WATER POLLUTION CONTROL RESEARCH SERIES
14010—06/69
     Oxygenation of  Ferrous Iron
U.S. DEPARTMENT OF THE INTERIOR • FEDERAL WATER QUALITY ADMINISTRATION

-------
       WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Reports describe
the results and progress in the control and abatement
of pollution in our Nation's waters. They provide a
central source of information on the research, develop-
ment, and demonstration activities in the Federal Water
Quality Administration, in the U. S. Department of the
Interior, through inhouse research and grants and contracts
with Federal, State, and local agencies, research institu-
tions, and industrial organizations.

Inquiries pertaining to Mater Pollution Control Research
Beports should be directed to the Head, ProjectsReports
System, Planning and Resources Office, Office of Research
and Development, Department of the Interior, Federal Water
Quality Administration, Room 1108, Washington, D. C, 2024.2.

-------
     Oxygenation  of Ferrous Iron
                      by
            Harvard University
      Cambridge, Massachusetts 02138
                    for the


    FEDERAL WATER QUALITY ADMINISTRATION

         DEPARTMENT OF THE INTERIOR
                  14010—06/69

             Contract PH 36-66-107
                  June, 1970
For sale by the Superintendent of Documents, U.S. Government Printing Office
        Washington, D.C. 20402 - Price $1.76 (paper cover)

-------
          FWQA Review Notice
This report has been reviewed by the Federal
Water Quality Administration and approved for
publication. Approval does not signify that
the contents necessarily reflect the views and
policies of the Federal Water Quality Adminis-
tration, nor does mention of trade names or com-
mercial products constitute endorsement or re-
commendation for use.

-------
                            ACKNOWLEDGMENTS

     This work is based to a large part on research supported by contract
PH 86-66-107 between the Federal Water Pollution Administration, Department
of the Interior, and Harvard University.  In the contract the scope of
the work was stated as follows:
Specific Aims;

     (1)  To determine rates of air oxidation of ferrous iron in the
                        _o
          presence of SO,  within the pH range 2-5;
     (2)  To determine rate of ferric iron hydrolysis within the pH
          range 2-5;
     (3)  To investigate the colloid-chemical properties of hydrolyzed
          iron  (III) Parameters:  (Fe+ ),  (Fe  ),  (H ),  (S04 ), PQ2;
     (4)  To investigate specific aims 1,  2, and 3 above under  the
                                                +2    +2
          effect of  the  following catalysts:  Mn   , Cu   , Si (QH)^,
          Si02(S), Fe203(S).

     The research  effort resulting  from this contract has become part
 of a thesis presented in April,  1969  by Philip Charles  Singer to the
 Division of Engineering and Applied Physics  of Harvard  University  in
 partial fulfillment of the requirements  for  the  degree  of Doctor of
 Philosophy.  In order to accomplish a well rounded research objective,
 the scope of the work was expanded to include  a comprehensive  treatment
 on the chemistry of aqueous iron, and to consider models describing
 pyrite oxidation.
      A good deal of experimental data was collected in order to transform
 the basic ideas and concepts of this research  into the usable  conclusions
 which have been reached.  The laboratory assistance of Karlene  Spencer,
 Gay Kunz, and Karl Schneider is an integral part of this report.
      Appreciation is also due to Mr.  Ronald Hill and Mr. Robert Scott
 and their staff of the Federal Water Pollution Control Administration
 who were instrumental in affording the authors opportunity and help'to
 conduct field studies relating to acidic mine  drainage.  Grateful acknowl-
 edgment is also extended to Professors J. Carrell Morris and Ralph Mitchell
 for their fruitful suggestions and for patiently reading and criticizing
 this manuscript.
                                   -iii-

-------
     Finally, Philip C. Singer is indepted to the United States Public
Health Service for a traineeship that provided financial support during
a major part of his graduate studies at Harvard University.
                                  -iv-

-------
                         TABLE OF CONTENTS
                                                                 page


ACKNOWLEDGEMENTS	iii

LIST OF FIGURES	,	    ix

LIST OF TABLES	xiii

SYNOPSIS   	xiv


Chapter 1:  INTRODUCTION

Chapter 2:  EQUILIBRIUM RELATIONSHIPS OF AQUEOUS IRON

            2-1  Introduction	2-1

            2-2  Iron(ll) Solubility  	  2-3

                 2-2.1  Solubility in Natural Waters 	  2-3

                 2-2.2  Recent Observations of Ferrous Iron
                        Solubility in Carbonate-Bearing Waters .2-6

                 2-2.3  Experimental  Determination of the
                        Solubility Product of Ferrous Car-
                        bonate (Siderite)	2-9

                        Experimental  Procedure  	 ....2-9
                        Experimental  Results and Discussion  . .  2 -12
                        X-Ray Analysis of Precipitate	2 -22
                        Stability Constant of FeHCO+   	  2-22
                        Summary  of Experimental Study   .....  2 -26

            2-3  Solubility of Ferric Iron	2 -27

                 2-3.1  Solubility in Natural Waters 	  2-27

                 2-3.2  Effect of Complex Formation on Fe(lll)
                        Solubility	2-30
                 2-3.3  Experimental  Determination of Sulfato-
                        Complex  of Fe(IIl)	2-32

                        Experimental  Procedure  ...,,....2 -33
                        Experimental  Results and Discussion. . .  2 -34

            2-4  Oxidation - Reduction Reactions of the
                 Iron (II)-Iron  (ill) System 	  2-40

            References	2 -45

-------
                                                                  Page
Chapter 3:  KINETICS OF OXYGENATION OF FERROUS IRON
            3-1  Introduction .......  	  ....3-1
            3-2  Oxygenation of Ferrous Iron at Neutral
                 pH-Values	3-2
                 3-2.1  Oxidation in Natural Groundwaters  ....  3 - 2
                 3-2.2  Oxidation of Fe(ll)  in the Presence
                        of Ferrous Carbonate Oversaturation .  .  .3-5
                        Experimental Procedures 	  3-6
                        Experimental Results and Discussion ...  3 - 7
            3-3  Oxygenation of Ferrous Iron in Acidic Systems.  .  3 -13
                 3-3.1  Experimenal Study of Kinetics of
                        Fe(Il) Oxidation at Acidic pH-Values  .  .  3 -13
                        Experimental Procedure  	  .3 -13
                        Experimental Results and Discussion ...  3 -16
            3-4  Oxygenation of Ferrous Iron as a Function
                 of pH	3-24
                 3-4.1   Summary of Experimental Results	3 -24
                 3-4.2   Kinetic Implications of Results 	  3 -27
            References	.3 -36
Chapter  4:  HYDROLYSIS OF FERRIC  IRON
            4-1  Introduction  	 ..... 	  4-1
            4-2  Kinetics of Ferric Iron Hydrolysis	4-2
                 4-2.1   Reactions of Fe+  with Water	4-2
                 4-2.2   Experimental Study  of the Kinetics
                         of'Fe(IIl) Hydrolysis  	  4-4
                         Experimental Procedui'e	4-4
                         Experimental Results  and Discussion . . .4-6
                         Solubility Product  of Amorphous
                         Ferric Hydroxide	4 -15
            4-3  Coagulative Properties of  Ferric  Iron	4 -15
            4-4  Removal of Phosphate  	 .....4 -19
                                   VI

-------
                                                                  Page

                4-4. L  Precipitation of  Phosphate by
                       Ferric Iron	 4 -21
                       Experimental Procedure	 4 -21
                       Experimental Results  and Discussion  .... 4 -22

            4-5 Summary  	  ...........4 -25

            References	4 -26

Chapter 5:   OXIDATION OF IRON PYRITE:  POLLUTION OF NATURAL WATERS
            BY COAL MINE DRAINAGE

            5-1 Introduction ........  	 5-1

            5-2 Thermodynamics and Stoichiometry of Reactions   .  . 5 - 2

            5-3 Previous Investigations  of the Kinetics and
                Mechanisms of Pyrite Dissolution  	 5-4

                5-3.1  Physical and Chemical Studies  	.5-4

                5-3.2  Microbiological Studies 	.5-7

            5-4 Purpose of Experimental  Study   .......... 5 -10

            5-5 Oxygenation of Ferrous Iron	...5 -11

                5-5.1  Experimental Procedure   	 5 -12
                5-5.2  Experimental Results  and Discussion  .... 5 -14

                       Effect of Sulfate	  ., 5 -14
                       Catalysis by Dissolved  Metal Ions	5 -17
                       The Effect of Clays	5-19
                       Catalysis by Powdered Charcoal   ...... 5 -24
                       Effect of Iron Pyrite	5 -25
                       Effect of Microorganisms	5 -26
                       Summary	..5 -26

            5-6 Field Investigations of  Pyrite Oxidation  in
                Natural Mine Waters	5 -26
                5-6.1  Collection and Analysis of Samples   .... 5 -27
                5-6.2  Results of Field  Investigation	5 -31
                       Stoichiometric Relationship Between
                       Sulfate Concentration and Acidity  	 5 -31
                       Rate of Oxidation of  Ferrous Iron	5 -33
                       Comparison with Laboratory Results   .... 5 -35
                       Implications of Field Results  .	5 -38

            5-7 Oxidation of Iron Pyrite	5 -44
                                  VlJ.

-------
                                                                 Page

                 5-7.1  Experimental  Procedures  	 5 -44

                 5-7.2  Results and Discussion	5 -46

                        Rate of Oxidation in Absence of Oxygen.  . 5 -46
                        Oxidation Rate in Presence of Oxygen   .  . 5 -51

            5-8  Conclusions	5 -53

                 5-8.1  Model Describing Pyrite  Oxidation
                        and Pollution by Coal Mine Drainage  ... 5 -53

            References	5 -58

Chapter 6:  CONCLUSIONS

            6-1  Principal Findings  	 6-1

            6-2  Practical Consequences and Implications
                 Resulting from this  Research	6-2

            References	6-7

APPENDICES

    A       Corrections of Experimental Solubility Data for
            Temperature and Activity

    B       Relative Significance of Soluble Phosphato-Complexes
            of Fe(lll)

    C       Derivation of Relationship Between Redox Potential
            and Sulfate Concentration for Determination of
            Stability Constant for FeSO +

    D       Thermodynamic Stability of Iron Pyrite

    E       Kinetics of Microbial Growth

    F       Autotrophic Iron Bacteria - Ratio of Ferrous Iron
            Oxidized to Organic Carbon Synthesized
                                  viii

-------
                          LIST OF FIGURES
                                                                  Page

Figure

    2-1   Solubility of ferrous iron in natural waters ....   2-5

    2-2   Solubility of ferrous iron in carbonate-
            bearing waters	2-5

    2-3   Solubility of ferrous iron in sulfide-bearing
            waters
                                                                  2 - 8
    2-4   Dual saturation-index diagram for calcite and
            siderite  in natural waters 	  2-8

    2-5   Determination of solubility product of ferrous
            carbonate	•	2 -14

     2-6   Conformance of  experimental solubility product
            to observations in natural groundwaters   	  2 -18

     2-7   Determination of solubility product of ferrous
            carbonate	2 -18

     2-8   Determination of solubility product of ferrous
             carbonate  ......  	  2 -20

     2-9    Determination of solubility product of ferrous
             carbonate	2 -20

     2 -10    X-Ray diffraction pattern of  experimental
             ferrous carbonate  	  2 -23

     2 -11    Standardization curve for divalent cation
             electrode in ferrous perchlorate solution  	  2 -24

     2 -12   Determination of free ferrous iron in bicar-
             bonate solution	2 -24

     2 -13   Solubility of  ferric iron	2 -28

     2 -14   Solubility of  ferric iron in phosphate solution. .  .   2  -28

     2-15   Experimental apparatus for potentiometric
             analyses	•   2  -35

     2 -16   Experimental data for determination of stability
             constant  of FeSO +	2 -37


                                   ix

-------
Figure                                                            Pa§e

    2 —17   Determination of stability constant of FeSO^'  «...  2 -37

    2 -18   p€ - pH diagram for  iron	«  2 -42


    3 - L   Oxidation  and removal of ferrous iron under
            conditions  favoring  precipitation of ferrous
                                                                  -in
            carbonate	J - ?

    3-2   Oxidation  and removal of ferrous iron in the presence
            of FeCO, oversaturation  ........ 	  3-9
     3-3   Effect  of  FeCO,  precipitation en Fe(Il) oxidation
            and  removal
3 -     -                                3 -11
     3-4  Effect of FeCO-  precipitation  on Fe(Il) oxidation
            and  removal  .............  ........ 3 -11

     3_5  U -  V absorbance spectra of  acidified  solutions
            of ferric perchlorate ...............  • 3 -14

     3-6  Relationship between absorbance of  acidified
             solutions of Fe(IIl) and Fe(lll)  concentration,
             at 272 mu ......................
     3 -- 7   Rate of oxygenation of Fe(ll) in bicarbonate-
             buffered systems  ...... .  ...........  3  -lc

     3-8   Rate of oxygenation of Fe(ll) ............  3  -19

     3-9   Rate of oxygenation of Fe(ll) ......  .  .....  3  -19

     3 -10   Oxygenation of Fe(Il) at pH 3 ...  .......  .  .  3  -20

     3 -11   Rate of Oxygenation of Fe(ll) at pH 2 .....  ...  3  -20

     3 -12   Oxygenation of ferrous iron at various initial
             concentrations of Fe(ll), at pH 3 ........  ,  .  3  -22

     3 -13   Rate of oxygenation of Fe(ll) over the pH-range
             of interest in natural waters ............  -  "25

     4-1   Logarithmic and reciprocal plots of the rate of
             hydrolysis of Fe+3   ......... . .......  4-8

     4-2   Logarithmic and reciprocal plots of the rate of
             hydrolysis of Fe+3   ....... . .........  4~9

     4-3   pH-dependence of  "first-order rate constant" for
             hydrolysis of Fa4"3   ............. . ...  4 -11

-------
Figure
    4-4   pH-dependence of "second-order rate constant"
            for hydrolysis of Fe+3 ................ 4 -11

    4-5   pH-dependence of "second-order rate constant"  for
            hydrolysis of F&   in the presence of sulfate   .... 4 -14
                                                         fj
    4-6   Comparison between rates of hydrolysis of Fe
            in the presence and absence of sulfate .  . ...... 4 -14

    4-7   Aggregation of colloidal silica dispersions by
            hydrolyzed ferric iron ............ .... 4 -18

    4-8   Solubility of ferric phosphate ............ 4 -18

    4-9   Precipitation of phosphate by homogeneously-
            generated ferric iron  ................ 4 -23

    5-1   Effect of sulfate on absorbance of Fe(lII) at
            272 mu ........................ 5-13

    5-2   Rate of oxygenation of ferrous iron as a. function
            of pH  ........................ 5-13

    5-3   Rate of oxidation of ferrous  iron in the presence
            of  sulfate  ...................... 5-15

     5-4  Effect of  sulfate on the oxidation rate of Fe(Il)
             at 50°C   .........  .  .............  5-15

     5-5  Effect of copper(ll)  on  oxidation of ferrous  iron  .  .  5  -18

     5-6  Rate of  oxidation of  Fe(ll)  in the  presence of
             suspended aluminum  oxide ....... ...,...,5  -18

     5-7   Oxidation of Fe(II)  as a function  of AJU°o
             concentration  . ................. .  .  5  -22

     5-8   Effect of pH on surface-catalytic  oxidation of
             Fe(ll)  ........................  5-23

     5-9   Rate of  oxidation of  Fe(Il)  in the presence of
             colloidal silica and  bentonite clay  .........  5 -23

     5 -10   Mining  sites for field investigations  of Fe(Il)
             oxidation,  near Elkins,  West Virginia ........  5 -28
                                    xi

-------
Figure
                                                                   Page
    5 -11   Stoichiometric relationship between acidity and
            aulfate concentration in mine drainage waters ....   5 -34

    5 -12   Chemical composition of drainage water through
            a strip mine	«	•	   -> -36

    5 -13   Oxidation of Fe(Il) in drainage water after
            leaving strip mine	5 -37

    5 -14   Rate of oxidation of Fe(Il)  in water collected
            from air-sealed underground mine	5 -37

    5 -15   Oxidation of Fe(ll) in water collected from air-
            sealed mine	5 - .2

    5 -16   Change in Fe(ll) concentration in  millipore
            filtered water collected  from air-sealed mine  ....   5 -42

    5 -17   Oxidation of Fe(ll) solutions inoculated with
            mine water	•	•* ~^

    5 -18   Reduction of ferric iron  by  iron  pyrite in the
            absence of oxygen  .........  	  5 -47

    5 -19   Rate of reduction  of  Fe(lll)  as  a function of
            pyrite concentration	->   f9

     5  -20   Effect of  initial  Fe(IIl) concentration on
            rate of reduction  of  Fe(IIl) by  pyrite	5 -50

     5  -21   Reduction  of Fe(IIl)  by  pyrite  in the  presence
             and absence  of oxygen	• •   -> "52


     B  - 1   Log concentration  diagram for phosphoric acid  ....   B - 2

     B  - 2   Distribution diagram for soluble hydroxo-species
             of ferric iron
                                                                    B -  2
                                    xii

-------
                           LIST OF TABLES

Table                                                           Page

    2-1   Equilibria Describing Fe(ll)  Solubility  	  2-4

    2-2   Experimental Determination of Solubility
            Product of FeC03	2-21

    2-3   Equilibria Describing Fe(lII)  Solubility  	  2  -29

    2-4   Experimental Data and Calculations in Determina-
            tion of Stability Constant for FeSO^+   	2  -38

    2-5   Equilibria for Construction of p£ - pH Diagram  ...  2  -43
     3-1   Kinetics of Oxidation of Ferrous Iron	3-3
     4-1    Check  on  the Solubility Product of Amorphous
             Ferric Hydroxide	4 -16
     5-1   Comparison of  Surface-Catalytic Rate Constants
             with Uncatalyzed Rate Constants	5 -21

     5-2   Chemical Catalysis  of Oxidation of Ferrous Iron ... 5 -27

     5-3   Summary of Field Data	5 -32
                                  Xlll

-------
                               SYNOPSIS



       The rate  of oxidation of iron(li)  by oxygen conforms  to  a rela-


tionship which is first-order in the concentrations of  ferrous  iron


and oxygen, and second-order in the concentration of hydroxide  ion,  at


pH-values between 6.0 and 7.5.  The reaction proceeds relatively ra-


pidly at pH-values greater than 6.5; the half-time of the reaction  is


4 minutes at pH 7.0, under a partial pressure of oxygen of 0.20 atmo-


spheres at 25°G.   Deferrization processes in water treatment employ


the rapidity of the oxidation reaction in order to remove the influent


iron(Il) as insoluble iron(lll) hydroxide.  Part of the iron(ll) may al-


so be removed as ferrous carbonate  (FeCO-), the solubility product  of


which is 6.0 x  10~   , as shown by this research.

                                                  _ 2
       The dependence of the oxidation rate on [OH ]  has been ob-


served,  in this study, at pH-values as low as 4.5, where the half-time


has  increased to approximately 300  days.  At lower pH-values, the  de-


pendence of the reaction rate on pH (or, more precisely, [OH ]) be-


comes less narked until at pH-values below 3.5, the oxidation proceeds


at a rate which is  independent of pH.  Here, a half-time of about  2000


days reflects the  slowness of the oxidation reaction.


        In  the acidic drainage waters issuing from coal mines, half of


the  acidity arises  from the  oxidation  of  the sulfide (S«(-II)) of iron


pyrite  (FeS_)  to  sulfate, and .half  stems  from the oxidation of  iron(ll)


to  iron(IIl) by oxygen and the  subsequent hydrolysis of the resultant
                                   xiv

-------
 Iron(lll).  Observations of the rate of oxidation of ferrous to ferric




 iron in these  acidic  streams  (at pH values close to 3) show it to pro-




 ceed considerably more rapidly than the laboratory studies at pH 3




 would predict.




        Several chemical agents which are indigenous to mine drainage




 waters have been cited in the literature, in various circumstances,




 as displaying catalytic properties in the oxidation of ferrous iron.




 These include  inorganic ligands, such as sulfate, which coordinate with




 iron(H) and iron(IIl), soluble metal ions, such as copper(ll), man-




 ganese(ll), and aluminum(lll), suspended material with large surface




 areas and high adsorptive capacities, such as clay particles,  and ma-




 terials which accelerate the decomposition of peroxides in the presence




 of iron(ll), such as charcoal.  An investigation into the catalytic




 capabilities of these chemical agents in synthetic mine waters demon-




 strates that clay particles, or their idealized counterparts alumina




 ^A^2°3^ and silica (Si°2^> exert the greatest influence on the rate of




 oxidation of iron(ll), but only at areal concentrations much larger than




 those encountered in most natural mine waters.   (The reaction  proceeds




 10-30  times more rapidly in the presence of 8000m /I of Al90-  than the




 uncatalyzed reaction.)




       Autotrophic microorganisms have frequently been implicated as




 the causative agents in the production of acidic mine drainage.   These




organisms are able to utilize the energy released by the oxidation of




 ferrous iron for their metabolic  processes.   A  study of the oxidation




of iron(ll) in natural mine streams near Elkins,  West Virginia shows
                                   xv

-------
the reaction to proceed at a rate which is zero-order in iron(II).




The field results suggest that the observed rapidity of the reaction




in acidic mine waters is apparently the result of microbial catalysis.




       Studies of the oxidation of iron pyrite in coal mine drainage




have usually considered only oxygen as the specific oxidant, with the




potentiality of iron(lll) as an oxidant having often been overlooked.




Experimental evidence obtained in laboratory systems shows that iron(IIl)




is rapidly reduced by iron pyrite both in the presence and absence of




oxygen.  There is virtually no difference between the rate of reduction




of iron(lll) by pyrite, or the rate of change of soluble iror.CII), under




aerobic  or anaerobic conditions  indicating that the  specific oxidant




of  iron  pyrite  is  ferric  iron.




       The time required  for  the reduction of 507« of the  initial




 iron(lll) concentration in contact  with 1 gram/liter of pyrite  is




 approximately 250  minutes which is  considerably  less than the half-




 time for the oxidation of iron(Il)  even when accelerated  by the chemi-




 cal catalysts  found in natural mine waters.  Consequently,  the rate-




 limiting step among the reactions involved  in the oxidation of  iron




 pyrite and the production of acidity in mine drainage  waters is the




 oxygenation of ferrous iron.




        Based upon the experimental  evidence presented,  the oxidation




 of iron pyrite in natural mine waters is shown  to  be compatible with a




 cyclical reaction model involving the slow oxygenation of iron(II)  to




 iron(IIl) followed immediately by the rapid reduction of iron(IIl)  by




 pyrite, generating in turn more iron(Il) and acidity:
                                    xv i

-------
                          slow
            Fe(ll)  + 02 	=v  Fe(lll)
                             fast                   „
            Fe(lll) -t ~ ~           -  " ""
The oxidation of iron(ll) by oxygen is the rate-determining reaction.

Oxygen is involved only indirectly in the oxidation of pyrite;  it

serves to regenerate iron(lll) which is itself the specific oxidant

of pyrite.  Precipitated iron(IIl) hydroxide within the mine serves as

a reservoir for soluble iron(IIl).

       The experimental results and the model are discussed from the

standpoint of evaluating the various control measures which have been

proposed  in order to deal with the costly problem of acidic mine drain-

age.   In  view of  this  research, emphases needs to be placed upon halt-

ing the catalytic oxygenation of  iron(Il).
                                   xv 11

-------
                                 CHAPTER 1







                               INTRODUCTION







       Aqueous iron assumes an important role in natural waters in the




limnological cycles of several key elements,  in water supplies and water




purification processes,  and in the formation  of several types of industrial




wastes.  Many natural phenomena can be explained and many practical appli-




cations can be derived as a result of investigations of the chemical be-




havior of aqueous iron.   An account of the equilibrium and kinetic relation-




ships which describe and control the distribution and activity of aqueous




iron in natural waters is reported in this thesis, with special emphasis




placed on the kinetics of oxygenation of iron(II), its application to de-




ferrization processes, and  its involvement in the formation of acidic coal




mine drainage.




       The  equilibrium relationships which govern the solubilities of ferrous




and  ferric  iron are  considered in Chapter  2  in order to gain some insight




as to  the concentrations of the various species  expected in natural waters.




Included is a redetermination of the solubility  product of ferrous carbonate




and  an estimate of the stability constant  for a  possible bicarbonate-complex




of Fe(II).




       The  kinetics  of the. oxygenation of  ferrous iron  are discussed in




Chapter 3,  in both neutral  and acidic systems.   In the  former case, attention




is paid to  the oxygenation  reaction  in bicarbonate solutions which are




supersaturated with  respect to ferrous carbonate.  The  rate of oxygenation

-------
                                                                       1-2





of ferrous iron over the entire pH-range of interest in natural  waters  is




described, and the results are analyzed in view of modern kinetic  theory




and the various mechanisms proposed for the reaction.




       The pH-dependence of the oxygenation reaction in neutral  and slightly




acidic systems, where hydrous ferric oxide was observed as the product  of




the reaction, gave reason to suspect that hydrolysis of ferric iron was




directly  involved in the oxygenation of ferrous iron.   Chapter 4 contains




an investigation of the kinetics of hydrolysis of Fe+  in systems over-




saturated with respect to ferric hydroxide.  The coagulative properties of




hydrolyzed ferric iron and phosphate removal by oxygenated ferrous iron are




also discussed.




       The chemical results obtained in Chapters 3  and 4 are applied in




Chapter  5, a study of the chemistry of coal mine drainage.  The kinetics




of the various  reactions  involved  in the  oxidation  of  iron pyrite and the




release  of acidity  are  investigated and their  relative rates compared in




order  to ascertain  which of the  sequential reactions is rate-limiting.   The




oxygenation  of  Fe(ll) is considered, subject to the catalytic influences of




several  agents  which are  indigenous to natural mine waters, including micro-




organisms.   A model is  proposed,  incorporating the  salient features of the




kinetic  study,  in order to describe the oxidation of iron pyrite  in mine




drainage waters.  The pertinent  consequences of the model are examined.




       The  significant  results of  the  research are  summarized in  Chapter




 6,  along with some  of the practical applications  of these results.

-------
                                  CHAPTER  2







                 EQUILIBRIUM RELATIONSHIPS  OF  AQUEOUS  IRON







2-1  Introduction






        The various species of iron which  exist in natural  waters  are




governed by solubility,  hydrolysis, complex-formation,  and  oxidation-




reduction relationships.  These relationships predict whether  or not a




given reaction will take place as written, and to what extent,  i.e.,




what concentration of a given species is expected.  In contrast, kinetic




relationships, which will be considered in Chapters 3 and 4,  are needed




to predict the rate at  which  equilibrium  is attained.




         In order to understand the behavior of iron in natural waters, a




theoretical treatment is  required in which various equilibria are assumed




to be  applicable in controlling  the different  species of iron.  Such a




theoretical treatment is  not  intended  to  provide an  all-inclusive chemical




description due  to the  complexity of the  natural system, but rather  is




 intended as an oversimplified version  to  gain  some insight as to which




 equilibria are relevant.   The thermodynamic data employed  have  been ob-




 tained in well-defined  systems where the  individual  variables were  iso-




 lated.  The combined  investigations of a  number  of such isolated  systems




 are then compared  to  the  natural system.   Deviations between  the  behavior




 predicted by  thermodynamic considerations and that which actually occurs




 in the real system can  be attributed to the existence of non-equilibrium

-------
                                                                       2-2




conditions where the kinetics of the reactions are limiting,  or  to  an  im-




proper description of the system due either to a lack of  dependable thermo-




dynamic data or to an oversimplification in predicting which  are the con-




trolling equilibrium relationships.   The occurrence of such deviations be-




tween the "predicted" and the "actual" requires an explanation.   A  portion




of the material presented in this chapter is intended for just  this pur-




pose.




       The material covered is limited to a discussion of the solubility




of iron compounds in natural waters.  In the case of ferrous  iron,  its




solubility is governed by the solubility of its respective carbonate,  hy-




droxide, or sulfide, depending upon the composition of the water.  The




solubility of ferric iron is correspondingly limited by its rather  in-




soluble oxides and hydroxides or, in the presence of high concentrations




of phosphate, by ferric phosphate.  The equilibrium concentration of Fe(lll)




may be increased as a result of complex-formation with chloride, sulfate,




phosphate, silicate, and organic matter.




       Experimental studies of the solubility of iron in natural waters




are also presented in this chapter.  These studies include a determination




of the ill-defined solubility product of ferrous carbonate, FeCO •  an  ex-




perimental analysis of the relevance of the bicarbonato-complex of  Fe(ll),




FeHCO- , in carbonate-bearing waters; and a determination of  the stability




constant of the sulfato-complex of Fe(lll), FeSO, , which was essential for




consideration of the catalytic effect of sulfate on the hydrolysis of




ferric iron, which is discussed in Chapter 4.




       Since the relevant, features of the Chemistry of aqueous iron per-




taining to natural waters have previously been discussed by Stumin and




Lee (1), only an overall review is presented here, including an up-dating

-------
                                                                         2-3






 of their discussion with more recent  data.   The  major  portion of  the chap-




 ter is devoted to the experimental  work undertaken  to  clarify our under-




 standing of the behavior of  iron in natural  waters.






 2-2 Iron(ll)  Solubility




         2-2.1  Solubility in Natural  Waters







         Under reducing conditions in  natural waters, as  in the bottoms of




 lakes  under conditions of stagnation,  and  in most groundwaters, ferrous




 iron,  in the -t- II oxidation  state,  is the  stable form  of iron.  In waters




 free of  dissolved carbon dioxide and  sulfide, the solubility of Fe(ll) is




 controlled  by solid  ferrous  hydroxide,  Fe(OR) , as shown in Figure 2-1.




 The equilibria utilized  in plotting this solubility diagram and all subse-




 quent  diagrams for Fe(ll) are  given in Table 2-1.  (The reader is referred




 to  Sillen's discussion (2) on  the graphic representation of equilibrium data




 for the  general principles in  the construction of such solubility diagrams.)




         In  natural groundv/aters, alkalinities often exceed 5 x 10   eq./l.




 (7).   Figure  2-2  indicates that, for  a water containing 5 x 10   moles/1.




 of  total carbonic  species CT,  the solubility of Fe(ll)  is markedly influ-




 enced  by formation of  ferrous  carbonate, FeCO-.   It is  immediately evident




 that at pH-values below  10.5,  ferrous carbonate controls the concentration




 of  Fe(ll) in  solution.




        In  hypolimnetic waters where the concentration  of sulfide species




 may be appreciable as  a result of anaerobic reduction of sulfate,  ferrous




 sulfide, FeS, as shown in Figure 2-3 for a total  sulfide content of  10~5




moles/1. (8),  limits the solubility  of Fe(ll) over  the  entire range  of  pH




 encountered in naLural systems.  It  should be noted  that even in the

-------
                                                           2-4
Table 2-1.  Equilibria Describing Fe(ll)' Solubility
Equat ion
Number
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
2-9
Equilibrium
Constant
Reaction at 25°C
Fe(OH)2(g) = Fe+2 + 20H~ 8 x 10"16
Fe(OH)2(g) . FeOH+ + OH" 4 x 10~10
Fe(OH).., , + OH" = Fe(OH)~ 8.3 x 10~6
£.\ S) J
FeC03(s) = Fe+2 + C03"2 2.1 x 10"11
H2C03 = H+ + HC03" 4.2 x 10~7
HC03~ = H+ + C03"2 4.8 x 10"11
FeS(g) = Fe+2 + S"2 6 x 10"18
H0S, x = H+ + HS" 1.0 x 10~7
2 (aq)
HS" = H+ + S"2 1.3 x 10"13
Refer enc
3
3
4
5
5
5
6
6
6

-------
                                                                     2-5
   -3
£
w
8
   -6
   -7
                 8
10       11
   PH
                                                12       13
     FIGURE 2-1.  Solubility  of  ferrous iron in waters free of
                  appreciable alkalinity and sulfide.
 C3
                     FeC03 + 2 OH  • Fe(OH),, + CO        /
                                                   x^    /
                   C - 5xlO"3M
                                                   1       /
                       /
                                                  -2
                                   3 (s)
                       8
               11
                         9        10
                              PH
FIGURE 2-2.  Solubility of ferrous iron in waters containing
             5x10* M of total carbonate species.

-------
                                                                       2-6
presence of a concentration of 10  ti total carbonic species, ferrous sul-



fide continues to be the stable solid phase.



       If equations 2-4 and 2-7 of Table 2-1 are combined to give




            FeS, x + CO."2 = FeCO.,, , + S~2                         (2-10)
               Is)     3         3v.s)



where the equilibrium constant for the reaction, as written, is
            K   =    =iO--   = 3xlO=
             •*   K4


it is seen that FeCO- becomes the  stable  solid phase (the reaction proceeds



as written) only when the ratio of  sulfide to carbonate is smaller than



3 x 10~  .  In. other words,  the solubility product of ferrous carbonate be-



comes operative only when the ratio of  alkalinity to total sulfide exceeds



2 x 10   at pH 6, or 8 x 10   at pH  7.  Hence,  at  the  extremely low concen-



trations of  sulfide found  in most  groundwaters,  ferrous carbonate governs



the solubility of  ferrous  iron,  as in Figure  2-2.





2-2.2   Recent Observations  of Ferrous Iron Solubility  in_ _P.arbQn,ate_-

        Bearing Waters
        The equilibrium constant for  the reaction



                           O       *?


             FeC03(s)  = F     + C03                                    (2"4)



 given in Table 2-1,  was computed from the free energy  data  tabulated  by



 Latimer (5), and is  based upon the experimental work of  Smith  (9)  iu  1913



 and the calculations of Kelley and Anderson (10).   Smith found the solu-


                                               +2      -2
 bility product of ferrous carbonate, K   - [Fe  ]  [CO, *"],  to  be  3.45 x
                                       SO            J


 10    at 30°C without correcting for activity.  Kelley and  Anderson modified



 that value for 25°C  and an ionic strength of zero  (although their method of



 correction is not immediately evident) and arrived at  a free energy of

-------
                                                                     2-7



solution for reaction 2-4 of 14.54 Kcal./mole, corresponding to a thermo-



dynamic equilibrium constant of K   = 2.1 x 10
         n                       so


       More recent reports, however, indicate some discrepancy between



field measurements of the solubility of Fe(ll) and the solubility predic-



ted using the accepted thermodynamic constant from the literature.



Hem (11), in examining twenty groundwaters for equilibrium with calcite



(a crystalline form of calcium carbonate) and siderite (the crystalline



form of ferrous carbonate), presents a "dual saturation-index" diagram



for the two minerals.  In general, one would expect a groundwater in



equilibrium with calcite to be in equilibrium with siderite, too, if both



minerals were present in the same geologic formation.  Furthermore,  one



would expect similar conditions of under- and oversaturation with respect



to the two solid phases.  Yet, the best straight line through Hem's  data



does not pass through the origin, although it has roughly a slope of



unity indicative of equivalent conditions of saturation (see Figure  2-4).



An error in the value of K   would result in such an observation.  (The
                          so


term "measured pH minus computed pH" is equal to the logarithm of the



degree of oversaturation.)



       Ghosh, O'Connor, and Engelbrecht (7) sampled the influent ground-



water at eight water treatment plants in Illinois and reported values of



oversaturation with respect to ferrous carbonate of from 20 to 40 times.



When these observations were corrected for temperature and activity  (12),



the oversaturation compared to the accepted value given by Latimer was



still on the order of five to ten times.



       Larson (13) has suggested that the existence of a bicarbonato-



complex of ferrous iron, FeHCO,  , may serve as an explanation for this



reported condition of apparent oversaturation, especially in view of Hem's

-------
                                                                     2-8
  -5
SB
o
M


I
-6
Ed
O
SB
o
  -7
2
  -8
                                                    Fe(OH)
                                     FeS
                                          (s)
                               M I  T I 11  i n i i  i i I I f  I i I \
                                                    i
      67         8        9        10       11         12       13

                                      PH                               .3

      FIGURE 2-3.   Solubility of ferrous iron in waters  containing 5x10  M

                   of total carbonate species and 10 M  of total sulfide

                   species.
            •f 2
               -3
       FIGUFvE 2-4.
                    -2      -1

                    measured pH
 0      +1

computed pH
                                                     +2
                 Dual  saturation-index  diagram for  calcite  and  siderite

                 in  natural  waters  (after  Hem (11)).   Measured pH -

                 computed  pK is  equivalent to the log of  the degree  of

                 oversaturation.

-------
                                                                        2-9
findings (14) that approximately 357o of the total manganous manganese,

Mn(ll), is present as the bicarbonate-complex,  MnHCO, ,  in a  grovmdwater

containing 5 x 10   eq./l. of alkalinity.


2-2.3  Experimental Determination of the Solubility Product of Ferrous
       Carbonate (Siderite)


        Since the solubility of ferrous iron in groundwaters  has been

shown to be limited by the solubility of ferrous carbonate, and since

the 50-year-old solubility product cannot account for the high concentra-

tions of Fe(ll) found in natural groundwaters,  this study was undertaken

in order to redetermine the  solubility product of ferrous carbonate,  and

to determine what effect,  if any, complexation of Fe(Il) by bicarbonate

has on the overall  solubility of Fe(ll).


Experimental Procedures

        Ferrous carbonate was prepared directly  in the laboratory by preci-

pitation  from  a solution  of  excess  ferrous perchlorate and sodium bicar-

bonate.   Ferrous  perchlorate was  added to  a pre-determined concentration of

perchloric  acid  in  a BOD  bottle and the  system was flushed with nitrogen

to remove all  traces of oxygen.   (Since conditions for the precipitation

of ferrous  carbonate were found to  be optimal  above  pH 6.5, the exclusion

of oxygen was  a prime requirement,  Fe(ll)  being rapidly  oxygenated at pH-

values greater than 6.0  (15).)   Sodium bicarbonate was added  and the vessel

was  immediately  stoppered to prevent contamination by oxygen.  A series of

 such bottles were stored  under  water at  constant temperature  again to pre-

vent seepage of oxygen  into the system,  the  exclusion of oxygen proving to

be the major experimental difficulty.

-------
                                                                   2-10







       The preparations of  ferrous  carbonate were made  in  a  constant




ionic medium to insure the  constancy of  the activity  coefficients  for




each series.




       After a period of storage of one  to five months,  the  samples




were removed from the water bath and aliquots were  taken from each for




the determination of alkalinity, soluble ferrous iron,  and concentration




pH.  This latter term, pCH, is a measure of the concentration of H  at  a




given  ionic  strength.  A combination pH electrode (Beckman Cat. No.




39142) was  standardized in a reference solution containing a known con-




centration  of HG1 and  the  same  ionic strength as the sample being ana-




lyzed.  The p°H of  the sample was  then measured (Corning Model 12 pH




meter) by immediately  inserting the pH  electrode into the vessel when it




was opened.  It was feared that rapid evolution of CO^  upon exposure of




the sample to the atmosphere  would raise  its p  H, but  the measured ? H




was found to remain relatively  constant.




        Alkalinity was determined by acidimetric titration with 0.1M HC1.




Since oxidation of Fe(ll)  and hydrolysis  of the resultant Fe(lII) pro-




 duces acidity which neutralizes a  portion of the alkalinity,  an  aliquot




 was rapidly titrated to pK 5  with  a pre-determined amount of HC1  and




 then slov/ly titrated to the endpoint at pH 4.3.  (Oxidation of Fe(ll)




 below pH 5  is relatively slow,  as  will  be seen in  Chapter 3.)




        The  determination of soluble ferrous iron was carried out by




 rapidly filtering an  aliquot of the sample into a test tube containing




 dilute acid in order  to quench any further reaction.  The filtration




 was performed under an atmosphere of carbon dioxide to prevent oxidation




 of Fed!)  and dissolution of FeCC>3.  100,  220, and 450 mu.  filter papers

-------
                                                                    2-11







(Millipore Filter Company)  were employed,  similar results  being  obtained




for each.   Filterable Fe(ll)  was measured  using the colorimetric reagent,




bathophenanthroline (4,7-diphenyl-l,10-phenanthroline)  (10).




       To  insure that the precipitate formed was crystalline  ferrous




carbonate  (siderite) and not  merely an amorphous intermediate,  the sedi-




ment was analyzed by X-ray diffraction.  The precipitate was  collected




following  filtration and air-dried overnight before determining  its




crystal structure using a Norelco X-ray diffractometer.




       For the analytical determination of the stability constant of the




bicarbonate-complex of Fe(ll), a specific  ion electrode (Orion  Research)




was employed.  The electrode contains a liquid ion-exchange resin having




a specificity  for various divalent cations, including Fe  , and is used




in conjunction with  any standard reference electrode.  The selectivity




coefficient for Fe    is high and in the absence of other divalent cations




the electrode  measures free Fe   directly.




       Various dilutions of a stock solution of ferrous perchlorate were




added  to oxygen-free samples under one atmosphere of carbon dioxide and




containing pre-set concentrations of  sodium bicarbonate.  Again, exclu-




sion of oxygen was mandatory to prevent oxidation of Fe(ll).   For a




given  alkalinity, the greatest concentration of Fe(ll) was added such




that the solubility  product of ferrous carbonate was not exceeded.  A




constant ionic medium was again maintained to  reduce variations in acti-




vity coefficients among the different systems.




       Following the addition of Fe(ll) to the  solutions of bicarbonate,




the samples were allowed one hour  to  reach equilibrium.  The divalent




cation electrode was calibrated using standardized  solutions of ferrous

-------
                                                                       2-12
perchlorate, free of CO  and at the same ionic strength as the samples.



(~"le solutions of Fe(C10,)_ had been standardized with permanganate,  which



itself had been previously standardized against oven-dried sodium oxalate.)



The potentials of the test solutions containing bicarbonate were measured



at the same time as the standard solutions, alternating between the un-



knowns and the standards since the electrode potential is prone to drift



with time.  The potentials were measured with a Corning Model 12 potentio-



meter.  The p H of the samples were recorded and aliquots were analyzed



for alkalinity and total ferrous iron, Fe(ll), by acidimetric titration



and by pote>  , -metric titration with standard permanganate, respectively.





Experimental Results and Discussion


        The solubility product of  ferrous  carbonate was computed utilizing



the following chemical equilibria:




             FeC03(s)  = Fe+2  + C03'2;      K°so =  [Fe+2][C03'2]        (2-4)




where KC    is the  concentration  solubility product of crystalline ferrous
         so


carbonate at  a known  ionic  strength and  temperature, and




             HC03" = H+ + C03"2;    KC2  =  [H+][C03"2]/[HC03"]         (2-6a)




where K% is  the second acidity  constant of  carbonic acid at  a given  ionic



strength.  It follows  that  the  equilibrium constant  for  the reaction
              FeCO ',  ,  +  H+ = Fe"1"2 +  HCO,"                            (2-11)
                  3(s)                   3
 is given by
                     KC       [Fe+2][HCO.

              Kc   _   so  _ 	:


                ^   K°2         Cir]

-------
                                                                      2-13
Rearranging and taking logarithms, one obtains
            pCH + pKCgo - PKC2 = -log ([Fe+2][HC03"])               (2-12)





where p- refers to the negative logarithm of that term.  Hence,  K°   can
                                                                  so
be readily calculated from a knowledge of the experimentally-determined




parameters p°H, soluble Fe(ll) (assumed to be equal to [Fe  ]),  and alka-




linity, and the known second acidity constant of carbonic acid under the




given experimental conditions.




       Equation 2-12 suggests that a plot of p°H versus -log([Fe+2][HCO "])




will yield a straight line of unit slope having an intercept at  p°H = 0




which is equal to pKC   - pK° .  Figure 2-5 is such a plot for a series of
                     SO      2.



preparations of ferrous carbonate at 22.5°C and an ionic strength of 0.1.




The best straight line of unit slope was fitted to the points  resulting




in an intercept of -0.57 ± .17, as shown.  The majority of the points fall




within the p H-range 6.5 to 7.5 where precipitation was found  to be optimal.




The four points at the lower p H contain the greatest experimental uncer-




tainty since, due to their low alkalinities, they required only  a small




amount of titrant before the endpoint was reached.  The fact that these




four points still approximate the linear plot is gratifying.




       The reason for the apparent scatter of the experimental points in


               ^

the region of p H 6.5 to 7.5 is due strictly to the experimental uncertain-




ty in measuring free Fe+ .  It has been assumed that filterable  Fe(ll)  is




equal to the concentration of free Fe+  in equilibrium with the  solid




phase.  The fact that the concentration of filterable Fe(ll) for a given




sample was constant for three filters of different pore size (100, 220,




and 450 mu,) lends credence to such an assumption in that all  solid Fe(ll)




is retained by the filter, i.e.,  filterable Fe(ll) equals soluble  Fe(ll).

-------
                                                                   2-14
    7.5
    7.0
   6.5
   6.0
- 5.5
60
O
   5.0
   4.0
                                             i	r
constant ionic medium
     I - 0.1

temp. - 22.5°C
     from Latimer (5)xx    °9
                       Intercept at p H « 0 is -0:57
       J	I         I         I
                5.0      5.5      6.0       6.5      7.0      7.5
                                   PCH
       FIGURE 2-5.  Experimental data for determination of solubility
                    product of ferrous carbonate (siderite).

-------
                                                                    2-15
On the other hand, it is certainly possible,  even probable,  that  some
dissolution of FeCO, and some oxidation of soluble Fe(ll)  does occur
during filtration.  These effects act in opposite directions and  tend
partially to cancel each other.  The extent of such dissolution and ox-
idation during filtration is manifested in the scatter in Figure 2-5.
The experimental uncertainty has been computed on the bases of the
greatest degree of scatter in the p H-region 6.5 to 7.5.
                                                    Q
       As shown by equation 2-12, the intercept at p H = 0 is equal to
the quantity pK°   - pKC_.  If one makes use of the Davies equation to
                SO      £-
estimate activity coefficients of single  ions as suggested by Schin-
dler  (17) for carbonates of bivalent metals, the thermodynamic second
acidity constant  of  carbonic  acid as a  function of temperature as re-
ported by Harned  and Scholes  (18),  and  the van't Hoff temperature rela-
tionship, the thermodynamic  solubility  product can be computed as shown
 in Appendix A.  The  resulting solubility  product, at  25°C and zero  ionic
 strength,  is 5.7  ±  2.3  x  10"11 or pK   is 10.24  ± 0.17.  The previously-
 accepted value  is 2.1 x 10~    or pK   =10.68,  as  shown in Table  2-1.
     r                              so
       Using the  experimental value of  pKC    = 10.24  ±  0.17, one  can
                                           SO
 compute the free  energy of formation of ferrous  carbonate  as  follows (the
 numbers in parenthesis  refer to the presently-accepted  values  (5)):

            FeC03U)  . Fe+2 + CO^2;  K^ - lO'10' "(UT10-68)      (2-4b)
          AF° = -1.364  log K   = 14.0 ± 0.2 K cal/mole  (14.54)  at 25°C
                             so
          AF° = 14.0 ^F°   „
                          Fe
                 14.0 = -20.3 - 126.2 -AF°
                                          FeC03
               AF°      = -160.5 - 0-2 Real./mole  (-161.06)
                   FeC03

-------
                                                                       2-16




                                      +2        -2
(The free energies of formation for Fe   and C0_   have been taken from



Latimer (5).)  Although the difference of only 0.6 K cal.  in 160 K. cal.



does not appear to be very significant, its importance is magnified when



considering the difference between two large numbers as is often done in



computing the change in free energy for a given reaction.   Such a small



error in the change in free energy can result in a much larger error when



the equilibrium constant for the reaction is computed.



       If the thermodynamic constant given in the literature is modified



to the experimental conditions of this study using the same temperature



and activity corrections described above, the resulting constant does not



fit the experimental data, as Figure 2-5 demonstrates.  The new solubility



product, however, can be shown to explain adequately the apparent over-



saturation described in the literature.  If Hem's data (11) for the reac-



tion




             CaC03,s) + H+  = Ca+2 + HCC>3~                            (2-13)




is considered, the degree  of oversaturation for calcite can be computed to



be
where Qr  is  the  actual reaction quotient  for  2-13  in the aquifer, arid KC




is the thermodynamic  equilibrium constant  for the  same reaction.  CL is




calculated  in  terms   of  the measured variables  as




                  (Ca+2)  (HGO  ~)

             Q^ =	E	L_B                                    (2-15)

             W       f U  "\
                      V. It  J
                         m




the  subscript  m  referring  to  measured  quantities,  and K  is given by

-------
                                                                    2-17



                 (Ca+2)  (HCO ~)

            ,,	m    3  m


             C =     (H+)
                         comp



where (H+)     refers to the computed activity of H+ in equilibrium with
          comp


the measured activities of calcium and bicarbonate.  Substitution of 2-15




and 2-17 into 2-14 yields
                                            (H+)
                                                measured





If the degree of oversaturation for siderite is assumed to be the same as




that for calcite, one obtains, in  similar fashion as with calcite




                      0     (Fe+2)  (HCO,") /(H+)

            <,    c   _ _I	m   3  m	S.                  (2-18)
            ^r  ~  v  ~ K   ~         K_
                      ^r           ^?



where Q   is the reaction quotient  for dissolution of siderite as  in re-




action 2-11,  again  in  terms of the measured parameters,  and K^  is the




 equilibrium constant.   It should be noted that


                 V

            v     so _ v                                         (2-19)
            jv  —     —~ **•
             T  K2     eq


 as in equation 2-lla.   Taking logarithms in equation 2-18,  one  obtains




             log Sc = log QF - log KF                              (2-20)




 If one now uses Hem's data (11)  to compute the oversaturation for cal-




 cite from 2-17 and the reaction quotient for  siderite from 2-18,  a plot




 of log S  versus log (X, should result in a straight line with a slope of

         \j             *


 unity.  The intercept at  log SQ = 0  should be equal to log Kp which is




 related to the solubility product of ferrous  carbonate by 2-19.




        Figure  2-6 is a  plot of Hen's data  on  the log SG - log QF coordi-




 nates.  Two lines are shown on the graph,  A corresponding to the solubil-




 ity product obtained in this  study,  and B  corresponding to the existing constant

-------
                                                                   2-18
    1.6
    0.8  —
cr*
    0.0
   -0.8
   -1.6
                                   from experimental;study / O  /
            -3.2

        FIGURE 2-6.
-2.4     -1.6      -0.8
             LOG S
Conformance of experimental solubility product to
data obtained by Hera (13) in natural grourjdwaters.
S is the degree of oversaturation with respect to
calcite and Q  is the reaction quotient  for  the
dissolution of siderite.
      4.0
          FIGURE 2-7.  Experimental data  for  the determination  of  the
                       solubility product of  ferrous  carbonate.

-------
                                                                      2-19






reported in the literature,  (In plotting these lines, a value of 10.33




was used for pK , at 25 C and zero ionic strength.)  Points 1 and 2,




which fall considerably off both lines, represent waters in contact with




rock formations which are quite low in carbonate content and where pH was




not measured directly at the time of collection.  Due to the low concen-




trations of carbonate in these waters, they are weakly buffered with




respect to pH so that pH may have changed considerably during storage.




(Hem suggests that the initial pH of these two waters could have been half




a unit or more higher than the pH which was measured in the laboratory




after storage.)  It is seen in Figure 2-6 that the experimental solubility




product obtained in this study conforms well to Hem's data.




       Ghosh, O'Connor, and Engelbrecht (19) have indicated that their




apparent oversaturation may partially be explained by their inability to




measure the pH in the aquifer at the depths from which the waters were de-




rived.  This redetermination of the solubility product additionally explains




their observations.




       The solubility product of siderite was also determined under dif-




ferent experimental conditions by varying the temperature of the system




and the concentration of the constant  ionic medium.  Figures 2-7 and 2-8




show the results at two other temperatures, and the findings at an ionic




strength of 0.05 are presented in Figure 2-9.  The data were treated in




the same manner as above and the outcome is summarized in Table 2-2.   It




will be noted that the solubility of FeCO, increases with decreasing tem-




perature.  The solubility products obtained in this study under various




experimental conditions are seen to be consistent among themselves and  are




approximately three times greater than the accepted value reported in the




literature.

-------
T
                                     I      r

-------
                                                                  2-21
Table 2-2.  Experimental Determination of Solubility Product  of  FeC03
Temperature
     C
            Ionic
           Strength
                       pK
                       ^
                         30
                               Experimental
                               temperature
                                               Thermo dynamic
                                              Solubility  Product
                                                     Corrected to
                                                     25°C, I = 0.0

17
22.5
30
22

Accepted

0.
0.
0.
0.


1
1
1
,05

Literature

-0.
-0.
-0.
-0.


72
57
46
,42


10.12
10.21
10.25
10.28

Value (5)

10.
10.
10.
10.

• i
21
24
20
31

10.68

6.
5.
6.
4.

2.

2 x
7 x
3 x
8 x

,1 x
„-!!
10
10"11
io"u
10"11
o-ll
10
- • —
      ^Activity corrections  were  made using the Davies equation
     -log tf
                    =   Az
                                              -0.31
for single ion activity  coefficients.

relationship

        K?   -AH!
          " ~  R
                                             The van't Hoff temperature
           i
           ln
       was  used  to  convert  the  experimental solubility products to 25°C.

       (AH° = -4630 cal./mole at.  25°C  (5).)  Sample calculations are given

       in Appendix A.

-------
                                                                       2-22
X-Ray Analysis of Precipitate



        The crystal structure of the precipitate was examined by x-ray



diffraction to establish whether the solution was in contact with an



amorphous deposit or with crystalline ferrous carbonate, i.e., siderite.



The diffraction pattern is  shown in Figure 2-10 along with a table de-



scribing the standard pattern for siderite given by the American Society



for Testing Materials (20).  The glancing angle, "6y corresponds to the



interplanar spacing, "d", and I/I-, is a measure of the relative intensity



of any single peak to the largest peak.  For example, at 26 = 40.6 , the



largest peak is obtained so that I/ 1. = 1007o, while the intensity of the



peak at 29 = 68.0   is only  45« of the intensity at 40.6°.  Comparison of



the diffraction pattern of  the precipitate with the ASTM standard shows



definitely that the precipitate was  crystalline ferrous carbonate so that



the solubility product obtained is the thermodynamic solubility product



of siderite.





Stability Constant  of FeHCO_+



        Using the divalent  cation electrode  and standardized  solutions of



ferrous perchlorate, one obtains a standard  curve relating the potential,
            *

E _ , to the concentration  of free ferrous iron, Fe  ,  in the constant
 DCE


ionic medium.  (See Figure  2-11.)  At concentrations of Fe    below



10~ M, E .,   approaches a limiting value due  to  selective exchange of Na
 which,  at that point,  is present  at  a concentration  three orders of mag-

                        t)

 nitude  greater than Fe  .   Consequently,  in  the  experimental study in


                                                       -4
 0.1M NaCIO, ,  a concentration of Fe(ll) in excess of  10 M was always



 employed.

-------
CO
W
i;
11
 40






 20 -.






  0


100






 80






 60





 40 —





 20  __






  0
                    FIGURE 2-10.  X-ray diffraction pattern^of experimental ferrous carbonate formed

                                  in solubility study.   Comparison with diffraction pattern of siderite.
                            X-ray diffraction data for siderite (ferrous carbonate. FeCOj) (20)

                     20  -  31.3    40.6    48.8    54.0    59.1    68.0    80.1    91.4    127.1


                    I/I  -   25     100      20      21      30      45      20      20      25
-• 40

                                                                                                               -- 20

-------
                                                                        2-24
8
§
w
1-3
w
s
H
I
 i
w
&
P-i
§
1
S
            constant ionic medium
                              Fe'~ CONCENTRATION
       FIGURE 2-11.   Standardization curve for divalent cation electrode
                     in ferrous perchlorate solution.
     -30
     -40  —
     -50   —
     -60
    -70
                                     Standardization Curve
                                                   constant ionic
                                                        medium
                                Fe  CONCENTRATION
FIGURE 2-12.  Determination of free ferrous iron, Fe
              bicarbonate solution.
                                                             +2
                                                                 in

-------
                                                                      2-25


       The measured potential of each of the three systems  investigated was

plotted against the total concentration of Fe(ll)  for that  system,  deter-

mined independently by titration with permanganate, the points being com-

pared to the calibration curve.  The results are presented  in Figure 2-12.
                                                                 +2
The standardization curve represents the concentration of free Fe " cor-

responding to the given potential, determined in a similar  manner as Figure

2-11.  It is seen that the total concentration of Fe(ll) in the sample is

equal to the concentration of Fe   corresponding to the measured potential.
                                                           ry
Any deviation between the concentrations of Fe(ll)  and Fe   would imply

formation of FeHCO* or some other soluble complex of ferrous iron and

would have been  indicated had  the three points fallen to the left of the

calibration curve,  i.e.,  less  free Fe   for a given Fe(ll).  For example,

the  second  sample contained  8.3 x 10~ M of total  ferrous iron as determined

by titration with permanganate. The potential of the sample was measured

as  -51.3 mv.   But according  to the  standard curve, this corresponds to a

concentration  of free Fe   of  8.3 x 10"   M.  Therefore, all of the total

Fe(ll)  is  present as  free Fe

        Six hours later,  the  procedure  was repeated and  with the  exception

of a slight shift  in  the  calibration curve,  the  results are  identical,

 i.e.,  the  experimental  points  fall  on  the calibration curve.  One must

                                               -2
conclude  that  even  in the presence  of  1.1 x 10    eq./l. of  alkalinity,
                                                                        f\
there are  no other  measurable  soluble  species  of  Fe(ll) besides  free Fe

        From these results and  the limitations  imposed by the  experimental

 technique,  it  may be assumed that the concentration  of Fe    is more than
                                              f\
 ten times  greater  than that  of FeHCO * at 10~   eq./l.  of alkalinity.  This

-------
                                                                     2-26


implies that the equilibrium constant (stability constant)  is less than

10 and the reaction is of no significance in natural waters.   (In more

dilute systems, of ionic strength less than 0.1, the stability constant

should be even smaller.)

       If FeHCO-  had been significant, one would have expected curvature

in Figures 2-5 and 2-7 to 2-9, the degree of curvature being  a function of

the concentration of HCO- .  As already seen, these data plot well as

straight lines.

       Hem (14) and Morgan (21) have investigated complex-formation of

Mn(II) by bicarbonate and found the solubility of Mn(ll) to be influenced

by such complexation.  For the reaction


            Mn"*"2 + HCO " = MnHC03+                                 (2-21)

Morgan reported an average thermodynamic equilibrium constant of 81,

while Hem found an average value of 63, indicating that 357* of the total
                                                                      ^
Mn(II) would be present as MnHCO-  in a groundwater containing 5 x 10

eq./l. of alkalinity.  However, no such complex of bicarbonate with

Fe(Il) was observed using the direct approach described above employing

the ion-sensitive electrode.


Summary of Experimental Study

       It can be concluded that the solubility product of ferrous carbo-

nate, which is based upon experimental data obtained 50 years ago, is in

error by a factor of 3, Fe(Il) being three times more soluble than the ac-

cepted value would predict.  The re-determined  solubility product accounts

for the recent, reports of apparent oversaturation of natural groundwaters

with respect to siderite.  The existence of a bicarbonato-complex of ferrous

-------
                                                                     2-27







iron to partially explain increased solubility of  Fe(Il)  has been  dis-




counted, the only soluble species of Fe(Il)  of any significance  in car-




bonate-bearing waters being free ferrous iron, Fe   .









2-3  Solubility of Ferric Iron




       2-3.1  Solubility in Natural Waters






        In oxygenated waters, ferric iron, in the + III oxidation state,




is the  stable form of  iron.  (Its rate of formation via the oxidation of




Fe(ll)  is discussed  in Chapter  3.)  Due to  its relatively great insolu-




bility,  ferric  hydroxide,  or ferric oxide-hydroxide, controls the concen-




tration of  soluble Fe(IH) in natural  waters.  Various structural forms




of insoluble ferric  hydroxide  are known to  exist  having  solubility products




 ranging from 10~35'5 to 10~44(22).  In the  experimental  study of  the kinetics




 of hydrolysis of Fe(lll) which is presented in  Chapter 4,  a solubility




 product of 10"38 was determined for freshly-prepared ferric hydroxide.




 For illustrative purposes, this value will  be used here.  Figure  2-13  is




 a solubility diagram for Fe(IIl) utilizing the equilibrium data presented




 in Table 2-3.  For the  sake of convenience, the simple case has been




 assumed in which Fe(OH)3  is in equilibrium only with its monomeric soluble




 hydroxo-ferric complexes, the presence of  multimers and other complex-




 formers, such  as silicate, sulfate, etc.,  being neglected for the time being.




         In waters containing relatively high concentrations of phosphate,




  insoluble ferric phosphate, FeP04> becomes operative in limiting the




  solubility  of  Fe(lll).  For a  water containing a total  concentration of




  all  phosphate  species of  10"4M,  Figure 2-14 demonstrates  that the influence

-------
                                                                     2-28
      -2 —
  §
  M


  I
  B   -6 -
  z
  8
  H
  M
  H
  >-^
  O
-8 —
     -10 —
             2468

                                      PH
         FIGURE 2-13.  Solubility of ferric iron.
    -4
    •«   -
I
o
^  -8   _
M
M
M
   -10  —
   -12
                                      FeP04 + 3 OH  -
        FIGURE 2-14.
                Solubility of ferric iron in the presence of 10  M
                of total phosphate species.

-------
                                                                     2-29
          Table 2-3.  Equilibria Describing Fe(lll) Solubility
Equation
No. Reaction
2-22 Fe(OH)3(g) = Fe+3 + 30H~
2-23 Fe+3 + H20 = Fe(OH)"1"2 + H+
2-24 FeOH1"2 + H^ = Fe(OH)2+ + H+
2-25 Fe(OH)3(g) + OH* = Fe(OH)4~
O ^
2-26 FeP04(g) = Fe+ + P04"
2-34 H3P04 = H2P04" + H+

2-35 H2.P04~ = HP04~2 + H+
2-36 HP^"2 = P04~3 + H+
Equilibrium
Constant
at 25°C
io-38
6.8 x IO"3
2.6 x IO"5
10~5
10"24
7.4 x IO"3
-8
6.4 x 10
5.0 x IO"13
Reference
expt'l, Ch. 4
23
23
27
25
22

22
22
of solid FeP04 is exerted only in the acidic pH-region below pH  5.  The



solubility product of FePO, is not a well-known quantity,  there  being


                                           -17 9
three different values for the constant: 10   '  (computed from  the tabula-



tion by Latimer (5)), 10"21'9(24), and IO"24 (25).  Again, for illustra-



tive purposes, the value of 10~   determined by Stumm and Galal-Gorchev (25)



has been utilized.



       One can derive an expression for the conversion of FePC>4 to



Fe(OH)_ in a  similar manner as was done for the system FeCO^ - FeS in



equation 2-10.  In this case,
            FeP04(s)  +  3 OH"  = Fe(OH)3(s)
(2-27)

-------
                                                                      2-30
where the equilibrium constant is

                  K0,     ..   [P0.~3]
            K   =^i=1014= 	1—                              (2-27a)
             **   K22           [OH"]3

For a system at pH 6, the total concentration of phosphate must exceed

2 x 10~ ti in order for solid FePO  to control the solubility of Fe(lll).

This is an unlikely situation in most natural systems but under localized

conditions where the composition of the water is non-uniform, FePO,  may be

influential.  Generally, however, the solubility of Fe(lll) is controlled

by its various oxides and hydroxides.


2-3.2  Effect of Complex Formation on Fe(lll) Solubility

       The presence of organic and inorganic ligands which are capable of

coordinating with Fe(lll) to form soluble complexes serves to increase the

solubility of Fe(lll)  in natural waters.  In contrast to the case of ferrous

iron where the tendency to  form complexes is insignificant, ferric iron has

a  strong  affinity for complexing ligands.  In the preceding section where

the solubility of Fe(OH), and FePO, were considered, the influence of
                                                                     o
complex-formation was neglected for reasons of simplicity.  Since Fe   has

an exceedingly strong affinity for the hydroxide ion, the relative affini-

ties of Fe+  for other ligands must be compared to its affinity for OH  to

evaluate  the extent of coordination of Fe(lll) by these other ligands.

Consequently, the relative  concentrations of the various complexes of

Fe(III) are pH-dependent.   This fact  is demonstrated in Appendix B where

it is shown that in the presence of phosphate, a rather strong complex-

former, the effect of  soluble phosphato-complexes of Fe(lll) is significant

only in the acidic pH-range below pH  4.  (It is probable that mixed

-------
                                                                      2-31




hydroxo-phosphate-complexes of Fe(lll) exist but there is insufficient


thermodynamic data to calculate their relevance.)


       It is apparent that, in natural waters, the major effect of ligands


other than OH~ is manifested in the acidic pH-range where the concentra-


tion of OH~ is inconsequential.


       A number of organic agents have a strong tendency to coordinate


with Fe+ , examples including EDTA and citrate (22).  Again, the existence


of mixed organo-hydroxo-complexes is likely.  Although Figure 2-13 implies

                                                          —8
that the concentration of soluble Fe(IIl) cannot exceed 10  M in the pH-


region 6 to 11,  significantly higher concentrations of soluble Fe(IIl) in


natural waters have often been reported.  Complex-formation with organic


material is usually cited  as  an  explanation,  Morgan (26) has considered a


hypothetical  system  involving nine metals and nine  ligands to demonstrate


the  significance of  complex-formation,  and  has  found that for the types of


ligands observed in  natural waters,  OH~ is  the  major ligand coordinated


with ferric  iron. The  discrepancy between  predicted concentrations of


soluble Fe(lll)  and  reported  concentrations can be  partially  explained by


the  analytical  difficulties  encountered in  distinguishing between soluble


Fe(IIl)  and  suspended colloidal  ferric  hydroxide.   Lengweiler,  Buser, and


Feitknecht (27), in  order  to  completely sediment colloidal  Fe(OH)_, demon-


strated  the  need to  resort to ultracentrifugation.  Hence,  it is  doubtful


that conventional methods  of  filtration are effective  in differentiating


between  soluble and  suspended Fe(lll).


       Furthermore,  as  Morgan (26) has  shown,  the concentration of organic


matter in natural waters is  insufficient to account for significant com-


 plexation of Fe(lll).  However,  in view of  the extremely high concentrations

-------
                                                                      2-32
of "solubilized" Fe(Hl) associated with organic color in natural  waters,

it has been suggested (28) that these color -causing organic  agents co-

ordinate with colloidal Fe(OH>3 forming a highly-dispersed peptized

colloid.

2-3.3   Experimental Determination of Sulfato -Complex of Fe(IIl)

       As  indicated in  the previous section, the influence of inorganic

ligands other than OH"  in coordinating with Fe(lll) is insignificant except

in the acidic pH-range, or in  the presence of relatively high concentra-

tions of the competing  ligand  compared to OH~.   In  the case of sulfate,

both these conditions  are fulfilled in the  acidic waters draining through

coal and copper mines  where  oxidation of sulf ide minerals releases  large

 concentrations of sulfate.   (Chapter 5  contains a  complete  discussion of

 the chemistry characterizing mine drainage.)   In these waters,  where  con-

 centrations of sulfate exceed 10"^ and PH-values  less than 3  are not un-

 common, complex-formation of Fe(III) by sulfate appears to  be  interrelated

 with the oxidation of ferrous iron and the hydrolysis of ferric iron.

        The stability constant for the reaction
                                                (FeSO/)
                                            .
  is fairly well-known (22), having been determined mainly by spectrophoto-

  metric techniques.  Potentiometry can also be conveniently applied to

  measure  such  stability constants (29, 30).  Since the rate of hydrolysis

  of ferric  iron  (Chapter 4) was  to be studied using a potent iometric method,

  the  investigation of complex-formation between  sulfate and Fe(III) served

  as a preparatory exercise in  gaining familiarity with the technique.

-------
                                                                      2-33
Furthermore,  the experimentally-determined stability constant  could  then

be applied, as needed,  in these future studies,  some of which  were con-

ducted in the presence of sulfate.


Experimental Procedure

       The following electrochemical cell was employed in the potentio-

metric study of complex-formation of Fe(lll) by sulfate:
   Pt
Fe(ll), Fe(lll), ClO, H+, Na+,
                                                 NaCl   Hg C12   Hg   (2-29)
                                               (sat'd)

The cell consisted of a bright platinum spiral indicator electrode  and a

calomel reference electrode separated by the test solution, contact between

the latter two being effected by a solution saturated with NaCl.  (NaCl

was used in place of KC1 to avoid possible precipitation of KCIO^ in the

event of leakage of K+ from the calomel electrode.)  The redox potential

is established by the electroactive Fe(Il)-Fe(111) couple  in accordance

with the Nernst Equation
             F
             b =
                                 ,
                             (Fe+3)
                                                                      (2-30)
 A constant ionic medium of 0.1M NaC104 was maintained and the system kept

 in a const ant -temperature water bath at 25°C so that equation 2-30 can


 be written as

                                      r\

             E « E°' - 0.0592 log -£sli                             C2-30a)
 E°  referring to the standard potential at the given ionic strength and

 temperature.

        The study was conducted  in the  pH-range 1 to 3 in order to avoid

 formation of higher-order  and polynuclear hydroxo -complexes cf Fe(lll)

-------
                                                                      2-34


other than FeOH+ ,  and to maintain the concentration of Fe(ll)  constant

as free ferrous iron, Fe  .  Ferrous and ferric perchlorate were added to

a solution of NaCIO  acidified with HC10, .   Nitrogen was bubbled through

the system to remove all traces of oxygen and the system was placed on a

magnetic stirrer.  Measurement of the potential of the Fe(Il) -Fe(IIl)

couple was effected using a Heath recording potentiometer (Model number

EUA 20-11), and the concentration pH was determined in the same manner as

previously described in the study of the solubility of FeCO., employing  a

Leeds and Northrup pH meter (Catalog number 7664) .  After observing con-

stant readings for the potential of the system in the absence of sulfate,

0.5 ml  aliquots of a pre-standardized  solution of Na-SO, were added from

a microburette.  Following  each  addition of  sulfate, the potential and p H

were  recorded;  stable readings being obtained within three minutes after

the  ligand was  added.  The experimental apparatus  is shown  in Figure  2-15.


Experimental  Results and Discussion

         The experimental system  can be represented  by  the following chemical

equilibria:
              Fe+3  +    SO.'2 = FeSO/;     K.  = - -r -           (2-28a)
                         4         4         l
                                                [H+][S04"2]

              HSO" = H+ + SO '*;           K   = -            (2-31)
                 44             a     [HS04']
                                                                     (2-23a)
                                                   [Fe  ]

-------
                                                                 2-35
                    3 ml. BURETTE
    NITROGEN DIFFUSER
   PLATINUM SPIRAL
 INDICATOR ELECTRODE

 CALOMEL REFERENCE
     ELECTRODE	^
     TEST
 SOLUTION'
                  GLASS ELECTRODE
                        CALOMEL REFERENCE
                            ELECTRODE
                             I    I    I
                                                           STIRRING
                                                             BAR
REACTION VESSEL WITH WATER JACKET
TO MAINTAIN CONSTANT TEMPERATURE
          FIGURE 2-15.
Experimental apparatus for potenttoraetrie
determination of stability constant of
sulfato- complex of ferric iron.

-------
                                                                      2-36
The equilibrium constants are defined for 25 C and an ionic strength of



0.1.  As derived in Appendix C, the potential can be related to the total



concentration of sulfate, S_, by the equation
                                                                    (2.32)
E refers to the difference  in potential between the system in the absence



of sulfate and that  after a given  addition of  sulfate, ST»  Having measured



E and  [H*] as a function of S_, one can compute and plot the left-hand-



side of the equation versus S  , Q, being  the well-known first hydrolysis



constant of Fe   .  (0.  = 2.89 x 10*   at 25°C and  an ionic strength of 0.1
                      n


(23).)  In  the absence of higher  order  sulfato-ferric complexes other than



that given  by equation 2-28a,  a  straight  line  should result, if  [H ] is



assumed to  remain relatively constant.

                                                         ^
        If  similar studies  are conducted for a  series of p H-values, the



slope  of  the  linear  plot  in each  case should be



                         K  K

            SLOPE =  n = —i-S	                                   (2-33)

                         [H+] + Ka




Rearranging terms, one obtains




            KlKa   - K   = [H+]                                     (2-33a)
            ——      a
              n


suggesting  that  if one plots — versus [H  ], the  intercept at [H  ] = 0


                   _1

                   K,
                             n

should be equal to —,  the reciprocal of the desired stability constant.
        Figure 2-16 shows the results of two experiments conducted at  p H



 1.02 ± 0.04 and 1.39 ± 0.07.  The raw data for curve A is presented in



 Table 2-4.  The linearity seems to validate the experimental  assumptions

-------
                                                                  2-37
      5.0
~    4.0
CM
Ov
tr>
o
CM
  '
     3.0
      2.0
      1.0
0.0
                                                        B
                                        5xlO"5M Fe(III)
                                               4xlO"4M Fe(III)
                                            p H -  1.02
                             J	J
    0        5        10       15        20    -  25

               TOTAL SULFATE CONCENTRATION, xlO M

    FIGURE 2-16.  Experimental data for determination of

                  stability constant of sulfato- complex

                  of Fe(III).
                                                                  30
          10
           8
        5 4
           2   —
                             Determination of stability constant
                             of FeSO
                                    4  '

-------
                                                                     2-38
    Table 2-4.  Experimental Data and Calculations in Determination
                    of Stability Constant for FeSO^+
(1)
Volume
Added,
ml
0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
[FeCllH
(2) (3)
Sulfate P°H
Concn. ,ST
moles/1.
0 0.98
2.49xlO"3
4.97
7.45
9.91
1.233xlO"2
1.478
1.720
1.961
2.20
2.44 1.06
r~ [Fe(Hl)]T~ 4
(4)
Potential,
mv.
+513.5
509.0
503.9
500.2
494.5
490.9
487.9
485.1
483.1
480.7
478.1
x 10 ~4M
(5)
+4.5
9.6
13.3
19.0
22.6
25.6
28.4
30.4
32.8
35.4

(6)
exp (59 2^~1 ^
.191
.452
.678
1.10
1.41
1.71
2.02
2.26
2.57
2.96

(7)
6)x[l + ^
+ [H+]
.197
.466
.698
1.13
1.45
1.76
2.08
2.33
2.65
3.05

       *Potential readings versus saturated calomel  reference electrode.

       +"£  is the equilibrium potential  of the cell  in  the  absence of
sulfate.  E9 refers to the potential  after the addition of  S  moles/1.
of sulfate.  (See Appendix C.)

-------
                                                                      2-39
made in deriving equation 2-32.  The slopes of a series of such plots at



different pCH-values were calculated and Figure 2-17 has been drawn in



accordance with equation 2-33a.  As seen from the figure, — at [H ] = 0



gives — = 1.9 x 10~3.  Consequently, K  = 5.3 x 102 or 102*   .  This

      Kl

agrees quite well with the values reported in "Stability Constants" (22),



which range from 102<3° (at 25°C and 0.5M NaClO^ to 103*°2 (at 18°C and



0.066M NaCIO,).  The two values given for 20°C and an  ionic strength of



0.1, conditions which most closely resemble those of this experiment, are


in2.62    . ,02.66
10     and 10



       The slope of Figure 2-17 which  is  equal to K K   is 14.9  If the
                                                   •L ct

                                                       -2       -1.55
experimental value for K   is  used, K  becomes  2.8 x 10  or 10


                                                   -1  59
This  is an excellent  agreement with the value  of 10  *   found by Reynolds



and Fukushima  (31) at  25°C and I  = 0.1, and  serves to  verify  further  the



experimental technique and procedure.



       One can now  apply  this stability constant for FeSO,  to the mine-



water system using  the same  technique as  that  employed for phosphate-



complexes of Fe(lH)  in lakes.  At  PH-values greater  than 2,  the  magnitude


                                                       -2            -2
of K  indicates that  most of the sulfate  exists as  SO    ,  i.e.,  [SO,   ] =
     a                                                <4             t


 S  = 10~  n.   If the desired  relationships are arranged based upon the



 previously-discussed equilibria (equations 2-28a and 2-23a) ,  then, for



 25°C, I =0.1, and S  = 10*M, the extent of coordination is given as



                    [FeS04+]             [FeOH+2]     Qh



             pH      ~=                   '"
              3      lO0'7              10°'54



              2      10° '7              10-° «4

-------
                                                                      2-40
Hence, it is evident that under conditions of low pH and high concentra-


tions of ligand, as in mine drainage waters, complex-formation can be


significant.





2-4  Oxidation-Reduction Reactions of the Iron(ll)-Iron(III) System




       In oxygenated waters, it is well-known that ferrous iron is


thermodynamically unstable, being oxidized to ferric iron.  Iron undergoes


reversible  oxidation and reduction reactions dependent upon given solution


conditions.  Oxidation-reduction equilibria relate the various oxidation


states of  a certain element and are characterized by the thermodynamic


redox potential, E, in  accordance with  the  classical Nernst Equation.  Such


a systematization  is  analogous to that  of considering acid-base equilibria


as ratios between  protonated  and deprotonated bases  characterized by an


acidity  potential  defined  in  terms of the change in  free  energy for the


proton  transfer reaction.   Hence, the electron  intensity  can be treated  in


the same manner as the  proton intensity or  the  activity of  the hydrogen


 ion (32).   For example,


                             AFU

             PH = -log (H+)  =  ^J-RT       P
-------
                                                                      2-41






changes in free energy for the proton and electron transfer  reactions,




respectively.   The term p£  is a convenient measure of electron  intensity,




and when seen in conjunction with an equivalent measure of proton  inten-




sity, ie., pH,  its value can be further appreciated.




        The distribution of the various species of iron which are  stable




under given solution conditions can be conveniently summarized with the




aid of a p£ - pH diagram in which the predominant species are presented as




a function of the two master variables, as shown, in Figure 2-18.  Such a




diagram can be made to be three-dimensional by incorporating a third axis




for total concentration of soluble iron, assumed to be 10  M in Figure 2-18.




Actually, the solubility diagrams showing log Fe versus pH which were pre-




sented earlier are two-dimensional variations of such a master plot, in




which p£  had arbitrarily been made constant.




        Since the p£  -pH diagram is a theoretical one based entirely upon




reversible Nernstian  redox potentials, only well-defined systems where




the measured potential  is known to be reversible can be  interpreted in




terms of  such a diagram.  Due to the complexity of  natural  systems where




several electroactive species exist, mere  insertion of an electrode into




the water will usually  yield a mixed potential  and, hence,  cannot be re-




lated or  defined  in terms of a p£ -pH diagram.  The diagram derives its




main value  in  serving as  a guide toward  expected behavior under various




conditions  of pH  and  p£ .




        The thermodynamic  equilibria diagramatically  shown  in Figure 2-18




are  presented  in  Table 2-5  together with the  corresponding  relationship




of p£  to  pH.  The values  for  the free  energy  used  in  these  calculations




were taken  from the tabulation by Latimer  (5)  and  the experimental

-------
                                                                    2-42
 15
 10
  5  —
 -5
-10
    FIGURE 2*18.  pfc ••-• j>H diagram foreiron.  Concentration of^soluble
                  iron species is lO'^M and alkalinity is 10  eq/1.

-------
                                                                    2-43
         Table 2-5.  Equilibria for Construction of p£ -pH Diagram
Reaction
2H.O = GV ,. + 4H+ + 4
L £.\&) e
2H.O + 2 - = H_, x + 20H~
2 e Zv.gJ
Fe+2 = Fe+3 + •'
Fe+2 + H_0 = FeOH+2 + H+ + e"

Fe+2 + 2K.O = Fe(OH)* + 2H+ + e"
L i.
Fe+2 + 3H20 = Fe(OH)3(g) + 3H+ + e"
Fe(OH)0, v + H.O = Fe(OH),, , + H+ +
2(s) 2 3(s)
FeC03(g) + 3H20 -
Fe(OH)3(s)+ HC03" + 2H+ + e"
r> j-7
Fe° = Fe+Z + 2e
Fe° + HCO ~ = FeCO,, , + H+ + 2e"
J J \ S )
Fe° + 2H20=Fe(OH)2(g) + 2H+ + 2e"
Reaction
FeC03( v + 2H 0 = Fe(OH)2, , + H+
+ HC03"
Fe+3 + H,0 , FeOH+2 + H+
L
FeOII+2 + H00 = Fe(OH)0+ + H+
L L
FeC03(g) +H+ =Fe+2 + HC03-
£07volts
-1.23
-0.828
-0.771
-0.914

-1.19

-1.06
e" -0.274

-1.08*

+0.440
+0.440*
+0.048
&

+13.6*
+ 2.17

+ 4.6

-0.09*
p£-pH Relationship
p£ = 20.8-pH
P£ = -PH
,, rt , (Fe+2)

r\£ 1 ^ L. nTT 1 no— \~.$--,. 	 	

(FeOH+Z)
T\C ?n i ^i-ifi TJ-IO. , , e.
pt — cu. i-ipii-iog
(Fe(OH)2+)
p€ = 17.9-3pH-log (Fe+2)
p€ = 4.63-pH

p£ = 15.3-2pH

p£ = -9.93
p£ = -5.93-1/2 pH
p£ = -0.97-pH
pH Relationship

pH = 13.6 + log(HC03")
IT "7 1-o
ptl - L. 1 - iOg „
(FeOH+Z)
oil 4 6 1-a (FeOH+2)

(Fe(OH)2+)
pH = 0.09 - log (Fe+2) -
log (HC03~)
       *Computed from the tabulation of free energies given by  Latimer  (5).
       ^Calculated usingAF
mentally in section 2-2.3.
FeCO,
      = -160.5 Kcal./mole,  determined experi-

-------
                                                                  2-44






solubility product of FeCO_ (see section 2-2.3).   It  is  clearly  seen




that in oxygenated natural waters,  where p( -values in the vicinity  of




the upper dotted line in Figure 2-18 are observed,  ferric  hydroxide  is




the predominant form of iron except in the acidic pH-range below 4,




where the solubility of Fe(lll) increases.  In the absence of  oxygen,




as in hypolimnetic waters, most natural groundwaters,  and  in anaerobic




biological systems such as sludge digesters, ferric iron is readily




reduced by organic matter and by sulfide.  The extent of the reduction




depends upon the p£ and pH of the system, the form of the  resultant




Fe(ll) also depending upon pH.  The p£ of such systems is  in the




region of the lower dotted line in Figure 2-18.




    Upon re-exposure to oxygen, Fe(lll) ,again becomes the stable form




of iron.  The rate at which Fe(lII) is formed, however,  cannot be in-




ferred from such thermodynamic considerations and requires investiga-




tion into the kinetics of oxidation of Fe(ll).  This is  the basis for




Chapter 3.




    The p£ - pH diagram presented has been simplified in order to




demonstrate the underlying principles describing the redox reactions




between Fe(ll) and Fe(lll) in natural waters.  For complex systems,




however, the various soluble and insoluble species of Fe-S, as well




as many of the Fe-organic complexes, should be superimposed on the




diagram.  Nevertheless, such simplified diagrams help to clarify the




chemistry at work in aqueous solutions.

-------
                                                                      2-45
                               References


1)  Stumm,  W.  and Lee,  G.  F.,  "The Chemistry of Aqueous Iron," Schweiz.
       Zeits.  Hydro1.,  22,  295 (1960)

2)  Sillen, L. G., "Graphical  Presentation of Equilibrium Data," Ch. 8,
       page 227,  in Part 1,  Volume  1,  Treatise on Analytical Chemistry,
       I.  M. Kolthoff and P. J. Elving,  editors, Interscience, New York,
       (1959)

3)  Leussing,  D.  L. and Kolthoff,  I.  M.,  "The Solubility Product of
       Ferrous Hydroxide and the lonization  of the  Aquo-Ferrous Iron,"
       Journ.  Amer. Chem. Soc., 75,  2476 (1953)

4)  Gayer, K.  H., and Woontner, L.,  "The Solubility of Ferrous Hydroxide
       and Ferric Hydroxide in Acidic and Basic  Media at  25  C," J. Phys.
       Chem., 60, 1509 (1956)

5)  Ringbom, A.,   Solubility of Sulfide, Analytical  Section,  IUPAC  (1953)

6)  Latimer, W.  E., The Oxidation States of the Elements  and Their Poten-
       tials  in  Aqueous Solutions, second edition,  Prentice-Hall  Inc.,
       Englewood Cliffs, N.J. (1952)

7)  Ghosh, M. M., O'Connor, J. T., and Engelbrecht, R. S.,  "Rate of
       Precipitation of  Iron  in Aerated Groundwaters," Journ. San. Eng.
       Div.,  Proc.  ASCE,  90,  199 (1966)

8)  Hem, J. D. ,  "Some  Chemical Relationships Among Sulfur Species .and
       Dissolved Ferrous  Iron," U. S. Geol. Surv. Water Supply Paper
       1459-C, Washington (1960)

9)  Srnith, H.  J., "On  Equilibrium  in the System: Ferrous Carbonate,
       Carbon Dioxide  and Water," Journ. Amer. Chem.  Soc. 40, 879
       (1918)

10)  Kelley, K. K.,  and Anderson, C.  T.,  "Contributions to the Data on
       Theoretical Metallurgy IV Metal  Carbonates-Correlations and
       Applications of Thermodynamic Properties,"  Bulletin  384, U. S.
       Bureau of Mines,  Washington  (1935)

11)  Hem,  J. D.,  "Restraints on Dissolved Ferrous Iron Imposed by Bi-
       carbonate, Redox Potential,  and pH," U.  S.  Geol. Surv. Water
       Supply Paper 1459-B, Washington (19601

12)   Stumm, W., and Singer,  P. C.,  "Precipitation of Iron in Aerated
        Groundwaters," discussion,  Journ.  San. Eng.  Div., Proc. ASCE, 92
        120 (1966)

-------
                                                                       2-46


13)  Larson, T. E., "Oxidation of Metals and Ions in Solution," p.  433  in
        Principles and Applications of Water Chemistry,  S.  D.  Faust and
        J. V. Hunter, editors, John Wiley and Sons, Inc., New York (1967)

14)  Hem, J. D. , "Manganese Complexes with Bicarbonate and  Sulfate in
        Satural Water," Journ. Chem. Eng. Data, 8, 99 (1963)

15)  Sttmm, W., and Lee G. F., "Oxygenation of Ferrous Iron," Ind.  En&.
        Chem., 53, 143 (1961)

16)  Leey G. F., and Stumm, W., "Determination of Ferrous Iron in the
        Presence of Ferric Iron," Journ. Amer. Wat. Works Asscn. 52,
        1567 (1960)

17)  Schindler, P. W., "Heterogeneous Equilibria  Involving Oxides,
        Hydroxides, Carbonates, and Hydroxide Carbonates,"  Ch. 9, p. 196
        in  Equilibrium Concepts in Natural Water  Systems, R. R. Gould,
        ed., Advances in  Chemistry Series 67, Amer. Chem. Soc., Washington
        (1967)

18)  Harned, H. S. ,  and Scholes,  S. R. ,  "The  lonization Constant of HC03
         from 0 to  50°," Journ. Amer. Chem.  Soc.,  63,  1706 (1941)

19)  Ghosh, M. M.,  O'Connor,  J. T.,  and Engelbrecht,  R. S., "Precipitation
         of Iron in Aerated Groundwaters," closure of  discussion, Journ.
         San. Eng.  Div.,  Proc. ASCE,  93,  118 (1967)

20)  Amer.  Soc. Testing Materials,  X-Ray Powder  Data  File,  Special Tech-
         nical Publication No. 48-L,  ASTM (1962)

21)  Morgan, J. J., "Chemistry of Aqueous Manganese II and  IV," Ph.D.
         Thesis, Harvard University  (1964)

22)  Sillen, L. G., and Martell,  E.  A.,  Stability Constants of Metal-Ion
         Complexes,  Special Publication No.  17, London, The  Chemical
         Society (l964)

 23)  Milburn,  R.  M., "A Spectrophotometric  Study of the Hydrolysis of
         Iron III Ion.  Ill Heats and Entropies of Hydrolysis," J. Amer.
         Chem.  Soc., 79,  537 (1957)

 24)   Zharovskii,  F. G.,  Trudy Kowissii Anolit Khim Akod Nank SSSR,  3,
         L01 (1951)

 25)   Galal-Gorchev, H.,  and Stumm,  W. , "The Reaction of  Ferric Iron
         with Ortho-Phosphate," Journ.  Inorg.  Nucl. Chera.,  25, 567  (1963)

 26)   Morgan, J. J., "Metal-Organic Complexes," paper presented at  Univ.
         of Alaska Symposium on Organic Matter in Natural Water,  Sept.
         2-4,  1968, Fairbanks, Alaska

-------
                                                                      2-47
27)   Lengweiler,  H.,  Buser W.,  and  Feitknecht, W.,  "Die Ermittling Der
        Loslichkeit von Eisen (III)  -  Hydroxiden Hit  59 Fe," Helv. Chim.
        Acta,  44,  pp.  796 and 805 (1961)

28)   Stumm, W., "Metal Ions  in Aqueous Solution," p.  520  in Principles
        and Applications of  Water Chemistry,  S. D.  Faust  and J. V. Hunter,
        editors,  John Wiley  and Sons,  Inc., New York  (1967)

29)   Willis, R. L. S., "Ferrous-Ferric Redox  Reaction in  the Presence
        of Sulfate Ion," Trans. Farod. Soc.,  59,  1315 (1963)

30)   Matoo, B. N., "Stability of Metal Complexes  in Solution.  III. Ion
        Association  in Ferric Sulfate  and Nitrate Solutions at Low Fe III
        Concentration," Zeits.  for  Phys.  Chem. Nerre  Folge, 19, 156  (1959)

31)   Reynolds, W.  L. ,  and Fukushima, S.,  "Iron  (II) and Iron (III) Isotope
        Exchange  in Presence of Sulfate Ions,"  Inorg. Chem.,2, 176  (1963)

32)   Stumm, W., "Redox Potential as an Environmental  Parameter, Conceptual
        Significance  and Operational Limitations,"  Proc.  Third Intl. Conf.
        Wat. Poll. Research, Munich,  Sept. 1966

-------
                                CHAPTER 3







                 KINETICS OF OXYGENATION OF FERROUS  IRON







3-1  Introduction







       Ferrous iron is thermodynamically unstable in the  presence of




oxygen.  The rate at which Fe(ll) is converted to Fe(lll) cannot be  in-




ferred from thermodynamic data but requires a thorough investigation  of




the kinetics of the oxidation, i.e., the mechanism by which the reaction




occurs and the various factors which influence such a mechanism.




       Conventional water treatment for the removal of iron consists  of




aeration of the raw water followed by  sedimentation and filtration.   The




former process allows for the escape of C0?) thus raising the pH,  and for




the introduction of oxygen which oxidizes Fe(ll) to Fe(lll).  The latter




hydrolyzes to form a precipitate which is  subsequently removed by sedi-




mentation and filtration.




       In natural waters, the cycles of phosphorus  and sulfur are inter-




related with the  iron cycle.  The  rate of  oxidation of ferrous to ferric




iron  during the  spring and  fall  overturn is, therefore,  partially respon-




sible for the rate of dissolution  and  oxidation  of  sulfide and for the




rate  of turnover,  of phosphate.




       This chapter  is comprised of  experimental studies of the rate of




oxygenation of  ferrous  iron in  the neutral pH-range encountered in natural




groundwaters, of heterogeneous  oxygenation in  the presence of precipitating

-------
                                                                    3-2





ferrous carbonate,  and of oxygenation in the acidic  pH-range  typical of



drainage waters from coal mines.   A model is presented  depicting  the



rate of oxidation of Fe(Il) over  the entire range of pH encountered in



natural waters.  The model is shown to be compatible with existing the-



ories describing the mechanism by which Fe(ll) is oxidized, and  its



characteristics and implications  are explained in accordance  with modern



kinetic theory.





3-2  Oxygenation of Ferrous Iron  at Neutral pH-Values



       3-2.1 Oxidation in Natural Groundwaters




       The deferrization of iron-bearing waters is customarily accom-



plished by oxidation of the ferrous iron in the raw water to  insoluble



ferric hydroxide utilizing dissolved oxygen.  The reactions describing



the process are




            Fe+2+  1/4 0£ + E+ = Fe+3 + 1/2 H^                     (3-D









            Fe+3 + 3H00 = Fe(OH)-, , + 3H+                         (3-2)
                     2          Jv. s;



It is well-established that the oxidation of Fe(ll) in the neutral  pH-



range proceeds at a rate which is dependent upon the concentrations of



Fe(ll), dissolved oxygen, and OH  , as shown in Table 3-1, in accordance



with the rate expression






            -dCFe(II)]  = k[Fe(ii)] Pn   [OH~J2                     (3-3)

                Qt                   (Jy




The rate constant k is reported by Stumm and Lee(3) to be 8.0 (1 2.5)  x


  13      2     -2    -1    -1
10   liter  mole   atm   min  ;   their results were derived for the

-------
                                                                    3-3
            Table 3-1.   Kinetics of Oxidation of Ferrous Iron
        Rate Equation                              Reference
-dCFe(ll)] = k[Fe(H)][0 ]
	   	£_                       Just  (1)

    dt         Cco232




           = k'[Fe(H)][Oj[OH~]2                Just  (1)
      1  = k[Fe+2][0_]                           Holluta and Eberhardt  (2)
                    _
   dt               2



where k  = function of pH
 -d[Fe(ll)]  = k[Fe(Il)][0.][OH"]2                Stumm and Lee
    at                    /
 pH-range 6.0  to  7.5.   Just  (1)  and  Stumm  and Lee (3) carried out their



 studies in  bicarbonate buffer systems,  the  latter work being conducted



 under conditions such that  interference by  precipitation of ferrous car-



 bonate was  precluded.  The  introduction of  Fe(lll)  at concentrations up



 to 10~St had  no  effect on the rate  of oxidation  (3).  The marked pH-



 dependence  of the reaction  rate should be noted; a  100-fold increase in



 rate was observed for each  increase of one  pH  unit.



        Ghosh, O'Connor, and Engelbrecht (4) conducted field studies of



 the rate of oxidation and removal of Fe(ll) from natural groundwaters



 at eight water treatment plants in  Illinois.   Their results corroborated



 the  first-order dependence of the reaction  rate  on  [Fe(ll)], but only a

-------
                                                                    3-4





remote relationship was observed between the rate and  the  pH of  the  system.



However, since the reported variation in pH among the  eight studies  was



only 0.3 pH units, one cannot justifiably conclude that  pH is  insignifi-



cant.  On the other hand, a definite correlation was noted between alka-



linity and rate of oxidation.  Stumm and Lee (3) also  noticed,  in addi-



tion to the second-order dependence on [OH ], that the reaction proceeded



at a slower rate in solutions of low alkalinity.  This discrepancy was



attributed to the slow response of the KCO ~ - CO  buffer  system to



localized changes in acidity brought about by the oxygenation  reaction,


                                             -2
or to possible base catalysis by HCO   or CO.,   .  Apparently,  higher



alkalinities  are associated with higher rates of oxidation.



       The actual rate of removal of Fe(ll)  from natural waters (4)  was



found to be  approximately one order of magnitude less than predicted by



the  studies  of oxidation in  synthetic systems (3).  Ghosh, O'Connor, and



Engelbrecht  (4) observed removal of Fe(Il) both by oxidation to insoluble



Fe(OH).  (reactions  3-1 and 3-2) and by precipitation as FeC03, in accord-



ance with




             Fe+2 + HC0" = FeC0.   + H+                           (3-4)
Despite the dual mechanism for removal, the rate was still less than



the  predicted rate.  Two  explanations were given (5) for the slowness of



the  reaction  in nature.   Firstly, the field studies were conducted in



waters of  alkalinity three to five times  less  than the laboratory studies.



Since the  rate of  the reaction is apparently accelerated by alkalinity,



the  reaction  should proceed  at a slower rate in the Illinois waters.



Secondly,  the presence of catalysts or  inhibitors present in the natural

-------
                                                                    3-5
system could account for variations not  only  between the field and

laboratory studies,  but also among the different groundwaters them-

selves.  Organic substances and sulfide  compounds,  in particular, were

cited as inhibitory agents (5) (6) in accordance with the  sequence


            Fe(ll) + 1/4 02 + ORG. = Fe(IIl)  - ORG. COMPLEX       (3-5a)


            Fe(lll) - ORG. COMPLEX = Fe(Il) + OXIDIZED  ORG.       (3-5b)


            Fe(ll) + 1/4 0  + ORG. = Fe(lll)  - ORG. COMPLEX       (3-5a)


In this scheme, the ferrous-ferric system functions as an electron-

transfer catalyst for the oxidation of organic material by oxygen.   The

overall rate of oxidation of  Fe(Il) may be retarded depending upon  the

rate of oxygenation of  Fe(ll)  in  the presence of organic matter in com-

parison to  the  rate of  reduction  of Fe(lll)  by  the organic matter.

       The  influence of such  factors  as these need to be quantitatively

evaluated before  the actual rate  of oxidation of Fe(ll) in natural

systems can be  predicted and  before an  efficient scheme for iron removal

can be designed.



 3-2.2 Oxidation  of Fe(ll)  in the Presence of  Ferrous  Carbonate Over-
        saturation


        It was indicated above that deferrization of  iron-bearing waters

 may be achieved by precipitation of Fe(ll) as FeCO , as well as by  its

 oxidation to Fe(OH)..   In fact, Hale (7)  effected  satisfactory removal

 of iron under anoxic conditions by the addition of hydrated  lime in a

 closed system to precipitate FeCO- along with

-------
                                                                     3-6


       For a groundwater previously in equilibrium with siderite  (FeCO-,

aee section 2-2.2),  aeration serves a dual purpose.   In addition  to

introducing oxygen for oxidation of Fe(ll), aeration allows  dissolved

carbon dioxide, with which the groundwater is oversaturated,  to escape.

Consequently, the pH of the system increases and the water becomes pro-

gressively oversaturated with respect to ferrous carbonate.   If the  de-

gree of oversaturation becomes such that the energy barrier  to nucle-

ation is overcome, crystallization of FeCO- takes place and  precipita-

tion follows.

       In this respect, it should be of interest to measure  the rate of

oxidation of Fe(ll) under the influence of oversaturated conditions

favoring precipitation of FeCO..


Experimental Procedure

       A known gas mixture of carbon dioxide and oxygen was  bubbled

through a series of flasks containing sodium bicarbonate, mounted on

magnetic stirrers.  The CO  - HCO ~ system buffers the solution with

respect to pH.  After the attainment of equilibrium, observed by  a con-

stancy in pH, various amounts of a stock solution of ferrous perchlorate

were added to the bicarbonate solutions.  (The stock solution of  Fe(II)
                                  •/
had previously been equilibrated with the same gas mixture.)  The con-

centration of Fe(ll) added was such that the resultant solution  was  over-

saturated with respect to ferrous carbonate.  The degree of  oversatura-

tion (S = Q/K, see section 2-2.2) was computed using the new solubility

product for  ferrous carbonate determined experimentally in Chapter 2.

The rates of oxidation and removal were measured by analyzing the system

for total and  filterable ferrous iron, respectively.

-------
                                                                    3-7







       For the determination of total Fe(ll),  aliquots were withdrawn




from the system at various intervals, and immediately added to  2 ml.




of concentrated HC10, in order to quench the reaction.  Solutions of




Fe(II) at concentrations greater than 10~ M were analyzed by titration




with standardized solutions of permanganate.  For concentrations of




Fe(ll) less than 10  M, the colorimetric reagent bathophenanthroline  was




used (8).




       Filterable Fe(Il) was  determined by  immediately filtering aliquots




of the suspension through  220 mu  filter paper (Millipore Filter Corpora-




tion, Bedford, Mass.)  into 2  ml.  of concentrated HCIO^, again to stop




the reaction.  Filtration  was rapid (less than  30  seconds  for 50 ml.  of




sample)  and was  conducted  under  a partial pressure of C(>2  of one atmo-




sphere  in order  to  prevent additional oxidation and  to avoid dissolution




of any  suspended FeCO~.   The filtrate was  analyzed for Fe(ll) by the




same  techniques  as  above.



        pH and alkalinity were also measured during the course of the




reaction, the latter determined by acidimetric  titration to pH  4.3




with standardized HC1.






 Experimental Results and Discussion




        In discussing oxidation and removal of ferrous iron under con-




 ditions favoring precipitation of ferrous carbonate, previous  workers




 have tended to oversimplify  the mathematical and chemical formulations




 of the process.  Ghosh, O'Connor,  and Engelbrecht (4) combined precipi-




 tation of ferrous  iron and oxidation of ferrous iron, arriving at a




 rate of  iron removal which was first-order in concentration of Fe(ll).

-------
                                                                     3-8
Conversely, in synthetic solutions having an initial oversaturation of




10 with respect to ferrous carbonate, Morgan and Birkner (9) observed




that precipitation and removal of Fe(ll) did not conform to first-order




kinetics.  They noted an immediate rapid decrease in filterable Fe(ll)




in their  supersaturated systems, corresponding to concurrent precipita-




tion of FeCO- and oxidation of Fe(ll).  Subsequently, the rate of dis-




appearance of Fe(ll) was in exact conformance with the first-order




relationship observed  in parallel studies conducted  in the  absence of




conditions of oversaturation.  The  latter description and the rate




constants reported were in agreement with those of Stumm and Lee (3).




        In the experimental study of heterogeneous oxidation described




 here, no such simple explanation was apparent.  Figure  3-1  demonstrates




 the compliance of both total Fe(ll) and filterable  Fe(Il)  to  the first-




 order formulation, despite the fact that the system was 60  times over-




 saturated with respect to FeCOr  The fact that only a  slight increase




 in total removal of Fe(Il) above that by oxidation  alone is observed,




 is indicative of little precipitation of the carbonate.   (Precipitation




 of FeCO   is manifested by the difference between the two curves.)   It




 would appear that a period of 20 minutes was not sufficient to allow for




 crystallization of FeCC>3 at pH 6.61.  The rate of oxidation is seen to




 agree with that predicted from the  rate formulation by Stumm and Lee (3)




 for equivalent conditions of pH, partial pressure of oxygen, and temper-




 ature.



        Correspondingly, at pH 6.25,  where the rate  of oxidation is




  slower,  Figure  3-2  shows  precipitation of FeCC>3 to  become  significant,




 but  only after  40  minutes have  elapsed.  Again, the rate of oxidation

-------
                                                                3-9
10

 8
  -3
 2
10
  -4
                          Total ferrous iron
                                      Filterable ferrous iron
          Rate of
         ferrous iron
         oxidation, after
         Stumm and Lee (3)
           0.30 atm. 0
           initial supersaturation
                  S ~ 60
                   o
                            10          15
                            TIME, minutes
    FIGURE 3-1.  Oxidation and removal of ferrous iron under condi-
                 tions favoring precipitation of ferrous carbonate.
 10

  8

  6
  -3
                   Total  ferrous  iron
                                        Filterable
                                           ferrous iron
          Rate of
        ferrous  iron
        oxidation^ after
        Stumm and  Lee  (3)
                         0.50 atm. 0
                             40          60          80         100
                            TIME,  minutes

    FIGURE  3-2.   Oxidation and removal of ferrous iron under condjL-
                  tions favoring precipitation of ferrous carbonate.

-------
                                                                    3-10







of Fe(Il) parallels that reported by Stumra and Lee(3)  and  is  apparently




unaffected by precipitation of FeCO,.




       As the pH is lowered still further, the elapsed time exceeds  the




induction time required for precipitation of FeCO-  and the precipitate




seems to exert a catalytic effect on the rate of oxidation of Fe(ll)^,




(see Figure 3-3).




       To demonstrate that this seemingly autocatalytic response in




both the rates of oxidation and removal of Fe(ll) is a. function of the




induction time for nucleation of ferrous carbonate, i.e.,  the time re-




quired to overcome the energy barrier preceding nucleation of the cry-




stalline phase, the studies at higher pH were repeated but the experi-




mental conditions were modified to  decrease the rate of the  oxidation




reaction.  Figure 3-4 resembles Figure  3-3, again  showing a rapid decay




in both  total and filterable Fe(Il)  after their conformance to the ac-




cepted first-order relationship for  the first 60 minutes.




       These  studies reflect the complex nature of heterogeneous reac-




tions.   For the case in question, the system consists of oxidation of




dissolved  ferrous iron, precipitation of ferrous iron as the carbonate,




heterogeneous oxidation of  solid ferrous  carbonate, and possible surface-




catalysis  of  the oxidation by  ferrous carbonate.   To represent such a




system by  a  simple relationship would indeed be a  mistake.




       Although no conclusions can  be drawn from this study, there are




some pertinent points worthy of consideration:




             i)  Supersaturation with respect  to  some  solid phase does




not imply  that precipitation  takes  place  immediately.  The magnitude of




 the activation energy barrier  to  the nucleation  process is inversely

-------
                                                                    3-L1
                                            Total Fe(II)
                              Filterable Fe(II)
                Rate  of Fe(II)  oxidation,
                      after  Sturara and Lee  (3)
          p H•- 5.84
          0.50 atm.  0
                            100        150
                              TIME, minutes
200
250
FIGURE 3-3.  Effect of FeCO^precipitation on Fe(II) oxidation and removal.
      2
     10
         0           25   "      50          75         100         125
                                 TIME, minutes

FIGURE 3-4.  Effect of FeCO, precipitation on Fe(II) oxidation and removal.

-------
                                                                    3-12
proportional to the supersaturation,  i.e.,  the activation energy de-




creases as the supersaturation increases.  Consequently,  the rate of




nucleation is a function of the degree of supersaturation,  there being




a critical supersaturation value below which nucleation is extremely




slow and above which nucleation is rapid.  Therefore,  the induction




time, i.e., the time required for formation of the critical-sized cluster,




decreases as the supersaturation increases (10).




            ii)  Precipitation of ferrous carbonate serves as a mechanism




for removal of Fe(ll), complementing removal by oxidation and hydrolysis.




The rate of removal, however, cannot be described by first-order kinetics.




            iii)  Ferrous carbonate appears to play a catalytic role  in




the oxidation  of Fe(ll).  The mechanism  for such an effect is uncertain




but could  conceivably be attributed to a specific surface reaction where-




by solid  ferrous carbonate  provides active  sites at which the concentra-




tion of Fe(Il) is greater than  in bulk solution, or sites at which the




reaction  is favored.  Although  it has not been demonstrated, one can




imagine that precipitation  of ferrous carbonate could, under certain




circumstances,  inhibit the  oxidation of Fe(ll) by lowering the concen-




tration of free ferrous  iron in  solution, or by decreasing the available




Fe(ll)  exposed to oxygen with the remainder being incorporated in the




interior  lattice of the  ferrous  carbonate crystal.  Such inhibition was




observed by Morgan  (11)  in  his  study of Mn(ll) oxygenation in the pre-




sence of  precipitating manganous carbonate.   In any case, even in the




event that oxidation  is  inhibited, the rate of removal of Fe(ll) either




by oxidation or precipitation should be  equally as great in the presence




of FeCO_  supersaturation as in  its absence.

-------
                                                                    3-13
3-3  Oxygenation of Ferrous Iron in Acidic Systems




       3-3.1  Experimental Study of Kinetics of Fe(ll)  Oxidation  at

              Acidic pH-Values
       The kinetic studies of the oxidation of ferrous iron  reported



above were confined to waters of pH greater than 6.   There are  a number




of instances in nature, however, where iron bearing  waters of pH con-



siderably below 6 are encountered.  Of special concern are those waters



in coal mining regions, where pH-values of 3 are not uncommon (see



Chapter 5).  It would be of interest to learn if iron oxidation in  such



acidic systems could be characterized by the same kinetic relationships



which describe the reaction in neutral waters.






Experimental Procedure



       The rate of oxidation of Fe(ll) was followed by measuring the




concentration of Fe(lll) with time and subtracting this quantity from



the initial concentration of Fe(ll).  An analytical procedure was



adopted whereby [Fe(lll) ] (the total concentration of all species  of



ferric iron) was determined spectrophotometrically at the isosbestic



point of an acidified solution where equilibrium had been established


                            +3                                  +2
between free ferric iron, Fe  , and monohydroxo-ferric iron, FeOH   ,



the only soluble species of Fe(lll) under the acidic conditions.  At



the isosbestic point of the system, at a wavelength of 272 mu.  both




species have identical molar absorptivities so that  for a given total



concentration of Fe(lll), regardless of the ratio <." Fe   to FeOH   ,




the same absorbance is recorded, independent of pH (see Figure  3-5).



Figure 3-6 demonstrates conformance of the absorbance of acidified  sclu-

-------
    40
                                                                    3-14
    30  	
I
to
I
    20
    10
  ISOSBESTIC POINT
      250


      FIGURE  3-5.
260           270
        WAVELENGTH,
                               280
290      300
 U-V absorbance  spectra  of acidified solutions of
 ferric perchlorate.
 o
 CO
     1.25
     1.00
     0.75
     0.50
     0.25
     0.00
absorbance readings
in 1.0 cm. cell
at 272 ni
                                      molar absorptivity

                                € » 1.55 x 103 liter/mole-
                                         cm
                4            6           8
         Fe(III)  CONCENTRATION,  xlO M
                                                                     10
     FIGURE 3-6.  Relationship between absorbance of acidified solutions
                  of Fe(III) and Fe(III) concentration, at 272 nvi.

-------
                                                                   3-15







tions of Fe(lll),  at 272 mu,  to the Beer-Lambert law.  The molar ab-




sorptivity is 1.55 x 10  liter -mole" -cm"  and is unaffected by the




presence of Fe(ll).




       The samples were prepared containing various dilutions of a




stock solution of ferrous perchlorate.  In order to determine the  de-




pendence of the reaction rate on [Fe(ll)], the studies were  performed




at constant pH and under constant partial pressures of oxygen.  The in-




vestigations at slightly acidic pH-values were conducted in  a CO^-HCO^




buffer system, as  in section  3-2.2,  in order to maintain constant  pH.




At pH-values below 5, however, the buffer capacity of the bicarbonate




system is  insufficient  to balance the acidity produced by hydrolysis of




the resultant Fe(lll)  (reaction 3-2), so  that the pH of the system tends




to be drastically  lowered.  To combat this effect,  smaller initial con-




centrations of Fe(ll)  were employed  (less than  10  M) to insure that,




for  a given percent oxidation, the corresponding change in pH would be




slight.



       At  still  lower  pH-values,  in  the vicinity of pH 3, no precautions




were necessary  since  the pH  was observed to  remain relatively constant.




 In this  range,  the pH was adjusted merely by dropwise addition of  con-




centrated  HCIO^.



        The samples were allowed  to equilibrate with the oxygen of  the




 atmosphere,  or,  in the case of the bicarbonate buffer systems, with the




 gas mixture of oxygen and carbon dioxide.  [Fe(IIl)]  was  determined by




 acidifying an aliquot with dilute HC104 in order to dissolve any  hydrous




 ferric oxide formed and then measuring its absorbance at  272 mu,  em-




 ploying a Beckman Model DU Spectrophotometer.  For the smaller range  of

-------
                                                                    3-16



concentrations, [Fe(ll)] was determined directly using the colorimetric


reagent bathophenanthroline (8).  After each reading,  the solutions were


re-equilibrated with their respective atmospheres.  The samples were


sealed and stored in an incubator at 25°C.  (In the lower pH-range, where


the rate of oxidation was observed to be slowest, a series of parallel


studies was conducted in the dark and exposed to light in order to test


for any photochemical effect on the rate of oxidation.)



Experimental Results and Discussion


       In order to describe the rate of oxygenation of ferrous iron, an


expression  similar to that of Stumm and Lee (3) was assumed, of the form



             -dfre(ll)]  = k  [Fe(ll)]mCOH-]n Pn                      (3-6)
                dt                          ««



where m and i\ are constants  to  be determined.   In the  rate  law of  Stumm


and Lee (3) at pH-values greater  than  6,  m = 1  and n = 2.   Since the


studies were conducted  at constant pH  and constant partial  pressure of


oxygen,  equation  3-6 can be  simplified to



                         = k' [Fe(ll)]m                            (3-7)
                 dt

 where


             k1  = k  [OH"]11 P0                                        (3-8)


 If the reaction were  first-order  in  [Fe(Il)],  i.e.,  if m =  1, then



             -d  log  [Fe(lD] B k«/2>3 , k.,                           (3-9)


 and a plot of log  [Fe(ll)] versus time should  be linear.

-------
                                                                   3-17




       Figure 3-7 presents some of the results obtained  in the bicar-


bonate-buffered system,  demonstrating the obedience of the data to


equation 3-9.  The concentrations of Fe(Il)  have  been carefully selected


so that the solubility of ferrous carbonate was not exceeded.  Equation


3-9 demands that the slope of the semilog plot be independent of the


concentration of Fe(Il) at any time, so that parallel  lines of slope k"


should result regardless of the initial concentration  of Fe(ll).  Figure


3-7 conforms to this requirement, too.


       At the lower pH-values where the buffer  capacity  of  the system


was low, the pH slowly declined as the reaction proceeded.  The  accom-


panying decrease  in pH was smallest for the smallest initial  concentra-


tion of Fe(ll), as planned.   In these  studies,  the course of  the reac-


tion was followed as  long as  the  pH did not differ greatly from its


starting value.   Figures  3-8  and  3-9  are  plots of log [Fe(ll)]  with time


and  also show  the corresponding  change in pH.  Conformance to the first-


order  expression  is demonstrated.   When  the studies were terminated,


approximately  25% of  the initial  concentration of Fe(ll) had been oxidized.


        Figures 3-10  and  3-11  depict the  rate of  oxidation at pH 3 and pH


 2, respectively.  These  results have also been fitted by a first-order


 rate expression,  but this has been done  only for convenience and for  the


 sake of comparison both with the above results and with those of Stumm


 and Lee (3).  For a reaction proceeding  as  slowly  as  this one does in


 the low pH-region,  where only 57» of the reaction is complete in 150 days,


 it becomes difficult to characterize the reaction with  respect to  its

                                                                       -4
 kinetic order.  The "first-order rate constant"  is approximately 1 x  10


 day"1 for both figures, even though the concentration of OH  differs  by

-------
                                                          3-18
                      Temp. 25 C

                            0.85 atm.
            25
50           75
 TIMS, minutes
100
125
FIGURE 3-7.  Rate of oxygenation of Fe(II)  in bicarbonate-
             buffered systems.

-------
                                                            -4.00
   -A.30
   -4.32  —
I
(d

O
H
O

s
-4.34 I—
-4.36 —
    -4.38
    -4.40  —
               pH constant at 4.80
                [Fe(II)J  ~ 5x10  M

                        O :
    -4.42  —





                     2           4

                     TIME/ days


 FIGURE 3-8.  Rate of oxygenation of Fe(II).
                                                            -4.02
                                                            -4.04
1

H
                                                        g   -4.06
w
u

§
o
                                                     01
                                                     h

                                                     o

                                                     s
                                                            -4.08
                                                        -4.10
                                                            -4.12
                                                        -4.14
                Temp.  25 C



                PQ2 "   °'90  atm'



                pH  4*70— 4.45







                [Fe(II)]o-v 10"4M
          0           2           4

                       TIME,  days


   FIGURE 3-9.   Rate  of oxygenation  of Fe(II)
                                                                                                        u>
                                                                                                        i

-------
                                                              3-20
                   pH 3.0
                   Temp.
                   P0  « 0.20 atm.
                    [Fe(II)l  ~ 9x10  M
                             75       100
                         TIME,  days
FIGURE 3-10.  Oxygenation of Fe(II) at pH 3
                                                       125
                 150
  -3.035
u
s
   -3. 050
T
                   [Fe(II)lo -9x10  M
                   25
100
                                                125
                   50        75
                        TIME,  days
FIGURE 3-11.  Rate of oxygenation of Fe(II)  at  pH  2,
                                                        150

-------
                                                                   3-21






an order of magnitude between them.  In fact,  additional  studies  in this




acidic pH-range show the rate of oxidation to  be relatively  independent




of pH.




       To gain some additional insight as to the order  of the reaction




rate with respect to [Fe(ll)] in this pH-range, farther studies were




conducted at differing initial concentrations  of Fe(Il).   The parallel




slopes in Figures 3-12a through d also imply that the reaction  is first-




order in [Fe(ll)], but this allegation is subject to the same reserva-




tions as above.




       Because of the slowness of the reaction, the analytical  procedure




consisting of the spectrophotometric measurement of Fe(lll)  proved to be




the most effective.  However, the technique is not amenable for use over




a wide range of initial concentrations of FeCll).  At low concentrations,




the amount of Fe(lll) produced by the oxidation reaction is so  small




that the measurements of absorbance become less precise.  At higher ini-




tial concentrations of Fe(II), a sufficient amount of Fe(lII)  forms that




even at pH 3, kinetically  irreversible hydrolysis of Fe(IIl) takes place




(see Chapter 4) and it becomes increasingly more difficult to recover




all of the Fe(lll) as Fe+  or FeOH*  .  For this reason,  the order of the




reaction with respect to Fe(ll) at this  low pH could not be adequately




verified by simply varying Fe(ll) over a wide  range of initial concentra-




tions.  Figure 3-12 shows  only a three-fold variation  in  initial con-




centration of Fe(ll).  One can conclude  from these results only that the




reaction  is extremely slow and, in this  range  of concentration, can be




represented by a  rate expression which is first-order  in  [Fe(ll)3.

-------
                                                                    3-22
-3.113
                  pH 3.0

             [Fe(lQ]o = 7.78xlO~ M
-3.109
                1	L
                                     -3.209  —
                                     -3.207 —
                                     -3.205
                                                [Fe(II)]   = 6.23x10
               40        80
                TIME, days
                                  120
40        80
 TIME, days
                                                                       120
 -3.414  _
        — 07      PH 3.0
            [Fe(II)]  - 3.89x10  M
                                 __ -3.639 _
                                  	  -3.635  — —
 -3.410
                                      -3.631
                                                     40       80
                                                      TIME,  days
                                                                       120
                40        80
                TIME, uays
FIGURE 3-12.  Oxygenation of ferroXis iron at various initial^oncentratior.s
              of Fe(Il).  Solution conditions are pH 3.0,  25 C,  and partial
              pressure of oxygen of 0.20 atm.

-------
                                                                   3-23
       George (12)  studied the oxidation of Fe(ll)  in perchloric acid



media and observed a rate law of the form





            -d C!e+2] = k    [Fe+2]2[0j                          (3-10)
                at       exp          L




where the rate constant, k   , increased slightly with a decrease  in



[H+].  The relative insignificance of pH upon the rate (from pH 0  to



pH 2) is similar to that obtained in this study, but George has charac-


                                            +2 2
terized the rate as being dependent upon [Fe  ] .



       Huffman and Davidson (13) have generalized from their own results



and those of others that the rate of oxidation is first-order in [Fe  ]



in the presence of strong complex-formers, such as pyrophosphate,  fluoride,



and dihydrogen phosphate.  For complexing agents of moderate strength,

                                                                 O *7

such as chloride and  sulfate, the reaction rate depends upon CFe  ] .



However, under extreme  conditions of temperature and  concentration of



ligand, the dependence  is again  first-order.  The results of George (12)



in perchlorate media are analogous  to  those  in  the  presence of moderate



complex-formers.  Although  sufficient  data are  not  available,  it  seems



that,  in the  presence of suitable  anions,  the relative  reaction rates



for  the bimolecular  mechanism (first-order in both  [Fe   ] and  [00])



generally  parallel  the  stability constants for  association of  the ligands



with Fe+3  (13).

-------
                                                                    3-24
3-4  Oxygenation of Ferrous Iron as a Function of pH




        3-4.1  Summary of Experimental Results






        By coupling the  experimental results obtained for acidic  systems




with those obtained by Stumm  and Lee (3) for neutral waters,  one  can




plot the rate of oxygenation  of ferrous  iron over the entire pH-range




of  interest  in natural waters, as  in Figure 3-13.  The rate of reaction




has been characterized by  the rate constant, k" = -d log [Fe(ll)]/dt,




and has been adjusted for  the conditions at 25 C and a partial pressure




of  oxygen  equal  to 0.20  attn.




         If one takes  the logarithm of  equation 3-8, substituting




k"  = k'/2.3  (from 3-9)  and making  use  of the  ion product of water, one




obtains






              log k"  = log C + n  pH                                (3-11)





where £ is a constant and n is the order of the reaction with respect




 to [OH~].   It  is readily apparent  that the instantaneous slope of the




 log k" versus  pH curve in Figure 3-13  corresponds to n.  The solid line




 above pH 6 derives from the experimental rate law of Stumm and Lee (3)




 (equation 3-3)  for the given conditions, with n  = 2.  The  dotted portion




 below pH 6 is an extrapolation of  their expression  to the  acidic pH-region




 of this study;  the rate diminishes by a factor of 100 for  each unit  de-




 crease in pH.   The experimental points are compatible with the formula-




 tion of Stumm and Lee at pH-values greater than  4.5, but at  lower pH-




 values, the points sytematically deviate from the extrapolated line.   At




 pH-values below 3, the rate becomes relatively constant and is no longer




 dependent upon pH, i.e., n = 0.

-------
                                                                 3-25
    +3.0
    +2.0
    +1.0
fr-
•o
     0.0
     -1.0
     -2.0
     -3.0
     -4.0
     -5.0
     -6.0
                                  i        r
                     k" " " d
                dt

     PQ  "0.20 atm.

     Temp. 25 C

  Experimental points
obtained in this study
 O  exposed to light
 D  in darkness
                    o    o
                   "D	
                                       Extrapolation of rate law
                                    /  of Stutran and Lee (3)  at
                                  /   25 C and 0.20 atm.  of oxygen
                                       J	L
           1        2         3        4         5        67

                                      pH
          FIGURE 3-13.  Oxygenation rate of ferrous iron as a function
                        of pll.

-------
                                                                    3-26




       With regard to the relative magnitude of George's results  (12)  in


comparison with those presented in Figure 3-13, one can approximate  his


rate expression (equation 3-10) by a pseudo-first-order formulation




                     = k'   [Fe+2]                                 (3-12)
               dt       exp


where


            k'   .k    [Fe+2]  P.                                (3-13)
             exp    exp       o  0^



 [Fe+2]  is  essentially constant for only 1% total reaction (the extent of



 the reaction followed by George) and  is equal to [Fe  ] , the initial



 concentration  of ferrous  iron.  This  approximation is not a mechanistic



 one but has been made solely for the  purpose of comparison.  Under a



 partial pressure of oxygen of 0.2  atm. and at 30 C and 10  n HC10,,



 George's results predict  that the  initial value of k" (k" =» -d log


                                                            -4   -1
 [Fe(H)]/dt as in  Figure  3-13) would  be approximately 1 x 10  day   if



 £Fe  J  = 10~T1, the concentration used by George.  This pseudo-first-



 order "rate constant"  is  of the  same  order of magnitude as that ob-



 tained in this study.



        The shape of the curve suggests that  there are two parallel re-



 action mechanisms; one  operative at higher pH which can be described by



 the rate law of Stumm and Lee,  and the other functioning under more



 acidic conditions  and  independent  of  pH.  Previous investigations of



 the kinetics of oxygenation of  Fe(ll) in  acidic media  are  in  accord with



 these latter results  in that the rate of  the reaction  is relatively  in-



 dependent of pH.   In  fact, George (12) observed that his k    (see



 equation 3-10) was proportional  to [H ]~     ,  increasing only slightly

-------
                                                                   3-27







with an increase in pH.   Alternatively,  one can consider the reaction




to be independent of pH  until the concentration of OH  becomes suffici-




ently large that OH* functions as an effective catalyst of  the oxidation




reaction.




       Figure 3-13 also  demonstrates that the oxidation of  Fe(ll) occurs




more rapidly in light than in darkness;  the reaction proceeds at  a rate




2 to 3 times faster in the presence of light.  In the  acidic region,




there appears to be some photooxidation of Fe(Il) taking place.




       In further studies of the oxidation of Fe(ll) under simulated




mine conditions (see Chapter 5), the investigations were conducted  in




the absence of light to avoid such photochemical effects.








3-4.2  Kinetic Implications of Results





       Consider first the oxidation of Fe(ll)  in the acidic region




where the reaction  proceeds  at a rate independent of pH.   Weiss (14)




proposed a chain mechanism involving one-electron oxidations in order




to describe the oxidation of  ferrous  iron  by molecular oxygen.   The




suggested  sequence  is:






            Fe+2  +  02 = Fe+3 + 0^             (k^  k^)        (3-14a,b)






                                                    )              (3-15)








            Fe+2 +  H02  = Fe+3 +  HO^           (k^1)         (3-16a,b)






            HV  * H+  . H202                 CK^)             (3-17)







             Fe+2 + H20   -> Fe+3  + OH~  +  OH     (^3)                (3-18)

-------
                                                                    3-28
            Fe+2 + OH •» Fe+3 + OH"              (k                (3-19)
The molecules (*) are free radicals or reactive intermediates.   Reaction

3-14a is believed to be the rate-determining step in the sequence.   The

rate of oxidation of Fe   can be derived using steady-state approxima-
                *   •
tions (15) for H02, OH, and HO- to give (16)



            "d ®   3 = k  [Fe+2] [0] R                          (3-20)
where
                      k2  [Fe+2] [H+;
                     [Fe+2]  [H+] + k   K   [Fe+3]
                                                                 (3-20a)
Equation 3-20a implies that the oxidation reaction is inhibited  by  Fe+

due to the back reaction 3-14b.  This accounts for the slowness  of  the

reaction in acidic solutions where the resultant Fe(lll)  is  present

predominantly as Fe  .  If conditions are such that


                     [Fe+3] \» k  [H+] [Fe+2]                    (3-20b)
then R, ^ 1, and the reaction is decelerated because of the relatively

rapid reduction of Fe   by 0 ~ in comparison to the oxidation of Fe
   •       *
by 02~ or H02 (equation 3-16a).

       If, on the other hand, anions are present which are capable of

forming strong complexes with Fe  , such complex-formation serves to

decrease the concentration of free Fe+  and thus inhibit the back reac-

tion (3-14b).  The net effect is to cause the oxidation to proceed more

rapidly since
            V  KHQ  CFe+] <
-------
                                                                   3-29
and R~l.  Hence, by equation 3-20






            ~* C^+2] - k. [Fe+2] [0,]                           (3-20d)
                at       l          L



       One serious drawback of the Weiss mechanism is its unlikelihood



from a coulombic standpoint.  Zwolinski, Marcus, and Eyring (17)  termed



the formation of oppositely-charged end-products, as in 3-l4a,  as



highly improbable.   In a  later paper, Weiss (18) modified his mechanism



in accordance with such reasoning to consider that the initial reaction,




3-14a, should be the formation of an ion-pair complex



            Fe+2 + 02 = (Fe+3'02~)                                (3-21)




stabilized by coulombic attraction between the oppositely-charged part-



ners.  In this scheme, the  association  does not violate the coulombic



restriction imposed  by Zwolinski, et al.  Again, the back reaction can




be  inhibited by  suitable  anions:




            (Fe+3-02~) +  X" * (X-'Fe+3-02-)                       (3-22)



where  the resultant  complex may  eventually dissociate




            (X--Fe+3-02-) » (X--Fe+3)  + 0^                       (3-23)




The original Weiss scheme continues  with reaction  3-15.  Under these



conditions, the  rate is  proportional to [Fe   ],  [C^],  and  [X ],  and



the anionic complex-former has served in the  same  canacity  as  in the




original scheme.



        In the absence of strong complex-formers,  the (Fe  '0^  )  com-



 plex  can be stabilized by Fe   (18),

-------
                                                                    3-30




                J.  -x     +2   ,  -j-3,_ -.„ +2-V   /•„ +3._ -2._  +3..   f.  _.^
                  O  ) + Fe   = (Fe   0?  Fe  ) = (Fe   (>„   Fe  )   (3-24)





This would explain the results of George (12) and Huffman and Davidson


                                                             +2 2
(13), where the observed rate is proportional to [0^] and [Fe  3 .



The new complex is again stabilized by coulombic forces and eventually



breaks up,





            (Fe+3'02"2'Fe+3) + H+  •» 2 Fe*3 + HC>2'                (3-25)




followed again by the same sequence  as above.



       If this were the mechanism describing the oxygenation of Fe(II)



and the effect of anionic complex-formers on the rate of reaction,  then



one should observe a decrease in rate with increasing concentration of



Fe   , as in 3-20.  None of the previous workers, however, have  observed



any inhibitory effects by the addition of Fe*  to their acidic  solutions.



       The situation can be  somewhat clarified by closer scrutinization



of 3-20a and b.   If the back reaction were relatively rapid and Fe



were  rapidly reduced by 0^ , then
            V KHO  CFe] ^ k2  CH]  CFe]                     C3-20b)




A quantitative comparison of the  two terms  is called for.  For the given



experimental conditions  ([H+J = 10~3M,  [Fe+Z] = 10"3M, and [Fe+3J~10"5M)



and using the value  approximated  by Benson  (19) that pK,-o  =12-4, one



obtains k ' /k- ^  1011.  Hence, if k^ /k. <^ 10U, the back reaction is



sufficiently slow that it can be  neglected  in 3-20a.  Barb, et al (20)



experimentally measured  k.. /k« =  1.0 at  pH  2.7 in perchlorate media.



This  implies that reduction of Fe  by 0_   is signficantly slower than



the corresponding oxidation of Fe  by 0_ .  Recent experimental evidence

-------
                                                                    3-31
indicates that the value for pK^.-  estimated by Benson is too  high and
                               *°2
that the proper value should be about pK.—.  = 5 (21).   Even if this

value is utilized, the reverse reaction under these conditions can still

be neglected.  Therefore, if the Weiss mechanism is valid, one should

observe the rate relationship 3-20d, i.e., a reaction rate which is

first-order in both [Fe  ] and [0_].  The experimental data presented

in Figure 3-13 do obey such a kinetic law in the acidic region.

       Another interesting observation with regard to the Weiss mechanism

is the hypothesis that the Fe  *()„  complex forms first and is subse-

quently stabilized by anionic ligands.  In view of the rapid nature of

simple complex-formation reactions (22), it is more likely that complex-

ation of Fe   by the ligand occurs first, followed by the reaction with

oxygen and, in the termolecular mechanism (13), by another Fe  .  Cher

and Davidson(23) have considered that complex-formation serves to make

AH and AF for reaction 3-14a less positive, with the net effect being

to lower the activation  energy of the reaction.

       The possibility of a 2-equivalent electron-transfer has been pro-

posed, largely as a result of the work of Cahill and Taube (24), who
              •                                              •
postulated Fe(lV) to be  a reaction intermediate in place of HO. in the

single-electron transfer.  Such an interpretation would be in accord with


            Fe(ll) + 02  = Fe(lV) + H^                           (3-26)


            Fe(il) + Fe(lV) = 2 FeClII)                           (3-27)


            Fe(ll) + H202= Fe(iv) +  2 OH~                         (3-28)

            Fe(ll) + Fe(lV) = 2 Fe(lll)                           (3-29)

-------
                                                                    3-32





The role of anionic complexes with the intermediate Fe(IV)  would be



similar to that observed for Fe(lll) in the single-electron-transfer



mechanism (23).



       Conocchioli, et al  (25) have begun to investigate the oxidation



of Fe(ll) by  2-equivalent  oxidants, their preliminary data corroborating

                                                                      •

the proposal  of Cahill and Taube with regard to the involvement of Fe(IV)



as an intermediate.  Their observation of the Fe(lll) dimer as the



primary  end-product led  to their formulation of the mechanism as



                „         2  equiv.

            Fe   + Ox. - > Fe(lV) + Red2                 (3-30)
             Fe(lV)  + Fe+2  raPid>   [Fe(lll)]                     (3-31)
 where [Fe(lll)32 refers to the dimer,  represented by FeFe  .   In




 studies employing HOC1 as the 2-equivalent oxidant, the dimer formed was




 observed to be TeFe+ ,  which slowly converted to the dihydroxo-dimer
                   Cl


 under acidic conditions.  In a similar vein,  it will be reported in the



 next chapter that the same inorganic ligands  which accelerate the rate of



 oxidation of Fe(ll)  by apparently stabilizing the transitory intermediate



 also accelerate  the  rate of hydrolysis of Fe   , which presumably proceeds



 through the dimer.



        The formation of a bridged activated complex, as proposed by



 Taube (26) to explain the role of complex formers in serving as a bridge



 between the oxidizing and reducing agents,  is unlikely in this situation;



 its major relevance  is in redox reactions between metal ions.



        If Figure 3-13 is considered, it is readily apparent that, in the



 acidic pH-range, the oxidation reaction proceeds relatively slowly, as

-------
                                                                    3-33



predicted by the Weiss mechanism.  Whether or not this type of mechanism


is valid, one could still postulate that as the pH is increased and [OH~]


Increases, the net effect is that the ligand OH~ behaves in the same


fashion as the other complexing ligands by coordinating with one of the


iron species and stabilizing the transitory complex.  If this were the


case, then one should not observe other anionic ligands to exert a


catalytic effect at the elevated pH-values.  Stumra and Lee (3), in a


preliminary survey, noted that chloride and sulfate exerted no such ac-


celerative effect at pH-values greater than 6, but H2PO ~ did.  Forma-


tion of mixed hydroxo-ligand complexes of Fe(lll) could account for such


an observation.



       At the higher pH-values investigated, care must be exercised in


interpreting the results because of the heterogeneity of the system.


The resultant Fe(lll) hydrolyzes quite rapidly (see Chapter 4), forming


insoluble ferric hydroxide.  Although it can be assumed that the cata-


lytic effect of OH  at higher pH-values is similar to that suggested


earlier, one must bear in mind that the previously-mentioned mechanisms


have all been derived for homogeneous, one-phase systems.


       Abel (27) has proposed that the second-order dependence on [OH~],


exhibited in Figure 3-13, arises from the following scheme:



            02 -f OH" = 02'OH"                                     (3-32)




            02'OH~ + OH" = 03~2 + H20                             (3-33)


                                                        _2
The two equilibria proceed rapidly with the resultant 0.,   slowly re-


acting with Fe   to form additional reactive oxo-complexes which propa-


gate the chain.

-------
                                                                   3-34
       Wells and Salam (28) have attached special  significance to


electrostatic considerations in that it is easier  to remove an elec-


tron (e~) from Fe+2 by decreasing its positive charge through complex-


formation.

       The second-order dependence on [OH"] is reminiscent of the

second-order dependence on [H2P04~] observed by Cher and Davidson  (23)

in their  study of the oxidation of Fe(ll)  in phosphoric acid solution.


An explanation was not given for  this second-order dependence, but in
     S
a later  paper King and Davidson (29) proved that  it was not due  to

                                -2
condensation of  H2P°4  to H2P2°7   *

       The dependence upon [OH"]  and the conversion of the solution to


 a two-phase system due  to hydrolysis of Fe(IIl) is  suggestive of a

 kinetic  dependence on hydrolyzed  species of Fe(lll).   It  is interesting

 to speculate with regard to such hydroxo-species, especially since the

 transition region observed in Figure 3-13 occurs  near  the pK-value for


 the reaction



             FeOH+2 + H20 = Fe(OH)2+ + H+    pK =  4.6             (3-34)



 Beyond pH 4.6, Fe(OH) * becomes the dominant soluble species of Fe(lll).

 The correspondence between the second-order hydroxide-dependence  of  the

 rate of oxidation of Fe(ll) and the dihydroxo-compos it ion of Fe(IIl)  is

  striking.  However,  there  is no evidence  to indicate that an autocatalytic


 mechanism is operative,  as in the oxygenation of Mn( II) (10);   the  addi-

  tion of Fe(lll) has no apparent  effect on the rate of oxidation of  Fe(ll).


  It is also unlikely that hydroxo-species  of Fe(ll) are involved,  since


 hydrolysis of Fe(ll) does not become significant until pH 6.5.

-------
                                                                    3-35
Consequently, with regard to Figure 3-13,  one can only emphasize the




significance of pH on the rate of oxidation of Fe(II) at. pH-values




greater than 4.5.  At this time, there is insufficient chemical evi-




dence on which to base any conclusive mechanistic interpretation.

-------
                                                                    3-36
                                References
 1)   Just,  G.,  "Kinetische Untersuchung der Autoxydation des  in Wasser
        gelosten Ferrobicarbonats," Z. Phys. Chem.,  63, 385 (1908)

 2)   Holluta,  J., and Eberhardt, M.,  "Uber geschlossene Enteisenung
        durch Schnellfiltration," Vom Wasser. XXIV,  79 (1957)

 3)   Stumn, W., and Lee, G. F., "Oxygenation of Ferrous Iron," Ind. Eng.
        Chem.,  53_, 143 (1961)

 4)   Ghosh, M.  M., O'Connor, J. T. , and Engelbrecht, R. S., "Precipita-
        tion of Iron in Aerated Groundwaters,* Journ. San. Eng. Div. ,
        Proc.  Amer. Soc. Civil Eng.,  92, 120 (1966)

 5)   Stumm, W., and Singer, P. C., "Precipitation of Iron  in  Aerated
        Groundwaters," discussion, Journ. San. Eng.  Div.,  Proc. Amer.
        Soc. Civil Eng., 92. 120 (1966)

 6)   Morgan, J. J., and Stumm, W. , "The Role of Multivalent Metal Oxides
        in Limnological Transformations, as Exemplified by Iron and
        Manganese," Proc. Second Intl. Conf. Water Poll. Res., Tokyo,
        p.  103 (19641'

 7)   Hale,  F.  E. , "Iron Removal Without Aeration - The Precipitation of
        Ferrous Carbonate in a Closed System," 3.^Amer. Wat.  Works Assn.
        28, 1577 (1936)

 8)   Lee, G. F., and Stumm, W., "Determination of Ferrous  Iron in the
        Presence of Ferric Iron," J.  Amer. Wat. Works Assn.,  52, 1567
        (1960)                     "

 9)   Morgan, J. J., and Birkner, F. B. , "Precipitation of  Iron in Aerated
        Groundwaters," discussion, Journ. San. Eng.  Div.,  Proc. Amer. Soc.
        Civil Eng., 9£, 137 (1966)

10)   Walton, A. G., The Formation and Properties of  Precipitates, Inter-
        science, New York (1967)

11)   Morgan, J. J., "Chemistry of Aqueous Manganese (II) and(lV)," Ph.D.
        thesis, Harvard University (1964)

12)   George, P., "The Oxidation of Ferrous Perchlorate by  Molecular
        Oxygen," J. Chem. Soc., 4349  (1954)

13)   Huffman,  R. E., and Davidson, N., "Kinetics of the Ferrous Iron -
        Oxygen Reaction in Sulfuric Acid Solution,"  J. Amer.  Chem. Soc.,
        28_, 4836 (1956)                             '	

-------
                                                                    3-37
14)   Weiss,  J.,  "Elektronenubergangsprozesse  im Mechanismus von Oxydations-
        und  Reduktions- Reaktionen in Losungen,"  Naturwissenschaften, 23^,
        64 (1935)

15)   Frost,  A. A.,  and Pearson,  R. G., Kinetics and Mechanism, John Wiley
        and  Sons, New York (1962)

16)   Weiss,  J.,  "Electron Transfer Reactions  in the Mechanism of Oxida-
        tion-Reduction Processes in Solution,"  J.  Ghim. Phys., 48, C-6
        (1951)

17)   Zwolinski,  B.  J., Marcus,  R.  J., and Eyring,  H.,  "Inorganic Oxidation-
        Reduction Reactions in Solution," Chem. Rev.,  55, 157 (1955)

18)   Weiss,  J.,  "The Autoxidation of Ferrous  Ions in Aqueous Solution,"
        Experentia, DC, 61 (1953)

19)   Benson, S.  W., The Foundations of Chemical Kinetics, McGraw-Hill
        Book Co., New York (1960)

20)   Barb, W. G.,  Baxendale, J.  H. , George, P., and Hargrave, K. R. ,
        "Reactions  of Ferrous and Ferric Ions with Hydrogen Peroxide,"
        Trans. Farad. Soc., 47_,  462 (1951)

21)   Morris, J.  C., Harvard University, Personal Communication

22)   Wendt,  H.,  and Strehlow, H. ,  "Schnelle lonenreaktionen  in Losungen.
        II.   Die Bildung Einiger Einfacher Komplexe  des Eisen-III-  Ions,"
        Z. Elektrochem, 66, 228 (1962)

23)   Cher, M.,  and Davidson, N., "The Kinetics of the  Oxygenation of
        Ferrous  Iron  in Phosphoric Acid Solution," J.  Amer. Chem. Soc.,
        7T_,  793  (1955)

24)   Cahill, A.  E., and Taube, H., "The Use of Heavy Oxygen  in the  Study
        of Reactions of Hydrogen Peroxide," J.  Amer. Chem. Soc.,  74,
        2312 (1952)

25)   Conocchioli, T. J., Hamilton, E. J., and Sutin, N. ,  "The Formation
        of Fe(lV) in the Oxidation of Iron (II)," J. Amer. Chem. Soc.,
        87,  926 (1965)

26)   Taube,  H.,  "Mechanisms of Redox Reactions of Simple  Chemistry,"
        Advances in Inorganic Chemistry and Radiochemistry,1,  1, Emelius,
        H. J., and Sharp, A. G.,  editors, Academic Press, New York  (1959)

27)   Abel, E., "Uber Autoxydation in Umbelichteter Homogener Wasseriger
        Losung.   Mit  Besonderer Berucksichtigung Anorganischer Systeme,"
        Z. Elektrochsm., 59, 903  (1955)

-------
                                                                     3-38
28)  Wells, C. F., and Salam, M. A., "A Kinetic Approach to the Nature
        of Ferrous Ions  in Aqueous Solution," Nature, 203, 751 (1964)

29)  King, J., and Davidson, N., "Kinetics of the Ferrous Iron - Oxygen
        Reaction  in Acidic Phosphate - Pyrophosphate Solutions," J.  _Ai?.er
        Chem. Soc., 80_,  1542 (1958)

-------
                               CHAPTER 4







                       HYDROLYSIS OF FERRIC IRON







4-1  Introduction





       The solubility of ferric iron in natural waters,  under most  con-




ditions, is controlled by the solubility of its various  oxides and  hy-




droxides.  Figure 2-13 demonstrated that the concentration of soluble




Fe(lll) is less than 10" M at pH 4.  When the solubility product of




ferric hydroxide is exceeded, a series of hydrolytic reactions takes




place as the formation of insoluble ferric hydroxide proceeds through




multiraeric and polymeric hydroxo-intermediates.  These kinetic inter-




mediates tend to be adsorbed at interfaces, thus accounting for the use




of Fe(lll) as a coagulant in water  treatment.




        It was also shown (section  2-3.2) that various ligands tend to




coordinate with Fe(IIl) and that the  degree of  coordination is a func-




tion of  the relative affinity of Fe(lll) for these various ligands ver-




sus its  affinity for OH~.  The  existence of mixed ligand-hydroxo-com-




plexes  was assumed to  be of relevance in natural waters.




        This chapter centers upon the  kinetics of hydrolysis of Fe(lII)




in systems oversaturated with respect to ferric hydroxide.  The effect




of sulfate on  the kinetics of hydrolysis was investigated, especially




due to  the high  concentrations  of  sulfate  found in mine drainage waters.




Sulfate also  serves as a representative ligand  in order to gain some in-




sight  as to  the  rate  of hydrolysis of Fe(lII)  in systems  containing

-------
                                                                  4-2





ligands which compete with OH~ for the coordination sites  of Fe(lll).



The coagulative properties of ferric iron are considered,  in brief,  to



demonstrate that it  is the hydrolytic intermediates which  are  respon-



sible for the destabilization of colloidal dispersions.  Finally,



phosphate removal by ferric  iron has been investigated,  both by the



direct  addition of Fe(lII) to a system containing phosphate and by the



addition of Fe(ll) which  is  subsequently oxidized, in situ, to Fe(lII).







4-2  Kinetics of Ferric  Iron Hydrolysis



        4-2.1  Reactions  of Fe*  with Water




        Ferric  iron  in  aqueous solution behaves as a multiprotic



BrBnsted  acid, with  protons  being  transferred from the coordinated



water molecules of Fe(lll) to the  solvent water  in the following step-



wise manner:




            Fe(H00)/3 + H00 = Fe(H_0),OH+2  + H,0+               (4-1)
                 L   D      c.         L   J        J




            Fe(H20)5OH+2 + H20 = Fed^O^OK^ + E^            (4-2)




            etc.



Such reactions can  proceed until  all  of  the  coordinated water molecules



have been deprotonated,  resulting  in  the formation of anionic hydroxo-



ferric  species.   In  addition to  these "aquo-acidity" reactions (l)



with mononuclear  products, these  simple  hydroxo-ferric complexes tend



to polymerize by  a  condensation  process,





            2 Fe(H00),OH+2 = Fe0(H,0)Q(OH).+4 +  2 H_0            (4-3)
                   L   D         L   L  o    /        /

-------
                                                                   4-3





where water is essentially squeezed out of the coordination shell.   The



resultant dimer has the structural configuration








            (H20)4Fe<^  ^>Fe(H20)4+4                         (4-3a)






The dimer is subject to additional hydrolytic reactions, again involv-



ing a proton-transfer






            Fe_(H~0)R(OH) +  + H-0 = Fe7(HJD)7(OHK   + H,0+     (4-4)






or additional condensation reactions by which the resultant molecule



is further dehydrated






            2 Fe_(H.O)7(OH).+3 = Fe.(H_0), _(OH),+6 + 2 H^O       (4-5)
                Lcl    j       "t  L   IZ    b        2




       In systems oversaturated with respect to insoluble ferric hy-



droxide, a series of such hydrolytic and  condensation reactions takes



place, the multimeric and polymeric hydroxo-species serving as kinetic



intermediates in the formation of the solid ferric hydroxide precipi-



tate.  (A useful review of the chemistry  of metal ions in aqueous



solution has recently been presented by Stumm (2).)  The solid phase,



however, continues to react as the amorphous freshly-precipitated hy-



droxide generally converts to  o(-FeOOH..   Under special conditions,



cX-Fe-0    $ -FeOOH, and  5-FeOOH can be  formed (3).  In addition to
     L j


these crystalline products, an inactive amorphous material is always



present, even after prolonged periods of  aging.  Under certain condi-



tions, a colloidal sol of ferric-oxide hydroxide can be maintained for



long periods of time.  As previously mentioned (section 2-3.2),  it is

-------
                                                                   4-4


analytically very difficult to distinguish between dissolved  and  sus-

pended ferric iron.

       Since the coagulative properties of Fe(lll) arise from the ten-

dency of the hydrolyzed polynuclear kinetic intermediates to  be adsorbed

at interfaces (4), and since Fe(OH)_ and its hydroxo-interraediates are

responsible, in part, for the transport of phosphate and organic  materi-

al in lakes (5), it would be desirable to elucidate the kinetic rela-

tionships which govern the hydrolysis of ferric iron.


4-2.2  Experimental Study of the Kinetics of Fe(lll)  Hydrolysis

Experimental Procedure

       The kinetics of hydrolysis of free ferric iron,  Fe+ ,  was  studied

potentiometrically, employing the ferro-ferri cell


            Pt/ Fe(ll),Fe(lIl),Na+,H+,C10.~/ NaCl  / Hg.Cl./  Hg   (4-6)
                                         4  (sat'd)    2  2

The cell has previously been described in section 2-3.3 and was shown

in Figure 2-15.  The reversible potential is established by the elec-

troactive ferrous-ferric couple in accordance with the  Nernst Equation

which, for a constant ionic medium (0.1 M NaCIO,) in a  constant temper-

ature water bath at 25 C, becomes


            E = E°' -0.0592 log [Fe ,]                           (4-7)
                                [Fe+3]

E   was determined in a preliminary experiment by measuring the poten-

tial of a known system at pH 1.0, where no hydrolysis or oxidation took

place, so that [Fe+3] = [Fe(lIl)]T and [Fe+2] = [Fe(ll)]T.  [Fe(ll)]T

and [Fe(lll)]  were determined using bathophenanthroline (6).

-------
                                                                  4-5



       Solutions containing various concentrations of Fe(ll), Fe(lII),


and H  were prepared, under a nitrogen atmosphere, at the same  ionic


strength and temperature as the standard system.  It was noted  that the


addition of even a weak alkali (dilute NaHCO-) to raise the pH  of  the


system brought about kinetically irreversible hydrolysis.  Apparently,


localized conditions of high pH (high concentrations of OH  and a  cor-
     t
responding large degree of oversaturation with respect to ferric hydrox-


ide) were produced in the vicinity of the tip of the burette from  which


the base was slowly dispensed.  Even under conditions of intense mixing,


basification of the system containing Fe(ll) and Fe(lll) resulted  in  a


sharp decline in potential.  Therefore, solutions containing Fe(lll)


had to be maintained under conditions such that the solubility of  fer-


ric hydroxide was never  exceeded until the time at which the kinetic


studies were begun.  In  most cases, Fe(IIl) was added directly to  the


system as the perchlorate  salt, at time zero.   In those  instances where


it was desired  to begin  the  study  at  a higher pH, solutions of Fe(ll)


were basified with dilute NaHCO  ,  under a nitrogen atmosphere, and then


ozonated (Sanders Ozonator Model No.  S-V-106, Triton Aquatics) in order


to generate ferric  iron  in situ  (7).  Ozone was bubbled  through the


system and the  extent of oxidation of Fe(ll)  was  followed by noting the


increase in potential of the ferrous-ferric couple.  Generally, one


minute was  sufficient  to  oxidize  about half  of the  initial concentra-


tion of Fe(ll), so  that  [Fe(Il)]T ^ [Fe(lIl)]T.   Any residual  traces of


ozone  or oxygen were removed by  the  nitrogen  which was  continuously bub-


bled through the solution.   [Fe(ll)]T was  determined before and after

-------
                                                                   4-6
ozonation, the difference being the total concentration of Fe(lll)




which was generated.




       The solutions were stored under an atmosphere of nitrogen and




the pH was kept sufficiently low so that no oxidation of Fe(ll)  took




place, and [Fe"1" ] =  [Fe(II)3_, throughout the duration of the experi-




ment.  [Fe   ], the concentration of unhydrolyzed iron (III) species at




any time, was calculated from the Nernst Equation (4-7) after measur-




ing the potential of the system to the nearest 0.1 mv., using a Model




D  Sargent Recording Titrator.   The liquid junction potential, E., has




been  shown (7) to be  strictly a function of [H ].  Extrapolating the




results of Biedermann  (7),  E. remains relatively constant during the




course of the experiment so that variations in E. can be neglected.




The concentration pH (p H)  was  measured using a Leeds and Northrup Po-




tentiometer  (Catalog No. 7664), calibrated  in the same manner as pre-




viously described  in section 2-2.3.  The  samples were analyzed for




 [Fe(ll)]  and [Fedll)]™ using  bathophenanthroline.






Experimental Results  and Discussion




       Due to the acidic nature of the  aquo-ferric  ion, the pH of the




system decreases when  ferric salts are  placed in solution.  The simple




proton-transfer reactions by which the  mononuclear  hydroxo-complexes




are formed (reactions  4-1 and 4-2) take place extremely rapidly (8).




When  the  solubility  product of  Fe(OH),  is  exceeded, the slower hydro-




lytic and condensation reactions proceed  whereby the multinucelar and




polynuclear  species  and, eventually, solid  ferric hydroxide are formed.




One would expect that  the rate  of disappearance of  free Fe   would be




dependent not only upon the concentration of Fe  ,  but also upon pH or,

-------
                                                                   4-7


more precisely, upon the concentration of OH .   Since hydrolysis tends

to decrease [OH 3, the reaction would be expected to decelerate as

hydrolysis progresses.  Furthermore,  the resultant decline  in pH also

causes a shift in the Fe   - FeOH   distribution so that  the concen-
              r\
tration of Fe   is partially buffered; on one hand, it  is diminished

in forming ferric hydroxide, and on the other hand, it  is augmented
                          O       *5
due to conversion of FeOH   to Fe  ,  i.e., the back reaction of 4-1.
This effect can be pictured as
                                  Fe\
                                   H+X                          (4-8)
                             FeOH+2   Fe (OH)
                                        x    y

Therefore, in order to derive any useful kinetic information,  the  system

must be maintained at constant pH.  Use of a pH-buffer is precluded be-

cause of possible kinetic effects the buffering agent might bring  about.

Alternatively, small concentrations of Fe   can be employed in order  to

minimize the pH-decrease accompanying hydrolysis, or one can follow

[Fe  3 with time and compute the instantaneous rate of decrease of

[Fe  3 at any given pH or in a region of constant pH.  These latter

two approaches were utilized in treating the experimental data.   In
                                                   •\
this manner, any change in the concentration of Fe   arises only from

hydrolytic reactions.

       Figures 4-1 and 4-2 show some of the experimental results.  In

each cases the data have been plotted assuming either a first- or

second-order dependence of the reaction rate on [jFe  3.  Rate laws

of the form

-------
       3.81  3.67  3.64
                         P'H
s
§
o     ,.
0  10'6
Q)
9

8

7

6   I—
                                 3.63
       0
   FIGURE 4-1.
                            (a)
                    Temp. 25 C

                constant ionic medium
                    I • 0.1 M NaClO,
               20          40
               TIME,  minutes
                                           60
                                                                  3.81 3.67 3.64
PCH
                                                                                     3.63
                                                                           20        40       60
                                                                          TIME,  minutes
                                                                                        +3
             (a) Logarithmic and  (b) reciprocal plots of the rate of hydrolysis of FeH
             (a) assumes a first-ord|r dependence and (b) a second-order dependence bf the
             reaction rate on the Fe*J concentration.                     epenaence or the
                                                                                                          00

-------
                              PCH
     3.52   3.47
DS

 o
 1-t
 X
  •\
 £3




 I
 55
 w
 0)
3.44
                                  3.41
                                                           3.52   3.47
PCH
                                                      3.44
                                                                                        3.41
                                                                                               150
                                                                                 200
0          50          100         150        200    0          50           100

                 TIME, minutes                                        TIME,  minutes                        ^


FIGURE 4-2.  (a) Logarithmic and (b) reciprocal plots of rate of hydrolysis  of the aquo-ferric  ion, Fe   .  ^

-------
                                                                 4-10




               pe+3]   = v  [Fe+3]                             (4-9a)
               dt          1



                                +3-2                           (4.9b)
                        =

               dt          t.


have been assumed, where ^ and k_ are functions  of pH.  The measured


p°H-values are noted.  Kinetic significance of [OH ]  is  implied by


the decreasing rate of hydrolysis for both the logarithmic  and recipro-



cal plots.  If the linear regions of each of the  respective logarith-


mic and reciprocal plots are considered, the instantaneous  rate of hy-


drolysis can be calculated for the pCH-range over which  such linearity


is observed to hold.  For each sample, the latter portion of the  curve


is relatively linear for it is in this region that the p H  remains es-



sentially constant.


       The  instantaneous slopes of the logarithmic and reciprocal plots



are equal to the  "rate constants," k. and k_, for the first- and  second-


order  relationships, respectively.   If k^ and k^ are assumed to be



functions of pH of the form




            kL .  kx'  [OH~]n                                    (4-lOa)




            k2 =  k2'  [OH']n                                    (4-10b)



then  a plot of log \a, or log  k« versus p H  should result in a straight


line  of  slope n.  Figures  4-3 and 4-4 are plots  of the instantaneous


slopes of the logarithmic  and reciprocal plots (Figures 4-1 and 4-2)


against  p^.  A linear regression analysis  was performed to compute


the best  straight line to  fit the experimental rate results.   The cor-


relation coefficients for  Figures 4-3 and 4-4 are 0.76 and 0.97,  re-


spectively, indicating that the data conform to  a second-order

-------
    -1.0
    -2.0 __
.*

9
    -3.0
    -4.0
       2.75


FIGURE 4-3.
   3.00
3.25      3.50
    P H
                               3.75
                                                                4.0 —
                                                            I
                                                            •4
                                                            I
                                                            
-------
                                                                  4-12
dependence on EFe  ]  considerably better than to a first-order depen-



dence.  The  slope of  Figure 4-4 is 3.7,  implying that the rate of hy-



drolysis  for the conditions investigated can be reasonably described



by  a kinetic relationship of the form
                       = k '  [Fe+3]2 [OH']4                     (4-11)
where the rate constant k«' is approximately 10  *  mole liter  rain



for the p°H range 2.8 to 3.8.



        The data were also tested for conformance  to the von Weimann



formulation, whereby the rate of precipitation (hydrolysis) is propor-



tional to the relative degree of over saturation,  i.e.,







             "d ffe* ]  = k (Q-K)/K                             (4-12)
                at



where Q is the reaction quotient and K is the equilibrium solubility



product for the formation of ferric hydroxide. The results might be



interpreted in terms of the kinetics of crystal growth but the compu-



tations were limited by the lack of a dependable  solubility product


                                  —38
for Fe(OH),.   A value of K   =10    was employed for the calculation
           J               SO


(see below),  but such treatment of the data met with little success.



        It should be emphasized that equation 4-11 is merely a data-



fitting relationship.   It is not intended to serve in any specific



mechanistic way since there are probably numerous other rate expressions



which could fit the experimental data.   Furthermore, equation 4-11 is



applicable only for the experimental conditions investigated; extra-



polation of this relationship to other  situations might result in



serious  error.

-------
                                                                   4-13



       Wendt observed (9) that the dunerization reaction  (4-3)  is con-


siderably slower than the simple proton-transfer reaction and that  the


former is accelerated in the presence of simple anions via a reaction


of the forn>

                                    OH
                            2 , Fe      Fe(5-n)                   (4-13)
Figure 4-5 contains plots cf the  second-order rate constant for the

                         "3
disappearance of free Fe+   in the presence of sulfate.  The figure was


developed by treating the experimental data  in the same fashion as de-


scribed above.  For the  three different  concentrations of sulfate


analyzed, th- average slope is  3.3.   The rate of hydrolysis is more


rapid  in  the presence of sulfate  than in its absence, as demonstrated


in Figure 4-6.  These results  are especially useful when applied to the


mine drainage  system considered in  Chapter  5.


       One  could  generalize,  on the basis of Wendt' s  findings  and those


reported  here,  that anions  which coordinate with Fe   tend to  increase


the rate  of hydrolysis  of Fe(lll).    Hence,  hydrolysis would be  ex-


pected to occur even  more rapidly in natural waters under  the  influ-


ence of  the many  natural anioas which are present,  such  as  sulfate,


chloride, phosphate,  orthosilicate, and the many organic  anions.


Schenk and  Weber  (10)  reported that silicic acid hindered the  hydroly-


 sis of ferric  iron, but their results are based upon  indirect  observa-


 tions  of free Fe+3, such as absorbance of the hydrolytic products at


 420 mu,  and residual [Fe(lll)! following filtration through an 0.45 u


 membrane filter.   It is suggested that  the  effects of such catalysts

-------
                                                                      4-14
   4.0

V
1-1
i
«L
   3.0  —
    2.0 -
 CM
    1.0
total sulfate
concentration

   10"4M
           total sulfate
           concentration

             5x!0"4M
       2.8
   3
3.6 2.8
                   total sulfate
                   concentration

                      10"3M
3.4
                                                 3.6  2.8
3c2
                                                           3.6
       FIGURE 4-5a,b,c.   pH-dependence of "second order  rate  constant"  for
                         hydrolysis of Fe   in the presence of sulfate.
        5.0
        4.0
     I  3.0
     o>
      CM
        2.0
         1.0
              total sulfate
              concentration\
                   .-3..      V
                          /
                              10
                            no sulfate


                              Temp. 25 C

                           I = 0.1M
                         3.0
                                     3.5
                                                                        4,0
                                                                    +3
           FIGURE  4-6.   Comparison between rates of hydrolysis of Fe
                         in  the  presence  and absence of sulfate.

-------
                                                                  4-15
or inhibitors be investigated by direct analytical methods such as



those employed in this study.





Solubility Product of Amorphous Ferric Hydroxide



       Immediately following the investigation of the kinetics of



hydrolysis of Fe(IIl), the remainder of each of the samples was re-



moved from the ferro-ferri cell and placed  in a BOD reaction bottle



under a nitrogen atmosphere.  The  BOD bottles were stored under water.



After a period of approximately 17 days,  the samples were removed from



storage and  again placed in  the ferro-ferri cell.  The reversible



Nernstian potential,  E,  pCH,  and  [Fe   ] were determined as before, and



the  solubility of ferric hydroxide was  computed.  The data and results



are  summarized  in Table  4-1.  The  pK  Q-value  indicated should not be



referred  to  as  the  solubility product of  ferric hydroxide, since  it is



probable  that  equilibrium had not  been attained,  and the  structural form



of the  end-product  was not analyzed.   In any  case,  the average value



determined,  pK.   =  38.1, does give some idea  as to  the magnitude  of
               SO


the  solubility product of freshly-precipitated ferric hydroxide.  Feit-



knecht  and  Schindler (3) report PK  -values for the solubility product
                                   SO


of amorphous Fe(OH)3 in 3M NaCK>4 of 38 to 39.1,  depending upon  the



 age  of  the precipitate,  i.e., the time elapsed following  precipitation



before  the measurements were made.







 4-3  Coagulative Properties of Ferric Iron



        O'Melia and Stumm (4) have shown that  destabilization of  col-



 loidal dispersions of silica by Fe(lll) is accomplished  by  the multi-



 meric and polymeric hydroxo ferric species which arise as unstable

-------
                                                                 4-16
Table 4-1.  Check on  the  Solubility Product of Ferric Hydroxide
Sample CR
No.
14
15
16
17
18
19
20
21
22
23
24


2
2
3
2
3
3
3
3
3
3
2

*
.89
.88
.00
.76
.21
.20
.31
.20
.20
.22
.85

E*
mv.
403.
391.
380.
416.
366.
369.
343.
364.
366.
372.
387.

Potential
NOTE:
CFe+2]
x 104M
6
8
0
2
9
2
7
6
0
4
9

4.52
5.60
5.15
5.46
1.95
2.28
2.25
2.04
2.10
1.88
5.64

readings
E°'
= 486
[Fe+2] CFe+3J
[Fe+3] x 106M 2
24.
39.
61.
15.
103.
93.
254.
112.
106.
83.
45.

versus
.0 mv.
6 18
1 14
7 8
1 36
1
5 2
0
1
1
0 2
6 12

saturated

.4
.3
.35
.2
.89
.44
.886
.82
.98
.27
.4

[OH"]3
c 103V
0.468
0.447
1.00
0.191
4.27
3.98
8.27
3.98
3.98
4.59
0.355
Average
K v
so pK
39 s°
x 10
8.60
6.39
8.35
6.92
8.07
9.71
7.33
7.25
7.88
10.42
4.40
pK
^ so
38.07
38.19
38.08
38.16
38.09
38.01
38.14
38.14
38.10
37.98
38.36
=38.1
calomel reference electrod




                  T = 25°C
                  I = 0.1
                  Readings taken 17 days  after precipitation

-------
                                                                 4-17



kinetic intermediates in the precipitation of ferric hydroxide.  The


destabilization process is attributed to specific adsorption  of  these


hydroxo-ferric species; this adsorption brings about aggregation


either by a bridging effect whereby the hydrolyzed Fe(lll)  forms


linkages between a number of dispersed particles, or by neutraliza-


tion of the negatively-charged colloidal particles by the positively-


charged hydrolyzed Fe(lll).


        The efficiency of destabilization depends upon the  hydrolytic


properties of Fe(lII) and is, therefore, influenced by pH.  Figure 4-7


shows the effect of pH on the concentration of iron (III) required for


aggregation of colloidal silica  (Ludox SM, E. I. Dupont de  Nemours and

                                                              2
Co., Wilmington, Delaware) at a.  surface concentration of 150  m /I.  The


experimental data for Figure 4-7 were obtained in the same  manner as


those obtained by O'Melia and Stumm (4) , usine light scattering  as an


indicator of coagulation.


        There are two results of the coagulation study that are directly


related to the kinetics of hydrolysis of Fe(IIl).  Firstly, O'Melia


and Stumm observed that prolonged  thermal aging of solutions of  ferric


perchlorate, in order to form polymeric hydroxo-ferric species,  in-


creased the critical coagulation concentration and the critical  re-


stabilization concentration for  the colloidal silica dispersion,i. e.,


greater concentrations of Fe(III)  were required for a given degree of


coagulation and restabilization, respectively.  This was attributed  to


partial precipitation of Fe(OH)  and a corresponding reduction in the


concentration of active hydrolytic intermediates.  Secondly,  the


plateau between pH 3 and 6 in Figure 4-7 suggests that in this region

-------
                                                                    4-18
    -3.0
    -3.5
 u
 .
 o
 •
 u
"I* -4.0
    -4.5
    -5.0
   Coagulation region
                                                I  I HI I 1  I  gl I  I  t  I,
                                    O    O
                                 Colloidal  suspension
              Ludox SM  -  150 m /I



              I	I
                    I
            3.0
4.0
                                        PH
5.0
6.0
         FIGURE 4-7.  Aggregation of  colloidal  silica dispersion by


                      hydrolyzed ferric  iron.


                      0
                  I
                  H
                  §
                  W
                  E-f
                  CO

                  O
                     -2
                           - Fe(OH)3


                       equilibrium
                                                    -24
                                            K  = 10    for FePO.  ,, N
                                             so                4  (s)

                                                    38
                                                 10    for Fe(OH)3
                     -8
                       FIGURf; 4-8.  Solubility of  ferric phosphate.

-------
                                                                  4-19






of  Fe(OH)3 insolubility, the polynuclear hydroxo-ferric species ex-




hibit equivalent coagulative behavior  independent of pH.  These studies




were conducted in pH-stats where constant pH was maintained by the ad-




dition of dilute sodium bicarbonate.  However, the localized gradients




of high pH which are set up by the external addition of a basifying




agent bring about kinetically irreversible hydrolytic reactions.  Such




effects were observed in the laboratory  investigations of the kinetics




of Fe(lll) hydrolysis (section 4-2.2), and have also been reported by




Biedermann and Schindler (11) and by Spiro, et al (12).  Since these




studies describe the coagulative properties of kinetic intermediates,




time of reaction and method of preparation of the hydrolytic reactants




should be of significance.  Similar comments have been made (13) re-




garding the coagulative behavior of activated (polymeric) silica.  The




pertinent variables include polymerization pH, polymerization concen-




tration, polymerization time, and age of diluted polymeric solution.




It is hoped that the kinetic studies reported in section 4-2.2 will




serve as a guide for further research designed to clarify the mechanism




of destabilization by hydrolyzed metal ions, and to identify the para-




meters which control the destabilization process.








4-4  Removal of Phosphate






        It was shown, in section 2-3.1, that the solubility of ferric




iron is controlled by the solubility of  its oxides and hydroxides,




except in waters containing appreciable concentrations of phosphate,




at pH-values below 5.  In a similar manner, the total concentration of




dissolved phosphate is governed by solid ferric phosphate and is lowest

-------
                                                                   4-20




at pH 3.5, where the 3olubility of FeP(>4 is at a minimum (see Figure


4-8).  The pH at which minimum solubility is exhibited depends upon


the solubility products of FeP04 and Fe(OH)3 which are utilized  in the


computations and the construction of Figure 4-8.  The critical pH is


not well-defined since neither solubility product is well-known  (see


section  2-3.1).  The boundary below pH 3.5 represents the equilibrium


between  dissolved phosphate  and solid ferric phosphate;  the boundary


above pH 3.5 corresponds  to  the control of soluble phosphate by  the


equilibrium between solid ferric phosphate and solid ferric hydroxide,


reaction 2-27.   (Soluble  phosphate-complexes of ferric  iron have been


neglected since they  do not  markedly enhance the solubility of FePO^.


In addition, other  phosphate minerals,  such as AlPO^, Ca^PO^^, and


Ca rt(PO.),(OH)0,  have been disregarded  although their significance in
   10  4 6    2

natural  systems should not be  overlooked.)


       Consequently,  under certain conditions, Fe(lll)  can serve as


an effective precipitant for the  removal of phosphate from waste waters.


Figure 4-8 suggests that the efficiency of phosphate removal by Fe(lll)


 is pH-dependent, since OH~ competes with the  soluble phosphate species


for the  metal  precipitant.  At pH -values greater than 3.5, the effi-


ciency is decreased due to formation of Fe(OH), and mixed hydroxo-


phosphato-precipitates of Fe(lll).  However,  by the judicious selec-


 tion of  pH and means of addition of Fe(lll),  the hydrolytic tendency


of Fe(lll) can be supressed  and the quantity  of precipitant required


 for removal of phosphate can be made to be stoichiometric  in accordance


 with the reaction


             Fe+3 + P0'3 = FePC                                  (4-14)

-------
                                                                   4-21




       Improvement of phosphate removal by Fe(IIl) can be achieved by



employing the techniques of homogeneous precipitation (14) whereby the



precipitant is not added directly to the system in the customary manner,



but Ls generated internally by a homogeneous chemical reaction within



the solution.  In this way, coprecipitation is minimized and undesirable



concentration effects are eliminated.  In the case of ferric iron,



homogeneous generation can readily be accomplished by the addition of



ferrous iron which is subsequently oxidized in situ.  The resultant



Fe(lll) is uniformly distributed throughout the system, promoting



direct contact between all of the phosphate and all of the iron.



       An experimental study of the removal of phosphate by  externally-



added and homogeneously-generated ferric iron was undertaken and the



results are presented below.





4-4.1  Precipitation of Phosphate by Ferric Iron



Experimental Procedure


       Various dilutions of a standardized  solution of ferric per-


                                                        -4
chlorate were added to a series of beakers  containing  10  M Na_HPO^.


(The stock solution of ferric perchlorate had been prepared in dilute



perchloric acid to prevent hydrolysis of Fe(IIl).)  Dilute sodium car-



bonate was simultaneously  added  in order to achieve and maintain the



desired pH.  The  solutions were  flash-mixed and then gently stirred for



fifteen minutes.  Samples  of  each were membrane filtered  (1.2 u pore



diameter) and the residual phosphate  in the filtrate was  determined



employing the method  recommended by the AASGP (15).

-------
                                                                 4-22




        In the study of  phosphate removal by homogeneous  precipita-



tion, acidified ferrous  perchlorate and dilute Na»CO, were  added to


                                         -4
a similar series of beakers containing 10  M NaJIPO,. The  solutions



were flash-mixed and gently stirred while the ferrous iron  was oxi-



dized, in situ, by atmospheric oxygen.  At pH-values  below  6 where the



oxygenation reaction is  slow  (see section 3-4.1),  ozone was bubbled



through the system.  (Ozone was generated by passing  a stream of oxygen



through a Sanders ozonator, Model No. S-V-106, Triton Aquatics.)  Ali-



quots were removed from  each  of the precipitating  systems at various



time intervals and were  analyzed for residual phosphate and Fe(II),



using the AASGP (15) and bathophenanthroline (6) procedures, respec-



tively.  The concentration of Fe(lll) at any given time was calculated



as the difference between the initial concentration of Fe(ll) and



[Fe(ll)] at that time.





Results and Discussion



        Figure 4-9a shows the residual concentration  of phosphate for



various concentrations of ferric iron applied directly from solutions



of ferric perchlorate.   The efficiency of Fe(lll)  as  a precipitant of



phosphate appears to improve  as the pH of the system  is lowered, con-



firming the predictions  made  in connection with Figure 4-8.  Stoichio-



metric removal is approached  at pH 5.



        Removal of phosphate  by homogeneously-generated ferric iron is



presented in Figure 4-9b.  Here again, removal is  improved  as the pH



decreases.  Comparison of Figures 4-9a and b demonstrates that, with



the exception of the results  at pH 7.0, the degree of removal is



enhanced by utilization  of the homogeneous precipitation  technique.

-------
                                                            4-23
o
.

 -

1

:

w
 .-
 W

 b
 .



 I

 a
 D
 o
 —
 : .

  w
  u

  :
  U
  I
  .--



  I

  -

  fcj
  .
           -5.0        -4.5         -4.0       -3.5        -3.0

                LOG [Fe(III)] APPLIED BY DIRECT ADDITION


                                              Y
            .5.0         -4.5        -4.0        -3.5       -3.0

              LOG [Fe(III)l APPLIED BY HOMOGENEOUS GENERATION
  FIGURE 4-9.
Precipitation of phosphate by  (a) direct addition

of Fe(III) and by  (b) homogeneous generation of


Fe(III).

-------
                                                                  4-24
The  improvement is exhibited most clearly at the lower pH-values



studied.   Stoichiometric removal is again observed in the lower pH-



range,  where partial removal is exhibited even at applied concentra-


                                      -4
tions of  Fe(lll)  considerably below 10  M.  It should be noted, how-



ever, that perfect Stoichiometric removal is not effected;  approxi-



mately  3  x 10 ~ M of phosphate still remains, even after the applica-



tion of 10~Si Fe(IIl).



        A distinction must  again be made between equilibrium and kine-



tics.   The problem at-hand-deals with a non-equilibrium situation and,



although  equilibrium considerations can serve as a guide, kinetic con-



siderations are required.   Phosphate may be removed from solution in



two  ways:  by formation  of insoluble ferric phosphate, or by incorpora-



tion in the ferric hydroxide network.  The latter is, by definition,



a  less  efficient process.   Conditions which favor direct iron-phosphate



interactions over hydrolytic reactions should lead to a more effectual



removal of phosphate.   At lower pH-values, hydrolysis is decelerated



and  precipitation of phosphate is enhanced.  However, given sufficient



time, the possibility of conversion of FePO, to Fe(OH)- cannot be dis-



counted under conditions where the latter is the thermodynamically



stable  solid phase.   The results reported in Figure 4-9 reflect removal



of phosphate after only fifteen minutes.  With longer periods of time,



phosphate might be released if such a solid conversion were to occur.



        The improvement in  the removal of phosphate by homogeneous



precipitation can also  be attributed, in part, to the catalytic influ-



ence which phosphate exerts on the oxidation of ferrous iron.   In



Chapter 3,  it  was indicated that the apparent catalytic effect of

-------
                                                                 4-25







phosphate may arise from its coordination with Fe  ,  in which case  it




would remain bound to the product Fe(lll), thus increasing its chances




of removal from solution.









4-5  Summary





        The rate of hydrolysis of  ferric  iron was  shown to be an influ-




ential parameter in coagulation and precipitation  phenomena in natural




systems.  Two  important features of the hydrolytic reactions were em-




phasized, namely the  decided dependence of  the  reaction rate on [OH ],




and  the kinetic  irreversibility of the reactions.  Consequently, the




creation of concentration gradients, such as those which  arise by the




addition of a basifying agent, may cause an undesirable effect in the




system under  investigation.  The techniques of homogeneous precipita-




tion are recommended as one means of overcoming such irreversible  in-




terferences.

-------
                                                                  4-26
                               References
 1)   Sillen,  L.G., "Quantitative Studies of Hydrolitic Equilibria,"
        Quart. Revs.,  Chem. Soc. London, 13, 146  (1959)

 2)   Stumm, W., "Metal Ions in Aqueous Solution," page 520 in Principles
        and Applications of Water Chemistry, S. D. Faust and J. V.
        Hunter, eds.,  John Wiley and Sons,  Inc.,  New York (1967)

 3)   Feitknecht, W., and Schindler, P., "Solubility Constants of Metal
        Oxides, Metal Hydroxides, and Metal Hydroxide Salts in Aqueous
        Solution," Pure Appl. Chem., j6, 132 (1963)

 4)   O'Melia, C. R. , and Stumm, W., "Aggregation  of Silica Dispersions
        by Iron (ill)," Journ. Coll. Inter. Sci., 23_, 437 (1967)

 5)   Morgan,  J. J., and Stumm, W., "The Role of Multivalent Metal
        Oxides in Limnological Transformations, as Exemplified by
        Iron  and Manganese," Proc. 2nd Intl. Wat. Poll. Res. Conf.,
        page  103, Tokyo (1964)

 6)   Lee,  G.  F., and Stumm, W., "Determination of Ferrous Iron in  the
        Presence of Ferric Iron," J. Amer.  Wat. Works Assn., 52, 1567
        (1960)

 7)   Biedermann, G.,  and Chow, J. T., "The  Hydrolysis of the Iron  (ill)
        Ion and the Solubility Product of Fe(OH)_ 7nCln ,n in 0.5M
        (Ha*)Cl~ Medium." Acta Chem. Scand., 20.   "   *    1376.  (1966)

 8)   Wendt, H., and Strehlow, H., "Schnelle lonenreaktionen in Losungen.
        II. Die Bildung einiger einfacher Komplexe des Eisen-III-ions,"
        Z. Elektrochem.,  £6, 228 (1962)

 9)   Wendt, H., "Schnelle lonenreaktionen in Losungen. III.  Die
        Kinetik der Bildung des binuklearen Eisen-III-hydroxokomplexes
        Fe(OH)_Fe  ,"  ~Z. Elektrochem., 66. 235 (1962)

10)   Schenk,  J. E., and Weber, W. J., "Chemical Interactions of Dis-
        solved Silica with Iron (II) and Iron (III),"  J. Amer. Wat.
        Works Assn.. 60_,  199 (1968)

11)   Biedermann, G., and Schindler, P., "On the Solubility Product of
        Precipitated Iron (III) Hydroxide," Acta  Chem.  Scand.,  11,
        731 (1957)

12)   Spiro, T. G., Allerton, S. E., Renner, J., Terzis,  A.,  Bils,  R.,
        and Saltman, P., "The Hydrolytic Polymerization of  Iron (III),"
        J. Amer. Ghem. Soc.. 88_ 2721 (1966)

-------
                                                                  4-27
13)  Stumm, W. , Huper, H., and Champlin, ,R. L. ,  "Formation of  Poly-
        silicates as Determined by Coagulation Effects,"  Environ.
        Sci. Tech., _!, 221 (1967)

14)  Gordon, L. , Salutsky, M. L., and Willard, H. H.,  Precipitation
        from Homogeneous Solution, John Wiley and Sons, Inc.,  New
        York (1959)

15)  Association of American Soap and Glycerine Producers, "Determina-
        tion of Orthophosphate, Hydrolyzable Phosphate, and Total
        Phosphate  in Surface Waters," J. Amer. Wat. Works Assn.,  50_,
        1563 (1958)

-------
                               CHAPTER 5
            OXIDATION OF IRON FYRITE:  POLLUTION OF NATURAL




                      WATERS BY COAL MINE DRAINAGE
5-1  Introduction
       Pollution by  coal  mine drainage  arises from the exposure of




sulfur-bearing minerals present in the  coal  strata to the natural




weathering process.   Mine drainage waters are characterized by low pH,




high acidity,  and  large concentrations  of sulfate and iron as well as




other dissolved  metals.  Various measures have  been  proposed to cope




with this problem, ranging from treatment of the ensuing wastewater to




abatement methodology  in which the weathering process is inhibited.




However, before considering  such  corrective measures,  attention must




 first be focused on  the  chemical  reactions which occur and upon the





 kinetics which govern  these  reactions.




        Despite previous  creditable efforts, no  unambiguous ansv;er has




 been found  as to  which of the steps in the  production of acid mine




 drainage determines the overall rate of dissolution of  the sulfuritic




 agglomerates.   To date,  a didactical approach  toward evaluating  the




 individual  factors controlling the kinetics of the  overall reaction




 has been lacking.  This chapter describes such an approach.  The rela-




 tive rates of the consecutive reactions involved have  been considered,




  as well as the importance of  each  as  it contributes to the problem of




  mine drainage.  The rate-determining  step of the overall sequence has

-------
                                                                   5-2
been ascertained and the physical, chemical, and biological factors




influencing this step have been quantitatively evaluated.  A model




is presented to describe the mechanism by which the sulfide minerals




are oxidized,  and the consequences of the model are discussed from




the standpoint of the various control methods which have been sug-




gested.









5-2  Thermodynamics  and Stoichiometry of Reactions







       The mine-water  system can  be  characterized by the following




overall  stoichiometric reactions:







            FeS2(s)  + |- 02 + H20  = Fe+2 + 2 S04"2 + 2 H+         (5-1)








            Fe+2 + i 02 + H+ = Fe+3  + \ H^                      (5-2)








            Fe+3 + 3 H20 = Fe(OH)3(s) + 3 H+                     (5-3)







            FeS2(s)  + 14 Fe+3 + 8 H20 = 15 Fe+2 -;- 2 SC>4~2 + 16 H+(5-4)







The sulfur-bearing minerals predominant in coal seams are the iron




sulfide  ores,  pyrite and marcasite.  Both have the same ratio of




sulfur to iron, but  their crystallographic properties are quite dif-




ferent.  Marcasite has an orthorhombic structure while pyrite is




isometric (1).  Marcasite is less stable and more easily decomposed




than pyrite.   The latter is the most widespread of all sulfide minerals




and, as  a result of  its greater abundance in the eastern United States




(2), pyrite is recognized as the  major source of acid mine drainage.

-------
                                                                    5-3
       During coal mining operations, pyrite is exposed  to  air and water




with atmospheric oxygen oxidizing the sulfide of the pyrite to sulfate




(5-1), releasing dissolved ferrous iron and acidity into the water.  The




dissolved ferrous  iron undergoes oxygenation to ferric iron (5-2) which




subsequently hydrolyzes  to form insoluble ferric hydroxide (5-3), re-




leasing more acidity to  the  stream and coating  the stream bed.   Ferric




iron  can  also  be reduced by  pyrite itself,  as  in reaction 5-4,  where




sulfide  is  again oxidized and acidity is  released  along with additional




ferrous  iron which may re-enter the reaction cycle via  5-2.




       The concentration of sulfate or acidity in  the water can be di-




rectly correlated to the amount of pyrite which has been dissolved. The




 introduction of acidity into the stream arises from the oxidation of




 S (-II) of the  iron pyrite  (reactions 5-1 and 5-4) and from the oxygena-




 tion of Fe(H)  and  the  ensuing hydrolysis of the resulting Fe(lll)




 (reactions 5-2  and  5-3).  There is  a stoichiometry which should not be




 overlooked: the dissolution of one  ncle of  iron pyrite leads ultimately




 to the release  of four  equivalents  of acidity-two  equivalents from




 the  oxidation of S^-II)  and two  from the  oxidation  of Fe(Il).   The




 decomposition of iron pyrite is among the  most acidic  of  all weathering




 reactions owing to the great  insolubility  of Fe(III) (3).




         Pyritic  agglomerates are thermodynamically unstable upon expo-




  sure to atmospheric oxygen, as demonstrated by calculating the change




  in free energy for  reaction 5-1 as  given in Appendix D.   The exclusion




  of oxygen from the system  may not be a practical  solution to the prob-




  lem from the standpoint of economics, technology, and, as will be  seen

-------
                                                                   5-4





below, chemistry.  Hence, efforts must be aimed at retarding  the  kine-



tics which control the dissolution of the pyritic  material.
5-3  Previous Investigations of the Kinetics and  Mechanism of Pyrite

     Dissolution



       5-3.1  Physical and Chemical Studies
       As a result of the complexity of the  reactions  and 'the  failure



of many previous observers to properly identify,  isolate,  and  control



the rate-determining variables,  much of the  data  reported  previously



is difficult to interpret.   The  complex nature of pyrite itself,  the



variety of forms that sulfur may take as an  intermediate,  and  the



failure to single out the oxidizing  agent all contribute to  the  com-



plexity of the system.



       Stokes (4)  studied the oxidation of pyrite and  marcasite  by



Fe(IIl) and,  by analyzing the end products,  concluded  that the reac-



tion proceeded by  a two-step mechanism




           FeS2(s)  + Fe (S04>3  = 3  FeS04 +  2 S°                  (5-5)




            2 S° -i-  6 Fe0(SO.), +  8 H00 = 12  FeSO.  +  8  H0SO.       (5-6)
                       243      2          4      24



Nelson and Snow (5)  observed that the rate of oxidation of pyritic



sulfur in coal by  oxygen was markedly influenced  by  the degree of



fineness of the coal, the rate being inversely proportional  to the



average diameter of the coal particles.  The addition  of ferric  sul-



fate was found to hasten, the oxidation process.



       Instead of looking at the overall reaction and  the  kinetic vari-



ables affecting it, Sato (6) employed a technique by which electrode

-------
                                                                  5-5






potentials of sulfide minerals were measured in order to elucidate




the oxidation mechanism by which the minerals dissolve.   When both the




oxidation potential of the sulfide test electrode and the oxidation




potential of the solution with which it is in contact are identical,




equilibrium has been reached, i.e., the potential corresponds to  the




equilibrium potential between the mineral and its dissolved oxidizing




ions.  Consequently, the equilibrium potential is controlled by the




first-step reaction: the heterogeneous reaction whereby the sulfide




dissolves.  By measuring the electrode potential of the mineral and




comparing it to the oxidation potentials of the various possible  half-




cell reactions of the sulfide using independent thermodynamic data,




Sato found the oxidation reaction for pyrite which best fits the




measured electrode potentials to be




            FeS2(s) = Fe+2 + S^g) + 2 e"                        (5-7)





for pH values below 2.  This technique, although it does not describe




the kinetics of the oxidation of pyrite, does give some insight as to




the mechanism by which pyrite is oxidized.  If the reduction reaction




could be identified then, by summation of the two half-reactions, one




would have the first step in the dissolution of iron pyrite. As  Sato




explains, once the heterogeneous reaction is established, the other




consecutive reactions for the oxidation of sulfur and ferrous iron can




be treated individually.




       Garrels and Thompson (7) studied the rate of reduction of  Fe(IIl)




by iron pyrite at pH values below 2 and observed that the instantaneous




rate of reduction of Fe(lll) decreased as the ratio of ferric to  ferrous

-------
                                                                  5-6







iron decreased.  The rate was assumed to be proportional  to  differen-




tial adsorption of ferric and ferrous iron on the surface of pyrite,




i.e., to the fraction of pyrite surface occupied by Fe(lll).   In  each




case, fifty percent of the ferric iron was reduced by  two grams of




crushed pyrite in less than one day.




       A number of hydrometallurgical studies have been conducted




dealing with the oxidation of iron sulfide minerals at elevated tem-




peratures and under pressure to further understand the kinetics of




the pressure-leach process employed in the treatment of pyritic ores.




McKay and Halpern (8) investigated the oxidation of iron  pyrite at




temperatures of 100 to 130°C and at partial pressures  of  oxygen of




0 to 4 atmospheres.  The reaction was found to be first order  in  sur-




face area of pyrite and first order in the partial pressure of oxy-




gen.  An attempt was made by McKay and Halpern to study the effect  of




Fe(lll) on pyrite in the absence of oxygen in which they  found that,




although FeS- was oxidized and leached, the quantity of FeS^ oxidized




was small compared to that amount oxidized by oxygen under  similar




conditions. However, since the concentration of Fe(lll) was limiting,




i.e., lO"1*! Fe(lll) was completely reduced by 0.33M FeS  within  two




hours, one should not expect more than r-; x 10" M of pyrite to  be leached




according to the stoichiometry of the reaction (5-4).   Hence,  the in-




ability of Fe(lll)  to compete with oxygen as an effective oxidant of




iron pyrite  was not  justifiably demonstrated,




       Pressure-leaching of  pyrite was  also  investigated by Gerlach,




et  al (9) over a wider  range of temperatures and  partial pressures of




oxygen.  Their results  are  similar to  those  of McKay  and Halpern (8)

-------
                                                                    5-7





with regard to the observed dependencies of the rate law and  the pro-




posed mechanism for the reaction.




       Neither study, however, considered in adequate detail  the oxi-




dation of pyrite by ferric iron which is inevitably produced  by the




process.  Barnes and Romberger (10) also fail to place ferric iron in




its proper perspective when they consider the concentration of dis-




solved Fe(lll) to be so small in acid mine drainage that its  signi-




ficance as an oxidant  is unlikely.  However, in view of the vast  amount




of Fe(ll) oxidized to Fe(lll), there certainly is a significant  supply




of Fe(lll) readily available  as  an oxidant of pyrite.  Even though the




standing concentration of Fe(lll) may be small, its reactivity in terms




of its rate of turnover  is relatively high.  Smith, Svanks, and  Shumate




(11) noted the significance of ferric iron after observing a similarity




between the rates of aerobic  and anaerobic oxidation of iron pyrite,




and postulated that  the ultimate oxidant for both aerobic and anaerobic




oxidation  is  ferric  iron.






5-3.2  Microbiological _Studies




       Since  the first  recorded  isolation  from acid mine waters of




bacteria capable of  influencing  the oxidation of ferrous iron to fer-




ric iron (12), numerous  biological reports dealing with microbial




catalysis  of  the oxidation of ferrous  iron and various sulfide miner-




als have flooded the literature. Temple and Colmer  (13) named their




"iron-oxidizing" autotroph Thiobacillus ferrooxjLdans^ it being capable




of  autotrophic growth  at  the  expense of ferrous  iron or thiosulfate.




This bacterium differs from the  autotroph  Thiobacillus thiooxidans

-------
                                                                   5-8





which also has oeen isolated from acid mine streams but which derives




its energy by catalyzing the oxidation of elemental sulfur.   After




summarizing the known facts regarding autotrophic bacteria and the




dissolution of iron pyrite, Temple and Delchamps (14)  postulated the




overall sequence describing the formation of acid in coal mines as




follows:



       The sulfide of finely-divided iron pyrite or marcasite is




chemically oxidized by oxygen to sulfate





            FeS2(s) + H20 + 3 1/2 DZ = Fe+2 + 2 S04"2 + 2 H+      (5-8)





Ferrous iron is oxygenated, under the catalytic influence of the auto-




troph Thiobacillus ferrooxidans, to  ferric iron





            Fe+2 + 1/4 02 + H+ = Fe+3 + 1/2 H20                   (5-9)





As rapidly as it is formed, ferric iron is chemically reduced by the




finely-divided iron pyrite





            2 Fe+3 + FeS2(s) = 3 Fe+2 + 2 S°                      (5-10)





The elemental sulfur liberated may be oxidized by ferric iron





            2 S° + 12 Fe+3 + 8 H20 = 12 Fe"1"2 + 2 S04~2 + 16 H+    (5-11)





or by oxygen, in which case the reaction is catalyzed by the autotroph




Thiobacillus thiooxidans





            S°.+  1 1/2 0   + H 0 = 2 H+ + S04"2                    (5-12)





The  ferrous  iron  resulting from the oxidation of pyrite by ferric iron




 is then subject to further microbial  action by Thiobacillus ferrooxi-




dans, as  in reaction  5-9.   A cycle  is established  involving formation

-------
                                                                  5-9







of Fe(lII) from Fe(il) by microbial catalysis,  and chemical oxidation




of iron pyrite by the resultant Fe(lII).




       This significant piece of work was obscured by later develop-




ments which questioned some of the physiological properties of  T_.




ferrooxidans,  in particular its ability toward mediating the  oxidation




of thiosulfate and not elemental sulfur.  Two additional autotrophic




organisms were proposed, confusing the  issue regarding autotrophic oxi-




dation of ferrous  iron and  its role  in  acid mine drainage:  Ferrobacil-




lus  ferrooxidans (15), which was able to grow on ferrous iron but not




on thiosulfate or  elemental sulfur,  and Ferrobacillus sulfoxidans (16),




which could utilize  either  ferrous iron or elemental sulfur as an




energy  source. Unz  and  Lundgren (17) concluded that the organisms




were all  nutritionally  similar and called for a re-evaluation of the




current classification procedure.




        This controversy  also hindered verification that the oxidation




of Fe(Il) was a  direct microbial effect and not an indirect one, i.e.,




whether or not the organisms were  true  chemo-autotrophs which derived




their  energy  from the direct oxidation  of Fe(ll).  Little quantitative




evidence  correlating the rate  of growth of the  organisms and the rate




of oxidation  of  ferrous  iron was  available until  Silverman and Lund-




gren (18)  showed that the change  in  the logarithm of the concentration




of F.  ferrooxidans paralleled  the  increase  in  the logarithm of the con-




centration of Fe(IIl) produced by  the oxidation of Fe(ll).  They also




observed  (19) that the quantity of carbon  assimilated was in accord-




 ance with the thermodynamic free energy available from  the oxidation




of  Fe(ll).   Schnaitman (20) demonstrated that  microbial mediation of




 the oxidation of ferrous iron by F.  ferrooxidans  conformed to Michael is-

-------
                                                                  5-10




Menton enzyme kinetics, and that the rate of oxidation was  proportional



to the relative concentration of bacterial cells.



       Numerous reports persist in the literature  concerning  the  nature



of the influence these chemo-autotrophic microorganisms exert on  the



oxidation of pyrite and marcasite, as well as other  sulfide minerals



encountered especially in the copper mines in the  western United  States.



Brynner, et al (21) have attributed leaching of chalcopyrite  (CuFeS^^



covellite (CuS), chalcocite (Cu-S), bornite (CuFeS,),  and tetrahedrite



(Cu0Sb0S_) to direct biological oxidation.   Molybdenite (MoS0)  (22)  and
   o  L  I                                                  t-


orpiment (As~S-) (23) have also been reported to be  subject to  direct



oxidation.  Silverman and Ehrlich (24)  reviewed the  subject of  microbi-



al catalysis of mineral transformations and indicated  that  the  action



of autotrophic microorganisms may be two-fold:   to regenerate ferric



iron from ferrous iron which then chemically oxidizes  the mineral sul-



fide; and to directly attack and oxidize the sulfide minerals inde-



pendent of the action of ferric iron.  Ehrlich (25)  has testified to



the likelihood of such a direct effect, but no mechanism has  yet  been



proposed.






5-4  Purpose of Experimental Study





       Since "at-source" control of coal mine drainage must depend



upon retardation of the kinetics controlling the oxidation  of pyrite,



it is necessary to know which of the sequential reactions  involved



controls the overall  rate of the reaction.  This study was  undertaken



to investigate the relative rates of the various reactions  producing



acid mine drainage,  to ascertain which of the steps is rate-limiting,

-------
                                                                 5-11





and to suggest measures for controlling this particular reaction  in




natural waters.









5-5  Oxygenation of Ferrous Iron






       The oxidation of ferrous iron was discussed in Chapter 3 and




it was seen that,  in the  acidic pH-range corresponding to conditions




encountered in mine drainage waters, the reaction proceeds relatively




slowly and is independent of pH.  However, the composition of actual




mine waters may be such that oxidation  is accelerated, i.e.,  physical




conditions or chemical and biological agents may be present which in-




fluence the rate of oxidation of Fe(ll).  Examples of these include  in-




organic ligands,  such as  sulfate, which complex ferrous and ferric




iron, soluble metal  ions  such as copper(ll), aluminum, and manganese(ll),




suspended material with large surface areas and high adsorptive capa-




cities, such as coal and  clay particles, and microorganisms.   All have




been  implicated in the literature,  in various circumstances,  as being




capable of accelerating the rate of oxidation of Fe(Il).  Consequently,




the oxidation was  investigated  in the presence of many of these chemi-




cal catalytic agents to observe their effect on the rate of oxidation




under synthetic mine conditions and to  compare the observed catalytic




rate with the actual rate of oxidation  of ferrous iron in natural mine




waters.




       The studies were  initially conducted in the absence of micro-




organisms in order to characterize  the  kinetics of the reaction in




purely chemical terms.

-------
                                                                  5-12




 5-5.1  Experimental Procedure




       The  experimental  procedure and analytical techniques were simi-




 lar  to those  employed  in Chapter 3 to follow the oxidation of ferrous




 iron with time.  For the slower long-term studies, the absorbances of




 acidified aliquots of  the  samples were measured at the isosbestic




 point of the  system (where Fe   and FeOH   have the same molar absorp-




 tivity)  as an  indication  of the quantity of ferrous iron oxidized.




 Figure 5-1  shows that  the  molar absorptivity of the acidified solution




 containing  Fe  and FeOH+  , at 272 mu,  is unaffected by concentrations




 of sulfate  as high as  10" M.




       For  the  short-term, more rapid catalytic studies, ferrous iron




 was  determined  directly  by titration with permanganate, or by chelation




 by the colorimetric reagent bathophenanthroline (4,7-diphenyl-l,10-




 phenanthroline) (26).  In  the surface-catalytic studies, aliquots of




 the  heterogeneous suspension were acidified with dilute perchloric acid




 and  filtered.  The solid material recovered was rinsed with dilute




 acid to remove  all traces of ferrous iron that might have been adsorbed




 on the surfaces of the particles.




       Samples were prepared containing various dilutions of a standard-




 ized solution of ferrous perchlorate and different concentrations of




 the  catalytic agents.   The effects of sulfate,  aluminum,  manganese(ll),




 copper(ll),  powdered charcoal,  alumina,  silica, aged ferric hydroxide,




 crushed iron pyrite,  and the natural clays bentonite and kaolinite were




 examined.  The pH was adjusted with concentrated perchloric acid and




 the  samples were allowed to equilibrate with the oxygen of the atmosphere.




After removing aliquots for the above-mentioned analyses,  the solutions

-------
                                                                5-13
   1.6
    1.2
s
    0.8
W
o
o
en
    0.4
                               [SO,"2] » 10"3M  A
                               [so ~2J = io"2M  n
                                            [SO,   1 •  10  M O
                  molar absorptivity in the absence  	
                  of  sulfate,  from Figure 2-2
                    2           4            "4
                         Fe(III) CONCENTRATION, x 10
        FIGURE 5-1.  Effect of  sulfate on absorbance of Fe(III) at
                     272 0*1.
       0.0
                     i—~T
      -i.o  K—
       -2.0
                     k" - - d log [Fe(II)l
                                 dt
                        p  «= 0.20 atm.

                        Temp. = 25 C
       -3.0
       -4.0
                                                      Stumm-Lee
                                                     rate law
                                                    extrapolation
                      j	L
0        1

FIGURE 5-2.
                                         3        4
                                        pH
                         Rate of oxygenation of ferrous iron as a
                         function of pH.

-------
                                                                  5-14





were re-equilibrated with the atmosphere.  The samples were sealed and




stored, quiescently, in an  incubator at 25°C in the dark to avoid any




photochemical  interference.  In the investigations involving hetero-




geneous catalysis  (clays, pyrite, and powdered charcoal), the suspen-




sions were stored  on magnetic stirrers and shielded from light.   One




study was conducted under sterile conditions to preclude the influence




of microbial catalysis which may have been accidentally effective




through contamination.






5-5.2  Experimental Results and Discussion




       Figure  5-2  summarizes the results  reported  in Chapter 3 for




the uncatalyzed rate of oxidation of  ferrous  iron.  This figure serves




as the basis for comparison with the  catalyzed rates.






Effect of Sulfate




       The rate of oxidation of  ferrous  iron  in the presence of sulfate




can be  satisfactorily  fitted to  a rate relationship which  is first-




order  in the concentration  of ferrous  iron, as shown in Figure 5-3.  The




rate constant  k" = -d  log,Q [Fe(ll)]/dt  is of the  same order of magni-




tude and is quite  similar to the rate constant for the uncatalyzed re-




action.  To magnify the effect of any catalytic dependence of the oxi-




dation on the  concentration of  sulfate,  the study  was repeated at 50 C




where the change in Fe(Il)  per unit  time is greater.  Figure 5-4




demonstrates more  vividly the catalytic  effect of  sulfate.




       Huffman and Davidson (27)  investigated the  oxidation of Fe(Il)




in  solutions of sulfuric  acid,  at 30.5°C  and  1 M H^SO,,  and observed




the reaction to be second-order  in Fe(ll) according to the rate law

-------
                                                                      5-15
z
B
H
8
   -3.010
   -3.014
P
M

T -3.018
to
   -3.022
pH 2.9

[Fe(II)] - 9.78x10  >[
    -2*     -3
      ] = 10 J
                   25°C


                   PQ » 0.20 atm.
                              :" - 1.2x!0"4 day"1
                      20
                                      80
                                                       100
                       40           60

                           TIME, days

FIGURE 5-3.  Rate of oxidation of ferrous iron in the presence of

             sulfate.
        -2.50  L
                         40
                                         160
                       80          120

                         TIME, days

FIGURE 5-4.  Effect of sulfate on the oxidation rate of

             ferrous iron at 50°c.
                                                              200

-------
                                                                    5-16
 where the termolecular rate constant k  = 2.8 x 10"  liter  mole



 atm  sec  .  Under a partial pressure of oxygen of 0.2 atm.  and  an



 initial concentration of Fe(ll) of 10~ M, and converting  the units of



 time, the initial rate "constant" of Huffman and Davidson can be approxi



 mated as k' = 4.8 x 10~5 day'* (k'= k [Fe(ll)] Prt  =* -d In  [Fe(lD] /dt)
                                      t        o 0_                o


 or  k" = 2.0 x 10   day  .   This pseudo-first-order rate  "constant"


 can now be compared with the rate constants obtained in this study.



        The observations of  Huffman  and Davidson in sulfuric  acid show



 the oxidation to occur more rapidly in a solution containing sulfate


 than in a medium of perchlorate as  investigated by George (28).  The



 latter conducted his  study  of  the oxidation of Fe(ll)  in  perchloric



 acid and  also observed  the  reaction to be second-order  in Fe(ll). (See



 Section 3-3.1.)  Performing a  similar calculation  as above,  the pseudo-



 first-order rate "constant" obtained by George  is  k" =  1.0 x 10   day ~*


 at 30°C and approximately 10~T1 HC10 .



       The results both of Huffman and Davidson and of George are of


 the  same order of magnitude as the results obtained in this  study,



 except that these authors have characterized the oxidation as being



 second-order in Fe(ll).  For such an investigation, where the reaction


 proceeds  so slowly that only 2-3% of the reaction is complete after



 three months, it becomes difficult to classify the reaction  with



 respect to  its order, as previously discussed in Chapter 3.   The



 agreement among the pseudo-first-order rate constants is gratifying


 in itself,  George and Huffman and Davidson also observed the rate  of



oxidation to increase only slightly with an increase in pH.

-------
                                                                 5-17


       Figure 5-4, and a comparison of the results of Huffman and

Davidson to those of George suggest that the oxidation of ferrous

iron takes place more rapidly in the presence of sulfate than in  per-

chlorate alone, the magnitude of the catalysis, however, being not

very great.


          'Dy Dissolved Metal Ions
       Of the dissolved heavy metals which are normally present in

natural mine waters,  copper (II)  exerted the strongest catalytic in-

fluence as  shown  in Figure 5-5.  For the sake of convenience, and since

the data fit the  formulation fairly well, the rate has been plotted as

a reaction  which  is first-order  in Fe(ll).  The pseudo-first-order

rate constant of  k" = 4 x 10" day~  indicates that 10~ M Cu(Il) ac-

celerates the rate of oxidation  of Fe(ll) approximately four-fold.

       The  cupric ion has been reported to be an efficient catalyst

in the oxidation  of Fe(Il) in solutions of phosphoric acid (29),

sulfuric acid (27), hydrochloric acid  (30), and in the neutral pH-

range corresponding to natural waters  (31).  Cher and Davidson (29)

account for the catalytic effect of copper (II) by the following

mechanism whereby Cu(ll)  serves  as  an  electron-transfer catalyst:


             Fe(Il) +  Cu(II)  = Fe(lll)  + Cu(I)                   (5-14)
                                    •
             Cu(l) 4- 0£ =  Cu(ll)  + H02                           (5-15)

                   <•
The free radical  H0_  reacts  further with  additional Fe(Il) as in the

Weiss scheme (32) previously discussed in Chapter 3.

       The  rate  law  for the  oxidation  of Fe(Il) in solutions of sulfuric

acid containing Cu(Il) was reported by Huffman and Davidson  (27) to be

-------
                                                                   5-13
55
s
I
g
W
K
8
fl-
    -2.05
    -2.07
-2.09
-2.35
     •2.37
     -2.39
                                   pH 3.0             3
                                   [Fe(II)]o= 8.74x10  M
                           10"AM Cu(II)
                                k" - 4.0xlO~4 day"1
                pH 3.0             3
                [Fe(II)lo= 4.49x10 M
                      20
                                                          80
                                                                 100
                           40          60
                             TIME,  days
    FIGURE 5-5.  Effect of copper(II)  on oxidation of ferrous iron.
  10

   9
   8

   7
I
                               D
                                         [Fe(II)]
                                             - 10"2M
        A1203

        pH 3.7

        k"
                     8250 ru /I
                                             - 5xlO"3M
                  S.OxlO"3 day"1
                        10
                                               25
                                                            30
                                                              35
                           15      20
                           TIME> days
FIGURE 5-6.  Rate of oxidation of ferrous iron in the presence of
             suspended aluminum oxide.

-------
                                                                 5-19


first-order in the concentrations of Fe(Il)  and Cu(II),  Figure 5-5


conforming to such a description.  The magnitude of  the  rate constant


obtained in this study, however, cannot be compared  with that of Huff-


man and Davidson in sulfuric acid since, in contrast to  solutions con-


taining no Cu(ll), the rate of oxidation increases markedly with  in-


creasing concentration of  acid (33).  Indeed, the'results of Huffman


and Davidson, in 0.11M (H+) and  a total activity of  sulfate of  0.58M,


show the rate of oxidation of Fe(ll)  in the presence of 10" M Cu(II)


to be about 100 times greater than  depicted by Figure 5-5.


       Although Stumm and  Lee (31)  observed other heavy metal  ions


(Mn   , Co+ )  to exhibit  similar  catalytic effects on the rate of  oxi-


dation of Fe(ll)  at neutral pH values,  it was found that Mn   and Al


showed no measurable  acceleration of  the reaction rate for the  acidic


conditions of this study.



The Effect of Clays


       On  the premise that surfaces of  clays play a significant role


in the oxidation  of  ferrous  iron in natural mine waters,  a study of


the  catalytic effect  of silica,  SiO_, and  alumina,  Al-O^, was under-


taken.   Silica  and alumina,  which are the  basic building  blocks of


all  clays  (clays  being composed of  two  dimensional  arrays of silicon-


oxygen tetrahedra and aluminum or magnesium-oxygen-hydroxide octa-


hedra (34)),  were selected as idealized representatives of natural


clays.   Hydration of the surface of the clay results in the formation


of  silanol,  Si-OH, and aluminol, Al-OH,groups.   Since the rate of oxi-

                                                                _ 2
 dation of ferrous iron at higher pH-values is  dependent upon ODH ]


 (see Chapter 3),  it was believed that the hydroxo-metal groups of the

-------
                                                                 5-20
hydrated clay surface played a specific catalytic role due to  the  ap-

parent localized high PH (high concentration of OH groups) at  the  par-


ticle surface.  Furthermore, since clays are strongly adsorptive and

exhibit  ion-exchange properties, a general catalytic influence was

thought  to occur as a result of adsorption of the reactants and local-

ized  increased  concentrations of reactants at the particle surface.

In  addition,  Schenk and Weber (35) observed orthosilicic acid, H4SiC>4,


with  which natural clays are in equilibrium, to increase the rate  of

oxidation of  Fe(ll) at pH-values greater than 5 in much the same manner


 as  other inorganic ligands, such as phosphate, chloride, and sulfate.


        Of the clay surfaces investigated, aluminum oxide  (Baymal


 colloidal alumina, manufactured by E. I. DuPont de Nemours and Co.,

 Wilmington, Delaware) exhibited the strongest catalytic properties.


 Figure 5-6 shows that, for  a given pH and concentration of alumina,

 the surface-catalytic oxidation can be described by a rate equation


 which is first-order  in concentration of Fe(ll). The slope of  the

 first-order plots  (which is, by definition,  the rate constant)  is  in-

 dependent of  the  initial concentration of Fe(Il) which  further confirms

 the order of  the reaction with  respect to Fe(Il).  The  rapidity of

 the surface-catalytic reaction  compared  to  the uncatalyzed oxidation


 is demonstrated by noting the half-time  of  the former (the time re-

 quired for 50% of  the  initial concentration of ferrous iron  to be

 oxidized) to  be only  about  40 days,  in contrast  to about 500  days for

 the latter.  Table 5-1  compares the surface-catalytic rate constants


 for alumina with the uncatalyzed rate constants for the oxidation of

                                                            2
 Fe(Il),  and  it is seen that in the presence of about 8000 m /I of

-------
                                                                5-21
     Table 5-1.   Comparison of Surface-Catalytic  Rate Constants
                  with Uncatalyzed Rate Constants


pH

3.5
3.8
4.0
log k"

Uncatalyzed
Reaction
-3.6
-3.4
-3.3

A12°3
8000 m*/l
-2.52
-2.05
-1.80
                    k"
-d log [Fe(II)]
       dt
         *Baymal Colloidal Alumina - manufactured by E.  I. DuPont
de Nemours and Co., Wilmington, Delaware.
Al O  surface, the catalyzed reaction is 10-30 times faster than  the

uncatalyzed reaction.

       Figure 5-7 demonstrates  further the direct catalytic dependence

on A1203, showing the  rate of oxidation of Fe(ll) to increase as  the

areal concentration  of the idealized clay is  increased.  The oxidation

rate  in  the presence of alumina is also dependent upon pH as seen in

Figure 5-8, with a  regular increase in rate with increasing pH being


observed.

-------
                                                                       5-22
1

§
g

I
8
M
>^

-------
                                                                5-23
     -1.8   	
     -2.0
   o
   s ..
2.2
      -2.4
      -2.6
                   k" - - d log  [Fe(II)l
                                 dt
            FIGURE 5-8.   Effect of pH on surface-catalytic oxidation
                         of ferrous iron.                   .
   -3.0
o
I
   -3.1
   -3.2
   -3.3
                        <
                           1800 m 11 Ludox SM
                                colloidal silica
                                          pH 3.65
                     m'/l Ludox SM
                       pH 4.1
                          10  gms/1  reagent
                          bentonite clay
                                 pH A.O
                                                             D
                       10
                                               30
                                                                   40
         FIGURE 5-9.
                20
           TIME, days
Rate of oxidation of ferrous iron in the presence
of colloidal silica and bentonite clay.

-------
                                                                   5-24
       (The catalytic effect of Al 0  cannot be attributed  to  specific




catalysis by dissolved Al(lll) in equilibrium with the solid since  it




was previously shown that dissolved Al(lll) had no measurable  effect




on the rate of oxidation of ferrous iron.)




       Ludox colloidal silica (Ludox SM is a colloidal silica  manu-




factured by E. I. DuPont de Nemours and Co., Wilmington,  Delaware)




and bentonite, a natural montmorillonlte clay (Bentonite Powder,




U.S.P., Fisher Scientific Co., Fair Lawn,  New Jersey) were also found




to catalyze the oxidation (Figure 5-9) but to a slightly lesser degree




than the alumina.  Kaolinite, another natural clay, demonstrated no




catalytic properties although it gradually neutralized the acidity.




       Colloidal ferric hydroxide, prepared by allowing a solution  of




Fe(lll) oversaturated with respect to the  hydroxide to hydrolyze,




showed no tendency toward catalyzing the oxidation of Fe(Il).







Catalysis by Powdered Charcoal




       Lamb and Elder (33) reported that granular, steam-activated




coconut charcoal markedly accelerated the  rate of oxygenation  of




ferrous iron.  This phenomenon was attributed to the ability of char-




coal to greatly enhance decomposition of hydrogen peroxide in  the




presence of Fe(ll), generating in turn the active free radicals which




oxidize Fe(ll) according to the Weiss mechanism (32).




       In this study, reagent-grade sugar  charcoal(purified reagent,




 Fisher Scientific Co., Fair Lawn, New Jersey) displayed no apparent




catalytic tendency toward the oxygenation of Fe(ll).  Again, however,




the catalytic properties of charcoal described by Lamb and Elder in-




creased considerably as the concentration of acid increased, thus

-------
                                                                 5-25



accounting in part for the discrepancy between their results  and those


reported here.  Furthermore, Lamb and Elder followed the oxidation of


Fe(ll) using an electrochemical technique whereby the change  in elec-


trochemical potential was correlated to a change in the Fe  /Fe


ratio by a modified Nernst equation.  In the heterogeneous system con-


taining suspended charcoal, the reliability of such measured  potentials


as indicators of the total concentration of Fe(ll), both in solution


and associated with the  surface of charcoal, must be questioned


especially when the rate of change of Fe(ll) is so small.  The presence


of impurities in the charcoal also tend to cast doubt on the experi-

                    *
mental reliability.



Effect of Iron Pyrite


       To complete  the  investigation of all natural chemical catalytic


agents which may be responsible for acceleration of the rate of oxi-


dation of ferrous iron  in mine waters, the effect of iron pyrite  itself


was studied.  It has been suggested by Smith, et al  (11) that the


surface of pyrite acts  as a catalyst to greatly increase the rate of


oxidation of adsorbed ferrous ions.


       Mineral iron pyrite  (Ward's Natural Science Establishment,


Rochester, New York) was ball-milled and a portion of the 200-250 mesh


fraction was suspended  in a solution of ferrous iron at pH 3.0.  The


system was left open to  the atmosphere and treated in the same fashion


as the suspensions  containing clays, aliquots being removed,  acidi-


fied, millipore-filtered, and titrated with permanganate.  The solid


material recovered  was  rinsed with dilute acid to remove any adsorbed


ferrous  iron.  No significant decrease in Fe(ll) was observed after

       3t
         Researchers at  Bituminous Coal Research, have found that  some

 activated carbons have  catalytic/ properties for ferrous  iron oxidation

 while others  do  not.

-------
                                                                      5-26


several days although in some instances the concentration of Fe(ll)


increased slightly.  Upon placing the system under an atmosphere of


nitrogen, a similar small increase in ferrous iron was observed, indica-


tive of a gradual dissolution of the pyrite itself.




Effect of Microorganisms

                                  tv<*£
       No microbial contaminations^observed in any of the above systems


as evidenced by similar rates of oxidation in both sterile and nonsterile


samples.  The sterile samples gave, in every case, analogous results to


those conducted under nonsterile conditions.  The above findings are,


therefore, the same as those obtained under sterile conditions.
       Table 5-2 summarizes the results of the catalytic studies with


the various chemical agents .  The greatest influence on the rate of


oxidation of Fe(ll) was exhibited by the clay particles or their


idealized counterparts, alumina and silica.  The areal concentrations,


however, are extremely large and are probably much greater than those


encountered in most natural waters .



5-6  Field Investigations of Pyrite Oxidation in Natural Mine Waters


       In order to compare these experimental results describing the


kinetics of oxidation of ferrous iron in synthetic mine waters with


the  rate of oxidation in natural mine drainage waters , field investi-


gations were  conducted in the bituminous coal region of West Virginia.


The  Federal Water  Pollution Control Administration has established a


demonstration project  in the Norton-Coalton  area in north-central  West

-------
                                                                5-27
     Table 5-2.  Chemical Catalysis of Oxidation of Ferrous  Iron

pH
3.0
3.5
3.8
4.0
log k"

Uncatalyzed
-3.8
-3.6
-3.4
-3.3
Catalysis By
10 "^ S04"2
at 50°C
-3.1
—
—
—
10 ~4M
Cu+2
-3.4
—
—
—
A12°3*
8000 m2/!.
—
-2.5
-2.1
-1.8
Si02t
3000 m2/!.
—
—
—
-2.2
Bentonite#
10 gms/1.
—
—
—
-2.2
                       k" =
-d log [Fe(Il)]
       dt
       *Baymal colloidal alumina, E. I. DuPont de Nemours and Co.,
Wilmington, Delaware
       *Ludox SM colloidal  silica, E. I. DuPont de Nemours and Co.,
Wilmington, Delaware

       ^Fisher Scientific Company, U.S.P., Fair Lawn, New Jersey
Virginia, near Elkins,  where various methods of pollution abatement

technology  are being examined.   These  attempts include air-sealing of

mines  and surface reclamation.


5-6.1   Collection and Analyses  of Samples

        Three sites were selected for  the field investigation.  Figure

5-10a shows an abandoned strip  (surface) mine which  is scheduled for

-------
                                                                    5-28
        Highwall
                                             Mine  wall
streams     Yellow
                                                         Wooded
                                                           area
                                                  (Steep downhill
                                                       grade)
                                                    To Tygart
                                                    Valley River
FIGURE 5-10a.
        Aerial  view of drainage  through  a  strip  (surface) mine  in
        the Mercer  seam near  Elkins, West  Virginia.  The points
        (•) refer to locations at which  samples  were collected
        for laboratory analyses.

-------
                                                                   5-29







reclamation.  Surface runoff flows through a vegetated area,  over  a




high-wall, and into the abandoned mine.  In mining terminology,  this




worked-out mine is designated as a "strip pit"; the piles of  waste




material previously dug out  are termed "spoil banks"; and the precipi-




tated ferric hydroxide or  sulfato-hydroxo intermediate is referred to




as "yellow boy."  Drainage water flowing through the mine deposits




thick sediments of yellow  boy at each juncture with drainage water




from the  spoil banks  which are  laden with pyritic material.  The water




exits the strip pit  at  site 7  and  flows  rapidly downhill through a




wooded  area to Roaring  Creek,  the  main interceptor for this mining




region.



        Figure 5-10b  represents a contour strip mine  where  drainage water




passes  between the mine wall and a spoil bank.  Water which  has drained




through the underground mine trickles out of the  mine opening at  site 2.




        The third sampling location, Figure 5-10c,  is a mine  entrance




 which has been sealed with concrete blocks and urethane foam to exclude




 air in an attempt to inhibit oxidation of pyrite inside the  underground




 mine.   Measurements by the FWPCA indicate that the partial pressure  of




 oxygen behind the wall has been reduced to 77. (36).  Drainage water




 flows out of the mine over  a weir,  and  is exposed to oxygen of the




 atmosphere.  After leaving  the mine tunnel,  the water enters a re-




 claimed  area which has been limed  in preparation for future vegetation.




 Deposits of yellow boy are  abundant behind  the weir  and on the floor





 of the tunnel.

-------
                                                                    5-30
                   Hlghwall
       Yellow boy
                                      Mine opening - source of
                                             acidic discharge
                                                    Mine
                                                    boundary
                                       Water goes
                                       underground
FIGURE 5-10b.
              Aerial view of surface mine near Elkins,  West Virginia
              (Site GT 7-2).  The points refer to sampling locations.
         Underground
             mine
             7% 0,
                         Urethane foam
                         and concrete blocks
                         to seal mine
                                    Weir
                                                Atmospheric
                                                  oxygen
         Mine floor
                                         Drainage
                                Yellow boy'
                                  deposits
                                                 water
Water flows
out of mine
   tunnel
FIGURE 5-10c.
                Profile of air-sealed entrance to- underground mine near
                Elkins, West Virginia (Site RT 9-11) .

-------
                                                                 5-31
       Samples from the three mining areas were analyzed  for ferrous

and total iron, sulfate, and acidity; temperature and pH  were re-

corded in the field at the time of collection.  Those samples taken

for iron analyses were acidified with dilute acid upon collection,

to quench any further reaction.  Ferrous and total iron were deter-

mined utilizing bathophenanthroline  (26).  The burbidimetric pro-

cedure described  in Standard Methods for the Examination of Water and

Wastewater  (37) was employed in the  analysis  for  sulfate, the percent

transmittance of  the  suspension of barium sulfate being measured using

a Bausch and Lomb spectrophotometer  (Model #340).  The method of

Salotto, et al (38),  in which the sample is  rapidly  titrated to pH 7.3

with standard sodium hydroxide, after addition of hydrogen  peroxide,

was employed for the determination of acidity.

 5-6.2  Results of Field Investigation
        Stoichiometric Relationship  Between Sulfate Concentration and
        Acidity

         It was indicated in  section  5-2 that the  stoichiometry  accom-

 panying the dissolution of  iron pyrite should conform to a definite

 pattern, the oxidation of one mole  of pyrite  causing  the release of

 two moles of sulfate  and four equivalents of  acidity.  Therefore, if

  sulfate and acidity were to behave  in  a  conservative  manner in mine

  waters,  the change  in either one or'both of  these products could be

  directly correlated  to the quantity of pyrite dissolved.
         A summary of  the data obtained by analysis of  the field samples

  is given in Table 5-3.  For site GT7-2,  a large increase in acidity,

  sulfate, and iron is seen at sampling site 4 after  water from the mine

-------
Table 5-3.  Summary of Field Data
Sample
Number
GT 7-2 1
3
4
5
6
2
	 i 	 . — _— •••••••
•••••{•••••i^^^^™™™™^^^^
MS - 1
—2
3
4 I
5
6
7
8
9
RT 9 -11
.1! 	
Distance,
Feet
0
50
75
300
500

^ — ~
0
100
300
425
550
800
1000
1200
1400
-
Temp. ,
°C
13
16
17
18
15
12
14
19
20
--
18
20
18
15
16
--
PH
5.2
4.2
3.7
3.7
3.8
2.8
3.9
3.6
3.4
3.4
3.4
3.3
3.2
3.5
3.3
3.2
Fellj
x 105M
0.12
0.15
1.74
1.67
0.80
7.92
0.07
3.49
5.63
6.50
8.54
5.94
15.5
11.3
7.90
78.6
FeT
x 105M
0.22
0.25
4.07
2.71
1.67
98.9
0.12
5.68
13.8
10.7
11.3
11.9
42.1
38.4
34.3
132.
-2
sv
x 10 M
0.76
0.86
2.49
4.44
4.32
29.0
-
7.1
19.9
25.8
33.8
34.3
33.7
57.8
49.0
56.8
78.0
Acidity
x 104 eq/1
i i
1.8
.0
. /
5.0
4.5
41.7
awMOBsnaiaBKas
7.5
24.8
26.0
21.7
22.4
23.1
56.5
4o. ->
60 .0
49.8
                                                                   I
                                                                   to

-------
                                                                 5-33
opening has drained into the stream.  The same observation can be




made in the drainage from the Mercer Seam immediately downstream




from the spoil bank.  The low PH of the  surface runoff upstream  from




the mines  should be noted.  Due to  the low buffer capacity of the




water, the mere  introduction of organic  acids  from the soil and  vege-




tation is  sufficient  to significantly  lower  the pH of the raw water.




       Figure 5-11 is an attempt  to correlate acidity to  concentra-




tion of  sulfate,  the lines drawn representing the theoretical 2/1




ratio.   As shown,  the field data do not conform to such  a correlation,




 the acidity being too low.  Some of the acidity is apparently being




 neutralized by clays which are invariably present in the drainage




 waters and coal strata.  The presence of dissolved aluminum in  the




 acidic waters is  indicative of decomposition  of the clays and neu-




 tralization  of  acidity by  the clay-water  interaction.  Hence, sulfate




 alone may be considered to behave  in  a  conservative fashion in acid




 mine waters, in contrast to  acidity and iron, and therefore  it may be




 used, by itself,  as an indicator of the quantity of pyrite dissolved.







 Rate of Oxidation of Ferrous Iron.





         Of the mining areas sampled, RT 9-11  and the Mercer  Seam proved




  to be the most informative.  RT 9-11 served  as a source of  mine water




  having a high concentration of  ferrous iron, while drainage from the




  Mercer Seam provided  an opportunity by which the change in concentra-




  tion of  Fe(Il) could  be followed  thorough a strip mine and downstream




  in an  attempt  to deduce the  rate  of  oxidation of iron.

-------
                                                                5-34
        50

        40

       30
Drainage":
from mine
opening
                                        GT 7-2
                                           I	I    J
          0        2         4        6 v   20  30   40   50
                    SULFATE CONCENTRATION,  x 10 M
          FIGURE 5-lla.  Stoichiometric relationship between
                         acidity and sulfate concentration
                         in drainage water  at location GT 7-2.
a1 4
«
X J
 •\
B2
a2
o

  1

  0
                   SULFAT.E CONCENTRATION,  x'10 M
     FIGURE 5-llb.   Stoichiometric relationship between acidity
                    and sulfate concentration in drainage  through
                    strip mine i«-Mercer seam.

-------
                                                                 5-35
Comparison with Laboratory Results




       The data from Table 5-3 are plotted in Figure 5-12 showing the




change in chemical composition of a  drainage stream flowing through




the Mercer mine and downstream before  entering Roaring Creek.  For the




initial 1000 feet, the  water  continues to collect leachate from the




spoil banks and from  the pyrite  with which  it  is  in contact.  The in-




crease  in the  concentration of  sulfate from 0  to  400  feet  is about 3 x




lO'2!-!.  This  implies  that the corresponding input of  ferrous iron would




increase  by  approximately 1.5 x lO'^l bearing  in mind the  stoichiometry




that one  mole of  pyrite releases two moles of  sulfate and  one mole of




ferrous iron.   However, most of this Fe(ll) is apparently  oxidized quite




 rapidly as evidenced by the  abundant  deposition of yellow boy  and by




 the low concentration  of Fe(Il) observed in the stream (only 8  x  10  M




 at 550 feet).



        Beyond 1000 feet, the drainage stream  flows outside the mine and




 is no longer  subject to continuous  pyritic  discharges, or to dilution




 from additional  drainage streams.   Figure  5-13  shows the oxidation of




 Fe(Il) in the non-pyritic  wooded  area to be essentially first-order in




 Fe(ll) with distance.   If  one  assumes a constant velocity of flow over




 this distance in the range of  1 ft. per min.  to 100  ft. per min., the




 range  of  the  first-order rate constant, k", is 7.4 x 10"  to 7.4 x 10




 min.'1,  where k" = -d log [Fe(ll)]/dt.  This  corresponds  to k"-1 x 1(T




  day"1  which is orders of magnitude greater than the laboratory experi-




  ments  at PH 3.3 would predict.  The  velocity of flow is probably not




  constant but for the  sake of comparison, in order to underscore  the




  rapidity of the reaction  in natural  mine waters, it is a useful  approxi
  mat ion.

-------
  JS
 t
  o
CM

 I
 CO
     0





     6
 >
 %
 s   2
 0




40








20
                                                                        5-36
                     300
                                 600
900
1200
                                      DISTANCE,  ft.J
          SULFATE
              ACIDITY
             TOTAL IRON
       0"            300            600           900            1200

      FIGURE 5rl2.  Chemical  composition, of drainage  water through strip

                    mine  in Mercer  seam.

-------
                                                           5-37
 £
n
 O
 ss
 o
 w
 I
 **•   7
 M   '
 
-------
                                                                  5-38





       Figure 5-14 presents the results obtained at the air-sealed



mine opening at KT 9-11.  Curves 2 and 3 show the decrease in Fe(Il)



by oxidation after collection of the samples and allowing them to



stand back in the laboratory exposed to the atmosphere.  Aliquots were



removed at various intervals and titrated with permanganate, sample



number 1 being  acidified at the time of collection to serve as a con-



trol.  Samples  2 and  3  were not acidified.



       The linearity  of the arithmetic plot indicates the oxidation of



Fe(ll) in  its native  solution to be zero-order in Fe(Il), i.e., the



instantaneous rate is independent of the concentration of ferrous iron.



The rate of  reaction  is dramatic when compared to the previous labora-



tory investigation.





 Implications of Field Results




       The zero-order nature of the oxidation reaction is suggestive



of a biological reaction  in which substrate is non-limiting and  in which



 the concentration of  microorganisms remains relatively constant.  The



 rate equation  for  such  conditions has been derived in Appendix E  as



             -dS   M,    B

             ^__ _ _292— _ constant                            (5-16)
               dt      y



 where S  is the concentration of  substrate (source of energy),  in



 this case ferrous iron.M     is  the maximum specific growth rate for
                          max


 the microorganisms,  v_ is  the yield of microorganisms per unit  of sub-



 strate utilized,  and ji  is the  instantaneous concentration of microorgan-



 isms, assumed here to be  constant.  Figure  5-14  satisfies equation 5-16



 and therefore the results imply  that  the  oxidation is  being catalyzed



 by microorganisms which are  utilizing the energy derived from the

-------
                                                                    5-39





oxidation of ferrous  iron  for cellular metabolism, i.e.,  catalysis




by autotrophic bacteria  is taking place.




       Microbial catalysis,  however,  is not the only explanation for




the zero-order nature of the observations.  A heterogeneous reaction




mechanism could be  invoked,  involving complete saturation of the solid




phase with the reactant, Fe(Il),  in order to account for the zero-order




dependence on Fe(Il). However,  for the additional reasons described




below, the autotrophic explanation  is an extremely plausible one.




       As shown  in  Appendix F,  a thermodynamic free  energy balance




appears  to negate the existence of  such autotrophic  iron bacteria;  only




1  gram of organic carbon is synthesized for every 250 gms of ferrous




iron oxidized.   In  fact, the autotrophic nature of these organisms




and their  ecological significance was initially doubted due to the




meager  amount of free energy available  from the oxidation of Fe(ll).




        Since the energy released is so  small,  one would not expect




the oxidation of 10~ M Fe(ll)  to significantly change the bacterial




concentration if a sufficient number of microorganisms were present in




the mine water at RT 9-11,  i.e., JB should remain  constant.  Figure  5-14




reflects such reasoning.  Kim (39), of the Pittsburgh Mining Research




Center  of the Bureau of Mines, has obtained similar  zero-order plots




of oxidation of Fe(ll)  in natural mine waters.




        If,  however,  the concentration of bacteria were diminished,  as




by filtration of the mine water, B^ would be expected to  increase




 logarithmically (see equation E-13 in the Appendix)  as the substrate,




Fe(Il), is utilized.  A sample of mine water from RT 9-11  was millipore-




 filtered (0.8 u pore diameter) and the resultant  degree  of oxidation  of

-------
                                                                    5-40



Fe(Il) was markedly reduced due to removal of a significant fraction of


the bacteria.  The unf iltered sample was completely oxidized within 20


hours, while the filtered sample displayed a lag before significant oxi-


dation began (see Figure 5-15).  If JB is not constant during the course


of the reaction, then



            -**   "maxn    "max11                                 (5-17)

            ~dt = T  o e


and,  as  derived in Appendix E

                     B      ju   t
            s  _ s = _°  e   "^                                  (5-18)
             o        y

where B    is the initial concentration of bacteria, and S  is the ini-


tial  concentration of substrate, Fe(II).  Taking logs of both sides one


obtains
             log (S  -S)  =  log  2. + J.                           (5-19)
                  o          y    £*• j

 Figure 5-16 is a semilogarithmic plot of the change in concentration of


 Fe(ll) with time, the linear nature of the plot confirming the logarith-


 mic change in B in  accordance with equations 5-17 and 5-19.  Equation


 5-19 indicates that the slope of the semilog plot is equal to  max/2. 3.


 Consequently, "max  is 0.076 hrs"  for the experimental conditions.


 The generation time, the  time required for the concentration of bacteria


 to double, is, by definition, equal to In 2/JU or 9.1 hours.  Silverman


 and Lundgren (18) observed generation times of about 7.0  hours  in their


 laboratory under ideal experimental conditions.


        To further substantiate  biological significance,  sterile solu-


 tions of ferrous sulfate were  inoculated with various  amounts of acid

-------
                                                                    5-41






mine drainage.  Two sterile controls were maintained:   one in which




aseptically filtered mine water (220 mu pore diameter) was added  to




sterile solutions of ferrous sulfate; and another containing sterile




ferrous sulfate alone.  Aliquots were removed aseptically and analyzed




for residual Fe(ll) by titration with permanganate.  Figure 5-17




shows a linear decrease in Fe(ll) for the arithmetic plot, as in




Figures 5-14, but only for the specimen containing microorganisms.   The




slowness of the oxidation reaction  in the laboratory  samples compared




to that in the field samples (Figure 5--17 compared to Figure 5-14)  may




have resulted from a decrease in the concentration of some essential




growth factor in the course of the  dilution.




       Therefore, the oxidation of  ferrous  iron occurs more rapidly in




natural mine waters than  in any of  the synthetic solutions investigated




in the laboratory subject to the various chemical catalytic additives.




The rapidity of the reaction in nature is apparently  the result of




microbial catalysis, as evidenced by Figures 5-14 to  5-17.

-------
                                                           5-42
      8
   X
    ^ 6
   55
   s
   w  4
   §
\
         0        10

         FIGURE 5-15.
                                sampled  from weir  and
                                millipore  filtered (0.8(j.)
                                sample from weir,
                                unacidified and  unfiltered
              20       30         40        50
              TIME,  hrs.
        Oxidation  of ferrous  iron in water
        collected  from air-sealed mine.
                       20          30
                         TIME, hrs.
FIGURE 5-16.  Change  in ferrous  iron concentration  in  milli-
              pore filtered water collected  from air-scaled
              mine.

-------
                                                               5-43
   12
   10
I
3
g
w
o
u
M
M
(U
                                  solutions inoculated with
                                  millipore filtered (0.22^)
                                  mine water
               solutions
               inoculated with
               untreated mine water
    4  	
                                                     160
      FIGURE 5-17.
         80          120
           TIME, hrs.
Oxidation of ferrous iron solutions inoculated
with mine water from RT 9-11.
200

-------
                                                                   5-44
5-7  Oxidation of Iron Pyrite






       The kinetics of oxidation of ferrous iron in the mine water




system (sections 5-5 and 5-6) and the kinetics of hydrolysis of  ferric




iron (Chapter 4) having been studied, the final step in the sequence




of chemical reactions describing the dissolution of pyritic agglomer-




ates is the oxidation of pyrite itself, both by oxygen and by ferric




iron.  With the exception of the study by Garrels and Thompson (7),




prior investigations of the oxidation of pyrite have been concerned




almost entirely with oxygen as the oxidant.  In many cases, the  poten-




tiality of ferric iron as an oxidant has been overlooked.  It was de-




sired to obtain some idea as to the rate of oxidation of pyrite  by




ferric iron relative to that by oxygen.






5-7.1  Experimental Procedures




       Several dilutions of a stock solution of ferric perchlorate were




prepared and adjusted to pH 1.0 with perchloric acid.   (Sato (6) and




Garrels and Thompson (7) noted the oxidation of pyrite to be indepen-




dent of pH below 2.0.)  Nitrogen was bubbled continuously through the




solution to remove all traces of oxygen,  any oxidation of pyrite then




being attributable only to the action of ferric iron.   Iron pyrite  from




Rico, Colorado (Ward's Natural Science Establishment,  Rochester, New




York) was ball-milled and screened, and the 200-250 mesh fraction was




selected for the experiment.  At time zero, various amounts of the




finely-divided pyrite were added to the solution of ferric iron, the




solid phase being uniformly dispersed by means of a magnetic stirrer.




The electrochemical potential of the system was measured at various

-------
                                                                5-45

intervals using a platinum spiral indicator electrode and a calomel

reference electrode saturated with NaCl, as described in section 2-3.3.

A Leeds and Northrup potentiometer (Cat. No. 7664)  was employed for

the potential measurements.

       The procedure adopted was similar to that used in Chapter 4.

In a well-defined system such as this, the reversible potential is
                                    O      *3
established by the electroactive Fe   - Fe+  couple in accordance  with

the Nernst Equation


            E = E° - .0592 log(Fe  *   at 25°C                 (5-20)
At PH 1.0, since  [Fe+2] =  [Fe(H)3T and [Fe"1"3] = [Fe(IIl)] ,  E° can

easily be computed  in a constant ionic medium by measuring the poten-

tial, E, and independently determining the total concentrations of

ferrous and ferric  iron.  By conducting all future studies under the

same experimental conditions of constant ionic strength and tempera-

ture (to insure that the activity coefficients remain constant), this

value for the equilibrium potential, E , can be utilized to compute the
          o       o
ratio [Fe  ]/[Fe  ] by simply measuring the potential of the system.

       After the  potential was recorded, aliquot s of the suspension

were removed and  filtered.  The concentration of total iron in the fil-

trate was determined using the bathophenanthroline procedure in which

10% hydroxylamine is utilized to reduce all ferric iron to the ferrous

state.  The individual concentrations of Fe(ll) and Fe(lll) were cal-
                                                            1      O
culated knowing the concentration of total iron and the [Fe  ]/[Fe  ]

ratio from the potential measurements.

-------
                                                                 5-46



5-7.2  Results and Discussion


       Rate of Oxidation in Absence of Oxygen


       The rate of disappearance of ferric iron in the presence of


finely -divided iron pyrite is remarkably rapid.  Since the rate of


oxidation of pyrite is independent of pH below pH 2.0 (6)  (7), a


simple rate law dependent upon the two reactants can be assumed, of


the form



            . d £Fe(IIl)]  = k [Fe(iii)]m[FeS_]n               (5-21)
                   at                        *-


If the concentration of pyrite is large compared to that of Fe(lll),


[FeS ] will remain relatively constant during the course of the reac-


tion.  (Note also, from the stoichiometry of the reaction (equation


5-4) f that the oxidation of one mole of pyrite consumes fourteen moles


of ferric iron.)  Therefore, under these conditions, the rate of oxi-


dation can be approximated by



            _ d CFe(lIl)] =    UeC m)]»                       (5-22)
                  at         1
 where
               = k  [FeS2]n                                      (5-23)
 and is constant.  Furthermore, if m = 1 and the reaction is first-


 order in the concentration of Fe(lll), a plot of log [Fe(lll)j versus


 time should be linear.  Figure 5-18 shows that the decrease in Fe(lll)


 conforms to such a  formulation.  It is also seen that the rate of de-


 crease of Fe(IIl) is a function of the concentration of pyrite, pre-


 sumably the surface concentration; for the same initial concentration


 of Fe(lll), the half-time is  about 250 minutes in the presence of

-------
•

 •
\ •!
                    50
100
                                                                    250
                                                300
                                                                                             350
                                   150         200

                                     TIME, min.

FIGURE 5-18.  Reduction of ferric iron by iron pyrite  (200-250 mesh)  in  the


              of oxygen.                                                                 «-n
                                                                                                   <
                                                                                                   • j

-------
                                                                 5-48







1 gm/1 of FeS7,  and only  25 minutes when 5  gms/1 of pyrite are pre-




sent.




       Taking the logarithm of  equation 5-23,  one obtains






            log  kx = log  k +  n  log [FeS2]                       (5-24)





A log-log plot of the pseudo-first-order rate  constants  (computed from




the slopes of the straight lines in Figure  5-18)  against the concentra-




tion of pyrite should yield  a straight line of slope  £,  the order of




the rate-dependence on FeS_,  if the assumed rate law, equation  5-21,  is




valid.  Two such plots are shown in Figure  5-19 for two  different  ini-




tial concentrations of Fe(lII).  The order, n, is seen to  be  about  1.3




and 1.1, implying a first-order-dependence  of  the rate on  the concen-




tration of iron pyrite.




       If the assumed rate law is valid and the first-order-dependence




upon Fe(lll) is correct,  as demonstrated in Figure 5-18, the  rate  con-




stant k1 should be  independent of the initial  concentration of  Fe(lll).




This is not the case, however,  as seen in Figure 5-20.  In fact, the




slope of the semilog plot increases as the initial concentration of




Fe(lll) decreases,  implying an inverse dependence of  the rate on Fe(lll),




An  understanding  of the kinetics of the reaction is further complicated




by  the  findings of  Garrels and Thompson (7) that the instantaneous rate




 of  reduction of Fe(lll) decreases with tirae (implying direct kinetic




 dependence on Fe(IIl)), but the  average rate of reduction is inde-




 pendent of the  initial concentration of Fe(lll) (suggestive of zero-




 order dependence  on Fe(lll)).  Furthermore, they present a figure




 similar to Figure 5-20,  showing  essentially a  logarithmic decrease in

-------
                                                                 5-49
   -3.0
c
•H
B
-2.5
   -2.0
            [Fe(III)]o= 3.4x10
                 pH  1.0
 1.3

-4
       -1.0
   -4.0
   -3.5
   -3.0
 a
•H
 e
 o
 3
    -2.5
    -2.0
        -1.0
                                                         (a)
                  -0.5            0.0
                           LOG FeS2 (gms/1)
                      0.5
                                                          (b)
                                       1.3xlO"3M
                                             pH  1.0
                                        1-1
                  d log fFedlPl
                        dt
                   -0.5
                      0.5
1.0
1.0
                             0.0
                      LOG FeS. (gms/1)
FIGURE 5-19a,b.  Rate of reduction of ferric  iron ad'a  function
                 of pyrite concentration.

-------
I
§
                                                                   5-50
    lo
      -3
   10
     -4
   ID
     '5
                                       [Fe(III))o = 2.65x10

                                       k, - 2.7xlO"4 min"1
                  [Fe(III)] = 1.28xlO"3M
                  k, - 1.6xlO~3 min"1
            .
            \
             \
                                   1.0 gm/1  FeS,
                                               i

                                       pH 1.0
               \
                                  3.26xlO"4M
          _     \   kl * 1'2xl°
                  \
                   \
                    \
                     A

                       \
                                "2
                    100
                                                       400
                                                                   500
                       200         300
                         TB4E, min.
FIGURE 5-20.  Effect of initial concentration of ferric iron
              on rate of reduction^of Fe(III) by pyrite, in
              the absence of oxygen.

-------
                                                                 5-51





Fe(IIl) with time, with the rate of decrease becoming  steeper as the




initial concentration of Fe(lll) decreases.




       (The stoichiometry of the reaction (equation 5-4)  was verified




by noting a Tk increase (15/14) in the concentration of total dissolved




iron (Fe(ll) and Fe(IIl)), using the bathophenanthroline procedure.)




       Since the mechanism of the oxidation of pyrite by Fe(lll) was




not the primary purpose of this study, but rather the relative  rate




of the reaction as compared to that of the oxidation of Fe(ll)  was of




major concern, it will be sufficient to note the rapidity by which




Fe(lll) is  reduced by pyrite.  Even for the  slowest case observed,




where  [Fe(llD]  ~  10~3M and  [FeS^ = 0.12  gms/1 (~  10" M),  the




half-time is  approximately  2 days  which  is considerably less than that




for  the oxidation of Fe(ll)  even when accelerated by any of the experi-




mental chemical catalysts  found  in natural waters.






Oxidation Rate in Presence  of  Oxygen




       After  finding that  there is no  appreciable  adsorption of dis-




 solved Fe(Il)  on  iron  pyrite,  and  that Fe(ll)  is not catalytically




oxidized  in the presence of FeS  (see section 5-5.2),  the rate of




 reduction of  Fe(lll) by pyrite in  the presence of  oxygen was investi-




 gated.   The experimental procedure was identical to that in the pre-




 ceding section except that  the system was left open to the  atmosphere.




 The results are shown in Figure 5-21,  indicating that  there is vir-




 tually no difference between the rate of reduction of  Fe(lll) by pyrite,




 or the rate of change of soluble Fe(ll), under aerobic or anaerobic




 conditions.  Hence,  the implication is that even in the presence of a




 partial pressure of oxygen of 0.2 atm. the oxidant of iron  pyrite  is





 ferric iron.   Fe(lll) at pH = 1 oxidizes FeSg faster than Og,

-------
                                                               5-52
                           Increase in Fe(II):
                             under one atmosphere of nitrogen
                              in the presence of  atmospheric
                              oxygen
                          change in Fe(III)  in the presence
                              of atmospheric oxygen
                in Fe(III) under
              tmosphere of nitrogen
       [Fe(III)]  ^3x10
            1.0 gm/1  FeS
                       80          120         160
                           TIKE, ruin.
FIGURE 5-21.  Reduction  of  ferric  iron and increase in
              ferrous  iron  in the  presence and absence of oxygen,

-------
                                                                 5-53


5-8  Conclusions


       5-8.1  Model Describing Pyrite Oxidation and Pollution by Coal
              Mine Drainage


       In accordance with the experimental results presented  in this

chapter, the following model is proposed to describe the oxidation of

iron pyrite in natural mine waters:

                                (+ 0 )
   Initiation Reaction; FeS^Cs) 	^Fe(ll) + S - compound    (5-24a)


   Propagation Cycle;
                                f          v\fast


                           -    ^            y T * ***** /} ^ ** '        \J — fc*rD j C /
                                              Fe(OH)3(s)         (5-24d)
The reactions  shown  are  schematic  and do not represent the exact

mechanistic  steps.   The  model  is similar to and carries with it the

same overall consequences  as that  suggested by Temple and Delchamps

(14).   In  this model,  the  rate-determining step is a reactive step in

the specific oxidation of  ferrous  iron, reaction 5-24b.  As this in-

vestigation  has  demonstrated,  the  rate of oxidation of ferrous iron

under  chemical conditions  analogous  to those found in mine waters is

very slow,  indeed  considerably slower  than the oxidation of iron py-

rite by ferric iron, reaction  5-24c.   At pH 3, half-times for the oxi-

dation of  Fe(Il) are on  the order  of 1000 days while  in the case of

oxidation  of pyrite  by Fe(lll), half-times on the order of 20 to 1000

minutes were observed.

-------
                                                                  5-54






       Reaction 5-24a serves only as an initiator of the overall  re-



action: ferrous iron may be released by simple dissociation of the



pyrite, or by oxidation of the pyrite by oxygen.  Once the sequence




has been initiated, a cycle is established in which ferric iron



rapidly oxidizes pyrite and is slowly regenerated through the oxy-



genation of the resultant ferrous iron, reactions 5-24b and c.  Oxygen




is  involved only  indirectly, in the regeneration of Fe(Hl).



       Precipitated  ferric  hydroxide deposited  in the mine and the



streams  serves  as a  reservoir for soluble Fe(lll), by reaction 5-24d.



If  the regeneration"of Fe(lll) by 5-24b  is halted so that the concen-




tration  of soluble Fe(IIl)  decreases,  it will be replenished by dis-



 solution of  the solid Fe(OH)_ and will be free  to act again should it


                                      *

come in  contact with additional  FeS2>  The presence of  sulfate in



 solution increases the concentration of  dissolved Fe(lll)  in equili-



brium with the precipitate, by complex formation  (see section 2-3.3).



 Barnes and Romberger's argument  (10) that there is  insufficient Fe(lII)




 available for the reaction  appears fallacious.



        Smith, et al  (13)  investigated the effect of pH  on the rate of




 oxidation of pyrite  by oxygen and observed the  reaction rate to be



 relatively independent of pH below pH 4, while  the  rate increased



 rapidly in a pH-dependent manner above pH 7.  Since this parallels the



 pH-dependence of the rate of oxidation of ferrous  iron  (Figure 5-2),



 this writer contends that Smith, et al actually observed  the  oxida-



 tion of pyrite by Fe(lll).   At higher pH-values,  the  rate of  oxida-



 tion of FeS_ increases because the  rate  of formation  of Fe(lll), via



 oxidation of Fe(ll), increases with increasing  pH.  These results  lend




 further support to the proposed  model.	^^__
        _    _


          The contact with Fe(llll however may be small  if the pyrite


 lies  on the  ceiling or on walls.

-------
                                                                 5-55






       The following pertinent consequences of this model  need  to be




emphasized:




       1)  Ferric iron cannot exist for long in contact with pyritic




agglomerates.  Fe(III) is rapidly reduced by iron pyrite.




       2)  The elimination of oxygen is of no consequence  with  regard




to the specific oxidation of iron pyrite.  However, the exclusion of




oxygen does inhibit regeneration of Fe(lll) through the oxidation of




Fe(ll), and will be of significance once the supply of available Fe(lll)




is exhausted.




       3)  The overall rate of dissolution of pyrite is independent of




its surface structure.   Interference with the surface of ^pyrite, such




as the application of  inhibitors which are adsorbed at the solid-




solution interface, is inconsequential since the oxidation of pyrite




is not the rate-limiting step.  If, on the other hand, such a technique




could make the rate of oxidation of pyrite less than the rate of oxi-




dation of Fe(H), then such a control measure may have merit.




       4)  Microorganisms can only be influential by mediating the




oxidation of ferrous  iron since it, alone, is the rate-determining




step.  Catalysis of the  specific oxidation of iron pyrite by micro-




organisms, even if it could be  unequivocally demonstrated, can have




no effect on the overall rate of dissolution of iron pyrite.




       It is probably this same cycle which is responsible for the




dissolution and leaching of other mineral sulfides as found in copper




and uranium mines.  Microbial leaching of these other minerals has




always been demonstrated in the presence of iron, pyrite being the




most abundant and widespread of all mineral sulfides (2).   Until it

-------
                                                                  5-56


can definitely be proven otherwise, cycle 5-24 adequately accounts for

the observed microbial leaching of other mineral sulfides, and direct

microbial oxidation must be discounted.

       The solution to the problem of  acid mine drainage, therefore,

appears to rest with methods  of controlling the oxidation of ferrous

iron.  Due to the cyclical nature of the process describing dissolu-

tion of pyrite, mere  treatment of the  resulting drainage water will

allow the problem to  compound and magnify  itself.  Hence, at-source

control measures  are  preferred.   One  such  method might  involve the

inhibition of  natural catalysts which are  of  significance in  accele-

rating  the rate of oxidation of Fe(Il) in  mine waters.   It  is of

primary concern to discover  which of  these catalysts causes the oxi-

dation to proceed as rapidly as it does in nature.   From the  prelimi-

nary analysis presented in this chapter,  the major  chemical catalysts

 appear to be alumino-silicate clays,  but only at  considerably larger

 surface concentrations than would be expected in  natural mine waters.

 Microbial catalysis, as by the autotrophic iron-bacteria Thiobacillus

 and Ferrobacillus ferrooxidans,  seems to be ecologically significant

 as evidenced by the few field investigations conducted.  Numerous ac-

 counts of autotrophic iron oxidation prevail, but only few quanti-

 tative reports of their actual activity in nature have appeared.
                                                 2      3
 Tuttle, Randies, and Dugan  (40)  observed only 10  to 10  iron-
                                                                  *
 oxidizers per ml. in an acid mine stream, using an MPN technique.

 This  is not a significant concentration when one considers the limited

  amount of iron which can  be oxidized  by these few microorganisms (see

  Appendix F).  However, there may be considerable surf ace ^growth, of

        ^
          These counts are probably not representative of the  concentra-
  tion near the pyrite surface.

-------
                                                                 5-57






these microorganisms associated with the available solid material




suspended or deposited in the drainage streams.   It is important to




know the actual concentration of these auto trophic microorganisms




found in mine waters in order to assess their relevance regarding




the rate of oxidation of ferrous iron.  Once the role of all cata-




lytic agents has been evaluated, methods can then be devised to con-




trol the oxidation  of ferrous  iron  and, hence, the dissolution of  iron




pyrite  and the  introduction of  acidity  into natural mine waters.

-------
                                                                  5-58
                               References


1)  Palathe,  C., Berman, H.,  and Frondel,  C.,   Dana' s  System of
       Mineralogy, 7th ed.,  vol. 1,  John Wiley  and Sons,  Inc., New
       York,  1944.

2)  Clark, C. S., "Oxidation of Coal Mine Pyrite," Journ. San. Eng.
       Div.,  Proc. Amer. Soc. Civil Eng..  9_2,  127 (1966).

3)  Krauskopf, K. B. , Introduction to Geochemistry, Ch. 18,  McGraw-
       Hill Book Company, New York, 1967.

4)  Stokes, H. N., "On  Pyrite and Marcasite," U. S. Geol. Surv.  Bull.
       186 (1901).

5)  Nelson, H.  W., Snow, R. D.,  and Keyes, D. B., "Oxidation of
       Pyritic  Sulfur in Bituminous Coal," Ind. Eng. Chem., 25_,
       1335  (1933).

6)  Sato, M. ,  "Oxidation  of Sulfide Ore Bodies.  II. Oxidation
       Mechanisms of Sulfide Minerals  at  25 C,"  Econ.  Geol., 55,
        1202  (1960).

 7)  Garrels, R. H.,  and Thompson, M. W.,  "Oxidation of Pyrite by
        Iron Sulfate Solutions,"  Amer.  Journ.  Sci. , 258-A,  57 (1960).

 8)  McKay,  0. R., and Halpern,  J.,  "A  Kinetic  Study of the  Oxidation
        of Pyrite in Aqueous Suspension,"  Trans.  Met.  Soc. AIMS, 212,
        301 (1958).

 9)   Gerlach, J., Hahne, H.,  and Pawlek, F.,  "Beitrag  zur Drucklangung
        von Eisensulfiden.  II.  Zur Kinetik der Drucklangung von
        Pyrit," Zeit. Erzb.  Metall.,  .19., 66 (1966).

10)  Barnes, H. L.,  and Romberger, S.  B.,  "Chemical Aspects  of Acid
        Mine Drainage," Journ.  Wat.  Poll.  Contr. Fed., 40_, 371  (1968).

11)  Smith, E. E.} Svanks, K.,  and Shumate, K., "Sulfide to Sulfate
        Reaction  Studies," Proc. 2nd Symp. on Coal Mine Drainage
        Research, Coal  Industry Advisory Committee to  ORSANCO,
        Pittsburgh,  May 1968.

12)  Colmer, A. R. ,  and Hinkle, M. E.,  "The Role of Microorganisms in
        Acid Mine Drainage: A Preliminary Report," Science, U)6_, 253
        (1947).

 13)  Temple, K.  L.,  and Colmer,  A. R.,  "The Autotrophic  Oxidation of
        Iron by a New Bacterium: Thiobacillus Ferrooxidans," J. Bact.,
        62,  605 (1951).

-------
                                                                  5-59


14)  Temple,  K.  L.,  and Delchamps,  E.  W.,  "Autotrophic Bacteria and
        the Formation of Acid in Bituminous Coal Mines," Appl. Micro-
        biol., 1,  255 (1953).

15)  Leathen, W. W.,  Kinsel, N.  A., and Braley,  S.  A., "Ferrobacillus
        Ferrooxidans: A Chemosynthetic Autotrophic  Bacterium," J.  Bac t.,
        22, 700 (1956).

16)  Kinsel, N.  A.,  "New Sulfur-Oxidizing Iron Bacterium:  Ferro-
        bacillus Sulfooxidans SP.N." J. Bact., 80,  628 (1960).

17)  Unz, R. F., and Lundgren, D. G., "A Comparative Nutritional  Study
        of Three Chemoautotrophic Bacteria: Ferrobacillus Ferrooxidans,
        Thiobacillus Ferrooxidans, and Thiobacillus Thiooxidans," Soil
        Science, 92, 302 (1961).

18)  Silverman, M. P.,  and  Lundgren,  D. G. , "Studies on the Chemo-
        autotrophic  Iron Bacterium Ferrobacillus Ferrooxidans: I.
        An Improved  Medium  and  Harvesting  Procedure for Securing High
        Cell Yields,"  J. Bact.,  77,  642 (1959).

19)  Silverman, M. P.,  and  Lundgren,  D. G., "Studies on the Chemo-
         autotrophic  Iron Bacterium Ferrobacillus Ferrooxidans.  II.
        Manometric  Studies," J. Bact.,  78, 325  (1959).

 20)   Schnaitman,  C.  A., "A Study of the Mechanism  of  Iron Oxidation by
         Ferrobacillus ferrooxidans," Ph.D. Thesis,  Syracuse University
         (1965).

 21)   Brynner, L.  C.  , Beck,  J. V., Davis,  D. B.,  and Wilson, D. G. ,
         "Microorganisms in Leaching Sulfide Minerals,"  Ind. Eng.  Chem.,
         46_,  2587 (1954).

 22)   Brynner, L.  C. , and Anderson, R. , "Microorganisms  in Leaching
         Sulfide Minerals,"  Ind. Eng. Chem., 49_,  1721  (1957).

 23)   Ehrlich, H.  L. , "Bacterial Action on Orpiment,"  Econ. Geol., 58_,
         991 (1963).

 24)   Silverman, M.  P., and  Ehrlich, H. L., "Microbial Formation and
         Degradation of Minerals," Advances in Appl. Microbiol.,  _6_,
         153 (1964).

 25)   Ehrlich, H. L, , "Observation on Microbial Association with Some
         Mineral Sulfides,"  p.  153  in Biogeochemistry of Sulfur Iso-
         topes, M. L. Jensen, ed., Nat'l Sci. Found. Symp.,  Yale Univ.,
         New Haven,  Conn. (1962).

 26)  Lee, G. F., and  Stumm,  W.,  "Determination of Ferrous Iron in the
         Presence of Ferric Iron," Journ.  Amer. Wat. Works Assoc., £2
         1567 (1960).

-------
                                                                   5-60


27)  Huffman,  R.  E.,  and Davidson,  N.,  "Kinetics of  the Ferrous Iron-
        Oxygen Reaction in Sulfuric Acid Solution,"  Journ. Amer. Chem.
        Soc.,  78, 4836 (1956),

28)  George, P.,  "The Oxidation of  Ferrous Perchlorate by Molecular
        Oxygen,"  Journ. Chem.  Soc., p.  4349 (1954).

29)  Cher, M. , and Davidson, N. "The Kinetics of the Oxygenation of
        Ferrous Iron in Phosphoric  Acid Solution," Journ. Amer. Chem.
        Soc.,  TTj 793 (1955).

30)  Crabtree, J. H., and Schaefer, W.  P., "The Oxidation of Iron  (II)
        by Chlorine," Inorg.  Chem., _5_,  1348 (1966).

31)  Stumm, W., and Lee, G. F., "Oxygenation of Ferrous  Iron,"  Ind. Eng.
        Chem.. 53_, 143 (1961).

32)  Weiss, J., "ElektronenUbergangsprozesse in Mechanismus von Oxyda-
        tions - und Reduktions-Reaktionen in LOsungen,"  Naturwissen-
        schaften, 23, 64 (1935).

33)  Lamb, A. B.  , and Elder, L. W., "The Electromotive Activation of
        Oxygen," Journ. Amer.  Chem. Soc., 53, 137  (1931).

34)  Van  Olphen, H., An Introduction to Clay Colloid Chemistry, Inter-
        science Publ., New York (1963).

35)  Schenk, J. E., and Weber, W.  J., "Chemical Interactions of Dis-
        solved Silica with Iron (II) and  (III),"  Journ.  Amer. Wat.
        Works Assn., 60_, 199 (1968).

36)  Scott, Robert, Personal Communication, Project Engineer, Federal
        Water Pollution Control Administration, Elkins,  West Virginia
        (1968).

37)  standard Methods  for the  Examination of Water and Wastewater, llth
        ed.,  American  Public Health Assn.,  Inc.,  New York (1960).

38)  Salotto,  B.  V., Earth, E. F.,  Ettinger, M. B., and Tolliver, W.  E.,
        "Determination of Mine Waste Acidity," submitted to Environ.
        Sci.  and Tech.  (1967)

39)  Kim, A.  G.,  Personal Communication,  Pittsburgh Mining Research
        Center,  U.  S.  Bureau of Mines,  Pittsburgh, Pa. (1968).

40)  Tuttle,  J.  H.,  Randies, C.  I.,  and Dugan, P. R., "Activity of
        Microorganisms in Acid Mine Water," Journ. Bact., 95, 1495
        (1968).

-------
                              CHAPTER 6






                             CONCLUSIONS






       Information concerning the chemistry of aqueous  iron can be ap-




plied to natural water systems in order to cope with the vast  and




costly problem of acid mine drainage, for the design of efficient de-




ferrization processes, and to provide for a better understanding of




the limnological cycles of a number of key elements.  This chapter sum-




marizes the relevant results realized  in this research, and  discusses




the practical consequences of these results.









6-1  Principal  Findings






        1)  Ferrous  iron  is thermodynamically  unstable  in the presence




of oxygen.  The rate at  which it is  oxidized  to ferric  iron is compa-




tible  with a rate law which  is  first-order in both  [Fe(ll)3 and [02],




and  second-order in [OH~] at pH-values above  4.5.   At  lower pH-values,




the  rate of oxidation is independent of pH.   The  reaction proceeds




extremely  slowly in the  acidic  region, but is catalyzed by inorganic




and  organic ligands capable of forming complexes  with  Fe(ll)  and




Fe(lll), heavy  metal ions,  clay particles, and bacteria.




        2)  The  oxidation of ferrous iron is the specific rate-




determining  step in the oxidation of iron pyrite and the subsequent




discharge  of  acidity into mine drainage waters.  The chemical oxygena-




 tion of Fe(Il)  at pH-values less than 4 takes place very slowly, even




 in the presence of the many chemical catalysts which are operative  in

-------
                                                                  6-2
nature.  The direct oxidation of iron pyrite by ferric iron is  quite

rapid and Fe(in) serves as the prime oxidant of iron pyrite.   A

cycle is established involving the rapid oxidation of pyrite by

Fe(lII), and the slow regeneration of Fe(lll) through the oxygenation

of Fe(ll), in the following schematic way:

                            fast                2
            FeS  + Fe(lll) - *• Fe(ll) + SC>4

                         slow
            Fe(ll) + 0  - »-Fe(lIl)
                            fast                  7
            FeS2 + Fe(IIl) - **  Fe(ll) + SO^


Microorganisms, presumably the autotrophic "iron bacteria," markedly

enhance the rate of oxidation of Fe(ll), and, therefore, accelerate the

overall rate of pyrite oxidation.  Oxygen participates in the  cycle

only in the regeneration of spent ferric iron.



6-2  Practical Consequences and Implications Resulting From This Research


       1)  The proposed cycle to describe the oxidation of iron pyrite

and the acidification of mine drainage waters can be utilized  in evalu-

ating  some of the methods recommended for the prevention of acid mine

drainage, and should serve as a guide to indicate the direction in

which  future control measures should be aimed.

       Mine-sealing and the application of bactericides are two pro-

posals, both directed at the retardation of the oxidation of ferrous

iron.  The former  involves the elimination of air and/or water from the

mine and  is intended to stop the reaction entirely.  The latter is

aimed  at  destroying the catalytic agent which is responsible for the

-------
                                                                     6-3

rapid rate at which Fe(ll)  is oxidized in mine drainage.   Previous asser-

tions that oxygen serves as the specifijg, direct oxidant  of iron pyrite

are unjustified in view of this research which indicates  that oxygen  is  in-

volved only indirectly, producing ferric iron which is, itself, the primary

oxidant of FeS .   Furthermore, this research has shown that in the absence

of microbial catalysis, the oxygenation of Fe(ll) is sufficiently  slow that

it is essentially halted.  The regeneration of Fe(lll) proceeds so slowly

that its effect on pyrite is of little consequence.

       Both mine-sealing and the application of "bactericides, however,  are

subject to similar limitations.  Firstly, neither will result in the immedi-

ate cessation of acidic discharges.  Ferric iron must first be flushed out

of the mine, or rendered inactive, before any beneficial  effects could be

realized.  Reports of previous mine-sealing operations demonstrate that in

spite of  significant  reductions  in the  concentration  of oxygen, there was

not always a marked  change in the  quantity of acid released  although the

discharge had  a higher  concentration of ferrous  iron  (l).   In the few suc-

cessful mine-sealing  operations, often little improvement in the quality
                                                                 £
of the effluent  from the mine was  observed  for several years (2).   These

observations can  readily be interpreted in  view of the proposed scheme;

the  elimination  of oxygen  stops  the  oxidation of ferrous  iron but the py-

rite  is still  subject to oxidation by the vast quantity  of Fe(lll) availa-

ble  in the mine.   Only after the active Fe(lll) is depleted should one  ob-

serve a decrease in the acidity of the drainage water.


          There are instances where mine sealing was followed by reduced
 acidity.   For a recent review see R. D. Hill, Acid Mine Water Control,
presented before the Mining Environmental Conference, University  of
Missouri, Rolla, Missouri.  April, 1969-

-------
                                                                   6-4







       Secondly, from a practical standpoint,  both methods are  of




doubtful applicability.  Total exclusion of oxygen is unlikely  due




to the many fractures and minute cracks in the mine wall and the fre-




quent collapses experienced in the mine itself.  Application of a




bactericide requires continual injection of the agent into the  system




at a suitable location where the entire inflow could be treated.   In




most mining areas, such a location is non-existent.  There are  some




situations, however, where a bactericide could be employed.  Although




it is not feasible to treat the mine itself, the various spoil  banks




exposing pyrite previously dug out of the mine are amenable to  such




treatment as they can be reached rather easily.  It is conceivable,




also, that strip mines would yield to such treatment since the  influ-




ent water can usually be located.




       Another proposal concerns the addition of an alkali to the




system  in order to raise the pH of the environment within the mine,




thus inhibiting the microorganisms which catalyze the oxidation of




ferrous  iron.  Here again, a suitable location for treatment is re-




quired.  Furthermore,  if the pK  is raised appreciably, this study  has




shown that the  chemical oxygenation of Fe(ll) will proceed sufficiently




fast by  itself.




       The introduction of organic material to chemically reduce ferric




iron and sulfate  has  also been suggested in order to promote preci-




pitation of  ferrous  sulfide.  The continuous  addition of organic mat-




erial  is necessary  and the elimination of oxygen  is mandatory  so that




Fe(ll)  and S(-II)  are maintained in their reduced  states.

-------
                                                                     6-5




       Inactivation of the pyrite surface through the  application of




chemical inhibitors which are adsorbed at the solid-solution interface




is unfeasible and inconsequential.  Since the specific oxidation of




pyrite is not the rate-determining step, partial coverage of the py-




ritic surface would not affect the overall rate of pyrite oxidation.




In addition, constant exposure of new pyritic surfaces would be expected




as a result of the frequent collapses inside the mine.




       This discussion has considered only control methods for the




abatement of pollution by  coal mine drainage. Proposals for the treat-




ment of  effluent waters  from  these mines  include  acid neutralization by




the addition of  lime,  reverse osmosis,  and ion  exchange.  Each produces




a voluminous or  a  concentrated waste which must ultimately be disposed of.




        This  research suggests that  treatment of acid  mine drainage re-




quires  treatment of the cycle by which  such  acidic  discharges  arise.  Al-




though  no specific schemes for  such treatment are proposed  as  a result of




this  study,  it is  emphasized that the catalytic oxidation of ferrous iron




need to be halted.  Because of the inaccessibility  of the pyrite oxida-




tion site, it is difficult to convert this theoretical suggestion  into  a




practical treatment.  (Catalytic oxidation of ferrous iron, however, is




 an asset, where mine drainage is treated for iron removal.}




        2)  This research has shown that the concentration of ferrous




 iron in natural groundwaters can often be predicted from considerations




 of solubility relationships, specifically the  solubility product of




 ferrous carbonate which was  obtained in  this study,  i.e., pKgo = 10.2k




 at 25°C and zero  ionic  strength.  The  expected concentration of Fe(ll)

-------
                                                                   6-6




can be calculated if the pH and alkalinity of the groundwater are



monitored.  The kinetic relationships describing the oxygenation of



Fe(ll) have also been derived.  If the solubility and kinetic relation-



ships are coupled, they can be applied to the design of an efficient



iron-removal facility in order to bring the raw water into conformance



with the Public Health Service drinking water standards for iron



(0.3ppm) (3).  This assumes an efficient filtration system for the



actual removal of the resultant hydrous ferric oxide.  For example, a



groundwater at pH 6.5 containing 5 x 10~  eq/1 alkalinity should con-



tain approximately 1.7 ppm of dissolved Fe(ll) (approximately 3 x 10  M),



and would therefore require 857. oxidation of the Fe(ll) in order to



provide an acceptable finished water having 0.3 ppm of iron.  At a



partial pressure of oxygen of 0.20 atm. and at 25 C, a detention time



of about 60 minutes would be necessary (toc = log (85/15)/k") if the
                                         o.)


oxidation were to proceed at pH 6.5.  Furthermore, iron-removal may



be aided by precipitation of Fe(ll) as ferrous carbonate.



       In addition, this study has demonstrated the capability of



many elements in natural waters to accelerate the oxidation of Fe(ll).



Retardation of the oxidation of ferrous iron in the presence of oxy-



gen can,  therefore, be attributed almost entirely to the presence of



reducing agents in natural systems, notably organic matter.  Ferric



iron is a potent oxidant of organic material, getting reduced to



Fe(ll) in the process.  The net effect appears as an inhibition of



the rate of oxidation of Fe(ll), whereas in fact, the Fe(Il)-Fedll)



reaction serves as an electron-transport mechanism between oxygen and



the organic material.  The stability of ferrous iron in the epilimnion

-------
                                                                  6-7
of lakes and reservoirs, in the presence of oxygen,  can,  in a similar

manner, only be interpreted in terms of a. steady-state condition

maintained by the two concurrent oxidations:  the oxidation of Fe(Il)

by oxygen and the oxidation of the organic matter by Fe(lll).
                              References
 1)  Scott, R., Project Engineer, Federal Water Pollution Control
        Administration, Elkins, West Virginia, Personal Communication
        (1968)

 2)  Moebs, N. N.,  "Mine Air Sealing: A Progress Report," Proc.  Second
        Svmo. Coal  Mine Drainage Res, , Coal Industry Advisory Committee
        to ORSANCO, Pittsburgh, May,  1968

 3)  United States  Public  Health Service Drinking Water Standards,
        Publication No.  956, Washington  (1962)

-------
                              APPENDIX A






Correction of Experimental Solubility Data for Temperature and Activity






       The experimental  equilibrium relationship, given by equation





(2-lla),  is




               +2^
                   CH+]           8q   KC2





 the superscript c_ referring to equilibrium constants  at a given ionic




 strength.   The corresponding thermodynamic equilibrium constant for the




 reaction,  at 25°C and zero ionic strength is
 The two equilibrium constants are related by the equation
                                                                   U-3)
 where the   Y ' s  are  single ion activity coefficients.  Schindler (1)




 has  suggested that for carbonates of bivalent metals in a constant




 ionic medium similar to 0.2M HaCl(>4> the  Davies  equation
  should be applied for the computation of activity coefficients.  I is




  the ionic strength of the system, z is the charge of the specific  ion




  under consideration, and A is a constant equal to 0.509 for water  at
  25°C.

-------
                                                                    A-2
       Taking logarithms of equation A-3, one obtains




            pK0^  - log tf     - log V      + log *   = pK         (A-5)

               eq        Fe L        HC03"        H+      q



where p-   refers to the negative logarithm of that term.  Substitu-



tion of the Davies equation into A-5 gives
                    0.509 | n  . /r ~ "•-••'•I \ *•    n

                                           Fe       HC03~     H*



                                                            pK    (A-6)
                                                              eq




If the proper charges for the ionic  species are used, the equation



describing the  experimental system at 0.1M NaCIO, reduces to




            pK°  + 0.107 (4 + 1 - 1) = pK                       (A-6a)
            F   eq                       F eq


or



            pKC  + 0.428 = pK                                   (A-6b)
                eq           F eq




Taking logarithms of the right-hand  equality in equation A-2, one



obtains



            pK   = pK   - pK0                                     (A-7)
            *^ eq   * so   * 2




Substitution of this quantity into A-6b gives, after rearrangement





            pK   = pK°   + 0.428 4- pK_                            (A-8)
              so      eq             L




       From the plot of p H versus -log[Fe  ][HCO_ ] in accordance


                                      c                   c
with  equation 2-12, the intercept at p  H = 0 is equal to pK



Figure 2-5 shows the desired  intercept  to be -0.57.  In the determina-



tion  of pK_ as  a function of  temperature, by Harned and Scholes (2),



it Is found that, at 22.5°C,  pK_ = 10.35.  Substitution into A-8 gives

-------
                                                                    A-3
            PK   = -0.57 + 0.43 + 10.35                           (A-8a)
              so
                                          -O,
            pK^ = 10.21            at 22.5°C                     (A-8b)




or          K   » 6.1 x 10"11
             so



       The solubility product can readily be converted to 25°C utili-



zing the van't Hoff temperature relationship



               K       °   T  - T
                                                                  ,    ~
                          ( -— — - 1                               (A-9)

               Kl     R   VT1T2  /
where K_ and K.. are the equilibrium constants at the absolute tempera-



tures T£ and T^, respectively.  R is the ideal gas constant,  equal to



1.987 cal./mole-K, and AH  is the change in enthalpy for reaction 2-4,



equal to -4639 cal./mole at 25°C (3).  (The literature value for AH° was



employed since the experimental temperature -dependence was not suffi-



ciently precise to extract a usable change in enthalpy for the reaction.



Only three experimental points were available for such a calculation,



giving AH°      .     . = 1800 1 1200 cal./mole.)
          experimental



Substitution of these values into A-9 yields
            K25°C - ^ZZ.sV1-068                                (A-9a)




which, combined with A-8b, gives




            K   = 5.7 x 10"U                                    (A-10)
             SO



            pK   = 10.24   at 25°C
            K so


the desired thermodynamic solubility product.

-------
                                                                  A-4
                             References
1) Schindler, P. W., "Heterogeneous Equilibria Involving Oxides,
       Hydroxides, Carbonates,  and Hydroxide Carbonates," Ch.  9,
       p. 196 in Equilibrium Concepts in Natural Water Systems,
       R. F. Gould, ed., Advances in Chemistry Series 67, Amer. Chem.
       Soc., Washington (1967)

2) Harned, H. S., and Scholes,  S. R. , "The lonization Constant of
       HCO," from 0 to 50°," Journ. Amer. Chem. Soc., 63, 1706 (1941)

3) Latimer, W. E. , The Oxidation States of the Elements and Their
       Potentials in Aqueous Solutions, 2nd ed. , Prentice-Hall Inc.,
       Englewood Cliffs, N.J.,  (1952)

-------
                             APPENDIX B
Relative Significance of Soluble Phosphato-Complexes of  Fe(lll)


       Consider the following equilibria, the equilibrium constants of

which were taken from Stability Constants (1):
Fe+3 + H20 = FeOH+2 + H+ ^ - 10'2'2
Fe+3 + 2H20 = Fe(0l02+ + 2H+ ^ = 10'6-8
+ -2 „ in7-2
H + HP04 = H2P04 K12 = 10

Fe+3 + HP04"2 = FeHP04+ ^ l = 1Q8'3
Fe+3 + H2P04- =FeH2P04+2 ^2 = 103'5
+ + _ po +2 K ^ = 102'4
6nrvJ4 ~ 2 4 122
VoPQ.. . = Fe + PO ~ K = 10~
(B-l)
(B-2)
(B-3)

(B-4)
(B-5)
(B-6)

(B-7)
           J4(s)

 Figures B-l and B-2 show pH-log concentration diagrams  for phosphate,

 and for Fe(lll) in the absence of phosphate,  respectively.

        Using equation B-6
                    4     . [H+] K12    - . 10-CH*]             (B-8)
              [FeHP04+3           1Z  91

 one sees that at pH-values greater than 2.4, FeHP04+  is the  predominant

 soluble phosphato-complex of Fe(IIl).  The following  relationships

 should also be noted:

-------
                                                                   B-2
-7
   02         4        6         8        10       12
                                   PH
   FIGURE B-l.  Log concentration diagram for phosphoric acid,
-12
                                6         8
                                    pH
    FIGURE B-2.  Distribution diagram £or soluble nonoroeric hydroxo-
                 species of ferric iron.

-------
                                                                   B-3
       A)  At pH>5, Fe(OH) "*" is the predominant soluble species  of


Fe(IIl).  Hence, by equation B-2,

            [Fe(OH)2+]   KXK2     1Q-6.8

             [Fe+3]    " CH+]2   " CH+]2


        B)   In the pH-range 2.5  to 4.5, FeOH1"  predominates.  Using
                                                                 (3-9)
equation B-l,

             Fe

              [Fe+3]     CH+]
            [FeOH+2]     Kl    IP"2'2
                             «
       C)  At pH> 7.5, [HP04" ]^PT, the total concentration  of  dis-


solved phosphate.  By equation B-4,


            CFeHPO.+2]             QO                           .
            - - £ - =  ^ LPT = lO8'3 x PT                      (




       D)  In the pH-range 2 to 7, using equation B-3,

                        p
             CHPO"2]=— - - ,  since  [H PO  "] ^ P .            (B-12)
                 4        +              L
       Hence,  by  equation B-4,



                     2]     iA_.10l.l  A-                (B-13)
              [Fe+33        CH+]K12           CH+]


       Making use of relationships (B-9) through (B-13) , one can pre-

 pare the following table showing the relative abundance of the various


 soluble  complex species (assuming P  = 10  M) :

-------
                                                                  B-4
       pH
[Fe(OH)2+]
[Fe+3]
io9-2
7 2
lO''^
105'2
103'2
101-2

[FeOH+2]
[Fe+3]


—
--
io2-6
101'6
io°-6
[FeHP04+]
CFe+3]
io4-3
4 1
10
io3a
lo2-1
401'1
10o.i
       8

       7

       6

       5

       4

       3

       It is,therefore,  apparent that the known soluble phosphate-

complexes of Fe(lll) become influential only below pH 4, but  to  a

very limited extent.
                             References


1)  Sillen, L. G.,  and Martell,  E-  A.,  Stability Constants  of Metal-
Ion Complexes, Special  Publication No.17,  London,  The Chemical Society
(1964)

-------
                              APPENDIX C


Derivation of Relations Between Redox Potential and Sulfate Concentration
for Determination of Stability Constant for FeSO^+


       The reactions pertaining to this study are


            Fe+3 + SO ~2 = FeSO.+                 K.              (C-l)
                     44                   1

            HS04" » H+ + S04"2                    KA              (C-2)


            Fe+3 + H0 = FeOH+2 + H+              Q               (C-3)

        In  a system  containing  ferrous and ferric iron at 25 C and a

 constant ionic  medium of 0.1M  NaCIO  , the redox potential is defined

 by  the Nernst Equation


             E = E°  - 0.0592 log Ssll                            (C-4)
                                 [Fe+3]

 It  will be convenient to refer to the system in the  absence of sulfate

 as  cell 1  so that

                                   [Fe+2L
             E,  = E°. - 0.0592 log - r-i                        (C-5)
              1     l              CFe+3]1

 If  the pH is maintained below 3, the total  concentration of  ferric

 Iron is given by


              CFeClII)-], =  CFe+3L + [FeOH+2]  =LFe+3^l + -J-)      (C-6)
                     T  1         i           l            [H+]


 and that of  ferrous  iron is


              [Fe(H)T]1  = CFe+2J1                                   (C-7)

-------
                                                                    C-2
Upon addition of sulfate, let us refer to the  system as cell 2, where



                                   TFe+2]

            E_ = E°- - 0.0592 log - - r-2-                         (G-8)

             2     2               [Fe+3]2




In the presence of sulfate,





            [Fe(IIl)T32 = [Fe+3]2 +  [FeOH+2]2  +  [FeSO^         (C-9)




or, substituting equations C-l and C-3,





            CFe(IIl)_]9 = [Fe+3],(l  + — £- + K. [SO  ~2])          (C-10)
                    T 2         Z     [H+:I     1   ^




where only the monosulf ato-complex of Fe(lll)  is assumed,  and





                    ]2 =  [Fe+2]2                                 (C-ll)





Sulfate, however, reacts  with water  in the acidic pH-range.  so  that






            CS°4"2]ADDED  - ST =  CS04"2]  + [HS°4"] + ^FeS04+]     (C"12)




If it is assumed that  [FeSO^] «  [S0^~2] +  [HSO^"], then, using



C-2, equation C-12 becomes





            ST = [S04"2](l + ^1)                               (C-13)

                               f\



Rearranging C-13




                 •>    KAST
            CSO. ~Z] = -^ -                                 (C-13a)
 and  substituting  into C-10, one obtains




                                             KKS
                             -u
                           [Fe+:](l
                                             KA
                                              A

-------
                                                                    C-3
       The difference in potential between the cell before and after

sulfate is added can be obtained by subtracting equation C-8 from

C-5 to give

                                      lFe+2]?[Fe+3L
            E. - E0 . E = 0.0592 log 	s-^	—              (C-15)
             1    £.                   r-w, +^T
since the standard potentials  are  equal,  i.e., E - » E _.  It is seen

also in this step that  the  liquid  junction potentials for the two

cells cancel out.  The  experiment  has been designed such that
 [Fe*2], =  DFe*2]..  Hence,
      *•          L
or
                               j.
                            [Fe+].
            E  » 0.0592 log - =-±                             (C-15a)
                            [Fe+3]2
             [Fe"1"3],
                                                                (c-15b)
 Furthernore,  by experimental design, [FeClIl),^ = [Fe(lIl)T]2  so that

 equation C-6 can be set equal to C-14,


             [Fe*3], (l + A.) . [Fe+3]2(l + A- + ^&_)    (C
                   1     CH+]          z     CH+]   KA + [H+]

 assuming the concentration of H+ to remain constant with the  addition

 of sulfate.  Equation C-16 can be rearranged and set equal to C-15b

 so that


 Considering the equality on the right,  after rearranging and simplifying,

 we obtain

-------
                                                                  C-4
                                            KA
                                                 +
                 CH+] v     °'0592     y  KA  +  CH+]
- l)- *l
  y
                                                                (C-17b)
Equation C-17b is now in a usable  form.   If  the potential  and  pH  are


measured as a function of the sulfate added,  and it is known that Q.


= 2.89 x 10"3 at 25°C and 0.1M NaClO^ (1), the left-hand side  can be


plotted against ST,  the slope of the resultant straight line being



K1KA
	.   Knowing K , the second acidity constant of sulfuric acid,

K. + [H+]              A
 A
and the pH at which the study was conducted, one can compute the  sta-


bility constant for the monosulfato-complex of Fe(lll).
                              References
 1)  Milburn, R. M.,  "A Spectrophotometric Study of the Hydrolysis of

       Iron (III) Ion.  III.   Heats and Entropies of Hydrolysis,"

       J. Amer. Soc.. 79,  537 (1957)

-------
                              APPENDIX D



Thermodynamic Stability of Iron Pyrite


       The change in free energy for the oxidation of iron  pyrite by

oxygen, using the data available in Latimer (l), is



          FeS2(s)   *  l°2  +   H2°   =   Fe+2  + 2S°4~2   +  ^  (D"1


      AF°  (-39.84)    I (0)  (-56.7)    (-20.3)  2C-177.34)  2(0)

        S. AF° = -278.4 kcal.

The reaction should take place  spontaneously, iron pyrite being thermo

dynamically unstable  in the presence of oxygen.
                               References
 1)   Latimer,  W. M. ,  The Oxidation  States of the Elements and Their
        Potentials  in Aqueous Solutions, second ed., Prentice-Hall, Inc.,
        Englewood Cliffs. New Jersey (1952)

-------
                              APPENDIX E





Kinetics of Microbial Growth (1)





       The change in concentration of microorganisms, B_,  with time  is



first -order in concentration of microorganisms,






            ~ = «B                                              (E-l)
            at



where ju is the specific  growth-rate constant.  For an enzymatic pro-



cess, the growth-rate  is given by the Michael is -Men ton equation,



                n   S
                 max                                             fv  n\

            " - K~TT                                          (E-2)
                 m


where ^ is the concentration of  substrate  (source of energy),  max

                                j^

is the maximum growth-rate, and  m   is  the Michael is -Men ton constant.



The  change in concentration of microorganisms  upon utilization of the



substrate is
 where Y_ is defined as the yield.   Equations E-l, E-2, and E-3 can be



 combined to give the change in substrate with time  as



                   A»   3
             -dS    max _ _                                    /-,
             "dT =(K  + S) Y                                        '
                    m



 If the concentration of substrate is large compared  to  the Michaelis-



 Menton constant, i.e., if substrate is non-limiting,  the  specific



 growth-rate is constant and equal to the maximum growth-rate, so that



 E-4 simplifies to

-------
                                                                   E-2




                  Ai   B
            -dS _  max                                            (E-5)

            dt  "   Y



the number of the microorganisms increasing logarithmically in accord-



ance with equation E-l.  If B remains relatively constant during the



course of the oxidation reaction, then






            •^— = constant                                        (.E-bJ
            dt



 If, on the other  hand, the concentration of microorganisms changes



 considerably, then  equation E-l  can be  integrated






            'B|B=Al 0   dt                                       (E-7)



            'B          /o



             °B  = B  eWt                                           
-------
                                                                 E-3
                  B

           s _s . J. e
            o     Y
/u   t)
V, max /
or, taking logarithms of both sides of the equation,  one obtains






           log (S -S) = log 5°  +  JSISI                     (E-13)
                 O          X       £* O



       These equations (E-6 and E-13) can now be used from a biologi-



cal viewpoint as models to account for the fate of Fe(ll)  in natural



waters.
                                References
1)  Monod, J., Recherces sur la Croissance des Cultures  Bacteriennes,

       Hermann and Cie, Paris (1942)

-------
                             APPENDIX F
Autotrophic Iron Bacteria - Ratio of Ferrous Iron Oxidized to Organic
Carbon Synthesized
       The free energy released by the oxidation of ferrous iron,

using the data available in Latimer (l), is


            4 Fe+2 + 02 +  4 H+  =  4 Fe+3  + 2 H 0               (F-,1)


      AF°  4C-20.3)  (0)   4(0)     4C-2.53)  2(-56.7)

          £AF° = -42.4 K cal. or

             -10.6 K cal./mole of Fe(ll) oxidized.

At pH 3, since

      AF =AF° + RT In Q                                          (F-2)

where Q is the reaction quotient, the free energy released per mole is


      AF = -10.6 + 1.364 log —~ = -6.5 K cal./mole            (F-2a)
                              10 "J

       For synthesis of cell material from CO- (assuming the assimi-

lated end product to be glucose), the free energy required is


           6 C02   +6 H20     = C6H12°6   * 6 °2                 (F"3)


      AF°  6C-92.3)  6C-56.7)     (-217.0)   6(0)

          2AF  = -!- 677 K cal./mole of glucose or

             +115 K cal./mole of carbon synthesized.

       Assuming a 367» efficiency for microbial conversion of energy as

is common in auto trophic processes (2), the stoichiometry of the

-------
                                                                   F-2
autotrophic oxidation of ferrous iron is

            6.5 K cal./55.8 gms. Fe(ll) oxidized     Q -&    1    ,_.
            115 K cal./l2 gms. carbon assimilated  X        250

Hence, one gram of organic carbon is synthesized for every 250 gms

of Fe(ll) oxidized.

       If one considers the thermodynamic relationships in another

way,

                                        6.5 K cai.         	
            1 mole of Fe(II) oxidized
                                        115 K cal./mole of carbon
                           0.36 x    Sm%	—                  (F-5)
                                  mole of carbon


                     = 0.25 gms of carbon synthesized


       Lamanna  and Mallette (3) approximate that  1 gm. of bacteria
            1 o       10
contains  10  to  10   bacterial cells.  Therefore, 1 mole of Fe(Il)

                       12
yields approximately 10   bacterial  cells or


            Y = ^1 = 1012 cells/mole of Fe(ll) oxidized          (F-6)
                  dS
                             References
 1)  Latimer, W. M., The Oxidation States of the Elements  and Their
       Potentials  in Aqueous Solutions, second ed., Prentice-Hall,
       Inc., Englewood Cliffs, New Jersey (1952)

 2)  McCarty, P. L., "Thermodynamics of Biological Synthesis and
       Growth," in Advances in Water Pollution Research,  J. K. Baars,
       ed., vol. 2, Pergamon Press, New York  (1965)

 3)  Lamanna, C., and Mallette, M. F., Basic Bacteriology,  third  ed.,
       The Williams and Wilkins Co., Baltimore (1965)
                                        V. S. GOVERNMENT PRINTING OFFICE : 1970 O - 401-741

-------