to their long-term
performance is limited. The field-scale process applications, including pilot tests, are summa-
rized as follows:
ALD Systems. Over the last decade, it has been reported that more than 30 ALD sys-
tems have been installed in the United States (Brodie et al., 1992; Hedin and Nairn, 1992).
Usually, ALD systems have been combined with an aerated wetland and have been used for
the removal of acidity and metals (mainly Fe and Mn).
SRB-Based Treatment Systems. The SRB-based processes are still under develop-
ment. There are some examples of pilot-scale applications. The technology has mainly been
incorporated with mine closure plans to eliminate maintenance requirements and the poten-
tial for future environmental liability and/or remedial action. Examples include:
suspended organic nutrients were placed in an abandoned mineshaft in Montana to
treat ARD in place (Canty, 1998);
a reactive wall (or barrier) has been implemented in the field for the treatment of seep-
age originating from a nickel mine in Sudbury, Canada (Blowes et al., 1998);
an anaerobic wetland has been constructed to treat seepage from heap-leach ponds in
Nevada (Homestake Mining); and
the treatment of seepage from waste-rock stockpile in Minnesota (LTV Steel Mining
Co.) with a combination of limestone-peat substrate and an aerated wetland (for five
years, the operation has been successfully treating 19 to 38 L/sec of flow containing Cu,
Ni, and Fe with a pH of 5) (Eger et al., 1997).
Aerated Wetlands. Most of the full-scale wetlands constructed for the treatment of
mine waters have consisted of aerated wetlands and lagoon components located in existing
stream channels. Several have been successfully used in Australia, South Africa, and the
United States. The performance of these systems is closely associated with the near-neutral
pH and low concentrations of metals in waters.
The inclusion of riffle zones and algae in addition to other aquatic plants has been found to be
effective in removing metals, particularly Mn. It has been reported that waters with flow rates up
to 3,000 m3/day (i.e., 35 L/sec) can be treated. Some examples include the removal of:
acidity, Zn, Fe, Mn, and thallium at Hilton Mine, Australia;
uranium at Ranger Mine, Australia;
183
arsenic, Zn, Cu, Cd, Pb, Fe, and Mn, at Tom’s Gully Gold Mine, Australia;
Cd, Pb, Cu, and Zn at Woodcutters Mine, Australia; and
Mn, Pb, Fe, and SO4 from the tailings dam at Hellyer Mine, Australia.
LIMITATIONS OF BIOLOGICAL PASSIVE TREATMENT SYSTEMS

Although passive treatment systems offer several benefits, their application is limited to site
conditions. Except for the ALD systems, their performance is highly dependent on tempera-
ture. At temperatures lower than 10°C, they do not function well (Kuyucak and St-Germain,
1994a, 1994b). Their performance is limited and dependent on the flow rate and the chemi-
cal composition of water being treated. These systems cannot handle high loading situations
and fluctuations in both flow rate and chemical composition. It has been reported that, dur-
ing spring runoff, an SRB system suspended in a mineshaft could not produce the same water
quality as it could before the spring runoff took place. The system was capable of recovering
to normal condition after a few months (Canty, 1998). An even flow distribution and the
contact of acid water with bacteria in the passive system are important factors affecting the
efficiency. Channelling in the system often occurs and causes inefficient use of the system
and, consequently, inefficient treatment of the mine water. The presence of alkalinity, sul-
fate, and nutrients in sufficient quantities and anaerobic and reducing conditions in the sys-
tem are especially necessary for the SRB based processes. The removal of Mn cannot be
achieved with an anaerobic process alone.
The performance of ALD depends mainly on the chemical composition of the mine water.
For instance, the presence of heavy metals, such as copper, ferric iron and sulfate in high con-
centrations, and dissolved oxygen in the acid water deteriorates the performance of ALD
because these compounds cause precipitation and clogging in the system.
The cost of passive systems is highly dependent on the site conditions, chemical composi-
tion of the acid water, and treatment requirements. In some cases, the application of passive
systems may be more expensive or as expensive as a chemical treatment option (Hedin et al.,
1994; Kuyucak and St-Germain, 1994). The size of the system, the simplicity of construction,
and the quantity and availability of the nutrients required dictate the cost of the process.
Therefore, the use of locally available materials as nutrient or organic substrate in these sys-
tems has been recommended.
In most cases, one method alone is not sufficient to achieve the treatment requirements at
the site. A combined passive system such as ALD with an aerobic and/or anaerobic wetland
could provide a continuous means of treatment to the site. Due to its constant flow rate, chem-
ical composition and temperature, passive treatment systems (e.g., reactive walls) have been
suggested as ideal for the treatment of seepage of wastewater.
ACKNOWLEDGMENTS
The author would like to express her appreciation to the staff at Golder Associates Ltd.,
Ottawa Office, for their support and for the encouragement that they provided during the
preparation of this manuscript. Specifically, the author would like to thank Sarah-Jane
Mosher and Ken Taylor for their assistance in editing the text and providing the figures for
this report, respectively.
REFERENCES
Aube, B.C., St-Arnaud, L.C., Payant, S.C., and Yanful, E.K., 1995, “Laboratory evaluation of the
effectiveness of water covers for preventing acid generation from pyritic rock,” In Proceedings
of Sudbury ’95: Mining and the Environment, May 1995, Sudbury, Ontario, Canada, Vol. 2,
pp. 495–504.
184
Bechard, G., 1993, “Microbiological Process for the Treatment of Acid Mine Drainage Using
Cellulosic Substrates,” Ph.D. Thesis, Carleton University, Ottawa, Canada.
Bechard, G., Yamazaki, H., Gould, D., and Bedard, P., 1994, “Use of Cellulosic Substrates for the
Microbial Treatment of Acid Mine Drainage,” Journal of Environmental Quality, Vol. 23, No. 1,
pp. 111–116.
Bechard, G., McCready, R.G.L., Koren, D.W., and Rajan, S., 1995, “Microbial Treatment of Acid
Mine Drainage at Halifax International Airport,” In Proceedings of Sudbury ’95: Mining and the
Environment, May 1995, Sudbury, Ontario, Canada, Vol. 2, pp. 545–554.
Belanger, M.-C., Ross, A., and Shoiry, J., 1995, “Traitement des eaux minieres acides du site Solbec-
Cupra par le procede de filtration MediaflexMC,” Biotechnology and the Mining Environment, In
Proceedings of the Eleventh Annual General Meeting of BIOMINET, pp. 110–117.
Bennett, P.G., Ferguson, C.R., and Jeffers, T.H., 1991, “Biological treatment of acid mine waters—
case studies,” In Proceedings of the Second International Conference on the Abatement of Acidic
Drainage, September 1991, Montreal, Quebec, Canada, Vol. 1, pp. 283–299.
Bhole, A.G., 1994, “Acid mine drainage and its treatment, Impact of Mining on the Environment,”
pp. 131–141.
Blowes, D.W., Ptacek, C.J., Cherry J.A., Gillham, R.W., and Robertson, W.D., 1995, “Passive
remediation of ground water using in-situ treatment curtains,” In Proceedings of
Geoenvironment 2000 Conference, Y.B. Acar and D.E. Daniel, eds., New Orleans, LA, Vol. 2, pp.
1590–1607.
Blowes, D.W., Ptacek, C.J., Benner, S.G., Waybrant, K.R., and Bain, J.G., 1998, “Porous reactive
walls for the prevention of acid mine drainage: a review,” In Mineral Processing and Extractive
Metallurgy Review, F.M. Doyle and N. Arbiter, eds. Montana, August 1996, pp. 25–38.
Brant, D.L., Ziemkiewicz, P.F., and Svonavec, J., 1995, “Anoxic limestone foundation drains for
AMD control in coal refuse piles,” In Proceedings of Sudbury ’95: Mining and the Environment,
May 1995, Sudbury, Ontario, Canada, Vol. 2, pp. 537–543.
British Columbia Task Force, 1989, “Draft Acid Rock Drainage Technical Guide,” Steffen
Robertson and Kirsten Inc., Vancouver, British Columbia, Canada, August.
Brodie, G.A., Britt, C.R., Tomaszewski, T.M., and Taylor, N.H., 1991, “Use of passive anoxic
limestone drains to enhance performance of acid drainage treatment wetlands,” Meeting of the
American Society of Surface Mining and Reclamation, Durango, CO, May 1991, pp. 211–228.
Brodie, G.A., Britt, C.R., Tomaszewski, T.M., and Taylor, H.N., 1992. “Anoxic lime drains to
enhance performance of aerobic acid drainage treatment wetlands,” Presented at the 1992 West
Virginia Surface Mine Drainage Task Force Symposium, Morgantown, WV, April 1992.
Broughton, L.M., and Robertson, MacG., 1992, “Acid rock drainage from mines—Where are we
now,” Minerals, Metals and the Environment Conference, February, Manchester, England.
Canty, M., 1998, “Overview of the sulfate-reducing bacteria demonstration project under the mine
waste technology program,” In Mineral Processing and Extractive Metallurgy Review, F.M. Doyle
and N. Arbiter, eds., Montana, August 1996, pp. 61–80.
Cohen, R.R.H., 1996, “The technology and operation of passive mine drainage treatment
systems,” In Managing Environmental Problems at Inactive and Abandoned Metal Mine Sites,
Chapter 5, EPA Seminar Publication, USEPA/625/R–95/007, pp. 18–29.
Darnall, D.W., McPherson, R.M., and Gardea-Torresday, J., 1989, “Metal recovery from geothermal
waters and groundwaters using immobilized algae,” Biohydrometallurgy, pp. 341–362.
Dave, N.K., 1992, “Water Cover on Acid Generating Mine/Mill Wastes: A Technical Review,”
CANMET Report, MSL 89–107, Ottawa, Ontario, Canada.
Eger, P., Lapakko, K., and Otterson, P., 1980, “Trace metal uptake by peat: interaction of a white
cedar bog and mining stockpile leachate,” In Proceedings of the Sixth International Peat
Congress, Duluth, MN, August 1980, pp. 542–547.
Eger, P., Wagner J.R., and Melchert G., 1997, “The use of a peat/limestone system to treat acid
rock drainage,” In Proceedings of the Fourth International Conference on Acid Mine Drainage,
Vancouver, Canada, May 1997, pp. 1195–1209.
185
Faulkner, B.B., and Skousen, J.G., 1994, “Treatment of acid mine drainage by passive treatment
systems,” First International Land Reclamation and Mine Drainage Conference and Third
International Conference on the Abatement of Acidic Mine Drainage, Pittsburgh, PA, April 1994,
pp. 250–257.
Fyson, A., Kalin, M., and Smith, M.P., 1995, “Nickel and arsenic removal from mine wastewater by
Muskeg sediments,” Biotechnology and the Mining Environment, Sudbury, Ontario, May 1995,
Vol. 2, pp. 459–466.
Hedin, R.S., Dvorak, D.H., Gustafson, S.L., Hyman, D.M., McIntire, P.E., Nairn, R.W., Neupert,
R.C., Woods, A.C., and Edenborn, H.M., 1991, “Use of Constructed Wetland for the Treatment
of Acid Mine Drainage at the Friendship Hill National Historic Site, Fayette County, PA,” Final
Report, US Bureau of Mines, Pittsburgh Research Center, Pittsburgh, PA.
Hedin, R.S., and Nairn, R.W., 1992, “Designing and sizing passive mine drainage treatment
systems,” Presented at the 1992 West Virginia Surface Mine Drainage Task Force Symposium,
April 1992, Morgantown, WV.
Hedin, R.S., and Nairn, R.W., 1993, “Containment removal capabilities of wetlands constructed to
treat coal mine drainage,” in Constructed Wetlands for Water Quality Improvement, Chapter 19
G.A. Moshiri, ed., pp. 187–195.
Hedin, R.S., Watzlaf, G.R., and Nairn, R.W., 1994, “Passive treatment of acid mine drainage with
limestone,” Journal of Environmental Quality, Vol. 23, pp. 1338–1345.
Hedin, R.S., Nairn, R.W., and Kleinmann, R.L.P., 1994, “Passive Treatment of Coal Mine
Drainage,” US Bureau of Mines, Information Circular 9389.
Hedin, R.S., 1996, “Environmental engineering forum: Long-term effects of wetland treatment of
mine drainage,” Journal of Environmental Engineering, Vol. 122, No. 1, pp. 83–84.
Jeffers, T.H., Ferguson, C.R., and Seidel, D.C., 1989, “Biosorption of metal contaminants using
immobilized biomass,” Biohydrometallurgy, pp. 317–327.
Johnson, D.B., 1995, “Acidophilic microbial communities: Candidates for bioremediation of acidic
mine effluents,” International Biodeterioration and Biodegradation, Vol. 35, pp. 41–58.
Jones, D.R., and Chapman B.M., 1995, “Wetlands to treat AMD – Facts and fallacies,” In
Proceedings of Second Australian Acid Mine Drainage Workshop, N.J. Grundon and L.C. Bell,
eds., Charters Towers, Queensland, Australia, March 1995, pp. 127–145.
Kalin, M., Scribailo, R.W., and Wheeler, W.N., 1990, “Acid mine drainage amelioration in natural
bog systems,” In Proceedings of the Seventh Annual General Meeting of BIOMINET, Mississauga,
Ontario, Canada, November 1988, pp. 59–77.
Kepler, D.A., and McCleary, E.C., 1994, “Successive alkalinity-producing systems (SAPS) for the
treatment of acidic mine drainage,” In Proceedings of the First International Land Reclamation
and Mine Drainage Conference and Third International Conference on the Abatement of Acidic
Drainage, Pittsburgh, PA, April 1994, Vol. 1, pp. 195–204.
Kilborn Inc., 1996, “Review of Passive Systems for Treatment of Acid Mine Drainage: Phase II,”
prepared for MEND.
Kleinmann, L.L.P., and Hedin, R.S., 1993, “Treat mine water using passive methods,” Pollution
Engineering, Vol. 54, No. 5, pp. 20–22.
Kuyucak N., 1987, “Algal Biosorbents for Gold and Cobalt,” Ph.D. Thesis, McGill University,
Montreal, Quebec, Canada.
Kuyucak, N., 1995, “Conventional and new methods for treating acid mine drainage,” In
Proceedings of CAMI’95, Montreal, Quebec, Canada, October 1995.
Kuyucak, N., and St-Germain, P., 1993, “Passive treatments methods for acid mine drainage,” In
EPD Congress 1993, The Minerals, Metals and Materials Society, J.P. Hager, ed., pp. 319–331.
Kuyucak, N., and St-Germain, P., 1994a, “In-situ treatment of acid mine drainage by sulphate
reducing bacteria in open pits: Scale-up experiences,” In Proceedings of the First International
Land Reclamation and Mine Drainage Conference and Third International Conference on the
Abatement of Acidic Drainage, Pittsburgh, PA, April 1994.
186
Kuyucak, N., and St-Germain, P., 1994b, “Possible options for in-situ treatment of acid mine
drainage seepages,” In Proceedings of the First International Land Reclamation and Mine
Drainage Conference and Third International Conference on the Abatement of Acidic Drainage,
Pittsburgh, PA, April 1994
Kuyucak, N., and St-Germain P., 1998, “Biologically supported water covers, conceptual process
development to prevent acid generation in tailings ponds,” Min. Pro. Ext. Met. Rev., December
1998, pp. 1–13.
Ledin M., and Pedersen, K., 1996, “The environmental impact of mine wastes—Roles of
microorganisms and their significance in treatment of mine wastes,” Earth Science Reviews,
Vol. 41, pp. 67–108.
MEND, 1990, “3.12.1: Assessment of Existing Natural Wetlands Affected by Low pH, Metal
Contaminated Seepages (Acid Mine Drainage),” March 1990.
MEND, 1993, “2.11.2a: Literature Review Report: Possible Means of Evaluating the Biological
Effects of Sub-Aqueous Disposal of Mine Tailings,” March 1993.
MEND, 1994, “2.20.1: Evaluation of Alternate Dry Covers for the Inhibition of Acid Mine Drainage
from Tailings,” March 1994.
Nicholson, R.V., 1994, “Iron-sulfide oxidation mechanisms: Laboratory studies,” In Environmental
Geochemistry of Sulfide Mine Wastes, Short Course Handbook, Chapter 6, Mineralogical
Association of Canada, Vol. 22, pp. 163–183.
Paine, P.J., 1987, “An historic and geographic overview of acid mine drainage,” Acid Mine
Drainage Seminar/Workshop, Halifax, Nova Scotia, Canada, March 1987, pp. 1–45.
Pett, R.J., MacKinnon, D.S., and Lane, P.A., 1990, “Natural wetlands fail to ameliorate acid mine
drainage in the fall,” In Proceedings of the Seventh Annual General Meeting of BIOMINET,
Mississauga, Ontario, Canada, November 1990, pp. 99–127.
Phillips, P., Bender, J., Simms, R., Rodriquez-Eaton, S., and Britt, C., 1995, “Manganese removal
from acid coal-mine drainage by a pond containing green algae and microbial mat,” Water
Science and Technology, Vol. 31, No. 12, pp. 161–170
Smith, M.P., and Kalin, M., 1989, “Biological polishing of mining waste waters: Bioaccumulation
by the Characease,” Biohydrometallurgy, pp. 659–670.
Smith, L.A., Alleman, B.C., and Copley-Graves, L., 1994, “Biological treatment options,” In Emerging
Technology for Bioremediation of Metals, J.L. Means and R.E. Hinchee, eds., pp. 12–25.
St-Germain, P., Larratt, H., and Prairie, R., 1997, “Field studies of biologically supported water
covers at two Noranda tailings ponds,” In Proceedings of the Fourth International Conference on
Acid Rock Drainage, Vancouver, British, Columbia, Canada, May 1997, Vol. 1, pp. 131–148.
St-Germain, P., and Kuyucak, N., 1998, “Biological water covers—a preliminary assessment,” In
Mineral Processing and Extractive Metallurgy Review, F.M. Doyle and N. Arbiter, eds. Montana,
August 1996, pp. 39–46.
Tarleton, A.L., Lang, G.E., and Wieder, R.K., 1984, “Removal of iron from acid mine drainage by
Sphagnum peat: Results from experimental laboratory mesocosms,” In Symposium on Surface
Mining, Hydrology, Sedimentology and Reclamation, University of Kentucky, Lexington, KY,
December, pp. 413–420.
US Environmental Protection Agency, 1995, “Workshop Report: Mine Waste Technical Forum,”
Las Vegas, NV, July 25–27.
Vatcharapijarn, Y., Graves, B., and Bender, J., 1994, “Remediation of mining water with microbial
mats,” In Emerging Technology for Bioremediation of Metals, J.L. Means and R.E. Hinchee, eds.,
pp. 124–129.
Waybrant, K.R., Blowes, D.W., and Ptacek, C.J., 1995, “Selection of reactive mixtures for the
prevention of acid mine drainage using porous reactive walls,” In The Proceedings of Sudbury
’95 Conference on Mining and the Environment, Sudbury, Ontario, Canada, May 1995, Vol. 3,
pp. 945–953.
“Wetland treatment for trace metal removal from mine drainage: The importance of aerobic and
anaerobic processes,” 1992, In Proceedings of Int’l Specialist Conference Wetlands Downunder-
Wetland Systems in Water Pollution Control, University of New South Wales, Sydney, NSW
Australia, November 1992, pp. 46.1–46.9.
187
Wheeler, W.N., Kalin, M., and Cairns, J.E., 1991, “The ecological response of a bog to acidic coal
mine drainage—deterioration and subsequent initiation of recovery,” In The Proceedings of the
Second International Conference on the Abatement of Acidic Drainage, Montreal, Quebec,
Canada, September, Vol. 2 pp. 449–464.
Willianen, S.P., Beckett, P., and Courtin, G., 1998, “Progress in establishing wetland plants in
permanently flooded uranium tailings,” In Mineral Processing and Extractive Metallurgy Review,
F.M. Doyle and N. Arbiter, eds., Montana, August, pp. 47–60.
Yanful, E., and Nicholson, R., 1991, “Engineered soil covers for reactive tailings management:
theoretical concepts and laboratory development,” In Proceedings of the Second International
Conference on the Abatement of Acidic Drainage, Vol. 1, pp. 461–487. September 16–18,
Montreal, Quebec, Canada.
188
Sawdust-Supported Passive
Bioremediation of Western
United States Acid Rock Drainage
in Engineered Wetland Systems
D.N. Thompson,* R.L. Sayer,* and K.S. Noah*
ABSTRACT
Economical remediation of acid rock drainage (ARD) at secluded high-altitude western ore mines is diffi-
cult given the current treatment technology. Several passive-engineered systems that are based on wet-
land technology were tested in the laboratory with three western ore mine ARDs (Co, Cu, and Fe mines)
and with a copper leachate effluent. These systems used well-weathered lodgepole pine sawdust as the
sole carbon source and mud from a pond at one of the mine sites as the source of sulfate-reducing bacte-
ria (SRB). Simple one-pass flow-through systems removed 44% to 99% of the various heavy metals from
the Co and Cu mine ARDs and increased the pH from 3 to 7 after 28 days of SRB acclimation with a five-
day residence time. Carbohydrate utilization rates were generally low, except for one 437-day run,
which consumed 25%, 30%, and 21% of the cellulose, hemicellulose and lignin, respectively. Initial pH,
amounts of toxic heavy metals, and the degree of predegradation of the sawdust were found to signifi-
cantly affect both remediation potential and acclimation time. Engineered SRB systems, alone or in com-
bination with other passive technologies, offer promise for the economical remediation of western US ore
mine ARDs.
INTRODUCTION
Thousands of miles of streams in the United States are seriously affected by acid rock drain-
age (ARD), which represents a major pollution problem (Emerick and Howard, 1988; Dietz
et al., 1994). ARD is formed when rain or groundwater seeps into a mine or through a tail-
ings pile. Acid generation occurs when this water comes into contact with sulfide minerals
and oxygen. The rate of sulfuric acid production can be accelerated more than 106 times by
bacterial action (Singer and Stumm, 1970). Iron-oxidizing bacteria, such as Thiobacillus
ferrooxidans, catalyze acid formation primarily by the oxidation of Fe2+, which is the rate-
limiting step in abiotic ARD formation (Singer and Stumm, 1970). The Fe3+, in turn, oxidizes
many sulfide minerals, producing acidity and additional Fe2+. The ARD ends up in streams,
where it lowers the pH and kills fish and other organisms. ARD flow rates typically range
* Biotechnologies Department, Idaho National Engineering and Environmental Laboratory, Idaho Falls, ID.
189
from 0.063 to 6.3 L/s (1 to 100 gpm), or more, depending on the mine site, geography, and
weather (Wakao et al., 1979; MacDonald et al., 1989; Gusek, 1995).
ARDs differ with geographical location and what is being mined (i.e., coal or metals). A
typical ARD from an ore mine in the western United States will contain high amounts of iron
and various heavy metals, with a pH ranging from 1 to around neutral (Wildeman et al.,
1993). ARDs from eastern US coal mines are often acidic, while ARDs from western US coal
mines are more likely to be near neutral (Hellier, 1989; Dietz et al., 1994). Current ARD
treatment technologies include conventional neutralization with an alkaline material such as
lime, passive treatment in constructed wetlands, anoxic limestone drains, and several less
common methods such as zeolites, ion exchange, microbial mats, and others (Kleinmann et
al., 1991; Kalin et al., 1993; Dietz et al., 1994; Skousen, 1997). These systems are generally
labor and/or capital intensive, a problem that is compounded for remote sites. The remedia-
tion method chosen will depend largely on geographical considerations such as the presence
or absence of flat areas for ponds, the availability and cost of electrical power, and the tem-
perature. Western US metal mines are typically at high altitudes, are relatively inaccessible,
have wide temperature swings, have little flat ground for ponds, and generally do not have a
cheap power supply (Emerick and Howard, 1988). Clearly, development of a passive system
that does not require substantial space or energy would be a significant improvement to the
economics of ARD remediation at western US mines, especially abandoned mines.
Lime precipitation, or liming, is one of the more-used methods in the western United
States. It is widely used because it works year-round, is less expensive than other chemical
treatment options, and is a proven technology (Siwik and Payant, 1970; MacDonald et al.,
1989; Hendricks, 1991; Gusek, 1995). However, one great disadvantage of liming is sludge
disposal. The sludge produced is high volume and is difficult to dewater (Ackman, 1982;
Kuyucak, Sheremata, and Wheeland, 1991). Liming also requires large investments in equip-
ment and labor (Siwik and Payant, 1970; MacDonald et al., 1989; Hendricks, 1991; Gusek,
1995). Wetland technology is a passive method that has gained wide acceptance at eastern
US mine sites, predominantly coal mines (Skousen, 1997). Wetlands remediate ARD through
the action of a complex consortium of microorganisms, including aerobic bacteria, fermenta-
tive bacteria, and SRB, as well as some plants (Emerick and Howard, 1988; Kleinmann et al.,
1991; Dietz et al., 1994). Wetlands, however, require large flat areas of land to treat typical
ARD flow rates. Therefore, they are generally unsuited to mountainous areas (Hellier et al.,
1994). However, wetland technologies might be economically applicable in the western
United States if the chemistry and system design were adapted to the more extreme condi-
tions and the lack of space (Emerick and Howard, 1988; Hellier et al., 1994).
Engineered systems that are based on wetland technologies, such as baffled ditches and
tanks in series, have been in development at various places in the United States and Canada
for a number of years (Béchard et al., 1989; Kalin et al., 1993). Their advantages include
minimal labor costs, minimal space, and minimal energy requirements, as well as the poten-
tial for easy scale-up (Wildeman et al., 1993; Gusek, 1995). The systems use sulfate-reducing
bacteria (SRB), as anaerobic wetlands do, to remediate the ARD. They contain three active
zones: aerobic, microaerobic, and anaerobic. In the aerobic zone, fungi and bacteria convert
lignocellulose or other wastes to soluble carbohydrates, some of which diffuse down into the
microaerobic zone. In the microaerobic zone, these carbohydrates are fermented to organic
acids by facultatively anaerobic bacteria. In the anaerobic zone, the SRB consume the organic
acids and sulfate to form hydrogen sulfide, and bicarbonate is formed from CO2 (Dvorak et
al., 1991; Kleinmann et al., 1991). The sulfides precipitate most metal contaminants. Excess
hydrogen sulfide gas escapes, causing a net consumption of protons, which raises the pH
(Dvorak et al., 1991). Most constructed wetlands, whether aerobic or anaerobic, have been
constructed at eastern US coal mines (Hellier, 1989; Dvorak et al., 1991; Hendricks, 1991;
Skousen, 1997; Rose et al., 1998). At western sites, however, space requirements have lim-
ited applications of SRB technologies to smaller constructed systems (Howard et al., 1989;
Wildeman et al., 1995) and anaerobic nonwetland SRB systems (Béchard et al., 1989; Kalin
et al., 1993; Wildeman et al., 1993; Gusek, 1995).
190
The purpose of this work was to evaluate the utility of engineered SRB systems in remedi-
ating western US ore mine ARDs. It was expected that some western ARDs may not be ame-
nable to this technology because of conditions not normally seen in eastern US ARDs,
including very low pH values, extremely high levels of various metals (such as Fe and Cu) or
the presence of other toxic metals (such as As). It was planned to use sawdust as the sole
source of carbon for the system. It was unknown whether softwood sawdust would supply
nutrients fast enough or for long enough to maintain the remediation potential, because soft-
wood sawdust is higher in lignin and generally degrades much more slowly than hardwoods
(Cowling and Kirk, 1976).
Several simple flow-through reactor designs were tested. The reactors were packed with
layered softwood sawdust (carbon source), mud (SRB source), and gravel (for improved
flow). Because the economics will inevitably be the deciding factor in choosing a system for
ARD remediation, care was taken to use only inexpensive raw materials that were available
at or near the mine sites. Simple designs that have been used in the past with eastern US
ARDs (Wildeman et al., 1993) were used to test their effectiveness with western ARDs. ARDs
from three different western ore mines (copper, cobalt, and iron mines) were tested, along
with a leachate effluent from a copper precipitation plant (not a typical ARD). Treatability
results varied substantially with initial ARD pH, the presence or absence of extremely high
levels of toxic metals, and the relative age of the sawdust used for the carbon source. This
indicates that site-specific factors may determine whether this technology can be applied at
particular western ore mines.
Lignocellulose
Partially degraded lodgepole pine sawdust was used as the lignocellulosic substrate. The
sawdust was scooped from just below the surface near the foot of the source pile to ensure
that it was old enough to be colonized with wood-degrading fungi and bacteria. Two sources
of sawdust were used. The first source (referred to as Batch 1 sawdust) was a well-weathered
sawdust obtained from Yellowstone Log Co., Rigby, ID. This batch was spread onto trays and
inoculated with Rigidoporus ulmarius (ATCC 26757), Rigidoporus vinctus (ATCC 32576) and
Rigidoporus vitreus (ATCC 32569), to further ensure the presence of wood-degrading fungi.
The second batch of lodgepole pine sawdust (referred to as Batch 2 sawdust) was obtained
from Call Forest Products, Idaho Falls, ID, and was significantly less weathered than Batch 1.
After addition to the reactors, this sawdust was inoculated with a suspension of Phanero-
chaete chrysosporium BKM-F–1767 (ATCC 24725); this source was used only for the iron
mine ARD run. There was no specific reason for adding different fungi to the two batches of
sawdust. Rather, the aim was to ensure the presence of large amounts of lignin degraders in
the sawdust, and genera that were readily available at the time of startup were used. Ideally,
sawdust to be used would be composted at the field site so that no inoculation would be nec-
essary. In addition, there was no specific reason for using two different sawdust batches
other than the fact that the first batch was depleted. Any system operating at a given mine
site would have to be robust enough to use different sources of sawdust, the choice of which
would depend on what was available at the time. Refer to Table 1 for the specific sawdust
that was used for each reactor run.
Sulfate-Reducing Bacteria
SRB were present in mud dug from the bottom of a small pond near the cobalt mine site.
Ponds such as this one are often present near mines due to runoff, and they provide a cheap
source of enriched SRB inoculum. In the absence of a nearby pond, another source of SRB,
such as waterlogged soil or sediment from a nearby stream, could be used because the initial
191
TABLE 1 Treatment systems, substrates, and residence times for each ARD and system tested
Residence
ARD System used Sawdust used* time, days
Copper Four troughs in series 100% fresh Batch 1 sawdust 4.7–26
leachate 25.4-cm-high × 15.2-cm-wide ×
61.0-cm-long
active volume of 22.4 L in each
89.6 L total active volume
liquid level at 11.8 L in each
Cobalt Single cylindrical reactor 50:50 (% w/w) mixture of: 5.1
mine 8.89-cm-I.D. × 34.3-cm-high fresh Batch 1 sawdust
ARD
active volume of 1.9 L pre-degraded Batch 1 sawdust
liquid level of 1.3 L
Copper Single cylindrical reactor 50:50 (% w/w) mixture of: 4.4
mine 8.89-cm-I.D. × 34.3-cm-high fresh Batch 1 sawdust
ARD
active volume of 1.9 L pre-degraded Batch 1 sawdust
liquid level at 1.3 L
Iron Four cylindrical reactors in series 100% fresh Batch 2 sawdust 5.0
mine 3.81-cm-I.D. × 34.3-cm-high
ARD
each with:
— each cylinder 0.39 L
— active volume of 0.35 L
— liquid level at 0.23 L in each
* A mixture of unused and used Batch 1 sawdust was employed in certain cases to further ensure high initial levels of wood
degrading fungi would be present, in an attempt to further shorten acclimation time. Where used, “pre-degraded Batch 1
sawdust” refers to sawdust that had been used in the copper leachate runs for 363 days.
SRB level affects mainly the startup (acclimation) time and not the final remediation poten-
tial. The mud was stored at 4°C in sealed 19-L (5-gal) buckets until needed. The mud was
added to the reactors without amendments.
Fertilizer
Because Batch 1 sawdust was well weathered, no fertilizer was added to runs where Batch 1
sawdust was used. Batch 2 sawdust, however, was essentially unweathered. Thus, it was
decided to add nitrogen and phosphorous fertilizers to the sawdust layer to help increase
fungal numbers quickly, given that the carbon-to-nitrogen (C:N) ratios in wood are very high
(Atlas and Bartha, 1987). Because only the iron mine ARD run used Batch 2 sawdust, this
was the only run that was fertilized. Pelleted nitrate and phosphate fertilizers, 34–0–0 and
0–45–0 (weight percent N:P:K) from Simplot Soilbuilders, Idaho Falls, ID, were mixed with
the sawdust before addition to the reactor to give a molar C:N:P ratio of 11:7:1 (Kuyucak et
al., 1991) in the final lignocellulose layer. The carbohydrate fractions of the sawdust were
used in determining the carbon content of the substrate for this calculation, neglecting car-
bon present in the lignin. The fertilizer pellets and sawdust were mixed by vigorous inversion
in a covered bucket.
192
TABLE 2 ARD sources and their chemical compositions
Species Copper leachate
or property Cobalt mine Copper mine Iron mine (undiluted)
pH 3.0 3.2 1.2 2.2
Eh (mV) 396 530 414 425
Fe (ppm) 105 4.3 10,700 1,100

Cu (ppm) 1.2 1.3 202 200
Mn (ppm) 4.2 75 11.7 440
Zn (ppm) 0.21 5.3 360 230
Co (ppm) 7.4 0.43 0.97 ND
Ni (ppm) 0.44 0.62 1.27 46
Mg (ppm) 27.3 220 475 11,750
Al (ppm) ND ND ND 11,000
As (ppm) 0.005 ND ND 0.088
2–
SO4 (ppm) 640 334 12,300 50,000
ND = Not determined.
Acid Rock Drainage

ARDs were obtained from four sites; their sources and initial chemical-characterization data
are summarized in Table 2. No carbon or nutritional amendments were made to the ARDs
before treatment in the reactor systems. Copper leachate was used as a 50:50 (v/v) mixture
with 18 MΩ nanopure water to reduce potential toxicity due to high metals content. Where
used, “synthetic ARD” refers to a pH 3.0 aqueous solution containing 100 ppm Fe2+, 165 ppm
Mg2+, 15 ppm PO43– and 1,500 ppm SO42–.
Reactor Designs
ARD treatment reactors were constructed from Plexiglas pipe and/or sheeting in two basic
designs, as shown in Figure 1. These designs were chosen because they are simple and have
been used successfully to remediate eastern US coal mine ARDs. The simplest design, shown
in Figure 1(a), was a cylindrical column (8.89-cm-I.D. × 34.3-cm-high) that was packed with
a layer of SRB-containing mud and overlain with approximately 20.3 cm of air-dried saw-
dust. A layer of gravel was placed below the mud to facilitate flow out of the reactor. The liq-
uid level was maintained in the sawdust layer by means of an overflow tube from the outlet,
which maintained the three necessary zones for the remediation as described above. The
active reactor volume was about 1.9 L (sawdust, mud, plus gravel) with liquid up to the
1.3-L mark or 20.3 cm, which covered about half of the sawdust layer. The total height was
30.5 cm. This simple design allowed for good contact between the ARD and the different lay-
ers, but other more complex flow-through designs would work as well.
The second design used is shown in Figure 1(b). The design consisted of a simple trough
(25.4-cm high, 15.2-cm wide and 61.0-cm long). The trough was packed in layers as described
above. ARD was sprinkled over the top surface of the sawdust layer and was collected at the
bottom beneath the gravel layer. The liquid level was maintained as for the cylindrical reactors.
(Refer to Table 1 for the particular system used for each ARD.) No particular criteria were used
to match an ARD with the geometry of the system used to treat it. ARD was added to the reac-
tor systems using a pump, although, at a mine site, the systems would be gravity fed. Residence
times, listed in Table 1, were set to help ensure good contact between the ARD and the solid
193
(a)
(b)
FIGURE 1 Schematics of passive ARD reactor systems (not drawn to scale): (a) single cylindrical
column and (b) single rectangular trough. Cylinders were used singly (copper mine and cobalt mine
ARDs) or in series (iron mine ARD) and troughs were used in series only (copper leachate). Zone A
is the aerobic zone, Zone B is the microaerobic zone, and Zone C is the anaerobic zone.
phases. Note that these retention times are based on the active reactor volume below the liquid
level, not the actual liquid volume contained in the voids.
Reactor Run Protocols

Reactors were run as follows: The cobalt mine ARD was treated for a total of 59 days with no
formal period of acclimation of the SRB to the cobalt mine ARD. The actual SRB acclimation
time, or time for the pH to begin to increase, was 25 days. The copper mine ARD was treated
in the cylindrical reactor system, including the sawdust and mud, used previously for the
cobalt mine ARD. Therefore, the acclimation period for this run was 59 days. It was expected
that the copper leachate would be toxic to the microbial populations because of its high Al
and Mg content. Therefore, preacclimation to synthetic ARD was performed. The SRB were
acclimated at total recycle until the pH increased (17 days). Then the system was run single
pass with synthetic ARD until day 70. After day 70, it was run single pass with the diluted
copper leachate. Finally, the iron mine ARD SRB acclimation period was begun with one liter
of undiluted iron mine ARD with total recycle; this run did not fully acclimate to the iron
mine ARD. Controls lacking SRB activity were not conducted, because the treatability tests
were not done to specifically attribute portions of the remediation potential to biotic and/or
abiotic sources.
194
Analytical
Physical and chemical parameters were monitored periodically follow:
Acidity and Reduction Potential. The pH and Eh were measured by electrodes in the
inlet and outlet reservoirs and in the outlet tubes from the reactors (see Figure 1).
Metals. Cu, Ni, Zn, Fe, Co, Al, Mn, and Mg were measured in 0.24-N HCl at room tem-
perature (23°C) by atomic absorption spectrophotometry (AA) using a Perkin-Elmer, Model
5100, atomic absorption spectrophotometer and the appropriate lamps. Calibrations were
done with standards for each metal made up at 0.500 to 5.00 ppm in 18 MΩ nanopure water.
Samples were diluted to this range with 0.24-N HCl.
Cellulose, Hemicellulose, and Lignin/Ash Compositions. These were determined
using the quantitative saccharification method (Saeman et al., 1945). Dried ground sawdust
samples (100 to 500 mg) were hydrolyzed first with 72% (w/v) H2SO4 at 30°C using 1 mL of
acid per 100 mg of sample and then with 4% (w/v) H2SO4 at 121°C in an autoclave. The
hydrolyzed samples were then collected quantitatively on a tared Gooch crucible, and the fil-
trate was brought up to a known volume. The remaining insoluble portion, defined as Klason
lignin with extractives and ash, was determined after drying at 105°C.
Carbohydrates were measured by HPLC using a BioRad HPX–87P carbohydrate column
with a guard and de-ashing system and were corrected for recovery losses incurred during
the high-temperature step. Cellulose was defined as the total anhydroglucose (glucan) con-
tent, and hemicellulose was defined as the sum of the total xylan, galactan, arabinan, and
mannan contents. The calculated hemicellulose values will inherently be low due to
unknown uronic acid contents. These data were obtained for the starting sawdust and the
finished reactor sawdust for each case. For the iron mine run lignocellulose samples, fertil-
izer remaining in the samples was removed by washing with distilled water before drying,
grinding, and analysis.
RESULTS
Results for the four ARD streams are presented below. Half of the ARD samples were remedi-
ated in the time allotted, while the other half were not remediated. The lack of remediation
in some cases was possibly due to very high metals concentrations, very low inlet pH, or
insufficiently weathered sawdust. No attempt was made to distinguish between biotic and
abiotic remediation, although black precipitates were observed in the reactors that showed
remediation, indicating SRB activity. However, some amount of precipitation of metals such
as iron due to oxidation is to be expected.
Initial Sawdust Composition

Results of the compositional analysis for both of the starting sawdust substrates are pre-
sented in Table 3. The initial compositions of the starting sawdust samples for the various
ARD runs are typical for softwoods (Cowling and Kirk, 1976), with the exception of the high
lignin/ash and low cellulose contents in Batch 1 sawdust. This high lignin/ash content was
probably the result of scooping some soil with the sawdust during collection. The low cellu-
lose content reflects the fact that Batch 1 sawdust was pre-degraded (weathered). Its cellu-
lose content is low compared to what would be expected for fresh softwood sawdust
(Cowling and Kirk, 1976). Because the lignin/ash content is so high, the cellulose numbers
are necessarily lower than if no soil were in the sawdust. However, the amounts of cellulose
and hemicellulose in the sample are about half those expected for a fresh softwood sawdust,
indicating substantial weathering (Cowling and Kirk, 1976). Batch 2 sawdust had a more
typical lignin/ash content and an overall composition that is normal for an unweathered soft-
wood sawdust (Cowling and Kirk, 1976).
195
TABLE 3 Initial and final sawdust compositions after use in the ARD reactors. Times of measure-
ment represent time at single-pass flow of ARD, after acclimation of the reactor, except for the iron
mine ARD, which was always at total recycle.
Composition*
Copper mine Copper leachate
(59 days (70 days
Cobalt mine acclimation)†† Iron mine acclimation)††
Component Day 0 Day 59 Day 0 Day 36 Day 0 Day 132 Day 0 Day 397
Cellulose† 19.1 20.6 20.6 19.1 45.8 42.4 21.8 16.4
±1.4 ±0.3 ±0.3 ±0.5 ±1.9 ±1.0 ±1.1 ±1.0
Hemicellulose‡ 9.8 10.2 10.2 9.7 25.0 22.9 11.6 8.1

±1.2 ±1.1 ±1.1 ±0.4 ±1.0 ±1.2 ±1.1 ±0.6
Lignin§ 55.3 57.7 57.7 60.4 32.2 35.5 61.9 48.8

±2.2 ±1.8 ±1.8 ±0.4 ±0.5 ±1.8 ±1.7 ±1.4
Sum** 84.2 88.5 88.5 89.1 103 101 95.3 73.3

±3.1 ±2.3 ±2.3 ±0.8 ±2.2 ±2.2 ±2.3 ±2.1
* Based on 100% dry weight of material.

† Defined as the total glucan content.
‡ Defined as xylan + galactan + arabinan + mannan.
§ Includes extractives and ash.
**Remaining fraction attributed to unknown uronic acid content and to recovery errors in analysis techniques.
††SRB acclimation as described above to an ARD other than the one being treated. The acclimation time is considered here
as prior to day 0. Where there is no acclimation time indicated (reactor acclimated to the same ARD that is being
treated), day 0 is considered to be at reactor startup.
Carbon Utilization
There was no measurable degradation of polysaccharides or lignin in the sawdust during the
cobalt mine ARD run, including the acclimation period (Table 3). There was SRB activity over
this period, as indicated by the mud and deeply submerged sawdust turning black from metal-
sulfide precipitates. The ultimate measure of biological activity in the system is sulfide produc-
tion using energy and carbon originating in the sawdust. The lack of measurable lignocellulose
degradation over more than a month while, at the same time, observing sulfide precipitation
suggests that the sawdust substrate should last for a long period. The copper mine ARD was
treated in the same reactor immediately following the completion of the cobalt mine ARD run.
Again, there was very little measurable degradation of the substrate, again suggesting that the
substrate will last in this process for a very long time. There was a slight degradation of the iron
mine ARD substrate sawdust (Batch 2). This probably corresponds to the highly accessible,
more easily released portions of the cellulose and hemicellulose (Thompson et al., 1992) which
were already removed from the well-weathered Batch 1 sawdust.
The copper leachate run, which was by far the longest, showed a 25% decrease in cellulose
content (as total glucan), a 30% decrease in hemicellulose (estimated as xylan + galactan +
arabinan + mannan) and a 21% decrease in lignin/ash. From this data, one can roughly esti-
mate the efficiency of shuttling the electrons stored in the sawdust to precipitated sulfides.
About 3,930 g of carbohydrate carbon were released from all four troughs over the 437-day
period. If a constant rate of carbohydrate release is assumed, this gives 0.375 g of carbohy-
drate carbon released per hour. An electron balance, assuming the ultimate product of the fer-
mentative step is acetate, yields four e– per acetate or 12 e–/C6 carbohydrate and 10 e–/C5
carbohydrate in the original lignocellulose. This value is consistent with a previously
196
suggested molar ratio of carbon substrate to sulfate to be reduced of 2:1 (Kuyucak et al.,
1991). Because eight e– are required to reduce SO42–to S2–, one mole of C6 or C5 carbohydrate
can supply enough e– to reduce 1.5 or 1.25 moles of SO42–and thus precipitate 1.5 or
1.25 moles of divalent metal, respectively.
It is unknown how much of the substrate carbon is consumed by the fungal and fermenta-
tive cells, but it is reasonable to expect they would be in near-maintenance metabolism
because of the high C:N ratio in wood. If the maintenance consumption in each microbial
zone is 1% of the total carbon in that zone, enough acetate would remain to convert 0.29 g
SO42– to S2– per dry gram of the initial weathered sawdust. At 5%, 10%, and 25% consump-
tion in each zone, this would be 0.25, 0.21, and 0.075 g SO42– per dry gram of the initial saw-
dust. Because one mole of SO42–can potentially precipitate one mole of divalent metal, there
is a very significant remediation potential for a reactor loaded with sawdust. If degradation
continued at a constant rate, the carbohydrates in the sawdust should last for 4.8 years, with
a potential precipitation of 0.87, 0.76, 0.63, and 0.22 kg of divalent metals per kg of sawdust
if the maintenance consumption rates in each microbial zone are 1%, 5%, 10%, or 25%,
respectively. Thus, assuming 1% consumption in each microbial zone, the laboratory system
should be able to effectively remediate a 0.063-L/s (1-gpm) stream at 1.3 ppm total metals, a
0.0063-L/s (0.1-gpm) stream at 13 ppm total metal, a 0.00063 L/s (0.01 gpm) stream at
130 ppm total metals or a 0.00011-L/s (0.0018-gpm) stream, the highest actual flow rate
used, containing 2,340 ppm total metals. Other maintenance levels would be scaled propor-
tionally to 1 + 3mƒ, where mƒ is the fractional maintenance consumption. The actual perfor-
mance would, of course, depend on the initial pH of the ARD, because some S2– would be
used in raising the pH. Finally, in an actual application outside the laboratory, environmental
conditions such as temperature will fluctuate widely, which will, in general, slow the ligno-
cellulose degradation (Atlas and Bartha, 1987).
ARD Remediation
Time courses of pH and Eh for the cobalt mine ARD run are shown in Figures 2(a) and
2(b). Note that the pH and metals concentration data given in this section are measure-
ments of the outlet reservoir ARD from the reactor, because that is what would be observed
downstream of an ARD remediation site. As mentioned above, the pH did not increase
until Day 25 because of SRB acclimation to the low pH and high metals concentrations. By
Day 28, the effluent pH was 7, and this pH was maintained until the run was stopped on
Day 59 to test the copper mine ARD in the reactor. The Eh, which was measured in the out-
let tube and in the outlet reservoir to indicate the redox environment seen by the SRB as well
as the downstream Eh, showed a slow general trend downward to more reducing conditions,
as would be expected if there were SRB activity.
The time courses of outlet levels of various metals in the cobalt mine ARD are shown in Fig-
ure 3. Iron and cobalt were remediated at 99% and 97%, respectively. Note that the effluent
metal concentrations were reduced during the acclimation period and were then maintained at
those levels. This reduction occurred while the pH was stable at its initial value. It is likely that
hydrogen sulfide gas was evolved (and protons were consumed, increasing the pH) only after
an excess of sulfide was produced, because sulfide would first precipitate with the divalent
metals (Wakao et al., 1979; Emerick and Howard, 1988). The same trend was also seen with
copper, nickel, and zinc, which were remediated at 98%, 68%, and 89%, respectively.
Because space is generally scarce at western US mine sites, the amount of surface area
required for the system is important. A simple sizing calculation for a typical eastern US wet-
land treating 4.4 L/s (70 gpm) of the cobalt mine ARD, assuming a loading factor of
2 m2/(mg/min) for iron (Wildeman et al., 1993), gives an area of 55,500 m2 or a surface to
flow ratio of about 13,000 m2/(L/s). In the single-pass cylindrical system (single cylinder),
the measured loading factor for iron was about 0.25 m2/(mg/min), which gives 6,900 m2 or
a surface to flow ratio of 1,600 m2/(L/s). Thus, a proportionally scaled system based on the
cobalt mine ARD treatment results would theoretically need only one-eighth of the area
197
(a)
(b)
FIGURE 2 Variation of outlet reservoir pH and outlet tube Eh with time for the cobalt mine ARD
run: (a) outlet reservoir pH and (b) outlet Eh values. Closed squares represent reservoir Eh, while
open squares are outlet-tube Eh.
needed for a typical eastern US wetland to treat an identical flow rate, a substantial space
savings.
Of the three additional streams treated, remediation was seen for the copper mine ARD,
but not for the iron mine ARD nor for the copper leachate. These data are summarized in
Table 4. The copper mine ARD, which was high in Zn and Cu and initially at pH 3.2, was suc-
cessfully remediated in 36 days to pH 6.2 with 44% to 99% metal removals in the preaccli-
mated column from the cobalt ARD runs.
The columns-in-series system was not successful in remediating the iron mine ARD, which
contained 10,700 ppm of iron and had a very low initial pH of 1.2. The pH and Eh were not
remediated to acceptable levels, nor were any of the metals.
Possibly, the lack of acceptable remediation was due to Fe, sulfate, and/or pH toxicity.
Other possibilities include incomplete SRB acclimation and/or insufficiently weathered saw-
dust. The answer is likely a combination of these because there was a 50% removal of iron
and some degradation of the sawdust (see Table 3). The low initial pH could be raised by the
addition of an anoxic limestone drain (Gusek, 1995) upstream of the reactor or the addition
of crushed limestone to the sawdust layer.
The troughs-in-series system used for the diluted copper heap leachate, which contained
5,500 ppm Al and 550 ppm Fe, was also not very successful, lowering only iron from its ini-
tial level. This same system was capable of remediating the synthetic ARD before adding the
leachate, but not after leachate addition. Neither pH nor Eh changed significantly during
198
(a)
(b)
FIGURE 3 Variation of outlet levels of various metals with time of treatment of a single-pass flow-
through stream: (a) iron and (b) cobalt, copper, nickel, and zinc. Inlet concentrations were:
105 ppm Fe, 7.4 ppm Co, 1.2 ppm Cu, 0.44 ppm Ni, and 0.21 ppm Zn.
passage through the system. Perhaps levels of one or more heavy metals were high enough
to be toxic to the microorganisms needed for the remediation. A fair amount of lignocellu-
lose degradation was observed, which might indicate a toxicity effect on the fermentative
bacteria or the SRB.
DISCUSSION
ARDs from the cobalt and copper mines were successfully remediated by the engineered SRB
systems tested. No measurable degradation of the sawdust was observed, although this was
not unexpected because the run time was short. The initial pH values of around 3 apparently
did not inhibit SRB acclimation substantially. Although there were metals present in both the
cobalt mine and copper mine ARDs, in levels as high as 100 and 220 ppm, respectively, these
were apparently not toxic enough to destroy remediation potential.
The iron mine ARD was not remediated within the allotted time. Possible reasons for the
lack of remediation of the iron mine ARD include insufficiently weathered sawdust, high iron
levels, and a very low pH of 1.2. It appears that the iron mine ARD run would have eventu-
ally acclimated, because the outlet pH slowly increased to 1.6 over the 132 days of the run.
Iron levels, also dropped by 50% over the course of the run, although this could have
occurred due to adsorption to the solid media. No reddish precipitate was observed in the
reactors, suggesting that the drop in Fe was not due to iron oxidation. The increase in Zn lev-
els could be due to solubilization of metals in the added fertilizers, because the system was
run with total recycle of 1.25 L of ARD over the entire run. Interestingly, the iron mine run
199
(a)
(b)
FIGURE 4 Variation of outlet reservoir pH and outlet tube Eh with time for the cobalt mine
ARD run: (a) outlet reservoir pH and (b) outlet Eh values. Closed squares represent reservoir
Eh, while open squares are outlet-tube Eh.
was the only one that showed fast initial degradation of the initial sawdust, which probably
reflects removal of the easily degraded, easily accessible fractions of the cellulose and hemi-
cellulose (Thompson et al., 1992). Although it is possible that this removal was biologically
mediated, it is also possible that the ARD may have been acidic enough to attack the easily
degraded portions of the sawdust to some extent at room temperature. This easily accessible
fraction was likely already removed during weathering of the Batch 1 sawdust.
Another run that did not indicate remediation potential was the copper leachate run. It is
not surprising that this stream was not remediated, because it is not a typical ARD. Given
that sulfate reduction below pH 4.5 is very slow unless sufficient alkalinity can be generated
to raise the pH, it is not surprising that SRB would acclimate more slowly to the lower pH
ARDs. However, the troughs-in-series system quickly remediated the pH 3.0 synthetic ARD
after only 17 days of acclimation. Observable lignocellulose degradation suggests that the
lack of remediation may not have been due to a limited carbon or nutrient supply. This sug-
gests that metal toxicity due to very high levels of Fe in the iron mine ARD and toxicity due to
Al and Mg in the copper leachate may be the mitigating factor in the case of these two
streams. This possibility is supported by the lack of a return of remediation potential after
leachate flow was stopped and replaced with the synthetic ARD.
From these data, it is clear that weathered or partially degraded lodgepole pine sawdust
can supply all the necessary carbon and nutrients not supplied by the mud. Partially degraded
200
TABLE 4 Remediation data for the copper mine ARD, iron mine ARD, and copper heap leachate in their
various treatment system designs. (Refer to Table 2 for initial values.) The copper leachate was diluted
50:50 with water before addition to the reactors. All data presented are from the outlet reservoir,
except for Eh, which is from the outlet tube to indicate the redox environment seen by the SRB.
Iron mine, Copper leachate,
Copper mine, outlet ARD outlet leachate
Species or outlet ARD after 36 days after 132 days after 397 days
property (59 days acclimation)* (132 days acclimation) (70 days acclimation)
pH 6.2 1.6 2.8
Eh (mV) –48 663 376
Fe (ppm) 2.43 5,000 174

Cu (ppm) 0.06 187 120
Mn (ppm) 63 13.9 220
Zn (ppm) 0.05 620 120
Co (ppm) 0.44 1.84 ND
Ni (ppm) 0.19 1.65 21.3
Mg (ppm) 200 501 4,680
Al (ppm) ND ND ND
2–
SO4 (ppm) ND ND ND
* SRB acclimation as described above to an ARD other than the one being treated. The acclimation time is considered here
as prior to Day 0. Where there is no acclimation time indicated (reactor acclimated to the same ARD that is being treated),
Day 0 is considered to be at reactor startup.
ND = Not determined.
sawdust provides a more easily degraded substrate, so acclimation is faster, and the substrate
should provide carbon for growth for a longer period. Fungi already established in sawdust
are more likely to efficiently release carbohydrates from the lignocellulose than newly inocu-
lated fungi (Béchard et al., 1989). The microbes present must have sufficient nutrients (N, P,
etc.) for remediation to occur; partially degraded sawdust seems to be able to supply these
nutrients over the time tested. Another benefit of using sawdust is that it is degraded slowly,
leading to a slow release of carbon, so substrate will not need to be added often.
The data also show that SRB from the mine site can be adapted to acidic, high-metal condi-
tions and used to catalyze the precipitation of divalent metals and increase the pH. However,
several factors can limit the applicability of SRB systems to specific ARDs or at specific sites.
Toxic metals can cause the SRB or another critical member of the consortium to die out and
end remediation potential for the system. If the initial pH is too low, the SRB may not be able
to acclimate in a reasonably short period. This is generally not much of a problem, because
very low inlet pH can be overcome by adding limestone to the sawdust or by another passive
technology (Dvorak et al., 1991; Wildeman et al., 1993). The systems were not tested at lower
temperatures, but it is well known that, at lower temperatures, microbial activity decreases
(Atlas and Bartha, 1987). Therefore, performance of the reactor would be expected to
decrease substantially during winter. This means that systems that can be insulated or can be
removed during the winter and replaced in the spring would be preferred. The next logical
step for this system would be to construct a larger volume system that maximizes contact with
the solids, has a large residence time and is suited for field trials at a mine site.
Economical remediation of ARD pollution at western ore mines is a challenge, given the
limitations of current treatment technologies (Siwik and Payant, 1970; Hellier et al., 1994;
Gusek, 1995). Technologies that are more efficient with respect to space, such as liming and
ion exchange, are inherently more labor and capital intensive. Technologies that are passive
201
and, therefore, more labor- and capital-efficient tend to use large areas due to low volumetric
productivity. Clearly, a system is needed that fosters both high surface/volumetric productiv-
ity and a passive nature (Wildeman et al., 1993; Gusek, 1995). Engineered SRB systems that
mimic wetlands seem to offer promise for this application (Gusek, 1995). Various consider-
ations that significantly affect wetland performance can logically be extended to engineered
systems. First, SRB require time to acclimate to the low pH seen in ARD, because they grow
best at pH values that are essentially near neutral (Postgate, 1984). This acclimation time,
which can take up to a year or more under certain conditions, is required to produce local
areas of higher pH. Thus, the acclimation period to an ARD with a very low pH can be short-
ened by artificially raising the pH above 3 during the initial startup period.
When using sawdust as the sole carbon source, it is reasonable to expect that the process
will be nitrogen-limited, given the very high C:N ratios seen in wood (Atlas and Bartha, 1987).
This is the proper situation to induce ligninolytic activity by white rot fungi, whose ligninolytic
enzymes are induced under conditions of carbon or nitrogen limitation (Boominathan and
Reddy, 1992). Partially degraded or weathered sawdust is a better substrate for this type of
ARD system than fresh sawdust, because the rate of supply of carbohydrates depends directly
on the accessibility of the polysaccharides to the extracellular hydrolytic enzymes (Thompson
et al., 1992) and because the microbial populations will be nitrogen limited. In a partially
degraded sawdust, substantial delignification and colonization with fungi has already
occurred. With degradation already underway, the startup time should be shorter. A mixture
of weathered and unweathered sawdust should provide an initially higher level of carbohy-
drates versus 100% weathered sawdust, as well as provide a larger fungal inoculum.
When designing an engineered SRB system, it is important to remember that kinetics is
limiting rather than transport (Wildeman et al., 1993), especially when sawdust is used as
the sole carbon source. Residence time and flow patterns should be set to maximize con-
tact of the ARD with the solid phases. Because flow is down through the sawdust and mud
layers, carbon, nutrients, metals, and sulfates are distributed through the system mainly by
convection rather than by diffusion, as in a wetland. Thus, microbial activity is more likely
to be rate-limiting in an engineered SRB system. This is in contrast to wetlands, where dif-
fusion of nutrients and contaminants into the sediments is more likely to be rate-limiting.
Flow, whether pumped or gravity fed, should be directed through the mud layer rather
than across the top of it, because a lack of flow through the mud layer in a wetland is the
main factor that leads to low surface and volumetric productivity (Wildeman et al., 1993).
The authors designed a passive system that meets these requirements (Noah et al., 1998).
The design, which is similar to that shown in Figure 1(b), is modular, so that scale-up is
based merely on the flow rate for a given ARD. Vertical baffles and weirs provide direc-
tional control for both horizontal and vertical flows and maximize contact between the liq-
uid and solid phases in the reactor.
There is clearly great potential for use of engineered SRB systems for reductive remedia-
tion of sulfate-contaminated streams, whether from mines or otherwise generated. There
may also be potential for use of other reductive systems for remediation of waste streams
from a variety of sources in these engineered systems. For example, in the author’s labora-
tory, the remediation of selenium-contaminated mine drainage streams from a gold mine
and a phosphate mine were also tested. The streams were treated in a cylindrical one-pass
flow-through system identical to that used to treat the cobalt mine ARD above (see Table 1).
To treat the selenium-contaminated streams, however, mud from the selenium-contaminated
sites was used in place of ARD-contaminated mud. The mud was found to contain measur-
able amounts of selenium-reducing bacteria (SeRB) and served as inoculum for the reactors.
This system decreased selenium levels in one gold mine leach pad drainage stream from 100
to 2 ppb with a six-day residence time (90 mL/day), and the system remediated its pH from
9.1 to 7.0. At a two-day residence time (270 mL/day), performance decreased with outlet
selenium levels ranging from 10 to 20 ppb. Another gold mine leach pad drainage stream
was remediated from pH 4.1 to 7.0, and its selenium levels were reduced from 40 ppb to less
than 5 ppb with a six-day residence time. Selenium in water from a phosphate mine was
202
reduced from 650 to 700 ppb to less than 10 ppb with a two-day residence time, and the pH
remained constant in the 7 to 8 range. These data show that other reductive enzyme systems,
such as for nitrates, would work in this type of system as well.
CONCLUSIONS
Metal removals of up to 99% and remediation of pH levels to neutral are attainable in engi-
neered passive SRB systems. However, in the case of extremely low pH ARDs or ARDs with
very high (greater than 5,000-ppm) metal contents, one-step remediation in an engineered
passive SRB system may not be economical. Pretreating these streams with other passive
technologies, such as limestone drains for pH, could allow treatment in the SRB systems.
Temperature fluctuations will also slow the remediation during the fall and winter months.
Local mud and softwood sawdust, preferably well weathered, provides the carbon, nutrients,
and microbial inocula necessary for the ARD remediation. The volumetric productivity in the
engineered SRB systems tested was high enough to decrease the land area required for reme-
diation by 88%, which is a major consideration when choosing a remediation technology at a
mountainous site. Finally, engineered passive SRB systems also show potential for the reme-
diation of selenium-contaminated streams, and they are likely candidates for remediation of
streams containing nitrates.
ACKNOWLEDGMENTS
The authors would like to thank Robert S. Cherry for his useful discussions and for his critical
editing of this manuscript. This work was supported by the INEEL Laboratory Directed
Research and Development Program under DOE Idaho Operations Office Contract DE-AC07-
99ID13727.
REFERENCES
Ackman, T.E., 1982, “Sludge disposal from acid mine drainage treatment,” US Bureau of Mines
Report of Investigations 8672.
Atlas, R.M., and Bartha, R., 1987, In Microbial Ecology: Fundamentals and Applications, 2nd Ed.,
Benjamin/Cummings Publishing Co., Inc., Menlo Park, CA.
Béchard, G., Rajan, S., Salley, J., and McCready, R.G.L., 1989, “Neutralization of acid mine
drainage using microbial processes,” Proceedings of the 1989 Annual General Meeting of
BIOMINET, Laval, Quebec, Canadian Government Publishing Centre, Ottawa, pp. 62–73.
Boominathan, K., and Reddy, C.A., 1992, “Fungal Degradation of lignin: Biotechnological
applications,” In Handbook of Applied Mycology, Vol. 4: Fungal Biotechnology, D.K. Arora, R.P.
Elander, and K.G. Mukerji, Eds., Marcel Dekker Inc., New York, Chapt. 25, pp. 763–822.
Cowling, E.B., and Kirk, T.K., 1976, “Properties of cellulose and lignocellulosic materials as
substrates for enzymatic conversion processes,” Biotechnol. Bioeng. Symp., pp. 95–123.
Dietz, J.M., Watts, R.G., and Stidinger, D.M., 1994, “Evaluation of acidic mine drainage treatment
in constructed wetland systems,” US Bureau of Mines Special Publication SP 06A–94, Mine
Drainage, Paper presented at the International Land Reclamation and Mine Drainage
Conference and the Third International Conference on the Abatement of Acidic Drainage,
Pittsburgh, PA. 1, pp. 70–79.
Dvorak, D.H., Hedin, R.S., Edenborn, H.M., and Gustafson, S.L, 1991, “Treatment of metal-
contaminated water using bacterial sulfate reduction: Results from pilot-scale reactors,”
Proceedings of the Second International Conference on the Abatement of Acidic Drainage,
Montreal, 1, pp. 301–314.
Emerick, J.C., and Howard, E.A., 1988, “Using wetlands for the control of western acid mine
drainage,” Paper presented at the SME Annual Meeting, Phoenix, AZ, Preprint No. 88–173.
203
Gusek, J.J., 1995, “Passive-treatment of acid rock drainage: What is the potential bottom line?”
Mining Engineering, Vol. 47, No. 3, pp. 250–253.
Hellier, W.W., Jr., 1989, “Constructed wetlands in Pennsylvania: An overview,” In
Biohydrometallurgy 1989, Proceedings of the International Symposium on Biohydrometallurgy,
Jackson Hole, WY, J. Salley, R.G.L. McCready, and P.L. Wichlacz, eds., CANMET SP89–10,
Energy, Mines, and Resources Canada, pp. 599–611.
Hellier, W.W., Giovannitti, E.F., and Slack, P.T., 1994, “Best professional judgment analysis for
constructed wetlands as a best available technology for the treatment of post-mining
groundwater seeps,” Bureau of Mines Special Publication SP 06A–94, Mine Drainage, Paper
presented at the International Land Reclamation and Mine Drainage Conference and at the Third
International Conference on the Abatement of Acidic Drainage, Pittsburgh, PA. 1, pp. 60–69.
Hendricks, A.C., 1991, “The use of an artificial wetland to treat acid mine drainage,” Proceedings
of the Second International Conference on the Abatement of Acidic Drainage, Montreal, 2,
pp. 549–558.
Howard, E.A., Emerick, J.C., and Wildeman, T.R., 1989, “Design and construction of a research
site for passive mine drainage treatment in Idaho Springs, Colorado,” In Constructed Wetlands
for Wastewater Treatment: Municipal, Industrial, and Agricultural,” D.A. Hammer, ed., Lewis
Publishers, Chelsea, MI.
Kalin, M., Fyson, A., and Smith, M.P., 1993, “ARUM—Acid reduction using microbiology,” In
Biohydrometallurgical Technologies, A.E. Torma, M.L. Apel, and C.L. Brierley, eds., The
Minerals, Metals and Materials Society, pp. 319–328.
Kleinmann, R.L.P., Hedin, R.S., and Edenborn, H.M., 1991, “Biological treatment of mine water—
An overview,” Proceedings of the Second International Conference on the Abatement of Acidic
Drainage, Montreal, 4, pp. 27–42.
Kuyucak, N., Sheremata, T.W., and Wheeland, K.G., 1991, “Evaluation of improved lime
neutralization processes. Part 3: Interpretation of properties of lime sludge generated by
different processes,” Proceedings of the Second International Conference on the Abatement of
Acidic Drainage, Montreal, 2, p. 31.
Kuyucak, N., Lyew, D., St-Germain, P., and Wheeland, K.G., 1991, “In situ bacterial treatment of
AMD in open pits,” Proceedings of the Second International Conference on the Abatement of
Acidic Drainage, Montreal, 1, pp. 335–353.
MacDonald, R.J.C., Kondos, P.D., Crevier, S., Rubinsky, P., and Wasselauf, M., 1989, “Generation
of and disposal options for Canadian Mineral Industry effluent treatment sludges,” In Tailings
and Effluent Management, Proceedings of the International Symposium on Tailings and Effluent
Management, Halifax, pp. 139–157.
Noah, K.S., Sayer, R.L., and Thompson, D.N., 1988, “Modular bioreactor for the remediation of
liquid streams and methods for using the same,” US Patent 5,772,887.
Postgate, J.R., 1984, The Sulphate-Reducing Bacteria, Second Ed., Cambridge University Press,
Cambridge.
Rose, A.W., Stalker, J., and Michaud, L., 1998, “Remediation of acid mine drainage within strip
mine spoil by sulfate reduction using waste organic matter,” In: Proceedings, 1996 American
Society for Surface Mining and Reclamation, May 18–23, 1996, Knoxville, TN, pp. 321–335.
Saeman, J.F., Bubl, J.L., Harris, E.E., 1945, “Quantitative saccharification of wood and cellulose,”
Ind. Eng. Chem., 17, p. 35.
Singer, P.C., and Stumm, W., 1970, “Acidic mine drainage: The rate-determining step,” Science,
167, pp. 1121–1123.
Siwik, R.S., and Payant, S., 1989, “Sizing of clarifiers and thickeners for lime neutralization
plants,” In Tailings and Effluent Management, Proceedings of the International Symposium on
Tailings and Effluent Management, Halifax, pp. 159–171.
Skousen, J., 1997, “Overview of passive systems for treating acid mine drainage,” Green Lands,
Vol. 27, No. 4, pp. 34–43.
Thompson, D.N., Chen, H.-C., and Grethlein, H.E., 1992, “Comparison of pretreatment methods
on the basis of available surface area,” Bioresource Technol., 39, pp. 155–163.
204
Wakao, N., Takahashi, T., Sakurai, Y., and Shiota, H., 1979, “A treatment of acid mine water using
sulfate-reducing bacteria,” J. Ferment. Technol., Vol. 57, No. 5, pp. 445–452.
Wildeman, T., Brodie, G., and Gusek, J., 1993, Wetland Design for Mining Operations, T.
Wildeman, ed., BiTech Publishers Ltd., Richmond, B.C., Canada. Chapters 12–16.
Wildeman, T. R., Gusek, J. J., Cevaal, J., Whiting, K., and Scheuring, J., 1995, “Bio treatment of
acid rock drainage at a gold-mining operation,” In: Bioremediation of Inorganics, R.E. Hinchee,
J.L. Means, and D.R. Burris, eds., Battelle Press, Columbus OH, pp. 141–148.
205
Biotechnologies for
Remediation and Pollution
Control in the Mining Industry
L. Bernoth,* I. Firth,† P. McAllister,‡ and S. Rhodes**
ABSTRACT
As biotechnologies emerge from laboratories into mainstream application, the benefits they offer are
judged against competing technologies and business criteria. Bioremediation technologies have passed
this test and are now widely used for the remediation of contaminated soils and groundwaters. Bioreme-
diation includes several distinct techniques that are used for the treatment of excavated soil and includes
other techniques that are used for in situ applications. They play an important and growing role in the
mining industry for cost-effective waste management and site remediation. Most applications have been
for petroleum contaminants, but advances continue to be made in the treatment of more difficult organic
and inorganic species. This paper discusses the role of biotechnologies in remediation and pollution con-
trol from a mining-industry perspective. Several case studies are presented, including the land applica-
tion of oily wastewater from maintenance workshops, the composting of hydrocarbon-contaminated
soils and sludges, the bioventing of hydrocarbon solvents, the intrinsic bioremediation of diesel hydrocar-
bons, the biotreatment of cyanide in water from a gold mine, and the removal of manganese from acidic
mine drainage.
INTRODUCTION
Mining, in common with most other industrial activities, has led to environmental pollu-
tion. Soils, groundwater, and surface waters have become contaminated, and there is
increasing pressure on miners to reduce the risks associated with this contamination. The
mitigation of mining’s environmental impacts is, therefore, a major challenge for the mod-
ern mining industry.
Contaminated mining and mineral-processing sites vary widely with respect to size, soil
type, type and concentration of pollutants, underlying geology, and so on. Consequently,
there is a need for innovative remedial technologies that are cost effective and reliable.
* Australian Wool Testing Authority, Melborne, Australia.

† Rio Tinto HSE, Melborne, Australia.
‡ Centre for Air Quality Studies, Environmental Proctection Authority, Melborne, Australia.
** Rio Tinto Technical Services, Sydney, Australia.
207
TABLE 1 Bioremediation technologies for soil and groundwater
In situ techniques Ex situ techniques
Bioventing Land farming (shallow mixed beds)
Biosparging Static vented piles (biopiles)
Bioflushing Composting in piles or windrows
Lagoon treatments Soil slurry reactors
Groundwater reinjection or reinfiltration White-rot fungal processes
Natural or intrinsic remediation
BIOREMEDIATION
Bioremediation is a family of technologies that rely on microbial processes to convert envi-
ronmental pollutants to harmless products such as carbon dioxide, water, and simple inor-
ganic salts (Thomas et al., 1992). The principles underlying bioremediation processes are
well understood. The role of microorganisms as catalysts in the degradation of natural and
xenobiotic compounds is well known, and a wide range of habitats and environmental condi-
tions support microbial growth and activity.
Both organic and inorganic compounds can be biodegraded or transformed by microbial
processes. In the most common bioremediation applications, microorganisms that occur natu-
rally in contaminated soil or waters are encouraged to accelerate the degradation of organic
contaminants such as petroleum hydrocarbons by the manipulation of environmental condi-
tions, e.g., oxygen supply, nutrient concentrations, and moisture content.
Bioremediation encompasses a range of distinct techniques for managing or reducing the
health and environmental risks associated with contaminated soil and water (Table 1).
Bioremediation processes can be applied either in situ or ex situ (e.g., after groundwater
pumping or soil excavation).
Some of the processes shown in Table 1, such as soil slurry reactors and white-rot fungi,
have not yet found significant large-scale application. Most of the techniques listed are, in
theory, amenable to enhancement by bioaugmentation, but this has not yet found wide
acceptance.
TARGET POLLUTANTS AND WASTE STREAMS FROM MINING AND

MINERAL PROCESSING
Both inorganic and organic pollutants are found at mine sites. Soil and groundwater contam-
ination can occur from the storage and distribution of fuels, and effluent streams containing
hydrocarbons are produced from cleaning and servicing of mining equipment. Streams con-
taining cyanide arise in processing of mineral ores, including gold, and from the aluminum
smelting process.
Waste rock and tailings may contain low levels of metals and the oxidation of the con-
tained sulfides in these materials can lead to the generation of acid rock drainage (ARD).
Biotechnologies can play a part in the treatment of all of these waste streams.
LAND FARMING AND BIOPILES

These processes are now mature technologies for applications involving petroleum hydrocar-
bons. Many fuel-storage and retail-distribution sites have been remediated using these tech-
niques. Land farming is a robust technique that is suited for remote sites with large areas of
available land, but they can allow significant losses of volatile compounds to the atmosphere.
Vented piles (biopiles) allow control of volatile losses during the operational phase, but
they are more complex to engineer and generally do not allow for mixing of the soil to
208
overcome heterogeneity and bioavailability limitations. They are, therefore, best used for
sandy soils. The author is aware of several examples where a vented pile design or operation
was inadequate, resulting in the failure to meet performance specifications or project reme-
diation targets, principally as a result of inadequate attention being paid to soil properties,
soil-contaminant interactions, and bioavailability.
Land farming can be used effectively for the on-site treatment of small-scale hydrocarbon
spills from various fuel-transfer and refueling operations (as occur daily at mine sites). Land
farming can also be used for contaminated sludges from silt traps and settling ponds. These
materials can readily be stabilized by simple biotreatment to minimize risks from leaching or
surface-water runoff. Centralized facilities can also be used for the biotreatment of some liq-
uid wastes.
Case Study 1: Biotreatment of Oily Wastewaters from a Maintenance Operation

The treatment and disposal of wastewater by irrigation has been accepted practice for many
years. This practice is applicable to many agricultural and industrial effluents, as well as to
treated sewage effluent. Maintenance operations in the mining industry can generate large
volumes of wastewater that contain low to moderate concentrations of detergents, solvents,
and other cleaning agents, together with emulsified oils and grease. The water is generally
treated using standard physical- and chemical-separation techniques.
Most mine sites now have access to used-oil recycling facilities. However, significant quan-
tities of emulsified and adsorbed oils are often present in these streams, resulting in poor oil
separation and a low-quality effluent containing 100 to 500 mg/L of oil and grease. If not
managed properly, such effluent can result in significant environmental contamination.
Rio Tinto has successfully used enhanced biotreatment processes involving land applica-
tion at remote mining sites in the Pilbara region of Western Australia. At one operation,
workshop wastewater was being produced at a rate of 20 kL/day. (This quantity was reduced
from 40 kL/day in 1991 due to new equipment and improved work practices.) The wastewa-
ter contains oil and grease in the range of 0.1% to 5% (v/v). The oil is comprised of medium
to heavy hydrocarbons, in the C11 to C26 range, a typical distribution for lubricating oils, and
the oil contains up to 260 mg/kg PAH (primarily 2- and 3-ring PAHs including naphthalene).
The bioremediation process was designed to utilize much of the existing operations equip-
ment. Oily wastewater is irrigated in a carefully controlled program to prepared one-hectare
areas with regular tillage (approximately monthly) and nutrient application (nitrogen and
phosphorous), based on regular monitoring data.
The moisture content of the soil is one of the most important properties influencing the
microbial activity in the soil. The moisture content at a given time reflects the water balance
for the system, i.e., the inputs from wastewater and rainfall balanced against evaporation
and downward migration through the soil. There is a net evaporative loss for this site, where
the yearly rainfall average is only 250 mm. Various operational strategies are used to control
soil moisture, which has ranged from 6% to 18%. At the upper range, this approaches the
plastic limit for this soil, making tillage of the bed impractical, but it is less than 70% of field
capacity. The lower moisture limit for microbial activity is not known for this site.
The process has been in continuous operation since 1991. Changes in site practices since
1991 have resulted in a net decrease in the total hydrocarbon loading and water application
onto the treatment area. Soil analyses show that there has been no significant accumulation
of PAHs or heavy metals in the soil, and the process still plays an important role in site waste
management.
SOIL REMEDIATION BY COMPOSTING

In most contaminated soils, organic contaminants typically occur in the range of 0.001% to
1.0%, but in composting processes, organic matter concentrations would more typically be
209
20% to 80%. This high loading of biodegradable organic matter normally results in the tem-
perature of compost heaps rising to 50°C or more as microbial decomposition proceeds.
Higher temperatures increase the rates of both biological and chemical reactions in the
compost. However, thermophilic conditions do not appear to always be a precondition for
efficient contaminant biodegradation, and other physico-chemical effects are likely to be
important (Mulvey and Pittman, 1996).
In well-managed compost systems, a succession of complex and active microbial popula-
tions occurs. These can include bacteria, actinomycetes, fungi, and higher organisms such as
protozoa. The microorganisms tend to be nutritionally and metabolically diverse, occupying
a wide range of microsites within the compost, each with very different physico-chemical
parameters (Miller, 1992). Higher rates of biodegradation, in combination with physical and
chemical effects, can achieve more rapid and extensive degradation of contaminants.
Typically, significant proportions of the contaminants are partially degraded and chemi-
cally incorporated in the stable, insoluble soil-humus fraction. Composting systems have
been recognized as having potential for the treatment of contaminated soil for some years
now (Mays et al., 1989; Taddeo et al., 1989; Woodward, 1990; Simpkin et al., 1992).
Composting may be used at mine sites for stabilization of soils and sludges contaminated
with medium to heavy petroleum hydrocarbons.
Composting may have advantages where:
the soil is of poor quality with low organic content,
the soil is clayey,
the populations of hydrocarbon-degrading bacteria are inadequate,
a more rapid treatment is required, and
only small volumes of contaminated material need to be treated.
Case Study 2: Composting Clay Soil with Heavy Hydrocarbons

Heavy clay soil at a fire training site was contaminated with petroleum hydrocarbons ranging
up to 7,000 mg/kg TPH. A total of 6,000 m3 of material was excavated and placed in a wind-
row formation. Organic matter (shredded green tree waste supplemented with manure and
grass clippings) was mixed with the soil at a ratio of up to 25% (by volume). A commercially
supplied nutrient formulation (MaxBac®) was also added. The moisture maintenance target
for the mixture was 20% to 25%.
A laboratory-scale (100-L) test on a sample of the field material was performed in parallel.
This test was used to simulate ideal composting conditions to determine the likely end point
of remediation for the full-scale treatment. A comparison of the laboratory and full-scale-
treatment results is shown in Table 2.
The laboratory and full-scale treatments were comparable with respect to a decrease in
concentrations and the TPH degradation rates. Substantial reductions in TPH concentrations
were observed at both small and large scales.
TABLE 2 Comparison of laboratory and full-scale soil composting results

Result Time, days Laboratory Full-scale
TPH concentration, mg/kg soil 0 2,000 ± 500 3,000 ± 1,200
60 610 ± 90 900 ± 360
113 650 ± 100 NA
TPH decrease, % — 68 70
Average TPH degradation rate, — 23 35
mg/kg soil.day–1
210
IN SITU BIOREMEDIATION
A number of approaches are available for in situ remediation, where excavation and treat-
ment are expensive or impractical. In situ remediations often involve simultaneous soil and
groundwater remediation using combinations of groundwater manipulation, vacuum extrac-
tion and product-collection systems. The significance of biodegradation mechanisms in the
subsurface is not always appreciated.
Bioventing and biosparging are now well-established techniques that require the efficient
delivery of oxygen to contaminated zones in the subsurface. Bioventing is better developed
than sparging, largely because of difficulties in predicting and monitoring the effects of air-
flow in the saturated zone during sparging.
Case Study 3: Bioventing for Hydrocarbon Solvents

An international multidisciplinary team was established for a site assessment and remedia-
tion project at an operating site in Germany (Peck et al., 1996). A detailed chemical and
hydrogeological investigation conducted by the German team members revealed a multi-
source aquifer contamination by aromatic and aliphatic hydrocarbons as well as chlorinated
solvents.
Initial remediation focussed on the removal of volatile contaminants (VOCs) by soil vapor
extraction (SVE), with the VOCs captured on granulated activated carbon (GAC). SVE oper-
ating costs for the system were high, with the regeneration of GAC accounting for the largest
proportion. In situ bioremediation was identified as a method that offered great potential for
cost savings.
A field test program involved air-injection testing, tracer testing and in situ respiration
studies. The testing was performed to assess:
the in situ biodegradation rates in the unsaturated zone,
the comparative performance of SVE for the removal of hydrocarbons by volatilization
and biodegradation, and
the means of optimizing SVE operation for biological hydrocarbon removal or bioventing.
This field test demonstrated that VOCs were being aerobically biodegraded in the vadose
zone and that contaminant mass removal via biodegradation pathways were, in fact, approx-
imately four times what was being achieved by the volatilization pathway at the time of the
trial. Subsequently, the operating regime of the SVE was modified, resulting in a substantial
decrease in GAC consumption at the cost of a small increase in projected remediation time.
INTRINSIC BIOREMEDIATION
Many contaminants in soil and groundwater are subject to transformation and degradation
mechanisms that result in partial or complete degradation and detoxification. A large effort
has been made in the United States in recent years to understand the natural, or intrinsic,
bioremediation of dissolved BTEX plumes from leaking underground storage tanks. This
effort appears to have reached the point where it is now clear that, even without active reme-
diation, such plumes will rapidly stabilize after source removal and then reduce in extent and
concentration. In most cases, active remediation of such plumes is not required, and inter-
vention is usually only needed to protect nearby receptors.
Case Study 4: Intrinsic Bioremediation of Diesel Hydrocarbons

Intrinsic remediation processes, including biodegradation, also occur in vadose zone soils,
although these processes are not nearly as well understood. A large (800-kL) diesel spill from
an underground transfer pipe at an industrial site in New Zealand resulted in a zone of con-
tamination covering an area of approximately 4.5 ha. Soil at the site consists almost entirely
211
of quartz pea gravel, with the predominant size fractions being 1 to 20 mm. Groundwater
exists as a very active unconfined aquifer, with the water table about 3 m below the ground
level and with a seasonal water-table fluctuation of up to 0.9 m.
A product-recovery scheme succeeded in recovering 520 kL of diesel over a 12-month
period. Subsequent operation of the recovery system yielded less than a 1,000 L of diesel,
confirming that very little additional product could be recovered. Initial remediation efforts
were directed at infiltration galleries, with later pilot-scale bioventing and aquifer-sparging
techniques proving to be very effective.
However, monitoring of the NAPL plume indicated that it continued to shrink from an
estimated 100 kL when product recovery ceased in late 1992, to 30 kL in mid-1996. It
became apparent that annual fluctuations in the water table were promoting bioremediation,
as the free product was smeared over the soil surfaces throughout the fluctuation zone, as
well as inducing some air exchange. Analysis of hydrocarbon concentrations showed that the
contamination remained fully confined to the site.
Soil TPH varied considerably, but generally remained less than 5,000 mg/kg; ground-
water TPH concentrations immediately downgradient of the plume were less than 7 mg/L;
and the maximum BTEX was 130 µg/L. Both TPH and BTEX were not detected in bores
100 m further downgradient. Intrinsic bioremediation was clearly evidenced by mapping
soil gas oxygen and carbon dioxide in the smear zone, as well as by changes in dissolved
oxygen and alternate electron acceptors (iron, nitrate and sulfate) in groundwater. This
data formed the basis of a monitored natural attenuation strategy for the site.
BIOREMEDIATION OF METALS AND OTHER INORGANICS

Bioremediation has been most commonly applied to the treatment of organic compounds
such as hydrocarbons because they are very widely used and are amenable to biodegrada-
tion. Inorganic compounds associated with the mining industry are also known to undergo
biological transformations.
The biodegradation of free cyanide has been widely reported, and there are numerous
reports on the more-stable cyanide complexes, such as nickel and copper, undergoing biolog-
ical transformation (Finnegan et al., 1991).
Biological reduction of selenium oxyanions has been studied extensively. Such a reduction
is considered more cost-effective than physical or chemical treatment technologies and has
been successfully tested on a pilot scale (Cantafio et al., 1996). Some other metals can be
reduced by bacteria from a soluble form to an insoluble one, e.g., uranium(VI) to ura-
nium(IV) (Gorby and Lovley, 1992). Such engineered bioprecipitation processes are still at
the bench stage (Truex et al., 1997). Microbially mediated sequestration is thought to be
responsible for reduction in bioavailability of lead in soil (Davis-Hoover, 1997). Another
approach for metals in solution is biosorption, i.e., the adsorption of metals to the surface of
microorganisms (e.g., microbial mats) or polymers produced by bacteria or other organisms.
Reduction of sulfate to sulfide can also be used to precipitate metals from water and, per-
haps, in soils (Groudev, 1997). Sulfate-reducing bacteria (SRB) play an important role in
natural metal-removal systems.
In many cases, the metal sulfides produced represent geostable metal species, thus
decreasing the possibility of further metal mobilization. The ability to exploit sulfide produc-
tion for immobilizing metals has been used passively in wetlands systems and in engineered
bioreactors (Wildeman et al., 1995).
These technologies for metals have considerable potential for use in the mining industry
and elsewhere.
Case Study 5: Biotreatment of Cyanides

Peak Gold Mine is an underground mine in western NSW. Tailings are slurried with process
water pumped from a storage dam for underground disposal. Water used for this sandfill
212
FIGURE 1 Comparison of indigenous population with known cyanide-degrading bacteria
operation must meet regulatory limits for free (10 mg/L) and weakly complexed (WAD)
cyanide (40 mg/L). Process water typically contains 80 to 100 mg/L WAD (mostly copper)
cyanide but contains little or no free cyanide. Therefore, it is necessary to remove complexed
copper cyanides from the process water. Preliminary studies identified carbon adsorption
and biotreatment as potential low-cost process options.
Initial screening tests compared indigenous bacteria in the process water dam with known
cyanide-degrading bacteria, including a commercially available strain of Pseudomonas
fluorescens. It was found that process water contained bacteria that can reduce WAD cyanide
concentrations to below 5 mg/L (Figure 1). Inoculation was recommended to reduce lag
times. A WAD cyanide degradation rate of 1.8 mg/L/day was measured in aerated reactors
after the addition of at least 2.5 mg/L phosphorus. No other additions were necessary.
Biotreatment was selected for implementation at the mine based on life-cycle cost and
engineering considerations. After gathering baseline microbiological and chemical data, the
mine is proceeding to install surface aerators and nutrient-dosing equipment.
Case Study 6: Manganese Removal from Acidic Mine Drainage

P.T. Kelian Equatorial Mining (KEM) is a gold mine that has been majority owned and oper-
ated by Rio Tinto since 1990. The mine is located in East Kalimantan in Indonesia. A single
large pit is open-cut mined at about 40 Mt/a, of which about 9 Mt/a is ore and 31 Mt/a is
waste rock. More than 4,000 people live downstream along the banks of the Kelian River, in
close proximity to the mine. These people use the water of the Kelian for mining, domestic
purposes, transport, and fishing.
KEM is proactively managing ARD produced from waste rock through lime treatment, selec-
tive placement, wet covers, and dry covers (Firth and van der Linden, 1997). Lime treatment
occurs at points above a dual system of “polishing” ponds, which discharge into the Kelian
River. During neutralizing reactions, manganese (Mn) is released from iron-manganese and
calcium-manganese carbonates. Manganese normally exists in mine drainage in the difficult-
to-remove form of Mn(II). The removal of Mn(II) requires oxidation to Mn(III) or Mn(IV),
which are usually insoluble. Manganese (II) is quite soluble at neutral and moderately alka-
line pH. Chemical treatment of Mn(II) for removal often requires a pH that is near 10 to ini-
tiate this oxidation quickly, although biological removal of Mn(II) requires only a near-
neutral pH (Watzlaf, 1985; Nealson et al., 1988; Duggan et al., 1993).
213
TABLE 3 Gross average manganese removal in bench-scale tests
Manganese removal, %
Treatment Floating Attached
Long residence time 7.9 23.1
Short residence time 10.8 12.8
A low-cost and low-maintenance means of controlling low-concentration manganese in

seepage waters was sought for mine closure. A maximum effluent concentration of 2.5 mg/L
for the water leaving the polishing pond was required.
A passive biological system was considered to have the best chance of meeting perfor-
mance and cost criteria. Biological oxidation of metals has long been known (Berger, 1937),
and manganese oxidation has been shown to increase when using an algal-bacterial associa-
tion (Stuetz et al., 1996). Such systems require control of the algal biomass in some manner
(Mouchet, 1986).
Microbial mats—microbial communities dominated by cyanobacteria (blue-green algae)—
are self-organized laminated structures bound tightly together (and to substrates) by slimy
secretions from various microbial components (Bender et al., 1995; Goodroad et al., 1995).
The removal of metals from water by microbial mats is related to deposition of metal com-
pounds outside cell surfaces as well as chemical modification of the aqueous environment
surrounding the mats (Bender et al., 1995).
Manganese removal (as manganese oxides) by microbial mats at near-neutral pH is
favored by high dissolved oxygen and redox levels (Phillips et al., 1995) and occurs via com-
plexation/precipitation by polysaccharide bioflocculants from cyanobacteria (Bender et al.,
1994b) and coprecipitation with iron oxides and hydroxides.
Manganese oxides can also autocatalyse the oxidation of manganese (Phillips et al., 1995)
and physically encapsulate metals such as Cu, Zn, Co, and Ni (Stuetz et al., 1996). Manga-
nese removal has been reported to occur on hard surfaces (e.g., rock beds) containing a black
slimy coating (manganese oxide and microbes) (Brant and Ziemkiewicz, 1997) and in micro-
bial mats.
Manganese in black slimy coatings on rocks was found in the Kelian River. Initial tests by
Rio Tinto suggested that manganese could be removed from KEM polishing-pond water
using site microorganisms. A research program was conducted to determine how effectively
manganese could be removed using biological methods to allow modification of the polish-
ing ponds to meet post-mine closure conditions. Work was carried out in two stages: labora-
tory optimization studies and site-based field validation and kinetic testing.
Laboratory Tests
A “floating” mat and an “attached” growth configuration were evaluated. The attached
growth system consisted of curtains of mesh in the solution, which bacteria can colonize
whilst allowing a reasonably high surface area to be exposed to the passing solution.
All tests were carried out using water with 20 mg/L manganese. The dependence of Mn
uptake on residence time and concentrations of nitrogen, phosphorous, and carbon was
examined. A partial factorial experimental design was used to allow the effects of individual
parameters and interactions on the Mn removal rates to be quantified.
Gross average removal rates for the floating and attached systems were 9.4% and 17.9%,
respectively. Residence time appeared to have little effect on Mn reduction for the floating
system (Table 3). However, longer residence time appeared to increase Mn removal for the
attached system, which is consistent with literature observations. For field application, greater
contact time with the biological oxidation system can be achieved primarily by increasing the
physical size of the biotreatment system or by ensuring suitably slow flow rates.
Increases in dissolved oxygen, redox potential and pH increased Mn removed from solution.
Increased nitrogen concentrations (to ∼5.4 mg/L) enhanced Mn removal for both systems.
214
FIGURE 2 Manganese concentrations in the three test ponds, flow rate 6 L/min
Increased phosphorous inputs improved Mn removal in the attached system, but not in the
floating system. There appeared to be no direct benefit from the addition of carbon to the sys-
tem. Based on measurements of KEM polishing-pond water, it was considered that phospho-
rous levels might require supplementation, at least initially.
Site Demonstration
Site trials involved the construction of three ponds (control, attached and floating mat) to
test likely final installation configurations for post-mine closure Mn removal. Test ponds
were 5-m wide, 6.5-m long and 0.75-m deep. Input water was pumped from the upper pol-
ishing pond via a stilling-well, which split and controlled the flow between the three test
ponds. Water passed through a gravel diffuser to prevent stratified flow.
Naturally occurring manganese-tolerant bacteria and blue-green algae were collected and
isolated from several seepage areas about the site. Once isolated, the bacteria were cultured
in a series of steps in the site laboratory to provide sufficient inoculum for the test ponds.
Two flow rates were tested at residence times of about 40 and 70 hours, reflecting achiev-
able residence times for the current polishing ponds. Results of tests from April 1997 through
August 1998 are summarized in Table 4, while manganese concentration trends are pre-
sented in Figure 2.
Manganese removal rates were dominated by the input concentrations, which were often
quite variable and tended to increase throughout the test series. Removal rates ranged from
about 44 to 3,100 mg/m2/h (compared with previously reported values of 34 to 460 mg/m2/h)
for manganese input concentrations of about 5 to 18 mg/L (Bender et al., 1994a; Goodroad et
al., 1995; Phillips et al., 1995).
Manganese removal improved significantly after inoculation. Prior to inoculation, Mn con-
centrations decreased primarily through physical removal of colloidal manganese particles,
although the ponds showed some bacterial and algae growth. After inoculation, biological
oxidation is apparent (Figure 3), with the floating system usually showing the highest Mn
removal rate. Manganese removal attributable to biological oxidation appeared to be about
6% to 16% for the control pond, 6% to 29% for the attached pond, and 26% to 44% for the
floating pond.
215
TABLE 4 Field test results
Average Removal Removal Removal
Test Mn input Mn rate, Mn rate, Mn rate,
period, conc.* removal, mg/L/h removal, mg/L/h removal, mg/L/h
Test type hours (mg/L) % (mg/m2/h) % (mg/m2/h) % (mg/m2/h)
No inoculation, 100 0.66 87 0.06 <0* <0* 95 0.06
10 L/min (0.2–1.1) (49) (<0) (49)
No inoculation, 140 5.8 26 0.22 41 0.30 — —

6 L/min (0.9–7.1) (167) (222)
No inoculation, 310 7.1 56 0.40 47 0.34 68 0.48

6 L/min (6.4–8.0) (299) (255) (363)
No inoculation, 229 6.2 58 0.36 56 0.36 36 0.22

10 L/min (5.2–6.9) (269) (269) (165)
Inoculation, 455 9.5 32 0.91 84 1.4 94 1.4

6 L/min (0.9–19) (684) (1,013) (1,048)
Inoculation, 334 14 48 1.8 66 2.3 91 3.3

10 L/min (0.4–21) (1,324) (1,748) (2,473)
Inoculation, 284 13 69 2.3 85 2.8 88 2.9

6 L/min (11–16) (1,733) (2,112) (2,190)
Inoculation, 292 11 74 1.9 75 2.0 80 2.0

10 L/min (7.5–22) (1,448) (1,524)
(1,463)
Inoculation, 404 18 48 2.2 47 2.1 93 4.1
10 L/min (5.4–27) (1,639) (1,604) (3,066)
* Net increase in manganese concentrations.
FIGURE 3 Average Mn removal rates for three test ponds, before and after inoculation
216
Flow rate also appeared to influence Mn removal, which averaged 84% and 91% for the
attached and floating ponds at 6 L/min, decreasing to 64% and 88%, respectively, at
10 L/min. Further, at the lower flow rate of 6 L/min, the systems readily attained an equilib-
rium manganese-removal level during the course of the tests, despite variable input. At the
higher flow rate, the ponds appeared to be more sensitive to perturbations.
The Future
Site trials confirmed that the KEM in-river standard could be met using biological Mn
removal with realistic flow rates and input manganese concentrations. This information will
be used for the design of post-mine closure polishing ponds or wetland structures. Although
the floating system performed best in the field trials, the attached system would be more eas-
ily engineered on site. Engineered structures supporting hanging nets that increase the sur-
face area for biological growth could be made from locally available materials. These
structures could also direct stream flow in the ponds to minimize short-circuiting to the dis-
charge weir.
CONCLUSIONS
The overview and case studies presented here demonstrate the potential of biotechnologies
in the mining industry’s efforts at improving its environmental performance. Biotreatment
processes for ARD, cyanide, and sulfate reduction for metals removal are all establishing
niches at operating mine sites. While much effort goes into pollution prevention programs,
various bioremediation techniques are also well established for the clean up of historical con-
tamination at reduced costs and with better environmental outcomes than previously
achieved.
REFERENCES
Bender, J., Gould, J.P., Vatcharapijarn, Y., Young, J.S., and Phillips, P., 1994a, “Removal of zinc
and manganese from contaminated water with cyanobacteria mats,” Water Environment
Research, Vol. 66, No. 5, pp. 679–683.
Bender, J., Rodriguez-Eaton, S., Ekanemesang, U.M., and Phillips, P., 1994b, “Characterisation of
metal-binding bioflocculants produced by the cyanobacterial component of mixed microbial
mats,” Applied & Environmental Microbiology, Vol. 60, No. 7, pp. 2311–2315.
Bender, J., Lee, R.F., and Phillips, P., 1995, “Uptake and transformation of metals and metalloids
by microbial mats and their use in bioremediation,” J. Industrial Microbiology, Vol. 14,
pp. 113–118.
Berger, H., 1937, “Die Biologie der Eisenbakterien (Die Eisenfallung),” cited in Hedberg, T. and
Wahlberg, T.A., 1998, “Upgrading of waterworks with a new biooxidation process for removal
of manganese and iron,” Water Sci. Tech., Vol. 37, No. 9, pp. 121–126.
Brant, D.L., and Ziemkiewicz, P.F., 1997, “Passive Removal of Manganese from Acid Mine
Drainage,” Poster presentation at the Fourth International Conference on Acid Rock Drainage,
Vancouver, B.C., Canada, May 30–June 6.
Cantafio, A.W., Hagen, K.D., Lewis, G.E., Bledsoe, T.L., Nunan, K.M., and Macy, J.M., 1996, “Pilot
scale selenium bioremediation of San Joaquin drainage water with Thauera selenatis,” Applied
and Environmental Microbiology, Vol. 62, No. 9, pp. 3298–3303.
Davis-Hoover, W.J., 1997, “Reduced bioavailability of lead by a lead-sequestering soil bacterium”
In situ and On site Bioremediation, Vol. 3, pp. 385–390, Battelle Press, Columbus, OH.
Duggan, L.A., Wildeman, T.R., and Updegraff, D.M., 1993, “Abatement of Manganese in Coal
Mine Drainages Through the Use of Constructed Wetlands,” US Bureau of Mines, Mining
Research Contract, J021002, January.
217
Finnegan, I., Joersien, S., Abbot, L., Smit, F., and Raubenheimer, H.G., 1991, “Identification of
and characterisation of an Acinetobacter sp. capable of assimilation of a range of cyano-metal
complexes, free cyanide ions and simple organic nitriles,” Applied Microbiology and
Biotechnology, Vol. 36, pp. 142–144.
Firth, I.C., and van der Linden, J., 1997, “ARD control at P.T. Kelian Equatorial Mining,”
Proceedings of the Fourth International Conference on Acid Rock Drainage, Vancouver, B.C.,
Canada, May 30-June 6. Vol. III, pp. 1231–1250.
Goodroad, L., Bender, J., Phillips, P., Gould, J., Hater, G., and Burrow, B., 1995, “Use of
constructed mixed microbial mats for landfill leachate treatment,” In Proceedings of the 18th
Madison Waste Conference—Municipal Industrial Waste: Where Are We Going,” Madison, WI, pp.
235–247.
Gorby, Y.A., and Lovley, D.R., 1992, “Enzymatic uranium precipitation,” Environmental Science
and Technology, Vol. 26, No. 1, pp. 205–207.
Groudev, S.N., 1997, “Microbial detoxification of heavy metals in soil,” In situ and On site
Bioremediation, Vol. 3, p. 409, Battelle Press, Columbus, OH.
Mays, M.K., Sikora, L.J., Hatton, J.W.E, and Lucia, S.M., 1989, Biocycle, Vol. 30, pp. 298–300.
Miller, F.C., 1992, Soil Microbial Ecology—Applications in Agricultural and Environmental
Management, F.B. Metting, ed., Marcel Dekker, Inc., New York, pp. 515–544.
Mouchet, P., 1986, “Algae reactions to mineral and organic micropollutants, Ecological
consequences and possibilities for industrial-scale application: a review,” Water Research, Vol.
20, No. 4, pp. 399–412.
Mulvey, P., and Pittman, C., 1996, “Treatment recovery and disposal technology: Bioremediation
techniques for service station facilities,” Proceedings of 3rd National Hazardous and Solid Waste
Convention, Sydney, Australia, pp. 439–446.
Nealson, K.H., Tebo, B.M., and Rosson, R.A., 1988, “Occurrence and mechanisms of microbial
oxidation of manganese,” Advances in Applied Microbiology, Vol. 33, pp. 279–318.
Peck, P.C., Rhodes, S.H., Anderson, B.N., and Henkler, R.D., 1996, “Bioremediation of solvent
hydrocarbons. Laboratory and in situ field studies,” Proc. 5th World Congress of Chemical
Engineering, San Diego, July, Vol. 3, pp. 737–742.
Phillips, P., Bender, J., Simms, R., Rodriguez-Eaton, S., and Britt, C., 1995, “Manganese removal
from acid coal-mine drainage by a pond containing green algae and microbial mat,” Water
Science Technology, Vol. 31, No. 12, pp. 161–170.
Simpkin, T.J., Walter, D., and Doesburg, J., 1992, Proceedings of the 85th Annual Air & Waste
Management Conference, Kansas City, MO, paper No. 92–27.06, June.
Stuetz, R.M., Greene, A.C., and Madgwick, J.C., 1996, “The potential use of manganese oxidation in
treating metal effluents,” Minerals Engineering, Vol. 9, No. 12, pp. 1253–1261.
Taddeo, A., Findlay, M., Dooley-Danna, M., and Fogel, S., 1989, Biotreatment, pp. 57–62.
Thomas, J.M., Ward, C.H., Raymond, R.L., Wilson, J.T., and Loehr, R.C., 1992, “Bioremediation,”
Encyclopaedia of Microbiology, Vol. 1, Academic Press.
Truex, M.J., Peyton, B.M., Valentine, N.B., and Gorby, Y.A., 1997, “Kinetics of U(VI) reduction by
a dissimilatory Fe(III)-reducing bacterium under non-growth conditions,” Biotechnology and
Bioengineering, Vol. 55, No. 3, pp. 490–496.
Watzlaf, G.R., 1985, “Comparative tests to remove manganese from acid mine drainage,” In
Control of Acid Mine Drainages, Technology transfer seminar, US Bureau of Mines, IC 9027, pp.
41–47.
Wildeman, T.R., Gusek, J.J., Cevaal, J., Whiting, K., and Scheuring, J., 1995, “Biotreatment of
acid rock drainage at a gold mining operation,” In Bioremediation of Inorganics, R.E. Hinchee,
J.L. Means, and D.R. Burris, eds., pp. 141–148, Battelle Press, Columbus, OH.
Woodward, R.E., 1990, USATHAMA Installation Restoration Program Environmental Technology
Development, Report No. TCN 89363.
218
SECTION 4
Biomineralization
Utility of Bioreagents in Mineral Processing 221
Effect of Mesophilic Microorganisms on the Electrochemical Behavior

of Galena 229
Ocean Manganese Nodules: Biogenesis and Bioleaching Possibilities 239
Immobilization of Free Ionic Gold and L-Asparagine-Complexed Ionic Gold by

Sporosarcina ureae: The Importance of Organo-Gold Complexes
in Gold Mobility 253
219
Utility of Bioreagents in
Mineral Processing
P. Somasundaran,* Namita Deo,* and K.A. Natarajan†
ABSTRACT
Biological processes have recently become more attractive in hydrometallurgy and mineral processing
due to their lower operating costs, potential for processing low-quality ores, and flexibility. Adhesion
of microbes and their metabolic products can modify the particles to render them hydrophobic or
hydrophilic. Such modifications of the surface properties are used in flotation and flocculation of min-
erals. Microbes are also able to dissolve different precious metals from their ores. In this paper, the role
of microbes and their metabolic products as surface modifying reagents in extraction processes is dis-
cussed. In addition, the ability of certain microbes to adsorb toxic metal ions or to degrade toxic chem-
icals discharged by industries is examined.
INTRODUCTION
The increasing demand for metals and the decreasing availability of high-grade ores have led
to numerous investigations to find better processing techniques and reagents. Reagents of
biological origin are of special interest because they exhibit specific interactions with miner-
als. Microbes and their metabolic products can modify the mineral surfaces, either directly or
indirectly. The direct mechanism involves adhesion of microbes to mineral particles, while
the indirect mechanism refers to the biological reagents acting as surface-active reagents.
Both types of interactions can lead to alteration of the mineral surfaces causing flocculation
or dispersion, depending upon the nature of alterations. Also, bacteria can degrade or adsorb
pollutants including toxic metal ions from tailing ponds. Biodegradation represents a poten-
tially important and effective treatment for many of the organic industrial wastes produced
by the various mineral industries.
Recent progress on these microbial processes in both hydrometallurgy and mineral pro-
cessing is reviewed in this paper. In addition, the role of the mechanisms of interactions
found in different biological processes is discussed.
Biomineral beneficiation can be defined as a beneficiation process brought about by
microorganisms that can mediate various surface chemical processes, including:
alteration of the surface characteristics of minerals;
* Langmuir Center for Colloid and Interfaces, Henry Krumb School of Mines, Columbia University,
New York, NY.
† Department of Metallurgy, Indian Institute of Science, Bangalore, India.
221
generation of surface-active reagents; and
sorption, accumulation, and precipitation of ions and compounds.
Several types of autotrophic and heterotrophic bacteria, fungi yeasts, and algae have been
implicated in mineral-beneficiation processes. Selective leaching, flotation, and flocculation are
some of the processes involved in biomineral processing (Natarajan, 1995).
BIOLEACHING
Thiobacillus ferrooxidans is the most widely applied microorganism in the oxidation of sev-
eral sulfide minerals. The following three major mechanisms have been proposed for the
bioleaching of sulfide minerals by Thiobacillus ferrooxidans, (Gottschalk and Buchler, 1912;
Dutrizac et al., 1971; Berry and Murr, 1978; Metha and Murr, 1983):
indirect
direct, and
galvanic interaction.
In most of the leaching processes, all three mechanisms may operate simultaneously. The
indirect attack mechanism involves the leaching of sulfides by ferric sulfate, which is a meta-
bolic product of the bacteria, according to
FeS2 + 31⁄2 O2 + H2O = FeSO4 + H2SO4 (EQ 1)
and
2FeSO4 + 1⁄2 O2 + H2SO4 = Fe2(SO4)3 + 2H2O (EQ 2)

The ferric sulfate produced is an efficient oxidant capable of dissolving most of the sulfide
minerals.
Several autotrophic and heterotrophic bacteria, algae, and fungi also promote dissolution
reactions. Fungi such as Aspergillus niger and Penicillum notatum degrade bauxite and clays.
Bacillus mucilaginous excretes polysaccharides that could interact with silica, silicates, iron
oxides, and calcium oxides (Groudev et al., 1983). Another approach for processing low-
grade bauxites is through selective dissolution of iron and calcium using Aspergillus niger,
Bacillus circulans, Bacillus polymyxa, or Pseudomonus aeroginosa.
Soil bacteria and fungi play a significant role in the solubilization of phosphates. Rock
phosphates can be solubilized by strains of Rhizobium and Bradyrhizobium. Hypomicrobium
spp. and Aspergillus niger are also efficient dephosphorisers (Halder et al., 1990). Biodegra-
dation of carbonates through direct and indirect biological activity is of relevance in the
removal of calcareous gangue from ores
CaCO3 + H+ → Ca++ + HCO3– (EQ 3)

Metabolic acid generation aids the above dissolution processes. In addition, CO2 generated
during bacterial respiration can produce similar effects
CO2 + H2O → H2CO3 (EQ 4)
CaCO3 + H2CO3 → Ca++ + 2HCO3– (EQ 5)

Both organic and inorganic reagents are produced through bacterial metabolism. Different
types of reagents, such as mineral acids, fatty acids, polymers, and chelating agents, are gen-
erated by bacteria, fungi, and yeasts. All these can markedly alter the efficiency of processing
of minerals as well as pollutants.
222
Direct attack involves bacterial attachment to mineral substrates and break-up of the lat-
tice by bacterial action. A simplified reaction describing the process is
MeS + 2O2 → MeSO4 (EQ 6)

where Me represents a bivalent metal.
The galvanic conversion mechanism, on the other hand, is based on electrochemical prin-
ciples. When two sulfide minerals with different rest potentials are contacted in a leaching
medium, a galvanic cell would be formed. In such a cell, the mineral with the lower potential
serves as the anode undergoing oxidation, while the nobler sulfide acts as the cathode. The
presence of Thiobacillus ferrooxidans accelerates the electrochemical oxidation process.
Anodic oxidation of chalcopyrite in contact with nobler pyrite, with oxygen reduction
occurring on pyrite surfaces, can be represented as
CuFeS2 → Cu++ + Fe++ + 2S0 + 4e (EQ 7)
O2 + 4H+ + 4e → 2H2O (EQ 8)

Refractory gold, where gold occurs finely disseminated within host sulfide minerals such as
arsenopyrite, pyrite, chalcopyrite, and sphalerite, is not amenable to direct cyanidation, and
conventional gold extraction processes become uneconomical. To liberate the interlocked
gold, host sulfide minerals are oxidized by bacterial pretreatment, leaving the native gold
free in the residue. The biochemical reactions involved in arsenopyrite and pyrite are repre-
sented by
T. ferrooxidans
FeAsS + 31⁄2 O2 + H2O FeAsO4 + H2SO4 (EQ 9)
T. ferrooxidans
2FeS2 + 71⁄2 O2 + H2O Fe2(SO4)3 + H2SO4 (EQ 10)
Access for cyanide to gold in many refractory ores can be enhanced by microbial leaching of
the matrix material. There are also microbes that produce enzymes that can complex with
gold. The potential for such direct microbial treatment of gold and other precious or rare
metals is very clear (Polkin et al.1982).
BIOFLOTATION
In addition to leaching, bacterial conditioning with T. ferrooxidans can produce significant
surface modification of sulfide minerals to affect their flotation behavior. For example,
T. ferrooxidans can enhance the flotation of pyrite in sulfuric acid medium at pH 2 with very
little effect on the flotability of galena. The availability of initial biomass and the period of
bacterial mineral interaction where variables influencing the flotability of both galena and
the sphalerite. With an increase in the cell concentration and reaction time, the flotability of
both galena and sphalerite were deleteriously affected. For sphalerite, 94% recovery was
obtained after treatment with 108 cells/mL. This decreased to about 53% when the cell con-
centration was increased ten fold. On the other hand, the flotability of galena was observed
to be affected more by an increase in the initial biomass (Yelloji Rao et al., 1991,1992). It
would thus be possible to separate sphalerite from mixed zinc-lead sulfide ores through bac-
terial pretreatment under controlled conditions.
Similar surface modification by T. ferrooxidans finds application also in coal cleaning pro-
cesses. Conditioning of coal with T. ferrooxidans can cause depression of pyrite without
affecting the flotability of coal (Botang, 1997). In addition to pyrite suppression, bacterial
conditioning with T. ferrooxidans can also help in the removal of ash from coal (Bos and
Kuenen, 1990).
223
It has been shown that the natural flotability of pyrite dropped from 85% to less than 10%
after conditioning with the bacteria. Traditional flotation reagents are known to increase the
flotation recovery of coal, total sulfur content, and total ash content. In contrast, bacterial
conditioning produces no effect on coal flotation but decreases the total sulfur content
(Botang, 1997).
Biomass can act as a flotation reagent for the beneficiation of nonsulfide minerals. Recent
studies have established that Bacillus polymyxa, a soil bacterium indigenously associated
with several iron ore and bauxite deposits, can significantly alter the surface chemistry of
minerals such as hematite, corundum, calcite, and quartz (Deo and Natarajan, 1997). The
flotability of quartz was found to increase significantly, attaining a steady-state recovery of
60% after three hours. On the other hand, the flotability of corundum, hematite, and calcite
was deleteriously affected due to such interactions. Thus, it is clear that bacterial interaction
can render the quartz surface more hydrophobic and hematite, corundum, and calcite sur-
faces more hydrophilic. It should then be possible to separate silica from iron oxide, alumina,
or limestone efficiently through biotreatment. It has also been observed that, in the presence
of small dosages of collector, close to 100% separation of quartz from hematite and corun-
dum is possible. The mechanisms for these interactions are, however, yet to be understood.
Opportunity exists to explore these phenomena and to achieve effective separation through
the control of cell biomass and flotation time.
BIOFLOCCULATION
There are yet other possible uses of microorganisms in mineral engineering that have
received little attention. Many microorganisms possess flocculent growth habit and produce
extracellular polymers such as polysaccharides, polypeptides, and polyglycans. Such biologi-
cally generated polymers can cause flocculation of the microorganisms as well as of the min-
eral particles. Mainly, the molecular architecture of the bacterial cell wall determines the
forces of dispersion and agglomeration.
Mycobacterium phlei, a gram-positive, rod-shaped, prokaryotic cell has been observed to
be useful as a flocculent for phosphate slimes, hematite, and coal fines (Smith et al., 1991;
Mishra et al., 1993). At mildly acidic pH values, this organism flocculated hematite but not
quartz. Substantial settling of hematite fines could be observed within a few minutes after
bacterial interaction.
Selective flocculation of fine coals with respect to the separation of pyritic sulfur could
also be achieved in the presence of M. phlei. More than 85% of pyritic sulfur could be
removed from coal by bioflocculation. Biosurfactants secreted by such organisms may be pri-
marily responsible for the flocculent action. The mechanisms involved in bioflocculation may
be similar to those encountered in other systems involving bridging, charge neutralization,
hydrophobic bonding and particle-to-particle contacts but relative roles of various factors are
not well established.
Bacillus polymyxa, a gram-positive rod, has been observed to enhance the flocculation of
hematite, corundum, and calcite (Deo and Natarajan, 1997). On the other hand, it disperses
quartz by making its surface hydrophobic. For example, while about 99% of hematite, cal-
cite, and corundum could be settled within two minutes in the presence of 109 cells/mL in
the pH range of 4 to 7, only about 10% of the quartz particles could be settled under the
above conditions. Selective bioflocculation could prove to be useful for the separation of sil-
ica from hematite, corundum, and calcite under the above conditions.
BIOENVIRONMENTAL CONTROL
Microorganisms are capable of concentrating metals from aqueous solutions through extra-
cellular precipitation, volatilization, complexation and accumulation, cellular binding, and
accumulation in intracellular spaces. Biosorption and bioaccumulation can be advantageously
224
used for the removal of toxic metals such as copper, lead, zinc, mercury, chromium, and ura-
nium from mineral processing effluents. Both living and dead biomass can be used for the
above purpose. The use of dead biomass has advantages in that:
no nutrient supply is needed,
the ready availability at very low costs as a waste material, and
the absence of metal-toxicity problems.
The type and amount of biomass, as well as the solution chemistry of the effluents, affect
metal uptake by biomass.
The following four major mechanisms have been identified for the removal of metal ions
from aqueous solutions by microbial actions (Kelly et al., 1979):
Bioadsorption consisting of the exocellular or pericellular binding of metals, which may
involve either:
— cation adsorption, which can be complexed to the behavior of commercial carboxylic
cation exchange resins and
— metal deposition precipitation at microbial surfaces, which may occur as surface com-
plexation or chelation or as entrapment by extracellular organelles.
Bioaccumulation consisting of intracellular uptake of metals, which may occur by:
— direct intracellular uptake of metals,
— uptake via monovalent cation transport systems,
— uptake via undefined processes, and
— uptake via anion transport pathways.
Production of insoluble compounds with precipitation of:
— metal sulfides and
— metal oxides
Miscellaneous microbe-metal interactions
Biosorption and bioaccumulation are processes where metal removal is a direct consequence
of the physical contact of microorganisms with metal ions. However, the metals may also be
removed from solutions without the need for this physical contact. This phenomenon is
encountered whenever metal ions combine with the direct or indirect products of microbial
metabolism and is, therefore, dependent on the existence of a viable biomass. In this case,
microorganisms can be considered as mere low-cost producers of the chemical compounds
that capture metal ions. These compounds can be either the products of microbial degrada-
tion of organic matter or metabolites. Heavy metals can be effectively removed from solu-
tions depleted of oxygen, for instance, by the activity of anaerobic, heterotrophic organisms.
This, in effect, reduces the environment thus created and favors the development of a micro-
flora of sulfate-reducers. The H2S produced by the microorganisms combines with heavy-
metal cations to form insoluble sulfides, which precipitate out of solution according to the
simple reaction:
Me+2 + H2S → MS + 2H+ (EQ 11)

Through their activity, microorganisms can also induce changes in the oxidation states of
some elements, thus producing less soluble ionic species. A typical example is the oxidation
of soluble ferrous to much less soluble ferric iron by T. ferrooxidans.
A number of different waste substances, both toxic and nontoxic, are discharged by vari-
ous mineral-industry operations. Many of these substances are complex organic chemicals
used in flotation or hydrometallurgical plants. Others include various petroleum products
used in general mill and mine equipment operation and maintenance. The handling,
225
treatment and proper disposal of these substances can be a difficult and expensive task.
Bioremediation represents a potentially effective treatment for many of the organic wastes
produced by the mineral industries.
Bioremediation is a process in which toxic components are degraded into harmless prod-
ucts rather than just adsorption of toxic ions on microbial surfaces. An interesting example is
the action of the cyanide-tolerant bacteria in the effluent from the gold-extraction processes
(Whitlock, 1989). There are also microorganisms that can break up natural petroleum prod-
ucts from groundwater and contaminated sites (Foght et al., 1990). There are now at least
50 companies in the United States alone and over 100 companies worldwide using this tech-
nology to clean up sites ranging from gasoline-soaked soil to supervened areas flooded with
carcinogens. Solozhenkin and Lyubavina (1980) investigated the biodegradation of thiol col-
lectors, such as potassium 2 21,6,61–tetramethyl-iminoxyl-xanthate (KTIX), potassium butyl
xanthate (KBX), and sodium diethyl dithiocarbamate (NaEC), by bacteria. In the presence of
bacteria, KTIX was 100% destroyed within 45 minutes, while, in the absence of bacteria, only
45% of the compound was destroyed. In another study, biodegradation of different collec-
tors, such as sodium oleate, sodium isopropyl xanthate, isododecyl oxypropyl aminopropyl
amine, and ammonium acetate by Bacillus polymyxa (Deo and Natarajan, 1998) was investi-
gated. In the presence of bacteria, about 150 ppm of amine collector was destroyed within
two hours, xanthate within five hours, and oleate in around six hours, while, in the absence
of bacteria, the above collectors remained unaltered. These studies showed that flotation
collectors can be effectively destroyed by bacterial degradation and the procedure may be
a viable method for treating such wastes.
From the viewpoint of microbial-mediated removal of chemicals from wastewaters,
manmade organic compounds can be classified into amenable and refractory to microbial
degradation. The organic compounds belonging to the first class can be eliminated by:
biotransforming them into innocuous forms,
degrading them by mineralization to carbon dioxide and water,
anaerobically decomposing them to carbon dioxide and methane, and
volatilizing them.
The organic compounds falling into the second class are removed by sorption (Rossi, 1990).
CONCLUSIONS
It is clear that microbes and their metabolite products can markedly alter the surface proper-
ties of mineral particles. Processes such as bioflotation and bioflocculation can take advan-
tage of such surface alteration to achieve selective separations. Although preliminary studies
have shown potential applications of biological processes in mineral processing, very few of
these are being practiced commercially. Furthermore, there is limited information available
on the mechanisms of the attachment of microbes and microbial products.
The wide use of microbes or their production on an industrial scale can be expected to occur
with improvement in the understanding of the mechanisms of bacterial/mineral interactions.
Microorganisms could also be efficiently used in environmental control in mining and mineral
processing. Biosorption, bioaccumulation, and biodegradation can be used with advantage to
detoxify liquid and solid wastes. The potential for use of microbe-mediated processes for both
extraction of values and removal of pollutants is clear. It will be useful to develop an under-
standing of the basic processes behind the microbial effects to fully realize such potential.
226
REFERENCES
Berry, V.K., and Murr, L.E., 1978, In Metallurgical Applications of Bacterial Leaching and Related
Microbiological Phenomena, L.E. Murr, A.E. Tormaand, and J.A. Brierley, eds., Academic Press,
New York, p. 103.
Bos, P., and Kuenen, J.G., 1990, In Microbial Mineral Recovery, H.L. Ehrlich and C.L Brierley, eds.,
McGraw-Hill, New York, p. 343.
Botang, D.A.D., and Philips, C.R., 1977, Sep. Sci., Vol. 12, p. 71.
Deo, Namita, and Natarajan, K.A., 1997, Minerals Engineering, Vol. 10, p. 1339.
Deo, Namita, and Natarajan, K.A., 1998, Minerals Engineering, Vol. 11, p. 717.
Dutrizac, J.E., MacDonald, R.A.J., and Ingraham, T.R., 1971, Can. Met. Qtly, Vol. 10, p. 3.
Foght, J.M., Fedorak, P.M., Gray, M.R., and Westlak, D.W.S., 1990, In Microbial Mineral Recovery,
H.L. Ehrlich and C.L. Brierley, eds., McGraw-Hill, New York, p. 379.
Gottschalk, V.H., and Buchler, H.A., 1912, Econ. Geology, Vol.7, p. 15.
Groudev, S.N., Groudev, V.I., and Petrov, E.C., 1983, Travaux ICSOBA, Vol. 13, p. 249.
Halder, A.K., Mishra, A.K., Bhattacharya, P., and Chakrabarty, P.K., 1990, J. Gen. Appl. Microbiol,
Vol. 36, p. 81.
Kelly, D.P., Norris, P.R., and Brierly, C.L., 1979, Microbiological Methods for the Extraction and
Recovery of Metals in Microbial Technology, Current State, Future Prospects, A.T. Bull, D.C.
Ellwood, and C. Ratledge, eds., Cambridge University Press, p. 263.
Metha, A.P., and Murr, L.E., 1983, Hydrometallurgy, Vol. 9, p. 235.
Mishra, M., Smith, R.W., Dubel, J., and Chen, S., 1993, Minerals and Metallurgical Processing, Vol.
10, p. 20.
Natarajan, K.A., 1995, In Selected Topics in Mineral Processing, Pradip and Rakesh Kumar, eds.,
New Age International (p) Ltd., Wiley Entern, New Delhi, p. 214.
Polkin, S.I., Adamov, E.V., and Panin, V.V., 1982, “Technology of bacterial leaching of nonferrous
and rare metals,” Nedra, Moscow, p. 242.
Rossi, G., 1990, Biohydrometallurgy, McGraw-Hill, Gmbh, Hamburg, p. 555.
Smith, R.W., Mishra, M., and Dubel, J., 1991, Minerals Engineering, Vol. 4, p. 1123.
Solozhenkin, P.M., and Lyubavina, L., 1980, “Biogeochemistry of ancient and modern
environments,” Proceedings Fourth International Symposium on Environmental Geochemistry,
P.A. Trundiger, M.R. Walter, and B.J. Kalph Springer, eds., New York, pp. 615.
Yelloji Rao, M.K., Natarajan, K.A., and Somasundaran, P., 1991, Mineral Bioprocessing, R.W.
Smith, and M. Mishra., eds., The Minerals Metals and Material Society, Warrendale, p. 105.
Yelloji Rao, M.K., Natarajan, K.A., and Somasundaran, P., 1992, Minerals and Metallurgical
Processing, Vol. 9, p. 95.
Whitlock, J.E., 1989, In Biohydrometallurgy, J. Salley, R.G.L. McCready, and P.L. Wichlacz, eds.,
CANMET, Ottawa, Canada, p. 613.
227
Effect of Mesophilic
Microorganisms on the
Electrochemical Behavior of Galena
J.L. González-Chávez,* F. González,† A. Ballester,† and M.L. Blázquez†
ABSTRACT
In recent years, the bioleaching of sulfidic raw materials has been of great interest. However, the appli-
cation of this biotechnology is now very limited for many different complex reasons. The mechanism of
the chemical and electrochemical reactions taking part in the process is not well understood. The
application of electrochemical techniques is useful in its elucidation. Among the common sulfides,
galena has been studied very little with respect to its behavior during bioleaching. Therefore, the elec-
trochemical evolution of massive electrodes of lead sulfide in the presence and absence of a mixed cul-
ture of mesophilic microorganisms was investigated by the application of potentiometry, anodic and
cathodic polarization, and cyclic voltammetry. The agitation rate, the use of pH control, and the com-
position of the nutrient mediums did not have any significant affects on the oxidation of galena. How-
ever, the temperature (25° to 45°C), the scan rate (4 to 100 mV/s), and the acid used to adjust the
pH were significant variables that influenced sulfide oxidation. Therefore, the peak (600 to 1,000 mV)
corresponding to the oxidation of sulfide ions from the mineral was only detected in the presence of a
complexing anion of lead. In the presence of bacteria, the peaks corresponding to galena oxidation
decreased and, at the same time, new peaks appeared both during the oxidation at –300 mV and dur-
ing the reduction of the mineral surface at 150 and –300 mV. These signals were more important
when the attack time increased. The rest potential of bioleached mineral increased with an increase in
leaching time, which was related to the electrochemical transformation of the galena surface.
INTRODUCTION
The use of microorganisms in metal solubilization processes has attracted a great deal of
attention in recent years. The leaching of metals using microorganisms is very attractive
because of its low cost and because the method has relatively fewer pollution problems com-
pared to conventional hydrometallurgical processes. Metals are commonly associated with
* Dpto. de Química Analítica, Facultad de Química, U. N. A. M. Cd. Universitaria, México, D.F. México.
† Dpto. de Ciencia de Materiales, Facultad de Ciencias Químicas, Universidad Complutense de Madrid.
Cd. Universitaria, Madrid, España.
229
various forms of inorganic sulfides in nature. It has been well documented that these metals
are constantly solubilized by microbiological actions (Bosecker, 1984; Rossi, 1990).
Some microorganisms can derive energy by oxidizing ferrous ions or reduced forms of sul-
fur compounds. The oxidation of sulfur compounds results in the solubilization of heavy met-
als and lowering the pH of the environment. These microorganisms are known to develop
extraordinary tolerance to metals, which are often toxic to other microorganisms. Numerous
researches (Lizama and Suzuki, 1989; Sand et al., 1992) have documented the application of
T. ferrooxidans and L. ferrooxidans in the leaching of heavy metals.
Galena (PbS) is the major commercial source of lead. The recovery of lead from galena via
hydrometallurgical processes has been studied by several researchers (Scott and Nicol, 1976;
Fuerstenau et al., 1986; Fuerstenau et al., 1987). However, few researchers have examined
the applications of microorganisms to the extraction of lead from lead-bearing ores. Torma
and Subramanian (1974) first reported the participation of T. ferrooxidans in the leaching of
galena concentrate. Later, Tomizuka and Yagisawa (1978) examined the role of bacteria and
proposed a basic scheme of galena leaching in the presence of T. ferrooxidans. In accordance
with these last authors, the bacterial leaching mechanism may be represented by
2PbS + 2H2SO4 + O2 → 2PbSO4 + 2H2O + 2S (EQ 1)
2S + 2H2O + 3O2 → 2H2SO4 (EQ 2)

Galena can serve as the sole energy source for bacteria at acidic pH values. Elemental sul-
fur generated as a result of Reaction (1) is further oxidized by bacteria to produce sulfuric
acid, as shown in Reaction (2).
Understanding the interactions of microorganisms with galena is important, although the
pertinent reaction mechanisms have received only limited attention. Nevertheless, Habashi
(1983) demonstrated that most metallic sulfides are dissolved by means of an electrochemi-
cal mechanism, but in the absence of bacteria.
This work examines the influence of microorganisms on galena attack by using rest poten-
tial measurements, anodic and cathodic polarization, and cyclic voltammetry. These tech-
niques were selected because they are among the most versatile electrochemical techniques
available for the mechanistic study of redox systems and because they help in identifying the
chemical species formed on the mineral surface. The primary aim of this work was to study
the electrochemical response of a massive galena electrode in a bacterial culture medium to
understand the dissolution and nature of the products formed during the bioleaching process.
EXPERIMENTAL
Material
Samples of massive galena from Compañía la Cruz (Linares, Jaen, Spain) were used. Chemi-
cal analysis, together with X-ray diffraction, showed a highly pure galena with very few
superficial defects. Galena was analyzed by SEM and EDS before electrochemical tests, and
no other mineral phases were detected on the samples.
Bacterial Cultures
A mixed culture of mesophilic microorganisms containing T. ferrooxidans, T. thiooxidans, and
L. ferrooxidans was used as inoculum. This culture has been grown and used for years in
Complutense University laboratories. Microorganisms were grown in 250-mL Erlenmeyer
flasks containing 9K nutrient medium without iron (II) (3.0 g·L–1 (NH4)2SO4, 0.5 g·L–1
MgSO4·7H2O, 0.5 g·L–1 K2HPO4 and 0.01 g·L–1 Ca[NO3]2) at pH 1.8 and nonsterile ground
galena (80 µm) at 5% (w/v) pulp density as energy source. Cultures were maintained in an
orbital shaker at 150 rpm and at 35°C.
230
In some tests, a different nutrient medium (diluted Norris) (Norris and Kelly, 1983) con-
taining 0.4 g·L–1 (NH4)2SO4, 0.5 g·L–1 MgSO4·7H2O and 0.2 g·L–1 K2HPO4 was used. All media
were prepared from analytical-grade chemicals and distilled water.
Before carrying out the actual test, all of the cell suspension used as inoculum were
adapted to galena by conventional techniques (Brierley, 1978).
The progress of leaching was monitored every 48 hours through the measurements of pH
and redox potential, in this case, using a platinum electrode.
Bacterial Adaptation
A mesophilic culture, adapted for 12 weeks on galena, was used as inoculum for the bacterial
attack of the tested electrodes during different periods. Bacteria were adapted to a 5% (w/v)
pulp density of galena for 12 weeks. In this period, bacteria were reinoculated on a new sam-
ple of galena every four weeks to provide fresh surfaces for the bacterial growth. At the end
of this adaptation period, the bacteria were used for electrochemical bioleaching tests.
Electrode Preparation
The electrodes were made from massive galena. The specimens were cube shaped (≈5 mm
on each edge) and, as far as possible, they had no visible imperfections. The mineral sample
was connected to a copper wire by means of tin welding, and the sample was isolated from
the liquid medium by a glass tube. The mineral electrode was embedded in epoxy resin to
obtain an exposed area of 0.5 cm2. Before each test, the working electrode surfaces were pol-
ished with #600 silicon carbide paper. This was followed by rinsing with distilled water.
To study the electrochemical behavior of the galena during bioleaching, the electrodes
were previously treated (for 7, 14, and 21 days) in an orbital shaker at 35°C with solutions
containing the above-mentioned nutrient mediums and the bacterial inoculum. The elec-
trodes were removed from the bioreactor at different time intervals and tested in the corre-
sponding electrochemical experiments.
Electrochemical Measurements
The electrochemical experiments were carried out at constant or programmed potentials
using a Wenking potentiostat (Model LB 81 M) and a Wenking voltage scanner (Model MVS
87), both from Bank Electronic. Data were acquired by an analogue board installed in a PC.
When necessary, stirring was accomplished with a magnetic-stirrer, with the working elec-
trode remaining stationary. Reported potentials are referred to the Ag/AgCl electrode
(+0.207 vs. SHE at 25°C). The following three types of electrochemical experiments were
performed: rest potential determination, measurement of potentiodynamic polarization, and
cyclic voltammetry.
The electrochemical cell was provided with a jacket connected to a recirculation pump for
circulating water of constant temperature (±0.1°C).
Electrochemical measurements were made with a conventional three-electrode system,
namely, a working electrode (galena), a counter electrode (platinum spiral wire) and a refer-
ence electrode (Ag/AgCl) provided with a Luggin-Haber capillary tip filled with a 3 M solu-
tion of potassium nitrate. Electrochemical experiments were carried out at 35°C with 100 mL
of electrolyte at pH 1.8 and in the same nutrient medium used for bacterial grown unless
otherwise stated.
EXPERIMENTAL RESULTS AND DISCUSSION
Electrochemistry of Galena Dissolution in the Absence of Microorganisms. Electro-

chemically, the dissolution of galena in an acid medium is characterized by three electroac-
tive zones:
231
FIGURE 1 Anodic polarization of galena at different scan rates
a low potential zone, ≈ +350 mV, where the galena is transformed into intermediate
products (So);
a medium potential zone, ≈ +700 mV, which is probably related to an adsorption or cat-
alytic wave; and
a high potential zone, ≈ +900 mV, which describes the oxidation of sulfur (elemental
sulfur and sulfide) to sulfate.
Figure 1 shows the anodic polarization curves of the galena under abiotic conditions in a
fresh nutrient medium at different scan rates. These curves show the expected wave at
potentials below 350 mV, corresponding to the oxidation of sulfide according to
PbS → Pb2+ + So + 2e– (EQ 3)

Above +900 mV, the dissolution of the galena takes place according to
PbS + 4H2O → PbSO4 + 8H+ + 8e– (EQ 4)

The rate of this dissolution process does not increase at higher potentials due to the forma-
tion of a layer of PbSO4 over the galena surface.
These curves, current density vs. potential (E), were obtained during a single triangular
potential sweep (STPS) between –200 mV and the selected anodic potential at three differ-
ent scan rates (v), namely, 100, 20, and 4 mVs–1.
Galena oxidation depends on scan rate. Three waves were obtained at the three scan rates,
but the current density increases with an increase of this variable. This is because, when the
scan is very low, the amount of sulfur produced is more important and this layer produced a
decrease in current density at higher potentials.
These waves presented a slight shift toward more positive potentials when the scan rate
increased.
To study the influence of the temperature, pH, agitation and different nutrient media, sim-
ilar tests were carried out at different scan rates using the following conditions:
temperature: 25°, 35°, and 45°C;
pH: 1.5, 1.8, and 2.5;
with and without agitation; and
with different nutrient medium (i.e., 9K, diluted Norris and sulfuric acid).
232
FIGURE 2 Anodic polarization of galena at different temperatures
FIGURE 3 Anodic polarization of galena at different pH values
The potentiodynamic curves (Figure 2) show that, during the anodic scan, the current
density, at ≈350 mV, increases when the temperature increases. At ≈900 mV, the same trend
is observed, although there is an inversion between the current densities at 35° and 45°C.
For the first wave, the potentials are less positive when the temperature increases, but, for
the third one, the potentials are more positive when the temperature increases. The second
wave almost disappears with an increase in temperature, which is in accordance with the
assumption that this wave corresponds to an adsorption or catalytic wave.
For different pH values (Figure 3), the current density trend was an increase when the pH
decreased. In the first wave, the potentials decrease with decreasing pH, and, for the third
wave, the potential decreases with increasing pH. For the second wave, its current density
significantly decreases with a pH increase.
Later, it will be shown that this effect is not due to pH, but it is due to the sulfate concen-
tration in the medium.
The acid used to fix the pH at 1.8 had a very important effect on the electrochemical sig-
nals (Figure 4). The peak corresponding to the oxidation of sulfide ions from the mineral was
only detected when the acid used was a complexing anion of lead, such as H2SO4 and HCl.
Perchlorate (ClO4–) does not complex lead.
233
FIGURE 4 Anodic polarization of galena at pH fixed with different acids
FIGURE 5 Anodic polarization of galena at a fixed pH (1.8) using HClO 4 and different
concentrations of Na2SO4
The presence of the complexing anions helps the oxidation of sulfide ions according to
PbS → Pb2+ + So + 2e– (EQ 5)
Pb2+ + SO42– → PbSO4 (EQ 6)
Pb2+ + 2Cl– → PbCl2 (EQ 7)

PbCl2 is significantly more soluble than PbSO4, which favors the sulfide oxidation in the pres-
ence of Cl– ions.
In Figure 5, the signals obtained for the galena dissolution at pH 1.8 in the presence of HClO4
and with additions of Na2SO4 are shown. The anodic signal increases with the addition of
Na2SO4. These results confirm the effect of complexing anions from the acid medium used to fix
the pH. The same results were obtained when NaCl was added to HClO4.
Agitation did not affect the galena dissolution greatly. Also, the different nutrient media
(9K, diluted Norris, and sulfuric acid) had no affect on dissolution. The current density did
increase with an increase in ionic strength (concentration of salts) of the medium, i.e.,
changing from a sulfuric medium to 9K medium (Figure 6).
234
FIGURE 6 Anodic polarization of galena at pH 1.8 in different media
FIGURE 7 Cyclic voltammetry of galena at pH 1.8 in 9K medium and 20 mV·s –1
Cyclic Voltammetry of Galena

Figure 7 shows the cyclic voltammetric curve for galena in 9K medium at pH 1.8 and 35°C.
These curves were obtained through a cyclic triangular potential sweep from –600 to
+1,400 mV at 20 mV.s–1.
Three peaks during oxidation and one corresponding to reduction of galena to metallic
lead could be scanned.
The first wave observed decreases after the first cycle, whereas the following two signals
almost disappear after the same cycle. This is attributable to the formation of sulfur layer on
the galena surface.
Electrochemistry of Galena Dissolution with Microorganisms

The anodic scan of galena (Figure 8) shows that, in comparison to the behavior of the min-
eral in the absence of bacteria, the first oxidation wave decreases with the duration of
bioleaching, and the second and third waves almost disappear.
As mentioned above, the first wave is related to the production of elemental sulfur,
which microorganisms help to oxidize to sulfate. The result is the formation of a mixed
235
FIGURE 8 Anodic polarization of galena after seven days of bioleaching at 35°C
FIGURE 9 Cyclic voltammetry of galena after bioleaching at 35°C
layer of elemental sulfur and lead sulfate on the galena surface. This layer prevents
further oxidation of the galena. The current density was not affected by the stirring of
the electrolyte, which suggest that a layer, possibly of elemental sulfur, was adsorbed
onto the electrode.
Pretreatment of the galena electrode with mesophilic bacteria increased its rest poten-
tial with respect to that measured in a fresh nutrient medium (Table 1). The reason was
the formation of different reaction products on the electrode surface in accordance with
an increase in the bioleaching time.
Figure 9 shows the curves corresponding to the cyclic voltammetry behavior of galena
after 7 and 14 days of bioleaching. These curves were obtained in the following condi-
tions: fresh nutrient medium (without microorganisms), at pH 1.8, and with a scan rate
of 20 mV/s.
The bioleached galena showed three new signals in the voltammetric scan, which
have been marked with an asterisk in Figure 9, one of them corresponding to an oxida-
tion and the other two to reductions. These signals became more intense as the duration
of bioleaching increased.
236
The oxidation wave in Figure 9 was related to the oxidation of hydrogen sulfide, with the
production of elemental sulfur on the galena surface. This wave appeared only after one
complete scan cycle.
The first reduction wave placed at ≈+250 mV would be related to the following reduction
reaction
So + 2e– → S2– (EQ 8)

The second reduction wave, placed at ≈–300 mV, was the most pronounced in the
cathodic scan and was probably due to the reduction of S° according to
So + 2H+ + 2e– → H2S (EQ 9)

where the E-pH dependency can be expressed as
E = E′ – 0.059·pH (EQ 10)

Each of the three new signals obtained increased with an increase in the duration of leaching.
When the cyclic voltammetry was carried out repeatedly on the bioleached galena elec-
trode, these new waves gradually disappeared, and the signals obtained were the same as
those obtained in the absence of microorganisms (Figure 7).
The oxidation peak at ≈+300 mV, corresponding to oxidation of PbS to elemental sulfur
(Reaction (3)), decreased due to the formation of a layer of elemental sulfur on the galena
surface as a consequence of the activity of the microorganisms.
CONCLUSIONS
Anodic dissolution of galena was not influenced by agitation and the different nutrient
media used.
Scan rate, temperature, pH and the acid used to fix pH were significant variables influ-
encing sulfide oxidation. A peak corresponding to the oxidation of sulfide ions from the
mineral was only detected in the presence of complexing anions of lead.
Very important differences in the electrochemical response of the massive galena elec-
trodes were noted after bioleaching at 35°C with mesophilic microorganisms.
Mesophilic microorganisms gradually modified the original galena surface, increasing
the rest potential of the mineral.
In the presence of bacteria, the peaks corresponding to the galena oxidation decreased
and, at the same time, new peaks appeared during both the oxidation and the reduction
of the mineral surface. These signals became more intense with an increase in bioleach-
ing duration.
ACKNOWLEDGMENTS
The authors wish to thank DGAPA-UNAM (Mexico) and CICYT (Spain) for partial financial
support to carry out this research.
REFERENCES
Bosecker, K., 1984, “Biodegradation of sulfur minerals and its applications for metal recovery,”
Studies in Inorganic Chemistry, Vol. 5, pp. 331–348.
Brierley et al., 1978, “Bacterial Leaching,” CRC Crit. Rev. Microbiol., Vol. 6, pp. 207–262.
Fuerstenau, M.C., Chen, C.C., Han, K.N., and Palmer, B.R., 1986, “Kinetics of galena dissolution in
ferric chloride solutions,” Met. Trans. B., Vol. 17B, pp. 415–423.
237
Fuerstenau, M.C., Nebo, C.O., Elango, B.V., and Han, K.N., 1987, “The kinetics of leaching of
galena with ferric nitrate,” Met. Trans. B., Vol. 18B, pp. 25–30.
Habashi, F., 1983, “Dissolution of minerals and hydrometallurgical processes,”
Naturwissenschaften, Vol. 70, pp. 403–411.
Lizama, H.M., and Suzuki, I., 1989, “Bacterial leaching of a sulfide ore by Thiobacillus
ferrooxidans,” Hydrometallurgy, Vol. 22, pp. 301–310.
Norris, P.R., and Kelly, D.P., 1983, “Iron and mineral oxidation with Leptospirillum-like bacteria,”
Recent Progress in Biohydrometallurgy, G. Rossi and A.E. Torma, ed., Asoc. Minera, Sarda,
Iglesias.
Rossi, G., 1990, Biohydrometallurgy, McGraw-Hill, New York.
Sand, W., Rohde, K., Sabotke, B., and Zenneck, C., 1992, “Evaluation of Leptospirillum ferrooxidans
for leaching,” Applied and Environmental Microbiology, Vol. June pp. 85–92.
Scott, P.D., and Nicol, M.J., 1976, “Kinetics of non-oxidative dissolution of galena in acidic
chloride solutions,” Transactions IMM, London, Vol. 85, pp. C40–44.
Tomizuka, N., and Yagisawa, M., 1978, “Optimum conditions for leaching of uranium and
oxidation of lead sulfide with Thiobacillus ferrooxidans and recovery of metals from bacterial
leaching solutions with sulfate reducing bacteria,” Metallurgical Applications of Bacteria
Leaching and Related Microbial Phenomena, L.E. Murr, A.E. Torma, and J.A. Brierley, eds.,
Academic Press, New York, pp. 321–344.
Torma, A.E., and Subramanian, K.N., 1974, “Selective bacterial leaching of a lead sulfide
concentrate,” Int. J. Miner. Process., Vol. 1, pp. 125–134.
238
Ocean Manganese Nodules: Biogenesis
and Bioleaching Possibilities
H.L. Ehrlich*
ABSTRACT
Marine manganese nodules harbor a mixed population of manganese-oxidizing and -reducing bacte-
ria, as well as bacteria that do not act on manganese. They also host phagotrophic protozoa. Out of
several laboratory studies, one line of investigation indicates that some kinds of bacteria found on
nodules can catalyze the oxidation of Mn 2+ bound to Mn(IV) oxide at the surface of nodules at in situ
temperature and pressure. In the absence of appropriate bacteria, the rate of oxidation of the surface-
bound Mn2+ is significantly slower. How Fe(III) is incorporated is still unclear. Base metals such as
Cu, Ni, and Co, which are usually found in marine nodules in various quantities, may be incorporated
through scavenging by Mn(IV) oxide. Laboratory evidence shows that some other bacteria on nodules
can reduce the Mn(IV) oxide to Mn2+ in the presence of an appropriate source of reducing power in the
form of organic carbon. The reduction of Mn(IV) oxide mobilizes manganese and trace metals such as
Cu, Ni, and Co in the nodules. These bacteria reduce Mn(IV) whether oxygen is present or not. They
could be harnessed in the bioleaching of nodules. The manganese(IV)-reducing bacteria and the bacte-
ria that neither oxidize nor reduce manganese are thought to keep the organic matter in the nodule
environment sufficiently low to favor bacterial manganese-oxidation, which in laboratory tests is
inhibited by excess organic carbon. The preying by the protozoa on the different kinds of bacteria is
thought to control the bacterial population on nodules and ensure continued growth of the manga-
nese-oxidizing bacteria and the nodules. To eliminate interference from seawater salts in the bioleach-
ing of Mn, Cu, Ni, and Co, it may be more advantageous to use terrestrial Mn(IV)-reducing bacteria
than the Mn(IV)-reducing bacteria resident on nodules. The terrestrial bacteria would allow bioleach-
ing in fresh water. Unlike the marine bacteria, most terrestrial bacteria studied to date reduce Mn(IV)
oxide only anaerobically. A general outline for an anaerobic bioleach process of marine nodules with
freshwater bacteria is discussed.
INTRODUCTION
The existence of manganese nodules (ferromanganese concretions) on the floor of the
world’s oceans was first reported by the Challenger Expedition, which was undertaken by the
British in 1873–1876 (Murray, 1891). They occur at various sites on the ocean floor and at
the sides of some sea mounts at water depths generally between 3,000 and 5,000 m. Of those
* Department of Biology, Rensselaer Polytechnic Institute, Troy, NY.
239
FIGURE 1 Marine manganese nodules from the Pacific Ocean. Note coin, a US penny (15-mm
diam), for size comparison.
that are associated with sediments, most reside on the surface, but some are buried. They are
not found at or near hot vents at mid-ocean ridge systems, which discharge dissolved manga-
nese, iron, and other solutes in hydrothermal solution. They vary in size from less than
10 mm to more than 250 mm, with an average size of 30 mm, and they may be oblate, dis-
coid, prolate, or nearly spherical (Figure 1) (Mero, 1962; Raab, 1972). Nodules smaller than
10 mm also occur, and these are called micronodules (e.g., Xavier, 1976).
The average nodule composition is: Mn = 24.2%, Fe = 14.0%, Cu = 0.53%, Ni = 0.99%, and
Co = 0.35% (all percentages by weight) (Mero, 1962). However, the metal content of individ-
ual nodules can diverge significantly from these values, being largely dependent on the envi-
ronment in which the nodules developed. The base-metal content of the nodules makes them
a potentially important industrial resource. The oxidation state of manganese in nodules is
mostly +4 and that of iron +3 (Murray et al., 1984; Piper et al., 1984; Pattan and Mudholkar,
1990). The manganese minerals that have been identified in nodules are todorokite, verna-
dite, and birnessite. The iron mineral has been described as oxyhydroxide (Piper et al., 1984).
The discovery of manganese nodules immediately raised the question of how they arose.
Although an abiotic origin has been favored by marine geochemists (e.g., Manheim, 1965;
Raab, 1972; Burns and Burns, 1975; Boudreau and Scott, 1978; Boulègue and Renard,
1980), biological involvement was suggested as early as the 1920s. Butkevitch (1928)
thought that Gallionella and Persius marinus were involved in nodule formation by precipitat-
ing manganese and iron. The basis for this claim was his finding of these organisms in the
Petchora and White Seas. Graham (1959) and Graham and Cooper (1959) proposed that
manganese nodules are formed as a result of the bacterial degradation of organic Mn(II)-
complexes followed by autoxidation of the freed Mn2+ ion in the alkaline seawater to Mn(IV)
oxide and subsequent accretion of the oxide to nuclei, such as arenaceous foraminiferal tests.
Some other suggestions of biological involvement in marine manganese nodule formation
include those of Kalinenko et al. (1962), Sorokin (1972), Monty (1973), Crerar and Barnes
(1974), Greenslate (1974a, 1974b), Dugolinsky et al. (1977); Schütt and Ottow (1977),
Ghiorse and Hirsch (1982), and Riemann (1983).
BIOGENESIS OF MARINE MANGANESE NODULES
Accretion of MnO2
Kalinenko et al. (1962) proposed that bacterial biofilms on dead plankton or mineral parti-
cles formed colloidal iron and manganese oxides, which were then agglutinated by slime
formed by the bacteria to form micronodules, from which macronocules could originate.
240
Sorokin (1972) observed that the activity of heterotrophic bacteria on nodules was more
intense than in the surrounding environment, but he found no special iron- or manganese-
oxidizing organisms among the bacteria. Like Graham (1959) and Graham and Cooper
(1959), Sorokin speculated that the role of the bacteria was to degrade the organic ligands of
iron and manganese complexes. But, according to Sorokin’s model, iron and manganese
were deposited on nodules by resident benthic fauna in their excretions. Sorokin suggested
that the freed Mn2+ then autoxidized to Mn(IV).
Monty (1973) concluded from electron microscopic examination of thin sections that nod-
ules are a form of stromatolite that result from manganese and iron oxide deposition in bac-
terial films in a seasonal growth pattern. Crerar and Barnes (1974) suggested, based on
thermodynamic and kinetic considerations, that nodules grow because of adsorption of Mn2+
and Fe3+ to their surface followed by autocatalysis and some microbially stimulated oxidation
of Mn2+ to Mn(IV) oxide. Greenslate (1974a, 1974b) concluded that nodule formation
involved manganese oxide deposition in planktonic microskeletal debris. This lead to the for-
mation of micronodules, which was then followed by the agglutination of such micronodules
to form macronodules through the agency of such benthic microorganisms as saccorhiza (see
also Dugolinsky et al., 1977).
Schütt and Ottow (1977) isolated a series of mesosphilic and psychrophilic bacterial
strains from uncontaminated manganese nodules, sediments, and bottom seawater that
readily oxidized MnCO3 in an agar medium to Mn(IV) oxide. Ghiorse and Hirsch (1982)
demonstrated the accumulation of iron and manganese oxides in the matrix of an acidic
(anionic) exopolymer of budding bacteria isolated from Baltic Sea manganese nodules, while
Riemann (1983) concluded that agglutinating rhizopodean protozoa on manganese nodules
release Mn(II) from Mn(IV) oxide associated with mineral particles ingested while feeding
on bacteria. According to Riemann (1983), this Mn(II) may then be reoxidized somehow and
the oxide accreted to the nodule surface. Small polychaete worms on nodule surfaces feeding
on the rhizopods may help to concentrate some of the inorganic components accumulated by
the rhizopods.
Accretion of Mn2+ to Nodules Followed by Its Oxidation

Several of the above proposals of biological involvement in marine manganese-nodule devel-
opment share some common features. All but the proposal of Crerar and Barnes (1974)
imply that manganese is accreted to nodules in the form of manganese oxide. Some other
studies suggest, however, that manganese is first sorbed by nodules in the form of Mn2+ and
only subsequently oxidized by certain bacteria to manganese(IV) oxide (Ehrlich, 1963, 1966,
1968). A reaction that describes Mn2+ sorption to nodule substance (H2MnO3) was formu-
lated by Ehrlich (1968) as
H2MnO3 + Mn2+ → MnMnO3 + 2H+ (EQ 1)
The H2MnO3 is meant to represent hydrated MnO2, recognizing its weak acidic properties.
Reaction (1), which involves proton displacement, is rapid and not directly catalyzed by
bacteria. It is readily demonstrable by immersing a nodule fragment in a 1-mM MnSO4
seawater-solution (Ehrlich et al., 1973).
In a subsequent reaction, the sorbed Mn2+ undergoes an oxidation, which was formulated
by Ehrlich (1968) as
MnMnO3 + 0.5 O2 + 2H2O → 2H2MnO3 (EQ 2)

In this reaction, theMn2+ bound to nodule substance is oxidized to Mn(IV), and it becomes
part of the nodule matrix (H2MnO3), ready to react with another Mn2+ according to Reaction
(1). Reaction (2) is catalyzed by specific kinds of bacteria found on nodules and adjacent
sediments and in seawater. Compared to Reaction (1), it is slow and, therefore, rate control-
ling in overall manganese uptake by nodules. This reaction sequence was first deduced from
241
experiments with unsterilized, crushed nodules in sterile seawater containing 0.1% peptone
with added Mn2+ (Ehrlich, 1963). In controls in which bacterial outgrowth was prevented by
the omission of peptone from the medium, manganese accretion to crushed nodule was sig-
nificantly slower than with bacteria. This was also noted in subsequent studies of Mn2+ accre-
tion to synthetic Mn-Fe oxide (e.g., Ehrlich, 1966, 1968) and to reagent-grade MnO2 with
and without intact cells or cell extracts of pure cultures of manganese-oxidizing bacteria
from nodules (e.g., Ehrlich, 1976; Arcuri and Ehrlich, 1980a). Indeed, Mn2+ uptake by MnO2
in the absence of the bacteria was usually unmeasurable.
Reaction (2) is catalyzed by specific kinds of bacteria, which contain a Mn2+-oxidizing
enzyme system (Ehrlich, 1968; Arcuri and Ehrlich, 1979, 1980a). These bacteria can oxidize
Mn2+ only if it is prebound to a sorbent such as ferromanganese in nodules, synthetic Mn-Fe
oxide, MnO2, or certain clays (Ehrlich, 1966, 1976, 1982). These findings suggest that bacte-
ria not only contribute to nodule growth but also to initiation of nodule formation around a
nucleus that binds Mn2+. Examinations of nodules since they were first discovered revealed
that they form around a nucleus. This nucleus may be microscopic in size (like a foramin-
iferal test) or large (like a piece of coral, a shark’s tooth, or often a piece of broken nodule)
(e.g., Burns and Brown, 1972; Raab, 1972).
The bacteria isolated by Ehrlich from nodules differ from the large majority of marine
manganese(II)-oxidizing bacteria, which he later isolated from water samples from around
some submarine hydrothermal vents (Ehrlich, 1983, 1985). The Mn2+-oxidizing enzyme sys-
tem in the bacteria from nodules, associated sediment, and water is constitutive and involves
an electron transport chain for conveying electrons from Mn2+ to O2 in their cell envelope.
The chain includes cytochromes and cytochrome oxidase (Arcuri and Ehrlich, 1979, 1980b).
Some biochemically useful energy is conserved by the organisms in this electron transfer
(Ehrlich, 1976). Manganese oxidation by these bacteria is inhibited by excess organic carbon
(peptone in experiments) (Ehrlich, 1966).
In contrast, all but one of the Mn(II)-oxidizing bacteria isolated by Ehrlich (1983, 1985)
from samples taken around deep-sea hydrothermal vents oxidize dissolved Mn2+ without
prior sorption. The manganese-oxidizing system of the isolates from around the vents
includes an electron transport chain in their cell envelope to which an energy conserving
mechanism is coupled, as in the manganese oxidizers from nodules (Ehrlich, 1983, 1985;
Ehrlich and Salerno, 1990). However, unlike the manganese-oxidizing system in bacteria
from nodules, the Mn2+-oxidizing system of the isolates from around the vents is inducible.
Deep-sea nodules occur at depths of 3,000 to 5,000 m, where the corresponding hydro-
static pressure ranges from 300 to 500 atm (30.3 to 50.5 MPa) and the temperature is 4°C or
lower. Therefore, it is important to know whether the manganese-oxidizing bacteria isolated
from samples taken at these depths and tested at atmospheric pressure and 15° or 25°C are
able to oxidize Mn2+ at in situ pressures and temperatures.
Ehrlich (1971, 1972) found that different isolates could indeed oxidize Mn2+ at elevated
hydrostatic pressure and lowered temperature, including those approximating in situ condi-
tions. However, the observed rates of oxidation and growth were significantly slower than at
1 atm of hydrostatic pressure and 15° or 25°C. One test strain showed some loss of cultura-
bility over 54 hours of incubation at 15°C and 408 atm (Ehrlich, 1972). The noticeable
decrease in growth and manganese oxidation rates under elevated pressure and low temper-
ature is consistent with reports of slow in situ growth of manganese nodules, estimated by
Bender et al. (1970) to be 0.2 to 1 mg cm–210–3y–1.
In evaluating metal tolerance of marine manganese-oxidizing bacteria from nodules, Yang
and Ehrlich (1976) found that, at 1 atm of hydrostatic pressure, growth of their test Strain
138 in seawater containing 0.1% glucose and 0.025% peptone was significantly less sensitive
to increasing concentrations of Mn2+ at 18°C than it was at 5°C. They ranked metal tolerance
of Strain 138 at 18°C and 1 atm (0.101 MPa) of hydrostatic pressure in 0.1% glucose—
0.005% peptone-seawater water medium as Mn2+> Co2+ >Ni2+ = Cu2+. Arcuri and Ehrlich
(1977) showed that growth of marine, manganese-oxidizing, nodule strain BIII39 in the
presence of Mn2+ (0 to 10 mg L–1) was unaffected by increased hydrostatic pressure from 1 to
242
340 atm (0.101 to 34.4 MPa) in seawater containing 0.1% glucose and 0.025% peptone.
However, growth was inhibited, and a slight loss in viability occurred at 408 atm (41.2 MPa).
When Cu2+ concentrations were 0, 0.1, and 1.0 mg L–1 in a similar test with strain BIII 39,
Arcuri and Ehrlich (1977) noted progressively greater growth inhibition at 1 atm (0.101
MPa) with increasing Cu concentration. A hydrostatic pressure of 272 atm (27.5 MPa) over-
came this inhibition. At a pressure of 340 atm (34.4 MPa), growth inhibition reappeared, and
at a pressure of 408 atm (41.2 MPa) no growth occurred; indeed a slight die-off was noted
with and without added copper.
Co2+ at 1 atm (0.101 MPa) was not inhibitory to strain BIII 39 at concentrations of 0.1 and
1 mg L–1, but it was very inhibitory at 10 mg L–1 (Arcuri and Ehrlich, 1977). Progressively
greater growth inhibition was noted when the pressure was raised to 272 and 408 atm (27.5
and 41.2 MPa). Growth stimulation or inhibition occurred at 340 atm (34.3 MPa), depending
on cobalt concentration. A fraction of the cell population lost viability at 408 atm (41.2 MPa).
Ni2+ at 1 atm (0.101 MPa) did not inhibit growth of strain BIII 39 at 0.1 and 1 mg L–1 con-
centrations but did inhibit it strongly at 10 mg L–1, like Co2+ (Arcuri and Ehrlich, 1977). At 272
and 340 atm (27.5 and 34.3 MPa), growth inhibition became noticeable as Ni2+ concentra-
tions increased. At 408 atm (41.2 MPa), neither growth nor loss in viability of cells was noted.
These and other experiments by Arcuri and Ehrlich (1977) showed that hydrostatic pressure
can modulate the toxic effect of metal ions on marine bacteria associated with nodules.
In the nodule-growth model under discussion in this section, the sorption of Mn2+ to ferro-
manganese by ion exchange (Reaction [1]) followed by bacterially catalyzed oxidation of the
sorbed Mn2+ (Reaction [2]) accounts for manganese incorporation into nodules. How iron,
which in the alkaline seawater must exist in ferric form, is incorporated into the nodule
matrix is still unclear. It may be by sorption from seawater as in the case of Cu2+, Ni2+, and
Co2+. Sorptive uptake of Cu2+, Ni2+, and Co2+ by nodule fragments was experimentally dem-
onstrated by Ehrlich et al. (1973). Because some bacteria have the ability to oxidize Co2+ to
Co3+ (Lee and Fisher, 1993; Lee and Tebo, 1994; Moffett and Ho, 1996; Souren, 1998; Tebo
1998; Moffett, 1998), it is possible that such bacteria may convert sorbed Co2+ to Co3+ in the
same way they oxidize Mn2+ to Mn(IV) oxide.
Some investigators have proposed that the source of the Mn and other trace metals for
nodules is the sediment in which these elements are thought to be mobilized under reducing
conditions (Raab, 1972). Others have proposed that the trace elements in nodules have a
variety of different sources (Cronan, 1978). Kasten et al. (1998) concluded that manganese
nodules from the Central Angola Basin in the south Atlantic Ocean derive Cu, Ni, Zn, and Mo
from seawater. Although nodules form at sites located far from hydrothermal discharge from
vents at mid-ocean ridges, some of the incorporated manganese may have originated in dis-
tant hydrothermal vent discharge carried away by ocean circulation. Average values for the
concentration of Mn, Cu, Ni, and Co in seawater have been given as 2, 3, 7, and 0.4 µg L–1,
respectively (Anonymous, 1971). Pore-water concentrations of Mn, Cu, and Ni, measured
just below the surface of clay sediment on the ocean floor, have been reported to range from
200 to 12,100 µg, 0.20 to 110 µg and 8 to 1,000 µg L–1, respectively (Raab, 1972). The con-
centrations of dissolved Mn2+ mostly are not high enough to be able to serve as an energy
source for manganese-oxidizing bacteria. However, sorption of such dissolved Mn2+ to nod-
ule surfaces results in its concentration to an extent required for its use as energy source by
manganese-oxidizing bacteria of the type described herein.
Assuming that the nodule radius increases at a rate of 7 to 1,200 mm 10–6y–1, as estimated
by Finney et al. (1984), nodules would be buried in accumulating sediment, which settles at
a rate of <1 to 3.5 mm 10–3y–1 in mid-ocean (Horn et al., 1972). Bioturbation has been
invoked to explain how most nodules remain at the sediment surface in the deep ocean and
also how nodules maintain their ovoid to spherical shape (Mero, 1962; Heath, 1979).
Nodules harbor not only significant numbers of manganese(II)-oxidizing bacteria, but also
manganese(IV)-reducing bacteria and bacteria that neither oxidize nor reduce manganese
(Ehrlich et al., 1972). The manganese(IV)-reducing bacteria have the ability to reduce
Mn(IV) in the presence of oxygen and can use organic carbon like glucose or acetate as an
243
electron donor (Trimble and Ehrlich, 1968; Ehrlich, 1988, 1993a). Because in situ nodules
exhibit a very gradual increase in mass over a long time span, Mn(II)-oxidizing activity must
be favored much of the time, whereas Mn(IV)-reducing activity must be minimal and spo-
radic, limited most likely by lack of an appropriate electron donor. Under laboratory condi-
tions, manganese-reducing bacteria from nodules are able to grow in appropriate media in
the absence of Mn(IV) oxide.
As already mentioned, nodules harbor benthic, shelter-building protozoans in addition to
bacteria on their surface. These protozoans are believed to feed on the bacteria by grazing
(Mullineaux, 1987). Their grazing activity helps in creating conditions that favor the contin-
ued multiplication of the bacteria to keep their population relatively constant, i.e., in a
steady state.
POTENTIAL FOR BIOLEACHING OF MARINE MANGANESE NODULES

Extraction by leaching Mn2+, Cu2+, Co2+, and Ni2+ from nodules can be accomplished chemi-
cally or microbiologically by the reduction of manganese(IV) oxides to soluble Mn2+ in aque-
ous solution. Depending on reaction conditions, the iron(III) oxide is not mobilized. An
example of chemical leaching is the treatment of nodules with SO2 and H2SO4 at elevated
temperature (e.g., Fuerstenau and Han 1977, p. 375). The SO2 serves as reductant. In
bioleaching of nodules, the transition metals can be mobilized by bacteria that reduce the
Mn(IV)-oxides with a suitable, externally supplied electron donor at ambient temperature
and pressure (Ehrlich et al., 1973). As previously mentioned, such bacteria are found among
the members of a typical manganese nodule flora. Appropriate external organic electron
donors include glucose and acetate.
Physiology of Mn(IV)-Reducing Bacteria from Marine Manganese Nodules

Insofar as tested, the Mn(IV)-reducing activity of bacteria from nodules and associated sedi-
ments isolated by Ehrlich and coworkers is inducible and about as great in the presence of
oxygen (air) as in its absence (Ehrlich, 1966; Trimble and Ehrlich, 1968, 1970; Ehrlich
1993a). Some of the ionic constituents of seawater appear to be required in manganese
reduction by the nodule bacteria so far tested (Ehrlich, 1966; Ghiorse and Ehrlich, 1974). An
overall reaction for MnO2-reduction by marine pseudomonad strain BIII 88 using acetate as
reductant can be written as (Ehrlich, 1993b)
CH3COO– + 4MnO2 + 7H+ → 2HCO3– + 4Mn2+ + 4H2O (EQ 3)

Some of the free energy released in this reaction is conserved (Ehrlich, 1993a). Ehrlich
(1993a, 1993c) proposed a biochemical mechanism by which this organism reduces MnO2.
This model features manganese bound in the cell envelope, which acts as an electron shuttle
between the surface of the Mn(IV) oxide to which the cell is attached, and the electron trans-
port system of the cell. The bound manganese is thought to undergo a reversible redox
change between the +2 and the +3 oxidation state in a dismutation reaction with the MnO2 at
the cell surface
{ Mn2+} + MnO2 + 4H+ → 2{Mn3+} + 2H2O (EQ 4)

The resultant bound {Mn3+}is then converted to Mn2+
by reducing power from an electron
donor used by the cell. The reducing power is transferred to {Mn3+} by the cell’s electron
transport system. Some of the Mn2+ that is thus formed remains bound to the cell envelope,
but a major portion enters the bulk phase. The overall process is not oxygen-sensitive.
The Mn(IV)-reducing marine bacteria isolated by Ehrlich and collaborators from manga-
nese nodules and associated sediments reduce Mn(IV) oxide at elevated hydrostatic pressure
and low temperature (Ehrlich, 1971, 1972). However, as with the Mn(II)-oxidizing bacteria
from nodules, the Mn(IV)-reducing activity of any of the test organisms decreases with
244
increasing pressure. Induction of the Mn(IV)-reducing system tested so far in only one strain,
BIII 41, was not inhibited at elevated hydrostatic pressure (Ehrlich, 1974).
The reduction of Mn(IV) oxide of manganese nodules results in the simultaneous solubili-
zation of base metals like Cu, Ni, and Co, but not Fe (Ehrlich et al. 1973). The concurrent sol-
ubilization of Cu, Ni, and Co is best explained if these metals are bound in the Mn(IV) oxide
matrix of the nodules. In the case of Co, a reduction of Co3+ to Co2+ may be involved, if the
dominant oxidation state of Co in nodules is +3. This reduction could be catalyzed by bacte-
ria, because, as mentioned above, bacteria are known to catalyze the oxidation of Co2+ to
Co3+. The free energy change for the reduction is favorable (∆Go′ = –213.1 kcal when acetate
is the electron donor).
Although some of the nodule bacteria also seem to have the ability to reduce Fe(III) in iron
minerals like goethite and limonite (iron oxyhydroxide) (De Castro and Ehrlich, 1970), the
lack of iron solubilization during microbial nodule leaching may be attributable to the unfa-
vorable bioenergetics of Fe(III) oxide reduction compared to that of Mn(IV) oxide. The ∆Go′
for the reduction of MnO2 by acetate, for instance, is –125.3 kcal whereas that for Fe(OH)3 is
–5.5 kcal (Ehrlich, 1993b).
The mobilization of Cu, Ni, and Co during bacterial reduction of Mn(IV) oxide by native
bacteria on nodules suggests that this process may be exploitable for bioleaching of marine
nodules. Although leaching of nodules by the native bacteria in preliminary laboratory tests
was slow (Trimble and Ehrlich, 1968; Ehrlich et al., 1973), it may be possible to improve the
rate significantly to make it competitive with available chemical hydrometallurgical pro-
cesses (e.g., Brooks et al., 1970; Han and Fuerstenau, 1975a, 1975b) on a cost basis.
Bioleaching of Marine Manganese Nodules

To exploit the metals contained in marine manganese nodules, they have to be collected on
the ocean floor and brought to the surface by special mining techniques (Mero, 1972;
Marchal, 1984). One approach to extracting the desired metals from the mined nodules is
bioleaching. This can be accomplished in two ways, one being enzymatic treatment, as
described above, and the other being microbial generation of lixiviants that solubilize manga-
nese and/or the other base metals. The lixiviants can be metabolic products that act as chemi-
cal reducing agents of Mn(IV) oxides and/or products that act as complexing agents of the
other base metals. The second approach has been tested with terrestrial manganese ores by
Gupta and Ehrlich (1989), using a strain of the fungus Penicillium. A thorough comparison of
effectiveness of enzymatic bioleaching with nonenzymatic bioleaching remains to be made,
but up to this point, enzymatic bioleaching looks more promising (Rusin and Ehrlich, 1995).
Although at first glance it may seem that enzymatic bioleaching of marine manganese
nodules on an industrial scale ought to utilize the Mn(IV)-reducing bacteria that are natu-
rally present on marine nodules, this may not be the best approach. Insofar as tested, sea-
water, even if diluted, seems to be needed by these bacteria for reduction of Mn(IV) oxide
and accompanying solubilization of base metals from nodules. For this reason, the use of
Mn(IV)-reducing bacteria from soil or freshwater, which do not need seawater salts, may be
preferable because bioleaching of nodules in freshwater eliminates possible interference of
seawater cations and/or anions in base-metal recovery from pregnant solution. Furthermore,
freshwater will be less corrosive than seawater toward steel reactor vessels. The extent that
nonferrous metals would be leached from nodules by freshwater or soil bacteria remains to
be determined.
A variety of bacteria from soil and freshwater that reduce Mn(IV) have been recently stud-
ied (Lovley and Phillips, 1988; Myers and Nealson, 1988; Lovley, 1991, 1992; Rusin and
Ehrlich, 1995). Some of these are strict anaerobes, others are facultative, but the majority
reduce Mn(IV) oxide only anaerobically (Troshanov, 1968; Di-Ruggiero and Gounot, 1990;
Lovley 1991). Most, if not all, anaerobic Mn(IV)-reducing bacteria also seem to have the abil-
ity to reduce iron(III). However, Mn(IV) reductase and Fe(III) reductase are distinct
enzymes, at least in Shewanella putrefaciens 200 (Burnes et al., 1998).
245
Detailed information about the physiology/biochemistry of Mn(IV) reduction by these
organisms is still incomplete (see Lovley and Phillips, 1988a; Myers and Nealson, 1988; Lov-
ley, 1991; Myers and Myers, 1992; 1993; Gaspard et al., 1998; Seeliger et al., 1998). Evi-
dence is emerging that the Mn(IV)- and Fe(III)-reductase in these gram-negative bacteria
probably resides at the cell surface, i.e., the outer membrane, and may be a form of c-type
cytochrome (Myers and Myers, 1992, 1993; Gaspard et al., 1998; Seeliger et al., 1998).
Much, if not all, of what is known about the biochemistry and physiology of anaerobic
iron(III) reduction seems to apply to anaerobic Mn(IV) reduction.
The possibility of leaching manganese from terrestrial ores by bacterial Mn(IV)-oxide
reduction on a commercial scale by soil/freshwater bacteria has been examined in bench-
scale experiments by several investigators (Perkins and Novielli, 1962; Babenko et al., 1983;
Holden and Madgwick, 1983; Kozub and Madgwick, 1983; Silverio, 1985; Ghosh and Imai,
1985a, 1985b; Noble et al., 1992; Rusin et al., 1992; Toro et al., 1993). Perkins and Novielli
(1962) attempted tank leaching of low-grade manganese ores (363.2 kg of Boulder City ore
with 2.0% Mn and 204.3 kg of Three Kids ore with 9.5% Mn, each at –65 mesh) from Nevada
in 7,560 L of molasses medium with a Bacillus culture. They agitated the reaction mixture
daily for 15 minutes and maintained the pH between 5 and 6 by additions of H2SO4. After 51
days of leaching the Boulder City ore, they found that 92.5% of the Mn in the ore had been
extracted, and, after 100 days of leaching the Three Kids ore, they found that 71.7% of the
Mn in the ore had been extracted. Mn was precipitated from each of the pregnant solutions
by raising the pH to 10.5.
Holden and Madgwick (1983) leached pyrolusite (42.9% Mn) and Kozub and Madgwick
(1983) leached similar material (44.7% Mn) from slime dam residue from the Groote
Eylandt Mine in Australia. Holden and Madgwick sterilized their material by autoclaving and
then leached it at 10% pulp density in a complex glucose medium (pH 6.0) inoculated with
soil. The best average rate of Mn2+ release they obtained was 55 mg L–1d–1. Kozub and
Madgwick (1983) leached their material microaerobically at 3% pulp density in a 10%
molasses medium, supplemented or not with a dried microalgal preparation. They depended
on active members of the bacterial flora present on their ore as leaching agents. In 35 days,
they extracted 39.3% of the Mn with microalgal supplement and 42.0% without it.
Babenko et al. (1983) leached sterilized ores from the Nikopol’ deposit in the former USSR
with Achromobacter delicatulus 183 in forcibly aerated glucose-ammonium sulfate medium
(pH 6.8 to 7) at a 2.5% pulp density. They extracted 94.1% of the Mn from the richest ore
(48.4% Mn) in 24 days.
Silverio (1985) compared leaching of Groote Eylandt ore in molasses and sucrose-yeast
extract media and found better manganese extraction in the latter. He attributed the better
extraction in sucrose medium to the probable presence of inhibitory substances in molasses.
Ghosh and Imai (1985) studied leaching of reagent-grade pyrolusite in mineral-salts
medium containing elemental S and ferric sulfate at pH 2.4 and inoculated with Thiobacil-
lus ferrooxidans AP–15. Best results were obtained when the pulp density of MnO2 was 4%,
the initial concentration of So was 1.0% and ferric sulfate 0.5%, using an inoculum of 5 ×
108 cells mL–1. Nearly 180 mg Mn2+ mL–1 were recovered in 15 days.
Rusin et al. (1992) worked with a manganiferous silver ore from Santa Cruz County, AZ.
The ore was leached in a proprietary medium in a continuously stirred tank reactor using
Bacillus MBX. In the presence of a proprietary organic collector, the organism solubilized
99.8% of the manganese and 86% of the silver anaerobically (under a nitrogen atmosphere)
in 14 days from sterilized ore at a pulp density of 15%. The ore had a particle size of 75 µm.
An additional 8.5% of the silver were recoverable from the ore residue by cyanidation.
Noble et al. (1992) leached a wad ore from the Three Kids deposit in Clark County, NV,
using the bacteria associated with the ore and any present in the nonsterile molasses
medium. They tested the ore in leach columns in 500 g and 9 kg quantities at a particle size
of –6.3 and –19 mm, respectively.
The leach medium was continually recycled. Although the authors mention no special provi-
sions for anaerobiosis, the interior of the columns probably turned anaerobic if the bacteria
246
consumed O2 faster than it could be replaced by diffusion from the air. Under the experimental
conditions, 63% of the Mn was mobilized from the –6.3-mm particles in 28 weeks, but only
20% of the Mn from the –19-mm particles was mobilized. They also studied “heap leaching”
with the ore, placing 35 kg of –19-mm particles in a 330-mm (I.D.) container. The ore bed had
a height of 400 mm and was fed a 3% molasses medium at a rate of 0.03 L min–1m–2. Under
these conditions, 10% of the manganese in the ore were mobilized in 57 weeks.
Toro et al. (1993) determined experimentally that manganese mobilization from an Ital-
ian pyrolusite-containing ore by bacteria was affected by ore particle size and by the nature
and amount of the organic matter in the medium. Oxygen seemed to be required for leaching
of this ore with the bacterial strains used.
A Possible Design for a Bioleach Process for Manganese Nodules

No tests of bioleaching marine nodules on a pilot or industrial scale seem to have been pub-
lished so far. Based on current knowledge, an ideal industrial leaching process for extracting
base metals from marine manganese nodules should be performed in reactors capable of sup-
porting the activity of an anaerobic culture. The reactors would be charged with a suspension
of ground nodules (–2 mm). Since nodules are very friable, grinding should not be an expen-
sive process. The optimal pulp density of the suspension would have to be experimentally
determined. The suspension medium should be a sterilized mineral salts solution containing
an experimentally determined, optimal concentration of a suitable electron donor that can
double as a sole source of carbon and that is preferably usable by the leaching organisms but
not by others.
Acetate is such a donor, and it could be produced from molasses or from CO2 + H2 in a spe-
cial reactor by homoacetogenic bacteria (Cheryan et al., 1977). The inoculum for leaching
should be a heavy suspension of a desired organism(s) prepared under axenic growth condi-
tions in a separate reactor. The inoculum should constitute about 10% (by volume) of the
nodule suspension to be able to outgrow any undesired organisms associated with the nod-
ules that might be able to grow in the leaching medium.
From an economic standpoint, it is impractical to sterilize the nodule material prior to
leaching. If the chosen primary leaching organism were a strict anaerobe like Geobacter met-
allireducens, it may be desirable to inoculate it together with a facultative organism like
Shewanella purtefaciens suspended in a dilute molasses solution. The facultative S. putre-
faciens could then scavenge any oxygen that was initially present in the medium, thereby cre-
ating conditions amenable to the growth of the strict anaerobe G. metallireducens. For
optimal leaching, it may be desirable to run the reactor in a semi-continuous mode.
Unlike Mn(IV)-reducing bacteria from nodules, freshwater bacteria like G. metallireducens
may also extract iron from the nodules if it produced a chemical lixiviant that mobilized the
ferric iron in nodules by complexation. Under such conditions any ferrous iron the organism
produced by reducing complexed ferric iron could, in turn, reduce Mn(IV) oxide chemically
(Lovley and Phillips, 1988b). Whether the quantity of iron that the bacteria mobilized would
have industrial value would have to be determined. No data are available yet on the extent of
mobilization of base metals such as Cu, Ni, and Co in nodules by these freshwater bacteria.
For recovery of base metals from pregnant solution, solvent extraction and electrowinning
should be considered. While Cu, Co, and Ni produced industrially from marine manganese
nodules should find a ready market, manganese produced from nodules, being twenty or
more times as abundant in nodules as Cu, Ni, and Co, may greatly exceed economic demand
unless new uses can be found for it.
CONCLUSION
Manganese nodules are a potentially important mineral reserve of base metals such as Mn,
Cu, Ni, and Co. Rapid metal extraction from nodules can be achieved by abiotic hydrometal-
lurgy, but it usually requires corrosive reagents and is energy intensive. Biohydrometallurgical
247
extraction of nodules is technically feasible. It does not require corrosive reagents, is not
energy intensive, and may offer an economic alternative.
REFERENCES
Anonymous, 1971, “Marine Chemistry. A report of the Marine Chemistry Panel of the Committee
on Oceanography,” National Academy of Sciences, Washington, DC, pp. 4–5
Arcuri, E.J., and Ehrlich, H.L., 1977, “Influence of hydrostatic pressure on the effects of the heavy
metal cations of manganese, copper, cobalt, and nickel on the growth of three deep-sea
bacterial isolates,” Appl. Environ. Microbiol., Vol. 33, pp. 282–288.
Arcuri, E.J., and Ehrlich, H.L., 1979, “Cytochrome involvement in Mn(II) oxidation by two marine
bacteria,” Appl. Environ. Microbiol., Vol. 37, pp. 916–923.
Arcuri, E.J., and Ehrlich, H.L., 1980a, “Electron transfer coupled to Mn(II) oxidation in two deep-
sea Pacific Ocean isolates,” In Biogeochemistry of Ancient and Modern Environments, Trudinger,
P.A., Walter, M.R., and Ralph, B.J., eds. Australian Academy of Sciences, Canberra, and
Springer-Verlag, Berlin, pp. 339–344.
Arcuri, E.J., and Ehrlich, H.L., 1980b, “A comparison of the cytochrome complements of seven
strains of marine manganese-oxidizing bacteria,” Z. allg. Mikrobiol., Vol. 20, pp. 583–586.
Babenko, Yu.S., Dolgikh, L.M., and Serebrayanaya, M.Z., 1983, “Characteristics of the bacterial
breakdown of primarily oxidized manganese ores from the Nikopol’ deposit,” Mikrobiologiya,
Vol. 52, pp. 851–856 (Engl. Transl. pp. 674–679).
Baglin, E.G., Noble, E.G., Lampshire, D.L., and Eisele, J.A., 1992, “Solubilization of manganese
from ores by heterotrophic micro-organisms,” Hydrometallurgy, Vol. 29, pp. 131–144.
Bender, M.L., Ku, T-L., and Broecker, W.S., 1970, “Accumulation rates of manganese in pelagic
sediments and nodules,” Earth Planet. Sci. Lett., Vol. 8, pp. 143–148.
Brooks P.T., Dean K.C., and Rosenbaum, J.B., 1970, “Experiments in processing marine nodules,”
Proc., Int. Miner. Process. Congr., 9th, pp. 328–333.
Boudreau, B.P., and Scott, M.R., 1978, “A model for the diffusion-controlled growth of deep-sea
manganese nodules,” Am. J. Sci., Vol. 278, pp. 903–929.
Boulègue, J., and Renard, D., 1980, “Catalyse bactérienne de l’oxydation du manganèse
manganeux dans les eaux. Conséquences géochimiques,” Comptes Rendus Acad. Sci., Paris, Vol.
290, serie D, pp. 1165–1168.
Burnes, B.S., Mulberry, M.J., and DiChristina, T.J., 1998, “Design and application of two rapid
screening techniques for isolation of Mn(IV) reduction-deficient mutants of Shewanella
putrefaciens,” Appl. Environ. Microbiol, 64, pp. 2716–2720.
Burns, R.G., and Burns, V.M., 1975, “Mechanism for nucleation and growth of manganese
nodules,” Nature, London, Vol. 255, pp. 130–131.
Burns, R.G., and Brown, B.A., 1972, “Nucleation and mineralogical controls on the composition of
manganese nodules,” In Ferromanganese Deposits on the Ocean Floor, Horn, D.R., ed., The
Office of the International Decade of Ocean Exploration, National Science Foundation,
Washington, DC, pp. 51–61.
Butkevitch, V.S., 1928, “The formation of marine iron-manganese deposits and the role of
microorganisms in the latter,” Wissenschaft. Meeresinst. Ber., Vol. 3, pp. 7–80.
Cheryan, M., Parekh, K.S., Shah, M., and Witjitra K., 1977, “Production of acetic acid by
Clostridium thermoaceticum,” Adv. Appl. Microbiol., Vol. 43, pp. 1–33.
Crerar, D.A., and Barnes, H.L., 1974, “Deposition of deep-sea manganese nodules,” Biochim.
Cosmochim. Acta, Vol. 38, pp. 279–300.
Cronan, D.S., 1978, “Manganese nodules: controversy upon controversy,” Endeavor (new series),
Vol. 2, pp. 80–84.
De Castro, A.F., and Ehrlich, H.L., 1970, “Reduction of iron oxide minerals by a marine Bacillus,”
Antonie v. Leeuwenhoek, Vol. 36, pp. 317–327.
Di-Ruggiero, J., and Gounot, A.M., 1990, “Microbial manganese reduction by bacterial strain
isolated from aquifer sediments,” Microb. Ecol., Vol. 20, pp. 53–63.
248
Dugolinsky, B.K., Margolis, S.V., and Dudley, W.C., 1977, “Biogenic influence on growth of
manganese nodules,” J. Sediment. Petrol., Vol. 47, pp. 428–445.
Ehrlich, H.L., 1963, “Bacteriology of manganese nodules. I. Bacterial action on manganese in
nodule enrichments,” Appl. Microbiol., Vol. 11, pp. 15–19.
Ehrlich, H.L., 1966, “Reactions with manganese by bacteria from marine ferromanganese
nodules,” Dev. Ind. Microbiol., Vol. 7, pp. 43–60.
Ehrlich, H.L., 1968, “Bacteriology of manganese nodules. II. Manganese oxidation by cell-free
extract from a manganese nodule bacterium,” Appl. Microbiol., Vol. 16, pp. 197–202.
Ehrlich, H.L., 1971, “Bacteriology of manganese nodules. V. Effect of hydrostatic pressure on
bacterial oxidation of MnII and reduction of MnO2 ,” Appl. Microbiol., Vol. 21, pp. 306–310.
Ehrlich, H.L., 1972, “Response of some activities of ferromanganese nodule bacteria to hydrostatic
pressure,” In Effect of the Ocean Environment on Microbial Activities, Colwell, R.R. and Morita,
R.Y., eds., University Park Press, Baltimore, MD, pp. 208–211.
Ehrlich, H.L., 1974, “Induction of MnO2-reductase activity under different hydrostatic pressures at
15°C,” Abstracts, Annual Meeting, American Society for Microbiology, Washington, DC, G–166.
Ehrlich, H.L., 1976, “Manganese as an energy source for bacteria,” In Environmental
Biogeochemistry, Nriagu, J.O., ed., Vol. 2, Ann Arbor Science Publishers, Ann Arbor, MI,
pp. 839–845.
Ehrlich, H.L., 1982, “Enhanced removal of Mn2+ from seawater by marine sediments and clay
minerals in the presence of bacteria,” Can. J. Microbiol., Vol. 28, pp. 1389–1395.
Ehrlich, H.L., 1983, “Manganese-oxidizing bacteria from a hydrothermally active area on the
Galapagos Rift,” Ecol. Bull., Stockholm, Vol. 35, pp. 357–366.
Ehrlich, H.L., 1985, “Mesophilic manganese-oxidizing bacteria from hydrothermal discharge areas
at 21° north on the East Pacific Rise,” In Planetary Ecology, D.E. Caldwell, J.A. Brierley, and
C.L. Brierley, eds., Van Nostrand Reinhold, New York, pp. 186–194.
Ehrlich, H.L., 1988, “Bioleaching of manganese by marine bacteria,” In Proceedings International
Biotechnology Symposium, Vol. II, 8th, Durand,G., Bobichon, L., and Florent, J., eds., July
17–22, 1988, Paris, France, pp. 1094–1105.
Ehrlich, H.L., 1993a, “Electron transfer from acetate to the surface of MnO2 particles by a marine
bacterium,” J. Industr. Microbiol., Vol. 12, pp. 121–128.
Ehrlich, H.L., 1993b, “Bacterial mineralization of organic carbon under anaerobic conditions,” In
Soil Biochemistry, Vol. 8: Bollag, M.-J., and Stotzky, G., eds., Marcel Dekker, New York,
pp. 219–247.
Ehrlich, H.L., 1993c, “A possible mechanism for the transfer of reducing power to insoluble
mineral oxide in bacterial respiration,” In Biohydrometallurgical Technologies, Torma, A.E.,
Apel, M.L., and Brierley, C.L., eds., The Minerals, Metals and Materials Society, Warrendale,
PA, pp. 415–422.
Ehrlich, H.L., and Salerno, J.C., 1990, “Energy coupling in Mn2+ oxidation by a marine
bacterium,” Arch. Microbiol., Vol. 154, pp. 12–17.
Ehrlich, H.L., Ghiorse, W.C., and Johnson, G.L., II., 1972, “Distribution of microbes in manganese
nodules from the Atlantic and Pacific Oceans,” Dev. Industr. Microbiol., Vol. 13, pp. 57–65.
Ehrlich, H.L., Yang, S.H., and Mainwaring, J.D., Jr., 1973, “Bacteriology of manganese nodules.
VI. Fate of copper, nickel, cobalt, and iron during bacterial and chemical reduction of the
manganese (IV),” Z. allg. Mikrobiol., Vol. 13, pp. 39–48.
Finney, B., Heath G.R., and Lyle, M., 1984, “Growth rates of manganese-rich nodules at MANOP
site H (Eastern North Pacific),” Geochim. Cosmochim. Acta, Vol. 48, pp. 911–919.
Fuerstenau D.W., and Han K.N., 1977. “Extractive metallurgy,” In Marine Manganese Deposits,
Glasby, G.P., ed., Elsevier Oceanographic Series, 15, Elsevier, Amsterdam, NL, pp. 357–390.
Gaspard, S., Vazquez, F., and Holliger, C., 1998, “Localization and solubilization of the iron(III)
reductase of Geobacter sulfurreducens,” Appl. Environ. Microbiol., Vol. 64, pp. 3188–3194.
Ghiorse, W.C., and Ehrlich, H.L., 1974, “Effects of seawater cations and temperature on manganese
dioxide-reductase activity in a marine Bacillus,” Appl. Microbiol., Vol. 28, pp. 785–792.
249
Ghiorse, W.C., and Hirsch, P., 1982, “Isolation and properties of ferromanganese-depositing
budding bacteria from Baltic Sea ferromanganese concretions,” Appl. Environ. Microbiol., Vol.
43, pp. 1464–1472.
Ghosh, J., and Imai, K. 1985, “Leaching of manganese dioxide by Thiobacillus ferrooxidans
growing on elemental sulfur,” J. Ferment. Technol., Vol. 63, pp. 259–262.
Graham, J.W., and Cooper, S., 1959, “Biological origin of manganese-rich deposits on the ocean
floor,” Nature, London, Vol. 183, pp. 1050–1051.
Greenslate, J., 1974a, “Manganese and biotic debris associations in some deep-sea sediments,”
Science, Vol. 186, pp. 529–531.
Greenslate, J., 1974b, “Microorganisms participate in the construction of manganese nodules,”
Nature, London, Vol. 249, pp. 181–183.
Gupta, A., and Ehrlich, H.L., 1989, “Selective and non-selective bioleaching of manganese from a
manganese-containing silver ore,” J. Biotechnol., Vol. 9, pp. 287–304.
Han, K.N., and Fuerstenau, D.W., 1975a, “Preferential acid leaching of nickel, copper, and
cobalt from ocean floor manganese nodules,” Inst. Min. Metall. Trans., Section C, Vol. 84,
pp. C105–C110.
Han, K.N., and Fuerstenau, D.W., 1975b, “Acid leaching of ocean manganese nodules at elevated
temperatures,” Int. J. Miner. Process., Vol. 2, pp. 163–171.
Heath, G.R., 1979, “Burial rates, growth rates, and size distributions of deep-sea manganese
nodules,” Science, Vol. 205, pp. 903–904.
Holden, P.J., and Madgwick, J.C., 1983, “Mixed culture bacterial leaching of manganese dioxide,”
Proc. Australas. Inst. Min. Metall., Vol. 286, pp. 61–63.
Horn, D.R., Horn, B.M., and Delach, M.N., 1972, “Distribution of ferromanganese deposits in the
world ocean,” In Ferromanganese Deposits on the Ocean Floor, Horn, D.R., ed., The Office of the
International Decade of Ocean Exploration, National Science Foundation, Washington, DC, pp.
9–17.
Kasten, S., Glasby, G.P., Schulz, H.D., Friedrich, G., and Andreev, S.I., 1998, “Rare earth elements
in manganese nodules from the South Atlantic Ocean as indicators of oceanic bottom water
flow,” Mar. Geol., Vol. 146, pp. 33–52.
Kalinenko, V.O., Belokopytova, O.V., and Nikolaeva, G.G., 1962, “Bacteriogenic formation of iron-
manganese concretions in the Indian Ocean,” Okeanologiya, Vol. 11, pp. 1050–1059.
Kozub, J.M., and Madgwick, J.C., 1983, “Microaerobic manganese dioxide leaching,” Australas.
Inst. Min. Metall., Vol. 288, pp. 51–54.
Lee, B.-G., and Fisher, N.S., 1993, “Microbially mediated cobalt oxidation in seawater revealed by
microtracer experiments,” Limnol. Oceanogr., Vol. 38, pp. 1593–1602.
Lee, B.-G., and Tebo, B.M., 1994, “Cobalt(II) oxidation by a marine manganese(II)-oxidizing
Bacillus sp. Strain SG–1,” Appl. Environ. Microbiol., Vol. 60, pp. 2949–2957.
Lovley, D.R., 1991, “Dissimilatory Fe(III) and Mn(IV) reduction,” Microbiol. Rev., Vol. 55,
pp. 259–287.
Lovley, D.R., 1992, “Microbial oxidation of organic matter coupled to the reduction of Fe(III) and
Mn(IV) oxides,” In Biomineralization Processes of Iron and Manganese, Modern and Ancient
Environments, Skinner, H.C.W., and Fitzpatrick, R.W., eds., Catena Supplement 21, Catena
Verlag, Cremlingen, Germany, pp. 101–114.
Lovley, D.R., and Phillips, E.J.P., 1988a, “Novel mode of microbial energy metabolism: Organic
carbon oxidation coupled to dissimilatory reduction of iron or manganese,” Appl. Environ.
Microbiol., Vol. 54, pp. 1472–1480.
Lovley, D.R., and Phillips, E.J.P., 1988b, “Manganese inhibition of microbial iron reduction in
anaerobic sediments,” Geomicrobiology J., Vol. 6, pp. 145–155.
Manheim, F.T., 1965, “Manganese-iron accumulations in the shallow marine environment,”
Symposium on Marine Geochemistry, Occ. Publ. No. 3–1965, Narragansett Marine Laboratory,
Univ. Rhode Island, pp. 217–276.
250
Marchal, P., 1984, “Préleveur libre autonome, véhicule experimental pour le ramassage des
nodules et la reconnaissance des fonds sous-marins,” In Deuxième Séminaire International Sur
Les Ressources Minérales Sous-Marines, Recherches et Exploitations Minères en Mer: Réalités et
Perspectives, Germinal, Orléans, France, pp 597–625.
Mero, J.L., 1962, “Ocean-floor manganese nodules,” Econ. Geol., Vol. 57, pp. 747–767.
Mero, J.L., 1972, “Potential economic value of ocean-floor manganese nodule deposits,” In
Ferromanganese Deposits on the Ocean Floor, Horn, D.R., ed., The Office for the International
Decade of Ocean Exploration, National Science Foundation, Washington, DC, pp. 191–203.
Moffett, J.W. 1998, “Reply to comment by A.W.M.G. Souren on ‘Oxidation of cobalt and
manganese in seawater via a common microbially catalyzed pathway’,” Geochim. Cosmochim.
Acta, Vol. 62, p. 359.
Moffett, J.W., and Ho, J., 1996, “Oxidation of cobalt and manganese in seawater via a common
microbially catalyzed pathway,” Geochim. Cosmochim. Acta, Vol. 60, pp. 3415–3424.
Monty, C., 1973, “Les nodules de manganèse sont des stromatolithes océaniques,” Comptes Rendu
Acad. Sci., Ser. D, Vol. 276, pp. 3285–3288.
Mullineaux, L.S., 1987, “Organisms living on manganese nodules and crusts: distribution and
abundance at three North Pacific sites,” Deep-Sea Research, Part A, Vol. 34, pp. 165–183.
Murray, J., 1891, “Report on the scientific results of the voyage of H.M.S. Challenger during the
years 1873–1876. Deep sea deposits,” H.M.S. Stationery Office, London.
Murray, J.W., Balistrieri, L.S., and Paul, B., 1984, “The oxidation state of manganese in marine
sediments and ferromanganese nodules,” Geochim. Cosmochim. Acta, Vol. 48, pp. 1237–1247.
Myers, C.R., and Myers, J.M., 1992, “Localization of cytochromes to the outer membrane of
anaerobically grown Shewanella putrefaciens MR–1,” J. Bacteriol., Vol. 174, pp. 3429–3438.
Myers, C.R., and Nealson, K.H., 1988, “Bacterial manganese reduction and growth with
manganese oxide as sole electron acceptor,” Science, Vol. 240, pp. 1319–1321.
Pattan, J.N., and Mudholkar, A.V., 1990, “The oxidation state of manganese in ferromanganese
nodules and deep-sea sediments from the Central Indian Ocean,” Chem. Geol., Vol. 85,
pp. 171–181.
Perkins, E.C., and Novielli, F., 1962, “Bacterial leaching of manganese ores,” US Bureau of Mines,
Report of Investigations 6102, 11 pp.
Piper, D.Z., Basler, J.R.F., and Bischoff, J.L., 1984, “Oxidation state of marine manganese
nodules,” Geochim. Cosmochim. Acta, Vol. 48, pp. 2347–2355.
Raab, W., 1972, “Physical and chemical features of Pacific deep sea manganese nodules and their
implications to the genesis of nodules,” In Ferromanganese deposits on the Ocean Floor, Horn,
D.R., ed., Office of the International Decade of Ocean Exploration, National Science
Foundation, Washington, DC, pp. 31–49.
Riemann, F., 1983, “Biological aspects of deep-sea manganese nodule formation,” Oceanology
Acta, Vol. 6, pp. 303–311.
Rusin, P., and Ehrlich, H.L., 1995, “Developments in microbial leaching—Mechanisms of
manganese solubilization,” Adv. Biochem. Eng./Biotechnol., Vol. 52, pp. 1–26.
Rusin, P., Sharp, J., Arnold, R., Sinclair, N.A., and Young, T.A., 1992, “Enhanced extraction of
silver and other metals from refractory oxide ores through bioreduction,” Mining Engineering,
Vol. 45, pp. 1467–1471.
Schütt, C., and Ottow, J.C.G., 1977, “Mesophilic and psychrophilic manganese-precipitating
bacteria in manganese nodules of the Pacific Ocean,” Z. Allg. Mikrobiol. Vol. 17, pp. 611–616.
Seeliger, S., Cord-Ruwisch, R., Schink, B., 1998, “A periplasmic and extracellular type cytochrome
of Geobacter sulfurreducens acts as a ferric iron reductase and as an electron carrier to other
acceptors or to partner bacteria,” J. Bacteriol., Vol. 180, pp. 3686–3691.
Silverio, C.M., 1985, “Microaerobic microbial manganese dioxide reduction,” NSTA Technology
Journal, Vol. 10, pp. 51–63
Sorem, R.K., and Foster, A.R., 1972, “Internal structure of manganese nodules and implications in
beneficiation,” In Ferromanganese deposits on the Ocean Floor, Horn, D.R., ed., Office of the
International Decade of Ocean Exploration, National Science Foundation, Washington, DC, pp.
167–181.
251
Sorokin, Yu.I., 1972, “Role of biological factors in the sedimentation of iron, manganese, and
cobalt in the formation of nodules,” Oceanology (Okeanologiya), Vol. 12, pp. 1–14.
Souren, A.W.M.G., 1998, “Comment on ‘Oxidation of cobalt and manganese in seawater via a
common microbially catalyzed pathway’ by J.W. Moffett and J. Ho,” Geochim. Cosmochim.
Acta, Vol. 62, pp. 351–355.
Tebo, B.M., 1998, “Comment on comment by A.W.M.G. Souren on ‘Oxidation of cobalt and
manganese in seawater via a common microbially catalyzed pathway’,” Geochim. Cosmochim.
Acta, 62, pp. 357–358.
Toro, L., Vegliò, F., Terreri, M., Ercole, C., and Lepidi, A., 1993, “Manganese bioleaching from
pyrolusite: Bacterial properties reliable for the process,” FEMS Microbiol. Rev., Vol. 11,
pp. 103–108.
Trimble, R.B., and Ehrlich, H.L., 1968, “Bacteriology of manganese nodules. III. Reduction of
MnO2 by two strains of nodule bacteria.,” Appl. Microbiol., Vol. 16, pp. 695–702.
Trimble, R.B., and Ehrlich, H.L., 1970, “Bacteriology of manganese nodules. IV. Induction of an
MnO2-reductase system in a marine Bacillus,” Appl. Microbiol., Vol. 19, pp. 966–972.
Troshanov, E.P., 1968, “Iron- and manganese-reducing microorganisms in ore-containing lakes of
the Karelian isthmus,” Mikrobiologiya, Vol. 37, pp. 934–940, (Engl. Transl. pp. 786–791).
Yang, S.H., and Ehrlich, H.L., 1976, “Effect of four heavy metals (Mn, Ni, Cu and Co) on some
bacteria from the deep sea,” In Proceedings of the Third International Biodegradation
Symposium, Sharpley, J.M., and Kaplan, A.M., eds., Applied Science Publishers, London,
pp. 867–874.
Xavier, A., 1976, “Sedimentologische und strukturelle Untersuchungen zur Genese der marinen
Eisen-Mangan-Akkretionen (‘Manganknollen’),” Senckenbergiana marit., Vol. 8, pp. 271–309.
252
Immobilization of Free Ionic Gold and
L-Asparagine-Complexed Ionic Gold by
Sporosarcina ureae: The Importance of
Organo-Gold Complexes in Gold Mobility
G. Southam,* W.S. Fyfe,† and T.J. Beveridge‡
ABSTRACT
Sporosarcina ureae was able to grow in the presence of up to 10-ppm L-asparagine complexed gold.
However, 10 ppm Au3+ was found to be highly toxic to Sporosarcina ureae, killing 109 bacteria/mL
within a few seconds. To compare the effects of these responses on gold mobility, an in vitro bacterial
biofilm model was developed to examine the interaction between Sporosarcina ureae and ionic gold
or L-asparagine gold complexes in gravity-fed columns. The immobilization of ionic gold was between
80% and 90% over a period of one week. But, as the available cellular reactive sites became saturated,
immobilization decreased to 14% within two weeks. In the L-asparagine-gold system, gold immobili-
zation in the cytoplasm occurred at a high rate (80% to 90%) throughout the experiment, with
limited toxicity toward Sporosarcina ureae. The bacterial immobilization and detoxification of
L-asparagine-complexed ionic gold was associated with a low molecular weight, i.e, <5,000 dalton
(atomic mass unit), intracellular protein fraction. This peptide-complexed gold would allow for the
continued biogeochemical cycling of gold under ambient (<100°C) surface or near-surface conditions.
INTRODUCTION
Eluvial and alluvial placer gold deposits contain varying amounts of single octahedral gold
crystals (secondary gold), crystal aggregates, and gold nuggets that are typically coated with
secondary gold (Giusti, 1986; Groen et al., 1990; Santosh et al., 1992; Craw and Youngson,
1993; Youngson and Craw, 1993). When a detrital gold specimen is examined in thin sec-
tion, it is typically comprised of a gold and silver core surrounded by secondary gold-bearing
material that is enriched in gold relative to the core (Giusti, 1986; Groen et al., 1990; Minter
et al., 1993). It logically follows that placer gold nuggets grow by the nucleation and precipi-
tation of octahedral gold on their outer surfaces.
* Department of Biological Sciences, Northern Arizona University, Flagstaff, AZ.

† Department of Geology, University of Western Ontario, London, Ontario, Canada.
‡ Department of Microbiology, University of Guelph, Guelph, Ontario, Canada.
253
Biological mechanisms for placer gold formation must consider the method of gold solubi-
lization and transport in the lithosphere or hydrosphere and must consider the method of
gold immobilization. Leaching of native gold by amino acid-producing bacteria (Lyalikova
and Mokeicheva, 1969; Korobushkina et al., 1974, 1976; Mineyev, 1976) has been demon-
strated. It can occur under normal environmental conditions, and it represents a novel mech-
anism for gold transport in the natural environment. In a Brazilian placer environment
examined by Freise (1931), new secondary gold was recovered several years after placer
mining, indicating that it had either formed in situ or had been transported into this system
from another source. Because bacterial organics have the ability to transform colloidal gold
into octahedral gold (Southam and Beveridge, 1994, 1996) and because of the abundant
plant organics described by Freise (1931) are presumably undergoing microbial degradation,
this “new” secondary gold likely formed in situ within the Brazilian placer system.
Sporosarcina ureae is a relatively rare soil bacterium that does not utilize carbohydrates
for growth (MacDonald and MacDonald, 1962), but it can be cultured exclusively on a chem-
ically defined growth medium containing L-asparagine (Asn), an amino acid (Goldman and
Wilson, 1977). Of particular importance, Asn was identified by Korobushkina et al. (1976) as
one of the most efficient amino acid compounds responsible for the solubilization and trans-
port of gold in the natural environment (Lyalikova and Mokeicheva, 1969). The goals of this
study were to examine the effect of Asn-complexed gold on the growth of S. ureae and to dis-
cuss the implications of the biogeochemical cycling of gold in geological systems.
EXPERIMENTAL PROCEDURES
Asn Complexation of Ionic Gold (Au3+): Asn-Au

Ionic gold (50, 100, and 250 ppm) from a 5,000 ppm(aq) solution of HAuCl4·3H2O was added
to an Asn solution (5 g/L, pH 9), and the gold complexation was then measured as a
decrease in absorbance at 395 nm (A395 nm), which is an absorbance peak for Au3+ (data not
shown). The percent stability of each reaction system was determined using atomic absorp-
tion spectroscopy. The stabilites were calculated by comparing the concentration of gold
found in the supernatant of a water control with that of the Asn test system after centrifuga-
tion (5 minutes at 14,000 G) to pellet any unstable precipitated gold. The effect of pH on
Asn-Au stability was also determined by adjusting the pH down to 4.
Metal Analysis
Soluble gold in the aqueous samples was analyzed using graphite furnace atomic adsorption
spectroscopy in a Perkin-Elmer, Model 2380, spectrophotometer. A Fisher Scientific gold ref-
erence standard was used. All quintuplicate analyses had standard deviations not exceeding
10%.
Growth of S. ureae in the Presence of Asn-Au

Sporosarcina ureae was routinely grown as shaker-flask (150 ppm) cultures in a liquid
medium containing Asn as the sole source of carbon, nitrogen and, energy (Goldman and
Wilson, 1976). The growth medium contained: 5.0 g/L L-asparagine, 3.4 g/L KCl, 0.25 g/L
K2HPO4, 0.25 g/L MgSO4·7H2O, 200 mg/L (NH4)2SO4, 5 mg/L L-cysteine, 2.5 mg/L
FeSO4·7H2O, 1 mg/L d-biotin, and 0.25 mg/L MnCl2·4H2O. The pH of the growth medium
was 8.7. For the Asn-complexed gold experiment, ionic gold (1, 5, 10, and 50 ppm) was
added to the Asn growth medium and allowed to complex overnight prior to inoculating the
bacterial culture. Bacterial growth was measured by optical density through a 1-cm light
path at 600 nm (OD600 nm).
254
SDS-PAGE Analysis of Asn-Au Grown S. ureae
Aliquots of cells were examined using 15% SDS-PAGE (sodium dodecyl sulfate-polyacryla-
mide gel electrophoresis) (Laemmli, 1970) to determine the soluble protein fraction respon-
sible for gold complexation. Cells (100 mg) were treated with lysozyme (2 mg), DNase
(0.5 mg), and RNase (0.5 mg), for one hour at room temperature to break open the bacteria
and then boiled in dissolution buffer for 5 minutes (Laemmli, 1970). This dissolution buffer
has the ability to release the soluble constituents from cell cytoplasm and any noncovalently
bound organic material from the cell envelope. These soluble constituents were then sepa-
rated from the particulate cell debris, which was removed by centrifugation at 14,000 G. The
SDS confers a constant net-negative charge per unit weight, allowing separation of proteins
in an electric field based on molecular size. After SDS-PAGE, the gels were cut into 1-mm
horizontal slices, and the soluble constituents were allowed to diffuse out of the gel for 24
hours at 4°C prior to protein and metal analysis. The molecular weights assigned to these
fractions were based on a standard curve produced by SDS-PAGE analysis of the following
low-range weight protein standards (BioRad®): phosphorylase B (97,400 Da), bovine serum
albumin (66,200 Da), ovalbumin (45,000 Da), carbonic anhydrase (31,000 Da), trypsin
inhibitor (21,500 Da), and lysozyme (14,400 Da).
Protein Assay
Soluble protein in the aqueous samples were analyzed in duplicate using the Bradford
(1976) protein assay with a lysozyme standard.
Biofilm Experiment
In all, five glass columns (10-mm diam, 200-mm high) containing 39.1 g of acid-washed ster-
ile silica (200 mesh; depth of silica = 150 mm; Fluka, Switzerland) were prepared as a sup-
porting matrix for biofilm growth. To initiate the biofilm experiment, the gold controls
(Columns 1 and 2) received 20 mL sterile dH2O, and the bacterial systems (Columns 3, 4,
and 5) received 20 mL of an early log phase S. ureae culture grown on the 5 g/L Asn medium
(OD600 nm = 2.775). After 15 minutes, these columns were allowed to drain by gravity, releas-
ing 10 mL of fluid possessing an average OD600 nm of 0.155, demonstrating that the S. ureae
cells were immobilized on the silica matrix and that the columns contained a 10-mL void vol-
ume. This 10-mL void volume was used for subsequent gold and media gravity fed pulses
(see Table 1), which displaced the “spent” fluid within the column for chemical analysis. The
biofilm system received a dilute (compared to the shaker-flask experiment) concentration of
Asn (0.5 g/L containing 1 ppm gold; pH 9) to promote biofilm formation.
TABLE 1 A general flowsheet for the biofilm experiment

Day Column(s) Pulse
0 1,2 20 mL sterile dH2O
3,4,5 20 mL S. ureae culture
1 1
2,5 10 mL Asn-Au solution*
3,4 10 mL Asn solution†
2 1,4 10 mL gold solution‡
2,3,5
3 Same as day 1
4 Same as day 2
* 0.5 g/L Asn, pH 9 containing 1 ppm Au.

† 0.5 g/L Asn, pH 9.
‡ 1 ppm of Au3+, pH 6.82.
255
FIGURE 1 Asn (5 g/L; pH 9) complexation of 50, 100, and 250 ppm Au 3+ measured at
A395 nm. Note the decrease in adsorbance over time as gold complexation occurs.
Gold retention in the columns was calculated by determining the absolute amount of gold
in the column washes and subtracting this value from the stock gold solution added to the
columns. After two weeks, the silica matrix in each of the columns was fractionated at 30 mm
intervals, sonicated as a 10–1 (wt./vol.) dilution in physiological saline (0.85 g/L NaCl), and
then plated from limiting dilutions onto on a solidified L-asparagine growth medium (see
recipe above) containing 15 g/L Bacto® Agar to determine the viable cell counts as colony
forming units (cfu) per gram of silica.

The Asn solution (5 g/L, pH 9) possessed the ability to complex up to 100 ppm of Au3+ in a
stable manner (Figure 1). However, in the 250-ppm system, only 45% of the gold remained
soluble (complexed), while the remaining gold precipitated as aggregates of fine-grained col-
loids (data not shown). Adjusting the pH of the stable Asn-gold solutions to pH 4 also
resulted in the precipitation of colloidal gold (data not shown).
Sporosarcina ureae was able to grow in the presence of up to 10 ppm Asn-Au (5 g/L Asn,
pH 9) (Figure 2). In contrast, 10 ppm Au3+ was highly toxic to Sporosarcina ureae, killing
109 bacteria/mL in seconds (data not shown). Bacteria from the 10-ppm Asn-Au system,
examined using transmission electron microscopy (Philips EM300), did not contain any gold
precipitates (data not shown). Examination of the 10-ppm Asn-Au grown S. ureae cells by
SDS-PAGE demonstrated that 93% of the gold was present in the <5,000 dalton, low molecu-
lar weight soluble peptide fraction (Figure 3).
The immobilization of Au3+ by S. ureae grown as a biofilm (Figure 4) initially occurred
with high efficiency (80% to 90%) for the first week. The efficiency then decreased (down to
14%), presumably as the cellular reactive sites became saturated with gold. The limited
availability of reactive sites is inferred because S. ureae was killed by the ionic gold. In the
Asn-Au biofilm system (Figure 4) gold immobilization occurred at a high rate (80% to 90%)
throughout the experiment (two weeks). While gold immobilization did demonstrate some
toxicity toward the bacterial cells when compared to the biofilm control, a viable population
256
FIGURE 2 Growth (OD600 nm) of S. ureae in the presence of L-asparagine complexed gold.
Note the low level of metal toxicity in the 10-ppm gold system compared to the lower gold
concentrations.
FIGURE 3 A comparison of gold complexation by S. ureae proteins ranging from ∼5 to 121 kDa.
Note the large gold peak associated with the low molecular weight peptide fraction.
257
FIGURE 4 The immobilization of Au3+ and L-asparagine complexed gold by the silica matrix
(chemical controls) and by the S. ureae biofilm (biogeochemical) receiving pulses of Asn
(0.5 g/L) and 1 ppm Au3+ on alternate days or 1 ppm pulses of Asn-gold every second day
(Table 1).
FIGURE 5 Viable cell counts of S. ureae recovered from the biofilm test systems after
sonication to remove the bacteria from the mineral surfaces. The ionic gold was extremely
toxic to S. ureae compared to the L-asparagine complexed gold.
258
was maintained in the Asn-Au system (Figure 5). Because the gold was distributed among a
wide range of proteins, bacterial biofilms may enhance the deposition of the Asn-complexed
gold by destroying the ligand after uptake of the Asn-gold.
Labile organic material, e.g., low molecular weight peptides, may serve as complexing
agents for the transport of gold in surface and near-surface natural systems. The role of amino
acids, specifically Asn, in the solubilization and complexation of gold (Korobushkina et al.,
1976) is an important link toward achieving a better understanding of gold transport pro-
cesses. Gold that is complexed by Asn can only be immobilized by bacteria that have the
capacity to utilize Asn as a source of nutrients (e.g., S. ureae; Goldman and Wilson, 1977).
Gold complexation by nontoxic bacterial organics would promote the biogeochemical cycling
of gold. This is because the bacterial cells that are responsible for gold immobilization only
have to deal with the toxicity of the gold and not a recalcitrant complexing agent (e.g., humic
and fulvic acids) (Baker, 1978; Fedoseyeva et al., 1986; Bergeron and Harrison, 1989; Bowell
et al., 1993). In the biofilm model presented here pulsed with Asn-Au, gold immobilization
occurred as protein complexes (Figure 3) in a highly efficient manner (Figure 4) by a popula-
tion of bacteria that were not killed by the process (Figure 5), suggesting that this process
could go on indefinitely. However, once the amount of gold reaches a critical concentration,
i.e., ∼100 µg Au/mg bacteria (Southam and Beveridge, 1994, 1996), octahedral gold forma-
tion can occur. The contribution of bacterial organic material to the biogeochemical cycling of
gold is supported by the detection of sulfur and phosphorus signals in octahedral gold formed
in vitro (Southam and Beveridge, 1996) and in Yukon placer gold (Southam, 1988).
In natural systems, the mechanism of gold transport and enrichment via octahedral gold
formation is not completely understood. Bacterial populations play an active role in mineral
weathering in subsurface and near-surface environments, e.g., supergene enrichment of Cu
at Morenci, AZ (Southam, 1999). Bacteria and byproducts of their metabolism also have the
ability to solubilize/transport gold (Lyalikova and Mokeicheva, 1969; Korobushkina et al.,
1974, 1976; Mineyev, 1976) and catalyze the formation of octahedral gold (Southam and
Beveridge, 1994, 1996).
Therefore, biogeochemical gold transformations may be responsible for the origin of
quartz-free gold nuggets in supergene-enrichment environments (Wilson, 1984), where gold
has been dissolved from its host material and redeposited in a pure form and where subsur-
face bacteria reside. Also, in groundwaters that possess very low (ppb) concentrations of sol-
uble gold (Bergeron and Harrison 1989), organic material and bacteria likely play an
important role in gold stabilization and transport and, therefore, should be considered in the
direct hydrogeochemical prospecting for gold (Hamilton et al., 1983). Low molecular
weight, labile organo-gold compounds have been generally overlooked in gold exploration
and in the supergene enrichment of gold. However, the interaction of gold with simple
organics may be an important link to a better understanding of gold transformations in
under ambient (<100°C), surface, or near-surface conditions.
ACKNOWLEDGMENTS
G.S. was supported by an Ontario Geoscience Research grant to W.S.F and T.J.B. The elec-
tron microscopy was performed at the NSERC Regional STEM facility in the Department of
Microbiology, University of Guelph, Guelph, Ontario, which is partially supported by an
Infrastructure NSERC (Canada) grant to T.J.B. We thank the four anonymous reviewers for
their constructive comments that have improved the presentation of this manuscript.
REFERENCES
Baker, W.E., 1978, “The role of humic acid in the transport of gold,” Geochimica et Cosmochimica
Acta, Vol. 42, pp. 645–649.
259
Bergeron, M., and Harrison, Y., 1989, “Le transport chimique de l’or dans les environnements de
surface: Formation d’un colloïde et complexation organique,” Canadian Journal of Earth
Science, Vol. 26, pp. 2327–2332.
Bowell, R.J., Gize, A.P., and Foster, R.P., 1993, “The role of fulvic acid in the supergene migration of
gold in tropical rain forest soils,” Geochimica et Cosmochimica Acta, Vol. 57, pp. 4179–4190.
Bradford, M.M., 1976, “A rapid and sensitive method for the quantitation of microgram quantities
of protein utilizing the principle of protein-dye binding,” Analytical Biochemistry, Vol. 72,
pp. 248–254.
Craw, D., and Youngson, J.H., 1993, “Eluvial gold placer formation on actively rising mountain
ranges, Central Otago, New Zealand,” Sedimentary Geology, Vol. 85, pp. 623–635.
Fedoseyeva, V.I., Fedoseyev, N.F., and Zvonareva, G.V., 1986, “Interaction of some gold complexes
with humic and fulvic acids,” Geochemistry International, Vol. 23, pp. 106–110.
Freise, F.W., 1931, “The transportation of gold by organic underground solutions,” Economic
Geology, Vol. 26, pp. 421–431.
Giusti, L., 1986, “The morphology, mineralogy, and behavior of ‘fine-grained’ gold from placer
deposits of Alberta: Sampling and implications for mineral exploration,” Canadian Journal of
Earth Science, Vol. 23, pp. 1662–1672.
Goldman, M., and Wilson, D.A., 1977, “Growth of Sporosarcina ureae in defined media,” Federation
of European Microbiological Societies Microbiology Letters, Vol. 2, pp. 113–115.
Groen, J.C., Craig, J.R., and Rimstidt, J.D., 1990, “Gold-rich rim formation on electrum grains in
placers,” The Canadian Mineralogist, Vol. 28, pp. 207–228.
Hamilton, T.W., Ellis, J., Florence, T.M., and Fardy, J.J., 1983, “Analysis of gold in surface waters
from Australian goldfields: An investigation into direct hydrogeochemical prospecting for
gold,” Economic Geology, Vol. 78, pp. 1335–1341.
Korobushkina, E.D., Chernyak, A.S., and Mineyev, G.G., 1974, “Dissolution of gold by
microorganisms and products of their metabolism,” Mikrobiologiya, Vol. 43, pp. 9–54.
Korobushkina, E.D., Mineyev, G.G., and Praded, G.P., 1976, “Mechanism of the microbiological
process of dissolution of gold,” Mikrobiologiya, Vol. 45, pp. 535–538.
Laemmli, U.K., 1970, “Cleavage of structural proteins during the assembly of the head of
bacteriophage T4,” Nature, Vol. 127, pp. 680–685.
Lyalikova, N.N., and Mokeicheva, L.Y., 1969, “The role of bacteria in gold migration in deposits,”
Mikrobiologiya, Vol. 38, pp. 805–810.
MacDonald, R.E., and MacDonald, S.W., 1962, “The physiology and natural relationships of the
motile, sporeforming Sarcinae,” Canadian Journal of Microbiology, Vol. 8, pp. 795–808.
Mineyev, G.G., 1976, “Organisms in the gold migration-accumulation cycle,” Geochemistry
International, Vol. 13, pp. 577–582.
Minter, M.G., Knight, J., and Frimmel, H.E., 1993, “Morphology of Witwatersrand gold grains from
the basal reef: evidence for their detrital origin,” Economic Geology, Vol. 88, pp. 237–248.
Santosh, M., Philip, R., Jacob, M.K., and Omana, P.K., 1992, “Highly pure placer gold formation in
the Nilambur valley, Wynad gold field, southern India,” Mineralum Deposita, Vol. 27,
pp. 336–339.
Southam, G., 1998, “Quantification of sulfur and phosphorus within secondary gold rims on
Yukon placer gold,” Geology, Vol. 26, pp. 339–342.
Southam, G., 1999, Unpublished observations.
Southam G., and Beveridge, T.J., 1994, “The in vitro formation of placer gold by bacteria,”
Geochimica et Cosmochimica Acta, Vol. 58, pp. 4527–4530.
Southam, G., and Beveridge, T.J., 1996, “The occurrence of bacterially derived sulfur and
phosphorus within pseudocrystalline and crystalline octahedral gold formed in vitro,”
Geochimica et Cosmochimica Acta, Vol. 60, pp. 4369–4376.
Wilson, A.F., 1984, “Origin of quartz-free gold nuggets and supergene gold found in laterites and
soils—A review and some new observations,” Australian J. of Earth Science, Vol. 31, pp. 303–316.
Youngson, J.H., and Craw, D. 1993, “Gold nugget growth during tectonically induced
sedimentary recycling, Otago, New Zealand,” Sedimentary Geology, Vol. 84, pp. 71.
260
Index
A Paenibacillus polymyxa in calcium removal

Acid rock drainage from bauxite ore 13–25
manganese removal by bioremediation sulfate-reducing bacteria 177–182
213–217 T. ferrooxidans in depression of pyrite
passive bioremediation by sulfate-reducing flotation 3–11
bacteria in sawdust-supported wetland Bauxite
systems 189–205 calcium removal using Paenibacillus
passive-biological control and treatment polymyxa 13–25
methods 169–188 Billiton Process Research 121
Aluminum Bioaccumulation 181–182
corundum-adapted Bacillus polymyxa in Biobeneficiation
separation of hematite and alumina corundum-adapted Bacillus polymyxa in
55–65 separation of hematite and alumina
microbial leaching from mineral raw 55–65
materials 43–46 microbial flotation of pyritic sulfur in
Amantaytau Goldfields Project (Uzbekistan) semicontinuous system 27–36
BIOX® plant 133 of miscellaneous mineral raw materials
Anoxic limestone drains 176–177 37–54
ARD. See Acid rock drainage Paenibacillus polymyxa in calcium removal
Ashanti (Ghana) BIOX® plant 131–132 from bauxite ore 13–25
Paenibacillus polymyxa in flotation of pyrite
B and chalcopyrite 67–81
Bacillus polymyxa. See also Paenibacillus T. ferrooxidans in depression of pyrite
polymyxa flotation 3–11
Bacillus polymyxa Bioenvironmental control 224–226
grown in the presence of corundum for Bioflotation 223–224
separation of hematite and alumina Bioleaching 222–223
55–65 advances in use of BIOX® Process for
Bacteria treatment of refractory gold ores
in biodegradation of metal cyanide 121–134
complexes formed during gold extraction electrochemical behavior of galena in
137–151 presence and absence of mesophilic
bioleaching of marine manganese nodules microorganisms 229–238
239–252 of marine manganese nodules 239–252
in environmental control 224–226 pilot testing to recover copper from
immobilization of ionic gold by chalcopyrite and nickel/cobalt from
Sporosarcina ureae 253–260 pentlandite/pyrrhotite 85–100
mesophilic microorganisms in bioleaching of uranium ores 101–119
of galena 229–238 Biological Oxidation. See BIOX® Process
261
Biomineralization desulfurization by semicontinuous
bioleaching of marine manganese nodules microbiological flotation system using
239–252 T. ferrooxidans 27–36
electrochemical behavior of galena during Cobalt
bioleaching with mesophilic bioleaching in recovery from pentlandite/
microorganisms 229–238 pyrrhotite 85–100
immobilization of ionic gold by Sporosarcina Composting in soil remediation 209–210
ureae 253–260 Copper
microbes as surface modifying reagents in bioleaching in recovery from chalcopyrite
extraction processes 221–227 85–100
Biopiles 208–209 Corundum
Bioreactive walls 181 in adaptation of Bacillus polymyxa for
Bioreactors 179–181 separation of hematite and alumina
fluidized-bed mode 13–25 55–65
total-recycle-slurry mode 13–25 Cu(I)–cyanide complex
Bioremediation biodegration of 137–151
ARD remediation by sulfate-reducing bacteria Cyanide
in sawdust-supported wetland systems biodegradation of metal cyanide complexes
189–205 137–151
biodegradation of metal cyanide complexes by bioremediation of 212–213
bacteria 137–151
biopiles 208–209 D
bioventing for hydrocarbon solvents 211 Diesel hydrocarbons 211–212
case studies of miscellaneous technologies
207–218 F
composting in soil remediation 209–210 Fairview (South Africa) BIOX® plant 127–128
of cyanides 212–213 Flotation
of diesel hydrocarbons 211–212 depression of pyrite flotation by yeast and
in situ 211 bacteria 3–11
intrinsic 211–212 microbial flotation of pyritic sulfur in
land farming 208–209 semicontinuous system 27–36
manganese removal from acidic mine drainage Fosterville (Australia) BIOX® plant 133–134
213–217
passive-biological control and treatment G
methods for acid rock drainage 169–188 Galena
removal of hazardous air pollutant precursors bioleaching with mesophilic microorganisms
from coal 153–167 229–238
treatment of oily wastewaters 209 GENCOR Process Research 121
Biosorption 181–182 Gold
Bioventing for hydrocarbon solvents 211 advances in use of BIOX® Process for
BIOX® Process treatment of refractory ores 121–134
advances in use for treatment of refractory biodegradation of metal cyanide complexes
gold ores 121–134 formed during extraction 137–151
immobilization of ionic gold by Sporosarcina
C ureae 253–260
Calcium
removal from bauxite ore using Paenibacillus H
polymyxa 13–25 HAP precursors. See Hazardous air pollutant
Chalcopyrite precursors
flotation tests with Paenibacillus polymyxa Harbour Lights (Australia) BIOX® plant 130
67–81 Hazardous air pollutant precursors
Clean Air Act (United States) 153 biochemical removal from coal 153–167
Coal Hematite
biochemical removal of hazardous air corundum-adapted Bacillus polymyxa in
pollutant precursors 153–167 separation of hematite and alumina 55–65
Hydrocarbon solvents 211
262
I S
Idaho National Engineering and Environmental Saccharomyces cerevisiae
Laboratory 153–167 in depression of pyrite flotation 3–11
Iron Sao Bento (Brazil) BIOX® plant 129–130
microbial removal from mineral raw materials Shale
38–43 bacterial leaching 49–50
Silicon
K microbial removal from low-grade bauxites
Kaolin 46–48
microbial beneficiation in improvement of Slurry column reactors
ceramic properties 48 for biochemical removal of hazardous air
pollutant precursors from coal 153–167
L Sporosarcina ureae
Land farming 208–209 immobilization of ionic gold by 253–260
Leptospirilum ferrooxidans SRB. See Sulfate-reducing bacteria
in bioleaching of uranium ores 101–119 Sulfate-reducing bacteria
in control and treatment of acid rock drainage
M 177–182
Manganese in passive bioremediation of ARD in sawdust-
biogenesis 239–244 supported wetland systems 189–205
bioleaching of marine manganese nodules Sulfur
239–252 removal of pyritic sulfur from coal by
removal from acid rock drainage 213–217 semicontinuous microbiological flotation
Mesophilic microorganisms in bioleaching of system using T. ferrooxidans 27–36
galena 229–238
Metal cyanide complexes T
biodegradation of 137–151 Tamboraque (Peru) BIOX® plant 132
Tetracyanonickelate
N biodegradation of 137–151
Nickel Thiobacillus ferrooxidans
bioleaching in recovery from pentlandite/ in bioleaching of uranium ores 101–119
pyrrhotite 85–100 in depression of pyrite flotation 3–11
in removal of pyritic sulfur from coal by
O semicontinuous microbiological flotation
Olympias Project (Greece) BIOX® plant 133 system 27–36
Thiobacillus thiooxidans
P in bioleaching of uranium ores 101–119
Paenibacillus polymyxa. See also Bacillus
polymyxa U
flotation tests with pyrite and chalcopyrite Uranium
67–81 bioleaching 101–119
Passive-biological systems
bioremediation of ARD by sulfate-reducing W
bacteria in sawdust-supported wetland Wastewater bioremediation 209
systems 189–205 Wetlands
control and treatment of acid rock drainage in ARD treatment 178–179
169–188 in passive bioremediation of ARD by sulfate-
Peptone reducing bacteria derived from sawdust
in biodegradation of metal cyanide complexes 189–205
formed during gold extraction 137–151 Wiluna (Australia) BIOX® plant 130–131
Pyrite
flotation depression by yeast and bacteria Y
3–11 Yeast
flotation tests with Paenibacillus polymyxa in depression of pyrite flotation 3–11
67–81
263

Mineral Biotechnology Microbial

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mineral Biotechnology Microbial

Uploaded by

Copyright:

Available Formats

Mineral Biotechnology:

Microbial Aspects of Mineral Beneficiation,

Society for Mining, Metallurgy, and Exploration, Inc.

Copyright © 2001 Society for Mining, Metallurgy, and Exploration, Inc.

All Rights Reserved. Printed in the United States of America

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in

Library of Congress Cataloging-in-Publication Data

TN688.3.B33 M57 2001

SECTION 3 BIOREMEDIATION 135

SECTION 4 BIOMINERALIZATION 219

Depression of Pyrite Flotation by Yeast and Bacteria 3

Calcium Removal from Bauxite Using Paenibacillus polymyxa 13

Desulfurization of Coal by Microbial Flotation in a Semicontinuous System 27

Biobeneficiation of Mineral Raw Materials 37

Role of Corundum-Adapted Strains of Bacillus polymyxa in the Separation of

Role of a Heterotrophic Paenibacillus polymyxa Bacteria in the Bioflotation of

* Department of Mining and Materials Process Engineering, Michigan Technological University.

MINERAL PYRITE AND COAL PYRITE FLOTATION PROCEDURE

ANALYSES AND CALCULATIONS

( %Wt. in Froth × Froth Wt. )

RESULTS AND DISCUSSION

Mineral Pyrite Flotation

COAL PYRITE FLOTATION

* Department of Chemical Engineering, Indian Institute of Science, Bangalore, India.

MATERIALS AND METHODS

TABLE 1 Chemical composition of bauxite ore used in biobeneficiation experiments

Leaching of Bauxite Ore

RESULTS AND DISCUSSION

MATERIALS AND METHODS

Microorganism and Culture Medium

Flotation Column Apparatus

Operation of the Semicontinuous Microbial Flotation System

Analysis of Sulfur and Ash

Continuous Cultivation of Thiobacillus ferrooxidans

Semicontinuous Microbial Column Flotation

FIGURE 3 Semicontinuous desulfurization of synthetic coal/pyrite mixture by microbial column

FIGURE 4 Semicontinuous desulfurization and deashing of Pittsburgh coal by microbial column

MICROBIAL REMOVAL OF IRON FROM MINERAL RAW MATERIALS

TABLE 1 Iron solubilization from kaolins by means of different microorganisms

MICROBIAL LEACHING OF ALUMINUM FROM MINERAL RAW MATERIALS

MICROBIAL REMOVAL OF SILICON FROM LOW-GRADE BAUXITES

*Ms is the silicon module, i.e., the Al2O3:SiO2 ratio.

The degradation of aluminosilicates by “silicate” bacteria greatly depends on the way of

IMPROVEMENT OF THE CERAMIC PROPERTIES OF KAOLINS

Type of Kaolin No. 3 Kaolin No. 5

BACTERIAL LEACHING OF SHALES

OTHER BIOBENEFICIATION PROCESSES

* Department of Metallurgy, Indian Institute of Science, Bangalore, India.

MATERIALS AND METHODS

Bacterial Strain, Growth, and Adaptation Conditions

TABLE 1 Composition of the control medium

Mineral and Ore Samples

Protein and Polysaccharide Production by Adapted and Unadapted Strains

Flocculation Studies Using Adapted Cells

Selective Flocculation of Hematite-Corundum Mixture by Corundum-Adapted Cells from

Separation of Hematite from Corundum by Different Concentrations of Corundum-Adapted

Protein and Polysaccharide Separation from Bacterial Metabolite and Characterization

Gel Electrophoresis for Isolation and Characterization of Bacterial Proteins

RESULTS AND DISCUSSION

TABLE 3 Separation of corundum-hematite-quartz mixture through selective flocculation using

TABLE 4 Separation of corundum-hematite-quartz mixture through selective flocculation using

To understand the beneficial role of corundum-adapted cells in the selective separation of

* Division of Mineral Processing, Luleå University of Technology, Luleå, Sweden.