USF Libraries
USF Digital Collections

Influence of solution and surface chemistry on yttrium and rare earth element sorption

MISSING IMAGE

Material Information

Title:
Influence of solution and surface chemistry on yttrium and rare earth element sorption
Physical Description:
Book
Language:
English
Creator:
Quinn, Kelly Ann
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Iron
pH
Carbonate
Ionic strength
ICP-MS
Dissertations, Academic -- Marine Science -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: The sorption behavior of yttrium and the rare earth elements (YREEs) was investigated using a variety of hydroxide precipitates over a range of solution conditions. Experiments with amorphous hydroxides of Al, Ga, and In were conducted at constant pH (~6.0) and constant ionic strength (I = 0.01 M), while YREE sorption by amorphous ferric hydroxide was examined over a range of ionic strength (0.01 M <̲ I <̲ 0.09 M), pH (3.9 <̲ pH <̲ 7.1), carbonate concentration (0 M <̲ CO32-T <̲ 150 micro-M), and temperature (10°C <̲ T <̲ 40°C). Sorption results were quantified via distribution coefficients, expressed as ratios of YREE concentrations between the solid and the solution, and normalized to concentrations of the sorptive solid substrate. Distribution coefficient patterns for Al, Ga, and In hydroxides were well correlated with the pattern for YREE hydrolysis.In contrast, amorphous ferric hydroxide developed a distinct pattern that was different than those for Al, Ga, and In precipitates but similar to the pattern predicted for natural marine particles. YREE sorption was shown to be strongly dependent on pH and carbonate concentration, significantly dependent on temperature, and weakly dependent on ionic strength. Distribution coefficients for amorphous ferric hydroxide (iKFe) were used to develop a surface complexation model that contained (i) two equilibrium constants for sorption of free YREE ions (M3+) by surface hydroxyl groups, (ii) one equilibrium constant for sorption of YREE carbonate complexes (MCO3+), (iii) solution complexation constants for YREE carbonates and bicarbonates, (iv) a surface protonation constant for amorphous ferric hydroxide, and (v) enthalpies for M3+ sorption. This quantitative model accurately described (i) an increase in iKFe with increasing pH, (ii) an initial increase in iKFe with increasing carbonate concentration due to sorption of MCO3+, in addition to M3+, (iii) a subsequent decrease in iKFe due to increasing YREE complexation by carbonate ions (especially extensive for the heavy REEs), and (iv) an increase in iKFe with increasing temperature.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2006.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Kelly Ann Quinn.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 158 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001909637
oclc - 173182081
usfldc doi - E14-SFE0001672
usfldc handle - e14.1672
System ID:
SFS0025990:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Influence of Solution and Surface Chemistry on Yttrium and Rare Earth Element Sorption by Kelly Ann Quinn A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Marine Science University of South Florida Major Professor: Robert H. Byrne, Ph.D. Peter R. Betzer, Ph.D. Eric H. De Carlo, Ph.D. Luis Garcia-Rubio, Ph.D. Edward S. Van Vleet, Ph.D. Date of Approval: July 7, 2006 Keywords: iron, pH, carbonate, ionic strength, ICP-MS Copyright 2006 Kelly Ann Quinn

PAGE 2

Acknowledgments I wish to thank my committee members, Peter Betzer, Eric De Carlo, Luis GarciaRubio, and Edward Van Vleet, for their support and guidance throughout the dissertation process. I want to especially thank my advisor, Robert Byrn e, for accepting me into the graduate program as his student and encourag ing me to pursue a doc toral degree. I would also like to thank Johan Schijf for teaching me how to operate th e ICP-MS and for the numerous conversations about my research an d life in general. Finally, I want to thank my family and friends, especially my husband Er ic and my parents, for all of their love and support over the years. This work was funded by a grant from the National Science Foundation (OCE0136333). Additional financial a ssistance was provided by the Von Rosenstiel, Riggs, and Gulf Oceanographic Charitable Trust Endow ed Fellowships. I am extremely grateful to all of the donors who help support the College of Marine Science.

PAGE 3

i Table of Contents List of Tables iii List of Figures vi List of Acronyms and Symbols x Abstract xii 1. Introduction 1 2. Comparative Scavenging of Yttrium and the Rare Earth Elements in Seawater: Competitive Influences of So lution and Surface Chemistry 19 2.1 Abstract 19 2.2 Introduction 20 2.3 Materials and Methods 23 2.4 Results and Discussion 28 2.4.1 Comparative log iKS Results 28 2.4.2 Linear Free-energy Relationships 31 2.4.3 Inter-element Patterns in YREE Solution Complexation and Surface Complexation 34 2.4.4 Oceanic log iKS Patterns 36 2.4.5 Comparative log iKFe Data Obtained for Freshly Precipitated Fe(III) Hydroxides 40 2.4.6 Comparative log iKS Behavior of Yttrium and the Rare Earth Elements 41 2.4.7 Critical Issues in YREE Surface Complexation Behavior 42 3. Sorption of Yttrium and Rare Earth Elements by Amorphous Ferric Hydroxide: Influence of pH and Ionic Strength 44 3.1 Abstract 44 3.2 Introduction 45 3.3 Materials and Methods 47 3.3.1 Materials and Preparation of the Experimental Solutions 47 3.3.2 pH Dependence of YREE Sorption 48 3.3.3 Ionic Strength Dependence of YREE Sorption 49 3.3.4 Sampling and Analysis 49 3.4 Data Analysis 51

PAGE 4

ii 3.4.1 Modeling of pH and Ionic Strength Effects 51 3.4.2 Surface Complexation Model 51 3.5 Results and Discussion 54 3.5.1 Empirical Model of the log iKFe Dependence on pH and Ionic Strength 54 3.5.2 Surface Complexation Model Results 61 3.5.3 Comparative log iKFe Predictions using SCM Results 65 3.5.4 log iKFe Predictions for Seawater 71 3.6 Conclusions 76 4. Sorption of Yttrium and Rare Earth Elements by Amorphous Ferric Hydroxide: Influence of Solution Complexation with Carbonate 78 4.1 Abstract 78 4.2 Introduction 79 4.3 Theory 82 4.4 Materials and Methods 86 4.5 Results and Discussion 89 4.5.1 Model Results Considering Sorption of Only Free YREEs 89 4.5.2 Model Results Including So rption of a YREE Carbonate Complex 92 4.5.3 Examination of the Competing Influences of Surface and Solution Complexation on T iFeK 99 4.6 Summary 102 5. Sorption of Yttrium and Rare Earth Elements by Amorphous Ferric Hydroxide: Influence of Temperature 104 5.1 Introduction 104 5.2 Materials and Methods 105 5.3 Results and Discussion 107 References 116 Appendices 127 Appendix A: Data for Freshly Prec ipitated Hydroxides of Trivalent Cations (Al3+, Ga3+, and In3+) 128 Appendix B: pH Dependent Data for Amorphous Ferric Hydroxide 131 Appendix C: Ionic Strength De pendent Data for Amorphous Ferric Hydroxide 139 Appendix D: Data for Amorphous Ferr ic Hydroxide Covering a Range of Carbonate Concentrations (2 3T[CO]) 142 Appendix E: Temperature Depende nt Data for Amorphous Ferric Hydroxide 155 About the Author End Page

PAGE 5

iii List of Tables Table 1.1 Some basic YREE properties, including name, symbol, atomic number (Z), atomic weight, a nd trivalent ionic radius for coordination number 6 (Shannon, 1976) 2 Table 1.2 YREE abundances (ppm and mol/kg) in mean shale (Haskin and Haskin, 1966), NASC (Haskin et al ., 1968), and PAAS (McLennan, 1989) 5 Table 2.1 Average log iKS results for iron, aluminum, gallium and indium (Figures 2.2A, 2.3A, 2.3B, and 2.2B, respectively) 31 Table 3.1 Results for the coefficients of equation (3.2) 57 Table 3.2 YREE surface complexation constants (Sn) determined with equation (3.19) and the data in Tables B.1B.6 63 Table 3.3 Estimated removal rates for YREEs via auth igenic iron 75 Table 4.1 YREE surface complexation constants (S1 and S2) determined using equations (4.12) and (4.13) with 3T[HCO] = 0 M, log SK1 = 4.76 (Quinn et al., 2006a), and th e experimental distribution coefficient results from carbonate-fr ee solutions in the present work (Tables B.1B.6 and D.4D.7) 90 Table 4.2 YREE surface complexation constants (S1, S2, and 3CO S1 ) determined with equations (4.12) and (4.13), log SK1 = 4.76 (Quinn et al., 2006a), and the experimental distribution coefficient results from the present work (Tables B.1B.6 and D.1D.7) 94 Table 5.1 YREE surface complexation constants (29815 S1 and 29815 S2 ) and enthalpy values (0 1H and 0 2H ; kcal/mol) determined with equations (5.4) and (5.6), log SK1 = 4.76 (Quinn et al., 2006a), and the distribution coefficient data in Tables B.1B.6, D.4D.7, and E.1E.4 112

PAGE 6

iv Table A.1 Distribu tion coefficient (log iKAl) results from the experiment performed at pH = 5.86 0.18 with an aluminum concentration of 1.00 0.05 mM 128 Table A.2 Distribu tion coefficient (log iKGa) results from the experiment performed at pH = 6.12 0.34 with a gallium concentration of 1.11 0.06 mM 129 Table A.3 Distribu tion coefficient (log iKIn) results from the experiment performed at pH = 6.08 0.04 with an indium concentration of 1.09 0.05 mM 130 Table B.1 Distribu tion coefficient (log iKFe) results from the experiment performed at pH = 5.15 0.02 with an iron concentration of 0.613 0.042 mM 131 Table B.2 Distribu tion coefficient (log iKFe) results from the experiment performed at pH = 6.12 0.05 with an iron concentration of 0.108 0.008 mM 132 Table B.3 Distribu tion coefficient (log iKFe) results from the experiment performed at pH = 7.06 0.05 with an iron concentration of 0.108 0.008 mM 133 Table B.4 Distribu tion coefficient (log iKFe) results from the experiment performed over the pH range 5.1 – 7.0 with an iron concentration of 0.108 0.008 mM 134 Table B.5 Distribu tion coefficient (log iKFe) results from the experiment performed over the pH range 5.1 – 7.0 with an iron concentration of 0.108 0.008 mM 135 Table B.6 Distribu tion coefficient (log iKFe) results from the experiment performed over the pH range 3.9 – 5.6 with an iron concentration of 10.0 0.7 mM 136 Table C.1 Distribu tion coefficient (log iKFe) results from the experiment performed over the ionic strength range 0.01 – 0.09 M with an iron concentration of 0.108 0.008 mM 139 Table C.2 Distribu tion coefficient (log iKFe) results from the experiment performed over the ionic strength range 0.01 – 0.09 M with an iron concentration of 0.108 0.008 mM 141

PAGE 7

v Table D.1 Distribu tion coefficient (log T iFeK) results from the experiment performed at pH = 5.38 0.02 and 30% CO2 142 Table D.2 Distribu tion coefficient (log T iFeK ) results from the experiment performed over the pH range 4.6 – 6.6 at 3% CO2 143 Table D.3 Distribu tion coefficient (log T iFeK) results from the experiment performed over the pH range 4.0 – 6.6 at 30% CO2 144 Table D.4 Distribu tion coefficient (log T iFeK) results from the experiment performed over the 2COP range 0% – 30% at pH = 6.52 0.01 145 Table D.5 Distribu tion coefficient (log T iFeK) results from the experiment performed over the 2COP range 0% – 30% at pH = 6.68 0.01 148 Table D.6 Distribu tion coefficient (log T iFeK ) results from the experiment performed over the 2COP range 0% – 30% at pH = 7.06 0.01 150 Table D.7 Distribu tion coefficient (log T iFeK) results from the experiment performed over the 2COP range 0% – 30% at pH = 7.10 0.03 152 Table E.1 Distribu tion coefficient (log iKFe) results from the experiment performed at T = 10.0oC over the pH range 4.7 – 6.9 with an iron concentration of 1.08 0.08 mM 155 Table E.2 Distribu tion coefficient (log iKFe) results from the experiment performed at T = 10.0oC over the pH range 5.0 – 7.1 with an iron concentration of 1.08 0.08 mM 156 Table E.3 Distribu tion coefficient (log iKFe) results from the experiment performed at T = 39.1oC over the pH range 4.9 – 6.8 with an iron concentration of 0.108 0.008 mM 157 Table E.4 Distribu tion coefficient (log iKFe) results from the experiment performed at T = 39.3oC over the pH range 5.3 – 7.1 with an iron concentration of 0.108 0.008 mM 158

PAGE 8

vi List of Figures Figure 1.1 (A) Dissolved YREE concentrations (pmol/kg) at a depth of 689 m in the Pacific Ocean (Zhang and Nozaki, 1996). (B) Shalenormalized pattern of the seawater sample shown in panel A using the PAAS values given in Table 1.2 (McLennan, 1989) 4 Figure 1.2 YREE speciation in seawater (pH 8.2 and 3T[HCO] = 2 mM) expressed as the ratio of species concentration to the total YREE concentration (log ([ML]/MT)) 10 Figure 2.1 Comparison between (A) predicted log iKS values (Byrne and Sholkovitz, 1996) and (B) directly measured log iKFe values using ferric oxyhydroxides (Bau, 1999) 24 Figure 2.2 log iKS results from filtered samples. (A) 100 M Fe(OH)3 (Table B.2). (B) 1 mM In(OH)3 (Table A.3). (C) 1 mM Al(OH)3 (Table A.1). (D) 1 mM Ga(OH)3 (Table A.2) 28 Figure 2.3 log iKS results normalized to pH 6.10. (A) 1 mM Al(OH)3. (B) 1 mM Ga(OH)3 30 Figure 2.4 Stability constants for YR EE complexes with hydroxide (Klungness and Byrne, 2000) and fluoride (Luo and Byrne, 2000), for the conditions of our experiments (T = 25oC; I = 0.014 M) 32 Figure 2.5 Linear free-energy relationships between the log iKS results from this work and log OH1 data from Klungness and Byrne (2000). (A) 100 M Fe(OH)3. (B) 1 mM In(OH)3. (C) 1 mM Al(OH)3. (D) 1 mM Ga(OH)3 33 Figure 2.6 Ratio of iKS (this work) and F1 (Luo and Byrne, 2000), shown as log (iKS/F1). (A) 100 M Fe(OH)3. (B) 1 mM In(OH)3. (C) 1 mM Al(OH)3. (D) 1 mM Ga(OH)3 35

PAGE 9

vii Figure 2.7 (A) Directly measured YREE concentr ations in seaw ater at three depths (Zhang and Nozaki, 1996), normalized to Post-Archean Australian Shale (PAAS) (McLennan, 1989). (B) The calculated solution complexation term from equa tion (2.2) at pH 7.6, 7.9, and 8.2 with a bicarbonate concentration of 210-3 M. (C) Predicted log iKS values, calculated by subtracting each of the curves in Figure 2.7A from the curve for pH 7.9 in Figure 2.7B 38 Figure 2.8 Comparison between the log iKFe values from this work and log iKS values for the three seawater samples from Figure 2.7C 39 Figure 2.9 Comparison between the log iKFe values from this work at pH 6.12, Bau (1999) at pH 5.97 (A), and Ohta and Kawabe (2000, 2001) at pH 6.01 (B) 40 Figure 3.1 Final regressions of log iKFe (Tables B.1–B.6; normalized to I = 0 M) versus pH for La, Sm, Dy, and Lu 55 Figure 3.2 Final regressions of log iKFe (Tables C.1 and C.2; normalized to pH 6.13) versus ionic strength (I) for La, Sm, Dy, and Lu 56 Figure 3.3 Coefficients of equation (3.2) 58 Figure 3.4 log iKFe(pred)/log iKFe(meas) versus pH for La, Sm, Dy, and Lu, where log iKFe(pred) are distribution coe fficients predicted from equation (3.2) using the coeffici ents listed in Table 3.1, and log iKFe(meas) are experimentally obser ved distribution coefficients (Tables B.1–B.6) 59 Figure 3.5 log iKFe results for experiments performed at (A) pH = 7.06 0.05 and I = 0.0109 M (Table B.3) and (B) pH = 6.10 0.03 and I = 0.0503 M (Table C.1) 60 Figure 3.6 log iKFe(pred)/log iKFe(meas) versus pH for Sm. (A) log iKFe(pred) are distribution coefficients predic ted from equation (3.19) with the assumption that SK1[H+] 1. (B) log iKFe(pred) are distribution coefficients predicted from equation (3.19) using log SK1 = 4.76 and the surface complexation constants (Sn) listed in Table 3.2 62 Figure 3.7 Surface stability constant s (equation (3.19)) for YREE sorption by amorphous ferric hydroxide 64

PAGE 10

viii Figure 3.8 Comparison between measured distribution coefficients and log iKFe values predicted from equation (3.19). (A) Result from Bau (1999) at pH 5.97. (B) Result from Ohta and Kawabe (2000, 2001) at pH 6.59. (C) Result from De Carlo et al. ( 1998) at pH 6.25 and I = 0.1 m 66 Figure 3.9 Comparison between the average log iKFe result at pH = 3.96 0.10 from the present work (Table B.6) and the log iKFe result at pH 3.91 from Bau (1999) 68 Figure 3.10 Regressions of log iKFe versus pH for La, Sm, Dy, and Lu 69 Figure 3.11 Regressions of log iKFe versus pH for La, Sm, Dy, and Lu 70 Figure 3.12 (A) Distribution coefficients (log iKFe) expressed in terms of free YREE concentrations ([M3+]) using equation (3.19) and the surface complexation constants (Sn) listed in Table 3.2. (B) Distribution coefficients (iFelogK ) expressed in terms of total YREE concentrations in seawater (MSW) using equations (3.19) and (3.22) and assuming 3 3T[HCO]210M. (C) Comparison between the predicted iFelogK pattern for seawater at pH 7.8 (equations (3.19) and (3.22)) and the measured iFelogK pattern at pH 7.8 from Koeppenkastrop and De Carlo (1992) 73 Figure 4.1 T iFelogK results at pH 7.06 and vari ous carbonate concentrations, 2 3T[CO], listed in the legend. (A) T iFelogK(pred) are distribution coefficients predicted from equation (4.12) using the S1 and S2 results listed in Table 4.1 and 3CO S1 = 0. (B) T iFelogK(meas) are directly measured distribution co efficients from an experiment performed at constant pH (7.06) and increasing 2COP (Table D.6) 91 Figure 4.2 T iFelogK(meas) versus T iFelogK(pred) for La, Sm, Dy, and Lu 93 Figure 4.3 Surface stability constant s (equation (4.12)) for YREE sorption by amorphous ferric hydroxide 95 Figure 4.4 Regressions of T iFelogK(meas) versus T iFelogK(pred) for La, Sm, Dy, and Lu 97 Figure 4.5 T iFelogK patterns covering a range of carbonate concentrations 98

PAGE 11

ix Figure 4.6 Patterns of the surface complexation term (log ([MSi]T[M3+]-1[Si]-1)) in equation (4.26) calculate d with equation (4.27) 100 Figure 4.7 Patterns of the solution complexation term (log (MT/[M3+])) in equation (4.26) calculate d with equation (4.8) 101 Figure 5.1 log iKFe results obtained over a range of temperatures, indicated in the legend, at pH 5.61 0.05 (A) and pH 7.06 0.03 (B) 108 Figure 5.2 log iKFe versus pH for La, Sm, Dy, and Lu at 10, 25, and 40oC 109 Figure 5.3 Regr essions of log iKFe(meas) versus log iKFe(pred) for La, Sm, Dy, and Lu 110 Figure 5.4 Enthalpy values (equatio n (5.4)) for YREE sorption by amorphous ferric hydroxide from 10 to 40oC 113 Figure 5.5 Regr essions of log iKFe(meas) versus log iKFe(pred) for La, Sm, Dy, and Lu 115

PAGE 12

x List of Acronyms and Symbols HREE heavy rare earth elements ICP-MS inductively-coupled plasma mass spectrometer ID-TIMS isotope dilution therma l ionization mass spectrometry INAA instrumental neut ron activation analysis LREE light rare earth elements MREE middle rare earth elements NASC North American Shale Composite PAAS Post-Archean Average Australian Shale SCM surface complexation model YREE yttrium and the rare earth elements iKS distribution coeffi cient, defined as iKS = [MSi]T/([M3+][Si]), where the subscript S can be Fe, Al, Ga, or In [MSi]T total concentration of sorbed YREE [M3+] concentration of a free hydrated YREE ion [Si] concentration of sorptive solid substrate T iFeK distribution coefficient expr essed in terms of total YREE concentration in solution, defined as T iFeK = [MSi]T/ (MT[Fe3+]S) MT total concentration of YREE in solution [Fe3+]S concentration of precipitated iron iFeK distribution coefficient expr essed in terms of total YREE concentration in seawater, defined as iFeK = [MSi]T/ (MSW[Fe3+]S) MSW total concentration of YREE in seawater I ionic strength QpH slope of the linea r regression of log iKFe with respect to pH QI slope of the linea r regression of log iKFe with respect to ionic strength log iKFe(pH 0, I = 0) intercept of the linear regression of log iKFe Sn stability constant for free YREE sorption by surface hydroxyl groups, defined as Sn = ([S–FeOn(OH)3-nM3-n][H+]n)/ ([M3+][S–Fe(OH)3]) where S– represents the bulk solid SK1 ferric hydroxide surface prot onation constant, defined as SK1 = 0 23[SFe(OH)]/[SFe(OH)][H]

PAGE 13

xiSK2 ferric hydroxide surface deprotonation constant, defined as SK2 = 0 43[SFe(OH)][H]/[SFe(OH)] 3CO S1 stability constant for sorpti on of YREE-carbonate complexes, defined as 3CO 0 S123[SFeO(OH)MCO][H]/ 33[MCO][SFe(OH)] in stability constant of the nth complex of metal M with solution ligand Li, defined as in = [M(Li)n]/([M3+][Li]n) where Li can be 2 3CO OH-, F-, 3HCO etc. 3H COn YREE complexation constant with carbonate expressed in terms of total bicarbonate concentrations, defined as 3H COn 32nn3n 3n3T[M(CO)][H]/[M][HCO] 0 nH enthalpy

PAGE 14

xii Influence of Solution and Surface Chemistry on Yttrium and Rare Earth Element Sorption Kelly Ann Quinn ABSTRACT The sorption behavior of yttrium and the rare earth elements (YREEs) was investigated using a variety of hydroxide prec ipitates over a range of solution conditions. Experiments with amorphous hydroxides of Al Ga, and In were conducted at constant pH (~6.0) and constant ionic strength (I = 0.01 M), while YREE sorption by amorphous ferric hydroxide was examined over a range of ionic strength (0.01 M I 0.09 M), pH (3.9 pH 7.1), carbonate concentration (0 M 2 3T[CO] 150 M), and temperature (10oC T 40oC). Sorption results were quantifie d via distribution coefficients, expressed as ratios of YREE concentrations between the solid and the solution, and normalized to concentrations of the sorptiv e solid substrate. Di stribution coefficient patterns for Al, Ga, and In hydroxides were well correlated with the pattern for YREE hydrolysis. In contrast, amorphous ferric hydr oxide developed a distin ct pattern that was different than those for Al, Ga, and In precipitates but similar to the pattern predicted for natural marine particles. YREE sorption was shown to be strong ly dependent on pH and carbonate concentration, significantly dependent on te mperature, and weakly dependent on ionic strength. Distribution coefficients for amorphous ferric hydroxide (iKFe) were used to develop a surface complexation model that co ntained (i) two equilibrium constants for sorption of free YREE ions (M3+) by surface hydroxyl groups, (ii) one equilibrium constant for sorption of YREE carbonate complexes (3MCO ), (iii) solution complexation constants for YREE carbonates and bicarbonate s, (iv) a surface protonation constant for amorphous ferric hydroxide, and (v) enthalpies for M3+ sorption. This quantitative model

PAGE 15

xiii accurately described (i) an increase in iKFe with increasing pH, (ii) an initial increase in iKFe with increasing carbonate conc entration due to sorption of 3MCO, in addition to M3+, (iii) a subsequent decrease in iKFe due to increasing YREE complexation by carbonate ions (especially extensive for th e heavy REEs), and (iv) an increase in iKFe with increasing temperature.

PAGE 16

1 1. Introduction Yttrium and the fourteen stable rare earth elements (YREEs) are extensively used to study geochemical processes in all types of na tural water (e.g., rivers, estuaries, and the ocean). The YREEs are ideal tools to expl ore fundamental aqueous reactions because they form a coherent series of elements whose chemical properties display small but systematic changes with increasing atomic nu mber. This chemical co herence is due to the gradual filling of their 4f electron shell. Because outer elec trons (n = 5, 6) shield this inner shell, there are only minor differences in the chemical reactivity along the series. The empty (La3+), half-filled (Gd3+), and completely filled (Lu3+) 4f electron shells have increased stability and therefore may display anomalous behavior rela tive to the rest of the YREE series (e.g., de Baar et al., 1991; McLennan, 1994). The dominant systematic change that is observed in YREE chemical pr operties, such as solution complexation, is caused by the decrease in ionic radius with in creasing atomic number, which is known as the lanthanide contraction. Table 1.1 lists th e ionic radii of the YR EEs, along with their atomic numbers and atomic weights. As can be seen in Table 1.1, Y has an ioni c radius almost equal to that of Ho and, therefore, the two elements ar e expected to display similar geochemical behaviors. It has been shown though that Y resembles a variety of REEs in its complexation characteristics (e.g., Moeller, 1963, 1972; Moeller et al., 1965; Byrne and Lee, 1993; Liu and Byrne, 1995). Because the 4f orbitals of the REEs influence bonding, the REEs display enhanced covalency compared to Y (Siekierski, 198 1; Borkowski and Siekierski, 1992). This delocalization of f orbitals in the REEs causes Y to behave as a light pseudolanthanide when complexing with soft ligands (e.g., orga nics). On the other hand, Y approaches the chemical behavior of Ho when participating in ionic interactions with hard ligands (e.g., carbonate).

PAGE 17

2Table 1.1. Some basic YREE properties, includi ng name, symbol, atomic number (Z), atomic weight, and trivalent ionic radius for coordination number 6 (Shannon, 1976). Element Symbol Z Atomic weight (g/mole) Ionic radius () yttrium Y 39 88.9059 0.900 lanthanum La 57 138.9055 1.032 cerium Ce 58 140.115 1.01 praseodymium Pr 59 140.9077 0.99 neodymium Nd 60 144.24 0.983 promethium Pm 61 (145) samarium Sm 62 150.36 0.958 europium Eu 63 151.96 0.947 gadolinium Gd 64 157.25 0.938 terbium Tb 65 158.9254 0.923 dysprosium Dy 66 162.50 0.912 holmium Ho 67 164.9304 0.901 erbium Er 68 167.26 0.890 thulium Tm 69 168.9342 0.880 ytterbium Yb 70 173.04 0.868 lutetium Lu 71 174.967 0.861 Another property that makes the YREEs a good probe of geochemical processes is oxidation state. All YREEs are present as trival ent ions in natural waters, with Ce and Eu also existing in the tetravalent and divalent states, respectively. Oxidation of Ce, which results in the formation of an insoluble oxide (CeO2), rapidly occurs in the upper water column of the ocean (e.g., Goldberg et al., 19 63; de Baar et al., 1983; German et al., 1995; Alibo and Nozaki, 1999). Pro cesses that convert dissolved Ce3+ to particulate Ce4+ include biologically mediated oxidation (M offett, 1990, 1994a,b) and abiotic oxidation on the surfaces of Mn oxides (Koeppenkastrop and De Carlo, 1992; Sholkovitz et al., 1994). Reduction of Eu, on the other hand, generally occurs at high temperatures and pressures, such as those found in hydrotherm al fluids (e.g., Michar d et al., 1983; German et al., 1990; Klinkhammer et al., 1994). The anomalous behaviors of Ce and Eu caused by redox reactions are quantified via Ce and Eu anomalies, defined as: Ce anomaly = 3Cen/(2Lan + Ndn) or 2Cen/(Lan + Prn) (1.1) and

PAGE 18

3 Eu anomaly = 2Eun/(Smn + Gdn), (1.2) where the subscript n represents shale-no rmalized concentrations. Depletions or enrichments of either element relative to neighboring elements yield values less than 1 (negative anomalies) or greater than 1 (positive anomalies). Differences in the ionic radii and the oxidation states of the YREEs lead to mass fractionation, defined by Byrne and Sholkovitz (1996) as “the variation in the relative lanthanide concentrations through biogeoc hemical reactions.” This fractionation in natural samples can be masked by the effect of the Oddo-Harkins Rule, which holds that elements with an even atomic number are more cosmogenically abundant than those with an odd atomic number. As an example of this odd-even pattern, Figure 1.1 shows the YREE abundances in a seawater sample (Zha ng and Nozaki, 1996). To remove this sawtooth distribution, samples are normalized to a chosen reference material, which is typically shale for samples obtained in the ma rine environment (Figure 1.1). As discussed by Piper (1974), REE patterns of marine samples more closely resemble the pattern of shale than that of chondrit e, the preferred reference material for normalization of terrestrial rocks and minerals Shale also represents the upper continental crust (Taylor and McLennan, 1985), which is considered to be the principal source material for REEs to the ocean. Several different shale values in the litera ture have been used for normalization, including (i) mean shale, an average of th e North American, European, and Russian shale composite (Haskin and Haskin, 1966; Piper, 1974; de Baar et al., 1985a), (ii) North American Shale Composite (NASC), an average of 40 shales mainly from North America (Haskin et al., 1968; Gromet et al., 1984; Go ldstein and Jacobsen, 1988), and (iii) PostArchean Average Australian Shale (PAAS), an average of 23 shales from Australia (Nance and Taylor, 1976; McLennan, 1989). Ta ble 1.2 lists the YREE abundances (ppm and mol/kg) in these three shales. By calculati ng anomalies in a manner similar to that shown in equations (1.1) and (1.2), Alibo and Nozaki (1999) showed for seve ral seawater samples that the magnitude and direction (pos itive or negative) of the anomaly for all REEs, except La and Ce, varied depending on the shale values used in the normalization.

PAGE 19

4 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu dissolved [YREE] (pmol/kg) (logarithmic scale) 0.1 1 10 100 1000 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu shale-normalized ratio (x108) (logarithmic scale) 0.1 1 10 100 A B Figure 1.1. (A) Dissolved YREE concentrations (pmo l/kg) at a depth of 689 m in the Pacific Ocean (Zhang and Nozaki, 1996). (B) Shale-normalized pattern of the seawater sample shown in panel A using the PAAS va lues given in Table 1.2 (McLennan, 1989). This indicates that interpretations of anomalies must be made cautiously. Despite the variation in specific anomalies, the major feat ures of shale-normalized patterns described below are maintained with different reference shales. As can be seen in Figure 1.1, the YREEs display a shale-normalized pattern in seawater that is enriched in heavy REEs (HREEs) relative to light REEs (LREEs). Additionally, there is a pronounced Ce depletio n relative to La and Pr creating a large

PAGE 20

5Table 1.2. YREE abundances (ppm and mol/kg) in mean shale (Haskin and Haskin, 1966), NASC (Haskin et al., 1968) and PAAS (McLennan, 1989). Mean shale NASC PAAS ppm mol/kg ppm mol/kg ppm mol/kg Y 36 405 27 304 27 304 La 41 295 32 230 38.2 275 Ce 83 592 73 521 79.6 568 Pr 10.1 71.7 7.9 56.1 8.83 62.7 Nd 38 263 33 229 33.9 235 Pm Sm 7.50 49.9 5.7 37.9 5.55 36.9 Eu 1.61 10.6 1.24 8.16 1.08 7.11 Gd 6.35 40.4 5.2 33.1 4.66 29.6 Tb 1.23 7.74 0.85 5.35 0.774 4.87 Dy 5.50 33.8 5.8 35.7 4.68 28.8 Ho 1.34 8.12 1.04 6.31 0.991 6.01 Er 3.75 22.4 3.4 20.3 2.85 17.0 Tm 0.63 3.73 0.50 2.96 0.405 2.40 Yb 3.53 20.4 3.1 17.9 2.82 16.3 Lu 0.61 3.49 0.48 2.74 0.433 2.47 negative Ce anomaly. The HREE enrichment an d Ce depletion increase with depth in the ocean and also increase from the Atlantic Ocean to the Pacific Ocean (Byrne and Sholkovitz, 1996, and references therein). As descri bed above, Ce exhibits anomalies due to its active redox chemistry. For elements not influenced by redox transformations, several descriptions of the processes that control the fr actionation between HREEs and LREEs in seawater have been considered over the past 40 years. Following one of the first measurements of REE concentrations in seawater, Goldberg et al. (1963) suggested that th e HREE enrichment may be due either to increasing stability of solution complexes acro ss the REE series or differential sorption by solids. Hgdahl et al. (1968) proposed th at REE fractionation may be caused either by redox reactions along with differential mineral uptake or by differential solubility due to variations in ionic radius. Howe ver, it was anticipated (Moell er et al., 1965; Hgdahl et al., 1968) that neither process would be effec tive for the REEs in seawater. Basing their

PAGE 21

6 interpretation on the proposition that REE di stributions are determined by the suspended matter in natural waters, Kolesov et al. (1975) stated that the relative abundances of the REE depend on the composition of the suspe nded matter, the formation of REE solution complexes with organic and inorganic ions, and REE solubility. Despite the fact that these initial studies were unable to establish th e cause of fractionation, they were able to provide relatively accurate descriptions of the overall REE fracti onation pattern (i.e., Figure 1.1). In addition to measuring REE concentratio ns in seawater, Goldberg et al. (1963) measured concentrations in a manganese n odule and suggested that incorporated REE were precipitated directly from seawater. Se veral studies were s ubsequently performed with oceanic ferromanganese nodules to dete rmine the mechanism of REE incorporation into nodules (Elderfield et al., 1981, and refe rences therein), but some researchers also used these studies to look at REE fractiona tion in seawater. Base d on observations that REE concentrations in ferromanganese nodules from depths greater than approximately 3500 m display a mirror-image pattern relative to seawater, Piper (1974) suggested that formation of deep-water nodul es fractionates the REE and th erefore could be responsible for the observed seawater fractionation pa ttern. Piper (1974) concluded that the mechanism controlling the fractionation mi ght involve REEs being released from dissolving biogenic tests and then being coprecipitated with Fe and Mn phases. Elderfield et al. (1981) agreed with Piper (1974) th at REE fractionation may be controlled by the formation of nodules but suggested that the pa ttern reflects the rela tive proportions of Ferich and P-rich carrier phases in the nodule with the Fe-rich phase producing the inverted seawater pattern. REE concentration determinations in seawat er experienced a hiat us of several years as methods more precise than instrumental neutron activation analysis (INAA) were developed to measure picomolar YREE concen trations. Compared to the REE chemical yield averaging about 90% in the work of Go ldberg et al. (1963), de Baar improved the analysis of seawater by INAA, demonstrating a chemical yiel d of 100% and precision of 2–5% (1 ) (de Baar et al., 1983, 1985a,b). A ma jor advancement came with the use of isotope dilution thermal ionization mass spectrometry (ID-TIMS) by Elderfield and

PAGE 22

7 Greaves (1982), which increased the precis ion and sensitivity of measurements to typically 1% (2 ). Over the next couple of years, additional measurements of REE concentrations in the ocean were interprete d as being indicative of fractionation during YREE scavenging by marine particulate matte r (Elderfield and Greaves, 1982; de Baar et al., 1983; Klinkhammer et al., 1983). It was al so proposed that since HREEs form more stable solution complexes in seawater, LREEs would be preferentially removed by the scavenging process (Elderfield and Greaves, 19 82; de Baar et al., 1983; Klinkhammer et al., 1983). Referring to the steady state sc avenging model of Schindler (1975) and Balistrieri et al. (1981), de Baar et al. (19 85b) stated that REE fractionation occurs during the “equilibration of REE(III) between inorga nic complexes in solution and surface sites on small suspended particles”. The work of de Baar et al. (1985a,b) may have been the first to explicitly attribute YREE fractionati on in seawater to th e competition between solution chemistry and surface chemistry. At the time of the de Baar et al. publications, YREE surface complexation characteristics were unknown and little was known about YREE solution complexation. Despite the fact that equilibrium constants had not yet been measured for REE solution complexes, Turner et al. (1981) assessed the speciation of trivalent REEs in seawater based on linear free-energy relationships for divalent metals. Using the observed correlation between the stability constant s of carbonate and oxalate complexes for divalent metals, Turner et al. (1981) calculate d that at a free carbonate ion concentration of 10-4.50 M, the fraction of free ion in seawater decreased along the REE series from 38% for La to 5% for Lu, while the fraction of carbonate complexes (3MCO only) generally increased from 22% for La to 81% for Yb. The remaining fraction of REEs was attributable to a combination of hydroxide, chloride, and sulfate complexes with fluoride complexes only accounting for approximately 1% of the total for each individual REE (Turner et al., 1981). Based on the fact that carbonate dominated the speciation of REEs in seawater, most of the initial studies involving YREE solution complexation focused on the carbonate ion. Initially, carbonate complexation consta nts were measured for individual REE, including La (Ciavatta et al., 1981), Eu (Lundqvist, 1982 ), Ce (Ferri et al., 1983), and Y

PAGE 23

8 (Spahiu, 1985). Since these four studies we re interested in groundwater speciation, experiments were performed at ionic strengths between 0.3 M and 3 M. In order to determine constants appropriate to seawater and also to improve the correlation used by Turner et al. (1981), Cantrell and Byrne (1987 a) measured carbonate and oxalate stability constants for Ce, Eu, and Yb at an ionic strength of 0.68 m. It was shown that 32M(CO) along with 3MCO, are the dominant species for REEs in seawater (Can trell and Byrne, 1987a), which significantly altered the specia tion scheme of Turner et al. (1981). Using the measured carbonate stabil ity constants for Ce, Eu, and Yb, along with those estimated for the remaining REEs via a quadratic func tion in atomic number, Cantrell and Byrne (1987a) calculated that at a tota l carbonate ion concentration of 10-3.86 m (free concentration = 10-4.71 m, using the ion pairing model of Millero and Schreiber, 1982), the fraction of carbonate complexes (3MCO and 32M(CO) ) increased along the YREE series from 86% for La to 98% for Lu, while the fraction of free ion decreased from 7% for La to 0.3% for Lu. Over the next several years, additional m easurements of carbonate complexation were carried out to further improve this speciatio n scheme, although the majority of studies involved only Eu (Thompson and Byrne, 1988; Chatt and Rao, 1989; Rao and Chatt, 1991). Since shale-normalized REE patterns in seawater exhibit an anomaly at Gd (de Baar et al., 1985b, 1991), Lee and Byrne (1993) determined carbonate complexation constants for Gd and its neighbors, Eu and Tb as well as representative light and heavy REEs (Ce and Yb). In addition to demonstratin g the “Gd-break” in the series, they also estimated carbonate stability constants for th e other REEs using the measured constants plus multiple linear regression analyses. By measuring carbonate stability constants for Y and Gd and then comparing the resulting Y and REE solution chemistries, Liu and Byrne (1995) showed that the solution complexation behavior of Y closely resembles that of Tb. The introduction of inductively-coupled plasma mass spectrometry (ICP-MS) allowed the entire YREE series to be studied simulta neously, greatly improving the understanding of comparative YREE aquatic geochemistry rela tive to perspectives that are gained by measurements of individual elements. Liu and Byrne (1998) determined carbonate stability constants for the entire YREE series by solvent exchange and ICP-MS. Stability

PAGE 24

9 constants for elements such as Ce, Eu, Gd, Tb, and Yb, which had been examined previously, were in good agreement with prior results. A pl ot of YREE carbonate complexation constants versus atomic number showed (i) a general increase from La to Lu, (ii) a negative anomaly at Gd, and (iii) a close relationship between Y and Eu (Liu and Byrne, 1998). Using adsorptive exchange analysis, Kawabe et al. (1999a) and Ohta and Kawabe (2000) also determined carbo nate complexation constants for the entire YREE series. However, their results were 1.0 – 1.5 log units higher than essentially all previously determined values obtained using a variety of techniques. In order to better constrain the results, Luo and Byrne (2004) utilized potentiometry to measure YREE carbonate stability constants. This work sh owed excellent agreement with the solvent exchange results of Liu and Byrne (1998). YREE behavior in natural waters has beco me increasingly well characterized through inclusion of YREE stability constant results for a variety of solu tion ligands, including bicarbonate (Luo and Byrne, 2004), fluoride (Schijf and Byrne, 1999; Luo and Byrne, 2000; Luo and Millero, 2004), hydroxide (Klungness and Byrne, 2000), chloride (Luo and Byrne, 2001), and sulfate (Schijf a nd Byrne, 2004). Using the above-referenced solution complexation constants, YREE speciati on for seawater can be calculated at pH 8.2 and a total bicarbonate concentration of 2 mM. Figure 1.2 shows the fraction of each species as log ([ML]/MT), where M represents a YREE ion and L represents a solution ligand. It can be seen that the results do not differ much from those of Cantrell and Byrne (1987a). In contrast to the extensive studies on in organic complexation, little is known about YREE solution complexation with natural organic matter. Byrne and Li (1995) summarized YREE complexation constants with 101 organic ligands, which were obtained from the compilations of Critical Stability Constants by Martell and Smith (1974, 1977, 1982) and Smith and Martell (1 975, 1989). More spec ific studies have determined YREE complexation constants with acetate (e.g., Kolat and Powell, 1962; Wood et al., 2000) and oxalate (e.g., Schijf and Byrne, 2001). These investigations have

PAGE 25

10 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log ([ML]/MT) -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 M(CO3)2 MCO3 + MOH2+ M3+ MSO4 + MCl2+ MF2+ MHCO3 2+ Figure 1.2. YREE speciation in seawater (pH 8.2 and 3T[HCO] = 2 mM) expressed as the ratio of species concentration to th e total YREE concentration (log ([ML]/MT)) (see text for details). utilized a variety of well-cha racterized organic ligands that may not accurately represent the complex properties of natural organic ma tter, which consists of humic acids, fulvic acids, and other biologically-derived materi al. Relatively few YREE binding constants with natural humic substances have been measured, and studies usually were conducted only for individual elements (s ee electronic annex of Sonke and Salters, 2006). Tang and Johannesson (2003) used the published data for Eu, Tb, and Dy plus linear free-energy relationships with the stability constants of lactic and acetic acids to estimate equilibrium constants for the entire YREE series. Combin ing these estimated constants with the Humic Ion-Binding Model V (Tipping, 1994), Tang and Johannesson (2003) predicted that organic matter dominates YREE speciatio n in circumneutral-pH river waters and organic-rich groundwaters (di ssolved organic carbon > 0.7 mg/L). To improve future calculations of YREE speciation in natural waters, Sonke and Salters (2006) determined humic substance binding constants for the entire YREE series.

PAGE 26

11 In addition to complexation by organic an d inorganic solution ligands, the other dominant process controlling YR EE distributions in natural waters is scavenging, or the sorptive removal of dissolved elements fr om solution by particles (Goldberg, 1954; Goldberg et al., 1963; Balistrieri et al., 1981; Bruland, 1983; Fowler and Knauer, 1986). YREEs are scavenged in estuaries, where fres hwater and seawater mix, by the formation of colloids (Sholkovitz, 1976, 1992; Sholkovitz and Elderfield, 1988). In the upper ocean, YREEs are predominantly scavenged by par ticles and particle coatings, which are composed of either organic matter or Fe oxid es (Balistrieri et al ., 1981; Hunter, 1983). Below approximately 300 m, concentrations of YREEs in seawater may also be controlled by YREE phosphate coprecipita tion (Jonasson et al., 1985; Byrne and Kim, 1993; Liu and Byrne, 1997; Liu et al., 1997). Before direct measurements were possibl e, the sorptive behavior of marine particulates was modeled based on the resi dence time of an element (Schindler, 1975). Two different methods were used for this modeling: (i) relative concentrations of dissolved YREEs were estimated using surf ace and solution complexation constants, and then compared to actual measurements, and (ii) surface complexation constants were estimated from shale-normalized YREE concentrations and solution complexation constants that had been directly measured. For the first method, the relevant surface groups and associated comple xation constants were not we ll known, so either YREE hydrolysis constants (Balistrie ri et al., 1981; Erel and Morg an, 1991; Erel and Stolper, 1993) or YREE complexation constants with dissolved monocarboxylic acids (Byrne and Kim, 1990) were used in the calculation. For the second method, Lee and Byrne (1993) and Byrne and Sholkovitz (1996) used the most accurate YREE ocean concentrations and solution complexation constants that were available at the time to calculate surface stability constants. These modeling efforts are described in more detail in Section 2.2. There have been relatively few measurem ents of REE concentrations in marine particulates, either sinking (Murphy and Dymond, 1984; Masuzawa and Koyama, 1989; Fowler et al., 1992; Tachikawa et al., 1997; Lerche and Nozaki, 1998) or suspended (Bertram and Elderfield, 1993; Sholkovi tz et al., 1994; Alibo and Nozaki, 1999; Tachikawa et al., 1999; Kuss et al., 2001). The shale-normalized REE patterns obtained

PAGE 27

12 in these studies show various degrees of enrichment in either HREEs (Murphy and Dymond, 1984; Kuss et al., 2001) or LREEs (F owler et al., 1992; Tachikawa et al., 1997, 1999), while some show an enrichment in middle REEs (MREEs) (Lerche and Nozaki, 1998; Alibo and Nozaki, 1999). In addition, Ce anomalies are both negative (Murphy and Dymond, 1984; Kuss et al., 2001) and positiv e (e.g., Sholkovitz et al., 1994; Tachikawa et al., 1997), and change with depth (Tachik awa et al., 1999) or particle flux (Fowler et al., 1992). The observed pattern variations may be caused by differences in (i) the shale values used for normalization, (ii) the ch emical methods used for leaching and/or digestion, and (iii) particle properties, such as composition, size, and sinking velocity. Because multiple parameters can influence compositional results, comparisons between different studies are difficult. In one of the first measurements of REE co ncentrations in sinking particles, Murphy and Dymond (1984) observed shale-normalized REE patterns similar to seawater (i.e., HREE enrichment and negative Ce anomaly). In sharp contrast, the shale-normalized REE patterns of Masuzawa and Koyama (1989) we re flat with positive Ce anomalies. Masuzawa and Koyama (1989) attributed their ob servations to the preferential removal of Ce, along with Mn, from seawater by settling particles. Although thei r patterns were quite different, Murphy and Dy mond (1984) and Masuzawa and Koyama (1989) both determined that the total R EE flux increases with depth, possibly due to ongoing sorption of REEs by the sinking particles. Using regressions between measured REE concentrations and percentages of end me mber components, Murphy and Dymond (1984) calculated that 30–60% of the REE flux co mes from the biogenic component, which is dominant in the upper water column and decreas es with depth. The remainder of the total flux was considered to be detrital, grad ually increasing with depth and becoming dominant in deeper water (Murphy and Dymond, 1984). Both of the above studies were performed during a single time period and could not be used to draw any conclusions about tempor al variations in particle fluxes. By using automated time-series sediment traps, Fowler et al. (1992) were able to observe changes in shale-normalized patterns of sinking partic les due to rapid variati ons in particle flux associated with the crash of a phytoplankton bloom. In the course of the Fowler et al.

PAGE 28

13 (1992) observations, REE patterns of particle s remained flat with negative Ce anomalies throughout the water column. In comparison, when particle fluxes were lower, REE patterns of particles were fl at in surface waters and developed LREE enrichments and positive Ce anomalies with depth. Fowler et al. (1992) concluded that slowly sinking particles are more effective at scavenging R EEs, especially LREEs and Ce, compared to fast sinking particles. This same effect of vertical flux on RE E sorption was observed by Tachikawa et al. (1997) in th eir comparison between a meso trophic site with large dust fluxes and an oligotrophic site with lo wer fluxes and greater LREE enrichments. Measuring REE concentrations in suspended particles, Bertram and Elderfield (1993) determined that < 5% of the total REE in s eawater is in particulate form, except for Ce, for which 20% of the total is particulate. Since the REE patterns of dissolved/ particulate ratios resembled the shale-norm alized pattern of seawater, Bertram and Elderfield (1993) suggested that particulate REEs are composed mainly of detrital matter. Sholkovitz et al. (1994) increased the unde rstanding of marine pa rticulate behavior by performing a series of three chemical di gestions on suspended particles. The first digestion used acetic acid to remove the su rface coatings, which consisted of organic matter and Mn oxides. The acetic acid dige st, containing 50–70% of the REEs and the main fraction of Mn, had a shale-normalized REE pattern that was the mirror-image of seawater (i.e., LREE enrichment and positive Ce anomaly). The second digestion, which involved strong acids, and the third digestion, which was carried out in a digestion bomb, both contained the major fraction of Al and Fe, indicating a detrital composition. The shale-normalized patterns for the strong acid an d bomb digests were relatively flat, which is consistent with a detrital source. Based on these three digesti ons, Sholkovitz et al. (1994) concluded that REEs are removed from seawater and fractionated by surface coatings, with Mn oxyhydroxides possibly cont rolling Ce oxidation and the preferential removal of LREEs compared to HREEs. Despite using a sequential digestion method similar to Sholkovitz et al. (1994), Lerche and Nozaki (1998) observed shale-norm alized patterns for sinking particles that had a MREE enrichment in all three digests. In addition, the Ce anomaly changed from negative in the acetic acid digest to unity in the acid leach to positive in the residual

PAGE 29

14 fraction. Lerche and Nozaki (1998) suggested that the adsorbed REEs are altered into a more refractory solid phase pr ior to or during particle si nking. The differences between the results of Sholkovitz et al. (1994) and Lerche and Nozaki (1998) indicate that the sorptive behavior of marine particulates is not yet fully understood. One of the major challenges in particulate REE studies is determining the association between the REEs and a specific carrier ph ase since particle composition can vary throughout the ocean. Several studies used observations of particulate Mn, Al, and Fe concentrations, which represent inorganic carr ier phases, to infer relationships with the REEs (Masuzawa and Koyama, 1989; Sholkovitz et al., 1994; Tachikawa et al., 1997, 1999). Kuss et al. (2001) extended this by also measuring concentrations of particulate organic carbon (POC), CaCO3, and opal, which are considered to be biogenic YREE carrier phases in suspended matter. Thr ough linear regression analyses of YREE concentrations versus carrier phase values Kuss et al. (2001) demonstrated strong fractionation between HREEs and LREEs for each particulate fraction. Y and the HREEs were strongly correlated with POC and Mn, i ndicating that organic lig ands form stronger complexes with increasing REE atomic number and that Mn may be associated with biogenic matter rather than oxyhydroxides. On the other hand, the LREEs were strongly correlated with Al and to a lesser extent Fe, which indicates a crustal or clay origin and some sorption onto an Fe-oxide phase. Since marine particles are composed of se veral different phases (e.g., Fe-Mn oxides, aluminosilicates, organic matter, etc.), surface complexation constants for discrete phases, which are similar to solution complexation constants, cannot be determined from the above-mentioned studies. In order to accu rately model YREE distributions in natural waters, a better understanding of YREE interactions with particle surfaces must be gained through observations of YREE associations with pure phases. To supplement the extensive database on YREE solution comple xation, an increasing number of studies have been devoted to examination of YREE surface chemistry. Investigations of YREE sorption have utilized a variety of pure substrat es, such as alumina (e.g., Fairhurst et al., 1995; Marmier et al., 1997; Rabung et al., 2 000), silica (e.g., Kosmulski, 1997; Takahashi et al., 1998; Marmier et al., 1999), rutile (e.g ., Ridley et al., 2005), sand (e.g., Tang and

PAGE 30

15 Johannesson, 2005), organics (e.g., Bingler et al., 1989; Stanley and Byrne, 1990), clays (e.g., Bradbury and Baeyens, 2002; Coppin et al., 2002; Rabung et al., 2005), manganese oxides (e.g., Bidoglio et al., 1992; De Carlo et al., 1998; Davranche et al., 2005), and iron oxides (e.g., Koeppenkastrop a nd De Carlo, 1992; Bau, 199 9; Ohta and Kawabe, 2001). Because iron is ubiquitous in natural waters and is the main substrate used in the present research, the following discussion will principally focus on YREE sorption by a variety of forms of particulate iron. Some of the first studies involving REE sorption by iron(III) hydroxide and oxide were interested in preconcentration an d decontamination of radionuclides. Musi et al. (1979) and Musi and Risti (1988) showed that REE so rption in low ionic strength solutions ( 0.15 M) increased as the pH increased from 4 to 7. The pH-adsorption edges for Ce, Gd, and Yb were very similar (Musi and Risti 1988), suggesting a lack of fractionation. Although they described the reactions involved in REE sorption, neither investigation modeled the process or related the results to seawater. Early investigations of REE sorption in seawater showed that LREEs are preferentially removed by most solids comp ared to HREEs, which leads to fractionation (Byrne and Kim, 1990; Koeppenkastrop et al., 1991; Koeppenkast rop and De Carlo, 1992). Despite the fact that surface comple xation appeared to be stronger for LREEs, differences among various solids were obse rved. Silica phases were unique in their sorption behavior because they displayed a greater affinity for HREEs, except when covered by a thin organic film, which enhan ced LREE sorption (Byrne and Kim, 1990). Koeppenkastrop and De Carlo (1992) observed that REE sorption by crystalline goethite was weaker than sorption by amorphous iron oxyhydroxide, but th at goethite created stronger fractionation between the LREEs and the HREEs. In addition, the residual seawater pattern from these experiments re sembled the shale-normalized REE pattern in the ocean (Koeppenkastrop and De Carlo, 1992). As was discussed earlier, se awater contains numerous solution ligands that compete with surfaces for free YREE ions. To determine YREE surface complexation constants, any effects from this strong solution comp lexation must be removed. YREE sorption onto amorphous ferric hydroxide in simple syntheti c solutions without complexing ligands has

PAGE 31

16 been investigated over a range of pH (3.5 – 9.0), ionic strength (0.0 – 0.7 M), and substrate loading (i.e., [YREE]/[Fe3+]T; 0.004 – 0.4) (De Carlo et al., 1998; Bau, 1999; Kawabe et al., 1999b; Ohta and Kawabe, 2001) These studies showed that in the absence of solution complexation, HREEs are pref erentially removed from solution but the magnitudes of estimated distribution coefficien ts for individual YREEs, at constant pH, varied by as much as a factor of 400. Similar to Musi et al. (1979) and Musi and Risti (1988), De Carlo et al. (1998) and Bau (1999) observed an increase in YREE sorp tion with increasing pH. In addition to increasing YREE sorption with increasing pH, Bau (1999) reported that the relative magnitudes of YREE sorption (i.e., fractio nation patterns) vary with pH. The pH dependence of YREE sorption can be de scribed via a surface complexation model (SCM), which provides a thermodynamic explanation for the competitive complexation of H+ and dissolved metal ions by surface hydrox yl groups (Schindler and Stumm, 1987; Dzombak and Morel, 1990). Several investigations of REE sorption have utilized a SCM to interpret the sorptive behavi or of individual REEs, such as Eu sorption onto hematite (Rabung et al., 1998a) and sorption of La and Yb by hematite and goethite (Marmier et al., 1997; Marmier and Fromage, 1999). As discussed by Rabung et al. (1998a), comparisons between these invest igations are difficult due to distinct differences in data interpretations along with SCM variations. By utilizing three different SCMs to interpret the same data, however, Marmier and Fromag e (1999) showed that the diffuse layer model and the constant capacitance model yielde d similar results and, at low loading, a non-coulombic SCM could satisfactorily model La sorption data. Since the SCM was applied to only a few individual REEs, vari ations in YREE sorption patterns with pH (Bau, 1999) were not addressed in these studies. In addition to modeling YREE sorption in terms of pH, the effect of solution complexation on YREE sorption needs to be included in SCMs. Stanley and Byrne (1990) examined REE sorption in seawater by Ulva lactuca L., a macroalga, over a range of carbonate concentrations. Th e variation in their calcul ated solution complexation intensity for Gd (i.e., ratio of free to total Gd ) was much greater than the variation in their observed Gd solid/solution distribution coe fficients. Stanley and Byrne (1990) suggested

PAGE 32

17 that this difference may be due to the sorp tion of complexed solution species such as 3GdCO, in addition to sorption of free dissolved metal (Gd3+), at high degrees of solution complexation. On the other hand Koeppenkastrop and De Carlo (1993) proposed that dissolved REEs dissociate from carbonate ligands before being sorbed as free ions onto a solid. Their conclusion wa s based on the observation that carbonate complexation slowed the rate of uptake of Eu by iron oxide in seawater. Kawabe et al. (1999a) and Ohta and Kawabe (2000) examined YREE sorption by amorphous ferric hydroxide in the absence and presence of carbonate (0 M 3[HCO] 12 mM). Their results showed that distribution coefficien ts increased along the YREE series in the absence of carbonate but this trend reversed when carbonate was present in solution. As was already stated, Ohta and Kawabe (2000) us ed their distribution coefficient results to calculate carbonate complexation constants, wh ich displayed a consistent pattern across the YREE series but were at least an order of magnitude greater than most literature data. Based on this review of YREE aqueous geoche mistry, the aim of my dissertation is to examine the influence of solution and surf ace chemistry on YREE sorption in aqueous solutions. My project has three main goals. The first goal is characterization of YREE sorption by the freshly precipitated hydroxide s of iron(III), aluminum, gallium, indium, and scandium. These experimental results will show whether different amorphous precipitates have similar or unique sorp tion patterns. Compar ison of distribution coefficient patterns for different solids with YREE solution complexation constant patterns allows assessment of the extent to which complexation properties of surface functional groups can be described in terms of solution complexation characteristics. The results of this assessment will be discussed in Chapter 2. The second goal of my project is characterization of the pH and ionic-st rength dependence of YREE sorption by amorphous ferric hydroxide. An SCM will be constructed from distribution coefficient measurements obtained over a wide range of conditions. The results will then be used for general predictions of environmental YREE be havior. This topic wi ll be discussed in Chapter 3. My third goal is characterizatio n of the influence of carbonate complexation on YREE sorption by amorphous ferric hydr oxide in simple aqueous solutions. Combining SCM parameters and previous characterizations of YREE carbonate

PAGE 33

18 complexation constants, predictions of YREE sorption in the presence of carbonate can be compared to experimental observations. An especially important issue here is whether YREE sorption behavior can be descri bed solely in terms of sorbed M3+ ions, or whether sorption of solution complexes such as 3MCO must also be considered. Chapter 4 will be devoted to the results of this objective. Finally, Chapter 5 will briefly describe observations from YREE sorption experime nts performed with amorphous ferric hydroxide over a range of temperatures. My studies will lead to a quantitative mo del of YREE sorption by an environmentally important sorptive substrate, amorphous ferri c hydroxide, and, in general, an improved quantitative model of YREE removal from seawater by marine particles. Since the processes that control YREE distributions in seawater are known to influence all metals in the ocean, this work will provide a better understanding of the general nature of metal cycling in the ocean.

PAGE 34

19 2. Comparative Scavenging of Yttrium and the Rare Earth Elements in Seawater: Competitive Influences of Solution and Surface Chemistry The following chapter has been peer-reviewed and published essentially in this form: Quinn K. A., Byrne R. H., and Schijf J. (2004) Aquatic Geochemistry 10, 59-80. 2.1 Abstract Distribution coefficients were obtained for yttrium and the rare earth elements (YREEs) in aqueous solutions containing fr eshly precipitated hydroxides of trivalent cations (Fe3+, Al3+, Ga3+, and In3+). Observed patterns of log iKS – where iKS = [MSi][M3+]-1[Si]-1, [MSi] is the concentration of a sorbed YREE, [M3+] is the concentration of a free hydrated YREE ion, and [Si] is the concentration of a sorptive solid substrate (Fe(III), Al, Ga In) – exhibited similarities to patterns of YREE solution complexation constants with hydroxide (OH1) and fluoride (F1), but also distinct differences. The log iKS pattern for YREE sorption on Al hydroxide precipi tates is very similar to the pattern of YREE hyd roxide stability constants (log OH1) in solution. Linear free-energy relationships between log iKS and log OH1 showed excellent correlation for YREE sorption on Al hydroxide precipitates, good correlation for YREE sorption on Ga or In hydroxide precipitates, yet poor co rrelation for YREE sorption on Fe(III) hydroxide precipitates. Whereas the correlation between log iKS and log F1 was generally poor, patterns of log (iKS/F1) displayed substantially incr eased smoothness compared to patterns of log iKS. This indicates that the conspicuous sequence of inflections along the YREE series in the patterns of log iKS and log F1 is very similar, particularly for In and Fe(III) hydroxide precipitates. While the log iKS patterns obtained with Fe(III) hydrox ide precipitates in this work are quite distinct from those obtained with Al, Ga, and In hydroxide precipitates, they are

PAGE 35

20 in good agreement with pa tterns of YREE sorption on ferr ic oxyhydroxide precipitates reported by others. Furthermore, our log iKS patterns for Fe(III) hydroxide precipitates bear a striking resemblance to predicted log iKS patterns for natura l surfaces that are based on YREE solution chemistry and shale-normalized YREE concentrations in seawater. Yttrium exhibits an itinerant behavior among the REEs: sorption of Y on Fe(III) hydroxide precipitates is intermediate to that of La and Ce, while for Al hydroxide precipitates Y sorption is similar to that of Eu. This behavior of Y can be rationalized from the propensities of di fferent YREEs for covalent vs. ionic interactions. The relatively high shale-normalized concentration of Y in seawater can be explained in terms of primarily covalent YREE interactions with scavenging particulate matter, whereby Y behaves as a light REE, and primarily ionic interactions with solution ligands, whereby Y behaves as a heavy REE. 2.2 Introduction One of the major objectives of chemical oceanography is to gain an understanding of the processes that control the concentrations and distributions of el ements in the oceans. Yttrium and the fourteen stable rare earth el ements are of unique value in this regard because of the coherence in their chemical properties. Chemical characteristics of the trivalent rare earth elements are sufficiently si milar that this coherence is intermediate to that of isotopes of a single element on the one hand and elements in the same group (column) of the Periodic Table on the other. On account of this, the oceanic abundances of rare earth elements, like isotopes, are ge nerally described in a comparative manner. In the same sense that the pattern of a fingerprint, rather than any quantitative aspect, is uniquely informative, abundance patterns of y ttrium and the rare earth elements (YREEs) provide sensitive measures of environmental processes. Although in all seawater below a certain depth the YREEs are near their solubility limits with respect to mixed-YREE phospha te precipitates (Byrne and Kim, 1993; Liu and Byrne, 1997; Liu et al., 1997), it is generally thought that scavenging processes – sorptive removal of elements from the wate r column by sinking particles (Goldberg,

PAGE 36

21 1954; Goldberg et al., 1963; Balistrieri et al., 1981; Bruland, 1983) play a dominant role in shaping both the absolute and relative abundances of YREEs in the oceans. In previous work (Byrne and Kim, 1990), the single-box residence time model of Schindler (1975) was used to derive a simple equatio n that describes relative YREE abundances in seawater: onstantc ])M[M( ])M[M( log )dtdA( A log3 S 3 T M M (2.1) where brackets [ ] denote the solution concentration of a chemical species, AM is the total amount (mol) of metal M in seawater and dAM/dt is its oceanic input (or removal) rate at steady state. The ratio AM/(dAM/dt) on the left hand side of equation (2.1) constitutes the equivalent of shale-normalized YREE concentr ations in seawater or, alternatively, the residence times of individual YREEs with respect to their total oceanic input (or removal). It is very important to note that the two sides of equation (2.1) are offset by a constant, independent of M, that incorporat es several poorly known quantities related to the surface characteristics, c oncentrations, and residence times of sorbing particles (Byrne and Kim, 1990). On the right hand side of equation (2.1), the numerator denotes the extent (intensity) of solution complexation for metal M, where MT represents the total dissolved metal concentration. The denominator denotes the a ffinities of particle surfaces for free metal ions M3+ in solution, where MS represents the total sorbed metal concentration. Thus, equation (2.1) predicts that shale-normalized YREE concentrations in seawater can be described as a direct competition between solution ligands and surface ligands for free hydrated metal ions. The solution complexation term, MT/[M3+], in equation (2.1) can be written as a summation over the contributions from all YREE solution complexes in seawater: )]L[1( ]M[ Mn in n,i i 3 T, (2.2) where in are stability constants of the nth complex of metal M with solution ligand Li: n i 3 ni ni]L][M[ ])L(M[. (2.3)

PAGE 37

22 Stability constants appropriate to seawater and freshwater have been determined for YREE complexation with fluoride (Schijf and Byrne, 1999; Luo and Byrne, 2000), chloride (Luo and Byrne, 2001), sulfate (Sch ijf and Byrne, 2004), hydroxide (Klungness and Byrne, 2000), and bicarbonat e and carbonate (Liu and By rne, 1998; Luo and Byrne, 2004). The surface complexation term (MS/[M3+]) in equation (2.1) can be written as a summation over different types of surface complexation sites: ]) S [ K ( ] M [ Mi i S i 3 S (2.4) where iKS is the affinity of surface ligand Si for element M: ] S ][ M [ ] MS [ Ki 3 i S i (2.5) The concentration [Si] of surface ligands in a solution, na tural or synthetic, is expressed in the same units as solution ligands but, of course, denotes moles of particulate ligands per kilogram of solution. The quantity iKS can be interpreted (as it will be here) as a distribution coefficient, where the ratio [MSi]/[Si] is calculated as the sorbed metal concentration per mole of solid substrate (s ince the exact type and density of surface ligands are unknown). We will occasionally disti nguish these distribution coefficients by replacing the subscript S with the chemical symbol of the trivalent cation in the hydroxide precipitate (e.g., iKGa for YREE sorption on Ga hydroxide). The properties of iKS relevant to natural marine surfaces are much less well understood than the properties of YREE co mplexation by solution ligands. Most prior work on YREE scavenging (Murphy and Dym ond, 1984; Koeppenkastrop et al., 1991; Fowler et al., 1992; Koeppenkastrop and De Carlo, 1993; Sholkovitz et al., 1993; Schijf et al., 1994; Bau et al., 1996; Lerche and Nozaki, 1998) involved measurements of YREE distributions between seawater and particulate matter and, as such, did not provide true estimates of iKS as defined in equation (2.5). Direct measurements of iKS are becoming increasingly common (Bau, 1999; Oh ta and Kawabe, 2000, 2001). However, measurements of iKS with a variety of particle surfaces are sufficiently scarce that the quality of such data is uncertain. Thus, additional iKS data are valuable for assessing the

PAGE 38

23 quality of previous iKS characterizations and extending such characterizations to different types of surfaces. Two indirect methods have been used to m odel the sorptive characteristics of natural marine particles. In the first case, linear free-energy relationships of the form b log a K g lo1 S iOH (2.6) where a and b are constants and OH1 is the YREE hydroxide st ability constant in solution, have been used to estimate iKS behavior based on YREE solution complexation behavior (Schindler, 1975; Ba listrieri et al., 1981; Erel and Morgan, 1991). As another means of evaluating the YREE sorptive character istics of natural marine particles, Lee and Byrne (1993) and Byrne and Sholkovitz (1996) used equation (2.1) to estimate average iKS values for marine particles based on (i) high-precision shale-normalized YREE concentrations in seawat er and (ii) the best soluti on complexation constant data available at the time. Figure 2.1 shows a co mparison of the Byrne and Sholkovitz (1996) iKS estimates and direct iKFe measurements obtained by Bau (1999) for YREE sorption on ferric oxyhydroxides at pH ~ 6. The patterns shown in Fi gure 2.1 are quite distinct from the pattern of log OH1 (Klungness and Byrne, 2000) and exhibit an intriguing similarity that has not been previously noted. In this work, we compare the YREE sorptive characteristics of a number of freshly precipitated trivalent metal hydroxides at pH ~ 6, namely Fe(OH)3, Al(OH)3, Ga(OH)3, In(OH)3, and Sc(OH)3. The resulting iKS data are discussed us ing the two approaches described above: linear free-energy relationships with OH1, and comparison with iKS estimates derived from observations of sh ale-normalized YREE concentrations plus independently measured YREE solution co mplexation constants. In addition, iKS data are discussed in terms of relative YREE ionicities with respect to solution and surface complexation (Martell and Hancock, 1996). 2.3 Materials and Methods Experiments were performed to study YREE sorption on a number of trivalent metal hydroxides, which were precipitated directly from an acidic solution containing the

PAGE 39

24 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKS 7.0 7.5 8.0 8.5 9.0 9.5 surface water deep water YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKFe 0.5 1.0 1.5 2.0 pH 5.97 pH 5.87 pH 5.87 A B Figure 2.1. Comparison between (A) predicted log iKS values (Byrne and Sholkovitz, 1996) and (B) directly measured log iKFe values using ferric oxyhydroxides (Bau, 1999). The distribution coefficient iKFe is equivalent to the quantity appDREY, defined by Bau (1999). trivalent metal and all YREEs by increasing the pH with ammoni a. Five separate experiments were performed. Solid substrat es included Fe(III) hydr oxide (a known strong YREE scavenger), hydroxides of the trival ent Group 13 elements Al, Ga, and In, and hydroxides of the YREE-like element Sc. All chemical manipulations were performed inside a class 100 cl ean air laboratory or laminar flow bench. Trace metal-clean wa ter (Milli-Q water) was produced with a Millipore (Bedford, MA) purification syst em. Teflon and polypropylene laboratory materials and polycarbonate filter membrane s were cleaned by soaking in HCl or HNO3

PAGE 40

25 for at least one week, followed by several th orough rinses with Milli -Q water. Certified 1.000 M hydrochloric acid was purchased from Sigma-Aldrich (St. Louis, MO), and TraceMetal Grade nitric acid fr om Fisher Scientific (Pitt sburgh, PA). A YREE stock solution, containing 66.7 ppm of each YREE in 2% HNO3, was prepared from single element ICP standards (SPEX CertiPrep, Metu chen, NJ). Salts of the trivalent metals were purchased from Sigma-Aldrich, except aluminum, for which an existing laboratory ICP standard was used (10,000 ppm in 10% HCl), and iron, which was purchased from Fisher Scientific as a ferric chloride solu tion (40% w/v in HCl). Indium was obtained as nitrate pentahydrate and scandium as chlori de hexahydrate. Anhydrous gallium chloride was shipped in an ampoule under argon. All salt s were of 99.999% pur ity with respect to metal content and were used as received. S candium chloride was added directly to the experimental solution. The other salts were di ssolved in a small quantity of Milli-Q water (gallium chloride under an argon atmosphere) and then diluted to a known volume with Milli-Q water, adding concentrated HCl or HNO3 as necessary, to make concentrated stock solutions of the trivalent metals. At the beginning of each experiment, a pH standard (pH 3.0, 0.001 M HCl) was prepared in 0.01 M NH4NO3 to match the ionic strength, I, of the experimental solution, which was calculated as I = 0.014 0.002 M. The experimental solution consisted of 100 M of the trivalent metal and 23.3 ppb of each YREE (total combined YREE concentration 2.36 M) in 0.01 M HCl, HNO3, or NH4NO3. Both solutions were equilibrated in a Teflon wide-m outh bottle inside a jacketed beaker thermostated at T = (25.0 0.1)C, and continuously stirred with a Teflon-coated ‘floating’ stir bar. The experimental solution was continuously bubbled with ultra-pure N2, which had been passed through an in-line trap (Supelco, Bellafonte, PA) th at removed all traces of CO2. The pH of the experimental solution was expressed on th e free hydrogen ion concentration scale. It was monitored regu larly by comparison with the pH standard, using a Ross-type combination pH electr ode (No. 810200) connected to a Corning 130 pH meter in the absolute millivolt mode. Linearity and Nernstian behavior of the electrode were verified by titrating a 0.3 M NaCl solution with concentrated HCl.

PAGE 41

26 After an equilibration period of at most 24 hours, a sample was taken at the initial solution conditions (pH 2, no YREE sorp tion) to determine the total YREE concentrations, MT, for calculation of the distribution coefficients, iKS. Precipitation of hydroxides was subsequently initiated by raising the pH of the experimental solution to about 6.0 with careful additions of 1 M NH4OH from a Gilmont micro-dispenser. The onset of precipitation was often accompanied by the appearance of finely dispersed solids in the solution and by slow fluctuations in the pH. Once a pH of about 6.0 was established, samples were taken with a pipette at fixed time interv als, increasing from minutes to hours to days. Two separate samp les were taken each time. One sample was filtered from a polypropylene syringe through a Nuclepore filter membrane (polycarbonate, 0.10 m pore size) mounted in a polypropylene filter holder. Five mL were used to rinse the syringe and membra ne, and discarded. The next 5 mL were collected in a pol ypropylene centrifuge tube. The second sample was centrifuged at about 4,000 rpm for one hour in a Centra-4B centr ifuge (International Equipment Company, Needham Heights, MA). In every experi ment, except the one with Fe(III), the concentration of the trivalent metal was increased to 1 mM after the first set of samples had been withdrawn. The pH wa s then readjusted to 6.0 to induce further precipitation of hydroxides and a second set of sample s was collected. Experiments with 100 M In, Al, and Ga, did not always produce a visibl e precipitate, although YREE sorption was observed. The experiments with 1 mM of these metals resulted in stronger YREE sorption and more precise iKS data. Higher concentrations of Fe(III) were not required because log iKFe is substantially larger than log iKS values for In, Al, and Ga. As a consequence, YREE sorption can be observed at lower substrate concentrations using Fe than is the case for In, Al, and Ga. The filtered samples and the supernatant of the centrifuged samples were diluted 5fold with 1% HNO3, and a small amount of internal st andard solution containing equal amounts of In, Cs, and Re was added. For the In experiment, the In of the internal standard was replaced with Rh. The resulting mixtures were analyzed for YREE with an Agilent Technologies 4500 Seri es 200 inductively-coupled pl asma mass spectrometer (ICP-MS). Solutions were in troduced into the ICP-MS w ith a Babington-type PEEK

PAGE 42

27 nebulizer and a double-pass (Scott-type) quartz spraychamber, Peltier-cooled to T = 2C. During instrument tuning, the formati on of oxide and doublecharged ions was minimized with a 10 ppb Ce solution. MO+ and M2+ peaks were always less than 1% and 3% of the corresponding M+ peak, respectively, and correction for this effect proved unnecessary. YREE concentrations were calcul ated from linear re gressions of four standards (0.5, 1, 2, and 5 ppb). A 1% HNO3 solution was run before and after the calibration line, to serve as a blank and to ri nse the instrument after the highest standard. In addition, after each autosampler position, Milli-Q water was aspirated for 10 s followed by a 1% HNO3 wash solution for 30 s, to rinse the outside of the autosampler probe and the sample introduction system. All standards and solutions were injected in triplicate. Ion count s were corrected for minor instrument drift by normalizing 89Y to 115In (or 103Rh), 139La–161Dy to 133Cs, and 163Dy–175Lu to 187Re. The HP ChemStation software does not allow a mass-dependent co rrection by interpolation between internal standards, yet a constant check on the valid ity of the drift correction was performed by comparing the Dy concentrations calculated from 161Dy and 163Dy, which were usually equal to within 2%. Blanks were generally below the instrument quantitation limit (0.01 ppb). Scandium concentrations were measured in samples from the Sc experiment, after an additional 10or 100-fold dilution, to verify that precipitation had occurred, because no visible solids were observed at any time a nd no YREE sorption could be detected. Since a suitable internal standard for Sc had not been added to the samples, Sc concentrations were determined with a semi -quantitative ICP-MS method, using a solution containing 100 ppb each of Be, Mg, Co, In, Bi, and U as a re ference. To correct for any difference in matrix between the samples and the referen ce, the sample respons e curve was normalized to the concentrations of the internal standards (0.25 or 2.5 ppb In, Cs, and Re). The accuracy of this semi-quantitative method is approximately 10%. Distribution coefficients were calculated using equation (2.5) and [M3+] = MT, which is a valid approximation, since the concen tration of solution species other than [M3+] was generally negligible ([MOH2+]/[M3+] 0.03; [M3+]/MT 0.97) for our experimental conditions (pH ~ 6). The concentration of surface ligands [Si] was set equal to the molar

PAGE 43

28 concentration of precipitated metal (100 M for Fe(III) and 1 mM for Al, Ga, and In). The concentration of sorbed YREE, [MSi], was calculated as the difference of the total initial concentration of each YREE measured at pH 2 and the concentration of each YREE in the filtrates at pH ~ 6.0. 2.4 Results and Discussion 2.4.1 Comparative log iKS Results The iKS results obtained with freshly precipi tated Fe(III), Al, Ga, and In hydroxides are shown in Figure 2.2. In the Sc experiments, no YREE sorption could be detected at all, even though measurement of the diss olved Sc concentration confirmed that precipitation of Sc hydroxides was complete within a few hours. YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKFe 3.2 3.4 3.6 3.8 4.0 4.2 4.4 15 min; pH 612 90 min; pH 616 5 hr; pH 617 24 hr; pH 610 46 hr; pH 607 48 hr; pH 607 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKAl 0.5 1.0 1.5 2.0 2.5 3.0 3.5 15 min; pH 606 44 hr; pH 581 68 hr; pH 570 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKIn 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 15 min; pH 613 90 min; pH 611 5 hr; pH 612 24 hr; pH 607 96 hr; pH 602 100 hr; pH 605 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKGa 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 15 min; pH 634 90 min; pH 635 5 hr; pH 633 24 hr; pH 618 44 hr; pH 604 141 hr; pH 547 Ga In Al Fe B A D C Figure 2.2. log iKS results from filtered samples. (A) 100 M Fe(OH)3 (Table B.2). (B) 1 mM In(OH)3 (Table A.3). (C) 1 mM Al(OH)3 (Table A.1). (D) 1 mM Ga(OH)3 (Table A.2).

PAGE 44

29 Distribution coefficient patte rns are highly consistent ove r periods of time between 15 minutes and as much as 141 hours. Solution pH in the Fe(III) and In experiments varied within a range of 0. 1 pH units and only small ch anges were observed in the absolute magnitudes of log iKS. In contrast, pH varied over a range of approximately 0.36 in the Al experiment and 0.88 in th e Ga experiment, and large changes in log iKS were observed. A plot of log iKS vs. pH for the Al experiment indicated that the influence of pH on log iKS was well described by the equation: log iKS(A) = log iKS(B) + Q (pHA – pHB). (2.7) where Q is a constant. For comparison with the Fe(III) and In expe riments, observations of log iKS(B) and pHB from the Al and Ga experiments were used to predict log iKS(A) values corresponding to pHA = 6.10. This was done with the Microsoft Excel routine Solver by changing Q to minimize the following sum of squares: 2 S i B S i)} A ( K g lo ) pH 10 6 ( Q ) B ( K {log (2.8) where S iK g lo indicates an average and the summation is over all patterns and all elements (YREEs). For Ga, the pattern at lo w pH (5.47) was discordant and therefore excluded. Best agreements were obtained with Q = 1.66 for Al and Q = 2.79 for Ga (Figure 2.3). Figure 2.3B shows a range in log iKGa on the order of 0.2. Table 2.1 provides averaged log iKS results for the data shown in Figures 2.2A,B and 2.3. The nearly constant magnitudes of log iKFe and log iKIn, as well as log iKAl corrected to constant pH, suggest that transformation from amorphous to more crystalline solids was a very slow process. Larger variations in log iKGa could be indicative of significant changes in the crystallinity or hydration state of the Ga(OH)3. However, since the log iKGa variations shown in Figure 2.3B are sm all with no consistent trend through time (uppermost to lowest log iKGa results were obtained at 90 minutes, 44 hours, 15 minutes, 5 hours, and 24 hours), it is possi ble that the observed variability is caus ed by a variety of factors.

PAGE 45

30 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKAl 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 A Al YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKGa 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 B Ga Figure 2.3. log iKS results normalized to pH 6.10. (A) 1 mM Al(OH)3. (B) 1 mM Ga(OH)3. See text for details.

PAGE 46

31Table 2.1. Average log iKS results for iron, aluminum, gallium, and indium (Figures 2.2A, 2.3A, 2.3B, and 2.2B, respectively). Results for Fe(III) and In did not vary significantly over the small rang e of experimental pH and we re averaged. Results for Al and Ga showed a linear dependence on pH an d were averaged after normalizing each measurement to pH 6.1 (see text). Uncertainties in pH and log iKS values represent one standard deviation of the mean. pH 6.12 0.04 pH 6.10 pH 6.10 pH 6.08 0.04 [M3+] log iKFe log iKAl log iKGa log iKIn Y 3.66 0.09 2.71 0.03 2.68 0.07 2.21 0.01 La 3.45 0.02 1.70 0.07 2.00 0.13 1.57 0.05 Ce 3.89 0.01 2.02 0.04 2.41 0.09 1.92 0.04 Pr 4.06 0.02 2.13 0.07 2.63 0.08 1.96 0.04 Nd 4.13 0.02 2.26 0.03 2.75 0.07 2.01 0.04 Pm Sm 4.30 0.02 2.64 0.02 3.02 0.07 2.30 0.02 Eu 4.26 0.02 2.72 0.02 3.05 0.07 2.37 0.02 Gd 4.06 0.02 2.68 0.01 2.95 0.07 2.28 0.02 Tb 4.13 0.03 2.86 0.02 3.06 0.07 2.50 0.02 Dy 4.13 0.03 2.95 0.03 3.08 0.07 2.56 0.01 Ho 4.06 0.03 2.97 0.03 3.06 0.07 2.52 0.01 Er 4.09 0.03 3.05 0.04 3.12 0.07 2.59 0.01 Tm 4.19 0.02 3.19 0.04 3.21 0.07 2.79 0.01 Yb 4.29 0.02 3.35 0.04 3.33 0.07 3.05 0.01 Lu 4.24 0.02 3.36 0.04 3.32 0.07 3.03 0.01 2.4.2. Linear Free-energy Relationships In previous work, the sorptive characteristic s of surfaces that coordinate with trace metals via O-donor groups have been assessed and modeled in terms of the characteristics of trace metal hydrolysis behavi or (Huang and Stumm, 1973; Balistrieri et al., 1981; Schindler and Stumm, 1987). Applyi ng this approach, sorption constants for a variety of trace metals have been estimated using equation (2.6) (Dzombak and Morel, 1990; Erel and Morgan, 1991) or simplified ve rsions of equation (2.6) wherein b = 0 (Erel and Stolper, 1993). Equation (2.6) indicates that log iKS patterns should resemble log OH1 patterns that are stretched or compressed with the factor “a”. The ability of

PAGE 47

32 equation (2.6) to estimate YREE sorptive charac teristics can be directly evaluated using the log iKS results in Table 2.1 a nd YREE hydroxide stability c onstants of Klungness and Byrne (2000). The log OH1 data shown in Figure 2.4 were selected for this analysis, because they are based on very coherent re sults obtained with both spectrophotometric and potentiometric techniques. Problems with older values in the literature were discussed by Klungness and Byrne (2000). The regressions of log iKS (Table 2.1) vs. log OH1 (Figure 2.4), shown in Figure 2.5, re veal that the YREE sorptive behaviors of Al, Ga, and In hydroxide precip itates are very well modeled in terms of YREE hydrolysis behavior. The log iKFe data, however, are poorly descri bed in terms of YREE hydrolysis. Thus, the comparative affinities of YREEs toward at least one abundant, naturally occurring substrate is poorly predicted usi ng the linear free-energy approach of equation (2.6). YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log 1 3.0 3.5 4.0 5.0 5.5 6.0 6.5 7.0 OH F Figure 2.4. Stability constants for YREE comp lexes with hydroxide (Klungness and Byrne, 2000) and fluoride (Luo and Byrne, 2000), for the conditions of our experiments (T = 25C; I = 0.014 M). Horizontal dotted lines were drawn through Y to emphasize its position with respect to the REE.

PAGE 48

33 The results in Figure 2.5 suggest that relationships between YREE solution chemistries and surface chemistries can vary si gnificantly in complexity. An appropriate model for YREE sorption on Fe(III) hydroxides must include factors that are not tightly coupled to YREE hydrolysis. As we will demons trate in Section 2.4.3, some aspects of the log iKS behavior shown in Figure 2.2 more clos ely resemble fluoride stability constant data (Schijf and Byrne, 1999; Luo and Byrn e, 2000), than the hydrolysis data (Figure 2.4). This is particularly true for heavy rare earth element (HREE) sorption on In and Fe(III) hydroxides. log OH1 4850525456586062646668 log iKFe 32 34 36 38 40 42 44 46 Ay = (0.370.11)x + (1.8.7) r2 = 0.478 Fe log OH1 4850525456586062646668 log iKAl 10 15 20 25 30 35 C Aly = (1.110.03)x (4.0.2) r2 = 0.991 log OH1 4850525456586062646668 log iKIn 12 14 16 18 20 22 24 26 28 30 32 Iny = (0.910.06)x (3.10.4) r2 = 0.949 B log OH1 4850525456586062646668 log iKGa 18 20 22 24 26 28 30 32 34 36 y = (0.790.06)x (1.90.4) r2 = 0.919 D Ga Figure 2.5. Linear free-energy relati onships between the log iKS results from this work and log OH1 data from Klungness and Byrne (2000). (A) 100 M Fe(OH)3. (B) 1 mM In(OH)3. (C) 1 mM Al(OH)3. (D) 1 mM Ga(OH)3. Dotted lines repres ent 95% confidence intervals.

PAGE 49

342.4.3. Inter-element Patterns in YREE Solution Complexation and Surface Complexation Patterns of YREE solution complexation data can be qualitatively described in terms of at least two components. One component is the overall intensity of complexation. Ligands such as hydroxide exhibit large complexation constants (log OH1(Gd) ~ 6) and, consequently, there is often a similarly large ch ange in stability constants for such ligands across the YREE series ( OH1 ~ 1.5). Another example is NTA (log NTA1(Gd) ~ 13), with a range of stability constants between La and Lu that spans more than two orders of magnitude (Li and Byrne, 1997). A second co mponent of YREE equilibrium constant behavior is embodied in the complex inter-e lement sequence of YR EE fluoride stability constants in solution (Schijf and Byrne, 1999; Luo and Byrne, 2000). This sequence, easily visible in Figure 2.4 (lower pattern ), is less discernable for YREE hydroxide stability constants (Figure 2. 4, upper pattern). YREE fluori de stability constants are relatively small (log F1(Gd) ~ 3) and increase relativel y little across the YREE series ( F1 ~ 0.6) hence the inflecti ons are more prominent. The surface complexation constant patterns seen in Figures 2.2 and 2.3 can be compared and contrasted with patterns of solution complexation constants. As is the case for complexation constants (log in) appropriate to a variet y of solution ligands (Li), the relative magnitudes of log iKS data between La and Lu vary substantially for different types of surfaces. The difference between log iKS values for La and Lu ( log iKS) is as small as 0.8 for Fe(III) hydroxide precipitates and as large as 1.7 in the case of Al hydroxide precipitates. In contrast to the general correspondence between log in and log in observed for solution complexation, log iKS is not tightly coupled to the overall magnitudes of the log iKS data for each solid substrate. In the case of Fe(III), log iKFe(Gd) ~ 4.0 and log iKFe ~ 0.8, while for Al, log iKAl(Gd) ~ 2.7 and log iKAl ~ 1.7. Also, in contrast to the general case for solution complexation, the sequence of inflections in log iKS data is clearly dis cernable not only when log iKS is small (Figure 2.2A) but also, in some cases (Figure 2.2B), when log iKS is comparatively large ( log iKIn ~ 1.5). It appears that some generalities appropriate to comparative YREE solution complexation behavior are not applicable to comparative observations of log iKS.

PAGE 50

35 The linear free-energy relationships shown in Figure 2.5 demonstrate that, in some cases, YREE surface chemistries are tightly correlated with YREE hydrolysis behavior. As a means of highlighting certain similarities in the characteristics of log iKS data (Figures 2.2A,B and 2.3) and log F1 characteristics (Figure 2.4), Figure 2.6 shows iKS/F1 ratios plotted against YREE identity (atomi c number). The relatively smooth plots of log (iKS/F1) vs. YREE atomic number in Figure 2.6 lack the patterns-of-four or tetrad effects (Monecke et al., 2002, and refe rences therein) that are seen in Figures 2.2.4. They are absent in Figure 2.6 because tetrad effects are present to a similar extent in the surface complexation term, log iKS, and the solution complexation term, log F1. A measure of the smoothness of the pattern s in Figures 2.2, 2.3, and 2.6 can be obtained by calculating the anomalies of the last three triads (EuGdTb, DyHoEr, YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log (iKAl/F1) -18 -16 -14 -12 -10 -08 -06 -04 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log (iKFe/F1) -06 -04 -02 00 02 04 06 08 A Fe Al C YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log (iKIn/F1) -22 -20 -18 -16 -14 -12 -10 -08 B In YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log (iKGa/F1) -17 -15 -13 -11 -09 -07 -05 Ga D Figure 2.6. Ratio of iKS (this work) and F1 (Luo and Byrne, 2000), shown as log (iKS/F1). (A) 100 M Fe(OH)3. (B) 1 mM In(OH)3. (C) 1 mM Al(OH)3. (D) 1 mM Ga(OH)3.

PAGE 51

36 and Tm–Yb–Lu), where the anomaly of the triad Eu–Gd–Tb is defined as log [2 iKS(Gd)/(iKS(Eu) + iKS(Tb))] and so forth. A “sm oothness index” can then be determined by summing the squares of these three anomalies for each pattern, with a smaller index signifying a smoother pattern. Smoothness indices for the iKS patterns in Figures 2.2A,B and 2.3 d ecrease in the order InFe>Al Ga. Smoothness indices for the F1-normalized patterns in Figure 2.6 are 3-8 tim es smaller than the indices calculated for Figures 2.2A,B and 2.3. This demonstrates that iKS patterns for all four solid substrates share features of the log F1 pattern (for elements heavier than Sm), but more so for In and Fe(III) than for Al and Ga. By combining equations (2.3) and (2.5) it can be shown that the ratio log (iKS/F1) is equal, within a constant, to log ([MSi]/[MF2+]). In other words, the patterns in Figure 2.6 are identical to the iKS patterns that would be observed if MF2+ were the dominant YREE solution species. Since free YREE ions are rarely the dominant species in natural waters, this analysis implies that the extent to which tetrad effects are observable in the aqueous environment can be diminished, or even entirely obscured, when dissolved YREE concentrations are controlled by opposing sorption/complexati on equilibria. It is likely that the inflections that are easily visible in Figures 2.2 and 2.3 are obscured in the aqueous environment when dissolved YREE con centrations are controlled by competitive solution and surface complexation. 2.4.4. Oceanic log iKS Patterns YREE concentrations in seawat er have evolved at least in large part through opposing equilibria in which partic le surfaces coordinate M3+ ions (surface complexation) and solution ligands (e.g., 2 3CO) strongly associate with M3+ ions (solution complexation). Equations (2.1) through (2.5 ) prescribe the use of observed shale-normalized YREE concentrations and characterizations of YR EE solution chemistry to model the average YREE surface complexation characteristics (i.e., log iKS characteristics) of marine particles. Byrne and Sholkovitz (1996) mo deled the surface chemistries of marine particulate matter using solution complexation da ta that involved direct measurements of

PAGE 52

37 carbonate complexation for five elements (C e, Eu, Gd, Tb, Yb) and linear free-energy relationships to model the remaining elements. It is currently possible to obtain calculations of YREE solu tion complexation using directly measured carbonate, hydroxide, sulfate, chloride, and fluoride asso ciation constants for all YREEs (Liu and Byrne, 1998; Schijf and Byrne, 1999, 2004; Klungness and Byrne, 2000; Luo and Byrne, 2000, 2001, 2004). Figure 2.7C shows calculated log iKS results that we re obtained using (i) the shale-normalized YR EE distributions obtained by Zhang and Nozaki (1996) (Figure 2.7A) and (ii) the above-referenced solution complexation constants for all YREEs (Figure 2.7B). Very similar results are obtained using published seawater YREE patterns for different oceanic regions. The log iKS results displayed in Figure 2.7C show a strong rese mblance to the distribution coefficient data obtained for Fe (III) (Figure 2.2A). This is very clearly demonstrated in Figure 2.8 where the seaw ater patterns for the three depths are individually compared with the log iKFe pattern of Figure 2.2A. A constant, calculated with the same sum-of-squares optimization technique described above (equation (2.8)), was added to each seawater pattern to maximi ze the degree of overlap. The agreement is most striking for the deepest sample (Figure 2.8C). These results indicate that the YREE sorption characteristics of natural marine pa rticles (corrected for solution complexation) are remarkably similar to those of Fe(III) hydroxides. It can also be seen that the correspondence between Figures 2.7C and 2.2A is better, in some respects, than the relationship between Figure 2.7C and the da ta of Bau (1999) (Figure 2.1B). The log iKFe data reported by Bau (1999) for the middle rare earth elements (MREEs i.e., Sm and Eu) are larger than distribution coefficients obt ained for the HREEs. In contrast, in Figures 2.1A, 2.2A, and 2.7C, the log iKS data for the MREEs are quite similar to the log iKS data for Yb and Lu. This is discussed further in Sections 2.4.5 and 2.4.7.

PAGE 53

38 log (REE/PAAS) -9.0 -8.5 -8.0 -7.5 -7.0 -6.5 -6.0 3936 m 689 m 46 m log (MT/[M3+]) 0.5 1.0 1.5 2.0 2.5 3.0 pH 8.2 pH 7.9 pH 7.6 A B YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKS 7.7 8.1 8.5 8.9 9.3 9.7 46 m 689 m 3936 m C Figure 2.7. (A) Directly measured YREE concentratio ns in seawater at three depths (Zhang and Nozaki, 1996), normalized to Po st-Archean Australian Shale (PAAS) (McLennan, 1989). (B) The calculated solution complexati on term from equation (2.2) at pH 7.6, 7.9, and 8.2 with a bicarbonate concentration of 210-3 M. (C) Predicted log iKS values, calculated by subtracti ng each of the curves in Figur e 2.7A from the curve for pH 7.9 in Figure 2.7B.

PAGE 54

39 log iKS 3.1 3.3 3.5 3.7 3.9 4.1 4.3 4.5 Fe 46 m log iKS 3.1 3.3 3.5 3.7 3.9 4.1 4.3 4.5 Fe 689 m A B YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKS 3.1 3.3 3.5 3.7 3.9 4.1 4.3 4.5 Fe 3936 m C Figure 2.8. Comparison between the log iKFe values from this work and log iKS values for the three seawater samples from Figure 2. 7C. A different constant was added to each of the seawater patterns. These constants we re optimized to attain maximum overlap for each pair of patterns (see text).

PAGE 55

40 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKFe 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 this work Bau (1999) YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKFe 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 this work Ohta and Kawabe (2000, 2001) A B Figure 2.9. Comparison between the log iKFe values from this work at pH 6.12, Bau (1999) at pH 5.97 (A), and Ohta and Kawabe (2000, 2001) at pH 6.01 (B). A different constant was added to the results of Bau ( 1999) and to those of Ohta and Kawabe (2000, 2001). These constants were optimized to at tain maximum overlap for each pair of patterns (see text). 2.4.5. Comparative log iKFe Data Obtained for Freshly Precipitated Fe(III) Hydroxides Figure 2.9 shows direct comparisons between the log iKFe results obtained in this work (Table 2.1), and the results of Bau (1999) and Ohta and Kawabe (2000, 2001). The log iKFe results shown in Figure 2.9 were collected between pH 5.9 and 6.1. The results obtained in the present work are in particular ly good agreement with the data of Ohta and Kawabe (2000, 2001). While noting that the re sults for all three studies have been brought into agreement with the same optimiza tion technique used for the comparisons in

PAGE 56

41 Figure 2.8, it is seen that our results for the MREEs are closer to those of Ohta and Kawabe (2000, 2001) than to the results of Bau (1999). In addition to the contrasting log iKFe characteristics shown in Figure 2.9, it should be noted that the log iKFe pattern obtained by Bau (1999) is substantially pH dependent. At pH values below about 4.5, the log iKFe patterns observed by Bau (1999) were relatively flat and positive Ce anomalies were observed. The appearance of Ce anomalies is in itself quite counterintuitive. It is we ll established that abiotic Ce oxidation occurs primarily in the presence of Mn oxide surfaces and Ce anomalies are generally not observed for YREE sorption onto ferric oxyhydroxides (Koe ppenkastrop and De Carlo, 1992; De Carlo et al., 1998; Ohta and Kawabe, 2001). Moreover, de Baar et al. (1988) and De Carlo et al. (1998) have both argued that abiotic Ce oxidation should be strongly suppressed at low pH. Ohta and Kawabe ( 2000, 2001) observed no changes in their patterns of log iKFe data over a pH range between approximately 5.6 and 6.6. The very small Ce anomalies in their patterns show no systematic behavior as a function of pH and may be due to a minor analytical artifact. 2.4.6. Comparative log iKS Behavior of Yttrium and the Rare Earth Elements The position of Y among the REEs is an im portant aspect of the data summarized in Table 2.1. The value of iKFe for Y falls between La and Ce (Figure 2.2A). In the remaining experiments, Y is positioned between Pr and Nd (Figure 2.3B, Ga), between Nd and Sm (Figure 2.2B, In), and near Eu (Fig ure 2.3A, Al). This it inerant behavior of Y among the REEs has been ascribed to the en hanced covalency of REEs relative to Y (Siekierski, 1981), and may involve deloca lization of electrons in lanthanide 4f orbitals (Borkowski and Siekierski, 1992). In ionic interactions, Y behaves as a h eavy REE. For example, in the case of complexation by fluoride, whose interacti ons are exceptionally ionic (Martell and Hancock, 1996), Y acts as a super-heavy REE with a F1 formation constant exceeding that of any REE (Figure 2.4). In interactions with more covalent ligands, the complexation constants of REEs are covalently enhanced relative to Y, and Y acts as a comparatively light REE. As such, the results in Figures 2.2 and 2.3 indicate that YREE

PAGE 57

42 interactions with Fe( III) hydroxides are comparatively cova lent and interactions with Al hydroxides are comparatively ionic. Shale-normalized Y concentrations in seaw ater are higher than those of any REE (Figure 2.7A). This behavior would be predic ted if interactions of YREEs with marine particles are relatively c ovalent (weaker Y sorption, Figure 2.7C) and solution complexation is comparatively ionic (stronge r Y complexation, Figure 2.7B). Inspection of Figures 2.2A and 2.7C indicates that Fe(III) hydroxides exhibit affinities for both Y and the REEs that closely resemble the modeled scavenging behaviors of natural particulate matter in the oceans. 2.4.7. Critical Issues in YREE Surface Complexation Behavior Given that ferric oxyhydroxides are demonstr ably important natural substrates for YREE scavenging in some, if not most, aqueous environments (Sholkovitz, 1976; Sholkovitz and Elderfield, 1988; Schijf et al., 1994; Johannesson and Lyons, 1995; Johannesson and Zhou, 1999; De Carlo et al., 2000), it is important that log iKS values are accurately characterized for such phases. In this regard, it has been noted above that there is significant disagreement in the log iKS characterizations of Bau (1999), on the one hand, and the results of Ohta and Kawabe (2000, 2001) and the present work, on the other. More importantly, however, the work of Bau (1999) indicated that the log iKFe pattern obtained for YREE sorption on ferric oxyhydroxides is substantially pH dependent. This result is in contrast with the observations of Ohta and Kawabe (2000, 2001) where such variations were not discernable over a pH range between approximately 5.6 and 6.6. This issue deserves careful investigation. If log iKS patterns, such as those in Table 2.1, are invari ant over a wide range of pH, then log iKS data might be regarded in much the same manner as conditional solution complexation constants. Like solution complexation constants, the log iKS data in Table 2.1 ar e expected to vary with temperature, pressure and ionic strengt h. However, in the same manner that YREE formation constant patterns are largely unaffected by medium composition (e.g., ionic strength), log iKS patterns of YREEs might also e xhibit medium-composition invariance. In this case the absolute magnitudes of YR EE associations with surfaces would, of

PAGE 58

43 course, be dependent on solution pH and co mposition, but relative YREE affinities for particle surfaces (the shape of the log iKS patterns) would be i ndependent of solution chemistry. The existence of such a simplifying characteristic would be quite important to models of YREE environmental chemistry.

PAGE 59

44 3. Sorption of Yttrium and Rare Earth Elem ents by Amorphous Ferric Hydroxide: Influence of pH and Ionic Strength The following chapter has been peer-reviewed and published essentially in this form: Quinn K. A., Byrne R. H., and Schijf J. (2006) Marine Chemistry 99, 128-150. 3.1 Abstract The sorption of yttrium and the rare ear th elements (YREEs) by amorphous ferric hydroxide at low ionic strength (0.01 M I 0.09 M) was investigated over a wide range of pH (3.9 pH 7.1). YREE distribution coefficients, defined as iKFe = [MSi]T/(MT[Fe3+]S), where [MSi]T is the concentration of YREE sorbed by the precipitate, MT is the total YREE concentration in solution, and [Fe3+]S is the concentration of precipitated iron, are weakly dependent on ionic strength bu t strongly dependent on pH. For each YREE, the pH dependence of log iKFe is highly linear over the investigated pH range. The slopes of log iKFe versus pH regressions range between 1.43 0.04 for La and 1.55 0.03 for Lu. Distribution coefficients ar e well described by an equation of the form iKFe = (S1[H+]-1 + S2[H+]-2)/(SK1[H+] + 1), where Sn are stability constants for YREE sorption by surface hydroxyl groups and SK1 is a ferric hydroxide surface protonation constant. Best-fit estimates of Sn for each YREE were obtained with log SK1 = 4.76. Distribution coefficient predictions, using this two-site surface complexation model, accurately describe the log iKFe patterns obtained in the present study, as well as distribution coefficient patterns obtained in previous studies at near-neutral pH. Modeled log iKFe results were used to predict YREE sorption patterns appropriate to the open ocean by accounting for YREE solution comple xation with the major inorganic YREE ligands in seawater. The predicted iFelogK pattern for seawater, while distinctly

PAGE 60

45 different from log iKFe observations in synthetic solutions at low ionic stre ngth, is in good agreement with results for natural seawater obtained by others. 3.2 Introduction Distributions of yttrium and the rare earth elements (YREEs) in natural waters have been intensively investigated for more than 40 years (Goldberg et al., 1963; Hgdahl et al., 1968; Kolesov et al., 1975). The absolute an d relative concentrations of the 15 stable YREEs have been determined in a variety of open ocean environments (e.g., de Baar et al., 1985a; German et al., 1995; Zhang a nd Nozaki, 1996; Nozaki and Alibo, 2003); estuaries (e.g., Sholkovitz and Elderfield, 1988; Sholkovitz et al., 1992; Sholkovitz, 1993, 1995); rivers (e.g., Goldstein and Jacobsen 1988; Sholkovitz, 1993; Zhang et al., 1998; Nozaki et al., 2000); lakes (e.g., Johannesson an d Lyons, 1994; Johannesson et al., 1994; De Carlo and Green, 2002); ground waters (e.g., Smedley, 1991; Johannesson et al., 1996, 1997; Duncan and Shaw, 2003); hydrothe rmal fluids (e.g., Klinkhammer et al., 1983, 1994; Michard, 1989; Bau an d Dulski, 1999); and pore wate rs (e.g., Elderfield and Sholkovitz, 1987; Sholkovitz et al., 1989; Haley et al., 2004). It is generally recognized that YREE distributions in natural waters ar e largely controlled by the interplay of YREE surface and solution chemistries. Quantitative investigations of YREE interactions with particle surfaces were preceded by decades of work describing the complexation of hydrated trivalent YREE cations with a vari ety of common inorganic anions, including carbonate, hydroxide, sulfate, fluoride, and chloride (see for instance Wood, 1990; Byrne and Sholkovitz, 1996, for referen ces). As a result, YREE inter actions with major solution ligands are much better characterized than YR EE interactions with particle surfaces. In order to enable descriptions, and accurate pr edictions, of YREE behavi or in terms of key environmental variables, such as pH and ioni c strength, it is esse ntial that YREE surface chemistry is modeled as quantitatively as YREE solution complexation. Toward this goal, investigations are increasingly bein g undertaken to examine the equilibrium distribution of YREEs between solutions and relevant mineral surfaces (Byrne and Kim, 1990; De Carlo et al., 1998; Bau, 1999; Ohta and Kawabe, 2000, 2001; Quinn et al., 2004).

PAGE 61

46 In modeling studies, Lee and Byrne ( 1993) and Byrne and Sholkovitz (1996) combined a quantitative model of seawater YREE speciation with measurements of seawater YREE concentrations to estimate the comparative, average affinities of YREEs for particle surfaces in the ocean. The result ant sorption pattern, expressed in terms of free ion concentrations (Byrne and Sholkov itz, 1996), could not be compared with any directly measured distribution coefficients at the time: while an early investigation of YREE interactions with mineral surfaces (Koeppenkastrop and De Carlo, 1992) had examined YREE distributions between Fe a nd Mn oxides and seawater, these authors presented their results only in gra phical form. The groundbreaking work of Koeppenkastrop and De Carlo (1992) showed that, for REE sorption by Fe and Mn oxides, light rare earth elements (LREEs) are preferentially removed from seawater compared to heavy rare earth elements (HR EEs). As their seawater experiments were performed at a single pH (viz., pH = 7.8), the influence of pH on YREE sorption by the minerals in seawater was not revealed. Howeve r, more recent inves tigations in synthetic solutions (De Carlo et al., 1998), demonstrated the profound effect of pH on the extent of REE sorption by ferric hydroxide s. Marmier and Fromage ( 1999) showed that, at low loading, the influence of pH on La sorption by hematite could be modeled satisfactorily with a non-coulombic surface complexation model (SCM). The experiments of Bau (1999) extended the work of De Carlo et al. (1998) to include Y, and although his sorption data, in contrast to Marmier and Fr omage (1999), were not used to derive a SCM, it was shown that not only the absolu te magnitudes, but also the relative magnitudes of YREE sorption (i.e., the distri bution coefficient pattern) vary with pH. Quinn et al. (2004) noted, for the first time, that the distribution coefficient patterns obtained at near-neutral pH by Bau (1999), a nd subsequently by Ohta and Kawabe (2000, 2001), closely resemble the modeled distri bution coefficient patterns of Byrne and Sholkovitz (1996), which are appropriate to ma rine particles. Quinn et al. (2004) also showed that the distribution coefficient pattern for YREE sorption by amorphous ferric hydroxide is quite distinct from patterns observed for YREE sorption by other trivalent amorphous hydroxides (aluminum, gallium, and indium).

PAGE 62

47 Whereas the most frequently utilized subs trate for investigation of YREE sorption appears to be amorphous ferric hydroxide, to date use of this substrate has included only a few investigations (De Carlo et al., 1998; Bau, 1999; Ohta and Kawabe, 2000, 2001; Quinn et al., 2004). In thes e studies, YREE sorption by amorphous ferric hydroxide was measured over a range of pH (3.5 – 9.0) but the magnitudes of estimated distribution coefficients for individual YR EEs, at constant pH, differ by as much as a factor of 400. Distribution coefficients have been obtained at ionic strengths close to or equal to that of seawater (De Carlo et al., 1998; Ohta and Kawabe, 2000, 2001) and at very low ionic strengths (De Carlo et al., 1998; Bau, 1999; Quinn et al., 2004). Although the pH dependence of YREE sorption by am orphous ferric hydroxide was modeled quantitatively by Ohta and Kawabe (2000, 20 01), their data were obtained at high substrate loading (i.e., high [YREE]/[Fe3+]T ratios) within a small pH range (approximately one unit). In the present work, we have produced a quantitative model of YREE sorption at low substrate loadings (sim ilar to those used by Bau, 1999), based on data obtained over an ionic strength range of 0.01 – 0.09 M and a relatively wide pH range of 3.9 – 7.1. We use this model to as sess the nature and the importance of YREE sorption by amorphous ferric hydroxides in the open ocean. 3.3 Materials and Methods 3.3.1. Materials and Preparation of the Experimental Solutions A class-100 clean air laborator y or laminar flow bench was utilized for all chemical manipulations. Teflon and polypropylene laboratory materials and polycarbonate filter membranes were cleaned by soaking in HCl or HNO3 for at least one week, followed by several thorough rinses with trace metal-clean water (Mil li-Q water) from a Millipore (Bedford, MA) purification system. Solution pH, on the free hydrogen-ion concentration scale, was measured using a Ross-type combination pH electrode (No. 810200) connected to a Corning 130 pH meter in the ab solute millivolt mode. Nernstian behavior of the electrode was verified periodically by titrating a 0.3 M NaCl solution with concentrated HCl. The electrode was calibra ted daily during each experiment using an HCl standard solution.

PAGE 63

48 Ammonium nitrate (99.999%), ammonium chloride (99.998%), and certified 1.000 M hydrochloric acid were purchased from Sigm a-Aldrich (St. Louis, MO). TraceMetal Grade nitric acid, TraceMetal Grade ammoni um hydroxide, and ferric chloride solution (40 3% w/v in HCl) were purchased from Fi sher Scientific (Pittsburgh, PA). A YREE stock solution, containing 66.7 ppm of each YREE in 2% HNO3, was prepared from single-element ICP standards (SPEX CertiPre p, Metuchen, NJ). All solutions were prepared with Milli-Q water. At the start of each experiment, an experi mental solution and a pH standard solution were prepared in Teflon wide-mouth bottles. The experimental solution consisted of 23.3 ppb of each YREE (total combined YREE concentration 2.36 M) and 0.10 – 10 mM iron in 0.01 M HCl. The concentration of iron utilized de pended on the desired experimental pH and was varied in order to obtain adequate YREE sorption. The pH standard solution consisted of 0.001 M HCl (pH 3.0) in 0.01 M NH4NO3. For all solutions containing 0.62 mM iron, the ionic streng th was initially 0.011 0.001 M. The experimental solution cont aining 10 mM iron had an ionic strength equal to 0.043 M. All solutions were equilibrated in a jacketed beaker, thermostated at T = (25.0 0.1)C, and continuously stirred with a Teflon-coated ‘floating’ stir bar. The experimental solution was bubbled throughout with ultra-pure N2, which had been passed through an in-line trap (Supelco, Bellafonte, PA) that removed all traces of CO2. 3.3.2. pH Dependence of YREE Sorption Six experiments were performed to study the pH dependence of YREE sorption by amorphous ferric hydroxide precipit ates. After the experimental solution had equilibrated for approximately 24 hours, a sample was take n at the initial solu tion conditions (pH 2; no YREE sorption) to determine total YREE concentrations, MT. The pH of the experimental solution was then increased by addition of 1 M NH4OH from a Gilmont micro-dispenser, initiating pr ecipitation of a yellow-brown ferric hydroxide colloid. Once the pH stabilized, samples were taken with a pipette. Experiments were performed either at constant pH (~5, 6, or 7) with severa l samples taken at fixed times within a 2-day

PAGE 64

49 period, or over a range of pH (3.9 – 5.6, or 5.1 – 7.0) with one or more samples taken at each half-unit pH increment. 3.3.3. Ionic Strength Dependence of YREE Sorption Two experiments were performed to study the ionic strength dependence of YREE sorption by amorphous ferric hydroxide. Both experiments covered the ionic strength range 0.01 – 0.09 M, in 0.02 M increments. An initial sample (pH 2, I = 0.01 M; no YREE sorption) was taken after a 24-hour equilibration period. The pH of the experimental solution was then increased to approximately 6.0 by addition of 1 M NH4OH. After sampling at these conditions (pH 6, I = 0.01 M), the ionic strengths of both the experimental solution and the pH st andard were increased by addition of 5 M NH4Cl. At each new ionic strength the pH of the experimental solution was readjusted to 6.0 with 1 M NH4OH. Several samples were taken after each pH readjustment. In the first experiment, four samples were taken at each ionic strength: one at 5 minutes, one at 4 hours, and two at either 90 minutes, 24 hour s, or 48 hours. In the second experiment, two samples were taken at each ionic streng th: one at 15 minutes and one at 60 minutes. 3.3.4. Sampling and Analysis Two sampling techniques were used concur rently. One sample aliquot was filtered using a polypropylene syringe with a Nuclepor e filter membrane (polycarbonate, 0.10 m pore size) mounted in a polypropyl ene filter holder. The syringe and membrane were first rinsed with 5 mL of solution and then anot her 5 mL were collect ed in a polypropylene centrifuge tube. A second sample aliquot wa s centrifuged using a Centra-4B centrifuge (International Equipment Company, Needha m Heights, MA) for one hour at about 4,000 rpm. A combination of these two tech niques was used in the experiment with 10 mM iron, because the copious precipitate was difficult to filter directly: after centrifuging 5 mL of sample, 3.5 mL of the resulting supernatant was filtered (using ~1 mL as a rinse). Because filtration provi ded better separation of particles from solutions containing the lowest concentration of iron (0.10 mM), centrifugation was not used in every experiment.

PAGE 65

50 The filtered samples and the supernatant of the centrifuged samples were diluted 5fold with 1% HNO3, and a small amount of internal st andard solution containing equal amounts of In, Cs, and Re was added. The re sulting mixtures were analyzed for YREEs with an Agilent Technologies 4500 Seri es 200 inductively-coupled plasma mass spectrometer (ICP-MS) following the procedure ou tlined in Quinn et al. (2004). In brief, all standards and sample solutions were injected in triplicate. During instrument tuning, the formation of oxide and double-charged ions was minimized with a 10 ppb Ce solution. MO+ and M2+ peaks were always less than 1% and 3% of the corresponding M+ peak, respectively, and correction for this effect proved unnecessary. YREE concentrations were calculated from linear regr essions of four standards (0.5, 1, 2, and 5 ppb). Ion counts were corrected for mi nor instrument drift by normalizing 89Y to 115In, 139La–161Dy to 133Cs, and 163Dy–175Lu to 187Re. To check the validity of the drift correction, a comparison was made of the Dy concentrations calculated from 161Dy and 163Dy, which were usually identical within 2%. Whenever this difference was significantly larger than 2%, a mass-dependent correction wa s performed by interpolating between the internal standards, 133Cs and 187Re, using Excel. Raw data from each experiment were correct ed by a dilution factor, which was based on the amount of NH4OH added to increase the pH. Di stribution coefficients were calculated from these corrected data using the following equation: iTiT iFe 3+3 iTS[MS][MS] K= [M][S]M[Fe], (3.1) where brackets denote the concentration of each indicated species. Over the range of experimental conditions employed in this work, the concentration of free YREE was set equal to the total dissol ved YREE concentration ([M3+] = MT). The concentration of sorptive solid substrate was set equal to the concentration of precipitated iron ([Si] = [Fe3+]S), which was assumed to be equal to the initial dissolved Fe concentration. The concentration of sorbed YREE, [MSi]T, was calculated as the difference between the YREE concentration of the initial sample at pH 2 and the YREE concentration of the filtrate at each subsequent time after a pH or ionic strength adjustment.

PAGE 66

513.4 Data Analysis 3.4.1. Modeling of pH and Ionic Strength Effects Observed distribution coefficient data we re initially modeled using an empirical equation wherein log iKFe was presumed to have simple linear dependences on pH and ionic strength (I): iFepHIiFelogKQpHQIlogK(pH 0, I = 0) (3.2) The slopes (QpH and QI) and intercept (log iKFe(pH 0, I = 0)) in equation (3.2) were determined by least squares analysis of the data given in Appendices B and C. QpH and log iKFe(pH 0, I = 0) were determined from fits of log iKFe versus pH at constant ionic strength (I ~ 0.011 M, Tables B.1–B.5; and I = 0.043 M, Table B.6) using equation (3.2) written as: iFeIpHiFelogKQIQpHlogK(pH 0,I0) (3.3a) QI was determined from log iKFe versus ionic strength datasets at pH 6.13 (Tables C.1 and C.2) using equation (3.2) written as: iFepHIiFelogKQpHQIlogK(pH 6.13,I0) (3.3b) After initially setting QI = 0 in equation (3.3a) iterations of equati ons (3.3a) and (3.3b) were performed until QpH and QI converged to constant values. Examination of equation (3.2) shows that log iKFe at I = 0 M and pH 6. 13 is larger than log iKFe at I = 0 M and pH 0 by the additive term QpH 6.13 (i.e., log iKFe(pH 6.13, I = 0) = log iKFe(pH 0, I = 0) + QpH 6.13). 3.4.2. Surface Complexation Model The surface complexation model (SCM) of Schindler and Stumm (1987) describes metal sorption onto hydroxide precipitates in term s of proton exchange with either one or two surface hydroxyl groups. The sorption r eactions are written in the form: 32 32SFe(OH)MSFeO(OH)MH, (3.4) 3 32SFe(OH)MSFeO(OH)M2H, (3.5) and

PAGE 67

523 26224SFe(OH)MSFeO(OH)M2H, (3.6) where S– represents the bulk solid and M3+ is a YREE ion. The total concentration of sorbed YREE can then be written as: 2 iT22224[MS][SFeO(OH)M][SFeO(OH)M][SFeO(OH)M] (3.7) Since there is no means of differentiating reactio ns (3.5) and (3.6) in our experiments, the last term in equation (3.7) can be omitted. Surface complexation constants, Sn, for reactions (3.4) and (3.5) can be written in the form: 3nn n3n Sn T3[SFeO(OH)M][H] M[SFe(OH)] (3.8) and then substituted into the first tw o terms of equation (3.7) to give: 12 iTT3S1S2[MS]M[SFe(OH)]([H][H]) (3.9) Equations (3.1) and (3.9) then yield an expression for iKFe in terms of surface complexation constants: 123 iFeS1S23SK([H][H])[SFe(OH)]/[Fe] (3.10) The total concentration of iron in each experi ment is equal to the sum of the dissolved and precipitated iron: 333 TDS[Fe][Fe][Fe] (3.11) where 3320 D23[Fe][Fe][FeOH][Fe(OH)][Fe(OH)] (3.12) and 30 S234[Fe][SFe(OH)][SFe(OH)][SFe(OH)] (3.13) Based on our experimental conditions ([Fe3+]T 0.10 mM and pH 4.0), the concentration of dissolved iron is quite small compared to that of the precipitated iron, which supports the assumption that [Fe3+]T = [Fe3+]S. Equation (3.13) partitions precipitated iron hydroxide am ong three types of charged su rface sites. At lower pH values, the first term dominates so the su rface is positively charged. Surface protonation at low pH is described by the following reaction: 0 322SFe(OH)HSFe(OH)HO, (3.14)

PAGE 68

53 with a stability constant defined by: 2 S1 0 3[SFe(OH)] K [SFe(OH)][H] (3.15) At higher pH values, the third term in e quation (3.13) dominates so the surface is negatively charged. The surface deprotonati on reaction at high pH is written as: 0 324SFe(OH)HOSFe(OH)H (3.16) with a stability constant defined by: 4 S2 0 3[SFe(OH)][H] K [SFe(OH)] (3.17) Combining equations (3.13), (3.15), and (3.17), gives: 301 S3S1S2[Fe][SFe(OH)](K[H]1K[H]) (3.18) Since the pH was always less than 7.2 in our experiments and the pristine point-of-zerocharge is approximately 8.0 for amorphous ferric hydroxide (Dzombak and Morel, 1990, and references therein), it was assumed that 1 S2K[H]1, hence the last term in equation (3.18) could be omitted. Using equati ons (3.10) and (3.18), the expression for iKFe (equation (3.1)) is then written as: 12 S1S2 iT iFe 3 TSS1[H][H] [MS] K M[Fe]K[H]1 (3.19) In contrast to equation (3.2), which is an empirical model, the SCM embodied in equation (3.19) provides a description of di stribution coefficient results in terms of specific chemical equilibria. Sin ce equation (3.19) expresses log iKFe solely as a function of pH, log iKFe data from experiments at constant ionic strength (I ~ 0.011 M, Tables B.1–B.5; I = 0.043 M, Table B.6) were used to determine the parameters in equation (3.19). In order to remove small variations in log iKFe attributable to the ~0.032 M ionic strength difference in the datasets of Tables B.1–B.5 and Table B.6 ( log iKFe 0.03), the QI results in Table 3.1, applied to the data in Tables B.1–B.6, were used to generate iKFe data appropriate to I = 0.025 M, the av erage ionic strength of these datasets. SigmaPlot (Version 8.02) was then used to solve equation (3.19) for S1, S2, and SK1 through minimization of the following re sidual sum of squares (RSS) function:

PAGE 69

54 2 12 1 S1S2 iFe S1[H][H] RSS1K(pH,I0.025 M) K[H]1 (3.20) where the summation was performed over all pH values. Values of SK1 for each YREE were obtained from equation (3.20), but since SK1 is a property of the iron hydroxide, a single value needed to be determ ined for all YREEs. The average SK1 value was therefore optimized by finding the minimum RSS over all pH values and all YREEs when SK1[H+] was kept constant in equation (3.20). Best-fit estimates for S1 and S2 were obtained for each individual YREE from equation (3.20) using this optimized SK1. 3.5 Results and Discussion 3.5.1. Empirical Model of the log iKFe Dependence on pH and Ionic Strength Based on the data compiled in Tables B.1–B .6, which were obtaine d over a range of pH, and the data compiled in Tables C.1 a nd C.2, which were obtained over a range of ionic strength, distribution coefficients (log iKFe) were observed to exhibit a strong dependence on pH and a much weaker depende nce on ionic strength. Four representative regressions for the final iteration of the pH data and the ionic strength data (i.e., using equations (3.3a) and (3.3b)) are shown in Fi gures 3.1 and 3.2. Based on these regressions for each YREE, the three coefficients of equation (3.2) (QpH, QI, and log iKFe(pH 0, I = 0)) are summarized in Table 3.1 and ar e depicted graphica lly in Figure 3.3. To examine the goodness-of-fit for the model, the ratios of pred icted distribution coefficients (equation (3.2)) to measured dist ribution coefficients were plotted against pH. This is shown in Figure 3.4 for four representative REEs. The generally random scatter around the horizontal line (log iKFe(pred)/log iKFe(meas) = 1) indicates that equation (3.2) satisfactorily models the data. In Figure 3.3, it can be seen that there are substantial uncertainties associated with each of the coefficients of equation (3.2). Th ese coefficient uncertainties are consistent with uncertainties in measured log iKFe patterns over time, which are on the order of 0.1 – 0.3 units. Figure 3.5 shows log iKFe observations from two re presentative experiments. The log iKFe patterns obtained in the experiment at pH = 7.06 0.05 are highly consistent

PAGE 70

55 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 y = (1.4260.04)x (5.150.2) r2 = 0.968La 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 7 y = (1.5400.03)x (4.940.2) r2 = 0.976Sm pH 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 7 y = (1.5840.03)x (5.410.2) r2 = 0.982Dy pH 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 7 y = (1.5510.03)x (5.110.1) r2 = 0.985Lu Figure 3.1. Final regressions of log iKFe (Tables B.1–B.6; normalized to I = 0 M) versus pH for La, Sm, Dy, and Lu. Dashed lines represent 95% confidence intervals.

PAGE 71

56 0.000.020.040.060.080.10 log iKFe 3.1 3.2 3.3 3.4 3.5 3.6 3.7 Lay = (0.430.3)x + (3.400.02) r2 = 0.062 0.000.020.040.060.080.10 log iKFe 4.0 4.1 4.2 4.3 4.4 4.5 4.6 Smy = (0.670.3)x + (4.300.02) r2 = 0.130 I (M) 0.000.020.040.060.080.10 log iKFe 3.9 4.0 4.1 4.2 4.3 4.4 Dyy = (0.430.3)x + (4.150.02) r2 = 0.070 I (M) 0.000.020.040.060.080.10 log iKFe 4.0 4.1 4.2 4.3 4.4 4.5 Luy = (0.900.3)x + (4.250.02) r2 = 0.234 Figure 3.2. Final regressions of log iKFe (Tables C.1 and C.2; normalized to pH 6.13) versus ionic strength (I) for La, Sm, Dy, a nd Lu. Dashed lines represent 95% confidence intervals.

PAGE 72

57 Table 3.1. Results for the coefficients of equation (3.2) (see text for details). Uncertainties represent one standard error. [M3+] QpH QI log iKFe(pH 0, I = 0) Y 1.534 0.05 0.64 0.5 -5.60 0.3 La 1.426 0.04 0.43 0.3 -5.15 0.2 Ce 1.459 0.04 0.73 0.3 -4.85 0.2 Pr 1.506 0.03 0.90 0.4 -4.99 0.2 Nd 1.509 0.04 0.98 0.4 -4.92 0.2 Pm Sm 1.540 0.03 0.67 0.3 -4.94 0.2 Eu 1.542 0.03 0.51 0.3 -5.00 0.2 Gd 1.522 0.03 0.44 0.3 -5.07 0.2 Tb 1.568 0.03 0.29 0.3 -5.29 0.2 Dy 1.584 0.03 0.43 0.3 -5.41 0.2 Ho 1.570 0.03 0.57 0.3 -5.39 0.2 Er 1.571 0.03 0.66 0.3 -5.37 0.2 Tm 1.572 0.03 0.76 0.3 -5.29 0.1 Yb 1.581 0.03 0.84 0.3 -5.25 0.1 Lu 1.551 0.03 0.90 0.3 -5.11 0.1

PAGE 73

58 1.35 1.40 1.45 1.50 1.55 1.60 1.65QpH QI -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu -6.0 -5.8 -5.6 -5.4 -5.2 -5.0 -4.8 -4.6 -4.4log iKFe(pH 0, I = 0) Figure 3.3. Coefficients of equation (3.2). Upper panel (QpH) is the slope of the linear regression with respect to pH. Middle panel (QI) is the slope of the linear regression with respect to ionic strength. Lower panel (log iKFe(pH 0, I = 0)) is the intercept of the linear regression. Error bars represent one standard error.

PAGE 74

59 3.54.04.55.05.56.06.57.07.5 log iKFe(pred) / log iKFe(meas) 0.0 0.5 1.0 1.5 2.0 2.5 La 3.54.04.55.05.56.06.57.07.5 log iKFe(pred) / log iKFe(meas) 0.8 0.9 1.0 1.1 1.2 1.3 Sm pH 3.54.04.55.05.56.06.57.07.5 log iKFe(pred) / log iKFe(meas) 0.8 0.9 1.0 1.1 1.2 Lu pH 3.54.04.55.05.56.06.57.07.5 log iKFe(pred) / log iKFe(meas) 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Dy Figure 3.4. log iKFe(pred)/log iKFe(meas) versus pH for La, Sm, Dy, and Lu, where log iKFe(pred) are distribution coefficients pr edicted from equation (3.2) using the coefficients listed in Table 3.1, and log iKFe(meas) are experimentally observed distribution coefficien ts (Tables B.1–B.6).

PAGE 75

60 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKFe 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 pH 613; 5 min pH 613; 90 min pH 613; 4 hrs pH 607; 46 hrs YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKFe 5.0 5.2 5.4 5.6 5.8 6.0 6.2 6.4 pH 710; 90 min pH 709; 5 hrs pH 704; 24 hrs pH 700; 46 hrs pH 700; 48 hrs A B Figure 3.5. log iKFe results for experiments performed at (A) pH = 7.06 0.05 and I = 0.0109 M (Table B.3) and (B) pH = 6.10 0.03 and I = 0.0503 M (Table C.1). over periods of time from 90 minutes up to 48 hours (Figure 3.5A). The same consistency over time (from 5 minutes up to 46 hours) is seen in the log iKFe patterns obtained at I = 0.05 M (Figure 3.5B). These observations indica te that the uncertainties shown in Figure 3.3 are strongly correlated acr oss the YREE series. Hence the absolute values of log iKFe, but not the relative values (patterns), are a ffected by these uncertainties. This ensures that, despite large standard errors, the coefficients of equation (3.2) (QpH, QI, and

PAGE 76

61 log iKFe(pH 0, I = 0)), summarized in Table 3.1, accurately represent measured log iKFe patterns as a function of pH and ionic strength. The large uncertainty associated with QI also results from the weak dependence of YREE sorption on ionic strength, as indica ted by the small (< 0.25) correlation coefficients in Figure 3.2. This lack of an ionic strength effect for sorption by ferric hydroxides, both amorphous and cr ystalline, has been observe d in previous experiments with other cations, including copper, lead, and cadmium (Swallow et al., 1980; Hayes and Leckie, 1987). Swallow et al. (1980) suggested that the lack of an ionic strength effect indicates there is no net change in charge during the sorption r eaction and therefore coulombic interactions do not play a role in the process. Using a generalized two-layer model with the best available estimates for zinc surface complexation constants, Dzombak and Morel (1990) indi cated that ionic strength ha s a minimal influence on zinc sorption by amorphous ferric hydroxide. Dzom bak and Morel (1990) explained the absence of an ionic strength effect in terms of the Gouy-Chapman and Debye-Hckel theories, stating that varia tions in ionic strength produce similar changes in the free energies of species at the surface and in the bulk solution, whereupon the standard free energy change of the sorption reaction remains nearly constant. 3.5.2. Surface Complexation Model Results Using Sm as an example, Figure 3.6 compares surface complexation model results obtained by (i) setting SK1[H+] in equation (3.19) equal to zero (Figure 3.6A) and (ii) determining the log SK1 value which minimized the RSS in equation (3.20) (Figure 3.6B). The ratios of predicted distri bution coefficients (equation (3. 19)) to measured distribution coefficients (log iKFe(pred)/log iKFe(meas)) exhibit systematic, pH-dependent variations when SK1[H+] is set equal to zero (Figure 3.6A ). In contrast, no such systematic deviations between predicted and measured log iKFe values (Figure 3.6B) are observed when using the log SK1 value that minimized the RSS in equation (3.20) (i.e., log SK1 = 4.76). Comparisons of log iKFe predictions and experimental observations, exemplified by the results shown for Sm in Fi gures 3.6A and 3.6B, clearly demonstrate the importance of variations in the [S–Fe(OH)2 +]/[S–Fe(OH)3 0] ratio (equation (3.15)) at low pH.

PAGE 77

62 pH 3.54.04.55.05.56.06.57.07.5 log iKFe(pred) / log iKFe(meas) 0.7 0.8 0.9 1.0 1.1 1.2 pH 3.54.04.55.05.56.06.57.07.5 log iKFe(pred) / log iKFe(meas) 0.8 0.9 1.0 1.1 1.2 A B Figure 3.6. log iKFe(pred)/log iKFe(meas) versus pH for Sm. (A) log iKFe(pred) are distribution coefficients predicted from equation (3.19) with the assumption that S1K[H]1. (B) log iKFe(pred) are distribution coeffici ents predicted from equation (3.19) using log SK1 = 4.76 and the surface complexation constants (Sn) listed in Table 3.2. For both panels, log iKFe(meas) are distribution coeffici ents from experiments at constant ionic strength (Tables B. 1–B.6), normalized to I = 0.025 M.

PAGE 78

63 The final best-fit results for the amorphous ferric hydroxide YREE surface complexation constants (S1 and S2) are summarized in the first two columns of Table 3.2 and are depicted graphically in Figure 3.7. It can be seen that the patterns for the two surface complexation constants are fairly sim ilar except for (i) the location of Y relative to the REEs and (ii) the sequence of inflec tions seen in the middle-to-heavy REEs, which is more pronounced for log S2. At low pH, S1[H+]-1 > S2[H+]-2 in equation (3.19) so reaction (3.4) represents the dominant so rption reaction. At high pH, the dominant sorption reaction is reaction (3.5 ), which is represented by the S2 term in equation (3.19) (Figure 3.7B). The pH at which [S–FeO2(OH)M+] = [S–FeO(OH)2M2+] is equal to the ratio S2/S1. This ratio is listed in the third column of Table 3.2 for each YREE and has an average value of 6.30. The distribution coefficients (log iKFe) used to determine S1, S2, and SK1 were expressed in terms of total (MT) rather than free ([M3+]) YREE concentrations. For most Table 3.2. YREE surface complexation constants (Sn) determined with equation (3.19) and the data in Tables B.1–B.6. Best fits of iKFe versus [H+] were obtained with log SK1 = 4.76 (see text for details). Uncertainti es represent one standard error. [M3+] log S1 log S2 -log (S2/S1) Y -2.96 0.06 -8.93 0.07 5.97 La -2.86 0.04 -9.49 0.11 6.63 Ce -2.37 0.04 -8.99 0.11 6.62 Pr -2.25 0.03 -8.73 0.08 6.48 Nd -2.16 0.03 -8.66 0.09 6.49 Pm Sm -2.05 0.03 -8.40 0.07 6.35 Eu -2.10 0.03 -8.42 0.07 6.32 Gd -2.27 0.03 -8.67 0.08 6.40 Tb -2.28 0.03 -8.49 0.06 6.21 Dy -2.31 0.03 -8.46 0.06 6.15 Ho -2.36 0.03 -8.55 0.06 6.20 Er -2.32 0.03 -8.52 0.06 6.20 Tm -2.23 0.03 -8.41 0.06 6.18 Yb -2.16 0.03 -8.27 0.05 6.11 Lu -2.15 0.03 -8.34 0.06 6.19

PAGE 79

64 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log S1 -3.2 -3.0 -2.8 -2.6 -2.4 -2.2 -2.0 -1.8 A YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log S2 -9.8 -9.6 -9.4 -9.2 -9.0 -8.8 -8.6 -8.4 -8.2 -8.0 B Figure 3.7. Surface stability constants (equation (3.19)) for YREE sorption by amorphous ferric hydroxide. Error bars represent one standard error. of our solution conditions (pH 6.5), this approximation is acceptable because M3+ is weakly hydrolyzed ([M3+]/MT 0.87). At pH 7, the propor tions of LREE hydroxides are small (e.g., [LaOH2+]/[La3+] 0.01) but the formation of HREE hydroxides is sufficiently large (e.g., [YbOH2+]/[Yb3+] 0.4) that it is necessary to assess the possible impact of YREE hydrolysis on our results. As such, dist ribution coefficient behavior was also modeled using free YREE concentrations. MT/[M3+] ratios were calculated using the hydrolysis constants (1 ) of Klungness and Byrne (2000). Distribution coefficients (iKFe) expressed in terms of MT and [M3+] are related as follows:

PAGE 80

6531 iFeiFeT1K([M])K(M)(1[H]) (3.21) Using distribution coefficien ts expressed in terms of free YREE concentrations, the equilibrium constants in equation (3.19) (S1, S2, and SK1) were recalculated using methods identical to those described above. Residual sum of square s (RSS) results were then compared for the model expressed in terms of MT and the model expressed in terms of [M3+]. For each YREE, the RSS was slightly larger (< 5% for LREEs and < 20% for HREEs) using log iKFe results expressed in terms of [M3+]. The lack of improvement when log iKFe was modeled in terms of free YREE concentrations rather than total concentrations may be caused by the rather small extent of YREE hydrolysis under the conditions used in this work. Since the magnitude of *1 1log(1[H]) in equation (3.21) was less than 0.15 for Yb at pH 7, and much smaller for most other elements, the extent of YREE hydrolysis appears to have been too small, relative to experimental uncertainties in log iKFe determinations, to distinguish YREE hydrolysis effects from experimental uncertainties in log iKFe. This issue will be addressed in future investigations of the effect of solution complexation on YREE sorption. 3.5.3. Comparative log iKFe Predictions using SCM Results The log iKFe patterns obtained in this work ar e compared, in Figure 3.8, with the log iKFe results obtained by Bau ( 1999), Ohta and Kawabe (2 000, 2001), and De Carlo et al. (1998). Figure 3.8A compares the log iKFe predictions of equation (3.19) with the observed result of Bau (1999) at pH 5.97, Figure 3.8B compares equation (3.19) predictions with the result of Ohta and Kawa be (2000, 2001) at pH 6.59, and Figure 3.8C compares the predictions of equation (3.19) with the result of De Carl o et al. (1998) at pH 6.25 and I = 0.1 m. In order to emphasize similarities/differences in log iKFe patterns, rather than absolute log iKFe magnitudes, the predicted patterns in Figure 3.8 were vertically adjusted w ith scaling constants (see also Quinn et al., 2004). The comparisons shown in Figure 3.8 demonstrate that equatio n (3.19), along with the regression data given in Table 3.2, accurately predicts the sh ape of patterns obtained at near-neutral pH by Bau (1999), Ohta and Kawabe (2000, 2001), and De Carlo et al. (199 8). Overall, there

PAGE 81

66 log iKFe 4.6 4.8 5.0 5.2 5.4 5.6 5.8 6.0 Bau (1999) this work (predicted) log iKFe 3.4 3.6 3.8 4.0 4.2 4.4 4.6 Ohta and Kawabe (2000, 2001) this work (predicted) B A YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKFe 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 De Carlo et al. (1998) this work (predicted) C Figure 3.8. Comparison between measured di stribution coefficients and log iKFe values predicted from equation (3.19). (A) Result from Bau (1999) at pH 5.97. (B) Result from Ohta and Kawabe (2000, 2001) at pH 6.59. (C) Result from De Carl o et al. (1998) at pH 6.25 and I = 0.1 m. The overlap of each pair of patterns was separately maximized by addition of a scaling constant determined by RSS analysis, to the predicted pattern only (see Quinn et al., 2004).

PAGE 82

67 is slightly better agreement between the predic tions of equation (3.19) and the results of Ohta and Kawabe (2000, 2001) than is observe d for comparisons with the results of Bau (1999) or the results of De Carlo et al. ( 1998). This is somewhat surprising because the degree of substrate loading ( [YREE]/[Fe3+]T) was much larger in the experiments of Ohta and Kawabe (2000, 2001) (0.4) than in the experiments of Bau (1999) (0.004), De Carlo et al. (1998) (0.06), or the present work (0.0002 – 0.02). Since the present study was not intended to examine the effect of loading, additional studies are needed to assess the influence of substrate loading on YREE log iKFe behavior. Although there is good genera l agreement between log iKFe patterns obtained at pH > 5.0 (i.e., Figure 3.8), the log iKFe results of Bau (1999) at lo w pH differ substantially from the log iKFe patterns obtained in the present study. Figure 3.9 compares, for example, the log iKFe pattern obtained by Bau (1999) at pH 3.91 with the average log iKFe pattern obtained in the present work at pH = 3.96 0.10 (Table B.6). The log iKFe pattern obtained by Bau (1999) is ve ry flat relative to log iKFe patterns observed in the present study. The cause of the obser ved differences in the log iKFe patterns at low pH (Figure 3.9) is unknown. Further investigations of YREE sorption onto amorphous ferric hydroxide at low pH are needed to resolve these differences. In addition to comparing dist ribution coefficient patterns obtained at specific pH values (Figures 3.8 and 3.9), it is informative to compare log iKFe versus pH relationships. In the present work, log iKFe is linearly dependent on pH (Figure 3.1) over the entire investigated pH range (3.9 – 7.1). Th is linearity is also seen in the results of Ohta and Kawabe (2000, 2001) between pH 5. 8 and 6.6 (Figure 3.10), and in the log iKFe results of De Carlo et al. (1998) between pH 4.0 and 7.0 (Figure 3.11). The results of De Carlo et al. (1998) show non-linearity only at higher pH (> 7.0, not shown), where a correction for hydrolysis (equation (3 .21)) is required. Although the log iKFe versus pH slope obtained by De Carlo et al (1998) is somewhat smaller than those obtained in Ohta and Kawabe (2000, 2001) and the present work, the results of Bau (1999) exhibit characteristics that differ both qualitatively and quantitatively from the log iKFe versus pH results obtained by others (Figures 3.10 and 3. 11). The data of Bau (1999) can be divided

PAGE 83

68 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKFe (this work) 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 log iKFe (Bau, 1999) 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 this work Bau (1999) Figure 3.9. Comparison between the average log iKFe result at pH = 3.96 0.10 from the present work (Table B.6) and the log iKFe result at pH 3.91 from Bau (1999). The values obtained in the present work ar e shown on the left axis whil e the values obtained by Bau (1999) are shown on the right axis.

PAGE 84

69 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 La1.290.14 1.430.04 pH 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 7 Dy1.540.10 1.580.03 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 7 Sm1.430.10 1.540.03 pH 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 7 Lu1.550.03 1.560.09 Figure 3.10. Regressions of log iKFe versus pH for La, Sm, Dy, and Lu. ( ) Results from the present work (Tables B.1–B .6; normalized to I = 0 M). ( ) Results from Ohta and Kawabe (2000, 2001) on the mole fraction s cale were corrected by a multiplicative factor (([Fe3+] + REE)/[Fe3+]) to make them consistent with iKFe as defined by equation (3.1). The slopes (QpH 1 standard error) are listed next to each regression line.

PAGE 85

70 3.04.05.06.07.08.0 log iKFe 0 1 2 3 4 5 6 La1.430.04 0.540.03 0.450.03 2.210.10 3.04.05.06.07.08.0 log iKFe 0 1 2 3 4 5 6 7 Sm1.540.03 0.780.03 0.780.08 2.380.09 pH 3.04.05.06.07.08.0 log iKFe 0 1 2 3 4 5 6 7 Dy1.580.03 0.930.03 0.630.06 2.310.06 pH 3.04.05.06.07.08.0 log iKFe 0 1 2 3 4 5 6 7 Lu2.200.07 0.710.07 1.550.03 0.720.04 Figure 3.11. Regressions of log iKFe versus pH for La, Sm, Dy, and Lu. () Results from Bau (1999) were normalized to his average iron concentration. ( ) Results from the present work (Tables B.1–B.6; normalized to I = 0 M). ( ) Results from De Carlo et al. (1998) at I = 0.1 m on the percent sorbed scale were c onverted to distribution coefficients as defined by equation (3.1), which included normalization to their iron concentration in moles of amorphous ferric hyd roxide per kilogram of solution. The slopes (QpH 1 standard error) are listed ne xt to each regression line. into two separate linear regressions (pH 3.6 – 5.0 and pH 5.3 – 6.2). The log iKFe versus pH slope in the upper pH region is two to four times larger than the slope obtained by De Carlo et al. (1998) and, near pH 6.0, the iKFe values of Bau (1999) are 2 to 3 orders of magnitude larger th an the values reported by De Carlo et al. (1998), Ohta and Kawabe (2000, 2001), and the present work.

PAGE 86

71 It should be noted here that the present work (Tables B.1–B.6) was performed at low ionic strength (I 0.043 M), whereas the study of Ohta and Kawabe (2000, 2001) was performed at I = 0.5 M. The good agreement be tween the results in the present work and the results in the work of Ohta and Kawa be (2000, 2001) supports the conclusion that ionic strength does not have a strong e ffect on YREE sorpti on by amorphous ferric hydroxide. The distribution coeffi cients for Eu and Tb of De Carlo et al. (1998), after correcting for nitrate complexation using the stability constants of Choppin and Strazik (1965), also showed a weak dependence on ionic strength over the range 0.1 – 0.7 m (similar to that seen in Figur e 3.2). In contrast, the log iKFe(Eu, Tb) results of De Carlo et al. (1998) at zero ionic strength were larger than their I = 0.1 m log iKFe(Eu, Tb) values by 0.6 log units over the pH range 4.0 – 7.0. While the origin of this difference is uncertain, the range of log iKFe values observed by De Carlo et al. (1998) between 0.0 and 0.7 m ionic strength may simply be attributab le to the typical experimental errors encountered in log iKFe determinations. 3.5.4. log iKFe Predictions for Seawater Distribution coefficients (iKFe) predicted from equation (3.1 9) at high pH (> 7.5) can be converted to results expressed in term s of total YREE concentration in seawater (iFeK ) using the following equation: 1 n iT iFeiFeini 3 i,n swS[MS] KK1[L] M[Fe] (3.22) where in are stability constants for formation of YREE solution complexes: in in 3n i[M(L)] [M][L], (3.23) and Msw represents the total concentr ation of each YREE in seawater: 32 sw3324M[M][MCO][M(CO)][MOH][MSO] (3.24) Stability constants used in equation (3. 22) include those for YREE complexation by chloride (Luo and Byrne, 2001), fluoride (L uo and Byrne, 2000), sulfate (Schijf and Byrne, 2004), hydroxide (Klungness and Byrn e, 2000), and carbonate and bicarbonate (Luo and Byrne, 2004).

PAGE 87

72 Figure 3.12A shows predicted (equation (3.19) and Table 3.2) distribution coefficients (log iKFe) expressed in terms of free YREE concentrations at several pHs relevant to seawater. Figure 3. 12B shows the Figure 3.12A log iKFe results transformed to iFelogK values using equation (3.22). The results shown in Figure 3.12B indicate that amorphous ferric hydroxide preferentially remo ves (scavenges) LREEs from seawater. Comparison of Figures 3.12A and 3.12B shows that log iKFe patterns vary little with pH while iFelogK patterns are strongly pH dependent. Th is is largely due to the strong pH dependence of carbonate complexation. Figure 3.12B also shows that the pH dependence of iFelogK is much larger for the LREEs than for the HREEs. Between pH 7.6 and 8.2, iFelogK for La increases by 0.6 units while iFelogK for Lu increases by only 0.2 units. Koeppenkastrop and De Carlo (1992) direct ly measured REE sorption by amorphous goethite in seawater. Figure 3.12C compares the iFelogK pattern predicted in the present work at pH 7.8 (equations (3.19) a nd (3.22)) with the pattern obtained by Koeppenkastrop and De Carlo (1992) at pH 7.8. In order to compare iFelogK results on the same concentration scale, the partiti on coefficients obtaine d by Koeppenkastrop and De Carlo (1992) were calculated in term s of moles of amorphous ferric hydroxide per kilogram of seawater. The patte rns shown in Figure 3.12C are qu ite similar, especially for the HREEs. The iFelogK results of Koeppenkastrop and De Carlo (1992) are approximately 0.6 units smaller than the re sults predicted from equations (3.19) and (3.22). This difference is generally simila r to observed experime ntal errors in log iKFe observations (e.g., Figure 3.1). On the other ha nd, it might also be expected that the absolute magnitude of iFelogK observations in seawater could be diminished by competitive sorption of major cations (e.g., Mg2+).

PAGE 88

73 log iKFe 5.5 6.0 6.5 7.0 7.5 8.0 8.5 pH 82 pH 79 pH 76 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iK'Fe 4.2 4.4 4.6 4.8 5.0 5.2 5.4 5.6 5.8 this work (predicted) Koeppenkastrop and De Carlo (1992) A log iK'Fe 4.5 5.0 5.5 6.0 pH 82 pH 79 pH 76 B C Figure 3.12. (A) Distribution coefficients (log iKFe) expressed in terms of free YREE concentrations ([M3+]) using equation (3.19) and the surface complexation constants (Sn) listed in Table 3.2. (B) Distribution coefficients (iFelogK ) expressed in terms of total YREE concentrations in seawater (Msw) using equations (3.19) and (3.22) and assuming 3 3T[HCO]210 M. (C) Comparison between the predicted iFelogK pattern for seawater at pH 7.8 (equations (3 .19) and (3.22)) and the measured iFelogK pattern at pH 7.8 from Koeppenkastrop and De Carlo (1992). (Continued on next page).

PAGE 89

74(Figure 3.12 caption – continued). The results of De Carlo et al. (1998) (not shown) are consistent with the results of Koeppenkastrop and De Carlo (1992), but exhibit a somewhat reduced precision, perhaps due to e xperimental differences such as a two-fold decrease (De Carlo et al., 1998) in the experiment’s total YREE concentration. Distribution coefficients expressed in terms of total YREE concentrations in seawater (equation (3.22)) can be used to calculate the fraction of each YREE that would be removed from the water column, on a mille nnial scale, by settling amorphous ferric hydroxide. Column 2 in Table 3.3 lists iFelogK values at pH 7.9. Equation (3.22) written in the form: iT iFesw 3 S[MS] KM [Fe] (3.25) provides YREE/iron molar ratios ([MSi]T/[Fe3+]S) for YREEs sorbed onto amorphous ferric hydroxide, presuming that YREEs maintain their average steady state concentrations (Msw) in the water column. Using the Msw concentrations obtained by Zhang and Nozaki (1996) for each YR EE at a depth of 2469 m (column 3), [MSi]T/[Fe3+]S values (column 4) were calculated vi a equation (3.25). The authigenic flux of iron (3flux S[Fe]) to the ocean floor is estimated to range from 14 to 50 mol cm-2 kyr-1 (Krishnaswami, 1976; Thomson et al., 1984) Columns 5 and 6 show the predicted number of moles of each YREE that would be associated with authigenic iron fluxes equal to 14 and 50 mol cm-2 kyr-1, respectively (33flux iTSS([MS]/[Fe])[Fe] ). These YREE fluxes can be compared with authigenic REE fluxes estimated by Thomson et al. (1984), where authigenic is defined as non-te rrigenous and therefore may include phases other than amorphous ferric hydroxide. The au thigenic fluxes calculated by Thomson et al. (1984) are 10 to 50 times larger, except for Ce, which is 3 orders of magnitude larger, than the rates of YREE removal by authigen ic iron estimated in the present work. The number of moles of each YREE associated with the flux of authigenic iron (i.e., columns 5 and 6) can also be compared to the total inventory ( Msw) of each YREE in

PAGE 90

75 Table 3.3. Estimated removal rates for YREEs via authigenic iron. iFelogK (column 2) are distribution coefficients expressed in term s of total YREE concentrations in seawater (Msw) at pH 7.9 (equations (3.19) and (3.22)). The Msw values in column 3 are directly measured seawater concentrations for each YREE at 2469 m (Zhang and Nozaki, 1996). The [MSi]T/[Fe3+]S values in column 4 are YREE/iron molar ratios for each YREE sorbed by amorphous ferric hydroxide. The [MSi]T values in columns 5 and 6 are the number of moles of each YREE that accumulate in 1 cm2 of ocean floor per millennium of amorphous ferric hydroxide sorption and deposition. The [MSi]T/ Msw values in columns 7 and 8 are the ratios (percentages) of the number of moles of each YREE removed per millennium of amorphous ferric hydroxide sorp tion to the total inventory of each YREE in the water column. (a) Authigenic flux of iron equal to 14 mol cm-2 kyr-1 (Krishnaswami, 1976). (b) Authigenic flux of iron equal to 50 mol cm-2 kyr-1 (Thomson et al., 1984). [M3+] log iFeK Msw [MSi]T/[Fe3+]S [MSi]T [MSi]T/ Msw pmol/L mol M/mol Fe pmol cm-2 kyr-1 % a b a b Y 5.13 228.9 30.69 429.6 1534 0.47 1.68 La 5.42 28.93 7.68 107.6 384.2 0.93 3.32 Ce 5.63 4.48 1.93 27.0 96.4 1.51 5.38 Pr 5.70 3.95 1.97 27.6 98.7 1.75 6.24 Nd 5.71 17.58 9.00 126.1 450.2 1.79 6.40 Pm Sm 5.72 3.17 1.66 23.3 83.2 1.84 6.56 Eu 5.64 0.86 0.38 5.3 18.8 1.53 5.46 Gd 5.51 4.82 1.57 22.0 78.7 1.14 4.08 Tb 5.50 0.87 0.27 3.8 13.6 1.10 3.91 Dy 5.41 6.64 1.70 23.8 85.2 0.90 3.21 Ho 5.27 1.88 0.35 4.9 17.5 0.65 2.33 Er 5.20 6.62 1.05 14.8 52.7 0.56 1.99 Tm 5.19 1.03 0.16 2.2 8.0 0.54 1.94 Yb 5.27 6.85 1.28 17.9 63.8 0.65 2.33 Lu 5.16 1.24 0.18 2.5 9.0 0.51 1.82

PAGE 91

76 the 4-km column of water overlying a sq uare centimeter of the ocean floor ( Msw = swM400 L ). The number of moles of each YR EE removed per millennium of sorption by amorphous ferric hydroxide relative to the to tal inventory of each YREE in the water column ([MSi]T/ Msw) is summarized in columns 7 and 8. This calculation shows that the number of moles of each YREE in the wate r column is between 15 and 200 times greater than the number of moles of each YREE de livered to the sea floor per millennium by settling amorphous ferric hydroxide. Despite the range in flux values, the small fractions (%) shown in columns 7 and 8 suggest that sorption by amorphous ferric hydroxide and subsequent removal to the ocean floor is not a significant sink for YREEs in the open ocean. The sorptive removal of YREEs by amorphous ferric hydroxides in estuaries, where extensive iron colloid formation is gene rated by mixing of freshwater and seawater (Sholkovitz, 1976, 1992; Sholkovitz and Elde rfield, 1988) should greatly dominate sorptive removal by amorphous ferric hydroxides in the open ocean. 3.6 Conclusions This investigation of YREE sorption by amorphous ferric hydroxi de, in conjunction with the works of De Carlo et al. (1998) and Ohta and Kawabe (2000, 2001), indicates that log iKFe is linearly dependent on pH over a wi de range of conditions. Plots of log iKFe versus pH in the present work have slopes that range between 1.43 0.04 for La and 1.55 0.03 for Lu. These results, obtained at low ionic strengths (0.011 M I 0.043 M), are in good agreement with the results obtained by Ohta and Kawabe (2000, 2001) at I = 0.5 M. In view of the very similar stab ility constants of trivalent YREEs and Cu2+ (Smith and Martell, 1976, 1989) it is interesting to note that the absolute magnitudes of log iKFe(Cu2+) obtained by Swallow et al (1980) for pH > 5.5 is very similar to the log iKFe(YREE) results in the presen t study, and the slope of log iKFe(Cu2+) versus pH ( log iKFe/ pH = 1.5) is very similar to the slopes obtained for YREEs in the present work. Over an ionic strength range where the activit y coefficients of dissolved ions exhibit strong variations (0.0 M I 0.1 M), the results of this study indicate that YREE sorption constants (log iKFe) are nearly constant. This is consistent with previous

PAGE 92

77 observations (Swallow et al., 1980; Hayes a nd Leckie, 1987; Dzombak and Morel, 1990) of very weak dependencies of metal so rption coefficients on ionic strength. The log iKFe results of this study can be used to predict the sorption behavior of YREEs in seawater. Predicted YREE sorption behavior for seawater is generally similar to the REE sorption results observed by Koeppe nkastrop and De Carl o (1992) in seawater at pH 7.8. Although the rem oval of YREEs from seawater via sorption onto amorphous ferric hydroxide does not appear to be significant in the open ocean, it should be expected that oceanic YREE patterns are strongly influenced by the YREE sorption onto amorphous ferric hydroxides that occurs in estuaries.

PAGE 93

78 4. Sorption of Yttrium and Rare Earth Elem ents by Amorphous Ferric Hydroxide: Influence of Solution Co mplexation with Carbonate The following chapter has been peer-reviewed and will be published essentially in this form: Quinn K. A., Byrne R. H., and Schijf J. (in press) Geochimica et Cosmochimica Acta. 4.1 Abstract The influence of solution complexation on th e sorption of yttrium and the rare earth elements (YREEs) by amorphous ferric hydroxide was investigated at 25oC over a range of pH (4.0 – 7.1) and car bonate concentrations (0 M 2 3T[CO] 150 M). Distribution coefficients, defined as T iFeK= [MSi]T/(MT[Si]), where [MSi]T is the total concentration of sorbed YREE, MT is the total YREE concentration in solution, and [Si] is the concentration of amorphous ferric hydroxide, initially increased in magnitude with increasing carbonate concentration, and th en decreased. The initial increase of T iFeK is due to sorption of YREE carbonate complexes (3MCO ), in addition to sorption of free YREE ions (M3+). The subsequent decrease of T iFeK, which is more extensive for the heavy REEs, is due to the increasing intensity of YREE solution complexation by carbonate ions. The competition for YREEs between solution complexation and surface complexation was modeled via the equation: 3 3 333CO 12H2 S1S2S1CO13T T iFe H1H22 S1HCO13TCO13TCO23T[H][H][HCO][H] K (K[H]1)(1[HCO][HCO][H][HCO][H]) where S1 and S2 are equilibrium constants for free YREE surface species, 3CO S1 is the equilibrium constant for the YREE-carbonate surface species, SK1 is the surface protonation constant for am orphous ferric hydroxide, and 3HCO1 3H CO1, and 3H CO2 are

PAGE 94

79 YREE solution complexation constants expresse d in terms of bicarbonate concentrations. The equation, which includes (i) a single new constant (3CO S1 ) for each YREE, (ii) previously published sorption coefficients (S1 and S2) determined in the absence of carbonate, and (iii) previously published so lution complexation constants, precisely predicts both the absolute magnitude of T iFeK and the pattern of T iFeK values over our range of experimental conditions. Experimentally observed T iFeK values, spanning more than five orders of magnitude, are accu rately described by our surface/solution complexation model. The 3CO S1log values determined for each YREE in this work are: Y(-1.300.04), La(-0.390.02), Ce(-0.210.02), Pr(-0.220.02), Nd(-0.200.02), Sm (-0.200.02), Eu(-0.260.02), Gd(-0.380. 02), Tb(-0.400.02), Dy(-0.510.02), Ho (-0.570.02), Er(-0.590.02), Tm(-0.560.02), Yb(-0.620.02), and Lu(-0.590.02). 4.2 Introduction It is generally recognized that distributions of yttrium and the rare earth elements (YREEs) in the ocean are controlled by co mpetition between solution complexation and surface complexation. Since YREE solution chemistry has been relatively well characterized (see for instance Wood, 1990; Byrne and Sholkovitz, 1996), recent studies of YREE fractionation processes have fo cused on YREE surface chemistry. Early investigations of REE sorption in seawat er utilized radiotracers and a variety of substrates, both organic (Bi ngler et al., 1989; Byrne and Kim, 1990; Stanley and Byrne, 1990) and inorganic (Byrne and Kim, 1990; Ko eppenkastrop et al., 1991). These studies showed that for most substrates, light REEs (LREEs) are preferen tially removed from seawater compared to heavy REEs (HREEs). Silica phases, which displayed a greater affinity for HREEs (Byrne and Kim, 1990), we re an exception to this generality. The major limitation of these early YREE sorption investigations was the omission of many REEs whose radionuclides were too short-lived or not commercially available. Toward a more comprehensive view of YREE sorption in seawater, Koeppenkastrop and De Carlo (1992) examined sorption of all REEs, except Pm and Sm, onto amorphous ferric hydroxide and crystalline FeOOH. Despite mo re extensive sorption by the amorphous

PAGE 95

80 phase, the crystalline phase produced stronger fractiona tion and a residual seawater pattern that resembled shale-normalized R EE patterns in the ocean (Koeppenkastrop and De Carlo, 1992). As noted by Koeppenkastrop and De Carl o (1992), interpretation of experiments performed in seawater is complicated by th e presence of strong solution complexation. As such, it was recognized that experiments should be undertaken in simple synthetic media in the absence of strongly complexing li gands. Starting with the work of De Carlo et al. (1998), (Y)REE sorption onto amorphou s ferric hydroxide in simple synthetic solutions (without complexing ligands) has been investigated over a range of pH (4.0 – 9.0) and ionic strength (0 – 0.7 M) (Bau, 1999; Kawabe et al., 1999b; Ohta and Kawabe, 2001; Quinn et al., 2004, 2006a). In general these experiments showed that, in the absence of solution complexation, sorption does not preferentially remove LREEs from solution. Quinn et al. (2004) showed that the YREE pattern obtained in experiments at near-neutral pH closely resembles the sorption pattern of natural marine particles that is predicted (Byrne and Sholkovitz, 1996) using shale-normalized oceanic YREE concentrations and a quantit ative model of YREE solution complexation in seawater. It has been well established that YREE so rption is strongly influenced by pH. In addition to an increase in the absolute magn itude of YREE sorption with increasing pH, Bau (1999) showed that there is a pH depe ndence in the pattern of YREE fractionation. Based on experimental results from Eu and La sorption onto hematite, Rabung et al. (1998a) and Marmier and Fromage (1999) used a surface complexation model to describe sorption intensity as a function of pH. Exte nding the work of Rabung et al. (1998a) and Marmier and Fromage (1999) to include th e entire YREE series, Quinn et al. (2006a) modeled YREE distribution coefficient results (3.9 pH 7.1) in terms of free ion (M3+) sorption with a two-site surface complexation model. Relatively few studies have compared YREE sorption in the absence and presence of solution complexation. Fairhurst et al. ( 1995) and Rabung et al. (1998b) showed that Eu3+ sorption onto hematite was suppressed at pH > 5.0 in the presence of humic acid and fulvic acid. At lower pH values, Eu3+ sorption was enhanced to varying degrees, depending on the concentration of humic acid (Fairhurst et al., 1995) Davranche et al.

PAGE 96

81 (2004) studied sorption of th e entire REE series onto ir on oxyhydroxide. A flat YREE sorption pattern was observed in the presen ce of humic acid, co mpared to an HREEenriched pattern in the absence of soluti on complexation (Davranche et al., 2004). YREE sorption in these studies was interpreted in terms of complexation w ith humate, with the latter being both dissolved in solution and sorbed onto hematite (Fairhurst et al., 1995; Rabung et al., 1998b; Davranche et al., 2004). Despite the fact that YREE solution complexation in the open ocean appears to be dominated by carbonate ions (B yrne and Sholkovitz, 1996), it s direct role in YREE sorption is poorly understood. Koeppenkast rop and De Carlo ( 1993) observed that carbonate complexation slowed the rate of upt ake of Eu by manganese and iron oxides. Based on their observations of sorption kine tics, Koeppenkastrop and De Carlo (1993) proposed that dissolved REEs dissociate from carbonate liga nds before being sorbed as free ions onto a solid. Kawabe et al. (1999a ) and Ohta and Kawabe (2000) investigated YREE sorption onto amorphous ferric hydroxide in the presence of carbonate over a narrow pH range (7.6 – 8.7) at an ionic stre ngth of ~0.5 M. Their results showed that HREE sorption was strongly suppressed in the presence of strong carbonate complexation. Despite the fact that YREE solution chemistry is relatively well understood compared to YREE surface chemistry, Ohta and Kawabe (2000) used their distribution coefficient results along with a theoretical model of surface complexation to derive YREE-carbonate solution complexation constants. As discussed by Luo and Byrne (2004), the results obtained by Ohta and Kawa be (2000) are approximately an order of magnitude larger than previous results obtain ed using a variety of procedures: solubility (e.g., Ferri et al., 1983), solvent exchange (e.g., Liu and Byrne, 1998), and potentiometry (e.g., Luo and Byrne, 2004). In the present study, we have examined th e effect of carbonate solution complexation on YREE sorption by amorphous ferric hydroxide at low ionic strength (I < 0.1 M) over a relatively wide range of pH (4 .0 – 7.1). Distribution coeffi cient results are quantitatively examined using the surface complexation model of Quinn et al. (2006a) and the carbonate complexation constants of Luo and Byrne (2004). Experimental results are

PAGE 97

82 used to extend the model of Quinn et al. (2006a) to include sorption of YREE solution complexes (i.e., 3MCO) in addition to sorption of free YREE ions (M3+). 4.3 Theory Measurements of YREE solution concentratio ns in the presence of freshly precipitated amorphous ferric hydroxide and dissolved carbonate (0 M 2 3T[CO] 150 M) were used to calculate distribution coefficients (T iFeK) in the following form: T iT iFe Ti[MS] K M[S] (4.1) where [MSi]T is the total molar concentration of a sorbed YREE, MT is the total molar concentration of a dissolved YREE, and [Si] is the total molar concentration of precipitated amorphous ferric hydr oxide. The total concentra tion of a sorbed YREE can be written as the sum of three or more terms. As one example, in solutions containing carbonate, [MSi]T can be written as: 2 0 iT2223[MS][SFeO(OH)M][SFeO(OH)M][SFeO(OH)MCO] (4.2) The first two terms on the right-hand side of equation (4.2) follow from the work of Quinn et al. (2006a) in carbonate -free solutions. The final term in equation (4.2) is one of a number of potentially important surfacebound YREE species. Equilibrium constants for the formation of S–FeO(OH)2M2+, S–FeO2(OH)M+, and 0 23SFeO(OH)MCO can be written, respectively, as: 2 2 S1 3 3[SFeO(OH)M][H] [M][SFe(OH)] (4.3) 2 2 S2 3 3[SFeO(OH)M][H] [M][SFe(OH)] (4.4) and 30 CO 23 S1 33[SFeO(OH)MCO][H] [MCO][SFe(OH)] (4.5) where brackets denote concentrations of the indicated species and S–Fe(OH)3 represents uncharged amorphous ferric hydroxide surf ace sites (as distinguished from 2SFe(OH)

PAGE 98

83 and 4SFe(OH) ). Under the conditions of our experiments, 4SFe(OH) is unimportant (Quinn et al., 2006a) and the concentration of S–Fe(OH)3 in equations (4.3), (4.4), and (4.5) can be expressed in terms of [Si] (equation (4.1)) via the equation: 1 3iS1[SFe(OH)][S](K[H]1) (4.6) where SK1 is the surface protonation consta nt for amorphous ferric hydroxide: 2 S1 0 3[SFe(OH)] K [SFe(OH)][H] (4.7) The value of SK1 used in this study (i.e., log SK1 = 4.76) was taken from the work of Quinn et al. (2006a). For carbonate-free solutions (i.e., 0 23[SFeO(OH)MCO] = 0 M), equations (4.1– 4.4), (4.6), and (4.7) were used by Quinn et al. (2006a) to model YREE sorption in the absence of significant solution complexa tion. In the presence of YREE carbonate complexation, additional sorbed species must be considered in equation (4.2) including the putative species 0 23SFeO(OH)MCO (equation (4.5)). Additionally, the sorption model of Quinn et al. (2006a) must be extende d to include the relationship between total dissolved YREE concentrations (MT) and free YREE concentrations ([M3+]) as follows: 3333H1H22 THCO13TCO13TCO23TM[M](1[HCO][HCO][H][HCO][H]), (4.8) where the YREE solution complexation constants (333HH HCO1CO1CO2, and ) are expressed in terms of bicarbonate co ncentrations (Luo and Byrne, 2004): 32 3 HCO1 3 3T[MHCO] [M][HCO] (4.9) 3H 3 CO1 3 3T[MCO][H] [M][HCO] (4.10) and 32 H 32 CO2 32 3T[M(CO)][H] [M][HCO] (4.11) and 3T[HCO] is the sum concentration of free bicarbonate ions (3HCO) and ion pairs (0 3NaHCO). A term for the formation of MOH2+ is not included in equation (4.8) since

PAGE 99

84 Quinn et al. (2006a) showed that, even in the absence of carbonate complexation, the influence of hydrolysis on the behavior of T iFeK at pH 7.0 is insignificant. Equations (4.1) through (4.11) can be comb ined to produce an equilibrium model for YREE sorption by amorphous ferric hydroxide in the presence of carbonate: 3 3 333CO 12H2 S1S2S1CO13T T iFe H1H22 S1HCO13TCO13TCO23T[H][H] [HCO][H] K (K[H]1)(1[HCO][HCO][H][HCO][H]) (4.12) Empirical T iFeK data, as defined by equation (4.1), we re fit using equati on (4.12) with the residual sum of squares (RSS) function as follows: 3 3CO 12H2 S1S2S1CO13T S1[H][H] [HCO][H] RSS1 K[H]1 (4.13) 2 3 1 T iFe T[M] K M Defined in this manner, the RSS provides equal weight to each experimental T iFeK result as distribution coefficients range over more than five orders of magnitude. The carbonate complexation constants in equation (4.12) were taken from the results of Luo and Byrne (2004): 3300 50 5 HCO1HCO1loglog3.066I/(11.269I)0.297I (4.14) 33HH 00 50 5 CO1CO1loglog4.088I/(13.033I)0.042I, (4.15) and 33HH 00 50 5 CO2CO2loglog4.088I/(13.033I)0.042I, (4.16) where the values of 30 HCO1log 3H0 CO1log and 3H0 CO2log for each YREE can be found in Table 5 of Luo and Byrne (2004) Bicarbonate concentrations were calculated from the equation: 21 3T01CO[HCO]KKP[H] (4.17) where the product 01KK describes the equilibrium:

PAGE 100

85223CO(g)HOHCOH. (4.18) The CO2 partial pressure (2COP ) in equation (4.17) is expressed in terms of the total atmospheric pressure (PT), the partial pressure of H2O at 25oC (2HOP), and the mole fraction of CO2/N2 gas mixtures (2COX) using the following equation: 222 2COCOTHO COPX(PP)0.969Xatm (4.19) 01KK data appropriate to equation (4.18) were taken from the results of Luo and Byrne (2004): 05 05 01logKK7.8291.022I/(11.390I)0.191I (4.20) For the purpose of creating graphs, carbonate concentrations were calculated using the following equation: 222 3T012CO[CO]KKKP[H] (4.21) where 2 3T[CO] is the sum concentration of free carbonate ions (2 3CO) and ion pairs (3NaCO), and 2K is the equilibrium constant fo r the dissociation of bicarbonate: 2 33HCOCOH. (4.22) 2K was calculated from the results of Luo and Byrne (2004): 05 05 2logK10.3312.044I/(11.060I)0.184I (4.23) Equations (4.14) through (4.23) explicitly show the substa ntial ionic strength dependences for equilibria in th e solution phase. In contrast, a wide variety of previous work has shown that the affinities of sorptive solid substrates for dissolved cations do not vary with ionic strength (Swallow et al., 1980; Hayes and Leckie, 1987; Dzombak and Morel, 1990; Quinn et al., 2006a). The data of Quinn et al. (2006 a) showed that the influence of ionic strength on S1 and S2 (equations (4.3) and (4.4)) was very weak. Based on these observations appropriate to the YREEs, and a vari ety of observations obtained using other cations (Swallow et al., 1980; Haye s and Leckie, 1987), it was assumed in this work that not only S1 and S2 but also 3CO S1 (equation (4.5)) was invariant over the range of ionic strength utilized in this investigation (0.01 M I

PAGE 101

86 0.1 M). It should be noted in this case that the product 3 3CO H S1CO1 has an ionic strength dependence identical to that of 3H CO1 (equation (4.15)). 4.4 Materials and Methods Three types of experiment were undertaken to investigate the influence of carbonate solution complexation on YREE sorption by amor phous ferric hydroxide. In one type of experiment, sorption was examined as a func tion of time at constant pH and constant 2COP In the other types of experiment, either solution pH was increased at constant 2COP or the 2COP was increased at constant pH. All solutions were prepared with trace metal-clean water (Milli-Q water) from a Millipore (Bedford, MA) purification system. Ammonium nitrate (99.999%) and certified 1.000 M hydrochloric acid were purchased fr om Sigma-Aldrich (St. Louis, MO). TraceMetal Grade nitric acid, TraceMetal Grade ammonium hydroxide, and ferric chloride solution (40% w/v in HCl) were pur chased from Fisher Scientific (Pittsburgh, PA). Sodium bicarbonate (Baker Analyzed ) was purchased from J.T. Baker Inc. (Phillipsburg, NJ). A YREE stock solution, containing 66.7 ppm of each YREE in 2% HNO3, was prepared from single-element ICP standards (SPEX CertiPrep, Metuchen, NJ). Ultra-pure N2 and various certified CO2/N2 gas mixtures (30%, 3%, 1%, 0.5%, 0.3%, 0.1%, and 0.01% CO2) were obtained from Airgas South Inc. (Clearwater, FL). All chemical manipulations were performe d in a class-100 clean air laboratory or laminar flow bench. Teflon and polypropylen e laboratory materials and polycarbonate filter membranes were cleaned by soaking in HCl or HNO3 for at least a week, followed by several thorough rinses with Milli-Q wa ter. Solution pH, on the free hydrogen ion scale, was monitored using a Ross-type combination pH electrode (No. 810200) connected to a Corning 130 pH meter in the ab solute millivolt mode. Nernstian behavior of the electrode was verified periodically by titrating a 0.3 M NaCl solution with concentrated HCl. At the beginning of each experiment, a pH standard solution and an experimental solution, both with an ionic strength (I) equa l to 0.011 M, were prepared in Teflon wide-

PAGE 102

87 mouth bottles. The pH standard solution was composed of 1 mM HCl (pH 3.0) in 0.01 M NH4NO3. The experimental solution was composed of 107.8 M ferric iron and 23.3 ppb of each YREE ([YREE]T = 2.36 M) in 0.01 M HCl. Both solutions were placed in jacketed beakers thermostated at T = (25.0 0.1)oC and were equilibrated for approximately 24 hours. Throughout each experiment, solutions were continuously stirred with a Teflon-coated ‘floating’ st ir bar and the experimental solution was continuously bubbled with a gas mixture, exce pt during titrant additions. Ultra-pure N2 gas was first passed through an in-line trap (Supelco, Bella fonte, PA) that removed all traces of CO2. After bubbling for one hour at pH 2.0, starting with ultra-pure N2 for experiments conducte d over a range of 2COP and with either 3% or 30% CO2 for experiments at constant 2COP, an initial solution sample wa s taken to determine the total dissolved YREE concentration, MT. Solution pH was then increased by addition of 0.7 M NaHCO3 with a Gilmont micro-disp enser, resulting in rapid formation of a yellow-brown Fe(OH)3 colloid. One experiment was performed at constant 2COP (30%) and constant pH (5.4), and samples were taken at 15 minutes, 90 minutes 5 hours, 24 hours, 46 hours, and 48 hours. Two experiments were performed at constant 2COP and increasing pH: one at 3% CO2 and the other at 30% CO2. Samples were taken at fixed pH increments between 4.0 and 6.6 after the solution had been equilibrate d with the gas mixture for one hour. Four experiments were performed at increasing 2COP and constant pH: two at pH 6.6 and two at pH 7.1. After taking the initial sample (pH 2.0), the pH was raised by addition of 1 M NH4OH using a Gilmont micro-dispenser. While bubbling with ultra-pure N2, four samples were taken: one at 15 minutes, one at 90 minutes, one at 5 hours, and one at 22 hours. Subsequently, CO2/N2 gas mixtures were used to progressively increase 2COP between 0.01% and 30% CO2 (two experiments) and between 0.3% and 30% CO2 (two experiments). After each 2COP increase, solutions were equilibrated for approximately one hour. The pH was then readju sted by addition of 1 M NaHCO3. At each 2COP, samples were taken at 15 minutes, 90 minutes, and either 22 hours or between 45 and

PAGE 103

88 70 hours. Occasionally, a fourth sample was take n at 47 or 66 hours. Because increases in the carbonate concentration caused increases in the ionic strength of the experimental solutions, the ionic strengths of pH sta ndard solutions were matched using 1 M NH4NO3. The sampling method was similar to that described in Quinn et al. (2006a). To summarize, during most experiments two samp les were collected, one filtered and one centrifuged. During a few experiments, only filtered samples were taken because better phase separation was achieved with filtration. Each filtered sample consisted of two 5mL aliquots of solution. The first was used to rinse the polypropylene syringe and the Nuclepore filter membrane (polycarbonate, 0. 10 m pore size). The second was collected in a polypropylene centr ifuge tube. The centrifuged sample consisted of one 5-mL aliquot of solution, which was centrifuged for one hour using a Centra-4B centrifuge (International Equipment Company, Nee dham Heights, MA) at about 2200 g. The filtered samples and the supernatant of the centrifuged samples were diluted 5fold with 1% HNO3 except where concentrations were below the lowest calibration standard (0.5 ppb), in which case no dilution was performed. A small amount of internal standard solution containing e qual concentrations of In, Cs and Re was added to each sample. The resulting mixtures were analy zed for YREEs with an Agilent Technologies 4500 Series 200 inductively-coup led plasma mass spectromet er (ICP-MS) following the procedure outlined in Quinn et al. (2004). In brief, all standards and sample solutions were injected in triplicate. During instrument tuning, the formation of oxide and doublecharged ions was minimized with a 10 ppb Ce solution. MO+ and M2+ peaks were always less than 1% and 3% of the corresponding M+ peak, respectively, and correction for this effect proved unnecessary. YREE concentrations were calculated from linear regressions of four standards (0.5, 1, 2, and 5 ppb). Ion counts were co rrected for minor instrument drift by normalizing 89Y to 115In, 139La–161Dy to 133Cs, and 163Dy–175Lu to 187Re. To check the validity of the drift correcti on, a comparison was made of the Dy concentrations calculated from 161Dy and 163Dy, which were usually identical within 2%. For each experiment, raw ICP-MS data we re corrected for dilution based on the volume of NH4OH and/or NaHCO3 titrants added to adjust the pH. Corrected data were then used to calculate di stribution coefficients (T iFeK) defined by equation (4.1). The

PAGE 104

89 concentration of sorbed YREE, [MSi]T, was calculated as the difference between the YREE concentration in the initial sample (pH 2.0) and the YREE concentrations in subsequent samples after a pH or 2COP adjustment. Based on the solubility behavior of Fe3+ (Liu and Millero, 1999), the concentrati on of precipitated iron at pH > 4.0 was assumed to be equal to the ini tial dissolved iron concentration ( 100 M). Quinn et al. (2006a) noted th at YREE equilibrium between experimental solutions and freshly preci pitated Fe(OH)3 is reached in about 15 mi nutes. In the present work, variations in T iFelogK for equilibration times 15 minutes are smaller than the uncertainty in experimental T iFelogK values. Therefore all data in Appendix D, which lists 111 T iFelogK observations for each rare earth el ement, were used in our data analysis, except for two observations identified in Table D.3. These were obtained under conditions (pH 3.98 and pH 4.49) that produced very weak so rption, and therefore poorly constrained T iFelogK data. This problem was expected a priori, from the work of Quinn et al. (2006a), in which well defined T iFelogK results at low pH ( 4.0) were obtained by conducting experiments using 10 mM concentr ations of precipitated amorphous ferric hydroxide. Utilization of the da ta in Tables B.1–B.6 along with the data shown in Appendix D resulted in regressions via equations (4.12) and (4 .13) that incorporated as many as 166 T iFelogK observations for each REE. 4.5 Results and Discussion 4.5.1. Model Results Considering Sorption of Only Free YREEs Since carbonate-free samples were in cluded in the present experiments, S1 and S2 values were recalculated using new non-carbonate T iFeK data (Tables D.4–D.7) plus the previous non-carbonate T iFeK data in Tables B.1–B.6. The S1 and S2 results calculated using equations (4.12) and (4.13) with 3T[HCO] = 0 M are listed in Table 4.1. Comparison of these results with the S1 and S2 results of Quinn et al. (2006a)

PAGE 105

90Table 4.1. YREE surface complexation constants (S1 and S2) determined using equations (4.12) and (4.13) with 3T[HCO] = 0 M, log SK1 = 4.76 (Quinn et al., 2006a), and the experimental distribu tion coefficient results from carbonate-free solutions in the present work (Tables B.1–B.6 and D.4–D.7). Un certainties represent one standard error. [M3+] s1log s2log Y -2.98 0.06 -8.86 0.05 La -2.87 0.04 -9.36 0.07 Ce -2.38 0.04 -8.86 0.08 Pr -2.26 0.04 -8.63 0.06 Nd -2.18 0.04 -8.55 0.07 Pm Sm -2.06 0.04 -8.31 0.06 Eu -2.11 0.04 -8.33 0.05 Gd -2.28 0.04 -8.56 0.06 Tb -2.29 0.04 -8.40 0.05 Dy -2.32 0.03 -8.38 0.04 Ho -2.37 0.03 -8.46 0.05 Er -2.33 0.03 -8.43 0.05 Tm -2.24 0.03 -8.32 0.04 Yb -2.17 0.03 -8.19 0.04 Lu -2.17 0.03 -8.26 0.05 demonstrates agreement within approximately 1%, well within the listed uncertainties for both constants. Distribution coefficient result s from experiments containi ng carbonate were initially modeled by assuming that only free YREE ions sorb onto am orphous ferric hydroxide. In this case, the 3CO S1 term in equation (4.12) is zero. Figure 4.1A shows T iFelogK patterns at pH 7.06 predicted using equation (4.12) with the S1 and S2 results listed in Table 4.1 and 3CO S1 = 0. Predicted T iFelogK values decrease mono tonically with increasing 2 3T[CO], and the decrease in T iFelogK for heavy REEs is approximately four orders of magnitude.

PAGE 106

91 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKT Fe(pred) 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKT Fe(meas) 4.5 5.0 5.5 6.0 6.5 7.0 7.5 A B 0.0 M 0.54 M 1.64 M 5.86 M 114.7 M Figure 4.1: T iFelogK results at pH 7.06 and va rious carbonate concentrations, 2 3T[CO], listed in the legend. (A) T iFelogK(pred) are distribution coefficients predicted from equation (4.12) using the S1 and S2 results listed in Table 4.1 and 3CO S1 = 0. (B) T iFelogK(meas) are directly measured distribution coefficients from an experiment performed at constant pH (7.06) and increasing 2COP (Table D.6). Each pattern represents an average over time for a single carbonate concentration. For clarity, the T iFelogK pattern at 2 3T[CO] = 0.88 M (2COP = 0.5%) is not shown.

PAGE 107

92 Figure 4.1B shows experimental T iFelogK results obtained at pH 7.06 for the same carbonate concentrations depicted in Figure 4.1A. In sharp contrast to the predicted behavior shown in Figure 4.1A, measured T iFelogK values at low carbonate concentrations are larger than those at 2 3T[CO] = 0 M. Furthermore, the range of T iFelogK values shown in Figure 4.1B is orders of magnitude smaller than the predictions shown in Figure 4.1A. Predicted and observed T iFelogK values are directly and quantitatively compared in Figure 4.2. In the absence of carbonate (open circles), T iFelogK values are well described using equa tion (4.12). In the presence of carbonate (closed circles), T iFelogK observations are uniformly larger than equation (4.12) T iFelogK predictions obtained assuming 3CO S1 = 0. Figure 4.2 clearly shows that YREE sorption data in the presence of carbonate cannot be a ppropriately modeled sole ly in terms of the sorption of free ions, M3+. 4.5.2. Model Results Including Sorption of a YREE Carbonate Complex Non-linear least squares regressions (equatio ns (4.12) and (4.13)) of the combined (carbonate plus non-carbonate) T iFelogK data obtained in this work (Tables B.1–B.6 and D.1–D.7) produced well-constrained estimates for 3CO S1 as well as S1 and S2 (Table 4.2). Figures 4.3A, 4.3B, and 4.3C provi de graphical representations of the S1, S2, and 3CO S1 data given in Table 4.2 (open circles). Al so shown in Figures 4.3A and 4.3B are the S1 and S2 data given in Table 4. 1 (closed circles). The 3CO S1 results in Figure 4.3C (closed circles) were obtained in fits (e quations (4.12) and (4.13)) of data at 2 3T[CO] > 0 M using the S1 and S2 values from Table 4.1. It is se en in Figures 4.3A and 4.3B that log S1 and log S2 results obtained in both two-parameter fits (S1 and S2 in Table 4.1) and three-parameter fits (S1, S2, and 3CO S1 in Table 4.2) are in very good agreement. Figure 4.3C shows that 3CO S1log results from three-parameter fits (S1, S2, and 3CO S1 in

PAGE 108

93 02468 log iKT Fe(meas) 0 2 4 6 8 La 02468 log iKT Fe(meas) 0 2 4 6 8 Sm log iKT Fe(pred) 02468 log iKT Fe(meas) 0 2 4 6 8 Dy log iKT Fe(pred) 02468 log iKT Fe(meas) 0 2 4 6 8 Lu Figure 4.2: T iFelogK(meas) versus T iFelogK(pred) for La, Sm, Dy, and Lu. T iFelogK(meas) are directly measured distribu tion coefficients from the present work (Tables B.1–B.6 and D.1–D.7). Observed T iFelogK values represent YREE sorption corresponding to 5.0% – 99.9% of the total YREE concentration. T iFelogK(pred) are distribution coefficients predicte d from equation (4.12) using the S1 and S2 results listed in Table 4.1 and 3CO S1 = 0. Open circles represent carbonate-free samples and closed circles represent samples containing carbona te. Diagonal lines represent perfect agreement between predicted and measured values (T iFelogK(pred) = T iFelogK(meas)).

PAGE 109

94 Table 4.2. YREE surface complexation constants (S1, S2, and 3CO S1 ) determined with equations (4.12) and (4.13), log SK1 = 4.76 (Quinn et al., 2006a), and the experimental distribution coefficient results from the present work (T ables B.1–B.6 and D.1–D.7). Uncertainties represent one standard error. [M3+] s1log s2log 3CO s1log Y -2.98 0.07 -8.82 0.05 -1.30 0.04 La -2.86 0.03 -9.34 0.06 -0.39 0.02 Ce -2.38 0.04 -8.84 0.07 -0.21 0.02 Pr -2.25 0.03 -8.60 0.06 -0.22 0.02 Nd -2.17 0.04 -8.53 0.06 -0.20 0.02 Pm Sm -2.05 0.04 -8.29 0.05 -0.20 0.02 Eu -2.11 0.04 -8.31 0.05 -0.26 0.02 Gd -2.28 0.03 -8.54 0.05 -0.38 0.02 Tb -2.29 0.03 -8.38 0.04 -0.40 0.02 Dy -2.32 0.03 -8.36 0.04 -0.51 0.02 Ho -2.36 0.03 -8.44 0.04 -0.57 0.02 Er -2.33 0.03 -8.40 0.04 -0.59 0.02 Tm -2.24 0.03 -8.29 0.04 -0.56 0.02 Yb -2.17 0.04 -8.16 0.04 -0.62 0.02 Lu -2.17 0.03 -8.23 0.04 -0.59 0.02

PAGE 110

95 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log S1 -3.2 -3.0 -2.8 -2.6 -2.4 -2.2 -2.0 -1.8 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log S2 -9.6 -9.4 -9.2 -9.0 -8.8 -8.6 -8.4 -8.2 -8.0 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log CO3S1 -1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 A B YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log (CO3S1 CO3H1 0) -4.4 -4.2 -4.0 -3.8 -3.6 -3.4 -3.2 -3.0 -2.8 C D Figure 4.3: Surface stability constants (equa tion (4.12)) for YREE sorption by amorphous ferric hydroxide. Op en circles in panels A, B, and C represent the S1, S2, and 3CO S1 results (Table 4.2) obtaine d in three-parameter fits using equations (4.12) and (4.13). Open circles in panel D represent the product 3 3CO H0 S1CO1 obtained by multiplying the 3CO S1 results from the present work and the 3H0 CO1 results from Luo and Byrne (2004). Closed circles in panels A and B represent the S1 and S2 results (Table 4.1) obtained in two-parameter fits us ing equations (4.12) and (4.13) with 3T[HCO] = 0 M. Closed circles in panel C represent the 3CO S1 results obtained in single-parameter fits using equations (4.12) and (4.13). See te xt for details. Error bars on the open circles represent standard errors. For 3CO S1 standard errors are within the size of the symbol. For 3 3CO H0 S1CO1 error bars were determined by statis tically combining the standard errors from both stability constants.

PAGE 111

96 Table 4.2) and from single-parameter fits (S1 and S2 taken from Table 4.1) are indistinguishable. The pattern in Figure 4.3D, which represents the product 3 3CO H S1CO1 used in equation (4.12), was obtained by mu ltiplying the formation constant for the surface species 0 23SFeO(OH)MCO (3CO S1 ; Table 4.2) and the formation constant for the solution species 3MCO at zero ionic strength (3H0 CO1 ; taken from Luo and Byrne, 2004). It should be noted that the pattern for 3 3CO H S1CO1 will not change as a function of ionic strength but the absolute magnitude has an ionic strength de pendence identical to that of 3H CO1 (equation (4.15)). As a visual demonstration of the goodness-of -fit for the model, Figure 4.4 compares observed T iFelogK data in the absence (open circles) and presence (closed circles) of carbonate with T iFelogK data predicted using equation (4 .12) and the parameters given in Table 4.2. The four REE shown in Figure 4. 4 are representative of the entire YREE series, which all display excellent fits with slopes close to one and intercepts close to zero. It can be seen that YREE sorption by amorphous ferric hydroxide in the presence of carbonate is well described by accounting for solution complexation (2 3MHCO, 3MCO and 32M(CO) formation), and the formation of three surface-bound YREE species (S–FeO(OH)2M2+, S–FeO2(OH)M+, and 0 23SFeO(OH)MCO ). In addition to predicted versus observed T iFelogK comparisons for individual YREEs, it is also informative to examine predicted versus observed patterns for the entire YREE series. In Figure 4.5, directly measured T iFelogK patterns are compared with T iFelogK patterns predicted from equation (4.12) using the S1, S2, and 3CO S1 results listed in Table 4.2. The T iFelogK patterns shown in Figure 4.5 were selected from Appendix D to represent the progression of shapes observed over the ra nge of carbonate concentrations used here. It can be seen that these shapes are generally well pred icted using equation (4.12) and the data given in Table 4.2. A lthough positive and negative deviations between

PAGE 112

97 02468 log iKT Fe(meas) 0 2 4 6 8 Lay = (0.970.01)x + (0.270.05) r2 = 0.983 02468 log iKT Fe(meas) 0 2 4 6 8 Smy = (1.010.01)x + (0.100.05) r2 = 0.984 log iKT Fe(pred) 02468 log iKT Fe(meas) 0 2 4 6 8 Dyy = (1.000.01)x + (0.120.05) r2 = 0.985 log iKT Fe(pred) 02468 log iKT Fe(meas) 0 2 4 6 8 Luy = (1.000.01)x + (0.110.05) r2 = 0.985 Figure 4.4: Regressions of T iFelogK(meas) versus T iFelogK(pred) for La, Sm, Dy, and Lu. T iFelogK(meas) are directly measured distribution coefficients from the present work (Tables B.1–B.6 and D.1–D.7). T iFelogK(pred) are distribution coefficients predicted from equation (4.12) using the S1, S2, and 3CO S1 results listed in Table 4.2. Open circles represent carbonate-free samples and clos ed circles represent samples containing carbonate. Dashed lines represen t 95% confidence intervals.

PAGE 113

98 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKT Fe 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 pH = 5.70 [CO3 2-]T = 9.64 nM YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKT Fe 3.8 4.0 4.2 4.4 4.6 4.8 5.0 5.2 pH = 5.70 [CO3 2-]T = 103 nM YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKT Fe 5.2 5.4 5.6 5.8 6.0 6.2 6.4 6.6 pH = 6.69 [CO3 2-]T = 974 nM YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKT Fe 5.4 5.6 5.8 6.0 6.2 6.4 6.6 6.8 7.0 pH = 7.06 [CO3 2-]T = 5770 nMA B C D Figure 4.5: T iFelogK patterns covering a range of carbonate co ncentrations. Closed circles represent distribution coefficients ex perimentally observed in the present work (Tables D.2, D.3, D.5, and D.6). Open circle s represent distribution coefficients predicted from equation (4.12) using the S1, S2, and 3CO S1 results listed in Table 4.2. predicted and measured values are seen in Fi gure 4.5, no systematic differences were observed for the T iFelogK patterns obtained in this investigation. Other than the species 0 23SFeO(OH)MCO two additional terms (23SFeO(OH)MHCO and 23SFeO(OH)MCO ) were considered in the equation (4.2) summation for [MSi]T. The surface complexation constants for these two species can be written as: 3HCO 23 S1 2 33[SFeO(OH)MHCO][H] [MHCO][SFe(OH)] (4.24)

PAGE 114

99 and 32 CO 23 S2 33[SFeO(OH)MCO][H] [MCO][SFe(OH)] (4.25) Replacement of the term for 0 23SFeO(OH)MCO formation (equation (4.5)) in equation (4.12) with terms for either 23SFeO(OH)MHCO or 23SFeO(OH)MCO formation (equations (4.24) and (4.25)) produced resi dual sum of squares (RSS) results much inferior to those obtained using equation (4 .13). Furthermore, inclusion of terms for 23SFeO(OH)MHCO and 23SFeO(OH)MCO in addition to the term for 0 23SFeO(OH)MCO led to insubstantial improvements relative to fits with only three surface terms (S1, S2, and 3CO S1 ). Equation (4.12) provides a robust descri ption of the data obtained in this investigation. Additional sorption terms may be requi red at higher pH and higher carbonate concentrations than were investigated in the present work. Tang and Johannesson (2005) reported 32M(CO) sorption on Carrizo sand for pH > 7.3. Under the conditions of our experiments (pH 7.15), sorption of 32M(CO) was not required to de scribe partitioning of YREEs between the aqueous pha se and amorphous ferric hydroxide. 4.5.3 Examination of the Competing Influences of Surface and Solut ion Complexation on T iFeK The distribution coefficients predicted from equation (4.12) can be separated into contributions from solution species (2 3MHCO 3MCO and 32M(CO)) and surface species (S–FeO(OH)2M2+, S–FeO2(OH)M+, and 0 23SFeO(OH)MCO ). This is shown by rearranging equation (4.1) as follows: T iTT iFe 33 i[MS]M logKloglog [M][S][M]. (4.26) The first term on the right-ha nd side of equation (4.26) describes the affinity of amorphous ferric hydroxide for free dissolved YREE ions (M3+). Using equations (4.2– 4.6) and (4.10), this term is written as:

PAGE 115

100 3 3CO 12H2 S1S2S1CO13T iT 3 iS1[H][H] [HCO][H] [MS] [M][S] K[H]1 (4.27) The second term in equation (4.26) is the complexation intensity of YREEs in solution (i.e., equation (4.8)). This term is a measur e of the relative proportions of YREEs that remain in solution as free ions. The compe titive influences of surface versus solution complexation on observed T iFelogK patterns (Figure 4.5) ar e shown in Figures 4.6 and 4.7, respectively. The patterns shown in Figure 4.6, which are calculated with equation (4.27), are relatively constant over a wide range of cond itions. The uniformity of these patterns is due to the fact that the terms S1, S2, and 3 3CO H S1CO1 in equation (4.27) have very YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log ([MSi]T[M3+]-1[Si]-1) 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log ([MSi]T[M3+]-1[Si]-1) 42 4.4 4.6 4.8 5.0 52 5.4 5.6 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log ([MSi]T[M3+]-1[Si]-1) 6.2 6.4 6.6 6.8 7.0 7.2 7.4 7.6 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log ([MSi]T[M3+]-1[Si]-1) 72 7.4 7.6 7.8 8.0 82 8.4 8.6 pH = 5.70 [CO3 2-]T = 9.64 nM pH = 5.70 [CO3 2-]T = 103 nM pH = 6.69 [CO3 2-]T = 974 nM pH = 7.06 [CO3 2-]T = 5770 nMA B C D Figure 4.6: Patterns of the surface complexation term (log ([MSi]T[M3+]-1[Si]-1)) in equation (4.26) calculate d with equation (4.27).

PAGE 116

101 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log (MT/[M3+]) 0.0 0.5 1.0 1.5 2.0 2.5 A B C D[CO3 2-]T = 974 nM [CO3 2-]T = 5770 nM [CO3 2-]T = 9.64 nM [CO3 2-]T = 103 nM pH = 7.06 pH = 6.69 pH = 5.70 pH = 5.70 Figure 4.7: Patterns of the solution complexation term (log (MT/[M3+])) in equation (4.26) calculated with equa tion (4.8). The symbols for each pattern co rrespond to the same symbols in Figure 4.6. Horizontal dot ted lines were drawn through Y to emphasize the relative slopes of the patterns, except for the pattern at 2 3T[CO] = 9.64 nM, which is relatively flat. For clarity, the vertical ax is is extended below zero although all of the solution complexation values are positive. similar patterns (Figures 4.3A, 4.3B, and 4. 3D) across the YREE series. In contrast, the patterns of the solution complexation term (MT/[M3+]) shown in Figure 4.7 exhibit large changes over the same range of conditions. Equation (4.26) indicates that the predicted T iFelogK patterns in Figure 4.5 can be obtained by subtracting the solution comple xation curves in Figure 4.7 from the corresponding surface complexation curves in Figure 4.6. Since the solution complexation term labeled A in Figure 4.7 is ve ry close to zero and displays a relatively flat pattern, the conjugate T iFelogK pattern at 2 3T[CO] = 9.64 nM (Figure 4.5A) closely resembles the pattern for the surface complexation term in Figure 4.6A. The T iFelogK values in Figure 4.5B are 0.5 to 0.6 units larger than the T iFelogK values shown in Figure

PAGE 117

102 4.5A. This is caused by a nearly one un it increase in the magnitude of the surface complexation term (Figure 4.6B) and a much smaller increase in the solution complexation intensity (Figure 4.7B). Comp ared to the pattern in Figure 4.5A, the T iFelogK pattern at 2 3T[CO] = 103 nM (Figure 4.5B) shows a small decrease in the HREEs (e.g., Lu) relative to the middle REEs (e.g., Sm). This is caused by larger increases in HREE solution complexation intensity than is the case for LREEs. The T iFelogK pattern at 2 3T[CO] = 974 nM (Figure 4.5C) displa ys a gradual decrease along the YREE series from Sm to Lu compared to the patterns in Figures 4.5A and 4.5B. This is due to the rapidly increasing significan ce in solution complexation for HREEs (Figure 4.7C). The T iFelogK pattern at 2 3T[CO] = 5,770 nM (Figure 4.5D) exhibits a pronounced decrease across the YREE series from Sm to Lu due to the sharp increase in the solution complexation term (Figure 4.7D ). Although the magnitudes of T iFelogK values increase for all YREEs between Figures 4.5A and 4.5D changes are smallest for the HREEs due to the stronger increase in intensity of HREE solution complexation. These results show that the somewhat complex T iFelogK behavior shown in Fi gure 4.1B has a relatively simple explanation in terms of compet itive solution and surface complexation. 4.6 Summary The present work describes the influe nce of carbonate complexation on YREE sorption by amorphous ferric hydroxide. In th e absence of carbonate, YREE sorption is well explained by complexation of free trivalent YREEs (M3+) at two surface sites (Quinn et al., 2006a). When carbonate is adde d to the system, YREE sorption behavior is well described by adding only one new term to the surface complexation model that is appropriate in the absence of solution complexation. The new term accounts for sorption of YREE carbonate complexes (3MCO ) by amorphous ferric hydroxide. The YREE sorption model developed in this work (e quation (4.12)), which incorporates the influences of both surface and solution comple xation, quantitatively predicts (i) the increase in T iFelogK that is caused by an increase in pH, (ii) the increase in T iFelogK that

PAGE 118

103 occurs at low carbonate concentrations due to sorption of 3MCO in addition to M3+, and (iii) the decrease in T iFelogK that occurs at high carbonate concentrations, especially for HREEs, due to increasing solution complexation.

PAGE 119

104 5. Sorption of Yttrium and Rare Earth Elem ents by Amorphous Ferric Hydroxide: Influence of Temperature 5.1 Introduction In previous work (Quinn et al., 2 006a,b), YREE sorption by amorphous ferric hydroxide was shown to be strongly dependent on pH a nd carbonate concentration (2 3T[CO]), but weakly dependent on ionic strength (I). Using the surface complexation model developed by Quinn et al. (2006a,b), YREE sorption can now be described over a wide range of solution conditions found in natural waters (3.9 pH 7.1; 0 M 2 3T[CO] 150 M; 0.01 M I 0.09 M). In addition to varia tions in pH, ionic strength, and carbonate concentration, however, natural waters exhi bit a range in temperature (-2oC T 400oC) that may also affect YREE sorption behavior. As an example, in mount ain streams it has been observed that diel (24-hour) fluctuations in temperature and pH produce ch anges in concentratio ns of trace metals, including iron, aluminum, zinc, and the REEs (e.g., Gammons et al., 2005a,b, and references therein). At a stream station with neutral pH, Gamm ons et al. (2005b) found that dissolved REE concentrati ons increased at night and pa rticulate REE concentrations increased during the day. Si nce the diel pH variation was very small (0.06 units) compared to the diel temperature change (11.7oC), these observations were attributed to temperature-dependent sorption of REEs by hydrous ferric and aluminum oxides. Using the linear relationship between dissolved REE co ncentrations and reciprocal temperature, Gammons et al. (2005b) estimated sorption en thalpies for each REE, which ranged from 57 to 120 kJ/mol (13.6 to 28.7 kcal/mol). B ecause stream particles are composed of several different phases, the enthalpies determined by Gamm ons et al. (2005b) should be considered conditional.

PAGE 120

105 Only a few studies have examined REE sorption by pure solids over a range of temperatures (Miller et al., 1982; Koeppenka strop, 1992; Ridley et al., 2005; Takahashi et al., 2005; Tertre et al., 2005 ). Despite the variety of substrates used, including montmorillonite (Miller et al., 1982; Tertre et al., 2005), amorphous FeOOH (Koeppenkastrop, 1992), bacteria (Takahashi et al., 2005), and rutile (Ridley et al., 2005), all of these studies showed that REE sorpti on increased with increasing temperature. Using either the van ’t Hoff equation or a modified form of the Clausius-Clapeyron equation, REE sorption enthalpies were determined to be positive and relatively small ( 45 kJ/mol or 11 kcal/mol) compared to other reaction enthalpies such as cation hydration, indicating a relatively weak temperature de pendence for REE sorption (Koeppenkastrop, 1992; Ridley et al., 2005; Tert re et al., 2005). Koeppenkastr op (1992) and Ridley et al. (2005) suggested that the endothermic sorpti on reaction for REEs is driven by a large entropy increase. Enthalpy values were only reported for a few i ndividual REEs, limiting interpretations of YREE behavior as a group. To further improve descrip tions of YREE behavior in natural waters, the current study investigates sorption of the entire YREE series by amorphous ferric hydroxide between 10 and 40oC. Observed temperature dependen ces are quantified in terms of sorption enthalpies that are incorporated in the surface complexation model originally developed by Quinn et al. (2006a). 5.2. Materials and Methods YREE sorption by amorphous ferric hydroxide was investigated at 10 and 40oC over a range of pH (4.7 – 7.1) using procedures essentially identical to those employed by Quinn et al. (2006a) at 25oC. Two experiments were performed at each temperature. A brief description of the experiment al procedure is provided below. All chemical manipulations were performe d in a class-100 clean air laboratory or laminar flow bench. Trace metal-clean water (Milli-Q water) from a Millipore purification system was used for all soluti on preparations. At the beginning of each experiment, a pH standard solution and an expe rimental solution, both with ionic strength I = 0.012 0.002, were prepared in Teflon wi de-mouth bottles. The pH standard solution

PAGE 121

106 was composed of 1.0 mM HCl (pH 3.0) in 0.01 M NH4NO3. The experimental solution was composed of 23.3 ppb of each YREE ([YREE]T = 2.36 M) and either 0.1 mM (experiments at 40oC) or 1.0 mM (experiments at 10oC) iron in 0.01 M HCl. The concentration of iron was increa sed for the experiments at 10oC in order to obtain adequate YREE sorption. The experimental solution wa s bubbled with ultra-pure N2 throughout each experiment. After equilibrating at the appr opriate temperature for approximately two hours, an initial sample (pH 2.0) was taken to determine the total YREE concentration. The pH of the experimental solutio n was then increased by addition of 1 M NH4OH using a Gilmont micro-dispenser. Tw o samples were taken at each half-unit pH increment: one at 15 minutes and one at either 60 minutes or 90 minutes. In contrast to previous experiments, only filtered samp les were taken because filtration provided better phase separation than centrifugation. The temperature of both solutions was m easured after the two-hour equilibration period, and also at the end of each experiment. The pH standard solution differed by 0.2o from the experimental solution, which was caused by differences in the physical configurations used for thermostating. The temperature changed approximately 0.5o during each experiment, which generally lasted < 15 hours. So lution pH on the free hydrogen-ion concentration scale was measur ed using a Ross-type combination pH electrode (No. 810200) connected to a Cornin g 130 pH meter in the absolute millivolt mode. Linearity and Nernstian be havior of the electrode was ve rified by titrating a 0.3 M NaCl solution with concentrated HCl at each temperature. The filtered samples were diluted 5-fold with 1% HNO3 except where concentrations were below the lowest calibration standard (0.5 ppb), in which case no dilution was performed. A small amount of internal standa rd solution containing equal concentrations of In, Cs, and Re was added to each sample. The resulting mixtures were analyzed for YREEs with an Agilent Technologies 4500 Se ries 200 inductively-coupled plasma mass spectrometer (ICP-MS) following the procedure ou tlined in Quinn et al. (2004). In brief, all standards and sample solutions were injected in triplicate. The instrument was tuned to minimize formation of oxide and double-ch arged ions using a 10 ppb Ce solution. MO+

PAGE 122

107 and M2+ peaks were always less than 1% and 3% of the corresponding M+ peak, respectively, and correction for this eff ect proved unnecessary. YREE concentrations were calculated from linear regressions of four standards (0.5, 1, 2, and 5 ppb). Ion counts were corrected for minor instrument drift by normalizing 89Y to 115In, 139La–161Dy to 133Cs, and 163Dy–175Lu to 187Re. To check the validity of the drift correction, a comparison was made of the Dy concentrations calculated from 161Dy and 163Dy, which were usually identical within 2%. Raw data from each experiment were correct ed by a dilution factor, which was based on the amount of NH4OH added to increase the pH. Di stribution coefficients were calculated from these corrected da ta using the following equation: iTiT iFe 3+3 iTS[MS][MS] K= [M][S]M[Fe], (5.1) where brackets denote the concentration of each indicated species. Over the range of experimental conditions employed in this work, the concentration of free YREE was set equal to the total dissolved YREE concentration ([M3+] = MT). The concentration of sorptive solid substrate was set equal to the concentration of precipitated iron ([Si] = [Fe3+]S), which was assumed to be equal to the initial dissolved Fe concentration. The concentration of sorbed YREE, [MSi]T, was calculated as the difference between the YREE concentration of the initial sample at pH 2 and the YREE concentration of the filtrate at each subsequent time after a pH adjustment. 5.3 Results and Discussion Distribution coefficient (log iKFe) results from experiments performed over a range of temperatures (10 – 40oC) are shown in Figure 5.1. It can be seen that as the temperature increased, YREE sorption increased at both pH 5.6 (Figure 5.1A) and pH 7.1 (Figure 5.1B). The log iKFe data displayed in Figure 5.1 were selected from Appendices B and E to emphasize the shape of the YREE pattern at each temperature. Because the absolute increase in log iKFe was very similar across the se ries, YREE fracti onation patterns remained relatively constant with increasi ng temperature between pH 5.6 and 7.1. This suggests that YREE fractionati on is temperature independen t, which was also observed

PAGE 123

108 by Tertre et al. (2005) in their study on YREE sorption by montmorillonite from 25 to 80oC. Figure 5.2 shows log iKFe results versus pH for four representative REEs. Despite the scatter, especially at 25oC, the log iKFe data in Figure 5.2 are clearly shifted to higher values with increasing temperature. YLaCePrNdPmSmEuGdTbDyHoErTmYbLulog iKFe 2.0 2.5 3.0 3.5 4.0 4.5 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu log iKFe 4.6 4.8 5.0 5.2 5.4 5.6 5.8 6.0 6.2 6.4 6.6 6.8 A B 10oC 25oC 40oC Figure 5.1. log iKFe results obtained over a range of temperatures, indicated in the legend, at pH = 5.61 0.05 (A) and pH = 7.06 0.03 (B).

PAGE 124

109 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 7 Sm 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 La pH 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 7 Dy pH 3.54.04.55.05.56.06.57.07.5 log iKFe 0 1 2 3 4 5 6 7 Lu Figure 5.2. log iKFe versus pH for La, Sm, Dy, and Lu at 10, 25, and 40oC. Closed triangles represent samples at 10oC, open circles represent samples at 25oC, and closed circles represent samples at 40oC. To describe the pH dependence of YREE sorption at 25oC, Quinn et al. (2006a) proposed a surface complexation model (SCM) in the form: 12 S1S2 iFe S1[H][H] K K[H]1 (5.2) where Sn are stability constants for sorption of free YREE ions and SK1 is the surface protonation constant for amor phous ferric hydroxide (log SK1 = 4.76; Quinn et al., 2006a). In order to highlight the influence of temperature on iKFe, equation (5.2), with the Sn values at 25oC calculated by Quinn et al. (2006b), was used to predict distribution coefficients at 10 and 40oC. Figure 5.3 compares directly measured log iKFe results at 10,

PAGE 125

110 02468 log iKFe(meas) 0 2 4 6 8 02468 log iKFe(meas) 0 2 4 6 8 log iKFe(pred) 02468 log iKFe(meas) 0 2 4 6 8 log iKFe(pred) 02468 log iKFe(meas) 0 2 4 6 8 La Sm Dy Luy = (1.000.02)x + (0.130.07) r2 = 0.948 y = (1.020.02)x + (0.010.08) r2 = 0.962 y = (1.030.02)x (0.050.08) r2 = 0.963 y = (1.010.02)x + (0.0040.08) r2 = 0.957 Figure 5.3. Regressions of log iKFe(meas) versus log iKFe(pred) for La, Sm, Dy, and Lu. log iKFe(meas) are directly measured distribution coefficients from the present work (Tables B.1–B.6, D.4–D.7, and E.1–E.4). log iKFe(pred) are distribution coefficients predicted from equation (5.2) using the S1 and S2 results listed in Table 4.1. Closed triangles represent samples at 10oC, open circles represent samples at 25oC, and closed circles represent samples at 40oC.

PAGE 126

111 25, and 40oC with distribution coefficient values predicted using equation (5.2). Despite the fact that the regressions are fairly good (r2 = 0.95 0.01 for all YREEs), data for 10oC consistently plot below the regression line while data for 40oC plot above the regression line, indicating the temperature effect on YREE sorption is not adequately modeled by equation (5.2). The influence of temperature on the stability constants, S1 and S2, that are used to describe YREE sorption, can be character ized using the van t Hoff equation: 2T 0 Sn n 29815 Sn2H 11 ln RT298.15 (5.3) where 29815 Sn is a stability constant at 298.15 K, 2T Sn is a stability constant at some absolute temperature (T2), 0 nH is an enthalpy change, and R is the gas constant (1.987 cal K-1 mol-1). Rearranging equation (5.3) and substituting into equation (5.2) for both S1 and S2 gives: 12pp 298151298152 S1S2 iFe S110[H]10[H] K K[H]1 (5.4) where 0 n n 2H11 p R(ln10)T298.15 (5.5) SigmaPlot (Version 8.02) was used to solve equation (5.4) for 29815 S1 29815 S2 0 1H and 0 2H through minimization of the following re sidual sum of squares (RSS) function: 122 pp 29815 129815 2 1 S1 S2 iFe S110[H]10[H] RSS1K K[H]1 (5.6) where the summation was performed over a ll pH values and all temperatures. The distribution coefficient (iKFe) data that were utilized for the determination of the parameters in equation (5.4) are listed in Tables B.1B.6, D.4D.7, and E.1E.4 (note that the fit exclusively used data at 2 3T[CO] = 0 M). YREE surface complexation constants (29815 S1 and 29815 S2 ) and enthalpy values (0 1H and 0 2H ) calculated using equations (5.4) and (5.6) are listed in Table 5.1. The

PAGE 127

112Table 5.1. YREE surface complexation constants (29815 S1 and 29815 S2 ) and enthalpy values (0 1H and 0 2H ; kcal/mol) determined with eq uations (5.4) and (5.6), log SK1 = 4.76 (Quinn et al., 2006a), and th e distribution coefficient da ta in Tables B.1B.6, D.4 D.7, and E.1E.4. Uncertainties represent one standard error. [M3+] 29815 S1 0 1H 29815 S2 0 2H Y -2.95 0.04 12.6 2.5 -8.92 0.04 9.9 1.8 La -2.90 0.03 12.9 1.6 -9.36 0.06 7.7 2.9 Ce -2.43 0.03 12.7 1.5 -8.85 0.06 8.0 2.9 Pr -2.30 0.03 12.5 1.4 -8.62 0.05 8.2 2.3 Nd -2.23 0.03 12.1 1.4 -8.55 0.05 8.2 2.4 Pm Sm -2.09 0.03 11.8 1.4 -8.32 0.04 8.3 2.1 Eu -2.14 0.03 12.0 1.5 -8.35 0.04 8.4 2.0 Gd -2.31 0.03 12.2 1.5 -8.58 0.05 8.4 2.2 Tb -2.31 0.03 12.5 1.6 -8.44 0.04 9.8 1.8 Dy -2.34 0.03 12.8 1.5 -8.43 0.04 10.6 1.7 Ho -2.40 0.03 13.0 1.4 -8.51 0.04 10.9 1.7 Er -2.37 0.03 13.2 1.4 -8.48 0.04 11.2 1.7 Tm -2.29 0.03 13.3 1.4 -8.38 0.04 11.8 1.6 Yb -2.22 0.03 13.4 1.4 -8.26 0.04 12.3 1.5 Lu -2.23 0.03 13.1 1.4 -8.32 0.04 12.2 1.6 29815 S1 and 29815 S2 results obtained in the present st udy differ by approximately 2% from those determined by Quinn et al. (2006b), wh ich is within the uncertainty of the estimates. In Figure 5.4, enthalpy values are plotted versus YREE atomic number. It can be seen that 0 1H for all YREEs are identical within the experimental error (average 0 1H = 12.7 kcal/mol). However, since the erro rs are strongly correlated across the YREE series, the observed pattern for 0 1H is likely to represent significant YREE trends. Enthalpy values determined with equations (5.4) and (5.6) can not be directly compared with previous REE thermodynamic re sults because either the model parameters (Ridley et al., 2005) or the solution chemi cal compositions (Koeppenkastrop, 1992) were distinctly different from those used in the current study. An indirect comparison may be

PAGE 128

113 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu H1 0 (kcal/mole) 9 10 11 12 13 14 15 16 YLaCePrNdPmSmEuGdTbDyHoErTmYbLu H2 0 (kcal/mole) 4 6 8 10 12 14 16 Figure 5.4. Enthalpy values (equation (5.4)) for YREE sorption by amorphous ferric hydroxide from 10 to 40oC. Error bars represent one standard error. made with the results of Tertre et al. ( 2005) in their investiga tion of YREE sorption by montmorillonite between 25 and 80oC. By plotting distribution coefficients (log Kd) versus reciprocal temperature, Tertre et al. (2005 ) calculated an apparent enthalpy of 9.3 2.4 kcal/mol for Eu sorption at pH 7.0. To make a similar a ssessment, distribution coefficients were predicted at pH 7.0 and 25, 40, and 80oC using equation (5.4) and the parameters listed in Table 5. 1. Plots of predicted values versus reciprocal temperature (not shown), yielded an av erage enthalpy of 10.4 1.2 kcal/mol for all YREEs. Despite the lack of comparable YR EE sorption enthalpies, the average 0 1H value obtained in the present study is on the same order of magnitude as the enthalpy for the

PAGE 129

114 first hydrolysis constant of iron (i.e., 1H = 10.2 0.3 kcal/mol; Byrne et al., 2000, and references therein), and also the av erage enthalpy for YREE hydrolysis (i.e., H0(M) = 11.3 kcal/mol for all YREEs; Klungness and Byrn e, 2000). The coherence of these values suggests that the enthalpies involved in YR EE sorption are related to the enthalpy of dissociation of H2O in the first coordination sp here of cations such as Fe3+ (Baes and Mesmer, 1981). To demonstrate the goodness-of-fit for the temperature-dependent SCM, distribution coefficients were pred icted at 10, 25, and 40oC using equation (5.4) and the parameters listed in Table 5.1. Figure 5.5 compares thes e predictions with directly measured log iKFe values for four representative REEs. The r2 values (0.98 0.01 for all YREEs) for the regressions in Figure 5.5 are si gnificantly better than those for the regressions in Figure 5.3. In addition, the data at 10 and 40oC are tightly grouped around the regression lines. These results indicate that the SCM was substantially improved by including terms (0 1H and 0 2H ) for the temperature depe ndence of YREE sorption. Based on the fact that sorption of free YREE ions (M3+) is significantly dependent on temperature, it is expected that sorption of YREE solution complexes, such as 3MCO may also be temperature dependent. In order to determine an enthalpy for 3MCO sorption, experiments n eed to be performed over a range of temperatures in the presence of carbonate. Additionally, the temp erature dependence of YREE carbonate complexation must be known. Cantrell and Byrne (1987b) calculated enthalpies for carbonate and bicarbonate stability constants of Eu between 15 and 35oC. Their results indicated that YREE complexation by 2 3CO is weakly dependent on temperature relative to the influence of temperature on YREE so rption. Once enthalpy va lues for carbonate complexation and 3MCO sorption are obtained for th e entire YREE series, a SCM can be produced that will accurately describe YREE sorption over a wide range of solution conditions (i.e., pH, carbonate co mplexation, and temperature).

PAGE 130

115 02468 log iKFe(meas) 0 2 4 6 8 02468 log iKFe(meas) 0 2 4 6 8 log iKFe(pred) 02468 log iKFe(meas) 0 2 4 6 8 log iKFe(pred) 02468 log iKFe(meas) 0 2 4 6 8 La Sm Dy Luy = (0.99.02)x + (0.18.05) r2 = 0.973 y = (1.02.01)x + (0.04.05) r2 = 0.984 y = (1.02.01)x + (0.03.06) r2 = 0.981 y = (1.01.01)x + (0.08.05) r2 = 0.984 Figure 5.5. Regressions of log iKFe(meas) versus log iKFe(pred) for La, Sm, Dy, and Lu. log iKFe(meas) are directly measured distribution coefficients from the present work (Tables B.1B.6, D.4D.7, and E.1E.4). log iKFe(pred) are distribution coefficients predicted from equation (5.4) using the 29815 S1 29815 S2 0 1H and 0 2H results listed in Table 5.1. Closed triangles represent samples at 10oC, open circles represent samples at 25oC, and closed circles represent samples at 40oC.

PAGE 131

116 References Alibo D.-S. and Nozaki Y. (1999) Rare earth el ements in seawater: Particle association, shale-normalization, and Ce oxidation. Geochim. Cosmochim. Acta 63(3/4), 363372. Baes C. F., Jr. and Mesmer R. E. (1981) The thermodynamics of cation hydrolysis. Am. J. Sci. 281, 935-962. Balistrieri L., Brewer P. G., and Murray J. W. (1981) Scavenging residence times of trace metals and surface chemistry of sinki ng particles in the deep ocean. Deep-Sea Res. 28A, 101-121. Bau M. (1999) Scavenging of dissolved y ttrium and rare earth s by precipitating iron oxyhydroxide: Experimental evidence for Ce oxidation, Y-Ho fractionation, and lanthanide tetrad effect. Geochim. Cosmochim. Acta 63(1), 67-77. Bau M. and Dulski P. (1999) Comparing yttriu m and rare earths in hydrothermal fluids from the Mid-Atlantic Ridge: implicati ons for Y and REE behaviour during nearvent mixing and for the Y/Ho ra tio of Proterozoic seawater. Chem. Geol. 155, 7790. Bau M., Koschinsky A., Dulski P., and Hein J. R. (1996) Comparison of the partitioning behaviours of yttrium, rare earth elem ents, and titanium between hydrogenetic marine ferromanganese crusts and seawater. Geochim. Cosmochim. Acta 60(10), 1709-1725. Bertram C. J. and Elderfield H. (1993) Th e geochemical balance of the rare earth elements and neodymium isotopes in the oceans. Geochim. Cosmochim. Acta 57, 1957-1986. Bidoglio G., Gibson P. N., Haltier E., Omenetto N., and Lipponen M. (1992) XANES and laser fluorescence spectroscopy for ra re earth speciation at mineral-water interfaces. Radiochim. Acta 58-9, 191-197. Bingler L. S., Byrne R. H., Vargo G. A., and Tomas C. R. (1989) Rare earth element uptake by the marine diatom Skeletonema costatum. Chem. Spec. Bioavailability 1(3), 103-110. Borkowski M. and Siekierski S. (1992) Factors affecting the position of Y and actinides(III) with resp ect to lanthanides in the NH4SCN Adogen-464SCN extraction system. Radiochim. Acta 56, 31-35. Bradbury M. H. and Baeyens B. (2002) Sorp tion of Eu on Naand Ca-montmorillonites: Experimental investigations and modell ing with cation exchange and surface complexation. Geochim. Cosmochim. Acta 66(13), 2325-2334. Bruland K. W. (1983) Trace elements in sea-water. In Chemical Oceanography, Vol. 8 (eds. J. P. Riley and R. Chester), pp. 157-220. Academic Press, London.

PAGE 132

117 Byrne R. H. and Kim K.-H. (1990) Rare earth element scave nging in seawater. Geochim. Cosmochim. Acta 54, 2645-2656. Byrne R. H. and Kim K.-H. (1993) Rare eart h precipitation and copr ecipitation behavior: The limiting role of PO4 3on dissolved rare earth c oncentrations in seawater. Geochim. Cosmochim. Acta 57, 519-526. Byrne R. H. and Lee J. H. (1993) Comparativ e yttrium and rare earth element chemistries in seawater. Mar. Chem. 44, 121-130. Byrne R. H. and Li B. (1995) Comparative complexation behavior of the rare earths. Geochim. Cosmochim. Acta 59(22), 4575-4589. Byrne R. H. and Sholkovitz E. R. (1996) Ma rine chemistry and geochemistry of the lanthanides. In Handbook on the Physics and Chemistry of Rare Earths, Vol. 23 (eds. K. A. Gschneidner Jr. and L. Ey ring), pp. 497-593. Elsevier, Amsterdam. Byrne R. H., Luo Y.-R., and Young R. W. (2000) Iron hydrolysis and so lubility revisited: observations and comments on ir on hydrolysis characterizations. Mar. Chem. 70, 23-35. Cantrell K. J. and Byrne R. H. (1987a) Rare earth element complexation by carbonate and oxalate ions. Geochim. Cosmochim. Acta 51, 597-605. Cantrell K. J. and Byrne R. H. (1987b) Te mperature dependence of europium carbonate complexation. J. Sol. Chem. 16(7), 555-566. Chatt A. and Rao R. R. (1989) Complexation of europium (III) with carbonate ions in groundwater. Mat. Res. Soc. Symp. Proc. 127, 897-904. Choppin G. R. and Strazik W. F. (1965) Comp lexes of trivalent lanthanide and actinide ions. I. Outer-sphere ion pairs. Inorg. Chem. 4(9), 1250-1254. Ciavatta L., Ferri D., Grenthe I., Salvator e F., and Spahiu K. (1981) Studies on metal carbonate equilibria. 3. The lanthanum(III) carbonate complexes in aqueous perchlorate media. Acta Chem. Scand. A 35, 403-413. Coppin F., Berger G., Bauer A., Castet S., and Loubet M. (2002) So rption of lanthanides on smectite and kaolinite. Chem. Geol. 182, 57-68. Davranche M., Pourret O., Gruau G., and Di a A. (2004) Impact of humate complexation on the adsorption of REE onto Fe oxyhydroxide. J. Colloid Interf. Sci. 277, 271279. Davranche M., Pourret O., Gruau G., Dia A ., and Le Coz-Bouhnik M. (2005) Adsorption of REE(III)-humate complexes onto MnO2: Experimental evidence for cerium anomaly and lanthanide tetr ad effect suppression. Geochim. Cosmochim. Acta 69(20), 4825-4835. de Baar H. J. W., Bacon M. P., and Brewer P. G. (1983) Rare-earth distributions with a positive Ce anomaly in the western North Atlantic Ocean. Nature 301, 324-327. de Baar H. J. W., Bacon M. P., Brewer P. G., and Bruland K. W. (1985a) Rare earth elements in the Pacific and Atlantic Oceans. Geochim. Cosmochim. Acta 49, 1943-1959. de Baar H. J. W., Brewer P. G., and B acon M. P. (1985b) Anomalies in rare earth distributions in seawater: Gd and Tb. Geochim. Cosmochim. Acta 49, 1961-1969. de Baar H. J. W., German C. R., Elderfie ld H., and van Gaans P. (1988) Rare earth element distributions in anoxic waters of the Cariaco Trench. Geochim. Cosmochim. Acta 52, 1203-1219.

PAGE 133

118 de Baar H. J. W., Schijf J., and Byrne R. H. (1991) Solution chemistry of the rare earth elements in seawater. Eur. J. Solid State Inorg. Chem. 28, 357-373. De Carlo E. H. and Green W. J. (2002) Rare earth elements in the water column of Lake Vanda, McMurdo Dry Valleys, Antarctica. Geochim. Cosmochim. Acta 66(8), 1323-1333. De Carlo E. H., Wen X.-Y., a nd Irving M. (1998) Th e influence of redox reactions on the uptake of dissolved Ce by suspende d Fe and Mn oxide particles. Aquat. Geochem. 3, 357-389. De Carlo E. H., Wen X. Y., and Cowen J. P. (2000) Rare earth element fractionation in hydrogenetic Fe-Mn crusts: The influence of carbonate complexation and phosphatization on Sm/Yb ratios. In Marine Authigenesis: From Global to Microbial, Vol. 66 (eds. C. R. Glenn, L. Prvt-Lucas, and J. Lucas), pp. 271285. SEPM (Society for Sedimentary Geology), Tulsa, OK. Duncan T. and Shaw T. J. (2003) The mobility of rare earth elements and redox sensitive elements in the groundwater/seawater mixi ng zone of a shallow coastal aquifer. Aquat. Geochem. 9, 233-255. Dzombak D. A. and Morel F. M. M. (1990) Surface Complexation Modeling. Hydrous Ferric Oxide. John Wiley & Sons. Elderfield H. and Greaves M. J. (1982) The rare earth elements in seawater. Nature 296, 214-219. Elderfield H. and Sholkovitz E. R. (1987) Rare earth elem ents in the pore waters of reducing nearshore sediments. Earth Planet. Sci. Lett. 82, 280-288. Elderfield H., Hawkesworth C. J., Greaves M. J., and Calvert S. E. (1981) Rare earth element geochemistry of oceanic ferromanganese nodules and associated sediments. Geochim. Cosmochim. Acta 45, 513-528. Erel Y. and Morgan J. J. (1991) The effect of surface reactions on the relative abundances of trace metals in deep-ocean water. Geochim. Cosmochim. Acta 55, 1807-1813. Erel Y. and Stolper E. M. (1993) Modeling of rare-earth element partitioning between particles and solution in aquatic environments. Geochim. Cosmochim. Acta 57, 513-518. Fairhurst A. J., Warwick P., and Richardson S. (1995) The influence of humic acid on the adsorption of europium onto inorgani c colloids as a function of pH. Colloids Surfaces A: Physicochem. Eng. Aspects 99, 187-199. Ferri D., Grenthe I., Hietanen S., and Salvatore F. (1983) Studies on metal carbonate equilibria. 5. The cerium(III) carbonate complexes in aqueous perchlorate media. Acta Chem. Scand. A 37, 359-365. Fowler S. W. and Knauer G. A. (1986) Role of large particles in the transport of elements and organic compounds through the oceanic water column. Prog. Oceanogr. 16(3), 147-194. Fowler S. W., Hamilton T. F., Peinert R. D ., La Rosa J., and Teyssie J.-L. (1992) The vertical flux of rare earth elemen ts in the northwestern Mediterranean. EROS 2000 (European River Ocean System): Proceed ings of the Third Workshop on the North-West Mediterranean Sea, 401-412.

PAGE 134

119 Gammons C. H., Nimick D. A., Parker S. R., Cleasby T. E., and McMcleskey R. B. (2005a) Diel behavior of ir on and other heavy metals in a mountain stream with acidic to neutral pH: Fish er Creek, Montana, USA. Geochim. Cosmochim. Acta 69(10), 2505-2516. Gammons C. H., Wood S. A., and Nimick D. A. (2005b) Diel beha vior of rare earth elements in a mountain stream with acidic to neutral pH. Geochim. Cosmochim. Acta 69(15), 3747-3758. German C. R., Klinkhammer G. P., Edmond J. M., Mitra A., and Elderfield H. (1990) Hydrothermal scavenging of rare -earth elements in the ocean. Nature 345, 516518. German C. R., Masuzawa T., Greaves M. J. Elderfield H., and Edmond J. M. (1995) Dissolved rare earth elements in the Southern Ocean: Cerium oxidation and the influence of hydrography. Geochim. Cosmochim. Acta 59(8), 1551-1558. Goldberg E. D. (1954) Marine geochemistry 1. Chemical scave ngers of the sea. J. Geol. 62, 249-265. Goldberg E. D., Koide M., Schmitt R. A., and Smith R. H. (1963) Rare-earth distributions in the marine environment. J. Geophys. Res. 68(14), 4209-4217. Goldstein S. J. and Jacobsen S. B. (1988) Rare earth elements in river waters. Earth Planet. Sci. Lett. 89, 35-47. Gromet L. P., Dymek R. F., Haskin L. A ., and Korotev R. L. (1984) The "North American shale composite": Its co mpilation, major and trace element characteristics. Geochim. Cosmochim. Acta 48, 2469-2482. Haley B. A., Klinkhammer G. P., and McManu s J. (2004) Rare eart h elements in pore waters of marine sediments. Geochim. Cosmochim. Acta 68(6), 1265-1279. Haskin L. A., Frey F. A., Schmitt R. A., a nd Smith R. H. (1968) Meteoritic, solar, and terrestrial rare earth distributions. In Physics and Chemistry of the Earth. Pergamon Press. Haskin M. A. and Haskin L. A. (1966) Rare earths in European shales: A redetermination. Science 154(3748), 507-509. Hayes K. F. and Leckie J. O. (1987) Modeling Ionic Streng th Effects on Cation Adsorption at Hydrous Oxide/Solution Interfaces. J. Colloid Interf. Sci. 115(2), 564-572. Hgdahl O. T., Melsom S., and Bowen V. T. (1968) Neutron acti vation analysis of lanthanide elements in sea water. In Trace Inorganics in Water, Vol. 73 (ed. R. A. Baker), pp. 308-325. American Chem ical Society, Washington, DC. Huang C.-P. and Stumm W. (1973) Specific adsorpti on of cations on hydrous -Al2O3. J. Colloid Interf. Sci. 43(2), 409-420. Hunter K. A. (1983) The adsorptive propertie s of sinking particles in the deep ocean. Deep-Sea Res. 30, 669-675. Johannesson K. H. and Lyons W. B. (1994) The rare earth element geochemistry of Mono Lake water and the importance of carbonate complexing. Limnol. Oceanogr. 39(5), 1141-1154. Johannesson K. H. and Lyons W. B. (1995) Ra re-earth element geochemistry of Colour Lake, an acidic freshwater lake on Ax el Heiberg Island, Northwest Territories, Canada. Chem. Geol. 119, 209-223.

PAGE 135

120 Johannesson K. H. and Zhou X. (1999) Origin of middle rare earth element enrichments in acid waters of a Ca nadian High Arctic lake. Geochim. Cosmochim. Acta 63(1), 153-165. Johannesson K. H., Lyons W. B., and Bi rd D. A. (1994) Rare earth element concentrations and speciation in al kaline lakes from the western U.S.A. Geophys. Res. Lett. 21(9), 773-776. Johannesson K. H., Stetzenbach K. J., Hodge V. F., and Lyons W. B. (1996) Rare earth element complexation behavior in circum neutral pH groundwaters: Assessing the role of carbonate and phosphate ions. Earth Planet. Sci. Lett. 139, 305-319. Johannesson K. H., Stetzenbach K. J., and Ho dge V. F. (1997) Rare earth elements as geochemical tracers of regional groundwater mixing. Geochim. Cosmochim. Acta 61(17), 3605-3618. Jonasson R. G., Bancroft G. M., and Nesbitt H. W. (1985) Solubilities of some hydrous REE phosphates with implications for diag enesis and sea water concentrations. Geochim. Cosmochim. Acta 49, 2133-2139. Kawabe I., Ohta A., and Miur a N. (1999a) Distribution coeffi cients of REE between Fe oxyhydroxide precipitates and NaCl so lutions affected by REE-carbonate complexation. Geochem. J. 33, 181-197. Kawabe I., Ohta A., Ishii S., Tokumura M ., and Miyauchi K. (1999b) REE partitioning between Fe-Mn oxyhydroxide precipitates and weakly acid NaCl solutions: Convex tetrad effect and fractionation of Y and Sc from heavy lanthanides. Geochem. J. 33, 167-179. Klinkhammer G., Elderfield H., and Hudson A. (1983) Rare earth elements in seawater near hydrothermal vents. Nature 305, 185-188. Klinkhammer G., German C. R., Elderfield H ., Greaves M. J., and Mitra A. (1994) Rare earth elements in hydrothermal fluids and plume particulates by inductively coupled plasma mass spectrometry. Mar. Chem. 45, 179-186. Klungness G. D. and Byrne R. H. (2000) Compar ative hydrolysis beha vior of the rare earths and yttrium: the influence of temperature and ionic strength. Polyhedron 19, 99-107. Koeppenkastrop D. (1992) Thermo dynamic and kinetic studies on the interaction of rare earth elements with metal oxides. Ph .D. thesis, University of Hawaii. Koeppenkastrop D. and De Carl o E. H. (1992) Sorption of rare-earth elements from seawater onto synthetic mineral part icles: An experimental approach. Chem. Geol. 95, 251-263. Koeppenkastrop D. and De Carlo E. H. (1993) Uptake of rare earth elements from solution by metal oxides. Environ. Sci. Technol. 27(9), 1796-1802. Koeppenkastrop D., De Carlo E. H., and Ro th M. (1991) A method to investigate the interaction of rare earth elements in aqueous solution with metal oxides. J. Radioanal. Nucl. Chem. Art. 152(2), 337-346. Kolat R. S. and Powell J.E. (1962) Acetate complexes of the rare earth and several transition metal ions. Inorg. Chem. 1, 293-296. Kolesov G. M., Anikiyev V. V., and Savenko V. S. (1975) On the geochemistry of rare earth elements in the waters of the Black Sea. Geochem. Int. 12(1), 82-88.

PAGE 136

121 Kosmulski M. (1997) Adsorption of trivalent cations on silica. J. Colloid Interf. Sci. 195, 395-403. Krishnaswami S. (1976) Authigenic transi tion elements in Pacific pelagic clays. Geochim. Cosmochim. Acta 40, 425-434. Kuss J., Garbe-Schnberg C.-D., and Kremling K. (2001) Rare earth elements in suspended particulate material of North Atlantic surface waters. Geochim. Cosmochim. Acta 65(2), 187-199. Lee J. H. and Byrne R. H. (1993) Complexation of trivalent rare eart h elements (Ce, Eu, Gd, Tb, Yb) by carbonate ions. Geochim. Cosmochim. Acta 57, 295-302. Lerche D. and Nozaki Y. (1998) Rare earth el ements of sinking particulate matter in the Japan Trench. Earth Planet. Sci. Lett. 159, 71-86. Li B. and Byrne R. H. (1997) Ionic strength dependence of rare earth NTA stability constants at 25 C. Aquat. Geochem. 3, 99-115. Liu X. and Byrne R. H. (1995) Compara tive carbonate complexation of yttrium and gadolinium at 25C and 0.7 mol dm-3 ionic strength. Mar. Chem. 51, 213-221. Liu X. and Byrne R. H. (1997) Rare earth a nd yttrium phosphate so lubilities in aqueous solution. Geochim. Cosmochim. Acta 61(8), 1625-1633. Liu X. and Byrne R. H. (1998) Comprehensiv e investigation of y ttrium and rare earth element complexation by carbonate ions using ICP-mass spectrometry. J. Sol. Chem. 27(9), 803-815. Liu X. and Millero F. J. (1999) The solubi lity of iron hydroxide in sodium chloride solutions. Geochim. Cosmochim. Acta 63(19/20), 3487-3497. Liu X., Byrne R. H., and Schijf J. (1997) Comparative coprecipitation of phosphate and arsenate with yttrium and the rare earths: The influence of solution complexation. J. Sol. Chem. 26(12), 1187-1198. Lundqvist R. (1982) Hydrophilic complexes of the actinides. I. Carbonates of trivalent americium and europium. Acta Chem. Scand. A 36, 741-750. Luo Y.-R. and Byrne R. H. (2000) The ioni c strength dependence of rare earth and yttrium fluoride complexation at 25C. J. Sol. Chem. 29(11), 1089-1099. Luo Y.-R. and Byrne R. H. (2001) Yttrium and rare earth elem ent complexation by chloride ions at 25C. J. Sol. Chem. 30(9), 837-845. Luo Y.-R. and Byrne R. H. (2004) Carbonate complexation of yttrium and the rare earth elements in natural waters. Geochim. Cosmochim. Acta 68(4), 691-699. Luo Y. and Millero F. J. (2004) Effects of temperatur e and ionic strength on the stabilities of the first and second fluoride complexes of yttrium and the rare earth elements. Geochim. Cosmochim. Acta. Marmier N. and Fromage F. (1999) Compari ng electrostatic and nonelectrostatic surface complexation modeling of the sorption of lanthanum on hematite. J. Colloid Interf. Sci. 212, 252-263. Marmier N., Dumonceau J., and Fromage F. (1997) Surface complexation modeling of Yb(III) sorption and desorption on hematite and alumina. J. Contam. Hydrol. 26, 159-167. Marmier N., Delise A., and Fromage F. (1999) Surface complexation modeling of Yb(III) and Cs(I) sorption on silica. J. Colloid Interf. Sci. 212, 228-233.

PAGE 137

122 Martell A. E. and Hancock R. D. (1996) Metal Complexes in Aqueous Solutions. Plenum Press. Martell A. E. and Smith R. M. (1974) Critical Stability Constant s, Vol. 1: Amino Acids. Plenum Press. Martell A. E. and Smith R. M. (1977) Critical Stability Constant s, Vol. 3: Other Organic Ligands. Plenum Press. Martell A. E. and Smith R. M. (1982) Critical Stability Constants, Vol. 5: First Supplement. Plenum Press. Masuzawa T. and Koyama M. (1989) Settling particles with positive Ce anomalies from the Japan Sea. Geophys. Res. Lett. 16(6), 503-506. McLennan S. M. (1989) Rare earth elemen ts in sedimentary rocks: Influence of provenance and sedimentary processes. In Geochemistry and Mineralogy of Rare Earth Elements, Vol. 21 (eds. B. R. Lipin and G. A. McKay), pp. 169-200. Mineralogical Society of America, Washington, DC. McLennan S. M. (1994) Rare earth element geochemistry and the "tetrad" effect. Geochim. Cosmochim. Acta 58(9), 2025-2033. Michard A. (1989) Rare earth element systematics in hydrothermal fluids. Geochim. Cosmochim. Acta 53, 745-750. Michard A., Albarde F., Michard G., Minster J. F., and Charlou J. L. (1983) Rare-earth elements and uranium in high-temperatur e solutions from East Pacific Rise hydrothermal vent field (13N). Nature 303, 795-797. Miller S. E., Heath G. R., and Gonzalez R. D. (1982) Effects of temperature on the sorption of lanthanide s by montmorillonite. Clays Clay Mineral. 30(2), 111-122. Millero F. J. and Schreiber D. R. (1982) Use of the ion pairing mode l to estimate activity coefficients of the ionic co mponents of natural waters. Am. J. Sci. 282, 15081540. Moeller T. (1963) The Chemistry of the Lanthanides. Reinhold. Moeller T. (1972) Complexes of the lanthanides. In Lanthanides and Actinides (ed. K. W. Bagnall), pp. 275-298. University Park, Baltimore, MD. Moeller T., Martin D. F., Thompson L. C., Ferrus R., Feistel G. R., and Randall W. J. (1965) The coordination chemistry of y ttrium and the rare earth metal ions. Chem. Rev. 65(1), 1-50. Moffett J. W. (1990) Microbially medi ated cerium oxidation in sea water. Nature 345, 421-423. Moffett J. W. (1994a) A radiotracer st udy of cerium and manganese uptake onto suspended particles in Chesapeake Bay. Geochim. Cosmochim. Acta 58(2), 695703. Moffett J. W. (1994b) The relationship betw een cerium and manganese oxidation in the marine environment. Limnol. Oceanogr. 39(6), 1309-1318. Monecke T., Kempe U., Monecke J., Sala M., a nd Wolf D. (2002) Tetr ad effect in rare earth element distribution patterns: A met hod of quantification with application to rock and mineral samples from granite-related rare metal deposits. Geochim. Cosmochim. Acta 66(7), 1185-1196. Murphy K. and Dymond J. (1984) Rare eart h element fluxes and geochemical budget in the eastern equatorial Pacific. Nature 307, 444-447.

PAGE 138

123 Musi S. and Risti M. (1988) Adsorption of trace el ements or radionuclides on hydrous iron oxides. J. Radioanal. Nucl. Chem. Art. 120(2), 289-304. Musi S., Gessner M., and Wolf R. H. H. (1979) Sorption of small amounts of europium(III) on iron(II I) hydroxide and oxide. J. Radioanal. Chem. 50(1-2), 91100. Nance W. B. and Taylor S. R. (1976) Rare ear th element patterns and crustal evolution I. Australian post-Archean sedimentary rocks. Geochim. Cosmochim. Acta 40, 1539-1551. Nozaki Y. and Alibo D. S. (2003) Dissolved ra re earth elements in the Southern Ocean, southwest of Australia: Unique patterns compared to the South Atlantic data. Geochem. J. 37, 47-62. Nozaki Y., Lerche D., Alibo D. S., and Tsut sumi M. (2000) Dissolved indium and rare earth elements in three Japanese rivers and Tokyo Bay: Evidence for anthropogenic Gd and In. Geochim. Cosmochim. Acta 64(23), 3975-3982. Ohta A. and Kawabe I. (2000) Rare earth element partitioning between Fe oxyhydroxide precipitates and aqueous NaCl solutions doped with NaHCO3: Determinations of rare earth element complexation constants with carbonate ions. Geochem. J. 34, 439-454. Ohta A. and Kawabe I. (2001) R EE(III) adsorption onto Mn dioxide ( -MnO2) and Fe oxyhydroxide: Ce(III) oxidation by -MnO2. Geochim. Cosmochim. Acta 65(5), 695-703. Piper D. Z. (1974) Rare earth elements in ferromanganese nodules and other marine phases. Geochim. Cosmochim. Acta 38, 1007-1022. Quinn K. A., Byrne R. H., and Schijf J. ( 2004) Comparative scavenging of yttrium and the rare earth elements in seawater: Competitive influences of solution and surface chemistry. Aquat. Geochem. 10, 59-80. Quinn K. A., Byrne R. H., and Schijf J. (2006a) Sorption of yttr ium and rare earth elements by amorphous ferric hydroxide: In fluence of pH and ionic strength. Mar. Chem. 99, 128-150. Quinn K. A., Byrne R. H., and Schijf J. ( 2006b) Sorption of yttrium and rare earth elements by amorphous ferric hydroxide: In fluence of solution complexation with carbonate. Geochim. Cosmochim. Acta in press. Rabung T., Geckeis H., Kim J.-I., and Beck H. P. (1998a) Sorption of Eu(III) on a natural hematite: Application of a surface complexation model. J. Colloid Interf. Sci. 208, 153-161. Rabung T., Geckeis H., Kim J. I., and Beck H. P. (1998b) The influence of anionic ligands on the sorption behaviour of Eu(III) on natural hematite. Radiochim. Acta 82, 243-248. Rabung T., Stumpf T., Geckeis H., Klenze R., and Kim J. I. (2000) Sorption of Am(III) and Eu(III) onto -alumina: experiment and modelling. Radiochim. Acta 88(9-11), 711-716. Rabung T., Pierret M. C., Bauer A., Geckeis H ., Bradbury M. H., and Baeyens B. (2005) Sorption of Eu(III)/Cm(III) on Ca-montmorillo nite and Na-illite. Part 1: Batch sorption and time-resolved laser fluor escence spectrosc opy experiments. Geochim. Cosmochim. Acta 69(23), 5393-5402.

PAGE 139

124 Rao R. R. and Chatt A. (1991) Studies on stab ility constants of eu ropium (III) carbonate complexation and application of SIT and ion-pairing models. Radiochim. Acta 54, 181-188. Ridley M. K., Machesky M. L., Wesolowski D. J., and Palmer D. A. (2005) Surface complexation of neodymium at the rutile-water interface: A potentiometric and modeling study in NaCl media to 250oC. Geochim. Cosmochim. Acta 69(1), 6381. Schijf J. and Byrne R. H. (1999) Determina tion of stability constants for the monoand difluoro-complexes of Y and the REE, using a cation-exchange resin and ICPMS. Polyhedron 18, 2839-2844. Schijf J. and Byrne R. H. (2001) Stability constants for monoa nd dioxalato-complexes of Y and the REE, potentially importa nt species in groundwaters and surface freshwaters. Geochim. Cosmochim. Acta 65(7), 1037-1046. Schijf J. and Byrne R. H. (2004) Determination of 4SO1 for yttrium and the rare earth elements at I = 0.66 m and T = 25C Implications for YREE solution speciation in sulfate-rich waters. Geochim. Cosmochim. Acta 68(13), 2825-2837. Schijf J., de Baar H. J. W., and Millero F. J. (1994) Kinetics of Ce and Nd scavenging in Black Sea waters. Mar. Chem. 46, 345-359. Schindler P. W. (1975) Removal of trace meta ls from the oceans: a zero order model. Thal. Jugosl. 11(1/2), 101-111. Schindler P. W. and Stumm W. (1987) The surface chemistry of oxides, hydroxides, and oxide minerals. In Aquatic Surface Chemistry. Chemical Processes at the Particle-Water Interface (ed. W. Stumm), pp. 83-110. John Wiley & Sons, New York, NY. Shannon R. D. (1976) Revised effective ionic ra dii and systematic studies of interatomic distances in halides and chalcogenides. Acta Crystallogr. A32, 751-767. Sholkovitz E. R. (1976) Flocculation of disso lved organic and inor ganic matter during the mixing of river water and seawater. Geochim. Cosmochim. Acta 40, 831-845. Sholkovitz E. R. (1992) Chemical evolution of rare earth elements : fractionation between colloidal and solution phases of filtered river water. Earth Planet. Sci. Lett. 114, 77-84. Sholkovitz E. R. (1993) The geochemistry of rare earth elements in the Amazon River estuary. Geochim. Cosmochim. Acta 57, 2181-2190. Sholkovitz E. R. (1995) The aquatic chemistr y of rare earth elem ents in rivers and estuaries. Aquat. Geochem. 1, 1-34. Sholkovitz E. R. and Elderfield H. (1988) Cy cling of dissolved rare earth elements in Chesapeake Bay. Glob. Biogeochem. Cycles 2(2), 157-176. Sholkovitz E. R., Piepgras D. J., and Jacobs en S. B. (1989) The pore water chemistry of rare earth elements in Buzzards Bay sediments. Geochim. Cosmochim. Acta 53, 2847-2856. Sholkovitz E. R., Shaw T. J., and Schneider D. L. (1992) The geochemi stry of ra re earth elements in the seasonally anoxic water column and porewaters of Chesapeake Bay. Geochim. Cosmochim. Acta 56, 3389-3402.

PAGE 140

125 Sholkovitz E. R., Church T. M., and Arimot o R. (1993) Rare earth element composition of precipitation, precipitati on particles, and aerosols. J. Geophys. Res. 98(D11), 20,587-20,599. Sholkovitz E. R., Landing W. M., and Lewis B. L. (1994) Ocean particle chemistry: The fractionation of rare earth elements be tween suspended particles and seawater. Geochim. Cosmochim. Acta 58(6), 1567-1579. Siekierski S. (1981) The position of americiu m within lanthanides in respect to unit cell volumes of isostructural compounds. A co mparison of covalency in lanthanide and actinide compounds. Inorg. Nucl. Chem. Lett. 17(5/6), 169-173. Smedley P. L. (1991) The geochemistry of ra re earth elements in groundwater from the Carnmenellis area, southwest England. Geochim. Cosmochim. Acta 55, 27672779. Smith R. M. and Martell A. E. (1975) Critical Stability Cons tants, Vol. 2: Amines. Plenum Press. Smith R. M. and Martell A. E. (1976) Critical Stability Constants, Vol. 4: Inorganic Complexes. Plenum Press. Smith R. M. and Martell A. E. (1989) Critical Stability Const ants, Vol. 6: Second Supplement. Plenum Press. Sonke J. E. and Salters V. J. M. (2006) Lanthanide-humic substances complexation. I. Experimental evidence for a lanthanide contraction effect. Geochim. Cosmochim. Acta 70, 1495-1506. Spahiu K. (1985) Studies on metal carbonate equilibria. 11. Yttrium(III) carbonate complex formation in aqueous perchlor ate media of various ionic strength. Acta Chem. Scand. A 39, 33-45. Stanley J. K., Jr. and Byrne R. H. (1990) The influence of solution chemistry on REE uptake by Ulva lactuca L. in seawater. Geochim. Cosmochim. Acta 54, 15871595. Swallow K. C., Hume D. N., and Morel F. M. M. (1980) Sorption of Copper and Lead by Hydrous Ferric Oxide. Environ. Sci. Technol. 14(11), 1326-1331. Tachikawa K., Jeandel C., and Dupr B. (1997) Distribution of rare earth elements and neodymium isotopes in settling particulate ma terial of the tropical Atlantic Ocean (EUMELI site). Deep-Sea Res. I 44(11), 1769-1792. Tachikawa K., Jeandel C., Vangriesheim A., and Dupr B. (1999) Distribution of rare earth elements and neodymium isotopes in suspended particles of the tropical Atlantic Ocean (EUMELI site). Deep-Sea Res. I 46, 733-755. Takahashi Y., Kimura T., Kato Y., Minai Y., and Tominaga T. (1998) Characterization of Eu(III) species sorbed on silica a nd montmorillonite by laser-induced fluorescence spectroscopy. Radiochim. Acta 82, 227-232. Takahashi Y., Chatellier X., Hattori K. H., Ka to K., and Fortin D. (2005) Adsorption of rare earth elements onto bacterial cell wa lls and its implication for REE sorption onto natural microbial mats. Chem. Geol. 219, 53-67. Tang J. and Johannesson K. H. (2003) Specia tion of rare earth elements in natural terrestrial waters: Assessing the role of dissolved organic matter from the modeling approach. Geochim. Cosmochim. Acta 67(13), 2321-2339.

PAGE 141

126 Tang J. and Johannesson K. H. (2005) Adsorp tion of rare earth el ements onto Carrizo sand: Experimental investigations and modeling with surface complexation. Geochim. Cosmochim. Acta 69(22), 5247-5261. Taylor S. R. and McLennan S. M. (1985) The Continental Crust: Its Composition and Evolution. Blackwell Science Publishers. Tertre E., Berger G., Castet S., Loubet M. and Giffaut E. (2005) Experimental sorption of Ni2+, Cs+ and Ln3+ onto a montmorillonite up to 150oC. Geochim. Cosmochim. Acta 69(21), 4937-4948. Thompson S. W. and Byrne R. H. (1988) Indica tor ligands in metal complexation studies: Role of 4-(2-pyridylazo)resorcinol in europium carbonate equilibrium investigations. Anal. Chem. 60, 19-22. Thomson J., Carpenter M. S. N., Colley S., Wilson T. R. S., Elderfield H., and Kennedy H. (1984) Metal accumulation rates in nor thwest Atlantic pelagic sediments. Geochim. Cosmochim. Acta 48, 1935-1948. Tipping E. (1994) WHAM-A chemical equilibri um model and computer code for waters, sediments, and soils incorporating a discrete site/electrostatic model of ionbinding by humic substances. Comput. Geosci. 20, 973-1023. Turner D. R., Whitfield M., and Dickson A. G. (1981) The equilibrium speciation of dissolved components in freshwater and seawater at 25C and 1 atm pressure. Geochim. Cosmochim. Acta 45, 855-881. Wood S. A. (1990) The aqueous geochemistry of the rare-earth elemen ts and yttrium. 1. Review of available low-temperature data for inorganic complexes and the inorganic REE speciation of natural waters. Chem. Geol. 82, 159-186. Wood S. A., Wesolowski D. J., and Palmer D. A. (2000) The aqueous geochemistry of the rare earth elements IX. A potentiometric study of Nd3+ complexation with acetate in 0.1 molal NaCl solution from 25C to 225C. Chem. Geol. 167, 231253. Zhang C., Wang L., Zhang S., and Li X. (1998) Geochemistry of rare earth elements in the mainstream of the Yangtze River, China. Appl. Geochem. 4, 451-462. Zhang J. and Nozaki Y. (1996) Rare earth el ements and yttrium in seawater: ICP-MS determinations in the East Caroline, Coral Sea, and South Fiji basins of the western South Pacific Ocean. Geochim. Cosmochim. Acta 60(23), 4631-4644.

PAGE 142

127Appendices

PAGE 143

128Appendix A: Data for Freshly Precipitate d Hydroxides of Trivalent Cations (Al3+, Ga3+, and In3+) Table A.1. Distribution coefficient (log iKAl) results from the expe riment performed at pH = 5.86 0.18 with an aluminum concentration of 1.00 0.05 mM. I (M) 0.016 0.016 0.016 pH 6.06 5.81 5.70 time 15 min 44 hrs 68 hrs Y 2.64 2.26 2.03 La 1.71 1.16 1.02 Ce 1.98 1.50 1.37 Pr 2.13 1.66 1.40 Nd 2.23 1.77 1.58 Pm Sm 2.55 2.18 1.97 Eu 2.64 2.26 2.06 Gd 2.61 2.21 2.00 Tb 2.78 2.41 2.18 Dy 2.86 2.50 2.29 Ho 2.86 2.51 2.31 Er 2.95 2.60 2.40 Tm 3.09 2.75 2.53 Yb 3.25 2.91 2.68 Lu 3.26 2.92 2.70

PAGE 144

129Appendix A (Continued) Table A.2. Distribution coefficient (log iKGa) results from the experiment performed at pH = 6.12 0.34 with a gallium concentration of 1.11 0.06 mM. I (M) 0.014 0.014 0.014 0.014 0.014 0.014 pH 6.34 6.35 6.33 6.18 6.04 5.47 time 15 min 90 min 5 hrs 24 hrs 44 hrs 141 hrs Y 3.30 3.47 3.26 2.86 2.56 2.27 La 2.56 2.79 2.49 2.22 2.00 1.54 Ce 3.01 3.22 2.97 2.60 2.34 1.90 Pr 3.24 3.44 3.21 2.79 2.52 1.98 Nd 3.38 3.56 3.34 2.91 2.62 2.06 Pm Sm 3.66 3.82 3.64 3.17 2.87 2.24 Eu 3.68 3.85 3.67 3.20 2.90 2.31 Gd 3.57 3.74 3.54 3.11 2.84 2.31 Tb 3.68 3.84 3.66 3.22 2.95 2.44 Dy 3.72 3.88 3.69 3.23 2.97 2.47 Ho 3.71 3.86 3.66 3.20 2.93 2.44 Er 3.77 3.92 3.72 3.25 2.98 2.51 Tm 3.87 4.01 3.83 3.35 3.07 2.59 Yb 4.00 4.12 3.95 3.46 3.19 2.70 Lu 3.99 4.11 3.95 3.44 3.16 2.69

PAGE 145

130Appendix A (Continued) Table A.3. Distribution coefficient (log iKIn) results from the experiment performed at pH = 6.08 0.04 with an indium concentration of 1.09 0.05 mM. I (M) 0.014 0.014 0.014 0.014 0.014 0.014 pH 6.13 6.11 6.12 6.07 6.02 6.05 time 15 min 90 min 5 hrs 24 hrs 96 hrs 100 hrs Y 2.21 2.22 2.22 2.22 2.20 2.20 La 1.50 1.54 1.56 1.63 1.58 1.63 Ce 1.85 1.90 1.93 1.94 1.94 1.98 Pr 1.89 1.95 1.96 1.99 1.99 2.01 Nd 1.93 1.99 2.00 2.04 2.05 2.05 Pm Sm 2.26 2.30 2.30 2.32 2.31 2.33 Eu 2.33 2.35 2.37 2.38 2.38 2.39 Gd 2.24 2.27 2.28 2.29 2.29 2.31 Tb 2.47 2.49 2.50 2.51 2.51 2.51 Dy 2.54 2.56 2.56 2.57 2.57 2.57 Ho 2.50 2.52 2.52 2.52 2.52 2.52 Er 2.57 2.59 2.59 2.59 2.59 2.60 Tm 2.77 2.79 2.79 2.79 2.80 2.80 Yb 3.02 3.04 3.04 3.05 3.06 3.06 Lu 3.01 3.03 3.03 3.04 3.04 3.05

PAGE 146

131Appendix B: pH Dependent Data for Amorphous Ferric Hydroxide Table B.1. Distribution coefficient (log iKFe) results from the expe riment performed at pH = 5.15 0.02 with an iron concentration of 0.613 0.042 mM. The precision in log iKFe was estimated to be 0.08 by statistically combini ng the precision of the YREE analyses with the uncertainty in the concentr ation of the ferric chloride solution. For Y the log iKFe precision is 0.2. I (M) 0.0122 0.0122 0.0122 0.0122 0.0122 0.0122 pH 5.17 5.17 5.17 5.13 5.12 5.12 time 15 min 90 min 5 hrs 24 hrs 46 hrs 48 hrs Y 2.38 2.31 2.28 2.10 1.81 1.67 La 2.21 2.15 2.19 2.17 2.15 2.14 Ce 2.63 2.61 2.62 2.62 2.62 2.60 Pr 2.75 2.75 2.77 2.79 2.79 2.77 Nd 2.82 2.82 2.85 2.87 2.88 2.87 Pm Sm 2.96 2.96 2.99 3.00 3.00 2.99 Eu 2.92 2.92 2.93 2.94 2.95 2.93 Gd 2.73 2.71 2.73 2.72 2.73 2.72 Tb 2.77 2.74 2.76 2.74 2.76 2.74 Dy 2.75 2.74 2.73 2.71 2.71 2.69 Ho 2.68 2.67 2.67 2.63 2.62 2.60 Er 2.69 2.69 2.67 2.66 2.63 2.61 Tm 2.75 2.75 2.73 2.73 2.72 2.70 Yb 2.82 2.82 2.81 2.82 2.82 2.80 Lu 2.79 2.79 2.78 2.77 2.78 2.75

PAGE 147

132Appendix B (Continued) Table B.2. Distribution coefficient (log iKFe) results from the expe riment performed at pH = 6.12 0.05 with an iron concentration of 0.108 0.008 mM. The precision in log iKFe was estimated to be 0.08 by statistically combini ng the precision of the YREE analyses with the uncertainty in the concentr ation of the ferric chloride solution. For Y the log iKFe precision is 0.2. I (M) 0.0107 0.0107 0.0107 0.0107 0.0107 0.0107 pH 6.12 6.16 6.17 6.10 6.07 6.07 time 15 min 90 min 5 hrs 24 hrs 46 hrs 48 hrs Y 3.74 3.71 3.71 3.55 3.55 3.54 La 3.43 3.40 3.42 3.38 3.40 3.39 Ce 3.85 3.87 3.88 3.86 3.88 3.87 Pr 4.00 4.03 4.06 4.04 4.05 4.06 Nd 4.06 4.11 4.12 4.11 4.12 4.13 Pm Sm 4.24 4.29 4.32 4.28 4.28 4.28 Eu 4.21 4.26 4.27 4.23 4.23 4.24 Gd 4.03 4.06 4.07 4.03 4.03 4.04 Tb 4.07 4.15 4.15 4.10 4.10 4.11 Dy 4.11 4.14 4.15 4.10 4.09 4.10 Ho 4.05 4.08 4.08 4.03 4.02 4.02 Er 4.07 4.10 4.11 4.05 4.05 4.05 Tm 4.16 4.20 4.21 4.16 4.15 4.16 Yb 4.25 4.30 4.31 4.26 4.26 4.26 Lu 4.21 4.25 4.26 4.21 4.21 4.21

PAGE 148

133Appendix B (Continued) Table B.3. Distribution coefficient (log iKFe) results from the expe riment performed at pH = 7.06 0.05 with an iron concentration of 0.108 0.008 mM. The precision in log iKFe was estimated to be 0.08 by statistically combini ng the precision of the YREE analyses with the uncertainty in the concen tration of the ferric chloride solution. The sample at 15 minutes was excluded from all calculations because it was anomalously low, possibly due to a dilution error. I (M) 0.0109 0.0109 0.0109 0.0109 0.0109 0.0109 pH 7.11 7.10 7.09 7.04 7.00 7.00 time 15 min 90 min 5 hrs 24 hrs 46 hrs 48 hrs Y 4.90 5.46 5.50 5.56 5.48 5.48 La 4.65 5.08 5.16 5.29 5.22 5.23 Ce 4.99 5.65 5.73 5.88 5.80 5.81 Pr 5.09 5.81 5.86 5.99 5.93 5.92 Nd 5.14 5.94 5.99 6.15 6.10 6.08 Pm Sm 5.26 6.14 6.17 6.28 6.27 6.21 Eu 5.23 6.09 6.11 6.22 6.22 6.16 Gd 5.13 5.90 5.93 6.05 6.00 5.98 Tb 5.18 6.01 6.04 6.14 6.11 6.08 Dy 5.18 5.98 6.00 6.10 6.07 6.03 Ho 5.14 5.88 5.92 6.00 5.96 5.94 Er 5.16 5.95 5.97 6.06 6.03 5.99 Tm 5.20 6.03 6.06 6.14 6.10 6.08 Yb 5.26 6.12 6.15 6.23 6.21 6.17 Lu 5.22 6.04 6.06 6.14 6.11 6.09

PAGE 149

134Appendix B (Continued) Table B.4. Distribution coefficient (log iKFe) results from the experiment performed over the pH range 5.1 – 7.0 with an iron concentration of 0.108 0.008 mM. The precision in log iKFe was estimated to be 0.08 by statistically combini ng the precision of the YREE analyses with the uncertainty in the con centration of the ferric chloride solution. I (M) 0.0107 0.0107 0.0107 0.0107 0.0107 pH 5.07 5.64 6.07 6.51 7.03 time 60 min 60 min 60 min 60 min 60 min Y 3.02 3.14 3.47 4.02 5.02 La 2.83 2.90 3.28 3.75 4.68 Ce 3.16 3.33 3.71 4.23 5.21 Pr 3.13 3.40 3.87 4.43 5.44 Nd 3.18 3.47 3.94 4.51 5.52 Pm Sm 3.32 3.62 4.11 4.70 5.74 Eu 3.30 3.59 4.06 4.66 5.71 Gd 3.22 3.44 3.87 4.44 5.46 Tb 3.24 3.49 3.94 4.54 5.60 Dy 3.19 3.45 3.92 4.54 5.61 Ho 3.10 3.34 3.83 4.47 5.54 Er 3.10 3.37 3.87 4.50 5.58 Tm 3.15 3.43 3.96 4.62 5.73 Yb 3.18 3.51 4.06 4.74 5.87 Lu 3.18 3.48 4.01 4.69 5.82

PAGE 150

135Appendix B (Continued) Table B.5. Distribution coefficient (log iKFe) results from the experiment performed ove r the pH range 5.1 – 7.0 with an iron concentration of 0.108 0.008 mM. The precision in log iKFe was estimated to be 0.08 by statistically combining the precision of the YREE analyses with the uncertainty in the concentration of the ferric chloride solution. I (M) 0.0103 0.0103 0.0103 0.0103 0.0103 0.0103 0.0103 0.0103 0.0103 0.0103 pH 5.11 5.11 5.59 5.58 6.03 6.01 6.60 6.58 7.03 7.02 time 15 min 60 min 15 min 60 min 15 min 60 min 15 min 60 min 15 min 60 min Y 2.95 2.91 3.12 3.08 3.39 3.42 4.16 4.19 5.09 5.17 La 2.60 2.69 2.86 2.82 3.10 3.18 3.83 3.86 4.67 4.75 Ce 3.07 3.07 3.27 3.26 3.57 3.61 4.32 4.35 5.25 5.34 Pr 3.07 3.07 3.35 3.35 3.72 3.76 4.51 4.54 5.42 5.51 Nd 3.16 3.15 3.44 3.44 3.79 3.84 4.59 4.62 5.51 5.59 Pm Sm 3.28 3.27 3.59 3.58 3.96 4.00 4.81 4.82 5.72 5.80 Eu 3.27 3.26 3.56 3.55 3.93 3.97 4.77 4.79 5.70 5.78 Gd 3.15 3.16 3.42 3.40 3.75 3.79 4.55 4.59 5.48 5.55 Tb 3.20 3.20 3.46 3.45 3.82 3.86 4.67 4.69 5.61 5.69 Dy 3.17 3.15 3.43 3.41 3.80 3.84 4.68 4.70 5.64 5.72 Ho 3.11 3.09 3.34 3.31 3.72 3.76 4.61 4.63 5.58 5.66 Er 3.12 3.09 3.35 3.33 3.75 3.78 4.64 4.67 5.62 5.70 Tm 3.15 3.13 3.41 3.40 3.83 3.87 4.76 4.77 5.74 5.83 Yb 3.20 3.17 3.49 3.48 3.92 3.97 4.87 4.89 5.89 5.98 Lu 3.17 3.16 3.47 3.45 3.89 3.93 4.83 4.84 5.85 5.95

PAGE 151

136Appendix B (Continued) Table B.6. Distribution coefficient (log iKFe) results from the experiment performed ove r the pH range 3.9 – 5.6 with an iron concentration of 10.0 0.7 mM. The precision in log iKFe was estimated to be 0.15 by statistically combining the precision of the YREE analyses with the uncertainty in the concentration of the ferric chloride solution. For Y the log iKFe precision is 0.3. n.v. = no value because measured YREE concentrations were indistin guishable from YREE concentra tions at t = 0 (i.e., [MSi]T = 0 in equation (3.1)). (Continued on next page) I (M) 0.0434 0.0434 0.0434 0.0434 0.0434 0.0434 0.0434 0.0434 0.0434 0.0434 pH 4.14 4.10 3.99 3.98 3.93 3.90 3.89 3.88 3.88 3.88 time 15 min 90 min 17.5 hrs 22 hrs 46 hrs 88 hrs 118 hrs 142 hrs 160 hrs 162 hrs Y 1.00 0.93 0.77 0.27 n.v. n.v. n.v. n.v. n.v. n.v. La 0.78 0.73 0.54 0.45 0.70 0.48 0.20 0.35 0.47 0.47 Ce 1.12 1.03 0.95 0.91 0.97 0.82 0.81 0.87 0.91 0.89 Pr 1.07 1.00 0.95 0.85 0.98 0.82 0.82 0.87 0.87 0.87 Nd 1.12 1.07 0.99 0.96 1.03 0.94 0.93 1.01 1.02 1.01 Pm Sm 1.23 1.18 1.10 1.06 1.10 1.02 1.04 1.05 1.07 1.10 Eu 1.17 1.13 1.01 0.99 1.08 0.96 0.98 1.02 1.07 1.06 Gd 1.05 1.00 0.93 0.86 0.95 0.86 0.78 0.84 0.93 0.93 Tb 1.06 1.00 0.86 0.81 0.93 0.76 0.77 0.89 0.92 0.94 Dy 1.04 0.97 0.81 0.77 0.81 0.60 0.80 0.82 0.84 0.83 Ho 1.08 1.01 0.81 0.77 0.82 0.45 0.78 0.73 0.73 0.65 Er 1.10 1.01 0.85 0.81 0.84 0.54 0.82 0.82 0.74 0.72 Tm 1.17 1.09 0.97 0.92 0.95 0.75 0.92 0.87 0.86 0.81 Yb 1.19 1.11 1.01 0.96 1.02 0.85 1.04 1.03 1.01 0.96 Lu 1.23 1.16 1.06 1.05 1.05 0.87 1.04 1.04 1.01 1.00

PAGE 152

137Appendix B (Continued) Table B.6. (Continued) (Continued on next page) I (M) 0.0434 0.0434 0.0434 0.0434 0.0434 0.0434 0.0434 0.0434 0.0434 0.0434 pH 4.15 4.14 4.11 4.67 4.64 4.62 4.60 5.14 5.10 5.09 time 60 min 24 hrs 5 days 60 min 24 hrs 48 hrs 6 days 60 min 24 hrs 48 hrs Y 0.40 0.63 n.v. 1.39 1.45 1.44 1.30 2.20 2.24 2.28 La 0.38 0.77 0.34 1.21 1.31 1.32 1.33 2.11 2.21 2.24 Ce 0.96 1.17 1.06 1.70 1.81 1.81 1.88 2.62 2.71 2.76 Pr 1.07 1.26 1.21 1.89 2.01 2.02 2.06 2.83 2.91 2.95 Nd 1.16 1.36 1.32 1.98 2.11 2.12 2.17 2.94 3.01 3.06 Pm Sm 1.25 1.44 1.41 2.10 2.22 2.22 2.28 3.06 3.12 3.15 Eu 1.20 1.39 1.36 2.03 2.15 2.16 2.21 2.98 3.05 3.09 Gd 1.02 1.22 1.18 1.84 1.95 1.96 2.01 2.75 2.83 2.87 Tb 1.02 1.18 1.13 1.83 1.92 1.93 1.98 2.73 2.80 2.84 Dy 1.02 1.16 1.10 1.81 1.89 1.90 1.97 2.71 2.76 2.80 Ho 1.03 1.18 1.08 1.80 1.88 1.88 1.94 2.66 2.72 2.76 Er 1.10 1.23 1.15 1.86 1.92 1.93 1.99 2.70 2.76 2.80 Tm 1.20 1.33 1.27 1.94 2.01 2.01 2.07 2.79 2.83 2.87 Yb 1.32 1.42 1.38 2.03 2.10 2.10 2.16 2.87 2.91 2.96 Lu 1.32 1.42 1.38 2.04 2.11 2.11 2.16 2.88 2.92 2.96

PAGE 153

138Appendix B (Continued) Table B.6. (Continued) I (M) 0.0434 0.0434 0.0434 0.0434 0.0434 pH 5.05 5.60 5.50 5.50 5.48 time 8 days 60 min 24 hrs 48 hrs 5 days Y 2.23 3.17 3.10 3.13 3.08 La 2.21 3.21 3.16 3.18 3.13 Ce 2.71 3.83 3.72 3.77 3.71 Pr 2.91 3.94 3.84 3.86 3.82 Nd 3.01 4.08 3.99 3.99 3.93 Pm Sm 3.11 4.13 4.02 4.06 4.01 Eu 3.05 4.13 4.00 4.04 3.98 Gd 2.83 3.81 3.72 3.75 3.69 Tb 2.80 3.77 3.69 3.72 3.67 Dy 2.76 3.73 3.64 3.68 3.63 Ho 2.73 3.64 3.58 3.61 3.56 Er 2.76 3.68 3.61 3.64 3.59 Tm 2.84 3.73 3.66 3.69 3.65 Yb 2.92 3.81 3.73 3.76 3.72 Lu 2.92 3.83 3.74 3.79 3.74

PAGE 154

139Appendix C: Ionic Streng th Dependent Data for Amorphous Ferric Hydroxide Table C.1. Distribution coefficient (log iKFe) results from the experiment performed over the ionic strength ra nge 0.01 – 0.09 M with an iron concentration of 0.108 0.008 mM. The precision in log iKFe was estimated to be 0.08 by statistically combining the precision of the YREE analyses with the uncertainty in the con centration of the ferric chloride solution. The 5 minute sample a t I = 0.03 M was excluded from all calculations because it was anomalously low, possibly due to a dilution error. (Continued on next page) I (M) 0.0101 0.0101 0.0101 0.0101 0.0302 0.0302 0.0302 0.0302 0.0503 0.0503 pH 6.07 6.15 6.08 6.07 6.19 6.18 6.18 6.15 6.13 6.13 time 5 min 4 hrs 24 hrs 44.5 hrs 5 min 90 min 4 hrs 22 hrs 5 min 90 min Y 3.57 3.61 3.50 3.43 3.45 3.57 3.58 3.59 3.65 3.66 La 3.29 3.36 3.32 3.29 3.32 3.43 3.43 3.45 3.44 3.45 Ce 3.75 3.86 3.83 3.80 3.81 3.93 3.94 3.97 3.94 3.96 Pr 3.90 4.04 4.00 3.98 3.96 4.10 4.13 4.15 4.13 4.14 Nd 3.97 4.12 4.07 4.05 4.03 4.18 4.20 4.22 4.20 4.22 Pm Sm 4.16 4.32 4.24 4.22 4.19 4.34 4.37 4.39 4.36 4.38 Eu 4.13 4.28 4.20 4.17 4.15 4.30 4.33 4.34 4.31 4.33 Gd 3.95 4.08 4.00 3.97 3.95 4.10 4.13 4.14 4.10 4.13 Tb 4.03 4.17 4.08 4.05 4.03 4.18 4.21 4.22 4.18 4.20 Dy 4.03 4.16 4.07 4.03 4.01 4.16 4.19 4.20 4.19 4.20 Ho 3.96 4.09 3.99 3.95 3.92 4.08 4.11 4.12 4.13 4.14 Er 3.99 4.11 4.01 3.98 3.95 4.11 4.14 4.16 4.16 4.17 Tm 4.07 4.22 4.11 4.08 4.05 4.21 4.25 4.27 4.26 4.28 Yb 4.16 4.33 4.22 4.19 4.15 4.32 4.36 4.38 4.37 4.39 Lu 4.13 4.27 4.17 4.13 4.10 4.28 4.31 4.33 4.32 4.34

PAGE 155

140 Appendix C (Continued) Table C.1. (Continued) I (M) 0.0503 0.0503 0.0700 0.0700 0.0700 0.0700 0.0900 0.0900 0.0900 0.0900 pH 6.13 6.07 6.12 6.13 6.13 6.09 6.15 6.15 6.15 6.10 time 4 hrs 46 hrs 5 min 90 min 4 hrs 22 hrs 5 min 90 min 4 hrs 23 hrs Y 3.66 3.56 3.60 3.64 3.63 3.61 3.64 3.67 3.69 3.66 La 3.46 3.37 3.41 3.44 3.44 3.42 3.46 3.48 3.46 3.45 Ce 3.96 3.90 3.92 3.95 3.95 3.94 3.96 3.99 4.00 3.98 Pr 4.14 4.07 4.10 4.13 4.13 4.12 4.14 4.18 4.18 4.16 Nd 4.22 4.15 4.17 4.22 4.21 4.19 4.22 4.26 4.26 4.25 Pm Sm 4.38 4.30 4.33 4.37 4.36 4.35 4.37 4.41 4.42 4.40 Eu 4.33 4.25 4.27 4.31 4.31 4.29 4.32 4.35 4.36 4.35 Gd 4.12 4.04 4.07 4.10 4.10 4.09 4.11 4.15 4.16 4.14 Tb 4.20 4.12 4.14 4.18 4.18 4.16 4.18 4.22 4.23 4.21 Dy 4.21 4.11 4.14 4.18 4.18 4.16 4.18 4.22 4.23 4.20 Ho 4.14 4.05 4.07 4.12 4.12 4.10 4.12 4.16 4.17 4.14 Er 4.17 4.08 4.11 4.15 4.15 4.13 4.15 4.19 4.20 4.17 Tm 4.28 4.18 4.21 4.26 4.26 4.24 4.26 4.30 4.31 4.28 Yb 4.38 4.29 4.32 4.37 4.36 4.35 4.37 4.41 4.42 4.39 Lu 4.34 4.25 4.27 4.32 4.32 4.30 4.32 4.37 4.38 4.35

PAGE 156

141Appendix C (Continued) Table C.2. Distribution coefficient (log iKFe) results from the experiment performed over the ionic strength ra nge 0.01 – 0.09 M with an iron concentration of 0.108 0.008 mM. The precision in log iKFe was estimated to be 0.08 by statistically combining the precision of the YREE analyses with the uncertainty in the concentration of the ferric chloride solution. I (M) 0.0105 0.0105 0.0301 0.0301 0.0500 0.0500 0.0700 0.0700 0.0945 0.0945 pH 6.10 6.13 6.03 6.04 6.13 6.13 6.12 6.12 6.13 6.13 time 15 min 60 min 15 min 60 min 15 min 60 min 15 min 60 min 15 min 60 min Y 3.65 3.71 3.58 3.67 3.84 3.66 3.67 3.69 3.69 3.70 La 3.37 3.42 3.29 3.40 3.54 3.35 3.36 3.38 3.37 3.38 Ce 3.82 3.89 3.76 3.86 4.01 3.85 3.84 3.86 3.86 3.88 Pr 3.97 4.05 3.92 4.02 4.18 4.02 4.01 4.04 4.03 4.05 Nd 4.04 4.12 3.99 4.10 4.25 4.09 4.08 4.11 4.11 4.12 Pm Sm 4.23 4.31 4.17 4.28 4.44 4.27 4.26 4.29 4.28 4.30 Eu 4.19 4.28 4.13 4.24 4.40 4.23 4.22 4.25 4.24 4.26 Gd 4.01 4.08 3.95 4.06 4.21 4.04 4.03 4.06 4.05 4.07 Tb 4.09 4.17 4.03 4.13 4.29 4.12 4.11 4.14 4.14 4.15 Dy 4.09 4.17 4.02 4.13 4.29 4.12 4.11 4.14 4.14 4.16 Ho 4.03 4.10 3.95 4.05 4.22 4.05 4.05 4.07 4.08 4.09 Er 4.05 4.13 3.97 4.07 4.25 4.08 4.07 4.10 4.11 4.12 Tm 4.14 4.22 4.06 4.16 4.34 4.18 4.17 4.20 4.20 4.22 Yb 4.23 4.32 4.16 4.26 4.44 4.28 4.27 4.30 4.31 4.32 Lu 4.19 4.28 4.12 4.22 4.40 4.24 4.23 4.26 4.26 4.28

PAGE 157

142Appendix D: Data for Amorphous Ferri c Hydroxide Covering a Range of Carbonate Concentrations (2 3T[CO]) Table D.1. Distribution coefficient (T iFelogK ) results from the experiment performed at pH = 5.38 0.02 and 30% CO2. I (M) 0.0118 0.0118 0.0118 0.0118 0.0117 0.0117 pH 5.41 5.39 5.39 5.38 5.36 5.36 time 15 min 90 min 5 hrs 24 hrs 46 hrs 48 hrs CO2 (%) 29.22 29.22 29.22 29.22 29.22 29.22 2 3T[CO] ( M) 0.0261 0.0238 0.0238 0.0227 0.0207 0.0207 Y 3.54 3.50 3.76 3.92 3.92 3.88 La 3.67 3.77 3.88 3.96 3.92 3.91 Ce 3.96 4.07 4.20 4.29 4.26 4.24 Pr 4.09 4.22 4.34 4.44 4.41 4.40 Nd 4.15 4.28 4.41 4.52 4.50 4.48 Pm Sm 4.27 4.40 4.53 4.64 4.62 4.61 Eu 4.23 4.36 4.49 4.60 4.58 4.56 Gd 4.08 4.21 4.33 4.44 4.41 4.40 Tb 4.09 4.22 4.35 4.46 4.44 4.43 Dy 4.05 4.17 4.30 4.42 4.41 4.39 Ho 4.00 4.11 4.25 4.37 4.36 4.35 Er 4.00 4.12 4.26 4.39 4.38 4.37 Tm 4.06 4.19 4.34 4.47 4.47 4.45 Yb 4.10 4.23 4.38 4.52 4.52 4.51 Lu 4.08 4.22 4.37 4.51 4.51 4.49

PAGE 158

143Appendix D (Continued) Table D.2. Distribution coefficient (T iFelogK) results from the experiment performed over the pH range 4.6 – 6.6 at 3% CO2. It should be noted that carbona te concentrations (2 3T[CO]) are listed in nM units. n.v. = no value because measured YREE concentrations were indistinguishable from YREE conc entrations at t = 0 (i.e., [MSi]T = 0 in equation (4.1)). I (M) 0.0106 0.0106 0.0107 0.0109 0.0111 0.0116 0.0128 pH 4.63 4.98 5.49 5.70 5.99 6.28 6.60 time 60 min 60 min 60 min 60 min 60 min 60 min 60 min CO2 (%) 2.904 2.904 2.904 2.904 2.904 2.904 2.904 2 3T[CO] (nM) 0.0693 0.347 3.65 9.64 36.9 142 638 Y n.v. n.v. 2.93 3.51 4.25 4.79 5.25 La 2.38 2.63 3.19 3.53 4.24 4.93 5.67 Ce 2.71 2.99 3.58 3.92 4.63 5.29 5.99 Pr 2.78 3.16 3.76 4.10 4.80 5.40 5.94 Nd 2.89 3.25 3.85 4.19 4.88 5.47 6.00 Pm Sm 2.95 3.38 4.00 4.34 5.00 5.53 5.96 Eu 2.93 3.34 3.96 4.29 4.96 5.48 5.92 Gd 2.85 3.23 3.79 4.12 4.79 5.34 5.80 Tb 2.93 3.28 3.85 4.17 4.82 5.34 5.77 Dy 2.78 3.18 3.81 4.15 4.79 5.29 5.70 Ho 2.60 3.06 3.74 4.09 4.73 5.22 5.63 Er 2.60 3.09 3.78 4.12 4.75 5.22 5.61 Tm 2.67 3.17 3.88 4.22 4.83 5.27 5.63 Yb 2.81 3.28 3.97 4.30 4.89 5.30 5.65 Lu 2.74 3.25 3.94 4.27 4.88 5.30 5.64

PAGE 159

144Appendix D (Continued) Table D.3. Distribution coefficient (T iFelogK) results from the experiment perfor med over the pH range 4.0 – 6.6 at 30% CO2. The samples at pH 3.98 and pH 4.49 for each YREE were excluded from any calculations due to weak sorption at low pH (see text for details). It should be noted th at carbonate concentrations (2 3T[CO]) are listed in nM units. n.v. = no value because measured YREE concentrations were indistinguishable from YREE concentrations at t = 0 (i.e., [MSi]T = 0 in equation (4.1)). I (M) 0.0106 0.0107 0.0109 0.0112 0.0120 0.0134 0.0162 0.0219 0.0342 pH 3.98 4.49 4.80 5.10 5.39 5.70 5.98 6.27 6.56 time 60 min 60 min 60 min 60 min 60 min 60 min 60 min 60 min 60 min CO2 (%) 29.22 29.22 29.22 29.22 29.22 29.22 29.22 29.22 29.22 2 3T[CO] (nM) 0.0350 0.367 1.54 6.17 23.9 103 398 1680 7570 Y n.v. n.v. 2.55 3.19 3.70 4.24 4.70 4.97 5.07 La 2.10 2.59 2.87 3.25 3.71 4.36 5.03 5.48 5.91 Ce 2.34 2.71 3.13 3.56 4.04 4.66 5.26 5.72 6.04 Pr 2.42 2.97 3.30 3.72 4.21 4.78 5.32 5.68 5.91 Nd 2.49 3.04 3.36 3.79 4.28 4.83 5.35 5.70 5.93 Pm Sm 2.30 3.13 3.47 3.91 4.40 4.92 5.40 5.71 5.88 Eu 2.39 3.11 3.45 3.88 4.36 4.87 5.33 5.62 5.76 Gd 2.35 2.95 3.30 3.72 4.19 4.73 5.22 5.53 5.69 Tb 2.39 3.03 3.34 3.77 4.22 4.73 5.18 5.47 5.59 Dy 2.19 2.96 3.32 3.74 4.19 4.67 5.10 5.35 5.44 Ho 2.05 2.92 3.27 3.70 4.14 4.62 5.04 5.28 5.35 Er 2.21 2.96 3.31 3.74 4.17 4.63 5.03 5.25 5.30 Tm 2.14 3.05 3.40 3.83 4.25 4.67 5.05 5.25 5.27 Yb 2.44 3.14 3.48 3.90 4.31 4.70 5.05 5.24 5.25 Lu 2.31 3.09 3.46 3.88 4.29 4.70 5.05 5.23 5.23

PAGE 160

145Appendix D (Continued) Table D.4. Distribution coefficient (T iFelogK) results from the experiment performed over the 2COP range 0% – 30% at pH = 6.52 0.01. (Continued on next page) I (M) 0.0107 0.0107 0.0107 0.0107 0.0107 0.0107 0.0107 0.0108 0.0108 0.0108 pH 6.51 6.53 6.54 6.50 6.52 6.54 6.56 6.52 6.53 6.53 time 15 min 90 min 5 hrs 21 hrs 15 min 90 min 19 hrs 15 min 90 min 21 hrs CO2 (%) 0.0 0.0 0.0 0.0 0.00971 0.00971 0.00971 0.0969 0.0969 0.0969 2 3T[CO] (M) 0.0 0.0 0.0 0.0 0.00140 0.00154 0.00168 0.0140 0.0147 0.0147 Y 4.23 4.25 4.26 4.26 4.51 4.54 4.59 4.80 4.89 4.97 La 3.89 3.91 3.94 3.97 4.19 4.22 4.29 4.50 4.61 4.72 Ce 4.37 4.41 4.44 4.49 4.72 4.75 4.82 5.04 5.14 5.23 Pr 4.54 4.60 4.63 4.67 4.93 4.96 5.01 5.24 5.34 5.41 Nd 4.61 4.67 4.70 4.75 5.01 5.04 5.09 5.33 5.42 5.49 Pm Sm 4.82 4.89 4.90 4.93 5.20 5.23 5.28 5.51 5.61 5.66 Eu 4.80 4.86 4.88 4.90 5.16 5.19 5.24 5.47 5.57 5.62 Gd 4.60 4.65 4.67 4.69 4.94 4.97 5.02 5.25 5.35 5.41 Tb 4.70 4.76 4.77 4.79 5.04 5.07 5.12 5.34 5.44 5.49 Dy 4.72 4.78 4.79 4.81 5.06 5.09 5.14 5.36 5.45 5.51 Ho 4.66 4.72 4.73 4.75 5.01 5.04 5.09 5.31 5.40 5.46 Er 4.69 4.75 4.76 4.78 5.05 5.08 5.13 5.34 5.43 5.50 Tm 4.79 4.86 4.87 4.90 5.17 5.20 5.25 5.46 5.55 5.61 Yb 4.89 4.96 4.98 5.01 5.28 5.32 5.37 5.56 5.66 5.70 Lu 4.85 4.91 4.92 4.95 5.23 5.26 5.32 5.52 5.61 5.66

PAGE 161

146Appendix D (Continued) Table D.4. (Continued) (Continued on next page) I (M) 0.0109 0.0109 0.0109 0.0109 0.0110 0.0110 0.0110 0.0113 0.0113 0.0113 pH 6.54 6.52 6.52 6.52 6.52 6.52 6.52 6.52 6.51 6.51 time 15 min 90 min 25 hrs 46.5 hrs 15 min 90 min 21.5 hrs 15 min 90 min 23 hrs CO2 (%) 0.291 0.291 0.291 0.291 0.485 0.485 0.485 0.961 0.961 0.961 2 3T[CO] (M) 0.0463 0.0422 0.0422 0.0422 0.0705 0.0705 0.0705 0.141 0.134 0.134 Y 5.11 5.16 5.20 5.21 5.24 5.28 5.27 5.28 5.34 5.37 La 4.89 4.96 5.04 5.06 5.12 5.18 5.19 5.25 5.34 5.39 Ce 5.41 5.46 5.52 5.54 5.60 5.65 5.64 5.70 5.76 5.81 Pr 5.59 5.65 5.69 5.71 5.77 5.81 5.81 5.86 5.91 5.95 Nd 5.67 5.72 5.76 5.79 5.84 5.86 5.87 5.91 5.96 6.00 Pm Sm 5.82 5.88 5.90 5.93 5.96 5.99 5.99 6.02 6.05 6.10 Eu 5.78 5.83 5.85 5.88 5.91 5.94 5.94 5.96 6.00 6.04 Gd 5.58 5.63 5.66 5.69 5.73 5.77 5.76 5.79 5.84 5.87 Tb 5.64 5.69 5.71 5.74 5.76 5.80 5.79 5.81 5.86 5.88 Dy 5.64 5.68 5.70 5.72 5.74 5.78 5.76 5.77 5.82 5.84 Ho 5.59 5.62 5.64 5.67 5.69 5.72 5.71 5.71 5.76 5.78 Er 5.62 5.65 5.68 5.70 5.71 5.75 5.74 5.73 5.77 5.80 Tm 5.71 5.75 5.76 5.79 5.79 5.82 5.81 5.80 5.84 5.86 Yb 5.79 5.82 5.84 5.87 5.86 5.89 5.87 5.85 5.88 5.91 Lu 5.75 5.79 5.80 5.84 5.83 5.86 5.85 5.82 5.86 5.88

PAGE 162

147Appendix D (Continued) Table D.4. (Continued) I (M) 0.0124 0.0124 0.0124 0.0290 0.0290 0.0290 pH 6.53 6.52 6.51 6.51 6.51 6.51 time 15 min 90 min 20.5 hrs 15 min 90 min 23.5 hrs CO2 (%) 2.904 2.904 2.904 29.22 29.22 29.22 2 3T[CO] (M) 0.458 0.437 0.417 5.63 5.63 5.63 Y 5.37 5.41 5.44 5.20 5.21 5.28 La 5.54 5.58 5.66 5.74 5.76 5.89 Ce 5.91 5.94 6.01 5.96 5.96 6.06 Pr 6.00 6.02 6.07 5.99 5.97 6.06 Nd 6.04 6.06 6.09 5.99 5.98 6.06 Pm Sm 6.11 6.12 6.16 5.99 5.98 6.04 Eu 6.04 6.06 6.09 5.91 5.90 5.96 Gd 5.90 5.92 5.96 5.80 5.79 5.86 Tb 5.88 5.90 5.93 5.71 5.72 5.78 Dy 5.82 5.84 5.86 5.61 5.62 5.67 Ho 5.77 5.79 5.81 5.53 5.54 5.59 Er 5.77 5.78 5.80 5.49 5.49 5.54 Tm 5.82 5.83 5.84 5.50 5.51 5.53 Yb 5.83 5.85 5.86 5.48 5.49 5.51 Lu 5.82 5.83 5.84 5.45 5.46 5.48

PAGE 163

148Appendix D (Continued) Table D.5. Distribution coefficient (T iFelogK) results from the experiment performed over the 2COP range 0% – 30% at pH = 6.68 0.01. (Continued on next page) I (M) 0.0107 0.0107 0.0107 0.0107 0.0110 0.0110 0.0110 0.0111 0.0111 0.0111 pH 6.65 6.68 6.71 6.68 6.68 6.67 6.69 6.68 6.67 6.68 time 15 min 90 min 5 hrs 21.5 hrs 25 min 2 hrs 70 hrs 15 min 90 min 22 hrs CO2 (%) 0.0 0.0 0.0 0.0 0.291 0.291 0.291 0.485 0.485 0.485 2 3T[CO] (M) 0.0 0.0 0.0 0.0 0.0884 0.0844 0.0925 0.148 0.141 0.148 Y 4.69 4.76 4.83 4.81 5.55 5.61 5.68 5.72 5.70 5.71 La 4.34 4.39 4.48 4.51 5.45 5.56 5.68 5.76 5.80 5.78 Ce 4.79 4.89 4.99 5.03 5.95 6.06 6.09 6.18 6.20 6.19 Pr 4.98 5.09 5.18 5.21 6.10 6.21 6.22 6.30 6.31 6.30 Nd 5.04 5.16 5.26 5.28 6.23 6.37 6.29 6.39 6.37 6.38 Pm Sm 5.26 5.38 5.48 5.47 6.42 6.56 6.41 6.47 6.48 6.46 Eu 5.24 5.35 5.44 5.44 6.29 6.44 6.35 6.41 6.41 6.40 Gd 5.04 5.15 5.24 5.23 6.13 6.24 6.19 6.25 6.26 6.24 Tb 5.14 5.26 5.35 5.33 6.11 6.20 6.20 6.24 6.23 6.23 Dy 5.15 5.27 5.36 5.35 6.13 6.21 6.17 6.21 6.20 6.20 Ho 5.10 5.21 5.30 5.29 6.05 6.13 6.10 6.14 6.12 6.12 Er 5.13 5.25 5.33 5.32 6.10 6.18 6.12 6.16 6.13 6.13 Tm 5.23 5.35 5.44 5.44 6.24 6.32 6.20 6.23 6.21 6.21 Yb 5.33 5.47 5.56 5.56 6.22 6.31 6.23 6.25 6.23 6.23 Lu 5.28 5.41 5.51 5.50 6.15 6.23 6.21 6.22 6.20 6.21

PAGE 164

149Appendix D (Continued) Table D.5. (Continued) I (M) 0.0116 0.0115 0.0116 0.0133 0.0132 0.0132 0.0387 0.0386 0.0387 pH 6.69 6.67 6.68 6.70 6.69 6.69 6.69 6.68 6.69 time 15 min 90 min 21.5 hrs 15 min 90 min 21 hrs 15 min 90 min 20.5 hrs CO2 (%) 0.961 0.961 0.961 2.904 2.904 2.904 29.22 29.22 29.22 2 3T[CO] (M) 0.310 0.282 0.296 1.02 0.974 0.974 14.5 13.8 14.5 Y 5.77 5.75 5.75 5.75 5.75 5.75 5.34 5.35 5.35 La 5.95 5.96 5.94 6.11 6.11 6.12 6.11 6.13 6.12 Ce 6.32 6.31 6.31 6.37 6.37 6.37 6.22 6.25 6.23 Pr 6.42 6.41 6.40 6.48 6.47 6.46 6.26 6.28 6.26 Nd 6.47 6.48 6.46 6.47 6.46 6.46 6.24 6.27 6.23 Pm Sm 6.54 6.56 6.52 6.48 6.48 6.46 6.17 6.19 6.16 Eu 6.47 6.47 6.44 6.42 6.42 6.40 6.09 6.11 6.09 Gd 6.32 6.33 6.31 6.31 6.30 6.29 5.99 6.01 5.99 Tb 6.30 6.29 6.27 6.25 6.25 6.24 5.88 5.89 5.88 Dy 6.24 6.23 6.21 6.17 6.17 6.15 5.74 5.75 5.74 Ho 6.17 6.15 6.14 6.11 6.10 6.10 5.65 5.66 5.64 Er 6.17 6.16 6.15 6.09 6.08 6.08 5.59 5.59 5.58 Tm 6.23 6.22 6.21 6.12 6.11 6.11 5.57 5.57 5.56 Yb 6.24 6.23 6.21 6.13 6.12 6.11 5.54 5.54 5.53 Lu 6.22 6.20 6.20 6.11 6.11 6.10 5.51 5.51 5.49

PAGE 165

150Appendix D (Continued) Table D.6. Distribution coefficient (T iFelogK) results from the experiment performed over the 2COP range 0% – 30% at pH = 7.06 0.01. (Continued on next page) I (M) 0.0107 0.0107 0.0107 0.0107 0.0113 0.0113 0.0113 0.0117 0.0117 0.0117 pH 7.04 7.06 7.08 7.05 7.07 7.08 7.07 7.06 7.06 7.07 time 15 min 90 min 5 hrs 21 hrs 15 min 90 min 21.5 hrs 15 min 90 min 45 hrs CO2 (%) 0.0 0.0 0.0 0.0 0.291 0.291 0.291 0.485 0.485 0.485 2 3T[CO] (M) 0.0 0.0 0.0 0.0 0.537 0.562 0.537 0.863 0.863 0.903 Y 5.48 5.63 5.68 5.70 6.10 6.18 6.20 6.12 6.19 6.13 La 5.10 5.25 5.31 5.37 6.27 6.37 6.39 6.45 6.51 6.49 Ce 5.62 5.80 5.88 5.97 6.86 6.94 7.01 7.01 7.11 7.06 Pr 5.83 5.99 6.06 6.05 6.81 6.88 6.94 6.91 6.98 6.93 Nd 5.91 6.09 6.16 6.22 6.91 7.02 7.12 7.01 7.10 7.10 Pm Sm 6.16 6.34 6.39 6.45 7.00 7.09 7.16 7.07 7.14 7.10 Eu 6.12 6.30 6.34 6.39 6.86 6.94 6.97 6.90 6.96 6.91 Gd 5.92 6.09 6.13 6.18 6.77 6.86 6.92 6.81 6.92 6.83 Tb 6.03 6.19 6.24 6.28 6.71 6.79 6.83 6.74 6.80 6.74 Dy 6.02 6.18 6.24 6.28 6.62 6.70 6.71 6.62 6.69 6.61 Ho 5.93 6.09 6.15 6.19 6.49 6.57 6.60 6.49 6.57 6.51 Er 5.96 6.12 6.18 6.22 6.47 6.54 6.56 6.46 6.52 6.47 Tm 6.07 6.24 6.30 6.35 6.53 6.59 6.63 6.51 6.58 6.52 Yb 6.18 6.35 6.44 6.49 6.56 6.64 6.68 6.54 6.61 6.55 Lu 6.13 6.31 6.38 6.42 6.54 6.61 6.65 6.51 6.59 6.54

PAGE 166

151Appendix D (Continued) Table D.6. (Continued) I (M) 0.0126 0.0125 0.0125 0.0167 0.0167 0.0167 0.0854 0.0844 0.0835 pH 7.06 7.04 7.04 7.07 7.06 7.06 7.07 7.06 7.05 time 15 min 90 min 22.5 hrs 15 min 90 min 20 hrs 15 min 90 min 20.5 hrs CO2 (%) 0.961 0.961 0.961 2.904 2.904 2.904 29.22 29.22 29.22 2 3T[CO] (M) 1.75 1.59 1.59 6.04 5.77 5.77 121 115 109 Y 6.07 5.99 6.08 5.82 5.90 5.85 4.69 4.73 4.71 La 6.52 6.37 6.56 6.47 6.46 6.52 6.58 6.69 6.75 Ce 7.04 6.85 7.06 6.81 6.88 6.91 6.44 6.53 6.57 Pr 6.91 6.79 6.93 6.71 6.79 6.77 6.55 6.61 6.60 Nd 7.02 6.90 7.06 6.77 6.85 6.83 6.51 6.61 6.56 Pm Sm 7.02 6.92 7.02 6.73 6.85 6.77 6.30 6.39 6.32 Eu 6.85 6.77 6.87 6.61 6.70 6.65 6.16 6.24 6.19 Gd 6.79 6.68 6.80 6.54 6.62 6.56 6.05 6.13 6.08 Tb 6.68 6.60 6.68 6.41 6.50 6.44 5.83 5.91 5.86 Dy 6.54 6.47 6.56 6.26 6.35 6.29 5.61 5.68 5.62 Ho 6.44 6.37 6.44 6.14 6.24 6.17 5.45 5.53 5.45 Er 6.39 6.33 6.40 6.09 6.18 6.12 5.33 5.41 5.33 Tm 6.43 6.37 6.43 6.09 6.20 6.13 5.27 5.34 5.28 Yb 6.45 6.38 6.45 6.08 6.18 6.11 5.21 5.28 5.22 Lu 6.43 6.37 6.43 6.05 6.16 6.09 5.16 5.23 5.16

PAGE 167

152Appendix D (Continued) Table D.7. Distribution coefficient (T iFelogK) results from the experiment performed over the 2COP range 0% – 30% at pH = 7.10 0.03. n.v. = no value because of an anomalous concentration reading. (Continued on next page) I (M) 0.0105 0.0105 0.0105 0.0105 0.0105 0.0105 0.0105 0.0105 0.0112 0.0112 pH 7.12 7.13 7.15 7.13 7.06 7.08 7.10 7.07 7.10 7.11 time 15 min 90 min 5 hrs 22 hrs 15 min 90 min 22.5 hrs 65.5 hrs 15 min 90 min CO2 (%) 0.0 0.0 0.0 0.0 0.00971 0.00971 0.00971 0.00971 0.291 0.291 2 3T[CO] (M) 0.0 0.0 0.0 0.0 0.0167 0.0184 0.0201 0.0175 0.614 0.643 Y 5.61 5.80 5.84 5.84 5.97 6.02 6.06 5.96 6.10 6.15 La 5.26 5.47 5.57 5.57 5.82 5.85 5.90 5.80 6.28 6.38 Ce 5.77 6.01 6.14 6.16 6.35 6.39 6.44 6.36 6.81 6.96 Pr 5.95 6.18 6.32 6.32 6.48 6.53 6.60 6.52 6.74 6.81 Nd 6.03 6.27 6.41 6.41 6.58 6.60 6.70 6.63 6.77 6.85 Pm Sm 6.23 6.49 6.68 6.69 6.77 6.83 6.91 6.84 6.88 6.97 Eu 6.22 6.50 6.71 6.70 6.81 6.87 6.98 6.88 6.82 6.90 Gd 6.04 6.29 6.46 6.45 6.58 6.63 6.72 6.61 6.66 6.73 Tb 6.11 6.34 6.48 6.48 6.56 n.v. 6.69 6.59 6.61 6.68 Dy 6.11 6.34 6.45 6.47 6.54 6.58 6.66 6.56 6.59 6.64 Ho 6.06 6.28 6.36 6.37 6.45 6.52 6.58 6.48 6.47 6.52 Er 6.08 6.29 6.37 6.38 6.45 6.52 6.58 6.48 6.47 6.51 Tm 6.18 6.39 6.46 6.48 6.53 6.61 6.67 6.58 6.54 6.57 Yb 6.27 6.50 6.58 6.61 6.62 6.72 6.77 6.69 6.53 6.57 Lu 6.22 6.44 6.50 6.54 6.58 6.65 6.72 6.63 6.51 6.54

PAGE 168

153Appendix D (Continued) Table D.7. (Continued) (Continued on next page) I (M) 0.0112 0.0116 0.0116 0.0116 0.0127 0.0127 0.0127 0.0176 0.0175 0.0175 pH 7.11 7.10 7.10 7.10 7.11 7.10 7.10 7.13 7.12 7.12 time 21.5 hrs 15 min 90 min 22 hrs 15 min 90 min 21 hrs 15 min 90 min 19.5 hrs CO2 (%) 0.291 0.485 0.485 0.485 0.961 0.961 0.961 2.904 2.904 2.904 2 3T[CO] (M) 0.643 1.03 1.03 1.03 2.20 2.11 2.11 8.10 7.72 7.72 Y 6.15 6.16 6.11 6.15 6.09 6.11 6.07 5.91 5.89 5.88 La 6.46 6.50 6.42 6.51 6.55 6.54 6.48 6.53 6.45 6.45 Ce 6.91 7.01 6.90 7.04 7.08 7.08 7.01 6.77 6.71 6.74 Pr 6.85 6.87 6.83 6.87 6.87 6.86 6.82 6.79 6.74 6.75 Nd 6.87 6.89 6.84 6.90 6.88 6.88 6.86 6.76 6.73 6.73 Pm Sm 6.96 6.97 6.93 6.99 6.96 6.95 6.89 6.72 6.69 6.72 Eu 6.91 6.90 6.85 6.91 6.87 6.88 6.82 6.62 6.59 6.61 Gd 6.75 6.75 6.72 6.75 6.72 6.72 6.67 6.49 6.45 6.46 Tb 6.68 6.67 6.64 6.67 6.62 6.64 6.59 6.40 6.39 6.39 Dy 6.65 6.62 6.58 6.62 6.56 6.56 6.52 6.33 6.30 6.31 Ho 6.51 6.50 6.46 6.49 6.43 6.45 6.40 6.22 6.20 6.21 Er 6.50 6.48 6.44 6.48 6.40 6.41 6.38 6.18 6.16 6.16 Tm 6.56 6.54 6.49 6.53 6.44 6.45 6.41 6.18 6.16 6.15 Yb 6.56 6.53 6.48 6.52 6.42 6.43 6.39 6.14 6.13 6.13 Lu 6.54 6.50 6.45 6.49 6.39 6.40 6.37 6.12 6.10 6.10

PAGE 169

154Appendix D (Continued) Table D.7. (Continued) I (M) 0.0811 0.0811 0.0811 0.0940 0.0930 pH 7.04 7.04 7.04 7.11 7.10 time 15 min 2.5 hrs 19 hrs 15 min 90 min CO2 (%) 29.22 29.22 29.22 29.22 29.22 2 3T[CO] (M) 103 103 103 152 145 Y 4.69 4.62 4.61 4.51 4.48 La 6.60 6.53 6.57 6.58 6.56 Ce 6.43 6.38 6.42 6.35 6.32 Pr 6.58 6.52 6.53 6.46 6.45 Nd 6.51 6.44 6.46 6.37 6.36 Pm Sm 6.33 6.28 6.29 6.19 6.17 Eu 6.22 6.15 6.16 6.06 6.04 Gd 6.08 6.01 6.01 5.92 5.90 Tb 5.90 5.82 5.82 5.71 5.70 Dy 5.70 5.62 5.60 5.49 5.48 Ho 5.54 5.44 5.43 5.31 5.30 Er 5.42 5.33 5.31 5.18 5.17 Tm 5.35 5.26 5.24 5.11 5.10 Yb 5.29 5.20 5.18 5.05 5.04 Lu 5.23 5.14 5.12 4.99 4.98

PAGE 170

155Appendix E: Temperature Dependent Da ta for Amorphous Ferric Hydroxide Table E.1. Distribution coefficient (log iKFe) results from the experiment performed at T = 10.0oC over the pH range 4.7 – 6.9 with an iron concentration of 1.08 0.08 mM. I (M) 0.0143 0.0143 0.0143 0.0143 0.0143 0.0143 0.0143 0.0143 0.0143 0.0143 pH 4.73 4.75 5.35 5.32 5.84 5.78 6.36 6.34 6.89 6.86 time 15 min 90 min 15 min 90 min 15 min 60 min 15 min 60 min 15 min 60 min Y 1.71 1.67 2.04 2.02 2.55 2.52 3.42 3.38 4.57 4.62 La 1.57 1.51 1.88 1.88 2.36 2.35 3.21 3.18 4.39 4.46 Ce 2.03 2.04 2.32 2.34 2.82 2.82 3.70 3.69 4.93 4.98 Pr 2.00 1.98 2.44 2.45 2.99 3.00 3.91 3.90 5.14 5.19 Nd 2.09 2.08 2.53 2.53 3.07 3.08 3.99 3.98 5.21 5.26 Pm Sm 2.24 2.21 2.69 2.69 3.25 3.26 4.18 4.18 5.42 5.47 Eu 2.22 2.19 2.64 2.65 3.20 3.21 4.13 4.12 5.37 5.41 Gd 2.11 2.09 2.48 2.49 3.00 3.01 3.92 3.90 5.15 5.19 Tb 2.12 2.10 2.51 2.50 3.03 3.04 3.96 3.94 5.18 5.22 Dy 2.07 2.04 2.46 2.45 3.00 3.00 3.93 3.91 5.13 5.17 Ho 1.98 1.94 2.37 2.37 2.92 2.92 3.85 3.82 5.05 5.09 Er 1.97 1.94 2.38 2.38 2.93 2.93 3.86 3.83 5.05 5.09 Tm 2.00 2.00 2.44 2.44 3.00 3.00 3.93 3.91 5.12 5.17 Yb 2.06 2.05 2.52 2.52 3.08 3.09 4.03 4.00 5.20 5.26 Lu 2.06 2.03 2.50 2.50 3.06 3.06 3.98 3.96 5.14 5.19

PAGE 171

156Appendix E (Continued) Table E.2. Distribution coefficient (log iKFe) results from the experiment performed at T = 10.0oC over the pH range 5.0 – 7.1 with an iron concentration of 1.08 0.08 mM. I (M) 0.0144 0.0144 0.0144 0.0144 0.0144 0.0144 0.0144 0.0144 0.0144 0.0144 pH 4.98 4.98 5.58 5.55 6.22 6.18 6.50 6.46 7.08 7.04 time 15 min 90 min 15 min 90 min 15 min 60 min 15 min 60 min 15 min 60 min Y 1.80 1.78 2.27 2.18 3.13 3.14 3.69 3.69 4.96 5.00 La 1.74 1.68 2.14 2.05 2.93 2.95 3.49 3.50 4.81 4.88 Ce 2.15 2.12 2.57 2.52 3.41 3.43 4.01 4.02 5.31 5.41 Pr 2.22 2.17 2.72 2.67 3.60 3.64 4.21 4.22 5.49 5.59 Nd 2.31 2.24 2.81 2.76 3.68 3.72 4.29 4.30 5.56 5.66 Pm Sm 2.45 2.39 2.98 2.94 3.87 3.90 4.49 4.50 5.72 5.84 Eu 2.43 2.37 2.93 2.89 3.83 3.86 4.45 4.45 5.71 5.84 Gd 2.30 2.26 2.75 2.70 3.61 3.65 4.23 4.22 5.48 5.58 Tb 2.32 2.27 2.78 2.72 3.66 3.69 4.26 4.26 5.52 5.61 Dy 2.26 2.20 2.73 2.67 3.63 3.65 4.24 4.23 5.50 5.58 Ho 2.12 2.05 2.64 2.57 3.54 3.56 4.14 4.14 5.43 5.50 Er 2.12 2.06 2.64 2.58 3.55 3.57 4.14 4.15 5.42 5.49 Tm 2.20 2.13 2.72 2.66 3.63 3.65 4.23 4.22 5.50 5.56 Yb 2.28 2.21 2.80 2.74 3.71 3.74 4.32 4.31 5.61 5.69 Lu 2.26 2.21 2.78 2.72 3.68 3.70 4.27 4.26 5.52 5.58

PAGE 172

157Appendix E (Continued) Table E.3. Distribution coefficient (log iKFe) results from the experiment performed at T = 39.1oC over the pH range 4.9 – 6.8 with an iron concentration of 0.108 0.008 mM. I (M) 0.0103 0.0103 0.0103 0.0103 0.0103 0.0103 0.0103 0.0103 0.0103 0.0103 pH 4.89 4.88 5.37 5.34 5.83 5.80 6.32 6.30 6.80 6.79 time 15 min 60 min 15 min 60 min 15 min 60 min 15 min 60 min 15 min 60 min Y 2.79 2.62 2.92 2.89 3.41 3.39 4.13 4.10 5.01 5.09 La 2.65 2.55 2.85 2.82 3.24 3.25 3.88 3.86 4.72 4.81 Ce 3.05 2.98 3.26 3.26 3.69 3.69 4.36 4.34 5.23 5.33 Pr 2.98 2.92 3.34 3.34 3.85 3.85 4.55 4.53 5.44 5.53 Nd 3.00 2.96 3.38 3.41 3.91 3.91 4.61 4.59 5.51 5.59 Pm Sm 3.11 3.09 3.53 3.53 4.06 4.07 4.80 4.77 5.70 5.80 Eu 3.13 3.08 3.49 3.50 4.03 4.03 4.76 4.74 5.66 5.76 Gd 3.06 3.01 3.36 3.37 3.84 3.85 4.56 4.53 5.45 5.54 Tb 3.07 2.98 3.38 3.38 3.91 3.91 4.66 4.64 5.56 5.65 Dy 3.03 2.96 3.34 3.34 3.90 3.90 4.66 4.64 5.58 5.65 Ho 3.00 2.90 3.27 3.27 3.83 3.82 4.60 4.58 5.51 5.58 Er 3.01 2.93 3.28 3.28 3.86 3.85 4.64 4.61 5.54 5.61 Tm 3.03 2.95 3.35 3.36 3.95 3.94 4.75 4.73 5.65 5.70 Yb 3.09 3.02 3.43 3.44 4.05 4.04 4.86 4.84 5.76 5.80 Lu 3.07 2.99 3.39 3.40 4.01 4.00 4.81 4.79 5.70 5.74

PAGE 173

158Appendix E (Continued) Table E.4. Distribution coefficient (log iKFe) results from the experiment performed at T = 39.3oC over the pH range 5.3 – 7.1 with an iron concentration of 0.108 0.008 mM. I (M) 0.0109 0.0109 0.0109 0.0109 0.0109 0.0109 0.0109 0.0109 0.0109 0.0109 pH 5.32 5.33 5.70 5.67 6.07 6.05 6.53 6.51 7.08 7.06 time 15 min 60 min 15 min 60 min 15 min 60 min 15 min 60 min 15 min 60 min Y 3.02 2.93 3.34 3.33 3.74 3.68 4.45 4.51 5.64 5.76 La 2.92 2.91 3.22 3.25 3.57 3.51 4.19 4.25 5.34 5.48 Ce 3.33 3.32 3.66 3.68 4.03 3.99 4.69 4.76 5.89 6.07 Pr 3.40 3.40 3.80 3.83 4.20 4.17 4.88 4.94 6.07 6.22 Nd 3.46 3.46 3.86 3.90 4.27 4.23 4.94 5.01 6.13 6.30 Pm Sm 3.61 3.61 4.02 4.05 4.44 4.40 5.13 5.20 6.33 6.53 Eu 3.57 3.57 3.99 4.01 4.39 4.36 5.09 5.16 6.29 6.49 Gd 3.42 3.42 3.80 3.82 4.20 4.15 4.87 4.96 6.08 6.25 Tb 3.46 3.46 3.87 3.89 4.28 4.25 5.00 5.07 6.20 6.36 Dy 3.44 3.43 3.85 3.88 4.28 4.25 5.01 5.08 6.22 6.36 Ho 3.38 3.36 3.78 3.80 4.21 4.17 4.95 5.02 6.16 6.30 Er 3.39 3.37 3.80 3.82 4.25 4.20 4.98 5.05 6.20 6.33 Tm 3.46 3.44 3.89 3.91 4.35 4.31 5.10 5.18 6.32 6.45 Yb 3.53 3.52 3.99 4.02 4.46 4.42 5.23 5.31 6.44 6.58 Lu 3.49 3.49 3.94 3.97 4.40 4.37 5.16 5.23 6.36 6.47

PAGE 174

About the Author Kelly Ann Quinn graduated Summa Cum Laude with a Bachelor of Science degree in Marine Biology and Chemistry from Fairlei gh Dickinson University, NJ in 2000. After working at the Hershey Foods Analytical Department for six months, she entered the M.S. program at the University of South Florid a in the College of Marine Science. Due to the quality and quantity of data collected, she switched to the Ph.D. program in the College of Marine Science in 2005. While pursuing her Ph.D., Kelly was a resear ch assistant, teaching assistant, and AAUS scientific diver. She also received the Von Rosensteil, Riggs, and Gulf Oceanographic Charitable Trust Endowed Fe llowships from the College of Marine Science. The results obtained in her research were published in se veral peer-reviewed journals and presented at the Goldschmidt Geochemistry Conference.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200397Ka 4500
controlfield tag 001 001909637
005 20080428124757.0
007 cr bnu uuuuu
008 070920s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001672
035
(OCoLC)173182081
040
FHM
c FHM
049
FHMM
090
GC11.2 (ONLINE)
1 100
Quinn, Kelly Ann.
0 245
Influence of solution and surface chemistry on yttrium and rare earth element sorption
h [electronic resource] /
by Kelly Ann Quinn.
260
[Tampa, Fla] :
b University of South Florida,
2006.
502
Dissertation (Ph.D.)--University of South Florida, 2006.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 158 pages.
Includes vita.
520
ABSTRACT: The sorption behavior of yttrium and the rare earth elements (YREEs) was investigated using a variety of hydroxide precipitates over a range of solution conditions. Experiments with amorphous hydroxides of Al, Ga, and In were conducted at constant pH (~6.0) and constant ionic strength (I = 0.01 M), while YREE sorption by amorphous ferric hydroxide was examined over a range of ionic strength (0.01 M < I < 0.09 M), pH (3.9 < pH < 7.1), carbonate concentration (0 M < [CO32-]T < 150 micro-M), and temperature (10C < T < 40C). Sorption results were quantified via distribution coefficients, expressed as ratios of YREE concentrations between the solid and the solution, and normalized to concentrations of the sorptive solid substrate. Distribution coefficient patterns for Al, Ga, and In hydroxides were well correlated with the pattern for YREE hydrolysis.In contrast, amorphous ferric hydroxide developed a distinct pattern that was different than those for Al, Ga, and In precipitates but similar to the pattern predicted for natural marine particles. YREE sorption was shown to be strongly dependent on pH and carbonate concentration, significantly dependent on temperature, and weakly dependent on ionic strength. Distribution coefficients for amorphous ferric hydroxide (iKFe) were used to develop a surface complexation model that contained (i) two equilibrium constants for sorption of free YREE ions (M3+) by surface hydroxyl groups, (ii) one equilibrium constant for sorption of YREE carbonate complexes (MCO3+), (iii) solution complexation constants for YREE carbonates and bicarbonates, (iv) a surface protonation constant for amorphous ferric hydroxide, and (v) enthalpies for M3+ sorption. This quantitative model accurately described (i) an increase in iKFe with increasing pH, (ii) an initial increase in iKFe with increasing carbonate concentration due to sorption of MCO3+, in addition to M3+, (iii) a subsequent decrease in iKFe due to increasing YREE complexation by carbonate ions (especially extensive for the heavy REEs), and (iv) an increase in iKFe with increasing temperature.
590
Adviser: Robert H. Byrne, Ph.D.
653
Iron.
pH.
Carbonate.
Ionic strength.
ICP-MS.
690
Dissertations, Academic
z USF
x Marine Science
Doctoral.
773
t USF Electronic Theses and Dissertations.
949
FTS
SFERS
ETD
GC11.2 (ONLINE)
4 856
u http://digital.lib.usf.edu/?e14.1672