Diss. ETH No. 8934
THE REDUCTIVE DISSOLUTION OF HEMATITE (a-Fe203) BY ASCORBATE
A dissertation submitted to the SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZURICH
for the degree of Doctor of Natural Sciences
presented by Steven Banwart
M.S. Civil and Environmental Engineering born 30 July 1959
citizen of USA
accepted on the recommendation of Prof. Dr. Werner Stumm, examiner
Prof. Dr. Walter Schneider, co-examiner 1989
Under heaven nothing is more soft and yielding than water. Yet for attacking the solid and strong, nothing is better; It has no equal. The weak can overcome the strong; The supple can overcome the the stiff. Under heaven everyone knows this, Yet no one puts it into practice. Therefore the sage says:
He who takes upon himself the humiliation of the people is fit to rule them.
He who takes upon himself the country's disasters deserves to be king of the universe.
The truth often sounds paradoxical.
Lao Tsu
ACKNOWLEDGEMENTS
There is a theory that we are born with all the knowledge of the universe and that we grow more ignorant with age. The first step in this regression is making the mistake of opening our mouths and learning to speak. The second step is making the mistake of learning to write, leaving a physical record of our ignorance which is not drowned out by the noise of a passing train (or jet). Every six years since I was six years old I have completed parts of my formal educa-tion, my own milestones in a wild pursuit to recapture the knowledge which I happily pondered in the womb. It would be a tough race if it weren't for the others who join in the watching, measuring, studying, and discussing. It may be a mad pursuit but at least it's not lonely and I want to thank my friends who were a part of my studies at the EA WAG.
Above all I thank my "doctor father" Werner Stumm and my mentor Jerry Schnoor, very simply, for giving me the chance to learn; I value it dearly. I also thank Simon Davies for helping me to get started on my lab work and Barbara Sulzberger for the great discus-sions we had- especially the week-end brunch sessions. I value the friendship of these two scientists, and the help they gave me, and wish them every success and happiness in their careers. I want to thank Prof. Walter Schneider, a man of many talents, for being on my exam-ining committee; and Laura Sigg for whom I worked as a teaching assistent- she's a great boss and I enjoyed working with her in the NDS labs.
My time at the EA WAG spanned, at least in part, the careers of twenty-one other PhD. students in the chemistry department. In the beginning there was... Geri, Bettina, Hans-Jakob, Bernhard, Erich, and Vreni- I thank them for tolerating my bad german. I especially want to thank Norbert and P.C. for their "sprachtips" and general help in settling into Switzerland. And those who came after ... Beat, Paul, Slavi, Yiwei and Yuegang, Jurg, Gianluca, Michael, Adrian, Annette, Magnus, and Joseph- I thank them for tolerating my bad german. I also had three amazing fellow students whose careers at EA WAG coincided
almost exactly with mine; Christophe, Daniel, and Dieter. Christophe is a true kindred spirit to me; one finds them seldom. What can I say? Our friendship began the day he shared his canned ratatouille with me when we both rented rooms at Frau Ringer's. We solved a lot of environmental problems together. Daniel, topping my list of Who's Who in Switzerland, helped me enormously in many ways- going beyond the call of duty to help me with my duties as a teaching assistent (equivalent to tolerating my bad german), and providing helpful discussions on work, politics, life, love, and the swiss army. I thank Dieter Raab for improving my bad german (or was it the beer that helped?) and for being the "glue" within our NDS class and later within the research group. These three proved their friendship to me many times, above all with actions, and I am very happy that I spent this time together with them. To the post-doc/visitor group I thank Litsa, Phillipe, Laurent, Reiner, and Jordi- all wonderful people to work with... and learn from. It was great fun. Steve McDow is the poet laureat of EA WAG and he and his guitar were an important part of the past few years, I'm grateful for the blues. I thank Janet and Sue for catering so many of our parties; folks can't live on theory alone. Thanks to Albert and Barbara for providing us with week-ends and evenings in Kanton Argau; a change is as good as a vacation.
Zurich is a cross-roads and I met many friends here- Larry and Jeanie, Vibeke, Antonio and Fabiolla, Silvio, Paul and Jane and Julie, Jim and Lynn and Beth, Susan M-D, Agatha and Norbert, Bruce H., Mike, Drew, Bruce F., Stan and Lisa, Priska, Lex, Marijka, Lizzy, Han Bin and Jian, Wong, Morgan and Anjou, Andree, Monika, Francis ... They all proved, in the most positive sense, the quote from Antonio "It is very important to eat well... and with the right people".
I thank all my fellow students and co-workers at EA WAG and want to especially acknowledge the help of Sonja Rex, Irma Kipfer, Hans-Ueli Laubscher, Maria Steiner, Ursula Mohlberg, Werner Roth, David Kistler, Max Reutlinger, Bruno Gisler, Claudia Maeder, and Elisabeth Stiissi.
Finally I wish to thank my throughout my education and Switzerland.
family for their love and support especially during my stay in
SUMMARY
The dissolution of iron(III) oxides and hydroxides plays an im-portant role in the cycling of nutrients and pollutants in natural wa-ters. One possible kinetic pathway for dissolution in the absence of light is the interaction between reducing organic compounds as electron donors with the iron(III)(hydr)oxide surface producing soluble iron(II). Laboratory experiments using hematite (a.-Fe203) as a model oxide surface and ascorbate as a reductant were carried out under controlled conditions in order to study the chemical processes taking place on the iron(III) oxide surface. The kinetic effects of protons and oxalate, a chelate ligand, on the the rate of reductive dissolution were assessed.
The reductive dissolution of hematite by ascorbate can be described by a direct reaction of ascorbate with the hematite surface through formation of a surface iron(III)-ascorbate complex, followed by electron transfer and subsequent release of iron(II) from the surface into solution. A 2: l stoichiometry is observed between iron(II) and ascorbate which reflects the two-electron loss from ascorbate going to dehydroascorbate, its first stable oxidation product. Comparative adsorption of ascorbate on hematite and l)-Al203 showed that ascorbate which disappears from solution in the presence of hematite particles is present predominantly as a surface iron(III)-ascorbate complex rather than being extensively oxidized.
The rate of reductive dissolution is proportional to the surface concentration of ascorbate and a first-order rate law can be written which explicitly includes the surface concentration of the reacting species.
RATE= ke{>FeIIIHA}
Adsorbed protons accelerate the dissolution by assisting the detach-ment of iron(II) from the hematite surface. The rate constant, ke, showed a fractional order dependence on proton concentration (approximately [H+]0.6) which, with help of a master isotherm for proton adsorption on metal oxides, can be interpreted as a third-order
dependence on surface proton concentration. Adsorption of oxalate, a ligand capable of forming mono-nuclear
bi-dentate surface complexes, in the presence of ascorbate leads to a significant increase in the rate of reductive dissolution even though adsorption of oxalate displaces some of the ascorbate from the surface. The rate of reductive dissolution in the presence of oxalate, however, depends on the surface concentrations of both ascorbate and oxalate. Most likely oxalate acts to accelerate the rate-determining detachment of iron(II) sites from the hematite surface.
ZUSAMMENFASSUNG
Die Auflosung von Eisen(III)oxiden und -hydroxiden spielt beim Kreislauf von Niihrstoffen und Schmutzstoffen in natilrlichen Gewiissern eine wichtige Rolle. Ein moglicher Mechanismus filr die Auflosung in der Abwesenheit von Licht ist die Wechselwirkung zwischen reduzierenden organischen Verbindungen als Elektronen-donoren und der Eisen(III)-(hydr)oxidoberfliiche, wodurch lOsliches Eisen(II) entsteht. Es wurden Laborexperimente mit Hiimatit (a-Fe20 3) als Modell-Oxid und Askorbat als Reduktionsmittel unter kontrollierten Bedingungen durchgefilhrt, um die chemischen Prozesse an der Eisen(IIl)oxidoberfliiche zu studieren. Die kinetischen Einflilsse von Protonen und Oxalat, einem Chelatliganden, auf die Geschwind-igkeit der reduktiven Auflosung wurden untersucht.
Die reduktive Auflosung von Hiimatit durch Askorbat kann als eine direkte Reaktion von Askorbat mit der Hiimatitoberfliiche beschrieben werden, wobei das Askorbat mit Eisen(III) an der Oberflliche einen Komplex bildet. Darauf folgt ein Elektronentransfer und das so produzierte Eisen(II) geht in Losung. Es wird eine 2: I St0chiometrie zwischen Eisen(II) und Askorbat beobachtet, was zeigt, dass Askorbat zwei Elektronen abgibt und in Hydroaskorbat, das erste stabile Oxidationsprodukt, ilbergeht. Ein Vergleich zwischen der Adsorption von Askorbat an Hiimatit und 8-Al20 3 zeigt, dass das aus der Losung verschwundene Askorbat an der Hiimatitoberfliiche vorwiegend als Eisen(III)-Askorbat Komplex vorliegt und nicht zu einem grossen Teil oxidiert ist.
Die Geschwindigkeit der reduktiven Auflosung ist proportional zur Oberflii.chenkonzentration von Askorbat und es kann ein Geschwin-digkeitsgesetz erster Ordnung aufgestellt werden, welches explizit die Oberfllichenkonzentration der reagierenden Spezies enthalt:
Auflosungsgeschwindigkeit ke{>FeillHA}
Adsorbierte Protonen beschleunigen die Auflosung, indem sie die Loslosung des Eisen(II) von der Hamatitoberflache beschleunigen. Die Geschwindigkeitskonstante fiir diese "protonenunterstiitzte" reduktive Auflosung, ke, hat eine nicht-ganzzahlige Abhlingigkeit von der Proto-nenkonzentration (ungefiihr [ff]0.6), welche mit Hilfe einer Standard-Adsorptionsisothermen fiir Protonen an Metalloxiden als Abhangigkeit dritter Ordnung von der Protonenkonzentration an der Oberflache in-terpretiert werden kann.
Die Adsorption von Oxalat, einem Liganden, der einkernige, bidentate Oberflachenkomplexe bilden kann, fiihrt zu einer betracht-lichen Beschleunigung der reduktiven Auflosung mit Askorbat, obwohl das adsorbierte Oxalat einen Teil des adsorbierten Askorbats ver-drlingt. Die Auflosungsrate ist also sogar erhoht, obwohl weniger Re-duktionsmittel an der Oberflache adsorbiert ist. Die Hypothese wird aufgestellt, dass das Oxalat die Auflosung beschleunigt, indem es die geschwindigkeitsbestimmende Loslosung des Eisen(II) von der Hlim-atitoberflache beschleunigt.
1
Table of Contents
SUMMARY/ ZUSAMMENFASSUNG LIST OF SYMBOLS CHAPTER 1. INTRODUCTION AND SYNOPSIS 1.1 THE TRANSFORMATIONS OF IRON OXIDES IN NATURAL
WATERS 1.2 MAIN RESULTS
CHAPTER 2. SURFACE CHEMISTRY AND THE
4 6
6 7
CHEMICAL KINETICS OF MINERAL DISSOLUTION 1 0 2.1 INTRODUCTION 1 0 2.2 THECHEMICALSTRUCTUREOFHEMATITE 10 2.3 THE CHEMICAL KINETICS OF DISSOLUTION 1 2 2.4 THE COMBINATION OF A REDUCT ANT AND CHELATE LIGAND 1 7 2.5 RA TE LAWS FOR SURFACE CHEMICAL REACTIONS 1 8
CHAPTER 3. THE REDUCTIVE DISSOLUTION OF HEMATITE BY ASCORBATE 25 3.1 INTRODUCTION 2 5 3.2 EXPERIMENTAL METHODS 25 3.3 RESULTS 2 7 3. 3 .1 Dissolution Kinetics of Hematite in the Presence of
Ascorbate 2 7 3 .3 .2 Determination of the Adsorption of Ascorbate 2 9 3.3.3 The Adsorption Kinetics of Ascorbate on Hematite 3 1 3.3.4 The Comparative Adsorption of Ascorbate on Hematite
and Aluminum Oxide 3 2 3 .3 .5 Estimation of the Surface Stability Constant for
the lron(Ul)-Ascorbate Complex 3 .3 .6 The Rate Dependence on Ascorbate Concentration 3.3.7 The Rate Dependence on pH 3.3.8 The Effect of pH on Ascorbate Adsorption
35 38 42 43
2 3.4 DISCUSSION 3 .4 .1 The Stoichiometric Reaction for the Reductive
Dissolution of Hematite by Ascorbate 3.4.2 Experimental Information for Formulation of a
Mechanism of Reaction 3.4.3 A Rate Expression for the Reductive Dissolution of
Hematite by Ascorbate 3 .4.4 The Rate Dependence on Surface Proton
Concentration 3.4.5 The Electron Transfer Reaction
CHAPTER 4. THE REDUCTIVE DISSOLUTION OF HEMATITE BY ASCORBATE IN THE PRESENCE OF OXALATE 4.1 EXPERIMENTAL METHODS 4.2 RESULTS AND DISCUSSION 4.2.1 The Adsorption of Oxalate on Hematite 4.2.2 Dissolution Kinetics of Hematite in the Presence of
Ascorbate and Oxalate 4.2.3 A Comparison of The Dissolution of Hematite by
Ascorbate in the Presence and Absence of Oxalate 4.2.4 The Effect of Ascorbate Concentration on the Rate of
Dissolution in the Presence of Oxalate 4.2.5 The Effect of Oxalate Concentration on the Rate of
Dissolution in the Presence of Ascorbate 4.2.6 The Effect of pH 4.2.7 The Role of Oxalate in Promoting Reductive
45
45
46
47
49 53
57 57 57 57
61
64
70
75 79
Dissolution 8 1 4.2.8 Protons and Ligands as Environmental Factors 8 4
CHAPTER 5. A COMPARISON OF THE EFFECTS OF OXALATE, PHOSPHATE, EDT A, AND ALUMINUM ON RATES OF REDUCTIVE DISSOLUTION 8 6 5.1 INTRODUCTION 8 6 5.2 EXPERIMENTAL METHODS 8 6 5.3 RESULTS AND DISCUSSION 8 7 5.3.1 The Effects of EDTA and Phosphate on the Rate of
Reductive Dissolution 87
3 5.3.2 The Reductive Dissolution in the Presence of
Aluminum 89
CHAPTER 6. CONCLUSIONS AND PERSPECTIVE 9 4 6.1 CONCLUSIONS 9 4 6.2 THE CYCLING OF IRON IN NA11JRAL WATERS 9 6 6.3 FROM CHEMICAL .KINETICS TO SYSTEM-LEVEL EFFECTS 9 8 6.4 FUWRE ENVIRONMENTS 100
APPENDIX I. ADSORPTION ISOTHERMS 101 APPENDIX II. EXPERIMENTAL DATA 110 LIST OF REFERENCES 1 I 6 CURRICULUM VITAE
SKsur ke RATE k'e
Xa Pj s ~l (intr)
[X]o [Xlsol [XlT Ca asp Cs Csites
c:q a b
4
LIST OF SYMBOLS
surface stability constant (l moI-1) first-order rate constant for reductive dissolution (hr· I) dissolution rate (mol m-2 hr 1) fourth-order rate constant for proton-assisted reductive dissolution (m6 moI-3 hr l) first-order rate constant for ligand-promoted reductive dissolution (hr- 1) surface concentration of species X (mol m·2) solution concentration of species X (mol i- l) adsorption maximum in Langmuir isotherm (mol m-2) first-order rate constant for ligand-promoted dissolution rate law written in terms of solution species (hr- 1) rate constant for proton-assisted dissolution (system dependent units) rate constant for catalytic reductive dissolution (hr· I) precursor complex concentration (mol m-2) mole fraction of reactive surface sites probability of precursor configuration for Cj total concentration of surface sites (mol m-2)
first intrinsic surface acidity constant (mol 1- l)
initial solution concentration of species X (mol 1- l) measured solution concentration of species X (mo! 1- l) total concentration of species X (mol 1-l) surface area concentration in colloid suspension (m2 1- l) specific surface area of colloid particles (m2 g· l) concentration of colloid in reacting suspension (g l- l) concentration of ion-exchange sites in reacting suspension (mo! 1- l) first distribution coefficient slope of double-reciprocal Langmuir plot (rn2 1- l) y-intercept of double-reciprocal Langmuir plot (rn2 mol-1) forward rate constant for ith step in reaction sequence
m pHzpc [H+lzpc ~pH
n ko
Re,T ke,H
ke,H,ox
I\ 1 \}'
R K F c
5 backward rate constant for ith step in reaction sequence Freundlich adsorption constant (system dependent units) Freundlich adsorption exponent pH of zero proton condition proton concentration corresponding to pHzpc pH difference between solution pH and pHzpc kinetic reaction order rate constant for reductive dissolution at unit surface proton concentration (m4 mo1-l hr 1) redox potential on hydrogen scale (V)
redox potential on hydrogen scale at pH 7 (V)
redox equilibrium constant second-order constant for oxalate-promoted reductive dissolution (m2 mol-1 hr-1) total rate of reductive dissolution (mol m-2 hr 1) pseudo first-order rate constant for proton-assisted reductive dissolution (hr 1) pseudo first-order rate constant for oxalate-promoted reductive dissolution (hr 1)
first conditional surface acidity constant (mol 1-1)
surface potential (V) ideal gas constant (8.314 J mo1-l K-1) temperature Kelvin Faraday constant (96,400 C mo1-l) surface capacitance (F m-2) surface charge (C m-2) Langmuir adsorption constant (mol 1-1)
6
Chapter 1. Introduction and Synopsis
1.1 THE TRANSFORMATIONS OF IRON OXIDES IN NATURAL WATERS The abundance of iron, its coordination chemistry, and its oxida-
tion-reduction properties place it in a central role in the chemistry of natural aquatic systems. Transformations between soluble iron(II) and solid iron(III)(hydr)oxides across redox boundaries, often associ-ated with large dissolved oxygen gradients in waters, soils, and sedi-ments create local redox cycles and chemical gradients which play an important role in controlling the availability of adsorbed nutrients (phosphate) and pollutants (both metals and organic compounds) to the biogeochemical cycles operating on larger scales. These processes are important for transformations of other reducible metal oxides such as manganese oxides as well. A general review of iron and manganese transformations at redox boundaries is provided by Davison (1985).
The interaction of both soluble and solid phase iron with bio-logically produced organic compounds may provide important kinetic pathways for the transformations of iron in nature. These biogenic exudates provide both reductants such as amino compounds or other reduced organic metabolytes and also chelating ligands such as simple carboxylic acids which can determine the predominant pathways for dissolution and alter the rates of reaction within these pathways (Sulzberger,1989). The production and supply of organic exudates in the soil environment has been reviewed by Stevenson (1967). The interaction of biologically produced compounds with dissolved or col-loidal iron may effect not only the chemical composition of natural waters but in turn the composition and productivity of the biological community. Extensive work on the biological production of iron spe-cific chelates (siderophores) exuded by algae, fungi, and bacteria, which enhance iron uptake by these organisms, has been carried out by Neilands (1967,1981,1982,1984). Iron (II) can be a growth-limit-ing nutrient in the marine environment (Lewin and Chen,1971) and can also determine the composition of algal communities in lake wa-ters under certain conditions (Murphy and Lean,1976). Such effects combined with the ability of solid-phase iron(III) oxides and hydrox-ides to scavenge toxic substances from surface waters, the marine en-
7
vironment, or in soil and groundwater systems creates a network of feed-back mechanisms to the biological community which plays a role in controlling the overall environmental quality of aquatic ecosystems. A broad goal of this work is to gain a better understanding of how chemical analogues to microbially produced reductants and chelate lig-ands effect the rates of dissolution of iron(Ill) oxides and hydroxides. Laboratory studies were carried out under controlled conditions (constant pH, ionic strength, temperature, exclusion of 02) using hematite as a model iron oxide, ascorbate as a reductant, and oxalate as a chelate ligand in order to study the basic chemical processes oc-curring on the iron oxide surface.
It is known that ligands able to form mono-nuclear chelate com-plexes with hydrated metal atoms on the surface of a metal oxide min-erals accelerate the dissolution (Furrer and Stumm,1986) and that re-ductants can act to accelerate the dissolution of iron(III)(hydr)oxides. It has been shown that the effect of a reductant acting together with a chelate ligand markedly accelerates the reductive dissolution of iron(III)(hydr)oxides (Zinder et. al., 1986). The main goal of this work is to better understand the reductive dissolution of hematite by ascorbate in the absence of light, and the acceleration of this reaction by protons and oxalate. Related research at the EA WAG includes the catalytic reductive dissolution of geothite by iron(II) complexes (Suter, 1989) and the photo-induced reductive dissolution of iron oxides (Siffert, 1989). Other areas of application where the reductive disso-lution of iron oxides is of interest include passive film formation in corrosive dissolution, the electrochemistry of metal oxide semi-conductors, and biochemical transformations of iron.
1.2 MAIN RESULTS The experimental results can be divided into four sections; the ad-
sorption of ascorbate on hematite and 8-Al203, the reductive dissolu-tion of hematite by ascorbate alone, the reductive dissolution of hematite in the presence of ascorbate and oxalate together, and a com-parison of the kinetic effects of oxalate, EDTA, phosphate, and alu-minum on the rate of reductive dissolution. A comparison of the ad-sorption of ascorbate on hematite and 8-Al203 showed that the ascor-bate which disappears from solution in the presence of hematite parti-
8
cles is present primarily as a surface iron(Ill)-ascorbate complex rather than being extensively oxidized. The stability constants for formation of surface iron(III)- and aluminum-ascorbate complexes were estimated from Langmuir adsorption isotherms at pH 3.
SKsur=9.6x105 1 mo1-l (1.1)
SKsur=7.5x105 1 mol-1 (1.2)
The rate of reductive dissolution of hematite by ascorbate alone was much faster than the acid dissolution of hematite in the absence of ascorbate. The rate of reductive dissolution was found to be first-order in surface ascorbate concentration. An empirical rate law explic-itly containing the surface concentration of a surface ascorbate com-plex can be written
RATE= ke{>FeIIIHA) (mol m-2 hr-1) (1.3)
The fractional-order dependence of the rate constant (ke, hr-1) on proton concentration (approximately [H+]0.6), with help from a master isotherm for proton adsorption on oxide surfaces, can be interpreted as a third-order dependence on surface proton concentration. A more general rate law for the "proton-assisted" reductive dissolution de-pending both on the surface concentrations of ascorbate and protons may be written.
(k'e=l.6xI015 m6 mo1-3 hrl) (1.4)
The rate of reductive dissolution was markedly accelerated in the presence of oxalate. A comparison of the empirical first-order rate constants (ke) at pH 3 and 4, in the absence and presence of oxalate, are listed in Table 1.1. The rate of dissolution was much faster than the sum of the rates of dissolution determined from experiments with ascorbate and oxalate reacting separately. Adsorbed oxalate accelerated the reductive dissolution while displacing ascorbate from the hematite surface.
9
Tabel 1.1 Empirical First-Order Rate Constants for the Reductive Dissolution of Hematite by Ascorbate in the Absence and Presence of
5. Ox 10-5 mol 1-1 Oxalate
pH 3 ascorbate 0.11 hr 1
ascorbate + 1.2 hr 1 oxalate
hematite=0.613 g 1-1
pH 4 0.045 hr 1
0.16 hr 1
The acceleration of the reductive dissolution in the presence of oxalate was only observed under conditions were oxalate, rather than ascorbate, was the predominant surface species. Under these conditions, the increase in the rate of "oxalate-promoted" reductive dissolution was approximately linear with the surface concentration of oxalate.
The rate of oxalate-promoted reductive dissolution showed a maximum with regard to pH at pH 3. At pH 2.5 the slower rate of dis-solution was due to much weaker adsorption of ascorbate. The oxalate-promoted reductive dissolution decreased with increasing pH above pH 3 with no observable rate of dissolution at pH 6. Adsorption of iron(II) may be occurring at pH 6 and inhibiting the reaction by blocking acitve sites for dissolution. The reductive dissolution was in-hibited by the presence of aluminum at pH 4 but not pH 3, coinciding with a strong adsorption edge for aluminum at pH 3.5. Aluminum would adsorb preferentially at crystal dislocations thus blocking the reactive surface sites. EDT A lead to a significant dissolution at pH 6 in the presence of ascorbate and may be acting both to accelerate the dis-solution through formation of surface complexes and possibly to pre-vent inhibition by keeping iron(II) in solution through formation of dissolved complexes. Phosphate, like oxalate, accelerated the reduc-tive dissolution at acid pH while displacing ascorbate from the hematite surface.
10
Chapter 2. Surf ace Chemical Kinetics of
2.1 INTRODUCTION
Chemistry and the Mineral Dissolution
The coordination chemistry of the interface between solid phases and aqueous solution has dealt mainly with the thermodynamics of acid/base equilibria and the adsorption of metal ions and anions from solution. A description of these processes as chemical coordination re-actions, combined with traditional and modified electrostatic models of the electrical double-layer have given rise to several surface specia-tion models (Westall and Hohl,1980; Sposito,1983). The fundamental chemical aspects, with attention to formulation of thermodynamic mass laws involving well-defined chemical species, have been re-viewed by Schindler and Stumm (1987).
The kinetics of protonation and de-protonation, surface metal complex formation, and surface metal-ligand complex formation have also been experimentally studied and included within the general the-oretical framework of surface coordination chemistry (Astumian et. al.,1981; Ikeda et. al.,1982; Hayes and Leckie,1986). The kinetics of metal oxide dissolution is frequently controlled by surface chemical reactions and can be described through formulation of kinetic rate laws which contain surface concentrations of reacting species related to the precursor complex for the rate-determining step in the overall reaction (Stumm and Furrer,1987).
2.2 THE CHEMICAL STRUCTURE OF HEMATITE The basic structural unit for all iron(III)(hydr)oxides is a central
iron(III) atom coordinatively bound to six o-2 or OH- in an octahedral or distorted octahedral arangement. The crystal structure of hematite is simply an infinite three-dimensional array of Fe06 octahedra (Schwertmann and Taylor,1977). These octahedra are slightly dis-torted giving a primitive crystal unit cell which is rhombohedral (a=5.427A, a.:::;55018'; Deer et. al.,1980). The geometry of an idealized hematite octahedron where the distortion has been neglected is shown in figure 2.1 Exact bond lengths for the hematite Fe06 -9 octahedron are given by Sherman (1985).
1 1
Figure 2 .1 The idealized geometry of a hematite octahedron. The 0-0 bond lengths in the four-coordinated plane are approximately 2.8 A. The Fe-0 bond lenghts perpendicular to the four-coordinated plane are approximately 2.0 A (Sherman,1985).
These octahedra taken as separate entities can be viewed as being analogous to iron(llI) octahedra in aqueous solution. When such an iron group at the hematite surface comes in contact with the aqueous phase the chemical adsorption of water gives rise to a surface iron(III)-aquo complex (Hair,1967). This solvation reaction can be written as
Fe06-9(s) + 2H+(aq) +:! Fe05(0H2)-7 (sur) (2.1)
Frequently such a surface group is written in short-hand with the total charge on the surface group being given a reference state of 0 when the surface iron(III) group is singly coordinated to a hydroxyl group. This is shown in figure 2.2. The chemisorbed water may undergo protonation/deprotonation and the resulting hydroxyl groups may participate in metal complexation reactions similar to hydroxyl ions in solution.
12
' a;(+1) Fe-/ 2
' a;(O) Fe-/
' ( • 1 ) Fe- 0 /
Figure 2.2 Simple structural formula for surface iron(lll) groups. The reference charge for the surface iron(///) mono-hydroxo complex is 0.
The surface iron(llI)-aquo complex may undergo ligand exchange where adsorbed water is replaced with ligands from solution. Such reactions are shown in Table 2.1 with the corresponding thermodynamic mass laws.
Surface groups may have more than one lattice oxide ion exposed to the solution. These groups, being doubly or triply coordinated with solution species, would be found at geometric dislocations on the sur-face such as "steps", or "corners" (Blum and Lasaga,1987) It is be-lieved that such physical dislocations on the surface are more chemi-cally "reactive". In a non-rigorous way this can be viewed as sites where the surface metal groups have a freer coordination environment for participation in reactions with the aqueous phase. These sites have surface metal complexes with fewer crystal lattice oxygens as coordi· nation partners (Valverde and Wagner,1976; Segal and Sellers,1984; Blesa and Maroto,1986; Blum and Lasaga,1987). A diagram of disloca-tions on a crystal surface is shown in figure 2.3.
2.3 THE CHEMICAL KINETICS OF DISSOWTIQN The dissolution of a mineral phase such as hematite is seen as a
change in the coordination environment of a specific surface species. The surface metal group exchanges crystal lattice oxide ions for sol-vent water or other dissolved ligands (Furrer and Stumm,1986). Chemical bonds must be broken and formed.
1 3
Table 2.1 Surface Acid-Base and Complexation Equilibria .
Acid-Base: >MOH2+ f2 >MOH+ H+ . s _{>MOH}[H+] ,Kai- {>MOH2+}
>MOH f2 >MO- + H+ s {>MO-}[H+]
;Ka2 = {>MOH}
Metal Ions: >MOH+ Me+2 f2 >MOMe+ + H+ ·Ks _ {>MOMe+)[H+] ' 1 - {>MOH}[Me+2]
Ligands:
s {(>M0)2Me}{[H+]2 2>MOH + Me+2 f2 (>M0)2Me + 2H+ · B ' 2 {>MOH}2[Me+2]
>MOH+ L-Z f2 >ML-(z-1) +OH- . Ks _ .._{ >_ML_-_(z_-1_) }..,.[O_H__...-] ' 1 - {>MOH}[L-Z]
>MOH+ HL-(z-1) f2 >ML(-z-1) + H20 . * l Ks __ .._{ >_M_L_-_(z_-_I ) .... }_ ' 1 - {>MOH}[HL-(z-1)]
The notation >MOH corresponds to one surface metal group such as shown in figure 2.2. Curved brackets, { } , indicate surface concentrations (mol m-2) and square brackets, [ ], indicate solution concentrations (mol 1-l ). Taken from Furrer and Stumm (1986).
14
Q-ll Q-l ~ • Q-l ~ . ,, ..
-- Fe ---- 0 Fe 11111 Q-l
~~ ~ ~-.....-- Fe ~ ~ O • "corner"
0 Fe ••Gl
~ "step"
0
/ 11111Q-l
Figure 2.3 Geometric arrangement of surface iron(///) groups at crystal dislocations
15
Reactants and products have specific chemical identities (species) and chemical rate laws including these species can be written and experimentally tested. Wieland et. al. ( 1988) have shown that the rate of dissolution for a variety of oxide minerals correlates to some extent with the strength of the metal-lattice oxygen bond. This indicates that bond breaking is important in the overall dissolution process. The ad-sorption of protons, ligands, and metals can enhance or inhibit this bond-breaking step in the dissolution (Valverde and Wagner,1976; Blesa and Maroto,1986; Stumm and Furrer,1987). Here some examples from published dissolution studies are presented in order to illustrate the kinetic effect of such adsorbed reactants.
It is known that the presence of F- greatly accelerates the rate of dissolution of aluminum oxide and feldspar minerals (Pulver et. al.,1984) and that the dissolution of aluminum oxide by F- may be so fast as to be diffusion controlled (Zutic and Stumm,1984). The effect of anions such a F- or CI- to accelerate the dissolution of metal oxide minerals has been explained by formation of surface complexes which "activate" the surface for dissolution. The leaching of iron from goethite and hematite by acids such as HCI and H2S04 is known to depend on the solution concentration of the anion and the ability of the anion to form complexes with iron (Surana and Warren,1969; Cornell et. al.,1976). Grauer and Stumm (1982), in a review of kinetic data for the acid dissolution of a variety of metal oxide minerals, showed that the fractional-order kinetics observed for anion concen-tration (RATE=k[L]n, n<l) could be explained by formation of surface complexes which control the rate of dissolution. The effect of surface chelate complexes to accelerate the dissolution of metal oxide minerals has also been observed (Rubio and Matijevic, 1979; Chang and Metijevic,1983; Furrer and Stumm,1986; Zinder et. al.,1986). These studies showed that adsorbed chelate ligands such as oxalate and sali-cylate, capable of forming mono-nuclear bi-dentate chelate complexes with the surface metal sites, greatly accelerate the rate of dissolution of metal oxides. It is believed that such surface complexes would po-larize the bond between the metal center and the lattice oxygen ions thus assisting the release of the metal group from the crystal surface. It was also proposed that ligands capable of forming multi-nuclear surface complexes may inhibit the dissolution due to a higher activa-
16
tion energy to release more than one metal center from the surface simultaneously.
In the case of complete transfer of an electron from a donor bound to the surface metal iron (reduction of the metal), the reduced metal group is., more easily released from the crystal surface due to the greater lability of the reduced metal-lattice oxygen bonds. For water exchange on dissolved iron-aquo complexes, a process whose energet-ics is dominated by bond-breaking, the observed rate constant is about four orders of magnitude larger for iron(II) aquo complexes than for iron(III)-aquo complexes (Basolo and Pearson,1967). An extreme ex-ample of such changes in water exchange rates upon change in oxida-tion state is given by the chrome-aquo complex which exhibits a water exchange rate with a half-life less than 10-9 seconds for Cr(II) ions and a half-life of about 106 seconds for Cr(III) ions (Taube,1968). Similar kinetic effects can be seen for the dissolution of metal oxide minerals. The rates of dissolution for higher valent metal oxide min-erals are frequently accelerated in the presence of reductants. There are several reviews concerning the reductive dissolution of mineral oxides (Gorchev and Kipryanov,1984; Segal and Sellers,1984; Stone, 1986).
It is also known that the rate of acid dissolution of metal oxides depends on proton concentration (Valverde and Wagner,1976; Blesa and Maroto,1986; Stumm and Furrer,1987). It has been shown that the rate of acid dissolution increases with increasing protonation of the oxide surface and depends directly on the concentration of protons on the mineral surface (Furrer and Stumm,1986; Carroll-Webb and Walther,1988; Schott,1989). Zinder et. al. (1986) also reported a de-pendence of the dissolution rate on surface proton concentration for reductive dissolution of goethite by ascorbic acid. One explaination is that protons on the surface are capable of attacking the crystal lattice oxygen ions, extracting electron density from, and thus weakening, the surface metal-oxygen bond (Wieland, 1987).
Another factor which plays a role in the chemical kinetics of min-eral dissolution is the number of "active sites'', the mole fraction of surface sites participating in the dissolution reaction. Active sites, surface sites at steps and corners, have a lesser degree of coordination to the crystal lattice (fewer bonds with the lattice to be broken) and
17
thus a lower activation energy for their release into the aqueous phase (Valverde and Wagner,1976; Blum and Lasaga,1987). The number of active sites depends upon the distribution of surface sites which are singly, doubly, and triply coordinated with aqueous phase reactants. On a very rough surface it is understood that there is a greater pro-portion of sites at discontinuities which are kinetically active as com-pared with a smooth surface (Blum and Lasaga,1987; Wieland et. al., 1988). In general the number of active sites is seen here to be a property of the mineral type, its history (formation, aging, weather-ing), and experimental treatment such as acid washing as a pre-condi-tioning prior to laboratory studies. Blocking of these sites by surface complexes which are not reactive would inhibit the rate of dissolution of the oxide mineral (Wieland et. al., 1988). Chou and Wollast (1985) observed that the rate of acid dissolution of albite (NaAlSi3 0 8) de-pends directly upon the concentration of surface aluminum complexes formed during the dissolution. The formation of these surface alu-minum complexes inhibited the dissolution at solution concentrations of aluminum well below the solubility limit for aluminum oxide phases. Metal complex formation on mineral surfaces would occur preferentially at the active sites (Blum and Lasaga,1987) thus blocking the surface from further reaction.
2.4 THE COMBINATION OF A REDUCT ANT AND CHELATE LIGAND The combination of a reductant and a chelate ligand leads to a
rapid dissolution of reducible metal oxide minerals (Schwertmann, 1964; McKeague and Day,1966). In published mechanistic studies of this type of system the reductant has often been a metal chelate com-plex which in some way reduces the oxide surface more efficiently than the metal ion (Fischer,1972; Valverde,1976; Segal and Sellers, 1982; Blesa et. al.,1984,1987; Cornell and Schindler,1987; Suter et. al.,1988). Reductive dissolution may also occur through specific ad-sorption of reducing organic compounds at the reducible metal oxide surface (Stone and Morgan,1984; Zinder et. al.,1986; LaKind and Stone,1988). It was pointed out earlier that the presence of chelate ligands can accelerate the non-reductive acid dissolution of oxide min-erals. Zinder (1985) showed that oxalate, a chelate ligand, greatly en-hances the reductive dissolution of goethite by ascorbate and proposed
18
that the role of oxalate is to increase the rate of detachment of iron(II) from the surface. The results of Stone and Morgan (1984) also suggest such a "chelate effect". For the systems with reducing metal-chelate complexes it is known that the chelate helps establish the reductive pathway for dissolution. The role of adsorbed chelate ligands in accelerating the reductive dissolution by reducing organic compounds as proposed by Zinder would be similar to the role of surface chelate complexes in accelerating the non-reductive dissolution of metal oxides
2.5 RATE LAWS FOR SURFACE CHEMICAL REACTIONS Based on the kinetic effects of protons, chelates, reductants, and
combinations of these reactants, four pathways for dissolution of iron(III) oxides and hydroxides in the absence of light have been pro-posed for conditions found in aquatic environments (Sulzberger et. al.,1989). These are shown in Table 2.2. The empirical rate law writ-ten for each mechanism explicitly includes the concentration of specific surface chemical species rather than dissolved concentrations of reac-tants. Such rate laws can also be formulated for solution concentra-tions of reacting species with the help of adsorption isotherms. For the case of mechanism (b), the ligand-promoted dissolution, the kinetic rate law for the dissolution is as follows. kL (hr-1) is the observed first-order rate constant and {>FelIIOx} is the surface concentration of mono-nuclear bi-dentate iron(IIl)-oxalate complex (mol m·2).
RATE = kL{>FeIIIOx} (mol m· 2 hr-1) (2.2)
The formation of the surface iron(llI)-oxalate complex may be de-scribed by the following surface chemical reaction. Surface species are written with concentration units as mol m-2 and solution species as mol 1-1.
SKsur (2.3)
The thermodynamic mass law can be written for this reaction.
19
Table 2.2 Dissolution Pathways for Iron(III) (Hydr)oxides in Natural Waters
PROTON-ASSISTED
al protons
ll
LIGAND-PROMOTED
bl
H20i w slow oxalate
+
ligands like oxalate
HO-...c,...o I
-o_.....c...,o
OH2 0 oJ.2n-3)-' / 'C' Fe111 +Fem,( 1 (aq) / 'OH '\---o,.C' 0 n
RATE= lq..{>FeIIIOx}
20
REDUCTIVE CATALYTIC REDUCTIVE
d) ligands like
cl reductant e.g. ascorbate
HO + _)(
oxalate + iron(ll) HO......c,..o
Fell + . I + aq -o,.....c"'o
0 0)
2n-3)-'C"" + Femf ' (aql
'\o...-C'o n
RATE= ke{>FelllHA} RATE= kc{>FelII(Ox)nFe(II)}
Curved brackets, { } , refer to surface concentrations of reacting species (mol nf 2 ). (from Sulzberger et. al., 1989)
21
s _ {>FeIIIOx} Ksur - {>FeIIIOH}[HOx-1 (1 mo1- l) (2.4)
This mass law combined with a mass balance for surface sites, Sm = {>FellloH} + {>FeIIIOx} (mol m-2), can be written as an isotherm. Here Sm is the adsorption maximum for oxalate (mol m-2) and SKsur (1 mo1- l) is the stability constant for the mass law (2.4). A derivation of this isotherm (Langmuir adsorption isotherm) is given in Appendix 1.
{ >FeIIlox} = _s_m_sK_su_r_[H_O_x_-_1 1 +sKsur[HOx-1
(mol m-2) (2.5)
This expression for the surface iron(III)-oxalate complex can be sub-stituted into the rate law (2.2) giving the following rate law for the solution species [HOx-1.
RATE k'LSmSKsur[HOx-] 1 +SKsur[HOx-]
(mol m-2 hr-1) (2.6)
k'L (hr 1) is the first-order rate constant corresponding to this form of the rate law.
This is a more complicated mathematical expression than equation (2.2). However, it is usually easier to analytically determine the con-centration of solution species than surface species. A greater problem is to differentiate between conflicting kinetic effects in the various steps of the mechanism. An excellent example is provided by the dis-solution of aluminum oxide in the presence and absence of oxalate (Furrer,1985). At pH values above the pK2=3.8 for oxalate, the ad-sorption of oxalate begins to decrease with pH. However, the acid dis-solution of aluminum oxide, in the absence of oxalate, decreases with increasing pH as well (figure 2.4 ). If the proton-assisted dissolution and oxalate-promoted dissolution react in parallel, how can one distin-guish between the two kinetic pathways and the kinetic effect of pro-ton concentration?
Furrer (1985) was able to establish that at a single pH the rate of dissolution was proportional to the surface concentration of oxalate as
' ..... ..c:. "" 'E
0 E .s ..
§ l..lJ I-ex: a:
22
25
20
15
10
5
0 2 3 4 5 6 7
pH Figure 2.4 Dissolution rate of o-Al20 3 at various pH values. added oxalate = lmM for pH 3 ,3.5. added oxalate = 5mM at pH 4,5,6 (Furrer,1985).
written in equation (2.2). The rate dependence on surface proton con-centration had also been determined. By observing both the rate of dissolution and the surface concentrations of oxalate and protons at various pH values, and calculating the rate of proton-assisted dis-solution due only to adsorbed protons acting alone, it was possible at
each pH to determine the first-order rate constant (kL {>::i~x}) for
the oxalate-promoted dissolution. Plotting this rate constant against pH (figure 2.5) demonstrated that the ligand-promoted dissolution of aluminum oxide was not acid catalyzed and that changes in the rate of dissolution at various pH values were due only to changes in oxalate adsorption and the rate of the parallel proton-assisted dissolution re-action. By determining surface concentrations of reactants for each experiment it was possible to differentiate between the possible ki-netic effects of less oxalate and lower proton concentration on the aluminum oxide surface at higher pH values.
-. ~ .c
'? 0 ~
..J .:.::
23
20
15
10
5
0 2 4 5 7
pH Figure 2.5 Dependence of Observed First-Order Rate Constants for the Oxalate-Promoted Dissolution of 8-Al203 (Furrer,1985).
Dissolution is a surface chemical process. The kinetics of dissolu-tion can be formulated as a chemical mechanism and rate expressions which specifically include the concentrations of surface species can be derived from the mechanism and compared with empirical rate laws. For such mechanistic investigations the analytical determination of the concentration of reactants on the surface is absolutely crucial.
One particular problem of formulating rate laws for dissolution, such as the example just given for oxalate reacting with aluminum oxide, is establishing the relation between a hypothetical mechanism and an empirical rate law which must be formulated in terms of surface species which are possible to measure in the laboratory. A generalized approach using surface coordination chemistry, activated complex theory, and surface lattice statistics has been developed by Wieland et. al. (1988). A general rate law for any dissolution mecha-nism can be written as follows.
Rate = kCj (mol m-2 hr-1) (2.7)
C j (mol m-2) is the concentration of the precursor complex in the rate-determining step of the overall dissolution reaction and is defined by
24
the proportionality Cj a xaP jS. xa is the mole fraction of reactive sites (fraction of the total number of surface sites participating in the reac-tion), Pj is the probability of a surface site having the necessary coor-dination environment to act as a precursor complex, and S (mol m-2) is the density of crystal lattice metal centers exposed to the aqueous phase. The subscript, j, refers to one of several reactions in parallel. The total rate of reaction is defined as the summation of equation (2.7) over j. This approach was applied by Wieland et. al (1988) to explain the integer reaction order often observed for surface proton concen-tration in the acid dissolution of metal oxide minerals and also to nor-malize the kinetic information from a number of dissolution studies in the literature in order to develop a Linear Free Energy Relation be-tween rate of dissolution and lattice binding energy for a variety of metal oxide minerals.
The mechanism for the reductive dissolution of iron(III) (hydr)oxides may be considered, as an initial hypothesis, to consist of three reaction steps; adsorption of reductant and formation of a sur-face complex, electron transfer within the surface complex, and release of reduced surface iron into solution (Table 2.2, mechanism c). It is known that adsorption of protons and ligands is usually very fast and comes into equilibrium within seconds to minutes (Ashida et. al.,1980; Astumian et. al.,1981; Ikeda et. al.,1982; Sasaki et. al.,1983; Hachiya et. al., 1984 ). The rates which have been reported by Zinder (1985) for reductive dissolution of goethite are much slower, on the order of 10-1 mol m-2 hr-1 (3 g 1- l goethite, 60 m2 1- l surface area); it has been assumed that detachment of iron(II) from the goethite surface was the rate determining step. The work presented in the following chapters includes the effect of protons and chelate ligands to accelerate the reductive dissolution of hematite, working under the initial hypothesis that adsorbed protons and chelate ligands enhance the release of iron(II) sites from the oxide surface.
25
Chapter 3. The Reductive Dissolution of Hematite by Ascorbate
3.1 IN1RODUCTION A general background for the chemical kinetics of mineral disso-
lution was presented in chapter 2. Many of the ideas from that chap-ter will be applied to the discussion of the experimental results for the ascorbate/hematite system. In this chapter the effect of changes in solution composition on the rate of dissolution for the reductive disso-lution of hematite by ascorbate alone is reported. Variation of ascor-bate and proton concentrations in solution and determination of sur-face concentrations of ascorbate, while observing the rate of dissolu-tion in each case, helps provide information for the formulation of an empirical rate law and to propose a kinetic mechanism.
3.2 EXPERIMENTAL METHODS The colloidal hematite used in all experiments was prepared in
our laboratory using the method of Matijevic and Scheiner (1978) as modified by Penners and Koopal (1986). Examination by electron mi-croscopy showed the resulting colloid to be a homodisperse suspension of smooth spherical particles with a uniform diameter of 50 nm. Analysis by x-ray diffraction confirmed the particles to be hematite (a-Fe203) with no crystal impurities. A pre-weighed portion of the dry solid phase was dissolved in concentrated HCl and then analyzed for Fe+3 by iodometric titration. The concentration of iron in the solid phase was 1.24x1Q-2 mol g-1, consistent with the stoichiometry of pure hematite (1.25x1Q-2 mol Fe(III) g-1 hematite). The ion-ex-change capacity as measured by fluoride adsorption (Faust,1985) was 1.78x1Q-4 mol g-1 and the specific surface area as calculated from particle geometry was 17.5 m2 g-1. Acid-base titrations at 1.0 mol 1-1 and 0.01 mol 1-1 ionic strength (NaN03) showed a zero point-of-charge
at pH=9.0 and an intrinsic surface acidity constant of pK~ 1 (intr)=7 .3
(see Appendix 1 for a discussion of proton adsorption on metal oxide surfaces). The B-Al203 (Degussa AG., aluminum oxide C) used in the ascorbate adsorption experiments had a surface area of 113 m2 g- 1
26
and an ion-exchange capacity of 2. lOx 10-4 mol g-1. The pH corre-
sponding to the zero point of charge was pHzpc =8.7 with
s pKal (intr)=7.4 (Kummert,1979; Furrer,1985). All solutions and sus-
pensions were prepared using de-ionized water (Barnstead Nanopur system) and analytical grade reagents.
All experiments were carried out in thermostatted reactors (250C) at an ionic strength of 0.01 mol 1-1 (NaN03) under the exclusion of light and with a partial pressure of 1 atm. N2 maintained inside the reactors. The colloidal suspensions were prepared by dilution of a portion of the original hematite batch together with a small amount of 1.0 mol 1-1 NaN03 to give an end concentration of 0.613 g 1- 1 hematite, 0.01 mol 1-1 NaN03. The pH was initially adjusted by a small addition of HN03. At low pH the proton concentration did not change measurably during the dissolution experiments. At higher pH values the reactor was kept at constant proton activity using a pH stat (a Metrohm 632 pH meter with combined glass electrode connected to two sets of a Metrohm 614 impulsomat with a Metrohm 535 dosimat, one for addition of acid, one for addition of base). Prior to each ex-periment a measured volume of the suspension (usually 100.0 ml) was added to the reactor and purged with the N 2 stream for at least 20
minutes. At time=O a known volume of ascorbate standard solution was added. Samples were periodically drawn with a syringe to mini-mize contact with atmospheric oxygen and filtered under nitrogen (20µm Sartorius membrane filter). The filter apparatus was kept cov-ered during filtration to eliminate photoreactions. Sample volumes were 10-15 ml. The filtrate was collected in glass vials to which 0.1 ml 1.0 mol 1-1 HN03 had been added to quench autooxidation reac-
tions. The filtrate was analyzed for iron(II) (or total iron in the case of
experiments without ascorbate) using the phenanthroline method (Tamura,1974)). Ascorbate was measured by UV spectrophotometry at 260 nm (Kontron Instruments UVIKON 860). A 2.00 ml aliquot of filtrate was added to a 1 cm cuvette. At time=O, 1.00 ml of 0.5 mol 1-1 phosphate buffer (pH=8) was added. The various ascorbate species have different absorptivities and peak absorbance wavelengths. The
27
increase in pH due to the buffer converted all ascorbate to hydrogen-ascorbate (pKl =4, pK2=l l). Hydrogen-ascorbate reacts with molecular oxygen and the measurement was corrected for the subsequent de-crease in concentration due to autooxidation. After addition of the buffer the absorbance of the hydrogen-ascorbate ion was measured at time=30,60, and 90 seconds. These absorbance measurements were extrapolated to time=O. The absorbance at time=O was corrected for the presence of iron which has an absorbance peak coinciding with that of hydrogen ascorbate. The absorbance due to iron was calculated using a calibration curve for iron absorbance at 260 nm with the iron concentration as measured independently using the phenanthroline method. This absorbance due to iron was subtracted from the total absorbance as measured above. The concentration of ascorbate was calculated from this corrected absorbance with the help of a calibra-tion curve. Calibration curves for ascorbate were quite linear over the concentration range 0-10-4 mol 1- l. Concentrations greater than 10- 4 mol 1-1 were analyzed after dilution. Calibration curves for ascorbate in the presence of 0,2.5, and 5.0xlQ-5 mol 1-1 iron(II) had slopes dif-fering by less than 2% over this concentration range with the slope slightly less at the higher concentration, presumably due to formation of an Fe(II)-ascorbate complex. The intercepts corresponded to the absorbances predicted by the calibration curve for the UV absorbance of iron(II) alone. The limit for detection of ascorbate was about lxl0-
6 mol 1-1 and the accuracy ~ 0.2x 10-6 mol 1-1. Adsorption experi-
ments were carried out in the same way as the dissolution experi-ments. A single sample was taken at t=60 minutes and the filtrate an-alyzed for ascorbate and iron(II).
3.3 RESULTS 3.3.1 Dissolution Kinetics of Hematite in the Presence of Ascorbate
The results for a single dissolution experiment are shown in figure 3.1. The kinetics of the reductive dissolution of hematite were fol-lowed by measuring the change in solution concentration of iron(II) with time. The dissolution was initially relatively rapid and then be-came slower. The hematite used in the experiments was not pre-treated in any way, such as acid washing. The initial rapid dissolution
28
is believed here to be the release of more easily detached surface sites into solution. After the first hour of the experiment the rate of disso-lution became linear with time. The rate of dissolution was taken to be the slope of the linear portion of the iron(II) vs. time curve (1.31xto-6 mol 1-l hrl, 0.613 g 1-1 hematite).
w ~ co a: § <( Cl w ::'.i 0 ~ i5
60
";"-' ....J 50 0
<D ::iE 40 \ 'o -r--- 30 -z 0 20 a: 0 10 z <(
0 0 2 4 6 8 10 12 14
TIME (hr)
Figure 3.1. Dissolved concentration of iron(ll) and ascorbate versus time in the presence of 5 .OxJ0-5 mol l-1 added ascorbate, pH 3, Hematite=0.613 g l-1. [H2AJ0 and [Fe+2 Jo are the initial concentrations obtained by extrapolation of the kinetic curves to time=O.
The disappearence of ascorbate from solution was also followed during the dissolution experiment. The disappearence of ascorbate took place in two stages; a rapid disappearence during the first hour of the experiment, due presumably to fast adsorption kinetics, and a sec-ond slower disappearence from solution due presumably to the subse-quent oxidation of adsorbed ascorbate by surface iron(Ill) sites. The slower rate of disappearence of ascorbate after the first hour of the experiment was linear with time and, as with the iron curve, the rate of disappearence was calculated as the slope of this linear portion of the ascorbate curve (6.4xI0-7 mo11-l hrl). The rate of disap-
29
pearence of ascorbate is approximately one-half the rate of dissolution reflecting the stoichiometry of a two-electron transfer from ascorbate to iron(III).
3.3.2 Determination of the Adsorption of Ascorbate The amount of ascorbate adsorbed at any sample time during the
experiment could be calculated by difference from a mass-balance using the added concentration of ascorbic acid as the total concentra-tion.
(mol 1-1) (3.1)
[H2Alsur is the concentration of ascorbate adsorbed, [H2A]T is the total concentration of ascorbate added, [H2A] sol is the concentration of all ascorbate species in solution (H2A,HA-,A-2), and [H2Aloxidized is the concentration of ascorbate which has been oxidized. The amount of oxidized ascorbate was estimated to be one-half the amount of iron(II) produced, based on the 2: 1 stoichiometry observed.
(mo11-l)
(3.2)
[Fe+2]sol is the concentration of all iron(II) species in solution and [Fe(II)lsur is the concentration of all surface iron(II) species. [H2A]T is known, and [H2Alsol and [Fe+2]sol were analytically determined. As a first approximation, the concentration of iron(II) on the surface was neglected. Using equation (3.2) and the assumption above, the con-centration of adsorbed ascorbate was calculated.
(mol 1- l) (3.3)
Using the specific area of the hematite (asp•m2 g-1) and the concen-tration of hematite in the suspension (Cs,g 1-1 ), the concentration of surface area in the suspension (Ca,m2 1-1) was calculated.
30
(m2 1-1) (3.4)
Using this surface area concentration the adsorption density of ascor-bate was calculated.
{>FeIIIHA} (mol m-2) (3.5)
Here the adsorption density of ascorbate is written as the surface con-centration of a surface iron(Ill)-hydrogen ascorbate complex. For the [Fe+2Jsol and [H2A1sol concentrations measured at each sample time and shown in figure 3.1, the corresponding surface concentrations of ascorbate were calculated using equations (3.3) and (3.5). The results of these calculations are listed in table 3.1.
Table 3.1. Change in [H2A1soh [Fe+2]soh and [H2A1sur with Time
Time (hours)
0 1.00 2.58 5.00 9.00 12.00
[Fe+2Jsol (1 o-6 mol 1- l)
0 7 .1 10.0 11.9 17 .1 22.1
[H2A1sol (I0-6 moI l-1)
50 (added) 36.5 34.8 33.0 30.7 29.3
1 [H2A1sur = [H2A]T - [H2A1sol 2fFe+2Jsol
{>FeIIIHA] = [H2~1sur
Hematite= 0.613 g 1-l, Ca=l0.73 m21-I
{>FeIIIHA} (Io-7 mol m-2)
0 9.3 9.5 10.3 10.0 9.0 (moI I-1)
(mol m-2)
31
Although the concentrations of solution species were changing with time the surface concentration of ascorbate remained virtually con-stant during the experiment, except for the increase during the initial rapid adsorption in the first hour.
The initial adsorption occuring in the first few minutes of the ex-periment was also calculated. The linear portions of the kinetic curves were extrapolated back to time=O. The initial concentration of ascor-bate remaining in solution just after the initial adsorption ([H2AJ0 in figure 3.1) and the concentration of iron(II) due to the initial rapid release ([Fe+2]o in figure 3.1) were estimated as the intercepts of the respective kinetic curves on the vertical axis. For the data plotted in figure 3.1, [H2AJ0=3.66x10-5 mol 1-l, [Fe+2Jo=5.91xI0-6 mo11-l. The total concentration of added ascorbate, [H2A]T, was 5.0xI0-5 mol 1-l. Equation (3.3) with [H2A1so1=[H2AJ0 and [Fe+2Jsol=[Fe+2Jo gives [H2A1sur=l.04xl0-5 mo11-l. Equation (3.4) (Ca=l0.73 m2 1-l) gives the surface concentration of ascorbate, {>Fe1IIHA}=9.73x10-7 mol m-2. This value is similar to the surface concentrations for ascorbate listed in Table 3.1. The initial concentration of adsorbed ascorbate calculated in this manner was determined from the experimental data for each reductive dissolution experiment carried out.
3.3.3 The Adsorption Kinetics of Ascorbate on Hematite Calculating the initial concentration of adsorbed ascorbate as pre-
sented in the previous section contains two assumptions which must be tested. Extrapolating the disappearence of ascorbate to time=O as-sumes that adsorption is rapid compared to the rate of the overall dis-solution reaction. The second assumption, used in formulating equa-tion (3.3), is that the concentration of iron(II) on the hematite surface is negligible compared to the other terms in the mass balance for ascorbate. The adsorption kinetics of ascorbate on hematite are shown in figure 3.2. The initial rapid disappearence of ascorbate was finished within 10-15 minutes. This demonstrates that extrapolating the plots of [H2AJsol and [Fe+2Jsol vs. time to time=O, for an experiment lasting several hours, is a good approximation of the initial concentrations where the disappearence of ascorbate from solution would be domi-nated by adsorption.
w ~ •....J m ....J a: 0 0 (0 ::E 0 ~
0 ..... -0 -w -~ z
~ ~ 0 i5 z <(
32
120
100 ASCORBATE
80
60
40
20 IRON(ll)
0 0 10 20 30 40 50
TIME (minutes)
Figure 3.2. Adsorption kinetics of ascorbate on hematite. pH=3, fH2AJr=l.OxJ0-4 mol /-1, hematite=0.613 g 1-l.
3.3.4 The Comparative Adsorption of Ascorbate on Hematite and Alu-minum Oxide
The assumption that there is relatively little iron(II) on the hematite surface was tested by comparing the adsorption of ascorbate on hematite with adsorption on a non-reducible metal oxide, o-Al203. The surface acid/base properties of the two oxides are virtually identi-
cal. Both have a pHzpc near pH 9, and both have a pK~ 1 (intr)=7. The
non-reducible oxide cannot contribute to the disappearence of ascor-bate through oxidation of the adsorbing ascorbate. All ascorbate which disappears from solution is present as a surface aluminum-ascorbate complex. For the adsorption of ascorbate on hematite, however, a por-tion of the ascorbate which disappears from solution would be oxidized by iron(III) surface sites, thereby producing iron(II). A part of the oxidized ascorbate can be accounted for by measuring the amount of
33
iron(II) in solution. The amount of iron(II) produced which remains on the surface is unknown. For the disappearence of ascorbate from solution in the presence of hematite, the mass balance for ascorbate (equation 3.2) can be written as follows.
1 [H2AJsur + 2 [Fe(Il)]sur (mol 1-1)
(3.6)
The unknown concentrations are written together on the left side of the equation and their sum can be calculated by the known values on the right side. For the non-reducible aluminum oxide this mass bal-ance is simplified.
(mol 1-1) (3.7)
If there is a significantly large amount of iron(Il) on the hematite surface (compared to the other terms in the mass balance) then the amount of ascorbate which disappears from solution in the presence of hematite particles, as calculated from equation (3 .6), should be larger than the amount disappearing in the presence of the non-reducible aluminum oxide particles, as calculated by equation (3.7). The two suspensions had equal concentrations of surface sites (Csites= 10-4 mot 1-1 ). The measured concentrations (from samples taken one hour after the addition of ascorbate) of [H2A]s0 i, [Fe+2]sol• the total added con-centration of ascorbate [H2A] T, and the calculated quantities [H2Alsur+[Fe(II)Jsur (in the presence of hematite) and [H2Alsur (in the presence of aluminum oxide) are listed in Table 1. In the presence of the reducible iron(III) oxide the sum of adsorbed ascorbate and sur-face iron(II) concentration is not significantly more, and in fact at 5x 10-5 mol 1- l and 10-4mol 1- l added ascorbate slightly less, than the amount of ascorbate adsorbed on the aluminum oxide. Evidently there is not a large amount of iron(II) produced on the hematite surface and the concentration of ascorbate adsorbed, as calculated by equation (3.3), is a reasonable approximation. The concentration of ascorbate remaining in solution after one hour is about the same for the two ox-ides.
34
Apparently oxidation of ascorbate does not contribute greatly to the disappearence of ascorbate from solution during the first hour of re-ductive dissolution.
Table 3.2. Mass Balance for Ascorbate in the Presence of Alz03 and Fe203 Particles
a-Fe203 (Csites=l.09xl o-4 mol i-1)
[HzA]T
(added)
1 [HzAlsol [Fe+2Jsol [H2A1sur+2fFe(Il)1sur
(measured) (measured) (calculated)
10.0 3.7 7 .9 2.4 50.0 39.4 6.9 7 .2 100.0 89.5 3.8 8.6
l [H2Alsur + [Fe+2Jsur = [HzA]T - [H2Alsol -2 [Fe+2Jsol
o-A12o3 (Csites=l.05x1Q-4 mol l-1)
[H2AlT [H2Alsol [HzAlsur (added) (measured)
10.0 50.0 100.0
all concentrations as 10-6 mol 1- l Hematite=0.613 g 1- l o-Al203=0.5 g 1-1
8.2 40.0 85.9
(calculated
1.8 10.0 14.1
35
3.3.5 Estimation of the Surface Stability Constant for the lron<Ill)-Ascorbate Complex
In addition to comparing the mass balance for ascorbate, isotherms for the adsorption of ascorbate on hematite and ~-Al203 were also compared. These are shown in figure 3.3. The surface sta-bility constants (SKsur) estimated from the adsorption data are nearly the same. Adsorption of anions on metal oxide surfaces can be ex-plained by formation of surface complexes. Formation of surface com-plexes is described by thermodynamic mass laws analagous to those written for solution complexes. In addition, the mass laws for surface complex formation often exhibit stability constants similar to those for the analagous reaction in solution (Kummert and Stumm,1980). The adsorption of ascorbate on hematite may be described by the following surface reaction. At pH values far below the first acidity constant for
the hematite surface (pK~ 1 (intr)=7) the predominant surface species
would be >FellloH2+. I have written the reacting ascorbate species as being the hydrogen ascorbate ion.
(3.8)
This equation describes the formation of the surface iron(Ill)-hydro-gen ascorbate complex as a ligand exchange reaction with hydrogen ascorbate ion replacing water within the inner coordination shell of the surface iron center. An analgous reaction can be written for formation of the surface aluminum-hydrogen ascorbate complex. In figure 3.3 the concentration of the surface iron(III)-hydrogen ascorbate and aluminum-hydrogen ascorbate complex is taken to be equal to the concentration of ascorbate adsorbed. These concentrations are plotted against the corresponding analytically determined concentration of ascorbate remaining in solution. This concentration is the sum of all solution ascorbate species (CT=[H2Al+[HA-]+[A-2], mol l-1). The for-mation reaction (equation 3.8) has been written with HA- as the re-acting ascorbate species. The concentration of [HA-] must be calculated for the conditions at which the experiment was carried out
-{\j
E 0 E
Q)
'o ..-........ <( :c <( /\ -
-{\j
'E 0 E .....
·o ..--<( :c
=(I) LL. /\ -
36
40 (a)
30
m 20 II
10
100 200 300 400 500 -6 -1
[HA] 1(10 moll ) 2 so
(b)
20
m
10 m II
0'--_...~....._~.....___,.___._~_._~....___.~__.___,
0 50 1 00 150 200 250 -6 -1
[H2
A] (10 mol I ) sol
Figure 3.3. Adsorption isotherms for ascorbate on hematite and S-Al203. pH=3.0, T=250C, (a) a-Fe203, Cs=0.613 g /-1. (b) S-A/203, Cs=2.0 g 1-l. Since all experiments were carried out at the same pH, the abcissa values can also be given as {HkJ=a1fH2Alsol• a1=0.078
37
(pH 3.0, ionic strength=0.01 mol 1- l) in order to estimate the value of the surface stability constant, SKsur· The thermodynamic mass law for formation of these complexes can be combined with the mass balance for surface sites to give an isotherm relating the concentration of the surface complex to the solution concentration of the reacting species (equation 3.9). This isotherm is identical to the Langmuir adsorption isotherm (see Appendix 1 for a derivation of the Langmuir isotherm for surface complex formation).
SK C:! [HA-] {>FeIIIHA) ,;, survm
l +SKsur[HA-J (mol m-2) (3.9)
Sm is the adsorption maximum for ascorbate (mol m-2). This equation can be linearized by taking the reciprocal of both sides.
0-=:;;'--~~-L-~~~-'-~~~--'~~~--'
0 2 3 4 -1 6 -1
[HA. j (10 I mol )
Figure 3.4. Double reciprocal plot for estimating the surface stability constants for formation of the hydrogen ascorbate complex on hematite and 6-Al20 3.
38
Table 3.3 Fit Parameters and Estimated Values for the Adsorption Maximum and Surface Stability Constants for the Adsorption of Ascor-
bate on Hematite and o-Al203.
parameter a-Fe203 o-Al203
a 1.03 8.306 m2 1- l b 9.867x105 6.241x106 m2 mol-1 r2 0.977 0.979 Sm 1.0lxI0-6 1.60x10-7 mol m-2 SKsur 9.58xl05 7.5lxl05 1 mo1-l
>FeIIIOHz+ +HA-<± >FeIIIHA + HzO SKsur = 9.6x105 1 mo1-l >AlOH2+ +HA-<± >AlHA + HzO SKsur 7.5x105 l mo1-l Hematite = 0.613 g 1-l, o-Al203 = 2.0 g 1-1. ascorbate acidity constants: logK1 =4.03, logK2=11.3 (250C, I=0.1 mol 1-l; Smith and Martell, 1976).
1 (m2 mo1-l) {>FeIIIHA}
(3.10)
1 a= andb 8KsurSm
1 S";; . The data from the isotherms can be plotted
as {>FeIIIHA}-1 and {>AIHA}-1 vs. [HA·]-1 as shown in figure 3.4. The adsorption maximum for ascorbate and the surface stability constants can be estimated from the constants obtained by a linear regression of
1 the double-recipricol plots. For the linear equation (3.10), Sm='b and
b SKsur=-a The regression constants and the estimated values for the
adsorption maximums and the stability constants are listed in Table 3.3. The surface stability constants differ by only a factor of 1.3 with the iron(III) complex being slightly more stable.
3.3.6 The Rate Di:;pendence on Ascorbate Concentration A series of experiments were carried out at pH 3 and pH 4 at various added concentrations of ascorbic acid. The linear portion of the disso-
39
lution curves for the series of dissolution experiments at pH 3 are plotted in figure 3.5. For all dissolution experiments reported here, the vertical axis is the change in iron(II) concentration after the first sample was taken, approximately 1 hour after the addition of ascor-bate.
0 E
"' ·o ,..--0 ......... "'
N .. (I)
LL .........
20
10
TIME (hr}
Figure 3.5. Dissolution at various concentrations of added ascorbate, pH=3, hematite=0.613 g l-1. [Fe+2Jsol is the change in iron(//) concentration in solution after the first sample taken 1-2 hours after the addition of ascorbate.
Due to the initial rapid release of iron(II) during the first hour there was usually l-5xl0-6 mol 1-1 iron(II) present in the first sam-ples taken. The higher concentrations in these samples corresponded to the experiments with higher added ascorbate concentrations. As in figure 3.1 the rates of dissolution were linear with time and were cal-culated as the slope of the linear plot of LFe+2Jsol vs. time. The slowest reaction was the acid dissolution of hematite in the absence of ascor-bate. Even relatively small additions of ascorbate accelerated the dis-solution markedly. The rate of dissolution increased with increasing added ascorbate concentration and demonstrated a strong dependence
40
on the amount of ascorbate added to the system. The rates of dissolution for the experiments at pH 3 are plotted
against the initial residual concentration of ascorbate in solution, [HzA]o (mo11-l) (see section 3.3.1) in figure 3.6. If the dissolution is
~
' .... .r:
(\j
'E 0 E ....
' 0 .,.... -w I-
~
3
" 2
0 0 100 200 300 400 500
-6 -1 [H
2A]
0 (10 mol I )
Figure 3.6. Dependence of dissolution rate on initial concentration of ascorbate in solution, pH=3, hematite=0.613 g l·l. fH2AJo is the total dissolved concentration of all ascorbate species ({H2AJ ,[HA·],{ A-2 ]).
proportional to the concentration of the surface iron(III)-hydrogen ascorbate complex then the rate dependence should follow the surface concentration of ascorbate as calculated by the isotherm equation used to estimate the surface stability constant for the ascorbate complex (equation 3.9).
(SKsurSm[HA ·]) RATE = k' __,:=;_....;;,; __
1 +SKsur[HA ·] (molm·2hf"l) (3.11)
k' (hr-1) is an empirical first-order rate constant. [H2Alo (mo11-l) is the experimentally determined sum of all dissolved as.corbate species
41
(CT=[H2Al+[HA-J+[A-2], mol 1-l ). Using the distribution coefficient for
h h d b . [HA-] . 3 11 b ' ' t e y rogen ascor ate ion a 1 = [OrJ , equation . can e wntten m
terms of [H2Alo=CT.
(mol m-2 hrl) (3.12)
This equation, like the isotherm (3.9), can be linearized by taking the reciprocal of both sides, RATE-l=a[H2AJ-l+b. a=(k'SKsurSmtq)·l (m2 hr 1-l) and b=(k'Sm)-1 (m2 hr mot-1 ). The dissolution data in figure
- I 3.6 were plotted as RATE-1 vs. [H2Alo and constants a and b were
obtained from a linear regression of the double-reciprocal plot. These constants were then used to plot an empirical rate law based on equa-tion (3.12) relating rate of dissolution to total dissolved concentration of ascorbate at pH 3.
[H2Alo RA TE == -a+,.,.b..,.,[H.,..2_,A,..,J-0 (mo! m-2 hr-1) (3.13)
This rate Jaw is plotted as the solid line in figure 3.6. The dependence of the dissolution rate on the surface concentra-
tion of ascorbate can be directly obtained by plotting the observed rate of reductive dissolution against the calculated concentration of ad-sorbed ascorbate as shown in figure 3.7. The data for the dissolution experiments at pH 3 and 4 are plotted along with the results for two additional dissolution experiments, one at pH 2.5 and one at pH 6. The dissolution at pH 6 was extremely slow with only about one mi-cromolar iron(II) present in solution after several days. It was not attempted to calculate a rate of dissolution quantitatively. The plot of iron(II) vs. time was used to estimate an upper bound on the rate of dissolution.
The rate of dissolution is proportional to the surface concentration of adsorbed ascorbate. An empirical first-order rate law can be writ-ten to include explicitly the concentration of the proposed reacting species, {>FelIIHA}, and the empirical rate constant for the reductive
42
dissolution (ke) can be estimated from the slope of the RATE vs. {>FeIIIHA} plots.
RATE = ke{>FelIIHA} (mot m-2 hr-1) (3.14)
The empirical rate constants are listed in figure 3.7.
-. .c N
'E 0 E
..... 'o .,.... -w !;( a:
3
2
0 0
pH-2.5 _1 k 9 •0.,45 h
I I I I • I
I I
I I
I I I
I I
I I
I
10 20
pH=4 _1 k -0.045 h e
pH .. 6 -1
k <0.001h e
Ill -7 -2 {>Fe HA} (10 mol m )
30
Figure 3 .7. Dependence of rate of dissolution on surface concentration of ascorbate, hematite=0.613 g 1-J.
3.3.7 The Rate De.pendence on pH The dissolution rates plotted in figure 3.7 are strongly dependent
on pH. The reaction order for proton concentration can be obtained by plotting logke against log[H+] (figure 3.8). The reaction order was es-timated as the slope of a linear regression through the three points. Although it is desirable to have more data over a wider range of pH values, it is possible to conclude that the dissolution reaction is acid catalyzed and fractional order in proton concentration; keCX[H+Jn. The slope of the plot is approximately n=0.6. Fractional order kinetics are typical for surface chemical reactions (Grauer and Stumm,1982;Blum
43
and Lasaga,1988) and indicate here that protonation of a surface species plays an important role in the rate-determining step in the overall dissolution reaction.
·0.5 ~
' .... :S
Q) ·1.0 x. O>
..Q
·1.5
+ log[H J
Figure 3.8. Dependence of observed first-order rate constants on solution proton concentration.
3.3.8 The Effect of pH on Ascorbate Ads01:ption Dissolution experiments in the presence of 10-4 mol t-1 added
ascorbate were carried out at pH 2.5, 3, and 4. The ascorbate adsorp-tion calculated from the kinetic data for these experiments was plotted against pH to obtain some idea of how the adsorption of ascorbate was changing over the pH range where the empirical rate constants were measured (figure 3.9). The adsorption edge for ascorbate appeared to be between pH 2 and 3 (approximated by the dotted line in figure 3.9) with a sharp increase in surface concentration of ascorbate seen be-tween pH 2.5 and 3. The adsorption was independent of proton con-centration between pH 3 and 4. The adsorption of ascorbate at pH 6 in the presence of 2.5x 10-4 mot 1- l ascorbate was 24.2x 10-7 mot m· 2 (taken from the data for the reductive dissolution experiment at pH 6), showing that ascorbate was still strongly absorbed at more neutral pH values. This is indicated by the horizontal dashed line in figure 3.9.
(\I
'E 0 E .....
'o ..... -< ::c ........
~ /\ -
44
15
10
5 . :
0 .. --2 3 4 5
pH Figure 3.9. The Effect of pH on Ascorbate Adsorption in the presence of 10-4 mol z-1 added ascorbate, T=250C. The dotted line indicates a possible adsorption edge for ascorbate below pH 2.5. The horizontal dashed line refers to the continued strong adsorption at higher pH values based on the adsorption of ascorbate at pH 6 in the presence of 2.5xJ0-4 mol z-1 added ascorbate (24.2xJ0-7 mol m-2 ).
45
3.4 DISCUSSION 3.4.1 The Stoichiometric Reaction for the Reductive Dissolution of Hematite by Ascorbate
The overall dissolution reaction may be written as follows.
(3.15)
Ascorbate, a di-protic acid (pKt =4, pKz=ll) would attack surface iron(III) sites reducing them to the more easily released iron(II) oxi-dation state. The first stable oxidation product of ascorbate is dehy-droascorbic acid (DHA).
11 0
ASCORBIC ACID
11 0
ASCORBATE RADICAL
Ff-I
II 0
H -o .. " .. ,
/ -0
Fe(lll)(OH 2 )4
IRON(lll)-HYDROGEN ASCORBATE
II 0
DEHYDROASCORBIC ACID
46
Ascorbate goes to DHA in a two-step electron transfer where ascorbate first gives up one electron forming the ascorbate radical which then reacts with a second ascorbate radical forming ascorbate and DHA through a rapid disproportionation reaction. The disproportionation reaction has an observed second-order rate constant which decreases from 108 M-1 s-1 at pH 4 to HP M-1 s-1 at pH 11 (Bielski, 1982). The loss of two electrons as ascorbate is oxidized to DHA is reflected in the 2: 1 stoichiometry seen in figure 3.1; the rate of dissolution is twice that of the rate of disappearence of ascorbate. lron(Ill) is the oxidant in this case with two iron(II) ions being produced for each molecule of ascorbic acid being oxidized. The structural formulas for the solution species of ascorbate, iron(Ill)-hydrogen ascorbate complex, DHA (Khan and Martell, 1976), and ascorbate radical (Bielski, 1982) are shown above.
3.4.2 Experimental Information for Formulation of a Mechanism of Reaction
The overall dissolution reaction may be written as three elemen-tary reaction steps; adsorption of ascorbate and formation of a surface iron(IIl)-hydrogen ascorbate complex, electron transfer within this surface complex producing iron(II) and ascorbate radical, and release of iron(II) from the crystal surface into solution.
kl >FeIIIOH2+ +HA- +z >FeIIIHA + H20
k_ 1
k2 >FeIIIHA + H20 <= >FeIIOH2+ + HA'
k_2
new surface site + Fe+2(aq)
(3.16)
(3.17)
(3.18)
47
The experimental measurements available for testing this mechanism are pH, dissolved iron(II) concentration vs. time, dissolved ascorbate concentration vs. time, and adsorbed ascorbate concentration calcu-lated for each experiment from the mass balance for ascorbate (equation 3.3). The following list of experimental observations, sum-marized from the previous section, must be consistent with the mech-anism.
1. The rate of dissolution is first-order in surface ascorbate concen-tration. RATE=ke{>FelIIHA}. 2. The empirical rate constants exhibit a fractional order dependence on proton concentration, approximately [H+]0.6. 3. The adsorption of ascorbate increases between pH 2.5 and pH 4 and continues to be strongly adsorbed at pH 6. 4. The amount of ascorbate oxidized appears to be estimated well from the concentration of iron(II) in solution suggesting that there is only a relatively small amount of iron(II) on the hematite surface. 5. The stability constants for formation of surface hydrogen ascorbate complexes, estimated from adsorption isotherms, are similar on hematite and aluminum oxide. The adsorption of ascorbate on hematite can be treated, as a first approximation, as adsorption on a non-reducible oxide surface. 6. The concentration of ascorbate on the hematite surface does not change greatly after the first hour of the dissolution reaction. This, along with the linear dissolution curves observed, indicates that the surface is at a steady-state with respect to the reacting species. 7. The reaction stoichiometry is 2: 1 between iron(II) and ascorbate, as expected from the two-electron loss for the oxidation of ascorbate to dehydroascorbic acid.
3.4.3 A Rate Expression for the Reductive Dissolution of Hematite by Ascorbate
A rate expression is a mathematical relation, corresponding to the elementary steps in a reaction mechanism, between the rate of a reac-tion and the concentration of the reactants. Here the rate expression begins with a kinetic assumption. As an initial hypothesis it is pro-posed that detachment of iron(II) from the hematite surface (equation
48
3.18) is the rate-limiting step in the overall reaction, as opposed to electron-transfer or adsorption. The rate expression for (3.18) is written as follows.
d[Fe+21 - k { F IIoH } dt - 3 > e 2 (mol m-2 hr-1) (3.19)
It is not possible to test this rate expression directly because the con-centration of iron(II) on the hematite surface was not experimentally determined. However, it is possible to derive, with help of the pro-posed mechanism, an expression relating the concentration of iron(II) on the hematite surface to the surface concentration of ascorbate. The rate expression for the production of iron(II) on the hematite surface can be written.
d{>FelIOH2} d t = k1{>FeIIIHA} - k.2{>FelIOH2}[HA·] - k3 {>FelIOH2} = 0
(mol m·2 hr-1) (3.20)
The resulting differential equation can be solved as a simple algebra problem if the reacting species are assumed to be at a steady-state concentration. This is a reasonable assumption for the dissolution re-action ocurring after the first hour of the reaction. The steady-state condition is indicated in (3.20) by setting the time differential equal to zero. The resulting simplified equation is written as follows.
(molm-2hr-1) (3.21)
This equation can be solved for {>FeIIOH2}.
(mol m·2) (3.22)
Equation (3.22) can be substituted into equation (3.19) to obtain a rate expression written in terms of the surface ascorbate complex.
d[Fe+2] dt
This rate expression can RA TE=ke { >FelllHA}. In this
((k_2~:~+k3)) is equal to ke·
49
(mol m-2 hr-1) (3.23)
explain the empirical rate law, case the pseudo first-order rate constant
Although (3.23) can be written to ap-
pear as a first-order rate expression the pseudo first-order constant has a concentration dependence on [HA·] which is unknown for the ex-perimental system studied. Equation (3.23) demonstrates that it is possible to derive a pseudo first-order expression for the reductive dissolution of hematite by ascorbate from the proposed mechanism.
3.4.4 The Rate Dwendence on Surface Proton Concentration The fractional order dependence of the empirical rate constants on
proton concentration indicates that protonation of a surface species is important in the rate-determining step of the overall dissolution reac-tion. The role of surface protonation in the release of the metal leav-ing group from the crystal lattice was discussed in Chapter 2. If de-tachment of iron(ll) is the rate-determining step in the dissolution re-action then protonation of the surrounding lattice oxygen ions should lower the activation energy for the release of the iron(ll) into solution (Wieland et. al.,1988). It is possible to estimate the reaction order for the rate of dissolution on surface proton concentration {H+} (mol m-2). Wieland et. al. (1988) have developed a relation to convert empirical rate laws for the acid dissolution of oxide minerals which are fractional order in proton concentration into rate laws expressed explicitly in surface proton concentration. They have shown that a general Freund-lich isotherm can describe the adsorption of protons on a variety of oxide mineral surfaces. In this approach the logarithm of the surface concentration of protons, log{>MOH2+}, was shown to depend linearly on the difference in pH between the solution pH and the pHzpc; L1pH=pHzpc-pH. pHzpc refers to the pH at which an oxide surface has no net charge in the absence of charge determining species other than protonated surface metal ions (>MOH2+) (Stumm and Morgan,1981).
50
(mol m-2) (3.24)
KF (mol m-2) is the Freundlich adsorption constant, [H+1zpc is the pro-ton concentration corresponding to the pHzpc• [H+] is the solution pro-ton concentration. This general isotherm for protons was found to have logKF=-6.51 and m=0.21. The isotherm predicted the surface protonation for seven different oxide minerals; beryllium, aluminum, and zirconium oxide; goethite, hematite, magnetite, anatase, and a latex surface, within a factor of 2 for values of ApH>l. Equation (3.24) can be solved for the solution proton concentration.
(moll-1) (3.25)
The fractional-order dependence of the empirical rate constants on proton concentration, kecx.[H+]n, can be converted into a dependence on surface proton concentration by substituting (3.25) into the propor-tionality.
(3.26)
A fractional-order dependence of n=0.6, and a value for the exponent in the Freundlich equation of m=0.21 as given by Wieland et. al. (1988), gives a third-order dependence for the rate of dissolution on surface proton concentration, k a. {>MOH2+}3. Zinder (1985) reported a third-order dependence on surface proton concentration for the reductive dissolution of goethite by ascorbic acid~ With a large excess of ascorbate in solution, the following logarithmic form of an empirical rate law was obtained ([H2AJT=l0-3 mo11-l, goethite=3 g 1-l ).
logRATE = logko +3log{>MOH2+) (mol m-2 hr-1) (3.27)
ko corresponded to a rate constant for reductive dissolution at unit surface activity of protons ({>MOH2+J=l mol m-2) extrapolated from the empirical expression. ko=l.23x109 m4 mot-2 hr-1 (Zinder,1985).
5 1
If the pH dependence of the empirical rate constants obtained here for the reductive dissolution of hematite are written as ke=keo[H+Jn (keo=12.5 hr-1 ), then the following pH dependence for the rate of reductive dissolution can be written.
(mol m-2 hr-1) (3.28)
keo is the empirical rate constant at unit proton activity. Equation (3.26) can be substituted into this expression to obtain a dependence of the rate of dissolution on surface proton concentration.
(3.29)
The terms in front of ( >FeIIIo H 2 +} on the right side can be grouped to-gether to calculate a value for ko for the reductive dissolution of hematite by ascorbate which can then be compared with the value of ko obtained by Zinder.
(m4 mo1-2 hr-1) (3.30)
The adsorption maximum for ascorbate on hematite, Sm=l .Oxto-6 mol m -2 is taken for the surface concentration of ascorbate with dissolved ascorbate in large excess. keo=l2.5 hr-I, [H+lzpc=lo-9 mol 1-1, K p= 1 o-6 .5, m=0.2, and n=0.6. This gives a value of ko= 1. 6x10 9 m4 mo 1-2 hr-1. This is about the same as the results of Zinder for the re-ductive dissolution of goethite by ascorbate (ko= 1. 23x1 o9 m4 mol- 2 hr-1 ). The empirical rate law for the "proton-assisted" reductive dis-solution of hematite by ascorbate may be written as follows.
(mol m-2 hr-1) (3.31)
52
For the dissolution of hematite studied here, the empirical rate con-stant would be k'e= l.6x 1015 m6 mo1-3 hr-1. The acid dissolution of hematite was almost two orders-of-magnitude slower than reductive dissolution in the presence of ascorbate at pH 3 (figure 3.5), and can be neglected as a parallel reaction.
It is believed that such integer values (n=2,3,4) for the reaction order in surface proton concentration correspond to a precursor com-plex, for detachment of the surface metal group as the rate-determin-ing step in the overall reaction, having a coordination environment of n-protonated oxygens ions surrounding the metal center in the leaving group (Furrer and Stumm,l986;Wieland et. al.,1988;Schott,1989). Other experimental evidence exists for integer values of the reaction order in surface proton concentration, and thus such precursor com-plexes, for the detachment reaction in the acid dissolution of oxide minerals (Furrer and Stumm,1986; Carroll-Webb and Walther,1988; Schott, 1989).
"" / Fe(lll)OH
OH+
~ /Fe(ll)OH2
OH+
~ I
Fe(lll)OH
Figure 3.10. Example of a triply-proton.ated precursor complex for the release of surface iron.(//) from the hematite crystal lattice.
The third-order dependence on surface proton concentration for the reductive dissolution of hematite by ascorbate possibly corresponds to a precursor configuration for the iron(II) leaving group as shown in figure 3.10.
56
~- {>FeIIIHA} k_ 1 - {>FeIIIOH2+}[HA-]
(1 mo1- l) (3.38)
As a first approximation for measuring the adsorption of ascor-bate on hematite, the electron-transfer kinetics can be neglected if they are sufficiently slow compared to adsorption/desorption. There was virtually no difference in the amount of ascorbate adsorbed on both hematite and aluminum oxide and the surface stability constants for the surface ascorbate complexes were similar. This demonstrates that there was not a significant error introduced into the calculation of adsorbed ascorbate on hematite by neglecting the electron-transfer kinetics. Thus the kinetics of adsorption/desorption are believed fast compared to electron-transfer and, as a first approximation, the ad-sorption of ascorbate on hematite can be treated as adsorption on a non-reducible metal oxide surface. The three step dissolution se-quence would procede by rapid adsorption followed by slower elec-tron-transfer and the rate-determining release of iron(II) sites from the hematite surface.
0 E
~ 0 ~ ---c: 0 ,_
.......::. -c 0 ,_
65
30 pH3
µM ascorbate µM oxalate
20 lJ 0.0 0.0 b. 100 0.0
• 0.0 50
• 100 50
10
TIME (hr)
Figure 4.3. Concentration of dissolved iron versus time in the presence or absence of ascorbate and oxalate. pH=3, hematite=0.613 g z-1.
Table 4.3 Rates of Dissolution and Surface Concentrations of Ascorbate and Oxalate at pH 3 in the Presence or Absence of Ascorbate and
Oxalate
Conditions RATE {>FeIIIHA}
HN03 oxalate
ascorbate
(lo-7 mol m·2 hr-1)(10-7 mol m·2)
0.017 0.21 1.48
0.0 0.0
12.6*
{FeIIIOx} (l0-7 mol m·2)
0.0 23.0**
0.0 ascorbate + 5. 97 4.1 * 19.6**
oxalate * [H2AJT = l.Ox to-4 mol 1- l [H20x]T = 5.0xI0-5 moI J-1, hematite= 0.613 g 1-l, **the surface concentration of oxalate in the absence of ascorbate was taken from the adsorption isotherm for oxalate on hematite.
51
If the pH dependence of the empirical rate constants obtained here for the reductive dissolution of hematite are written as ke=keo[H+]n (keo=l2.5 hr-1 ), then the following pH dependence for the rate of reductive dissolution can be written.
RATE ::: keo[H+]n{>FeIIIHA} (mol m-2 hr-1) (3 .28)
keo is the empirical rate constant at unit proton activity. Equation (3.26) can be substituted into this expression to obtain a dependence of the rate of dissolution on surface proton concentration.
(3.29)
The terms in front of { >Felllo H 2 +} on the right side can be grouped to-gether to calculate a value for ko for the reductive dissolution of hematite by ascorbate which can then be compared with the value of ko obtained by Zinder.
(m4 mo1-2 hr-1) (3.30)
The adsorption maximum for ascorbate on hematite, Sm= I .Ox 1 o-6 mol m -2 is taken for the surface concentration of ascorbate with dissolved ascorbate in large excess. keo=I2.5 hr-1, [H+lzpc=Io-9 mol 1-1, Kp=Io-6.5, m=0.2, and n=0.6. This gives a value of ko=l.6xI09 m4 mo 1-2 hr-1. This is about the same as the results of Zinder for the re-ductive dissolution of goethite by ascorbate (ko= l.23x 109 m4 mol-2 hr-1 ). The empirical rate law for the "proton-assisted" reductive dis-solution of hematite by ascorbate may be written as follows.
(mol m-2 hr-1) (3.31)
53
For the dissolution at pH 6, ascorbate is present at high surface concentration ((>FeillHA}=2.4xI0-6 mol m-2). The practically non-existent dissolution at pH 6, even with a high surface concentration of ascorbate, shows that ascorbate is not acting as a surface chelate ligand to accelerate the dissolution of hematite. Coordination of the hydrogen ascorbate ion (pK2=1l) with surface iron(III) should help deprotonate the coordinated hydrogen ascorbate, due to the high positive charge on the central iron(III) ion. This effect is evidently not strong enough in the case of the surface iron(IIl)-hydrogen ascorbate complex to form a chelate complex at pH 6. In addition, the rate of dissolution in the presence of 10-4 mol 1-1 ascorbate was fastest at pH 2.5, in spite of a much lower surface concentration of ascorbate than at pH 3 or 4. This is explained by protons assisting the release of iron(II) sites from the surface but cannot be explained by ascorbate acting as a chelate ligand to enhance the rate of release of iron(II) sites into solution.
No significant dissolution was seen at pH 6. Two effects may play a role. Low proton concentration on the hematite surface would lead to slow dissolution kinetics. In addition, it has been proposed by LaKind and Stone (1989) that adsorption of iron(II) may inhibit the reductive dissolution. Published studies of iron(II) adsorption on metal oxide minerals (hematite, magnetite, anatase) all show a strong adsorption edge beginning at pH 5 (Blesa et. al., 1984; Mulvaney et. al.,1988; Wehrli, 1989). As mentioned in Chapter 2., adsorbed metal ions such as aluminum may block active sites on the mineral surface, thus inhibiting the dissolution. This effect here would be analogous to passive film formation in corrosive dissolution.
3.4.5 The Electron Transfer Reaction Although electron-transfer is probably not the rate-determining
step for the dissolution reaction reported here, it is possible to define some boundaries for the possible redox reaction occurring on the hematite surface. From the results for the comparative adsorption of ascorbate on hematite and aluminum oxide is appears that the con-centration of surface iron(II) is much lower than the concentration of the surface iron(III)-hydrogen ascorbate complex. Equation (3 .23) was derived from the proposed mechanism in order to obtain an ex-pression for the concentration of iron(II) in ·terms of the iron(IIl)-hy-
54
drogen ascorbate complex, {>FeIIOH2} k2 {>FeIIIHA}. If (k.2[HA· ]+k3)
{>FeIIOH2}<<{>FeIIIHA} then the coefficient on the right side must be small.
(dimensionless) (3.32)
This inequality can be rewritten.
(hr 1) (3.33)
kz (hr- I) is the forward rate constant for electron transfer, k.2 (hr-1 l mo1- l) is the rate constant for the back-reaction in the electron-transfer step, [HA·] (mol 1-1) is the concentration of ascorbate radical, and k3 (hr-1) is the rate constant for the release of iron(II) from the surface. Release of iron(II) from the hematite surface is most-likely the rate-determining step in the overall dissolution reaction. If re-lease of iron(II) is much slower than the electron-transfer reactions (k3<<k.2[HA·]), then k2<<k.2[HA·]. There may be a significant back-re-action for electron-transfer.
If electron-transfer is reversible it should be possible to define a redox potential for the reaction. The reduction of a surface site may be written as the following half-reaction.
Eh =: 0.36 V, logKe 6.1 (3.34)
The exact redox potential for this reaction is unknown. It was as-sumed that the Eb can be approximated by the redox potential for the oxygenation reaction of iron(II) adsorbed at the iron(Ill) oxide sur-face. This redox potential for the oxidation of ferrous iron adsorbed on goethite has been estimated by Wehrli (1989) from oxygenation ki-netic data using a Linear Free Energy Relation (LFER) for the oxygena-tion kinetics of reduced metal ions in solution.
The redox reaction for the oxidation of ascorbate radical to ascor-bate has also been described. The redox potential for the system was given as 0.3V at pH 7. Here the radical is written as the protonated
55
species.
HA·+ e- ~ HA- pH7 Eh = 0.3 V, logKe = 5.1 (3.35)
The redox potential of the ascorbate system depends upon the solution speciation and changes with pH. Ascorbate becomes less reducing at lower pH values (Steenken and Neta, 1979). At pH 7 the redox poten-tial of ascorbate is about the same as for the surface iron(IIl)/iron(II) couple as calculated by Wehrli.
The relative rates of the adsorption step and electron-transfer can also be compared using the proposed mechanism. The similar stability constants for the formation of the surface ascorbate complexes on hematite and aluminum oxide indicate that adsorption/desorption ki-netics are fast compared to electron-transfer. The rate expression for the surface concentration of iron(III)-hydrogen ascorbate, based on the proposed mechanism, is written as follows.
d(>FeIIIHA} dt
(mol m-2 hr 1) (3 .36)
Using the steady-state assumption once again, this rate expression can be set equal to zero and rearranged to give the following expression.
k 1 _ (>FeIIIHA} k_ 1 + k2 - (>FeIIIOH2+)[HA-] + k_2(>FeIIOH2}[HA·]
(I mo1-l) (3.37)
If the rate constants for the electron-transfer reaction are much smaller than for the surface complexation reaction (k2,L2<<k 1,L I) and if the concentration of surface iron(II) and ascorbate radical are small, then k2 and k_2(>FeIIOH2}[HA·] can be set to zero on the left and right side of the equation respectively. In this case, equation (3.37) collapses to the equilibrium expression for the surface complex
ki formation, SKsur=-k .
- 1
Chapter Hematite
4. by
57
The Reductive Dissolution Ascorbate in the Presence
Oxalate
4.1 EXPERIMENTAL METIIODS
of of
The dissolution and adsorption experiments were carried out in the same way as those for the reactions of hematite and ascorbate in the absence of oxalate. Oxalate concentrations were determined by scintillation counting using oxalate standard solutions labeled with C-14 oxalic acid (Amersham International). Stock solutions of sodium oxalate were prepared oo-1, 10-2, 10-3 mo11-l) and an identical quantity of C-14 oxalate was added to each stock solution to give an activity of 2.5x10-4 Ci 1-1 (250 nCi/ml) in each stock solution. One Curie (Ci) corresponds to 3.7xlQ-10 radioactive decays per second (1 Curie=3. 7x1 o-10 Bequerel). In order to determine oxalate concentra-tion in a reacting hematite suspension, a known volume of stock ox-alate solution was added at time=O to give the desired oxalate con-centration and a total C-14 activity in the reacting suspension of 1-3 nCi/ml. One ml aliquots were taken from the sample filtrate or stan-dard solutions and mixed with 9 ml xylene scintillation fluid (Luma Gel, LUMAC) and counted (Betamatic, Kontron Instruments) for either 20 or 50 minutes depending on the total C-14 activity added to the reacting suspension. Calibration curves were determined from stan-dard solutions prepared from the same stock solution that was used in the reacting hematite suspension.
4.2 RESULTS AND DISCUSSION 4.2.1 The Adsorption of Oxalate on Hematite
The stability constant for formation of the surface iron(III)-oxalate complex was estimated from the isotherm for adsorption of oxalate on hematite in the same way as was done for the adsorption of ascorbate on hematite and aluminum oxide (section 3.3.6). The for-mation reaction was written as follows.
58
-"' 'E a 0
..... E 0 ,... --x 0 10 -ar u.. /\ - O"--_,_~_,_~....___,_~_._~..____.~_,
0 50 100 150 200
-6 • 1 [H2 Ox] (1 O mol I )
sol
0.20 ~
'o E
"'e 0.15
..... 0 ,... 0.10 -x Q_ -Q) u.. 0.00 /\ 0 1 0 20 30 40 50 -
- -1 6 -1 [HOx ] (10 I mol )
Figure 4.1. a: Adsorption isotherm for oxalate on hematite, pH3, hematite=0.613 g 1-l. Since all experiments were made at the same pH the abscissa value may be given as [HOx-]=a1fH20xJsol· (a1=0.82; 1=0.01 mol z-1). b: Double reciprocal plot of {>FellloxJ-1
-1 vs. [HOx-]801
..
59
>FeIIIOH + HOx- ;::t. >FelIIOx + H10 (4.1)
The isotherm is shown in figure 4.1 a and the corresponding double-reciprocal plot of (>FelIIOxJ-1 versus [HOx·J-1 is shown in figure 4.lb. The fit parameters, the estimate of the surface stability constant and the adsorption maximum for oxalate on hematite are listed in Table 4.1.
Table 4.1 Surface Stability Constant, Adsorption Maximum, and Fit Parameters from the Double-Reciprocal Plot
a=0.02316 m2 J- l b=4.801x105 m2 moJ-1 r2=0.94 SKsur=2.07x 107 1 moI-1 Sm=2.08x10-6 mol m-2
l a _l_ + b S - b-1 SK b (>FeIIIOx) - [HA·] ' m - ' sur a acidity constants for oxalate: logKi=l.04, logK2=3.82 (250C, 1=0.1 mol 1-l; Smith and Martell,1976)
The adsorption maximum for oxalate is 2.0SxI0-6 mol m-2, com-pared with l.OlxI0-6 mol m-2 for ascorbate (section 3.3.5). Both of these values are much smaller than the ion-exchange capacity for hematite (l.02xt0-5 mol m-2) as measured by fluoride adsorption. The fluoride ion is much smaller and exhibits less steric crowding at high surface densities (mol m-2) of the adsorbed ion. The adsorption maximum for fluoride, oxalate (pH 3), and ascorbate (pH 3) are com-pared along with a characteristic dimension and projected surface area for each of the adsorbing ions in Table 4.2. The surface area of an iron site is taken here to be the area of the four-coordinated plane of a hematite octahedron. The projected surface area of an oxalate ion (28 A2) is approximately 3.5x the area of an iron site (8 A2), and
60
Table 4.2 *Comparison of Ion Dimension, Projected Surface Area, and Adsorption Maximum for Fluoride, Ascorbate, and Oxalate
adsorbing ion characteristic projected surface area adsorption length surface per adsorbing maximum
area molecule
A A2 A2 mol m-2
surface iron 2.8 8 site
fluoride 1.4 6 16 1x 10-5
oxalate 3 28 80 2xl0-6
ascorbate 6 113 160 lx 10-6
iron(III) site fluoride oxalate ascorbate 0
t~ t~ rFe<:c fF•, H II
*Characteristic lengths were estimated from the ionic radius in the case of fluoride, and from carbon bond lengths and structural geome-try for oxalate and ascorbate. Projected areas were calculated using the characteristic lengths as radii. Surface area per adsorbing molecule was calculated from the reciprocal of the adsorption density.
61
the projected area of an ascorbate ion (113 A2) is approximately equal to 14x the surface area of an iron site. Fluoride has a smaller projected surface area (6 A2) than an iron site and would not be af-fected by neighboring adsorbed fluoride ions. Thus it is believed that fluoride adsorbs approximately 1: 1 with the number of iron(III) surface sites at high concentrations of F-. The larger oxalate and ascorbate ions, on the other hand, may be influenced by crowding from neighboring adsorbed ions. The effective concentration of sur-face iron(III) sites available for reactions with adsorbed oxalate or ascorbate is only 10-20% of the adsorption maximum for fluoride.
4.2.2 Dissolution Kinetics of Hematite in the Presence of Ascorbate and Oxalate
The dissolution kinetics of a hematite/ascorbate/oxalate mixture were observed by following the dissolved concentration of iron(II), ascorbate, and oxalate with time. The concentration versus time curves for a single dissolution experiment are shown in figure 4.2. As was observed for the dissolution of hematite in the presence of ascor-bate without oxalate, the dissolution was initially relatively rapid and then became slower and linear with time. The rate of dissolution was determined by calculating the slope of the linear portion of the iron(II) versus time curve, after the first sample point, approximately one hour after addition of the reactants. For the results plotted in fig-ure 4.2 the rate of dissolution is calculated to be 4.7x10-6 mol 1-1 hr-1. The corresponding hematite concentration was 0.613 g 1-1.
The ascorbate concentration versus time curve also shows an ini-tially rapid reaction followed by a relatively slower and linear de-crease in ascorbate concentration with time. The rate of disap-pearence of ascorbate was calculated as the slope of the linear portion of the kinetic curve after the first sample point. This gave a reaction rate of -2.5x10-6 mol 1-1 hr 1, or about one-half the rate of dissolu-tion.
It was shown in Chapter 3 that the adsorption kinetics of ascor-bate in the presence of hematite were fast, coming to a steady-state in the first 10-15 minutes of the reaction.
0 E -CD
0 T"""
62
100
..... 80 -2.5
60
40
20
[H 20x) sol
0 0 2 3 4 5
TIME (hr)
Figure 4.2. Dissolved concentrations of iron(ll), ascorbate, and oxalate versus time in the presence of 10-4 mo11-l added ascorbate and 2.0xJ0-5 mo11-l added oxalate, pH=3, hematite=0.613 g 1-1.
The rapid increase in dissolved iron(II) concentration at the beginning was believed to be due to the release of more reactive surface sites. This initial rapid increase accounted for 5-15x1Q-6 mol 1-1 iron(II) in solution with the higher concentrations corresponding to higher con-centrations of ascorbate and oxalate. A 2: 1 reaction stoichiometry was reported for the dissolution of hematite in the presence of ascorbate without oxalate, and the stoichiometry for the results plotted in figure 4.2 were also approximately 2: 1 for the rate of dissolution compared to the rate of disappearence of ascorbate from solution. This sto-chiometry is expected from the two-electron loss of ascorbate going to its first stable oxidation product, dehydroascorbic acid (DHA) (section 3.4.1).
63
The ascorbate concentration was calculated in the same way as was described in section 3.3.2.
(mo11-l) (4.2)
Equation 4.2 is a mass balance for ascorbate with the amount of oxi-dized ascorbate estimated to be one-half of the solution concentration of iron(II). The amount of adsorbed ascorbate ( mol 1-1) calculated by (4.2) can be converted to a surface concentration (mol m-2) by divid-ing [H2A] sur by the amount of surface area in the reacting suspension (Ca, m2 1-1 ).
(mol m-2) (4.3)
The initial dissolved concentrations of ascorbate ([H2AJ0) and iron(II) ([Fe+ 2 Jo), corresponding to the initial reaction conditions where ad-sorption dominates the disappearence of ascorbate from solution, were estimated by extrapolating the linear portion of the respective kinetic curves to time=O. For the results plotted in figure 4.2 [H2Alo=8.85xl0-5 moI I-1, [Fe+2]o=l.13xl0-5 mol I-1, and [H2A]T=l0-4 mol 1-1. These concentrations can now be substituted into equation 4.2.
(mol 1-l) (4.4)
Equation (4.4) gives [H2Alsur=5.85xl0-5 mol 1-1 which is equivalent to a surface concentration of {>FelIIHA)=5.45xI0-7 mol m-2 (Ca=l0.73 m2 I-l).
The kinetic curve for dissolved oxalate concentration versus time also exhibits two distinct parts. The concentration initially decreases more rapidly, as with ascorbate, presumably due to rapid adsorption kinetics. Oxalate appeared to react no further after the first sample point, remaining at constant solution concentration throughout the remainder of the experiment. The adsorption of oxalate, [H20xJsur•
64
was calculated to be the difference in added concentration, [H20x]T, and the average of the oxalate concentrations determined for the sample points, [H20xlsol·
(mol l-1) (4.5)
The concentration of adsorbed oxalate (mol 1-1) was then converted to the surface density of oxalate (mol m-2) by dividing [H20xlsur by the
[H20xlsur surface area of the hematite suspension, {>FeIIIox} = Ca . For the
results plotted in figure 4.2, [H2 0 x] T = 2. 0 x 10 -5 mol 1-1 and [H20xlso1=3.0x1Q-6 mol l-1 giving [H20xlsur=l.7x10-5 mo11-l. This is equal to a surface concentration of {>FeIIIox}=l.58xl0-6 mol m-2 (Ca=l0.73 m2 1-1 ).
4.2.3 A Comparison of the Dissolution of Hematite by Ascorbate in the Presence and Absence of Oxalate
The results of four experiments at pH 3 in the presence and ab-sence of ascorbate and oxalate are shown in figure 4.3. For all of the dissolution experiments reported, the iron concentration on the verti-cal axis is the change in dissolved iron concentration after the first sample point. The slowest dissolution is the acid dissolution of hematite in the absence of ascorbate (reductant) or oxalate (chelate ligand). In the presence of 5.0xlQ-5 mol 1-1 oxalate, the oxalate-pro-moted dissolution of hematite is somewhat faster. In the presence of 1 o-4 mol 1-1 ascorbate without oxalate, the proton-assisted reductive dissolution is much faster than the oxalate-promoted dissolution. The fastest dissolution takes place in the presence of a combination of ascorbate and oxalate. The accelerated reductive dissolution due to oxalate is referred to here as the oxalate-promoted reductive dissolu-tion. The rates of dissolution and the corresponding surface concen-trations of ascorbate and oxalate for each experiment are listed in Table 4.3.
There is an order-of-magnitude increase between each of the rates of dissolution in going from the acid-dissolution (2x1Q-9 mol m-2 hrl) to the oxalate-promoted dissolution (2x10-8 mol m-2 hr-1) to the proton-assisted reductive dissolution (1.5x10-7 mol m-2 hr 1).
66
The rate of reductive dissolution in the presence of ascorbate and ox-alate together, 6x10- 7 mol m-2 hr 1, is about four times faster than the rate of reductive dissolution in the absence of oxalate. In addition, the surface concentrations of ascorbate and oxalate are significantly less than for the dissolution experiments carried out in the presence of ascorbate or oxalate alone.
The dissolution by oxalate alone and the acid dissolution without ascorbate or oxalate are much slower than reductive dissolution and are considered negligible as parallel reactions. In mixtures of ascor-bate and oxalate, the "proton-assisted" reductive dissolution and the "oxalate-promoted" reductive dissolution may be reacting in parallel. The total rate of dissolution is the sum of the rates for the two mecha-nisms. The name proton-assisted reductive dissolution refers to the mechanism of dissolution due to ascorbate alone as discussed in Chapter 3.
Re,T = Re,H + Re,Ox (mol m-2) (4.6)
Re, T is the total rate of dissolution, Re,H is the rate of proton-assisted reductive dissolution, and Re,Ox is the rate of oxalate-promoted re-ductive dissolution.
There are four possible explanations for the acceleration of the reductive dissolution in the presence of oxalate. 1. Oxalate accelerates the release of iron(III) sites from the hematite surface which are then reduced in solution. 2. Iron(II) produced by reductive dissolution of hematite forms solu-tion complexes with oxalate that react, in addition to ascorbate, as a more efficient reductant. 3. Oxalate itself acts in some way as an electron donor on the hematite surface creating a higher concentration of iron(II) and thus a faster rate of reductive dissolution. 4. Oxalate acts as a surface chelate to enhance the release of iron(II) sites into solution, similar to its role believed to accelerate non-reduc-tive dissolution.
The first explanation can be ruled out. If oxalate were acceler-ating the release of iron(III) sites from the surface then the rate of dissolution in a mixture of oxalate and ascorbate could never be faster
67 than the sum of the rates of dissolution due to ascorbate and oxalate reacting alone. Here, the rate of dissolution in the presence of ascor-bate and oxalate together is about 3.5 times faster than the sum of the rates of oxalate-promoted dissolution and proton-assisted reductive dissolution.
Explanation 2. is based on a number of published studies of the dissolution of iron(Ill)-oxide minerals catalyzed by combinations of iron(II) and a chelate ligand (Fisher,1972; Blesa et. a.,1984,1987; Cor-nell and Schindler,1987; Suter et. al.,1988). Suter et. al. observed the acceleration of the dissolution of hematite in the presence of oxalate upon addition of iron(II). They explained the acceleration as forma-tion of iron(II)-oxalate complexes that adsorb and undergo an inner-sphere electron exchange with the surface iron(llI) site at which the reducing complex adsorbs. The adsorbed complex is envisioned as being bound with oxalate acting as a bridging ligand between surface iron(III) and adsorbed iron(II). Additional coordination of oxalate to the iron(II) ion would lower the reduction potential sufficiently to create a favorable driving force for electron-transfer, via the oxalate bridge. The oxidized iron complex would then desorb leaving a more easily released iron(II) site on the surface. Upon release, this iron(II) would then participate in formation of the reducing complex. This was viewed as a catalytic process, producing iron(III) in solution but with no net increase in the concentration of iron(II). For the results shown in figure 4.3 it is possible that the iron(II) produced by ascor-bate reacting with the hematite surface could participate in such a catalytic process with oxalate.
This hypothesis was tested by adding various amount of iron(II) to a hematite/oxalate mixture at pH 3. The results are plotted in fig-ure 4.4 as the total dissolved iron concentration versus time in the presence of 0, 5x 10-5, and 1.5xto-4 mol 1-1 added iron(II). There was no acceleration of the rate of dissolution and the total concentra-tion of iron in solution did not change significantly during the dissolu-tion experiments. For the experimental system investigated by Suter et. al. (1988) the concentration range of oxalate was 0.33-3.3x 10- 3 mol 1-1 total added oxalate.
0 E
~o
150
68
150 µM iron(ll)
:s. 100
"""---~---~~---s_o.µ_M_ir_o.n(~ll~)----50
___..__.------1--- 0 iron(ll) oi..::::::..... __ __._ __ ....._ __ .J_ __ ....__J, __ _.... __ -'---....___.
0 10 20 30 40 50
Time (hr)
Figure 4.4 Effect of dissolved iron(II) on the dissolution of hematite in the presence of 5.0xJ0-5 mol l-1 oxalate, pH 3, hematite=0.613 g z-1. (Data provided by C. Siffert,1989)
For the conditions under which the experiments reported here were carried out ([Fe+2]so1<5x 10-5 mol 1-1, [Hz Ox] so1<5x 10-5 mol 1-1) the catalytic mechanism involving iron(II) and oxalate is not operative and cannot explain the marked acceleration of the rate of reductive dissolution observed in the presence of ascorbate and oxalate to-gether.
The third possible explaination for the enhanced rate of reductive dissolution is that oxalate itself acts as an electron donor, in addition to ascorbate, to help reduce the hematite surface. Oxalate itself, in the absence of light, does not lead to a reductive dissolution of hematite (Siffert,1989). It may be possible, however, that adsorbed oxalate could act as a radical quencher to enhance reduction of the hematite surface by ascorbate Such a reduction reaction may be written as follows.
>FeIIIHA +HzO +::± >FeIIOHz + HA· (4.7)
69
The electron-transfer would occur within the iron(III)-hydrogen ascorbate complex producing iron(II) and ascorbate radical. If oxalate were to react with the ascorbate radical then the back-reaction would be inhibited and the oxidation-reduction steady-state would shifted to the right resulting in an enhanced concentration of iron(II) on the hematite surface, thus accelerating the rate of reductive dissolution.
For the experimental results plotted in figure 4.2 the concentra-tion of oxalate decreases only during the first hour of the experiment, presumably due to fast adsorption. After the first hour the concen-tration of oxalate remains nearly constant in solution. If oxalate were being oxidized it is expected that C02 would be produced. The reac-tors were purged continually with nitrogen gas (N2) during the ex-periments and most of the C-14 labeled C02 produced would have been purged with the outflowing Ni stream and not conserved within the reactor. It would be expected that the measured concentration of oxalate would decrease continuously during an experiment, not re-main constant as seen in figure 4.2.
If oxalate were acting as an additional electron donor in addition to ascorbate then the measured stoichiometry between iron(II) and ascorbate should appear significantly greater than 2: 1; ascorbate would appear to give up more than two electrons during its oxidation. For the results plotted in figure 4.2 the stoichiometry between iron(II) and ascorbate is 1.88:1. Based on the conservation of oxalate in solu-tion after the first hour of the reaction, and the approximately 2:1 stoichiometry between iron(II) and ascorbate, it is not believed that oxalate participates directly as an electron donor in the oxalate-promoted reductive dissolution.
The fourth possible explanation for the acceleration of the re-ductive dissolution in the presence of oxalate is that oxalate enhances the rate-determing release of iron(II) ions from the hematite surface in much the same way chelate ligands are believed to accelerate the release of surface iron(III) sites in non-reductive dissolution. In this case oxalate would lower the activation energy for the detachment of surface iron(II) ions by forming a chelate complex which concentrates electron density onto the iron(II) metal center, thus polarizing and weakening the iron(II)-lattice oxygen bonds. Such an effect may oc-
70
cur upon reduction of a surface iron(III)-oxalate complex by ascorbate possibly through formation of a ternary surface ion pair (>Felllox,HA) or adsorption of oxalate at surface iron(II) sites (>Fello H + H 0 x -<=!>FeIIOx+H20). In either case the effect of ascorbate and oxalate to-gether would be synergistic leading to a lower activation energy for the release of the surface iron(II)-oxalate complex than would be ex-pected for dissolution by ascorbate or oxalate reacting alone. It is also expected that the rate of dissolution would show a dependence on both surface ascorbate and surface oxalate concentration.
4.2.4 The Effect of Ascorbate Concentration on the Rate of Dissolution in the Presence of Oxalate
A series of experiments were carried out at various added con-centrations of ascorbate in the presence of 5.0xIQ-5 mol 1-1 added ox-alate. Dissolved iron(II) concentration versus time is plotted for some of the experiments in figure 4.5. The iron(II) concentration plotted on the vertical scale is the change in dissolved iron(II) concentration af-ter the first sample of an experiment was taken. Iron(II) concentra-tions in the first sample taken were 5-15 mol 1-1 with the higher con-centrations corresponding to the higher added ascorbate concentra-tions at pH 3. For both the experiments at pH 3 and pH 4 the rates of dissolution increase with increasing concentration of added ascorbate. The rates of dissolution are significantly slower at pH 4 than at pH 3.
The rate of reductive dissolution in the absence of oxalate fol-lowed a first-order rate law in adsorbed ascorbate concentration.
(mol m-2 hr-1) ( 4.8)
Such a rate law was shown to be equivalent to the following rate law for the dependence of dissolution rate on the solution concentration of the adsorbing reactant species (section 3.3.6).
(mol m-2 hr-1) (4.9)
-. "' .
7 1
50
0 40 E 0 ..... 30 -........ ::::;:;-- 20 Q) u.. ........
........ 10 --Q) u.. ........ 10 20 30
TIME (hr) Figure 4.5. Dissolved iron concentration versus time for dissolution at various added concentrations of ascorbate in the presence of 5.0xJ0-5 mol 1-l oxalate, pH 3 and 4., hematite=0.613 g 1-l. For the slowest dissolution (without ascorbate) the total iron concentration was determined
72
k' (hr-1) is a first-order rate constant, SKsur (I mol-1) is the stability constant for formation of the surface iron(III)-hydrogen ascorbate complex, a 1 is the distribution coefficient for the reacting solution
[HA-] species (at= Cr , CT=[HzA]+[HA ·]+[A-2]), and [H2AJ0 is the ex-
perimentally determined total concentration of dissolved ascorbate species ([H2Alo=CT). Data which can be linearized by a double-recip-
rocal plot, RATE-l=a[HzAl{/ +b, can be fit to equation (4.9) .with
a=(k'SKsurSmal)-1 and b=(k'Sm)-1. The linear regression of a plot of 1 RA TE-1 versus [H2Alo was used to obtain the values of a and b.
[H2AJ0 These constants were then used to plot RATE-a+b[H
2A]o' correspond-
ing to equation (4.9) with the equation coefficients written in terms of a and b. This empirical rate law for the dependence of dissolution rate on dissolved ascorbate concentration is plotted as the solid line, along with the experimental data for the dissolution experiments, in figure 4.6a. The double-reciprocal plots of the data are shown in fig-ure 4.6b.
The rate of dissolution is plotted against surface ascorbate con-centration in figure 4.7. The surface concentrations of ascorbate are much lower here than for the reductive dissolution in the absence of oxalate, at similar concentrations of added ascorbate. The detection limit for ascorbate was about 10-6 mol 1- l and the lowest concentra-tions of surface ascorbate (2xl0· 7 mot 1-1) correspond to about a 2x 10-6 mol l-1 change in dissolved ascorbate concentration in the mass balance used to calculate adsorbed ascorbate. It is not possible to obtain an accurate correlation between rate of dissolution and sur-face concentration of ascorbate from this data. The slope of a linear regression through the data was used to estimate values for first-or-der rate constants (ke,hr-1) as was done for the in the absence of oxalate, corresponding RATE=ke{>FeIIIHA} (section 3.3.6).
reductive dissolution to the rate law
73
a ' ... .c ~ ·- 6
w ~ 2 a:
-~ 3
.... .c <O 0
-i.u I-< a:
2
20 40 60 80 100 -6 -1
[H2A] 0 (10 mol I }
0.05 0.10 0.15
Figure 4.6. a: Dependence of dissolution rate on the initial concentration of total ascorbate in solution in the presence of 5.0xJ0-5 mo/ [-l oxalate, and b: the double recipricol plot of RATE-I vs. [HzAJo-1, pH=3 and 4, hematite=0.613 g z-1.
74
10
8 ' .... .r;;
':'E 6
0 E ...
·o 4 ~ w pH4 ~ 2
a: 0
0 4 6 8 10
Figure 4.7. Dependence of dissolution rate on surface concentration of ascorbate in the presence of 5.0xJ0-5 mol z-1 oxalate, pH=3 and 4, hematite=0.613 g z-1.
Table 4.4 Comparison of the Empirical First-Order Rate Constants for the Reductive Dissolution by Ascorbate in the Absence and
Presence of 5xI0-5 mol i-1 Oxalate
ascorbate alone ascorbate + oxalate
hematite = 0.613 g 1-1
pH 3
0.11 hr 1 1.2 hr I
pH 4
0.045 hr-1 0.16 hr-1
75
These calculated rate constants from figure 4.7 and the empirical first-order rate constants for the reductive dissolution in the absence of oxalate are compared in Table 4.4.
4.2.5 The Effect of Oxalate Concentration on the Rate of Dissolution in the Presence of Ascorbate
A series of experiments were carried out at various concentra-tions of added oxalate in the presence of 10-4 mol 1-1 ascorbate at pH 3 and pH 4. The rate of dissolution is plotted against total dissolved
[H20xlsol oxalate concentration in figure 4.8. The relation, RATE-a+b[H20xJsol,
is plotted as a solid line as well. The constants a and b were calcu-lated as before from a linear regression of the double-reciprocal plot
1 -1 of RATE- vs. [H20x~01 .
' ... .c .-0 E
<D ·o
T'"
w ~ a:
8
7
6 I pH3
s
4
3
2 pH4
10 20 30 40 so -6 ·1
[H . Ox] ( 1 O mol I ) 2 sol
Figure 4.8. Rate of dissolution versus total dissolved oxalate concentration in the presence of 10-4 mol 1-J ascorbate, pH 3 and 4, hematite=0.613 g 1-l.
76
The rates of dissolution are plotted against the surface concen-tration of oxalate in figure 4.9. The rate of dissolution appears to be almost independent of adsorbed oxalate concentration below ap-proximately l o-6 mol m-2. There is also a significant reductive dis-solution due to ascorbate alone at {>FelIIQx}=O. Above {>FelIIQx} = I o-6 mol m-2 the increase in rate of dissolution is approximately proportional to the surface oxalate concentration.
7 ...... .c 6
"' 'E 5 0 E 4
.... ·o 3 .... - 2 w I-~ 0
0 5 10 15 20 25
{>Fe(lll)Ox} (10-7 mol m·2)
Figure 4.9. Dependence of dissolution rate on adsorbed oxalate concentration in the presence of 10-4 mo/ 1-l ascorbate, pH 3 and 4, hematite=0.613 g 1-l.
The surface concentration of ascorbate was also measured in this se-ries of experiments. In figure 4.10 the surface concentrations of ascorbate and oxalate are plotted against the corresponding dissolved oxalate concentration for each experiment. In figure 4.10 a transition can be seen between the dominant adsorbed species that depends on the concentration of dissolved oxalate. At low dissolved oxalate con-centrations in the presence of 10-4 mol 1- l added ascorbate, ascorbate is the dominant surface species.
77
30 ->< a pH3 0 -~~-; 20 {>Fe(lll)Ox} A - 0
<(' ..... E :::c :::::-::=:,.. Q) u. A -
->< Q. -Q) u. A --<( :::c ---Q) u. A -
'o 10 ,.... {>Fe(lll)HA} -0
0 10 20 30
·6 • 1 [H 20X]
601(10 mol I )
30
b pH4 N
'E 20 {>Fe(lll)Ox}
..... '
0 E 0 10 ,....
{>Fe(lll)HA}
10 20 30
·6 • 1 [H Ox] (10 mol I )
2 sol
Figure 4.10. Adsorption of ascorbate and oxalate at various solution concentrations of oxalate in the presence of J0·4 mol /·1 ascorbate, a: pH 3 and b: pH 4. Hematite==0.613 g /·1.
78
Above the transition, adsorbed oxalate predominates. The surface concentration where the two surface species exist at equal concentra-tions is about 10-6 mol m-2. Thus, a significant acceleration of the reductive dissolution in the presence of oxalate was only seen under conditions where oxalate was the predominant surface species. A pos-sible explanation of the displacement of ascorbate from the hematite surface is adsorption competiton between ascorbate and oxalate. The Langmuir competitive adsorption isotherm is discussed in Appendix 1 with the data from the experiments at pH 4 used for an example cal-culation.
It was shown in the previous section that the oxalate-promoted reductive dissolution depends on the surface concentration of ascor-bate. For the region of the plotted curves in figure 4.9 where the rate of dissolution is proportional to surface oxalate concentration, the surface concentration of ascorbate is almost constant and the disso-lution may also be approximately described by a second-order rate law.
7
6 ~
' .... ..c: C\I 5 ·E 0 4 E ,._
'o 3
w 2
I-<( a:
o ........................ _._ ........ ~..._ ...................... _. 0 2 4 6 B 10 12
{>Fe(lll)OX}{>Fe(lll)HA} -13 2 -4
(10 mol m J
Figure 4.11. The rate of oxalate-promoted reductice dissolution as a second-order function of surface concentration of both ascorbate and oxalate, pH 3, hematite=0.613 g l-1.
79
This can be seen for the data at pH 3 but not at pH 4.
RATE= ke,ox{>FeIIIHA) {>FeIIIOx} (mol m-2 hrl) (4.10)
The data for the oxalate-promoted reductive dissolution at pH 3 have been plotted as an example of such a relation in figure 4.11. No sim-ple rate law was found to describe the dissolution data over the entire range of surface oxalate concentration.
4.2.6 The Effect of pH The rates of dissolution in the presence of 10-4 mol 1- l ascorbate
and 5xl0-5 mol 1-1 oxalate were compared at various pH values. Dis-solution at pH 6 proved to be very slow and higher concentrations of reactants were used (2.5x 10-4 mol 1-1 ascorbate, 10-4 mol l-1 ox-alate). The rate of dissolution is plotted against pH in figure 4.12. The rates of dissolution and the corresponding surface concentrations of ascorbate and oxalate are listed in Table 4.5. Here, the maximum rate of dissolution was observed at pH 3 with lower rates of dissolution at pH 2.5 and 4.
·~ .c N
'E 0 E
.... •O c w .... <( a:
6
6
4
2
0 2 3 4 5 8 7
pH
Figure 4.12. Dependence of dissolution rate on pH in the presence of 10-4 mol 1- l ascorbate and 5xl0-5 mol 1-l oxalate, hematite=0.613 g 1-l. At pH 6: 2 .5xJ0-4 mol 1-l ascorbate, 10-4 mol 1-l oxalate.
80
Table 4.5. Dissolution Rates and Surface Ascorbate and Oxalate Concentrations in the Presence of 1 o-4 mol 1-1 Ascorbate and
5x10-5 mo11-l Oxalate at Various pH Values
pH RATE oo-7 mol m-2 hr-1)
{>FelIIHA} (10-7 mol m·2)
2.5 3.53 3.0 5.96 4.0 1.47 6.0 <0.01
* 2.5xI0-4 mol 1- l ascorbate added ** 10-4 mol 1-1 oxalate added hematite = 0.613 g 1- l
<1 4.10 4.64 11.8*
(>FeIIIOx} (1 o-7 mol m-2)
21.6 19.6 20.5
13.0**
There was virtually no observable rate of dissolution at pH 6. The slow dissolution at pH 2.5 may be explained by the low surface con-centration of ascorbate ({>FelIIHA}<I0-7 mol m-2). Similar dissolu-tion maxima have been seen for the reductive dissolution of goethite by hydroquinone and methylhydroquinone (LaKind and Stone,1989). Both reductants exhibited a dissolution maximum with respect to pH at pH 4. In both cases there was no dissolution observed at pH 6 or above. Low surface concentrations of reductant at pH values below pH 4 and adsorption of iron(II) at higher pH values were given as pos-sible explanations for the observed maxima.
The slow rates of dissolution observed here at pH 4 and 6 cannot be explained by low surface concentrations of either ascorbate or ox-alate. Both are adsorbed at significant concentrations at these pH val-ues. However, as in the explanation proposed by LaKind and Stone, iron(II) adsorbs strongly on hematite and other metal oxide surfaces above pH 5 (Blesa et. al.,1984; Mulvaney et. al.,1988; Wehrli,1989) and may inhibit the reductive dissolution by blocking active dissolu-tion sites on the hematite surface.
8 1
4.2.7 The Role of Oxalate in Promotini Reductive Dissolution The accelerating effect of oxalate on the rate of reductive dissolu-
tion is believed to arise from the destabilization of a surface site through complexation, acting synergistically with the increased !abil-ity of the site due to reduction by ascorbate. There are two possible pathways which may lead to this accelerating effect; reduction of surface iron(IIl)-oxalate complexes or adsorption and complexation by oxalate at iron(II) sites. Reduction of surface iron(III)-oxalate complexes will be discussed first.
Electron-transfer to surface iron(IIl)-oxalate complexes may oc-cur either directly or upon reduction of a neighboring iron site with a subsequent "self-exchange" reaction occuring between the two surface sites. The mobility of electrons on the surface is unknown and move-ment of electrons within the crystal lattice due to the semi-conductor properties of hematite may play a role (Mulvaney et. al.,1988). Direct electron-transfer may occur from ascorbate to a surface iron(III)-ox-alate complex through the conjugated carbon-bonding system of ox-alate acting as an electron bridge. One possible configuration for such a reaction is shown in figure 4.13. This configuration is pictured here as a surface iron(Ill)-oxalate, -ascorbate ion pair. The formation of such an ion pair may be effected by the overall charge on the hematite surface, or the charge on neighboring surface sites. Even if the ion pair were present in small concentrations, it could lead to a significant acceleration of the rate of dissolution if the release of the resulting iron(II)-oxalate complex were much faster relative to the release of surface iron(II) ions into solution. Electron-transfer may be slow for such a reaction. If this were the case, and the release of the resulting iron(Il)-oxalate complex from the hematite surface would be relatively fast, then electron-transfer may become the rate-determing step in the overall reaction.
The second possible pathway for formation of surface iron(Il)-ox-alate complexes is adsorption of oxalate at iron(II) sites subsequent to electron-transfer. The detachment of iron(II) would occur in a two-step sequence. Oxalate would first adsorb, forming the surface iron(II) complex, followed by a relatively slower release of the surface iron(II) complex into solution.
82
Fe(lll)OH
/ // 0 H ""' /oIO o
Fe( Ill) ( eo ""' e-
Fe(lll)OH
/ Figure 4.13 Possible configuration for reduction of a surface iron(lll)-oxalate complex by ascorbate with oxalate as an electron bridge.
If the release of surface iron(II)-oxalate complexes is significantly faster than the release of surface iron(II) ions, then an acceleration of the rate of dissolution would be seen.
Using the dissolution mechanism presented in Chapter 3 for the reductive dissolution of hematite by ascorbate, an overall mechanism for dissolution in mixtures of ascorbate and oxalate would be as fol-lows. Surface complexation at iron(II) sites is taken as the pathway for acceleration by oxalate.
83 kl
>FeIIIOH2+ + HA- f2 >FeIIIHA + H20 k_ 1
k2 >FeIIIOH + HOx- f2 >FeIIIOx + H20
k_2
k3 >FeIIIHA + H20 f2 >FeIIOH2 + HA'
k_3
k4 >FelIOH + HOx- f2 >FelIOx H20
k_4
{H+),k5 >FeIIOH2 -+ new surface site + Fe+2(aq)
slow
new surface site + FeOx(aq)
(4.11)
(4.12)
(4.13)
(4.14)
(4.15)
(4.16)
Rapid adsorption and electron-transfer are assumed to be in a steady-state followed by slow release of surface iron(II)-ions and -oxalate complexes into solution. Equation (4.15) corresponds to the proton-assisted reductive dissolution and (4.16) to the oxalate-promoted re-ductive dissolution. These two reactions in parallel give rise to the additive rate law for the total rate of dissolution, Re,T = Re,H + Re,Ox· Because the concentration of surface iron(llI) sites is much greater than the concentration of iron(II) sites, the adsorption of reactants would be dominated by formation of surface iron(III) complexes. The more labile surface iron(II) complexes, although present in much smaller concentrations, would be the precursor complexes for the ki-netic-controlling detachment of surface sites in the reductive dissolu-tion.
84
4.2.8 Protons and Lii.tands as Environmental Factors Both of the proposed pathways for formation of surface iron(II)-
oxalate complexes may occur in parallel to the proton-assisted reduc-tive dissolution. A general rate law was written as the sum of proton-assisted reductive dissolution (Re,H) and oxalate-promoted reductive dissolution (Re,Ox). Re,T = Re,H + Re,Ox· In Chapter 3, Re,H was writ-ten to reflect a dependence both upon the surface concentration of ascorbate and the surface concentration of protons. In this chapter, the rate of oxalate-promoted reductive dissolution depended upon the surface concentrations of ascorbate and oxalate. A general approach to the kinetics of chemical transformations in natural waters can be applied by writing a pseudo first-order rate law for the reacting sub-strate, in this case ascorbate. Effects which alter the rate of reaction, environmental factors (temperature, inhibitors, catalysts), are lumped together in a pseudo first-order rate constant (Hoigne, 1989). The rate of proton-assisted, and oxalate-promoted, reductive dissolution de-pend on the surface concentration of ascorbate and can be written as pseudo first-order rate laws in surface ascorbate concentration.
Re,T = ke,H{>FeIIIHA} + ke,H,Ox{>FeIIIHA) (mol m-2 hr-I) (4.17)
The rate constants can be summed together.
Re,T = (ke,H + ke,H,OxH>FeIIIHA) (mol m-2 hr-1) (4.18)
This rate law clearly separates the role of the substrate (ascorbate) from the rate constants which depend upon the environmental factors. The pseudo first-order rate constants, ke,H and ke,H,Ox• vary with the concentrations of protons and oxalate. Under certain conditions, such as low surface oxalate concentration, ke,H may be larger than ke,H,Ox· At other concentrations, ke,H,Ox may be larger. In this way, depend-ing on the concentrations of the environmental factors, the predomi-nant mechanism of dissolution may shift between proton-assisted re-ductive dissolution and oxalate-promoted reductive dissolution.
Changes in concentrations of the environmental factors are be-lieved to alter the activation energy for the reaction discussed here. This would have an exponential effect on the rate of reaction. Changes
85
in substrate concentration, on the other hand, have only a linear effect on the rate of reaction. Reductive dissolution is significantly acceler-ated in the presence of oxalate, even with less ascorbate on the hematite surface. It is believed, for both the non-reductive and re-ductive dissolution, that formation of mono-nuclear bi-dentate surface complexes by ligands such as oxalate would lower the activation en-ergy for the release of surface metal sites into solution. For the ox-alate-promoted reductive dissolution, the accelerating effect due to surface chelate formation would, in most cases, outweigh the linear effect of a decrease in surface ascorbate concentration. Thus the overall rate of dissolution depends upon the interplay between envi-ronmental factors and substrate, and between competing mechanisms reacting in parallel.
86
Chapter 5. A Comparison of the Effects of Oxalate, Phosphate, EDT A and Aluminum on Rates of Reductive
Dissolution
5.1 INTRODUCTION Three other reactants were chosen to compare their kinetic effect
on the rate of reductive dissolution with that of oxalate. EDTA was chosen because it forms very stable complexes with iron. Phosphate was chosen as a reactant because it is an inorganic chelate ligand of environmental significance due to its role in the eutrophication of nat-ural waters. Aluminum was chosen as a reactant because it is redox inert under the conditions of these experiments and adsorbs strongly on metal oxides, possibly inhibiting the dissolution. Mobilization of aluminum due to watershed acidification is a question of environ-mental interest due to the possible biologically toxic effects of alu-minum ions (Schnoor and Stumm, 1987). It may also be interesting to consider the effect of increased aluminum mobilization on geochemical kinetics with respect to dissolution of reducible metal oxides. Exam-ining the effects of these three reactants on rates of reductive dissolu-tion will hopefully provide some contrast with the kinetic effect of ox-alate and also act as a conceptual bridge to the chemistry of natural waters discussed in Chapter 6.
5.2 EXPERIMENT AL METHODS The dissolution experiments were carried out as described in
Chapter 3. Experiments at pH 3 required no pH stat whereas ex-periments at higher pH were carried out at constant proton activity as described in Chapter 3. For the effect of aluminum on reductive dissolution it was necessary to determine the adsorption of aluminum at various pH values. The pH of a suspension of 0.613 g 1-1 hematite was initially adjusted by a small addition of HN03. A known volume of aluminum standard solution was then added to give a total concentration of 2.5x1Q-5 mol 1-1 A1+3 in the reactor (T=250C, I=0.01 mol 1-1 ). The suspension was filtered after 60 minutes and the filtrate analyzed for Al+3 and pH. The aluminum concentration in the filtrate was determined spectrophotometrically (Dougan and Wilson, 197 4)
87 and the adsorbed concentration determined by difference from the total added concentration and the filtrate concentration.
5.3 RESULTS AND DISCUSSION 5.3.1 The Effects of EDTA and Phosphate on the Rate of Reductive Dissolution
For both the reductive dissolution by ascorbate alone and in the presence of oxalate there was no observable dissolution at pH 6. A ligand which forms more stable complexes with iron(III) and iron(II) may lead to an acceleration of the reductive dissolution at more
40 250 µM ascorbate pH 6 +
100 µM EDTA 30
0 <D
E •O
20
0 ~.
N + Q)
!::!:.
en
10
250 µM ascorbate
0 0 20 40 60 80 100
Time (hr)
Figure 5.1 The EDTA-promoted reductive dissolution compared with dissolution by ascorbate or EDTA alone. The bottom dissolution curve is the reductive dissolution in the presence of oxalate. Total iron was measured for the dissolution with EDTA alone. Hematite=0.613 g l-1.
88 neutral pH values. Two experiments were carried out with EDTA at pH 6, one with EDTA alone and one with EDTA in the presence of ascor-bate. The results in figure 5.1 show a significant dissolution due to EDT A alone and also a marked acceleration of this dissolution in the presence of ascorbate. The rate of dissolution and the surface con-centration of ascorbate for each experiment are listed in Table 5 .1.
The reactivity of EDTA toward the hematite surface is due to sev-eral effects. EDTA alone leads to a significant dissolution producing iron(III) which may be reduced in solution in the presence of ascor-bate. EDTA may also accelerate the reductive dissolution in the same way as oxalate, through surface complex formation with iron(II) sites. In addition, EDTA may participate in the catalytic dissolution of hematite by forming reductive iron(II) complexes (Dos Santos Afonso et. al., 1988).
Table 5.1. Dissolution at pH 6 in the Presence of Ascorbate and/or Oxalate or EDT A
Reactants (l0-4 mo11-l)
ascorbate, 2.5
ascorbate, 2.5 oxalate, 1.0
EDTA, 1.0
ascorbate, 2.5 EDTA,1.0
Hematite=0.613 g 1-1
Rate (l0-7 mol m-2 hrl)
<0.02
<0.01
0.17
0.62
{>Fe(III)HA} (IO-7 mol m-2)
24.2
11.8
5.97
89 As discussed in the previous chapters, iron(II) adsorbs strongly
above pH 5 and may inhibit the reductive dissolution by blocking ac-tive sites for dissolution. Oxalate does not form very stable complexes with iron(II) in solution (logKS=3.05, 25oc, 1=1.0 mol 1-1; Smith and Martell,1976). Under the conditions for the experiments reported in Chapter 4 ([H20xJso1<5x10-5 mo11-l, [Fe+2Jso1<5x10-5 mol l-1, 1=0.01 mol 1-1) this gives a concentration for dissolved iron(Il)-oxalate complexes on the order of 10-6 mol 1-1 or less. EDT A, on the other hand, forms very stable complexes with iron(II) (logKS=14.3, 20oc, l=O.l mol I-1; Smith and Martell, 1976). By competing with the hematite surface for iron(II), EDTA would prevent the possible inhibi-tion of the dissolution due to adsorption of iron(II). Published data of the adsorption of iron(II) on magnetite in the absence and presence of 5x10-4 mol 1- l EDT A shows a strong adsorption edge at pH 5 in the absence of EDT A and a weak adsorption edge at pH 7 in the presence of EDTA (Blesa et. al.,1984). Here it is likely that EDTA acts both to ac-celerate the reductive dissolution directly, and perhaps more impor-tantly, to prevent the inhibition by keeping iron(II) in solution as an iron(Il)-EDT A complex.
In addition to the results for oxalate and EDT A, the kinetic effect of ortho-phosphate (H3P04) on the reductive dissolution of hematite by ascorbic acid was tested. The effect of phosphate on the rate of re-ductive dissolution in the presence of 10-4 mol 1- l ascorbate is shown if figure 5.2. At both pH 3 and pH 4 the addition of 10-4 mol l- 1 phosphate approximately doubled the rate of reductive dissolution. There was no measurable ascorbate adsorbed (<I0-7 mol m-2) in the presence of phosphate. As with oxalate and EDTA, phosphate leads to an acceleration of the reductive dissolution while displacing ascorbate from the hematite surface. Phosphate, like oxalate and EDTA is capa-ble of forming surface chelate complexes which may accelerate the dissolution.
5.3.2 The Re<luctive Dissolution in the Presence of Aluminum The effect of certain chelating ligands to accelerate the reductive
dissolution of hematite by ascorbate has been demonstrated. In figure 5.3 the effect of aluminum on the rate of reductive dissolution at pH 3 and 4 is shown. There was no difference in the rate of reductive
0 E
~o ..... -0 "' .......
t\I +
Q)
!:!::..
90
50
pH3
40
30
20
10
100 µM phosphate, no ascorbate, pH 3 0
0 10 20 30 40 50
Time (hr)
Figure 5.2. Effect of phosphate on the rate of reductive dissolution in the presence of 10-4 mol l-1 ascorbate, pH 3 and 4. Hematite=0.613 g 1-l. Total iron was measured for the dissolution by phosphate alone.
91
20
·-0 E
<D ·0 ..-- 10
"' +
0 Ill
Q)
!:!:..
10 20 30
Time (hr)
Figure 5.3. Effect of aluminum on the rate of dissolution in the presence of 10-4 mol l-1 ascorbate at pH 3 and 4, hematite=0.613 g l-1.
dissolution in the presence or absence of aluminum at pH 3. At pH 4 however, the rate of dissolution by ascorbate in the presence of alu-minum is only one-third that in the absence of aluminum. The presence of aluminum clearly inhibits the reductive dissolution at pH 4. The adsorption of ascorbate in the presence of aluminum (15.8x10- 7 mol m-2) is about the same as in the absence of aluminum (13.6x1Q-7 mol m-2). The slower dissolution at pH 4 cannot be explained by a lower surface concentration of ascorbate.
The pH dependence of the adsorption of aluminum on hematite, in the pH range of the dissolution experiments, was also determined. The adsorption of aluminum over the pH range 2.5-4.5 is shown in figure 5.4. The strong adsorption edge beginning at pH 3.5 coincides with the inhibition of dissolution at pH 4 but not at pH 3.
-C\I
'E 0 E ....
'O T"" -<')
+ <( -
10
8
6
4
2
0 2
92
+3 -5 -1 [Al ] = 2.5x10 mol I
T
3
pH 4 5
Figure 5.4. Adsorbed concentration of aluminum versus pH. Hematite=0.613 g 1-I.
This suggests that formation of surface aluminum complexes slows the rate of dissolution. One explanation may be that aluminum forms bi-nuclear surface complexes which can anchor iron(II) to the surface. As mentioned in Chapter 2, the greater activation energy required to release two surface metal centers simultaneously into solution would inhibit the release of these complexes from the crystal surface. For-mation of such a complex is schematically shown as follows.
"Fe(ll)OH-
/ '-....
Fe(lll)OH
I
+3 + Al(OH
2 )6
" Fe(ll)O '-...
"' I Al(OH24
'-.,..Fe(lll)O /
/ (5.1)
93
A more general explanation is that metal ions form surface com-plexes preferably at crystal discontinuities such as steps and corners (Blum and Lasaga,1987). As explained in Chapter 2 these sites are also the most easily detached because of the lower degree of coordi-nation to the crystal lattice. Blocking of these sites through formation of surface metal complexes would decrease the mole fraction of active surface sites, thus lowering the overall rate of dissolution. This may account for the inhibition of reductive dissolution at pH 6. Adsorbing iron(II) would play the same role as aluminum. The adsorption edge for iron(II) on hematite is about pH 5, for aluminum pH 3.5. For both metals, the reductive dissolution is strongly inhibited at pH values above the adsorption edge.
A parallel can be drawn between aluminum or iron(II) adsorption in the inhibition of reductive dissolution and passive film formation in the corrosive dissolution of elemental iron. In both cases the adsorp-tion of a redox stable metal ion or complex (Al+3, iron(III) in an oxy-genated environment, iron(II) in a reducing environment), and even-tual transition at high surface concentrations to a new thermodynami-cally stable surface phase, would protect the bulk solid phases against attack from the reductant or oxidant.
94
Chapter 6. Conclusions and Perspective
6.1 CONCLUSIONS A summary of the results and conclusions from the experimental
work are listed below.
1. The reductive dissolution of hematite by ascorbate, its accelera-tion due to protons and chelate ligands, and its inhibition by metal ions is a surface chemical process. In order to gain some understanding of the chemical processes occurring on the hematite surface, it is neces-sary to consider surface speciation. Here surface concentrations (mol m-2) of ascorbate and oxalate were calculated from mass balances us-ing analytically determined solution concentrations in the presence and absence of hematite particles.
2. The rate of dissolution, at a single pH, is proportional to the con-centration of ascorbate. A first-order rate law can be written which specifically includes the concentration of a surface iron(III)-ascorbate complex, RATE=ke{>FeIIIHA} (mol m·2 hr-1 ).
3. The empirical rate constant, ke (hr-1) showed a fractional-order dependence on proton concentration (approximately [H+ ]0.6). With help of a master isotherm for proton adsorption on metal oxides this rate dependence can be interpreted as a third-order dependence on surface proton concentration. This result provides evidence that re-lease of iron(II) sites from the hematite surface, rather than electron-transfer, is the rate-determining step in the overall reaction. Protona-tion of neighboring surface sites would lower the activation energy for release of the iron(II) site into solution.
4. A comparison of the adsorption of ascorbate on hematite and o-A I 2 0 3 indicated that the ascorbate which disappears from solution in the presence of hematite particles is present predominantly as a sur-face iron(IIl)-ascorbate complex rather than being oxidized. Results from these experiments also suggested that the surface concentration
95 of iron was small compared to the surface concentration of ascorbate. The stability constants for the surface iron(III)-hydrogen ascorbate complex (logSKsur=6.0) and the surface aluminum-hydrogen ascorbate complex (logSKsur=5.9), were estimated from Langmuir adsorption isotherms at pH 3 and found to be similar. These results indicated that the adsorption of ascorbate on hematite, as a first approximation, can be treated as adsorption on a non-reducible oxide surface.
5. The rate of reductive dissolution was increased significantly in the presence of oxalate, a ligand capable of forming mono-nuclear bi-den-tate surface complexes. The empirical first-order rate constants for the reductive dissolution in the presence of oxalate were significantly larger than the rate constants for proton-assisted reductive dissolu-tion. In addition, the acceleration of the dissolution was significantly greater than the sum of the individual rates of dissolution in experi-ments with ascorbate and oxalate reacting alone. The acceleration was only seen under conditions where oxalate, rather than ascorbate, was the predominant surface species. Under these conditions, the rate of oxalate-promoted reductive dissolution showed an approximately lin-ear increase with the surface concentration of oxalate. The kinetic effect of ascorbate and oxalate together can be explained by the desta-bilization of a surface site through formation of a mono-nuclear surface chelate complex, acting synergistically with the increased la-bility of the surface site due to reduction by ascorbate. The precursor complex for the oxalate-promoted reductive dissolution would be a surface iron(II)-oxalate complex which can be formed either through reduction of a surface iron(llI)-oxalate complex, or adsorption and complexation by oxalate at a surface iron(Il) site subsequent to electron-transfer.
6. There was no observable rate of reductive dissolution at pH 6, even in the presence of oxalate. Iron(Il) adsorbs significantly on hematite at pH 6 and may inhibit the reductive dissolution by blocking active sites for dissolution.
7. The reductive dissolution at pH 6 was significantly accelerated in the presence of EDT A, again with less ascorbate adsorbed in the pres-
96 ence of EDTA. EDTA can act both to acclerate the reductive dissolution through surface complex formation, and to prevent the possble inhibi-tion due to adsorption of iron(II) by complexing iron(II) in solution. The reductive dissolution was also accelerated in the presence of ortho-phosphate, presumably through formation of mono-nuclear surface chelate complexes as proposed for the kinetic effect of oxalate.
8. The presence of aluminum inhibited the reductive dissolution at pH 4 but not at pH 3. This inhibition coincided with the adsorption edge for aluminum on hematite (about pH 3.5). Aluminum would tend to form surface complexes preferentially at crystal discontinuities, thus decreasing the number of "active sites" for dissolution. The role of aluminum and iron(II) in inhibiting the reductive dissolution through adsorption is seen as an analogy to passive film formation in corrosive dissolution.
6.2 THE CYCLING OF IRON IN NATURAL WATERS Although many of the experiments described here were carried
out at conditions very different from those found in surface waters (highly crystalline solid phase, low pH, mixtures of only two or three reactants), the chemical processes observed are believed relevent to natural waters. The purpose of these model laboratory experiments is not to quantitatively predict the behaviour of natural systems, but to shed light on basic chemical process which may be occuring in nature. Such knowledge helps pose better hypotheses for field research and aids in the interpretation of data. Figure 6.1 shows a schematic dia-gram of iron cycling in a lake. A dominant feature is the change in pE with depth in a density stratified lake. The pE is controlled most strongly by microbial kinetics. In the photic zone 02 is produced as a consequence of photosynthesis. In the deeper waters within the hy-polimnion the reduced carbon in the detritus settling from the photic zone is oxidized by microbial processes where 02 acts as the prefered oxidant yielding C02. This microbially mediated oxidation leads to a depletion of oxygen in the hypolimnion. The dissolved oxygen profile is an important indicator of the redox conditions present within the water column.
97
photosynthesis
hv
0) ~ Fe(lll)(OH) Fe(ll)
~ g epllimnion ·;;; 02 .2 Fe(ll) ~ Fe(lll) oxide
;; the•moclloe c J ~ hypolimnion 5l ) v Fe(lll) oxide + organics
reductive dissolution
Figure 6.1 Schematic Diagram of Iron Transformations in a Stratified Lake.
The presence or relative absence of dissolved oxygen dictates the thermodynamic boundary conditions for the dissolution of iron(III) (hydr)oxides via a reductive pathway and also the oxidation of dis-solved iron(II) and subsequent precipitation of sparingly soluble iron(III) oxide solid phases. The following thermodynamic mass law describes the dependence of reduction-dissolution and oxidation-precipitation on the redox conditions present. For this reaction, the pE""O at neutral pH (Sulzberger et. al.,1988).
Fe+2 + 3Hz0 <=! Fe(OH)3(s) +3H+ + e- (6.1)
Cycling of iron in natural waters arises from changes in pE within the water column. The oxygen rich upper zone is a source of iron(III) oxide solid phases which settle downward and the anoxic bottom wa-ters are a source of dissolved iron(II) which diffuses upward. The cy-cling of iron and the transformation between soluble iron(II) and
98 sparingly soluble iron(III) solid phases is important because of the ability of the solid phases to adsorb nutrients (phosphate) or toxic metals and toxic organic compounds. Thus precipitation of solid phases is associated with the scavaging of pollutants and nutrients from the water column and dissolution is associated with a release of pollutants and nutrients into the water column. A small steady-state concentration of iron(Il) may exist in the photic zone due to photo-reduction and dissolution of solid phase iron(III) oxide particles. The concentration would depend upon the rate of the photo-reaction com-pared to the rate of the re-oxidation of the resulting iron(Il).
It has been shown in the previous chapters that analogs to micro-bial metabolic products (simple organic acids) can effect the rate of dissolution. Phosphate was shown also to have a kinetic effect at acid pH values. Aluminum acts to inhibit the reductive dissolution of hematite above pH 3.5. Phosphate and aluminum can both be re-garded as "environmental factors" arising from human activity. The problem of lake eutrophication as a result of excessive phosphate loading is well known (Stumm and Morgan,1981). Watershed acidifi-cation and subsequent mobilization of aluminum from minerals (due to accelerated weathering of aluminum containing mineral phases under increasing proton concentration) is seen as a result of accelerated oxi-dation of the reduced carbon reservoir by burning of fossil fuels for energy production (Schnoor and Stumm, 1987). These two environ-mental problems can be used here to help illustrate the relation of chemical kinetics to the system level biogeochemical cycles.
6.3 FROM CHEMICAL KINETICS TO SYSIBM LEVEL EFFECTS It is important to understand the impact of various environmental
factors such as phosphate and aluminum. Phosphate can lead to an in-crease in the rates of algal growth, and thus the primary productivity of an aquatic ecosystem. Phosphate may also lead to an acceleration of the rates of reductive dissolution. Adsorption of phosphate onto iron oxide particles and incorporation of dissolved phosphate by organisms into cell material followed by burial of these inorganic and organic particles in the lake sediments is seen as a negative feedback mecha-nism which helps reduce the impact of phosphate loading on an aquatic system. This is seen as the chief removal mechanism for
99 phosphate within a lake. If phosphate acts to accelerate the reductive dissolution, however, the impact of this negative feedback on the sys-tem is reduced. In an extreme case, greatly accelerated reductive dis-solution would serve to keep adsorbed phosphate in the water column and act to concentrate phosphate within the system. This would be an example of phosphate acting as an environmental factor to create pos-itive feedback helping drive the system to a steady-state condition different from the orginal state of the system.
Aluminum was seen to inhibit the reductive dissolution of hematite above pH 3.5. Lake acidification and aluminum mobilization may possibly act to decrease the productivity of the aquatic system as well. In this case aluminum would serve as an environmental factor which acts an inhibitior on both the biological and geochemical kinetics in the system. Such an environmental factor ensures that changes in one part of the biogeochemical cycles will act as positive feedback on the other parts of the system. Reduced productivity acts to decrease the availability of the organic metabolytes thus slowing rates of dis-solution. Aluminum itself acts to inhibit the reductive dissolution which may then reduce the availability of nutrients to the biological community. Such effects create a cycle of positive feedback which would serve again to drive the system to a condition different from the original state. Such an "aluminum effect" would push the system to a more oligotrophic state.
It is not the intent of this final section to suggest that the experi-mental results from the previous chapters will serve to predict the future environmental state of surface waters which are acidified or subject to excessive phosphate loading. Rather it is intended to em-phasize, using the examples just presented, that the study of the ki-netics of chemical transformations in natural waters is a study of sys-tem evolution. Within a selected range of concentrations for environ-mental factors (pH, DOC, phosphate, aluminum) the condition of the system will not vary greatly. A significant change in one or several environmental factors may drive the system to a new state where the biogeochemical dynamics may serve to maintain the new state of the system rather than return the system to its original condition.
It is of utmost importance to understand the chemical and trans-port processes in natural aquatic systems which serve to amplify dis-
100 turbances. The hypothetical cases for phosphate and aluminum de-scribed above are both examples of chemical mechanisms which serve to amplify disturbances caused by environmental factors. Such ampli-fication mechanisms are the connection between how processes at the molecular scale may drive a system to a new state. The end result of such imposed changes in environmental factors may be a radically different sort of system than the one that initially felt the impact of the changes. Biological species critical to a particular pathway for an important chemical transformation may no longer exist. Certain seg-ments of the biogeochemical cycles may be greatly accelerated or in-hibited leading to a different steady-state composition of the chemical millieu and the biological community. The system may evolve to a very different state. This new state will be characterized by the chemical water quality of the aquatic system and will certainly be re-flected at the macroscopic level by what sort of biological community the system supports, including human activity.
6.4 FUTURE ENVIRONMENTS Buffering factors which provide negative-feedback in response to
changes in environmental factors will help a system maintain its origi-nal state. Understanding such buffering processes is important but there is also an urgent need to consider where an aquatic system may be heading when changes in environmental factors are sufficiently large or sudden so that the buffering mechanisms cannot maintain the original state of the system. Environmental problems such as water-shed acidification, lake eutrophication, large scale ocean disposal of municipal waste, and global climate change have given great impetus to the study of natural aquatic systems. This brief list of problems caused by human activity leaves no doubt that we are changing our environment. The goal of aquatic chemistry in the face of such prob-lems goes beyond understanding the processes which maintain the state of aquatic environments. It is necessary, as well, to seek an un-derstanding of the processes that are leading to a different future en-vironment; which are creating change within the system. Hopefully a better understanding of the pathways to future environments will help us to make better decisions today concerning the impact of our activity on the present environment.
101
Appendix 1. Adsorption Isotherms
The first section is a discussion of the constant capacitance model for proton adsorption and its relation to the Frumkin-Fowler-Guggenheim adsorption isotherm and the Langmuir adsorption isotherm. The second section is a derivation of the Langmuir ad-sorption isotherm used to calculate the surface stability constants for ascorbate and oxalate. The third section is a discussion of the Langmuir competitive adsorption isotherm using ascorbate and ox-alate adsorption as an example. The Freundlich adsorption isotherm is discussed in the final section. A general review of the acid/base chemistry of metal oxides is provided in Stumm and Morgan (1981). The discussion of adsorption isotherms for protons is taken largely from Wieland et. al. (1988).
1. The constant capacitance model for the adsorption of protons on a metal oxide surface.
A common approach to the formulation of acidity constants for the surface acid/base equilibria of metal oxide surfaces is to separate the total free energy of protonation into a chemical contribution and an electrostatic contribution.
(i)
~ 1 (intr) is the first intrinsic surface acidity constant and corre-
sponds to the chemical thermodynamic constant for the dissociation
reaction. K: 1 is a conditional constant and depends upon the surface
potential as well as the the intrinsic acidity constant. The exponential term is the Holtzman factor for the electrostatic energy required to remove a proton from the surface into the bulk solution. F (96,400 C mol-1) is the Faraday constant, R (8.314 J mol-l K-1) is the ideal gas constant, K is temperature Kelvin, and '¥ (V) is the potential difference between the oxide surface and the solution. The constant capacitance
102
model assumes that surface potential is proportional to surface charge, 'P=Co (C is the capacitance, F m-2; o is the surface charge, C m-2). If adsorbed protons are the charge-determining species, then the electrostatic term can be written as a function of surface proton concentration (mol m-2). Here the surface proton concentration has been written as the concentration of proton-ated iron(Ill) sites ( o=F { >FeIIIOH2+} ).
(ii)
s It can be seen that Ka 1 changes with the surface concentration of
protons. If the mass law for the dissociation reaction (>Felllo H 2 + +:! >FellloH + H+) is substituted into (ii) then the following expression re-sults. Curved brackets, { }, refer to surface concentration (mol m-2) and square brackets, [ ], to solution concentration (mol 1-1 ).
(iii)
A mass balance for surface sites which assumes that the concentration of fully deprotonated sites is negligible ({>FelII0-}=0) can be used to write an equation expressed only in terms of {>FelllOH2+} and [H+].
Such an assumption corresponds to pH values below the pK: 1 (in tr)
where the predominant surface species are >FellloH2+ and >FeillQH. Sm (mol m-2) is the total number of surface sites.
(mol m-2) (iv)
This mass balance is solved for {>FellIQH} and substituted into the mass law.
103 (Sm-(>FeIIIOH2+})[H+] s . ( F2 )
(>FeIIIOH2
+} = Kal (mtr) exp CRT(>FeIIIOH2+} (v)
Solving this equation for [H+] gives the Frumkin-Fowler-Guggenheim isotherm for proton adsorption.
(mo11-l)
(vi)
If the effects of surface charge are neglected (the exponential term is set equal to unity), then the FFG isotherm collapses to the Langmuir adsorption isotherm.
Sm[H+] (>FeIIIOH2+} = ---==----
~l (intr) + [H+]
(mo11-l) (vii)
(mol m-2) (viii)
A basic assumption in the Langmuir adsorption isotherm is that the energy of adsorption does not change with surface coverage of ad-sorbing species. By neglecting the changes in electrostatic energy due to changes in surface charge arising from protonation of the surface, the constant capacitance model for proton adsorption (equivalent to the FFG isotherm) simplifies to the Langmuir isotherm.
2. The Langmuir adsorption isotherm for ascorbate Adsorption of ascorbate from solution can be written as a surface
complex formation reaction. The protonated surface iron(Ill) sites and the hydrogen ascorbate ion are written here as the reacting species.
SKsur (ix)
104 s _ {>FelIIHA} Ksur - {>FelIIOHz+}[HA+] (l mo1- l) (x)
It was mentioned in the previous section that one of the assumptions implicit in the Langmuir adsorption isotherm is that the energy of ad-sorption does not change with surface concentration of the adsorbing species. This is reflected in the mass law written above by omitting a correction term such as the Holtman factor for changes in surface po-tential due to changes in surface charge upon adsorption of charged species.
Another assumption used in deriving the Langmuir adsorption equation is that the surface is homogeneous, the energy of adsorption is equal at all surface sites. An oxide surface, in the absence of charge-determining species other than protons, can be described by three types of surface sites with different degrees of protonation; >FelIIOHz+, >FellloH, >FeIIIo-. Because the charge for each type of site is different, the electrostatic contribution to the adsorption energy is different for each type of site. If the effects of surface charge on the adsorption energy are neglected, then the adsorption energy is the same for each site and the surface is treated as being homogeneous. The concentration of "unoccupied" sites can be grouped into one term regardless of the charge on the surface site. In this case the mass bal-ance for surface sites is simplified. Here the concentration of unoccupied surface sites is written as {>FelIIOHz+} in order to be consistent with the mass law written above.
(mol m-2) (xi)
This can be used to write the mass law in terms of (>FelllHA} and [HA·].
SK S [HA-] {>FeIIIHA} = sur m
1 + SKsurfHA-] (mol m-2) (xii)
If the numerator and denominator are multiplied by (SKsur)-1, then this isotherm takes the familiar form of the Langmuir adsorption
-1 isotherm; KL=S~ur .
S [HA-] {>FeIIIHA) =_m...._ __ KL+ [HA-]
105
(mol m-2) (xiii)
The Langmuir adsorption constant, KL (mol 1- l ), has the same units as an acidity (dissociation) constant. Because the surface complex for-mation equilibrium was written as an association reaction, the result-ing Langmuir constant is the reciprocal of the surface stability con-
-1 1 stant, KL=s~ur (mol 1- ).
3. The Lani:muir competitive adsorption isotherm This isotherm is also referred to as the Langmuir multi-compo-
nent isotherm. It is used to model adsorption from solutions contain-ing several adsorbing species. Here it is applied to formation of sur-face complexes by adsorption from solutions containing more than one ligand. The basic assumption of surface homogeneity is still used. As stated in the previous section, this is equal to neglecting the effects of surface charge arising from surface sites with differing degrees of protonation. An assumption in the competitive adsorption isotherm is that only one species can adsorb at a single site. This is formulated by including the concentration of all adsorbing species in the mass bal-ances for the individual species. Solution species which form surface complexes of different stabilities "compete" for available surface sites. The stability constant and adsorption maximum for each surface com-plex can be estimated from the individual adsorption isotherms for each species. These equations have been written for surface ascorbate and oxalate species with hydrogen ascorbate and hydrogen oxalate as the reacting dissolved species. The following equations can be derived in a similar way as was done for the Langmuir adsorption isotherm by combining the mass balances for ascorbate and oxalate with the ther-modynamic mass laws. The subscripts HA and Ox refer to ascorbate and oxalate respectively.
106 S HASK HA[HA-l {>FeIIIHA} = m sur
1 + SKsurHA[HA-l + SKsurox[HOx-1 (mol m-2) (xiv)
III SmoxSKsurox[HOx-1 {>Fe Ox}=
1 + SKsurHA[HA-l + SKsuroxlHOx-1 (mol m-2) (xv)
These equations are used here as an example of how adsorption data from isotherms of individual species can be incorporated to describe adsorption in more complex systems. The surface concentration of the respective complexes, the surface stability constants, the adsorption maxima, and the reacting solution species are defined the same as for the individual adsorption isotherms for ascorbate and oxalate on hematite. The stability consants and adsorption maxima are listed in Table I.I
Table I.I Surface Stability Constants and Adsorption Maxima Estimated from the Adsorption Isotherms for Ascorbate and Oxalate
Ascorbate Surface Stability Constant 9. 6x l o5 l mol- l
Adsorption Maximum 1.01x1 o-6 mol m-2
Oxalate 2.lxl07 l mo1-l
2.lxlo-6 mol m-2
Surface concentrations of ascorbate and oxalate were experimen-tally determined for a series of dissolution experiments in the pres-ence of oxalate and ascorbate together. The solution concentrations of ascorbate and oxalate determined in these experiments were used to calculate surface concentrations from the adsorption competition isotherms above. The solution concentrations were analytically de-termined as total concentrations (CT, mol 1-1) and the concentration of hydrogen ascorbate and hydrogen oxalate must be calculated from the total dissolved concentration using the distribution coefficients,
107 (HA-J [HOx-J
<X}HA=cTHA; a.1ox= Crox. . The concentrations of [HA-] and [HOx-1 for
the experiments at pH 4 (figure 4.lOb) were put into the isotherms above, along with the surface stability constants and adsorption maxima, to calculate surface concentrations of ascorbate and oxalate. These calculated values are plotted in figure I.I against the solution concentration of hydrogen oxalate along with the experimentally de-termined surface concentrations of ascorbate and oxalate. The nu-merical values used in the calculations are listed in Table I.II
- 25 )(
0 {>Fe(lll)Ox} :::::::- 20 • :::,. N -Q) E • LL 15 • " - 0
....:.,...E <( • 10. :::c --Q) LL
" -0 ..-- {>Fe(lll)HA}
5 • 0
0 2 4 6 8 1 0 12 -1 -6 1
[HOx ] (10 mol I ) Figure I.I Adsorption competition between ascorbate and oxalate. The open symbols are the values calculated from the Langmuir competitive adsorption isotherm. The filled symbols are the corresponding experimental points. 10-4 mol /·1 added ascorbate, pH 4, hematite=0.613 g 1-1. a10x=0.33 (pH 4, l=0.01 mol 1-1 ).
108 Table I.II Numerical Values Used to Calculate the Surface
Concentrations of Ascorbate and Oxalate
[HA-] [HOx-] {>FeIIIOx} {>FelIIHA} {>FeIIIOx} {>FeIIIHA} calculated
--oo-6 mol 1- l )-- ------------(1o-7 44.0 0 0 47.4 0.6 4.4 49.1 1.1 6.6 49.7 1.7 8.6 50.3 4.31 13.3 51.2 9.97 16.7
a1HA=0.46 (pH 4, 1=0.01 mot 1-1) a10x=0.33 (pH 4, I=0.01 mol 1-1) [H2A1T=I0-4 mo11-l
l 0.1 7.9 6.9 5.9 3.7 2.0
measured mol m-2)------------
0 12.7 8.1 11.0
11.6 6.8 14.7 6.4 16.9 4.91 19.7 3.9
The Langmuir competitve adsorption isotherm only qualitatively models the adsorption of ascorbate and oxalate together. Adsorbing oxalate causes ascorbate to be displaced from the hematite surface. For the competitve adsorption isotherm, this preferential adsorption of oxalate is explained by the larger stability constant for the surface iron(Ill)-oxalate complex. Formation of surface oxalate complexes would block surface sites from formation of ascorbate complexes.
4. The Freundlich adsorption isotherm The Freundlich adsorption isotherm is an empirical relation be-
tween surface concentration and dissolved concentration of an ad-sorbing species. The adsorption of oxalate is used here as an example.
{>FeIIIOx} = Kp[HOx·]m (mol m-2) (xvi)
The value of the exponent, m, depends upon the system being studied. The usefulness of the Freundlich isotherm lies in the fact that adsorp-tion data can frequently be linearized by a log-log transformation which corresponds to the transformed Freundlich isotherm.
109 log(>FeIIIOx} = logKF + mlog[HOx-] (xvii)
One application of this isotherm is the interpretation of kinetic rate laws for surface controlled reactions which are written in terms of solution species.
RATE = k[HOx-]n (mol m-2 hr-1) (xviii)
Solving the Freundlich isotherm for [HOx-J and substituting this di-rectly into equation (ixx) allows the rate law to be written in terms of the reacting surface species.
(mol m-2 hr-1) (xix)
Because the Freundlich exponent, m, is often less than one, fractional-order kinetics (n<l) for rate laws expressed in solution concentrations of reacting species is an indication that a surface controlled reaction is taking place. For example, if RATE=k[HOx-J0.3 were experimentally observed for a dissolution reaction and m=0.3 was found for the Freundlich adsorption isotherm for oxalate on the reacting surface, then the rate law can be interpreted as being first-order in {>Felllox}.
110
Appendix 2. Summary of Experimental Data for Dissolution and Adsorption
Experiments
DISSOLUTION EXPERIMENTS
1. Ascorbate, pH=3, hematite=0.613 g 1- l [H2AlT RAIB
oo-6 mo11-l) oo-7 mol m-2 hr-1) 0 0.0125 0 0.0174 10 0.717 10 0.760 50 1.22 50 1.22 50 1.12 100 1.58 100 1.48 100 1.30 500 2.13 500 2.21 500 2.22
2. Ascorbate, pH=4, hematite=0.613 g 1- l [H2A]T RAIB
oo-6 mo11-l) oo-7 mol m-2 hr-1) 10 0.190 30 100 250
0.282 0.562 1.02
(>FeIIIHA} oo-7 mol m-2)
0 0
4.19 4.19 9.69 9.69 10.3 15.3 12.6 12.3 19.0 18.5 18.4
(>FeIIIHA} oo-7 mol m-2)
4.26 9.42 12.7 22.8
111
3. Ascorbate, 5.0xI0-5 mol 1-l oxalate, pH=3, hematite=0.613 g 1- l [H2A]T RATE (>FeIIIHA}
(I0-6 mol I-2) (I0-7 mol m-2 hrl) (I0-7 mol m-2) 0 0.211 0 0 0.205 0 25 25 50 50 100 100
2.63 2.88 3.85 4.69 6.10 5.97
1.96 1.58 2.42 1.96 5.22 4.10
4. Ascorbate, 5.0xI0-5 mol 1-l Oxalate, pH=4, hematite=0.613 g 1- l [H2A]T RATE (>FeIIIHA}
(I0-6 moI I-1) (I0-7 mol m-2 hrl) (I0-7 mol m-2) 10 0.286 1.86 20 0.428 4.29 30 0.810 2.42 50 0.820 4.57 60 1.10 5.97 100 1.15 3.92 100 1.37 4.66
5. Oxalate, 10-4 mol t-1 Ascorbate, pH=3, hematite=0.613 g 1- l [H20x]T RATE (>FeIIIHA) (>FeIIIOx)
oo-6 mol 1-2) (I0-7 mol m-2 hrl) oo-7 mol m-2) 0 1.58 15.3 0 0 1.48 12.6 0 0 1.29 12.3 0 10 2.08 5.89 9.04 10 2.16 7.09 9.04 20 4.34 5.45 15.8 20 3.45 3.88 16.0 50 6.10 5.17 . 19.6 50 5.96 4.10 19.6
112
6. Oxalate, 10-4 mol 1- l Ascorbate, pH=4, hematite=0.613 g 1- l [H20x]T RAIB (>FeIIIHA} (>FeIIIOx}
(I0-6 mol I-l)(I0-7 mol m-2 hr-1) (I0-7 mol m-2) 0 0.562 12.7 0 10 0.583 11.0 8.11 1 5 0.629 6.82 11.6 20 0.749 6.43 14.7 30 0.951 4.91 16.9 50 1.14 3.90 19.7 50 1.37 4.64 10.5
7. Ascorbate and Oxalate, pH=2.5, hematite=0.613 g 1- l [H2A]T [H20x]T RA1E (>FeIIIHA} (>FeIIIOx}
(lo-6 mo11-l) (lo-7 mol m-2 hr-1) (lo-7 mol m-2) 100 0 1.84 3.92 0 100 50 3.53 <l 21.6
8. Ascorbate and Oxalate or EDTA, pH=6, hematite=0.613 g 1-1 (>FeIIIHA}
(IO-7 mol m-2) 24.2
Chelate [ligand]T [H2A]T RAIB
none oxalate EDTA EDTA
oo-6 mol 1-l) oo-7 mol m-2 hr 1) 0 250 <0.02
100 100 100
250 0
250
<0.01 0.154 0.576
11.8 0
5.97
9. Phosphate and Ascorbate, pH=3 and 4, hematite=0.613 g 1-1 [H2A]T [H3P04]T pH RAIB (>FeIIIHA}
(lo-6 mol 1- l) (IO-7 mol m-2 hr-I) (IO-7 mol m-2) 0 100 3 0 0
100 100 3 2.44 <l 100 100 4 1.82 <l
113 10. 2.5x1Q-4 mol 1-l Aluminum,I0-4 mol 1-I Ascorbate, hematite=0.613 g 1- I
pH
3 4
ADSORPDONIS011IERMS
RATE oo-7 mol m-2 hr-1)
1.67 0.252
{>FeIIIHAJ oo-7 mol m-2)
7.56 15.8
11. Ascorbate Adsorption on Hematite, pH 3, hematite=0.613 g 1- l [H2AlT [Fe+2Jsol,t=lhr [H2AJsol,t=lhr {>FelIIHA}
(to-6 mol l-1) oo-7 mol m-2) 10 7.9 3.7 2.19 20 9.5 10.1 4.80 30 10.0 19.1 5.50 50 6.9 39.4 6.66 70 10.5 56.8 7.41 100 3.8 89.5 8.01 150 9.9 134 10.3 250 16. l 228 13.0
12. Ascorbate Adsorption on o-Al203, pH 3, o-Al203=2.0 g 1- l [H2AJT [H2Alsol,t=l hr {>AlHA}
oo-6 mol 1-l) oo-6 mo11-l) oo-9 mol m-2)
10 3.3 2.68 20 7.9 4.84 30 13.9 6.44 50 29.4 8.24 100 68.3 12.6 250 197 21.2 500 444 22.4
114
13. Oxalate Adsorption on Hematite, pH 3, hematite=0.576 g I-1 [H20x]T [H20xlsol,t=lhr (>FeIIIOx}
(10-6 mol I-1) (lQ-6 mol I-1) (I0-7 mol m-2) 10 0.03 7 .10 20 0.51 13.9 30 3.39 18.9 40 9.25 21.9 50 17.4 23.0 100 63.8 23.6 200 163 25.2
14. Aluminum Adsorption on Hematite, [AI+3]T=2.5x1Q-5 mol 1-l, hematite=0.613 g i-1
pH [Al+3lsol {AI+3lsur (10-6 mol i-1) (IQ-7 mol m-2)
2.50 26.9 0 3.03 24.9 0.09 3 .51 25.9 0 4.03 22.1 2.7 4.52 15.4 8.9
EXPERIMENT AL DATA NOT REPORTED IN CHAPTERS 15. Various Hematite Concentrations in the presence of 10-4 mol i-1 ascorbate, pH=3
RATE hematite [H2Alsur (l0-6 mol i-1) (g i-1) (lQ-6 mol I-1)
0.603 13.6 0.613 0.850 35.5 1.23 1.25 46.5 1.84
115 16. Various Hematite Concentrations in the Presence of 10-4 mol 1- l Ascorbate and 5.0xI0-5 moI J-1 Oxalate, pH=3
RA'IE hematite {>FelIIHA} {>FeIIIOx} oo-8 mol m-2 hr-1) (g t-1) oo-7 mot m-2)
11.5 0.613 3.91 19.7 13.7 0.613 4.66 20.5 5.89 1.23 7.97 16.8 4.94 3.43 5.19
1.84 2.45 3.06
7.99 9.14 9.69
13.9 11.3 9.17
116
LIST OF REFERENCES
Ashida M., Sasaki M., Hachiya K., and Yasunaga T.(1980). Kinetics of adsorption-desorption of OH- on Ti02-H20 interface by means of pressure-jump technique. J. Colloid Int. Sci.1A;l_,512
Astumian R., Sasaki M., Yasunaga T., and Schelly Z.(1981). Proton adsorption-desorption kinetics on iron oxides in aqueous suspensions, using the pressure-jump method. J. Phys. Chem . .8.,i,3832
Basolo F. and Pearson R.(1967). Mechanisms of Inorganic Reactions; a Study of Metal Complexes in Solution. Wiley, New York
Bielski B.(1982). Chemistry of ascorbic acid radicals in: Ascorbic Acid: Chemistry, Metabolism, and Uses. American Chemical Society, Washington, D.C.
Blesa M., Borghi E., Maroto A., and Regazzoni A.(1984). Adsorption of EDT A and iron-EDT A complexes on magnetite and the mechanism of dissolution of magnetite by EDTA. J. Colloid Int. sci . .2.8.;l..2 9 5
Blesa M. and Maroto A.(1986). Dissolution of metal oxides. Journal de Chimie PhysicqueU, 7 57
Blesa M., Marinovich H., Baumgartner E., and Maroto A.(1987). Mechanism of dissolution of magnetite by oxalic acid-ferrous ion solution. Inorg. Chem.22_,3113
Blum A. and Lasaga A.(1987). Monte-Carlo simulation of surface reaction rate laws in: Aquatic Surface Chemistry (ed. W. Stumm). Wiley-Interscience, New York
Blum A. and Lasaga A.(1988). Role of surface speciation in the low-temperature dissolution of minerals. Naturelll..,431
Carroll-Webb S. and Walther J.(1988). A surface complex reaction model for the pH-dependence of corundum and kaolinite reaction rates. Geochim. Cosmochim. Acta~,2609
Chang H. and Matijevic E.(1983). Interactions of metal hydrous oxides with chelating agents IV. dissolution of hematite. J. Colloid Int. Sci.~419
Chou L. and Wollast R.(1985). Steady-state kinetics and dissolution mechanisms of albite. Am. J. Sci.lli,963
117 Cornell R., Posner A., and Quirk J.(1976). Kinetics and mechanisms
of the acid dissolution of goethite (a-FeOOH). J. Inorg. Nucl. Chem.lS.,563
Cornell R. and Schindler P.(1987). Photochemical dissolution of goethite in acid/oxalate solution. Clays and Clay MineralsJ.5..;.i,34 7
Davison W.(1985). Conceptual models for Transport at a redox boundary in: Chemical Processes in Lakes (ed. W. Stumm), Wiley-Interscience, New York
Deer W., Howie R., and Zussman J.(1980). An Introduction to the Rock-Forming Minerals. Longman, London
Dos Santos Afonso M., Morando P., and Blesa M., and Banwart S. and Stumm W.(1988). The reductive dissolution of iron oxides by ascorbate. the role of carboxylate anions in accelerating reductive dissolution. J. Colloid Int. Sci., submitted for publication
Dougan W. and Wilson A.(1974). The absorptiometric determination of aluminum in water. a comparison of some chromogenic reagents and the development of an improved method. A nalyst22.,,413
Faust B.(1985). PhD Thesis: Photo-Induced Reductive Dissolution of Hematite (a-Fe203) by S(IV) Oxyanion. Caltech, Pasadena
Fischer W.(1972). die Wirkung von zweiwertigem Eisen auf Losung und Umwandlung von Eisen(III)-hydroxiden in: Genesis and Use of Hydromorphic Soils (eds. E. Schlichtung and U. Schwetmann). Verlag Chemie, Weinheim
Furrer G.(1985). Die oberflachenkontrollierte Auflosung von Metalloxiden: ein koordinationschemischer Ansatz zur Verwitterungskinetic. Ph.D. Thesis-number 7737, ETH-Zurich
Furrer G. and Stumm W.(1986). The coordination chemistry of weathering: I. dissolution kinetics of o-Al203 and BeO. Geochim. Cosmochim. Acta.iQ.,1841
Gorichev I. and Kipriyanov A.(1984) Regular kinetic features of the dissolution of metal oxides in acidic media. Russian Chemical Reviewsll, 1039
118 Grauer R. and Stumm W.(1982). Die K<>ordinationschemie
ox.idischer Grenzflachen und ihre Auswirkung auf die Auflosunguskinetic oxidischer Festphasen in wasserig Losungen. Colloid Polymer. Sci.2,fill,959
Hair M.(1967). Infrared Spectroscopy in Surface Chemistry. Marcel Dekker, New York
Hachiya K., Sasaki M., Saruta Y., Mikami N., and Yasunaga T.(1984). Static and kinetic studies of adsorption-desorption of metal ions on a y-Al203 surface 1. static study of adsorption-desorption. /. Phys. Chem . .8.8.,23
Hachiya K., Sasaki M., Saruta Y., Mikami N., and Yasunaga T.(1984). Static and kinetic studies of adsorption-desorption of metal ions on a y-Al203 surface 2. kinetic study by means of pressure-jump technique. /. Phys. Chem . .8.8.,21
Hayes K. and Leckie J.(1986). Mechanism of lead ion adsorption at the goethite-water interface in: Geochemical Processes at Mineral Surfaces (eds. J. Davis and K. Hayes). ACS, Washington
Hoigne J.(1989). Formulation and calibration of environmental reaction kinetics; oxidation by aqueous photooxidation as an ex.amplification in: Aquatic Chemical Kinetics (ed. W. Stumm), in preparation. Wiley-Interscience, New York
Ikeda T., Sasaki M., Hachiya K., Astumian R., Yasunaga T., and Schelly Z.(1982). Adsorption-desorption kinetics of acetic acid on silica-alumina particles in aqueous suspensions, using the pressure-jump relaxation method. /. Phys. Chem.,8..6.,3 861
Khan M. and Martell A.(l 967). Metal ion and metal chelate catalyzed oxidation of ascorbic acid by molecular oxygen. I. cupric and ferric ion catalyzed oxidation. /. Am. Chem. soc . .8.2.;12,, 4176
Khan M. and Martell A.(1967). Metal ion and metal chelate catalyzed oxidation of ascorbic acid by molecular oxygen. II. cupric and ferric chelate catalyzed oxidation. /. Am. Chem. Soc.~.7104
Kummert R.(1979). Die Oberfliichenkomplexbildung von organischen Siiuren mit Gamma-A/uminumoxid and ihre Bedeutung fur naturliche Gewiisser. Ph.D. Thesis-number 6371, ETH-Zurich
119 Kummert R. and Stumm W.(1980) The surface comlexation of
organic acids on hydrous y-Al203. J. Colloid Int. Sci.,U,;2,373 LaKind J. and Stone A.(1988). Reductive Dissolution of Goethite by
Phenolic Reductants. Geochim. Cosmochim. Acta.il,96 l Lewin J. and Chen C.(1976). Available iron: a limiting factor for
marine phytoplankton. Limnol. Oceanog.1..6.,610 Matijevic E. and Scheiner P.(1978). Ferric Hydrous Oxide Sols III.
preparation of uniform particles by hydrolysis of Fe(IIl)-chloride, -nitrate, and -perchlorate solutions. J. Colloid Int. sci . .2.l;_l, 5 09
McKeague J. and Day J.(1966). Dithionite and oxalate-extractable Fe and Al as aids in differentiating various classes of soils. Can. J. Soil Sci.!Q., 13
Mulvaney P., Cooper R., Grieser F., and Meisel D.(1988). Charge trapping in the reductive dissolution of colloidal suspensions of iron(III) oxides. Langmuir!, 1206
Murphy T. and Lean D.(1976). Blue-green algae: their excretion of iron-selective chelators enables them to dominate other algae. Sciencefil,900
Neilands J.(1967). Hydroxamic acids in nature. Science ill, 1443 Neilands J.(1981). Microbial iron compounds. Ann. Rev.
Biochem.~,115
Neilands J .( 1982). Microbial envelope proteins related to iron. Ann. Rev. Microbiol.J.§.,285
Neilands J.(1984). Siderophores of bacteria and fungi. Microbiological Sciencesl,9
Penners N. and Koopal L.(1986). Preparation and optical properties of homodisperse haematite hydrosols. Colloids Surfacesl!l.331
Pulfer K., Schindler P., Westall J., Grauer R.(1984). Kinetics and mechanism of dissolution of bayerite (y-Al(OH)3) in HN03-HF solutions at 298.2K. J. Colloid Int. Sci.lfil.;.Z,554
Rubio J. and Matijevic E.(1979). Interactions of metal hydrous oxides with chelating agents 1. B-FeOOH-EDTA. J. Colloid Int. Sci.2.R..;l,408
120 Sasaki M., Moriya M., Yasunaga T., and Astumian R.(1983). A
kinetic study of ion-pair formation on the surface of a-FeOOH in aqueous suspensions using the electric field pulse technique. J. Phys. Chem.ll,1449
Schott J.(1989). Modelling of the dissolution of strained and unstrained multiple oxides: surface speciation approach in: Aquatic Chemical Kinetics (ed. W. Stumm), in preparation, Wiley-Interscience, New York
Schindler P. and Stumm W.(1987). The surface chemistry of oxides, hydroxides, and oxide minerals in: Aquatic Surface Chemistry (ed. W. Stumm). Wiley-Interscience, New York
Schnoor J. and Stumm W.(1985). Acidification of aquatic and terrestrial systems in: Chemical Processes in Lakes (ed. W. Stumm). Wiley-Interscience, New York
Schwertmann U .(1964 ). Differenzierung der Eisenoxide des Bodens durch Extraktion mit Ammoniumoxalat-Losung. Z. Pflanzenernaehr. Bodenk.1.fli, 194
Schwertmann U. and Taylor R.(1977). Iron oxides in: Minerals in Soil Environments (eds. J. Dixon and S. Weed). Soil Science Society of America, Madison
Segal M. and Sellers R.(1982). Kinetics of metal oxide dissolution. reductive dissolution of nickel ferrite by tris(picolinato)vanadium(II). J. Chem. Soc., Faraday Trans.1..£,1149
Segal M. and Sellers R.(1984). Redox reactions at solid-liquid interfaces. Advances in Inorganic and Bioinorganic Mechanisms. l!JJL.L.91
Sherman D.(1985). The electronic structures of Fe+3 coordination sites in iron oxides; applications to spectra, bonding, and magnetism. Phys. Chem. MineralsU, l 6 l
Siffert C.(1989). L'effet de la Lumiere sur la Dissolution des Oxydes de fer(lll) dans les Milieux Aqueux. PhD. Thesis-number 8852, ETH-Zurich
Smith R. and Martell A.(1976). Critical Stability Constants, .Y.Q.l....l. Plenum Press
Sposito G.(1983). On the surface complexation model of the oxide-aqueous solution interface. J. Colloid Int. Sci.21.i.2.,329
121 Steenken S. and Neta P.(1979). Electron transfer rates and
equilibria between substituted phenoxide ions and phenoxy radicals. J. Phys. Chem . .al.;.2., 1134
Stevenson F.(1967). Organic acids in soil in: Soil Biochemistry (eds. A. McLaren and G. Peterson), Marcel Dekker, New York
Stone A.(1986). Adsorption of Organic Reductants and Subsequent Electron Transfer on Metal Oxide Surfaces in; Geochemical Processes at Mineral Surfaces (eds. J. Davis and K. Hayes), American Chemical Society, Washington
Stone A. and Morgan J.(1984). Reduction and dissolution of manganese(III) and manganese(IV) oxides by organics. 1. reaction with hydroquinone. Environ. Sci. Technol.1.8.,450
Stone A. and Morgan J.(1984). Reduction and dissolution of manganese(III) and manganese(IV) oxides by organics: 2. survey of the reactivity of organics. Environ. Sci. Technol.1.8.,61 7
Stumm W. and Furrer G.(1987). The dissolution of oxides and aluminum silicates; examples of surface-coordination-controlled kinetics in: Aquatic Surface Chemistry (ed. W. Stumm), Wiley-Interscience, New York
Stumm W. and Morgan J.(1981) Aquatic Chemistry, Wiley-Interscience,New York
Sulzberger B., Suter D., Siffert C., Banwart S., and Stumm W. ( 1989). Dissolution of Fe(III) (hydr)oxides in natural waters; laboratory assessment on the kinetics controlled by surface speciation. Marine Chemistry, in press
Surana V. and Warren 1.(1969). The leaching of goethite. Trans. Inst. Min. Metall.1..8.,C 133
Suter D.(1989). Die reduktiv katalysierte Aufliisung von Goethit. Ph.D. Thesis-number 8933, ETH-Zilrich
Suter D., Siffert C., Sulzberger B., and Stumm W.(1988). Catalytic dissolution of iron(III)(hydr)oxides by oxalic acid in the presence of iron(II). Naturwissenschaftenll,511
Tamura H., Goto K., Yotsuyanagi T., Nagayama M.(1974). Talentall_,314
Taube H.(1968). Mechanisms of oxidation-reduction reactions. J. Chem. Ed.~,452
122 Valverde N.(1976). Investigations on the rate of dissolution of
metal oxides in acidic solutions with additions of redox couples and complexing agents. Ber. Bunsenges. Physik. C hem . .8.Q.;,±,333
Valverde N. and Wagner C.(1976). Considerations on the kinetics and the mechanism of the dissolution of metal oxides in acidic solutions. Ber. Bunsenges. Physik. Chem.8.Q.,330
Westall J. and Hohl H.(1980). A comparison of electrostatic models for the oxide/solution interface in: Advances in Colloid and Interface Sciencell,265
Wehrli B.(1989). Redox reactions of metal ions at mineral surfaces in: Aquatic Chemical Kinetics (ed. W. Stumm), in preparation. Wiley-lnterscience, New York
Wieland E.(1988). Die Verwitterung schwerlOslicher Mineralien-ein koordinationschemischer Ansatz zur Beschreibung der Auflosungskinetik. Ph.D. Thesis-number 8532, ETH-Zurich
Wieland E., Wehrli B., and Stumm W.(1988). The coordination chemistry of weathering: III. a generalization on the dissolution rates of minerals. Geochim. Cosmochim. Acta.51., 1969
Zinder B.(1985). PhosphatrucklOsung aus Sedimenten als Folge der Reduktion von Eisenoxides. Ph.D. Thesis-number 7738, ETH-Ziirich
Zinder B., Furrer G., and Stumm W.(1986). The coordination chemistry of weathering: II. dissolution of Fe(III) oxides. Geochim. Cosmochim. ActaiQ., 1861
Zutic V. and Stumm W.(1984). Effect of organic acids and flouride on the dissolution kinetics of hydrous alumina. model using the rotating disc electrode. Geochim. Cosmochim. Acta~, 1493
CURRICULUM VITAE
I was born on July 30, 1959 in Des Moines County, Iowa, USA. I attended primary and secondary school at the Mediapolis Community School and entered the University of Iowa in 1977. I received my Bachelors Degree in Civil Engineering in 1981 and my Masters Degree in Civil and Environmental Engineering in 1983. I left the USA for Europe in 1983, spending my first year at the Von Karmen Institute for Fluid Mechanics in Rhode-St.-Genese, Belgium, then coming to the EA WAG in the summer of 1984 to begin my doctoral studies with Prof. Werner Stumm. Since being at the EA WAG I have participated in the Post-Graduate Course for Water Supply and Water Resource Protection, both as a student and later as a teaching assistent.
