OCR error (C:\Conversion\JobRoot\000007P8\tiff\2000SUG4.tif): Unspecified error

-------
                   .JPPi'IFPLAE S f-ARPlJ. 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  Hater
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.

A tripilicate abstract card sheet is included in the
report to facilitate information retrieval.  Space is
provided on the card for the user's accession number and
for additional uniterms.

Inquires pertaining to Water Pollution Control Research
Reports should be directed to the Head,  Project Reports
System, Planning and Resources Office, Office of Research
and Development, Department of the Interior, Federal
Water Quality Administration, Room 1108, Washington, D. C.
20242.

-------
     Sulfide to Sulfate  Reaction Mechanism
A Study of the  Sulfide  to Sulfate Reaction Mechanism
 as it Relates  to  the Formation of Acid Mine Waters
                  Ohio State University
                   Research Foundation
                  Columbus,Ohio   43210
                         for the

     FEDERAL WATER POLLUTION CONTROL ADMINISTRATION

                 DEPARTMENT OF THE INTERIOR
                     Program Number
                FWPCA Grant No. 14.010 FPS
                      February 1970
       For sale by the Superintendent of Documents, U.S. Government Printing Office
                   Washington, D.C. 20402 - Price $1.50

-------
This report has been reviewed by the Federal
Water Pollution Control Administration and
approved for publication.   Approval does not
signify that the contents necessarily reflect
the views and policies of the Federal Water
Pollution Control Administration.
                       ii

-------
                               ABSTRACT

                    THE SULFIDE TO SULFATE  REACTION

                    by E.  E.  Smith and K. S.  Shumate

     A detailed study of the mechanisms  and kinetics of the chemical
reactions responsible for acid mine drainage has been made.

     The mineralogical features of the solid phase reactant (pyrite)
that determine its reactivity were described.  The rate-limiting reac-
tions and variables affecting the rate of these reactions were identi-
fied.

     It was found that two basic oxidation modes are important:   oxygen-
ation, in which oxygen is the immediate  oxidizing agent; and ferric ion
(or microbiologically catalyzed) oxidation, in which ferric ions are
the oxidants.  From a knowledge of the dissolved oxygen, ferric/ferrous
ratio, and total iron ion content at the reaction site, the reaction
regime can be determined.

     Kinetic equations were derived for  both reaction modes.  From these
basic relationships the oxidation rate in real pyritic systems can be
accurately predicted when conditions at  the reaction site are known.

     For a given pyrite surface, oxygenation rate is, for all practical
purposes, dependent only on the oxygen concentration in the aqueous
phase surrounding the reactive site on the  pyrite surface.  Ferric ion
oxidation rate is determined by the ferric/ferrous ratio and free ferric
ion concentration in solution, and is not affected by dissolved oxygen.

     In a real system, the ferric/ferrous ratio is determined by a
"relative" microbial activity; i.e., the number of bacteria per  exposed
pyrite surface.  Bacterial activity is limited by pH and oxygen concen-
tration as well as nutrient levels.

     This report was submitted in fulfillment of Research Grant  No.
1^010 FPS between the Federal Water Pollution Control Administration
and The Ohio State University Research Foundation.
Key Words:   Mine Drainage/Coal Mine Drainage/Sulfides/Iron Sulfides/
            Pyrite/Ferrobacillus/Pollution Abatement/Industrial Wastes/
            Reaction Kinetics
                                 iii

-------
                            TABLE OF CONTENTS
Section 1     Conclusions and Recommendations

Section 2     Introduction

Section 3     Mineralogy

                 Classification of Pyritic Materials
                 Surface Area and Pore Size Distribution
                 Preparation of Sulfur Ball Samples
                 Identification of Intermediate Products

Section h     Oxygenation

                 Experimental Equipment
                 Results of Preliminary Runs
                 Discussion of Preliminary Runs

                    Role of Water
                    Oxygen Concentration

                 High Pressure Oxygenation
                 Effect of pH
                 Effect of Various Anions and Cations

Section 5     Microbiology

                 Preliminary Experimental Phase

                    Apparatus and Growth Media
                    Results
                    Conclusions

                 Warburg Respirometer Oxidation Experiments

                    Objectives
                    Materials and Procedures
                    Results

Section 6     Ferric Ion Oxidation

                 Experimental Equipment
                 Experimental Procedure
                 Experimental Results

                    Low pH, High Sulfate Runs
                    Treatment of Data
                    Interpretation of Data
Page

  1

  3

  7

  7
 10
 10
 12

 13

 13
 15
 18

 18
 20

 22
 27
 29

 31

 31

 32
 35
 38

 38

 39
 39
 39
 51
 52

 52
 52
 58

-------
                    TABLE OF CONTENTS - (Continued)
                 Combined Oxygenation and Ferric  Ion Oxidation
                 Varying pH, Type and Concentration of Anion

                    Ionic Composition of Reaction Solution
                    Iron Complexes in Sulfate Solution
                    Iron Complexes in Chloride Solution
                    Anion Concentration and pH Effect on
                     Kinetics

Appendix I    Related Studies

                 Determination of Ferric and Ferrous Ion Adsorp-
                  tion

                    Direct Measure of Adsorption
                    Adsorption from Streaming Current Measure-
                     ment
                    Equation for Streaming Current
                    Experimental
                    Conclusions

                 Application of the Potentiostat to Pyrite Oxi-
                  dation Studies
                 Inhibition Studies
                 Adsorption of Oxygen, Nitrogen,  and Water on
                  Pyrite

Appendix II   Calculation of Experimental Data for Ferric Ion
               Oxidation

Appendix III  Calculation of Ionic Composition
                 Sulfate  Solution

                     Determination of H+ Concentration
                     Calculation of HS04" Concentration
                     Calculation of Fe+++, Fe++,  and Their
                      Complexes

                  Chloride Solution

Appendix IV   Run Data, Ferric Ion Oxidation

Appendix V    Characteristics of Pyrite Samples

Acknowledgement s
 60
 62

 67
 67
 68

 69

 73


 73

 73

 73
 7^
 76
 76


 76
 77

 80


 91

 93
 93
 97

 98

103

107

109
                                  vi

-------
                     TABLE OF CONTENTS - (Continued)




                                                                    Page




References                                                           111




List  of Publications                                                113




Glossary of Terms                                                    115




Abstract Cards
                                  Vll

-------
                                FIGURES

Figure

   1       Pore Volume Distribution, Museum Grade Pyrite and
           Sulfur Ball                                              11

   2       Vapor Phase Oxidation Apparatus                          ih

   3       Liquid Phase Oxidation Apparatus                         lU

   h       Variable Pressure Apparatus                              ih

   5       Oxidation Rate vs. Partial Pressure of Water             19

   6       Oxidation Rate vs. % Relative Saturation                 19

   7       Oxidation Rate vs. Temperature                           19

   8       Oxidation Rate vs. Oxygen Concentration in Aqueous
           Phase                                                    19

   9       Oxidation Rate vs. Oxygen Concentration                  25

  10       Oxidation Rate vs. Mole % Nitrogen in Liquid             26

  11       Oxidation Rate vs. pH                                    28

  12       Culture Apparatus                                        33

  13       Bacterial Count and Oxygen Uptake Rate vs. Time          36

  lU       Oxygen Uptake Curve                                      UO

  15       Oxygen Uptake Rate                                       Ul

  16       Change in Cell Population with Time                      U3

  17       Change of Microbial Oxidation Rate with Time             kk

  18       Oxygen Uptake Rate vs. Total Cell Population             H5

  19       Oxygen Uptake Rate vs. Pyrite to Water Ratio             Vf

  20       Ferric Ion Oxidation Apparatus                           50

  21       Reciprocal Rate vs. Reciprocal (Ferric Concentration)'72   56

  22       Rate vs. MF, McDaniels Sulfur Ball                      56

-------
                         FIGURES - (Continued)

Figure                                                             Page

  23       Rate vs. EMF for Three Pyrite Samples                     59

  2k       Rate vs. EMF in Sulfate Solution                         63

  25       Effect of Sulfate Concentration                          6k

  26       Comparison of Rates in Sulfate and in  Chloride
           Solution                                                 65

  27       Comparison of Rates in Sulfate and in  Chloride
           Solution                                                 66

  28       Reciprocal Rate vs. Reciprocal (Free Ferric  Ion
           Concentration)'72                                          70

  29       Potentiometer Setting vs.  Current                        78

  30       Pyrite to Reference Voltage vs.  Potentiometer
           Setting                                                  79

  31       Nitrogen Adsorption Isotherm on Museum Grade Pyrite
           at 77.3K                                                81

  32       Nitrogen Adsorption Isotherm on Sulfur Ball  at
           77-3K                                                   82

  33       Nitrogen Adsorption Isotherm on Museum Grade Pyrite
           at 25.0C                                                83

  3^       Nitrogen Adsorption Isotherm on Sulfur Ball  at
           25.0C                                                   8^

  35       Water Adsorption Isotherm  on Museum  Grade Pyrite at
           25.0C                                                   85

  36       Water Adsorption Isotherm  on Sulfur  Ball at  25.0C        86

  37       Oxygen Adsorption Rate on  Sulfur Ball  at 25C             87
                                  IX

-------
                                 TABLES

 Table                                                             Page

    I     Liquid Phase Runs                                          16

   II     Vapor Phase Runs                                           16

  III     Variable Pressure Runs  at  25 C                             17

   IV     High Pressure Runs  at 25C                                23

    V     Oxidation Rate vs.  pH                                      28

   VI     Rate vs. Concentration  at  EMF = 0.700                      53

  VII     Rate vs. EMF (Museum Grade Pyrite)                         53

 VIII     Rate vs. Iron Concentration (Sulfur Ball No. 2)            5*4-

   IX     Rate vs. Iron Concentration (McDaniels Sulfur Ball)        5k

    X     Rate vs. EMF (Sulfur Ball  Wo. 2)                           55

   XI     Combined Oxygenation and Ferric Ion Oxidation              60

  XII     Chemical Inhibition Tests                                  77

 XIII     Oxygen Adsorption,  Absorption and Desorption on
          Sulfur Ball                                               89

  XIV   '  Stability Constants for Sulfate Solutions                  97

   XV     Stability Constants for Chloride  Solutions                101

  XVI     Rate Data, Sulfate  Solution                              103

 XVII     Rate Data, Sulfate  Solution,  Iron Concentration
          Varies                                                   105

XVIII     Rate Data, Sulfate  Solution,  Sodium Sulfate Added        105

  XIX     Rate Data, Chloride Solution                              106

   XX     Rate Data, Chloride Solution, Iron Concentration
          Varies                                                   106

-------
                               Section 1

                    CONCLUSIONS AND RECOMMENDATIONS


                              Conclusions

     The following major conclusions may be drawn from information
obtained during the course of this project:

     1.  The principal mineralogical form of naturally occurring
         iron disulfide materials in Eastern coal mining areas  is
         pyrite.

     2.  The major factor influencing reactivity of  various  types
         of pyritic material is surface area (texture),  not  trace
         elements or included minerals.

     3.  The rate-determining reaction in "natural"  systems, in
         the absence of mass transport limitations,  is an electron
         transfer between the oxidizing agent (oxygen or ferric
         ion) absorbed on the surface of pyrite, and the pyrite
         itself.

     U.  There are two modes of pyrite oxidation as  determined  by
         the immediate oxidizing agent; i.e., oxygen or ferric
         ions.  Each oxidation mode is independent of the other.
         For a given pyrite surface, oxygenation rate is, for all
         practical purposes, dependent only on the oxygen concen-
         tration in the aqueous phase surrounding the reactive
         site on the pyrite surface and independent  of other solute
         concentration.  Ferric ion oxidation is primarily deter-
         mined by the ferric/ferrous ratio and free  ferric ion  con-
         centration, and is not affected by dissolved oxygen.

     5-  Kinetics for the oxygenation of pyrite have been experi-
         mentally determined as a function of (a)  temperature,
         (b) oxygen concentration, (c) pH, (d)  water partial pres-
         sure, (e) surface area (or texture), and (f) concentra-
         tion of iron, sulfate, and other ions.

     6.  Kinetics of pyrite oxidation by ferric ions have been
         determined as a function of (a)  ferric/ferrous  ratio,
         (b) total iron concentration, and (c)  free  ferric ion
         and hydrogen ion concentration (qualitatively).  A  mech-
         anism has been proposed, based on the competitive adsorp-
         tion of ferric and ferrous ions, which correlates experi-
         mental rate data.

-------
     7.  Ferric ion oxidation of pyrite is the chemical analogy
         of microbial-enhanced pyrite oxidation.

     8.  Significant microbial catalysis will occur only if the
         water-to-pyrite ratio is sufficiently high to provide
         for the required population of organisms.

     9-  Those factors which determine the reaction regime, either
         chemical or microbiological, have been described.

    10.  lyrite oxidation rates can now be accurately predicted
         in terms of known and measureable variables.  Basic rela-
         tionships are now available for calculating the oxidation
         rate under natural conditions, regardless  of oxygen con-
         centration, pH, bacteria, and other environmental condi-
         tions.  Major variables have been identified, and measure-
         ments necessary to permit calculation of oxidation rates
         described.
                            Recommendat ions

     With the information now available,  kinetics of pyrite oxidation
in well-defined laboratory systems are reasonably well known.   Applica-
tion of these basic data are hampered by lack of knowledge of  the physi-
cal, chemical, and biological conditions  at the reactive site  in real
pyritic systems.  Continuation and extension of basic field or "pilot"
scale studies in which the kinetics of formation, desorption,  and dis-
persal of oxidation products in a natural environment are being examined,
is recommended.

     Several aspects of the work described in this report require further
investigation.  The reactive specie in ferric ion oxidation has not been
positively identified.  Considering the reliability of stability con-
stants and experimental data, free ferric ions or one of the ferric
hydroxide complexes may be the reactive specie; neither can be elimin-
ated from consideration on the basis of available data.

     Additional information on the effect of pH and oxygen concentration
on microbial activity is needed.  The water-to-pyrite ratio at which
microbial catalysis becomes a significant factor in pyrite oxidation
should be better defined.  These data are needed in order to designate
more precisely the regime (oxygenation or ferric ion oxidation) that is
determined by environmental conditions.

-------
                               Section 2

                             INTRODUCTION

     Effective methods for alleviating acid mine drainage can best be
found and developed if an understanding of the kinetics and reactions
responsible for acid mine drainage is available.  Therefore the objec-
tive of this study was to determine the basic rate-limiting mechanism(s)
and kinetics of pyrite oxidation.

     The significance of this simple statement of objective can be
illustrated by assuming that the objective was achieved.  If this were
so, all the variables that influence the rate of oxidation would be
known, logical approaches to inhibiting the reaction could be deduced,
and evaluation of oxidation rates for different pyritic systems could
be precisely calculated in terms of the environmental conditions which
exist at the reaction site.

     It must be noted that a complete evaluation of a pyritic system
(i.e., strip mine, gob pile, drift mine, etc.) also requires a descrip-
tion of the environment and kinetics of the dispersal of oxidation pro-
ducts.  This expresses the relationship between this project and the
"pilot scale" or basic field studies of the type now being conducted at
The Ohio State University in which the environmental conditions at the
reaction sites are being examined.  Neither project can be effective
without the other.  The necessity of more basic information to help
plan experiments and evaluate field data has become more evident as
analysis of pyritic systems becomes more sophisticated.  For example,
the ferric/ferrous ratio and its relation to the kinetics of microbial-
catalyzed systems is an essential bit of basic information for inter-
preting data from a natural system.

     The experimental work performed under this project was limited to
laboratory-scale operations.  Physical rate-limiting parameters (other
than reactive surface area) were eliminated or controlled.  Experimental
conditions were varied to develop kinetics of pyrite oxidation, not
necessarily to duplicate "natural" environments.  Although not specifi-
cally related to any one real system, the data obtained from this pro-
ject provide the necessary input data for analysis of real systems.

     Chronologically, mineralogical investigations were performed con-
currently with a preliminary kinetic study to identify the reactants as
well as the physical, chemical, and biological parameters influencing
rate of pyrite oxidation.  Chemical (or direct oxidation by oxygen,
hereafter termed "oxygenation") as opposed to microbiological (oxidation
by ferric ions) reactions, were examined initially.  As the kinetics of
the oxygenation reactions were established, microbial-enhanced reactions
were compared to the chemical reactions from both a mechanistic and
kinetic reference.

-------
     Since we believe that systems having ferric ions as the oxidizing
medium are chemically analogous to microbiologically controlled systems,
the kinetics of ferric ion systems were studied in detail.

     The following report is separated into sections arranged in the
same order as emphasis developed in the course of the project.   This
is not to imply that these works are unrelated, but is presented in
this manner in order to most clearly discuss the results and signifi-
cance of these areas of research.

     To assist in following the subsequent discussions, it  may be help-
ful to keep in mind the purpose behind each major phase of  the project.

     Phase One (Mineralogy) was undertaken to describe the  materials
and their properties that influence the oxidation of pyrite.

     Phase Two (Oxygenation) was an attempt to define the chemical
kinetics of pyrite oxygenation; first, to determine the nature of the
rate-limiting reaction, and second, to quantitatively evaluate the
physical and chemical parameters that influence the oxidation reaction.

     The third phase (Microbiological, Ferric Ion Oxidation) was a
study of the kinetics of microbial-enhanced reactions, the  development
of an analogous chemical system, and a comparison of the rate-limiting
reactions for chemical and microbiological systems.

-------
                              MINERALOGY

     This phase of the study applied mineralogical concepts and tech-
niques to the study of iron sulfide oxidation.

     Iron sulfide exists in coal formations in two structural states 
pyrite and marcasite.  Various authorities differ on which is the most
common of the two types.  Inasmuch as the oxidation rate is strongly
dependent upon the structural state of the sulfide, the determination
of the -abundances of the two modifications was undertaken first.   Repre-
sentative sulfide samples were collected from coal "beds of various ages
in eastern and southeastern Ohio.  Of the 30 samples collected and
examined by x-ray diffraction, all but two were determined to contain
pyrite as the major constituent.  In the majority of the samples, mar-
casite was either not present, or was present in amounts less than five
per cent.  The details of this study are described by Birle.2

     It has been shown that marcasite has a considerably higher oxida-
tion rate than pyrite, and the high rate of oxidation of sulfides in
coals has been commonly attributed to marcasite.  Textural studies on
pyrite samples indicate that this mineral may in some cases show
extremely high reactivity as a result of textural features (such as
grain size, distribution of impurities, grain shape, etc).  Hence the
oxidation rates obtained by various investigators show considerable
differences, because of both textural and structural variations of the
iron sulfides.

     A survey of mineralogical literature reveals that the various crys-
tal faces of pyrite have different oxidation resistances, the octahe-
dron(lll) being most susceptible to attack.  A study was initiated to
determine, quantitatively, the rates for different crystal faces.

     Although most crystal faces consist of plane surfaces, it is gen-
erally necessary to polish these for detailed examination.  A mechanical
polish would, in almost every case, change the orientation of the sur-
face; in order to avoid this, the metallurgical technique of electro-
polishing was used.  The technique has been previously used only on
metals, and this constituted the first such approach on sulfides.  A
proper choice of solutions compositions and concentrations with the pro-
per electrical potential was established.  It is possible to control
the conditions under which pyrite can be made to polish or etch by vary-
ing the electrical potential.  Inasmuch as this technique, if properly
developed, may have considerable significance to ore microscopy,  a
report on this application was published by The American Mineralogist
(see Ehlers and Birle, List of Publications, page 113).

     Examination of polished and etched pyrite from Pennsylvanian-age
coal beds by means of the electron microscopy has revealed the presence
of fossil bacteria.  Although there have been several reported instances
of fossil bacteria, none has been discovered with such remarkably good

-------
preservation.  These bacteria, which lived approximately 200-million
years ago in the coal swamps were instrumental in precipitating iron
hydroxide (which was later converted to pyrite by H2S).   A strong simi-
larity can be seen between these and present-day swamp bacteria.  A
study of these bacteria types yields considerable information as to the
precise conditions of formation of these ancient coal-forming environ-
ments.  The interpretation of the various bacteria typed was aided con-
siderably by Dr. James Schopf, of the Coal Petrography Branch of the
U. S. Geological Survey.

     A third aspect involved a study of the hydrated sulfates associated
with exposed coal surfaces on, and adjacent to, coal mining areas.
White, yellow, and greenish crusty or fibrous materials  quite commonly
are present on pyrite-containing material.  These materials were col-
lected at several localities and subjected to x-ray diffraction analysis.
The most abundant mineral found was melanterite FeS04-7H20.  Halotrichite,
FeS04'Ale(S04)3'22H20, which had been found previously in several Ohio
localities was also obtained.  Another compound was identified as
FeSO "UHpO  by diffraction and polarizing microscopy.   This material,
although known as an artifically formed material, has never been ob-
tained from a natural geological environment.  Study of  this material
under laboratory conditions revealed that it was sensitive to changes
in humidity.

     Chemically pure samples were examined by x-ray diffraction at
various fixed humidities.  It was determined that above  approximately
75%, humidity this material exists as melanterite, FeS04'7H20.  Below
this level, and to at least 35$> humidity, the sample loses some of its
combined water and exists as FeS04-UH20.  This transition appears to
take place within about 30 minutes, although there is some variation as
a function of grain size.  This transition has been studied in some
detail.  This failure of previous investigators to observe the tetrahy-
drat-e form can probably be ascribed to the fact that most sample collect-
ing is done during the summer months under conditions of relatively high
humidity.

      It is also known that the mineral szmolnokite, FeS04-H20, exists
in nature.  Attempts were made, without success, to synthesize this
phase and determine its stability field at low humidities.

      This aspect of the study is extremely interesting mineralogically,
since the effect of humidity on such transitions has not been examined
in any detail for most hydrated phases.  The detailed results of this
study can be found in a paper by Ehlers and Stiles 7
                                   6

-------
Classification of Pyritic Materials

     Naturally occurring pyrite found associated in coals and partings
in coal beds were classified according to nature and sequence of deposi-
tion.  The following classification has been proposed by David Stiles:

     A.  Primary Pyrite

         1.  Sulfur Ball
         2.  Disseminated Pyrite
         3.  Primary Replacement Pyrite

     B.  Secondary Pyrite

         1.  Secondary Replacement Pyrite
         2.  Fracture-filling Pyrite

     Characteristics of each class may be described as shown below.

A.  Primary Pyrite

     Primary pyrite was deposited contemporaneously with the peat that
was converted into coal.  Some of this pyrite could have been formed by
bacterial metabolism as is shown by the bacterial remains within the
" sulfur-ball."  Part of the primary pyrite was deposited by replacement
of plant fragments.  It is certain that the replacement took place
early in the coalification cycle because the fragments are only slightly
distorted.  Had the replacement occurred after coalification, the plant
fragments would have been greatly flattened.  Evidence of pyrite deposi-
tion and replacement was observed.

     1.  Sulfur Ball

          Sulfur ball is found as flattened circular masses ranging  -in
size from 1 inch to 30 inches in diameter.  The longer dimensions are
always parallel to the bedding plane of the coal seam.  It is found  as
lenses varying from 1/2 inch to 2 inches thick and 1 foot to 10 feet
long and wide.  These lenses had previously been thought to be different
from the sulfur ball but their gain size and texture are very similar
so they are grouped under the classification of sulfur ball.  Pyrite
comprises 90 to 98$ of the sulfur ball, coal is the remaining percent-
age.  Small amounts of quartz, clays, and clays, and calcite are also
found in the sulfur ball masses.

-------
          The pyrite in the sulfur ball masses is found as small grains
ranging in size from 2 to 5 M-m in diameter.   Usually,  these small grains
agglomerate into spheres 10 |_im to 30 JJJD. in diameter.   Individual grains
are subhedral to anhedral with only a few good crystals for outlines
apparent.  These masses are a "brassy-gray color.   The  grayish tint of
the pyrite is caused "by minute inclusions of coal, since pyrite is
normally a bright brassy yellow.

          It appears that these masses were formed at  the same time the
peat was being deposited.  Bacteria nearly perfectly preserved are
found within the "sulfur ball" masses.  Had the masses been formed by
replacement after compaction of the peat had started,  the bacteria
would have been distorted.  Since the bacteria are only slightly dis-
torted, it is thought that the pyrite containing  the bacteria was depos-
ited along with peat formation very early in the  coal-forming process.
For this reason sulfur ball has been placed in the catagory of primary
pyrite.

     2.  Disseminated pyrite

          This type of pyrite is found mainly in  the  lower-grade coals
at the bottom and top of the beds and also in shaley  layers.  Dissemi-
nated pyrite is seldom visible to the unaided eye because of the low
percentage present and very small grain size.  When seen, it is a
brassy-gray color quite similar to the "sulfur-ball."   As in "sulfur
ball," the grayish cast of the pyrite is caused by minute coal inclu-
sions.

          The individual pyrite grains are 1 to 5 M-i"1 in diameter and
these are quite frequently agglomerated into masses 5  to 25 |om in diam-
eter.  The coal inclusions are found along the grain boundaries of the
pyrite.  A few crystal faces are evident on some  of the grains but none
are apparent on most of the grains.  These primary grains have fairly
regular and smooth boundaries forming an equidimensional to slightly
elongated body.

     3.  Primary Replacement Pyrite

          This type is found replacing plant parts and is very frequently
found associated with "sulfur ball."  Some larger plant fragments, as
tree trunks and branches, are found completely replaced by pyrite and
completely isolated from other pyrite.  Smaller pyrite-replaced plant
parts (such as leaves, stems, and seeds) are frequently found in masses.
These masses vary in size from 1/2 to 10 pounds and occur in no special
shape.  Pyrite comprises 25 to 50 per cent of the mass with coal as the
remaining percentage.

          The color of this pyrite is a bright brassy-yellow because
little coal is found within the actual pyrite mass.  The large percent-
age of coal in these masses is found between the  actual pyrite-replaced

-------
plant fragments.  Pyrite grains of this type vary in size from 50 um to
1 mm.  The shape of the grains depends somewhat upon the size of the
material replaced; for example, plant cells that have been replaced con-
tain pyrite grains the shape of the cell.  But large plant fragments
that are replaced by many pyrite grains contain some rather well-shaped
crystals.

B.  Secondary Pyrite

     Secondary pyrite was deposited after the coalification cycle was
complete.  A variety of secondary pyrite (i.e., fracture-filling pyrite)
is found in fractures that could only be formed after the coal had been
created.  Peat and the intermediate products in the conversion to coal
are all quite plastic and would give, rather than fracture, as coal does.
Secondary pyrite is characterized by larger grain size than primary
pyrite.  Some fractures are found in primary pyrite that have been
filled with the larger-grained secondary pyrite.

     1.  Secondary Replacement Pyrite

          This type of replacement is commonly found replacing "sulfur
balls" and associated with plant replacement.  On "sulfur balls" the
very small primary crystals are replaced with crystals up to 1 mm in
diameter.  Frequently, small "sulfur balls" are nearly completely
replaced but the larger ones are replaced only near the surface.  Frac-
tures in the sulfur balls are filled with this same type of pyrite.

          Plant fragments replaced by primary pyrite are found that have
undergone secondary pyrite replacement.  They are seldom recognized since
the secondary replacement tends to destroy the original identity.

          Secondary replacement pyrite appears much brighter and more
brassy-yellow than the primary pyrite varieties because of the larger
grain size and lower percentage of included coal particles.  These
crystals are 0.25 to 2 mm in diameter and usually have fairly good
crystal shape.

     2.  Fracture-Filling Pyrite

          As the name implies, this type of pyrite is found filling
fractures in coal.  Fracture-filling pyrite occurs randomly through a
coal seam with the exception of the very top and bottom.   This is true
probably because the shale layers tend to keep the coal from fracturing
and also they are quite impermeable to solutions.  Quite frequently this
type is found surrounding larger sulfur balls and pyrite lenses.  As
the coal is compressed, fractures are formed around the incompressible
sulfur ball or pyrite lense.  The fractures then are filled with pyrite
and as the pyrite growth continues, the fractures are lengthened and
branching fractures are formed.  When a polished section is made of coal
containing fracture-filling pyrite, it resembles a dineric pattern.

-------
          The pyrite grains in the form of flakes approximately 0.1 to
0.3 mm thick and 0.7 to 15 mm in diameter, are anhedral,  bright brassy-
yellow, seem not to incorporate any coal,  and are of secondary origin.
The larger surfaces of the flakes are fairly smooth, since they form
against the coal fracture surfaces.

Surface Area and Pore Size Distribution of Pyrite

     In order to determine significant differences between museum grade
pyrite and the sulfur ball material, measurements of the  surface area
and pore volume distribution of both materials were made.  Surface area
was determined using the B.E.T. method.  The pore volume  distribution
was obtained using nitrogen desorption and the method of  calculation
presented in Orr and Dallavalle.15  The specific surface  of museum-
grade pyrite in the size range from 60 to 150 mesh was found to be
0.12 + 0.01 m2/g, and that of the sulfur ball material in the same size
range to be 1.12  0.02 m2/g.  The corresponding oxidation rates were
5 and 130 |_tg/hr/g pyrite.

     The pore volume distributions of the two materials are presented
in Fig. 1, as the change in pore volume per unit change in pore radius
in cc/A versus the pore radius in A.  The curves show a much larger
total pore volume for the sulfur ball material, and a fairly flat dis-
tribution of pore volumes, except near the lower limit of detection at
about 25 A radius.

     The greater surface area and pore volume of the sulfur ball material
as predicted from its structure definitely will account to some extent
for its greater reactivity, since the rate-controlling step involves a
solid surface.

Preparation of "Sulfur Ball" Samples

     All coal and shale samples were crushed to pass a l/l+-inch screen.
They were then separated in a mixture of carbon tetrachloride-bromoform
(or tetrabromoethane) having a specific gravity of 2.00.   The portion
which sank to the bottom was removed, dried, and crushed  to pass a
60-mesh screen.  All particles smaller than 150 mesh were saved.  The
60- to 150-mesh portion was placed in a bath of pure bromoform and
allowed to stand overnight.  The float portion was discarded and the
sink dried at 100C  The enriched pyrite was separated into closer-
sized fractions for subsequent experimental work, and stored in tightly-
stoppered bottles purged with nitrogen.  Immediately prior to use, the
samples were washed with warm 2$ hydrochloric acid, then  washed with
distilled water, dried under vacuum, and weighed in the form it was
charged to the reactor.

     Characteristics of the several different pyrite samples used "are
given in Appendix V.
                                  10

-------
                                            o<

                                              CO
                                              Z>
                                              O

                                              S

                                              LJ
                                              cr
                                              o
                                              Q_
                                                    H
                                                    H
                                                    oi
                                                    pq
                                                    DD
                                                     0)
                                                    -P
                                                    H
s
0)
w
                                                     o
                                                    H
                                                    -P
                                                     w
                                                    H
                                                     O
                                                    PH
11

-------
Identification of Intermediate Products

     With the equipment and facilities available, no mineralogical
species, other than "final" products, were observed on the surface of
pyrite.  X-ray and electron diffraction studies failed to reveal any
unexpected intermediates.  It should "be noted that no detailed micro-
probe analyses were made, nor was low-energy electron diffraction equip-
ment available for use on this project.

     It therefore appears that the rate-controlling step is an electron
transfer reaction on the surface of pyrite itself (or through an absorbed
ionic specie) since no other solid phase is present in identifiable
quantities.
                                   12

-------
                               Section h-

                              OXYGENATION

     The first oxygenation (i.e., oxygen as the immediate oxidizing
agent) studies were designed to rationalize the conflicting reports on
the effect of oxygen concentration and the role of water in the kinetics
of pyrite oxidation.  The influence of different anions and cations, as
well as concentration of reaction products in the liquid phase, was
then determined in order to establish a firm base for quantitative
kinetic studies.

     The kinetics of pyrite oxidation were evaluated in two types of
environment:  (l) liquid phase, in which the weight of aqueous solution
in contact with the pyrite was 20 to 1000 times greater than the weight
of pyrite; and (2) vapor phase, in which the pyrite was suspended in a
gas phase of controlled humidity and oxygen concentration.

Experimental Equipment

     In addition to standard Warburg equipment and techniques, equipment
shown in Figs. 2, 3, and k were used.  The apparatuses shown in Figs. 2
and 3 are similar in concept; i.e., recirculating tne vapor phase
(Fig. 2) and the liquid phase (Fig. 3) through a bed of pyrite.  In each
case, oxidation of pyrite was monitored by measuring the quantity of
make-up oxygen required to maintain a constant pressure in the vapor
space.

     This type of equipment was used for several reasons:  (l) to estab-
lish influence of diffusional resistances and desorption of reaction
products on reaction rates, (2) to allow the control of concentration
of various components in the recirculating fluid, and (3) to permit
periodic sampling of streams to follow build-up of reaction products.

     Figure k is a modified differential Warburg unit used to determine
oxidation rates over comparatively wide pressure ranges.  It consists
of two identical flasks on either arm of a manometer.  The same quantity
of water was added to each flask; then pyrite was placed in the right-
hand flask.  A small amount of manometer fluid was added to the top of
the mercury.  To fill the system with vapor, the mercury column was
lowered so that the water level was at the right-hand stopcock, the
system was evacuated, and then the vapor was admitted through the vacuum
line.  The left stopcock was then closed and the mercury leveling bottle
raised or lowered, depending on the pressure in the system, so that the
water level rose half-way in the manometer  The entire assembly was
agitated during liquid phase runs.

     The pyrite sample used for these preliminary runs was a "sulfur ball"
material collected from the Middle Kittanning No. 6 coal seam in Vinton
County, Ohio.  Several stages of float-and-sink separation were used to
                                   13

-------
            To metered oxygen supply
                     L

Saturated
Salt
Solution
                                  Fig.  2 - Vapor  Phase  Oxidation Apparatus
                          Reactor
  Peristaltic
    Pump
            I mm I.D. Capillary Tubing
                                     To metered oxygen supply
                                                                          *- Aerator
                                                                        Reactor
                                           Peristaltic
                                             Pump


                             Erlenmeyer      Fig. 3
                          (with Center Well)
                                          -  Liquid Phase  Oxidation  Apparatus
To charge gas
  or vacuum
                             Hg

                           \V      To
                            N~-3 Leveling
                                 Bottle
Fig.  k -  Variable  Pressure Apparatus

-------
obtain an 85% FeS2 material.  This was then carefully screened and the
70- to 100-mesh fraction used as the sample was designated "Sulfur Ball
Wo. 1."

Results of Preliminary Runs

     Data on liquid phase oxidation using equipment shown in Fig.  3
showed that flow rates from three or four ml/hr to 30 1/hr of fluid
through the bed had no effect on rate of oxidation.  Also, the rate
remained constant for the duration of the runs (one to two weeks).

     Two runs were made with a column of "Amberlite" IRA-120 in series
with the reactor to remove all iron from the circulating fluid. With
"zero" iron in solution, the rate remained the same.
     The relation between the increase of iron in solution and oxygen
absorbed was frequently checked and found to be approximately 3-5 moles
oxygen consumed per mole of soluble iron produced.  The relative ferrous-
ferric ion concentration was found to vary between 85$ and 95$> ferrous
for runs listed in Table 1.

     A series of liquid phase runs was made in which oxygen concentra-
tion in the vapor space and temperature were varied.  The results of
these runs, together with those obtained with the ion-exchange resin in
the system, are listed in Table I.

     A similar study was made using the vapor phase apparatus.  As in
the liquid phase runs, flow rates had no effect on the rate of oxygen
absorption, which remained constant throughout the run periods of from
several days to two weeks.  Temperature and humidity of the recycled
vapor were varied; the results are shown in Table II.

     Since the other oxidation units were not capable of operating over
wide pressure ranges, the equipment shown in Fig. U was constructed.
Runs were made at different oxygen pressures and vapor compositions to
study the effect of dissolved oxygen and nitrogen concentrations on
oxidation rate.  These data are shown in Table III.

     All rate data in Tables I and III were recalculated to the same
reference; i.e., 100 |_ig02/hr/g of pyrite at 25C and oxygen pressure of
760 mmHg.  This was necessary in order to obtain consistent results
from one column packing or washing to the next.  The principal problem
was the tendency of the sulfur ball particles to break down during
agitation.  For one continuous series of runs, the change from run to
run was small.  But after washing or repacking, a reference run was
necessary to obtain comparative rates.
                                  15

-------
             Table I.   Liquid Phase Runs
Vapor
Cone.
(% 02)
100
100
100
100
10
26
54
79
100
100
100
Temp.
(c)
20
25
30
35
25
25
25
25
25
25
35
Oxidation
Rate*
58
100
138
215
23
47
79
85
100
102**
210**
           *Rate = i_ig 02/g pyrite/hr
           v v
            With "Amberlite"
             Table II.   Vapor Phase Runs
Temp.
(c)
25



35



^5



Partial Pressure
H20
(mmHg)
22.8
17-8
12.6
7.6
40.5
31.2
21.5
13-1
69
53.2
33
22.3
Relative
Saturation
(*)
96
75
53
32
96
75
51
31
96
75
46
31
Oxidation
Rate*
85
50
25
16
168
90
57
33
400
152
70
48
*.
 Rate = |ag 02/g pyrite/hr
                           16

-------
        Table III.  Variable Pressure Runs at 25 C
02 Pressure
(cm Hg)
76
183
170
170
1^3
91
1+8
38
28
21.5
20
1^
15
10
3^
15
ik
15.5
10
7.1
N2 Pressure
(cm Hg)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
128
100
67
59
70
27
Liquid 02
Concentration
(ppa)
3k.h
95
88
88
7k
kl
25
19-5
1^.5
11.2
10. k
7.3
7.8
5.2
17-5
7.8
7-3
8.0
5.2
3.7
Oxidation
Rate*
100
165
162
156
Ikk
110
68
62
52
39.5
37
3^
32
25
^7
30.5
30
26.5
2U.5
19
Rate = |ag 02/g pyrite/hr
                             17

-------
Discussion of Preliminary Runs

     Based on the observations that (l)  flow rates  of the recirculated
fluid had no effect on oxygen absorption rates,  (2)  rates were constant
throughout the test period, and (3) removal of iron from solution did
not affect rates, it seems evident that  reaction rates in the
recirculating-type equipment, for the reaction periods used,  are not
influenced by diffusional resistances nor desorption of oxidation pro-
ducts.  This conclusion is supported by  the value of Ik kcal/g-mol for
the activation energy determined from an Arrhenius  plot of data pre-
sented in Table I.  This value for activation energy is typical of a
reaction-controlled, rather than a physically controlled, rate-limiting
step.

     Role of Water

          In studying the role of water, Kim11 conducted a series of
vapor phase oxidations using the equipment shown in Fig. 2.  Over the
limited temperature range studied, he observed that the rate  varied
linearly with absolute humidity (or partial pressure) of water in the
vapor pha.se, apparently indicating that  the rate of oxidation was first-
order with respect to water and suggesting that water is a reactant.

          When this study was extended to cover a wider temperature
range, a distinct temperature dependence was observed, as shown in
Fig. 5.  It was also noted that in the temperature  range covered by
Kim both the oxidation rate and partial  pressure of water vapor (over
a saturated salt solution) doubled with  a 10C temperature rise.  In
other words, the influence on rate observed by Kim may be interpreted
as a change caused by temperature at a constant relative humidity.  The
runs plotted in Fig. 5 are replotted as  Rate vs. % Relative Saturation
in Fig. 6.  Note that at a given value of relative  saturation, the rate
nearly doubles with each 10C increase in temperature, the same as
observed in liquid phase oxidation.  When curves in Fig. 6 are compared
to the adsorption isotherms for water on sulfur ball and museum grade
pyrite (see Figs. 35 and 36) the shapes  of the curves are very similar,
strongly suggesting that the rate is dependent on the quantity of water
adsorbed or, in other words, the number  of reactive sites which are
covered by water.  Note that even the so-called vapor phase oxidation
really occurs in a liquid phase formed by adsorption of water condensed
from the vapor phase.

          When the isotherms of Fig. 6 are extrapolated to 100$ Relative
Saturation and the rates thus obtained plotted,  together with those
from the liquid phase runs (Fig. 7) 5 the similarity of temperature
effect in both vapor and liquid phase oxidations is apparent.

          While it is not possible to describe the  role of water in all
phases of the reaction, for the rate-limiting reaction water  is -involved
as a reaction medium rather than a reactant  If water were one of the
                                   18

-------
  400
-c 300
o>
^
o>

- 200
UJ
   100
            D 25C
            O 35C
            A 45 C
      400
      300
      200
      100
      0      20     40     60     80
           H20 Partial Pressure ( mm Hg)

  Fig. 5 - Oxidation  Rate vs  Partial
           Pressure of Water
                                       0
                D 25C
                O 35C
                A 45C
                25      50     75
                 % Relative Saturation
                                                                     100
           Fig. 6  -  Oxidation Rate vs %
                     Relative Saturation
  400



~ 300
_c
o>

 4.200
jj

c
   100
           O-Liquid Phase
           A-Vapor Phase
         20
     200


      160


      120


      80


      40
                                                        	Eq.(l)
                                                        Eq.(2)
50
                30      40
              Temperature (C)

        Fig.  7 - Oxidation Rate vs
                 Temperature
      20   40   60    80    100
    Oxygen Concentration p.p.m.

Fig.  8  - Oxidation Rate vs Oxygen
         Concentration in Aqueous
         Phase
                                    19

-------
reactants in the rate-limiting step,  the rate of reaction would level
off as the partial pressure of water  approached saturation pressure
(i.e., since water is no longer the limiting reactant).   Instead,  the
rate of increase is greater as water  concentration nears saturation,
as seen in Fig. 7-

          Even in vapor phase oxidation, rates remain constant over
long periods of time.  This can be explained by products "salting-out"
of a condensed liquid phase as the content of dissolved reaction pro-
ducts reaches saturation.  This process was qualitatively observed by
Birle2 from photomicrographs showing  the build-up of salts around  the
edges of pores or etch marks which contained water.

          The basic function of water at least for the rate-limiting *
reaction, is to provide a means by which the oxidation products are
desorbed (dissolved) for the pyrite surface.  Normal oxidation, which
may occur in the dry state on clean surfaces, can be stopped by build-
up of products on the "reactive sites" of pyrite.  Data for oxygen
adsorption on dry pyrite can be interpreted in this manner (see Fig.  37
and Table XIII).

     Oxygen Concentration

          Possible kinetic mechanisms by which oxygen enters the rate-
limiting reaction are suggested by the data presented in Table III.

          These data show a quantitative relation between oxidation
rate and oxygen concentration, and indicate that the concentration of
inert gas (nitrogen) affects oxidation rate.

          Using procedures and modifications suggested by Hougen and
Watson,10 rate equations for different assumed mechanisms can be derived.
For the type of reaction under study, simplifications can be made.,  It
has been shown that neither desorption of products nor diffusional
resistances influence the rate of reaction under the laboratory condi-
tions used in this study.  Therefore, the experimental rates can be
considered "initial rates" and derivations simplified by eliminating
consideration of both product composition and concentration.

          Data at low pressures (shown in Fig. 8) may be correlated by
assuming the following mechanism:

          (l) oxygen is adsorbed on "reactive site" of pyrite;

          (2) oxygen is dissociated,  forming an activated complex;

          (3) the activated complex decomposes to form an oxidation
              product; and

          (k) the oxidation product is desorbed, forming another reactive
              site.

                                   20

-------
          If decomposition of the activated complex is the rate-limiting
step, the adsorption and desorption step would be in equilibrium.
The following rate equation can be derived:


                    r = kCA/(l + /K^GA. + K^)2                   (1)
or


                   r' = k'CA/(l + KACA + KjCj)                      (2)


where

           r = reaction rate for "dual site" adsorption,

          r1 = reaction rate for "single site"  adsorption,

          CA = concentration of dissolved oxygen,

          GI = concentration of dissolved inert gas,

          KA = adsorption equilibrium constant  for oxygen, and

          KI = adsorption equilibrium constant  for inert gas.
          Plotting the data of Table III together with the derived
equations (Fig. 8) indicates that the adsorption mechanism is consist-
ent with experimental data.  However, it must be emphasized that this
does not prove the validity of the mechanism.  It merely indicates
that the proposed mechanism is possible according to this set of experi-
mental data.  The single-site adsorption mechanism, although showing
greater deviation, can not be eliminated on the basis of these data.

          Oxidation rates in both vapor and liquid phase runs do not
appear to be limited by oxidation products.  Rates are constant for
long periods of time.  Visible build-up of oxidation products in vapor-
phase oxidations do not influence rate.  The desorption of oxidation
products by adsorbed water is a continuous and effective mechanism for
renewing "reactive site" on the pyrite surface.

          The presence of dissolved nitrogen reduces the rate of oxida-
tion slightly as shown in Table III.  The rate data are not so consist-
ent that a reliable value for KI can be calculated.

          Since the adsorption of oxygen occurs in an aqueous medium
and since the presence of nitrogen reduces oxidation rate, it is likely
                                  21

-------
that physical rather than chemical adsorption on "reactive sites" is
involved.  It is unusual for chemisorption to occur from highly polar
solvents.  Also, if oxygen were chemisorbed, nitrogen would have no
effect on rate.  At the same time it may not seem likely that nitrogen
could compete with oxygen in an aqueous solution for physical adsorption
on a relatively nonadsorptive material like pyrite.  However, adsorption
isotherms for nitrogen on sulfur ball pyrite (see Fig. 314-) show a rela-
tively strong adsorption, compared to adsorption on museum grade pyrite.

          It was recognized that a more critical study was needed of
the effect of nitrogen on the reaction since the manner in which nitro-
gen and oxygen are adsorbed is of fundamental importance to inhibition
and mechanism studies.   Therefore a high-pressure liquid-phase oxida-
tion investigation was  made to quantitatively evaluate the effect of
nitrogen partial pressure on oxidation rate.

High Pressure Oxygenation

     Equipment similar in concept to that shown in Fig. k was redesigned
to operate at 25 atm.  Instead of a leveling bottle to control the mano-
meter liquid level, a displacement piston was used to position the mano-
meter fluid.

     The experiment was divided into two phases.  During the first phase,
oxidation rates as a function of oxygen dissolved in the liquid were
determined.  The second phase consisted of studying the effect of an
inert gas (nitrogen) on the oxidation rate of pyrite.

     All experimental runs were carried out at 25C.  The oxidation
rates were recorded as |_ig02/g/hr consumed, where gram refers to one
gram of pyritic material.  The runs are numbered in the sequence in
which they were performed.

     First, the agitation rate which would eliminate the diffusional
resistance was determined.  No direct measurement was taken of the
degree of agitation, but it was set at a rate such that an increase in
agitation did not change the oxidation rate.  The agitation rate, which
was determined at the highest oxidation rate, was then held constant
throughout the remainder of the experimental runs.

     Table IV lists the experimental runs and the resulting oxidation
rates expressed in |j.g02 consumed/hr/g pyrite.  One should notice the
two columns of oxidation rates; i.e., Oxidation Rate and Recalculated
Oxidation Rate.  Since the base rate runs indicated that the oxidation
rate changed slowly over a period of weeks, because of a loss of pyrite
and attrition of pyrite particles, the series of runs (Runs #3 - #1?
and #21 - #2?) involving pure oxygen were recalculated to the .same
reference.  The reference used was the rate of 335 M-gOa consumed/hr/g
pyrite at 25C and 38^- cm Hg oxygen pressure.  Once the material was
contaminated with nitrogen the recalculation was not made because the
original base rate could not be duplicated.


                                   22

-------
                Table  IV.   High-Pressure Runs at 25C
Run #
3
4
5
11
12
13
14
15
16
17
17a
18
20
new sample
21
22
23
25
26
27
28
29
31
32
33
34
35
36
37
38
39
ho
41
42
Oxygen
Pressure
(cm Hg)
384
596
74.5
384
157^
384
384
384
596
1068
1068
1068
849

384
849
1574
188
240
384
384
380
380
380
384
384
229
384
1600
384
384
384
304
384
Nitrogen
Pressure
(cm Hg)
0
0
0
0
0
0
0
0
0
0
241
536
0

0
0
0
0
0
0
69
212
707
1245
0
1024
0
0
0
0
0
1240
0
1240
Liquid
02 Cone.
(ppm)
186
284
40
186
626
186
186
186
284
460
460
460
382

186
380
626
118
142
186
186
184
184
184
186
186
138
186
636
186
186
186
186
186
Oxidation
Rate*
285
394
93
334
711
325
320
332
404
590
530
498
340

313
565
827
193
251
362
355
349
316
273
318
292
202
300
612
362
328
298
319
285
Recalculated
Oxidation
Rate**
335
450
104
335
711
325
320
332
404
590




335
564
811
186
236
334














#-*.
I_ig02 consumed/hr/g Pyrite
Based on reference rate of 335 ng02  consumed/hr/g at 25C and
384 cm Hg
                                   23

-------
     A further calculation was necessary at the end of Pun #38-  Since
some of the pyritic material was lost in the washing prior to this run,
the rate was assumed to be equal to Run #2? and the amount of pyritic
material left was back-calculated.

     Figure 9 is a plot of the rate of oxygen consumption versus dis-
solved oxygen concentration.  The points in the lower concentration
region represent data shown in Fig. 8.

     Figure 10 is a plot of the oxygen consumed per hour per gram pyrite
versus mole percent of nitrogen in the total dissolved gas.  Partial
pressure and the dissolved oxygen concentration were held constant at
38*4- cm Hg and 186 ppm, respectively.

     Attempts to check the original base rate, after each sample was
contaminated with nitrogen, were not successful, but an equilibrium
oxidation rate of a lower value was established.  However, after this
initial contamination and decrease in rate, there was still a definite
reduction in rate with increased dissolved nitrogen concentration.  This
was established by making several repetitive runs (Runs #38 - #^2) using
pure 02 and N2 mixtures.  The effect of nitrogen on the oxidation rate
of pyrite is revealed to some degree in Fig. 10, where oxygen consump-
tion is plotted versus mole % nitrogen in the dissolved gases for a
fixed dissolved oxygen concentration.  Runs at higher mole ratios of
nitrogen to oxygen were restricted by pressure limitations of the equip-
ment.

     After the run at 6l.5 mole % nitrogen, the system was evacuated
and a base run was made.  A comparison of the oxidation rate of the
original base run, made before the sample was subjected to nitrogen
(Run #38), and this run (Run #39) (362 and 3l8 M.g02 consumed/g pyrite/
hr, respectively) showed a significant difference.  A certain amount of
nitrogen contamination was apparent.  This contaminating effect of
nitrogen was also noted in the first sample used and was the reason for
discarding the material.

     Since the nitrogen could not be removed by an ordinary vacuum, it
was reasoned that the nitrogen may be rather strongly adsorbed on the
pyrite.  In an effort to check this, the flask was cut off and the
pyritic material was washed, dryed, and heated to 100C under vacuum
in an attempt to drive off the absorbed nitrogen.

     The results of Runs #38 and #39 made after the flask was rejoined
to the apparatus indicates that nitrogen was desorbed upon heating under
vacuum.

     Although it was shown that nitrogen has an initial contamination
effect, it is of interest to know the decrease in oxidation rate, if
any, brought about by dissolved nitrogen alone.  The consistent results
of a series of runs (#39 - #^2) indicate that an increase in dissolved

-------
                                                                   o
                                                                   o
                                                                   O
                                                                   8
                           8
                           10
                                                                   8
                                                                   CVJ
                                                                   o
                                                                   o
                                                                           E
                                                                           Q.
                                                                           Q.
                                                                          IT
                                                                          fe
                                                                          ui
                                                                          o
                                                                          z
                                                                          o
                                          sd
                                          o
                                          H
                                          -P
                                          05
                                                                                   CD
                                                                                   o

                                                                                   o
                                                                                  o

                                                                                   c
                                                                                   0)
                                                                                   X
                                                                                  O

                                                                                   w
                                  -P
                                  cti
                                  PS

                                  fi
                                  o
                                   H
                                  -P
                                  cd
                                  T3
                                  H
                                  X
                                  O

                                   i

                                  cr\
00
ro
CVJ

-------
                                           o
                                           o
                                          o
                                          or
                                                 Q

                                                 ^>
                                                 O
                                     oi
                                    -P
                                     ra
                                     S
                                     O
                                    o
O
O
CVJ
O
O
                                                 CO
                                                 UJ
                                                 CO
                                                 <
                                                 o

                                                 Q
                                                 UJ
                                                 O
                                                 CO
                                                 CO
                                                 Q

                                                 U.
                                                 o
                                                  oJ
                                                 UJ
                                                 _l
                                                 O
                                          O
                                          CM
                                                          e\j
                                                         O
                                    d
                                    !
                                    H
                                    J

                                     C!
                                    H

                                     CM
                                                         rH
                                                         q
                                                         o
                                                         H
                                                         Ml
                                                         H
             26

-------
nitrogen does decrease the oxidation rate over and beyond the first con-
tamination effect.

     It does not seem likely that an inert gas such as nitrogen would
be strongly adsorbed.  However, adsorption studies (see Fig. 3*0 indi-
cate that the structure of sulfur ball pyrite is such that a signifi-
cant number of reactive sites are found in small parting planes or
ledges having a spacing or configuration that produces a relatively
strong adsorption of nitrogen.  The nitrogen adsorption isotherm (see
Fig. 3*0 shows a decided hysteresis loop which can account for the
nitrogen contamination observed.  When nitrogen occupies the stronger
reaction sites, reaction rate is decreased.  Oxygen and nitrogen are
in dynamic equilibrium in competition for the less active adsorption
(reaction) sites.

Effect of pH

     By controlling the pH of recirculated water in the apparatus shown
in Fig. 35 the effect of pH on oxidation rate, as measured by oxygen
consumption, was determined.  As shown in Table V and Fig. 11, the reac-
tion rate increases with pH  slowly up to a pH 3 and rapidly as pH
exceeds 6.

     These results were surprising in view of the generally held opinion
that the higher the acidity, the greater the oxidation rate.  Exactly
the opposite is true.

     A proven explanation for this phenomenon is not available.  It is
not due to the increased rate of ferrous-to-ferric ion oxidation in
solution.  The concentration of ferric ion in solution above a pH of
3.5 is too small to account for the oxidation rates observed.  Assuming
the adsorption mechanism for oxygen is relevant, it may be that altera-
tion of the pyrite surface with pH increases the number of adsorption
sites, or the adsorption equilibrium constant for oxygen.
                                   27

-------
Table V.  Oxidation Rate vs. pH;
          Sample:  Sulfur Ball #2
          Oxygen Concentration = 39
      PH
Oxidation Rate
      1.5
      2.0
      U.O
      6.0
      6.0
      7.0
      8.0
      9.0
     10.0
     2.10
     2.2
     3.2
     5.5
     7.9

    13.5
                                   pH
                                     Fig. 11 - Oxidation Rate vs. pH
                                  28

-------
Effect of Various Anions and Cations

     Varieties of materials have been added to the circulated water to
study their effects on oxidation rate.  Concentration of sulfate and
iron has a negligible effect at normal pH's.  Humic acids;  chromium;
copper; manganese; nickel cations; and nitrate, chloride, and phosphate
anions had no effect on rate when added to the solution in  concentration
many times greater than could be expected in natural conditions.  High
concentrations of phosphates did have an inhibitory effect, but negli-
gible inhibition was noted in concentrations of less than 200 ppm.  The
oxygenation reaction is remarkably insensitive to concentration of
material other than oxygen.

     There is no evidence that trace impurities or features, other than
those that can be described by texture of the pyrite agglomerate, have
a major effect on reactivity of pyritic materials.
                                  29

-------
                               Section 5

                             MICROBIOLOGY

     The role of autotrophic bacteria of the Ferrobacillus-Thiobacillus
group in the kinetics of pyrite oxidation has been a subject of conjec-
ture for many years, but little quantitative kinetic information appli-
cable to pyrite oxidation in field environments has been available.
The possible implication of the various species of organisms belonging
to this group in pyrite oxidation kinetics is based on several observa-
tions.  First, these organisms have the ability to oxidize reduced
inorganic sulfur and/or iron species, particularly elemental sulfur,
ferrous iron, and thiosulfate as their sole source of energy for the
support of their metabolic processes.  Second, they generally tolerate
an environmental pH of less than 2, and appear to be almost universally
found in acid coal mine drainages and in association with pyritic-
containing materials in and around mines.  Third, the presence of iron-
oxidizing strains in growth media containing ferrous iron has been shown
to greatly increase the rate of iron oxidation, as compared to sterilized
controls.3

     Although several of the strains which have been isolated and studied
have been identified as separate species, there appears to be increasing
opinion that these strains may be, in fact, variants of a single species8
and the organisms are collectively referred to in this report as the
Ferrobacillus-Thiobacillus group.  Recent literature relating to the
Ferrobacillus-Thiobacillus group are available6'21 and such information
will not be repeated here.  In this study, the intent has been to inves-
tigate the kinetic effect of organisms from this group on the oxidation
of pyrite under simulated natural environmental conditions.  In all of
the experiments described below, the only source of ferrous iron and
reduced sulfur available to the organisms was pyritic material, and the
aqueous salt solution used in the various runs was a simulated ground
water.  Possible carbon sources were carbonaceous materials in the
pyritic, carbonate species in the simulated ground water, and C02 in
the atmosphere of the various apparatuses.  Simulation of a natural
flow of water past the pyrite  underwent a progressive development dur-
ing the course of the experimentation, and will be discussed for each
specific apparatus used.

Preliminary Experimental Phase

     At the time this phase of the project was begun, most of the data
available in the literature on the various strains of the Ferrobacillus-
Thiobacillus group were derived from experiments in which the organisms
were grown in media containing elemental sulfur, thiosulfate, or ferrous
iron as the energy source; little work had been published in which
pyritic material was described as the sole energy source.  Braley3 grew
organisms on pyritic material, as well as on the above materials, and
found the sulfur-oxidizing strains to produce appreciable amounts of
                                  31

-------
sulfuric acid from elemental sulfur,  but to have little effect  on the
rate of oxidation of the iron sulfides of coals.  The iron-oxidizing
bacteria increased the rate of acid production from the iron sulfides
to more than three times the rate in the sterile solution,  and  together
the iron oxidizing bacteria and the sulfur oxidizing bacteria increased
the rate ten-fold.  Braley postulated that the role of the  bacteria was
to catalyze the oxidation of ferrous  ions to ferric ions, which in turn
oxidized the pyrite  to ferrous and sulfate ions,  in accordance with
the reactions first presented by Stokes24:


              FeS2 + 7Fe2(S04)3 + 8H20 -15FeS04 +  8H2S04
Since these organisms are strict aerobes, oxygen is the only electron
acceptor which will support the bacterial oxidation of ferrous to
ferric iron in this scheme.  If this bacterial oxidation of ferrous
iron occurs in solution, then  there is no necessity for the bacteria
themselves to come into physical contact with the pyrite.  This mecha-
nism will be referred to as the indirect mechanism.

     In contrast to the indirect mechanism model, other researchers5'19
have presented evidence that the bacteria exert their catalytic effect
while absorbed on the pyrite surface.  From a practical standpoint, the
important difference between these two conceptual models lies in the
fact that in the former case the total number of bacteria in a given
system is potentially a rate-determining factor, while in the latter
model, only those bacteria absorbed on the pyrite surface will directly
affect oxidation rates.

     The specific objectives of this preliminary phase of research were
to develop a culture apparatus which would simulate field conditions,
and to quantitatively measure pyrite oxidation rates and bacterial growth
under controlled environmental conditions with pyritic material as the
sole source of iron in the system.  Hopefully, the experimental results
would tend to confirm one of the above mechanistic models.  The appara-
tus development and experimental results are described in detail in the
M.S. thesis by Bailey.1  The pertinent procedures and results are given
below.

     Apparatus and Growth Media

          The final apparatus design used by Bailey is shown in Fig. 12.
Essentially, this apparatus is a respirometer in which the gas and water
volumes are kept constant during a run, and oxygen uptake is calculated
from observed changes in the gas pressure within the system, with due
regard to exterior pressure and temperature changes.  The entire appara-
tus could be autoclaved, all liquid additions and removals could be
made aseptically, and gas inlets and outlets were protected by sterile
filters.  During an experimental run, a measured sample of pyrite was
                                  32

-------
                                       CO

                                      I
                                      o
                                       I
                                      CM
33

-------
placed, in the pyrite bed, and ten liters of simulated ground water
(iron free) were placed in the main culture vessel,  which is a 20-liter
carboy.  A recirculating pump passed the water intermittently through
the pyrite bed, at one-minute cycles, and the bed was allowed to drain
freely between cycles.  The recirculation rate used gave a retention
time in the carboy of approximately 28 hours.  The magnetic stirrer
insured complete mixing and oxygen saturation of the water in the carboy.

     The hydraulic design of this system was based on the belief that
water entering an underground mine and traveling through many intricate
passages at varying speeds will, at those points where the flow is
relatively continuous, reach a steady-state chemical composition.  Local
pyrite oxidation rates and solute concentrations will differ depending
on the distance traveled through the pyritic material.  Thus, this labo-
ratory simulation, in which an equilibrium will be attained between the
reaction rates and dilution by daily replacement of sample volumes with
an equal volume of the inorganic salt solution, should approximate the
conditions at certain points in a mine.

     The normal sequence for a run was to operate the system under ster-
ile conditions for a period sufficient to give the base rate of pyrite
oxidation due to oxygen alone, without bacterial catalysis.  Then, the
system was inoculated with a culture of bacteria, and oxidation rates,
bacterial numbers, and iron concentration in the recirculated water were
followed until the system reached steady state.

     The system was difficult to operate, and only the results of the
most meaningful run are reported herein.  The operating parameters for
this run were as follows:

     Temperature - 20C

     Organism - Iron oxidizing organism of the Ferrobacillus-
                Thiobacillus group, isolated from acid drainage
                from the McDaniels Mine, Vinton Co., Ohio by
                serial dilution and growth in Silverman's 9-K
                medium.22  Possibly a mixed culture.

     Liquid Medium - Simulated ground water (iron-free),
                     composition given in Appendix.

     Pyrite - 25 grams of crushed sulfur ball, 35-^-8 mesh.
              Sulfur ball taken from McDaniels Mine, Vinton
              Co., Ohio.

     Atmosphere Composition - Normal atmosphere (~ 21% 02).

     The apparatus was operated for 150 hours under sterile conditions,
at which time an inoculum containing approximately I05 cells was intro-
duced into the system, and the run was continued for a total of

-------
1030 hours.  During the entire course of the run, oxygen uptake rate
and soluble ferrous and ferric iron concentrations in the recirculated
ground water were monitored, as well as pH of the water.  Following
inoculation with the bacterial culture, bacterial numbers in the recir-
culated water were monitored.  Iron determinations were made colorimetri-
cally with a Beckman DU spectrophotometer, using a modification of the
phenanthroline method as described in Standard Methods.23  The procedure
was modified by omitting the boiling step in the sample preparation pro-
cedure.  Bacterial populations were determined by the use of a multiple-
tube serial dilution method, with population estimates based on the
Most Probable Number tables given in Standard Methods.  The medium used
was Silverman's 9-K, with positive growth being indicated by a color
change to rust-red.  Incubation of the tubes was at 20C for a period
of two weeks.

     Results

          For the run described above, plots of oxygen uptake rate and
cell concentration versus time are shown in Fig. 13.  Referring to
Fig. 135 it is seen that following inoculation with the organisms at
150 hours, the oxygen uptake rate remained at the base (chemical oxida-
tion) rate until about 325 hours, at which time it began to increase,
ultimately leveling off at about 825 hours.

          In regard to the growth of bacteria following the inoculation
at 150 hours, the cells appeared to increase in a relatively constant
logarithmic manner until reaching a population of about k.3 x 10s
organisms/ml, as indicated by the Most Probable Number test.  During
the entire course of the run, the pH of the recirculated water dropped
from an initial value of k to a terminal value of 2.7.  Thus, it is
improbable that limitation of the bacterial numbers resulted from an
excessively low pH.  Since the air space of the apparatus was flushed
with fresh air daily, and approximately 100 ml of fresh ground water
medium containing NaHC03 was added daily to make up sampling volume
losses, the availability of C02 for autotrophic cell growth was far in
excess of the amount used, based on observed population increases.  The
limitation of bacterial population may have resulted from a deficiency
of available nitrogen, although trace element availability, or the
buildup of metabolic products are other possible limiting factors.  It
is interesting to note that the bacterial population reached a level of
about 104 per ml before there was any appreciable effect on oxygen up-
take rate.  If the rate increase at 325 hours is interpreted as evidence
of the beginning of biological catalysis, then the catalysis may be
related to the number of organisms in the recirculated water.  If the
indirect model for bacterial catalysis is correct, then the observed
level of 104 organisms per ml may represent the population required to
supply ferric ions at a sufficient rate to support a detectable oxida-
tion of pyrite by ferric ions, as compared to the base rate for oxygena-
tion.  On the other hand, if the bacterial catalysis is due to bacteria
adsorbed on the pyrite particles, then this level of 104 organisms per
                                  35

-------
             10s r
_QJ
o
E
LJ

a:
Ld
 CM
O
1-1
   io-2L
                    LEFT SCALE
                    FOR BACTERIAL
                    COUNT
*-  RIGHT SCALE
   FOR BACTERIAL
   COUNT
     I0r    I03 -
                                                                                     c
                                                                                     o
                                                                                     a>
                                                                                     6
                                              o
                                              o
                                              CL
                                              UJ
                                              h-
                                              O
                                              <
                                              m
                                     400                  800

                                             TIME  (hr)

                     Fig. 13 - Bacterial Count and 02 Uptake Rate vs Time
                                        36

-------
ml would be the concentration of bacteria necessary to provide an equili-
brium with a minimum critical population adsorbed on the pyrite surfaces.

          Other observations may provide an indication of which model
is more likely to be correct.  The soluble iron in the recirculated
water was predominately in the ferrous form until about 320 hours, at
which time the ferric-ferrous ratio began to rise continuously for the
remainder of the run.  This suggests that the increase in pyrite oxida-
tion rate may be related to the increase in ferric-ferrous ratio in
solution.  The application of the indirect model, in light of the ferric
ion oxidation kinetics discussed in Section 6, would require a sufficient
population of organisms in the recirculated water to increase the ferric-
ferrous ratio to the point (at any given total soluble iron concentra-
tion) at which the rate of oxidation of iron by organisms is equal to
the rate of reduction of ferric iron at the pyrite surface  Further,
the ferric ion oxidation kinetic model predicts a maximum rate of pyrite
oxidation when the ferric-ferrous ratio approaches infinity.  The pres-
ence of a population of organisms in excess of the number required to
keep all iron in the ferric state would result in no further increase
in pyrite oxidation rate.  The data shown in Fig. 13 could be explained
in terms of the indirect model and ferric ion oxidation kinetics, if the
continued increase in oxygen uptake rate between U60 and 820 hours could
be accounted for.  Since the bacterial population leveled off at about
h60 hours, and since an increase in specific oxidative capacity per
bacterial cell would be unlikely during this stationary phase of the
bacterial culture, it is likely that the increasing oxidation rate is
related instead to the total iron concentration.  For the data gathered
during this run, a plot of log (Fe+++) versus pH yields a straight line
with a slope of -2, indicating that the ferric iron solubility was con-
trolled by the equilibrium of Fe(OH)++ with Fe(OH)3, with Fe(OH)++ being
the principal ferric specie  in solution.  Since pyrite oxidation rate
by ferric ions is a function of both ferric-ferrous ratio and total
ferric ion concentration, the gradual lowering of the pH during the
course of the run, and the consequent increase in soluble ferric iron,
could have resulted in an increased pyrite oxidation rate.

          The above interpretation of the data in terms of the indirect
model does not prove this model to be correct.  Presumably, the data
could possibly be explained in terms of the model in which it is
assumed that the bacteria responsible for the catalysis action are
adsorbed on the pyrite surface.  The coincidence of increasing the
ferric-ferrous ratio in solution with observation of the beginning of
the catalyzed oxidation rate, and the continued increase in oxidation
rate after the bacterial population became stabilized,  are, however,
more difficult to explain if one assumes that adsorbed bacteria are the
primary catalytic agent.
                                  37

-------
     Conclusions

          The conclusions of this preliminary work are:

          1.  The iron-oxidizing autotrophs can be very  satisfactorily
              grown in an environment in which the only  iron source
              is pyrite, and at sufficiently high bacterial populations,
              will significantly increase the rate of pyrite oxidation
              over the base chemical rate.

          2.  Bacteria may function catalytically by merely oxidizing
              ferrous iron in solution.  The ferric iron produced then
              oxidizes pyrite in accordance with the anaerobic pyrite
              oxidation kinetics presented in Section 6.

Warburg Respirometer Oxidation Experiments

     The preliminary work described above showed that it is practical
to work with organisms grown only in a mixture of pyrite and simulated
ground water, and that the culture under study could be  grown directly
in the respirometer.  However, a decision was made to alter the method
of experimentation for two reasons.  First, when considering water and
oxygen transport mechanisms in natural systems such as underground
mines and gob piles, both the continuous flushing of the pyrite bed and
the water-pyrite ratio inherent to the apparatus used by Bailey appear
to be far out of proportion compared to probable natural conditions.
In any natural environment, continuous flushing of pyrite surfaces with
large volumes of water would, in general, require that the pyrite be
completely inundated.  Because of the low diffusivity of oxygen in water,
this would effectively block the availability of oxygen  in quantities
sufficient to support commonly observed pyrite oxidation rates.  It is
much more likely that the rate of flow of water past actively oxidizing
pyrite in natural systems is very slow, or that there is no continuous
flow at all other than the seepage induced by the water  condensed by
hygroscopic salts of pyrite oxidation under the high humidities present
in underground systems.  Intermittent flushing of surfaces will occur
by changing ground water levels in underground mines, and by infiltra-
tion of rain water or snow melt in the case of gob piles.  Thus, a
natural environment is more accurately approximated by a batch system,
readily obtainable in a simple Warburg respirometer.  The water-pyrite
ratio, which is determined in a natural system by the nature and per-
meability of the mineral matrix surrounding the pyrite,  can also be
readily varied in the Warburg apparatus.

     The second major problem encountered in the preliminary work was
with the analytical methods.  The nature of the carboy respirometer
made accurate oxygen uptake rate determinations difficult, and results
tended to be erratic.  Further, bacterial counts by serial dilution
were time-consuming, and results were subject to a two-week incubation
delay.  Use of the Warburg eliminated the former problem; to cope with


                                   38

-------
the "bacterial count problem, it was decided to employ direct counts,
using phase contrast microscopy and a standard Lineweaver-Burke counting
chamber.

     Objectives

          The specific objectives of this phase of the work were to
describe the growth of organisms of the Ferrobacillus-Thiobacillus group
in a system consisting of varying amounts of crushed pyrite and simulated
ground water, to determine the relationship between bacterial activity
and overall pyrite oxidation kinetics, and to determine whether the
indirect model or the bacterial adsorption model is more likely to be a
satisfactory model of the bacterial catalysis mechanism in natural
systems.

     Materials and Procedures

          All Warburg runs were made at 20C, using 1^0 ml flasks and
a shake rate of 80/min, with a 1-1/2 inch stroke measured at the manom-
eter stopcock.  No C02 absorbent was used, and carbon uptake due to cell
growth was estimated to be insignificant compared to the gross gas up-
take observed.  Thus, gross uptakes were taken to be equal to oxygen
uptakes.

          The seed bacterial culture used was obtained by incubating
acid drainage from the McDaniels Mine in Vinton Co., Ohio, with a mix-
ture of simulated ground water and crushed pyrite.

          The pyrite used was 80-100-mesh crushed sulfur ball, taken
from the McDaniels Mine and washed with dilute HC1 and distilled water
before use to remove fine carbonaceous material and iron salts.

          In all experiments, the initial gas phase in the flask was
air, and care was taken to assure that the partial pressure of oxygen
did not drop below 0.2 atm during the course of a run.  Daily flushings
with fresh air ensured both a constant oxygen content and the continued
availability of atmospheric C02.

     Results

          Figure Ik is a typical cumulative oxygen uptake versus time
curve for a mixture of O.U g pyrite and 20 ml of ground water, inocu-
lated with bacteria at zero time.  For the first 100 hours of the run,
the rate of oxygen uptake was 10.k ng/hr.  Between 100 and 180 hours,
the rate increased steadily, leveling off at 180 hours at a rate of
150 |_ig/hr.  An indication of the significance of this type of curve is
seen in Fig. 15.  Here, the first 180 hours of data from Fig. lU are
replotted, together with data from an identical reaction mixture which
was not inoculated with bacteria.  Thus,  it is seen that the 10.k |_ig/hr
base rate is apparently the rate due to chemical oxidation by oxygen.
                                  39

-------
g

o
o
 CM
O
LU
I
    18
    16
    14
    10
8
    0,
     0       40      80     120      160
                             TIME(hr)
                                          200
240
               Fig.  1^ - Oxygen Uptake Curve
                        (0.4 g pyrite in 20 ml solution)

-------
                     o
                     liJ
                     X
                     o
                                     O
                                     CD
                                                        0)

                                                       -P
                                     O
                                     CM
                                     O
                                     00
                                                     o>  CM
                                                               

                                                               03 -H

                                                              U
                                                              HD  ft
                                                    LTN
                                     O
ID
N">
     CJ
o
111

-------
 The  increase  in rates past 100 hours must then be due to bacterial
 activity.  Figure  16 is a semilogarithmic plot of bacterial concentra-
 tion versus time.  Bacterial growth appears to follow a normal first-
 order  relationship up to a bacterial concentration of about 2 x 10s
 cells/ml,  at  which time the population appeared to reach its maximum
 limit.   It is significant to note that the time at which the oxidation
 rate stabilized at 150 p.g/hr coincides with the time at which the
 bacterial  population ceased to increase, indicating that the bacterial
 population was controlling the rate of pyrite oxidation in this case.

           If  one analyses these data in terms of the indirect model for
microbial  catalysis, then the chemical base rate (oxygenation rate) and
 the  oxidation by bacterially produced ferric ions will proceed indepen-
 dently,  with  the total observed rate being the sum of the two.
 Figure 17  shows a  plot of bacterial oxygen uptake rate versus time,
 obtained from the  data of Fig. 15 by subtracting out the base chemical
 rate of  10. U  |_ig/hr.  Although the slope of the curve is not indentical
 with that  of  Fig.  16, it is seen that the bacterial oxygen uptake rate
 is approximately proportional to the population of bacteria.  In regard
 to the limiting bacterial population of 2 x 10s per ml, this value was
 a characteristic maximum population when the simulated ground water was
 used.  Growth of the organisms in a mixture of unwashed pyrite and dis-
 tilled water  gave  about half this value, or 5 x 107 per ml, and commonly
 quoted maximum populations in Silverman's 9~K media are on the order of
 108  per  ml.   Whether this limiting population is due to limited nitrogen
 source,  or to toxic product buildup or to some other limitation, is not
 known.   It should  be mentioned, however, that if the nitrogen source in
 Bailey's groundwater medium is increased by a factor of ten, the limit-
 ing  population is  increased by a factor of approximately ten.

           Further  insight into the relationship between bacterial popu-
 lation and the catalyzed oxygen uptake rate may be gained by evaluating
 the  rate versus population data in light of a simple model.  If it is
 assumed  that  catalyzed oxygen uptake rate is first-order with respect
 to cell  population, then

                d02/dt = kH  (N = bacterial population)

 and

            log d02/dt = log N + log k.

 Thus,  oxygen  uptake versus total cell population should plot on log-log
 paper  as a straight line with a slope of unity.  Figure 18 is a plot of
 oxygen uptake rate versus total cell population in the flask.  While
 the  data show considerable scatter, there is a distinct tendency toward
 a unit slope  at lower cell populations, where the organisms are still
 in the log growth  phase.  During the 1:1 section of the curve in
 Fig. 18, the  oxygen uptake rate per cell is from 2.5 to 5 x 10~7 fig
 02/cell-hr, depending on whether the upper or lower envelope is used

-------
    I08

-------
    200
    100
01

It.



LU


5
(T


UJ
 CM

O
o
<
GO
      10
      0
        0
50
100
150
200
250
                              TIME(hr)

         Fig. 17 - Change of Microbial Oxidation Rate with Time

-------
Q.

2
UJ
o
          o
         CM
         K

          
-------
for the calculation.  At the higher cell populations,  when the organisms
are approaching the stationary phase of growth,  the apparent respiratory
activity per cell decreases considerably below these values.

          The above data give some insight into the nature of growth of
autotrophic iron-oxidizing organisms in a pyrite-ground water system,
but they do not definitely prove whether bacterial catalysis is indirect
through the oxidation of ferrous iron in solution, or  direct through
adsorption of the cells on the pyrite surface.  Figure 19 presents
evidence that the catalysis is through the indirect mechanism.   Here,
the chemical and the maximum biological oxygen uptake  rates are plotted
against the pyrite-to-water (wt = volume) ratio in the Warburg flask.
The water volume was held constant at 20 ml in each flask, and the
pyrite samples were 0.1, 0.2, O.U, and 0.6 g.  While the chemical rate
increases linearly with pyrite weight increase in the  flask, the bio-
logical rate shows a linear increase initially,  but a  tendency to
approach a constant limiting rate at the higher pyrite-to-water ratios.
While this phenomenon cannot be explained readily in terms of an adsorp-
tion mechanism, it is entirely compatible with the indirect mechanism.
At a constant volume of water in the four flasks, and  a limiting bacte-
rial population of about ID3 per ml, the total bacterial population and
hence the total oxidative capacity of the bacteria has a specific maxi-
mum limit.  In accordance with the ferric ion kinetics presented in the
following section of this report, a given type and grain size of pyrite
has a maximum specific oxidation rate by ferric iron.   If the oxidative
capacity of the available bacteria surpasses this rate, then the weight
of pyrite limits the oxygen uptake rate.  As the weight of pyrite
increases, however, the ferric ions cannot be supplied rapidly enough
by the available bacteria, and the maximum rate of pyrite oxidation is
equal to the maximum rate of ferrous iron oxidation by the bacteria.
The chemical rate of pyrite oxidation by oxygen will continue to climb,
however, and at some pyrite-to-water ratio the chemical rate will sur-
pass the biological rate.  From the available data this ratio is calcu-
lated to be about UOO grams pyrite to one liter of water.  This, of
course, depends on the type and grain size of the pyrite.

          The implications of bacterial catalysis through the indirect
mechanism are very significant.  In a natural system having a low
pyrite-to-water ratio, the overall oxidation might be  due primarily to
bacterial catalysis.  However, in case of a high pyrite-to-water ratio,
there will be insufficient water to support the necessary bacterial
population, and the reaction will be essentially chemically controlled.
Oxygen transport considerations indicate that this is  probably the  case
in underground mining systems.

-------
        cr
        cr
        UJ

        I
        o
        LJ
        H
        (T

        Q_
                K

                 M
                 0)
                 O
                p
                -p
                H
(D

8
 cl
 0>
 bD
 >>
 X
O
                 bO
                H
o

-------
                               Section 6

                         FERRIC ION OXIDATION

     In the preceding section, microbial catalysis is shown to operate
through a mechanism involving the oxidation of ferrous ions to ferric.
The ferric ions then become the immediate oxidizing agent for pyrite.
These experimental data also show that indirect oxidation is of primary
importance, although they do not rule out a minor contribution by direct
oxidation.  If the indirect mechanism is the major microbial rate-
determining mechanism, then ferric ion oxidation is the chemical anal-
ogy of microbial-enhanced oxidation.

     From  a kinetic point of view, the overall reaction rate is a
steady-state value determined by the relative rates of two reactions:
(a) oxidation of ferrous ions to ferric  (Fe++  - e ->Fe+4~+) and (b)
reduction  of ferric ions by the pyrite oxidation reaction (l^Fe++  +
FeS2 - 15Fe+++ + 2S6+).

     In the microbiological system, the rate of ferrous ion oxidation
to ferric, and therefore the ferric-ferrous ratio reached, is determined
by microbial activity.  In the experimental chemical system,  the oxida-
tion of ferrous ions is controlled chemically by addition of an oxidiz-
ing agent at such a rate to maintain a constant ferric-ferrous ratio.

     The real system is self-regulating.  The higher the microbial
activity (or higher the ferrous ion oxidation rate) the higher the
ferric-ferrous ratio becomes.   With an increase in ferric-ferrous ratio,
the rate of ferric ion reduction by pyrite increases.

     The ferrous-to-ferric oxidation is a solution reaction.   The ferric
ion reduction reaction is a heterogenous reaction so that the reduction
rate is a function of surface area of pyrite.  Therefore the ferric-
ferrous ratio reached in a given pyritic system depends on the exposed
surface area of pyrite.

     The chemical (as compared to microbiological)  system is much more
reproducible and easily controlled.  It is possible to evaluate the
microbial oxidation rate in a real system from a knowledge of the rate
of ferrous oxidation or ferric reduction.  Therefore a study has been
made of the kinetics of the ferric ion oxidation reaction to help
evaluate the relative effectiveness of microbial and oxygenation systems.
Results of this study are given below.

Experimental Equipment

     The equipment shown in Fig. 20 was developed to measure oxidation
rate at any selected value of EMF (or

-------
                                  UJ
                                  cr
                                  a:
                                  UJ
                                  0:  _J
                                  5  o
                                  uj  oc
UJ
h-

LU
CC

CO
V.i..,.i..,Ti
                                                                   ft
                                                                     g
                                                                     H

                                                                     8
                                                                   o
                                                                   H
                                                                   (U
                                                                   O
                                                                   CM
                                                                   UD
                                                                   H
                                UJ
                                 I

-------
     A mixing vessel contained the E^ electrodes (platinum, and calomel
reference cell), stirrer, automatic titrator delivery tube, inlet and
outlet lines for the recirculated fluid, and a nitrogen purge line.
The liquid to the bed of pyrite (reactor) was recirculated by a peri-
staltic pump.  The EMF of solution was controlled by an automatic titra-
tor which added either Fe2(S04)3 or KMn04 to maintain a constant ferric-
ferrous ratio in solution.

     When combined oxygenation and ferric ion oxidation runs were made,
all lines and electrodes to the mixing vessel were sealed in a tight-
fitting cover, the stirrer was removed since adequate agitation was
obtained by directing the flow of inlet recirculated solution, and the
nitrogen purge line was replaced by a line to a metered supply of oxygen.
The top of the burette of the automatic titrator was connected to the
vapor space above the solution so the total vapor volume of the system
remained the same as the titrant was added to the mixing vessels  In
this way the oxygen consumed could be determined by the quantity of
oxygen added to maintain a constant pressure on the system.

Experimental Procedure

     Initial ferric ion oxidations were run at low pH's or, put in
another way, at high sulfate concentrations since the low pH values
were reached by addition of sulfuric acid.  In light of more recent
work, the important point to be made is that these initial runs were
made at essentially constant sulfate concentration and constant ionic
strength.  Later studies were made to determine the effect of pH on
ferric ion oxidation rates.   Interpretation of these data indicate that
the major variable involved is anion concentration, not pH.

     Experimental results are therefore presented in several data sets:

       I.  For high sulfate concentration, low pH,  and constant
           ionic strength as a function of

            A.  EMF or ferric-ferrous ratio,
            B.  total iron concentration at high ferric-ferrous
                ratios,  and
            C.  type of pyrite.

      II.  Combined oxygenation and ferric ion oxidation runs.

     III.  At varying pH  and anion concentration as a function
           of

            A.  EMF or ferric-ferrous ratio and
            B.  total iron concentration.

     All rates for ferric ion oxidations are based on the observed
rate of ferric ion reduction and a stoichiometery of lU moles  of ferric

-------
ions to oxidize one mole of pyrite (with iron ion as ferrous).   The
stoichiometery was checked experimentally and found to agree with the
theoretical value of 1^, within limits of experimental error.   Rates
are calculated in terras of g. mole pyrite oxidized per hour per  gram of
pyrite.  It should be noted that oxidation rates expressed as (igOp con-
sumed per hour per gram of pyrite multiplied by 0.0098 is equivalent  to
micromoles of pyrite oxidized per hour per gram of pyrite (when a stoi-
chiometry of 3~l/2 mole of oxygen per mole of pyrite oxidized  is assumed).

Experimental Results

     Low pH, High Sulfate Runs

          Several series of runs at low pH and high sulfate concentra-
tions were made, using three different pyrite samples.  Experimental
data are given in Tables VI-X.

          For some as yet unknown reason, data on the museum grade
pyrite could not be consistently reproduced.  Consistent data were
obtained on a set of runs (e.g., a series of runs at different  ferric-
ferrous ratios) if operated continuously.  But if the reactor were shut
down overnight, the rate the following day would often be higher or
lower by 25 to 50$  Sulfur ball material was much more predictable;
the overall rate decreased slowly and regularly with reaction  time as
the pyrite was consumed.

          Two consistent sets of data for the museum grade pyrite are
given in Tables VI and VII.

          Similar data for McDaniels Sulfur Ball sample and Sulfur Ball
Ho. 2 are given in Tables VIII and X and graphically in Figs.  21 and  22.

          The ferric-ferrous ratio was calculated from the equation
EMF = O.U30 + 0.059 log (Fe"H"+/Fe++) as determined experimentally for
these series of runs at low pH and high sulfate concentrations.

          All rates within a series were recalculated to the same base
rate for this series because of a small but consistent change  in base
rate during the course of the series.  The change was determined experi-
mentally by periodically performing a base rate run under base conditions
and noting the change in rate over the operating period from one base
run to the next.  The change was assumed to be due to consumption of
pyrite through oxidation.  In some cases, more than 25$ of the original
pyrite charge was consumed.

     Treatment of Data

          A simple "dual-site" adsorption model gives a rate equation
that correlates experimental data within limits of experimental error.
                                  52

-------
     Table VI.  Museum Grade Pyrite

 Rate vs. Cone, at EMF = 0.700
Total Iron 1*
(gA ) ^ VFe^TT
0.125 0.83 21.1
0.250 1.25 ll*.9
0.500 1.6 10.6
1.00 2.2 7.^5
2.00 3-1 5-3
l+.OO l+.O 3.73
8.00 i*.9 2.6U
10.00 5.2 2.36
Fe+++ in moles/liter
Note: "r" (rate) in units of
oxidized per hour per 
1
r
1.21
0.80
0.63
0.1*5
0.3^
0.25
0.20
0.195
l_ig-mole FeS2
rram of sample
     Table VII.  Museum Grade Pyrite

Rate vs. FMF; Iron Cone. = 10 g/l	
            p +++
EMF         TTT      rexp      rcalc^
            Fe
0.700
0.650
0.600
0.550
0.500
0.1*77

calc

Note:

37,700
5,360
760
108
15. 1*
6.3
20 -
~~ r
/ 1
1 Fe+++
5-2 5.2
3.8 2.6
1.36 1.35
0.51* 0.59
0.23 0.23
0.15 O.il*
6.7YFe++/Fe+++

Fe++/Fe+++
"r" (rate) in units of ^g-mole FeS2
oxidized per
hour per gram of sample
                    53

-------
     Table VIII.   Rate vs.  Iron Concentration

EMF = 0.6^0 V, pH = 0.2;  Sulfur Ball #2
Concentration
(mg-mole/^)
0.097
0.15
0.21
0.28
0.^2
0.75
l.la
2.06
2.76
3.28
u.u
5-3
5.8
10.9
Note: "r" (rate)
Oxidation
Rate
l.U
1.9
2.U
2.6
3-2
3-96
5.1+
6.3
6.6
7.7
8.9
9-3
10.3
13.7
in units of
oxidized per hour per
1/Conc.
10.3
6.7
U.7
3.6
2.U
1.33
0.71
0.148
0.36
0.31
0.23
0.19
0.17
0.092
l_ig-mole FeS2
gram of sample.
I/Rate
0.72
0.52
0.1+2
0.39
0.31
0.25
0.19
0.16
0.15
0.13
0.11
0.1075
0.097
0.073


      Table IX.   Rate vs.  Iron Concentration

EMF = 0.650 V; pH = 0.5; McDaniels Sulfur Ball
        Run 23                      Run 50
      Fe Concentration            Fe Concentration
19.3
18.7
1U.3
10.9
8-5
6.5
5-3

5.0
3-85
2.15
1.10
0.59
0.31
0.183

19-1
1^.2
11.3
9-1
U.8
11.7
8.2
6.2
U.3
2.26
1.26
0.675
0.193
1.28
0.5^
0.29
Note: "r" (rate) in units of |ag-mole FeS2 oxidized
      per hour per gram of sample.

-------
            Table X.  Rate vs EMF, Sulfur Ball #2, Ref. 16

          Iron Cone. = 1 g/l, pH = 0.2
EMF
0.65
0.60
0.55
0.50
0.477
0.460
0.430
165 - 5
l +
rexp
14
12.5
11
6.1
3.6
2.1
0.56

5^Fe++/Fe+++
2.3 + 7oJFe++/Fe++
rcalc*
14
12.3
9.2
5.4
3.6
2.7
1.3

+
          As one possibility, assume a ferric ion is adsorbed, on two
"reactive sites" of pyrite (dual-site adsorption).  Also assume ferrous
ions  (Fe++) compete for these dual sites.  The activated complex formed
by ferric ion adsorption is decomposed by electron transfer from one of
the reactive sites to the ferric complex, thereby forming an adsorbed
ferrous ion, which is then desorbed.

          Using Hougen-Watson10 concepts, the following rate equation
was derived:
                                                                    (3)
where

           k^f = rate constant, electron transfer reaction

             K = equilibrium constant, electron transfer reaction

            ],_ = adsorption equilibrium constant for ferric ions

            K2 = adsorption equilibrium constant for ferrous ions

    Fe++,Fe+++ = concentration of ferrous and ferric ions


                                   55

-------
                                                                                       I

                                                                                       i
                                                                                       w
                                                                                       H
                                                                                       Q)
                                                                                       H
                                                                                       id
                                                                                       cd
                                                                                       cci  cd
                                                                                       rt FQ
                                                                                       OJ
                                                                                       OJ
                                                                                       bD
                                                                                       H
                                   UOI)
                       O
                       in
O
T
R
O
00
o.
ID
o

-------
          At high EMF's, where the  square root of the ferrous-ferric
 ratio  (Fe++/Fe+++)  is negligible, this equation may be written
                           r =
                                 Fe+++
                                        -JsT
 or
                                           t/^T
                                                                     (5)
          A plot of 1/r vs . I/ v Fe+++ for ferric ion oxidation at high
 IMF's  should  give a straight line with a slope equal to 1/k and an
 intercept equal t
          These data for the three pyrite samples shown in Tables VI,
VIII, and IX are plotted in Fig. 21.
          After k and \K  are calculated from the slope and intercept ,
 k'  and -/Kg can be determined in the following manner:  Assume that the
 reaction rate goes to  zero as EMF approaches 0.370, or in other terms,
 as  the ferrous -ferric  ratio approaches 9-0-  (note discussion of
 potentiostat measurements on page 79 regarding pyrite  reference cell
 voltage as current goes to zero) .  These data were interpreted as an
 indication of a zero oxidation rate at some low oxidation-reduction
 potential rather than  a means of determining the ferrous -ferric ratio
 when the rate drops to zero.  The value of 3-0 for the ferrous-ferric
 ratio  is taken as a single correlating factor to give best data fit
 (although potentiostat data indicate this value is a function of iron
 and sulfate concentration) .

          By setting the numerator of Eq. (3) equal to zero when
N/Fe++/Fe+++ = 3.0, k is found to equal k/3-

          K2 can be determined from one value of the Rate vs. EMF data
where the ferric-ferrous ratio is significant.

          The data presented in Table VIII for Sulfur Ball #2 may be
used to illustrate the calculation of the Rate Equation.

          From Fig. 21, the slope of the line for Sulfur Ball #2 = 0.0061
and Intercept = 0.02, then
                                   57

-------
               k = 1/0.0061 = 165

              k' = 165/3 = 55

                 = Intercept x k = 3.3
          (Note:  since /Fe++/Fe+++ is not negligible at EMF = 0.650,
the intercept value of 33 includes  K2 times ^Fe++/Fe+++ .   By trial
and error, Kx = 3.3 - 70 x 0.0137 = 2.3, when K> is calculated from
rate at EMF = 0.1+77.  Therefore the rate equation, as calculated using
Fig. 21 and one point from Rate vs. EMF data is,
                         165 - 55 yFe++/Fe+++
                                                                   (6)
                        1Fe+++
                               + 2.3 + 70
          Table X compares calculated and experimental rates for these
data.

          In like manner, the rate equations for the experimental points
given in Fig. 22 can be calculated.  The solid lines on Fig. 22 are loci
of the calculated rates for different total iron concentrations.  Calcu-
lated and experimental rates for museum grade pyrite are compared in
Table VII.

     Interpretation of Data

          The excellent correlation of experimental data by Eq. (3)
indicates that the form of the adsorption equation, if not the specific
mechanism used to derive it, is relevant.  It appears that the relative
adsorption of ferrous and ferric ions is rate-determining in ferric ion
oxidation.  A comparison of adsorption equilibrium constants for ferrous
and ferric ions is surprising.  For sulfur ball pyrite the ratio of
Kg/K-L (ratio of adsorption equilibrium constants for ferrous and ferric
ions, respectively) is 900 for Sulfur Ball #2, 2500 for McDaniels Sulfur
Ball, and ^5?000 for Museum Grade pyrite.  In other words, the relative
adsorption of ferrous ions is much greater than ferric for all types of
pyrite.  The selective adsorption of ferrous ions is particularly great
on museum grade pyrite; i.e., more than 20 to 50 times greater than on
the sulfur ball.  It is also interesting to note that the reactivity
of the pyrite samples is inversely related to Kg/K^.  The rate curves
for the three pyrite samples are compared in Fig. 23.

          The data show that ferric ion oxidation rates are a function
of total iron concentration, and are determined by the ferrous-ferric
ratio, not E^.  As the E^ of the solution is raised to the point where
all iron is ferric, the rate becomes constant.  Further increase in E^
has a negligible effect on rate.

                                  58

-------
LU
      on
      01
       bO
      H

-------
           It should be  again  noted that all these runs were made at
 low pH (0.2 to 0.5) and similar sulfate content and ionic strength.

 Combined Oxygenation  and Ferric Ion Oxidation

      Results of runs  in which both oxygen and ferric ions were present
 are given in Table  XI.   The concentration of each oxidizing agent and,
 in the case of ferric ions, the ferric-ferrous ratio also were selected
 to give similar oxidation rates.  McDaniels Sulfur Ball was the pyrite
 material used.
         Table XI.   Combined Oxygenation and Ferric  Ion .Oxidation:
                    McDaniels Sulfur  Ball
Run
No.
31
33
3U
35
38
1+1
^
EMF
_
500
-
500
500
550
550
Iron Cone.
(g/*)
_
1.00
-
1.00
o.Uo
0.20
0.20
Rate
(by lOO/o 02)
3.6
3.6
3.5
3.U
3.5
3.U
3.5
Rate
(oy Fe+++ ions)
(aerobic only)
U.5
(aerobic only)
l+.U
3.5
5-1
5.0
     These combined oxygenation-ferric ion oxidation runs show the
independence of the two reaction modes.  The oxygenation rate is not
influenced by solution EMF or ferric-ferrous ratio, and the ferric ion
rate is not changed by the partial pressure of oxygen.  This leads to
the conclusion that the reactive sites for the two oxidation mechanisms
are not the same

     It is interesting to note that the increase in oxidation rate, over
the oxygenation rate, is approximately the same for samples heavily
inoculated with Ferrobacillus ferrooxidans12 and samples subject to
ferric ion oxidation at high ferric-ferrous ratios.  This observation,
together with those discussed in the Microbiology section, leads to the
conclusion that bacteria such as Ferrobacillus ferrooxidans function
to generate a high ferric-ferrous ratio in solution.  The rate of oxida-
tion by ferric ions would then be the same in both a biological or chemi-
cal system, and determined by the ferric-ferrous ratio and total iron
concentration.  In real systems, the ferric-ferrous ratio is a steady-
state value determined by the relative rates of (l) microbial oxidation
of ferrous ions to ferric and (2) reduction of ferric ions by reaction
with pyrite.
                                   60

-------
     These data provide a basis for defining the reaction regime (i.e.,
oxygenation or ferric ion oxidation if data on oxygen concentration,
total iron, and the ferric-ferrous ratio at the reaction site are known.
Oxygenation and ferric ion oxidation rates are approximately the same
at oxygen partial pressures of 0.21 atm, when EMF = O.U^O.  This cor-
responds to a ferric-ferrous ratio of 2.2 or 70$ of iron ions in the
ferric state.  At an EMF of O.UO, where 2U$ of iron is ferric, the
ferric ion rate is one-fifth to one-tenth the oxygenation rate in air.
On3y with microbial-enhanced oxidation can ferric-ferrous ratios this
high be attained.  In other words, if the ferric-ferrous ratio is less
than 0.3 (2.k% ferric) and the partial pressure of oxygen is 15 to 20$,
the system is in an oxygenation regime; i.e., the oxidation rate is
determined by the chemical oxygenation mechanism.  If oxygen vapor con-
centration is less than 2% and 70$ of the iron is in the ferric state,
the system is in a ferric ion oxidation regime, generated by microbial
activity.  Although this last observation is true regardless of micro-
biological activity, it may have little practical significance.  The
observation is based on the premise that microbial rate is not influenced
by oxygen concentration above a dissolved oxygen concentration of 0.1 ppm;
that is, a 70$ ferric concentration can be microbially generated at low
oxygen levels.  More recent data indicate that microbial activity may
be limited by oxygen concentration at much higher levels than 0.1 ppm.

     It seems evident that no significant amount of pollution is pro-
duced by microbial activity from a pyritic system which has an aerobic
effluent with a ferric-ferrous ratio less than 0.2 (15$ ferric ions).

     Since ferrous ion oxidation is a solution reaction, the majority
of bacteria will be found in solution.  Even if the solution should
pass through a section of highly pyritic material (i.e., where the
number of bacteria per unit exposed pyrite surface is low because of
large pyrite surface) the ferric-ferrous ratio would not drop much
below 1.0 since the rate of ferric ion oxidation of pyrite (or ferric
ion reduction) decreases rapidly as the ferric-ferrous ratio decreases.
The bacterial ferrous ion oxidation would remain the same.

     At the same time it is not necessarily true that acid effluents
having high (more than 70$ ferric ions) ferric-ferrous ratios were
generated by microbial action.  In underground mines for example, the
effluent water may have a long residence time (ponding) after leaving
the reaction site.  The ferrous ions may be microbially oxidized after
contacting pyrite so that the ferric ions are formed in a pyrite-free
environment and do not contribute to the actual oxidation of pyrite.

     In other words, significant microbial-enhanced pyrite oxidation may
be discounted in pyritic systems having a low ferric-ferrous ratio aero-
bic effluent (less than 0.2).   A high ferric-ferrous ratio in low-pH
mine waters does not prove that microbial activity was important in the
actual oxidation of pyrite.   It only proves that microbial oxidation of
ferrous ions to ferric has occurred.   If a high ferric-ferrous ratio


                                  61

-------
is generated while the solution is in the presence of exposed pyrite,
ferric ion oxidation will occur.

     The controversy that has developed over the relative significance
of oxygenation or ferric ion oxidation may lie in the difference
between museum grade and sulfur ball pyrite.

     As noted earlier, the ratio of adsorption equilibrium constants
(K2/K1) for museum grade and sulfur ball pyrite differ by a factor of
20 to 50j indicating the selective adsorption of ferrous ions is greater
on museum grade pyrite.  The difference in oxygenation rates for the
same samples is equally significant.  As given in Appendix V, the
oxygenation rates for museum grade pyrite and McDaniels sulfur ball are
0.03 and 3-5 M-g mole/hr,g.  The corresponding maximum ferric ion oxida-
tion rates for museum grade pyrite and McDaniels sulfur ball are (from
Fig. 23) 6 and 12 \j.g mole/hr,g.

     The ratio (maximum ferric ion oxidation rate/oxygenation rate) for
museum grade pyrite is 6/0.03 = 200.  The same ratio for the sulfur ball
sample is 12/3-5 = 3.^.  If one were to judge the relative importance
of ferric ion oxidation and oxygenation on data obtained on museum
grade pyrite, an erroneous conclusion could be reached in respect to
relative rates which exists in real pyritic systems.

     Also note that the maximum ferric ion oxidation rates quoted above
are higher than could be reached in real systems even under ideal con-
ditions.

Varying pH, Type and Concentration of Anion

     Reports9 on oxidation by ferric ions noted a negligible effect of
pH on rate for pH's below 2.0.  An investigation of the accuracy of
this report has led to an interesting study involving effect of pH or
anion concentration on the kinetics of ferric ion oxidation.  Data are
presented in Appendix TV.

     As shown in Figs. 2^-27, oxidation rate changes greatly in both
sulfuric and hydrochloric acid solutions, as the pH is varied from 0.5
to 2.0.  In both solutions, the rate is higher when the pH is higher.
Since the pH was adjusted by adding either sulfuric acid or hydrochloric
acid, a decrease in pH was accompanied by an increasing in anion con-
centration.  Therefore, it is necessary to determine which of the two
variables, pH or anion concentration, affects the rate or whether both
variables are influential.

     Figure 25 shows that at equal pH values the reaction rates are
always lower when HS04~ concentrations are higher.  In solutions having
pH = 2.0 and HS04~ concentrations of 0.32 g-mole/g, the rates are com-
parable to those in solutions having pH n 0.5 and HS04~ concentration
of 0.^9 g-mole/g.  It therefore appears that the sulfate concentration


                                  62

-------
                                              CM


                                              to
                                              H
(9oi*)3ivy  NOIIVQIXO

-------
                                                 
                                                H
                                                s
                                                 0)
                                                 O

                                                 g
                                                o

                                           --^       -p
                                           UJ
                                                CQ
                                                p
                                                 o
                                                 Q)
                                                Lf\

                                                CM
                                      oo
31VH  NOI1VQIXO

-------
            p  in   in
          j  <\i  d   o
                                                                    g
                                                                    H


                                                                    H
                                                                    o
                                                                    CO
                                                                    o

                                                                    a
                                                                    
-------
                                                                        I
                                                                        o
                                                                        CO
                                                                        (D
                                                                        H
                                                                        O
                                                                        3
                                                                    2   
                                                                    V)
                                                                    O
                                                                    tr
                                                                    tr
                                                                    o
                                                                    tr
                                                                    cc
                                                                    LJ
                                                                        C!
                                                                        H

                                                                        W
                                                                        0)
                                                                        
                                                                        W
                                                                        H
                                                                        m
                                                                        cd
                                                                        PH
                                                                        O

                                                                         I

                                                                        c-
                                                                        CM
00
 10      ^r       n      CM
(QOI*)31Vd  NOI1VQIXO
                                   66

-------
is the primary influence on rate and not pH.  Following this premise
further, it is probable that the increase in rate with increasing pH
in chloride solutions actually results from the decrease in chloride
concentration.

     There are two possible mechanisms by which anion concentration can
affect the oxidation rate; i.e., (l) the anions may be adsorbed on the
pyritic surface, reducing the available sites for adsorption of ferric
ions, or (2) complexes of ferric ions are formed with the anions, and
the complexes are not as readily adsorbed on the pyrite surface.

     It is possible that sulfate ions can be adsorbed on the pyrite sur-
face since sulfate is an oxidation product of pyrite  It is unlikely
that chloride ions would be adsorbed to the same degree, at least to the
extent they need be considered in competitive adsorption for reactive
sites.

     In sulfate as well as chloride solutions, several complexes of
ferric ions can be formed.  Whatever form of complex the ions have with
the anion, the formation of complexes cause a reduction in the net posi-
tive charge of the group.  Since the reduction of ferric ions requires
the transfer of an electron to the ferric ions, it is reasonable that
Such transfer will be retarded because of the negative charges of the
complexed anion (or anions) surrounding the ferric ion.  Furthermore,
the diameter of the complexed ions would be larger than that of non-
complexed ferric ions.  In the case of sulfate complexes, the diameter
of the complexed ferric ion would be even larger than that of hydrated
ferric ion.  This fact can very likely account for the difference in
degree of adsorption of complexed ions compared to free ferric ion.

     Based on the above conditions, one can postulate that the only
ferric ions which are active in the reaction with pyrite are those that
are not complexed.  In order to evaluate this postulate, it is necessary
to determine the composition of the ionic species in the reaction solu-
tion.

     Ionic Composition of the Reaction Solution

          To be able to determine the ionic composition of the reaction
solution,  we must know the type of iron complexes in the solution and
the ionic  equilibrium relationship between the species in the solution.
A brief description of iron complexes in sulfate and chloride solutions
is given below.

     Iron Complexes in Sulfate Solution

          Several complexes of ferric and ferrous ions have been
identified.13'25>27   The  equations  below describe the  complex formations
and association reactions of the iron ion in sulfate solution in the  pH
range involved.


                                 67

-------
                         Fe+++ + S04~ = FeS04+


                        Fe+++ + HS04" = FeHSO^

                       FeS04+ + S04" = Fe(S04)2~


                                 HS04 = H+ + S04~~


                          Fe+++ + H20 = FeOH++ + H+

                                FeS04 = Fe++ + S04~~

          Considering the equilibrium relationship of each reaction,
the composition of the ionic species in solution can be calculated.   In
Appendix III a detailed description of the calculation procedure is
given.  For the calculation, equilibrium constant values (also known  as
stability constants) were taken from the literature.20  These values
are also tabulated in Appendix III.  To be used in the calculation, the
values of the stability constants were corrected to values correspond-
ing to the ionic strength of the reaction solutions.  The equation pro-
posed by Davies4 for calculating activity coefficient was used for this
purpose.

     Iron Complexes in Chloride Solution

          For the chloride solutions, the equilibrium reactions below
were considered.


                          Fe+++ + el" = FeCl++

                         FeCl++ + Cl" = FeClp+

                         FeCl2+ + Cl" = FeCl3

                           Fe++ + Cl" = FeCl+

                          FeCl+ + Cl" = Feel,

                          Fe+++ + H20 = FeOKf+

The complexes formed in the reactions listed above have been identified
by Rabinowitz and Stockmayer.1Y  Their stability constants are given
in Appendix III.
                                  68

-------
          Procedures similar to those used to calculate the ionic com-
position in sulfate solution were applied.  A detailed description is
given in Appendix III.

          Comparing the free ferric ion concentration in chloride and
sulfate solutions, it can be seen that at a given ferric-ferrous ratio
and pH value the free ferric ion concentration is the same.  As shown
in Fig. 26, however, the rates in chloride solution are always higher
than the rates in sulfate solution.  This can be explained if it is
further assumed that, in addition to forming complexes with ferric ions,
the sulfate ions also directly inhibit the oxidation reaction.  A pos-
sible explanation for this is that some sulfate ionic species are capable
of being adsorbed on the reactive sites of the pyrite surface.  Unfor-
tunately the experimental data to evaluate this assumption directly are
not available.

     Anion Concentration and pH Effect on Kinetics

          Based on present information, detailed evaluation of pH or
anion concentration effects on ferric ion oxidation kinetics is not
warranted.

          In the first place the stability constants for the assumed
complexes are only approximate.  Literature values, in addition to being
somewhat inconsistent, are not directly applicable to solutions of high
ionic strength used in this work.  Depending on the values used for the
stability constants, wide variations in ion concentrations result from
calculations.

          In addition, most of the original experimental data did not
include a direct analysis of anion concentration or ferric-ferrous
ratio.  For example, sulfate ions were added as sulfuric acid to con-
trol pH, not to provide a sulfate concentration.  At the low pH's used,
this is not a good quantitative method for determining sulfate content.

          Considering these facts, the best quantitative interpretation
can be made from sets of runs at fixed ferric-ferrous ratios, or con-
stant ionic strengths.  Under these conditions, actual changes in a
variable will be (approximately) porportional to the calculated change
of the variable being studied, even if the absolute value of the variable
is in error.  For example, we may not know the actual value of the free
ferric ion concentration because of errors in stability constants, but
the change in ferric ion concentration with change in total iron con-
centration (at a given ferric-ferrous ratio) will be in the same pro-
portion as the actual value.

          This is illustrated in Fig. 28 where the reciprocal of the
free ferric ion concentration is plotted against the reciprocal oxida-
tion rate; this is the same type of plot as Fig. 21.  Here again an
excellent correlation of experimental data is obtained.  However, this


                                  69

-------
    60 r
    50
    40
~.  30
    20
     10
     0
                                        i           i
      0
20
40
60
                                       Te
            Fig. 28 - Reciprocal Rate vs Reciprocal of the  Square Root  of
                     Free Ferric Ion Concentration
                                       70

-------
does not help to identify the active oxidizing ionic species because
other ions would undergo a similar change in concentration with change
in total iron concentration.

          A series of ferric ion oxidation runs was made by Sasmojo18
at pH = ^.0 in a solution saturated with ferric ions.  The EMF was
varied from 0.500 to 0.650.  As may be expected from the kinetic model
previously developed, the oxidation rate was much lower than rates for
runs at lower pH and an iron concentration of 0.01 moles per liter.
The obvious reason is the low ferric ion concentration that can exist
in a solution of such a high pH.  Because of the low iron concentration,
it was difficult, experimentally, to maintain a reasonably constant EMF,
but in runs where well over 99-9^  the iron was ferric, the maximum
oxidation rate observed was 0.59 M-g mole pyrite oxidized per hour per
gram pyrite.  This is less than jfo of the maximum rate in sulfate
solutions of 0.01 molar iron solutions.  This rate is also approximately
    of the oxygenation rate.
          As the pH increases, the ferric ion oxidation rate becomes
limited by the (free) ferric ion concentration in solution.  For a
solution of pH = 3-0 and above, it is highly unlikely that the
microbial- enhanced reaction rate could be greater than the oxygenation
rate simply because there can not be enough ferric ions in solution at
high ferric-ferrous ratios to produce that significant an oxidation
rate.
                                  71

-------
                              APPENDIX I

                            RELATED STUDIES

     Specific research projects were undertaken on subjects directly
related to the sulfide-to-sulfate reaction.  In a few cases, no positive
contribution to the overall project was obtained.  These studies are
reported to describe the difficulties encountered and, hopefully, to
assist future researchers in avoiding the same pitfalls and "blind
alleys."

Determination of Ferric and Ferrous Ion Adsorption

     Kinetic data strongly indicate that competitive adsorption of
ferrous and ferric ions is a factor determining the rate of ferric ion
oxidation of pyrite.  An attempt was made to measure directly the
adsorption equilibrium constants for these ions to compare with the
corresponding constants obtained from kinetic data using Eq. (3)-

     Direct Measurement of Adsorption

          The attempt to determine the adsorption isotherm was first
made by measuring the amount of ferrous and ferric ions adsorbed on
pyrite surfaces.  The experiment was conducted by equilibrating a
ferric or ferrous solution with the pyrite particles.  The amount of
iron adsorbed was to be determined by the difference between iron con-
centrations before and after the solution was equilibrated with the
particles.

          This method was not successful because (a) the amount of
ferric or ferrous ions adsorbed appeared to be very small, so that the
change in iron concentrations before and after the solution was con-
tacted with the pyrite particles did not show any change in iron con-
centration; and (b) complications arose because of reaction of the
ferric ion with pyrite.  Furthermore, there was always pyrite itself,
from which iron could be desorbed upon contacting the pyrite particles
with a solution.

          Because of the difficulties encountered, the direct experi-
mental approach to determining the adsorption of ferric and ferrous ion
was discontinued.

     Adsorption from Streaming Current Measurement

          The adsorption of ions from a solution on solid surfaces in
contact with the solution will result in the formation of an electrical
double layer at the solid-liquid interface.  The formation of, such a
double layer causes several phenomena to occur when there is a relative
motion between the solid surface and the solution.  Such phenomena are
known as electrokinetic phenomena.  One is the development of a streamin
                                  73

-------
current when a liquid is flowing past solid surfaces with large surface
areas .

          Based on the consideration that adsorbed ionic species could
cause a streaming current, a method was developed to determine the
adsorption of ferric and ferrous ions from measurements of streaming
current .

          The electrical double layer is an array of charged particles
which exist at every interface.  This array may consist of a layer of
adsorbed ions, known as the compact layer, followed by a diffuse layer
consisting of an ionic atmosphere, in which ions of one sign are in
excess relative to those of the opposite sign.  This excess charge will
balance the electrical charges of the adsorbed ions.  Therefore there
is an unsymmetric distribution of charges in the neighborhood of the
solid- liquid interface.  Detailed discussion of the double layer theory
can be found elsewhere16 and will not be presented here.

          If we let a solution in contact with solid particles flow
while maintaining the solid particles at their fixed position, the dif-
fuse layer will be carried away from the particles' surface.  Since the
diffuse layer contains excess charged particles of one sign, the flow
of the solution causes transport of electrical charges; i.e., a current
flow is developed.  This transport of electrical charges due to the flow
of a solution past solid surfaces in known as streaming current.  If we
can measure this streaming current, we would know the amount of charged
particles in the diffuse layer that are transported by the flowing
liquid.  Since the charges in the diffuse layer are balanced by the
charges of the adsorbed ions, the amount of the adsorbed ions can be
determined if we know the valence of the adsorbed ions, and if there
is only one type of adsorbed ion.

     Equation for Streaming Current

          The streaming current is the number of electrical charges of
charged particles in an electrolyte solution transferred per unit of
time attributable to the flow of the solution past solid surfaces.  If
we have a packed bed of solid particles, the streaming current can be
expressed by the equation,
where

               = streaming current (A) ,
                 charge density in the diffuse double layer
                 (C/cm2 of interfacial area) ,

-------
            a^_ = interfacial area per unit volume of the packed
                 bed (cm2/cm3), and

             q. = volumetric flow rate of the liquid (cm3/sec).


          The charge density of the compact layer, TJC, is given by


                               nc = -nd                            (8)

since the compact layer and the diffuse layer have charges of opposite
sign.

          In order to obtain the value of T]C, an expression for q. should
be known.  For flow through a packed bed, we can use the Blake-Kozeny
equation;

                               6
                            -
                          1-25
where
          AP = pressure drop across the packed bed (dyn/cm2),

           $ = void fraction of the packed bed,

           A = superficial cross-sectional area of the packed
               bed (cm2),

           (j. = viscosity of the solution (poise), and

           L = length of the packed bed (cm) .


          Combining Eqs. (7), (8) and (9), we obtain
All the quantities on the right hand side of Eq. (10)  can be determined
experimentally, and therefore ru can be calculated.,
                               L.

-------
     Experimental

          The equipment -used to measure the streaming current consisted
of a packed bed of pyrite particles, assembled in a constant head arrange-
ment, so that liquid flow through the packed bed could be maintained
constant.  A stop-clock was provided in the flow line to regulate the
liquid flow rate.  A rotameter was used to measure the flow rate.  The
pressure drop across the packed bed was obtained from a prepared cali-
bration curve, relating pressure drop to flow rate.

          To measure the streaming current, a platinum electrode was
provided at each end of the packed bed.  Each electrode was constructed
from platinum gauze welded to a platinum wire, which functioned as a
frame and leads for external connections.  Outside the packed bed the
platinum electrodes were connected by a resistor having resistance much
less than the resistance of the solution.  The ratio of the solution
resistance to that of the resistor was 1000:1.  With this arrangement,
most of the streaming current developed would be conducted through the
external resistor.  By using a resistor of known resistance, the stream-
ing current can be determined from the potential drop across the resis-
tor.

          Preliminary measurement of the streaming current showed that
its value was of the order of 10"6 A.  A consistent and reproducible
result, however, could not be obtained  For an unknown reason, erratic
results were always obtained.  Since the case of the erratic behavior
could not be found, the experiment was discontinued.

     Conclusions

          It is very possible that an extremely small number of adsorp-
tion sites make a direct determination of adsorption equilibrium con-
stants difficult from an experimental point of view.  Measurement of
adsorption isotherms for liquid phase systems is difficult under the
best conditions.  Pyrite, with its low surface area, is an extremely
difficult material to work with.  We do not believe that the failure to
experimentally observe adsorption of ferric or ferrous ions should in
itself rise doubts as to the adsorption mechanism proposed to explain
the kinetics of pyrite oxidation by ferric ions.

Application of the Potentiostat to Pyrite Oxidation Studies

     A potentiostat was used as a tool to study the sulfide-to-sulfate
reaction.  We hoped that it would prove to be an effective instrument
that could quickly and accurately determine the oxidation rate of pyrite,
and aid in finding the reaction mechanism.

     Unfortunately, no relationship between current flow, pyrite-
electrode potential, and conventional oxidation rates could be found.
Only at, or approaching, equilibrium conditions were results reproducible
and capable of rationalization.

                                   76

-------
     A crystal of museum pyrite was the anode and the potential between
the anode  (pyrite)  and reference cell ( saturated KCl) was controlled by
the potentiostat.   The anodic current was determined as a function of
the pyrite-to-reference cell voltage for different solution concentra-
tions; results are  plotted in Fig. 29.  When the potentiometer setting
of the potentiostat was plotted as a function of pyrite-to-reference
cell voltage the data shown in Fig. 30 were obtained.  Note that  the
voltage at which current flow goes to zero is the same in each case, as
would be expected,  but the extrapolation to the potential corresponding
to zero current is  much easier using Fig. 30.

     A detailed discussion of the equipment and procedure used in this
study is given by Wendschuh.26

Inhibition Studies

     An effort was  made to find a chemical inhibitor for  the pyrite
oxidation reaction.  Physical methods for preventing the  reaction, such
as impervious coatings, elimination of oxygen or water, etc., are
dependent on the physical properties of the pyritic system itself and
must therefore be examined on a pilot or field scale.

     For the chemical oxygenation reaction, several ions  are known to
decrease oxidation  rate; e.g., phosphate, thiocyanate, phenathroline,
and bipyridine.  The  plan was to evaluate these and other materials in
an attempt to resolve the mechanism by which inhibition occurs.  Hope-
fully, it would then  be possible to specify a material which could be
added to the pyritic  system as a vapor, thereby reaching  the majority
of the reactive areas in an underground mine.

     Other ions, as listed in Table XII, were selected on the basis of
their adsorption characteristics, iron complexing ability, and oxidation-
reduction values in solution.
                      Table XII.  Chemical Inhibition Tests
                       Ion Studied	Effect on Rate

                      Fluoride           increase
                      Qxalate
                      Chloride
                      Sulfate
                      Thiocyanate         50$ decrease
                      Phenanthroline       50$ decrease
                      Bipyridine          50$ decrease
                      Hydroxylamine        90$ decrease
                      Nitrate
                      Iodide
                      Borate
                      Chlorate           increase
                      Phosphate           80$ decrease
                                    77

-------
-

o
(/>
 
;.! s
*  -2
e^tf
Q. C -
i- 6
0  Q.
P P CL
N- ; f^-

u> in o>
2 3 3
Q. Q. Q.
cfcf c?c?
^^ ^^ V^ ^^
to co to to
< I^X^^x^
2| 3 Zl Zl
R
Q oon<




o
<
<
OR <
<> D ^0 <0 rt
i i ^xx PI ip^
fn
&
O
CO
 . t>

-------
UJ
or
UJ
u_
UJ
o:

o
    07r
    0.6
    0.5
S  0.4

UJ
o
    0.3
UJ
K

a:  0.2

Q.
    0 I
     0
                        0
                                   o
                            o    o    o
                                                                 a
                                                                 o
                                                     a
                                                  0  o
                                                  A
                                                            a
                                       A
                                                       a
                                                   A
                     D
              a
                      A
A
                                           DAD
                                                 o

                                           o
                                      O
          O
         O
                                O
       O   IN. H2S04

       0   2.N H2S04 plus  I70O ppm Fe ions

       D   IjN H2S04 plus 170ppm Fe ions

       A   |_N H2S04 plus 17 ppm  Fe ions
       0.5                  0.25                 0.45


                         POTENTIOMETER  SETTING (V)


           Fig. 30 - Pyrite to Reference Voltage vs Potentiometer Setting
                                                      0.65
                                   79

-------
     It should be noted that all of the above ions had to be present in
concentrations greater than 250 ppm before any appreciable effect on
rate was observed.  Eate was determined by charge in iron concentrations
of solution, not oxygen absorption.  In several of the organic ions,
oxidation of these ions consumed oxygen, and in the case of hydroxylamine
hydrochloride, probably reduced the dissolved oxygen concentration to
produce the low oxidation rate.

     Since oxygenation rates are independent of iron and sulfate concen-
trations, it was not surprising to observe that there is no correlation
between complexing strength for ferrous or ferric ions and change in
oxidation rate.

     No quantitative measure of adsorption was possible so that relative
adsorption of the inhibitory ions is not known.  However, based on
structural considerations, relative adsorption does not appear to be a
logical explanation for the observed data.

Adsorption of Oxygen, Nitrogen, and Water on lyrite

     The gas phase adsorption isotherms reported in this section were
determined for two reasons:  first, to evaluate that physical property
of pyrite (or pyritic material) that determines adsorption capacity of
the different gases; and second, to try to find an independent method
of evaluating adsorption equilibrium constants previously determined
from Eq. (l) or (2).  Unfortunately the total adsorption of oxygen in
relation to that adsorbed by the reaction was too small to obtain a
reliable value as a difference.

     Adsorption isotherms for oxygen, nitrogen and water were determined
for both Museum Grade and Sulfur Ball pyrite at 25 C.  Curves and data
points are given in Figs. 31-37-  Surface area was determined for each
material by nitrogen adsorption at 77.3K.  The 100 - oo mesh Museum
Grade sample had a surface area of 0.70 m2/g, and the 70-100 mesh Sulfur
Ball material had an area of 0.82 m2/g as determined from a B.E0T. plot
of nitrogen adsorption data.

     The nitrogen adsorption curves at 77-3K (Figs. 31 and 32) are
normal for materials of low surface area.

     The difference in the Museum Grade and Sulfur Ball samples is
apparent in the nitrogen adsorption curves at 25C.  The adsorption of
nitrogen on the Sulfur Ball is almost a hundred times greater than for
Museum Grade pyrite.  This is no doubt due in part to adsorption on
the small quantity of  carbonaceous and siliceous material included in
the Sulfur Ball agglomerates.  However, it is equally apparent, when
considering the difference in reactivity of Sulfur Ball and Museum
Grade pyrite, that a significant fraction of this increased adsorption
is due to pyrite itself.  The small-grained agglomerates composing
Sulfur Ball would present many more adsorption sites per unit surface


                                   80

-------
                                                                                                             o
                                                                                                             o
                                                                                                             ro
                                                        *JT .	s

                                                         >
                                                            o
                                            CO H ^ O  C-M3 CO
                                            oo ir\ iTv t-co H oo
                                            H H H H  H CO OJ

                                            O O O O  O O O
                                                                             ON C-  t- H
                                                                 oo ir\ t- ON oo H co
                                                                             H   -

                                                                                                                    -    8
                                                                                                                       
                                                                                                                        w
                                                                                                                        o ro
                                                                                                                        M   .
                                                                                                                         I

                                                                                                                        rH
                                                                                                                        ro
                                                                                                                        M
                                                                                                                       H
                                                    I	I
in
OJ
O
CVJ
(>

O
O

O
lO
o
                                                           81

-------
                                                   PL.
                                                         lf\ [ CO ONi-1 CM J" \O CO
                                                         HHHHOJWCMOJW
                                                                     o
                                                      a
                                                      o
                                                      H
                                                      O
                                                      10
                                                              B
                                                             H
                                                             -P

                                                              i
                                                              a
                                                              a;
                                                  '-^    CO O\cO oovD H c\l O tf>
                                                   M    vDOvOJ-rOOD-HOJ
                                                                  ON O I> 00
                                                                  cxD CM irvH
                                                                     H H C\J (M CO
                                    J	I
m
CM
ci
o
CM
ci
I    I
in

CD
I    I
    o

    o
I    I    I
    m
    o
    d
                                                                                                       j	L
                                                                                                                  o
                                                                                                                  o
                                                                                                                  ro
                                                                                                      oo
                                                                                                      t^-
                                                                                                      t-
                                                                                                                               0) CO
                                                                                                Q-   ^30
                                                                                                      -p i/^\
                                                                                                      o  .
                                                                                                      M ir\
                                                                                                      M CM



                                                                                                      II
                                                                                                      -P -rl


                                                                                                      O
                                                                                                      ra  
-------
O CT\_d- ONvD VD CTN H
vDcoTiSoOCOvDcooO
                                                                           ft
                                                                           ra
                                                                           Q)
                                                                           -p
                                                                           H
                                                                           PH



                                                                           "<1J


                                                                           |


                                                                           O
                                                                           Ji
                                                                           8
                                                                           w
                                                                           S
                                                                           H
                                                                           o o
                                                                           w   
                                                                           Td ITS
                                                                           < CO
                                                                           I
                                                                           ro
                                                                           oo
                                 83

-------
 CM
 CO
 0
                       COCO C\J LT\ (
                          8O H CM.
                          O O O c
                                                                                     OCTN
                                                                                     C\J CO
                                                                                  vo oo vo
                                                                                  O O  O
                                                                                              -
                                                                                               OOOOOO
                                                                                                                 . .
                                                                                                                 d d
                                                                                                               O
                                                                                                               O
                                                                                                               00
                                                                   OOOOOOOO    OOOOOO

                                                                                            Q
                                      o

                                      H

                                      B
                                      o
                                      -3
                                      <
                                                                                            f
                                                                   (M
                                                                   co
                                                                   U>O LT\ ^O CO J-  IA J-"    CO CO fO OO CO . .
                                                                      i-IMOAlJ-vocOvO    J-CvlJ-COHi-l
                                                                                                                 O\LT\
                                                                                                                 '^) co
 I   \A

s
                                                                                                                                         O
                                                                                                                                         O
                                                                                                                                         O
                                                                                                                                         O
                                                                                                                                         CM
q
d
ID
O
d
P
d
q
d
Q
d
                                                      OJ
                                                      o
q
d
                                                                                                                                                      u
                                                                                                                                                      b
                                                                                                                                                       id

                                                                                                                                                      JS
                                                                                                                                                       a
                                                                                                                                                       PQ
                                                                                                                                                       I

-------
 1.5
 1.0
0.5
  0
                   O.Sh
                   0.85
                   3.08
                     33
                     61
                     21
                     10
 8.M
13.79
17.37
20.72
20.2k
22.29
                  (mmHg)    [cc/g(STP)]
0.102
0.105
0.205
0.216
0.297
0.317
O.IH9
Q.k20
0.612
0.801*
1.088
1.089
1.378
                                                                            E
                                                                            6
                                                                           rd
                                                                           C\J
                                                                             CJ
                                                                           oT
   0
         8
        12
16
20
24
28
     Fig. 35 - H20-Adsorption Isotherm on "Museum Grade"  Pyrite  100-00 Mesh
               at 25.0C
                                      85

-------
 t- c- <
 t- H ^
i ONCO U"N ONOO .
i OJ i  "  ~
CO ONOO


      d
                                            VD
                                                             CO OJ ITv
                                                             H ONCU
                                                             O ON ON
                                                             rH
   g
  H
                                          g
                                          H
                                                           o
                                                           CO
bO
     O HCO  tJ- ir\rot>-OMr\cn^Dj-
     ro oo D~  H ro ONCO ir\ o -* O W H
                                                       t O\ 1TN
                                                       LTNOO CO
     O O O  H OJ OOVO H  ir\OO H W OO
                         H  H H PJ OJ (M
                                H
                                OJ
                                           
-------
             ID
             ro
                             OJ
               oo
               OJ
i	r
T	r
          SCOOOJt^J-J-OOr-iCr-
          CMHCOWCTMr\_d-J-CO(MOJ CM C\J HHH

        r-1 ^ ON j-' ojoddododdddod
         -
                                         O O t

                                         COC-^
                                    c\j cvi CM c\j oj c\j
              OJ
                             O
              CO             CO

                    31VH
                                                                                          OJ
O
o
OJ
                                                                                                           o
                                                                                                           o
                                                                                                       o-
                                                                                                           O
                                                                                                           in
                                                                                                                      o

                                                                                                                      o
                                                                                                                         o
                                                                                                                        -P
                                                         87

-------
area than Muse-urn Grade pyrite.  Also note the hysteresis  loop in Fig.  3^>
which could account for the nitrogen "contamination" reported on page  22.

     Water adsorption isotherms are presented in Figs.  35 and 36.  The
adsorption isotherm for Sulfur Ball appear  to have a hysteresis loop,
but examination of desorption and re-adsorption data indicate that the
initial adsorption cycle probably opened new capillaries  and irreversibly
increased adsorption capacity.

     The oxygen adsorption data are particularily interesting since they
can be interpreted in several ways.  Physical adsorption  is extremely
rapid, and if it does occur it takes place in a few seconds (assuming
no mass transfer limitations).  The relatively slow uptake of oxygen
shown in Fig. 37 and Table XIII could be attributed to  chemisorption
and/or chemical reaction.

     Considering both these adsorption data and reaction  kinetics of
the oxygenation reaction when water is present, an adsorption-reaction
mechanism can be described which explains the experimental observations
below.

     Assume that oxygen is first physically adsorbed; if  adsorbed on a
dual site the adsorbed oxygen dissociates and reacts with pyrite.  When
water is absent, oxidation products are not removed from  reactive sites
and the reaction rate decreases rapidly.  When water is present, oxidation
products are removed from reactive sites by dissolving  in the surround-
ing aqueous phase.

     The assumption that the initial adsorption of oxygen is physical
can explain why nitrogen affects oxidation rate by competitive adsorp-
tion on reactive sites.  Data of Table XIII show that a small but
significant amount of oxygen was removed during the desorption cycle.
This oxygen was physically rather than chemically adsorbed.

-------
            Table XIII.  Oxygen Adsorption, Absorption,  and
                         Desorption on Sulfur Ball
Adsorption
Desorption
Equilibration
Time (hr)
0.5
1.0
2.0
3.0
19.0
20.0
0.5
3-0
0.5
2.0
0.5
1.0
21.0
45.0
70.0
17.0
24.0
45.0
0.1
20.0
78.0
Equilibrium
Pressure (mmHg)
44.04
43-93
43.81
43.69
43.18
43.18
80.93
80.82
120.63
120.55
221.79
221.68
217.39
214.46
211.83
694.65
691.91
691. 70
184.30
184.09
183.47
VA/W
(cc/g)
0.00067
0.00117
0.00173
0.00228
o.oo462
o.oo462
0.00509
0.00537
0.00562
0.00575
0.00588
0.00591
0.00979
0.01240
0.01474
0.01782
0.01873
0.02143
0.02036
0.02132
0.02417
                                  89

-------
                              APPENDIX II

      CALCULATION OF EXPERIMENTAL DATA FOR FERRIC IRON OXIDATION


     Derivation of Eqs.  (ll)-(l6) may be found in the dissertation of
S. Sasmajo.18  The notation below is used.

     r  = rate of pyrite oxidation (g-mole/hr-g pyrite)

     v  = volumetric rate of titration (^/hr)

     C3 = concentration of ferric ions (g-mole/^)

     C2 = concentration of ferrous ions (g-mole/^)

     Cfc = concentration of total iron (g-mole/e)

     R  = ferric-ferrous ratio

     R  = average ferric-ferrous ratio between the inlet and
           outlet of the packed bed

     N  = normality of KMn04 (g equivalent/,0 ) or concentration
          of ferric ion of titrant (g-mole/,0)

     q  = volumetric flow rate of the reaction solution through
          the reactor (,0/hr)

     t  = time (hr)

     V  = total volume of the reaction solution

     m  = mass of pyrite sample (g)

     r* = rm

     Conditions at time t = 0 will be denoted by a subscript 0.   Condi-
tions at the inlet of the pyrite-packed bed will be denoted by a super-
script i and at the outlet of the packed bed by a superscript 0.

Reaction Rate ("r") with Potassium Permanganate as Titrant
                           r =                                   (11)
                               (15R
                                  91

-------
Normality of KMn04 Solution to Maintain Constant
Total Iron Concentration
                             (15R
N =
                                   R + 1

Reaction Rate with Ferric Ion Solution as Titrant
                                                                 (12)
                           r =
                                     Nv
                               (15R
Normality of Ferric Ion Solution to Maintain Constant

Total Iron Concentration
Calculation of Average Ferric/Ferrous Ratio

     For large changes in R:
            R =
                          111
                          15   (15R1 + Ik)     (Rl+ 1)
- 0/n-
                                        (13)
                                                                 (1*0
                                        (15)
where
                     Ru =
                          (CitqRi)/(Ri+D -
                     r
-------
                             APPENDIX III

                   CALCULATION OF IONIC COMPOSITION
     The method of calculation to determine the concentration of free
Fe+++ ions and free Fe++ ions and their complexes,  as well as major
anions that are present in the solution, are described in this appendix.
Separate descriptions for sulfate solution and for  chloride solution
are given below.
                           Sulfate Solution

     In our sulfate solution, the following chemical equilibria between
the ionic species were considered:

          1.  Fe+++ + SOI"  = FeSOt

          2.  Fe+++ + HS04  = FeHSOt+
          3.   FeS04 + SO  = Fe(S04)

          k.   HS04          = H+ + SO"

          5.   Fe+++ + H20   = Fe(OH)++ + H+

          6.   Fe++  + SO"  = FeS04

          7.   Na+   + SO"   = NaS04

     If Q1, Q2, Q3, etc.  are the stability constants  for each of the
equilibrium reactions listed above, we have the following equilibrium
conditions:


                  (FeSOt)
               (Fe+++)(S04-)


          Qa =	                                      (l8)
               (Fe+++)(HS04)


          Q3 =   Fe S4 2	                                      (19)
                                                                 (20)
                 (HSO;)
                                  93

-------
          Q5 -
          Q6 =     e                                             (22)
               (Fe++)(SO-4-)
               (Na+)(S04-)


Using Eqs. (l?)-(23), the ionic composition in the solution can be de-
termined.  The calculation procedure is described below.


Determination of H4 Concentration

     The H+ concentration may be determined as the difference between
the total hydrogen added to the solution (from a charge balance)  and
the hydrogen combined as HSO^, or,
where, from charge balance, total hydrogen added (Hj)  equals


               Hj = 2(S04-)T - 3(Fem) -  2(Pei:E)  -  Na1          (25}


Calculation of HS04 Concentration

     The equation to calculate the HSC>4 concentration  is derived below.
To simplify the expressions, the following notations will be  used:

               Fe    = total ferric ion concentration

               Fe    = total ferrous ion concentration

               Fe    = concentration of free ferric  ions

               Fe+   = concentration of free ferrous ions

               L     = concentration of anion; in this case,
                       L is the sulfate ion concentration

               H     = concentration of hydrogen ion

               HL     = concentration of HS04 ions

-------
     The total sulfate ions in the solution is  equal to the sum of free
SO" and all the S04~ ions that are associated.  Therefore,
          (L)total = L + HL + Fe+++L +  2Fe+++L2 + Fe+++ + Fe

                                                                (26)

     A balance for ferric ion concentration gives

          Fe111 = Fe+++ + Fe+++L+ Fe+++L2  + Fe+++HL + Fe+++OH    (27)
Using the equilibrium relations,  Eqs.  (17)- (23), Eqs.  (26) and (27)
can be transformed to
                                    (    giGi(HL)Si) + Fe++ - G5(HL)
                                    \i=l           /
                                                                (28)

and
                    Fe+++ = _    _                     (29)
where

                    G0  =  Q4/H
                    G2  =

                    G3  =

                    G4  =  Q5/H
and
                    G5 =
The values of the g^'s are
                     g2 = 1
                     gs = 2

                     g4 = 0
and                       n
                     g5 = 1
                                 95

-------
Combining Eqs. (28) and (29), we obtain

               (L)  ^   - G Fe111 - G Fe11
          HL = 	total	&	5	                       (30)
where                     ^


                    Ga = J=l	                        (31)
                              k
                           1.   '^ ^  e~i f TJT i Q*l
                              /  \JTn I J~LJ^j j
                             1=1



                                                                 (32)
                         1 +

and
                    Gc =
                             G6(HL)
     From a knowledge of (l) total sulfate [LTJ, (2) total ferric ion
LFe111], (3) total ferrous ion [Fe11], and (4) total sodium [ua1] as
obtained from analysis of solution or from inventory of quantities
added to the solution, the ionic concentration can be calculated using
the following procedure:

          A.  Assume H+ = 1/2 H , HL = 1/2 H , and calculate
              Gi's.            T            T

          B.  Solve for G&, G^, and GC; substitute into Eq. (30)
              and solve for HL.

          C.  Using relation, H1" = H^ - H^, recalculate G^'s
              using new H+ value.

          D.  Again solve for G , G^, and GC using corrected
              values for HL and G^'s; substitute in Eq. (30)
              and solve for new HL.

          E.  Repeat Steps C and D until HL values agree.

          F.  When HL and H+ values are known, ionic composition
              can be calculated from stability constant equations.

-------
Calculation of Fe+++, Fe++ and Their Complexes

     After the HSO^ and H4" concentrations are known, the values of Fe+++,
Fe++ and their complexes can be calculated from Eqs. (l7)-(23).

     In order to perform the calculations as outlined above, values of
stability constants should be known.  Table XIV summarizes the values
of stability constants used in the calculation.

         Table XIV.   Stability Constants  for Sulfate Solution20
Constants
Qi
Q2
Q3
Q4
Q5
Qe
Values of
log (Qi)
2.23
0.78
0.97
-1.99
-2A3
O.OU
Ionic
Strength
1.2
1.2
1.0
1.2
0.0
1.1
     Since the ionic strengths of the solutions used are not the same
as those listed in Table XIV, corrections are needed.   The calculation
method to correct the stability constant is as follows:

     (l)  Calculate the values of the Q's at zero ionic strength from
the values given in Table XIV, using the relation
                             Q1 = Qo,iQf ,i                       (3*0

where

          QJ = stability constant at a given ionic strength
               for equilibrium reaction i,

        Qo i = stability constant at zero ionic strength for
               equilibrium reaction i, and

        CL. . = activity coefficient quotient in terms of activ-
          '    ity coefficients for the equilibrium reaction i.
                                   97

-------
     (2)  To obtain the value of Qf ^, the activity coefficients should
be known.  Values of activity coefficients were estimated from the
empirical relation proposed by Davies4; i.e.,


     - log f = 0.50 Zi2[lV(l+l'/2) - 0.20 I]                     (35)

where

     f^ = activity coefficient of ionic species i

     z^ = valence of the ionic species i, and

     I  = ionic strength of the solution.

     (3)  Calculate the value of 0^ at any ionic strength from Eq.. (3U),
using the value of Qo,i obtained in step 1.


                           Chloride Solution

     In chloride solution, the following chemical equilibria between
the ionic species are considered:

               1.  Fe+++ + Cl"  = FeCl++

               2.  FeCl++ + Cl" = FeCls

               3-  Fed* + Cl"  = FeCl3

               h.  Fe+++ + H20  = FeOH++ + H+

               5.  Fe++ + Cl~   = FeCl+

               6.  FeCl+ + Cl~  = FeCl2

     The equilibrium relations for the above equations are

                             )                                   (36)
                    (FeCl)(Cl-)
                                                                 (37)
                                                                 (38)
                                   98

-------
                                                                 (39)
               Q5 =
                      (FeCl+)
               Qs =
The Q's in the above equations are the stability constants.

     The calculation procedure to determine the  ionic  composition  in
the solution is similar to that in sulfate  solution.   Only a brief out-
line will therefore be given.

Determination of H* Concentration

     The hydrogen ion concentration was calculated  in  the same manner
as in the case of sulfate solution [Eqs.  (2k) and (25)].

Calculation of Cl~ Concentration

     Following a similar method as that used in  calculating HS04,  and
also using the same notations, the following equations can be derived:
                            D  iGi
                          Li=l
(L)
                                       gi
+ Fe4
                     Fe
                                 Fe
                                    III
                             1 +
                                 1=1
r6
  V   r
  L 1%
L 1=5
           ,    and
                            1 +
                                FeIII
                                                                (kk)
                                 99

-------
and
where

                    L = chloride ion concentration;

                   G! = Q15g! = 1

                   G2 = Q,iQe>g2 = 2

                   G3 = Qi^Qajgs = 3

                   G4 = Q4/H,g4 = 0

                   G5 = Q5,g5 = 1

                   G6 = QsQsjgs = 2  .


Combining,

                              1+  E Gi(L)gl
                                   6
                              1 +  E Gi(L)gi
                                  i=5
     Iterative calculation similar to that used to calculate the HSC>4
concentration will give the Cl~ concentration.  Note that in Eqs.  (U2)-
     L is used to represent Cl~ concentration.
Calculation of Fe+++, Fe"1"1", and their Complexes

     The calculation for iron ions and their complexes in chloride
solution is exactly the. same as that for sulfate solution.

     The stability constants for the complexes in chloride  solution
are given in Table XV.  Corrections for stability constants due to
                                 100

-------
ionic strength were made in exactly the same manner as in sulfate solu-
tion.
         Table XV.   Stability Constants in Chloride Solution20


            ,  ,                Values of                Ionic
       Constants	log  (Qj)	Strength

          Q1                      1.48                    0.0

          Q2                      0.65                    0.0

          Q3                     -1.0                     0.0

          Q.4                     -2.^3                    0.0

          Q5                      0.36                    2.0

          0,6                      0.0k                    2.0
                                   101

-------
                              APPENDIX IV

                             TABLE HEADINGS
RUN          -  run number
PH           -  solution pH
EMF          -  measured EMF of solution (volts)
EMF CORK     -  corrected EMF (volts)
IRON CONG.   -  total iron concentration x 102; multiply tabulated values
                by 10~2 to obtain iron concentration in g-moles/
GRAM SAMPLE  -  weight of sample (grams)
FLOW RATE    -  flow rate of recirculated solution (,0/hr.)
RATE         -  oxidation rate; multiply tabulated values by 10~6 to
                obtain reaction rate in g-mole pyrite oxidized per hour
                per gram of pyrite sample (g-mole/hr, gm)
                                  103

-------
RUN DATA, FERRIC ION OXIDATION
         Table XVI
                 (*)
RUM

F018 S
F013 -&
F020 3
F021 --3--
F022 B
FO23 B
F024 B
f051
F051 S
RR-
RR1
4*32
STD1
STO2
STD3
F05S
F057
F034 a
F097 8
FOS3 S-
F0100B
-F0131
F0102
FO4B3
F0104
-G 1 -0- 5 -
F0106
F025
T02S
F027
FQ2&
F029
^03G
F031
F032 	
F052
Jta&2^
PH -

D.sa
-C.-50
D.50
Q.5G-
C.50
C.50
C.50
\j  StJ
0.50
-0,50- -
G.5G
C.5G
0.50
C.50----
0.50
0.50 -
0.50
C.50
C.50
G.50
O.SQ
--G-.5G- -
G.50
G.5G--
0.50
-6.50
0.50
1.00
--4.-00-
1.00
-1^00 -
1.00
LOG
1.00
-1-.OG--
1.00
-1 . OO-
EMF-- -

0. 450
044G
C. 425
0. 4G^
0. 375
0.460
G.470
0.450
0. 450
0.4S&
0. 450
CU-4S-0--
0.450
S.-45C
C. 450
0.400
0. 425
0, 6G G
0. 450
0.550
0. 575
0. 75-0
0. 700
G, &S-3
0.475
0, 5GG
0. 525
0. 361
&. -34 5
0.409
0.424
0. 435
D. 443
0.453
0.45-8--
0.458
O,i.^
EMF
CORR.
0.450
0.43S
0.425
0.393
G.374
0,453
0.469
0.445
0.445
G448
0.448
0.448
0.442
0.444
0.451
0.392
0.422
G.5^37
0.445
&.S49
0.573
0.669
0.656
0.&3-3
0.475
G.500
0.525
0.349
O.-J-90
0.410
0.425
0.433
0.^446
0.459
^.460-
0.463
0.456
IRON
CONC.
0.991
1.013
1.013
1^091
1.C45
1.000
O.S87
0.971
1.041
1.041
1.023
1.023
1.068
1.068
1.091
1.060
1.045
1.182
0.983
0.367
G.999
0.3S3
0.972
0.956
0.983
0.959
0.975
1.016
1.091
1.038
1.038
1.038
0.978
1.061
1.056
0.952
1.QO5
GRAM
SAMPLE
17.30
17.30
17.30
17.3G
17. 3G
17.30
17.30
17.30
17.30
17.30
17.30
17.30
17.30
17.30
17.30
17.30
17.30
4.50
4.50
4.50
4.50
4.50
4.50
4 . 5O
4.50
4. SO
4.50
17.30
17.30
17.30
17.30
17.30
17.30
17.30
17.30
17.30
17.30
FLOW
RATE
41.4
41.4
41.4
41.4
41.4
41.4
41.4
41.4
33.6
4i.4
41.4
41.4
41.4
41.4
41.4
41.4
41.4
52.3
52.3
52.3
52.3
50.3
50.3
51.7
52.3
52.3
52.3
41.4
43.9
43.3
43.9
43.9
43.9
43.9
43.9
41.4
39.0
RATE

1.10
0.63
0.52
0.32
0.16
1,43
2.06
1.13
1.12
1.23
1 .12
1.19
1.31
0.65
1.12
0.26
0.42
7.07
0.8G
5.73
6.52
8.35
7.74
7. 60
1.37
2.84
4.90
0.09
0.27
0.41
0.70
0.92
1.19
1.66
1.85
1.53
2.03
TION
           THE &E6INING OF THIS APPENDIX

-------
                       Table XVI
(*) -
Continued
RUN
 	 	
J533 _
F034
F.03S
F03S
FQ37 .
F03S
-Fa33
FQ40
--F-OJU,
F042
~F-a3~
F058
FO-S-9
F060
-FC~
F0125
 F-O 1 2 sa -
F0126
-FO 1-2-7
F0128
-F4r29--.
F013C
-E01-3-1-- -
F0132
^F~O4-J-3--
F0134
F0135
F043
FQ44
-F-34-&- 	
F048
F-O4-7 -
F048
-F049
F050
F&&4
PH

--2.-D0-
2.00
- 2.00-
2.0Q
--2--CQ
2.0C
- 2-* 00-
2. DC
2.00-
2.00
-2 . DC
2.00
2*OO
2.00
2.OO
2.00
-2.-G O
2.00
--2-,-DQ-
2.00
-2.00
2.00
2, CO
2.00
2*00
2.00
2.0Q
2 .50
2.EO
2. SO
2. 50
--2* &G
2.50
-2.50
2.50
-2-. 50
EMF
- 	 	 	 	
 0*3
-------
                        Table XVII
                                 (*)
             RATE  DATAt  SULFATE  SOLUTION
             -WITHOUT  ADDITION OF  SODIUM SULFATE
             IRON  CONC.       VARIES
RUN
	
F0111
f ^H 2
F0113
F0114
F0115
F^m
F0117
F011S
FQ1G7
Foioe
F0139
F011C
F0119
FQ12G
F0121
-F0122
F0123
~f 0-1-2*
PH
	
C.50
---0.50
C.50
- ~Q . 50
D.EO
e.so
G.5Q
0 . 50
0 . 50
0.50
--G.50
O.EO
2.00
2.00
2.00
- 2 . 00
2.00
2.00
EMF
	
0.800
-- 0.800
Q.80D
- 0. 6GO
o.eoo
- -G.8GO
0.8CC
-- D.&QO
0.750
0. 750
0.750
C.75G
C.SOO
-- &.750
0. 750
- 0, 7&G
0.750
G.75G
MF
C OR R .
0.705
G-.699
O.S95
0.690
0.585
0-.S83-
0.576
O.670 -
0.672
O.S65
0.661
0.556
0.698
0.595
0.575
0.583
0.596
0,606
IRON
CONC.
4.471
2.866
1.99<*
1.300
0.897
0.508
O.<*02
0.249
0.956
0.530
0.405
0.258
3.266
0.110
0.040
0.053
0.090
0.112
GRAM
SAMPLE
4.50
4.50
4.50
4.50
t.5D
4. SO
4.50
4.50
4.50
4.50
4.50
4.50
4.50
4.50
4.50
4.50
4.50
4.50
FLOW
RATE
52.3
52.3
52.3
52.3
52.3
52. 3
52.3
52.3
52.3
52.3
52.3
52.3
47.2
46.0
46.0
46.0
44.7
45.3
RATE

12.44
10.32
8.50
7. 03
5.94
4.97
3.35
3.16
7.26
5.59
5.00
4.04
11.18
18.96
15.63
14. S3
14.41
12.40
    SE- EXPLANATION AT  THE 8E6ININ&OF THIS  APPENDIX
                       Table XVIII(*)
              RATE  DATA t SULFATE  SOLUTION
              SODIUM SULFATE   0.20 GM-MOLES/L
              IROR CONC. 	   0.01 GK-MOLES/L
-Run

F061
F-Q&2
F063
-F044
-C06S
F056
_F 0X
F063
	 PH _

C.EO
	 C.5Q-
0.50
	 0*50
	 2 .-00
2.00
	 -2 .-00-
2.00
-_J1E ..

0. 400
0.425
0.45C
0. 4-4-5
-.0.-4CS--
0.350
^0. Z7 5 _
0.418
--EMF -
CORR.
0.402
0.4.30
0.456
{U444
0.436
0.381
0.411 __.
0.451
IRON
CONC.
1.051
- 1.O66
1.084
0.997
1.016
1.084
-4.124
1.057
GRAM
SAMPLE
17.30
17.30
17.30
17.30
- 17.30
17.30
- 17.30
17.30
J-LOW
RATE
43.9
43.9
42.6
43.9
41.5
41.4
41.4
41. 4
RATE

0.08
0.31
0.72
0.58
0.77
0.13
0.35
1.29
(*)  SEE EXPLANATION AT
THE BEGINING  OF THIS APPENDIX
     - 106

-------
                         Table XIX
                                  (*)
RATE DATA* CHLORIDE SOLUTION
IRON CONC. 0.01 GM-MOLES/L
RUN

-FG7S-
F077
F073
F079
F^Q SG
F031
F032
F089
F07Q
FQ71
F072
F-073
F074
F075
PH
-- -
0.50
R 0.50
 -5Q
C.50
	 G5G-
0.50
-0 . 5G
0.50
2.0G
- -2. GO
2. '00
2 . GG-
2.00
2,00
EMF
--
-G, 4GG
0.425
D.47-G
0. 440
G. 45G
D.46C
0.475
0. 450
0. 425
0.44C
0. <5C
G, 47 5
0. ^6C
3,470
EMF
CO^R.
G.393
O.l38
0.43
0.4U7
G.452  -
0.459
0,467
0.451
Q.<*28
-0.441
0.444
0.479
0.459
C.4G7
IRON
CONC.
0.3-83
0.983
0.991
0.997
1.013
1.025
1.C16
C.912
1.069
1.072
1.069
1.G78
1.087
1.084
GRAM
SAMPLE
17.30
17.30
17.30
17.30
17.50
17.30
17.30
17.33
17.30
17.30
17.30
17.30
17.30
17.30
FLO*
RATE
41.4
41.4
41.4
41.4
41.4
41.4
41.4
41.4
41.4
41.4
37.8
41.4
41.4
41,4
RATE

0.21
0.36
1.40
0.65
0.75
C.92
1.28
0.71
0.99
1.53
2.00
4.31
2.86
3.53
<*)-S-EE --EXPLANATION AT   THE BEGINING OF THIS APPENDIX
                         Table XX
                                 (*)

RUN
F^9G
F091
FOS2
F093

PH
 -- 0^-50--
0.5G
- .-G-.5C -
0.50
IRON C
EMF
B. 45G
0.450
- O. 4SG
0.450
ONC.
EMF
-CGRR.
-CU45Q
0.450
Q.45Q --
0.450
VARIE!
IRON
CONC.
0.578
0. 344
G.225
0.186

GRAM
SAMPLE
17.30
17.30
17.30
17.30

FLOW
RATE
41.4
41.4
41.4
41.4

RATE
0.81
0.7S
0,58
0.53
(*) SEE  EXPLANATION AT   THE  BEGINING OF THIS  APPENDIX
                                 10?

-------
            APPENDIX V.   Characteristics of Pyrite Samples
                                      Sample Designation
                       Museum
                        Grade
                       Sulfur
                       Ball #1
                    Sulfur
                    Ball #2
                         McDaniel's
                         Sulfur Ball
Mesh Size
Reactivity
.#*
60-100
 0.03
60-100
 1.0
60-100
 2.0
60-100
 3-5
Chemical Analysis
Fe
S
Si02
A1203
Carbonaceous
^3-2
^9-7
6.0
41.65
46.72
2.3
6^5
35.1*
38.5*
6.0*
1.85*
17. 9* (by
dif f . )
  Chemical analysis on a 35-50 mesh sample
**Note: Reactivity given in term of oxygenation rate (as |ag-mole of
        pyrite oxidized per hour per gram of pyrite) in 100$ oxygen
        for the mesh size listed.  Other particle sizes of the same
        type of pyrite were used as noted in the text.

          The reactivities are nominal values.   Reactivities vary by
        as much as 50$ from one reactor loading to the  next, depend-
        ing on how well the "fines," which formed in storage and
        handling, were washed from the charge to the reactor bed.
        Also, method of packing the bed affected reactivity.  When
        interpreting experimental data taken from different series
        of runs comparative rather than absolute rate data must be
        used.  In many cases a reliable basis for comparison is not
        available.
                                 108

-------
                            ACKNOWLEDGEMENTS
     A large number of faculty, staff, and students at The Ohio State
University have contributed to this report.

     The co-supervisor of the project for the last several years and
co-author of this report Dr. Kenesaw S. Shumate, Associate Professor,
Department of Civil Engineering, The Ohio State University, is respon-
sible for the microbiological studies conducted during the course of
this project and the written discussion of this phase of work.

     Dr. Karlis Svanks, Assistant Professor, Department of Chemical
Engineering, developed equipment and procedures for the ferric ion
oxidation studies and made the adsorption isotherm determinations.

     Dr. Ernest Ehlers, Professor, Department of Mineralogy supervised
the mineralogical investigations.

     The many hours of stimulating discussions with Professors
P. R. Dugan and C. I. Randies of the Department of Microbiology; and
Russell Brant, Geologist, The Ohio River Valley Water Sanitation Com-
mission, were very helpful in developing the program and interpreting
results.

     A major contribution was made by the graduate students who worked
on the project.  These students and the title of their theses are
listed on page 113. Birle, Kim, Morth, Konecik, Jutte, Wendschuh, Halko,
and Sasmojo were in the Department of Chemical Engineering.  Bailey was
in the Department of Civil Engineering with Dr. Shumate as advisor.

     The financial support of the Federal Water Pollution Control
Administration, and formerly the Department of Health, Education, and
Welfare, through Research Grant WP-003^-0 is gratefully acknowledged.
                                  109

-------
                               REFERENCES
 1.  Bailey, J. R. , "Biological Oxidation of Pyrite,"  M.  S.  Thesis,  The
       Ohio State University (1968).

 2.  Birle, J. D., "Sulfide to Sulfate Reaction Mechanism in Pyrite
       Materials," M. S.  Thesis, The Ohio State University (1963).

 3.  Braley, S. A., Research Summary Report on Fellowship No.  326B to
       Mellon Institute (1953).

 h.  Davies, C. W., "The Extent of Dissociation of Salts  in Water,
       Part VIII," J. Chem. Soc., (1938), 2093.

 5.  Delchamps, E. W., and Temple, K.  L., "Autotrophic Bacteria and  the
       Formation of Acid in Bituminous Coal Mines," Applied Microbiol.
       1 (1953), 255-

 6.  Dugan, P. R., and Randies, C. I., "The Microbial  Flora of Acid
       Mine Water and Its Relationship to Formation and Removal of Acid,"
       Project A-002-OHIO Completion Report, The Ohio  State University
       Water Resources Center, (Oct.  1968).

 7.  Ehlers, E. G., and Stiles, D. V., "Melanterite-Rozenith Equilibrium,"
       Am.  Mineralogist 50 (1965), 1^57-

 8.  Gale,  N.  L., and Beck, J. V., "Evidence for Calvin Cycle and Hexose
       Monophosphate Pathway in Thiobacillus ferrooxidans,"  J.  Bacteriol.
       9k (1967), 1059-

 9.  Garrels,  R. M., and Thompson, M.  E., "Oxidation of Pyrite by Iron
       Sulfate Solutions," Am. J. Sci., Bradley Volume 258A (i960),  57.

10.  Hougen, 0. A., and Watson, K. M., Chemical Processes Principles,
       Vol. Ill, John Wiley & Sons, Inc., New York (19^7).

11.  Kim, H. W., "Vapor Phase Oxidation of Pyrite," M.  S.  Thesis, The
       Ohio State University (1966).

12.  Konicek,  M. G., "The Biological Oxidation of Pyrite," M.  S.  Thesis,
       The  Ohio State University (1966).

13-  Lister, M. W., and Rivington, D.  E., "Ferric Halide  Complexes,"
       Canad.  J. Chem.  ^3 (1955), 1603.

1^.  Nancolas, G. H., Interactions in Electrolyte Solutions,  Elsevier,
       Amsterdam (196671

15.  Orr, C.,  and Dallavalle, J. M.,  Fine Particle Measurement:  Size,
       Surface, and Pore Volume, MacMillan (1959).


                                   Ill

-------
16.  Parsons, R. ,  Modern Aspects of Electrochemistry,  (J.  0.  M.  Brockris,
       edit or ) ,  Butt erworth ,  London ( 195 ^ ) 

17.  Rabinowitch,  E., and Stockmeyer,  W.  H. ,  "Association of  Ferric
       Ions with Chloride, Bromide and Hydroxyl Ions,"  J.  Am.  Chem.
       Soc. 6k,  (19^2),  335-

18.  Sasmojo, S.,  "Oxidation  Kinetics  of  Pyritic Materials in Aqueous
       Media," Ph.D. Dissertation, The Ohio  State University  (1969).

19.  Schaeffer,  W.  I.; Holbert,  P. E. ; and Umbreit, W.  W. , "Attachment
       of Thiobacillus thiooxidans to  Sulfur Crystals," J. Bacteriol.
       85 (1963),  137-

20.  Si lien, L.  G., and  Mart ell, E., "Stability Constants of  Metal Ion
       Complexes," Special Publication No. 17,  The Chemical Society,
       London (196^).

21.  Silverman,  M.  P., "Mechanism of Bacterial  Pyrite  Oxidation,"  J.
       Bacteriol.  9  (1967),
22.   Silverman, M.  P., and Lundgren,  D.  G. ,  "Studies  on the Chemoauto-
       trophic Iron Bacterium Ferrobacillus  ferrooxidans -  I.   An
       Improved Medium and Harvesting Procedure for Securing High
       Cell Yields," J.  Bacteriol.  77, (1959),  6^2.

23.   Standard Methods for the Examination of Water and Waste Water,
       12th Ed., Am Public Health Assoc. , New York, New York (1965).

24.   Stokes, H. N. , "On Pyrite and Marcasite,"  U.S.G.S.  Bulletin 186,
       (1901).

25.   Sykes, K. W. ,  "The Structure and Reactivity of the Complexes of
       Ferric Ions  with Some Simple Anions," Chem.  Soc.  (London)
       Spec. Publ. , Wo.  1 (195^ ), 6U.

26.   Wendschuh, P.  H. , "The Use of the Potentiostat to Study the Sulfide
       to Sulfate Reaction Mechanism," M. S. Thesis,  The Ohio State
       University (196?).

27.   Whiteker, R. A., and Davidson, W. , "Iron (ill) Sulfate Complex
       Ions: Ion-Exchange and Spectra," J. Am.  Chem.  Soc. _75 (1953),
       3081.
                                  112

-------
                         LIST OF PUBLICATIONS
Master of Science Theses

     Birle, J. D., "Sulfide to Sulfate Reaction Mechanism in Pyritic
       Materials," The Ohio State University (1963).

     Kim, H. W., "Vapor Phase Oxidation of Pyrite," The Ohio State
       University (196U).

     Morth, A. H., "Reaction Mechanism of the Oxidation of Iron Pyrite,"
       The Ohio State University (1965).

     Konicek, M.  G.,  "The  Biological Oxidation of Pyrite," The Ohio
       State University (1966).

     Jutte, B. B. Jr., "Effect of Oxygen and Nitrogen in Aqueous Solu-
       tion on the Kinetics of the Sulfide-to-Sulfate Reaction," The
       Ohio State University (1966).

     Wendschuh, P.  H., "The Use of the Potentiostat to Study the Sul-
       fide to Sulfate Reaction Mechanism," The Ohio State University
       (1967).

     Bailey, J. R.,  "Biological Oxidation of Pyrite," The Ohio State
       University (1968).

     Halko, E., "Aerobic-Anaerobic Oxidation of Pyrites," The Ohio
       State University (1968).

Ph.D. Dissertation

     Sasmojo, S., "Oxidation Kinetics of Pyritic Materials in Aqueous
       Media," The Ohio State University (1969).
Papers
     Ehlers, E.  G.,  and Birle,  J.  D.,  "Electropolishing of Pyrite,"
       Am.  Mineralogist 49 (196U), 800.

     Ehlers, E.  G. ;  Stiles,  D.  V.; and Birle,  J. D.,  "Fossil Bacteria
       on Pyrite,"  Science 1^8  (1965), 1719-

     Schopf, J.  M.;  Ehlers,  E.  G.; Stiles,  D.  V.;  and Birle,  J. D.,
       "Fossil Iron Bacteria Preserved in Pyrite,"  Proc.  Am.  Phil. Soc.
       109 (1965),  288.

     Ehlers, E.  G.,  and Stiles, D. V., "Melanterite-Rozenith Equili-
       brium," Am. Mineralogist 50 (1965),
                                  113

-------
Morth, A. H., and Smith, E.  E.,  "Kinetics of the Sulfide-to-Sulfate
  Reaction," Paper to 151st National Meeting, Am.  Chem.  Soc.,
  Pittsburgh, Pa. (March, 1966).

Smith, E. E., "Engineering Aspects of Acid Mine Drainage,"  Proc.
  Second Annual Symposium, Water Resources Research,  The Ohio
  State University (1966).

Smith, E. E., "Acid Mine Drainage Research at the Ohio State
  University," Proc. 22nd Purdue Industrial Waste Conference
  (May, 1967).

Smith, E. E.; Svanks, K.; and Shumate, K. S., "Sulfide to Sulfate
  Reaction Studies,"  Second Symposium on Coal Mine Drainage
  Research, Mellon Inst., Pittsburgh, Pa. (May, 1968).

Smith, E. E. ; Svanks, K.; and Halko, E.,  "Aerobic-Anaerobic Oxida-
  tion of Pyrite,"  157th National Meeting, Am. Chem.  Soc.,
  Division of Fuel Chemistry, Minneapolis, Minn. (April, 1969).

-------
                           GLOSSARY OF TERMS
EMF - Electromotive Force (voltage) as observed using  a  Platinum-
      Sat-urated Calomel electrode pair for oxidation-reduction
      potential measurements.

Ferric Ion Oxidation - Oxidation of pyrite where immediate  oxidizing
      agent is the ferric ion.  The ultimate electron  acceptor  in
      natural systems is oxygen.

Free Ferric Ion - Non-complexed ferric ion in solution.

Museum Grade Pyrite - Large crystalline mass of pyrite,  usually from
      igneous deposits; the type obtained from mineral supply houses.

Oxygenation - Oxidation of pyrite by direct reaction with oxygen.

Sulfur Ball - Pyritic material found in coal measures; present  in
      finely disseminated particles up to massive agglomerates.
                                    ^      &U. S. GOVERNMENT PRINTING OFFICE : 1970 O - 384-187

-------
BIBLIOGRAPHIC:  The Ohio State University
Research Foundation.
A Study of the Sulphide-to-Sulphate Reaction
Mechanism.  FWPC Publication No. WP-

  A detailed study of the mechanisms and kinetics
of the chemical reactions responsible for acid
mine drainage has been made.  The mineralogical
features of the solid phase reactant (pyrite) th
that determine its reactivity were described.
The rate-limiting reactions and variables affect-
ing the rate of these'reactions were identified.
  It was found that two basic oxidation modes are
important: oxygenation, in which oxygen is the
immediate oxidizing agent; and ferric ion (or
microbiologically catalyzed) oxidation, in which
ferric ions are the oxidants.  From a knowledge
of the dissolved oxygen, ferric/ferrous ratio,
and total iron ion content at the reaction site,
ACCESSION NO:

KEY WORDS:

Mine Drainage
Coal Mine Drainage
Sulfides
Iron Sulfides
Pyrite
Ferrobacillus
Pollution Abatement
Industrial Wastes
Reaction Kinetics
BIBLIOGRAPHIC:  The Ohio State University
Research Foundation.
A Study of the Sulphide-to-Sulphate Reaction
Mechanism.  FWPC Publication No. WP-

  A  detailed study of the mechanisms and kinetics
of the chemical reactions responsible for acid
mine drainage has been made.  The mineralogical
features of the solid phase reactant (pyrite) th
that determine its reactivity were described.
The  rate-limiting reactions and variables affect-
ing  the rate of these'reactions were identified.
  It was found that two basic oxidation modes are
important: oxygenation, in which oxygen is the
immediate oxidizing agent; and ferric ion (or
microbiologically catalyzed) oxidation, in which
ferric ions are the oxidants.  From a knowledge
of the dissolved oxygen, ferric/ferrous ratio,
and  total iron ion content at the reaction site,
ACCESSION NO:

KEY WORDS:

Mine Drainage
Coal Mine Drainage
Sulfides
Iron Sulfides
Pyrite
Ferrobacillus
Pollution Abatement
Industrial Wastes
Reaction Kinetics
BIBLIOGRAPHIC:  The Ohio State University
Research Foundation.
A Study of the Sulphide-to-Sulphate Reaction
Mechanism.  FWPC Publication No. WP-

  A detailed study of the mechanisms and kinetics
of the chemical reactions responsible for acid
mine drainage has been made.  The mineralogical
features of the solid phase reactant (pyrite) th
that determine its reactivity were described.
The rate-limiting reactions and variables affect-
ing the rate of these'reactions were identified.
  It was found that two basic oxidation modes are
important: oxygenation, in which oxygen is the
immediate oxidizing agent; and ferric ion (or
microbiologically catalyzed) oxidation, in which
ferric ions are the oxidants.  From a knowledge
of the dissolved oxygen, ferric/ferrous ratio,
and total iron ion content at the reaction site,
ACCESSION NO:

KEY WORDS:

Mine Drainage
Coal Mine Drainage
Sulfides
Iron Sulfides
Pyrite
Ferrobacillus
Pollution Abatement
Industrial Wastes
Reaction Kinetics

-------
the reaction regime can be determined.
  Kinetic equations were derived for both reac-
tion modes.  From these basic relationships the
oxidation rate in real pyritic systems can be
accurately predicted when conditions at the
reaction site are known.
  For a given pyrite surface oxygenation rate is,
for all practical purposes, dependent only on the
oxygen concentration in the aqueous phase sur-
rounding the reactive site on the pyrite surface.
Ferric ion oxidation rate is determined by the
ferric/ferrous ratio and free ferric ion concen-
tration in solution, and is not affected by dis-
solved oxygen.
  In a real system, the ferric/ferrous ratio is
determined by a "relative" microbial activity;
i.e., the number of bacteria per exposed pyrite
surface.  Bacteria concentration is limited by pH
and oxygen concentration as well as nutrient
levels.
the reaction regime can be determined.
  Kinetic equations were derived for both reac-
tion modes.  From these basic relationships the
oxidation rate in real pyritic systems can be
accurately predicted when conditions at the
reaction site are known.
  For a given pyrite surface oxygenation rate is,
for all practical purposes, dependent only on the
oxygen concentration in the aqueous phase sur-
rounding the reactive site on the pyrite surface.
Ferric ion oxidation rate is determined by the
ferric/ferrous ratio and free ferric ion concen-
tration in solution, and is not affected by dis-
solved oxygen.
  In a real system, the ferric/ferrous ratio is
determined by a "relative" microbial activity;
i.e., the number of bacteria per exposed pyrite
surface.  Bacteria concentration is limited by pH
and oxygen concentration as well as nutrient
levels.
the reaction regime can be determined.
  Kinetic equations were derived for both reac-
tion modes.  From these basic relationships the
oxidation rate in real pyritic systems can be
accurately predicted when conditions at the
reaction site are known.
  For a given pyrite surface oxygenation rate is,
for all practical purposes, dependent only on the
oxygen concentration in the aqueous phase sur-
rounding the reactive site on the pyrite surface.
Ferric ion oxidation rate is determined by the
ferric/ferrous ratio and free ferric ion concen-
tration in solution, and is not affected by dis-
solved oxygen.
  In a real system, the ferric/ferrous ratio is
determined by a "relative" microbial activity;
i.e., the number of bacteria per exposed pyrite
surface.  Bacteria concentration is limited by pH
and oxygen concentration as well as nutrient
levels.

-------