IIIIIIIIIIIIIIIIIIIIIIIIII» - 32m Ir'w THEE. 1lJ£3éLfik. I 1 MkMQOn State 1_ University MICHIGAN STATE UNIVERQITYI l8 WAR 1111111111111 1111111111 1. 111111111 3 1293 00569 4512 This is to certify that the thesis entitled The Oxidation of Chromium by Manganese Oxide: the Nature and Controls of the Reaction presented by Michael James Takacs has been accepted towards fulfillment of the requirements for Masters degree in Geology [1 (‘1 //‘7/15/ Major professor/ Date 5/520/88 1 f 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution MSU LIBRARIES .—:_. RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. THE OXIDATION or CNROHIUN BY MANGANESE OXIDE: THE NAIURB AND CONTROLS or T3] REACTION By Michael James Takacs A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Geological Sciences 1988 ABSTRACT TEN OXIDATION OF CNRONIUN BY NANGANRSN OXIDE: TN! NATURE AND CONTROLS or T33 REACTION BY Michael James Takacs The trace metal chromium (Cr) occurs in aqueous systems in two oxidation states: the non-toxic trivalent species and the toxic, mobile hexavalent species. One mechanism by which Cr may be oxidized from the trivalent to the hexavalent state is through an oxidation-reduction reaction with manganese oxides. This research explored the general nature and chemical controls of Cr oxidation by a synthetic manganese oxide. The results of laboratory experiments conducted at pH 4 . 5 show the oxidation occurs through the reduction of surficial tetravalent manganese (Mn) ions. The reaction appears to be preceded by the adsorption of Cr onto the oxide surface and new surface sites for the reaction are generated by the reduction/dissolution of the Mn oxide surface. The reaction is inhibited at neutral pH's by the formation of Cr hydroxide but is otherwise unaffected by changing solution chemistry. The reaction also proceeded in mixed manganese-iron oxide systems. To my wife Audrey and to my parents, without whose support and patience I could not have maintained the will to see this project to its conclusion. ii I would like to acknowledge my fellow geochemists with whom I shared an office, a lab, in some cases a house and many fun and interesting times, both in academic life and across the street. In particular, Tim Wilson, John Marsh, Jim Tolbert and Rico (even though you're only a geophysicist) all helped contribute to the quality of my education and to the length of time it took to finish my thesis. I would like to thank my committee members Dr. Larson and Dr. Ellis for some good insights for my manuscript. And finally I would like to thank my instructor, advisor and friend Dr. David Long for helping me along throughout the duration of my Master's work. Dave, it was a privilege and a pleasure to work in your program and within the camaraderie of your group. iii TABLE OF CONTENTS LIST OF TABLES O O O O ....... O ..... O O O O O O O O OOOOOOOOOOOO Vi LIST OF rIGURES O O O OOOOOOO O O O O O O O O O O O O O O O O O O O O O O O O O O Vi i-X INTRODUCTION 1.0 GENERAL INTRODUCTION.........................1-5 2. 0 AQUEOUS GEOCHEMISTRY OF CHROMIUM.............6-14 3. 0 STATEMENT OF PURPOSE.................. ...... 3.1 Approach to the Problem........ ............ 15- -16 3. 2 The Goals of this Research.......... ..... ..17- -19 METHODS 1.0 GENERAL APPROACH AND LIMITATION....... ....... 20-21 2.0 PREPARATION AND CHARACTERIZATION OF Mn OXIDE 2.1 Introduction ........... ... ................. 22-23 2. 2 Preparation of Mno ........................23-24 2. 3 Characterization o MnO ...................24—27 2.4 Stability of the Mno Sgock Suspension.....27-35 3.0 PREPARATION AND CHARAC ERIZATION 0F Fe OXIDE 3.1 Introduction...............................35-36 3.2 Preparation of Hydrous Ferric Oxide........36-37 3.3 Characterization of Fe oxide...............37-38 GENERAL EXPERIMENTAL PROCEDURES Clean Techniques...........................38-40 Reaction Vessel.................. ....... ...40-42 General Experimental Procedures............42-43 Sample Filtrat'on......6...................43-45 Analysis of Cr + and Cr in Solution......45-53 Atomic Absorption Spectrophotometry........54 Experimental Reproducibility....... ...... ..54-55 4. 0 4 4. 4. 4 4 4 4 OOOO O \JO‘U’ID‘JNH RESULTS AND DISCUSSION 1.0 THE OXIDATION OF Cr(III) BY MnO 1.1 THE GENERAL NATURE or THE REAC ION.........56-63 1.2 THE STOICHIOMETRY OF THE REACTION. ....... ..64-71 1.3 THE EFFECT OF CHANGING REACTANT CONCENTRATIONS................... .......... 72-80 iv 1.4 KINETIC CONSIDERATIONS............. ...... ..80-89 1.5 SUMMARY AND CONCLUSIONS.............. ...... 89-93 2.0 THE EFFECT OF SOLUTION CHEMISTRY ON THE OXIDATION OF Cr(III) BY MN OXIDE INTRODUCTION...............................94 THE EFFECTS OF SOLUTION pH.................94-98 THE OXIDATION OF SOLID Cr-HYDROXIDE........98-103 THE EFFECT OF DISSOLVED OXYGEN.............103-104 THE EFFECT OF IONIC STRENGTH...............104-110 EFFECT OF THE NATURE OF THE ELECTROLYTE....110-116 THE EFFECT OF ADSORBED COPPER..............116-122 SUMMARY AND CONCLUSIONS....................122-124 mfimmbUNt-J NNNNNNNN O 3.0 THE COMPETITION BETWEEN MN AND FE OXIDES FOR Cr(III) 3.1 INTRODUCTION....................... ...... ..125-128 3.2 PROCEDURES...................... ........ ...128-134 3.3 RESULTS AND DISCUSSION........ .......... ...134-149 3.4 SUMMARY AND CONCLUSIONS............ ...... ..150-151 SUMMARY AND CONCLUSIONS 1.0 SUMMARY............................. ......... 152-154 2.0 DISCUSSION...................................154-159 3.0 ENVIRONMENTAL SIGNIFICANCE...................159-162 4.0 SUGGESTIONS FOR FURTHER RESEARCH.............162-163 APPENDIX I. PAST RESEARCH............. ...... . ...... 164-208 APPENDIX II. EXPERIMENTAL DATA............. ...... ..209-223 APPENDIX III. RATE DATA CALCULATIONS..... .......... 224-263 BIBLIOGRAPHY................................. ...... 264-271 TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE LIST OF TABLES REACTION ORDER WITH RESPECT TO MnOz............86 EXPERIMENTAL RATE CONSTANTS....................86 ADSORPTION OF Cr SPECIES BY Fe(OH)3 AT pH 405000OOOOOOOOOOOOOOOOOOOO OOOOOOOOOOOOOOO 13o SELECTIVE DESORPTION OF Cr6+ FROM 1 x 10'3M Fe(OH)3 BY RAISING pH To 10 FOR 45 MINUTES.....13O REPLICATE ANALYSIS OF THE DESORPTION OF Cr6+ FRgM Fe(OH)3 BY §AISING pH TO 10 + — - o - -3 cr _ 906 x 10 I Fe(OH)3 -' 1 x 10 M00000132 SELECTIVITY SEQUENCES OF ADSORBED IONS........177 SUMMARY OF MULTI-SORBATE ADSORPTION STUDIES...181 SUMMARY OF IONIC STRENGTH EFFECTS ON ADSORPTIONOOOOOOOOOOOOOOOOOOOOOOOOOOOOO OOOOO OO183 APPARENT DIFFERENCES IN ADSORPTION BEHAVIOR OF Mn AND Fe OXIDES BASED ON PAST RESEARCHOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO0187 PAST STUDIES OF ADSORPTION REACTION STOICHIOMETRYOOOOOOOOOOOOOOOOOOOOOOOOOO0.0.00.191 PAST RESEARCH ON THE ADSORPTION OF Cr(III)....197 PAST RESEARCH ON THE ADSORPTION OF Cr(VI).....200 PAST RESEARCH ON THE OXIDATION OF Cr3+ BY Mn OXIDESOOOOOOOOOOOOOOOOOOOOO OOOOOOOOOOOOO 202 vi Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 4. 5O 6O 7. 8. 9. 10. 11. 12. 13. 14. 15. LIST OF FIGURES Conceptual Model of Chromium Geochemistry in Oxic Aqueous Systems................ ....... 11 Schematic of Chromium/Mn-Fe Oxide Interactions Addressed in Present Study.. ..... 16 Effect of Oxide Aging and Sonic Vibration Pre-treatment on the OXidation Rate Of cr by "n02.00000.0000000000030 Concentration of Mn in Aliquots Of Mnoz Stock Solution Sampled Over Time..............33 Stability of MnOZ Suspension at Various pH's..................................34 Experimental Reaction Vessel.......... ........ 41 Efficiency of Unmodified Cr6+ Extraction Procedure from Martin and Riley (1982)........48 Efficiency of Modified (2 X) Cr6+ Extraction Procedure from Martin and Riley (1982)........49 Efficiency of Modified (3 X) Cr6+ Extraction Procedure from Martin and Riley (1982)........49 Reproducibility of Oxidation/Reduction Experiment............................ ....... 55 Oxidation of Cr3+ by MnOZ Under the Conditions of the Standard Oxidation ExperimentOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO0.0.59 Oxidation of Cr3+ in Molar Excess by MnOz....59 Stoichiometric Ratio of Reaction Products....66 Adsorption of Mn2+ by MnOz at Selected Metal/Oxide Ratios................... ........ 68 Effect of MnOz Loading on Cr Oxidation Rate.................. ........ .....73 vii Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. Effect Of MnOz Loading on MnO2 DiSSOIution RateOOOOOOOOOOOOOOOOOOOOOOOOOOOOO73 Amount of Cr Oxidized at 1 and 4 Minutes as a Function of MnOZ Loading................75 Amount of MnOZ Reduced at 1 and 4 Minutes as a Function of MnOZ Loading................75 Effect of Cr3+ Loading on Cr Oxidation Rate WithmnOZ in ExceSSOOOOOOOOOOOOOOOOOOOOO78 Amount of Cr Ogidized in 30 Seconds as a Function of Cr + Loading with Mnoz in Excess............................... ..... 78 Effect of Cr3+ Loading on Cr Oxidation Rate with Similar Reactant Concentrations.........79 Effect of Cr3+ Loading on MnO2 Dissolution Rate with Similar Reactant Concentrations....79 Typical Plot for Determination of Experimental Rate Constant...................88 Efgect of pH on the Oxidation Rate of cr+byMHOZ.......................... ....... 97 Effect of pH on the Dissolution Rate Of MnOZ During the Oxidation Of Cr +...... ...... 97 Efgect of pH on the Oxidation Rate Of cr+by”n02......................... oooooooo 99 Effect Of pH on the Dissolution Rage of MnO During the Oxidation of Cr + with B ank Corrections.............. ......... 99 Oxidation of Fresh and Aged Cr(OH)3 by MnOz............................... ....... 102 Rate of Cr Oxidation by Mnoz in Open-Air and N2 AtmosphereSOOOOOQOOOOOoooeeeeee ooooooo 105 Reduction by Cr3+ in Rate of MnO 5 N2 Atmospheres..................105 Open-Air an Effect of Ionic Strength on the Oxidation Rate as set by NaNO3eeeeeoeeoeoee00.000.00.00108 Effect of Ionic Strength on the Reduction Rate as set by NaNO3oooooooeoeooeoeoooo oooooo 108 viii Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. Effect of Ionic Strength on the Oxidation Rate as set by C3(N03)2......................109 Effect of Ionic Strength on the Reduction Rate as set by C3(NO3)2............... ooooooo 109 Reduction Rate of MnOz in Various Background Electrolytes.............. ........ 112 Reduction Rate Of MnO in CaClZ Solutions of Various Concentrat1ons....................114 Oxidation Rate of Cr3+ in CaClz Solutions of Various Concentrations....................114 Reduction Rate of Mnoz in Ca(CO3)2, NaCl and caC12000OOOOOOOOOOOOOOOOOOOOOOOOOOO OOOOOO 115 Reduction Rate of MnOZ in KCl, CaClZ and MgCIZOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOlls Adsorption of 0.50 mg/l Copper, Manganese, Zinc and Nickel by Various Amounts of MnOz...118 Rate of Oxidation/Reduction with Adsorbed Copper..............................120 Rate Of Oxidation/Reduction in Control Experiment with NO Copper............. ....... 120 Rates Of Cr Oxidation with Adsorbed Copper and in Control Experiment....... ...... 121 Rate of Mn Solubilization with Adsorbed Copper and in Control Experiment.............121 Amount of Cr Oxidized in Filtered Versus pH Stabilized Replicates.....................135 Reproducibility of Mixed-Oxide Oxidation RateSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO135 Percent Cr Oxidized in Mixed-Oxide Systems asaFunCtion Of “1102 Loadingeooeeeoeeoeeeeo0137 Total Amount of Cr Oxidized in Mixed-Oxide Systems Versus Initial MnOZ Concentration....137 Percent Cr Oxidized Versus the Initial MOlar Ratio Of Fe(OH)3 t0 MUOZ...............139 Amount of Cr3+ in Solution and Cr Oxidized Versus Time in a MIxed-Oxide System..........139 ix Figure Figure Figure Figure Figure Figure Figure 51. 52. 53. 54. 55. 56. 57. Kinetics of Adsorption by Fe(OH)3 and OXidation byMnOZOOOOOOOOOOOOOOOOOOOOOO OOOOOO 143 Percent Cr Oxidized in Mixed-Oxide System When Fe(OH)3 is Precipitated on 203X10 MMnOZOOOOOOOOOOOOOOOOOOOOOOOOO145 Percent Cr Oxidized in Mixed-Oxide System When Fe(0§)3 is Precipitated on 902x10 MMDOZOOOOOOOOOOOOOOOOOOOOOOOOOOO.145 Percent Er Oxidized in Mixed-Oxide System When Cr3 is Adsorbed to Fe(OH)3 Prior to Addition of 2. 3 x 10 5 M MnOz................147 Percent Cr Oxidized in Mixed-Oxide System When Cr3+ is Adsorbed to Fe(OH)3 Prior to Addition of 9. 2 x 10 5 M MnOz................147 Percent Cr Oxidized in Long Term Mixed- Oxide System When Cr is Adsorbed to Fe(OH)3 Prior to Addition of MnOz..... ....... 149 Conceptual Model of the Adsorption and Oxidation of Cr by Mn Oxide............. ..... 160 INTRODUCTION The following section of this report provides an introduction to the topic addressed by this research. After a general discussion of the historical development and relevance of this field of research, a brief summary of the aqueous geochemistry Of chromium (Cr) is provided. A detailed summary of past research which is pertinent to the present project is presented as Appendix I. The introduction is concluded with a discussion of the specific goals of this research. 1.0 GENERAL INTRODUCTION The realization of the importance Of trace metals to human health, in both a detrimental and essential role, has generated immense interest in the chemical behavior of these elements in man's environment. As aqueous systems are a major point of interaction between man and his environment, both directly, and through the food chain, much effort has been expended to define the sources, sinks and pathways Of these trace metals in the OCeans, atmosphere, streams, lakes and ground waters. It is essential to understand the geochemical and biochemical processes which control the mobility and bioavailability of these elements in order to predict the fate of the increasing anthropogenic sources to our environment. In general, the environmental fate of trace ‘metals is influenced. by' a *variety of processes; chemical, physical and biological. The important processes are known to include the weathering of rocks and sediments, mineral precipitation and dissolution, complexation by both inorganic and organic ligands, adsorption onto mineral and solid organic surfaces and the uptake by organisms and consequent cycling through the food chain. These processes are in turn controlled by external variables such as pH, the redox state of the system, the partial pressure of gas phase components, temperature and any other condition which affects the chemical speciation or bioavailability of the metal ion in the environment. The trace metal Cr is an element that has been shown to be both toxic and a dietary essential. In general, trivalent Cr is considered to be essential for animal and human health. There is also some evidence for essentiality to plant life (Mertz, 1971), although this point has not been entirely resolved. In high levels, there is evidence for toxic effects of the trivalent form to aquatic life (McKee and Wolf, 1963), although in general, this form is not considered highly toxic to animals or plants (Mertz, 1971). The hexavalent form of Cris considered to be highly toxic to both. animals and. plants, and. is a suspected carcinogen (Beliles, 1979). Chromium is used in a variety of industries including tannery, platingy metallurgic, and. chemical. Wastes from these industries, as well as municipal waste, are commonly landfilled, applied. toi agricultural land or released, to 3 rivers or other water bodies. The natural source of Cr to the environment occurs primarily through the weathering Of ultramafic rocks. The mineral chromite is the most important commercial source of this metal. Garrels et al., (1974) estimated that the interference index for man's effects on the natural exogenic cycle for Cr was 71%, which suggests that it is important to evaluate the fate of this increasing anthropogenic source. Much of the recent research on the environmental fate Of metals such as Cr has focused on the role of adsorption onto sediment or suspended particles for controlling trace metal concentrations in aqueous systems. Studies of metal partitioning in sediments, soils and suspended particulate materials have shown that hydrous oxides of iron (Fe) and manganese (Mn), organic materials and clays are the most important adsorbing phases. The idea that adsorption by hydrous oxides is an important process for controlling the distribution of trace metals in sea water was first suggested by Krauskopf (1956). Jenne (1968) proposed that adsorption by these species is a principle control of trace element mobility in soil and fresh water environments. Iron and. Mn oxides have jbeen shown to scavenge ‘metals more effectively than other co-existing mineral species (Chao and Theobald, 1976). Reasons for this include, 1) the occurrence Of’ these oxides as coatings on clays or other' mineral grains, which allows greater chemical reactivity than their concentrations would suggest (Jenne, 1968) , 2) the large 4 surface areas of these oxides, which can exceed 100 mZ/g, and 3) the high amount of available surface binding sites capable of fOrming strong covelant bonds with trace metals (Stumm and Mbrgan, 1981). Besides adsorption, reactions at solid surfaces may also include redox reactions which can further influence the behavior Of metal species (Hem, 1978). At the present time, adsorption by hydrous oxides is widely recognized as an important process, but the quantitative role in geochemical cycles is poorly understood. It has become apparent, that any attempt to model trace metal behavior in the aqueous environment, will have to include both the solution chemistry, in order to predict the speciation and solubility controls, and some way of assessing the role Of adsorption and co-precipitation reactions. One approach to assess the role of adsorption has been to determine the thermodynamic data for the reactions between metals and model adsorbents, such as synthetic Mn and Fe oxides, in the laboratory. The ultimate goal Of this approach is to use the experimentally determined binding constants in a predictive equilibrium model, such as those currently used to model solution chemistry. This approach is hindered by the difficulty in Characterizing the adsorbents present in the complex natural environment, and by the limitations Of the experimental data itself over a wide range Of environmental conditions. Still, this approach holds some promise for predicting trace metal behavior in at least a semi-quantitative sense. 5 In the case of Cr, the interaction with solid surfaces, such as those of Mn and Fe oxides, is more complex than for most other metals. Chromium occurs in two oxidation states which exhibit very different chemical Characteristics, including very different adsorptive behaviors. Furthermore, the controls on the valence state transformations in the environment are poorly understood. The adsorptive behavior Of this metal is also complicated by redox reactions which are catalyzed by adsorption onto Mn oxides and possibly other solid species. A model to describe the environmental behavior of Cr will thus be more complex than those for other metals. In order to predict the oxidation state of this element it is likely that the model will involve kinetic considerations, as well as thermodynamic data, if a semi-quantitative approach is to be attained. The research presented in this report is an attempt to provide data that will contribute to an overall model of Cr geochemistry. The focus of this research is on the interaction Of Cr with a hydrous Mn oxide similar to those found naturally occurring. Specifically, this research will attempt. to 1) provide a fundamental description. of the oxidation Of Cr by Mn oxide, 2) determine the effect of solution chemistry on this reaction, and 3) study the oxidation of Cr by Mn oxide in the presence of an additional solid species, Fe oxide, which is Often associated with Mn oxide in the environment. 6 2.0 AQUEOUS GEOCHEMISTRY OF CHROMIUM The element Cr (atomic number 24), is a VI—B group transition metal. The ground state electron configuration is [argon] 3d54s1 (Douglas et al., 1983), thus all orbitals are half-filled. Chromium can lose any' number of 'these six electrons. The two oxidation states which are stable in 6+ 3+ and Cr . Trivalent Cr has natural aqueous systems are Cr three 3d electrons, which are in the high spin state when in octahedral coordination, and the resulting octahedral site preference energy is the largest of the transition metals (Murray et al., 1983). In the aqueous environment, Cr3+ is amphoteric, and is usually a positively charged or neutral aqua-species, unless the pH is very high. The hexavalent form is very acidic, and is an anionic aquo-species under most conditions. This form is also a strong oxidizing agent. In the aqueous environment the trivalent form, Cr3+, is octahedrally coordinated with hydration water, which in a kinetic sense is "permanent" (Elderfield, 1970). These water molecules are hydrolyzed to various extents, dependent on pH. A recent study by Rai et al., (1986) experimentally evaluated the hydrolysis and solubility constants for Cr3+. It was shown that the thermodynamic data previously reported in the literature were both inadequate and inaccurate. Their research demonstrated that CrOH2+, Cr(OH)3° and Cr(OH)4- are the dominant hydroxy species, and that polynuclear species are not important, which is in contrast to some previously 7 reported data. These authors also evaluated the solubility of Cr(OH)3-solid and (Cr,Fe)(OH)3-solid solution, which 3+ solubility in most systems. It was probably control Cr shown that the kinetics of precipitation-dissolution Of these compounds is very fast, and that when sufficient Fe is present, (Cr,Fe)(OH)3-ss should control Cr solubility. It is apparent that over the pH range of most natural waters, the 3+ is very low; even lower than previously 3+_ solubility of Cr reported. Several studies have demonstrated that Cr complexes are unimportant compared to the hydroxy species, even in sea water conditions (Rai et al., 1986: Elderfield, 1970; Van Der weijden and Reith, 1982). The importance of Cr3+-organic complexes has not been satisfactorily evaluated. Hexavalent chromium, Cr6+, is present in natural aqueous systems primarily as the tetrahedrally coordinated chromate ion, Cr042', with. minor amounts of HCrO4', HZCrO4, and Crzo72" (Elderfield, 1970; Leckie et al., 1983). This species is very soluble and is generally considered to be very mobile under many environmental conditions. Based on the available thermodynamic data, Cr6+ should be the dominant species for most oxygenated systems. Under sea water conditions (pH=8.1 and pe=12.5) the calculated ratio Of Cr6+ to Cr3+ is about 1021 (Elderfield, 1970: Nakayama et al., 1981). Reducing the pe to 8.5 only changes this ratio to 109 (Nakayama et a1., 1981). Cr6+ has also been shown to be stable in oxic, alkaline ground waters (Robertson, 1975). 8 In contrast to the calculated predictions, many 3+ in analyses of investigators have found considerable Cr fresh water and sea water (Pankow et al., 1977; Nakayama et al., 1981; Elderfield, 1970; Fukai, 1967). Potential reasons for this discrepancy include; 1) analytical techniques do not adequately distinguish between Cr3+ and Cr6+, 2) published stability constants are in error, 3) estimates of the redox conditions in environments such as sea water are incorrect, 4) the oxidation state of Cr is kinetically controlled, and 5) there have been important stable species Of Cr3+ that have been overlooked (Elderfield, 1970). Although analytical techniques may be responsible for some inaccurate data, 4) and 5) above seem likely to be responsible for this Observation. The slow kinetics of oxidation of Cr3+ to Cr6+ by 02 have been demonstrated in several studies. Schroeder and Lee, (1975) found only 3% Of an initial 0.125 mg/L Cr3+ spike was oxidized by 02 in 30 days, in a solution buffered at pH=8.6. Nakayama et al., (1981) did not find any detectable oxidation in sea water over 300 hours. Similarly, Van Der Weijden and Reith, (1982) could not detect any oxidation of 0.100 mg/L Cr3+ in either fresh water or sea water (pH 5.5-8.0) over a six week period. The reaction is thought to be very slow, due in part, to the change in coordination. number 'upon. change. in. oxidation state (Elderfield, 1970). The ability of thermodynamic equilibrium to describe the oxidation state of Cr will depend on the 9 residence time of the dissolved species in the particular environment. Of interest, and the competition from. other reactions. Cranston and Murray (1978) have shown that the oxidation state of Cr does seem to respond to redox conditions in the environment. Rate data for the reduction Of Cr6+ in a reducing environment is apparently scarce. Oxidation. and reduction. reactions Ibetween. Cr land a variety of other species, which may influence the behavior of this element in the environment, have been demonstrated to occur in the laboratory. Oxidation of Cr3+ has been shown to occur via adsorption onto Mn oxides, apparently by the reduction of surface Mn4+ (Schroeder' and. Lee, 1975: and others). Reduction of the highly oxidizing Cr6+ has been shown 'to occur 'with a ‘variety of substances, including reducible organics, such as ascorbic, humic and gallic acids (Nakayama et al., 1981; James and Bartlett, 1983), organic compounds containing sulfhydryl groups (Schroeder and lee, 1975), hydrogen sulfide (Smillie et al., 1981) and ferrous iron (Schroeder and Lee, 1975). It is difficult to predict the environmental importance of these reactions as the rate of these processes are not well known. Another generally unresolved component Of the overall geochemical behavior of Cr is the possible formation Of "inert" complexes of Cr3+ with organic ligands. Nakayama et a1. (1981) have shown that the Cr3+-citric acid complex does not coprecipitate with Fe oxide and is not oxidized by Mn oxides. James and Bartlett (1983) have shown that under 10 certain conditions in soils, citric acid will increase the solubility of Cr3+ by the formation of soluble Cr complexes. The geochemical behavior of Cr, as with any trace metal, is also influenced by adsorption. The speciation of a dissolved metal species must be considered in adsorption reactions, and this is especially true for Cr, with its two valence states. Cr3+ would be expected to adsorb strongly over the pH range of natural waters on many solid surfaces (see Appendix I). As an anion, Cr6+ adsorption on many common adsorptive substrates would be limited (see also Appendix I.). For instance, adsorption on Fe oxides would be important only at pH's in the lower range for natural waters, and adsorption of this species by most Mn oxides would not be expected to occur at all (see Appendix I). Assembling the data from these previous investigations into a qualitative description of Cr geochemistry results in a complicated model. Figure 1 is a schematic of the geochemical processes which are thought to control the behavior of Cr in oxygenated aqueous systems. In solution, Cr3+ (as a kinetically stable hydroxy species) can exist only in very low concentrations, possibly unless complexed with organic ligands. Some Cr3+ may become oxidized by 02 by this kinetically slow reaction, but the tendency will probably be for this species to precipitate as Cr(OH)3 or (Fe,Cr)(OH)3, or to adsorb on particulate Fe oxide, organics, or clays. Thus Cr3+ species will tend to accumulate in sediments. The potential of these sediment 11 OINPO>O maOm=O( O.XO :- >¢Pm.3m:OOmO 3.4-80:10 ...O .5003 ..(q-FANOZOO .—. ”£30...— 11. ZO.hn¢Omn—3 ZO-POOOMK mmxmadaou 02.3.5-5 ' . .. .O . ...o ...:o. o IO-IOGZO ND-XOCIU h2m4<>=uh PZMJ(>44<23<2>003¢N2P « O >fl >44<0-.—.m2.¥ ZO.hO=PO hammmcn :- awammcno< m20..—.0<¢mhz_ NEXO .mlca\3:.XO¢—¢O uO O.._. ZO_._.Omo. 02_m¢_._.m DEM—504‘: C(DIKCOW/ O_km20(a 8‘ It! 1: m5)“. 1 1.11 40.3.... L: n-.— _ .l 1.: _ _I 1 _ll $3.3m meE S: _ 3.355-085... z : ._ _ _ _ 2 z _ _ _ _ _ _ 3.3.. do» _ : ... _ _1_ 1.2.30 0.. 52-02:... . 1:; 00.. 39.5.4.» . 2...... 3.5.3.... 1220.59 m2... 93 N2 0 3053.5 .... wan... 204nm». Oh Om10w 0213.245 moomhommm «o #2333: 42 atmosphere (by bubbling N2 through the system), and with ease of reagent addition and sampling. For some of the experiments of this project, this entire system, including the magnetic stirring device, was also used inside a glove- box so that the atmosphere outside the reaction vessel could also be purged with N2. 4.3 General Experimental Procedures The following procedures are a general description of the methods used for most of the short-term (0.25-4 hrs.) experiments used to study the oxidation of Cr3+ by Mnoz. Special procedures used in mixed Fe-Mn oxide systems, kinetic studies, long term experiments, etc. are described in the appropriate results sections of this report. A typical oxidation experiment was conducted as follows: 1. The proper volume (usually 500 or 1000 m1) of double- distilled water containing the desired background electrolyte was added to the reaction vessel and adjusted to a pH near that desired for the experiment with the appropriate acid or base (i.e. HNO3 and NaOH for an NaN03 background solution). 2. The Mno stock solution was placed in the sonic vibration water ath for' 20 :minutes and then returned to the magnetic stirring device so that any suspension that had settled during vibration would be resuspended. 3. An aliquot of Mnoz suspension, containing the desired amount of Mnoz, was removed from the stock solution by glass or eppendorf pipette ‘while the suspension was stirring briskly on the magnetic stirrer. 4. The aliquot of Mno suspension was added to the reaction vessel and the p adjusted to that desired for the reaction. 43 5. The system was allowed to equilibrate for at least 1 hour at this pH and the pH was re-adjusted to the desired level if drift occurred. 6. Cr3+ was added to the reaction vessel as a small aliquot of 1000 mg/l Cr(NO3) stock solution to obtain the desired initial concen ration of Cr +. The Cr was added via an eppendorf pipette to ensure that the addition was essentially instantaneous. The pH was re-adjusted to the desired level if drift occurred. 7. The system was then sampled over time by removing an aliquot (15-20 ml) by either pipette or sampling syringe connected to a Teflon tube. 8. The samples were then immediately filtered through a 0.22um millipore filter (see below). 9. The resulting supernatant was used for anaéysis of tLe desired parameters, usually Cr (total), Cr , and Mn by atomic absorption spectrophotometry (AAS). 4.4 Sample Filtration The filtering of aliquots of the experimental system to separate the solution (and solutes) from the oxide(s) (and hence terminate the reaction between the two phases) was an important step in the experimental procedure. This procedure had to meet the following requirements: 1) there had to be no 'measurable uptake or release of chemical species that would effect the analytical results, 2) there had to be essentially complete removal of the solid phases from the sample, and 3) the process had to be fast with respect to the rate of reaction. Two different filtration techniques were used in this study. The first, used only in early experiments, was the removal of an aliquot by glass pipette followed by delivery 44 to a vacuum driven millipore glass filtering device, where the sample was collected in a glass test tube connected to the end of the filter stem. This method worked well but was much too slow (approximately 30-45 seconds) for the early part of the oxidation reaction and inadequate for the kinetic studies. Since a faster process was necessary, a second technique was developed. This consisted of sampling the system with a plastic syringe connected to a Teflon tube and then injecting the sample through a Millipore Swinnex polypropylene filter holder. The sample was forced through the filter' by pressure rather than drawn through. by a vacuum. This whole process took only about 10 seconds which is relatively fast with respect to the reaction rate for most of the experiments performed in this study. For consistency, the time of the sample with respect to the start of the experiment was always recorded as the end of the filtering process. The use of Millipore mixed cellulose acetate and nitrate filters did not appear to affect the sample chemistry by retention or release of chemical species. This determination was made by analyzing filter blanks and metal standards that were filtered. For a solution with pH 4.5 and 0.10M NaNO3 the filtering process did not effect levels of Cr, Mn, Cu, or Zn as measured by flame AAS. Both 0.45pm and 0.22pm filters were tested for the ability to remove iron and manganese oxides from solution. 45 Both filter sizes appeared to be adequate for removal of freshly precipitated amorphous iron oxide as the resulting supernatants were below detection for Fe by flame AAS methods. However, the 0.45pm filter did not appear to remove as much MnO2 from solution as the 0.22pm filter. In four 0.45pm filtered sample replicates, detectable Mn was measured in all four replicates with an average value of 0.03 mg/l. In four replicates filtered by a 0.22pm filter the Mn levels were at or below detection level (0.01 mg/l). Furthermore, the Mn levels of samples passed through the 0.45pm filter could be reduced to detection by re-filtering through the 0.22pm filter, suggesting that small Mnoz particles do pass through the 0.45am filter. It is not known whether the 0.22pm filters allow passage of a small quantity of finely particulate Mnoz or whether the detection 2"" in the limit values (if real) were due to dissolved Mn system. At these levels the Mn does not effect the experimental results. The 0.22pm filters were used for all experimental data presented in this report. 3+ 6+ 4.5 Analysis of Cr and Cr in Solution The success of this research, in part, depended on the ability to accurately determine the oxidation state of Cr in solution. The method to be employed had to be effective for small sample sizes, relatively simple due to the large number of samples, and accurate through the necessary concentration ranges of the experiments. 46 There have been various analytical methods of Cr 3+/Cr6+ determination reported in the literature. These include: cation/anion exchange resins (Pankow et.a1.,1977 and Smillie et.al., 1981), the diphenylcarbazide colorimetric method for Cr6+ described in Standard Methods (APHA et.a1.,1975) and used by Schroeder and Lee (1975) and Bartlett and James 3+ (1979), the coprecipitation of Cr (and exclusion of Cr6+) with amorphous iron hydroxide (Fukai, 1967; Nakayama et.al., 1981; Cranston and Murray, 1978), and the precipitation of the Pb-Cr6+ complex with 8042' (Martin and Riley, 1982). Because of the reported accuracy, working concentration range, and equipment requirements, the method of Martin and Riley (1982) was investigated for use in this research. The rationale for this method is based on the strong complex formed between Pb and hexavalent Cr in solution at low pH's. After this complex forms, it can be coprecipitated with lead-sulfate. Hence after centrifugation, the precipitate will contain all of the Cr6+ and the Cr3+ will remain in solution. After separation from the solution, the precipitate can then be resolubilized, diluted to the original sample volume and measured for Cr by AAS. Most of the experiments in this study required accurate Cr6+ determinations over the range of 0.01 to 0.50 mg/l. Hence the Pb-SO4 method of Martin and Riley (1982) was tested over this range for the efficiency of Cr6+ recovery 3+ and Cr exclusion. Figure 7 shows the efficiency of Cr6+ recovery for six standards, all with 0.50 mg/l total Cr, 47 with various amounts of the two oxidation states, using the unmodified Martin-Riley method. The 45° line represents 100% recovery and points plotting below the line represent incomplete recovery. As the data demonstrate, this method was not able to recover all of the Cr6+, with the recovery 6+ concentration. rate getting worse with increasing Cr Before abandoning this method, however, an attempt was made to modify it by adding more lead and more sulfate, as following' the exact. method resulted in only ‘very small amounts of precipitate. Figure 8 shows the efficiency of recovery using twice the amount of Pb(NO3)2 and (NH4)2804 (i.e. 200ul vs. 10011.1). This resulted in a much better recovery rate as only the samples containing 0.40 and 0.50 mg/l Cr6+ were slightly depleted with respect to the standard. The use of three times the amount of these reagents resulted in a ‘very' good recovery rate in all standards as shown on Figure 9. Thus, this modification of the Martin-Riley method resulted in a potentially useful procedure. The points on these graphs are the average of four measurements, with two replicates taken of each standard and each replicate measured twice by AAS. There was very little variation between measurements suggesting good accuracy as well as good precision. This modified method was studied further for recovery efficiency in a variety of background electrolytes and trace metals, for optimization of the amount of reagents necessary to resolubilize the precipitate, and for the best method for 48 _cr(vn) RECOVERY (mg/l) 0.60 uuuoonneo Cr(Vl) mum: 0.50 ‘ 0.40 " D D 0.30 d D 0.20 ‘ D 0.10 ‘ U 0100 I 1 r I I 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Cr(V1) STANDARD (mg/l) 6+ Figure 7. Efficiency of Unmodified Cr Extraction Procedure from Martin and Riley (1982) 49 0.60 MODIFIED Cr(Vl) monou 0.50. Pb AND 504 DOUBLED 0.20 d 0'00) RECOVERY (mg/0 O 8 0.10 1 0.00 I Y T T I 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Cr(\fl) STANDARD (mg/l) Figure 8. Efficiency of Modified (2 X) Cr6+ Extraction Procedure from Martin and Riley (1982) 0.50 MODIFIED Cr(V|) mow 0,50 . Pb mo $04 TRIPLED 0.30 ‘ 0.20 -‘ 0:011) RECOVERY (mg/l) 0.10 . 0.00 r fi ,4 r “r v 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Cr(VI) STANDARD (mg/I) Figure 9. Efficiency of Modified (3 X) Cr6+ Extraction Procedure from Martin and Riley (1982) 50 the preparation of AAS standards. Since some of the adsorption/oxidation experiments were designed to explore the effects of different electrolytes and ionic strengths, it had to be shown that this modified method of Martin and Riley (1982) (hereafter referred to as MMR) would work in a variety of experimental matrices. To address this concern, 0.50 mg/l Cr6+ and Cr3+ standards were made in various matrices and the MMR extraction was used to test for 5+ 3+ complete recovery of Cr and complete exclusion of Cr . The results for the 0.50 mg/l Cr3+ standards showed that there was no removal of Cr3+ by the Pb-SO4 precipitate in any of the matrices tested. The results for Cr6+ recovery from 0.50 mg/l standards showed there was complete recovery, or near complete recovery, in all of the matrices tested except two. The only potential problems occurred with the 0.10 M NaZSO4 matrix and 0.5 M NaCl, both of which resulted in a low recovery rate. In the first case, the presence of a high amount of $042- in the matrix caused a precipitate to form upon addition of the Pb(N03)2. The PbSO4 precipitate thus forms at an earlier step in the procedure than required. This may explain the low recovery and suggests that this method may not be useful for solutions containing high levels of sulfate. Similarly, a precipitate also formed upon Pb addition in matrices with a high level of C1. The formation of this precipitate (probably PbClz) also caused a low recovery rate. Lower levels of Cl salts did not affect the reliability of this method. 51 To ensure that the presence of other trace metals did not effect the MMR technique, 0.50 mg/l Cr6+ solutions containing 1.00 mg/l Mn2+, Cu2+, or Zn2+ in 0.10 M NaNO3 were also tested for interference with the recovery of Cr. The results showed that there was no interference by these trace metals, and thus the MMR method was suitable for experiments designed to explore the effects of competing trace metals on the oxidation rate of Cr by Mnoz. The modification of the method of Martin and Riley (1982) by increasing the amount of Pb-SO4 precipitate resulted in a fairly concentrated solution matrix to be analyzed by AAS. To minimize any error due to matrix effects when comparing samples to standards and blanks, the standards and blanks were prepared exactly like the samples, except all solution and reagent volumes were increased by a factor of 10, and the Cr was added to the standard after the Pb-SO4 precipitate was resolubilized. Thus the standard preparation was not actually an extraction procedure but the resulting matrix was very similar to that of the samples. The exact procedure of standard preparation is listed in detail below. With the modifications and careful standard/blank preparations this method for determining the concentration of Cr6+ was found to be adequate for use in this research. An error analysis of 10 replicates showed that over the range of 0.02 to 0.50 mg/l that the precision (standard 52 error) was approximately +/- 0.02 mg/l. The exact MMR method as used in this research is listed as follows: 1. A 10ml aliquot of filtered sampled was placed into a 50 ml Pyrex centrifuge tube. 2. The pH of the sample was adjusted to 3.5 +/- 0.2 by the addition of 10% v/v acetic acid, usually about 10ul for samples starting at a pH of 4.5. If the sample pH was lower than 3.5 or inadvertently lowered below 3.3, it was raised to the proper range by 10% v/v NH40H. 300111 of 1.0 M Pb(N03) was added to the sample by eppendorf pipette and t e solution mixed on a vortex stirrer for about 15 seconds. This was allowed to set for 3 minutes to allow the Pb-Cr complex to form. 4. 500u1 of concentrated acetic acid was added to the sample and the solution mixed 8n the vortex stirrer. This was done to make sure all Cr + is solubilized. 300ul of 0.20 M (NH4) SO was added to the sample by eppendorf pipette and2 t e solution stirred for 15 seconds on the vortex stirrer. The Pb-Cr-SO4 precipitate begins to form. The centrifuge tubes were capped and placed in the centrifuge. The rpms were increased slowly over the first 5 minutes to a speed of 10,000 rpm. The samples were spun at this rate for 10 minutes. The samples were removed from the centrifuge and the solution was removed from the precipitate by a vacuum driven suction device. To the precipitate was added 1.0 ml of concentrated nitric acid, 100ul of 30% hydrogen peroxide, and 100 pl of 0.5 M Ca(NO3)2 . The volume of the sample 'was then returned to its original 10 ml with double-distilled water and stirred on the vortex stirrer until all of the precipitate was resolubilized. 10.The sample was then analyzed for Cr by AAS. The left over sample from which the initial 10 ml aliquot was taken was analyzed for total Cr (and Mn or any other(fipecies) and thus Cr + was determined by [Cr] total - [Cr ]. 53 The preparation of blanks and standards was conducted as follows: 1. Two flasks of 100 ml of double-distilled water containing the same background. electrolytes as 'the experimental system were poured into 200 ml glass centrifuge tubes. 2. The pH of each was adjusted to 3.5 with 10% v/v acetic acid. 3. 3 ml of 1.0 M Pb(N03)2 was added and the solutions stirred. 4. 5 ml of concentrated acetic acid was added to each followed by stirring. 5. 3 ml of 0.20 M (NH4)ZSO4 was added and the solutions stirred and then centrifuged at 10,000 rpm for 15 minutes. 6. The solution was removed from the precipitate as described above and then to both was added 10 ml concentrated nitric acid, 1 ml 30% hydrogen peroxide, and 1 ml 0.50 M Ca(N03)2 . 7. One of these was then returned to the original volume of 100 ml with double-distilled water and stirred until all of the precipitate dissolved. This was the blank matrix for the AAS analysis. 8. To the other was added 5 ml of freshly prepared 10 mg/l Cr standard. The volume was then returned to its original 100 ml with double distilled water. This resulted in a 0.50 mg/l Cr standard. 54 4.6 Atomic Absorption Spectrophotometry The samples were analyzed for CrT, Cr6+, MnT or any other metal parameter by flame atomic absorption spectrophotometry with a Perkin Elmer 560 atomic absorption unit. The flame was an air/acetylene mix. The wavelength used for Cr analysis was 357.9 nm with a slit width of 0.7. For Mn analysis the wavelength was 279.5 nm with a slit width of 0.2. 4.7 Experimental Reproducibility Several experiments were repeated to determine whether the data were reproducible. This was done over a period of several weeks to avoid the possibility of changing oxide reactivity due to an increase in particle size. Figure 10 shows the results of two identical oxidation experiments performed about two weeks apart. Some of the samples were collected at slightly different time intervals, such that the data points should fall along the same curve. The data show the reproducibility is very good, especially for the measured Cr6+. The amount of solution Mn was slightly higher for one experiment, however the agreement is within analytical error. Comparisons of other experimental data from duplicate experiments, presented in the results section also show good reproducibility. 55 1.0 Mn02-1.9X10-5M 0'9 3‘ Cr-9.6X10-0H ‘ . x X A 0.5 x A x A A A 0.7 « E. XA v 0.5 4 1 05 4 K o g . o D 0 D o g D g 0.4 - 0C! 03 -' U - x :1 Cr (2/5/87) A 0.2 4'6 AW! (2/5/85) oer (1/17/85) °-‘ ‘ XMn (1/17/85) 0.0 . . r . . o 15 30 45 oo 75 so ELAPSED 11m: (MINUTES) Figure 10. Reproducibility of Oxidation/Reduction Experiment RESULTS AND DISCUSSION In this section of the report, the results of the experiments conducted to investigate reaction between Cr and Mnoz are presented, discussed and interpreted. The results are presented in three main parts: 1) the general nature, stoichiometry and kinetics of the reaction, 2) the effect of solution chemistry on this reaction, and 3) the competition between Fe and Mn oxides for dissolved Cr. In each part, the results are discussed and interpreted in light of current adsorption theories in an attempt to describe the interaction between Cr and hydrous oxide surfaces in some detail. The experimental data presented on each figure are listed in Appendix II. PART 1.0 THE OXIDATION OF Cr(III) BY M1102 1.1 THE GENERAL NATURE OF THE REACTION 1.10 Introduction Preliminary experiments were conducted to optimize the experimental parameters so that the reaction between Cr3+ and the Mnoz surface could be most effectively studied with the analytical techniques developed. In particular, these experiments determined the reactant concentrations and solution parameters necessary to obtain reliable rate data and to ensure that all reaction products remained in solution so that reaction stoichiometry could be determined. 56 57 Quality control and repeatability analyses were also done at this time. Data from these experiments were used to describe the general reaction characteristics. The adsorption experiments described in this section of the report were all conducted at pH 4.5 in a solution matrix of 0.10 M NaNO3, in an open to the atmosphere environment at room temperature. At this pH, there was no precipitation of Cr hydroxides and the adsorption of reaction products was minimal. The concentration of reactants was varied, however 3+' and initial reactant concentrations of 9.6 X 10'6 M Cr 1.9 X 10"5 M Mnoz were used as a control, or base-line experiment for describing the reaction, for comparing rates under different solution conditions and for repeatability analyses. Experiments conducted with these reactant concentrations, for convenience, are referred to as the "standard" oxidation experiment in the following text. 1.12 Results and Discussion The addition of Cr3+ to an Mnoz suspension resulted in visual changes to the system within minutes. The suspension became lighter and took on a yellowish brown tint. Measurement of solution Cr6+ and total dissolved Mn typically’ yielded. a jpattern of rapid increase. of these parameters over the first several minutes, followed by a decreasing rate of increase as these reaction products approached a constant level. For the standard oxidation experiment, the reaction was usually nearly complete within 58 30 minutes. Total solution CrT (i.e. dissolved Cr3+ plus Cr6+) remained very nearly constant at the initial concentration of Cr3+, except for a very small depletion at the very beginning of the reaction. A graph of solution Mn, Cr6+ and CrT versus time for the "standard" oxidation experiment is shown in Figure 11. As shown on the figure, the concentrations of solution Mn and Cr6+ increase together as the reaction proceeds and, except for a slight depletion at. the beginning' of the. experiment, CrT .remains nearly constant throughout the reaction. The most surprising result of these initial experiments was the high rate and completeness of the reaction, even at this relatively low initial Mnoz concentration. under the conditions of the "standard" oxidation experiment, it took only 30 minutes to oxidize more than 90% of the Cr3+ in solution, with over half of the Cr oxidized within the first 5 minutes. The initial molar ratio of MnOz to Cr3+ in this experiment is only about 2 to 1. Based on past research, the adsorption of other trace metals such as Cu, Ni and Zn at pH 4.5 with this metal to oxide ratio ions would be minimal. The quantity of Cr oxidized even at this low Mn02 loading would seem to indicate that Cr3+ has an unusually high affinity for the Mn02 surface. When the initial Mn02 loading was less than the initial Cr3+ concentration the reaction proceeded until nearly all of the Mnoz was depleted, and may have gone to completion if given. enough. time. IFigure 12 shows the results of an 59 16.0 12.0 - i b 1 S 5 .1 \ E 4.0 C] Cr(Vl) OTOTAL Cr A Mn(ll) 0.0 1 T U o 20 4o 60 00 100 120 110 160 ELAPSED TIME (MINUTES) Figure 11. Oxidation of Cr3+ by MnOZ Under the Conditions of the Standard Oxidation Experiment 5.0 4.0 . I0 :5 '- 3.0 .. x i E __. g 2.0 -I t. .3 A UN) 1.0 H o Mn(||) -|NIT|AL Mn02 0.0 I I 1 T T r 010 203040900070 00 ELAPSED TIME (MINUTES) 3+ Figure 12. Oxidation of Cr in Molar Excess by MnO2 60 experiment with the initial MnOZ concentration less than the 3+ concentration. After several hours all but a initial Cr small fraction of the initial MnOz was reduced. In addition to an apparent high affinity of Cr3+ for the MnOZ surface, these results suggest. that the oxidation- reduction process cannot be explained with an adsorption model which assumes a fixed quantity of adsorption sites. The reaction is obviously not limited by the adsorption capacity of the initial oxide suspension, which on a molar basis usually represents only a relatively small fraction of initial oxide concentration. Rather, it appears that there are nearly unlimited sites for adsorption of Cr, up to the predicted molar stoichiometric ratio of the reaction. This behavior is not "adsorption-like" in the sense that similar concentrations of other trace metals would not even adsorb to any significant extent under these conditions. Based on these initial observations, two working hypotheses were formulated at this point. The first was that the redox reaction is not adsorption controlled, at least in the sense of the formation of covelant bonds between the oxide surface and Cr3+. Possibly, the Cr ion need only to enter the electrical double layer (EDL) of the oxide, with oxidation taking place without direct coordination with surface oxygens. In this case the interaction between Cr3+ and oxidized Mn would exhibit characteristics of a solute-solute type interaction. This could explain the oxidation of Cr under conditions where 61 adsorption reactions should not occur. The second working hypothesis was that the reaction is adsorption controlled, with the reaction driven by a high affinity of Cr3+ for surface sites on MnOz, and that the oxidation-reduction reaction is autocatalytic, generating new surface sites for adsorption as the reaction proceeds. The idea of autocatalysis is suggested because the reduction of Mn4+ to Mn2+, which would be unstable in the Mn06 octahedron, must cause partial dissolution of the Mnoz structure. As the MnOz particulates dissolve it is feasible that new surface sites are created. If it is assumed that the reduction and desorption of an Mn ion produces one new surface site, the total number of surface sites would not begin to decrease until the reduction and release of Mn to solution no longer produced a new site (or sites). This would probably occur only' as the MnO2 particles reached, a critical size or dissolved completely. It would also be possible for the oxidation-reduction reaction to cause an increase in the surface area of the remaining oxide or to otherwise influence its reactivity; In other words, as the concentration of the oxide decreases it is possible that the specific surface area (and number of surface sites) of the remaining oxide may remain constant or even increase. The results. of subsequent. experimentation ‘were 'used. to ‘test these working hypotheses. Even if adsorption is a required reaction step, the 3+ constant concentration of CrT in solution shows that Cr is 62 not removed from solution for very long before it is oxidized and returned to solution. As described above, only a small depletion of total solution Cr is measured at the very beginning of the reaction (see Figure 11). This depletion (4-8% of an initial 0.50 mg/l) however, was seen in all standard experiments and, although barely outside of the expected analytical variation, does appear to be real. If this depletion is real, it may indicate that Cr3+ is removed from solution (adsorbed) for a very short time prior to oxidation. As there is no evidence for the adsorption of Cr by MnOZ without consequent oxidation, the rate of oxidation appears to be dependent on the ability of Cr to come in contact with the oxidizing agent. As demonstrated on Figure 11, the kinetics of the reaction are somewhat parabolic, with the initial rate nearly linear and the rate quickly slowing down with time. Since the Cr3+ does not adsorb without being oxidized, the decreasing reaction rate must reflect an increasing amount 3+ in solution to come of time necessary for the remaining Cr in contact with the oxidizing agent. Experimentally induced parabolic kinetics in heterophase reactions have been caused by grinding up the solid phase to decrease the grain size. Unless pre-treated, solids prepared in this manner may contain fine particles which are highly reactive as compared to the bulk solid (Holdren and Berner, 1979). Although this particular situation is not applicable to the present research, it is possible that the fast part of the reaction 63 3+ and Mn02 occurs with small, highly reactive between Cr MnOZ particulates (high surface area) and the slower part occurs with the remaining larger particulates which have lower surface area. Without a study of grain size distribution this possibility can not be evaluated. The decreasing reaction rate with time could also be caused by several properties of adsorption reactions. The sorbent/sorbate ratio is continuously decreasing during the reaction, as three moles of MnOz are dissolved for every two moles of Cr oxidized (see next section). This decreasing ratio may cause a lower fractional adsorption over time thus reducing the oxidation rate. If the reaction is adsorption controlled the reaction rate would also decrease as the particulates become very small or dissolve, as the rate of production of new sites would not be proportional to amount of Mn dissolved. In this case there would simply be less adsorption sites as the reaction proceeds. The kinetics of adsorption reactions on hydrous oxides typically exhibit a very rapid uptake of the metal ion, followed by a period of very slow uptake. This slow part of the reaction is thought to be due to the diffusion of the metal ions to binding sites within the oxide structure. However, if the oxidation reaction is generating new sites on the oxide surface, it seems unlikely that diffusion would become rate limiting. 64 1.2 THE STOICHIOMETRY OF THE REACTION 1.21 Introduction The general hypothesized reaction between Cr3+ and MnO2 is given by the equation: 3Mno2 + zcr3+ + 1211+ = 3Mn2+ + zcr6+ + 6H20 At pH 4.5 the predominant Cr species are Cr(OH)2+ and HCrO4- (Leckie et al., 1983). Using these species and the 2+ unhydrolyzed Mn ion would result in a reaction that is proton neutral: 3Mno2 + 2Cr(OH)2+ = 3Mn2+ + 2HCro4' From this reaction the predicted ratio of products would be 3:2 Mn2+ to Cr6+ with no change in pH. This does not include the stoichiometry of any adsorption reactions in terms of balancing the hydrogen ions. The stoichiometry of the reaction was evaluated by calculating the molar ratio of total dissolved Mn to Cr6+ from the experimental data. Thus Mn/Cr6+ could be plotted over time for each experiment. The experimental conditions were designed to assure that re-adsorption of products did not affect the stoichiometry, at least after the reaction had neared completion (see below). The value of total dissolved Mn represents the Mn that passed through the 0.22pm filter. It is assumed that this is primarily 2+ dissolved Mn as filtration studies (see Methods) showed 65 that only' very small amounts of’ MnOz pass through the filters, and other oxidation states of Mn are unstable in water as free ions (Hem, 1978). 1.22 Results and Discussion 6+ values, in general, confirm The measured molar Mn/Cr the predicted stoichiometry. Although there was some variation, most of the experimental data exhibit the pattern of Mn/Cr6+ ratios with time shown in Figure 13. The ratio in the first several minutes of the reaction (when the reaction is fastest) was typically well below the predicted ratio. The ratio then quickly increased to a value very near the predicted 1.5. By the time the reaction was essentially complete, the Min/Cr6+ ratio was almost always between 1.5 and 1.6. This would appear to confirm that the 4+ is reduced to Mn2+ during the oxidation of Cr, surface Mn rather than to Mn3+ as was shown to occur during the oxidation of Co2+ by MnOz (Crowther et al., 1983). 3+ Reduction to Mn might result in the disproportionation to Mn2+ and MnOZ, however this would not produce the present Mn/Cr ratio unless exactly one half of the Mn3+ became Mn2+. This does not fit the stoichiometry of the probable disproportionation reactions described by Hem (1978). However, this possibility can not be entirely ruled out based on the data from this study. Assuming' the 'hypothesized reaction stoichiometry is indeed correct, the possible causes for deviation of the 66 1.8 1.7 1.6 1.5 MUD/CFO") 1.4 1.3 112 l T T T T I T O 20 40 60 80 100 1 20 1 40 1 60 ELAPSED TIME (MINUTES) Figure 13. Stoichiometric Ratio of Reaction Products \\ 67 Mn/Cr ratio from the predicted 1.5 include; 1) analytical error, 2) adsorption of reaction products, 3) different rates of release to solution of the reaction products, 4) 2+ desorption of any adsorbed Mn (residual from oxide 3+ or dissolution of preparation) upon the adsorption of Cr the oxide structure, and 5) an average oxidation state of Mn in the oxide of less than or greater than four (i.e. the presence of Mn3+, etc.). The reason for the low Mn/Cr ratio at the beginning of the reaction is not clear. The analytical data from this part of the experiment has a large relative error due to the low levels of both Cr and Mn. However it seems unlikely that the error would be systematic, always resulting in a low ratio. If the low ratios in the early part of the reaction are not caused by analytical procedures, they could be caused either by adsorption of the Mn produced, or by the slow release (as compared to Cr6+) of Mn to solution. If adsorption of Mn2+ does occur in the early stages of the reaction, it may become desorbed as the sorbent/sorbate ratio decreases as the reaction proceeds. This would explain the increase of the ratio with time. To address this possibility, experiments were conducted to determine the extent of' IMn2+ adsorption. at MnOz and Mn2+ concentrations similar to those encountered over time during the "standard" oxidation experiment. In these experiments, the concentrations of solution Mn and MnOz were chosen to :represent the quantities that should. be present in the 68 5° ‘1 4o EITIME EQUNALENT I MIN. ATIME EQUIVALENT 5 MIN. 8 mm EQUIVALENT 30 MIN. m 30- g 2;" ml 2 I 10 ~ 0 0 IL, I . ¢ 0 1 2 3 4 5 Mn(II) 10 mm RATIO Figure 14. Adsorption of Mn2+ by Mn02 at Selected Metal/Oxide Ratios 69 system during the standard oxidation experiment at time intervals 1.0, 5.0, and 30.0 minutes, respectively, based on the amount of Cr that was oxidized. The amount of Mn adsorbed was determined from the amount deleted from solution after 5 minutes. The results are shown on Figure 14. The results indicate that adsorption only occurs at the conditions found very early in the reaction. This small percentage of adsorption could contribute to the low Mn/Cr6+ ratios seen early in the oxidation experiments, but is does not appear to be a satisfactory explanation by itself. The experiments show that by the end of the experiment there should be no measurable adsorption of Mn2+, suggesting the final stoichiometric ratio is accurate. As re-adsorption of solution Mn does not appear to be significant, differential desorption kinetics could be the 6+ ratios. The initial deviation and reason the low Mn/Cr subsequent adjustment of the Mn/Cr ratio to the predicted stoichiometric value during the reaction could be explained if Cr6+ is released to solution faster than Mn2+. During the fast part of the reaction the difference would result in a significant deviation in the ratio. As the reaction slows 2+ down and nears completion the Mn in solution would have a chance to "catch up" to its stoichiometric equivalent of Cr5+. The unfavorable nature of the MnOZ surface for Cr6+ adsorption (see Appendix I) would likely cause rapid desorption of this species. It does not seem unreasonable that this may be faster than the dissolution of the MnOZ 70 structure and release of Mn2+ to solution following reduction. The surface charge, at least, would be favorable for the attraction of the positively charged Mn2+ ions. If 3+ adsorption is a the formation of new surface sites for Cr required reaction step, the dissolution and release of Mn to solution may be the rate limiting step in the reaction. As shown on Figure 13, the Mn/Cr6+ ratio when the reaction is complete is slightly above the predicted ratio of 1.50 (1.51-1.59). Possible explanations for this include: 1) the average oxidation state of structural Mn in the oxide may be less than 4.0, and some oxidation of Cr by Mn3+ may take place, and 2) the total solution Mn may include small amounts of’ Mn not produced. by’ the REDOX reaction. Based on the iodometric titration the average oxidation state of the Mn in the stock suspension is about 3.9. The predicted reaction stoichiometry, assuming both Mn3+ and Mn4+ oxidize Cr in proportion to their respective mole fractions, would result in a Mn/Cr6+ ratio of 1.54. Thus the experimentally determined ratio may be reflecting the actual reaction stoichiometry. It is also possible that some particulate Mn02 and/or residual Mn2+ released upon oxide dissolution may be included in the total dissolved Mn analysis. The particle size decreases during the reaction which may cause some Mn02 to pass through the filter. This probably causes some of the variation in the data. Despite these minor variations in the reaction product stoichiometry 71 with time, the experimental results confirm the predicted reaction stoichiometry, as least with respect to Cr and Mn. As described earlier, the specific reaction predicted to occur at pH 4.5 is balanced with respect to protons. The change in the proton concentration during the reaction, as measured by a pH electrode, was not adequate to confirm this prediction. The experiments were conducted either open to the atmosphere or in an N2 purged system. However no measures were taken to ensure the absence of C02 species in the system. Without using C02 free reagents, an accurate measurement of proton consumption or generation can not be obtained. The pH was monitored in both open air and N2 systems to measure any pH changes which occurred throughout the reaction so that a qualitative assessment of the role protons in the reaction could be made. The "standard" oxidation experiment could not be used for this purpose, 3+ aliquot results in a small because the addition of the Cr pH drop (0.1 to 0.2 units) from the acidified Cr3+ stock solution. To avoid this problem, several experiments were conducted in which the aliquot of MnOZ stock suspension was added to the Cr3+ solution with the pH pre-set at 4.5. Experiments done in this reverse order did not show any appreciable change in pH upon addition of the Mn02 or as the reaction proceeded. This would seem to support the hypothesis of a proton neutral reaction at pH 4.5, however this conclusion is tenuous because of the potential buffer effect of the C02 system. 72 1.3 THE EFFECT OF CHANGING REACTANT CONCENTRATIONS 1.31 Introduction The effect on the reaction rate of sorbent/sorbate ratios was studied by holding either the initial Cr3+ or Mn02 concentration. constant, and. then 'varying' the other parameter. This was done in systems where 1) the molar concentrations of both reactants was similar, and 2) where one reactant was in a large excess (essentially constant). Most of the data from 2) is presented in the section on reaction. kinetics. The purpose of this section is to explore qualitatively, the effects on the oxidation rate of varying either the sorbent (MnOz) or the sorbate (Cr3+). From this data, the implications with respect to the nature of the reaction were evaluated. 1.32 Results and Discussion The effect of the Mnoz loading, when the concentrations of reactants are similar, on the oxidation of 9.6 X 10"6 M (0.50 mg/l) Cr3+ is shown on Figure 15. The amount (rate) of Cr oxidized increases as the initial MnOZ concentration is increased from 4.6 X 10’6 to 2.9 X 10"4 M. This result is confirmed from the measurements of Mn in solution as the reaction proceeds, shown on Figure 16. The dissolved Mn data for the highest MnOZ loading was not included on this figure because significant adsorption of released Mn2+ 73 12.0 _ 04.6x1o-6M A9.1x10-6M 01.9x10-5M x2.9x10-5M +2.9xm-4M 10.0 -I a: 8.0 . 9. f 5.0 - \ S L A {r 4.0 - 0 240 d f __E a 040 I I I f T r I 0 5 1O 15 20 25 30 35 40 ELAPSED TIME (MINUTES) Figure 15. Effect of MnOz Loading on Cr Oxidation Rate 20.0 04.64 x 10-6M A9.1o x 10-6M 018.6 x 10-6M x29.1 x 10-6M 16.0 ~ A 7‘ 1D 3, 12.0 J X E V 8.0 “ A _.=. e 1.: C 2 4.0 ‘ f2: {.1 00 r T 1 rk T r 1 0 5 10 15 20 25 30 55 40 ELAPSED TlME (MINUTES) Figure 16. Effect of MnOz Loading on Mnoz Dissolution Rate 74 occurs at this loading. The rate of the oxidation-reduction reaction increases nearly proportionally to MnOz loading until the Mn02 loading exceeds 1.9 X 10"5 M beyond which a further increase in the MnOz loading results in a smaller increase in the reaction rate. This relationship is more clearly illustrated in Figures 17 and 18. These graphs show the amount of Cr oxidized and Mn reduced, respectively, at the 1.0 and 4.0 minute time intervals (i.e. the fast part of the reaction) for the various MnOz loadings. The graphs show that the amount of Cr oxidized, or Mn reduced, is linearly dependent on the MnOz loading up until the initial MnO2 loading is about 1.9 X 10"5 M, at which time the molar Mn02 to Cr3+ ratio is about 2. Increasing the initial oxide loading further results in an increase in the initial oxidation rate that is not proportional to the increased loading. As shown on Figure 17, after this MnOZ to Cr3+ ratio is attained, an order of magnitude increase in MnOz loading results in only a small increase in the amount of Cr oxidized at either time interval. The distinct change in the linear relationship between reaction rate and MnOZ is not "solution-like" reaction behavior and provides evidence that the reaction is adsorption controlled. The linear rate dependence at low MnO2 loadings (below MnOZ/Cr3+ =2) suggests the reaction rate in this region is limited by the initially available surface area (and hence the concentration of surface sites). The break in this linear behavior for Mn02:Cr3+ ratios 75 10.0 800 ‘ ..- ..- —— ... a, .— "A A“ -' "' " "1’ 2 6.0 .. x i _________ :1 ,~ 40« ,D""'“”‘ $2 0 2.0 . 00100000 firm 1 minut0 A0lopsed firm 4 minutu 000 T T I T T 0.0 50.0 100.0 150.0 200.0 250.0 500.0 Mn02 LOADING (M/L X 10-5) Figure 17. Amount of Cr Oxidized at 1 and 4 Minutes as a Function of MnOz Loading 15.0 00100000 firm 1 m1nut0 180100000 111110 4 minut00 12.0 ~. ,A a- ’ ’ ’ A / «0 A 9.0 « x < 3. so . A . ’- “ ‘0 E , ~— —- z 3.0 .. f 9/0 0.0 T T I T I 0.0 5.0 10.0 r 15.0 20.0 25.0 30.0 141102 LOADING (M/L x 10-5) Figure 18. Amount of Mn02 Reduced at 1 and 4 Minutes as a Function of MnOZ Loading 76 greater than two suggests that there may be a relationship to the initial binding capacity of the MnOz for Cr. By assuming the reaction is adsorption controlled, the following explanation is proposed. Below the critical oxide loading there are insufficient sites for the adsorption of Cr, at that particular pH. So increasing the oxide loading in this range causes an increase in adsorption (and thus oxidation rate) pr0portional with the increase in binding sites. As the oxide loading is increased, there reaches a 3+ which can be point where the fraction of initial Cr adsorbed at pH 4.5 becomes a maximum. Increasing the MnOz loading beyond this point would then result in very little increased oxidation, as there is already enough oxide to adsorb the maximum amount of Cr at this pH. Therefore, even large increases in the MnOZ loading will result in very little change in the oxidation rate. There are some difficulties with this interpretation. A molar MnOz to Cr3+ ratio of 2.0 as the condition necessary for maximum fractional adsorption would suggest that the binding capacity of MnOZ for Cr3+ at pH 4.5 is much greater than for other trace metals. However, the kinetics of adsorption, which may be slow as compared to the oxidation reaction, and the formation of new oxidation sites during the reaction may be complicating the actual relationship between the original binding capacity of the oxide suspension and the rate of the oxidation reaction. In other words, the quantitative relationship between maximum 77 fractional adsorption and the initial oxidation rate is not clear. It is clear however, that the rate dependence is very responsive to the amount of Mnoz present, and that this rate dependence changes greatly at a critical value, the latter at least suggests the reaction is adsorption controlled. The experiments described above were conducted with a 3+ constant initial Cr concentration with the MnOz loading varied. Experiments were also conducted to explore the 3+ concentrations , both effects of varying the initial Cr with MnOZ in excess and with both reactants in similar concentrations. The effect of the initial Cr3+ loading on the oxidation rate, in the presence of excess MnOz, is shown on Figure 19. The data demonstrate that rate of oxidation increases with increasing Cr loading. Figure 20 shows a graph of the amount of Cr oxidized in 30 seconds versus initial Cr. There is a linear trend between the Cr loading and the initial oxidation rate. With excess MnOz, the fraction of adsorption remains at a maximum so the addition of more Cr3+ results in an equivalent increase in the amount adsorbed. When the surface sites are not in great excess, the rate dependence on Cr loading shows a more complicated relationship. Using an initial MnOz concentration of 9.1 X 10"6 M, experiments were conducted in which the Cr concentration was varied from 3.8 to 11.5 X 10'6 M. The results are shown on Figures 21 and 22 for the Cr oxidized 78 15.0 006 - 3.5X10-5M ACri - 7.7X10-6M ._ O o rt a <>/ 2' 9.0 '1 / : ___._.. A A’— / 3 6.0 . A ’52 3.0 /0 000 T T U r r T 0 1 2 5 4 5 5 ELAPSED TIME (MINUTES) Figure 19. Effect of Cr3+ Loading on Cr Oxidation Rate with MnOZ in Excess 10.0 ELAPSED TIME 30 SECONDS 0.0 « 8‘ § 6.0 « X § v 4.0 ‘ 2’: 2.0 . 0.0 . . 0.0 2.0 4.? 5.0 0.0 10.0 12.0 14.0 Cr(lll) LOADING (M/L x 10-6) Figure 20. Amount of Cr Oxidized in 30 Seconds as a Function of Cr + Loading with Mn02 in Excess 79 30 03.65 x 10-6M J A169 x 10-6M 25 011.5 x 10-6M 3 ii 20 4 " 1.5 « <. c 5 §;14>4 ’ 0.5 4 000 T T U TU 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 6.0 ELAPSED TIME (MINUTES) Figure 21. Effect of Cr3+ Loading on Cr Oxidation Rate with Similar Reactant Concentrations 5.0 03.55 X 10-6M £57.59 X 10-6M 011.5 X 10-6M 3.0 -‘ 2.0 7 . M001) (H/L X 10—6) 1.0T 0.0 I 0.0 2.0 , 4.0 6.0 5.0 cI-(III) LOADING (M/L x 10-6) Figure 22. Effect of Cr3+ Loading on MnOZ Dissolution Rate with Similar Reactant Concentrations 80 and MnOZ reduced, respectively. Increasing the Cr loading from 3.8 to 7.7 X 10'6 M resulted in a small increase in the oxidation rate, although the amount oxidized within the first minute was almost identical. A further increase in the Cr loading to 12.0 X 10'6 M resulted in a significant decrease in the oxidation rate. This decrease in reaction rate with an increase in one of the reactants is in certainly in direct contrast to solute- solute reactions as increasing one of the reactants would drive the reaction forward. The only apparent explanation for this type of behavior is that the reaction is adsorption controlled. At a given pH, the amount of adsorption (fractional adsorption) is controlled by the sorbent/sorbate ratio. As this ratio decreases the fractional adsorption at this pH will decrease (i.e. the adsorption edge shifts to a higher pH), even though the concentration of one of the reactants has increased. This relationship may explain the decreasing rate of oxidation with the decrease in the sorbent/sorbate ratio. 1.4 KINETIC CONSIDERATIONS 1.41 Introduction It is clear from the preceding discussions that the oxidation-reduction reaction between Cr3+ and MnOZ is not easily' described by* adsorption theory. Because. of the apparent autocatalytic nature of this reaction a 81 thermodynamic description is not entirely adequate. To attempt to gain further insight into this reaction, a kinetic analysis of some of the rate data was conducted. 1.42 Methods From the general stoichiometry of the reaction and the irreversible nature of the reaction, the experimental rate law can be written as follows: (1) d[MnOz]/dt = 2/3d[Cr3+]/dt = -k[Mn02]a[Cr3+]b where; [Mngz] is the concentration of Mngz 1 [Cr is the concentration of Cr k is the rate constant a,b are exponents reflegting the reaction order of MnOz and Cr +, respectively An attempt was made to determine k, a and b. The reaction order with respect to MnO2 (a) was determined by the method of initial rates (Gardiner, 1972). This method requires that all but one reactant remain essentially constant during the reaction so that the initial rate of change of this reactant can be used to determine its reaction order. To determine the reaction order with respect to MnOz, experiments were conducted in which the initial Cr3+ concentrations were in large molar excess to the MnOZ concentrations. Thus the Cr3+ concentration remained essentially constant during the reaction while the Mnoz was reduced. 131 this case, the rate law in ‘terms of the dissolution of the MnOz can be represented as: 82 (2) d[Mn2+]/dt = -k[Mn02]ta where: [Mn2+] is the amount of oxide dissolved [MnO ] is amount of MnO remaining 2 t . 2 at any time, t The dissolution of MnOZ was measured in a set of 3+ where the initial 5 experiments containing 9.6 x 10'5 M Cr MnOz concentration was varied from 4.8 x 10"6 to 1.4 x 10' M. From this data, the initial rate of the dissolution reaction was determined. The initial rate was estimated by two different methods to ensure accuracy. The first method was to plot the curve of Mn2+ measured versus the elapsed time of the reaction and then determine the slope of the line drawn through the linear portion of the reaction. In most cases the first three data points yielded a roughly linear relationship. The data used for determining the initial rates and the calculations of the slope (via linear regression) are shown on pages 224 to 227 in Appendix III. The second method of estimating the initial rate of dissolution with this data set was to fit a curve to the dissolution ‘versus time data and then to take the lst derivative of the equation for the curve at time zero. Curve fitting was conducted with the computer program STATGRAPHICS, 1985. Several curve fitting equations were tried but the data were best fit with the formula: (3) [Mn2+] = A - B * EXP(-C * time) The plots of the fitted curves with correlation coefficients, the values of the curve equation coefficients 83 and the first derivatives at time zero are presented in Appendix III, pages 228 to 233. The order of the reaction with respect to MnOz was determined from these initial rates from the slope of the best-fit line through the plot of the log of the initial rate versus the log of the initial MnOz loading. These plots and the regression calculations are presented in Appendix III on pages 234 and 235, for the two methods of determining the initial rates respectively. The experimental rate constant, kex' was also determined from ‘the experimental data above, with Cr3+' present in excess. This data set will be referred to as data set A in subsequent discussions. Additionally, two other sets of 3+ experiments were conducted with Cr present in excess. One contained 9.6 x 10'5 M, identical to the experiments described above, the other contained twice as much Cr3+, or 1.9 x 10'4 M. The initial MnOz loadings were varied from 4.7 x 10""6 M to 5.0 x 10'5 M in both sets of experiments. These data sets will be referred to as data sets B and C respectively. The reaction order with respect to Mnoz was also determined for these data sets (via the curve fitting method). The fitted curves and calculations are presented in Appendix III, pages 236 to 241 and 242 to 247, respectively. The rate constant was determined by the method used by Stone and Morgan (1984) from the integrated form of the rate law. The amount of MnOz remaining at any time t ([MnOZJt), 84 is simply the initial oxide loading less the amount of Mn2+ in solution as the result of reduction, or: (4) [Mnozlt = [Mnozli - [Mn2+] Substituting for [Mn021t in the rate law presented in equation (2) and integrating yields the equation: (5) ln([MnOz]i - [Mn2+]) - ln[MnOz]i = -kt If this equation is valid, then a graph of ln([Mn02]i [Mn2+]) - [Mn02]i versus time should result in a straight line with a slope equal to the experimental rate constant, kex' The graphs for all three sets of experimental data are presented in Appendix III, pages 248-250. The reaction order with respect to Cr3+ was more difficult to address. An attempt was made to employ the initial rates approach by conducting experiments with MnOZ in large excess in order to maintain a constant MnOz concentration. Because it must actually be the number of surface sites that are constant, a very large molar excess of MnO2 is required to use this technique. Under these conditions, the oxidation reaction proceeded too quickly to obtain enough data to determine the initial rate. Because of this result this approach was abandoned. 3+ was evaluated The reaction order with respect to Cr using a method similar to that presented in Stone and Morgan (1984), which was a study of the reduction of Mn oxide by the organic molecule hydroquinone. These investigators 85 determined the reaction order with respect to hydroquinone, which was present in excess in their experiments, by determining the effect on the experimental rate constant of varying the amount of excess hydroquinone. The reaction order with respect to hydroquinone was then determined from the slope of a plot of the log of the rate constant versus the log of the hydroquinone loading. 3+ was determined The reaction order with respect to Cr by the change in the log of kex' divided by the log of the change in the Cr3+ loading, using the data from data sets B and C described above. 1.43 Results The calculated reaction orders with respect to MnOZ in 3+ are shown on Table 1. All of the presence of excess Cr the slopes were ‘very close to 1.0 indicating that the reaction is lst order with respect to MnOz. The reaction orders determined from data set A, 1.03 and 1.08, show that both methods chosen to determine the initial reaction rate were adequate. A first order rate dependence with respect to the adsorbing substrate has been observed in other investigations of multi-phase oxidation-reduction reactions (Stone and Morgan, 1984; Postma, 1985). This result is expected for simple adsorption controlled reactions as the amount of adsorption is directly proportional to the surface area of the sorbent in the system. 86 TABLE 1 REACTION ORDER WITH RESPECT TO HnOZ Data Range of Mnoz Regggégn Method of s t [Cr3+] Loading Initial Rate e (M) T Determination Slope r2 A 9.6x10’5 4.8x10‘5 to 1.03 0.963 slope of 1.5x10'5 first 3 data points A 9.6x10"S 41.13::10'6 to 1.08 0.967 lst derivative 1.5x10'5 of fitted curve B 9.6x1o‘5 4.7x10‘5 to 0.90 0.998 lst derivative 5.0x10’5 of fitted curve c 1.9x10‘4 4.71:10’6 to 0.99 0.996 lst derivative 5.0::10“5 of fitted curve TABLE 2 EXPERIMENTAL RATE CONSTANTS Data c 3+ R f k A k Set [ r ] ange 0 ex verage ex A 9.6x10'5 0.141-0.19s 0.166 B 9.6x1o'5 0.143-0.173 0.159 c 1.9x1o‘4 0.075-o.134 0.111 \\ 87 The plots of ln([Mn02]i - [Mn2+]) — ln[MnOZ] versus time were only linear for about the first half of the reaction. Thus the experimental rate constant, k was determined ex' from this part of the reaction. A typical plot of this type is shown on Figure 23. The value of kex was determined from a linear regression of the first part of the reaction as depicted on the Figure. This calculation was conducted for data sets A, B and C. The average value and range of values for kex determined from each data set are summarized on Table 2. The plotted data and regression calculations are presented in Appendix III, pages 251 to 263. The non- linear behavior exhibited by the data after the reaction had prOceeded demonstrates that the rate law is not valid after considerable dissolution of the oxide had taken place. Interestingly, Stone and Morgan (1984) reported very similar results for the reduction of Mn oxide by hydroquinone. It is likely that this result is caused by a change in the surface properties of the Mn oxide after significant dissolution has taken place. 3+ The reaction order with respect to Cr was computed by the equation: (change in log kex) Reaction Order = -------------------------- (change in log Cr loading) The values of kex for the 9.6 x 10"5 M and 1.9 x 10'4 M Cr3+ loadings were 0.159 and 0.111, respectiveLy. These result 3+ in a reaction order of 0.5 with respect to Cr . A reaction order less than 1.0 was also found in the study by Stone and 88 0.0 .1 SLOPE 8 EXPERIMENTAL RATE CONSTANT -O.5 it? I; z -1.0 d D I E v D 33 -L5‘ 0 -2.0 T T I 0.0 5.0 10.0 15.0 20.0 800000 Time (minutes) DMnO2I - 1.23X10-5 M Figure 23. Typical Plot for Determination of Experimental Rate Constant 89 Morgan (1984) for systems with hydroquinone in large excess. Based on these results, the experimental rate law fOr the reduction. of jMnOZ by the oxidation of Cr3+, with Cr3+ present in large excess, can be written as follows: (6) d[Mn02]/dt = “kex[Mn02]1'°[Cr3+]°°5 where: ke depends on tbs initial [Cr3+$ and has units of (moles)- ' (minutes)'1° The lst order reaction with respect to [MnOz] is in agreement with other data from adsorption controlled reactions. The square root dependence on the Cr3+ concentration presents some interesting implications with respect to the reaction mechanism. It may be related to the fact that each Cr3+ ion must lose 3 electrons and each Mn4+ can only except 2. This suggests there may be intermediate reaction products and several reaction steps which are not considered in the overall thermodynamic reaction equation. A detailed analysis of reaction mechanisms and rate- controlling steps was outside of the scope of the present research. 1.5 SUMMARY AND CONCLUSIONS The general nature, stoichiometry and kinetics of the oxidation of Cr3+ by MnOZ has been evaluated by experiments under controlled laboratory conditions. Experiments were conducted to determine the rate and extent of the reaction by measuring the amount of Cr6+ formed by oxidation and the 90 amount of Mn2+ solubilized upon reduction for a variety of initial reactant concentrations. The important results from these experiments are summarized as follows: 1. The oxidation of Cr 3+ is a relatively fast reaction, with most of an initial 9.6 X 10"6 M Cr3+ spike being oxidized within 30 minutes in a 1.9 X 10'"5 M MnOZ suspension. The oxidation of Cr3+ occurs even in systems with very low MnO2 concentrations; which means the reaction occurs when the potential for adsorption (as compared to other metals) should be very low. The reaction proceeds until all of the Cr is oxidized or nearly all of the MnOZ has been reduced. The reaction exhibits parabolic kinetics, with much of the reaction occurring at a nearly linear rate within the first several minutes, followed by a decreasing rate of reaction with time. Except for the initial stage of the reaction, the stoichiometry of the reaction products shows that 3 moles of MnOz are reduced for every' 2 ‘moles of Cr oxidized. This confirms the predicted stoichiometry for 3+ the reaction of Cr with Mn4+ from the oxide surface. The rate of the reaction is dependent on the initial ratio of the reactants. Increasing the concentration of 91 MnOz results in a linear increase in reaction rate up until the initial MnOZ to Cr3+ molar ratio is approximately 2. Further increasing the MnOZ concentration to higher MnOZ to Cr3+ ratios results in increasingly smaller increases in the reaction rate. 6. With Cr3+ present in large molar excess to MnOz, the reaction exhibits first order rate dependence with respect to MnO2 and a square root dependence with respect to Cr3+. From the above results two general conclusions were made. The first is that Cr3+ is oxidized by Mn"+ ions which are part of the oxide structure. This conclusion is supported by the strong adherence of the reaction products to the predicted reaction stoichiometry; The predicted stoichiometry is maintained throughout the reaction except in the first several minutes when the Mn2+ to Cr6+ ratio is lower than the predicted 1.5. It is proposed that the release of reduced Mn to solution is slow with respect to the overall reaction causing this deviation from stoichiometry when the reaction is fast. The second main conclusion drawn from the data was that if the reaction is adsorption controlled, then 1) the affinity of Cr3+ for the Mn02 surface must be very high, and 2) the reaction must be autocatalytic, with new surface sites being generated by the reduction and dissolution of the Mn oxide. This conclusion was drawn from the oxidation 92 data at low MnOz to Cr3+ ratios. The adsorption of metals, especially at low pH, is very much dependent on the ratio of the oxide to the metal. The adsorption of metals such as copper and zinc by MnOZ at low pH's is minimal, or even undetectable when the oxide concentrations are less than or equal to the metal concentration. However, the oxidation of Cr proceeded strongly under these conditions. This suggests that the affinity of Cr3+ for the Mnoz surface is much higher than for other metals. Even if the affinity of Cr3+ for the MnOz surface were high, an adsorption controlled reaction could not proceed to the extent shown by the experimental data unless new surface sites were generated by the reaction. Conceptually, this result is consistent with the predicted reaction mechanism. 4+ ion in the oxide structure is reduced, it becomes As an Mn unstable within the mineral structure and is released to solution. Effectively, the oxide particulate begins to dissolve. The loss of a coordinated Mn ion must result in a vacancy within the mineral lattice, which is essentially just another surface site. From a thermodynamic standpoint, the Cr3+ ions in solution are not only reactive with respect to the binding sites initially present on the oxide surface, but also with respect to all of the potential surface sites which can be generated during the reaction. An alternative hypothesis to the second conclusion above, which explains the apparent discrepancies between the experimental data and available. data on adsorption 93 equilibria was also investigated and rejected. The alternative hypothesis was that the reaction was not adsorption controlled and the reaction was comparable to a solute-solute, or metal-ligand type interaction. This hypothesis was rejected based on the influence of the initial MnOz to Cr3+ ratio on the reaction rate. The reaction rate slowed. down with an increased Cr3+ concentration when Mn02 was not present in a large excess. This result is consistent with adsorption equilibria and in direct contrast to solute-solute interactions. Investigation of the experimental rate law showed that the reaction is first order with respect to MnOz and has a square root dependence with respect to Cr3+. The lst order rate dependence with respect to MnOz is consistent with other surface controlled reactions with MnOz. The square root dependence with respect to Cr suggests the reaction may be complex, with several elementary steps. PART 2.0 THE EFFECT OF SOLUTION CHEMISTRY ON THE OXIDATION OF CR(III) BY MN OXIDE 2.1 INTRODUCTION This part of the report describes the effects of the solution chemistry on the rate and extent of the oxidation/reduction reaction between Cr3+ and the MnOZ surface. Experiments were conducted to explore the role of solution composition which may influence the rate and extent of this reaction in natural aqueous systems. In particular, the effects of pH, dissolved oxygen (presence or absence), ionic strength, nature of the swamping electrolyte and a competing adsorbed trace metal (copper) were examined. This part of the project was designed to provide further insight into the potential importance of this redox reaction in the environment, as well as to further define the fundamental nature of the reaction. 2.2 THE EFFECTS OF SOLUTION pH 2.21 Introduction The potential effects of hydrogen ion activity (pH) on the adsorption-oxidation-desorption of Cr at the MnOz surface are numerous. The potential effects on the system upon a change in pH include: 1) changes in the speciation of the oxide surface sites (i.e. surface hydrolysis), 2) a change in the electrostatic properties of the oxide surface, 3) changes in the speciation of the dissolved metal species, 94 95 4) the decrease in Cr3+ solubility with increasing pH and 5) the instability of the MnOZ at low pH. With the Cr concentrations used in the experiments in this research, effects 1-3 above could only be studied in the pH range of approximately 3.0 to 5.0, as the formation of Cr hydroxide at higher pH's overwhelms all other pH effects. This narrow working range presents some difficulty in determining whether there is a pH dependence in the oxidation rate. The effects of 4 and 5 above were studied outside of this pH range at pH's less than 3.0 and greater than 5.0. From past work on the adsorption of other trace metals on MnOz (McKenzie, 1980; Gadde and Laitinen, 1974), it has been shown that the quantity of metals adsorbed per unit weight of oxide generally increases within the pH interval of 3.0 to 5.0. This suggests there are more sites available with increasing pH and/or the energetics are becoming more favorable for adsorption. If Cr3+ adsorbs similarly to these other metals, it might be expected that the oxidation rate of adsorbed Cr would increase as the pH increased. However, given that the oxidation reaction exhibits characteristics considered atypical compared to the adsorption of other metals, it is not clear whether a pH dependence should be detected experimentally. To explore the effect pH on the oxidation rate at low pH's, duplicate experiments were performed at pH's 1.55, 3.00 and 4.50. The concentrations of reagents used were: Mno2 = 1.9 x 10‘5 M, Cr3+ = 9.6 x 10‘6 M and the matrix was 96 0.10 M NaNO3. These experiments were conducted open to the atmosphere. The pH was adjusted to the desired level with 0.10 N HNO3 approximately 10 minutes before the addition of 6+ 2+ the Cr3+ spike. Dissolved Cr and Mn were then measured in filtered aliquots withdrawn over time. 2.22 Results and Discussion The amount of Cr oxidized and the amount of Mn solubilized for each experiment are shown on Figures 24 and 25 respectively. The amount of Cr oxidized over the first ten minutes of the reaction at pH 3.0 and 4.5 was nearly the same. The amount oxidized at pH 1.55 was significantly less over this same time period. The similar rates of oxidation at pH 3.0 and 4.5 was supported by the amount of Mn2+ formed during the reaction (see Figure 25). However, the quantity of Mn solubilized at pH 1.55 was much greater than that at the higher pH's, inconsistent with the lesser amount of Cr oxidized. This result was not unexpected as MnOZ has been shown to be unstable at pH's lower than 3.5 (Murray, 1974). To verify the role of oxide dissolution in this apparent discrepancy, another set of identical experiments were performed at pH 1.65 and 4.5. A blank (not spiked with Cr) MnOZ suspension was also set to pH 1.65 and aliquots were withdrawn over the time of the experiment, filtered and measured for Mn2+. From the blank the amount of Mn2+ formed from the dissolution of the oxide was subtracted from the total Mn solubilized during the redox reaction with Cr. The 97 « 5.0 . Cr(Vl) MOLAR (x 10—6) 0.0 q T T o 2 4 6 a 10 12 TIME (MINUTES) d .4 Figure 24. Effect of pH on the Oxidation Rate of Cr3+ by Mn02 10.0 8.0 . ’15 S? 6.0 - 5. g 40 4 5 2 2.0 « 00 T T I; T fi f o 2 4 6 a 10 12 TIME (MINUTES) Figure 25. Effect of pH on the Dissolgtion Rate of MnO2 During the Oxidation of Cr + 98 quantity of Cr oxidized and the corrected values of Mn solubilized are shown on Figures 26 and 27, respectively. In agreement with the previous experiments, the rate of Cr oxidation was significantly less at the lower pH. The corrected rate of Mn2+ production (Figure 27) is also lower at pH 1.65 than at pH 4.50. The molar ratio of Mn2+ to Cr6+ for the reaction products at both pH's is very close to the predicted 1.5. Thus the correction obtained from the blank appears to result in the true rate of reduction caused by Cr oxidation at the low pH. Approximately 20% of the original MnO2 suspension dissolved in the blank over the time interval of the experiment. This rate of dissolution can probably account for the lower oxidation rate at pH 1.65, especially since the smaller more reactive particles may be the first to dissolve. With the exception of this dissolution effect, there does not appear to be a measurable pH effect on the reaction rate at pH's below 5. 2.3 THE OXIDATION OF SOLID CR-HYDROXIDE 2.31 Introduction The oxidation of Cr by MnOz at higher pH's (i.e. greater than 6.0) will be influenced by the formation of Cr(OH)3 solid. The formation of this species has been ignored in some previous studies of Cr-Mn oxide redox reactions. Studies by Rai et al. (1986), using reagent MnOZ, and James and Bartlett (1983), using Mn oxide bearing soils, have 99 6.0 DpH-1.65 6.0 23‘ I o 5. g 4.0 .. O z 3 0 2.0 4 0.0 I T T T T7 0 2 4 6 5 1O 12 TIME (MINUTES) Figure 26. Effect of pH on the Oxidation Rate of Cr3+ by Mno2 12.0 H-1.65 100 . Up A 0H-4.5 Mn(II) MOLAR (x 10-6) 4.4 .1 0.0 I T o 2' 4 6 e 10 12 TIME (MINUTES) Figure 27. Effect of pH on the Dissolution Rate of MnO2 During the Oxidation of Cr I with Blank Corrections 100 shown that oxidation of chromium hydroxide does occur, but at a much slower rate and reduced extent than dissolved 3+ is believed to be the chromium species. As dissolved Cr species interacting with the MnOz, the formation of a solid Cr phase with very low solubility should severely limit the oxidation/reduction reaction. Since Cr solubility should be controlled by hydroxides under the pH conditions of most natural systems, the oxidation by MnO2 may not be environmentally significant. However, if oxidation by Mn02 can proceed under these conditions, this reaction may be very environmentally significant. To determine the extent and rate of the oxidation of Cr hydroxide by the MnOz used in the current investigation, a set of long term experiments were conducted. Both aged and freshly precipitated Cr(OH)3 solid were used. For the freshly precipitated hydroxide, one liter of 1.92 X 10'5M Cr(NO)3/0.10 M NaNO3 solution was adjusted to pH 8.2 with NaOH. A light greenish precipitate was seen to form upon adjustment of the pH. This solution was allowed to set for approximateky 60 minutes with occasional pH adjustments to maintain the pH above 7.5. After this time, an aliquot of MnO2 stock solution was added to bring the total Mn concentration to approximately 9 X 10'3 M. The suspension was sampled periodically over the next 48 days and analyzed for Cr6+. The aged Cr(OH)3 was prepared as above but was allowed to set for 48 hours with occasional pH adjustments 101 before the addition of the MnOz. The solution was sampled periodically over the next 46 days. 2.32 Results and Discussion The results of these experiments are shown on Figure 28. As can be seen from this graph, small amounts of Cr were oxidized in both experiments, with a maximum of 10.6 X 10'7 and 2.9 X 10'7 moles oxidized in the fresh and aged systems, respectively. More oxidation occurred with freshly precipitated Cr(OH)3, similar to the results of James and Bartlett (1983). These investigators interpreted this result as evidence for a residual positive charge on the fresh hydroxide, allowing adsorption onto the Mn oxide surface. Although this cannot be verified by the present results, the fact that the oxidation was not complete until at least 10 days time would suggest that the larger quantity oxidized in the freshly precipitated hydroxide solution was not simply due to incomplete precipitation of Cr hydroxide, as any remaining dissolved Cr species would have likely been oxidized very quickly. It should be noted that only about 5% of the freshly precipitated Cr(OH)3 was oxidized throughout the duration of the experiment. Still, this quantity is almost two orders of magnitude higher than the solubility of Cr3+ in this pH range. In summary, the formation of Cr hydroxides greatly reduces the rate and extent of the oxidation reaction. However, the extent of 102 12.0 C] CI 10.0 . 77‘ 8.0 - 9 x g 6.0 30 g I: g 4.0 4 Y o A A 2.0 - A DFRESH Cr(OH)3 AAGEO Cr(OI-I)3 00 i T T T T 0 10 20 50 40 50 ELAPSED TIME (DAYS) Figure 28. Oxidation of Fresh and Aged Cr(OH)3 by MnO2 103 the reaction is still sufficient to increase the solubility of Cr significantly. 2.4 THE EFFECT OF DISSOLVED OXYGEN 2.41 Introduction Research by Nakayama et al. (1981) on the oxidation of Cr3+ by a trivalent Mn oxide, MnO(OH) , showed that the presence of dissolved oxygen had a large effect on the oxidation rate, with the reaction rate much slower in the absence of oxygen. These investigators did not propose an explanation for this behavior, but termed the role of the Mn oxide as having a "catalytic" effect in the oxidation of Cr. In contrast, Rai et al. (1986) did not find any effect of dissolved oxygen content on the oxidation of Cr by reagent MnOz. To determine whether dissolved oxygen would influence the oxidation rate of Cr in the present study, duplicate experiments were performed in open air and in Nz-purged experimental systems. The 02 deficient conditions were established by bubbling N2 gas through the solution in the reaction vessel, which itself was contained in an N2 purged glove box. The dissolved 02 was measured by an Orion model 97-08-99 dissolved oxygen electrode. The concentration of dissolved oxygen was below 0.10 mg/l during the duration of the experiment. The MnO2 suspension was shown to be stable under these conditions (see Methods). The concentration of 104 reactants used were: Cr3+ = 9.6 x 10"6 M, Mno2 = 1.9 x 10‘5 M. The solution matrix was 0.10 M NaNO3. 2.42 Results and Discussion The rates of Cr oxidized and Mn solubilized for both open-air and Nz-purged systems are shown in Figures 29 and 30, respectively. The oxidation rate of Cr can be seen to be very similar in air and N2. The rate in the absence of 02 actually appears to be slightly greater. However, the rate of Mn2+ produced is almost identical for both experiments, suggesting the slight difference in the oxidation rates may have been analytical. From these results it is concluded that dissolved oxygen does not play an important role in the redox reaction, at least under the experimental conditions investigated. These results are consistent with the proposed reaction mechanism, which does not include a role for dissolved oxygen. 2.5 THE EFFECT OF IONIC STRENGTH 2.51 Introduction There are several potential effects of changing ionic the strength with an inert electrolyte on adsorption reactions. These include: 1) changes in the activity of dissolved species, 2) changes in the electrostatic potential at the oxide surface and 3) changes in colloid stability (Swallow et al, 1980 and Morgan and Stumm, 1964). In 105 12.0 10.0 - A B e D 5 I? 8.0 r (a 9 A 5 CI § 60~[9 ‘2’ A c 4.0 -D > Y 0 A DOPEN TO ATMOSPHERE 20 P AN2 PURGED 0.0 k T I I T I T O 20 40 60 80 100 120 140 TIME (MINUTES) Figure 29. Rate of Cr Oxidation by MnOz in Open-Air and N2 Atmospheres 20.0 160~ A e C] §~ [9 8 120 - m 5. E 0: 5 [9 g 80 - Q Q g DOPEN TO ATMOSHERE 4.0 ~13 AN2 PURGED 0.0 g 1 I T fi 1 I o 20 4o 60 80 100 120 140 TIME (MINUTES) Figure 30. Rate of MnOZ Reduction by Cr3+ in Open-Air and N2 Atmospheres 106 general, past studies have shown a general tendency for decreased. adsorption.*with increasing ionic strength (see Appendix I). However, the adsorption of many trace metals on oxide surfaces is apparently unaffected by changes in ionic strength. In particular, there appears to be little competition for surface sites between trace metals and background electrolytes and the changing electrostatic conditions induced. by 'these electrolytes appear' to Zhave little influence on specifically adsorbed trace metals (Swallow et al. 1980). Several studies (Morgan and Stumm, 1964: Posselt et al., 1968) have shown that adsorption kinetics on Mn oxide surfaces are controlled by ionic strength. It is believed that the flocculation of the oxide, brought about by the collapse of the diffuse layer with increasing ionic strength, necessitates that ions diffuse through the flocculated suspension to find an available site for adsorption. To determine whether the rate of oxidation of Cr by MnOz is influenced by ionic strength, several sets of duplicate experiments were conducted in which the concentration of the swamping electrolyte was varied over several orders of magnitude. The first set of experiments measured the rate of oxidation of Cr3+ and the solubilization of Mn in solutions of 0.005, 0.05 and 0.50 M NaNO3, where in this case the molarity of the electrolyte solution is equivalent to the ionic strength. The concentrations of reactants were 1.9 X 10'5 M and 9.6 X 10'6 107 M for MnOZ and Cr3+, respectively. Another set of experiments were conducted with 0.01 and 0.10 M Ca(NO3)2 (ionic strength 0.03 and 0.30, respectively) to set the ionic strength. All experiments were conducted at pH 4.5. 2.52 Results and Discussion The results for oxidation and reduction rates in the NaNO3 matrices are shown on Figures 31 and 32. The oxidation rate of Cr3+ and the corresponding reduction of Mn is significantly slower as ionic strength increases. Similar results are seen for the rate of reaction in the Ca(NO3)2 matrices (Figures 33 and 34) although the rate difference is not as pronounced and is only apparent during the early stages of the reaction. It should be noted that these two sets of experiments ‘were not conducted close enough together in time to allow direct comparison between the two electrolytes. Given the data from past research, the most likely cause of the decreasing oxidation rate with increasing ionic strength appears to be the flocculation of the oxide. The flocculation of the MnO2 suspension was easily visible in the concentrated matrices and the flocs became larger as the ionic strength was increased. This effect was more pronounced for the NaNO3 solutions which is somewhat surprising since divalent ions usually produce a much higher degree of flocculation than monovalent ions (Drever, 1982). 2+ A possible explanation for this result is that Ca adsorbs 108 10.0 8.0 4 "I? 9 6.0 ~ 2‘, g 4.0 . ?( 00.005M NaN03 o 2 o A0.05M NDNOJ ' 00.50M NaN03 0.0 . . . 0 5 10 15 20 25 TIME (MINUTES) Figure 31. Effect of Ionic Strength on the Oxidation Rate as Set by NaNO3 15.0 5 12.0 « T ‘ , $2 9.0 . 5, - 7‘ c: . 3 3 6.0 - g '1 ’1 C 2 D0.005M NONO3 3.0 4 A0.05M NaN03 I ; ' /’ 00.50M NaN03 00 I T T 7 T 1— o 5 10 15 20 25 TIME (MINUTES) Figure 32. Effect of Ionic Strength on the Reduction Rate as Set by NaNO3 109 10.0 v 00.01 M Co(N03)2 8.0 . A0.10 M CO(N03)2 Cr(v1) MOLAR (x 10-6) 0.0 T T T 0 5 10 15 20 25 TIME (MINUTES) Effect of Ionic Strength on the Oxidation Rate Figure 33. as Set by Ca(NO3)2 12.0 10.0 - ’cf 8.0 - 2 5. g 6.0 4 S 4.0 00.01 M CD(N03)2 § AO.10 M 00(N03)2 2.0 0.0 r I 10 1'5 20 25 TIME (MINUTES) 0 (IO-I Figure 34. Effect of Ionic Strength on the Reduction Rate as Set by Ca(NO3)2 110 on MnOz by a different mechanism than Na+, resulting in different effects on the electrostatic properties. The adsorption capacity of an oxide surface for trace metals is generally unaffected by ionic strength. However, a study by Posselt. et al., (1968) demonstrated. that the adsorption capacity of MnOZ for Ca2+ was reduced by increasing the ionic strength with NaClO4. The possibility of effects other than the flocculation of the oxide, therefore, cannot be ruled out by the present results. 2.6 EFFECTS OF THE NATURE OF THE ELECTROLYTE 2.61 Introduction In addition to the effects of ionic strength discussed above, the composition of the matrix solution may influence an adsorption reaction by "specific" effects, which are based on the composition of the solution. These specific effects may include the competition for surface sites on the oxide, and more importantly in the case of trace metals, the complexation of trace :metals by“ dissolved ligands. Complexation of the adsorbing metal can cause an increase or a decrease in adsorption of a trace metal (see Appendix I). Rai et al., (1986) showed that the solubility of Cr(OH)3- solid was unaffected by low concentrations of common ligands such as 8042', N03”, 01" and HCO3', which suggests that complexing is unimportant under these conditions. Elderfield (1970), using published stability constants, 111 concluded that complexing by ligands other than OH' would be insignificant even under sea water conditions. Although the data is scarce, all indications are that complexing of Cr should be minimal in most aqueous systems. One possible 3+ exception may be the complexing of Cr with organic ligands. Nakayama et al. (1981) demonstrated that both 3+ 3+ coprecipitation of Cr with Fe oxide and oxidation of Cr with Mn oxide could be reduced by the formation of a Cr- citric acid complex. Based on available data, it was hypothesized that changing the swamping electrolyte solution would not effect rate of the redox reaction between Cr3+ and the MnOZ surface. To test this hypothesis, duplicate experiments were performed in various swamping electrolyte solutions of identical concentrations. The first test compared the rate of MnOZ dissolution upon addition of a Cr3+ spike in 0.05 M solutions of CaClz, Ca(NO3)2, NaNO3 and NaCl. The quantity of Cr6+ was not measured in most of these experiments due to the problems created by C1 in the Cr6+ extraction procedure. The concentration of reactants were that of the "standard" oxidation experiment and the pH was 4.5. 2.62 Results and Discussion Figure 35 shows the results of these experiments. Only slight differences were found between electrolyte solutions. The reaction appears to be slightly faster in Ca(NO3)2 than in either sodium salt and slightly slower in CaClz. The 112 20.0 150 J ‘ ‘1’ 2 ‘/’/’ 3 x ’. 5 10.0 " 5 ‘§ 8 A.05 M 00(N05)2 c ,6 1 5.0 . //r. 0.05 M 00012 0.06 M NON05 X.05 M NDCI 0.0 l T T T7 T I 0 5 10 15 20 25 50 55 ELAPSED TIME (MINUTES) Figure 35. Reduction Rate of MnOz in Various Background Electrolytes ‘\ 113 slightly faster rate of reaction in Ca(NO3)2 solution is not readily explainable. One possible explanation is the apparent increased degree of flocculation of the MnOZ which seems to occur in Na salt solutions. The slower rate of reaction in CaClz matrix was not anticipated. To further explore the rate of reaction in CaCl2 solutions, duplicate experiments were performed in 0.005, 0.05 and 0.50 M CaClz solutions. The results, in terms of Mn2+ solubilized, are shown on Figure 36. There was a slight decrease in the rate of reduction as the concentration of CaClz solution was increased from 0.005 to 0.05 M. However, as the concentration was increased to 0.50 M, the rate of oxide dissolution was decreased significantly. The measurement of oxidized Cr confirmed these results for the 0.005 and 0.05 M CaC12 solutions (Figure 37). Two possibilities were considered as an explanation for this reduced reaction rate in CaClz; 1) the competition for binding sites by the high Ca concentrations and 2) the complexing of Cr3+ by C1". These possibilities were tested by comparing the reduction rate in 0.50 M CaClz to that in 1.0 M NaCl (same Cl concentration, different cation) and in 0.50 M Ca(N03)2 (same Ca concentration, different ligand). The results are shown on Figure 38. It can be seen from the graph that only CaC12 causes this large reduction in the reaction rate. This unique behavior of CaC12 was further 114 15.0 00.005 M COCI2 12.0 4 A0.05 M COCI2 a 00.50 M COCIZ ‘ T a 9 9.0 4 ‘ Z." 3 r4 g 6.0 - / 2 '11 3.0 ~ " 0.0 - '3 ' . r ‘ 1 0 5 10 15 20 25 TIME (MINUTES) Figure 36. Reduction Rate of MnO2 in CaCl2 Solutions of Various Concentrations 10.0 I 8.0 4 G c'. 4- 6.0 4 5. 3 40 4 9 00.005 M (2002 65 A005 M COCI2 2.0 0.0 T ,4 1 1 o 5 10 15 20 25 TIME (MINUTES) Figure 37. Oxidation Rate of Cr3+ in CaClz Solutions of Various Concentrations 115 200 00.50 Ca(N03)2 A1.0 M NDCI 16‘0 A 00.50 M COC|2 7.3 9 120 4 35, g 8.0 4 T? 2 4.0 4 0.0 " T T T T T T 0 10 20 30 4o 50 60 70 TIME (MINUTES) Figure 38. Reduction Rate of MnOZ in Ca(CO3)2, NaCl and CaC12 15.0 01.0 M KG! 120 . A050 M COCI2 00.50 M MgCl2 9.0 6.0 Mn(II) MOLAR (x 10-6) 3.0 0.0 T T 0 5 10 15 20 25 50 55 TIME (MINUTES) Figure 39. Reduction Rate of Mn02 in KCl, CaCl2 and MgCl2 116 verified by comparing reduction rates in 1.0 M KCl and 0.50 M MgClz solutions. The results are shown in Figure 39. From these results it does not appear that either the blocking of surface sites or the complexing of Cr3+ by C1- can explain the decreased oxidation rate in concentrated CaCl2 solution. The dramatic decrease in the rate of reaction as the Ca012 concentration is increased from 0.05 to 0.50 M as compared to the minor effect caused by increasing the CaClZ concentration from 0.005 to 0.05 M would suggest that there is a solubility control. At the present time no explanation is proposed to explain these results. 2.7 THE EFFECT OF ADSORBED COPPER 2.71 Introduction Past studies have shown that the competition between trace metals for adsorption sites on an oxide surface does not follow a simple relationship (see Appendix I). In some cases, a particular metal may adsorb at the expense of another, while other metals can co-adsorb with no detectable effects on each other. Benjamin and Leckie (1981a) presented evidence that Fe oxides have high energy sites that are specific for each metal, resulting in little competition between metals on this oxide surface. Limited research on metal competition on Mn oxides suggests that some competition may occur (Dempsey and Singer, 1980; Gadde 117 and Laitinen, 1974). If there is competition between Cr 3+ and other trace metals for the surface sites on MnOz, it seems likely that this would influence the rate of oxidation of Cr. There is no reason to pre-suppose, however, that Cr adsorbs at the same surface sites as other metals on Mn oxide. Although a rigorous treatment of this problem was outside the scope of this research, a preliminary investigation was conducted to determine whether the adsorption of a similar quantity of a competing trace metal would effect the oxidation rate of Cr. The adsorptive behavior of 0.50 mg/l Cu, Zn, Mn and Ni was investigated for a range of Mn02 concentrations at pH 4.5. The results are shown on Figure 40. From these results it is apparent that Cu adsorbs strongest under these conditions. Based on this result, and the need to maintain reasonably low MnO2 concentrations, Cu was chosen to study potential competitive effects. It should be noted that an oxide concentration of 1.4 X 10'4 M ‘was required to ensure greater than 90% adsorption of 7.9 X 10'6 M Cu. This is an order of magnitude more MnOZ than is required to completely oxidize 9.6 x 10'6 M Cr3+. To study the potential competitive effects of Cu2+ adsorption on Cr oxidation, an MnOz suspension with Mn = 1.4 X 10'4 M, was allowed to equilibrate with 7.9 X 10'6 M Cu2+ at pH 4.5. A sample was withdrawn after an hour, filtered and analyzed for Cu2+. From this 118 10.0 000 AMI) OZn XNI 504 CO I 9 x 604 2 0 1.4.1 0:1 m 8 404 i 0 2 20 4 0.0 T T T T 00 So 100 150 200 250 Mn02 m/I (X 10-5) Figure 40. Adsorption of 0.50 mg/l Copper, Manganese, Zinc and Nickel by Various Amounts of MnOz 119 2+ adsorbed was calculated. The measurement the amount of Cu system was then spiked with Cr(NO3)3 to obtain an initial Cr3+ concentration of 9.6 X 10'"6 M. The system was then sampled over small time intervals and analyzed for dissolved Mn2+, Cu2+ and Cr6+. A control experiment was also conducted which was identical to that above except no Chz+ was added to the system. 2.72 Results and Discussion The rate of Cr oxidized and Mn reduced for the Cu spiked experiment are presented on Figure 41. The oxidation of Cr was almost completed within 15 minutes. The amount of Mn2+ released was less than the predicted stoichiometric amount due to adsorption by the remaining oxide. As shown on Figure 41, a considerable amount of Cu was desorbed during the reaction. The results from the control experiment are shown on Figure 42. These results show that the rate of oxidation was very similar to that in the Cu spiked system. The rate of Mn released, however, was significantly lower. This relationship is shown more clearly on Figures 43 and 44, which directly compare the results of the control experiment with the Cu spiked experiment for Cr oxidation and Mn released, respectively. The above results seem to indicate that the rate of 3+ oxidation of Cr is not effected by the presence of 2+ adsorbed Cu . This suggests that the adsorption of Cr3+ does not take place at the expense of Cu2+, unless this 120 12.0 0 Cr(VI) A Mn(II) 0 Cu(ll) cL A 10.0 8.0 6.0 4.0 DIS. METAL (M x 10-6) 2.0 0.0 T T “IT T T T T T 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 16.0 TIME (minutes) Figure 41. Rate of Oxidation/Reduction with Adsorbed Copper 12.0 A Mn(II) DIS. METAL (M x 10-6) 00 w . r 0 . 2 4 6 6 10 12 14 16 16 TIME ' (MINUTES) T T T T T T Figure 42. Rate of Oxidation/Reduction in Control Experiment with No Copper 121 10.0 £50 6.0 4 {la 9 6.0 4 5- g 4.0 4 ’s‘ ‘5 2.0 .. . 0 Cu -0.00 A CU -0.50 mg/l 0.0 '4: T T I T I T I o 2 4 6 0 10 12 14 16 18 TIME (minutes) Figure 43. Rates of Cr Oxidation with Adsorbed Copper and in Control Experiment 12.0 10.0 4 §‘ 60 4 2 5. g 6.0 4 x 9 4.0 4 ‘E 2 0 CU -0.00 2.0 A CU -o.50 mg/l 0.0 T T T T l T 0 2 4 6 8 10 12 14 16 18 TIME (MINUTES) Figure 44. Rate of Mn Solubilization with Adsorbed Copper and in Control Experiment 122 competition can occur but not influence the rate of the oxidation reaction. However, significant amounts of Cu were desorbed during the reaction suggesting Cu was somehow displaced from the Mnoz surface. The desorption of Cu may have been caused, at least in part, by the adsorption of released Mn2+. The competition of Mn2+ for surface sites with Cu2+ would explain why more Mn remains in solution in the presence of Cu than was found in the control (see Figure 44). It thus appears that these two metals may compete for sites on the Mnoz surface. 2.8 Summary and Conclusions The effect of changing solution parameters on the reaction rate has been evaluated by conducting duplicate experiments in which one solution parameter was varied. In particular, the effects on the oxidation rate of Cr of solution pH, dissolved 02, ionic strength, identity of the swamping electrolyte and the presence of adsorbed copper were investigated. The results of this part of this research can be summarized as follows: 1. There is no measurable pH dependence on the reaction rate over the pH range 3.0 to 5.0. 2. Only small quantities of Cr(OH)3-solid are oxidized by Mnoz, however the amount oxidized exceeds the solubility of Cr at pH's above 7.0. 123 3. The concentration of dissolved oxygen does not appear to have any measurable effect on the oxidation rate of Cr. 4. The oxidation rate decreases as the ionic strength, as set by a swamping inert electrolyte, increases. 5. Of a variety of swamping electrolyte solutions, only high concentrations of CaC12 appears to have specific effects on the oxidation rate. The reaction proceeds at a significantly reduced rate in the presence of 0.50 M 6. The presence of adsorbed Cu2+ on the MnOz suspension does not result in any measurable change in the oxidation rate as compared to a control experiment. From the above results the following conclusions can be made: 1. The absence of a pH dependence on the oxidation rate is somewhat surprising considering’ the variety' of roles that. protons ‘must. play in ‘the ‘various steps of the reaction. If the hypothesized adsorption-oxidation- desorption reaction is indeed correct, then it must be concluded that the effects of pH are probably obscured in this relatively fast reaction. 2. The oxidation of solid phase Cr(OH)3 by MnOz occurs at a much reduced rate compared to soluble Cr3+. The occurrence of this reaction is significant as this solid 124 species controls the solubility of Cr in most natural systems. The results suggest that the presence of Mn oxides in soils or sediments may serve to increase the solubility and potential mobility of this trace metal. 3. The dependence of the oxidation rate on ionic strength is interpreted to result from increased oxide flocculation with increasing ionic strength. This suggests the oxidation reaction is surface area controlled, and may be diffusion influenced when the ionic strength is high. 4. In general, only very small effects on the oxidation rate are induced by changing the nature of the swamping electrolyte. The exception is the significant decrease in the oxidation rate induced by high concentrations of CaClz. It does not appear that the decreased rate is 2+ 3+ due to the adsorption of Ca or the complexing of Cr by C1”. 5. Although Cu2+ is desorbed from the MnOz surface during the reaction, there is no effect on the rate of Cr3+ oxidation. unless the competition does not effect the rate, it does not appear that the sites for Cr oxidation are the same as those for Cu adsorption. However, it must be noted that the current experiments do not provide nearly enough information to make any firm conclusions on this issue. Additional experimentation is necessary. PART 3.0 THE COMPETITION BETWEEN MN AND FE OXIDES FOR CR(III) 3.1 INTRODUCTION Due to their similar geochemical properties and their tendency to occur together in the natural environment, Mn and Fe oxides are often considered as a single component in metal/sediment partitioning studies. Describing the general adsorptive behavior of most trace metals in natural systems does not require a rigorous evaluation of the partitioning between these two oxides. In the case of chromium however, this is important to consider, as Fe oxides tend to remove Cr from solution while Mn oxides oxidize Cr to the mobile, toxic hexavalent form. In the few attempts to apply adsorption reactions to chemical modeling (Davies-Colley et al., 1984; Vuceta and Morgan 1978), the presence of two (or more) adsorbing substrates has been handled as a cumulative increase in the adsorption capacity of the solid phase as a whole, with each substrate treated like a ligand with an experimentally determined binding constant. However, there has been little published research which has attempted to experimentally measure the adsorptive properties of a mixture of solid surfaces, or to determine the effect of one additional solid component on the adsorption equilibria of another. Because Cr which adsorbs to Mn oxides is oxidized while Cr' adsorbed to Fe oxides is stable , experiments were 125 126 conducted to explore the "competition" between synthetic MnOz and Fe(OH)3-am for Cr3+. The experiments were conducted using a suspension containing both MnOz and Fe(OH)3-am oxides. The data were interpreted while making one major assumption; that the two solid phases did not chemically react with one another. A precise measurement of metal partitioning between Fe and‘ Mn oxides is normally very difficult to attain because selective chemical extractions do not adequately distinguish between Mn oxides and amorphous Fe oxides. However in the present study, the contrasting behavior of Cr3+ on Mn versus Fe oxides allows a relatively straightforward assessment of Cr partitioning between these two oxides. It has demonstrated that any Cr3+ that is adsorbed by MnOz will be quickly oxidized and desorbed, 3+ while Cr adsorbed on Fe oxides is not oxidized. The absence of analytically detectable adsorption on MnOZ, due to the rapid oxidation and desorption of Cr6+, permits the determination of how much Cr has interacted with each oxide, without actually having to chemically separate the oxide phases. In other words, the amount of Cr6+ formed is a convenient measurement of the amount of Cr3+ that has reacted with MnOz. Any un-oxidized Cr remaining in the system must be adsorbed by the Fe oxide. Hence, there was no need to rely on selective chemical extractions to determine the amount of Cr adsorbed by each oxide. However, the Cr6+ that is formed can be re-adsorbed by Fe(OH)3-am. 127 Thus procedures were developed to ensure that all Cr6+ was selectively desorbed prior to analysis. The amount of Cr3+ adsorbed by Fe(OH)3-am could then be assessed indirectly. Describing the competition of multiple solid phases for trace metals is a very complicated process. Factors which may influence metal partitioning; 1) the binding capacity of each solid phase, 2) the energy of interaction between the adsorbate and the various surface sites (i.e. the relative affinities of each surface for the metal), 3) the kinetics of adsorption onto each surface, and 4), the potential effects of the solid phases on one another (i.e. one solid may tend to "coat" the other). The present experiments were designed to quantitatively measure the competition of MnOZ and Fe(OH)3-am for“ Cr3+} The results *were interpreted qualitatively in terms of the factors listed above. The specific problems addressed in this part of the study are summarized as follows: 1. The nature of the competition between MnOz and Fe(OH)3-am for trivalent Cr. In particular, the relative 3+ effectiveness of each oxide to react with Cr over a range of reactant concentrations. 2. The ability of Mnoz to oxidize Cr3+ which is adsorbed on Fe(OH)3-am). This area of research provides insight into determining whether adsorptive equilibria will respond to the addition of an additional reacting surface. 128 3. The effect on the competition for Cr3+ of precipitating the iron oxide onto the MnOZ prior to spiking the suspension ‘with Cr3+. This set of experiments was designed to determine whether Fe oxide may influence the reactivity of MnOz when it occurs as a layer, or coating on the MnOz surface. 3.2 PROCEDURES The procedures employed in this part are generally the same as those employed in the rest of the study, except for several modifications and additional techniques. To ensure proper formation of the iron oxide suspension, all of the experiments using multi-sorbent systems were conducted under an N2 atmosphere. The Fe(OH)3-am was prepared in the glove box in an N2 atmosphere as described in the Methods section. All other reagents, glassware, etc. were also placed within the glove box before it was sealed and purged. Thus the experiments were all conducted under nitrogen a atmosphere from the synthesis of the Fe oxide through the filtration of the sample aliquots. Before the multi-sorbent experiments could be conducted, it was necessary to characterize the adsorptive behavior of Cr3+ and Cr6+ on Fe(OH)3-am. The adsorption of 9.6 X 10'6 M Cr (both III and VI) at pH 4.5 by various concentrations of Fe(OH)3-am was determined by measuring the decrease of Cr concentration in solution after one hour of equilibration. 129 It was determined that a 1 X 10"3 M Fe(OH)3-am suspension (measured as FeT) was sufficient to adsorb nearly 100% of 3+ or as Cr6+ (see Table 3). the initial Cr, either as Cr This oxide concentration. could also adsorb all of both oxidation states together, suggesting that the adsorption of one oxidation state did not effect the adsorption of the other under these conditions. This concentration (10'3 M) of Fe(OH)3—am was used in all of the multi-sorbent experiments. The Fe oxide to Mn oxide ratios were varied by adjusting the amount of MnOz. 6+ produced by the In order to measure the amount of Cr adsorption onto Mn02, it was necessary to ensure that any Cr6+ adsorbed by the Fe(OH)3-am was desorbed prior to analysis of the sample. This was accomplished by utilizing the contrasting pH dependent adsorption behavior of cations and anions. Anions, in general, tend not to adsorb above the pH of the oxide (see Appendix I). As adsorption is zpc a reversible process, the desorption of Cr6+ was accomplished by raising the pH above the szpc of the Fe oxide and allowing sufficient time for desorption to proceed. In a system containing 10"3 M Fe(OH)3-am and 9.6 X 10'6 M Cr6+, complete desorption was attained by raising the pH of the suspension to 10.0 with 1.0 M NaOH and allowing the system to equilibrate for 45 minutes. This desorption process was shown to be specific for Cr6+ by repeating the previous experiment with both oxidation states adsorbed to the Fe oxide. The data from this procedure is presented on 130 ADSORPTION OP TABLE 3 Cr SPECIES BY Pe(OH)3 AT pH 4.5 Conc. of Fe(OH)3(M) % Cr3+ Adsorbed* % Cr6+ Adsorbed* 1 x 10“3 5 x 10‘4 1 x 10‘4 100 98 49 96 82 28 * Initial [Cr] - 9.6 x 10’5N TABLE 4 SELECTIVE DESORPTION OF Cr5+ FROM 1 x 10‘3N Fe(OH)3 BY RAISING pH TO 10 FOR 45 MINUTES CrT Desorbed (mg/1) Cr6+ Desorbed (mg/1) Cr Adsorbed (mg/1) Replicate Replicate 1 2 1 2 0.50 Cr3+ 0.02 0.00 0.00 0.00 0.25 Cr3+/0.25 Cr5+ 0.24 0.27 0.24 0.25 0.50 Cr5+ 0.50 0.49 0.50 0.48 131 Table 4. All of the desorbed Cr was verified to be 6+ hexavalent by the Cr extraction. Five replicate samples (approximately 20 ml each) of a suspension containing 6+ were also collected, adjusted to pH 10, and adsorbed Cr analyzed for Cr6+ to ensure that this procedure was reproducible and the aliquots were representative of the system. The results, shown on Table 5, show that the technique could be employed with good accuracy and precision. 6+ and total Cr it was possible With a measurement of Cr to determine the concentration of all forms of Cr species in an experimental system, including the concentrations of Cr3+ and Cr6+' as both dissolved and adsorbed species. For each time interval in which a sample was collected, the following set of procedures were applied in order to determine all forms of Cr in the sample: 1. Immediate filtration (IF). An aliquot system was immediately passed through a 0.22pm filter. This allowed for the measurement of total dissolved Cr (i.e. Cr3+ and Cr6+ that was not adsorbed). 2. Immediate filtration followed by Cr6+ extraction (IFE). A split portion of the (IF) sample was used to determine the concentration of dissolved Cr6+. The Cr3+ in solution was then computed as (IF) - (IFE). 3- Immediate Base Addition followed by Cr6+ extraction (IBE). 100u1 of 1.0 N NaOH was added to an unfiltered aliquot of sample (to pH = 10) and the sample allowed to 132 REPLICATE ANALXSIS OF THE DESORPTION 0F Cr5+ FROM FeéOH)3 BY RAISING pH TO 10 Fe(OH)3 = 1 x 10"3 = 9.6 x 10‘ . TABLE 5 M Replicate % Cr Desorbed* 96 98 100 99 97 * Average of two readings on AAS 133 set for 45 minutes. After this time the sample was filtered and subjected to Cr6+ extraction. This procedure caused the desorption of all Cr6+ from the Fe oxide and the precipitation and/or adsorption of any Cr3+ remaining in solution (as verified by measuring total Cr in solution as well). Thus the Cr6+ adsorbed on Fe oxide was computed by (IBE) - (IFE). The Cr3+ adsorbed onto Fe oxide was then the initial Cr spike minus the amount of solution Cr and the amount of 6+ adsorbed Cr , or Cr3+ (IF) - (IBE). initial ' This entire sampling scheme was not necessary for every sample taken throughout the experiment. In fact, there was little or no solution Cr present after the first ten minutes of the reaction, after which time the Cr was adsorbed either 6+ 3+ or Cr . At this point in the reaction, Cr6+ was as Cr 3+ adsorbed was measured by the (IB)E procedure and the Cr the difference between this and the original Cr concentration. While Cr3+ remained in solution there was the potential for continuing oxidation after the pH was adjusted to 10. The potential for this occurrence was investigated by an oxidation experiment (conducted without iron oxide) in which two replicates were taken; one of which was filtered immediately, the other of which was raised to pH 10 (to precipitate Cr3+ hydroxide) and allowed to set for 45 minutes before filtration. A comparison of three sets of replicates taken at different time intervals is shown on 134 Figure 45. The oxidation reaction clearly did not stop instantaneously when the pH was raised. Thus, until the reaction is essentially complete, the amount of Cr6+ measured is probably too high. The presence of iron oxide probably' further' decreases the discrepancy, as Cr3+ may adsorb or’ precipitate on the oxide surface as ‘well as precipitate as a hydroxide. HOwever, the oxidation versus time curves presented below may not be accurate in the early time intervals. At the completion of the reaction, when no Cr is present in solution, the amount of Cr oxidized accurately reflects the amount of Cr which has interacted with MnOz. Because of the complexity’ of 'the reaction and the experimental procedures used to obtain the data, the reproducibility of the experimental results were tested. Figure 46 shows the results of two identical multi-sorbent experiments (FeT = 10"3 M, MnOz = 2.3 X 10'5 M, pH = 4.5) conducted exactly one month apart. The amount of Cr6+ formed in the multi-sorbent system is highly reproducible. Although complex, the reactions are apparently well defined given a certain set of conditions. 3.3 RESULTS AND DISCUSSION 3.31 The General Nature of the Competition for Cr(III) The first multi-sorbent experiments conducted were dGSigned to determine the nature of the "competition" 135 050 AIMMEDIATE FILTRATION 0.40 q DBASE ADDED D 030« 8 0 A N a g . 02o~ C] A U A 010« 000 .7 I T 0.00 2.00 4.00 6.00 8.00 TIME (minutes) Figure 45. Amount of Cr Oxidized in Filtered Versus pH Stabilized Replicates 100 F0(0H)3 .. 10-3 M - 2.35-5 M 80 . M002 0 u N § 25 —=-=B n: O E u 8 w a o T T T T 0 1O 20 30 4O 50 TIME (minutes) Figure 46. Reproducibilty of Mixed-Oxide Oxidation Rates 136 between Fe and Mn oxides for Cr3+ with as little potential for one oxide affecting the other as possible. In these experiments the 10'3 M Fe(OH)3 suspension was precipitated and aged for four hours followed by the addition of an aliquot of MnOZ suspension. This multi-component suspension was allowed to "equilibrate" for only a few minutes before 3+ 3+ an aliquot of Cr was added to bring the initial Cr concentration to 9.6 X 10'6 M. As a starting point, the initial MnOz concentrations chosen were similar to those studied in single sorbent systems, even though this meant that on a molar basis the Fe oxide would be in much higher concentrations than the MnOz. Somewhat unexpectedly, these low levels of MnOz (relative to the Fe(OH)3-am) were able to oxidize significant amounts of the Cr spike. The amount of Cr oxidized (as a percentage) versus time for five Mn02 concentrations, ranging from 1.2 to 9.2 X 10'5 M, are shown on Figure 47. As the graph shows, increasing the amount of MnOz increases the 3+ oxidized (and consequently decreases the amount of Cr amount adsorbed onto Fe(OH)3-am) . This point is further illustrated in Figure 48 in which the molar amount of Cr oxidized at the end of the experiment is plotted against the initial MnO2 loading. There is a linear relationship between the amount of MnOz and the amount of Cr oxidized. Even in the presence of a large excess of Fe(OH)3-am, much of the Cr was oxidized, which suggests MnOz competes very favorably with Fe(OH)3 for Cr3+. This is further 137 100 C] C] D D x 801 D x x S D + g: 60 j + i X 0 0 O 5 O U 404 O 8 A A If A x 5.9x10—5 M X D9.2x10-5 M 204 A 1.2x10-5 M 02.3x10-5 M +4.6X10-5 M O F V I T T 0.0 10.0 20.0 30.0 40.0 50.0 TIME (minutes) Figure 47. Percent Cr Oxidized in Mixed-Oxide Systems as a Function of MnOz Loading 10.0 I I: 8.0 4 C1 ".5 9 6.0 4 D x C] ‘5‘ 40 § :1 3:: O 2.0 4 0.0 I TI l T 0.0 2.0 4.0 6.0 5.0 10.0 M002 Molor ( X 10-5) Figure 48. Total Amount of Cr Oxidized in Mixed-Oxide Systems Versus Initial Mn02 Concentration 138 emphasized on Figure 49 which shows the percent Cr oxidized versus the molar ratio of Fe(OH)3-am to MnOZ. Even with an 83-fold excess of Fe(OH)3-am, 38% of the Cr3+ was still oxidized by MnOZ and 62% adsorbed to the Fe oxide. When the Fe(OH)3 excess is reduced to 11 times, 94% of the Cr is oxidized by the MnOz. 3+ The measurement of solution Cr shows that the reaction 3+ is depleted from seems to be nearly complete when Cr solution, although a small quantity of Cr is oxidized after Cr3+ is no longer measurable in solution. This relationship is shown on Figure 50 which is a graph of the percent of the initial Cr oxidized and the percent of Cr remaining in solution. This result suggests that the Cr oxidation occurs primarily through the successful competition for solution Cr by MnOZ, rather than by the oxidation of Cr adsorbed to the Fe(OH)3-am surface. However the potential for oxidation of Cr adsorbed onto the Fe(OH)3 surface was further tested (see below). The significant oxidation of Cr by MnOZ in the presence of a nearly two order of magnitude molar excess of Fe(OH)3 presents an interesting problem. Quantitative modeling of adsorption reactions has shown that the binding capacity of an oxide, usually expressed as moles of surface sites per mole of oxide, and not the molar quantity of the oxide, is the correct means of comparing oxide reactivities. The binding capacities of Mn and amorphous Fe oxides reported in other studies (see Appendix I) shows that these two oxides 139 100 D a 604 8 111 C] 9 604 6 a (I U E ‘0‘ a O m Lu Q 204 O T *T T T 0 20 40 60 80 100 Fe(OH)3 t0 M002 Molor Rotio Figure 49. Percent Cr Oxidized Versus the Initial Molar Ratio of Fe(OH)3 to Mn02 1000‘ ‘ A Cr(III) in solution 60 . C] Cr oxIdIzed C): o 3’ 60 a *2 .1 4° LA. 0 .. 4O 4 2 U o a: LJJ O. 20 . 0 5.1 I T T T A 0 10 20 so 40 50 TIME (minutes) Figure 50. Amount of Cr3+ in Solution and Cr Oxidized Versus Time in a MIxed-Oxide System 140 have similar binding capacities. However, the adsorption capacity of an oxide is a pH dependent quantity and most binding capacities are determined at higher pH's. Mn oxides have been shown to have a higher adsorption capacity than Fe oxides at low pH's (see Appendix I) so it is conceivable that the molar excess of Fe(OH)3-am is deceiving. The binding capacity of the Fe(OH)3-am used in this study for 9.6 X 10'6 M Cr3+ at pH 4.5 was found to be approximately 0.07 moles of Cr per mole of Fe(OH)3-am. The binding capacity of MnO2 for Cr was not measured as new surface sites are generated during the redox reaction, making this measurement impossible. The binding capacity of MnOz, even if taken to represent the total of all sites present initially plus those created during the reaction, cannot exceed 0.67 moles Cr/moles MnOZ, as this represents the stoichiometric oxidation equivalent. In other words, only 0.67 moles of Cr can be adsorbed per mole of MnOz as all of the MnOz would be subsequently reduced and dissolved. This would suggest there can be at most, an order Of magnitude difference in the binding capacities of the two oxides for Cr. From this argument it is apparent that the difference in binding capacities alone cannot account for the strong competition by MnOz, as significant amounts of Cr is oxidized when the Fe oxide is in a nearly two-order of magnitude excess. Besides the number of surface sites available on each oxide, the energy and rate of the reaction between Cr and 141 the various surface sites may influence the partitioning between these two sorbents for Cr. If a higher energy bond is formed between Cr and the surface sites on one oxide as compared to the other, it would be expected that this oxide would compete favorably. This is analogous to solution complexes in multi-ligand systems, where metals in solution show a higher degree of coordination with the ligand that forms the higher energy bond. Qualitatively, past research would suggest that the MnOZ-Cr bond should be stronger than the Fe(OH)3-Cr bond at the pH's used in these experiments. This results from the favorable surface charge on MnOz at pH 4.5 and the apparent ability of this surface to bond with unhydrolyzed metal species (see Appendix I). The kinetics of adsorption may also influence the resulting distribution of Cr between the two oxide surfaces. Although adsorption reactions have been included in equilibrium models, it has not been demonstrated that the various surface sites in multi-sorbent systems behave like ligands in solution, where equilibrium speciation is attained. based on ‘the concentration. of each ligand and coordination energy of each complex. The adsorptive equilibria in these multi-phase systems could be much slower to respond to the addition of new species than metal-ligand equilibria. If this is true, the adsorption kinetics of Cr onto the Fe oxide and Mn oxide surfaces may influence the 3+ resulting distribution. If the adsorption of Cr onto MnOz 142 is faster than the adsorption onto Fe(OH)3, this may explain why MnOz can compete so effectively. To explore this possibility, the kinetics of adsorption of 9.6 X 10’6 M Cr3+ on 10'3 M Fe(OH)3-am was compared to the oxidation rate of the same amount of Cr3+ by 2.3 X 10"5 M MnOz. In a mixed-oxide system containing the same concentrations of these components, 54% of the initial Cr3+ was oxidized and 46% adsorbed by the Fe oxide (see Figure 47). As shown on Figure 51, the adsorption by Fe(OH)3-am was significantly faster than the oxidation by MnOz. For this set of conditions, the kinetics of the single sorbent 3+ systems would predict that more Cr should have been adsorbed by Fe(OH)3. 3.32 The Effects of the Order of Addition of the Reagents In all previously described experiments, the oxide phases were mixed after they were precipitated, and the Cr3+ was added last. In the following experiments the competition of these two oxides for Cr was investigated for the following changes in the order of addition of the reagents: 1) the Fe(OH)3 was precipitated on the MnOZ, and 2) the Cr was added to the Fe(OH)3 before the addition of MnOz. Since the kinetics suggest that Fe oxide should compete favorably for solution Cr, then either 1) the oxides interact such that the reactivity of one oxide surface (i.e. Fe(OH)3-am) is reduced, or 2) the MnOZ can simply "out- 143 100 80 ‘ m 0 g 60 '1 ’2. 2: LA. 0 p— 40 “ Z LAJ U 33 a ACr SORBED BY Fe(OH)3 2O -1 _ DCr OXIDIZED BY M002 0 j , , 0 1O 20 3O 40 TIME (minutes) Figure 51. Kinetics of Adsorption by Fe(OH)3 and Oxidation by MnOz 144 compete" the Fe(OH)3 due to a higher affinity of Cr for the MnO2 surface. If the Mn oxide were to coat the Fe oxide and block binding sites, this could help explain the ability of MnOz to compete favorably. This particular situation was not addressed in the present research. HOwever, the potential of Fe oxide to affect the reactivity of MnOZ was investigated. To explore this problem, the aliquot of the MnO2 stock suspension was added to the Fe(NO3)3 solution prior to raising the pH and hydrolyzing the Fe. This, at least theoretically, should have caused some Fe oxide to precipitate on the MnOz particles. Multi-substrate suspensions were prepared in this manner for two different MnOz loadings. Then, Cr3+ was added to the system, samples were collected and analyzed as described above. The data for the amount of Cr oxidized are shown on Figures 52 and 53. Also shown on these graphs is the amount of Cr oxidized for experiments with identical oxide concentrations, in which the MnO2 was added to the system after the Fe(OH)3-am had been precipitated and aged. There is a significant decrease in the amount of Cr that was oxidized for both M002 loadings when the iron oxide is precipitated on the MnOz surface. This clearly demonstrates that there is a potential for one oxide to influence the reactivity Of the other. It is still not known whether the addition of MnOz to pre-formed Fe(OH)3-am causes a reduction of the reactivity of the Fe oxide, but this possibility cannot be PERCENT CR OXIDIZED Figure 52. PERCENT CR OXIDIZED Figure 53. 145 100 80* AFe(0H)3 FIRST D M002 FIRST C] 0 I Y j— T 0 10 20 30 4o 50 TIME (minutes) Percent Cr Oxidized in Mixed-Oxide System When Fe(OH)3 is Precipitated on 2.3 X 10 M MnOz 100 n s in 604 604 40 4 204 EJFe(OH)3 FIRST AMn02 FIRST 0 T 477 T T 0 10 20 30 40 50 TIME (minutes) Percent Cr Oxidized in Mixed-Oxide System When Fe(OH)3 is Precipitated on 9.2 X 10 M Mn02 146 ruled out- ILf the MnOz does not effect the reactivity of the Fe(OH)3, then Cr3+ must have a higher affinity for the MnOz surface. As described above, the oxidation of Cr3+ by MnOz is nearly complete when there is no longer any Cr3+ in solution. This suggests (but does not prove) that Cr3+ adsorbed onto Fe(OH)3-am is no longer reactive with respect the MnOZ surface. To determine whether Cr3+ adsorbed to the Fe(OH)3 surface can be oxidized by MnOz, an experiment was conducted in which the initial Cr3+ spike (9.6 X 10’6 M) was allowed to adsorb onto the Fe(OH)3-am (10"3 M) at pH 4.5, prior to the addition of the MnOz suspension (2.3 X 10"5 M). If oxidation occurs under these conditions it would suggest Cr3+ adsorbed to Fe(OH)3 becomes desorbed upon the addition of MnOz. This would strongly suggest that adsorptive equilibria does respond to the addition of a new solid phase. The results of this experiment are shown on Figure 54. The data show that approximately 6% of the initial Cr spike was oxidized after 1.0 hour. The reaction does not appear to have been completed during this time interval. Also shown on this figure is the oxidation rate in a control experiment containing the same reactant concentrations in which the Cr spike was added to the mixed oxide suspension. Increasing the amount of MnOZ to 9.2 X 10"5 M and increasing the reaction time ‘to 2.0 Ihours showed. that significant oxidation does occur (see Figure 55) although at a much 147 100 DCr ADDED LAST 80 . A M002 ADDED LAST PERCENT CR OXIDIZED L A I.) O . T l T r 1 0 10 20 30 4O 50 50 TIME (minutes) Figure 54. Pegcent Cr Oxidized in Mixed-Oxide System When Cr I is Agsorbed to Fe(0H)3 Prior to Addition of 2.3 x 10‘ M 14002 100 ACr ADDED LAST 80 4 DM002 ADDED LAST 0 LAJ N . § 604 X 0 m U E 404 1.1.! O a: LIJ CL 0 I T r ' ' 0 20 4o; 60 80 100 120 TIME (minutes) Figure 55. Peggent Cr Oxidized in Mixed-Oxide System When Cr 1 is A sorbed to Fe(OH) Prior to Addition 9.2 x 10'g M 14002 3 Of 148 slower rate than under than the control. These results demonstrate that MnO2 can oxidize Cr that was adsorbed to Fe(OH)3. Since the reaction was apparently incomplete in both of the above experiments, a long term experiment was conducted to see if the amount of Cr oxidized would eventually reach that measured in the experiments where Cr was the last reagent added. An experiment with the same initial concentration of reactants as the one described on Figure 54 was continued for 43 days. The results are shown on Figure 56. The amount of Cr oxidized reached a maximum within 19 days. At this time 50% of the initial Cr spike had become oxidized. This is only slightly less than the 54% oxidized in the identical experiment in which Cr was added as the last reactant. This result would appear to suggest that the amount of Cr oxidized in the multi-sorbent system is dependent on an equilibrium distribution between the Mn and Fe oxides. When all of the initial Cr is first adsorbed to the Fe(OH)3 the adjustment to equilibrium conditions occurs much slower. This further suggests that the binding energies and binding capacities, rather than the kinetics of adsorption, influence the partitioning of Cr in this multi-sorbent system. If this interpretation is correct, it would suggest that Cr3+ has an ‘unusually' high affinity for the MnOz surface as compared to Fe(OH)3. It would also suggest that 149 100 80 d o u '1‘ Q 60 4 x 0 E: C} D o E 404 LAJ 8 a: U 20 1F] 0 ' I r l l 0 10 20 30 40 50 TIME (days) Figure 56. Percent Cr Oxidized in Long Term Mixed-Oxide System When Cr is Adsorbed to Fe(OH)3 Prior to Addition of Mn02 \\ 150 the modeling of multi-sorbent adsorption equilibria can be treated in a similar manner to metal complexation. 3.4 SUMMARY AND CONCLUSIONS The competition between synthetic Mn02 and Fe(OH)3-am for Cr3+ has been evaluated in multi-sorbent experiments, with the amount of Cr oxidized as a measure of the Cr adsorbed by the MnOz. The results show that considerable Cr3+ was oxidized even in the presence of a large molar excess of Fe(OH)3, suggesting that MnOz competes successfully for this trace metal. Reasons for these results were not clear and the following ideas were explored through additional experimentation: 1) the binding capacity of Mn02 at pH 4.5 is much higher than that of Fe(OH)3, 2) the Cr-MnOz bond is a higher energy bond and the results reflect the equilibrium speciation between Cr and the various sites, 3) the kinetics of the oxidation reaction are faster than those for adsorption onto Fe(OH)3, and. 4) the presence of Mn02 reduces the reactivity of the Fe(OH)3 surface by blocking surface sites. Although the surface capacity of MnO2 is likely to be considerably higher than that of Fe(OH)3 at pH 4.5, especially considering that the oxidation reaction may generate new sites, it does not appear that this difference in binding capacity actually offsets the large molar excess of Fe(OH)3 in these experimental systems. In other words, there was probably more available binding sites on the Fe 151 oxide than the MnOz. Furthermore, comparison of single- sorbent kinetics shows that the oxidation reaction is actually slower than the adsorption of Cr onto (Fe(OH)3. This suggests the effective competition by MnOZ is not simply a result of kinetic considerations. 3+ in mixed sorbent The strong competition by MnOz for Cr systems is apparently a result of a higher energy Of interaction between Cr and MnOz surface sites than Cr and Fe(OH)3 surface sites. This is supported by the fact that there is an equilibrium distribution between oxidized Cr and Cr adsorbed on Fe(OH)3, for a given set of experimental conditions, regardless of the order of addition of the reactants. In other words, the same amount of Cr is oxidized whether the Cr spike is added to a mixed oxide suspension or allowed to adsorb to Fe(OH)3 prior to the addition of MnOz. The potential of one oxide to affect the reactivity of the other was demonstrated, although it seems unlikely that the M002 added as a solid could have a large effect on the Fe(OH)3. This is supported by the fact that Mn02 competes favorably even when the Cr is adsorbed onto the Fe(OH)3 prior to the addition of the MnOz. SUMMARY AND CONCLUSIONS Research has been conducted to explore the nature and controls of the oxidation of chromium by manganese oxide. This reaction was investigated under controlled laboratory conditions using a synthetic manganese oxide which has similar properties to manganese oxides found in soils and sediments. This section of the report begins with a summary of the principal findings of this investigation. This summary is followed by a brief discussion of these findings in terms of the current knowledge of surface Chemistry. Finally, the jpractical environmental significance. of 'the results are discussed and suggestions for continued areas of research are proposed. 1.0 SUMMARY The primary findings of this research can be summarized as follows: 3+ 3+ 1. At low pH's, where Cr is soluble, the oxidation of Cr by low crystallinity Mn02 is very rapid and the reaction proceeds to completion or near completion. The completeness of the reaction suggests the reaction is autocatalytic. 2. The reaction stoichiometry, as determined by the measurement of reaction products, confirms the general 152 153 hypothesized reaction stoichiometry for pH 4.5 given by the equation: 2 Cr(OE)2+ + 3 Mno2 = 2 HCr04’ + 3 Mn2+ 3. The reaction rate is sensitive to changes in the relative 3+ present. At low MnOz to Cr3+ amounts of Mn02 and Cr ratios the reaction rate increases linearly to increases in the MnOZ concentration. At an oxide to metal molar ratio of approximately 2, the change in the reaction rate no longer increases in proportion to the increase in oxide concentration. At high MnOZ to Cr3+ ratios, large increases in the concentration of MnOz cause only small increases in the oxidation rate. 4. At low MnOZ to Cr3+ ratios, the reaction can be described 5. by the following rate equation: d[Mn02]/dt = -k[Mn02]1’o[Cr3+]o‘5 This rate equation breaks down after considerable MnOZ has been dissolved, suggesting the surface properties of the oxide or the reaction mechanisms are changing. There is no measurable influence on the reaction rate induced by changing the solution pH when the solution is below saturation with respect to Cr(OH)3. At higher pH's, the formation of Cr(OH)3 causes a large decrease in the rate and extent of the reaction. However, the 154 amount of Cr oxidized exceeds the reported solubility of Cr(OH)3. 6. Increasing the ionic strength with an inert electrolyte causes a decrease in the reaction rate, probably due to the increased flocculation of the oxide particulates. 7. The reaction rate is not sensitive to copper adsorbed on the MnOZ surface prior to the addition of Cr3+, suggesting that there was either an abundance of sites such that Cu adsorption did not affect the oxidation rate or competition for sites does not influence the oxidation rate. 8. MnOZ competes effectively for Cr3+ with Fe(OH)3 even when the Fe oxide is present in a large molar excess to the MnOz. The amount of Cr which is oxidized by MnOz in a mixed-oxide system does not depend on the order of addition of the reagents, although the rate of reaction is highly dependent on the experimental procedure employed. These preliminary results suggest that the amount of Cr3+ oxidized versus the amount adsorbed by Fe(OH)3 is controlled by an equilibrium distribution of the Cr between the two oxides. 2 . 0 DISCUSSION Past research (Schroeder and Lee, 1975; Van der Weijden and Reith, 1982) has suggested that the oxidation of Cr by 155 Mn oxides takes place following the adsorption of this metal on the oxide surface. One of the primary goals Of this research was to determine whether the experimental data can be explained in terms of adsorption theory. The results listed above can, in some instances, be adequately explained with surface Chemistry models. In other instances, the data are not readily interpretable in terms of an adsorption controlled reaction. Qualitatively, adsorption models predict the influence of the oxide to metal ratio on the reaction rate which was observed in the present experiments. At low MnOz to Cr3+ 3+ was shown to ratios, an increase in the amount of Cr actually cause a decrease in the reaction rate. This behavior is consistent with the prediction for an adsorption controlled reaction, as the fraction of metal which is adsorbed decreases with an increasing metal to oxide ratio. The data from this research also demonstrate that the reaction rate becomes insensitive to increasing MnOZ concentrations at high MnOZ to Cr3+ ratios. This is also predicted for an adsorption controlled reaction. When enough oxide is present to adsorb the maximum amount of the metal which can be adsorbed at that particular' pH, an increase in the oxide concentration would be predicted to result in little increased adsorption. If the reaction is adsorption controlled, it stands to reason that the oxidation rate should also experience only a minor increase. 156 Another finding which would appear to be consistent with past research on adsorption is the decreasing reaction rate with increasing ionic strength. The rate of adsorption of trace metals, particularly on Mn oxides (Morgan and Stumm, 1964) has been shown to decrease with increasing ionic strength. Although this phenomena has not been adequately investigated, the slower adsorption rate with increasing ionic strength appears to be caused by a decrease in the effective surface area of the oxide due to the increased particle size caused by flocculation. Only the rate of the reaction, not the extent was decreased by increasing the ionic strength. The finding most difficult to reconcile with adsorption theory is that the reaction proceeds at pH 4.5 with metal to oxide ratios that are insufficient to cause measurable adsorption of other trace metals. However, there are several points to consider which may resolve this question. The first point is that Cr3+ may have a much stronger affinity for oxide surfaces than other trace metals. Indeed, Cr3+ was found to adsorb the strongest of a series of transition metals on amorphous iron hydroxide (Leckie et al., 1980). This suggests that Cr may inherently adsorb strongly to many surfaces, possibly because this metal has a higher positive charge than most other stable metal species. Thus a comparison of the adsorptive behavior of Chromium to other metals on Mn oxide may not be appropriate. 157 Besides the general tendency for chromium to adsorb more strongly than other metals, there may an intrinsic affinity of Cr for the MnOZ surface in particular. Both Mn4+ and Cr3+ are 3d transition metals which have very high crystal field stabilization energies in octahedral coordination. Given that 1) several mechanisms of adsorption, including lattice substitution, have been demonstrated to occur on Mn oxides (see Appendix I) and 2) as many as half of the octahedral sites may be vacant in poorly ordered MnOZ (Burns and Burns, 1979), it is possible that Cr not only adsorbs to surface sites, but also becomes incorporated. within. the oxide structure by displacing water or electrolytes from the vacant octahedral sites. This too could contribute to a uniquely high affinity of Cr3+ for the MnO2 surface which may drive the reaction under these experimental conditions. The extensive oxidation of trace metals at the MnOz surface at low pH's and oxide to metal ratios is not unique to this study. Postma (1985) has shown that ferrous iron (Fe2+) is oxidized to ferric iron by MnOZ at low pH's and low MnOz to Fe2+ ratios. For instance, at pH 3.0, 100% of the Mn oxide was reduced with initial Fe2+ and MnOZ concentrations of 1 X 10'3 moles and 1.5 X 10'5 moles, respectively. This study did not address the problem from an adsorption perspective, although adsorption was presumed to have preceded oxidation. Perhaps the reason that these results do not appear to conform to the predictions of an adsorption controlled 158 oxidation reaction is actually more fundamental. The thermodynamic model of adsorption assumes that the surface sites are a fixed quantity. In the reaction between Cr3+ and MnOz the amount of surface sites are not fixed as the reduction and release of structural Mn must generate additional adsorption sites. When this autocatalytic nature is considered it becomes apparent only a small fraction of the Cr3+ need be adsorbed by the initial MnOz surface. After this initial Cr3+ is adsorbed, oxidized and desorbed, new sites are available for the next fraction of the initial Cr3+. Thus even if the initial adsorption capacity of the oxide is small, the reaction can proceed until there are not enough sites left to cause adsorption. From a thermodynamic standpoint, the Cr3+ in solution may be reacting to the concentration of all potential surface sites, not just the sites initially present. An oxidation reaction which is proceeded by adsorption imposes certain limitations on the mechanism of the oxidation reaction. Based on the model of the oxide/metal interactions presented in current adsorption theories, electron transfer between the oxide metal and the adsorbed metal would occur after an inner-sphere complex is formed. Zinder, et al., (1986) have shown that reductive dissolution of Fe oxides is facilitated by the coordination of ligands at the oxide surface and that the electron transfer takes place ‘with the: hydroxide ions as an. electron "bridge". Based on these ideas, a conceptual model of Cr oxidation 159 following adsorption onto MnOZ is presented on Figure 57. As Cr3+ at pH 4.5 is primarily Cr(OH)2+ (Rai et al., 1986), this would suggest that a bidentate bond would be formed between Cr3+ and the surface hydroxyls. This model suggests that each Cr3+ ion would have to donate electrons to at least 2 adjacent surface Mn4+ ions. It is apparent that the adsorption of 2 adjacent Cr ions to reduce three Mn ions, as shown in the figure, is not plausible, as such a restriction would cause surface sites to become limited quickly. MOre likely, intermediates are formed and the reaction has several steps. Whether or not the conceptual model presented above represents an approximation of the actual reaction mechanism, the overall evidence suggests that the reaction is preceded by adsorption. 3.0 ENVIRONMENTAL SIGNIFICANCE The ultimate goal of research such as described in this report is to increase our knowledge of the fundamental reactions which may be important in understanding and predicting the fate of trace metals in the environment. The results of this study, which represent only a small step towards this goal, have provided some additional insight on the potential controls of chromium geochemistry' in environment. The findings of this study have demonstrated that the potential for oxidation of Cr to the toxic, mobile hexavalent form by Mn oxides certainly does exist. At pH's 160 / \ \ / W4 0\@E<°"I \ 0/ 6 OH Figure 57. Conceptual Model of the Adsorption and Oxldatlon of Or by M0 Oxide 161 representing the very low range of natural waters, but which are close to the values found in some in soils, this reaction would appear to be potentially important. Additionally, the reaction could be important in polluted waters with low pH's. The lowering of the pH can solubilize Cr3+ which may then be oxidized if sufficient Mn oxides are present. The preliminary results of this research also suggest that Cr solubility can be increased in the presence of Mn oxides. This may be an important reaction in the soils and sediments of environments with more neutral pH's. The insolubility of the hydroxide species is responsible for the immobility of Cr in many environments. If Mn oxides can cause dissolution of Cr hydroxide by oxidizing the Cr3+ in solution, then the mobility of this species may be increased. The oxidation of Cr adsorbed to iron oxides, followed by 6"' suggests that this reaction the re—adsorption of the Cr may contribute to the depletion of Cr in Mn rich sediments in the ocean as compared to low Mn sediments. This reaction may also explain the partitioning of chromium in other sediments. Thus, although the rate and extent of the reaction appear to be limited by the formation of the hydroxide species, the reaction may be important in controlling the geochemical behavior of Chromium in environments which are enriched in Mn oxides, such as ocean floor sediments. If this reaction is important in these 162 environments then Mn oxides may exert an influence on the overall global geochemical cycle for chromium. The preliminary result of the mixed-oxide experiments also suggest that an equilibrium approach to the partitioning of metals based on laboratory determined binding constants may be useful for modeling trace elements in aqueous systems. However, much further research is necessary to establish the feasibility of this approach. 4.0 SUGGESTIONS FOR FURTHER RESEARCH The continued development of a model of Cr geochemistry in sediment/water systems can be advanced by continued research in several areas, including further study of valence state transformations of this metal, by Mn oxides as well as other oxidizing/reducing agents. The research begun here could be expanded in the following areas: 1. A mechanistic approach to interpreting the reaction rate data. 2. Further efforts to delineate the effects of Mn oxides on the solubility of Cr hydroxides. 3. A study of the role of organic complexes in the geochemistry of chromium. This should include the characterization of such complexes and the effects on the oxidation and adsorption of these complexes by oxides and other adsorbents. 163 4. Further studies of Cr uptake and valence state transformations using naturally occurring sediments and soils. APPENDICES APPENDIX I APPENDIX I PAST RESEARCH Since the focus of this research is on adsorption and adsorption controlled oxidation, it was necessary to explore adsorption models, particularly those which have been found useful in describing surface reactions on hydrous oxides. Hence, this section on past research begins with a summary of recent developments in adsorption modeling, including mechanistic interpretations as well as mathematical formulations. This is followed by a summary of previous research on Cr(III) and Cr(VI) adsorption onto solid surfaces and on the oxidation of Cr(III) by surface reactions on various Mn oxides. ADSORPTION MODELS Adsorption of metals from solution involves the removal of metal solutes by accumulation at a solid surface. Any solid in solution experiences an imbalance of forces at the solid/solution interface. This imbalance of forces will tend to be reduced by adsorption of ions from solution (Schindler, 1981). At any solid/solution interface there is a surface Charge and an electrical potential gradient extending from the surface into the solution phase (Leckie et al. 1980). The origin of this surface charge can be a result of several processes including, 1) chemical reactions 164 165 at the solid surface, such as non-stoichiometric dissolution or ionization of surface functional groups, 2) substitution Of’ unlike charged. particles in the lattice, or 3) ion adsorption from solution (Stumm and Morgan, 1981). Thus the forces which cause the adsorption of metals may come from chemical interactions, such as the formation of covelant bonds, from electrostatic interactions, such as ion- exchange, and from other forces such as van der Waal's forces or hydrogen bonds (Stumm and Morgan, 1981). The surfaces of hydrous oxides such as Fe and Mn oxides can be pictured as an amphoteric substance consisting of surface hydroxyls (Schindler, 1981). The charge of the oxide results from a reduced coordination number of the metal ions on the surface layer of the oxide. Thus the surface exhibits acidity (Schindler, 1981). The charge is controlled by the hydration. and hydrolysis of this surface and the participation of surface hydroxyl groups in acid-base reactions (Murray, 1974). The dissociation of these surface groups can be represented by the following reactions: SOH2+ ==== SOH° + H+ SOHO ==== so' + H+ where S represents a surface layer metal ion such as Fe or Mn. It can be seen from these reactions that the surface charge is dependent on the pH of the solution. At equilibrium, the electrical potential of a chemical species must be the same in all phases of the system. Hence the ions 166 H+ and OH' are usually chosen as the potential-determining ions (PDI) for calculating the surface potential (Leckie et al. 1980). Since the surface charge is influenced by the pH of the solution, there is a unique pH for each solid where the surface is uncharged. This pH is commonly termed the pH of zero-point-charge (szpc) (Stumm and Morgan, 1981). This pH, which is characterized by the condition: + _ _ [SOH2 ] - [so ] will depend on the acidity of the metal ion and the electrostatic field strength of the solid (Murray , 1974) , and will thus be unique for each oxide. Healy et al.,(1966) demonstrated that the pH for a series of Mn oxides was zpc related to the crystallinity, in that as the atomic packing of the lattice increases the PHzpc increases. When the pH of the solution environment is above the PHzpc of an oxide surface, the oxide surface will have a negative charge, and conversely a positive charge when the solution pH is less than the pH c (Parks, 1965). The pH ZP of Mn oxides (MnOZ species) are about 1.5-2.7 and for Fe zpc oxides (Fe(OH)3-am: FeOOH) are about 7.9-8.5 (Kinniburgh and Jackson, 1981). Thus in many natural waters Fe oxides will have a positive surface Charge and Mn oxides a negative Charge. Although the electrostatic attraction/repulsion is important in adsorption energetics, adsorption does take place on surfaces with an unfavorable charge. In other words 167 a cationic species can adsorb onto an oxide species below the pH where the net Charge on the oxide surface is zpc positive. This shows that. other’ bonding’ mechanisms contribute to the energy of adsorption. The typical pattern of cation adsorption with variable pH is marked by an abrupt increase in adsorption from near 0% to near 100% over a several unit increase in pH. This area of abrupt increase in. adsorption is termed the adsorption edge (Leckie et al.,1980). Anion adsorption patterns are a mirror image of cation adsorption with the abrupt adsorption occurring over a several unit decrease in pH. These patterns of adsorption have led to analogies of solution processes: hydrolysis of metals for cation adsorption and protonation of bases for anion adsorption. The location of the adsorption edge for various metals on an oxide surface is a function of the properties of the metal in solution (i.e. size, pH of hydrolysis) and on the amount of solid substrate in the system (Kinniburgh and Jackson, 1981), but is apparently only weakly influenced by the identity of the solid (Leckie et a1. 1980). For cation adsorption, an increase in the metal/oxide ratio causes the adsorption edge to shift to a higher pH. In other words, increasing the amount of metal in solution while keeping pH and the oxide concentration constant will result in a decreasing fraction of adsorption, which is the opposite effect in precipitation reactions (Leckie et al.,1980). Anion adsorption is affected in the same manner but in a 168 mirror image, that is, increasing the metal oxide ratio causes the adsorption edge to shift to a lower pH. While the analogies to solution chemistry are obvious, adsorption is set apart thermodynamically from solute reactions by the electrostatic interactions between surface and solute. The total free energy of adsorption must be broken down into its Chemical and electrostatic components: AGadsorp = AG’chem + AGcoul because the electrostatic term AG 1 changes with the cou reaction of ions with the surface and with the electrostatic interactions of adsorbed species (Morel et a1., 1981). The IASGChem contains the energy of chemical bonding and other non-electrostatic components. This free energy expression is the basis for all models put forward to explain and predict the adsorption process. In the last twenty years a variety of adsorption models have been developed to explain experimental adsorption data. The ultimate aim of such models is to be able to predict the adsorption of various species over changing conditions of pH, sorbate/sorbent concentrations, and solution chemistry. Any model that will be useful to model natural systems must also be able to account for the effects of competing cations and ligands. Early models of the solid-solution interface, such as the model of Guoy' and. Chapman, considered. an. entirely electrostatic theory of adsorption based on an electrical 169 double layer (EDL) model of the interface. The EDL concept defines the area at the interface which has different properties than bulk solution as having two layers or planes each with a distinct surface potential. The Guoy-Chapman model considered ions as point charges that were held to the surface by electrostatic forces in the diffuse, outer layer of the EDL. The Guoy-Chapman model was modified by Stern and later Graham to include non-electrostatic specific adsorption, such that specifically bound ions were bonded to the surface in the compact layer (Stern layer) nearest to the surface (Stumm and Morgan, 1981). This model works well for predicting electrical phenomena at some interfaces but results in anomalies when applied to the metal oxide/solution interface (Leckie et al.,1980). The evolution of adsorption models which describe adsorption on hydrous oxides has involved the consideration of discreet surface species reacting with solution species. Such models must consider both chemical and electrostatic interactions and are still constrained by the charge-balance and charge-potential relationships of the electrostatic EDL models (James, 1981). .A model described by James and Healy (1972) recognized that the adsorption edge of hydrolyzable metals is strongly dependent on the pH of hydrolysis of the metal. This model does not attempt to demonstrate direct surface-metal bonds but suggests that adsorption is due to the preferential uptake of hydrolyzed metal species. This hydrolysis- 170 adsorption reaction may take place at pH's below the pH of hydrolysis in solution thus may also be thought of as the surface hydrolysis of adsorbed metals. The free energy of adsorption in this model is then broken down such that: AGsorp = AGchem T AC;coul + AGsolv where the A Gsolv term is the energy to replace secondary hydration water of a metal with interfacial water of low dielectric (James and Healy, 1972). Thus this model predicts that the hydrated metal species is adsorbed. The A Gchem term is undefined and used to correct for the discrepancy in electrical potential in the EDL caused by considering discreet ions and real surfaces (James and Healy, 1972). Recent developments in adsorption models have attempted to define this reaction as a chemical rather than physical process, although all such models are defined by a consistent set of stoichiometric and energetic expressions rather than by any proven mechanism (Morel et al.,198l). Models generally applied to adsorption onto hydrous oxides consider the oxide surface as a polyprotic acid (Schindler, 1981) and the adsorption of trace metals as a complexation reaction with the formation of covelant bonds between the surface site and metal ion (Leckie et al., 1980). These models must still consider the electrostatic energy term separately from the other energy components. The way in which the various models are different from each other has to do with how the chemical and electrical components of the total free energy are formulated. The 171 species predicted to react at the surface and the location Of the adsorbed ions with respect to the EDL can be modeled by a variety of theoretical formulations, any of which can result in the ability to model experimental data. Morel et al., (1981) compared several recent adsorption models and concluded that all were too flexible, and that the models differ from each other in three general areas: 1) the set of surface species and surface reactions considered, 2) the expression of the "mass law", and 3) the calculation of the coulombic term. These differences can be demonstrated by comparing two recent models of adsorption on hydrous oxides: the model of Stumm et al., (1970) and Schindler et a1., (1976) hereafter referred to as the surface-complexation model, and the model of Davis and Leckie (1978) hereafter referred to as the site-binding model. Both are "chemical" rather than "physical" models, each formulating a distinct set of reactions between the solute ions and discreet surface sites. The set of surface species considered (i.e. SO”, SOH, SOHZT) are the same for both models but there are differences in the surface reactions considered, including both trace metal and electrolyte interactions. The surface- complexation model predicts that the unhydrolyzed free metal is the primary adsorbed species (SO-Me+) while the site- binding model predicts simultaneous adsorption and hydrolysis such that (SO-MeOH) is generally the primary 172 adsorbed species. By considering different solution species and different degrees of hydrolysis or protonation of the surface species it is possible to write 'more than one reasonable reaction to describe the observed stoichiometry of experimental data. Many times more than one reaction is required to model the experimental data. The two models also differ with respect to ionic strength effects. The surface-complexation model does not consider electrolytes to be specifically adsorbed. Hence the adsorption constants generated are conditional constants containing all the effects of the background electrolytes, which are not clearly understood, but may include activity coefficient corrections, compression Of the EDL, and those dependent on the particular electrolyte (Morel et al.,198l). The site-binding model allows for the specific adsorption of weakly bound electrolyte ions and thus adsorption constants applicable over varying ionic strength. Mass law expressions can vary both in units and in the treatment of polydentate species (Morel et al., 1981). The units may be based on the number of surface sites (moles of sites per liter) or on surface area (area of sites per total surface area). If the area of all surface sites are identical then the adsorption constants will be independent of the units chosen. Models may also differ in how polydentate surface reactions are handled. The surface- complexation model considers a bidentate bond as a bond between the solute ion and two identical adjacent surface 173 sites while the site-binding model considers distinct monodentate and bidentate surface sites analogous to bidentate complexes in solution (Morel et al.,1981 and Leckie et al., 1980). The resultant mass law expressions: ,. g Eififlz’ffl K 3533121331 [SO]2[Me] [(801211Me1 show that there is a square dependence on surface site concentration in the surface-complexation model and not in the site-binding model. There are many different ways to formulate the coulombic term, Accoul, of the adsorption process. Most are a variation of the general expression: 10Kcoul = (-l/RT)AGcou1 = -AZ(-F/RT)r where )k is the electrical potential at the locus of adsorption andWZXZ is the net change in charge number of the surface species due to adsorption (Morel et al., 1981). The differences between the various models arises from the different "pictures" of the EDL in terms of where a particular ion (PDI, specifically sorbed ion, electrolytes) is considered to be adsorbed. In other words, the models differ' in ‘the assignment. of ‘the electrostatic jpotential experienced by an adsorbed species and hence in the equations used to relate surface potential to surface charge (Westall and Hohl, 1980). For example, the surface- complexation model suggests that the PDI's and specifically 174 sorbed ions experience the same potential (i.e. both are adsorbed in the compact layer of the EDL) and that electrolytes experience a different potential in the diffuse layer. In contrast, the site-binding model, sometimes referred to as a triple-layer model, pictures the PDI's (H and OH) in an inner compact layer (i.e. essentially part of the solid) and the adsorption of both specifically sorbed species and weakly bound electrolytes to occur in an outer compact layer experiencing a different electrical potential. It must be noted that the formulation of the coulombic term must be consistent with the formulation of the physical picture Of the adsorption model (i.e. the predicted surface species). As stated previously, the purpose of adsorption models is to be able to explain real data over a variety Of conditions. Studies comparing the ability of several models to predict experimental data (Westall and Hohl, 1980 and Morel et al., 1981) have shown that there are so many adjustable parameters in these models, that nearly all of them can be made to work quite well. This shows that the models are all too flexible. The fundamental problem remains to be the inability to separate the AGcoul term from the AGchem term in the expression of the free energy of adsorption. The approach has been to isolate the AGchem term by extrapolating measured experimental data to a condition of zero surface potential and surface charge (i.e. A Gcoul=°) (Anderson et 175 al., 1981: Westall and Hohl, 1980). However, the path chosen for' this extrapolation. is :model dependent, resulting in variations in the A Gchem term from model to model. Upon determination. of the [X GChem. term, which is a true thermodynamic constant, the adsorption model attempts to predict adsorption over a range of solution parameters by its theoretical formulations of the electrostatic interactions. To do this the surface Charge is calculated based on the requirements of the particular model, and then, by its relationship to capacitance, the surface potential is calculated. However, the capacitance is merely a fitting parameter, and hence the surface potential necessary to fit the experimental data can always be determined. Hence each model can calculate the "correct" free energy of adsorption but with different contributions of AGchem and AGcoul. So what is lacking for the formation of a single, less general model is independent verification of the true surface speciation of the adsorbed species and of the true surface potential (Anderson et al., 1981; Leckie et al., 1980). Analytical determination of these parameters is very difficult or impossible by present methods. CHEMICAL CONTROLS OF ADSORPTION BEHAVIOR Given the nature of the present research there are two topics with respect to adsorption behavior that will be discussed. The first is a summary of research on the competition of sorbate ions for the surface of hydrous 176 oxides. This discussion of multi-sorbate interactions will be broken down into four sections: 1) comparisons of the relative affinities, or selectivity sequences for a group of ions as determined from single sorbate/sorbent systems, 2) experimental results of the actual competition between metals for surface sites, 3) the effects of ionic strength on metal adsorption, and 4) the effect of complexing ligands On metal adsorption. The second topic of consideration, which is necessary background for mixed sorbent experiments, is the comparison of the adsorptive properties of Mn and Fe oxides. Of special interest are the unique properties of the MnOz surface. There have been many studies that have measured the relative adsorptive properties of a series of ions on a particular oxide surface. By comparing the amount of metal adsorbed at a certain pH and metal/oxide concentration, or by determining the pH required to sorb a certain amount of metal at constant oxide concentration, the selectivity sequence of a group of metals can be determined. The lower the pH that a cationic metal ion adsorbs, the greater affinity that metal has for the surface. Table 1.1 summarizes the selectivity sequences for alkali metals, alkali earth metals, and transition metals for oxides of Mn and Fe. It should be noted that comparisons between these groups yields the following relationship: transition metals > alkali earth metals > alkali metals in terms of the affinity for oxide surfaces (Kinniburgh and Jackson, 1981). 177 TABLE I.1 SELECTIVITY SEQUENCES OF ADSORBED IONS Investigation Substrate Selectivity Sequence Stumm et a1, 1970 Mn02* Cs > Na Posselt et a1, 1968 MnOz Ba > Sr > Ca > Mg Murray, 1975 MnOz Ba > Sr > Ca > Mg Kinniburgh et al, Fe(OH)3 Ba > Ca > Sr > Mg 1976 McKenzie, 1980 MnOZ Pb > Cu > Mn > Co > Zn > Ni Gadde and Laitinen, MnOz Pb > Zn > Cd 1974 Murray, 1975 MnOz Co 2 Mn > Zn > Ni Loganathon and MnOZ Co > Zn Burau, 1973 Kinniburgh, et a1, Fe(OH)3 Pb > Cu > Zn > Ni > Cd > Co 1976 . Gadde and Laitinen, Fe(OH)3 Pb > Zn > Cd 1974 Leckie et a1, 1983 Fe(OH)3 Cr > Pb > Cu > Zn > Cd > Ni Forbes et a1, 1976 FeOOH Cu > Pb > Zn > Co > Cd McKenzie, 1980 FeOOH Cu > Pb > Zn > CO > Ni > Mn * all MnOz species are low PHzpc forms 178 The selectivity sequence for monovalent cations, which are thought to adsorb due primarily to electrostatic forces, is found to vary from surface to surface, apparently due to the acidity and pH of the surface considered (Kinniburgh and zpc Jackson, 1981). The. selectivity sequence for' the. alkali metals on. Mn02 (Cs+ >..>Li+) is thought to be a size dependent sequence with the larger ions (i.e. Cs) having a smaller hydrated radius. The smaller, more polarizable hydration spheres are also more distortable allowing closer approach to the surface (Kinniburgh and Jackson, 1981) . Another factor that may influence the selectivity sequence is the polarity of the surface. Solid surfaces with high polarity (i.e. high pH tend to have a high degree of zpc) structure in the surficial water molecules. The selectivity sequences on these types of solids tends to be opposite that on MnOz (i.e. Li+ >..> Cs+) depending on the tendency of the ion to break up this structured water. This relationship is not, however, unequivocal (Kinniburgh and Jackson, 1981). The selectivity sequences for the alkali earth metals varies greatly from surface to surface (see Kinniburgh and Jackson, 1981). From Table 1.1 it can be seen that the sequence .for low pH Mn02 (Ba > Sr > Ca > Mg) is zpc consistent in different studies and also different from the sequence on amorphous iron oxides (Ba > Ca > Sr > Mg). This suggests there must be some factors involved that are specific to the surface considered. For MnOz the sequence follows a size relationship with increasing affinity with 179 decreasing hydrated radius (Posselt, et. al., 1968). There is also a general relationship with the tendency to hydrolyze (Kinniburgh and Jackson, 1981). There appears to be a strong dependence on the ability to hydrolyze in the selectivity sequences for the transition metals (Leckie et a1., 1983 and others). However, for MnOZ and amorphous Fe oxides, the order of Pb and Cu is switched compared to crystalline forms of Fe oxides suggesting there are surface specific effects (McKenzie, 1980). There are also discrepancies in the Mn oxide system, probably due to redox reactions with species such as Mn and Co (Hem, 1978: Murray and Dillard, 1979). However, despite these deviations and some pH dependent order changes, there is certainly a strong dependence on the ability to hydrolyze. In general, the greater the electronegativity of the metal ion, the greater the tendency to form covelant bonds with surface oxygen atoms and hence the greater the affinity for the surface (Kinniburgh. et. al., 1976). This relationship is consistent with the adsorption models discussed previously. Research in single sorbate systems which has generated selectivity sequences suggests that there are different binding energies for different metals and that some metals should be able to "out-compete" other metals for surface sites. Many studies have documented the adsorption of trace metals in the presence of a large excess of Ca, Mg, or Na which suggests either 1) trace metals bond more strongly to the available surface sites and compete favorably, or 2) 180 these ions bind to different sites on the oxide surface. In order to determine the nature of "competition" amongst metals and whether adsorption models can be used to predict competition in mmlti-sorbate systems, several studies have measured adsorption in such experimental systems. Table 1.2 summarizes some of these studies. From these results it does not appear that competition among adsorbates can be easily predicted or that it is simply the case of the more strongly bound species adsorbing at the expense of the weaker binding species. In fact, despite 'the 'very’ different binding' energies for 'various metals, competitive effects are weak or non-existent in some cases. This suggests that metals may not compete for the same binding sites on the oxide surface. Benjamin and Leckie, (1981b) studied the adsorption of Cd onto Fe203-H20 am in the presence of other stronger binding metals in a 10- 100 times excess. There was only limited competition under these: conditions. These authors suggest. that this oxide surface consists of a variety of sites, most of which will bind to a many different metals, but with different binding strengths unique for each metal. In other words, a site which is a high energy site for Cu adsorption may be only a weak energy site for Pb. Evidence for different types of surface sites with different bonding energies has been provided in single sorbate experiments as well (Benjamin and Leckie, 1981a). These investigators showed that the average binding constant 181 TABLE 1.2 SUMMARY OF HULTI-SORBATE ADSORPTION STUDIES . . Primary Competing Investigation Substrate Metal Metal Results Dempsey & Fe(OH)3 Zn Ca NO competition Singer, 1980 Mn02 Zn Ca Slight competition at low pH only Balistrieri & FeOOH Pb, Cu, the other No competition ' Murray, 1982 Zn or three except when Cd is Cd primary metal Swallow et al, Fe(OH)3 Pb,Cu Cu,Pb No competition 1980 Gadde & .Mn02 Cd or Pb Decreased sorption Laitinen, 1974 Zn of Cd and Zn by Pb Benjamin & Fe203-H20 Cd Cu,Pb Slight or absent Leckie, 1981 amorph. or Zn competitive effect Cu Pb Weak competition 182 for Cu, Cd and other metals on Fe203-HZO am changed as adsorption density reached a certain level, even though there was still a large excess of unoccupied sites. They argued that electrostatic effects due to EDL or sorbate- sorbate interactions could not be responsible for this phenomena, and suggested the hypothesis of different sites with different binding energies to explain the data. The change in the average binding constant was attributed to the condition where the high energy sites (which preferentially fill up first) become limiting. The "concentration“ of these high energy sites appears to be unique for different metals. This is an important concept in terms of the application of existing adsorption models to real systems, as all of the models described previously consider only one type (energy) of surface site and thus may have limited utility in describing adsorption over a wide range of sorbate/sorbent conditions. It also suggests that adsorption onto Fe oxides may be accurately modeled without considering competition between trace metals which would greatly simplify this task (Benjamin and Leckie, 1981b). However, based on the limited data for competition studies on Mn oxides (see Table 1.2) it appears that this may not be the case for all surfaces. Studies which have explored the effects of the swamping electrolyte (i.e. ionic strength effects) have also resulted in a variety of results (see Table 1.3). Although electrolytes have been shown to bind to the surface less strongly than trace metals it is clear that these species 183 TABLE I.3 SUMMARY OF IONIC STRENGTH EFFECTS ON ADSORPTION Investigation Sorbent Metal Electrolyte Results (conc. range M) Kinniburgh Fe(OH)3 Ca,Sr NaNO3 small decrease in et a1, 1975 (0.4 - 2.0) adsorption Posselt et al, M002 Ca NaClO4 proportionate 1968 (.004 - .134) decrease in log of sorption capacity with ionic strength Swallow et al, Fe(OH)3 Pb,Cu NaClO4 NO effect 1980 (.005 - .50) 5 NaCl and Pb adsorption SOW decreased by presence of Cl Balistrieri FeOOH Pb,Cu, .1 NaNO3 to SOW increase in Cu and Murray, Zn,Cd by stepwise adsorption 1982 Of .53 NaCl, increase in Pb adsorption at low pH and less at high pH decrease in Zn and Cd adsorption 184 must be considered as they are usually present in great excess. The effects of ionic strength on the adSorption of trace metals are unclear but may include, 1) electrostatic effects caused by changes in the charge of the EDL, 2) activity coefficient effects (decreasing metal activity with increasing ionic strength), 3) competition for binding sites on the oxide surface, and 4) competition with the surface via the formation of solution complexes (which may or may not adsorb). Results from studies employing "inert" electrolytes (such as NaNO3 and NaClO4) to set ionic strength show that these species have only minor effects on the adsorption of trace metals, at least. on Fe oxides (Kinniburgh. et al.,1975: Swallow et al., 1980). This suggests that the electrostatic EDL effects of ionic strength on adsorption of trace metals are minor, possibly because adsorption of trace metals result in no net change in surface charge (Swallow et al., 1980). Balistrieri and Murray, (1982) explored the influence of the components of major ion sea water (NaCl, NaZSO4 and MgClz) on the adsorption of trace metals by geothite. Although not explicitly a study of ionic strength, this research demonstrated both electrostatic and competition effects may have been responsible for the difference in adsorptive behavior in 0.10 M NaNO3 compared to synthetic major ion seawater. Apparently Mg and 804 caused both increases and decreases in adsorption depending on the 185 metal. It should be noted that these researchers found no effect by C1 on the adsorption of Pb, while Swallow et al., (1980) found a pronounced decrease in adsorption on amorphous Fe oxide. It may well be that electrolytes may have different effects on different surfaces, or possibly even variable effects on a single surface under different solution conditions. Another process which must certainty be considered for predicting adsorption in complicated natural systems is solution complexation of metal ions by complexing ligands. The potential effects on adsorption include, competition for metal ions (i.e. the complex does not adsorb), the adsorption of ligands onto the surface resulting in a change in surface charge, and the adsorption of complexes. The latter is indistinguishable from the complexation of a metal by an adsorbed ligand. Studies by Davis and Leckie (1978b) and Benjamin and Leckie (1981c) have shown that complexing ligands cause only slight effects on the charge of the EDL in the presence of a swamping electrolyte and that complete non-adsorptive behavior of complexes is probably rare. These studies indicate that in general a metal-ligand complex exhibits either "metal-like" or "ligand-like" adsorptive behavior and this can cause either an increase or decrease in adsorption. The effect on adsorption may also be pH dependent and will almost certainly depend on the identity of the solid. The difference between "metal-like" and "ligand-like" complexes 186 is apparently due to the stereochemical arrangement of the complex at the surface rather than the overall charge on the complex, with ”metal-like" complexes adsorbed with the metal closer to the surface (Leckie et al.,1980). In general, simple, soluble ligands like $04 and Cl tend to form metal- like complexes which tend to decrease metal adsorption (Leckie et al., 1980). It is certain that surface speciation is vital for the development of an adsorption model that will work well in predicting the behavior of metals in the environment. ADSORPTION ON MN VS FE OXIDES From the preceding discussions it is evident that the adsorption of trace metals is somewhat unique for each oxide surface considered. A good case in point are the apparent differences between hydrous iron and manganese oxides. Because some of the experiments in the present research are designed to examine the competition for Cr3+ by these two oxide species, it is valuable to summarize the apparent differences based on past research. It must be noted that only a few studies (McKenzie, 1980: Balistrieri and Murray 1982; Gadde and Laitinen, 1974: Dempsey and Singer, 1980) have evaluated adsorption on these two oxides for the purpoSe of making a comparison. Some apparent differences may be due to widely different experimental parameters, procedures and interpretations. Table 1.4 summarizes the main differences in adsorption behavior 187 TABLE 1.4 APPARENT DIFFERENCES IN ADSORPTION BEHAVIOR OF Mn AND Fe OXIDES BASED ON PAST RESEARCH Component of Behavior Mn Oxides Fe Oxides 1. Adsorption of anionic species 2. pH dependence of metal adsorption 3. Stoichiometry of the reaction in terms of proton release 4. Capacity and intensity of adsorption 5. Type of bonding mechanism 6. Competition amongst adsorbates Anions do not adsorb Adsorption edge is gradual and begins at very low (<2) pH's Variable but most report near 1 H+ released per metal adsorbed - May increase with increasing pH Higher amount of surface sites per mole of solid and some evidence for higher energy of interaction Several have been suggested. More difficult to model by current models maybe due to difficulty in characterizing the surface Although data is scarce, it appears that some metals may compete for the same sites suggesting that stronger binding metals should sorb at the expense of other species. Anions adsorb at most pH's Adsorption edge is abrupt with very little or no adsorption at very low pH's. 1.2 to 2 H... released per metal adsorbed with most studies >1.5 Amorphous iron oxides have similar surface site conc. as M002 - Geothite has much lower capacity Can be modeled by surface-complexation models which suggests that binding to surface hydroxls is dominant Rather extensive study indicates that competition among metals is minimal which suggests metals adsorb mainly to specific sites 188 between Mn02 (low pH variety) and Fe oxides (geothite and zpc amorphous iron hydroxide). The lack of anion adsorption on Mn oxides compared to Fe oxides has been demonstrated by Balistrieri and Murray (1982). This can be explained partly in terms of the vastly different electrostatic conditions at the surfaces of these two oxides. The surface of Mn oxide is negative for most of the pH ranges studied, and electrostatic repulsion of anions would be expected. However, Leckie et al., (1980), have shown that the apparent lack of adsorption of anions above the PHzpc of amorphous iron hydroxide is an experimentally induced condition rather than an indication of complete electrostatic adsorption behavior of anions. This suggests that anion adsorption on Mn oxides is not theoretically impossible under certain experimental conditions, however it is apparently unimportant under most conditions. The pH dependent adsorption behavior on MnOZ appears to be unique compared to most other hydrous oxides. This Observation is based on the location and shape of the adsorption edge for trace metal adsorption. First of all, significant adsorption takes place on Mn02 at pH's below 2.0, near the pH and increases with a gradual slope as zpc' pH increases (McKenzie, 1980: Gadde and Laitinen, 1974). This is in contrast to pH dependent adsorption on Fe oxides which follows the classic adsorption edge pattern and rarely begins at pH's lower than 3.0. The adsorption of metals at very low pH by MnOz has not been satisfactorily resolved, 189 but could be related to, 1) the low PHzpc (i.e. the lack of an unfavorable positive surface Charge), 2) different mechanisms of bonding (i.e. exchange of metals for structural Mn) and/or 3) the relatively high dielectric constant of this oxide surface. Although the electrostatic term is certainly more favorable for adsorption on MnOZ at low pH's, Murray (1975) suggested that the szpC should not be the major consideration in the adsorption behavior of oxides. MnOZ was shown to exhibit similar adsorption behavior with TiOz, which has a high pH and high zpc dielectric, while adsorption on SiOz, which also has a low szpC but a low dielectric, exhibited quite different behavior. This was interpreted to show that the dielectric of the solid influences the adsorption behavior. The MnOz surface, which has a relatively high dielectric, would be more favorable for the adsorption of the unhydrolyzed species due to the relatively small change in the solvation energy required to move the ion to the surface. The exchange of :metals for structural Mn. has been demonstrated to occur by measuring the Mn released to solution upon the adsorption of metals (Loganathon and Burau, 1973: McKenzie, 1979: Murray, 1975). However, the amount of Mn released does not suggest that this is an important mechanism for metal adsorption. Many investigators have measured the release of protons upon the adsorption of metals from solution in order to demonstrate the reaction stoichiometry. As with many other 190 results of adsorption research, there is a wide variety in the experimental results (see Table 1.5) many of which reflect the lack of standardized experimental methods. In very general terms, it appears that the release of protons upon adsorption of divalent trace metals (hereafter referred to as HT/MeZT) is lower for adsorption on Mn oxides (1.0) than for adsorption onto Fe oxides (1.5-2.0). There are several issues concerning the interpretation of this experimental data that have not been resolved. One very significant problem is whether the release of electrolytes should be included in adsorption stoichiometry. McKenzie (1979) , showed that the release of 2 protons per metal adsorbed at low ionic strength could be systematically reduced to 1 by increasing the ionic strength with KNO3. He suggested that at low ionic strength there was one proton released from the surface by exchange with the metal and that the other proton was released from the diffuse layer to balance the charge. At higher ionic strength the diffuse layer counterions were primarily K" which explains the reduction of the H"'/Me2+ ratio. It was suggested that the H+/Me2+ ratio of 2 determined by Loganathon and Burau (1973), was due to the use of an acid washed sample which left H+ as the diffuse layer counterion. The interpretation of McKenzie (1979) does not consider specific adsorption of electrolytes, which could also be used to explain these results. For instance, the work of 191 TABLE 1.5 PAST STUDIES OF ADSORPTION REACTION STOICHIOMETRY A + . . H Investigator pH Ox1de Metals Released Comments Gadde and 6.0 N002 Pb 1.4 Laitinen Cd 1.3 .(1974) Zn 1.1 Gadde and 5.0-6.0 HFO Pb 1.2-1.6 Increased Laitinen with pH (1973) Loganathon 4.0 MnOZ Co 2.0 2 = H-K-Na and Burau Oxide was (1973) acid washed McKenzie 4.0 Mn02 Pb,Cu 1.0-2.0 Dependent (1979) Mn,Zn on ionic strength Morgan and 4.5-8.2 MnOZ Mn 1.0-1.7 Increased Stumm (1964) with pH McKenzie 4.0 MnOz Pb,Cu 1.0 Several (1980) Mn,Zn forms of Mn oxide 5.0 FeOOH Pb,Cu 1.3-2.0 Mn,Zn Forbes et 4.5-9.4 FeOOH Cd,Co,Cu 2.0 Used two a1 (1976) Pb,Zn methods Balistrieri 3.4-5.3 Mnoz Ca,Mg <1.0 However and Murray H+Na+K = 2.0 (1982) Murray 3.5-7.0 MnOz CO,Mn,Ni 1.0 Increased (1975) with pH Leckie et Fe203-am Cu 1.9 pH = 5.47 a1 (1980) Pb 1.65 pH = 5.35 Cd 1.8 pH = 7.0 Zn 3.2 pH = 6.75 192 Balistrieri and Murray (1982) using the site-binding model shows that the ratio of (H + Na + K) released to (Ca + Mg) adsorbed onto MnOz equals two, which suggests that these species adsorb to bidentate surface sites (HT/MeZT < 1.0) and that the charge in the inner adsorption plane is conserved. It is not known whether the adsorption of trace metals also results in this behavior, but these authors have suggested that adsorption of Ca and Mg on iron oxide occurs on monodentate sites (Balistrieri and Murray, 1981). Further evidence for the model of McKenzie (1979) which predicts that the inner layer of the MnOz surface should become positively charged with increasing adsorption has been provided by studies which have measured the sign of the surface charge as a function of adsorption. Loganathon and Burau (1973), Morgan and Stumm (1964) and Murray (1974) have demonstrated that the MnOZ surface becomes less negative and eventually experiences a charge reversal with increasing adsOrption of metals. This phenomena could be due to the exchange of a proton for a divalent metal, resulting in an accumulation of positive charge in the sinner plane of adsorption. This would strongly suggest that a HT/Me2+ ratio Of 1.0 is correct, at least for some experimental conditions. McKenzie (1980) suggests that the adsorption of metals at low pH's by MnOz is due to the adsorption Of the free unhydrolyzed metal ion, and thus the release of one proton by the reaction: 193 SOH° + Me2+ = SOMe+ + H+ At higher pH's the adsorption of the hydroxo species (or the hydrolysis of the ion at the surface) by the reaction: SOH° + Me2+ + H20 = SOMeOH + 2H+ also becomes important, which would cause an increase in the H+/Me2+ ratio. The final reaction at even higher pH's (if the metal concentration is high enough) would be the surface precipitation of the metal species: 30‘ + Me2+ + 2H20 = SOMe(0H)2 + 2H+ 2+ should increase with This scenario suggests that the HT/Me pH to a ratio greater than 1.0. Several investigations have demonstrated this relationship (Morgan and Stumm, 1964: Murray, 1975). Applying a consistent approach to adsorption on Fe oxides would suggest that there is no appreciable adsorption of only the unhydrolyzed species on these oxides. The location of the adsorption edge for various metals on iron oxides does appear to be related to the appearance of the hydroxo metal complex, although the free metal is still commonly in great excess (McKenzie, 1980). This would suggest that the hydrolyzed species may be preferentially adsorbed. The H+/Me2+ ratio for adsorption on iron oxides supports this idea, as this ratio is usually reported to range from 1.5 to 2.0 (Leckie et al.,1980: McKenzie, 1980: Forbes et al., 1976). One could speculate that the lack of adsorption at low pH's (i.e. the unhydrolyzed species) on 194 iron oxides may be related to the inability to overcome the solvation energy. On a weight basis, the adsorption capacities of MnO2 and amorphous Fe oxide are on the same order of magnitude, although. the crystalline Fe oxides have a significantly lower capacity (Louma and Davis, 1983). Based on the ability to adsorb at lower pH's, MnOZ would appear to be a more efficient adsorbent than Fe oxides (Gadde and Laitinen, 1974: McKenzie, 1980). It must be emphasized however, that multi-sorbent experiments have not been reported and the competition. between solids for metals ‘will likely' be a complicated process. Based on the above discussion the following points would appear to summarize the pertinent points to consider for the mixed oxide experiments presented in this research: 1. Although. the. adsorption capacities are similar; MnOz should be a more efficient adsorbent than amorphous Fe oxides at lower pH's, such as those used in the present research. This may be due to the ability of MnOz to adsorb the unhydrolyzed metal ion. 2. The adsorption of the oxidized, anionic form of Cr by Mn02 should be negligible: but the adsorption of this species by Fe oxides in multi-sorbent systems will be significant. 3. Although the data is scarce, it appears that adsorption on MnO2 may be more susceptible to competitive effects 195 in multi—sorbate systems than Fe oxides. This, however, undoubtedly will be strongly dependent on solution characteristics. ADSORPTION OF CHROMIUM BY HYDROUS OXIDES Compared to other hydrolyzable metals there has been 3+ onto oxide relatively little study of the adsorption of Cr surfaces. This is probably due, in part, to the low solubility of this species in the pH range of natural waters. Some of the previous research on this subject is summarized on Table 1.6. A recent study by Leckie et al., (1983) is the most complete study to date of Cr3+ adsorption. The important conclusions from this and other studies can be summarized in the following list: 3+ on Fe and Al oxides 1. The pH dependent adsorption of Cr is typical for hydrolyzable metals, with the adsorption edge occurring at around pH 3-5 for all of the substrates studied. The adsorption edge experiences the normal shift to higher pH's as the concentration of the oxide is decreased. 2. The adsorption edge of Cr3+ is at a lower pH than most other metal ions, suggesting a relatively stronger binding interaction (Leckie et al., 1983). 3. Based on the multi-site model of Benjamin and Leckie (1981a), it would appear that there are an abundance of 196 high energy sites for Cr3+ adsorption on Fe203‘H20 am, as the adsorption edge does not shift as Cr3+ concentrations are increased from 10'”7 to 10'5 M. This is in contrast to the behavior Of other metals under the same experimental conditions, and provides further evidence for a strong adsorptive bond. 4. The presence of the complexing ligands SO42”, Cl' and A5043" had little effect on the adsorption of Cr3+ on Fe203'H20 am. However, in fly-ash transport water, the adsorption was increased relative to a "clean" system, suggesting that Cr3+-ligand complexes adsorb in a ligand-like manner (Leckie et al., 1983). 5. Cr3+ adsorption by clay minerals follows that predicted by the CEC of the mineral at low pH's. However, at pH=4 the CEC was exceeded on montmorillonite. Removal of Cr from solution by Clays is high compared to other metals (Griffin et al., 1977). Due to the strong binding of Cr3+ to the substrates that have been studied and the unique adsorption characteristics of MnOz, it is predicted that adsorption of Cr3+ by this oxide would be very strong even at low pH's. The effects of 3+ competition on this surface between Cr and other metals, however, are difficult to predict. There has been considerably more research on the 6+ adsorption of Cr on hydrous oxides, probably because this anionic form is both highly toxic and mobile due to its high 197 TABLE 1.6 PAST RESEARCH ON THE ADSORPTION OF Cr(III) Substrate Cr3+ pH Investigator (conc.) (conc.) (range) Comments James and SiOz 2X10"4 M 3-6 Adsorbed at Healy, 1972 lower pH's than (75m2/L) most other ions Leckie et Fe203'H20 5x10“S to 3-6 Studied ion/ al, 1983 oxide ratiOs, (4x10'4 to 5x10’7 M kinetics and effect of 1x10'3 M) ligands A1203 1x10"s to 3-10 Adsorption edge similar to that (Z-ZOg/L) 1x10’7 M on Fe oxide Griffin et clays 80ppm 1-5 Uptake related al, 1977 0.1 g to CEC except montmor. for mont. at kaolinite pH = 4 198 solubility. Some of the studies of this topic are summarized on Table 1.7. As opposed to the trivalent form, hexavalent chromium is a relatively weak binder as compared to other anions. The pertinent aspects of the adsorption behavior of this species are summarized as follows: 1. The adsorption of Cr6+ follows the typical pattern for anions: the adsorption edge in a mirror image of cation adsorption. 2. Cr6+ also behaves as a typical anion in terms of adsorption capacity. Since anions are generally larger than cations, each adsorbed species covers more surface area, causing the surface to become saturated at a certain solute/solid ratio. 3. There is no evidence for appreciable adsorption of Cr6+ above the pH of Fe or Al oxides, which suggests the zpc electrostatic contribution to the adsorptive bond is very important (Leckie et al., 1983: Rai et al., 1986; MacNaughton, 1974). This idea is enhanced by the rather large decrease is adsorption as a function of increasing ionic strength with "inert" electrolytes (Rai et al., 1986: MacNaughton, 1974). 4. The adsorption of Cr6+ is greatly reduced by the presence of other anions. The effect has been attributed to both the competition for surface sites and the decrease of the positive surface charge caused by the adsorption of 199 the anions (Aoki and Munemori, 1982; Rai et al., 1986; Leckie et al., 1983). 5. The adsorption of cations results in only a very small increase in the adsorption of Cr6+. This suggests that these species are not adsorbed on the same sites and that the increase is due to the increase in positive surface charge and/or the adsorption of Cr-metal complexes. If the pH and cation concentrations are high enough to cause the surface precipitation of metal 6+ hydroxides, the adsorption of Cr is usually enhanced. This is due to the high pH of the metal hydroxide zpc relative to the oxide it has precipitated on (Aoki and Munemori, 1982; Benjamin, 1983). These results affirm that adsorption of Cr6+ by MnOz is very unlikely. It also suggests that hexavalent Cr formed by the oxidation of the trivalent species would tend to be removed from solution by Fe oxides, provided more strongly bound anions are not present in significant amounts. OXIDATION OF CR(III) BY MN OXIDES The ability of Mn oxides to participate in various redox reactions has been known for some time. Redox reactions between Mn oxides and metals such as Co, Ni, and Pb (Hem, 1978: Crowther et al., 1983), Fe2+ (Postma, 1985) and various organic compounds (Stone and Morgan, 1984) have been demonstrated in the laboratory. Several studies have also 200 TABLE I.7 PAST RESEARCH ON THE ADSORPTION OF Cr(VI) . Substrate Cr6+ pH Investigator (conc.) (conc.) (range) Comments Davis and Fe(OH)3 5x10"7 M 5-9 both HCrO4' and Leckie 1980 (10'3M) CrO42‘ adsorbed MacNaughton A1203 2x10"4 M 3-9 increase conc. 1974 of KNO3 reduces 12g/L adsorption Benjamin Fe203°H20 10"5 to 4.5-9 effects of 1983 trace metals on 10‘3-10"4 M 10"6 M adsorption Aoki and Fe(OH)3 10‘3 M 4-9 effects of both Munemori metals and 1982 5x10”3 to ligands 10'2 M Mayer and alumina 10"6 M 8 effects Of Schick kaolinite salinity 1981 sediment Rai et al, Fe203°H20 10’4 to 3-10 effects of a 1986 wide variety of .87 to 17.4 5x10'6 M cations & anions Leckie et Fe203'H20 10"5 to 3-11 effects of a al, 1983 wide variety of 8x10"4 to 10"6 M cations & anions 5x10‘5 M 201 shown ‘that. Mn oxides have the ability to catalyze the oxidation of Cr3+ to Cr6+ (see Table 1.8). It has been suggested that this process is a three step reaction consisting of adsorption, oxidation and desorption (Rai et al., 1986: Bartlett and James, 1979: Schroeder and Lee, 1975). From the following review of this past research it can be seen that some studies are of questionable merit and that there are still many unanswered questions concerning the reaction mechanism and whether this process may be environmentally significant. Schroeder and Lee (1975) presented some of the earliest 3+ is much faster in work to show that the oxidation of Cr the presence of Mn oxide than by 02 gas only. Although these authors reported their oxide only as "MnOZ" (reagent grade?) they demonstrated that in the presence of this oxide, 100% of their initial 0.125mg/L Cr3+ became oxidized within seven days, with 89% of the oxidation occurring within the first day. They also showed that the rate of oxidation increased as the concentration of MnOZ was increased. This reaction was strongly inhibited when natural lake water was used in place Of distilled water. The authors attributed this to competition for adsorption sites by cations such as Ca and Mg, and suggested that the oxidation reaction requires adsorption onto surface sites which are few in number. One problem with this study, is, that at a pH of 8.6, the formation of Cr(OH)3-solid may have been overlooked. The solubility of this species is only 7.4 X 10’8 M from pH 6.3 .202 TABLE I.S SUMMARY OF PAST RESEARCH OF THE OXIDATION OF Cr3+ BY Mn OXIDES Mn Oxide Cr(III) pH Solution Investigator (conc.) (conc.) Range Matrix Schroeder and unspecified 2.4x10'5M 8.6 KHCO3 Lee (1975) MnOz buffer (2.9):10‘3 to 2.9x10’4 M) Nakayama et a1 MnO(OH) 1x10‘5M 8.1 seawater (1982) (30mg/L) natural Mn nodules (SOmQ/L) Van Der Weijden reagent 1.9x10’5M 5.5 freshwater and Reith (1982) grade and 6.3 freshwater "natural" 8.1 seawater "1102 (1.15x10‘3) Bartlett and natural 1x10'5M 3.2-9.0 distilled James (1979) soil water: pH (2000:1) adjusted solution:soil with HCl, KHCO3 James and natural 1x10’3M 6.7 Cr(OH)3 or Bartlett soil Cr-citrate (1983) (3.09) formed in solution Rai et al., M002 1.9-38.5 3.0-4.7 distilled (1986) 6.3,8.3 water (1.4x10"1 to x 10‘5N 10.1 1.4x10‘3 M) 203 to 11.0 (Rai et al., 1986), which suggests the formation of this species may have formed to some extent and complicated the reaction process. Nakayama et al., (1981) studied the oxidation of Cr3+ by MnO(OH) (a trivalent Mn oxide) and by powdered, naturally occurring, deep-sea Mn nodules in a natural sea water matrix at pH 8.1. They found small percentages of the original Cr3+ spike (1 X 10'5 M) became oxidized over several hundred hours. They also demonstrated that the reaction was slower in the absence of 02 (i.e. N2 purged system) and that the presence of 10’3 M citric acid prevented the reaction from occurring. Again, the use of 10"5 M Cr3+ at pH 8.1 probably resulted in hydroxide precipitation. A similar study by Van Der Weijden and Reith (1982) using both reagent grade and "natural" MnOZ also demonstrated the oxidation of Cr3+ in seawater (pH=8.1), although Cr(OH)3- solid probably influenced these experiments as well. Their experiments were run for 1000 hours, but like the other studies, most of the oxidation appeared to occur in the first 10 hours or less. These investigators also ran experiments in fresh water at pH 5.5 and 6.3 and made measurements of the amount of Cr adsorbed during the duration of the experiments. They found the highest rate of oxidation at pH 5.5, with about 80% oxidized at the time of their first 'measurement. (within. 1 hour?). There ‘was no detectable adsorption at this pH over the entire duration of the experiment. At pH 6.3 the redox reaction was slightly 204 slower and a small percentage of the original Cr was found to be adsorbed. This adsorbed fraction decreased as the percent oxidized increased. This could be explained by the very slow oxidation of Cr adsorbed onto the Mn oxide, or possibly the adsorbed Cr is actually Cr(OH)3-solid which oxidizes at a much slower rate. At pH 5.5, the solution was probably undersaturated with respect to the hydroxide, which would explain why there was no "adsorption". The methods used in this study were not explained in detail, making their conclusions difficult to evaluate. Bartlett and James (1979) demonstrated the oxidation of Cr3+ by naturally occurring soils. The evidence provided for this oxidation being caused by oxidized manganese in the soil included; 1) there was no oxidation by low Mn soils, 2) there was no oxidation in acid soil samples where Mn is in its reduced form, 3) the oxidation of Cr was accompanied by an increase in the extractable Mn from the soil, and 4) the oxidation of Cr was proportional to the amount of Mn that could be reduced by hydroquinone. These authors recognized the solubility problem of Cr3+ at pH's greater than about 5.5 and studied the reaction at both low and high pH's. They demonstrated that the amount and rate of Cr oxidized decreased with increasing pH and attributed this to the decrease in solubility and the reduced rate of oxidation of the hydroxide precipitate. In a later paper (James and Bartlett, 1983) these authors also showed that small amounts of Cr(OH)3-solid did oxidize in Mn-rich soils and that the 205 aging of this species before the addition of the soil caused a reduction in the amount oxidized. They speculated that the fresh precipitate may have contained a residual positive charge that caused adsorption onto the Mn oxide surface. The most complete and experimentally sound description of the oxidation of Cr3+ by Mn oxide in a simple system is provided in a recent paper by Rai et a1., (1986). Using reagent Mnoz (pyrolusite) , the authors measured the oxidation of Cr as a function of pH, atmospheric composition r3+ concentration, and surface area of (02 vs. N2 purged), C MnOz. They also measured the Mn released to solution during the redox reaction and were able to gain insight into the reaction mechanism. The conclusions of this report can summarized as follows: 1. The experimental results are consistent with surface Mn4+ as the oxidizing agent for Cr3+ oxidation. 2. The oxidation of Cr3+ by Mn02 in acidic conditions slows down with time, and follows the average rate law: dfCr/dt = k (A/V) [CrT1-1 (1_fCr)3.2(+-0.08) fraction of oxidized Cr to total Cr surface area Of MnO to solution volume total dissolved Cr concentration where: dfc (A/V CrT 3. The rate of oxidation demonstrates a small (and questionable) decrease as the pH is increased from 3.0 to 4.7. 206 4. At higher pH's (6.3, 8.3, 10.1) where Cr(OH)3-solid was allowed to form, the amount of oxidation was greatly decreased, but the amount oxidized exceeded Cr3+ solubility. 5. There was no difference in the oxidation rate in an 02 versus an N2 purged system, suggesting [02] does not influence the reaction at low pH's. 3+ 6. The amount of Cr oxidized is nearly proportional to the surface area of MnOZ present. 7. The rate of oxidation was independent of the initial Cr concentration from 1.9 to 38.5 X 10'5 M. The MnOz dissolution rate (Mn released to solution) however, 3+ decreased with increasing initial Cr concentrations. 8. The ratio of Mn to Cr6+ released was always greater than the ratio of 1.5 predicted by the reaction stoichiometry, due to acid dissolution of the MnOz: and this ratio decreased with increasing Cr3+. In their mechanistic interpretations, the authors pointed to 7) and 8) above to suggest that the desorption of Cr6+ 6+ as the rate limiting step. The adsorption of Cr on this form of MnOZ would be expected to take place under acidic conditions because the PHzpc is about 7.3. This adsorption was interpreted to limit the sites available for oxidation and for acidic dissolution. This argument assumes that all of' the initial Cr3+ concentrations saturate the surface 207 sites where oxidation takes place. Thus the initial rates 3+ are independent of the initial Cr concentration, and the rate as the. experiment progressed is controlled. by the desorption of adsorbed Cr6+. This also suggests that acid dissolution sites are different from Cr oxidation sites. It is certainly apparent that the adsorption of the reaction product (Cr6+) does have an effect on the overall reaction. This suggests that oxidation of Cr by Mn oxides with a low pH may display a quite different reaction rate zpc and rate-controlling step. In acidic solutions, none of the reaction products would be expected to be adsorbed, and the dissolution of the oxide should result in more active sites for continued oxidation. Based on all of ‘the above summarized research the following summary describes some of what is known and also what needs to be determined by further research: 1. The oxidation of Cr3+ by Mn oxides does occur and it is faster than oxidation by 02. 2. The rate of reaction follows a trend of decreasing with time in most of the previous experiments described. 3. The rate depends on the amount of Mn oxide present suggesting the reaction is surface area controlled. 4. Much more work is needed where care is taken to remain below Cr(OH)3 saturation in order to unambiguously describe the reaction of the dissolved species. Further 208 study is also warranted at higher (more realistic) pH's to describe the influence of Mn oxides on Cr solubility. 5. The role of dissolved 02 gas remains equivocal, as two studies have demonstrated conflicting results. More work is clearly needed to describe the reaction mechanism, effects of solution chemistry and the extent this oxidation reaction would be expected to be important in nature. This includes the need for research using low pH ZpC' low crystallinity Mn02 such as those forms commonly found in nature. APPENDIX II APPENDIX II EXPERIMENTAL DATA TABLE OF DATA SHOWN ON FIGURE 3 2/5/85 6/30/85 6/30/85 Fresh Oxide No Vibration Sonic Vibration Time Cr6+ Time Cr5+ Time Cr5+ (min) (mg/1) (min) (mg/1) (min) (mg/1) 1.50 0.19 1.00 0.10 1.00 0.15 4.75 0.33 2.75 0.17 2.75 0.24 9.25 0.40 5.00 0.25 5.10 0.32 15.50 0.42 8.00 0.29 8.00 0.38 25.00 0.46 15.00 0.37 15.25 0.41 45.00 0.49 30.00 0.45 31.00 V 0.48 80.00 0.50 60.00 0.48 60.00 0.49 120.00 0.50 125.00 0.50 135.00 0.50 TMEE c69'0mv1) Mn Mn(fiano 0.080 0.105 0.035 0.030 0.015 0.040 0.155 0.260 0.105 0.130 0.065 0.125 0.220 0.420 0.190 0.305 0.125 0.290 0.320 0.575 0.270 0.455 0.230 0.425 0.380 0.695 0.365 0.610 0.310' 0.580 21f7 TABLE OF DATA SHOWN ON FIGURES 33 AND 34 MATRIX T1me 0.01 M Ca(N03)2 Solution 0.10 M Ca(CO3)2 Solution (min) cr6+ (mg/1) Mn (mg/1) cr6+ (mg/1) Mn (mg/1) 0.000 0.01 0.00 0.05 0.030 0.02 0.02 0.06 0.070 0.09 0.06 0.13 0.134 0.20 0.13 0.25 10.0 0.220 0.36 0.25 0.39 20.0 0.310 0.51 0.33 0.57 TABLE OF DATA SHOWN ON FIGURE 35 MATRIX Time ‘ _ 0.05 M NaCl 0.05 M NaNO3 0.05 M Ca(N03)2 0.05 M CaC12 ‘33") Mn (mg/1) Mn (mg/l) Mn (mg/1) Mn (mg/1) 0 0.000 0.050 0.00 0.015 2 0.170 0.150 0.13 0.150 5 0.365 0.325 0.35 0.295 15 0.610 0.595 0.66 0.540 30 0.730 0.725 0.84 0.655 1218 MOFMSUNGIFIGJRE36AND37 MNHUX 0m5Mmm2$mum omsMaazmmum mstmzwmum Cr“ (mg/1) m (mg/1) a“ (rug/1) Mn (mg/1) Mn (mg/1) 0.045 0.030 0.050 0.04 0.01 0.120 . 0.180 0.125 0.18 0.03 0.265 0.380 0.255 0.36 0.05 0.385 0.540 0.350 0.49 0.09 0.485 0.675 0.435 0.61 0.13 TABLE or DATA snown ON FIGURE 38 MATRIx Time 0.5 M Ca(NO3)2 0.5 M 06012 1.0 M NaCl (min) Mn (mg/1) Mn (mg/1) Mn (mg/1) 0 0.040 0.08 0.01 5 0.375 0.14 0.41 30 0.810 0.30 0.80 60 0.870 0.40 0.85 TABLE or DATA snowu ON FIGURE 39 MATRIX Time (min) Mn (mg/1) Mn (mg/1) Mn (mg/1) 0 0.035 0.090 0.00 2 0.200 0.100 0.17 5 0.390 0.115 0.37 15 0.650 0.155 0.62 30 0.770 0.210 0.75 .219 TABLE OF DATA SHOWN ON FIGURE 40 Mn02 Loading Metal Absorbed (molar) (molar) Cu Mn Zn Ni 4.73 10"5 3.3 x 10"6 4.60 x 10'7 -- -— 4.77 10'5 -— -- 3.80 x 10"7 6.80 x 10"7 9.47 10‘5 5.7 x 10'6 2 73 x 10'6 -- ~- 1.42 10" 7.1 x 10"6 -- -- -- 1.43 10"4 -- -- 2.14 x 10'6 3.07 x 10‘6 1.89 10‘4 7.5 x 10"6 6 46 x 10‘6 -- -- 2.37 10'4 -- -- 4.13 x 10‘6 4.60 x 10‘6 TABLE OF DATA SHOWN ON FIGURES 41, 43 AND 44 Metal Released Time TO Solution (mg/l) (min) Cr6+ Mn Cu 0.50 0.11 0.09 0.06 1.00 0.17 0.19 0.11 2.33 0.35 0.45 0.15 4.50 0.42 0.53 0.17 8.00 0.47 0.56 0.18 15.50 0.47 0.57 0.19 220 TABLE ON DATA SHOWN ON FIGURES 42, 43 AND 44 Metal Released Time TO Solution (mg/1) (min) Cr5+ Mn 0.50 0.13 0.06 1.00 0.23 0.16 2.25 0.37 0.38 4.50 0.42 0.44 7.75 0.46 0.46 16.00 0.46 0.46 TABLE OF DATA SHOWN ON FIGURE 45 Percent Cr Oxidized Time (min) Immediate Base Filtration Added 1 0.14 0.21 3 0.21 0.29 7 0.29 0.33 TABLE OF DATA SHOWN ON FIGURE 46 Time Du licate 1 Du licate 2 (min) Cr + (mg/1) Cr + (mg/l) 2.0 --- 0.120 5.0 0.210 0.200 10.0 0.230 0.225 20.0 0.265 0.265 21.0 0.275 0.270 45.0 -—- --- 221. TABLE OF DATA SHOWN ON FIGURES 47, 48 AND 49 Amount Of Cr Oxidized (mg/1) Time (min) MnOz Loading (M) 1.2 x 10‘5 2.3 x 10'5 4.6 x 10‘5 6.9 x 10‘5 9.2 x 10‘5 1.0 --- --- --- --—“ 0.330 1-5 --- --- --- 0.270 --- 2.0 0.105 0.120 --- --- --- 3-0 “-7 --- --- --- 0.395 5.0 --- 0.200 --- 0.355 --- 7.0 0.155 --- --— --- --- 8-0 7“ --- --- --- 0.450 10.0 --- 0.225 0.290 --- --- 15-0 “*7 --- -—* -—- 0.465 20-0 --- --- -—- 0.380 --- 21.0\\ 0.185 0.265 0.305 --- --- 30-0 --- --- --- --- 0.470 45.0* 0.190 0.270 --- 0.420 0.470 * Values used for Figures 48 and 49 222 TABLE OF DATA SHOWN ON FIGURE 50 Time Cr Total Cr6+ (min) (mg/1) (mg/1) 2.0 0.12 0.120 5.0 0.05 0.200 10.0 0.03 0.225 21.0 0.02 0.265 45.0 0.00 0.270 TABLE OF DATA SHOWN ON FIGURE 51 Time Cr Adsorbed Cr Oxidized (min) (mg/1) (mg/1)A 1.0 0.22 0.070 2.0 0.28 0.130 3.0 0.32 0.180 5.0 0.36 0.245 7.0 0.39 0.300 10.0 0.42 0.360 20.0 0.45‘ 0.425 30.0 0.47 0.470 223 TABLE OF DATA SHOWN ON FIGURES 52 AND 54 Fe(OH)3 First MnOz First MnOZ Last Time (min) Cr6+ (mg/l) Cr6+ (mg/l) Cr5+ (mg/1) 2.0 0.120 0.060 '- 5.0 0.200 0.060 '- 10.0 0.225 0.070 0.04 20.0 --- 0.065 '- 21.0 0.265 -'- '- 30.0 --- --- 0.05 45.0 0.270 0.065 '- 60.0 "' "‘ 0.06 TABLE OF DATA SHOWN ON FIGURES 53 AND 55 Fe(OH)3 First MnOz First MnOZ Last Time (min) Cr6+ (mg/1) Cr5+ (mg/l) Cr6+ (mg/l) 1.0 0.330 --- --- 2.0 --- 0.125 --- 3.0 0.395 --- --- 7.0 --- 0.155 --- 8.0 0.450 --- --- 10.0 --- --- 0.055 15.0 0.465 --- --- 20.0 --- 0.175 --- 30.0 0.470 --- 0.075 45.0 0.470 0.195 --- 60.0 --- --- 0.110 TABLE OF DATA SHOWN ON FIGURE 56 Time (days) cr6+ (mg/1) 0.2 1.0 19.0 43.0 0.095 0.150 0.250 0.250 APPENDIX I I I APPENDIX III RATE DATA CALCULATIONS nwc'non m as “ rum-Ion or 2200071014007 1702 mm: mm mm l,(X 1H) 0x1»: LOADING . ‘X “8'0 0' REACTION 15’IIIIIVTIIIITIT'VIIYI'II E (""!""l""!rth‘ . . . . P F . S » 12 0 ’ {a 9’- g I D H I n 6f' n f. ‘ M 0 3} 0 L 4 L 9 > . . . . A _ . . . R 0’ 11111111111111111111011‘ R o 11111111111111111111 ’ 0 4 8 12 16 z) o 1 2 3 4 TIME (MINUTES) IMPSED 1'"! (MINUTES) Data Set A. Points used to determine initial r regression Of first three data points. Cr = Simple Regression of mnSO on time ate by 1i ear 9. 6 X 10 M. Standard T Prob. Parameter Estimate Error Value Level Intercept 1.133915-7 1.30764E-8 8.67148 0.0730925 Slope 5.71046E—7 5.76167E-9 99.1112 6.42307E-3 Analysis of Variance Source Sum of Squares Df Mean Square F—Ratio Model .0000 1 .0000 9823.0306 Error .0000000 1 .0000000 Total (Corr.) .0000000 2 Correlation Coefficient = 0.999949 Stnd. Error of Est. = 1.03897E-8 Do you want to plot the fitted line? (Y/N): lHELP 2LABEL BSAVSC 4RECORD 5 6 7 PRINT SAT JUN 13 1987 12:11:00 PM VERSION 1.1 Linear Regression MnOz initial = 4.8 224 x 10"6 M 225 m IT'I'UIII‘T'YTIW?‘ W33 ; o 1 2 3 4 the ( Simple Regression of mn75 on time Standard T Prob. Parameter Estimate Error Value Level Intercept -2.13119E-8 .284316-8 -0.339129 0.791853 Slope 1.21404E-6 2.76897E-8 43.8444 0.0145175 Analysis of Variance Source Sum of Squares Df Mean Square F-Ratio Model .0000 1 .0000 1922.3293 Error .0000000 1 .0000000 Total (Corr.) .0000000 2 Correlation Coefficient = 0.99974 Stnd. Error of Est. = 4.99312E-8 Do you want to plot the fitted line? (Y/N): lHELP 2LABEL BSAVSC 4RECORD 5 6 7 B PRINT SAT JUN 13 1987 12:14:00 PM VERSION 1.1 Re ssion 0! 0075 on time (lefifNP ;TfifIITTI1TTTYIUY‘r I 0 n D a o 0 L I i O ‘ h s a o L L : : : 3E ..... 1 ..... 5” H; ..... . fi 0 O U was N p tr. ,f- O O l . O I I I O . O . . O Q . 0 O I I l; . I l o 11 LlllJAJLJLALLLlJl‘ 0 1 2 3 4 the ( Linear Regression Mnoz initial = 7.3 X 10“6 M 226 r‘dr diueber CSL nuance :r r U!“ le ue Level Intercept 1.79952E-7 3.14866-7 0.571529 0.669453 Slope 1.35496-6 1.387335-7 9.76626 0.0649592 Analysis of Variance Source Sum of Squares Df Mean Square F-Ratio Model .000000 1 .000000 95.379837 Error .0000000 1 .0000000 Total (Corr.) .0000000 2 Correlation Coefficient = 0.994799 Stnd. Error of Est. = 2.50168E-7 Do you want to plot the fitted line? (YIN): IHELP 2LABEL SSAVSC 4RECDRD 5 6 7 8 PRINT SAT JUN 13 1987 12:16:00 PM VERSION 1.1 Regression of lh100 on tine (X 11-6 ’IIIV!IUIU!UIII!I II I 2 4L ..... , ..... ',.... . ..... I : a 1 n 31 ........... 5H.H.H.Hl n I 6 b 4 0 2:- ..... \ .......... I ...... I 1 1.. ..H.u.n.u.n.“1 . . . I 0 1141111111111114111‘ 0 1 2 3 4 use Simple Regression of mn125 on time Standard T Prob. Parameter Estimate Error Value Level Intercept 2.697396-7 1.35176E-7 1.99546 0.295746 Slope 1.641296-6 5.95609E-8 27.5564 0.0230923 Analysis of Variance Source Sum of Squares Df Mean Square F-Ratio Model .00000 1 .00000 759.35788 Error .0000000 1 .0000000 ‘Total (Corr.) .0000000 2 IZorrelation Coefficient = 0.999342 Stnd. Error of Est. =- 1.07402E-7 Linear Regression MnOZ initial = 9.7 X 10"6 M 227 \A 1“) ’V'UIIIITUIUUUIIU 7" b e I I 5} ..... i ..... . .......... 1 4} fl , n D 1 33 g t 21 . . . O'IILIIHAIIIILIIIIII‘ 0 1 2 3 4 fine Simple Regression of mn150 on time Standard T Prob. Parameter Estimate Error Value Level Intercept 1.15042E-6 1.43538E-7 8.01474 0.0790227 Slope 1.92674E-6 6.324526-8 30.4645 0.0208896 Analysis of Variance Source Sum of Squares Df Mean Square F-Ratio Model .00000 1 .00000 928.08860 Error .0000000 1 .0000000 Total (Corr.) .0000000 2 Correlation Coefficient = 0.999462 Stnd. Error of Est. = 1.14046E-7 Linear Regression MnOz initial = 1.2 X 10"5 M 1HELP 2LABEL BSAVSC 4RECORD 5 PRINT SAT JUN 13 1987 12:21:00 PM 3 ssion of 00150 on time (xggq 'UUIIIUUIIUIIUI 1' O O I 65 55 I 45 009‘: 3 33 0 1 2 3 4 tn» Linear Regression MnOZ 6 7 8 VERSION 1.1 initial = 1.5 x 10"5 M 228 Model fitting Results estiaate stnd.err0r ratio Coefficient 1 .00000356 .00000009 1000000 Coefficient 2 .00000357 .00000011 1000000 Coefficient 3 .23005280 .01821823 1000000 Total iterations = 9 Total function evaluations t 43 Analysis of Variance for the full Regression source sun of squares df aean square ratio Model .0000 3 .0000 1476.9650 trror .0000000 5 .0000000 Total .0000000 8 Total (corr.) .0000000 7 R-squared 8 0.995835 39 mtual U1I1FIUIIUVIIIUI IIIWIITIjTrTIIF d g 4 q -1 lllllLLLllllLlllLll 0 4 8 12 16 20 time ‘\ Data Set A. Curve Fitting Results, MnO2 = 4.8 x 10"6 M 229 estiaate stnd.err0r ratio Coefficient 1 .00000590 .00000017 35.0040 Coefficient 2 .00000602 .00000023 25.8496 Coefficient 3 .29502163 .02960444 9.9655 Total iterations I 8 Total function evaluations 8 38 Analysis of Variance for the full Regression source sun of squares of aean seuare ratio Model .00000 3 .00000 878.33480 lrror .0000000 5 .0000000 Total .0000000 8 Total (corr.) .0000000 7 R-squared 8 0.992601 (x ”49,161 of Fitted M09911 m d 78 Hintual [ITIUIIITVIIIIUI' L 58 38 LII-5330.33 18 -2 LJlLLllllllJllllLll 0 4 8 12 16 20 time Data Set A. Curve Fitting Reéults, MnO2 = 7.3 X 10’6 M 230 Model fitting Results estiaate stnd.error ratio Coefficient 1 .00000786 .00000012 67.7931 Coefficient 2 .00000794 .00000015 53.9750 Coefficient 3 .24945122 .01227581 20.3206 Total iterations 8 8 Total function evaluations 8 38 Analgsis of Variance for the full Regression source sun of squares df eean square ratio Model .0000 3 .0000 3721.1481 Rrror .0000000 5 .0000000 Total .0000000 8 Total (corr.) .0000000 7 R-squared 8 0.998322 mtual 99 UFU'U‘I'TTI'II‘T 79 ......;....; ..... j LLI 59 Col-‘3 ID 3 39 .. ..... lLlllllLllJl RLJllllllLJlllllllL 048121620 time 5 Data Set A. Curve Fitting Results, MnO2 = 9.7 x 10‘6 M 223]. Model fitting Results . . ._ _. .. --_..—_---——-—_— . estiaate stnd.err0r ratio Coefficient 1 .00000986 .00000022 45.6571 Coefficient 2 .00000989 .00000027 37.1122 Coefficient 3 .23665703 .01707132 13.8628 Total iterations 8 8 Total function evaluations 8 38 Analysis of Variance for the full Regression source sun of squares df mean square ratio Model .0000 3 .0000 1768.0200 lrror .0000000 5 .0000000 Total .0000000 8 Total (corr.) .0000000 7 R-squared 8 0.996491 (X 1E_é’)lot of Fitted ModelFitted 11 ”Uri l I I I I I 1 Fri I I 1 Jtual - l 9 :- --3 L 4 m 7 r '1 a - . n D d 1 5 r . 2 : : 5 3 r 1 1 .... ............. ,13 -1 7 l R l 1 1L] 11 1'11 1 111 Rd 0 4 8 12 16 20 time ’ Data Set A. Curve Fitting Results, MnO2 = 1.2 x 10‘5 M 232 Model fitting Results estiaate stnd.err0r ratio Coefficient 1 .00001253 .00000038 32.9308 Coefficient 2 .00001228 .00000048 25.4835 Coefficient 3 .24857830 .02592314 9.5891 Total iterations 8 7 I Total function evaluations 8 33 Analysis of Variance for the full Regression source sun of squares df aean square ratio Model .00000 - 3 .00000 894.47340 trrcr .0000000 5 .0000000 Total .0000000 8 Total (corr.) .0000000 7 R-squared 8 0.992517 (x ”43161 of Fitted miei'itted 15 ..VHPHIHWWitual <:>(J1F‘*:3 DD 53 O'JllllllLljjlllllLRJ 0 4 8 12 1620 time , I Data Set A. Curve Fitting Results, MnO2 = 1.5 X 10"5 M 2233 Cursor at Rou: 1 Data fditor Maxi-us Roast 8 Column: 1 Hunter of Cols! 5 Rou deriviOO deriv125 deriv150 deriv50 deriv75 1 11.980642-6 2.340542-6 3.052542-6 8.212885-7 1.776038-6 l 2 l1.64269R-6 1.959891-6 2.533351-6 6.911351-7 1.423498-6 I 3 11.202643-6 1.458R-6 1.856732-6 5.184111-7 9.844592-7 I 4 l8.27242!-7 1.022332-6 1.278841-6 3.6712R-7 6.32424l-7 I 5 I5.34621R-7 6.756581-7 8.277362-7 2.454512-7 3.773872-7 I 6 l 2.69232-7 3.52441-7 4.178428-7 1.30381-7 1.676648-7 I 7 I7.73477R-8 1.07942f-7 1.205682-7 4.127221-8 3.835388-8 l 8 11.349272-8 2.059391-8 2.116111-8 8.246741-9 4.863233-9 l 9 l l 10 I I 11 I I 12 I I 13 I I 14 l I length 8 8 8 8 8 Type M N N N M Cursor at low: 1 Data fditor Maxinua Ross: 8 Column: 1 Number of Ccls! 5 Rou ld100 ld125 ld150 ldSO ld75 1 1 -13.1321 -12.9651 -12.6995 -14.0124 -13.2411 I 2 I -13.3192 -13.1426 -12.886 -14.1849 -13.4624 I 3 I 813.631 -13.4384 -13.1967 -14.4725 -13.8312 I 4 I -14.0052 -13.7934 -13.5696 -14.8176 -14.2737 I 5 I -14.4417 ~14.2076 -14.0046 -15.2202 -14.79 I 6 I -15.1277 ~14.8584 -14.6882 -15.8528 -15.6013 l 7 l -16.375 -16.0417 -15.9311 -17.0031 -17.0764 I 8 l -18.1211 -17.6983 -17.6711 -18.6134 -19.1416 I 9 I I 10 I I 11 I I 12 I I 13 I I 14 l I Length 8 8 8 8 8 Type M M M N N Data Set A. 1st derivatives and In Of lst derivatives for fitted curves. Row 1 5 Time 0.00 234 Simple Regression of rate on loading Standard ' T Prob. Parameter Estimate Error Value Level Intercept -1.58573 1.93216 -0.820707 0.471954 Slope 1.03465 0.166327 6.22058 8.374918-3 Analysis of Variance Source Sum of Squares Df Mean Square F-Ratio Model .824925 1 .824925 38.695639 Error .0639549 3 .0213183 Total (Corr.) .8888800 4 Correlation Coefficient = 0.963354 Stnd. Error of Est. = 0.146008 Do you want to plot the fitted line? (YIN): 1HELP 2LABEL 3SAVSC 4RECORD 5 6 7 8 PRINT SAT JUN 13 1987 12:54:00 PM VERSION 1.1 Regression of rate on loading -13.° ’TI'UIUVUYIU'I'IIIIVIVVU I vvvv' ' *ALAA-A r -1308 L A AAAA 01*. -14. 2 ’ oooooooooooooooooooooooooo l. r i : : : : -14,5E....'...s....2....$... L t o ALAJeJAAJ A -15.° jlliLLllillllillllimL -13.0-12.6-12.2-11.8-11.4-11.0 lmflflw Data Set A. Reaction order with respect to Mno by regression of initial rate from 810 ’ ' - On Mnoz Loading ( pe Of lnltlal paints) 235 Regression Analysis - Linear aodel: Y 8 aibx Dependent variable: irate Independent variable: [18173.loading Standard 7 Prob. Paraaeter Istiaate trror Value Level Intercept -0.612672 1.91661 -0.319665 0.77019 Slope 1.08487 0.164988 6.57544 7.155953-3 Analysis of Variance Source Sun of Squares DI Hean Square I-Ratio Prob. Level Model .906950 1 .906950 43.236450 .00716 trror .0629296 3 .0209765 Total (Corr.) .9698800 4 Correlation Coefficient - 0.967014 R-squared a 93.51 percent Stnd. Irror oi lst. 8 0.144833 ’Regression of lnitial reduction Rate on initial 1102 Loading -12e6 e e. I... .l lllllllllmlml' -12.6 -11.8 -11 -13 -12.2 -11.4 "102 Loading (log no: es/liter) Issac :o-wocpnu —--«-:~ no I p . C to Data Set A. Reaction order with respect to Mno by regression of initial rate (from curve fitting? on Mnoz Loading 236 Hodel fitting Results estiaate stnd.error ratio Coefficient 1 .00000336 .00000019 17.4739 Coefficient 2 .00000324 .00000019 17.1935 Coefficient 3 .29250717 .05015567 5.8320 Total iterations 8 8 Total function evaluations 8 37 Analysis of variance for the full Regression source sun of squares df aean square ratio Hodel .00000 3 .00000 444.09281 trror .0000000 3 .0000000 Total .0000000 6 Total (corr.) .0000000 5 R-squared 8 0.993694 Plot of fitted Model (X 1l-6) "0284.731-611 .. pm“ W l 33 noc—onn-uo 0 3 4 6 8 10 time (minutes) Data Set B. Curve Fitting Results, Mnoz = 4.7 X 10"6 M 237 Model fitting Results . estiaate stnd.error ratio Coefficient 1 .00000690 .00000043 16.0566 Coefficient 2 .00000677 .00000041 16.3469 Coefficient 3 .27170512 .04773283 5.6922 Total iterations 8 8 Total function evaluations 8 37 Analusis of Variance for the full Regression source sun of squares df aean square ratio Model .00000 3 .00000 424.28288 lrror .0000000 3 .0000000 Total .0000000 6 Total (corr.) .0000000 5 R-squared 8 0.993808 Plot of Fitted Model (X “-6) H102 8 8565-6 ILFitte 9 WW ua 33 soc—ovum": o ALLIAAALAAAIAALIAJAV 0 2 4 6 8 10 tile (Innutes) Data Set B. Curve Fitting Results, Mnoz = 9.6 X 10"6 M ‘\ 238 Model fitting Results estiaate stnd.error ratio Coefficient 1 .00001390 .00000044 31.8949 Coefficient 2 .00001372 .00000042 32.6356 Coefficient 3 .27411479 .02420291 11.3257 Total iterations 8 8 Total function evaluations 8 37 Analysis of Variance for the full Regression source sun of squares df mean square ratio Model .0000 3 .0000 1637.5181 Rrror .0000000 3 .0000000 Total .0000000 6 Total (corr.) .0000000 5 R-squared 8 0.998415 Plot of Fitted Model (x 11-6) moz . mstee amt... 25 'W‘Iual D b i 12 D aaaaaaaaaaaaaaa ad ; : : o 9 p ................... .1 i b - 4 v n ' d e 6 p ........... I ......... d b ' 8 b J M 3 _ ................... .. I'I . . o JAAlAAjiAAAlAAAlAJ‘ 0 2 I 6 8 10 time (minutes) Data Set 8. Curve Fitting Results, Mno2 = 1.9 x 10‘5 M 239 Model Fitting Results estiaate Coefficient 1 .00003949 Coefficient 2 .00003888 Coefficient 3 .20213162 stnd.error ratio .00000166 23.7602 .00000155 25.0105 .02007746 10.0676 Total iterations 8 8 Total function evaluations 8 37 Analysis of Variance for the full Regression source sun of squares df aean square ratio Model .0000 3 .0000 1712.9383 Error .0000000 3 .0000000 Total .0000000 6 Total (corr.) .0000000 5 R-squared 8 0.998563 Plot of Fitted Ibdel (Xibm noesfing-rnu? ‘ ' l 3 2 p I S i V Q d M "1 I / 1 0 0 2 4 6 8 time (minutes) 10 Data Set B. Curve Fitting Results, Mnoz 5.0 x 10’5 M 240 Cursor at Roe: 1 Data Lditor Maxiaua Rows: Coluan: 1 Number of Cols: Rou ld51 ld52 ld53 ld54 1 I -13.8692 -13.206 -12.4909 -11.7539 I 2 I -13.957 ~13.2876 812.5731 -11.8145 I 3 I -14.1617 -13.4778 -12.765 -11.956 I 4 I -14.7467 -14.0212 -13.3132 -12.3603 I 5 I -15.6242 -14.8363 -14.1356 -12.9667 I 6 I -16.7943 -15.9231 -15.232 -13.7752 I 7 I I 8 I I 9 I I 10 l l 11 I I 12 I I 13 I I 14 I I Length 6 6 6 6 0 0 Type M N M N M M Data Set B. In of lst derivatives for fitted curves. Row 1 = Time 0.00 241 Regression Analysis 8 Linear aodel: Y 8 aibx Dependent variable: inrate5 independent variable: lload Standard T Prob. Paraneter Rstinate Irror Value Level intercept 82.81136 0.413184 86.80413 0.0209245 Slope 0.899377 0.0369774 24.3224 1.68612883 Analysis of Variance Source Sun of Squares Df Mean Square 78Ratio Prob. Level Model 2.48589 1 2.48589 591.57790 .00169 lrror .0084043 2 .0042021 Total (Corr.) 2.4942927 3 Correlation Coefficient 8 0.998314 R-squared = 99.66 percent Stnd. Rrror of Rst. 8 0.0648239 Regression of initial rate on initial M102 Loading -1106 vvvr‘lvrrfrTvvv.‘ e0 -12 :. ..... ..... ....' ..‘zl'... . . {'7' 812.4 “12.8 432 fi . . -13.5 -.zfiii.§{1.; ..... L....1 7" -14 an.;lL...1....1iJ..‘ 813 812 811 810 89 Log initial M102 (Cr85m/l) 0 09 '8 —3 ”'W'“: 9" In 0 I" Data Set B. Reaction order with respect to Mno regression of initial rate (from curve fitting r“ on Mnoz Loading 2412 Model fitting Results estinate stnd.error ratio Coefficient 1 .00000327 .00000043 7.57298 Coefficient 2 .00000311 .00000040 77.69695 Coefficient 3 .21230551 .07042296 3.01472 Total iterations 8 9 Total function evaluations 8 41 Analysis of Variance for the full Regression source sun of squares ' df mean square ratio Model .00000 3 .00000 161.66454 frror .0000000 3 .0000000 Total .0000000 6 Total (corr.) .0000000 5 R-squared 8 0.983243 Plot of fitted Model (X “-5) ”284.7318“ .. fitte? Di 33 BO I J D . 4 D : 4 L ...................... ‘ . 1 1 D 8 4 > ‘ I y I b ........................ I I D e a 4 >llllllllllllllllllllllll +I‘1I!UIIY11'YU!IIVIIVUTTJ r4 - . , e i z 4 he. mmmmmmmmmmmmmm . oooooooooo q D . 4 D 8 4 D ‘ 4 . ; 4 h OI.I m . . 1 D o 4 D ' I L ' 3 ‘ eeeeeeeeeeeeee .uueaoaeoeq D . 4 p a L ............. .- ......... .1 > I i D 8 4 D 4 D I ........................... D 4 D D e e o e 1 D LlllllllllllllllllLlllll 0 2 4 6 810 Elapsed time (minutes) O 2 4 6 810 llapsed Time (minutes) to determine Kex 250 Reaction Kinetics W21 8 9.6186 Cri 8 1.9184 l '0a01 II ( M 80.11 i -o.21 n 80.41 n-o.31j 1'11111111111111111'1111 D o o e e i . I m e 4 D Z 4 b eeeeeeeeeeeeeeeeeeeeeeeeeee , ° 4 D . e 3 D ° 4 D 1 4 he no I r , . D O 4 D a ' I b eeeeeeeeeeeeeeeeeeeeeeeee d z 4 i ; : i D o P ooooooooooooooo a oooooooooo # D e 4 D ' I I I O I ‘ 4 llLLlJlJllllllllllllle vo-ewoa 3\ 80. 51 0246810 [lapsed Time (minutes) Reaction Ki netios M1021 8 1.9185 Cri 8 1.9184 0.02 88-8-1003 3\: 3 I were: 388: ~— e e e a 8: on 8 u- a: co co .6 s VIIYTT'VY'I'IIIIIVU'UIIV 4 I I C A O J 4 4 4 4 b eeeaeeeee eeeaeaaad D a a e m ’ e e e e 1 o a I o 4 4 D O O O O ’lLllllllllllLllllLlllIll 0 2 4 6 8 10 Elapsed Time (minutes) Data Set C. VMN033\33IMNOSIA3fl Reaction Ki neti cs _0 H1021 8 4.7186 Cri 81.9184 LVIITTIUI'IUVVVTIIVVIIT ‘ D e a e a . e e e a D D 4 D 4 ............................. D 4 D 4 , 4 D I 4 b a o a m m q D 4 D 4 D . 4 D I be 1 D 4 D 4 D 4 D 4 D - 4 ’ i D 4 D . O I I D lllllllllllllllllllllll 0 2 4 6 8 10 flapsed Time (minutes) Re ti Kinetics M102i 8.31.0133 Cri . 1.9284 0.13 . :8 .6 -b N 0 ~41 «a «I VVVVVTV‘V" v' 0 O vwwo: 3\: 3 I w-No: 3A: -- as q 55 s ’IUUUlUI'I'UI'IIIIUF'IUTI‘ O O O O 4 4 4 D e a a ’ a a e o ’emee: eeeee I eeeeeeeeeee ...-6* o AAJ‘AAAA A A m AAIAA O I D I blelllllllllIJlLJljllllj 0 2 4 6 8 10 flapsed Time (minutes) Plots used to determine Kex 2551. Sinaie Regression of nn45 on time Standard T Preb. Paraeeter Estimate Error Value Level Intercept 80.0252762 0.0200301 82.26191 0.334213 Slope 80.140773 6.01373E83 823.4087 1.81995E83 Anaivsis of Variance Source Sum of Squares Di flean Sousre F-Ratio Hodei .22418 1 .22418 547.96524 Error . 0008162 2 . 3004091 Tate} 'Ccrr.) .2250000 3 t 8 80.99818 0.6302266 Correi t‘on Coefficien a . F F I '- :tnd. :rror 5? ast. Do you want to oiot the fitted line? iY/N): ifiELP ZLRBEL SSQVSC 43ECJRD 5 6 7 3 FRifiT SUN 59R 2? 3993 03:21:00 F5 VERSISN 1.1 Regression of mn48 en time z’i 1‘1: v0‘--"-.ro-I ..... leullliiriril' ..... 1 i f ' ' J .K J a ' : i~. ' 1 . "4 ' d .91 ‘=..1 ....... 4' 'Jl‘h' \ ’ ' r 1‘\. Z ‘ ‘ o ' \‘ . J x: ' ’11 I \v - 1 P \ . . 4 not“: ....... - .. :1 V .....r .\. . e . ' 2 ° 2 ‘ .3 l \\. . . 1 r 5x . : 1 . '\ . . 4 \ CC . . . d"".“lr ......... .' %\‘L.-te.on.q e ‘3 ' ' o \ e o ‘ ' Q o o L .1 , 4‘ 1 4‘ cqu._-¥- A 1111‘ '111411144‘11: uuv '— 0 1 a 3 4 5 6 time 9REVIE¥ Data Set A. Reggession to Determine Kex " I! Mno2 = 4.8 x 10 f 23552 Sinole Regression of on73 on time tandard T Prob. Parameter Estimate Error Value Level Intercept 80.0133425 0.0862069 -0.154773 0.391203 Slooe 80.195359 0.0258323 87.54798 0.0171035 Anaivsis of Variance Source Son of Squares 01 Mean Souare F-Ratio Model .431744 1 .431744 56.971963 Error .0151564 2 .0075782 Total (Corr.) .4469000 3 Correlation Coefficient = 80.932397 Stnd. Error of Est. = 0.0870527 Do .34 haet to o.ot :19 ‘.tte: .ine? (YiN): 1rELP ZLABEL SSAVSC 4RECORD 5 6 7 9 9REVlEfi 100017 PRINT SUN EAR 27 1985 03:23:00 ?5 VERSION 1.1 REC:0FF Regression of nn73 on time t .5. (\2 .. VIV fwifit‘tu ...—n! .EHH; t‘\.: . . j .013...\: ........ '. ........... a; 1 \\ : . i - .33; ..... ‘:\;.,.; ........... 3 m 1.4 L ‘ Q ‘ f\\ : 1 .1 ' \1 1 5 .r.33. ....... 3..\o. . ....... 4 r \ 1 D 4. «8.73% ............. EX; ....... .‘J‘ , N t 3\ . i -_-‘"93{ulle-v11§1'1'11'1111'1u5LLJ 3 1 2 3 4 5 6 Data Set A. Reggession to Determine Kex Mno2 = 7.3 x 10' M 12553 Siooie Regression of nn97 on tine Standard 7 Prob. Parameter Estiaate Error ’v’aiue Level interceot 80.024116 0.0360788 80.663026 0.572679 Siooe 80.165525 0.0108321 815.2809 4.25521E-3 Analysis of Variance Source Sun of Souares 0f nean Souare F-Ratio model .30995 1 .30995 233.50718 Error .0026547 2 . 0013273 Total (Corr.) .3126000 3 Correlation Coefficient = 80.995745 Stnd. Error of Est. 8 0.0364328 Do you want to plot the fitted line? (Yfiii (D 4) 31.? "I <1 e». r" 3:. lfiEL? 2LABEL SSAVSC 4RECGRD 5 a 7 PRINT 50% NAR 27 1985 03:25:00 PH VERSIGN 1.1 Regression of mn97 on time :Ve;ntfilile|;-‘iI;IWo'iluii ‘g. . . . . \' . . . . --.u::) 3:? Data Set A. Reggession to Determine Kex MnOZ = 9.7 X 10- M r 225i4 Siaoie Reoress:on of on123 on time Standard 1 Prob. Fara-eter Estimate Error Value Level interceot 80.019558 0.019665 80.99456 0.424749 Elooe 80.162762 5.904115-3 827.5676 1.31324E83 Analysis of Variance Source Sun of Squares Oi fiean Souare F-Ratio nodal .29969 1 .29969 759.97513 Error .0007887 2 .0003943 lotal (Corr. ) .3004750 3 Correlation Coefficient = 80.998687 Stnd. Error of Est. = 0.0198579 00 you want to olot the fitted line? (YIN): lfiELP BLABEL 350988 4RECORD 5 6 7 8 9REVIEH 10001? PRENT SUN 80R 27 1988 03:27:00 PH VERSIGN 1.1 REC:0FF Regression of mn123 on time 80.06, . .I;-.I...:ll¥f|t. T.-.illir '\. 1 t 1 1: Data Set A. Reggession to Determine Kex Mno2 = 1.2 x 10‘ M / I 255 Siaoie Reoressmn of moi-37 or. time Standard 7 Prob . Paraaeter Estioate Error Value Level intercept 80.0785083 0.0288175 82.721133 0.112459 Slooe 80.171823 8. 65201E-3 819. 8593 2. 5259493 Analysis of Variance Source Sun of Souares 0f fiean Square F-Ratio Nodel .33398 1 .33398 394.39325 Error .0016936 2 .0008-168 's'otal (Corr. 1 . mm (A Correlation Coefficient = 80.997474 Stnd. Error of Est. = 0.0291002 00 you want to olot the fitted line? (YIN): ma? zoo 3mm mamas a 7 s wwmwmmn PRINT sou man 27 1993 03:29:00 en veaszcx 1.: REC:0FF Regression of moi-47 on time fleié‘ 80. 35 810FF‘3 5 In 0‘ Data Set A. Reggession to Determine Kex Mno2 = 1.5 x 10‘ M 256 Simple Regression of mnSO on time Standard T Prob. Parameter Estimate Error Value Level Intercept 80.078455 0.0372117 82.10834 0.281947 Slope 80.149915 0.0202906 —7.38841 0.0856442 Analysis of Variance _-—-—-_-—~————~-~.—— ...-..._ Source Sum of Squares Df Mean Square F-Ratio Model .088250 1 .088250 54.588602 Error .0016166 1 .0016166 Total (Corr.) .0898667 2 Correlation Coefficient = 80.990965 Stnd. Error of Est. = 0.0402074 Do you want to plot the fitted line? (Y/N): 1HELP ZLABEL SSAVSC 4RECORD 5 6 7 8 PRINT FRI SEP 4 1987 09:50:00 PM VERSION 1.1 Regression of mn50 on time ' ’TVfiIYYYYFIIYIYYVIIIIIIIIUIY . e . . . . I I C O I AA..- 4J2} m .0022 n 880.32 80A2 4.52 uuiiauiaaaaiauiinaaaiaa .0 .5 LOINSZJ)ZJS&O the Data Set B. Reggession to Determine Kex Mno2 = 4.7 x 10‘ M 257 Simple Regression of mn100 on time Standard T Prob. Parameter Estimate Error Value Level Intercept 80.0317827 5.9323E-3 85.35757 0.117475 Slope 80.173175 3.23473E83 853.5361 0.01189 Analysis of Variance Source Sum of Squares Df Mean Square F-Ratio Model .1178 1 .1178 2866.1157 Error .0000411 1 .0000411 Total (Corr.) .1178000 2 Correlation Coefficient = 80.999826 Stnd. Error of Est. = 6.40988E-3 Do you want to plot the fitted line? (Y/N): 1HELP ELABEL SSAVSC 4RECORD 5 6 7 8 PRINT FRI SEP 4 1987 09:52:00 PM Regression of mn100 on time -mos 445’ m-azi OOH: .0 .5 L01n524924510 the VERSION 1.1 Data Set B. Reggession to Determine Kex M Mno2 = 9.6 x 10 258 Simple Regression of mn200 on time Standard T . Prob. Parameter Estimate Error Value Level Intercept 80.0241681 7.81985E-3 83.09061 0.199217 Slope 80.169185 4.26396E83 839.6779 0.0160413 Source Sum of Squares Df Mean Square F8Ratio Model .1124 1 .1124 157 .335 Error .0000714 1 .0000714 Total (Corr.) .1124667 2 Correlation Coefficient = 80.999683 Stnd. Error of Est. = 8.44939E-3 Do you want to plot the fitted line? (Y/N): 1HELP 2LABEL SSAVSC 4RECORD 5 6 7 8 PRINT FRI SEP 4 1987 09:57:00 PM VERSION 1.1 Regression of ngOO on time .0003’TYTIYIIIIIIIIIITUUIIITIIFIIT i = s 2 s s b . . . . . ‘ i 4 m-mzal ....................... J n ’ ‘ 2 I + 0 L < o -0.33:. ....... s .......... (...-1 I I -0A31 ........ . .............. . r 1 I 2 : : : : 1 4.53 ’llllLLlLlllelLlllllllljllll‘ .0 .5 L01n52J>LSELO the Data Set B. Reggession to Determine Kex Mno2 = 1.9 x 10‘ M 259 Simple Regression of mn500 on time Standard T Prob. Parameter Estimate Error Value Level Intercept 80.0310357 6.4716E-3 84.79567 0.130874 Slope 80.143463 3.5288E83 840.6551 0.0156559 Source Sum of Squares Df Mean Square F-Ratio Model .0808 1 .0808 1652.8356 Error .0000489 1 .0000489 Total (Corr.) .0808667 2 Correlation Coefficient = 80.999698 Stnd. Error of Est. = 6.9926E-3 Do you want to plot the fitted line? (Y/N): 1HELP ZLABEL SSAVSC 4RECORD 5 6 7 8 PRINT FRI SEP 4 1987 10:00:00 PM VERSION 1.1 Regression of nnSOO on time 0'06 T IIF‘IUI‘IYYTIYUTY'UUIIIII‘I I Q I O C O I I O .o o Loixszozsao the Data Set B. Reggession to Determine Rex Mno2 = 5.0 x 10‘ M 260 aluminum: U I I l -_°u. Parameter Estimate Error - Value Level Intercept 80.0583616 2.42685E83 824.0483 0.0264573 Slope 80.133701 1.3233E83 8101.036 6.3007E83 Source Sum of Squares Df Mean Square F-Ratio Model .070 1 .070 10208.333 Error .0000069 1 .0000069 Total (Corr.) .0702000 2 Correlation Coefficient = 3 Stnd. Error of Est. = 2.62222E-3 Do you want to plot the fitted line? (YIN): 1HELP ZLABEL SSAVSC 4RECORD 5 6 7 8 PRINT FRI SEP 4 1987 09:39:00 PM VERSION 1.1 Regression of mn4? on time ‘0'06 n Yflt VIII IIUI 11“ VIII . ! ! ! ! ! 1 . . . . . L i b -0A6» .................... . r 4 m . I g.m25h ..................... . i p .ogml ...................... . 4.46 11111.11L1iniiiiiiimii“jinn;W .0 .5 L01u524)2J510 the Data Set C. Reggession to Determine Kex Mno2 = 4.7 x 10‘ M 261 Simple Regression of mn96 on time Standard T ' Prob. Parameter Estimate Error Value Level Intercept 80.0592275 0.0186058 83.18327 0.193774 Slope 80.0749576 0.0101453 87.38841 0.0856442 Analysis of Variance Source Sum of Squares Df Mean Square F8Ratio Model .022063 1 .022063 54.588602 Error .0004042 1 .0004042 Total (Corr.) .0224667 2 Correlation Coefficient = 80.990965 Stnd. Error of Est. = 0.0201037 Do you want to plot the fitted line? (YIN): 1HELP ZLABEL SSAVS PRINT 8&04 .Qog} ;.H3H.U.K.H3H -0A2’ .0.24; 8&28 .0 FRI SEP 4 19 C 4RECORD 5 6 7 8 87 09:42:00 PM VERSION 1.1 Regression of mn96 on time ’IIUTIIUIIITVITITYIT‘IIU‘IYVII . . o . o 4 d 4 ... I 4 J 4 ......................... .......... ..... ooooo ooooooooooooooooooooooo ..... O O O I I lllllllllhlllllllllllllllll fine .5 1J)1£524)2Ji&0 Data Set C. Reggession to Determine Kex Mno2 = 9.6 x 10' M 262 Simple Regression of mn195 on time Standard T Prob. Parameter Estimate Error Value Level Intercept 80.0429626 0.0159093 82.70047 0.225777 Slope 80.123514 8.67496E83 814.238 0.0446393 ~-_-~”_-~~H”Hmu__fl~~—-~-—_n_”-~-*—-“**nm~“*—m.“—n—~-~~—_~-~‘h-n—--"H-—”_ Source Sum of Squares Df Mean Square F-Ratio Model .05990 1 .05990 202.72192 Error .0002955 1 .0002955 Total (Corr.) .0602000 2 Correlation Coefficient = 80.997543 Stnd. Error of Est. = 0.0171901 Do you want to plot the fitted line? (Y/N): 1HELP ZLABEL SSAVSC 4RECORD 5 6 7 8 PRINT FRI SEP 4 1987 09:44:00 PM VERSION 1.1 Regression of mn195 on time .IO'OI TITI IIII IIII TIII IIII IIII . 1 ! ! ! 1 .041. ....... é ............... 1 m E i j n - . 14.211 ....... g ................ . 9 . 5 > .¢31. ...................... . .o.‘1‘llllillljillllillllilllIi111 .0 .5 LOIHSZJ)&51LO the Data Set C. Reggession to Determine K Mno2 = 1.9 x 10 M ex 263 Simple Regression of mn505 on time Standard T Parameter Estimate Error Value Intercept 80.0202886 7.0109E-3 82.89387 Slope 80.113752 3.82286E83 829.7557 n—N_“-_-—--~___———-_—_------—-—_--—-----_-—-—_---u-—~——. —m——-————-—-——_—-——-——-—-———--————--—“-.——-_-————-—_-—--_— Source Sum of Squares Df Mean Square Model .05081 1 .05081 Er r or . 0000574 1 . 0000574 Total (Corr.) .0508667 2 Correlation Coefficient = 80.999436 Stnd. Error of Est. = 7.57532E-3 Do you want to plot the fitted line? (Y/N): 1HELP ELABEL SSAVSC 4RECORD 5 6 7 PRINT FRI SEP 4 1987 09:47:00 PM VERSION 1.1 Regression of mn505 on time 0'04 LTIITIHIIIHIIIIFIIIIITVIIIII 1 P I -0fi6. ...................... i 4 m 1 . n . ‘ 5 —0.16 .. ....................... . 0 r ‘ S ‘ ‘ p 4 b 1 .mge. ...................... . h 4 1 . . . . . 4 .0. 36 ‘llllllllllLillllllllljlllllj .0 .S L01n524>2J510 the Data Set C. Reggession to Determine Rex Mno2 = 5.0 x 10' M Prob. Level 0.211811 0.0213868 F—Ratio 885.40434 BIBLIOGRAPHY BIBLIOGRAPHY Anderson, M.A., Jenne, E.A. and Chao, T.T. (1973) The sorption of silver by poorly crystallized manganese oxides. Geochim. Cosmochim. Acta, 37, 611-622. Anderson, M.A., Bauer, C., Hansmann, D., Loux, N. and Stanforth, R. (1981) Expectations and limitations for aqueous adsorption chemistry. In: Adsorption of inorganics at liquid:§glid interfaceg (Anderson, M.A. and Rubin, A.J., eds.) Ann arbor Science, 327-347. Aoki, T. and Munemori, M. (1982) Recovery of chromium VI from waste waters with iron III hydroxide. Wat. Res., 16, 7938796. Balistrieri, L.S. and Murray, J.W. (1982) The surface chemistry of delta-Mno in major ion seawater. Geochim. Cosmochim. Ac a, 46, 1041-1052. Balistrieri, L.S. and Murray, J.W. (1982) The adsorption of Cu, Pb, Zn and Cd on geothite from major ion sea water. Geochim. Cosmochim. Acta, 46, 1253-1265. Balistrieri, L.S. and Murray, J.W. (1981) The surface chemistry of geothite in major ion seawater. Am. J. Sci., 281, 7888806. Bartlett, R. and James, B. (1979) Behavior of chromium in soils: III. Oxidation. J. Env. Qual., 8, 31835. Beliles, R.P. (1979) The lesser metals. In: The Toxicitv of Heavy Metals in the EnvironmentLAPt. 2 (F.W. Oehme, ed.) Marcel Dekker, Inc. NY Benes, P. and Steinnes, E. (1975) Migration forms of trace elements in natural fresh waters and the effect of water storage. Wat. Res., 9, 741-749. Benjamin, M.M. and Leckie, J.0. (1981a) Multi-site adsorption of Cd, Cu, Zn and Pb on amorphous iron oxyhydroxide. J. Colloid Interface Sci., 79, 209-221. 264 265 Benjamin, M.M. and Leckie, J.G. (1981b) Competitive adsorption of Cd, Cu, Zn and Pb on amorphous iron hydroxide. J. Colloid Interface Sci., 83, 410-419. Benjamin, M.M. and Leckie, J.O. (1981c) Conceptual model of metal-ligand surface interactions during adsorption. Env. Sci. Tech., 15, 1050-1057. Benjamin, M.M. (1983) Adsorption and surface precipitation of metals on amorphous iron oxyhydroxide. Env. Sci. Tech., 17, 6868692. Bricker, O.P. (1965) Some stability relations in the system Mn-OZ—H 0 at 25° and one atmosphere total pressure. Amer. M1n., 50, 1296-1354. Burns, R.G and Burns V.M. (1979) Manganese oxides. In: Marine Minegals (R.G. Burns, ed.) Min. Soc. Amer. Short Course Notes Vol. 6, 1846. Campbell, J.A. and Yeats, P.A. (1981) Dissolved chromium in the northwest Atlantic Ocean. Earth and Plan. Sci. Letters, 53, 4278433. Carpenter, R.E. and Hayes, W.B. (1980) Annual accretion of Fe-Mn oxides and certain associated metals in a stream environment. Chem. Geol., 29, 2498259. Chao, T.T and Theobald, P.K. (1976) The significance of secondary iron and manganese oxides in geochemical exploration. Econ. Geol. 71, 1560-1569. Cranston, R.E. and Murray, J.W. (1978) The determination of chromium species in natural waters. Anal. Chim. Acta, 49, 275-282. Cranston, R.E. and Murray, J.W. (1980) Chromium species in the Columbia River and estuary. Limnol. Oceanogr. 25, 1104-1112. Crerar, D.A., Fischer, A.G., and Plaza, C.L. (1980) Geology and geochemistgy of manganese (Varentsov, I.M. and Grassely, G., eds.) Vol. I. E Schweizebart'sche Verlagsbuchhandlung, Stutgart., 898334. Crowther, D.L., Dillard, J.G. and Murray, J.W. (1983) The mechanism of Co(II) oxidation on synthetic birnessite. Geochim. Cosmochim. Acta, 47, 1399-1403. 266 Davis, J.A. and Leckie, J.G. (1978) Surface ionization and complexation at the oxide/water interface. II. Surface properties of amorphous iron oxyhydroxide and adsorption of metal ions. J. Colloid Interface Sci., 67, 908107. Davis, J.A. and Leckie, J.G. (1978b) Effect of adsorbed complexing ligands on trace metal uptake by hydrous oxides. Env. Sci. Tech., 12, 1309-1315. Davis, J.A. and Leckie, J.G. (1980) Surface ionization and complexation at the oxide/water interface. 3. adsorption of anions. J. Colloid Interface Sci., 74, 32843. Davies-Colley, R.J., Nelson, P.0. and Williamson, K.J. (1984) Copper and cadmium uptake by estuarine sedimentary phases. Env. Sci. Tech. 18, 4918499. Dempsey, B.A. and Singer, P.C. (1980) Effects of calcium on the adsorption of zinc by MnO (3) and Fe(0H) (am). In: Contaminants and Sediments, ( .A. Baker, ed. Ann Arbor Pub., 333-355. Douglas, B., McDaniel, D.H. and Alexander, J.J. (1983) Concepts and Methods of Inorganic Chemistgy 2nd ed. John Wiley 8 Sons, New York. Drever, J.I. (1982) The Geochgmistry of Natural Waters. Prentice-Hall, Inc., New Jersey, 388 p. Elderfield, H. (1970) Chromium speciation in sea water. Earth and Plan. Sci. Letters, 9, 10816. Emerson, S., Cranston, R.E. and Liss, P.S. (1979) Redox species in a reducing fjord: equilibrium and kinetic considerations. Deep-Sea Research, 26A, 859-878. Forbes, E.A., Posner, A.M., and Quirk, J.P. (1976) The specific adsorption of divalent Cd, Co, Cu, Pb and Zn on geothite. J. Soil Sci., 27, 154-166. Fukai, R. (1967) Valency state of chromium in seawater. Nature, 213, p. 901. Gadde, R.R. and Laitinen, B.A. (1974) Studies of heavy metal adsorption by hydrous iron and manganese oxides. Anal. Chem., 46, 202282026. Gadde, R.R. and Laitinen, B.A. (1973) Study of sorption of lead by hydrous ferric oxide. Env. Letters, 5, 223-235. 267 Gardiner, W.C. (1972) at Mechanis s of hemica Reactions W.A. Benjamin, Inc., Menlo Park, CA, 284 p. Garrels, R.M., McKenzie, F.T. and Hunt, C. (1974) Chemical c s and he oba env 0 me ' assessin human influgnces, William Kaufmann, Inc. Los Altos, CA. 206 p. Gephart, C.J. (1982) The relative importance of iron oxides, manganese oxides and organic material on the adsorption of chromium in natural water sediment systems. Unpublished Master's Thesis, Michigan State University. Griffin, R.A., Au, A.K. and Frost, R.R. (1977) Effect of pH on the adsorption of chromium from land-fill leachate by clay minerals. J. Env. Sci. Health, A12, 431-449. Healy, T.W., Herring, A.P., and Fuerstenau, D.W. (1966) The effect of crystal structure on the surface properties of a series of manganese dioxides. J. Colloid Interface Sci., 21, 435-444. Hem, J.D. (1978) Redox processes at surfaces of manganese oxide and their effects on aqueous metal ions. Chem. Geol. 21, 1998218. Holdren, G.R. and Berner, R.A. (1979) Mechanism of feldspar weathering --I. Experimental studies. Geochim. Cosmochim. Acta, 43, 1161-1171. James, B.R. and Bartlett, R.J. (1983) Behavior of chromium in soils. VI. Interactions between oxidation-reduction and organic complexation. J. Env. Qual. 12, 173-176. James, R.G. (1981) Surface ionization and complexation at the colloid/aqueous electrolyte interface. In: Adsogptiog of inorgagics at the solid-liggid integface. (Anderson, M.A. and Rubin, A.J., eds.) Ann Arbor Science, 219-262. James, R.G. and Healy, T.W. (1972) Adsorption of hydrolyzable metal ions at the oxide-water interface. J. Colloid Interface Sci., 40, 42-81. Jenne, B.A. (1968) Controls on Mn, Fe, Co, Ni, Cu, and Zn concentrations in soils and water: the significant role of hydrous Fe and Mn oxides. Am. Chem. Soc. Adv. in Chem. Series 73, 337-383. Kinniburgh, D. G. and Jackson, M. L. (1981) Cation adsorption by hydrous metal oxides and clay. In: Adsorption of ingrganic cg at the solid- -liggid interface (Anderson, M.A. and Rubin, A. J., eds.) Ann Arbor Science, 91- 160. 268 Kinniburgh, D.G., Jackson, M.L. and Syers, J.K. (1976) Adsorption of alkaline earth, transition and heavy metal cations by hydrous oxide gels of iron and aluminum. Soil Sci. Soc. Am. J., 40, 796-799. Kinniburgh, D.G., Syers, J.K. and Jackson, M.L. (1975) Specific adsorption of trace amounts of calcium and strontium by hydrous oxides of iron and aluminum. Soil Sci. Soc. Am. Proc., 39, 464-470. Krauskopf, K.B. (1956) Factors controlling the concentrations of thirteen rare metals in sea-water. Geochim. Cosmochim. Acta 9, 1-32. Leckie, J.G., Benjamin, M.M., Hayes, K.F. Kaufman, G. and Altman, S. (1980) AdsotptionZQOQrecipitation of trace metais ftom water with irog oxyhydtoxide. Electric Power Research Institute, Project 910-1. Leckie, J.G., Appleton, A.R., Ball, N.B., Hayes, K.F. and Honeyman, 3.0. (1983) Adsgrptive temgvai gf ttece meteis item tiy-eeh effluents onto iton gxynydroxide. Electric Power Research Institute, Project 910-1. Lingane, J.J. and Karplus, R. (1946) New method for determination of manganese. Ind. Eng. Chem., 18, 191. Loganathon, P. and Burau, R.G. (1973) The sorption of heavy metal ions by a hydrous manganese oxide. Geochim. Cosmochim. Acta, 37, 1277-1293. Loring, D.H. (1979) Geochemistry of Co, Ni, Cr and V in the sediments of the estuary and open Gulf of St. Lawrence. Can. J. Earth Sci., 16, 1197-1209. Louma, S.N. and Davis, J.A. (1983) Requirements for modeling trace element partitioning in oxidized estuarine sediments. Mar. Chem. 12, 159-181. MacNaughton, M.G. (1974) The adsorption of chromium (III) at the oxide/water interface. Interim report TR-75-17, Air Force Civil Engineering Center, Tyndall AFB Fl 32401. Martin, T.D. and Riley, J.K. (1982) Determining dissolved hexavalent chromium in water and wastewater by electrothermal atomization. Atomic Spectroscopy, 3, 1748179. Mayer, L.M. and Schick, L.L. (1981) Removal of hexavalent chromium from estuarine waters by model substrates and natural sediments. Env. Sci. Tech., 15, 1482-1484. 269 McKee, J.E. and Wolf, H.W. (1963) Watet Quaiity Qtiteria, 2nd edition, St. Wat. Qual. Cont. Bd., Sacramento, CA McKenzie, R.M. (1971) The synthesis of birnessite, cryptomelane and some other oxides and hydroxides of manganese. Min. Mag., 38, 493-502. McKenzie, R.M. (1980) The adsorption of lead and other heavy metals on oxides of manganese and iron. Aust. J. Soil Res., 18, 61-73. McKenzie, R.M. (1979) Proton release during adsorption of heavy metal ions by a hydrous manganese dioxide. Geochim. Cosmochim. Acta, 43, 1855-1857. Mertz, Walter (1971) Chromium: the relation of selected trace elements to health and disease. In: Envitgnmegtei Geoghemistty in Heeitn agg Qieeese, Geol. Soc. Am. Men. No. 123, Geological Society of America, Boulder CO, 29-35. Morel, F.M.M., Yeasted, J.C. and Westall, J.C. (1981) Adsorption models: a mathematical analysis in the framework of general equilibrium calculations. In: Adsor tion of nor anics at the solid-li id t rfa e. (Anderson, M.A. and Rubin, A.J., eds.) Ann Arbor Science, 263-294. Morgan, J.J. and Stumm, W. (1964) Colloid-chemical properties of manganese dioxide. J. Colloid Sci., 19, 3478359. Murray, J.W., Spell, B. and B. Paul (1983) The contrasting geochemistry of manganese and chromium in the eastern tropical Pacific ocean. From: Itace Metals in Seawate; (Wong, C.S. et al., eds.) Plenum Press, 643-669. Murray, J.W. (1974) The surface chemistry of hydrous manganese dioxide. J. Colloid Interface Sci., 46, 3578371. Murray J.W. (1975) The interaction of ions at the manganese dioxide-solution interface. Geochim. Cosmochim. Acta., 39, 505-519. Murray, J.W., Balistrieri, L.S. and Paul, B. (1984) The oxidation state of manganese in marine sediments and ferromanganese nodules. Geochim. Cosmochim. Acta, 48, 123781248. Murray, J.W. and Dillard, J.G. (1979) The oxidation of coba1t(II) adsorbed on manganese dioxide. Geochim. Cosmochim. Acta, 43, 781-787. 270 Nakayama E., Kuwamoto, T. Tsurubo, S. and Fujinaga, T. (1981) Chemical speciation of chromium in sea water: Part 2. Effects of manganese oxides and reducible organic materials on the redox processes of chromium. Anal. Chim. Acta, 30, 401-404. Pankow, J.F., Leta, D.P., Lin, J.W., Ohl, S.E, Shum, W.P. and Janaurer, G.E. (1977) Analysis for chromium traces in the aquatic ecosystem. Sci. Tot. Env., 7, 17-26. Parks, G.A. (1965) The isoelectric points of solid oxides, solid hydroxides and aqueous hydroxo complex systems. Chem. Rev., 65, 177-198. Posselt, H.S., Anderson, F.J. and Weber, W.J. (1968) Cation adsorption on colloidal hydrous manganese dioxide. Env. Sci. Tech., 2, 1087-1093. Postma, D. (1985) Concentration of Mn and separation from Fe in sediments--1. Kinetics and stoichiometry of the reaction between birnessite and dissolved Fe(II) at 10°C. Geochim. Cosmochim. Acta, 49, 1023-1033. Potter, R.M. and Rossman, G.R. (1979) Mineralogy of manganese dendrites and coatings. Amer. Min., 64, 1219-1226. Rai, D., Zachara, J.M., Eary, L.E., Girvin, D.C., Moore, D.A., Resch, C.T., Sass, B.M. and Schmidt, R.L. (1986) Geochemicei Behavior of Chtomium Species. Electric Power Research Institute, Project 2485-3. Robertson, F.N. (1975) Hexavalent chromium in ground water in Paradise Valley, Arizona., 516-527 Ross, S.J., Franzmeier, D.P. and Roth, C.B. (1976) Mineralogy and chemistry of manganese oxides in some Indiana soils. Soil Sci. Soc. Am. J., 40, 137-143. Schindler, P.W. (1981) Surface complexes at oxide-water interfaces. In: Adsorption of inorganics at solid- ligpid intertaces (Anderson, M.A. and Rubin, A.J. eds.) Ann Arbor Science, 1-50. Schindler, P.W., Walti, E., and Furst, B. (1976) The role of surface hydroxyl groups in the surface chemistry of metal oxides. Chimia, 30, 107-109. Schroeder, D.C. and Lee, G.F. (1975) Potential transformations of chromium in natural waters. Wat., Air and Soil Poll., 4, 355-365. 271 Smillie, R.H., Hunter, K., and Loutit, M. (1981) Reduction of chromium (VI) by bacterially produced hydrogen sulfide in a marine environment. Wat. Res., 15, 1351- 1354. Standard Methods (1975) American Public Health Association, 14th Edition, 1193p. Stone, A.T. and Morgan, J.J. (1984) Reduction and dissolution of manganese (III) and manganese (IV) oxides by organics: I. Reaction with hydroquinone. Env. Sci. Tech., 18, 450-456. Stumm W. and Morgan J.J. (1981) Agpatic Chemistty 2nd Ed. Wiley, 780 p. Stumm, W., Huang, C.P. and Jenkins, S.R. (1970) Specific chemical interactions affecting the stability of dispersed systems. Croat. Chem. Acta, 42, 223-245. Swallow, K.C., Hume, D.N. and Morel, F.M.M. (1980) Sorption of copper and lead by hydrous ferric oxide. Env. Sci. Tech., 14, 1326-1331. Turekian, K.K., (1978) Chemical Oceanography vol. 2 (J.P Riley and R. Skirrow, eds.) Academic Press, New York. Van den Berg, C.M.G. and Kramer, J.R. (1979) Determination of complexing capacities and conditional stability constants for copper in natural waters using MnOz. Anal. Chim. Acta, 106, 113. Van der Weijden, C.H. and Reith, M. (1982) Chromium (III)- chromium (VI) interconversions in seawater. Mar. Chem. II, 5658572 Vuceta, J. and Morgan, J.J. (1978) Chemical modeling of trace metals in fresh waters: role of complexation and adsorption. Env. Sci. Tech., 12, 1302-1308. Westall, J.C. and Hohl, H. (1980) A comparison of electrostatic models for the oxide/water interface. Adv. Coll. Interface Sci., 12, 265-294. Zasoski, R.J. and Burau, R.G. (1979) A technique for studying the kinetics of adsorption in suspensions. Soil Sci. Am. J., 42, 372-374. Zinder, B., Furrer, G. and Stumm, W. (1986) The coordination chemistry of weathering: II. Dissolution of Fe(III) oxides. Geochim. Cosmochim. Acta, 50, 1861- 1869.