WATER POLLUTION CONTROL RESEARCH SERIES
14010—06/69
Oxygenation of Ferrous Iron
U.S. DEPARTMENT OF THE INTERIOR • FEDERAL WATER QUALITY ADMINISTRATION
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Reports describe
the results and progress in the control and abatement
of pollution in our Nation's waters. They provide a
central source of information on the research, develop-
ment, and demonstration activities in the Federal Water
Quality Administration, in the U. S. Department of the
Interior, through inhouse research and grants and contracts
with Federal, State, and local agencies, research institu-
tions, and industrial organizations.
Inquiries pertaining to Mater Pollution Control Research
Beports should be directed to the Head, ProjectsReports
System, Planning and Resources Office, Office of Research
and Development, Department of the Interior, Federal Water
Quality Administration, Room 1108, Washington, D. C, 2024.2.
-------
Oxygenation of Ferrous Iron
by
Harvard University
Cambridge, Massachusetts 02138
for the
FEDERAL WATER QUALITY ADMINISTRATION
DEPARTMENT OF THE INTERIOR
14010—06/69
Contract PH 36-66-107
June, 1970
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C. 20402 - Price $1.76 (paper cover)
-------
FWQA Review Notice
This report has been reviewed by the Federal
Water Quality Administration and approved for
publication. Approval does not signify that
the contents necessarily reflect the views and
policies of the Federal Water Quality Adminis-
tration, nor does mention of trade names or com-
mercial products constitute endorsement or re-
commendation for use.
-------
ACKNOWLEDGMENTS
This work is based to a large part on research supported by contract
PH 86-66-107 between the Federal Water Pollution Administration, Department
of the Interior, and Harvard University. In the contract the scope of
the work was stated as follows:
Specific Aims;
(1) To determine rates of air oxidation of ferrous iron in the
_o
presence of SO, within the pH range 2-5;
(2) To determine rate of ferric iron hydrolysis within the pH
range 2-5;
(3) To investigate the colloid-chemical properties of hydrolyzed
iron (III) Parameters: (Fe+ ), (Fe ), (H ), (S04 ), PQ2;
(4) To investigate specific aims 1, 2, and 3 above under the
+2 +2
effect of the following catalysts: Mn , Cu , Si (QH)^,
Si02(S), Fe203(S).
The research effort resulting from this contract has become part
of a thesis presented in April, 1969 by Philip Charles Singer to the
Division of Engineering and Applied Physics of Harvard University in
partial fulfillment of the requirements for the degree of Doctor of
Philosophy. In order to accomplish a well rounded research objective,
the scope of the work was expanded to include a comprehensive treatment
on the chemistry of aqueous iron, and to consider models describing
pyrite oxidation.
A good deal of experimental data was collected in order to transform
the basic ideas and concepts of this research into the usable conclusions
which have been reached. The laboratory assistance of Karlene Spencer,
Gay Kunz, and Karl Schneider is an integral part of this report.
Appreciation is also due to Mr. Ronald Hill and Mr. Robert Scott
and their staff of the Federal Water Pollution Control Administration
who were instrumental in affording the authors opportunity and help'to
conduct field studies relating to acidic mine drainage. Grateful acknowl-
edgment is also extended to Professors J. Carrell Morris and Ralph Mitchell
for their fruitful suggestions and for patiently reading and criticizing
this manuscript.
-iii-
-------
Finally, Philip C. Singer is indepted to the United States Public
Health Service for a traineeship that provided financial support during
a major part of his graduate studies at Harvard University.
-iv-
-------
TABLE OF CONTENTS
page
ACKNOWLEDGEMENTS iii
LIST OF FIGURES , ix
LIST OF TABLES xiii
SYNOPSIS xiv
Chapter 1: INTRODUCTION
Chapter 2: EQUILIBRIUM RELATIONSHIPS OF AQUEOUS IRON
2-1 Introduction 2-1
2-2 Iron(ll) Solubility 2-3
2-2.1 Solubility in Natural Waters 2-3
2-2.2 Recent Observations of Ferrous Iron
Solubility in Carbonate-Bearing Waters .2-6
2-2.3 Experimental Determination of the
Solubility Product of Ferrous Car-
bonate (Siderite) 2-9
Experimental Procedure ....2-9
Experimental Results and Discussion . . 2 -12
X-Ray Analysis of Precipitate 2 -22
Stability Constant of FeHCO+ 2-22
Summary of Experimental Study ..... 2 -26
2-3 Solubility of Ferric Iron 2 -27
2-3.1 Solubility in Natural Waters 2-27
2-3.2 Effect of Complex Formation on Fe(lll)
Solubility 2-30
2-3.3 Experimental Determination of Sulfato-
Complex of Fe(IIl) 2-32
Experimental Procedure ...,,....2 -33
Experimental Results and Discussion. . . 2 -34
2-4 Oxidation - Reduction Reactions of the
Iron (II)-Iron (ill) System 2-40
References 2 -45
-------
Page
Chapter 3: KINETICS OF OXYGENATION OF FERROUS IRON
3-1 Introduction ....... ....3-1
3-2 Oxygenation of Ferrous Iron at Neutral
pH-Values 3-2
3-2.1 Oxidation in Natural Groundwaters .... 3 - 2
3-2.2 Oxidation of Fe(ll) in the Presence
of Ferrous Carbonate Oversaturation . . .3-5
Experimental Procedures 3-6
Experimental Results and Discussion ... 3 - 7
3-3 Oxygenation of Ferrous Iron in Acidic Systems. . 3 -13
3-3.1 Experimenal Study of Kinetics of
Fe(Il) Oxidation at Acidic pH-Values . . 3 -13
Experimental Procedure .3 -13
Experimental Results and Discussion ... 3 -16
3-4 Oxygenation of Ferrous Iron as a Function
of pH 3-24
3-4.1 Summary of Experimental Results 3 -24
3-4.2 Kinetic Implications of Results 3 -27
References .3 -36
Chapter 4: HYDROLYSIS OF FERRIC IRON
4-1 Introduction ..... 4-1
4-2 Kinetics of Ferric Iron Hydrolysis 4-2
4-2.1 Reactions of Fe+ with Water 4-2
4-2.2 Experimental Study of the Kinetics
of'Fe(IIl) Hydrolysis 4-4
Experimental Procedui'e 4-4
Experimental Results and Discussion . . .4-6
Solubility Product of Amorphous
Ferric Hydroxide 4 -15
4-3 Coagulative Properties of Ferric Iron 4 -15
4-4 Removal of Phosphate .....4 -19
VI
-------
Page
4-4. L Precipitation of Phosphate by
Ferric Iron 4 -21
Experimental Procedure 4 -21
Experimental Results and Discussion .... 4 -22
4-5 Summary ...........4 -25
References 4 -26
Chapter 5: OXIDATION OF IRON PYRITE: POLLUTION OF NATURAL WATERS
BY COAL MINE DRAINAGE
5-1 Introduction ........ 5-1
5-2 Thermodynamics and Stoichiometry of Reactions . . 5 - 2
5-3 Previous Investigations of the Kinetics and
Mechanisms of Pyrite Dissolution 5-4
5-3.1 Physical and Chemical Studies .5-4
5-3.2 Microbiological Studies .5-7
5-4 Purpose of Experimental Study .......... 5 -10
5-5 Oxygenation of Ferrous Iron ...5 -11
5-5.1 Experimental Procedure 5 -12
5-5.2 Experimental Results and Discussion .... 5 -14
Effect of Sulfate ., 5 -14
Catalysis by Dissolved Metal Ions 5 -17
The Effect of Clays 5-19
Catalysis by Powdered Charcoal ...... 5 -24
Effect of Iron Pyrite 5 -25
Effect of Microorganisms 5 -26
Summary ..5 -26
5-6 Field Investigations of Pyrite Oxidation in
Natural Mine Waters 5 -26
5-6.1 Collection and Analysis of Samples .... 5 -27
5-6.2 Results of Field Investigation 5 -31
Stoichiometric Relationship Between
Sulfate Concentration and Acidity 5 -31
Rate of Oxidation of Ferrous Iron 5 -33
Comparison with Laboratory Results .... 5 -35
Implications of Field Results . 5 -38
5-7 Oxidation of Iron Pyrite 5 -44
VlJ.
-------
Page
5-7.1 Experimental Procedures 5 -44
5-7.2 Results and Discussion 5 -46
Rate of Oxidation in Absence of Oxygen. . 5 -46
Oxidation Rate in Presence of Oxygen . . 5 -51
5-8 Conclusions 5 -53
5-8.1 Model Describing Pyrite Oxidation
and Pollution by Coal Mine Drainage ... 5 -53
References 5 -58
Chapter 6: CONCLUSIONS
6-1 Principal Findings 6-1
6-2 Practical Consequences and Implications
Resulting from this Research 6-2
References 6-7
APPENDICES
A Corrections of Experimental Solubility Data for
Temperature and Activity
B Relative Significance of Soluble Phosphato-Complexes
of Fe(lll)
C Derivation of Relationship Between Redox Potential
and Sulfate Concentration for Determination of
Stability Constant for FeSO +
D Thermodynamic Stability of Iron Pyrite
E Kinetics of Microbial Growth
F Autotrophic Iron Bacteria - Ratio of Ferrous Iron
Oxidized to Organic Carbon Synthesized
viii
-------
LIST OF FIGURES
Page
Figure
2-1 Solubility of ferrous iron in natural waters .... 2-5
2-2 Solubility of ferrous iron in carbonate-
bearing waters 2-5
2-3 Solubility of ferrous iron in sulfide-bearing
waters
2 - 8
2-4 Dual saturation-index diagram for calcite and
siderite in natural waters 2-8
2-5 Determination of solubility product of ferrous
carbonate • 2 -14
2-6 Conformance of experimental solubility product
to observations in natural groundwaters 2 -18
2-7 Determination of solubility product of ferrous
carbonate 2 -18
2-8 Determination of solubility product of ferrous
carbonate ...... 2 -20
2-9 Determination of solubility product of ferrous
carbonate 2 -20
2 -10 X-Ray diffraction pattern of experimental
ferrous carbonate 2 -23
2 -11 Standardization curve for divalent cation
electrode in ferrous perchlorate solution 2 -24
2 -12 Determination of free ferrous iron in bicar-
bonate solution 2 -24
2 -13 Solubility of ferric iron 2 -28
2 -14 Solubility of ferric iron in phosphate solution. . . 2 -28
2-15 Experimental apparatus for potentiometric
analyses • 2 -35
2 -16 Experimental data for determination of stability
constant of FeSO + 2 -37
ix
-------
Figure Pa§e
2 —17 Determination of stability constant of FeSO^' «... 2 -37
2 -18 p€ - pH diagram for iron « 2 -42
3 - L Oxidation and removal of ferrous iron under
conditions favoring precipitation of ferrous
-in
carbonate J - ?
3-2 Oxidation and removal of ferrous iron in the presence
of FeCO, oversaturation ........ 3-9
3-3 Effect of FeCO, precipitation en Fe(Il) oxidation
and removal
3 - - 3 -11
3-4 Effect of FeCO- precipitation on Fe(Il) oxidation
and removal ............. ........ 3 -11
3_5 U - V absorbance spectra of acidified solutions
of ferric perchlorate ............... • 3 -14
3-6 Relationship between absorbance of acidified
solutions of Fe(IIl) and Fe(lll) concentration,
at 272 mu ......................
3 -- 7 Rate of oxygenation of Fe(ll) in bicarbonate-
buffered systems ...... . ........... 3 -lc
3-8 Rate of oxygenation of Fe(ll) ............ 3 -19
3-9 Rate of oxygenation of Fe(ll) ...... . ..... 3 -19
3 -10 Oxygenation of Fe(Il) at pH 3 ... ....... . . 3 -20
3 -11 Rate of Oxygenation of Fe(ll) at pH 2 ..... ... 3 -20
3 -12 Oxygenation of ferrous iron at various initial
concentrations of Fe(ll), at pH 3 ........ , . 3 -22
3 -13 Rate of oxygenation of Fe(ll) over the pH-range
of interest in natural waters ............ - "25
4-1 Logarithmic and reciprocal plots of the rate of
hydrolysis of Fe+3 ......... . ....... 4-8
4-2 Logarithmic and reciprocal plots of the rate of
hydrolysis of Fe+3 ....... . ......... 4~9
4-3 pH-dependence of "first-order rate constant" for
hydrolysis of Fa4"3 ............. . ... 4 -11
-------
Figure
4-4 pH-dependence of "second-order rate constant"
for hydrolysis of Fe+3 ................ 4 -11
4-5 pH-dependence of "second-order rate constant" for
hydrolysis of F& in the presence of sulfate .... 4 -14
fj
4-6 Comparison between rates of hydrolysis of Fe
in the presence and absence of sulfate . . ...... 4 -14
4-7 Aggregation of colloidal silica dispersions by
hydrolyzed ferric iron ............ .... 4 -18
4-8 Solubility of ferric phosphate ............ 4 -18
4-9 Precipitation of phosphate by homogeneously-
generated ferric iron ................ 4 -23
5-1 Effect of sulfate on absorbance of Fe(lII) at
272 mu ........................ 5-13
5-2 Rate of oxygenation of ferrous iron as a. function
of pH ........................ 5-13
5-3 Rate of oxidation of ferrous iron in the presence
of sulfate ...................... 5-15
5-4 Effect of sulfate on the oxidation rate of Fe(Il)
at 50°C ......... . ............. 5-15
5-5 Effect of copper(ll) on oxidation of ferrous iron . . 5 -18
5-6 Rate of oxidation of Fe(ll) in the presence of
suspended aluminum oxide ....... ...,...,5 -18
5-7 Oxidation of Fe(II) as a function of AJU°o
concentration . ................. . . 5 -22
5-8 Effect of pH on surface-catalytic oxidation of
Fe(ll) ........................ 5-23
5-9 Rate of oxidation of Fe(Il) in the presence of
colloidal silica and bentonite clay ......... 5 -23
5 -10 Mining sites for field investigations of Fe(Il)
oxidation, near Elkins, West Virginia ........ 5 -28
xi
-------
Figure
Page
5 -11 Stoichiometric relationship between acidity and
aulfate concentration in mine drainage waters .... 5 -34
5 -12 Chemical composition of drainage water through
a strip mine « • -> -36
5 -13 Oxidation of Fe(Il) in drainage water after
leaving strip mine 5 -37
5 -14 Rate of oxidation of Fe(Il) in water collected
from air-sealed underground mine 5 -37
5 -15 Oxidation of Fe(ll) in water collected from air-
sealed mine 5 - .2
5 -16 Change in Fe(ll) concentration in millipore
filtered water collected from air-sealed mine .... 5 -42
5 -17 Oxidation of Fe(ll) solutions inoculated with
mine water • •* ~^
5 -18 Reduction of ferric iron by iron pyrite in the
absence of oxygen ......... 5 -47
5 -19 Rate of reduction of Fe(lll) as a function of
pyrite concentration -> f9
5 -20 Effect of initial Fe(IIl) concentration on
rate of reduction of Fe(IIl) by pyrite 5 -50
5 -21 Reduction of Fe(IIl) by pyrite in the presence
and absence of oxygen • • -> "52
B - 1 Log concentration diagram for phosphoric acid .... B - 2
B - 2 Distribution diagram for soluble hydroxo-species
of ferric iron
B - 2
xii
-------
LIST OF TABLES
Table Page
2-1 Equilibria Describing Fe(ll) Solubility 2-4
2-2 Experimental Determination of Solubility
Product of FeC03 2-21
2-3 Equilibria Describing Fe(lII) Solubility 2 -29
2-4 Experimental Data and Calculations in Determina-
tion of Stability Constant for FeSO^+ 2 -38
2-5 Equilibria for Construction of p£ - pH Diagram ... 2 -43
3-1 Kinetics of Oxidation of Ferrous Iron 3-3
4-1 Check on the Solubility Product of Amorphous
Ferric Hydroxide 4 -16
5-1 Comparison of Surface-Catalytic Rate Constants
with Uncatalyzed Rate Constants 5 -21
5-2 Chemical Catalysis of Oxidation of Ferrous Iron ... 5 -27
5-3 Summary of Field Data 5 -32
Xlll
-------
SYNOPSIS
The rate of oxidation of iron(li) by oxygen conforms to a rela-
tionship which is first-order in the concentrations of ferrous iron
and oxygen, and second-order in the concentration of hydroxide ion, at
pH-values between 6.0 and 7.5. The reaction proceeds relatively ra-
pidly at pH-values greater than 6.5; the half-time of the reaction is
4 minutes at pH 7.0, under a partial pressure of oxygen of 0.20 atmo-
spheres at 25°G. Deferrization processes in water treatment employ
the rapidity of the oxidation reaction in order to remove the influent
iron(Il) as insoluble iron(lll) hydroxide. Part of the iron(ll) may al-
so be removed as ferrous carbonate (FeCO-), the solubility product of
which is 6.0 x 10~ , as shown by this research.
_ 2
The dependence of the oxidation rate on [OH ] has been ob-
served, in this study, at pH-values as low as 4.5, where the half-time
has increased to approximately 300 days. At lower pH-values, the de-
pendence of the reaction rate on pH (or, more precisely, [OH ]) be-
comes less narked until at pH-values below 3.5, the oxidation proceeds
at a rate which is independent of pH. Here, a half-time of about 2000
days reflects the slowness of the oxidation reaction.
In the acidic drainage waters issuing from coal mines, half of
the acidity arises from the oxidation of the sulfide (S«(-II)) of iron
pyrite (FeS_) to sulfate, and .half stems from the oxidation of iron(ll)
to iron(IIl) by oxygen and the subsequent hydrolysis of the resultant
xiv
-------
Iron(lll). Observations of the rate of oxidation of ferrous to ferric
iron in these acidic streams (at pH values close to 3) show it to pro-
ceed considerably more rapidly than the laboratory studies at pH 3
would predict.
Several chemical agents which are indigenous to mine drainage
waters have been cited in the literature, in various circumstances,
as displaying catalytic properties in the oxidation of ferrous iron.
These include inorganic ligands, such as sulfate, which coordinate with
iron(H) and iron(IIl), soluble metal ions, such as copper(ll), man-
ganese(ll), and aluminum(lll), suspended material with large surface
areas and high adsorptive capacities, such as clay particles, and ma-
terials which accelerate the decomposition of peroxides in the presence
of iron(ll), such as charcoal. An investigation into the catalytic
capabilities of these chemical agents in synthetic mine waters demon-
strates that clay particles, or their idealized counterparts alumina
^A^2°3^ and silica (Si°2^> exert the greatest influence on the rate of
oxidation of iron(ll), but only at areal concentrations much larger than
those encountered in most natural mine waters. (The reaction proceeds
10-30 times more rapidly in the presence of 8000m /I of Al90- than the
uncatalyzed reaction.)
Autotrophic microorganisms have frequently been implicated as
the causative agents in the production of acidic mine drainage. These
organisms are able to utilize the energy released by the oxidation of
ferrous iron for their metabolic processes. A study of the oxidation
of iron(ll) in natural mine streams near Elkins, West Virginia shows
xv
-------
the reaction to proceed at a rate which is zero-order in iron(II).
The field results suggest that the observed rapidity of the reaction
in acidic mine waters is apparently the result of microbial catalysis.
Studies of the oxidation of iron pyrite in coal mine drainage
have usually considered only oxygen as the specific oxidant, with the
potentiality of iron(lll) as an oxidant having often been overlooked.
Experimental evidence obtained in laboratory systems shows that iron(IIl)
is rapidly reduced by iron pyrite both in the presence and absence of
oxygen. There is virtually no difference between the rate of reduction
of iron(lll) by pyrite, or the rate of change of soluble iror.CII), under
aerobic or anaerobic conditions indicating that the specific oxidant
of iron pyrite is ferric iron.
The time required for the reduction of 507« of the initial
iron(lll) concentration in contact with 1 gram/liter of pyrite is
approximately 250 minutes which is considerably less than the half-
time for the oxidation of iron(Il) even when accelerated by the chemi-
cal catalysts found in natural mine waters. Consequently, the rate-
limiting step among the reactions involved in the oxidation of iron
pyrite and the production of acidity in mine drainage waters is the
oxygenation of ferrous iron.
Based upon the experimental evidence presented, the oxidation
of iron pyrite in natural mine waters is shown to be compatible with a
cyclical reaction model involving the slow oxygenation of iron(II) to
iron(IIl) followed immediately by the rapid reduction of iron(IIl) by
pyrite, generating in turn more iron(Il) and acidity:
xv i
-------
slow
Fe(ll) + 02 =v Fe(lll)
fast „
Fe(lll) -t ~ ~ - " ""
The oxidation of iron(ll) by oxygen is the rate-determining reaction.
Oxygen is involved only indirectly in the oxidation of pyrite; it
serves to regenerate iron(lll) which is itself the specific oxidant
of pyrite. Precipitated iron(IIl) hydroxide within the mine serves as
a reservoir for soluble iron(IIl).
The experimental results and the model are discussed from the
standpoint of evaluating the various control measures which have been
proposed in order to deal with the costly problem of acidic mine drain-
age. In view of this research, emphases needs to be placed upon halt-
ing the catalytic oxygenation of iron(Il).
xv 11
-------
CHAPTER 1
INTRODUCTION
Aqueous iron assumes an important role in natural waters in the
limnological cycles of several key elements, in water supplies and water
purification processes, and in the formation of several types of industrial
wastes. Many natural phenomena can be explained and many practical appli-
cations can be derived as a result of investigations of the chemical be-
havior of aqueous iron. An account of the equilibrium and kinetic relation-
ships which describe and control the distribution and activity of aqueous
iron in natural waters is reported in this thesis, with special emphasis
placed on the kinetics of oxygenation of iron(II), its application to de-
ferrization processes, and its involvement in the formation of acidic coal
mine drainage.
The equilibrium relationships which govern the solubilities of ferrous
and ferric iron are considered in Chapter 2 in order to gain some insight
as to the concentrations of the various species expected in natural waters.
Included is a redetermination of the solubility product of ferrous carbonate
and an estimate of the stability constant for a possible bicarbonate-complex
of Fe(II).
The kinetics of the. oxygenation of ferrous iron are discussed in
Chapter 3, in both neutral and acidic systems. In the former case, attention
is paid to the oxygenation reaction in bicarbonate solutions which are
supersaturated with respect to ferrous carbonate. The rate of oxygenation
-------
1-2
of ferrous iron over the entire pH-range of interest in natural waters is
described, and the results are analyzed in view of modern kinetic theory
and the various mechanisms proposed for the reaction.
The pH-dependence of the oxygenation reaction in neutral and slightly
acidic systems, where hydrous ferric oxide was observed as the product of
the reaction, gave reason to suspect that hydrolysis of ferric iron was
directly involved in the oxygenation of ferrous iron. Chapter 4 contains
an investigation of the kinetics of hydrolysis of Fe+ in systems over-
saturated with respect to ferric hydroxide. The coagulative properties of
hydrolyzed ferric iron and phosphate removal by oxygenated ferrous iron are
also discussed.
The chemical results obtained in Chapters 3 and 4 are applied in
Chapter 5, a study of the chemistry of coal mine drainage. The kinetics
of the various reactions involved in the oxidation of iron pyrite and the
release of acidity are investigated and their relative rates compared in
order to ascertain which of the sequential reactions is rate-limiting. The
oxygenation of Fe(ll) is considered, subject to the catalytic influences of
several agents which are indigenous to natural mine waters, including micro-
organisms. A model is proposed, incorporating the salient features of the
kinetic study, in order to describe the oxidation of iron pyrite in mine
drainage waters. The pertinent consequences of the model are examined.
The significant results of the research are summarized in Chapter
6, along with some of the practical applications of these results.
-------
CHAPTER 2
EQUILIBRIUM RELATIONSHIPS OF AQUEOUS IRON
2-1 Introduction
The various species of iron which exist in natural waters are
governed by solubility, hydrolysis, complex-formation, and oxidation-
reduction relationships. These relationships predict whether or not a
given reaction will take place as written, and to what extent, i.e.,
what concentration of a given species is expected. In contrast, kinetic
relationships, which will be considered in Chapters 3 and 4, are needed
to predict the rate at which equilibrium is attained.
In order to understand the behavior of iron in natural waters, a
theoretical treatment is required in which various equilibria are assumed
to be applicable in controlling the different species of iron. Such a
theoretical treatment is not intended to provide an all-inclusive chemical
description due to the complexity of the natural system, but rather is
intended as an oversimplified version to gain some insight as to which
equilibria are relevant. The thermodynamic data employed have been ob-
tained in well-defined systems where the individual variables were iso-
lated. The combined investigations of a number of such isolated systems
are then compared to the natural system. Deviations between the behavior
predicted by thermodynamic considerations and that which actually occurs
in the real system can be attributed to the existence of non-equilibrium
-------
2-2
conditions where the kinetics of the reactions are limiting, or to an im-
proper description of the system due either to a lack of dependable thermo-
dynamic data or to an oversimplification in predicting which are the con-
trolling equilibrium relationships. The occurrence of such deviations be-
tween the "predicted" and the "actual" requires an explanation. A portion
of the material presented in this chapter is intended for just this pur-
pose.
The material covered is limited to a discussion of the solubility
of iron compounds in natural waters. In the case of ferrous iron, its
solubility is governed by the solubility of its respective carbonate, hy-
droxide, or sulfide, depending upon the composition of the water. The
solubility of ferric iron is correspondingly limited by its rather in-
soluble oxides and hydroxides or, in the presence of high concentrations
of phosphate, by ferric phosphate. The equilibrium concentration of Fe(lll)
may be increased as a result of complex-formation with chloride, sulfate,
phosphate, silicate, and organic matter.
Experimental studies of the solubility of iron in natural waters
are also presented in this chapter. These studies include a determination
of the ill-defined solubility product of ferrous carbonate, FeCO • an ex-
perimental analysis of the relevance of the bicarbonato-complex of Fe(ll),
FeHCO- , in carbonate-bearing waters; and a determination of the stability
constant of the sulfato-complex of Fe(lll), FeSO, , which was essential for
consideration of the catalytic effect of sulfate on the hydrolysis of
ferric iron, which is discussed in Chapter 4.
Since the relevant, features of the Chemistry of aqueous iron per-
taining to natural waters have previously been discussed by Stumin and
Lee (1), only an overall review is presented here, including an up-dating
-------
2-3
of their discussion with more recent data. The major portion of the chap-
ter is devoted to the experimental work undertaken to clarify our under-
standing of the behavior of iron in natural waters.
2-2 Iron(ll) Solubility
2-2.1 Solubility in Natural Waters
Under reducing conditions in natural waters, as in the bottoms of
lakes under conditions of stagnation, and in most groundwaters, ferrous
iron, in the -t- II oxidation state, is the stable form of iron. In waters
free of dissolved carbon dioxide and sulfide, the solubility of Fe(ll) is
controlled by solid ferrous hydroxide, Fe(OR) , as shown in Figure 2-1.
The equilibria utilized in plotting this solubility diagram and all subse-
quent diagrams for Fe(ll) are given in Table 2-1. (The reader is referred
to Sillen's discussion (2) on the graphic representation of equilibrium data
for the general principles in the construction of such solubility diagrams.)
In natural groundv/aters, alkalinities often exceed 5 x 10 eq./l.
(7). Figure 2-2 indicates that, for a water containing 5 x 10 moles/1.
of total carbonic species CT, the solubility of Fe(ll) is markedly influ-
enced by formation of ferrous carbonate, FeCO-. It is immediately evident
that at pH-values below 10.5, ferrous carbonate controls the concentration
of Fe(ll) in solution.
In hypolimnetic waters where the concentration of sulfide species
may be appreciable as a result of anaerobic reduction of sulfate, ferrous
sulfide, FeS, as shown in Figure 2-3 for a total sulfide content of 10~5
moles/1. (8), limits the solubility of Fe(ll) over the entire range of pH
encountered in naLural systems. It should be noted that even in the
-------
2-4
Table 2-1. Equilibria Describing Fe(ll)' Solubility
Equat ion
Number
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
2-9
Equilibrium
Constant
Reaction at 25°C
Fe(OH)2(g) = Fe+2 + 20H~ 8 x 10"16
Fe(OH)2(g) . FeOH+ + OH" 4 x 10~10
Fe(OH).., , + OH" = Fe(OH)~ 8.3 x 10~6
£.\ S) J
FeC03(s) = Fe+2 + C03"2 2.1 x 10"11
H2C03 = H+ + HC03" 4.2 x 10~7
HC03~ = H+ + C03"2 4.8 x 10"11
FeS(g) = Fe+2 + S"2 6 x 10"18
H0S, x = H+ + HS" 1.0 x 10~7
2 (aq)
HS" = H+ + S"2 1.3 x 10"13
Refer enc
3
3
4
5
5
5
6
6
6
-------
2-5
-3
£
w
8
-6
-7
8
10 11
PH
12 13
FIGURE 2-1. Solubility of ferrous iron in waters free of
appreciable alkalinity and sulfide.
C3
FeC03 + 2 OH • Fe(OH),, + CO /
x^ /
C - 5xlO"3M
1 /
/
-2
3 (s)
8
11
9 10
PH
FIGURE 2-2. Solubility of ferrous iron in waters containing
5x10* M of total carbonate species.
-------
2-6
presence of a concentration of 10 ti total carbonic species, ferrous sul-
fide continues to be the stable solid phase.
If equations 2-4 and 2-7 of Table 2-1 are combined to give
FeS, x + CO."2 = FeCO.,, , + S~2 (2-10)
Is) 3 3v.s)
where the equilibrium constant for the reaction, as written, is
K = =iO-- = 3xlO=
•* K4
it is seen that FeCO- becomes the stable solid phase (the reaction proceeds
as written) only when the ratio of sulfide to carbonate is smaller than
3 x 10~ . In. other words, the solubility product of ferrous carbonate be-
comes operative only when the ratio of alkalinity to total sulfide exceeds
2 x 10 at pH 6, or 8 x 10 at pH 7. Hence, at the extremely low concen-
trations of sulfide found in most groundwaters, ferrous carbonate governs
the solubility of ferrous iron, as in Figure 2-2.
2-2.2 Recent Observations of Ferrous Iron Solubility in_ _P.arbQn,ate_-
Bearing Waters
The equilibrium constant for the reaction
O *?
FeC03(s) = F + C03 (2"4)
given in Table 2-1, was computed from the free energy data tabulated by
Latimer (5), and is based upon the experimental work of Smith (9) iu 1913
and the calculations of Kelley and Anderson (10). Smith found the solu-
+2 -2
bility product of ferrous carbonate, K - [Fe ] [CO, *"], to be 3.45 x
SO J
10 at 30°C without correcting for activity. Kelley and Anderson modified
that value for 25°C and an ionic strength of zero (although their method of
correction is not immediately evident) and arrived at a free energy of
-------
2-7
solution for reaction 2-4 of 14.54 Kcal./mole, corresponding to a thermo-
dynamic equilibrium constant of K = 2.1 x 10
n so
More recent reports, however, indicate some discrepancy between
field measurements of the solubility of Fe(ll) and the solubility predic-
ted using the accepted thermodynamic constant from the literature.
Hem (11), in examining twenty groundwaters for equilibrium with calcite
(a crystalline form of calcium carbonate) and siderite (the crystalline
form of ferrous carbonate), presents a "dual saturation-index" diagram
for the two minerals. In general, one would expect a groundwater in
equilibrium with calcite to be in equilibrium with siderite, too, if both
minerals were present in the same geologic formation. Furthermore, one
would expect similar conditions of under- and oversaturation with respect
to the two solid phases. Yet, the best straight line through Hem's data
does not pass through the origin, although it has roughly a slope of
unity indicative of equivalent conditions of saturation (see Figure 2-4).
An error in the value of K would result in such an observation. (The
so
term "measured pH minus computed pH" is equal to the logarithm of the
degree of oversaturation.)
Ghosh, O'Connor, and Engelbrecht (7) sampled the influent ground-
water at eight water treatment plants in Illinois and reported values of
oversaturation with respect to ferrous carbonate of from 20 to 40 times.
When these observations were corrected for temperature and activity (12),
the oversaturation compared to the accepted value given by Latimer was
still on the order of five to ten times.
Larson (13) has suggested that the existence of a bicarbonato-
complex of ferrous iron, FeHCO, , may serve as an explanation for this
reported condition of apparent oversaturation, especially in view of Hem's
-------
2-8
-5
SB
o
M
I
-6
Ed
O
SB
o
-7
2
-8
Fe(OH)
FeS
(s)
M I T I 11 i n i i i i I I f I i I \
i
67 8 9 10 11 12 13
PH .3
FIGURE 2-3. Solubility of ferrous iron in waters containing 5x10 M
of total carbonate species and 10 M of total sulfide
species.
•f 2
-3
FIGUFvE 2-4.
-2 -1
measured pH
0 +1
computed pH
+2
Dual saturation-index diagram for calcite and siderite
in natural waters (after Hem (11)). Measured pH -
computed pK is equivalent to the log of the degree of
oversaturation.
-------
2-9
findings (14) that approximately 357o of the total manganous manganese,
Mn(ll), is present as the bicarbonate-complex, MnHCO, , in a grovmdwater
containing 5 x 10 eq./l. of alkalinity.
2-2.3 Experimental Determination of the Solubility Product of Ferrous
Carbonate (Siderite)
Since the solubility of ferrous iron in groundwaters has been
shown to be limited by the solubility of ferrous carbonate, and since
the 50-year-old solubility product cannot account for the high concentra-
tions of Fe(ll) found in natural groundwaters, this study was undertaken
in order to redetermine the solubility product of ferrous carbonate, and
to determine what effect, if any, complexation of Fe(Il) by bicarbonate
has on the overall solubility of Fe(ll).
Experimental Procedures
Ferrous carbonate was prepared directly in the laboratory by preci-
pitation from a solution of excess ferrous perchlorate and sodium bicar-
bonate. Ferrous perchlorate was added to a pre-determined concentration of
perchloric acid in a BOD bottle and the system was flushed with nitrogen
to remove all traces of oxygen. (Since conditions for the precipitation
of ferrous carbonate were found to be optimal above pH 6.5, the exclusion
of oxygen was a prime requirement, Fe(ll) being rapidly oxygenated at pH-
values greater than 6.0 (15).) Sodium bicarbonate was added and the vessel
was immediately stoppered to prevent contamination by oxygen. A series of
such bottles were stored under water at constant temperature again to pre-
vent seepage of oxygen into the system, the exclusion of oxygen proving to
be the major experimental difficulty.
-------
2-10
The preparations of ferrous carbonate were made in a constant
ionic medium to insure the constancy of the activity coefficients for
each series.
After a period of storage of one to five months, the samples
were removed from the water bath and aliquots were taken from each for
the determination of alkalinity, soluble ferrous iron, and concentration
pH. This latter term, pCH, is a measure of the concentration of H at a
given ionic strength. A combination pH electrode (Beckman Cat. No.
39142) was standardized in a reference solution containing a known con-
centration of HG1 and the same ionic strength as the sample being ana-
lyzed. The p°H of the sample was then measured (Corning Model 12 pH
meter) by immediately inserting the pH electrode into the vessel when it
was opened. It was feared that rapid evolution of CO^ upon exposure of
the sample to the atmosphere would raise its p H, but the measured ? H
was found to remain relatively constant.
Alkalinity was determined by acidimetric titration with 0.1M HC1.
Since oxidation of Fe(ll) and hydrolysis of the resultant Fe(lII) pro-
duces acidity which neutralizes a portion of the alkalinity, an aliquot
was rapidly titrated to pK 5 with a pre-determined amount of HC1 and
then slov/ly titrated to the endpoint at pH 4.3. (Oxidation of Fe(ll)
below pH 5 is relatively slow, as will be seen in Chapter 3.)
The determination of soluble ferrous iron was carried out by
rapidly filtering an aliquot of the sample into a test tube containing
dilute acid in order to quench any further reaction. The filtration
was performed under an atmosphere of carbon dioxide to prevent oxidation
of Fed!) and dissolution of FeCC>3. 100, 220, and 450 mu. filter papers
-------
2-11
(Millipore Filter Company) were employed, similar results being obtained
for each. Filterable Fe(ll) was measured using the colorimetric reagent,
bathophenanthroline (4,7-diphenyl-l,10-phenanthroline) (10).
To insure that the precipitate formed was crystalline ferrous
carbonate (siderite) and not merely an amorphous intermediate, the sedi-
ment was analyzed by X-ray diffraction. The precipitate was collected
following filtration and air-dried overnight before determining its
crystal structure using a Norelco X-ray diffractometer.
For the analytical determination of the stability constant of the
bicarbonate-complex of Fe(ll), a specific ion electrode (Orion Research)
was employed. The electrode contains a liquid ion-exchange resin having
a specificity for various divalent cations, including Fe , and is used
in conjunction with any standard reference electrode. The selectivity
coefficient for Fe is high and in the absence of other divalent cations
the electrode measures free Fe directly.
Various dilutions of a stock solution of ferrous perchlorate were
added to oxygen-free samples under one atmosphere of carbon dioxide and
containing pre-set concentrations of sodium bicarbonate. Again, exclu-
sion of oxygen was mandatory to prevent oxidation of Fe(ll). For a
given alkalinity, the greatest concentration of Fe(ll) was added such
that the solubility product of ferrous carbonate was not exceeded. A
constant ionic medium was again maintained to reduce variations in acti-
vity coefficients among the different systems.
Following the addition of Fe(ll) to the solutions of bicarbonate,
the samples were allowed one hour to reach equilibrium. The divalent
cation electrode was calibrated using standardized solutions of ferrous
-------
2-12
perchlorate, free of CO and at the same ionic strength as the samples.
(~"le solutions of Fe(C10,)_ had been standardized with permanganate, which
itself had been previously standardized against oven-dried sodium oxalate.)
The potentials of the test solutions containing bicarbonate were measured
at the same time as the standard solutions, alternating between the un-
knowns and the standards since the electrode potential is prone to drift
with time. The potentials were measured with a Corning Model 12 potentio-
meter. The p H of the samples were recorded and aliquots were analyzed
for alkalinity and total ferrous iron, Fe(ll), by acidimetric titration
and by pote> , -metric titration with standard permanganate, respectively.
Experimental Results and Discussion
The solubility product of ferrous carbonate was computed utilizing
the following chemical equilibria:
FeC03(s) = Fe+2 + C03'2; K°so = [Fe+2][C03'2] (2-4)
where KC is the concentration solubility product of crystalline ferrous
so
carbonate at a known ionic strength and temperature, and
HC03" = H+ + C03"2; KC2 = [H+][C03"2]/[HC03"] (2-6a)
where K% is the second acidity constant of carbonic acid at a given ionic
strength. It follows that the equilibrium constant for the reaction
FeCO ', , + H+ = Fe"1"2 + HCO," (2-11)
3(s) 3
is given by
KC [Fe+2][HCO.
Kc _ so _ :
^ K°2 Cir]
-------
2-13
Rearranging and taking logarithms, one obtains
pCH + pKCgo - PKC2 = -log ([Fe+2][HC03"]) (2-12)
where p- refers to the negative logarithm of that term. Hence, K° can
so
be readily calculated from a knowledge of the experimentally-determined
parameters p°H, soluble Fe(ll) (assumed to be equal to [Fe ]), and alka-
linity, and the known second acidity constant of carbonic acid under the
given experimental conditions.
Equation 2-12 suggests that a plot of p°H versus -log([Fe+2][HCO "])
will yield a straight line of unit slope having an intercept at p°H = 0
which is equal to pKC - pK° . Figure 2-5 is such a plot for a series of
SO 2.
preparations of ferrous carbonate at 22.5°C and an ionic strength of 0.1.
The best straight line of unit slope was fitted to the points resulting
in an intercept of -0.57 ± .17, as shown. The majority of the points fall
within the p H-range 6.5 to 7.5 where precipitation was found to be optimal.
The four points at the lower p H contain the greatest experimental uncer-
tainty since, due to their low alkalinities, they required only a small
amount of titrant before the endpoint was reached. The fact that these
four points still approximate the linear plot is gratifying.
The reason for the apparent scatter of the experimental points in
^
the region of p H 6.5 to 7.5 is due strictly to the experimental uncertain-
ty in measuring free Fe+ . It has been assumed that filterable Fe(ll) is
equal to the concentration of free Fe+ in equilibrium with the solid
phase. The fact that the concentration of filterable Fe(ll) for a given
sample was constant for three filters of different pore size (100, 220,
and 450 mu,) lends credence to such an assumption in that all solid Fe(ll)
is retained by the filter, i.e., filterable Fe(ll) equals soluble Fe(ll).
-------
2-14
7.5
7.0
6.5
6.0
- 5.5
60
O
5.0
4.0
i r
constant ionic medium
I - 0.1
temp. - 22.5°C
from Latimer (5)xx °9
Intercept at p H « 0 is -0:57
J I I I
5.0 5.5 6.0 6.5 7.0 7.5
PCH
FIGURE 2-5. Experimental data for determination of solubility
product of ferrous carbonate (siderite).
-------
2-15
On the other hand, it is certainly possible, even probable, that some
dissolution of FeCO, and some oxidation of soluble Fe(ll) does occur
during filtration. These effects act in opposite directions and tend
partially to cancel each other. The extent of such dissolution and ox-
idation during filtration is manifested in the scatter in Figure 2-5.
The experimental uncertainty has been computed on the bases of the
greatest degree of scatter in the p H-region 6.5 to 7.5.
Q
As shown by equation 2-12, the intercept at p H = 0 is equal to
the quantity pK° - pKC_. If one makes use of the Davies equation to
SO £-
estimate activity coefficients of single ions as suggested by Schin-
dler (17) for carbonates of bivalent metals, the thermodynamic second
acidity constant of carbonic acid as a function of temperature as re-
ported by Harned and Scholes (18), and the van't Hoff temperature rela-
tionship, the thermodynamic solubility product can be computed as shown
in Appendix A. The resulting solubility product, at 25°C and zero ionic
strength, is 5.7 ± 2.3 x 10"11 or pK is 10.24 ± 0.17. The previously-
accepted value is 2.1 x 10~ or pK =10.68, as shown in Table 2-1.
r so
Using the experimental value of pKC = 10.24 ± 0.17, one can
SO
compute the free energy of formation of ferrous carbonate as follows (the
numbers in parenthesis refer to the presently-accepted values (5)):
FeC03U) . Fe+2 + CO^2; K^ - lO'10' "(UT10-68) (2-4b)
AF° = -1.364 log K = 14.0 ± 0.2 K cal/mole (14.54) at 25°C
so
AF° = 14.0 ^F° „
Fe
14.0 = -20.3 - 126.2 -AF°
FeC03
AF° = -160.5 - 0-2 Real./mole (-161.06)
FeC03
-------
2-16
+2 -2
(The free energies of formation for Fe and C0_ have been taken from
Latimer (5).) Although the difference of only 0.6 K cal. in 160 K. cal.
does not appear to be very significant, its importance is magnified when
considering the difference between two large numbers as is often done in
computing the change in free energy for a given reaction. Such a small
error in the change in free energy can result in a much larger error when
the equilibrium constant for the reaction is computed.
If the thermodynamic constant given in the literature is modified
to the experimental conditions of this study using the same temperature
and activity corrections described above, the resulting constant does not
fit the experimental data, as Figure 2-5 demonstrates. The new solubility
product, however, can be shown to explain adequately the apparent over-
saturation described in the literature. If Hem's data (11) for the reac-
tion
CaC03,s) + H+ = Ca+2 + HCC>3~ (2-13)
is considered, the degree of oversaturation for calcite can be computed to
be
where Qr is the actual reaction quotient for 2-13 in the aquifer, arid KC
is the thermodynamic equilibrium constant for the same reaction. CL is
calculated in terms of the measured variables as
(Ca+2) (HGO ~)
Q^ = E L_B (2-15)
W f U "\
V. It J
m
the subscript m referring to measured quantities, and K is given by
-------
2-17
(Ca+2) (HCO ~)
,, m 3 m
C = (H+)
comp
where (H+) refers to the computed activity of H+ in equilibrium with
comp
the measured activities of calcium and bicarbonate. Substitution of 2-15
and 2-17 into 2-14 yields
(H+)
measured
If the degree of oversaturation for siderite is assumed to be the same as
that for calcite, one obtains, in similar fashion as with calcite
0 (Fe+2) (HCO,") /(H+)
<, c _ _I m 3 m S. (2-18)
^r ~ v ~ K ~ K_
^r ^?
where Q is the reaction quotient for dissolution of siderite as in re-
action 2-11, again in terms of the measured parameters, and K^ is the
equilibrium constant. It should be noted that
V
v so _ v (2-19)
jv — —~ **•
T K2 eq
as in equation 2-lla. Taking logarithms in equation 2-18, one obtains
log Sc = log QF - log KF (2-20)
If one now uses Hem's data (11) to compute the oversaturation for cal-
cite from 2-17 and the reaction quotient for siderite from 2-18, a plot
of log S versus log (X, should result in a straight line with a slope of
\j *
unity. The intercept at log SQ = 0 should be equal to log Kp which is
related to the solubility product of ferrous carbonate by 2-19.
Figure 2-6 is a plot of Hen's data on the log SG - log QF coordi-
nates. Two lines are shown on the graph, A corresponding to the solubil-
ity product obtained in this study, and B corresponding to the existing constant
-------
2-18
1.6
0.8 —
cr*
0.0
-0.8
-1.6
from experimental;study / O /
-3.2
FIGURE 2-6.
-2.4 -1.6 -0.8
LOG S
Conformance of experimental solubility product to
data obtained by Hera (13) in natural grourjdwaters.
S is the degree of oversaturation with respect to
calcite and Q is the reaction quotient for the
dissolution of siderite.
4.0
FIGURE 2-7. Experimental data for the determination of the
solubility product of ferrous carbonate.
-------
2-19
reported in the literature, (In plotting these lines, a value of 10.33
was used for pK , at 25 C and zero ionic strength.) Points 1 and 2,
which fall considerably off both lines, represent waters in contact with
rock formations which are quite low in carbonate content and where pH was
not measured directly at the time of collection. Due to the low concen-
trations of carbonate in these waters, they are weakly buffered with
respect to pH so that pH may have changed considerably during storage.
(Hem suggests that the initial pH of these two waters could have been half
a unit or more higher than the pH which was measured in the laboratory
after storage.) It is seen in Figure 2-6 that the experimental solubility
product obtained in this study conforms well to Hem's data.
Ghosh, O'Connor, and Engelbrecht (19) have indicated that their
apparent oversaturation may partially be explained by their inability to
measure the pH in the aquifer at the depths from which the waters were de-
rived. This redetermination of the solubility product additionally explains
their observations.
The solubility product of siderite was also determined under dif-
ferent experimental conditions by varying the temperature of the system
and the concentration of the constant ionic medium. Figures 2-7 and 2-8
show the results at two other temperatures, and the findings at an ionic
strength of 0.05 are presented in Figure 2-9. The data were treated in
the same manner as above and the outcome is summarized in Table 2-2. It
will be noted that the solubility of FeCO, increases with decreasing tem-
perature. The solubility products obtained in this study under various
experimental conditions are seen to be consistent among themselves and are
approximately three times greater than the accepted value reported in the
literature.
-------
T
I r
-------
2-21
Table 2-2. Experimental Determination of Solubility Product of FeC03
Temperature
C
Ionic
Strength
pK
^
30
Experimental
temperature
Thermo dynamic
Solubility Product
Corrected to
25°C, I = 0.0
17
22.5
30
22
Accepted
0.
0.
0.
0.
1
1
1
,05
Literature
-0.
-0.
-0.
-0.
72
57
46
,42
10.12
10.21
10.25
10.28
Value (5)
10.
10.
10.
10.
• i
21
24
20
31
10.68
6.
5.
6.
4.
2.
2 x
7 x
3 x
8 x
,1 x
„-!!
10
10"11
io"u
10"11
o-ll
10
- • —
^Activity corrections were made using the Davies equation
-log tf
= Az
-0.31
for single ion activity coefficients.
relationship
K? -AH!
" ~ R
The van't Hoff temperature
i
ln
was used to convert the experimental solubility products to 25°C.
(AH° = -4630 cal./mole at. 25°C (5).) Sample calculations are given
in Appendix A.
-------
2-22
X-Ray Analysis of Precipitate
The crystal structure of the precipitate was examined by x-ray
diffraction to establish whether the solution was in contact with an
amorphous deposit or with crystalline ferrous carbonate, i.e., siderite.
The diffraction pattern is shown in Figure 2-10 along with a table de-
scribing the standard pattern for siderite given by the American Society
for Testing Materials (20). The glancing angle, "6y corresponds to the
interplanar spacing, "d", and I/I-, is a measure of the relative intensity
of any single peak to the largest peak. For example, at 26 = 40.6 , the
largest peak is obtained so that I/ 1. = 1007o, while the intensity of the
peak at 29 = 68.0 is only 45« of the intensity at 40.6°. Comparison of
the diffraction pattern of the precipitate with the ASTM standard shows
definitely that the precipitate was crystalline ferrous carbonate so that
the solubility product obtained is the thermodynamic solubility product
of siderite.
Stability Constant of FeHCO_+
Using the divalent cation electrode and standardized solutions of
ferrous perchlorate, one obtains a standard curve relating the potential,
*
E _ , to the concentration of free ferrous iron, Fe , in the constant
DCE
ionic medium. (See Figure 2-11.) At concentrations of Fe below
10~ M, E ., approaches a limiting value due to selective exchange of Na
which, at that point, is present at a concentration three orders of mag-
t)
nitude greater than Fe . Consequently, in the experimental study in
-4
0.1M NaCIO, , a concentration of Fe(ll) in excess of 10 M was always
employed.
-------
CO
W
i;
11
40
20 -.
0
100
80
60
40 —
20 __
0
FIGURE 2-10. X-ray diffraction pattern^of experimental ferrous carbonate formed
in solubility study. Comparison with diffraction pattern of siderite.
X-ray diffraction data for siderite (ferrous carbonate. FeCOj) (20)
20 - 31.3 40.6 48.8 54.0 59.1 68.0 80.1 91.4 127.1
I/I - 25 100 20 21 30 45 20 20 25
-• 40
-- 20
-------
2-24
8
§
w
1-3
w
s
H
I
i
w
&
P-i
§
1
S
constant ionic medium
Fe'~ CONCENTRATION
FIGURE 2-11. Standardization curve for divalent cation electrode
in ferrous perchlorate solution.
-30
-40 —
-50 —
-60
-70
Standardization Curve
constant ionic
medium
Fe CONCENTRATION
FIGURE 2-12. Determination of free ferrous iron, Fe
bicarbonate solution.
+2
in
-------
2-25
The measured potential of each of the three systems investigated was
plotted against the total concentration of Fe(ll) for that system, deter-
mined independently by titration with permanganate, the points being com-
pared to the calibration curve. The results are presented in Figure 2-12.
+2
The standardization curve represents the concentration of free Fe " cor-
responding to the given potential, determined in a similar manner as Figure
2-11. It is seen that the total concentration of Fe(ll) in the sample is
equal to the concentration of Fe corresponding to the measured potential.
ry
Any deviation between the concentrations of Fe(ll) and Fe would imply
formation of FeHCO* or some other soluble complex of ferrous iron and
would have been indicated had the three points fallen to the left of the
calibration curve, i.e., less free Fe for a given Fe(ll). For example,
the second sample contained 8.3 x 10~ M of total ferrous iron as determined
by titration with permanganate. The potential of the sample was measured
as -51.3 mv. But according to the standard curve, this corresponds to a
concentration of free Fe of 8.3 x 10" M. Therefore, all of the total
Fe(ll) is present as free Fe
Six hours later, the procedure was repeated and with the exception
of a slight shift in the calibration curve, the results are identical,
i.e., the experimental points fall on the calibration curve. One must
-2
conclude that even in the presence of 1.1 x 10 eq./l. of alkalinity,
f\
there are no other measurable soluble species of Fe(ll) besides free Fe
From these results and the limitations imposed by the experimental
technique, it may be assumed that the concentration of Fe is more than
f\
ten times greater than that of FeHCO * at 10~ eq./l. of alkalinity. This
-------
2-26
implies that the equilibrium constant (stability constant) is less than
10 and the reaction is of no significance in natural waters. (In more
dilute systems, of ionic strength less than 0.1, the stability constant
should be even smaller.)
If FeHCO- had been significant, one would have expected curvature
in Figures 2-5 and 2-7 to 2-9, the degree of curvature being a function of
the concentration of HCO- . As already seen, these data plot well as
straight lines.
Hem (14) and Morgan (21) have investigated complex-formation of
Mn(II) by bicarbonate and found the solubility of Mn(ll) to be influenced
by such complexation. For the reaction
Mn"*"2 + HCO " = MnHC03+ (2-21)
Morgan reported an average thermodynamic equilibrium constant of 81,
while Hem found an average value of 63, indicating that 357* of the total
^
Mn(II) would be present as MnHCO- in a groundwater containing 5 x 10
eq./l. of alkalinity. However, no such complex of bicarbonate with
Fe(Il) was observed using the direct approach described above employing
the ion-sensitive electrode.
Summary of Experimental Study
It can be concluded that the solubility product of ferrous carbo-
nate, which is based upon experimental data obtained 50 years ago, is in
error by a factor of 3, Fe(Il) being three times more soluble than the ac-
cepted value would predict. The re-determined solubility product accounts
for the recent, reports of apparent oversaturation of natural groundwaters
with respect to siderite. The existence of a bicarbonato-complex of ferrous
-------
2-27
iron to partially explain increased solubility of Fe(Il) has been dis-
counted, the only soluble species of Fe(Il) of any significance in car-
bonate-bearing waters being free ferrous iron, Fe .
2-3 Solubility of Ferric Iron
2-3.1 Solubility in Natural Waters
In oxygenated waters, ferric iron, in the + III oxidation state,
is the stable form of iron. (Its rate of formation via the oxidation of
Fe(ll) is discussed in Chapter 3.) Due to its relatively great insolu-
bility, ferric hydroxide, or ferric oxide-hydroxide, controls the concen-
tration of soluble Fe(IH) in natural waters. Various structural forms
of insoluble ferric hydroxide are known to exist having solubility products
ranging from 10~35'5 to 10~44(22). In the experimental study of the kinetics
of hydrolysis of Fe(lll) which is presented in Chapter 4, a solubility
product of 10"38 was determined for freshly-prepared ferric hydroxide.
For illustrative purposes, this value will be used here. Figure 2-13 is
a solubility diagram for Fe(IIl) utilizing the equilibrium data presented
in Table 2-3. For the sake of convenience, the simple case has been
assumed in which Fe(OH)3 is in equilibrium only with its monomeric soluble
hydroxo-ferric complexes, the presence of multimers and other complex-
formers, such as silicate, sulfate, etc., being neglected for the time being.
In waters containing relatively high concentrations of phosphate,
insoluble ferric phosphate, FeP04> becomes operative in limiting the
solubility of Fe(lll). For a water containing a total concentration of
all phosphate species of 10"4M, Figure 2-14 demonstrates that the influence
-------
2-28
-2 —
§
M
I
B -6 -
z
8
H
M
H
>-^
O
-8 —
-10 —
2468
PH
FIGURE 2-13. Solubility of ferric iron.
-4
•« -
I
o
^ -8 _
M
M
M
-10 —
-12
FeP04 + 3 OH -
FIGURE 2-14.
Solubility of ferric iron in the presence of 10 M
of total phosphate species.
-------
2-29
Table 2-3. Equilibria Describing Fe(lll) Solubility
Equation
No. Reaction
2-22 Fe(OH)3(g) = Fe+3 + 30H~
2-23 Fe+3 + H20 = Fe(OH)"1"2 + H+
2-24 FeOH1"2 + H^ = Fe(OH)2+ + H+
2-25 Fe(OH)3(g) + OH* = Fe(OH)4~
O ^
2-26 FeP04(g) = Fe+ + P04"
2-34 H3P04 = H2P04" + H+
2-35 H2.P04~ = HP04~2 + H+
2-36 HP^"2 = P04~3 + H+
Equilibrium
Constant
at 25°C
io-38
6.8 x IO"3
2.6 x IO"5
10~5
10"24
7.4 x IO"3
-8
6.4 x 10
5.0 x IO"13
Reference
expt'l, Ch. 4
23
23
27
25
22
22
22
of solid FeP04 is exerted only in the acidic pH-region below pH 5. The
solubility product of FePO, is not a well-known quantity, there being
-17 9
three different values for the constant: 10 ' (computed from the tabula-
tion by Latimer (5)), 10"21'9(24), and IO"24 (25). Again, for illustra-
tive purposes, the value of 10~ determined by Stumm and Galal-Gorchev (25)
has been utilized.
One can derive an expression for the conversion of FePC>4 to
Fe(OH)_ in a similar manner as was done for the system FeCO^ - FeS in
equation 2-10. In this case,
FeP04(s) + 3 OH" = Fe(OH)3(s)
(2-27)
-------
2-30
where the equilibrium constant is
K0, .. [P0.~3]
K =^i=1014= 1— (2-27a)
** K22 [OH"]3
For a system at pH 6, the total concentration of phosphate must exceed
2 x 10~ ti in order for solid FePO to control the solubility of Fe(lll).
This is an unlikely situation in most natural systems but under localized
conditions where the composition of the water is non-uniform, FePO, may be
influential. Generally, however, the solubility of Fe(lll) is controlled
by its various oxides and hydroxides.
2-3.2 Effect of Complex Formation on Fe(lll) Solubility
The presence of organic and inorganic ligands which are capable of
coordinating with Fe(lll) to form soluble complexes serves to increase the
solubility of Fe(lll) in natural waters. In contrast to the case of ferrous
iron where the tendency to form complexes is insignificant, ferric iron has
a strong affinity for complexing ligands. In the preceding section where
the solubility of Fe(OH), and FePO, were considered, the influence of
o
complex-formation was neglected for reasons of simplicity. Since Fe has
an exceedingly strong affinity for the hydroxide ion, the relative affini-
ties of Fe+ for other ligands must be compared to its affinity for OH to
evaluate the extent of coordination of Fe(lll) by these other ligands.
Consequently, the relative concentrations of the various complexes of
Fe(III) are pH-dependent. This fact is demonstrated in Appendix B where
it is shown that in the presence of phosphate, a rather strong complex-
former, the effect of soluble phosphato-complexes of Fe(lll) is significant
only in the acidic pH-range below pH 4. (It is probable that mixed
-------
2-31
hydroxo-phosphate-complexes of Fe(lll) exist but there is insufficient
thermodynamic data to calculate their relevance.)
It is apparent that, in natural waters, the major effect of ligands
other than OH~ is manifested in the acidic pH-range where the concentra-
tion of OH~ is inconsequential.
A number of organic agents have a strong tendency to coordinate
with Fe+ , examples including EDTA and citrate (22). Again, the existence
of mixed organo-hydroxo-complexes is likely. Although Figure 2-13 implies
—8
that the concentration of soluble Fe(IIl) cannot exceed 10 M in the pH-
region 6 to 11, significantly higher concentrations of soluble Fe(IIl) in
natural waters have often been reported. Complex-formation with organic
material is usually cited as an explanation, Morgan (26) has considered a
hypothetical system involving nine metals and nine ligands to demonstrate
the significance of complex-formation, and has found that for the types of
ligands observed in natural waters, OH~ is the major ligand coordinated
with ferric iron. The discrepancy between predicted concentrations of
soluble Fe(lll) and reported concentrations can be partially explained by
the analytical difficulties encountered in distinguishing between soluble
Fe(IIl) and suspended colloidal ferric hydroxide. Lengweiler, Buser, and
Feitknecht (27), in order to completely sediment colloidal Fe(OH)_, demon-
strated the need to resort to ultracentrifugation. Hence, it is doubtful
that conventional methods of filtration are effective in differentiating
between soluble and suspended Fe(lll).
Furthermore, as Morgan (26) has shown, the concentration of organic
matter in natural waters is insufficient to account for significant com-
plexation of Fe(lll). However, in view of the extremely high concentrations
-------
2-32
of "solubilized" Fe(Hl) associated with organic color in natural waters,
it has been suggested (28) that these color -causing organic agents co-
ordinate with colloidal Fe(OH>3 forming a highly-dispersed peptized
colloid.
2-3.3 Experimental Determination of Sulfato -Complex of Fe(IIl)
As indicated in the previous section, the influence of inorganic
ligands other than OH" in coordinating with Fe(lll) is insignificant except
in the acidic pH-range, or in the presence of relatively high concentra-
tions of the competing ligand compared to OH~. In the case of sulfate,
both these conditions are fulfilled in the acidic waters draining through
coal and copper mines where oxidation of sulf ide minerals releases large
concentrations of sulfate. (Chapter 5 contains a complete discussion of
the chemistry characterizing mine drainage.) In these waters, where con-
centrations of sulfate exceed 10"^ and PH-values less than 3 are not un-
common, complex-formation of Fe(III) by sulfate appears to be interrelated
with the oxidation of ferrous iron and the hydrolysis of ferric iron.
The stability constant for the reaction
(FeSO/)
.
is fairly well-known (22), having been determined mainly by spectrophoto-
metric techniques. Potentiometry can also be conveniently applied to
measure such stability constants (29, 30). Since the rate of hydrolysis
of ferric iron (Chapter 4) was to be studied using a potent iometric method,
the investigation of complex-formation between sulfate and Fe(III) served
as a preparatory exercise in gaining familiarity with the technique.
-------
2-33
Furthermore, the experimentally-determined stability constant could then
be applied, as needed, in these future studies, some of which were con-
ducted in the presence of sulfate.
Experimental Procedure
The following electrochemical cell was employed in the potentio-
metric study of complex-formation of Fe(lll) by sulfate:
Pt
Fe(ll), Fe(lll), ClO, H+, Na+,
NaCl Hg C12 Hg (2-29)
(sat'd)
The cell consisted of a bright platinum spiral indicator electrode and a
calomel reference electrode separated by the test solution, contact between
the latter two being effected by a solution saturated with NaCl. (NaCl
was used in place of KC1 to avoid possible precipitation of KCIO^ in the
event of leakage of K+ from the calomel electrode.) The redox potential
is established by the electroactive Fe(Il)-Fe(111) couple in accordance
with the Nernst Equation
F
b =
,
(Fe+3)
(2-30)
A constant ionic medium of 0.1M NaC104 was maintained and the system kept
in a const ant -temperature water bath at 25°C so that equation 2-30 can
be written as
r\
E « E°' - 0.0592 log -£sli C2-30a)
E° referring to the standard potential at the given ionic strength and
temperature.
The study was conducted in the pH-range 1 to 3 in order to avoid
formation of higher-order and polynuclear hydroxo -complexes cf Fe(lll)
-------
2-34
other than FeOH+ , and to maintain the concentration of Fe(ll) constant
as free ferrous iron, Fe . Ferrous and ferric perchlorate were added to
a solution of NaCIO acidified with HC10, . Nitrogen was bubbled through
the system to remove all traces of oxygen and the system was placed on a
magnetic stirrer. Measurement of the potential of the Fe(Il) -Fe(IIl)
couple was effected using a Heath recording potentiometer (Model number
EUA 20-11), and the concentration pH was determined in the same manner as
previously described in the study of the solubility of FeCO., employing a
Leeds and Northrup pH meter (Catalog number 7664) . After observing con-
stant readings for the potential of the system in the absence of sulfate,
0.5 ml aliquots of a pre-standardized solution of Na-SO, were added from
a microburette. Following each addition of sulfate, the potential and p H
were recorded; stable readings being obtained within three minutes after
the ligand was added. The experimental apparatus is shown in Figure 2-15.
Experimental Results and Discussion
The experimental system can be represented by the following chemical
equilibria:
Fe+3 + SO.'2 = FeSO/; K. = - -r - (2-28a)
4 4 l
[H+][S04"2]
HSO" = H+ + SO '*; K = - (2-31)
44 a [HS04']
(2-23a)
[Fe ]
-------
2-35
3 ml. BURETTE
NITROGEN DIFFUSER
PLATINUM SPIRAL
INDICATOR ELECTRODE
CALOMEL REFERENCE
ELECTRODE ^
TEST
SOLUTION'
GLASS ELECTRODE
CALOMEL REFERENCE
ELECTRODE
I I I
STIRRING
BAR
REACTION VESSEL WITH WATER JACKET
TO MAINTAIN CONSTANT TEMPERATURE
FIGURE 2-15.
Experimental apparatus for potenttoraetrie
determination of stability constant of
sulfato- complex of ferric iron.
-------
2-36
The equilibrium constants are defined for 25 C and an ionic strength of
0.1. As derived in Appendix C, the potential can be related to the total
concentration of sulfate, S_, by the equation
(2.32)
E refers to the difference in potential between the system in the absence
of sulfate and that after a given addition of sulfate, ST» Having measured
E and [H*] as a function of S_, one can compute and plot the left-hand-
side of the equation versus S , Q, being the well-known first hydrolysis
constant of Fe . (0. = 2.89 x 10* at 25°C and an ionic strength of 0.1
n
(23).) In the absence of higher order sulfato-ferric complexes other than
that given by equation 2-28a, a straight line should result, if [H ] is
assumed to remain relatively constant.
^
If similar studies are conducted for a series of p H-values, the
slope of the linear plot in each case should be
K K
SLOPE = n = —i-S (2-33)
[H+] + Ka
Rearranging terms, one obtains
KlKa - K = [H+] (2-33a)
—— a
n
suggesting that if one plots — versus [H ], the intercept at [H ] = 0
_1
K,
n
should be equal to —, the reciprocal of the desired stability constant.
Figure 2-16 shows the results of two experiments conducted at p H
1.02 ± 0.04 and 1.39 ± 0.07. The raw data for curve A is presented in
Table 2-4. The linearity seems to validate the experimental assumptions
-------
2-37
5.0
~ 4.0
CM
Ov
tr>
o
CM
'
3.0
2.0
1.0
0.0
B
5xlO"5M Fe(III)
4xlO"4M Fe(III)
p H - 1.02
J J
0 5 10 15 20 - 25
TOTAL SULFATE CONCENTRATION, xlO M
FIGURE 2-16. Experimental data for determination of
stability constant of sulfato- complex
of Fe(III).
30
10
8
5 4
2 —
Determination of stability constant
of FeSO
4 '
-------
2-38
Table 2-4. Experimental Data and Calculations in Determination
of Stability Constant for FeSO^+
(1)
Volume
Added,
ml
0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
[FeCllH
(2) (3)
Sulfate P°H
Concn. ,ST
moles/1.
0 0.98
2.49xlO"3
4.97
7.45
9.91
1.233xlO"2
1.478
1.720
1.961
2.20
2.44 1.06
r~ [Fe(Hl)]T~ 4
(4)
Potential,
mv.
+513.5
509.0
503.9
500.2
494.5
490.9
487.9
485.1
483.1
480.7
478.1
x 10 ~4M
(5)
+4.5
9.6
13.3
19.0
22.6
25.6
28.4
30.4
32.8
35.4
(6)
exp (59 2^~1 ^
.191
.452
.678
1.10
1.41
1.71
2.02
2.26
2.57
2.96
(7)
6)x[l + ^
+ [H+]
.197
.466
.698
1.13
1.45
1.76
2.08
2.33
2.65
3.05
*Potential readings versus saturated calomel reference electrode.
+"£ is the equilibrium potential of the cell in the absence of
sulfate. E9 refers to the potential after the addition of S moles/1.
of sulfate. (See Appendix C.)
-------
2-39
made in deriving equation 2-32. The slopes of a series of such plots at
different pCH-values were calculated and Figure 2-17 has been drawn in
accordance with equation 2-33a. As seen from the figure, — at [H ] = 0
gives — = 1.9 x 10~3. Consequently, K = 5.3 x 102 or 102* . This
Kl
agrees quite well with the values reported in "Stability Constants" (22),
which range from 102<3° (at 25°C and 0.5M NaClO^ to 103*°2 (at 18°C and
0.066M NaCIO,). The two values given for 20°C and an ionic strength of
0.1, conditions which most closely resemble those of this experiment, are
in2.62 . ,02.66
10 and 10
The slope of Figure 2-17 which is equal to K K is 14.9 If the
•L ct
-2 -1.55
experimental value for K is used, K becomes 2.8 x 10 or 10
-1 59
This is an excellent agreement with the value of 10 * found by Reynolds
and Fukushima (31) at 25°C and I = 0.1, and serves to verify further the
experimental technique and procedure.
One can now apply this stability constant for FeSO, to the mine-
water system using the same technique as that employed for phosphate-
complexes of Fe(lH) in lakes. At PH-values greater than 2, the magnitude
-2 -2
of K indicates that most of the sulfate exists as SO , i.e., [SO, ] =
a <4 t
S = 10~ n. If the desired relationships are arranged based upon the
previously-discussed equilibria (equations 2-28a and 2-23a) , then, for
25°C, I =0.1, and S = 10*M, the extent of coordination is given as
[FeS04+] [FeOH+2] Qh
pH ~= '"
3 lO0'7 10°'54
2 10° '7 10-° «4
-------
2-40
Hence, it is evident that under conditions of low pH and high concentra-
tions of ligand, as in mine drainage waters, complex-formation can be
significant.
2-4 Oxidation-Reduction Reactions of the Iron(ll)-Iron(III) System
In oxygenated waters, it is well-known that ferrous iron is
thermodynamically unstable, being oxidized to ferric iron. Iron undergoes
reversible oxidation and reduction reactions dependent upon given solution
conditions. Oxidation-reduction equilibria relate the various oxidation
states of a certain element and are characterized by the thermodynamic
redox potential, E, in accordance with the classical Nernst Equation. Such
a systematization is analogous to that of considering acid-base equilibria
as ratios between protonated and deprotonated bases characterized by an
acidity potential defined in terms of the change in free energy for the
proton transfer reaction. Hence, the electron intensity can be treated in
the same manner as the proton intensity or the activity of the hydrogen
ion (32). For example,
AFU
PH = -log (H+) = ^J-RT P =
pH = -E p£ = -ER
G/0.0592 K/0.0592
The quantity 0.0592 assumes one mole of protons or electrons transferred
at 25°C. E refers to the acidity potential, the potential at the glass
G
electrode or some other electrode for measuring pH, ER refers to the rever
sible redox potential at an inert electrode, and AFR andAF£ refer to the
-------
2-41
changes in free energy for the proton and electron transfer reactions,
respectively. The term p£ is a convenient measure of electron intensity,
and when seen in conjunction with an equivalent measure of proton inten-
sity, ie., pH, its value can be further appreciated.
The distribution of the various species of iron which are stable
under given solution conditions can be conveniently summarized with the
aid of a p£ - pH diagram in which the predominant species are presented as
a function of the two master variables, as shown, in Figure 2-18. Such a
diagram can be made to be three-dimensional by incorporating a third axis
for total concentration of soluble iron, assumed to be 10 M in Figure 2-18.
Actually, the solubility diagrams showing log Fe versus pH which were pre-
sented earlier are two-dimensional variations of such a master plot, in
which p£ had arbitrarily been made constant.
Since the p£ -pH diagram is a theoretical one based entirely upon
reversible Nernstian redox potentials, only well-defined systems where
the measured potential is known to be reversible can be interpreted in
terms of such a diagram. Due to the complexity of natural systems where
several electroactive species exist, mere insertion of an electrode into
the water will usually yield a mixed potential and, hence, cannot be re-
lated or defined in terms of a p£ -pH diagram. The diagram derives its
main value in serving as a guide toward expected behavior under various
conditions of pH and p£ .
The thermodynamic equilibria diagramatically shown in Figure 2-18
are presented in Table 2-5 together with the corresponding relationship
of p£ to pH. The values for the free energy used in these calculations
were taken from the tabulation by Latimer (5) and the experimental
-------
2-42
15
10
5 —
-5
-10
FIGURE 2*18. pfc ••-• j>H diagram foreiron. Concentration of^soluble
iron species is lO'^M and alkalinity is 10 eq/1.
-------
2-43
Table 2-5. Equilibria for Construction of p£ -pH Diagram
Reaction
2H.O = GV ,. + 4H+ + 4
L £.\&) e
2H.O + 2 - = H_, x + 20H~
2 e Zv.gJ
Fe+2 = Fe+3 + •'
Fe+2 + H_0 = FeOH+2 + H+ + e"
Fe+2 + 2K.O = Fe(OH)* + 2H+ + e"
L i.
Fe+2 + 3H20 = Fe(OH)3(g) + 3H+ + e"
Fe(OH)0, v + H.O = Fe(OH),, , + H+ +
2(s) 2 3(s)
FeC03(g) + 3H20 -
Fe(OH)3(s)+ HC03" + 2H+ + e"
r> j-7
Fe° = Fe+Z + 2e
Fe° + HCO ~ = FeCO,, , + H+ + 2e"
J J \ S )
Fe° + 2H20=Fe(OH)2(g) + 2H+ + 2e"
Reaction
FeC03( v + 2H 0 = Fe(OH)2, , + H+
+ HC03"
Fe+3 + H,0 , FeOH+2 + H+
L
FeOII+2 + H00 = Fe(OH)0+ + H+
L L
FeC03(g) +H+ =Fe+2 + HC03-
£07volts
-1.23
-0.828
-0.771
-0.914
-1.19
-1.06
e" -0.274
-1.08*
+0.440
+0.440*
+0.048
&
+13.6*
+ 2.17
+ 4.6
-0.09*
p£-pH Relationship
p£ = 20.8-pH
P£ = -PH
,, rt , (Fe+2)
r\£ 1 ^ L. nTT 1 no— \~.$--,.
(FeOH+Z)
T\C ?n i ^i-ifi TJ-IO. , , e.
pt — cu. i-ipii-iog
(Fe(OH)2+)
p€ = 17.9-3pH-log (Fe+2)
p€ = 4.63-pH
p£ = 15.3-2pH
p£ = -9.93
p£ = -5.93-1/2 pH
p£ = -0.97-pH
pH Relationship
pH = 13.6 + log(HC03")
IT "7 1-o
ptl - L. 1 - iOg „
(FeOH+Z)
oil 4 6 1-a (FeOH+2)
(Fe(OH)2+)
pH = 0.09 - log (Fe+2) -
log (HC03~)
*Computed from the tabulation of free energies given by Latimer (5).
^Calculated usingAF
mentally in section 2-2.3.
FeCO,
= -160.5 Kcal./mole, determined experi-
-------
2-44
solubility product of FeCO_ (see section 2-2.3). It is clearly seen
that in oxygenated natural waters, where p( -values in the vicinity of
the upper dotted line in Figure 2-18 are observed, ferric hydroxide is
the predominant form of iron except in the acidic pH-range below 4,
where the solubility of Fe(lll) increases. In the absence of oxygen,
as in hypolimnetic waters, most natural groundwaters, and in anaerobic
biological systems such as sludge digesters, ferric iron is readily
reduced by organic matter and by sulfide. The extent of the reduction
depends upon the p£ and pH of the system, the form of the resultant
Fe(ll) also depending upon pH. The p£ of such systems is in the
region of the lower dotted line in Figure 2-18.
Upon re-exposure to oxygen, Fe(lll) ,again becomes the stable form
of iron. The rate at which Fe(lII) is formed, however, cannot be in-
ferred from such thermodynamic considerations and requires investiga-
tion into the kinetics of oxidation of Fe(ll). This is the basis for
Chapter 3.
The p£ - pH diagram presented has been simplified in order to
demonstrate the underlying principles describing the redox reactions
between Fe(ll) and Fe(lll) in natural waters. For complex systems,
however, the various soluble and insoluble species of Fe-S, as well
as many of the Fe-organic complexes, should be superimposed on the
diagram. Nevertheless, such simplified diagrams help to clarify the
chemistry at work in aqueous solutions.
-------
2-45
References
1) Stumm, W. and Lee, G. F., "The Chemistry of Aqueous Iron," Schweiz.
Zeits. Hydro1., 22, 295 (1960)
2) Sillen, L. G., "Graphical Presentation of Equilibrium Data," Ch. 8,
page 227, in Part 1, Volume 1, Treatise on Analytical Chemistry,
I. M. Kolthoff and P. J. Elving, editors, Interscience, New York,
(1959)
3) Leussing, D. L. and Kolthoff, I. M., "The Solubility Product of
Ferrous Hydroxide and the lonization of the Aquo-Ferrous Iron,"
Journ. Amer. Chem. Soc., 75, 2476 (1953)
4) Gayer, K. H., and Woontner, L., "The Solubility of Ferrous Hydroxide
and Ferric Hydroxide in Acidic and Basic Media at 25 C," J. Phys.
Chem., 60, 1509 (1956)
5) Ringbom, A., Solubility of Sulfide, Analytical Section, IUPAC (1953)
6) Latimer, W. E., The Oxidation States of the Elements and Their Poten-
tials in Aqueous Solutions, second edition, Prentice-Hall Inc.,
Englewood Cliffs, N.J. (1952)
7) Ghosh, M. M., O'Connor, J. T., and Engelbrecht, R. S., "Rate of
Precipitation of Iron in Aerated Groundwaters," Journ. San. Eng.
Div., Proc. ASCE, 90, 199 (1966)
8) Hem, J. D. , "Some Chemical Relationships Among Sulfur Species .and
Dissolved Ferrous Iron," U. S. Geol. Surv. Water Supply Paper
1459-C, Washington (1960)
9) Srnith, H. J., "On Equilibrium in the System: Ferrous Carbonate,
Carbon Dioxide and Water," Journ. Amer. Chem. Soc. 40, 879
(1918)
10) Kelley, K. K., and Anderson, C. T., "Contributions to the Data on
Theoretical Metallurgy IV Metal Carbonates-Correlations and
Applications of Thermodynamic Properties," Bulletin 384, U. S.
Bureau of Mines, Washington (1935)
11) Hem, J. D., "Restraints on Dissolved Ferrous Iron Imposed by Bi-
carbonate, Redox Potential, and pH," U. S. Geol. Surv. Water
Supply Paper 1459-B, Washington (19601
12) Stumm, W., and Singer, P. C., "Precipitation of Iron in Aerated
Groundwaters," discussion, Journ. San. Eng. Div., Proc. ASCE, 92
120 (1966)
-------
2-46
13) Larson, T. E., "Oxidation of Metals and Ions in Solution," p. 433 in
Principles and Applications of Water Chemistry, S. D. Faust and
J. V. Hunter, editors, John Wiley and Sons, Inc., New York (1967)
14) Hem, J. D. , "Manganese Complexes with Bicarbonate and Sulfate in
Satural Water," Journ. Chem. Eng. Data, 8, 99 (1963)
15) Sttmm, W., and Lee G. F., "Oxygenation of Ferrous Iron," Ind. En&.
Chem., 53, 143 (1961)
16) Leey G. F., and Stumm, W., "Determination of Ferrous Iron in the
Presence of Ferric Iron," Journ. Amer. Wat. Works Asscn. 52,
1567 (1960)
17) Schindler, P. W., "Heterogeneous Equilibria Involving Oxides,
Hydroxides, Carbonates, and Hydroxide Carbonates," Ch. 9, p. 196
in Equilibrium Concepts in Natural Water Systems, R. R. Gould,
ed., Advances in Chemistry Series 67, Amer. Chem. Soc., Washington
(1967)
18) Harned, H. S. , and Scholes, S. R. , "The lonization Constant of HC03
from 0 to 50°," Journ. Amer. Chem. Soc., 63, 1706 (1941)
19) Ghosh, M. M., O'Connor, J. T., and Engelbrecht, R. S., "Precipitation
of Iron in Aerated Groundwaters," closure of discussion, Journ.
San. Eng. Div., Proc. ASCE, 93, 118 (1967)
20) Amer. Soc. Testing Materials, X-Ray Powder Data File, Special Tech-
nical Publication No. 48-L, ASTM (1962)
21) Morgan, J. J., "Chemistry of Aqueous Manganese II and IV," Ph.D.
Thesis, Harvard University (1964)
22) Sillen, L. G., and Martell, E. A., Stability Constants of Metal-Ion
Complexes, Special Publication No. 17, London, The Chemical
Society (l964)
23) Milburn, R. M., "A Spectrophotometric Study of the Hydrolysis of
Iron III Ion. Ill Heats and Entropies of Hydrolysis," J. Amer.
Chem. Soc., 79, 537 (1957)
24) Zharovskii, F. G., Trudy Kowissii Anolit Khim Akod Nank SSSR, 3,
L01 (1951)
25) Galal-Gorchev, H., and Stumm, W. , "The Reaction of Ferric Iron
with Ortho-Phosphate," Journ. Inorg. Nucl. Chera., 25, 567 (1963)
26) Morgan, J. J., "Metal-Organic Complexes," paper presented at Univ.
of Alaska Symposium on Organic Matter in Natural Water, Sept.
2-4, 1968, Fairbanks, Alaska
-------
2-47
27) Lengweiler, H., Buser W., and Feitknecht, W., "Die Ermittling Der
Loslichkeit von Eisen (III) - Hydroxiden Hit 59 Fe," Helv. Chim.
Acta, 44, pp. 796 and 805 (1961)
28) Stumm, W., "Metal Ions in Aqueous Solution," p. 520 in Principles
and Applications of Water Chemistry, S. D. Faust and J. V. Hunter,
editors, John Wiley and Sons, Inc., New York (1967)
29) Willis, R. L. S., "Ferrous-Ferric Redox Reaction in the Presence
of Sulfate Ion," Trans. Farod. Soc., 59, 1315 (1963)
30) Matoo, B. N., "Stability of Metal Complexes in Solution. III. Ion
Association in Ferric Sulfate and Nitrate Solutions at Low Fe III
Concentration," Zeits. for Phys. Chem. Nerre Folge, 19, 156 (1959)
31) Reynolds, W. L. , and Fukushima, S., "Iron (II) and Iron (III) Isotope
Exchange in Presence of Sulfate Ions," Inorg. Chem.,2, 176 (1963)
32) Stumm, W., "Redox Potential as an Environmental Parameter, Conceptual
Significance and Operational Limitations," Proc. Third Intl. Conf.
Wat. Poll. Research, Munich, Sept. 1966
-------
CHAPTER 3
KINETICS OF OXYGENATION OF FERROUS IRON
3-1 Introduction
Ferrous iron is thermodynamically unstable in the presence of
oxygen. The rate at which Fe(ll) is converted to Fe(lll) cannot be in-
ferred from thermodynamic data but requires a thorough investigation of
the kinetics of the oxidation, i.e., the mechanism by which the reaction
occurs and the various factors which influence such a mechanism.
Conventional water treatment for the removal of iron consists of
aeration of the raw water followed by sedimentation and filtration. The
former process allows for the escape of C0?) thus raising the pH, and for
the introduction of oxygen which oxidizes Fe(ll) to Fe(lll). The latter
hydrolyzes to form a precipitate which is subsequently removed by sedi-
mentation and filtration.
In natural waters, the cycles of phosphorus and sulfur are inter-
related with the iron cycle. The rate of oxidation of ferrous to ferric
iron during the spring and fall overturn is, therefore, partially respon-
sible for the rate of dissolution and oxidation of sulfide and for the
rate of turnover, of phosphate.
This chapter is comprised of experimental studies of the rate of
oxygenation of ferrous iron in the neutral pH-range encountered in natural
groundwaters, of heterogeneous oxygenation in the presence of precipitating
-------
3-2
ferrous carbonate, and of oxygenation in the acidic pH-range typical of
drainage waters from coal mines. A model is presented depicting the
rate of oxidation of Fe(Il) over the entire range of pH encountered in
natural waters. The model is shown to be compatible with existing the-
ories describing the mechanism by which Fe(ll) is oxidized, and its
characteristics and implications are explained in accordance with modern
kinetic theory.
3-2 Oxygenation of Ferrous Iron at Neutral pH-Values
3-2.1 Oxidation in Natural Groundwaters
The deferrization of iron-bearing waters is customarily accom-
plished by oxidation of the ferrous iron in the raw water to insoluble
ferric hydroxide utilizing dissolved oxygen. The reactions describing
the process are
Fe+2+ 1/4 0£ + E+ = Fe+3 + 1/2 H^ (3-D
Fe+3 + 3H00 = Fe(OH)-, , + 3H+ (3-2)
2 Jv. s;
It is well-established that the oxidation of Fe(ll) in the neutral pH-
range proceeds at a rate which is dependent upon the concentrations of
Fe(ll), dissolved oxygen, and OH , as shown in Table 3-1, in accordance
with the rate expression
-dCFe(II)] = k[Fe(ii)] Pn [OH~J2 (3-3)
Qt (Jy
The rate constant k is reported by Stumm and Lee(3) to be 8.0 (1 2.5) x
13 2 -2 -1 -1
10 liter mole atm min ; their results were derived for the
-------
3-3
Table 3-1. Kinetics of Oxidation of Ferrous Iron
Rate Equation Reference
-dCFe(ll)] = k[Fe(H)][0 ]
£_ Just (1)
dt Cco232
= k'[Fe(H)][Oj[OH~]2 Just (1)
1 = k[Fe+2][0_] Holluta and Eberhardt (2)
_
dt 2
where k = function of pH
-d[Fe(ll)] = k[Fe(Il)][0.][OH"]2 Stumm and Lee
at /
pH-range 6.0 to 7.5. Just (1) and Stumm and Lee (3) carried out their
studies in bicarbonate buffer systems, the latter work being conducted
under conditions such that interference by precipitation of ferrous car-
bonate was precluded. The introduction of Fe(lll) at concentrations up
to 10~St had no effect on the rate of oxidation (3). The marked pH-
dependence of the reaction rate should be noted; a 100-fold increase in
rate was observed for each increase of one pH unit.
Ghosh, O'Connor, and Engelbrecht (4) conducted field studies of
the rate of oxidation and removal of Fe(ll) from natural groundwaters
at eight water treatment plants in Illinois. Their results corroborated
the first-order dependence of the reaction rate on [Fe(ll)], but only a
-------
3-4
remote relationship was observed between the rate and the pH of the system.
However, since the reported variation in pH among the eight studies was
only 0.3 pH units, one cannot justifiably conclude that pH is insignifi-
cant. On the other hand, a definite correlation was noted between alka-
linity and rate of oxidation. Stumm and Lee (3) also noticed, in addi-
tion to the second-order dependence on [OH ], that the reaction proceeded
at a slower rate in solutions of low alkalinity. This discrepancy was
attributed to the slow response of the KCO ~ - CO buffer system to
localized changes in acidity brought about by the oxygenation reaction,
-2
or to possible base catalysis by HCO or CO., . Apparently, higher
alkalinities are associated with higher rates of oxidation.
The actual rate of removal of Fe(ll) from natural waters (4) was
found to be approximately one order of magnitude less than predicted by
the studies of oxidation in synthetic systems (3). Ghosh, O'Connor, and
Engelbrecht (4) observed removal of Fe(Il) both by oxidation to insoluble
Fe(OH). (reactions 3-1 and 3-2) and by precipitation as FeC03, in accord-
ance with
Fe+2 + HC0" = FeC0. + H+ (3-4)
Despite the dual mechanism for removal, the rate was still less than
the predicted rate. Two explanations were given (5) for the slowness of
the reaction in nature. Firstly, the field studies were conducted in
waters of alkalinity three to five times less than the laboratory studies.
Since the rate of the reaction is apparently accelerated by alkalinity,
the reaction should proceed at a slower rate in the Illinois waters.
Secondly, the presence of catalysts or inhibitors present in the natural
-------
3-5
system could account for variations not only between the field and
laboratory studies, but also among the different groundwaters them-
selves. Organic substances and sulfide compounds, in particular, were
cited as inhibitory agents (5) (6) in accordance with the sequence
Fe(ll) + 1/4 02 + ORG. = Fe(IIl) - ORG. COMPLEX (3-5a)
Fe(lll) - ORG. COMPLEX = Fe(Il) + OXIDIZED ORG. (3-5b)
Fe(ll) + 1/4 0 + ORG. = Fe(lll) - ORG. COMPLEX (3-5a)
In this scheme, the ferrous-ferric system functions as an electron-
transfer catalyst for the oxidation of organic material by oxygen. The
overall rate of oxidation of Fe(Il) may be retarded depending upon the
rate of oxygenation of Fe(ll) in the presence of organic matter in com-
parison to the rate of reduction of Fe(lll) by the organic matter.
The influence of such factors as these need to be quantitatively
evaluated before the actual rate of oxidation of Fe(ll) in natural
systems can be predicted and before an efficient scheme for iron removal
can be designed.
3-2.2 Oxidation of Fe(ll) in the Presence of Ferrous Carbonate Over-
saturation
It was indicated above that deferrization of iron-bearing waters
may be achieved by precipitation of Fe(ll) as FeCO , as well as by its
oxidation to Fe(OH).. In fact, Hale (7) effected satisfactory removal
of iron under anoxic conditions by the addition of hydrated lime in a
closed system to precipitate FeCO- along with
-------
3-6
For a groundwater previously in equilibrium with siderite (FeCO-,
aee section 2-2.2), aeration serves a dual purpose. In addition to
introducing oxygen for oxidation of Fe(ll), aeration allows dissolved
carbon dioxide, with which the groundwater is oversaturated, to escape.
Consequently, the pH of the system increases and the water becomes pro-
gressively oversaturated with respect to ferrous carbonate. If the de-
gree of oversaturation becomes such that the energy barrier to nucle-
ation is overcome, crystallization of FeCO- takes place and precipita-
tion follows.
In this respect, it should be of interest to measure the rate of
oxidation of Fe(ll) under the influence of oversaturated conditions
favoring precipitation of FeCO..
Experimental Procedure
A known gas mixture of carbon dioxide and oxygen was bubbled
through a series of flasks containing sodium bicarbonate, mounted on
magnetic stirrers. The CO - HCO ~ system buffers the solution with
respect to pH. After the attainment of equilibrium, observed by a con-
stancy in pH, various amounts of a stock solution of ferrous perchlorate
were added to the bicarbonate solutions. (The stock solution of Fe(II)
•/
had previously been equilibrated with the same gas mixture.) The con-
centration of Fe(ll) added was such that the resultant solution was over-
saturated with respect to ferrous carbonate. The degree of oversatura-
tion (S = Q/K, see section 2-2.2) was computed using the new solubility
product for ferrous carbonate determined experimentally in Chapter 2.
The rates of oxidation and removal were measured by analyzing the system
for total and filterable ferrous iron, respectively.
-------
3-7
For the determination of total Fe(ll), aliquots were withdrawn
from the system at various intervals, and immediately added to 2 ml.
of concentrated HC10, in order to quench the reaction. Solutions of
Fe(II) at concentrations greater than 10~ M were analyzed by titration
with standardized solutions of permanganate. For concentrations of
Fe(ll) less than 10 M, the colorimetric reagent bathophenanthroline was
used (8).
Filterable Fe(Il) was determined by immediately filtering aliquots
of the suspension through 220 mu filter paper (Millipore Filter Corpora-
tion, Bedford, Mass.) into 2 ml. of concentrated HCIO^, again to stop
the reaction. Filtration was rapid (less than 30 seconds for 50 ml. of
sample) and was conducted under a partial pressure of C(>2 of one atmo-
sphere in order to prevent additional oxidation and to avoid dissolution
of any suspended FeCO~. The filtrate was analyzed for Fe(ll) by the
same techniques as above.
pH and alkalinity were also measured during the course of the
reaction, the latter determined by acidimetric titration to pH 4.3
with standardized HC1.
Experimental Results and Discussion
In discussing oxidation and removal of ferrous iron under con-
ditions favoring precipitation of ferrous carbonate, previous workers
have tended to oversimplify the mathematical and chemical formulations
of the process. Ghosh, O'Connor, and Engelbrecht (4) combined precipi-
tation of ferrous iron and oxidation of ferrous iron, arriving at a
rate of iron removal which was first-order in concentration of Fe(ll).
-------
3-8
Conversely, in synthetic solutions having an initial oversaturation of
10 with respect to ferrous carbonate, Morgan and Birkner (9) observed
that precipitation and removal of Fe(ll) did not conform to first-order
kinetics. They noted an immediate rapid decrease in filterable Fe(ll)
in their supersaturated systems, corresponding to concurrent precipita-
tion of FeCO- and oxidation of Fe(ll). Subsequently, the rate of dis-
appearance of Fe(ll) was in exact conformance with the first-order
relationship observed in parallel studies conducted in the absence of
conditions of oversaturation. The latter description and the rate
constants reported were in agreement with those of Stumm and Lee (3).
In the experimental study of heterogeneous oxidation described
here, no such simple explanation was apparent. Figure 3-1 demonstrates
the compliance of both total Fe(ll) and filterable Fe(Il) to the first-
order formulation, despite the fact that the system was 60 times over-
saturated with respect to FeCOr The fact that only a slight increase
in total removal of Fe(Il) above that by oxidation alone is observed,
is indicative of little precipitation of the carbonate. (Precipitation
of FeCO is manifested by the difference between the two curves.) It
would appear that a period of 20 minutes was not sufficient to allow for
crystallization of FeCC>3 at pH 6.61. The rate of oxidation is seen to
agree with that predicted from the rate formulation by Stumm and Lee (3)
for equivalent conditions of pH, partial pressure of oxygen, and temper-
ature.
Correspondingly, at pH 6.25, where the rate of oxidation is
slower, Figure 3-2 shows precipitation of FeCC>3 to become significant,
but only after 40 minutes have elapsed. Again, the rate of oxidation
-------
3-9
10
8
-3
2
10
-4
Total ferrous iron
Filterable ferrous iron
Rate of
ferrous iron
oxidation, after
Stumm and Lee (3)
0.30 atm. 0
initial supersaturation
S ~ 60
o
10 15
TIME, minutes
FIGURE 3-1. Oxidation and removal of ferrous iron under condi-
tions favoring precipitation of ferrous carbonate.
10
8
6
-3
Total ferrous iron
Filterable
ferrous iron
Rate of
ferrous iron
oxidation^ after
Stumm and Lee (3)
0.50 atm. 0
40 60 80 100
TIME, minutes
FIGURE 3-2. Oxidation and removal of ferrous iron under condjL-
tions favoring precipitation of ferrous carbonate.
-------
3-10
of Fe(Il) parallels that reported by Stumra and Lee(3) and is apparently
unaffected by precipitation of FeCO,.
As the pH is lowered still further, the elapsed time exceeds the
induction time required for precipitation of FeCO- and the precipitate
seems to exert a catalytic effect on the rate of oxidation of Fe(ll)^,
(see Figure 3-3).
To demonstrate that this seemingly autocatalytic response in
both the rates of oxidation and removal of Fe(ll) is a. function of the
induction time for nucleation of ferrous carbonate, i.e., the time re-
quired to overcome the energy barrier preceding nucleation of the cry-
stalline phase, the studies at higher pH were repeated but the experi-
mental conditions were modified to decrease the rate of the oxidation
reaction. Figure 3-4 resembles Figure 3-3, again showing a rapid decay
in both total and filterable Fe(Il) after their conformance to the ac-
cepted first-order relationship for the first 60 minutes.
These studies reflect the complex nature of heterogeneous reac-
tions. For the case in question, the system consists of oxidation of
dissolved ferrous iron, precipitation of ferrous iron as the carbonate,
heterogeneous oxidation of solid ferrous carbonate, and possible surface-
catalysis of the oxidation by ferrous carbonate. To represent such a
system by a simple relationship would indeed be a mistake.
Although no conclusions can be drawn from this study, there are
some pertinent points worthy of consideration:
i) Supersaturation with respect to some solid phase does
not imply that precipitation takes place immediately. The magnitude of
the activation energy barrier to the nucleation process is inversely
-------
3-L1
Total Fe(II)
Filterable Fe(II)
Rate of Fe(II) oxidation,
after Sturara and Lee (3)
p H•- 5.84
0.50 atm. 0
100 150
TIME, minutes
200
250
FIGURE 3-3. Effect of FeCO^precipitation on Fe(II) oxidation and removal.
2
10
0 25 " 50 75 100 125
TIME, minutes
FIGURE 3-4. Effect of FeCO, precipitation on Fe(II) oxidation and removal.
-------
3-12
proportional to the supersaturation, i.e., the activation energy de-
creases as the supersaturation increases. Consequently, the rate of
nucleation is a function of the degree of supersaturation, there being
a critical supersaturation value below which nucleation is extremely
slow and above which nucleation is rapid. Therefore, the induction
time, i.e., the time required for formation of the critical-sized cluster,
decreases as the supersaturation increases (10).
ii) Precipitation of ferrous carbonate serves as a mechanism
for removal of Fe(ll), complementing removal by oxidation and hydrolysis.
The rate of removal, however, cannot be described by first-order kinetics.
iii) Ferrous carbonate appears to play a catalytic role in
the oxidation of Fe(ll). The mechanism for such an effect is uncertain
but could conceivably be attributed to a specific surface reaction where-
by solid ferrous carbonate provides active sites at which the concentra-
tion of Fe(Il) is greater than in bulk solution, or sites at which the
reaction is favored. Although it has not been demonstrated, one can
imagine that precipitation of ferrous carbonate could, under certain
circumstances, inhibit the oxidation of Fe(ll) by lowering the concen-
tration of free ferrous iron in solution, or by decreasing the available
Fe(ll) exposed to oxygen with the remainder being incorporated in the
interior lattice of the ferrous carbonate crystal. Such inhibition was
observed by Morgan (11) in his study of Mn(ll) oxygenation in the pre-
sence of precipitating manganous carbonate. In any case, even in the
event that oxidation is inhibited, the rate of removal of Fe(ll) either
by oxidation or precipitation should be equally as great in the presence
of FeCO_ supersaturation as in its absence.
-------
3-13
3-3 Oxygenation of Ferrous Iron in Acidic Systems
3-3.1 Experimental Study of Kinetics of Fe(ll) Oxidation at
Acidic pH-Values
The kinetic studies of the oxidation of ferrous iron reported
above were confined to waters of pH greater than 6. There are a number
of instances in nature, however, where iron bearing waters of pH con-
siderably below 6 are encountered. Of special concern are those waters
in coal mining regions, where pH-values of 3 are not uncommon (see
Chapter 5). It would be of interest to learn if iron oxidation in such
acidic systems could be characterized by the same kinetic relationships
which describe the reaction in neutral waters.
Experimental Procedure
The rate of oxidation of Fe(ll) was followed by measuring the
concentration of Fe(lll) with time and subtracting this quantity from
the initial concentration of Fe(ll). An analytical procedure was
adopted whereby [Fe(lll) ] (the total concentration of all species of
ferric iron) was determined spectrophotometrically at the isosbestic
point of an acidified solution where equilibrium had been established
+3 +2
between free ferric iron, Fe , and monohydroxo-ferric iron, FeOH ,
the only soluble species of Fe(lll) under the acidic conditions. At
the isosbestic point of the system, at a wavelength of 272 mu. both
species have identical molar absorptivities so that for a given total
concentration of Fe(lll), regardless of the ratio <." Fe to FeOH ,
the same absorbance is recorded, independent of pH (see Figure 3-5).
Figure 3-6 demonstrates conformance of the absorbance of acidified sclu-
-------
40
3-14
30
I
to
I
20
10
ISOSBESTIC POINT
250
FIGURE 3-5.
260 270
WAVELENGTH,
280
290 300
U-V absorbance spectra of acidified solutions of
ferric perchlorate.
o
CO
1.25
1.00
0.75
0.50
0.25
0.00
absorbance readings
in 1.0 cm. cell
at 272 ni
molar absorptivity
€ » 1.55 x 103 liter/mole-
cm
4 6 8
Fe(III) CONCENTRATION, xlO M
10
FIGURE 3-6. Relationship between absorbance of acidified solutions
of Fe(III) and Fe(III) concentration, at 272 nvi.
-------
3-15
tions of Fe(lll), at 272 mu, to the Beer-Lambert law. The molar ab-
sorptivity is 1.55 x 10 liter -mole" -cm" and is unaffected by the
presence of Fe(ll).
The samples were prepared containing various dilutions of a
stock solution of ferrous perchlorate. In order to determine the de-
pendence of the reaction rate on [Fe(ll)], the studies were performed
at constant pH and under constant partial pressures of oxygen. The in-
vestigations at slightly acidic pH-values were conducted in a CO^-HCO^
buffer system, as in section 3-2.2, in order to maintain constant pH.
At pH-values below 5, however, the buffer capacity of the bicarbonate
system is insufficient to balance the acidity produced by hydrolysis of
the resultant Fe(lll) (reaction 3-2), so that the pH of the system tends
to be drastically lowered. To combat this effect, smaller initial con-
centrations of Fe(ll) were employed (less than 10 M) to insure that,
for a given percent oxidation, the corresponding change in pH would be
slight.
At still lower pH-values, in the vicinity of pH 3, no precautions
were necessary since the pH was observed to remain relatively constant.
In this range, the pH was adjusted merely by dropwise addition of con-
centrated HCIO^.
The samples were allowed to equilibrate with the oxygen of the
atmosphere, or, in the case of the bicarbonate buffer systems, with the
gas mixture of oxygen and carbon dioxide. [Fe(IIl)] was determined by
acidifying an aliquot with dilute HC104 in order to dissolve any hydrous
ferric oxide formed and then measuring its absorbance at 272 mu, em-
ploying a Beckman Model DU Spectrophotometer. For the smaller range of
-------
3-16
concentrations, [Fe(ll)] was determined directly using the colorimetric
reagent bathophenanthroline (8). After each reading, the solutions were
re-equilibrated with their respective atmospheres. The samples were
sealed and stored in an incubator at 25°C. (In the lower pH-range, where
the rate of oxidation was observed to be slowest, a series of parallel
studies was conducted in the dark and exposed to light in order to test
for any photochemical effect on the rate of oxidation.)
Experimental Results and Discussion
In order to describe the rate of oxygenation of ferrous iron, an
expression similar to that of Stumm and Lee (3) was assumed, of the form
-dfre(ll)] = k [Fe(ll)]mCOH-]n Pn (3-6)
dt ««
where m and i\ are constants to be determined. In the rate law of Stumm
and Lee (3) at pH-values greater than 6, m = 1 and n = 2. Since the
studies were conducted at constant pH and constant partial pressure of
oxygen, equation 3-6 can be simplified to
= k' [Fe(ll)]m (3-7)
dt
where
k1 = k [OH"]11 P0 (3-8)
If the reaction were first-order in [Fe(Il)], i.e., if m = 1, then
-d log [Fe(lD] B k«/2>3 , k., (3-9)
and a plot of log [Fe(ll)] versus time should be linear.
-------
3-17
Figure 3-7 presents some of the results obtained in the bicar-
bonate-buffered system, demonstrating the obedience of the data to
equation 3-9. The concentrations of Fe(Il) have been carefully selected
so that the solubility of ferrous carbonate was not exceeded. Equation
3-9 demands that the slope of the semilog plot be independent of the
concentration of Fe(Il) at any time, so that parallel lines of slope k"
should result regardless of the initial concentration of Fe(ll). Figure
3-7 conforms to this requirement, too.
At the lower pH-values where the buffer capacity of the system
was low, the pH slowly declined as the reaction proceeded. The accom-
panying decrease in pH was smallest for the smallest initial concentra-
tion of Fe(ll), as planned. In these studies, the course of the reac-
tion was followed as long as the pH did not differ greatly from its
starting value. Figures 3-8 and 3-9 are plots of log [Fe(ll)] with time
and also show the corresponding change in pH. Conformance to the first-
order expression is demonstrated. When the studies were terminated,
approximately 25% of the initial concentration of Fe(ll) had been oxidized.
Figures 3-10 and 3-11 depict the rate of oxidation at pH 3 and pH
2, respectively. These results have also been fitted by a first-order
rate expression, but this has been done only for convenience and for the
sake of comparison both with the above results and with those of Stumm
and Lee (3). For a reaction proceeding as slowly as this one does in
the low pH-region, where only 57» of the reaction is complete in 150 days,
it becomes difficult to characterize the reaction with respect to its
-4
kinetic order. The "first-order rate constant" is approximately 1 x 10
day"1 for both figures, even though the concentration of OH differs by
-------
3-18
Temp. 25 C
0.85 atm.
25
50 75
TIMS, minutes
100
125
FIGURE 3-7. Rate of oxygenation of Fe(II) in bicarbonate-
buffered systems.
-------
-4.00
-A.30
-4.32 —
I
(d
O
H
O
s
-4.34 I—
-4.36 —
-4.38
-4.40 —
pH constant at 4.80
[Fe(II)J ~ 5x10 M
O :
-4.42 —
2 4
TIME/ days
FIGURE 3-8. Rate of oxygenation of Fe(II).
-4.02
-4.04
1
H
g -4.06
w
u
§
o
01
h
o
s
-4.08
-4.10
-4.12
-4.14
Temp. 25 C
PQ2 " °'90 atm'
pH 4*70— 4.45
[Fe(II)]o-v 10"4M
0 2 4
TIME, days
FIGURE 3-9. Rate of oxygenation of Fe(II)
u>
i
-------
3-20
pH 3.0
Temp.
P0 « 0.20 atm.
[Fe(II)l ~ 9x10 M
75 100
TIME, days
FIGURE 3-10. Oxygenation of Fe(II) at pH 3
125
150
-3.035
u
s
-3. 050
T
[Fe(II)lo -9x10 M
25
100
125
50 75
TIME, days
FIGURE 3-11. Rate of oxygenation of Fe(II) at pH 2,
150
-------
3-21
an order of magnitude between them. In fact, additional studies in this
acidic pH-range show the rate of oxidation to be relatively independent
of pH.
To gain some additional insight as to the order of the reaction
rate with respect to [Fe(ll)] in this pH-range, farther studies were
conducted at differing initial concentrations of Fe(Il). The parallel
slopes in Figures 3-12a through d also imply that the reaction is first-
order in [Fe(ll)], but this allegation is subject to the same reserva-
tions as above.
Because of the slowness of the reaction, the analytical procedure
consisting of the spectrophotometric measurement of Fe(lll) proved to be
the most effective. However, the technique is not amenable for use over
a wide range of initial concentrations of FeCll). At low concentrations,
the amount of Fe(lll) produced by the oxidation reaction is so small
that the measurements of absorbance become less precise. At higher ini-
tial concentrations of Fe(II), a sufficient amount of Fe(lII) forms that
even at pH 3, kinetically irreversible hydrolysis of Fe(IIl) takes place
(see Chapter 4) and it becomes increasingly more difficult to recover
all of the Fe(lll) as Fe+ or FeOH* . For this reason, the order of the
reaction with respect to Fe(ll) at this low pH could not be adequately
verified by simply varying Fe(ll) over a wide range of initial concentra-
tions. Figure 3-12 shows only a three-fold variation in initial con-
centration of Fe(ll). One can conclude from these results only that the
reaction is extremely slow and, in this range of concentration, can be
represented by a rate expression which is first-order in [Fe(ll)3.
-------
3-22
-3.113
pH 3.0
[Fe(lQ]o = 7.78xlO~ M
-3.109
1 L
-3.209 —
-3.207 —
-3.205
[Fe(II)] = 6.23x10
40 80
TIME, days
120
40 80
TIME, days
120
-3.414 _
— 07 PH 3.0
[Fe(II)] - 3.89x10 M
__ -3.639 _
-3.635 — —
-3.410
-3.631
40 80
TIME, days
120
40 80
TIME, uays
FIGURE 3-12. Oxygenation of ferroXis iron at various initial^oncentratior.s
of Fe(Il). Solution conditions are pH 3.0, 25 C, and partial
pressure of oxygen of 0.20 atm.
-------
3-23
George (12) studied the oxidation of Fe(ll) in perchloric acid
media and observed a rate law of the form
-d C!e+2] = k [Fe+2]2[0j (3-10)
at exp L
where the rate constant, k , increased slightly with a decrease in
[H+]. The relative insignificance of pH upon the rate (from pH 0 to
pH 2) is similar to that obtained in this study, but George has charac-
+2 2
terized the rate as being dependent upon [Fe ] .
Huffman and Davidson (13) have generalized from their own results
and those of others that the rate of oxidation is first-order in [Fe ]
in the presence of strong complex-formers, such as pyrophosphate, fluoride,
and dihydrogen phosphate. For complexing agents of moderate strength,
O *7
such as chloride and sulfate, the reaction rate depends upon CFe ] .
However, under extreme conditions of temperature and concentration of
ligand, the dependence is again first-order. The results of George (12)
in perchlorate media are analogous to those in the presence of moderate
complex-formers. Although sufficient data are not available, it seems
that, in the presence of suitable anions, the relative reaction rates
for the bimolecular mechanism (first-order in both [Fe ] and [00])
generally parallel the stability constants for association of the ligands
with Fe+3 (13).
-------
3-24
3-4 Oxygenation of Ferrous Iron as a Function of pH
3-4.1 Summary of Experimental Results
By coupling the experimental results obtained for acidic systems
with those obtained by Stumm and Lee (3) for neutral waters, one can
plot the rate of oxygenation of ferrous iron over the entire pH-range
of interest in natural waters, as in Figure 3-13. The rate of reaction
has been characterized by the rate constant, k" = -d log [Fe(ll)]/dt,
and has been adjusted for the conditions at 25 C and a partial pressure
of oxygen equal to 0.20 attn.
If one takes the logarithm of equation 3-8, substituting
k" = k'/2.3 (from 3-9) and making use of the ion product of water, one
obtains
log k" = log C + n pH (3-11)
where £ is a constant and n is the order of the reaction with respect
to [OH~]. It is readily apparent that the instantaneous slope of the
log k" versus pH curve in Figure 3-13 corresponds to n. The solid line
above pH 6 derives from the experimental rate law of Stumm and Lee (3)
(equation 3-3) for the given conditions, with n = 2. The dotted portion
below pH 6 is an extrapolation of their expression to the acidic pH-region
of this study; the rate diminishes by a factor of 100 for each unit de-
crease in pH. The experimental points are compatible with the formula-
tion of Stumm and Lee at pH-values greater than 4.5, but at lower pH-
values, the points sytematically deviate from the extrapolated line. At
pH-values below 3, the rate becomes relatively constant and is no longer
dependent upon pH, i.e., n = 0.
-------
3-25
+3.0
+2.0
+1.0
fr-
•o
0.0
-1.0
-2.0
-3.0
-4.0
-5.0
-6.0
i r
k" " " d
dt
PQ "0.20 atm.
Temp. 25 C
Experimental points
obtained in this study
O exposed to light
D in darkness
o o
"D
Extrapolation of rate law
/ of Stutran and Lee (3) at
/ 25 C and 0.20 atm. of oxygen
J L
1 2 3 4 5 67
pH
FIGURE 3-13. Oxygenation rate of ferrous iron as a function
of pll.
-------
3-26
With regard to the relative magnitude of George's results (12) in
comparison with those presented in Figure 3-13, one can approximate his
rate expression (equation 3-10) by a pseudo-first-order formulation
= k' [Fe+2] (3-12)
dt exp
where
k' .k [Fe+2] P. (3-13)
exp exp o 0^
[Fe+2] is essentially constant for only 1% total reaction (the extent of
the reaction followed by George) and is equal to [Fe ] , the initial
concentration of ferrous iron. This approximation is not a mechanistic
one but has been made solely for the purpose of comparison. Under a
partial pressure of oxygen of 0.2 atm. and at 30 C and 10 n HC10,,
George's results predict that the initial value of k" (k" =» -d log
-4 -1
[Fe(H)]/dt as in Figure 3-13) would be approximately 1 x 10 day if
£Fe J = 10~T1, the concentration used by George. This pseudo-first-
order "rate constant" is of the same order of magnitude as that ob-
tained in this study.
The shape of the curve suggests that there are two parallel re-
action mechanisms; one operative at higher pH which can be described by
the rate law of Stumm and Lee, and the other functioning under more
acidic conditions and independent of pH. Previous investigations of
the kinetics of oxygenation of Fe(ll) in acidic media are in accord with
these latter results in that the rate of the reaction is relatively in-
dependent of pH. In fact, George (12) observed that his k (see
equation 3-10) was proportional to [H ]~ , increasing only slightly
-------
3-27
with an increase in pH. Alternatively, one can consider the reaction
to be independent of pH until the concentration of OH becomes suffici-
ently large that OH* functions as an effective catalyst of the oxidation
reaction.
Figure 3-13 also demonstrates that the oxidation of Fe(ll) occurs
more rapidly in light than in darkness; the reaction proceeds at a rate
2 to 3 times faster in the presence of light. In the acidic region,
there appears to be some photooxidation of Fe(Il) taking place.
In further studies of the oxidation of Fe(ll) under simulated
mine conditions (see Chapter 5), the investigations were conducted in
the absence of light to avoid such photochemical effects.
3-4.2 Kinetic Implications of Results
Consider first the oxidation of Fe(ll) in the acidic region
where the reaction proceeds at a rate independent of pH. Weiss (14)
proposed a chain mechanism involving one-electron oxidations in order
to describe the oxidation of ferrous iron by molecular oxygen. The
suggested sequence is:
Fe+2 + 02 = Fe+3 + 0^ (k^ k^) (3-14a,b)
) (3-15)
Fe+2 + H02 = Fe+3 + HO^ (k^1) (3-16a,b)
HV * H+ . H202 CK^) (3-17)
Fe+2 + H20 -> Fe+3 + OH~ + OH (^3) (3-18)
-------
3-28
Fe+2 + OH •» Fe+3 + OH" (k (3-19)
The molecules (*) are free radicals or reactive intermediates. Reaction
3-14a is believed to be the rate-determining step in the sequence. The
rate of oxidation of Fe can be derived using steady-state approxima-
* •
tions (15) for H02, OH, and HO- to give (16)
"d ® 3 = k [Fe+2] [0] R (3-20)
where
k2 [Fe+2] [H+;
[Fe+2] [H+] + k K [Fe+3]
(3-20a)
Equation 3-20a implies that the oxidation reaction is inhibited by Fe+
due to the back reaction 3-14b. This accounts for the slowness of the
reaction in acidic solutions where the resultant Fe(lll) is present
predominantly as Fe . If conditions are such that
[Fe+3] \» k [H+] [Fe+2] (3-20b)
then R, ^ 1, and the reaction is decelerated because of the relatively
rapid reduction of Fe by 0 ~ in comparison to the oxidation of Fe
• *
by 02~ or H02 (equation 3-16a).
If, on the other hand, anions are present which are capable of
forming strong complexes with Fe , such complex-formation serves to
decrease the concentration of free Fe+ and thus inhibit the back reac-
tion (3-14b). The net effect is to cause the oxidation to proceed more
rapidly since
V KHQ CFe+] <
-------
3-29
and R~l. Hence, by equation 3-20
~* C^+2] - k. [Fe+2] [0,] (3-20d)
at l L
One serious drawback of the Weiss mechanism is its unlikelihood
from a coulombic standpoint. Zwolinski, Marcus, and Eyring (17) termed
the formation of oppositely-charged end-products, as in 3-l4a, as
highly improbable. In a later paper, Weiss (18) modified his mechanism
in accordance with such reasoning to consider that the initial reaction,
3-14a, should be the formation of an ion-pair complex
Fe+2 + 02 = (Fe+3'02~) (3-21)
stabilized by coulombic attraction between the oppositely-charged part-
ners. In this scheme, the association does not violate the coulombic
restriction imposed by Zwolinski, et al. Again, the back reaction can
be inhibited by suitable anions:
(Fe+3-02~) + X" * (X-'Fe+3-02-) (3-22)
where the resultant complex may eventually dissociate
(X--Fe+3-02-) » (X--Fe+3) + 0^ (3-23)
The original Weiss scheme continues with reaction 3-15. Under these
conditions, the rate is proportional to [Fe ], [C^], and [X ], and
the anionic complex-former has served in the same canacity as in the
original scheme.
In the absence of strong complex-formers, the (Fe '0^ ) com-
plex can be stabilized by Fe (18),
-------
3-30
J. -x +2 , -j-3,_ -.„ +2-V /•„ +3._ -2._ +3.. f. _.^
O ) + Fe = (Fe 0? Fe ) = (Fe (>„ Fe ) (3-24)
This would explain the results of George (12) and Huffman and Davidson
+2 2
(13), where the observed rate is proportional to [0^] and [Fe 3 .
The new complex is again stabilized by coulombic forces and eventually
breaks up,
(Fe+3'02"2'Fe+3) + H+ •» 2 Fe*3 + HC>2' (3-25)
followed again by the same sequence as above.
If this were the mechanism describing the oxygenation of Fe(II)
and the effect of anionic complex-formers on the rate of reaction, then
one should observe a decrease in rate with increasing concentration of
Fe , as in 3-20. None of the previous workers, however, have observed
any inhibitory effects by the addition of Fe* to their acidic solutions.
The situation can be somewhat clarified by closer scrutinization
of 3-20a and b. If the back reaction were relatively rapid and Fe
were rapidly reduced by 0^ , then
V KHO CFe] ^ k2 CH] CFe] C3-20b)
A quantitative comparison of the two terms is called for. For the given
experimental conditions ([H+J = 10~3M, [Fe+Z] = 10"3M, and [Fe+3J~10"5M)
and using the value approximated by Benson (19) that pK,-o =12-4, one
obtains k ' /k- ^ 1011. Hence, if k^ /k. <^ 10U, the back reaction is
sufficiently slow that it can be neglected in 3-20a. Barb, et al (20)
experimentally measured k.. /k« = 1.0 at pH 2.7 in perchlorate media.
This implies that reduction of Fe by 0_ is signficantly slower than
the corresponding oxidation of Fe by 0_ . Recent experimental evidence
-------
3-31
indicates that the value for pK^.- estimated by Benson is too high and
*°2
that the proper value should be about pK.—. = 5 (21). Even if this
value is utilized, the reverse reaction under these conditions can still
be neglected. Therefore, if the Weiss mechanism is valid, one should
observe the rate relationship 3-20d, i.e., a reaction rate which is
first-order in both [Fe ] and [0_]. The experimental data presented
in Figure 3-13 do obey such a kinetic law in the acidic region.
Another interesting observation with regard to the Weiss mechanism
is the hypothesis that the Fe *()„ complex forms first and is subse-
quently stabilized by anionic ligands. In view of the rapid nature of
simple complex-formation reactions (22), it is more likely that complex-
ation of Fe by the ligand occurs first, followed by the reaction with
oxygen and, in the termolecular mechanism (13), by another Fe . Cher
and Davidson(23) have considered that complex-formation serves to make
AH and AF for reaction 3-14a less positive, with the net effect being
to lower the activation energy of the reaction.
The possibility of a 2-equivalent electron-transfer has been pro-
posed, largely as a result of the work of Cahill and Taube (24), who
• •
postulated Fe(lV) to be a reaction intermediate in place of HO. in the
single-electron transfer. Such an interpretation would be in accord with
Fe(ll) + 02 = Fe(lV) + H^ (3-26)
Fe(il) + Fe(lV) = 2 FeClII) (3-27)
Fe(ll) + H202= Fe(iv) + 2 OH~ (3-28)
Fe(ll) + Fe(lV) = 2 Fe(lll) (3-29)
-------
3-32
The role of anionic complexes with the intermediate Fe(IV) would be
similar to that observed for Fe(lll) in the single-electron-transfer
mechanism (23).
Conocchioli, et al (25) have begun to investigate the oxidation
of Fe(ll) by 2-equivalent oxidants, their preliminary data corroborating
•
the proposal of Cahill and Taube with regard to the involvement of Fe(IV)
as an intermediate. Their observation of the Fe(lll) dimer as the
primary end-product led to their formulation of the mechanism as
„ 2 equiv.
Fe + Ox. - > Fe(lV) + Red2 (3-30)
Fe(lV) + Fe+2 raPid> [Fe(lll)] (3-31)
where [Fe(lll)32 refers to the dimer, represented by FeFe . In
studies employing HOC1 as the 2-equivalent oxidant, the dimer formed was
observed to be TeFe+ , which slowly converted to the dihydroxo-dimer
Cl
under acidic conditions. In a similar vein, it will be reported in the
next chapter that the same inorganic ligands which accelerate the rate of
oxidation of Fe(ll) by apparently stabilizing the transitory intermediate
also accelerate the rate of hydrolysis of Fe , which presumably proceeds
through the dimer.
The formation of a bridged activated complex, as proposed by
Taube (26) to explain the role of complex formers in serving as a bridge
between the oxidizing and reducing agents, is unlikely in this situation;
its major relevance is in redox reactions between metal ions.
If Figure 3-13 is considered, it is readily apparent that, in the
acidic pH-range, the oxidation reaction proceeds relatively slowly, as
-------
3-33
predicted by the Weiss mechanism. Whether or not this type of mechanism
is valid, one could still postulate that as the pH is increased and [OH~]
Increases, the net effect is that the ligand OH~ behaves in the same
fashion as the other complexing ligands by coordinating with one of the
iron species and stabilizing the transitory complex. If this were the
case, then one should not observe other anionic ligands to exert a
catalytic effect at the elevated pH-values. Stumra and Lee (3), in a
preliminary survey, noted that chloride and sulfate exerted no such ac-
celerative effect at pH-values greater than 6, but H2PO ~ did. Forma-
tion of mixed hydroxo-ligand complexes of Fe(lll) could account for such
an observation.
At the higher pH-values investigated, care must be exercised in
interpreting the results because of the heterogeneity of the system.
The resultant Fe(lll) hydrolyzes quite rapidly (see Chapter 4), forming
insoluble ferric hydroxide. Although it can be assumed that the cata-
lytic effect of OH at higher pH-values is similar to that suggested
earlier, one must bear in mind that the previously-mentioned mechanisms
have all been derived for homogeneous, one-phase systems.
Abel (27) has proposed that the second-order dependence on [OH~],
exhibited in Figure 3-13, arises from the following scheme:
02 -f OH" = 02'OH" (3-32)
02'OH~ + OH" = 03~2 + H20 (3-33)
_2
The two equilibria proceed rapidly with the resultant 0., slowly re-
acting with Fe to form additional reactive oxo-complexes which propa-
gate the chain.
-------
3-34
Wells and Salam (28) have attached special significance to
electrostatic considerations in that it is easier to remove an elec-
tron (e~) from Fe+2 by decreasing its positive charge through complex-
formation.
The second-order dependence on [OH"] is reminiscent of the
second-order dependence on [H2P04~] observed by Cher and Davidson (23)
in their study of the oxidation of Fe(ll) in phosphoric acid solution.
An explanation was not given for this second-order dependence, but in
S
a later paper King and Davidson (29) proved that it was not due to
-2
condensation of H2P°4 to H2P2°7 *
The dependence upon [OH"] and the conversion of the solution to
a two-phase system due to hydrolysis of Fe(IIl) is suggestive of a
kinetic dependence on hydrolyzed species of Fe(lll). It is interesting
to speculate with regard to such hydroxo-species, especially since the
transition region observed in Figure 3-13 occurs near the pK-value for
the reaction
FeOH+2 + H20 = Fe(OH)2+ + H+ pK = 4.6 (3-34)
Beyond pH 4.6, Fe(OH) * becomes the dominant soluble species of Fe(lll).
The correspondence between the second-order hydroxide-dependence of the
rate of oxidation of Fe(ll) and the dihydroxo-compos it ion of Fe(IIl) is
striking. However, there is no evidence to indicate that an autocatalytic
mechanism is operative, as in the oxygenation of Mn( II) (10); the addi-
tion of Fe(lll) has no apparent effect on the rate of oxidation of Fe(ll).
It is also unlikely that hydroxo-species of Fe(ll) are involved, since
hydrolysis of Fe(ll) does not become significant until pH 6.5.
-------
3-35
Consequently, with regard to Figure 3-13, one can only emphasize the
significance of pH on the rate of oxidation of Fe(II) at. pH-values
greater than 4.5. At this time, there is insufficient chemical evi-
dence on which to base any conclusive mechanistic interpretation.
-------
3-36
References
1) Just, G., "Kinetische Untersuchung der Autoxydation des in Wasser
gelosten Ferrobicarbonats," Z. Phys. Chem., 63, 385 (1908)
2) Holluta, J., and Eberhardt, M., "Uber geschlossene Enteisenung
durch Schnellfiltration," Vom Wasser. XXIV, 79 (1957)
3) Stumn, W., and Lee, G. F., "Oxygenation of Ferrous Iron," Ind. Eng.
Chem., 53_, 143 (1961)
4) Ghosh, M. M., O'Connor, J. T. , and Engelbrecht, R. S., "Precipita-
tion of Iron in Aerated Groundwaters,* Journ. San. Eng. Div. ,
Proc. Amer. Soc. Civil Eng., 92, 120 (1966)
5) Stumm, W., and Singer, P. C., "Precipitation of Iron in Aerated
Groundwaters," discussion, Journ. San. Eng. Div., Proc. Amer.
Soc. Civil Eng., 92. 120 (1966)
6) Morgan, J. J., and Stumm, W. , "The Role of Multivalent Metal Oxides
in Limnological Transformations, as Exemplified by Iron and
Manganese," Proc. Second Intl. Conf. Water Poll. Res., Tokyo,
p. 103 (19641'
7) Hale, F. E. , "Iron Removal Without Aeration - The Precipitation of
Ferrous Carbonate in a Closed System," 3.^Amer. Wat. Works Assn.
28, 1577 (1936)
8) Lee, G. F., and Stumm, W., "Determination of Ferrous Iron in the
Presence of Ferric Iron," J. Amer. Wat. Works Assn., 52, 1567
(1960) "
9) Morgan, J. J., and Birkner, F. B. , "Precipitation of Iron in Aerated
Groundwaters," discussion, Journ. San. Eng. Div., Proc. Amer. Soc.
Civil Eng., 9£, 137 (1966)
10) Walton, A. G., The Formation and Properties of Precipitates, Inter-
science, New York (1967)
11) Morgan, J. J., "Chemistry of Aqueous Manganese (II) and(lV)," Ph.D.
thesis, Harvard University (1964)
12) George, P., "The Oxidation of Ferrous Perchlorate by Molecular
Oxygen," J. Chem. Soc., 4349 (1954)
13) Huffman, R. E., and Davidson, N., "Kinetics of the Ferrous Iron -
Oxygen Reaction in Sulfuric Acid Solution," J. Amer. Chem. Soc.,
28_, 4836 (1956) '
-------
3-37
14) Weiss, J., "Elektronenubergangsprozesse im Mechanismus von Oxydations-
und Reduktions- Reaktionen in Losungen," Naturwissenschaften, 23^,
64 (1935)
15) Frost, A. A., and Pearson, R. G., Kinetics and Mechanism, John Wiley
and Sons, New York (1962)
16) Weiss, J., "Electron Transfer Reactions in the Mechanism of Oxida-
tion-Reduction Processes in Solution," J. Ghim. Phys., 48, C-6
(1951)
17) Zwolinski, B. J., Marcus, R. J., and Eyring, H., "Inorganic Oxidation-
Reduction Reactions in Solution," Chem. Rev., 55, 157 (1955)
18) Weiss, J., "The Autoxidation of Ferrous Ions in Aqueous Solution,"
Experentia, DC, 61 (1953)
19) Benson, S. W., The Foundations of Chemical Kinetics, McGraw-Hill
Book Co., New York (1960)
20) Barb, W. G., Baxendale, J. H. , George, P., and Hargrave, K. R. ,
"Reactions of Ferrous and Ferric Ions with Hydrogen Peroxide,"
Trans. Farad. Soc., 47_, 462 (1951)
21) Morris, J. C., Harvard University, Personal Communication
22) Wendt, H., and Strehlow, H. , "Schnelle lonenreaktionen in Losungen.
II. Die Bildung Einiger Einfacher Komplexe des Eisen-III- Ions,"
Z. Elektrochem, 66, 228 (1962)
23) Cher, M., and Davidson, N., "The Kinetics of the Oxygenation of
Ferrous Iron in Phosphoric Acid Solution," J. Amer. Chem. Soc.,
7T_, 793 (1955)
24) Cahill, A. E., and Taube, H., "The Use of Heavy Oxygen in the Study
of Reactions of Hydrogen Peroxide," J. Amer. Chem. Soc., 74,
2312 (1952)
25) Conocchioli, T. J., Hamilton, E. J., and Sutin, N. , "The Formation
of Fe(lV) in the Oxidation of Iron (II)," J. Amer. Chem. Soc.,
87, 926 (1965)
26) Taube, H., "Mechanisms of Redox Reactions of Simple Chemistry,"
Advances in Inorganic Chemistry and Radiochemistry,1, 1, Emelius,
H. J., and Sharp, A. G., editors, Academic Press, New York (1959)
27) Abel, E., "Uber Autoxydation in Umbelichteter Homogener Wasseriger
Losung. Mit Besonderer Berucksichtigung Anorganischer Systeme,"
Z. Elektrochsm., 59, 903 (1955)
-------
3-38
28) Wells, C. F., and Salam, M. A., "A Kinetic Approach to the Nature
of Ferrous Ions in Aqueous Solution," Nature, 203, 751 (1964)
29) King, J., and Davidson, N., "Kinetics of the Ferrous Iron - Oxygen
Reaction in Acidic Phosphate - Pyrophosphate Solutions," J. _Ai?.er
Chem. Soc., 80_, 1542 (1958)
-------
CHAPTER 4
HYDROLYSIS OF FERRIC IRON
4-1 Introduction
The solubility of ferric iron in natural waters, under most con-
ditions, is controlled by the solubility of its various oxides and hy-
droxides. Figure 2-13 demonstrated that the concentration of soluble
Fe(lll) is less than 10" M at pH 4. When the solubility product of
ferric hydroxide is exceeded, a series of hydrolytic reactions takes
place as the formation of insoluble ferric hydroxide proceeds through
multiraeric and polymeric hydroxo-intermediates. These kinetic inter-
mediates tend to be adsorbed at interfaces, thus accounting for the use
of Fe(lll) as a coagulant in water treatment.
It was also shown (section 2-3.2) that various ligands tend to
coordinate with Fe(IIl) and that the degree of coordination is a func-
tion of the relative affinity of Fe(lll) for these various ligands ver-
sus its affinity for OH~. The existence of mixed ligand-hydroxo-com-
plexes was assumed to be of relevance in natural waters.
This chapter centers upon the kinetics of hydrolysis of Fe(lII)
in systems oversaturated with respect to ferric hydroxide. The effect
of sulfate on the kinetics of hydrolysis was investigated, especially
due to the high concentrations of sulfate found in mine drainage waters.
Sulfate also serves as a representative ligand in order to gain some in-
sight as to the rate of hydrolysis of Fe(lII) in systems containing
-------
4-2
ligands which compete with OH~ for the coordination sites of Fe(lll).
The coagulative properties of ferric iron are considered, in brief, to
demonstrate that it is the hydrolytic intermediates which are respon-
sible for the destabilization of colloidal dispersions. Finally,
phosphate removal by ferric iron has been investigated, both by the
direct addition of Fe(lII) to a system containing phosphate and by the
addition of Fe(ll) which is subsequently oxidized, in situ, to Fe(lII).
4-2 Kinetics of Ferric Iron Hydrolysis
4-2.1 Reactions of Fe* with Water
Ferric iron in aqueous solution behaves as a multiprotic
BrBnsted acid, with protons being transferred from the coordinated
water molecules of Fe(lll) to the solvent water in the following step-
wise manner:
Fe(H00)/3 + H00 = Fe(H_0),OH+2 + H,0+ (4-1)
L D c. L J J
Fe(H20)5OH+2 + H20 = Fed^O^OK^ + E^ (4-2)
etc.
Such reactions can proceed until all of the coordinated water molecules
have been deprotonated, resulting in the formation of anionic hydroxo-
ferric species. In addition to these "aquo-acidity" reactions (l)
with mononuclear products, these simple hydroxo-ferric complexes tend
to polymerize by a condensation process,
2 Fe(H00),OH+2 = Fe0(H,0)Q(OH).+4 + 2 H_0 (4-3)
L D L L o / /
-------
4-3
where water is essentially squeezed out of the coordination shell. The
resultant dimer has the structural configuration
(H20)4Fe<^ ^>Fe(H20)4+4 (4-3a)
The dimer is subject to additional hydrolytic reactions, again involv-
ing a proton-transfer
Fe_(H~0)R(OH) + + H-0 = Fe7(HJD)7(OHK + H,0+ (4-4)
or additional condensation reactions by which the resultant molecule
is further dehydrated
2 Fe_(H.O)7(OH).+3 = Fe.(H_0), _(OH),+6 + 2 H^O (4-5)
Lcl j "t L IZ b 2
In systems oversaturated with respect to insoluble ferric hy-
droxide, a series of such hydrolytic and condensation reactions takes
place, the multimeric and polymeric hydroxo-species serving as kinetic
intermediates in the formation of the solid ferric hydroxide precipi-
tate. (A useful review of the chemistry of metal ions in aqueous
solution has recently been presented by Stumm (2).) The solid phase,
however, continues to react as the amorphous freshly-precipitated hy-
droxide generally converts to o(-FeOOH.. Under special conditions,
cX-Fe-0 $ -FeOOH, and 5-FeOOH can be formed (3). In addition to
L j
these crystalline products, an inactive amorphous material is always
present, even after prolonged periods of aging. Under certain condi-
tions, a colloidal sol of ferric-oxide hydroxide can be maintained for
long periods of time. As previously mentioned (section 2-3.2), it is
-------
4-4
analytically very difficult to distinguish between dissolved and sus-
pended ferric iron.
Since the coagulative properties of Fe(lll) arise from the ten-
dency of the hydrolyzed polynuclear kinetic intermediates to be adsorbed
at interfaces (4), and since Fe(OH)_ and its hydroxo-interraediates are
responsible, in part, for the transport of phosphate and organic materi-
al in lakes (5), it would be desirable to elucidate the kinetic rela-
tionships which govern the hydrolysis of ferric iron.
4-2.2 Experimental Study of the Kinetics of Fe(lll) Hydrolysis
Experimental Procedure
The kinetics of hydrolysis of free ferric iron, Fe+ , was studied
potentiometrically, employing the ferro-ferri cell
Pt/ Fe(ll),Fe(lIl),Na+,H+,C10.~/ NaCl / Hg.Cl./ Hg (4-6)
4 (sat'd) 2 2
The cell has previously been described in section 2-3.3 and was shown
in Figure 2-15. The reversible potential is established by the elec-
troactive ferrous-ferric couple in accordance with the Nernst Equation
which, for a constant ionic medium (0.1 M NaCIO,) in a constant temper-
ature water bath at 25 C, becomes
E = E°' -0.0592 log [Fe ,] (4-7)
[Fe+3]
E was determined in a preliminary experiment by measuring the poten-
tial of a known system at pH 1.0, where no hydrolysis or oxidation took
place, so that [Fe+3] = [Fe(lIl)]T and [Fe+2] = [Fe(ll)]T. [Fe(ll)]T
and [Fe(lll)] were determined using bathophenanthroline (6).
-------
4-5
Solutions containing various concentrations of Fe(ll), Fe(lII),
and H were prepared, under a nitrogen atmosphere, at the same ionic
strength and temperature as the standard system. It was noted that the
addition of even a weak alkali (dilute NaHCO-) to raise the pH of the
system brought about kinetically irreversible hydrolysis. Apparently,
localized conditions of high pH (high concentrations of OH and a cor-
t
responding large degree of oversaturation with respect to ferric hydrox-
ide) were produced in the vicinity of the tip of the burette from which
the base was slowly dispensed. Even under conditions of intense mixing,
basification of the system containing Fe(ll) and Fe(lll) resulted in a
sharp decline in potential. Therefore, solutions containing Fe(lll)
had to be maintained under conditions such that the solubility of fer-
ric hydroxide was never exceeded until the time at which the kinetic
studies were begun. In most cases, Fe(IIl) was added directly to the
system as the perchlorate salt, at time zero. In those instances where
it was desired to begin the study at a higher pH, solutions of Fe(ll)
were basified with dilute NaHCO , under a nitrogen atmosphere, and then
ozonated (Sanders Ozonator Model No. S-V-106, Triton Aquatics) in order
to generate ferric iron in situ (7). Ozone was bubbled through the
system and the extent of oxidation of Fe(ll) was followed by noting the
increase in potential of the ferrous-ferric couple. Generally, one
minute was sufficient to oxidize about half of the initial concentra-
tion of Fe(ll), so that [Fe(Il)]T ^ [Fe(lIl)]T. Any residual traces of
ozone or oxygen were removed by the nitrogen which was continuously bub-
bled through the solution. [Fe(ll)]T was determined before and after
-------
4-6
ozonation, the difference being the total concentration of Fe(lll)
which was generated.
The solutions were stored under an atmosphere of nitrogen and
the pH was kept sufficiently low so that no oxidation of Fe(ll) took
place, and [Fe"1" ] = [Fe(II)3_, throughout the duration of the experi-
ment. [Fe ], the concentration of unhydrolyzed iron (III) species at
any time, was calculated from the Nernst Equation (4-7) after measur-
ing the potential of the system to the nearest 0.1 mv., using a Model
D Sargent Recording Titrator. The liquid junction potential, E., has
been shown (7) to be strictly a function of [H ]. Extrapolating the
results of Biedermann (7), E. remains relatively constant during the
course of the experiment so that variations in E. can be neglected.
The concentration pH (p H) was measured using a Leeds and Northrup Po-
tentiometer (Catalog No. 7664), calibrated in the same manner as pre-
viously described in section 2-2.3. The samples were analyzed for
[Fe(ll)] and [Fedll)]™ using bathophenanthroline.
Experimental Results and Discussion
Due to the acidic nature of the aquo-ferric ion, the pH of the
system decreases when ferric salts are placed in solution. The simple
proton-transfer reactions by which the mononuclear hydroxo-complexes
are formed (reactions 4-1 and 4-2) take place extremely rapidly (8).
When the solubility product of Fe(OH), is exceeded, the slower hydro-
lytic and condensation reactions proceed whereby the multinucelar and
polynuclear species and, eventually, solid ferric hydroxide are formed.
One would expect that the rate of disappearance of free Fe would be
dependent not only upon the concentration of Fe , but also upon pH or,
-------
4-7
more precisely, upon the concentration of OH . Since hydrolysis tends
to decrease [OH 3, the reaction would be expected to decelerate as
hydrolysis progresses. Furthermore, the resultant decline in pH also
causes a shift in the Fe - FeOH distribution so that the concen-
r\
tration of Fe is partially buffered; on one hand, it is diminished
in forming ferric hydroxide, and on the other hand, it is augmented
O *5
due to conversion of FeOH to Fe , i.e., the back reaction of 4-1.
This effect can be pictured as
Fe\
H+X (4-8)
FeOH+2 Fe (OH)
x y
Therefore, in order to derive any useful kinetic information, the system
must be maintained at constant pH. Use of a pH-buffer is precluded be-
cause of possible kinetic effects the buffering agent might bring about.
Alternatively, small concentrations of Fe can be employed in order to
minimize the pH-decrease accompanying hydrolysis, or one can follow
[Fe 3 with time and compute the instantaneous rate of decrease of
[Fe 3 at any given pH or in a region of constant pH. These latter
two approaches were utilized in treating the experimental data. In
•\
this manner, any change in the concentration of Fe arises only from
hydrolytic reactions.
Figures 4-1 and 4-2 show some of the experimental results. In
each cases the data have been plotted assuming either a first- or
second-order dependence of the reaction rate on [jFe 3. Rate laws
of the form
-------
3.81 3.67 3.64
P'H
s
§
o ,.
0 10'6
Q)
9
8
7
6 I—
3.63
0
FIGURE 4-1.
(a)
Temp. 25 C
constant ionic medium
I • 0.1 M NaClO,
20 40
TIME, minutes
60
3.81 3.67 3.64
PCH
3.63
20 40 60
TIME, minutes
+3
(a) Logarithmic and (b) reciprocal plots of the rate of hydrolysis of FeH
(a) assumes a first-ord|r dependence and (b) a second-order dependence bf the
reaction rate on the Fe*J concentration. epenaence or the
00
-------
PCH
3.52 3.47
DS
o
1-t
X
•\
£3
I
55
w
0)
3.44
3.41
3.52 3.47
PCH
3.44
3.41
150
200
0 50 100 150 200 0 50 100
TIME, minutes TIME, minutes ^
FIGURE 4-2. (a) Logarithmic and (b) reciprocal plots of rate of hydrolysis of the aquo-ferric ion, Fe . ^
-------
4-10
pe+3] = v [Fe+3] (4-9a)
dt 1
+3-2 (4.9b)
=
dt t.
have been assumed, where ^ and k_ are functions of pH. The measured
p°H-values are noted. Kinetic significance of [OH ] is implied by
the decreasing rate of hydrolysis for both the logarithmic and recipro-
cal plots. If the linear regions of each of the respective logarith-
mic and reciprocal plots are considered, the instantaneous rate of hy-
drolysis can be calculated for the pCH-range over which such linearity
is observed to hold. For each sample, the latter portion of the curve
is relatively linear for it is in this region that the p H remains es-
sentially constant.
The instantaneous slopes of the logarithmic and reciprocal plots
are equal to the "rate constants," k. and k_, for the first- and second-
order relationships, respectively. If k^ and k^ are assumed to be
functions of pH of the form
kL . kx' [OH~]n (4-lOa)
k2 = k2' [OH']n (4-10b)
then a plot of log \a, or log k« versus p H should result in a straight
line of slope n. Figures 4-3 and 4-4 are plots of the instantaneous
slopes of the logarithmic and reciprocal plots (Figures 4-1 and 4-2)
against p^. A linear regression analysis was performed to compute
the best straight line to fit the experimental rate results. The cor-
relation coefficients for Figures 4-3 and 4-4 are 0.76 and 0.97, re-
spectively, indicating that the data conform to a second-order
-------
-1.0
-2.0 __
.*
9
-3.0
-4.0
2.75
FIGURE 4-3.
3.00
3.25 3.50
P H
3.75
4.0 —
I
•4
I
-------
4-12
dependence on EFe ] considerably better than to a first-order depen-
dence. The slope of Figure 4-4 is 3.7, implying that the rate of hy-
drolysis for the conditions investigated can be reasonably described
by a kinetic relationship of the form
= k ' [Fe+3]2 [OH']4 (4-11)
where the rate constant k«' is approximately 10 * mole liter rain
for the p°H range 2.8 to 3.8.
The data were also tested for conformance to the von Weimann
formulation, whereby the rate of precipitation (hydrolysis) is propor-
tional to the relative degree of over saturation, i.e.,
"d ffe* ] = k (Q-K)/K (4-12)
at
where Q is the reaction quotient and K is the equilibrium solubility
product for the formation of ferric hydroxide. The results might be
interpreted in terms of the kinetics of crystal growth but the compu-
tations were limited by the lack of a dependable solubility product
—38
for Fe(OH),. A value of K =10 was employed for the calculation
J SO
(see below), but such treatment of the data met with little success.
It should be emphasized that equation 4-11 is merely a data-
fitting relationship. It is not intended to serve in any specific
mechanistic way since there are probably numerous other rate expressions
which could fit the experimental data. Furthermore, equation 4-11 is
applicable only for the experimental conditions investigated; extra-
polation of this relationship to other situations might result in
serious error.
-------
4-13
Wendt observed (9) that the dunerization reaction (4-3) is con-
siderably slower than the simple proton-transfer reaction and that the
former is accelerated in the presence of simple anions via a reaction
of the forn>
OH
2 , Fe Fe(5-n) (4-13)
Figure 4-5 contains plots cf the second-order rate constant for the
"3
disappearance of free Fe+ in the presence of sulfate. The figure was
developed by treating the experimental data in the same fashion as de-
scribed above. For the three different concentrations of sulfate
analyzed, th- average slope is 3.3. The rate of hydrolysis is more
rapid in the presence of sulfate than in its absence, as demonstrated
in Figure 4-6. These results are especially useful when applied to the
mine drainage system considered in Chapter 5.
One could generalize, on the basis of Wendt' s findings and those
reported here, that anions which coordinate with Fe tend to increase
the rate of hydrolysis of Fe(lll). Hence, hydrolysis would be ex-
pected to occur even more rapidly in natural waters under the influ-
ence of the many natural anioas which are present, such as sulfate,
chloride, phosphate, orthosilicate, and the many organic anions.
Schenk and Weber (10) reported that silicic acid hindered the hydroly-
sis of ferric iron, but their results are based upon indirect observa-
tions of free Fe+3, such as absorbance of the hydrolytic products at
420 mu, and residual [Fe(lll)! following filtration through an 0.45 u
membrane filter. It is suggested that the effects of such catalysts
-------
4-14
4.0
V
1-1
i
«L
3.0 —
2.0 -
CM
1.0
total sulfate
concentration
10"4M
total sulfate
concentration
5x!0"4M
2.8
3
3.6 2.8
total sulfate
concentration
10"3M
3.4
3.6 2.8
3c2
3.6
FIGURE 4-5a,b,c. pH-dependence of "second order rate constant" for
hydrolysis of Fe in the presence of sulfate.
5.0
4.0
I 3.0
o>
CM
2.0
1.0
total sulfate
concentration\
.-3.. V
/
10
no sulfate
Temp. 25 C
I = 0.1M
3.0
3.5
4,0
+3
FIGURE 4-6. Comparison between rates of hydrolysis of Fe
in the presence and absence of sulfate.
-------
4-15
or inhibitors be investigated by direct analytical methods such as
those employed in this study.
Solubility Product of Amorphous Ferric Hydroxide
Immediately following the investigation of the kinetics of
hydrolysis of Fe(IIl), the remainder of each of the samples was re-
moved from the ferro-ferri cell and placed in a BOD reaction bottle
under a nitrogen atmosphere. The BOD bottles were stored under water.
After a period of approximately 17 days, the samples were removed from
storage and again placed in the ferro-ferri cell. The reversible
Nernstian potential, E, pCH, and [Fe ] were determined as before, and
the solubility of ferric hydroxide was computed. The data and results
are summarized in Table 4-1. The pK Q-value indicated should not be
referred to as the solubility product of ferric hydroxide, since it is
probable that equilibrium had not been attained, and the structural form
of the end-product was not analyzed. In any case, the average value
determined, pK. = 38.1, does give some idea as to the magnitude of
SO
the solubility product of freshly-precipitated ferric hydroxide. Feit-
knecht and Schindler (3) report PK -values for the solubility product
SO
of amorphous Fe(OH)3 in 3M NaCK>4 of 38 to 39.1, depending upon the
age of the precipitate, i.e., the time elapsed following precipitation
before the measurements were made.
4-3 Coagulative Properties of Ferric Iron
O'Melia and Stumm (4) have shown that destabilization of col-
loidal dispersions of silica by Fe(lll) is accomplished by the multi-
meric and polymeric hydroxo ferric species which arise as unstable
-------
4-16
Table 4-1. Check on the Solubility Product of Ferric Hydroxide
Sample CR
No.
14
15
16
17
18
19
20
21
22
23
24
2
2
3
2
3
3
3
3
3
3
2
*
.89
.88
.00
.76
.21
.20
.31
.20
.20
.22
.85
E*
mv.
403.
391.
380.
416.
366.
369.
343.
364.
366.
372.
387.
Potential
NOTE:
CFe+2]
x 104M
6
8
0
2
9
2
7
6
0
4
9
4.52
5.60
5.15
5.46
1.95
2.28
2.25
2.04
2.10
1.88
5.64
readings
E°'
= 486
[Fe+2] CFe+3J
[Fe+3] x 106M 2
24.
39.
61.
15.
103.
93.
254.
112.
106.
83.
45.
versus
.0 mv.
6 18
1 14
7 8
1 36
1
5 2
0
1
1
0 2
6 12
saturated
.4
.3
.35
.2
.89
.44
.886
.82
.98
.27
.4
[OH"]3
c 103V
0.468
0.447
1.00
0.191
4.27
3.98
8.27
3.98
3.98
4.59
0.355
Average
K v
so pK
39 s°
x 10
8.60
6.39
8.35
6.92
8.07
9.71
7.33
7.25
7.88
10.42
4.40
pK
^ so
38.07
38.19
38.08
38.16
38.09
38.01
38.14
38.14
38.10
37.98
38.36
=38.1
calomel reference electrod
T = 25°C
I = 0.1
Readings taken 17 days after precipitation
-------
4-17
kinetic intermediates in the precipitation of ferric hydroxide. The
destabilization process is attributed to specific adsorption of these
hydroxo-ferric species; this adsorption brings about aggregation
either by a bridging effect whereby the hydrolyzed Fe(lll) forms
linkages between a number of dispersed particles, or by neutraliza-
tion of the negatively-charged colloidal particles by the positively-
charged hydrolyzed Fe(lll).
The efficiency of destabilization depends upon the hydrolytic
properties of Fe(lII) and is, therefore, influenced by pH. Figure 4-7
shows the effect of pH on the concentration of iron (III) required for
aggregation of colloidal silica (Ludox SM, E. I. Dupont de Nemours and
2
Co., Wilmington, Delaware) at a. surface concentration of 150 m /I. The
experimental data for Figure 4-7 were obtained in the same manner as
those obtained by O'Melia and Stumm (4) , usine light scattering as an
indicator of coagulation.
There are two results of the coagulation study that are directly
related to the kinetics of hydrolysis of Fe(IIl). Firstly, O'Melia
and Stumm observed that prolonged thermal aging of solutions of ferric
perchlorate, in order to form polymeric hydroxo-ferric species, in-
creased the critical coagulation concentration and the critical re-
stabilization concentration for the colloidal silica dispersion,i. e.,
greater concentrations of Fe(III) were required for a given degree of
coagulation and restabilization, respectively. This was attributed to
partial precipitation of Fe(OH) and a corresponding reduction in the
concentration of active hydrolytic intermediates. Secondly, the
plateau between pH 3 and 6 in Figure 4-7 suggests that in this region
-------
4-18
-3.0
-3.5
u
.
o
•
u
"I* -4.0
-4.5
-5.0
Coagulation region
I I HI I 1 I gl I I t I,
O O
Colloidal suspension
Ludox SM - 150 m /I
I I
I
3.0
4.0
PH
5.0
6.0
FIGURE 4-7. Aggregation of colloidal silica dispersion by
hydrolyzed ferric iron.
0
I
H
§
W
E-f
CO
O
-2
- Fe(OH)3
equilibrium
-24
K = 10 for FePO. ,, N
so 4 (s)
38
10 for Fe(OH)3
-8
FIGURf; 4-8. Solubility of ferric phosphate.
-------
4-19
of Fe(OH)3 insolubility, the polynuclear hydroxo-ferric species ex-
hibit equivalent coagulative behavior independent of pH. These studies
were conducted in pH-stats where constant pH was maintained by the ad-
dition of dilute sodium bicarbonate. However, the localized gradients
of high pH which are set up by the external addition of a basifying
agent bring about kinetically irreversible hydrolytic reactions. Such
effects were observed in the laboratory investigations of the kinetics
of Fe(lll) hydrolysis (section 4-2.2), and have also been reported by
Biedermann and Schindler (11) and by Spiro, et al (12). Since these
studies describe the coagulative properties of kinetic intermediates,
time of reaction and method of preparation of the hydrolytic reactants
should be of significance. Similar comments have been made (13) re-
garding the coagulative behavior of activated (polymeric) silica. The
pertinent variables include polymerization pH, polymerization concen-
tration, polymerization time, and age of diluted polymeric solution.
It is hoped that the kinetic studies reported in section 4-2.2 will
serve as a guide for further research designed to clarify the mechanism
of destabilization by hydrolyzed metal ions, and to identify the para-
meters which control the destabilization process.
4-4 Removal of Phosphate
It was shown, in section 2-3.1, that the solubility of ferric
iron is controlled by the solubility of its oxides and hydroxides,
except in waters containing appreciable concentrations of phosphate,
at pH-values below 5. In a similar manner, the total concentration of
dissolved phosphate is governed by solid ferric phosphate and is lowest
-------
4-20
at pH 3.5, where the 3olubility of FeP(>4 is at a minimum (see Figure
4-8). The pH at which minimum solubility is exhibited depends upon
the solubility products of FeP04 and Fe(OH)3 which are utilized in the
computations and the construction of Figure 4-8. The critical pH is
not well-defined since neither solubility product is well-known (see
section 2-3.1). The boundary below pH 3.5 represents the equilibrium
between dissolved phosphate and solid ferric phosphate; the boundary
above pH 3.5 corresponds to the control of soluble phosphate by the
equilibrium between solid ferric phosphate and solid ferric hydroxide,
reaction 2-27. (Soluble phosphate-complexes of ferric iron have been
neglected since they do not markedly enhance the solubility of FePO^.
In addition, other phosphate minerals, such as AlPO^, Ca^PO^^, and
Ca rt(PO.),(OH)0, have been disregarded although their significance in
10 4 6 2
natural systems should not be overlooked.)
Consequently, under certain conditions, Fe(lll) can serve as
an effective precipitant for the removal of phosphate from waste waters.
Figure 4-8 suggests that the efficiency of phosphate removal by Fe(lll)
is pH-dependent, since OH~ competes with the soluble phosphate species
for the metal precipitant. At pH -values greater than 3.5, the effi-
ciency is decreased due to formation of Fe(OH), and mixed hydroxo-
phosphato-precipitates of Fe(lll). However, by the judicious selec-
tion of pH and means of addition of Fe(lll), the hydrolytic tendency
of Fe(lll) can be supressed and the quantity of precipitant required
for removal of phosphate can be made to be stoichiometric in accordance
with the reaction
Fe+3 + P0'3 = FePC (4-14)
-------
4-21
Improvement of phosphate removal by Fe(IIl) can be achieved by
employing the techniques of homogeneous precipitation (14) whereby the
precipitant is not added directly to the system in the customary manner,
but Ls generated internally by a homogeneous chemical reaction within
the solution. In this way, coprecipitation is minimized and undesirable
concentration effects are eliminated. In the case of ferric iron,
homogeneous generation can readily be accomplished by the addition of
ferrous iron which is subsequently oxidized in situ. The resultant
Fe(lll) is uniformly distributed throughout the system, promoting
direct contact between all of the phosphate and all of the iron.
An experimental study of the removal of phosphate by externally-
added and homogeneously-generated ferric iron was undertaken and the
results are presented below.
4-4.1 Precipitation of Phosphate by Ferric Iron
Experimental Procedure
Various dilutions of a standardized solution of ferric per-
-4
chlorate were added to a series of beakers containing 10 M Na_HPO^.
(The stock solution of ferric perchlorate had been prepared in dilute
perchloric acid to prevent hydrolysis of Fe(IIl).) Dilute sodium car-
bonate was simultaneously added in order to achieve and maintain the
desired pH. The solutions were flash-mixed and then gently stirred for
fifteen minutes. Samples of each were membrane filtered (1.2 u pore
diameter) and the residual phosphate in the filtrate was determined
employing the method recommended by the AASGP (15).
-------
4-22
In the study of phosphate removal by homogeneous precipita-
tion, acidified ferrous perchlorate and dilute Na»CO, were added to
-4
a similar series of beakers containing 10 M NaJIPO,. The solutions
were flash-mixed and gently stirred while the ferrous iron was oxi-
dized, in situ, by atmospheric oxygen. At pH-values below 6 where the
oxygenation reaction is slow (see section 3-4.1), ozone was bubbled
through the system. (Ozone was generated by passing a stream of oxygen
through a Sanders ozonator, Model No. S-V-106, Triton Aquatics.) Ali-
quots were removed from each of the precipitating systems at various
time intervals and were analyzed for residual phosphate and Fe(II),
using the AASGP (15) and bathophenanthroline (6) procedures, respec-
tively. The concentration of Fe(lll) at any given time was calculated
as the difference between the initial concentration of Fe(ll) and
[Fe(ll)] at that time.
Results and Discussion
Figure 4-9a shows the residual concentration of phosphate for
various concentrations of ferric iron applied directly from solutions
of ferric perchlorate. The efficiency of Fe(lll) as a precipitant of
phosphate appears to improve as the pH of the system is lowered, con-
firming the predictions made in connection with Figure 4-8. Stoichio-
metric removal is approached at pH 5.
Removal of phosphate by homogeneously-generated ferric iron is
presented in Figure 4-9b. Here again, removal is improved as the pH
decreases. Comparison of Figures 4-9a and b demonstrates that, with
the exception of the results at pH 7.0, the degree of removal is
enhanced by utilization of the homogeneous precipitation technique.
-------
4-23
o
.
-
1
:
w
.-
W
b
.
I
a
D
o
—
: .
w
u
:
U
I
.--
I
-
fcj
.
-5.0 -4.5 -4.0 -3.5 -3.0
LOG [Fe(III)] APPLIED BY DIRECT ADDITION
Y
.5.0 -4.5 -4.0 -3.5 -3.0
LOG [Fe(III)l APPLIED BY HOMOGENEOUS GENERATION
FIGURE 4-9.
Precipitation of phosphate by (a) direct addition
of Fe(III) and by (b) homogeneous generation of
Fe(III).
-------
4-24
The improvement is exhibited most clearly at the lower pH-values
studied. Stoichiometric removal is again observed in the lower pH-
range, where partial removal is exhibited even at applied concentra-
-4
tions of Fe(lll) considerably below 10 M. It should be noted, how-
ever, that perfect Stoichiometric removal is not effected; approxi-
mately 3 x 10 ~ M of phosphate still remains, even after the applica-
tion of 10~Si Fe(IIl).
A distinction must again be made between equilibrium and kine-
tics. The problem at-hand-deals with a non-equilibrium situation and,
although equilibrium considerations can serve as a guide, kinetic con-
siderations are required. Phosphate may be removed from solution in
two ways: by formation of insoluble ferric phosphate, or by incorpora-
tion in the ferric hydroxide network. The latter is, by definition,
a less efficient process. Conditions which favor direct iron-phosphate
interactions over hydrolytic reactions should lead to a more effectual
removal of phosphate. At lower pH-values, hydrolysis is decelerated
and precipitation of phosphate is enhanced. However, given sufficient
time, the possibility of conversion of FePO, to Fe(OH)- cannot be dis-
counted under conditions where the latter is the thermodynamically
stable solid phase. The results reported in Figure 4-9 reflect removal
of phosphate after only fifteen minutes. With longer periods of time,
phosphate might be released if such a solid conversion were to occur.
The improvement in the removal of phosphate by homogeneous
precipitation can also be attributed, in part, to the catalytic influ-
ence which phosphate exerts on the oxidation of ferrous iron. In
Chapter 3, it was indicated that the apparent catalytic effect of
-------
4-25
phosphate may arise from its coordination with Fe , in which case it
would remain bound to the product Fe(lll), thus increasing its chances
of removal from solution.
4-5 Summary
The rate of hydrolysis of ferric iron was shown to be an influ-
ential parameter in coagulation and precipitation phenomena in natural
systems. Two important features of the hydrolytic reactions were em-
phasized, namely the decided dependence of the reaction rate on [OH ],
and the kinetic irreversibility of the reactions. Consequently, the
creation of concentration gradients, such as those which arise by the
addition of a basifying agent, may cause an undesirable effect in the
system under investigation. The techniques of homogeneous precipita-
tion are recommended as one means of overcoming such irreversible in-
terferences.
-------
4-26
References
1) Sillen, L.G., "Quantitative Studies of Hydrolitic Equilibria,"
Quart. Revs., Chem. Soc. London, 13, 146 (1959)
2) Stumm, W., "Metal Ions in Aqueous Solution," page 520 in Principles
and Applications of Water Chemistry, S. D. Faust and J. V.
Hunter, eds., John Wiley and Sons, Inc., New York (1967)
3) Feitknecht, W., and Schindler, P., "Solubility Constants of Metal
Oxides, Metal Hydroxides, and Metal Hydroxide Salts in Aqueous
Solution," Pure Appl. Chem., j6, 132 (1963)
4) O'Melia, C. R. , and Stumm, W., "Aggregation of Silica Dispersions
by Iron (ill)," Journ. Coll. Inter. Sci., 23_, 437 (1967)
5) Morgan, J. J., and Stumm, W., "The Role of Multivalent Metal
Oxides in Limnological Transformations, as Exemplified by
Iron and Manganese," Proc. 2nd Intl. Wat. Poll. Res. Conf.,
page 103, Tokyo (1964)
6) Lee, G. F., and Stumm, W., "Determination of Ferrous Iron in the
Presence of Ferric Iron," J. Amer. Wat. Works Assn., 52, 1567
(1960)
7) Biedermann, G., and Chow, J. T., "The Hydrolysis of the Iron (ill)
Ion and the Solubility Product of Fe(OH)_ 7nCln ,n in 0.5M
(Ha*)Cl~ Medium." Acta Chem. Scand., 20. " * 1376. (1966)
8) Wendt, H., and Strehlow, H., "Schnelle lonenreaktionen in Losungen.
II. Die Bildung einiger einfacher Komplexe des Eisen-III-ions,"
Z. Elektrochem., £6, 228 (1962)
9) Wendt, H., "Schnelle lonenreaktionen in Losungen. III. Die
Kinetik der Bildung des binuklearen Eisen-III-hydroxokomplexes
Fe(OH)_Fe ," ~Z. Elektrochem., 66. 235 (1962)
10) Schenk, J. E., and Weber, W. J., "Chemical Interactions of Dis-
solved Silica with Iron (II) and Iron (III)," J. Amer. Wat.
Works Assn.. 60_, 199 (1968)
11) Biedermann, G., and Schindler, P., "On the Solubility Product of
Precipitated Iron (III) Hydroxide," Acta Chem. Scand., 11,
731 (1957)
12) Spiro, T. G., Allerton, S. E., Renner, J., Terzis, A., Bils, R.,
and Saltman, P., "The Hydrolytic Polymerization of Iron (III),"
J. Amer. Ghem. Soc.. 88_ 2721 (1966)
-------
4-27
13) Stumm, W. , Huper, H., and Champlin, ,R. L. , "Formation of Poly-
silicates as Determined by Coagulation Effects," Environ.
Sci. Tech., _!, 221 (1967)
14) Gordon, L. , Salutsky, M. L., and Willard, H. H., Precipitation
from Homogeneous Solution, John Wiley and Sons, Inc., New
York (1959)
15) Association of American Soap and Glycerine Producers, "Determina-
tion of Orthophosphate, Hydrolyzable Phosphate, and Total
Phosphate in Surface Waters," J. Amer. Wat. Works Assn., 50_,
1563 (1958)
-------
CHAPTER 5
OXIDATION OF IRON FYRITE: POLLUTION OF NATURAL
WATERS BY COAL MINE DRAINAGE
5-1 Introduction
Pollution by coal mine drainage arises from the exposure of
sulfur-bearing minerals present in the coal strata to the natural
weathering process. Mine drainage waters are characterized by low pH,
high acidity, and large concentrations of sulfate and iron as well as
other dissolved metals. Various measures have been proposed to cope
with this problem, ranging from treatment of the ensuing wastewater to
abatement methodology in which the weathering process is inhibited.
However, before considering such corrective measures, attention must
first be focused on the chemical reactions which occur and upon the
kinetics which govern these reactions.
Despite previous creditable efforts, no unambiguous ansv;er has
been found as to which of the steps in the production of acid mine
drainage determines the overall rate of dissolution of the sulfuritic
agglomerates. To date, a didactical approach toward evaluating the
individual factors controlling the kinetics of the overall reaction
has been lacking. This chapter describes such an approach. The rela-
tive rates of the consecutive reactions involved have been considered,
as well as the importance of each as it contributes to the problem of
mine drainage. The rate-determining step of the overall sequence has
-------
5-2
been ascertained and the physical, chemical, and biological factors
influencing this step have been quantitatively evaluated. A model
is presented to describe the mechanism by which the sulfide minerals
are oxidized, and the consequences of the model are discussed from
the standpoint of the various control methods which have been sug-
gested.
5-2 Thermodynamics and Stoichiometry of Reactions
The mine-water system can be characterized by the following
overall stoichiometric reactions:
FeS2(s) + |- 02 + H20 = Fe+2 + 2 S04"2 + 2 H+ (5-1)
Fe+2 + i 02 + H+ = Fe+3 + \ H^ (5-2)
Fe+3 + 3 H20 = Fe(OH)3(s) + 3 H+ (5-3)
FeS2(s) + 14 Fe+3 + 8 H20 = 15 Fe+2 -;- 2 SC>4~2 + 16 H+(5-4)
The sulfur-bearing minerals predominant in coal seams are the iron
sulfide ores, pyrite and marcasite. Both have the same ratio of
sulfur to iron, but their crystallographic properties are quite dif-
ferent. Marcasite has an orthorhombic structure while pyrite is
isometric (1). Marcasite is less stable and more easily decomposed
than pyrite. The latter is the most widespread of all sulfide minerals
and, as a result of its greater abundance in the eastern United States
(2), pyrite is recognized as the major source of acid mine drainage.
-------
5-3
During coal mining operations, pyrite is exposed to air and water
with atmospheric oxygen oxidizing the sulfide of the pyrite to sulfate
(5-1), releasing dissolved ferrous iron and acidity into the water. The
dissolved ferrous iron undergoes oxygenation to ferric iron (5-2) which
subsequently hydrolyzes to form insoluble ferric hydroxide (5-3), re-
leasing more acidity to the stream and coating the stream bed. Ferric
iron can also be reduced by pyrite itself, as in reaction 5-4, where
sulfide is again oxidized and acidity is released along with additional
ferrous iron which may re-enter the reaction cycle via 5-2.
The concentration of sulfate or acidity in the water can be di-
rectly correlated to the amount of pyrite which has been dissolved. The
introduction of acidity into the stream arises from the oxidation of
S (-II) of the iron pyrite (reactions 5-1 and 5-4) and from the oxygena-
tion of Fe(H) and the ensuing hydrolysis of the resulting Fe(lll)
(reactions 5-2 and 5-3). There is a stoichiometry which should not be
overlooked: the dissolution of one ncle of iron pyrite leads ultimately
to the release of four equivalents of acidity-two equivalents from
the oxidation of S^-II) and two from the oxidation of Fe(Il). The
decomposition of iron pyrite is among the most acidic of all weathering
reactions owing to the great insolubility of Fe(III) (3).
Pyritic agglomerates are thermodynamically unstable upon expo-
sure to atmospheric oxygen, as demonstrated by calculating the change
in free energy for reaction 5-1 as given in Appendix D. The exclusion
of oxygen from the system may not be a practical solution to the prob-
lem from the standpoint of economics, technology, and, as will be seen
-------
5-4
below, chemistry. Hence, efforts must be aimed at retarding the kine-
tics which control the dissolution of the pyritic material.
5-3 Previous Investigations of the Kinetics and Mechanism of Pyrite
Dissolution
5-3.1 Physical and Chemical Studies
As a result of the complexity of the reactions and 'the failure
of many previous observers to properly identify, isolate, and control
the rate-determining variables, much of the data reported previously
is difficult to interpret. The complex nature of pyrite itself, the
variety of forms that sulfur may take as an intermediate, and the
failure to single out the oxidizing agent all contribute to the com-
plexity of the system.
Stokes (4) studied the oxidation of pyrite and marcasite by
Fe(IIl) and, by analyzing the end products, concluded that the reac-
tion proceeded by a two-step mechanism
FeS2(s) + Fe (S04>3 = 3 FeS04 + 2 S° (5-5)
2 S° -i- 6 Fe0(SO.), + 8 H00 = 12 FeSO. + 8 H0SO. (5-6)
243 2 4 24
Nelson and Snow (5) observed that the rate of oxidation of pyritic
sulfur in coal by oxygen was markedly influenced by the degree of
fineness of the coal, the rate being inversely proportional to the
average diameter of the coal particles. The addition of ferric sul-
fate was found to hasten, the oxidation process.
Instead of looking at the overall reaction and the kinetic vari-
ables affecting it, Sato (6) employed a technique by which electrode
-------
5-5
potentials of sulfide minerals were measured in order to elucidate
the oxidation mechanism by which the minerals dissolve. When both the
oxidation potential of the sulfide test electrode and the oxidation
potential of the solution with which it is in contact are identical,
equilibrium has been reached, i.e., the potential corresponds to the
equilibrium potential between the mineral and its dissolved oxidizing
ions. Consequently, the equilibrium potential is controlled by the
first-step reaction: the heterogeneous reaction whereby the sulfide
dissolves. By measuring the electrode potential of the mineral and
comparing it to the oxidation potentials of the various possible half-
cell reactions of the sulfide using independent thermodynamic data,
Sato found the oxidation reaction for pyrite which best fits the
measured electrode potentials to be
FeS2(s) = Fe+2 + S^g) + 2 e" (5-7)
for pH values below 2. This technique, although it does not describe
the kinetics of the oxidation of pyrite, does give some insight as to
the mechanism by which pyrite is oxidized. If the reduction reaction
could be identified then, by summation of the two half-reactions, one
would have the first step in the dissolution of iron pyrite. As Sato
explains, once the heterogeneous reaction is established, the other
consecutive reactions for the oxidation of sulfur and ferrous iron can
be treated individually.
Garrels and Thompson (7) studied the rate of reduction of Fe(IIl)
by iron pyrite at pH values below 2 and observed that the instantaneous
rate of reduction of Fe(lll) decreased as the ratio of ferric to ferrous
-------
5-6
iron decreased. The rate was assumed to be proportional to differen-
tial adsorption of ferric and ferrous iron on the surface of pyrite,
i.e., to the fraction of pyrite surface occupied by Fe(lll). In each
case, fifty percent of the ferric iron was reduced by two grams of
crushed pyrite in less than one day.
A number of hydrometallurgical studies have been conducted
dealing with the oxidation of iron sulfide minerals at elevated tem-
peratures and under pressure to further understand the kinetics of
the pressure-leach process employed in the treatment of pyritic ores.
McKay and Halpern (8) investigated the oxidation of iron pyrite at
temperatures of 100 to 130°C and at partial pressures of oxygen of
0 to 4 atmospheres. The reaction was found to be first order in sur-
face area of pyrite and first order in the partial pressure of oxy-
gen. An attempt was made by McKay and Halpern to study the effect of
Fe(lll) on pyrite in the absence of oxygen in which they found that,
although FeS- was oxidized and leached, the quantity of FeS^ oxidized
was small compared to that amount oxidized by oxygen under similar
conditions. However, since the concentration of Fe(lll) was limiting,
i.e., lO"1*! Fe(lll) was completely reduced by 0.33M FeS within two
hours, one should not expect more than r-; x 10" M of pyrite to be leached
according to the stoichiometry of the reaction (5-4). Hence, the in-
ability of Fe(lll) to compete with oxygen as an effective oxidant of
iron pyrite was not justifiably demonstrated,
Pressure-leaching of pyrite was also investigated by Gerlach,
et al (9) over a wider range of temperatures and partial pressures of
oxygen. Their results are similar to those of McKay and Halpern (8)
-------
5-7
with regard to the observed dependencies of the rate law and the pro-
posed mechanism for the reaction.
Neither study, however, considered in adequate detail the oxi-
dation of pyrite by ferric iron which is inevitably produced by the
process. Barnes and Romberger (10) also fail to place ferric iron in
its proper perspective when they consider the concentration of dis-
solved Fe(lll) to be so small in acid mine drainage that its signi-
ficance as an oxidant is unlikely. However, in view of the vast amount
of Fe(ll) oxidized to Fe(lll), there certainly is a significant supply
of Fe(lll) readily available as an oxidant of pyrite. Even though the
standing concentration of Fe(lll) may be small, its reactivity in terms
of its rate of turnover is relatively high. Smith, Svanks, and Shumate
(11) noted the significance of ferric iron after observing a similarity
between the rates of aerobic and anaerobic oxidation of iron pyrite,
and postulated that the ultimate oxidant for both aerobic and anaerobic
oxidation is ferric iron.
5-3.2 Microbiological _Studies
Since the first recorded isolation from acid mine waters of
bacteria capable of influencing the oxidation of ferrous iron to fer-
ric iron (12), numerous biological reports dealing with microbial
catalysis of the oxidation of ferrous iron and various sulfide miner-
als have flooded the literature. Temple and Colmer (13) named their
"iron-oxidizing" autotroph Thiobacillus ferrooxjLdans^ it being capable
of autotrophic growth at the expense of ferrous iron or thiosulfate.
This bacterium differs from the autotroph Thiobacillus thiooxidans
-------
5-8
which also has oeen isolated from acid mine streams but which derives
its energy by catalyzing the oxidation of elemental sulfur. After
summarizing the known facts regarding autotrophic bacteria and the
dissolution of iron pyrite, Temple and Delchamps (14) postulated the
overall sequence describing the formation of acid in coal mines as
follows:
The sulfide of finely-divided iron pyrite or marcasite is
chemically oxidized by oxygen to sulfate
FeS2(s) + H20 + 3 1/2 DZ = Fe+2 + 2 S04"2 + 2 H+ (5-8)
Ferrous iron is oxygenated, under the catalytic influence of the auto-
troph Thiobacillus ferrooxidans, to ferric iron
Fe+2 + 1/4 02 + H+ = Fe+3 + 1/2 H20 (5-9)
As rapidly as it is formed, ferric iron is chemically reduced by the
finely-divided iron pyrite
2 Fe+3 + FeS2(s) = 3 Fe+2 + 2 S° (5-10)
The elemental sulfur liberated may be oxidized by ferric iron
2 S° + 12 Fe+3 + 8 H20 = 12 Fe"1"2 + 2 S04~2 + 16 H+ (5-11)
or by oxygen, in which case the reaction is catalyzed by the autotroph
Thiobacillus thiooxidans
S°.+ 1 1/2 0 + H 0 = 2 H+ + S04"2 (5-12)
The ferrous iron resulting from the oxidation of pyrite by ferric iron
is then subject to further microbial action by Thiobacillus ferrooxi-
dans, as in reaction 5-9. A cycle is established involving formation
-------
5-9
of Fe(lII) from Fe(il) by microbial catalysis, and chemical oxidation
of iron pyrite by the resultant Fe(lII).
This significant piece of work was obscured by later develop-
ments which questioned some of the physiological properties of T_.
ferrooxidans, in particular its ability toward mediating the oxidation
of thiosulfate and not elemental sulfur. Two additional autotrophic
organisms were proposed, confusing the issue regarding autotrophic oxi-
dation of ferrous iron and its role in acid mine drainage: Ferrobacil-
lus ferrooxidans (15), which was able to grow on ferrous iron but not
on thiosulfate or elemental sulfur, and Ferrobacillus sulfoxidans (16),
which could utilize either ferrous iron or elemental sulfur as an
energy source. Unz and Lundgren (17) concluded that the organisms
were all nutritionally similar and called for a re-evaluation of the
current classification procedure.
This controversy also hindered verification that the oxidation
of Fe(Il) was a direct microbial effect and not an indirect one, i.e.,
whether or not the organisms were true chemo-autotrophs which derived
their energy from the direct oxidation of Fe(ll). Little quantitative
evidence correlating the rate of growth of the organisms and the rate
of oxidation of ferrous iron was available until Silverman and Lund-
gren (18) showed that the change in the logarithm of the concentration
of F. ferrooxidans paralleled the increase in the logarithm of the con-
centration of Fe(IIl) produced by the oxidation of Fe(ll). They also
observed (19) that the quantity of carbon assimilated was in accord-
ance with the thermodynamic free energy available from the oxidation
of Fe(ll). Schnaitman (20) demonstrated that microbial mediation of
the oxidation of ferrous iron by F. ferrooxidans conformed to Michael is-
-------
5-10
Menton enzyme kinetics, and that the rate of oxidation was proportional
to the relative concentration of bacterial cells.
Numerous reports persist in the literature concerning the nature
of the influence these chemo-autotrophic microorganisms exert on the
oxidation of pyrite and marcasite, as well as other sulfide minerals
encountered especially in the copper mines in the western United States.
Brynner, et al (21) have attributed leaching of chalcopyrite (CuFeS^^
covellite (CuS), chalcocite (Cu-S), bornite (CuFeS,), and tetrahedrite
(Cu0Sb0S_) to direct biological oxidation. Molybdenite (MoS0) (22) and
o L I t-
orpiment (As~S-) (23) have also been reported to be subject to direct
oxidation. Silverman and Ehrlich (24) reviewed the subject of microbi-
al catalysis of mineral transformations and indicated that the action
of autotrophic microorganisms may be two-fold: to regenerate ferric
iron from ferrous iron which then chemically oxidizes the mineral sul-
fide; and to directly attack and oxidize the sulfide minerals inde-
pendent of the action of ferric iron. Ehrlich (25) has testified to
the likelihood of such a direct effect, but no mechanism has yet been
proposed.
5-4 Purpose of Experimental Study
Since "at-source" control of coal mine drainage must depend
upon retardation of the kinetics controlling the oxidation of pyrite,
it is necessary to know which of the sequential reactions involved
controls the overall rate of the reaction. This study was undertaken
to investigate the relative rates of the various reactions producing
acid mine drainage, to ascertain which of the steps is rate-limiting,
-------
5-11
and to suggest measures for controlling this particular reaction in
natural waters.
5-5 Oxygenation of Ferrous Iron
The oxidation of ferrous iron was discussed in Chapter 3 and
it was seen that, in the acidic pH-range corresponding to conditions
encountered in mine drainage waters, the reaction proceeds relatively
slowly and is independent of pH. However, the composition of actual
mine waters may be such that oxidation is accelerated, i.e., physical
conditions or chemical and biological agents may be present which in-
fluence the rate of oxidation of Fe(ll). Examples of these include in-
organic ligands, such as sulfate, which complex ferrous and ferric
iron, soluble metal ions such as copper(ll), aluminum, and manganese(ll),
suspended material with large surface areas and high adsorptive capa-
cities, such as coal and clay particles, and microorganisms. All have
been implicated in the literature, in various circumstances, as being
capable of accelerating the rate of oxidation of Fe(Il). Consequently,
the oxidation was investigated in the presence of many of these chemi-
cal catalytic agents to observe their effect on the rate of oxidation
under synthetic mine conditions and to compare the observed catalytic
rate with the actual rate of oxidation of ferrous iron in natural mine
waters.
The studies were initially conducted in the absence of micro-
organisms in order to characterize the kinetics of the reaction in
purely chemical terms.
-------
5-12
5-5.1 Experimental Procedure
The experimental procedure and analytical techniques were simi-
lar to those employed in Chapter 3 to follow the oxidation of ferrous
iron with time. For the slower long-term studies, the absorbances of
acidified aliquots of the samples were measured at the isosbestic
point of the system (where Fe and FeOH have the same molar absorp-
tivity) as an indication of the quantity of ferrous iron oxidized.
Figure 5-1 shows that the molar absorptivity of the acidified solution
containing Fe and FeOH+ , at 272 mu, is unaffected by concentrations
of sulfate as high as 10" M.
For the short-term, more rapid catalytic studies, ferrous iron
was determined directly by titration with permanganate, or by chelation
by the colorimetric reagent bathophenanthroline (4,7-diphenyl-l,10-
phenanthroline) (26). In the surface-catalytic studies, aliquots of
the heterogeneous suspension were acidified with dilute perchloric acid
and filtered. The solid material recovered was rinsed with dilute
acid to remove all traces of ferrous iron that might have been adsorbed
on the surfaces of the particles.
Samples were prepared containing various dilutions of a standard-
ized solution of ferrous perchlorate and different concentrations of
the catalytic agents. The effects of sulfate, aluminum, manganese(ll),
copper(ll), powdered charcoal, alumina, silica, aged ferric hydroxide,
crushed iron pyrite, and the natural clays bentonite and kaolinite were
examined. The pH was adjusted with concentrated perchloric acid and
the samples were allowed to equilibrate with the oxygen of the atmosphere.
After removing aliquots for the above-mentioned analyses, the solutions
-------
5-13
1.6
1.2
s
0.8
W
o
o
en
0.4
[SO,"2] » 10"3M A
[so ~2J = io"2M n
[SO, 1 • 10 M O
molar absorptivity in the absence
of sulfate, from Figure 2-2
2 4 "4
Fe(III) CONCENTRATION, x 10
FIGURE 5-1. Effect of sulfate on absorbance of Fe(III) at
272 0*1.
0.0
i—~T
-i.o K—
-2.0
k" - - d log [Fe(II)l
dt
p «= 0.20 atm.
Temp. = 25 C
-3.0
-4.0
Stumm-Lee
rate law
extrapolation
j L
0 1
FIGURE 5-2.
3 4
pH
Rate of oxygenation of ferrous iron as a
function of pH.
-------
5-14
were re-equilibrated with the atmosphere. The samples were sealed and
stored, quiescently, in an incubator at 25°C in the dark to avoid any
photochemical interference. In the investigations involving hetero-
geneous catalysis (clays, pyrite, and powdered charcoal), the suspen-
sions were stored on magnetic stirrers and shielded from light. One
study was conducted under sterile conditions to preclude the influence
of microbial catalysis which may have been accidentally effective
through contamination.
5-5.2 Experimental Results and Discussion
Figure 5-2 summarizes the results reported in Chapter 3 for
the uncatalyzed rate of oxidation of ferrous iron. This figure serves
as the basis for comparison with the catalyzed rates.
Effect of Sulfate
The rate of oxidation of ferrous iron in the presence of sulfate
can be satisfactorily fitted to a rate relationship which is first-
order in the concentration of ferrous iron, as shown in Figure 5-3. The
rate constant k" = -d log,Q [Fe(ll)]/dt is of the same order of magni-
tude and is quite similar to the rate constant for the uncatalyzed re-
action. To magnify the effect of any catalytic dependence of the oxi-
dation on the concentration of sulfate, the study was repeated at 50 C
where the change in Fe(Il) per unit time is greater. Figure 5-4
demonstrates more vividly the catalytic effect of sulfate.
Huffman and Davidson (27) investigated the oxidation of Fe(Il)
in solutions of sulfuric acid, at 30.5°C and 1 M H^SO,, and observed
the reaction to be second-order in Fe(ll) according to the rate law
-------
5-15
z
B
H
8
-3.010
-3.014
P
M
T -3.018
to
-3.022
pH 2.9
[Fe(II)] - 9.78x10 >[
-2* -3
] = 10 J
25°C
PQ » 0.20 atm.
:" - 1.2x!0"4 day"1
20
80
100
40 60
TIME, days
FIGURE 5-3. Rate of oxidation of ferrous iron in the presence of
sulfate.
-2.50 L
40
160
80 120
TIME, days
FIGURE 5-4. Effect of sulfate on the oxidation rate of
ferrous iron at 50°c.
200
-------
5-16
where the termolecular rate constant k = 2.8 x 10" liter mole
atm sec . Under a partial pressure of oxygen of 0.2 atm. and an
initial concentration of Fe(ll) of 10~ M, and converting the units of
time, the initial rate "constant" of Huffman and Davidson can be approxi
mated as k' = 4.8 x 10~5 day'* (k'= k [Fe(ll)] Prt =* -d In [Fe(lD] /dt)
t o 0_ o
or k" = 2.0 x 10 day . This pseudo-first-order rate "constant"
can now be compared with the rate constants obtained in this study.
The observations of Huffman and Davidson in sulfuric acid show
the oxidation to occur more rapidly in a solution containing sulfate
than in a medium of perchlorate as investigated by George (28). The
latter conducted his study of the oxidation of Fe(ll) in perchloric
acid and also observed the reaction to be second-order in Fe(ll). (See
Section 3-3.1.) Performing a similar calculation as above, the pseudo-
first-order rate "constant" obtained by George is k" = 1.0 x 10 day ~*
at 30°C and approximately 10~T1 HC10 .
The results both of Huffman and Davidson and of George are of
the same order of magnitude as the results obtained in this study,
except that these authors have characterized the oxidation as being
second-order in Fe(ll). For such an investigation, where the reaction
proceeds so slowly that only 2-3% of the reaction is complete after
three months, it becomes difficult to classify the reaction with
respect to its order, as previously discussed in Chapter 3. The
agreement among the pseudo-first-order rate constants is gratifying
in itself, George and Huffman and Davidson also observed the rate of
oxidation to increase only slightly with an increase in pH.
-------
5-17
Figure 5-4, and a comparison of the results of Huffman and
Davidson to those of George suggest that the oxidation of ferrous
iron takes place more rapidly in the presence of sulfate than in per-
chlorate alone, the magnitude of the catalysis, however, being not
very great.
'Dy Dissolved Metal Ions
Of the dissolved heavy metals which are normally present in
natural mine waters, copper (II) exerted the strongest catalytic in-
fluence as shown in Figure 5-5. For the sake of convenience, and since
the data fit the formulation fairly well, the rate has been plotted as
a reaction which is first-order in Fe(ll). The pseudo-first-order
rate constant of k" = 4 x 10" day~ indicates that 10~ M Cu(Il) ac-
celerates the rate of oxidation of Fe(ll) approximately four-fold.
The cupric ion has been reported to be an efficient catalyst
in the oxidation of Fe(Il) in solutions of phosphoric acid (29),
sulfuric acid (27), hydrochloric acid (30), and in the neutral pH-
range corresponding to natural waters (31). Cher and Davidson (29)
account for the catalytic effect of copper (II) by the following
mechanism whereby Cu(ll) serves as an electron-transfer catalyst:
Fe(Il) + Cu(II) = Fe(lll) + Cu(I) (5-14)
•
Cu(l) 4- 0£ = Cu(ll) + H02 (5-15)
<•
The free radical H0_ reacts further with additional Fe(Il) as in the
Weiss scheme (32) previously discussed in Chapter 3.
The rate law for the oxidation of Fe(Il) in solutions of sulfuric
acid containing Cu(Il) was reported by Huffman and Davidson (27) to be
-------
5-13
55
s
I
g
W
K
8
fl-
-2.05
-2.07
-2.09
-2.35
•2.37
-2.39
pH 3.0 3
[Fe(II)]o= 8.74x10 M
10"AM Cu(II)
k" - 4.0xlO~4 day"1
pH 3.0 3
[Fe(II)lo= 4.49x10 M
20
80
100
40 60
TIME, days
FIGURE 5-5. Effect of copper(II) on oxidation of ferrous iron.
10
9
8
7
I
D
[Fe(II)]
- 10"2M
A1203
pH 3.7
k"
8250 ru /I
- 5xlO"3M
S.OxlO"3 day"1
10
25
30
35
15 20
TIME> days
FIGURE 5-6. Rate of oxidation of ferrous iron in the presence of
suspended aluminum oxide.
-------
5-19
first-order in the concentrations of Fe(Il) and Cu(II), Figure 5-5
conforming to such a description. The magnitude of the rate constant
obtained in this study, however, cannot be compared with that of Huff-
man and Davidson in sulfuric acid since, in contrast to solutions con-
taining no Cu(ll), the rate of oxidation increases markedly with in-
creasing concentration of acid (33). Indeed, the'results of Huffman
and Davidson, in 0.11M (H+) and a total activity of sulfate of 0.58M,
show the rate of oxidation of Fe(ll) in the presence of 10" M Cu(II)
to be about 100 times greater than depicted by Figure 5-5.
Although Stumm and Lee (31) observed other heavy metal ions
(Mn , Co+ ) to exhibit similar catalytic effects on the rate of oxi-
dation of Fe(ll) at neutral pH values, it was found that Mn and Al
showed no measurable acceleration of the reaction rate for the acidic
conditions of this study.
The Effect of Clays
On the premise that surfaces of clays play a significant role
in the oxidation of ferrous iron in natural mine waters, a study of
the catalytic effect of silica, SiO_, and alumina, Al-O^, was under-
taken. Silica and alumina, which are the basic building blocks of
all clays (clays being composed of two dimensional arrays of silicon-
oxygen tetrahedra and aluminum or magnesium-oxygen-hydroxide octa-
hedra (34)), were selected as idealized representatives of natural
clays. Hydration of the surface of the clay results in the formation
of silanol, Si-OH, and aluminol, Al-OH,groups. Since the rate of oxi-
_ 2
dation of ferrous iron at higher pH-values is dependent upon ODH ]
(see Chapter 3), it was believed that the hydroxo-metal groups of the
-------
5-20
hydrated clay surface played a specific catalytic role due to the ap-
parent localized high PH (high concentration of OH groups) at the par-
ticle surface. Furthermore, since clays are strongly adsorptive and
exhibit ion-exchange properties, a general catalytic influence was
thought to occur as a result of adsorption of the reactants and local-
ized increased concentrations of reactants at the particle surface.
In addition, Schenk and Weber (35) observed orthosilicic acid, H4SiC>4,
with which natural clays are in equilibrium, to increase the rate of
oxidation of Fe(ll) at pH-values greater than 5 in much the same manner
as other inorganic ligands, such as phosphate, chloride, and sulfate.
Of the clay surfaces investigated, aluminum oxide (Baymal
colloidal alumina, manufactured by E. I. DuPont de Nemours and Co.,
Wilmington, Delaware) exhibited the strongest catalytic properties.
Figure 5-6 shows that, for a given pH and concentration of alumina,
the surface-catalytic oxidation can be described by a rate equation
which is first-order in concentration of Fe(ll). The slope of the
first-order plots (which is, by definition, the rate constant) is in-
dependent of the initial concentration of Fe(Il) which further confirms
the order of the reaction with respect to Fe(Il). The rapidity of
the surface-catalytic reaction compared to the uncatalyzed oxidation
is demonstrated by noting the half-time of the former (the time re-
quired for 50% of the initial concentration of ferrous iron to be
oxidized) to be only about 40 days, in contrast to about 500 days for
the latter. Table 5-1 compares the surface-catalytic rate constants
for alumina with the uncatalyzed rate constants for the oxidation of
2
Fe(Il), and it is seen that in the presence of about 8000 m /I of
-------
5-21
Table 5-1. Comparison of Surface-Catalytic Rate Constants
with Uncatalyzed Rate Constants
pH
3.5
3.8
4.0
log k"
Uncatalyzed
Reaction
-3.6
-3.4
-3.3
A12°3
8000 m*/l
-2.52
-2.05
-1.80
k"
-d log [Fe(II)]
dt
*Baymal Colloidal Alumina - manufactured by E. I. DuPont
de Nemours and Co., Wilmington, Delaware.
Al O surface, the catalyzed reaction is 10-30 times faster than the
uncatalyzed reaction.
Figure 5-7 demonstrates further the direct catalytic dependence
on A1203, showing the rate of oxidation of Fe(ll) to increase as the
areal concentration of the idealized clay is increased. The oxidation
rate in the presence of alumina is also dependent upon pH as seen in
Figure 5-8, with a regular increase in rate with increasing pH being
observed.
-------
5-22
1
§
g
I
8
M
>^
-------
5-23
-1.8
-2.0
o
s ..
2.2
-2.4
-2.6
k" - - d log [Fe(II)l
dt
FIGURE 5-8. Effect of pH on surface-catalytic oxidation
of ferrous iron. .
-3.0
o
I
-3.1
-3.2
-3.3
<
1800 m 11 Ludox SM
colloidal silica
pH 3.65
m'/l Ludox SM
pH 4.1
10 gms/1 reagent
bentonite clay
pH A.O
D
10
30
40
FIGURE 5-9.
20
TIME, days
Rate of oxidation of ferrous iron in the presence
of colloidal silica and bentonite clay.
-------
5-24
(The catalytic effect of Al 0 cannot be attributed to specific
catalysis by dissolved Al(lll) in equilibrium with the solid since it
was previously shown that dissolved Al(lll) had no measurable effect
on the rate of oxidation of ferrous iron.)
Ludox colloidal silica (Ludox SM is a colloidal silica manu-
factured by E. I. DuPont de Nemours and Co., Wilmington, Delaware)
and bentonite, a natural montmorillonlte clay (Bentonite Powder,
U.S.P., Fisher Scientific Co., Fair Lawn, New Jersey) were also found
to catalyze the oxidation (Figure 5-9) but to a slightly lesser degree
than the alumina. Kaolinite, another natural clay, demonstrated no
catalytic properties although it gradually neutralized the acidity.
Colloidal ferric hydroxide, prepared by allowing a solution of
Fe(lll) oversaturated with respect to the hydroxide to hydrolyze,
showed no tendency toward catalyzing the oxidation of Fe(Il).
Catalysis by Powdered Charcoal
Lamb and Elder (33) reported that granular, steam-activated
coconut charcoal markedly accelerated the rate of oxygenation of
ferrous iron. This phenomenon was attributed to the ability of char-
coal to greatly enhance decomposition of hydrogen peroxide in the
presence of Fe(ll), generating in turn the active free radicals which
oxidize Fe(ll) according to the Weiss mechanism (32).
In this study, reagent-grade sugar charcoal(purified reagent,
Fisher Scientific Co., Fair Lawn, New Jersey) displayed no apparent
catalytic tendency toward the oxygenation of Fe(ll). Again, however,
the catalytic properties of charcoal described by Lamb and Elder in-
creased considerably as the concentration of acid increased, thus
-------
5-25
accounting in part for the discrepancy between their results and those
reported here. Furthermore, Lamb and Elder followed the oxidation of
Fe(ll) using an electrochemical technique whereby the change in elec-
trochemical potential was correlated to a change in the Fe /Fe
ratio by a modified Nernst equation. In the heterogeneous system con-
taining suspended charcoal, the reliability of such measured potentials
as indicators of the total concentration of Fe(ll), both in solution
and associated with the surface of charcoal, must be questioned
especially when the rate of change of Fe(ll) is so small. The presence
of impurities in the charcoal also tend to cast doubt on the experi-
*
mental reliability.
Effect of Iron Pyrite
To complete the investigation of all natural chemical catalytic
agents which may be responsible for acceleration of the rate of oxi-
dation of ferrous iron in mine waters, the effect of iron pyrite itself
was studied. It has been suggested by Smith, et al (11) that the
surface of pyrite acts as a catalyst to greatly increase the rate of
oxidation of adsorbed ferrous ions.
Mineral iron pyrite (Ward's Natural Science Establishment,
Rochester, New York) was ball-milled and a portion of the 200-250 mesh
fraction was suspended in a solution of ferrous iron at pH 3.0. The
system was left open to the atmosphere and treated in the same fashion
as the suspensions containing clays, aliquots being removed, acidi-
fied, millipore-filtered, and titrated with permanganate. The solid
material recovered was rinsed with dilute acid to remove any adsorbed
ferrous iron. No significant decrease in Fe(ll) was observed after
3t
Researchers at Bituminous Coal Research, have found that some
activated carbons have catalytic/ properties for ferrous iron oxidation
while others do not.
-------
5-26
several days although in some instances the concentration of Fe(ll)
increased slightly. Upon placing the system under an atmosphere of
nitrogen, a similar small increase in ferrous iron was observed, indica-
tive of a gradual dissolution of the pyrite itself.
Effect of Microorganisms
tv<*£
No microbial contaminations^observed in any of the above systems
as evidenced by similar rates of oxidation in both sterile and nonsterile
samples. The sterile samples gave, in every case, analogous results to
those conducted under nonsterile conditions. The above findings are,
therefore, the same as those obtained under sterile conditions.
Table 5-2 summarizes the results of the catalytic studies with
the various chemical agents . The greatest influence on the rate of
oxidation of Fe(ll) was exhibited by the clay particles or their
idealized counterparts, alumina and silica. The areal concentrations,
however, are extremely large and are probably much greater than those
encountered in most natural waters .
5-6 Field Investigations of Pyrite Oxidation in Natural Mine Waters
In order to compare these experimental results describing the
kinetics of oxidation of ferrous iron in synthetic mine waters with
the rate of oxidation in natural mine drainage waters , field investi-
gations were conducted in the bituminous coal region of West Virginia.
The Federal Water Pollution Control Administration has established a
demonstration project in the Norton-Coalton area in north-central West
-------
5-27
Table 5-2. Chemical Catalysis of Oxidation of Ferrous Iron
pH
3.0
3.5
3.8
4.0
log k"
Uncatalyzed
-3.8
-3.6
-3.4
-3.3
Catalysis By
10 "^ S04"2
at 50°C
-3.1
—
—
—
10 ~4M
Cu+2
-3.4
—
—
—
A12°3*
8000 m2/!.
—
-2.5
-2.1
-1.8
Si02t
3000 m2/!.
—
—
—
-2.2
Bentonite#
10 gms/1.
—
—
—
-2.2
k" =
-d log [Fe(Il)]
dt
*Baymal colloidal alumina, E. I. DuPont de Nemours and Co.,
Wilmington, Delaware
*Ludox SM colloidal silica, E. I. DuPont de Nemours and Co.,
Wilmington, Delaware
^Fisher Scientific Company, U.S.P., Fair Lawn, New Jersey
Virginia, near Elkins, where various methods of pollution abatement
technology are being examined. These attempts include air-sealing of
mines and surface reclamation.
5-6.1 Collection and Analyses of Samples
Three sites were selected for the field investigation. Figure
5-10a shows an abandoned strip (surface) mine which is scheduled for
-------
5-28
Highwall
Mine wall
streams Yellow
Wooded
area
(Steep downhill
grade)
To Tygart
Valley River
FIGURE 5-10a.
Aerial view of drainage through a strip (surface) mine in
the Mercer seam near Elkins, West Virginia. The points
(•) refer to locations at which samples were collected
for laboratory analyses.
-------
5-29
reclamation. Surface runoff flows through a vegetated area, over a
high-wall, and into the abandoned mine. In mining terminology, this
worked-out mine is designated as a "strip pit"; the piles of waste
material previously dug out are termed "spoil banks"; and the precipi-
tated ferric hydroxide or sulfato-hydroxo intermediate is referred to
as "yellow boy." Drainage water flowing through the mine deposits
thick sediments of yellow boy at each juncture with drainage water
from the spoil banks which are laden with pyritic material. The water
exits the strip pit at site 7 and flows rapidly downhill through a
wooded area to Roaring Creek, the main interceptor for this mining
region.
Figure 5-10b represents a contour strip mine where drainage water
passes between the mine wall and a spoil bank. Water which has drained
through the underground mine trickles out of the mine opening at site 2.
The third sampling location, Figure 5-10c, is a mine entrance
which has been sealed with concrete blocks and urethane foam to exclude
air in an attempt to inhibit oxidation of pyrite inside the underground
mine. Measurements by the FWPCA indicate that the partial pressure of
oxygen behind the wall has been reduced to 77. (36). Drainage water
flows out of the mine over a weir, and is exposed to oxygen of the
atmosphere. After leaving the mine tunnel, the water enters a re-
claimed area which has been limed in preparation for future vegetation.
Deposits of yellow boy are abundant behind the weir and on the floor
of the tunnel.
-------
5-30
Hlghwall
Yellow boy
Mine opening - source of
acidic discharge
Mine
boundary
Water goes
underground
FIGURE 5-10b.
Aerial view of surface mine near Elkins, West Virginia
(Site GT 7-2). The points refer to sampling locations.
Underground
mine
7% 0,
Urethane foam
and concrete blocks
to seal mine
Weir
Atmospheric
oxygen
Mine floor
Drainage
Yellow boy'
deposits
water
Water flows
out of mine
tunnel
FIGURE 5-10c.
Profile of air-sealed entrance to- underground mine near
Elkins, West Virginia (Site RT 9-11) .
-------
5-31
Samples from the three mining areas were analyzed for ferrous
and total iron, sulfate, and acidity; temperature and pH were re-
corded in the field at the time of collection. Those samples taken
for iron analyses were acidified with dilute acid upon collection,
to quench any further reaction. Ferrous and total iron were deter-
mined utilizing bathophenanthroline (26). The burbidimetric pro-
cedure described in Standard Methods for the Examination of Water and
Wastewater (37) was employed in the analysis for sulfate, the percent
transmittance of the suspension of barium sulfate being measured using
a Bausch and Lomb spectrophotometer (Model #340). The method of
Salotto, et al (38), in which the sample is rapidly titrated to pH 7.3
with standard sodium hydroxide, after addition of hydrogen peroxide,
was employed for the determination of acidity.
5-6.2 Results of Field Investigation
Stoichiometric Relationship Between Sulfate Concentration and
Acidity
It was indicated in section 5-2 that the stoichiometry accom-
panying the dissolution of iron pyrite should conform to a definite
pattern, the oxidation of one mole of pyrite causing the release of
two moles of sulfate and four equivalents of acidity. Therefore, if
sulfate and acidity were to behave in a conservative manner in mine
waters, the change in either one or'both of these products could be
directly correlated to the quantity of pyrite dissolved.
A summary of the data obtained by analysis of the field samples
is given in Table 5-3. For site GT7-2, a large increase in acidity,
sulfate, and iron is seen at sampling site 4 after water from the mine
-------
Table 5-3. Summary of Field Data
Sample
Number
GT 7-2 1
3
4
5
6
2
i . — _— •••••••
•••••{•••••i^^^^™™™™^^^^
MS - 1
—2
3
4 I
5
6
7
8
9
RT 9 -11
.1!
Distance,
Feet
0
50
75
300
500
^ — ~
0
100
300
425
550
800
1000
1200
1400
-
Temp. ,
°C
13
16
17
18
15
12
14
19
20
--
18
20
18
15
16
--
PH
5.2
4.2
3.7
3.7
3.8
2.8
3.9
3.6
3.4
3.4
3.4
3.3
3.2
3.5
3.3
3.2
Fellj
x 105M
0.12
0.15
1.74
1.67
0.80
7.92
0.07
3.49
5.63
6.50
8.54
5.94
15.5
11.3
7.90
78.6
FeT
x 105M
0.22
0.25
4.07
2.71
1.67
98.9
0.12
5.68
13.8
10.7
11.3
11.9
42.1
38.4
34.3
132.
-2
sv
x 10 M
0.76
0.86
2.49
4.44
4.32
29.0
-
7.1
19.9
25.8
33.8
34.3
33.7
57.8
49.0
56.8
78.0
Acidity
x 104 eq/1
i i
1.8
.0
. /
5.0
4.5
41.7
awMOBsnaiaBKas
7.5
24.8
26.0
21.7
22.4
23.1
56.5
4o. ->
60 .0
49.8
I
to
-------
5-33
opening has drained into the stream. The same observation can be
made in the drainage from the Mercer Seam immediately downstream
from the spoil bank. The low PH of the surface runoff upstream from
the mines should be noted. Due to the low buffer capacity of the
water, the mere introduction of organic acids from the soil and vege-
tation is sufficient to significantly lower the pH of the raw water.
Figure 5-11 is an attempt to correlate acidity to concentra-
tion of sulfate, the lines drawn representing the theoretical 2/1
ratio. As shown, the field data do not conform to such a correlation,
the acidity being too low. Some of the acidity is apparently being
neutralized by clays which are invariably present in the drainage
waters and coal strata. The presence of dissolved aluminum in the
acidic waters is indicative of decomposition of the clays and neu-
tralization of acidity by the clay-water interaction. Hence, sulfate
alone may be considered to behave in a conservative fashion in acid
mine waters, in contrast to acidity and iron, and therefore it may be
used, by itself, as an indicator of the quantity of pyrite dissolved.
Rate of Oxidation of Ferrous Iron.
Of the mining areas sampled, RT 9-11 and the Mercer Seam proved
to be the most informative. RT 9-11 served as a source of mine water
having a high concentration of ferrous iron, while drainage from the
Mercer Seam provided an opportunity by which the change in concentra-
tion of Fe(Il) could be followed thorough a strip mine and downstream
in an attempt to deduce the rate of oxidation of iron.
-------
5-34
50
40
30
Drainage":
from mine
opening
GT 7-2
I I J
0 2 4 6 v 20 30 40 50
SULFATE CONCENTRATION, x 10 M
FIGURE 5-lla. Stoichiometric relationship between
acidity and sulfate concentration
in drainage water at location GT 7-2.
a1 4
«
X J
•\
B2
a2
o
1
0
SULFAT.E CONCENTRATION, x'10 M
FIGURE 5-llb. Stoichiometric relationship between acidity
and sulfate concentration in drainage through
strip mine i«-Mercer seam.
-------
5-35
Comparison with Laboratory Results
The data from Table 5-3 are plotted in Figure 5-12 showing the
change in chemical composition of a drainage stream flowing through
the Mercer mine and downstream before entering Roaring Creek. For the
initial 1000 feet, the water continues to collect leachate from the
spoil banks and from the pyrite with which it is in contact. The in-
crease in the concentration of sulfate from 0 to 400 feet is about 3 x
lO'2!-!. This implies that the corresponding input of ferrous iron would
increase by approximately 1.5 x lO'^l bearing in mind the stoichiometry
that one mole of pyrite releases two moles of sulfate and one mole of
ferrous iron. However, most of this Fe(ll) is apparently oxidized quite
rapidly as evidenced by the abundant deposition of yellow boy and by
the low concentration of Fe(Il) observed in the stream (only 8 x 10 M
at 550 feet).
Beyond 1000 feet, the drainage stream flows outside the mine and
is no longer subject to continuous pyritic discharges, or to dilution
from additional drainage streams. Figure 5-13 shows the oxidation of
Fe(Il) in the non-pyritic wooded area to be essentially first-order in
Fe(ll) with distance. If one assumes a constant velocity of flow over
this distance in the range of 1 ft. per min. to 100 ft. per min., the
range of the first-order rate constant, k", is 7.4 x 10" to 7.4 x 10
min.'1, where k" = -d log [Fe(ll)]/dt. This corresponds to k"-1 x 1(T
day"1 which is orders of magnitude greater than the laboratory experi-
ments at PH 3.3 would predict. The velocity of flow is probably not
constant but for the sake of comparison, in order to underscore the
rapidity of the reaction in natural mine waters, it is a useful approxi
mat ion.
-------
JS
t
o
CM
I
CO
0
6
>
%
s 2
0
40
20
5-36
300
600
900
1200
DISTANCE, ft.J
SULFATE
ACIDITY
TOTAL IRON
0" 300 600 900 1200
FIGURE 5rl2. Chemical composition, of drainage water through strip
mine in Mercer seam.
-------
5-37
£
n
O
ss
o
w
I
**• 7
M '
-------
5-38
Figure 5-14 presents the results obtained at the air-sealed
mine opening at KT 9-11. Curves 2 and 3 show the decrease in Fe(Il)
by oxidation after collection of the samples and allowing them to
stand back in the laboratory exposed to the atmosphere. Aliquots were
removed at various intervals and titrated with permanganate, sample
number 1 being acidified at the time of collection to serve as a con-
trol. Samples 2 and 3 were not acidified.
The linearity of the arithmetic plot indicates the oxidation of
Fe(ll) in its native solution to be zero-order in Fe(Il), i.e., the
instantaneous rate is independent of the concentration of ferrous iron.
The rate of reaction is dramatic when compared to the previous labora-
tory investigation.
Implications of Field Results
The zero-order nature of the oxidation reaction is suggestive
of a biological reaction in which substrate is non-limiting and in which
the concentration of microorganisms remains relatively constant. The
rate equation for such conditions has been derived in Appendix E as
-dS M, B
^__ _ _292— _ constant (5-16)
dt y
where S is the concentration of substrate (source of energy), in
this case ferrous iron.M is the maximum specific growth rate for
max
the microorganisms, v_ is the yield of microorganisms per unit of sub-
strate utilized, and ji is the instantaneous concentration of microorgan-
isms, assumed here to be constant. Figure 5-14 satisfies equation 5-16
and therefore the results imply that the oxidation is being catalyzed
by microorganisms which are utilizing the energy derived from the
-------
5-39
oxidation of ferrous iron for cellular metabolism, i.e., catalysis
by autotrophic bacteria is taking place.
Microbial catalysis, however, is not the only explanation for
the zero-order nature of the observations. A heterogeneous reaction
mechanism could be invoked, involving complete saturation of the solid
phase with the reactant, Fe(Il), in order to account for the zero-order
dependence on Fe(Il). However, for the additional reasons described
below, the autotrophic explanation is an extremely plausible one.
As shown in Appendix F, a thermodynamic free energy balance
appears to negate the existence of such autotrophic iron bacteria; only
1 gram of organic carbon is synthesized for every 250 gms of ferrous
iron oxidized. In fact, the autotrophic nature of these organisms
and their ecological significance was initially doubted due to the
meager amount of free energy available from the oxidation of Fe(ll).
Since the energy released is so small, one would not expect
the oxidation of 10~ M Fe(ll) to significantly change the bacterial
concentration if a sufficient number of microorganisms were present in
the mine water at RT 9-11, i.e., JB should remain constant. Figure 5-14
reflects such reasoning. Kim (39), of the Pittsburgh Mining Research
Center of the Bureau of Mines, has obtained similar zero-order plots
of oxidation of Fe(ll) in natural mine waters.
If, however, the concentration of bacteria were diminished, as
by filtration of the mine water, B^ would be expected to increase
logarithmically (see equation E-13 in the Appendix) as the substrate,
Fe(Il), is utilized. A sample of mine water from RT 9-11 was millipore-
filtered (0.8 u pore diameter) and the resultant degree of oxidation of
-------
5-40
Fe(Il) was markedly reduced due to removal of a significant fraction of
the bacteria. The unf iltered sample was completely oxidized within 20
hours, while the filtered sample displayed a lag before significant oxi-
dation began (see Figure 5-15). If JB is not constant during the course
of the reaction, then
-** "maxn "max11 (5-17)
~dt = T o e
and, as derived in Appendix E
B ju t
s _ s = _° e "^ (5-18)
o y
where B is the initial concentration of bacteria, and S is the ini-
tial concentration of substrate, Fe(II). Taking logs of both sides one
obtains
log (S -S) = log 2. + J. (5-19)
o y £*• j
Figure 5-16 is a semilogarithmic plot of the change in concentration of
Fe(ll) with time, the linear nature of the plot confirming the logarith-
mic change in B in accordance with equations 5-17 and 5-19. Equation
5-19 indicates that the slope of the semilog plot is equal to max/2. 3.
Consequently, "max is 0.076 hrs" for the experimental conditions.
The generation time, the time required for the concentration of bacteria
to double, is, by definition, equal to In 2/JU or 9.1 hours. Silverman
and Lundgren (18) observed generation times of about 7.0 hours in their
laboratory under ideal experimental conditions.
To further substantiate biological significance, sterile solu-
tions of ferrous sulfate were inoculated with various amounts of acid
-------
5-41
mine drainage. Two sterile controls were maintained: one in which
aseptically filtered mine water (220 mu pore diameter) was added to
sterile solutions of ferrous sulfate; and another containing sterile
ferrous sulfate alone. Aliquots were removed aseptically and analyzed
for residual Fe(ll) by titration with permanganate. Figure 5-17
shows a linear decrease in Fe(ll) for the arithmetic plot, as in
Figures 5-14, but only for the specimen containing microorganisms. The
slowness of the oxidation reaction in the laboratory samples compared
to that in the field samples (Figure 5--17 compared to Figure 5-14) may
have resulted from a decrease in the concentration of some essential
growth factor in the course of the dilution.
Therefore, the oxidation of ferrous iron occurs more rapidly in
natural mine waters than in any of the synthetic solutions investigated
in the laboratory subject to the various chemical catalytic additives.
The rapidity of the reaction in nature is apparently the result of
microbial catalysis, as evidenced by Figures 5-14 to 5-17.
-------
5-42
8
X
^ 6
55
s
w 4
§
\
0 10
FIGURE 5-15.
sampled from weir and
millipore filtered (0.8(j.)
sample from weir,
unacidified and unfiltered
20 30 40 50
TIME, hrs.
Oxidation of ferrous iron in water
collected from air-sealed mine.
20 30
TIME, hrs.
FIGURE 5-16. Change in ferrous iron concentration in milli-
pore filtered water collected from air-scaled
mine.
-------
5-43
12
10
I
3
g
w
o
u
M
M
(U
solutions inoculated with
millipore filtered (0.22^)
mine water
solutions
inoculated with
untreated mine water
4
160
FIGURE 5-17.
80 120
TIME, hrs.
Oxidation of ferrous iron solutions inoculated
with mine water from RT 9-11.
200
-------
5-44
5-7 Oxidation of Iron Pyrite
The kinetics of oxidation of ferrous iron in the mine water
system (sections 5-5 and 5-6) and the kinetics of hydrolysis of ferric
iron (Chapter 4) having been studied, the final step in the sequence
of chemical reactions describing the dissolution of pyritic agglomer-
ates is the oxidation of pyrite itself, both by oxygen and by ferric
iron. With the exception of the study by Garrels and Thompson (7),
prior investigations of the oxidation of pyrite have been concerned
almost entirely with oxygen as the oxidant. In many cases, the poten-
tiality of ferric iron as an oxidant has been overlooked. It was de-
sired to obtain some idea as to the rate of oxidation of pyrite by
ferric iron relative to that by oxygen.
5-7.1 Experimental Procedures
Several dilutions of a stock solution of ferric perchlorate were
prepared and adjusted to pH 1.0 with perchloric acid. (Sato (6) and
Garrels and Thompson (7) noted the oxidation of pyrite to be indepen-
dent of pH below 2.0.) Nitrogen was bubbled continuously through the
solution to remove all traces of oxygen, any oxidation of pyrite then
being attributable only to the action of ferric iron. Iron pyrite from
Rico, Colorado (Ward's Natural Science Establishment, Rochester, New
York) was ball-milled and screened, and the 200-250 mesh fraction was
selected for the experiment. At time zero, various amounts of the
finely-divided pyrite were added to the solution of ferric iron, the
solid phase being uniformly dispersed by means of a magnetic stirrer.
The electrochemical potential of the system was measured at various
-------
5-45
intervals using a platinum spiral indicator electrode and a calomel
reference electrode saturated with NaCl, as described in section 2-3.3.
A Leeds and Northrup potentiometer (Cat. No. 7664) was employed for
the potential measurements.
The procedure adopted was similar to that used in Chapter 4.
In a well-defined system such as this, the reversible potential is
O *3
established by the electroactive Fe - Fe+ couple in accordance with
the Nernst Equation
E = E° - .0592 log(Fe * at 25°C (5-20)
At PH 1.0, since [Fe+2] = [Fe(H)3T and [Fe"1"3] = [Fe(IIl)] , E° can
easily be computed in a constant ionic medium by measuring the poten-
tial, E, and independently determining the total concentrations of
ferrous and ferric iron. By conducting all future studies under the
same experimental conditions of constant ionic strength and tempera-
ture (to insure that the activity coefficients remain constant), this
value for the equilibrium potential, E , can be utilized to compute the
o o
ratio [Fe ]/[Fe ] by simply measuring the potential of the system.
After the potential was recorded, aliquot s of the suspension
were removed and filtered. The concentration of total iron in the fil-
trate was determined using the bathophenanthroline procedure in which
10% hydroxylamine is utilized to reduce all ferric iron to the ferrous
state. The individual concentrations of Fe(ll) and Fe(lll) were cal-
1 O
culated knowing the concentration of total iron and the [Fe ]/[Fe ]
ratio from the potential measurements.
-------
5-46
5-7.2 Results and Discussion
Rate of Oxidation in Absence of Oxygen
The rate of disappearance of ferric iron in the presence of
finely -divided iron pyrite is remarkably rapid. Since the rate of
oxidation of pyrite is independent of pH below pH 2.0 (6) (7), a
simple rate law dependent upon the two reactants can be assumed, of
the form
. d £Fe(IIl)] = k [Fe(iii)]m[FeS_]n (5-21)
at *-
If the concentration of pyrite is large compared to that of Fe(lll),
[FeS ] will remain relatively constant during the course of the reac-
tion. (Note also, from the stoichiometry of the reaction (equation
5-4) f that the oxidation of one mole of pyrite consumes fourteen moles
of ferric iron.) Therefore, under these conditions, the rate of oxi-
dation can be approximated by
_ d CFe(lIl)] = UeC m)]» (5-22)
at 1
where
= k [FeS2]n (5-23)
and is constant. Furthermore, if m = 1 and the reaction is first-
order in the concentration of Fe(lll), a plot of log [Fe(lll)j versus
time should be linear. Figure 5-18 shows that the decrease in Fe(lll)
conforms to such a formulation. It is also seen that the rate of de-
crease of Fe(IIl) is a function of the concentration of pyrite, pre-
sumably the surface concentration; for the same initial concentration
of Fe(lll), the half-time is about 250 minutes in the presence of
-------
•
•
\ •!
50
100
250
300
350
150 200
TIME, min.
FIGURE 5-18. Reduction of ferric iron by iron pyrite (200-250 mesh) in the
of oxygen. «-n
<
• j
-------
5-48
1 gm/1 of FeS7, and only 25 minutes when 5 gms/1 of pyrite are pre-
sent.
Taking the logarithm of equation 5-23, one obtains
log kx = log k + n log [FeS2] (5-24)
A log-log plot of the pseudo-first-order rate constants (computed from
the slopes of the straight lines in Figure 5-18) against the concentra-
tion of pyrite should yield a straight line of slope £, the order of
the rate-dependence on FeS_, if the assumed rate law, equation 5-21, is
valid. Two such plots are shown in Figure 5-19 for two different ini-
tial concentrations of Fe(lII). The order, n, is seen to be about 1.3
and 1.1, implying a first-order-dependence of the rate on the concen-
tration of iron pyrite.
If the assumed rate law is valid and the first-order-dependence
upon Fe(lll) is correct, as demonstrated in Figure 5-18, the rate con-
stant k1 should be independent of the initial concentration of Fe(lll).
This is not the case, however, as seen in Figure 5-20. In fact, the
slope of the semilog plot increases as the initial concentration of
Fe(lll) decreases, implying an inverse dependence of the rate on Fe(lll),
An understanding of the kinetics of the reaction is further complicated
by the findings of Garrels and Thompson (7) that the instantaneous rate
of reduction of Fe(lll) decreases with tirae (implying direct kinetic
dependence on Fe(IIl)), but the average rate of reduction is inde-
pendent of the initial concentration of Fe(lll) (suggestive of zero-
order dependence on Fe(lll)). Furthermore, they present a figure
similar to Figure 5-20, showing essentially a logarithmic decrease in
-------
5-49
-3.0
c
•H
B
-2.5
-2.0
[Fe(III)]o= 3.4x10
pH 1.0
1.3
-4
-1.0
-4.0
-3.5
-3.0
a
•H
e
o
3
-2.5
-2.0
-1.0
(a)
-0.5 0.0
LOG FeS2 (gms/1)
0.5
(b)
1.3xlO"3M
pH 1.0
1-1
d log fFedlPl
dt
-0.5
0.5
1.0
1.0
0.0
LOG FeS. (gms/1)
FIGURE 5-19a,b. Rate of reduction of ferric iron ad'a function
of pyrite concentration.
-------
I
§
5-50
lo
-3
10
-4
ID
'5
[Fe(III))o = 2.65x10
k, - 2.7xlO"4 min"1
[Fe(III)] = 1.28xlO"3M
k, - 1.6xlO~3 min"1
.
\
\
1.0 gm/1 FeS,
i
pH 1.0
\
3.26xlO"4M
_ \ kl * 1'2xl°
\
\
\
A
\
"2
100
400
500
200 300
TB4E, min.
FIGURE 5-20. Effect of initial concentration of ferric iron
on rate of reduction^of Fe(III) by pyrite, in
the absence of oxygen.
-------
5-51
Fe(IIl) with time, with the rate of decrease becoming steeper as the
initial concentration of Fe(lll) decreases.
(The stoichiometry of the reaction (equation 5-4) was verified
by noting a Tk increase (15/14) in the concentration of total dissolved
iron (Fe(ll) and Fe(IIl)), using the bathophenanthroline procedure.)
Since the mechanism of the oxidation of pyrite by Fe(lll) was
not the primary purpose of this study, but rather the relative rate
of the reaction as compared to that of the oxidation of Fe(ll) was of
major concern, it will be sufficient to note the rapidity by which
Fe(lll) is reduced by pyrite. Even for the slowest case observed,
where [Fe(llD] ~ 10~3M and [FeS^ = 0.12 gms/1 (~ 10" M), the
half-time is approximately 2 days which is considerably less than that
for the oxidation of Fe(ll) even when accelerated by any of the experi-
mental chemical catalysts found in natural waters.
Oxidation Rate in Presence of Oxygen
After finding that there is no appreciable adsorption of dis-
solved Fe(Il) on iron pyrite, and that Fe(ll) is not catalytically
oxidized in the presence of FeS (see section 5-5.2), the rate of
reduction of Fe(lll) by pyrite in the presence of oxygen was investi-
gated. The experimental procedure was identical to that in the pre-
ceding section except that the system was left open to the atmosphere.
The results are shown in Figure 5-21, indicating that there is vir-
tually no difference between the rate of reduction of Fe(lll) by pyrite,
or the rate of change of soluble Fe(ll), under aerobic or anaerobic
conditions. Hence, the implication is that even in the presence of a
partial pressure of oxygen of 0.2 atm. the oxidant of iron pyrite is
ferric iron. Fe(lll) at pH = 1 oxidizes FeSg faster than Og,
-------
5-52
Increase in Fe(II):
under one atmosphere of nitrogen
in the presence of atmospheric
oxygen
change in Fe(III) in the presence
of atmospheric oxygen
in Fe(III) under
tmosphere of nitrogen
[Fe(III)] ^3x10
1.0 gm/1 FeS
80 120 160
TIKE, ruin.
FIGURE 5-21. Reduction of ferric iron and increase in
ferrous iron in the presence and absence of oxygen,
-------
5-53
5-8 Conclusions
5-8.1 Model Describing Pyrite Oxidation and Pollution by Coal
Mine Drainage
In accordance with the experimental results presented in this
chapter, the following model is proposed to describe the oxidation of
iron pyrite in natural mine waters:
(+ 0 )
Initiation Reaction; FeS^Cs) ^Fe(ll) + S - compound (5-24a)
Propagation Cycle;
f v\fast
- ^ y T * ***** /} ^ ** ' \J — fc*rD j C /
Fe(OH)3(s) (5-24d)
The reactions shown are schematic and do not represent the exact
mechanistic steps. The model is similar to and carries with it the
same overall consequences as that suggested by Temple and Delchamps
(14). In this model, the rate-determining step is a reactive step in
the specific oxidation of ferrous iron, reaction 5-24b. As this in-
vestigation has demonstrated, the rate of oxidation of ferrous iron
under chemical conditions analogous to those found in mine waters is
very slow, indeed considerably slower than the oxidation of iron py-
rite by ferric iron, reaction 5-24c. At pH 3, half-times for the oxi-
dation of Fe(Il) are on the order of 1000 days while in the case of
oxidation of pyrite by Fe(lll), half-times on the order of 20 to 1000
minutes were observed.
-------
5-54
Reaction 5-24a serves only as an initiator of the overall re-
action: ferrous iron may be released by simple dissociation of the
pyrite, or by oxidation of the pyrite by oxygen. Once the sequence
has been initiated, a cycle is established in which ferric iron
rapidly oxidizes pyrite and is slowly regenerated through the oxy-
genation of the resultant ferrous iron, reactions 5-24b and c. Oxygen
is involved only indirectly, in the regeneration of Fe(Hl).
Precipitated ferric hydroxide deposited in the mine and the
streams serves as a reservoir for soluble Fe(lll), by reaction 5-24d.
If the regeneration"of Fe(lll) by 5-24b is halted so that the concen-
tration of soluble Fe(IIl) decreases, it will be replenished by dis-
solution of the solid Fe(OH)_ and will be free to act again should it
*
come in contact with additional FeS2> The presence of sulfate in
solution increases the concentration of dissolved Fe(lll) in equili-
brium with the precipitate, by complex formation (see section 2-3.3).
Barnes and Romberger's argument (10) that there is insufficient Fe(lII)
available for the reaction appears fallacious.
Smith, et al (13) investigated the effect of pH on the rate of
oxidation of pyrite by oxygen and observed the reaction rate to be
relatively independent of pH below pH 4, while the rate increased
rapidly in a pH-dependent manner above pH 7. Since this parallels the
pH-dependence of the rate of oxidation of ferrous iron (Figure 5-2),
this writer contends that Smith, et al actually observed the oxida-
tion of pyrite by Fe(lll). At higher pH-values, the rate of oxida-
tion of FeS_ increases because the rate of formation of Fe(lll), via
oxidation of Fe(ll), increases with increasing pH. These results lend
further support to the proposed model. ^^__
_ _
The contact with Fe(llll however may be small if the pyrite
lies on the ceiling or on walls.
-------
5-55
The following pertinent consequences of this model need to be
emphasized:
1) Ferric iron cannot exist for long in contact with pyritic
agglomerates. Fe(III) is rapidly reduced by iron pyrite.
2) The elimination of oxygen is of no consequence with regard
to the specific oxidation of iron pyrite. However, the exclusion of
oxygen does inhibit regeneration of Fe(lll) through the oxidation of
Fe(ll), and will be of significance once the supply of available Fe(lll)
is exhausted.
3) The overall rate of dissolution of pyrite is independent of
its surface structure. Interference with the surface of ^pyrite, such
as the application of inhibitors which are adsorbed at the solid-
solution interface, is inconsequential since the oxidation of pyrite
is not the rate-limiting step. If, on the other hand, such a technique
could make the rate of oxidation of pyrite less than the rate of oxi-
dation of Fe(H), then such a control measure may have merit.
4) Microorganisms can only be influential by mediating the
oxidation of ferrous iron since it, alone, is the rate-determining
step. Catalysis of the specific oxidation of iron pyrite by micro-
organisms, even if it could be unequivocally demonstrated, can have
no effect on the overall rate of dissolution of iron pyrite.
It is probably this same cycle which is responsible for the
dissolution and leaching of other mineral sulfides as found in copper
and uranium mines. Microbial leaching of these other minerals has
always been demonstrated in the presence of iron, pyrite being the
most abundant and widespread of all mineral sulfides (2). Until it
-------
5-56
can definitely be proven otherwise, cycle 5-24 adequately accounts for
the observed microbial leaching of other mineral sulfides, and direct
microbial oxidation must be discounted.
The solution to the problem of acid mine drainage, therefore,
appears to rest with methods of controlling the oxidation of ferrous
iron. Due to the cyclical nature of the process describing dissolu-
tion of pyrite, mere treatment of the resulting drainage water will
allow the problem to compound and magnify itself. Hence, at-source
control measures are preferred. One such method might involve the
inhibition of natural catalysts which are of significance in accele-
rating the rate of oxidation of Fe(Il) in mine waters. It is of
primary concern to discover which of these catalysts causes the oxi-
dation to proceed as rapidly as it does in nature. From the prelimi-
nary analysis presented in this chapter, the major chemical catalysts
appear to be alumino-silicate clays, but only at considerably larger
surface concentrations than would be expected in natural mine waters.
Microbial catalysis, as by the autotrophic iron-bacteria Thiobacillus
and Ferrobacillus ferrooxidans, seems to be ecologically significant
as evidenced by the few field investigations conducted. Numerous ac-
counts of autotrophic iron oxidation prevail, but only few quanti-
tative reports of their actual activity in nature have appeared.
2 3
Tuttle, Randies, and Dugan (40) observed only 10 to 10 iron-
*
oxidizers per ml. in an acid mine stream, using an MPN technique.
This is not a significant concentration when one considers the limited
amount of iron which can be oxidized by these few microorganisms (see
Appendix F). However, there may be considerable surf ace ^growth, of
^
These counts are probably not representative of the concentra-
tion near the pyrite surface.
-------
5-57
these microorganisms associated with the available solid material
suspended or deposited in the drainage streams. It is important to
know the actual concentration of these auto trophic microorganisms
found in mine waters in order to assess their relevance regarding
the rate of oxidation of ferrous iron. Once the role of all cata-
lytic agents has been evaluated, methods can then be devised to con-
trol the oxidation of ferrous iron and, hence, the dissolution of iron
pyrite and the introduction of acidity into natural mine waters.
-------
5-58
References
1) Palathe, C., Berman, H., and Frondel, C., Dana' s System of
Mineralogy, 7th ed., vol. 1, John Wiley and Sons, Inc., New
York, 1944.
2) Clark, C. S., "Oxidation of Coal Mine Pyrite," Journ. San. Eng.
Div., Proc. Amer. Soc. Civil Eng.. 9_2, 127 (1966).
3) Krauskopf, K. B. , Introduction to Geochemistry, Ch. 18, McGraw-
Hill Book Company, New York, 1967.
4) Stokes, H. N., "On Pyrite and Marcasite," U. S. Geol. Surv. Bull.
186 (1901).
5) Nelson, H. W., Snow, R. D., and Keyes, D. B., "Oxidation of
Pyritic Sulfur in Bituminous Coal," Ind. Eng. Chem., 25_,
1335 (1933).
6) Sato, M. , "Oxidation of Sulfide Ore Bodies. II. Oxidation
Mechanisms of Sulfide Minerals at 25 C," Econ. Geol., 55,
1202 (1960).
7) Garrels, R. H., and Thompson, M. W., "Oxidation of Pyrite by
Iron Sulfate Solutions," Amer. Journ. Sci. , 258-A, 57 (1960).
8) McKay, 0. R., and Halpern, J., "A Kinetic Study of the Oxidation
of Pyrite in Aqueous Suspension," Trans. Met. Soc. AIMS, 212,
301 (1958).
9) Gerlach, J., Hahne, H., and Pawlek, F., "Beitrag zur Drucklangung
von Eisensulfiden. II. Zur Kinetik der Drucklangung von
Pyrit," Zeit. Erzb. Metall., .19., 66 (1966).
10) Barnes, H. L., and Romberger, S. B., "Chemical Aspects of Acid
Mine Drainage," Journ. Wat. Poll. Contr. Fed., 40_, 371 (1968).
11) Smith, E. E.} Svanks, K., and Shumate, K., "Sulfide to Sulfate
Reaction Studies," Proc. 2nd Symp. on Coal Mine Drainage
Research, Coal Industry Advisory Committee to ORSANCO,
Pittsburgh, May 1968.
12) Colmer, A. R. , and Hinkle, M. E., "The Role of Microorganisms in
Acid Mine Drainage: A Preliminary Report," Science, U)6_, 253
(1947).
13) Temple, K. L., and Colmer, A. R., "The Autotrophic Oxidation of
Iron by a New Bacterium: Thiobacillus Ferrooxidans," J. Bact.,
62, 605 (1951).
-------
5-59
14) Temple, K. L., and Delchamps, E. W., "Autotrophic Bacteria and
the Formation of Acid in Bituminous Coal Mines," Appl. Micro-
biol., 1, 255 (1953).
15) Leathen, W. W., Kinsel, N. A., and Braley, S. A., "Ferrobacillus
Ferrooxidans: A Chemosynthetic Autotrophic Bacterium," J. Bac t.,
22, 700 (1956).
16) Kinsel, N. A., "New Sulfur-Oxidizing Iron Bacterium: Ferro-
bacillus Sulfooxidans SP.N." J. Bact., 80, 628 (1960).
17) Unz, R. F., and Lundgren, D. G., "A Comparative Nutritional Study
of Three Chemoautotrophic Bacteria: Ferrobacillus Ferrooxidans,
Thiobacillus Ferrooxidans, and Thiobacillus Thiooxidans," Soil
Science, 92, 302 (1961).
18) Silverman, M. P., and Lundgren, D. G. , "Studies on the Chemo-
autotrophic Iron Bacterium Ferrobacillus Ferrooxidans: I.
An Improved Medium and Harvesting Procedure for Securing High
Cell Yields," J. Bact., 77, 642 (1959).
19) Silverman, M. P., and Lundgren, D. G., "Studies on the Chemo-
autotrophic Iron Bacterium Ferrobacillus Ferrooxidans. II.
Manometric Studies," J. Bact., 78, 325 (1959).
20) Schnaitman, C. A., "A Study of the Mechanism of Iron Oxidation by
Ferrobacillus ferrooxidans," Ph.D. Thesis, Syracuse University
(1965).
21) Brynner, L. C. , Beck, J. V., Davis, D. B., and Wilson, D. G. ,
"Microorganisms in Leaching Sulfide Minerals," Ind. Eng. Chem.,
46_, 2587 (1954).
22) Brynner, L. C. , and Anderson, R. , "Microorganisms in Leaching
Sulfide Minerals," Ind. Eng. Chem., 49_, 1721 (1957).
23) Ehrlich, H. L. , "Bacterial Action on Orpiment," Econ. Geol., 58_,
991 (1963).
24) Silverman, M. P., and Ehrlich, H. L., "Microbial Formation and
Degradation of Minerals," Advances in Appl. Microbiol., _6_,
153 (1964).
25) Ehrlich, H. L, , "Observation on Microbial Association with Some
Mineral Sulfides," p. 153 in Biogeochemistry of Sulfur Iso-
topes, M. L. Jensen, ed., Nat'l Sci. Found. Symp., Yale Univ.,
New Haven, Conn. (1962).
26) Lee, G. F., and Stumm, W., "Determination of Ferrous Iron in the
Presence of Ferric Iron," Journ. Amer. Wat. Works Assoc., £2
1567 (1960).
-------
5-60
27) Huffman, R. E., and Davidson, N., "Kinetics of the Ferrous Iron-
Oxygen Reaction in Sulfuric Acid Solution," Journ. Amer. Chem.
Soc., 78, 4836 (1956),
28) George, P., "The Oxidation of Ferrous Perchlorate by Molecular
Oxygen," Journ. Chem. Soc., p. 4349 (1954).
29) Cher, M. , and Davidson, N. "The Kinetics of the Oxygenation of
Ferrous Iron in Phosphoric Acid Solution," Journ. Amer. Chem.
Soc., TTj 793 (1955).
30) Crabtree, J. H., and Schaefer, W. P., "The Oxidation of Iron (II)
by Chlorine," Inorg. Chem., _5_, 1348 (1966).
31) Stumm, W., and Lee, G. F., "Oxygenation of Ferrous Iron," Ind. Eng.
Chem.. 53_, 143 (1961).
32) Weiss, J., "ElektronenUbergangsprozesse in Mechanismus von Oxyda-
tions - und Reduktions-Reaktionen in LOsungen," Naturwissen-
schaften, 23, 64 (1935).
33) Lamb, A. B. , and Elder, L. W., "The Electromotive Activation of
Oxygen," Journ. Amer. Chem. Soc., 53, 137 (1931).
34) Van Olphen, H., An Introduction to Clay Colloid Chemistry, Inter-
science Publ., New York (1963).
35) Schenk, J. E., and Weber, W. J., "Chemical Interactions of Dis-
solved Silica with Iron (II) and (III)," Journ. Amer. Wat.
Works Assn., 60_, 199 (1968).
36) Scott, Robert, Personal Communication, Project Engineer, Federal
Water Pollution Control Administration, Elkins, West Virginia
(1968).
37) standard Methods for the Examination of Water and Wastewater, llth
ed., American Public Health Assn., Inc., New York (1960).
38) Salotto, B. V., Earth, E. F., Ettinger, M. B., and Tolliver, W. E.,
"Determination of Mine Waste Acidity," submitted to Environ.
Sci. and Tech. (1967)
39) Kim, A. G., Personal Communication, Pittsburgh Mining Research
Center, U. S. Bureau of Mines, Pittsburgh, Pa. (1968).
40) Tuttle, J. H., Randies, C. I., and Dugan, P. R., "Activity of
Microorganisms in Acid Mine Water," Journ. Bact., 95, 1495
(1968).
-------
CHAPTER 6
CONCLUSIONS
Information concerning the chemistry of aqueous iron can be ap-
plied to natural water systems in order to cope with the vast and
costly problem of acid mine drainage, for the design of efficient de-
ferrization processes, and to provide for a better understanding of
the limnological cycles of a number of key elements. This chapter sum-
marizes the relevant results realized in this research, and discusses
the practical consequences of these results.
6-1 Principal Findings
1) Ferrous iron is thermodynamically unstable in the presence
of oxygen. The rate at which it is oxidized to ferric iron is compa-
tible with a rate law which is first-order in both [Fe(ll)3 and [02],
and second-order in [OH~] at pH-values above 4.5. At lower pH-values,
the rate of oxidation is independent of pH. The reaction proceeds
extremely slowly in the acidic region, but is catalyzed by inorganic
and organic ligands capable of forming complexes with Fe(ll) and
Fe(lll), heavy metal ions, clay particles, and bacteria.
2) The oxidation of ferrous iron is the specific rate-
determining step in the oxidation of iron pyrite and the subsequent
discharge of acidity into mine drainage waters. The chemical oxygena-
tion of Fe(Il) at pH-values less than 4 takes place very slowly, even
in the presence of the many chemical catalysts which are operative in
-------
6-2
nature. The direct oxidation of iron pyrite by ferric iron is quite
rapid and Fe(in) serves as the prime oxidant of iron pyrite. A
cycle is established involving the rapid oxidation of pyrite by
Fe(lII), and the slow regeneration of Fe(lll) through the oxygenation
of Fe(ll), in the following schematic way:
fast 2
FeS + Fe(lll) - *• Fe(ll) + SC>4
slow
Fe(ll) + 0 - »-Fe(lIl)
fast 7
FeS2 + Fe(IIl) - ** Fe(ll) + SO^
Microorganisms, presumably the autotrophic "iron bacteria," markedly
enhance the rate of oxidation of Fe(ll), and, therefore, accelerate the
overall rate of pyrite oxidation. Oxygen participates in the cycle
only in the regeneration of spent ferric iron.
6-2 Practical Consequences and Implications Resulting From This Research
1) The proposed cycle to describe the oxidation of iron pyrite
and the acidification of mine drainage waters can be utilized in evalu-
ating some of the methods recommended for the prevention of acid mine
drainage, and should serve as a guide to indicate the direction in
which future control measures should be aimed.
Mine-sealing and the application of bactericides are two pro-
posals, both directed at the retardation of the oxidation of ferrous
iron. The former involves the elimination of air and/or water from the
mine and is intended to stop the reaction entirely. The latter is
aimed at destroying the catalytic agent which is responsible for the
-------
6-3
rapid rate at which Fe(ll) is oxidized in mine drainage. Previous asser-
tions that oxygen serves as the specifijg, direct oxidant of iron pyrite
are unjustified in view of this research which indicates that oxygen is in-
volved only indirectly, producing ferric iron which is, itself, the primary
oxidant of FeS . Furthermore, this research has shown that in the absence
of microbial catalysis, the oxygenation of Fe(ll) is sufficiently slow that
it is essentially halted. The regeneration of Fe(lll) proceeds so slowly
that its effect on pyrite is of little consequence.
Both mine-sealing and the application of "bactericides, however, are
subject to similar limitations. Firstly, neither will result in the immedi-
ate cessation of acidic discharges. Ferric iron must first be flushed out
of the mine, or rendered inactive, before any beneficial effects could be
realized. Reports of previous mine-sealing operations demonstrate that in
spite of significant reductions in the concentration of oxygen, there was
not always a marked change in the quantity of acid released although the
discharge had a higher concentration of ferrous iron (l). In the few suc-
cessful mine-sealing operations, often little improvement in the quality
£
of the effluent from the mine was observed for several years (2). These
observations can readily be interpreted in view of the proposed scheme;
the elimination of oxygen stops the oxidation of ferrous iron but the py-
rite is still subject to oxidation by the vast quantity of Fe(lll) availa-
ble in the mine. Only after the active Fe(lll) is depleted should one ob-
serve a decrease in the acidity of the drainage water.
There are instances where mine sealing was followed by reduced
acidity. For a recent review see R. D. Hill, Acid Mine Water Control,
presented before the Mining Environmental Conference, University of
Missouri, Rolla, Missouri. April, 1969-
-------
6-4
Secondly, from a practical standpoint, both methods are of
doubtful applicability. Total exclusion of oxygen is unlikely due
to the many fractures and minute cracks in the mine wall and the fre-
quent collapses experienced in the mine itself. Application of a
bactericide requires continual injection of the agent into the system
at a suitable location where the entire inflow could be treated. In
most mining areas, such a location is non-existent. There are some
situations, however, where a bactericide could be employed. Although
it is not feasible to treat the mine itself, the various spoil banks
exposing pyrite previously dug out of the mine are amenable to such
treatment as they can be reached rather easily. It is conceivable,
also, that strip mines would yield to such treatment since the influ-
ent water can usually be located.
Another proposal concerns the addition of an alkali to the
system in order to raise the pH of the environment within the mine,
thus inhibiting the microorganisms which catalyze the oxidation of
ferrous iron. Here again, a suitable location for treatment is re-
quired. Furthermore, if the pK is raised appreciably, this study has
shown that the chemical oxygenation of Fe(ll) will proceed sufficiently
fast by itself.
The introduction of organic material to chemically reduce ferric
iron and sulfate has also been suggested in order to promote preci-
pitation of ferrous sulfide. The continuous addition of organic mat-
erial is necessary and the elimination of oxygen is mandatory so that
Fe(ll) and S(-II) are maintained in their reduced states.
-------
6-5
Inactivation of the pyrite surface through the application of
chemical inhibitors which are adsorbed at the solid-solution interface
is unfeasible and inconsequential. Since the specific oxidation of
pyrite is not the rate-determining step, partial coverage of the py-
ritic surface would not affect the overall rate of pyrite oxidation.
In addition, constant exposure of new pyritic surfaces would be expected
as a result of the frequent collapses inside the mine.
This discussion has considered only control methods for the
abatement of pollution by coal mine drainage. Proposals for the treat-
ment of effluent waters from these mines include acid neutralization by
the addition of lime, reverse osmosis, and ion exchange. Each produces
a voluminous or a concentrated waste which must ultimately be disposed of.
This research suggests that treatment of acid mine drainage re-
quires treatment of the cycle by which such acidic discharges arise. Al-
though no specific schemes for such treatment are proposed as a result of
this study, it is emphasized that the catalytic oxidation of ferrous iron
need to be halted. Because of the inaccessibility of the pyrite oxida-
tion site, it is difficult to convert this theoretical suggestion into a
practical treatment. (Catalytic oxidation of ferrous iron, however, is
an asset, where mine drainage is treated for iron removal.}
2) This research has shown that the concentration of ferrous
iron in natural groundwaters can often be predicted from considerations
of solubility relationships, specifically the solubility product of
ferrous carbonate which was obtained in this study, i.e., pKgo = 10.2k
at 25°C and zero ionic strength. The expected concentration of Fe(ll)
-------
6-6
can be calculated if the pH and alkalinity of the groundwater are
monitored. The kinetic relationships describing the oxygenation of
Fe(ll) have also been derived. If the solubility and kinetic relation-
ships are coupled, they can be applied to the design of an efficient
iron-removal facility in order to bring the raw water into conformance
with the Public Health Service drinking water standards for iron
(0.3ppm) (3). This assumes an efficient filtration system for the
actual removal of the resultant hydrous ferric oxide. For example, a
groundwater at pH 6.5 containing 5 x 10~ eq/1 alkalinity should con-
tain approximately 1.7 ppm of dissolved Fe(ll) (approximately 3 x 10 M),
and would therefore require 857. oxidation of the Fe(ll) in order to
provide an acceptable finished water having 0.3 ppm of iron. At a
partial pressure of oxygen of 0.20 atm. and at 25 C, a detention time
of about 60 minutes would be necessary (toc = log (85/15)/k") if the
o.)
oxidation were to proceed at pH 6.5. Furthermore, iron-removal may
be aided by precipitation of Fe(ll) as ferrous carbonate.
In addition, this study has demonstrated the capability of
many elements in natural waters to accelerate the oxidation of Fe(ll).
Retardation of the oxidation of ferrous iron in the presence of oxy-
gen can, therefore, be attributed almost entirely to the presence of
reducing agents in natural systems, notably organic matter. Ferric
iron is a potent oxidant of organic material, getting reduced to
Fe(ll) in the process. The net effect appears as an inhibition of
the rate of oxidation of Fe(ll), whereas in fact, the Fe(Il)-Fedll)
reaction serves as an electron-transport mechanism between oxygen and
the organic material. The stability of ferrous iron in the epilimnion
-------
6-7
of lakes and reservoirs, in the presence of oxygen, can, in a similar
manner, only be interpreted in terms of a. steady-state condition
maintained by the two concurrent oxidations: the oxidation of Fe(Il)
by oxygen and the oxidation of the organic matter by Fe(lll).
References
1) Scott, R., Project Engineer, Federal Water Pollution Control
Administration, Elkins, West Virginia, Personal Communication
(1968)
2) Moebs, N. N., "Mine Air Sealing: A Progress Report," Proc. Second
Svmo. Coal Mine Drainage Res, , Coal Industry Advisory Committee
to ORSANCO, Pittsburgh, May, 1968
3) United States Public Health Service Drinking Water Standards,
Publication No. 956, Washington (1962)
-------
APPENDIX A
Correction of Experimental Solubility Data for Temperature and Activity
The experimental equilibrium relationship, given by equation
(2-lla), is
+2^
CH+] 8q KC2
the superscript c_ referring to equilibrium constants at a given ionic
strength. The corresponding thermodynamic equilibrium constant for the
reaction, at 25°C and zero ionic strength is
The two equilibrium constants are related by the equation
U-3)
where the Y ' s are single ion activity coefficients. Schindler (1)
has suggested that for carbonates of bivalent metals in a constant
ionic medium similar to 0.2M HaCl(>4> the Davies equation
should be applied for the computation of activity coefficients. I is
the ionic strength of the system, z is the charge of the specific ion
under consideration, and A is a constant equal to 0.509 for water at
25°C.
-------
A-2
Taking logarithms of equation A-3, one obtains
pK0^ - log tf - log V + log * = pK (A-5)
eq Fe L HC03" H+ q
where p- refers to the negative logarithm of that term. Substitu-
tion of the Davies equation into A-5 gives
0.509 | n . /r ~ "•-••'•I \ *• n
Fe HC03~ H*
pK (A-6)
eq
If the proper charges for the ionic species are used, the equation
describing the experimental system at 0.1M NaCIO, reduces to
pK° + 0.107 (4 + 1 - 1) = pK (A-6a)
F eq F eq
or
pKC + 0.428 = pK (A-6b)
eq F eq
Taking logarithms of the right-hand equality in equation A-2, one
obtains
pK = pK - pK0 (A-7)
*^ eq * so * 2
Substitution of this quantity into A-6b gives, after rearrangement
pK = pK° + 0.428 4- pK_ (A-8)
so eq L
From the plot of p H versus -log[Fe ][HCO_ ] in accordance
c c
with equation 2-12, the intercept at p H = 0 is equal to pK
Figure 2-5 shows the desired intercept to be -0.57. In the determina-
tion of pK_ as a function of temperature, by Harned and Scholes (2),
it Is found that, at 22.5°C, pK_ = 10.35. Substitution into A-8 gives
-------
A-3
PK = -0.57 + 0.43 + 10.35 (A-8a)
so
-O,
pK^ = 10.21 at 22.5°C (A-8b)
or K » 6.1 x 10"11
so
The solubility product can readily be converted to 25°C utili-
zing the van't Hoff temperature relationship
K ° T - T
, ~
( -— — - 1 (A-9)
Kl R VT1T2 /
where K_ and K.. are the equilibrium constants at the absolute tempera-
tures T£ and T^, respectively. R is the ideal gas constant, equal to
1.987 cal./mole-K, and AH is the change in enthalpy for reaction 2-4,
equal to -4639 cal./mole at 25°C (3). (The literature value for AH° was
employed since the experimental temperature -dependence was not suffi-
ciently precise to extract a usable change in enthalpy for the reaction.
Only three experimental points were available for such a calculation,
giving AH° . . = 1800 1 1200 cal./mole.)
experimental
Substitution of these values into A-9 yields
K25°C - ^ZZ.sV1-068 (A-9a)
which, combined with A-8b, gives
K = 5.7 x 10"U (A-10)
SO
pK = 10.24 at 25°C
K so
the desired thermodynamic solubility product.
-------
A-4
References
1) Schindler, P. W., "Heterogeneous Equilibria Involving Oxides,
Hydroxides, Carbonates, and Hydroxide Carbonates," Ch. 9,
p. 196 in Equilibrium Concepts in Natural Water Systems,
R. F. Gould, ed., Advances in Chemistry Series 67, Amer. Chem.
Soc., Washington (1967)
2) Harned, H. S., and Scholes, S. R. , "The lonization Constant of
HCO," from 0 to 50°," Journ. Amer. Chem. Soc., 63, 1706 (1941)
3) Latimer, W. E. , The Oxidation States of the Elements and Their
Potentials in Aqueous Solutions, 2nd ed. , Prentice-Hall Inc.,
Englewood Cliffs, N.J., (1952)
-------
APPENDIX B
Relative Significance of Soluble Phosphato-Complexes of Fe(lll)
Consider the following equilibria, the equilibrium constants of
which were taken from Stability Constants (1):
Fe+3 + H20 = FeOH+2 + H+ ^ - 10'2'2
Fe+3 + 2H20 = Fe(0l02+ + 2H+ ^ = 10'6-8
+ -2 „ in7-2
H + HP04 = H2P04 K12 = 10
Fe+3 + HP04"2 = FeHP04+ ^ l = 1Q8'3
Fe+3 + H2P04- =FeH2P04+2 ^2 = 103'5
+ + _ po +2 K ^ = 102'4
6nrvJ4 ~ 2 4 122
VoPQ.. . = Fe + PO ~ K = 10~
(B-l)
(B-2)
(B-3)
(B-4)
(B-5)
(B-6)
(B-7)
J4(s)
Figures B-l and B-2 show pH-log concentration diagrams for phosphate,
and for Fe(lll) in the absence of phosphate, respectively.
Using equation B-6
4 . [H+] K12 - . 10-CH*] (B-8)
[FeHP04+3 1Z 91
one sees that at pH-values greater than 2.4, FeHP04+ is the predominant
soluble phosphato-complex of Fe(IIl). The following relationships
should also be noted:
-------
B-2
-7
02 4 6 8 10 12
PH
FIGURE B-l. Log concentration diagram for phosphoric acid,
-12
6 8
pH
FIGURE B-2. Distribution diagram £or soluble nonoroeric hydroxo-
species of ferric iron.
-------
B-3
A) At pH>5, Fe(OH) "*" is the predominant soluble species of
Fe(IIl). Hence, by equation B-2,
[Fe(OH)2+] KXK2 1Q-6.8
[Fe+3] " CH+]2 " CH+]2
B) In the pH-range 2.5 to 4.5, FeOH1" predominates. Using
(3-9)
equation B-l,
Fe
[Fe+3] CH+]
[FeOH+2] Kl IP"2'2
«
C) At pH> 7.5, [HP04" ]^PT, the total concentration of dis-
solved phosphate. By equation B-4,
CFeHPO.+2] QO .
- - £ - = ^ LPT = lO8'3 x PT (
D) In the pH-range 2 to 7, using equation B-3,
p
CHPO"2]=— - - , since [H PO "] ^ P . (B-12)
4 + L
Hence, by equation B-4,
2] iA_.10l.l A- (B-13)
[Fe+33 CH+]K12 CH+]
Making use of relationships (B-9) through (B-13) , one can pre-
pare the following table showing the relative abundance of the various
soluble complex species (assuming P = 10 M) :
-------
B-4
pH
[Fe(OH)2+]
[Fe+3]
io9-2
7 2
lO''^
105'2
103'2
101-2
[FeOH+2]
[Fe+3]
—
--
io2-6
101'6
io°-6
[FeHP04+]
CFe+3]
io4-3
4 1
10
io3a
lo2-1
401'1
10o.i
8
7
6
5
4
3
It is,therefore, apparent that the known soluble phosphate-
complexes of Fe(lll) become influential only below pH 4, but to a
very limited extent.
References
1) Sillen, L. G., and Martell, E- A., Stability Constants of Metal-
Ion Complexes, Special Publication No.17, London, The Chemical Society
(1964)
-------
APPENDIX C
Derivation of Relations Between Redox Potential and Sulfate Concentration
for Determination of Stability Constant for FeSO^+
The reactions pertaining to this study are
Fe+3 + SO ~2 = FeSO.+ K. (C-l)
44 1
HS04" » H+ + S04"2 KA (C-2)
Fe+3 + H0 = FeOH+2 + H+ Q (C-3)
In a system containing ferrous and ferric iron at 25 C and a
constant ionic medium of 0.1M NaCIO , the redox potential is defined
by the Nernst Equation
E = E° - 0.0592 log Ssll (C-4)
[Fe+3]
It will be convenient to refer to the system in the absence of sulfate
as cell 1 so that
[Fe+2L
E, = E°. - 0.0592 log - r-i (C-5)
1 l CFe+3]1
If the pH is maintained below 3, the total concentration of ferric
Iron is given by
CFeClII)-], = CFe+3L + [FeOH+2] =LFe+3^l + -J-) (C-6)
T 1 i l [H+]
and that of ferrous iron is
[Fe(H)T]1 = CFe+2J1 (C-7)
-------
C-2
Upon addition of sulfate, let us refer to the system as cell 2, where
TFe+2]
E_ = E°- - 0.0592 log - - r-2- (G-8)
2 2 [Fe+3]2
In the presence of sulfate,
[Fe(IIl)T32 = [Fe+3]2 + [FeOH+2]2 + [FeSO^ (C-9)
or, substituting equations C-l and C-3,
CFe(IIl)_]9 = [Fe+3],(l + — £- + K. [SO ~2]) (C-10)
T 2 Z [H+:I 1 ^
where only the monosulf ato-complex of Fe(lll) is assumed, and
]2 = [Fe+2]2 (C-ll)
Sulfate, however, reacts with water in the acidic pH-range. so that
CS°4"2]ADDED - ST = CS04"2] + [HS°4"] + ^FeS04+] (C"12)
If it is assumed that [FeSO^] « [S0^~2] + [HSO^"], then, using
C-2, equation C-12 becomes
ST = [S04"2](l + ^1) (C-13)
f\
Rearranging C-13
•> KAST
CSO. ~Z] = -^ - (C-13a)
and substituting into C-10, one obtains
KKS
-u
[Fe+:](l
KA
A
-------
C-3
The difference in potential between the cell before and after
sulfate is added can be obtained by subtracting equation C-8 from
C-5 to give
lFe+2]?[Fe+3L
E. - E0 . E = 0.0592 log s-^ — (C-15)
1 £. r-w, +^T
since the standard potentials are equal, i.e., E - » E _. It is seen
also in this step that the liquid junction potentials for the two
cells cancel out. The experiment has been designed such that
[Fe*2], = DFe*2].. Hence,
*• L
or
j.
[Fe+].
E » 0.0592 log - =-± (C-15a)
[Fe+3]2
[Fe"1"3],
(c-15b)
Furthernore, by experimental design, [FeClIl),^ = [Fe(lIl)T]2 so that
equation C-6 can be set equal to C-14,
[Fe*3], (l + A.) . [Fe+3]2(l + A- + ^&_) (C
1 CH+] z CH+] KA + [H+]
assuming the concentration of H+ to remain constant with the addition
of sulfate. Equation C-16 can be rearranged and set equal to C-15b
so that
Considering the equality on the right, after rearranging and simplifying,
we obtain
-------
C-4
KA
+
CH+] v °'0592 y KA + CH+]
- l)- *l
y
(C-17b)
Equation C-17b is now in a usable form. If the potential and pH are
measured as a function of the sulfate added, and it is known that Q.
= 2.89 x 10"3 at 25°C and 0.1M NaClO^ (1), the left-hand side can be
plotted against ST, the slope of the resultant straight line being
K1KA
. Knowing K , the second acidity constant of sulfuric acid,
K. + [H+] A
A
and the pH at which the study was conducted, one can compute the sta-
bility constant for the monosulfato-complex of Fe(lll).
References
1) Milburn, R. M., "A Spectrophotometric Study of the Hydrolysis of
Iron (III) Ion. III. Heats and Entropies of Hydrolysis,"
J. Amer. Soc.. 79, 537 (1957)
-------
APPENDIX D
Thermodynamic Stability of Iron Pyrite
The change in free energy for the oxidation of iron pyrite by
oxygen, using the data available in Latimer (l), is
FeS2(s) * l°2 + H2° = Fe+2 + 2S°4~2 + ^ (D"1
AF° (-39.84) I (0) (-56.7) (-20.3) 2C-177.34) 2(0)
S. AF° = -278.4 kcal.
The reaction should take place spontaneously, iron pyrite being thermo
dynamically unstable in the presence of oxygen.
References
1) Latimer, W. M. , The Oxidation States of the Elements and Their
Potentials in Aqueous Solutions, second ed., Prentice-Hall, Inc.,
Englewood Cliffs. New Jersey (1952)
-------
APPENDIX E
Kinetics of Microbial Growth (1)
The change in concentration of microorganisms, B_, with time is
first -order in concentration of microorganisms,
~ = «B (E-l)
at
where ju is the specific growth-rate constant. For an enzymatic pro-
cess, the growth-rate is given by the Michael is -Men ton equation,
n S
max fv n\
" - K~TT (E-2)
m
where ^ is the concentration of substrate (source of energy), max
j^
is the maximum growth-rate, and m is the Michael is -Men ton constant.
The change in concentration of microorganisms upon utilization of the
substrate is
where Y_ is defined as the yield. Equations E-l, E-2, and E-3 can be
combined to give the change in substrate with time as
A» 3
-dS max _ _ /-,
"dT =(K + S) Y '
m
If the concentration of substrate is large compared to the Michaelis-
Menton constant, i.e., if substrate is non-limiting, the specific
growth-rate is constant and equal to the maximum growth-rate, so that
E-4 simplifies to
-------
E-2
Ai B
-dS _ max (E-5)
dt " Y
the number of the microorganisms increasing logarithmically in accord-
ance with equation E-l. If B remains relatively constant during the
course of the oxidation reaction, then
•^— = constant (.E-bJ
dt
If, on the other hand, the concentration of microorganisms changes
considerably, then equation E-l can be integrated
'B|B=Al 0 dt (E-7)
'B /o
°B = B eWt
-------
E-3
B
s _s . J. e
o Y
/u t)
V, max /
or, taking logarithms of both sides of the equation, one obtains
log (S -S) = log 5° + JSISI (E-13)
O X £* O
These equations (E-6 and E-13) can now be used from a biologi-
cal viewpoint as models to account for the fate of Fe(ll) in natural
waters.
References
1) Monod, J., Recherces sur la Croissance des Cultures Bacteriennes,
Hermann and Cie, Paris (1942)
-------
APPENDIX F
Autotrophic Iron Bacteria - Ratio of Ferrous Iron Oxidized to Organic
Carbon Synthesized
The free energy released by the oxidation of ferrous iron,
using the data available in Latimer (l), is
4 Fe+2 + 02 + 4 H+ = 4 Fe+3 + 2 H 0 (F-,1)
AF° 4C-20.3) (0) 4(0) 4C-2.53) 2(-56.7)
£AF° = -42.4 K cal. or
-10.6 K cal./mole of Fe(ll) oxidized.
At pH 3, since
AF =AF° + RT In Q (F-2)
where Q is the reaction quotient, the free energy released per mole is
AF = -10.6 + 1.364 log —~ = -6.5 K cal./mole (F-2a)
10 "J
For synthesis of cell material from CO- (assuming the assimi-
lated end product to be glucose), the free energy required is
6 C02 +6 H20 = C6H12°6 * 6 °2 (F"3)
AF° 6C-92.3) 6C-56.7) (-217.0) 6(0)
2AF = -!- 677 K cal./mole of glucose or
+115 K cal./mole of carbon synthesized.
Assuming a 367» efficiency for microbial conversion of energy as
is common in auto trophic processes (2), the stoichiometry of the
-------
F-2
autotrophic oxidation of ferrous iron is
6.5 K cal./55.8 gms. Fe(ll) oxidized Q -& 1 ,_.
115 K cal./l2 gms. carbon assimilated X 250
Hence, one gram of organic carbon is synthesized for every 250 gms
of Fe(ll) oxidized.
If one considers the thermodynamic relationships in another
way,
6.5 K cai.
1 mole of Fe(II) oxidized
115 K cal./mole of carbon
0.36 x Sm% — (F-5)
mole of carbon
= 0.25 gms of carbon synthesized
Lamanna and Mallette (3) approximate that 1 gm. of bacteria
1 o 10
contains 10 to 10 bacterial cells. Therefore, 1 mole of Fe(Il)
12
yields approximately 10 bacterial cells or
Y = ^1 = 1012 cells/mole of Fe(ll) oxidized (F-6)
dS
References
1) Latimer, W. M., The Oxidation States of the Elements and Their
Potentials in Aqueous Solutions, second ed., Prentice-Hall,
Inc., Englewood Cliffs, New Jersey (1952)
2) McCarty, P. L., "Thermodynamics of Biological Synthesis and
Growth," in Advances in Water Pollution Research, J. K. Baars,
ed., vol. 2, Pergamon Press, New York (1965)
3) Lamanna, C., and Mallette, M. F., Basic Bacteriology, third ed.,
The Williams and Wilkins Co., Baltimore (1965)
V. S. GOVERNMENT PRINTING OFFICE : 1970 O - 401-741
------- |