Fractionation behavior of the rare earths in natural waters : the role of carbonate complexation and phosphate coprecipitation

Citation
Fractionation behavior of the rare earths in natural waters : the role of carbonate complexation and phosphate coprecipitation

Material Information

Title:
Fractionation behavior of the rare earths in natural waters : the role of carbonate complexation and phosphate coprecipitation
Creator:
Liu, Xuewu
Place of Publication:
Tampa, Florida
Publisher:
University of South Florida
Publication Date:
Language:
English
Physical Description:
x, 129 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Rare earth metals -- Solubility ( lcsh )
Yttrium -- Solubility ( lcsh )
Seawater -- Composition ( lcsh )
Dissertations, Academic -- Marine Science -- Doctoral -- USF ( FTS )

Notes

General Note:
Includes vita. Thesis (Ph. D.)--University of South Florida, 1997. Includes bibliographical references (leaves 111-116).

Record Information

Source Institution:
University of South Florida
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
024649384 ( ALEPH )
39175658 ( OCLC )
F51-00206 ( USFLDC DOI )
f51.206 ( USFLDC Handle )

Postcard Information

Format:
Book

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

FRACTIONATION BEHAVIOR OF THE RARE EARTHS IN NATURAL WATERS: THE ROLE OF CARBONATE COMPLEXATION AND PHOSPHATE COPRECIPITATION by / XUEWU LIU A di s sertation s ubmitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Marine Science University of South Florida December 1997 Major Professor: Robert H Byrne, Ph D.

PAGE 2

Graduate School University of South Florida Tampa, Florida CERTIFICATE OF APPROVAL Ph.D. Dissertation This is to certify that the Ph D. Dissertation of XUEWULIU with a major in Marine Science has been approved by the Examining Conunittee on October 30, 1997 as satisfactory for the dissertation requirement for the Doctor of Philosophy degree Examining Committee: Major Professor: Robert H. Byrne, Ph.D Member: Benjamin P. Flower, Ph.D. Member: Paula G. Coble Ph.D. Member : Luis H Garcia-Rubio Ph.D. Member: EdwardS. Van Vleet, Ph.D

PAGE 3

DEDICATION This dissertation is dedicated to my wife, Bin Peng, and to my parents Yuqin Liu and Fengzhen Song

PAGE 4

ACKNOWLEDGMENTS Foremost I would like to thank my major professo r Dr. Robert H. Byrne, for his guidance and generou s support both academically and financially in each stage of this work I would also like to thank my committee members, Drs. Paula G Coble Benjamin P. Flower, Luis H Garcia-Rubio and EdwardS. Van Vleet for their time, patience, and advice. I am very grateful for all the assistance from my colleagues, Dr. Wensheng Yao, Dr. Jong Hyeon Lee and all faculty staff and students at the Department of Marine Science University of South Florida Special thanks go to Dr. Johan Schijf for his time and as s i s tance in ICP MS analy s es and experimental techniques I would al s o like to thank Ms Joan He s le r for processing all kind s of paperwork and lab orders. I would like to acknowledge my wife Bin for her love and s upport throughout the years while I worked toward this degree. My sincerest thanks are due to my parent s my inlaw s and my uncle This dissertation s tudy was supported by funds from the National Scienc e Foundation.

PAGE 5

LIST OFT ABLES LIST OF FIGURES ABSTRACT TABLE OF CONTENTS 1. INTRODUCTION: AN OVERVIEW OF YITRIUM AND RARE EARTH ELEMENT GEOCHEMISTRY 1.1 General Rare Earth Chemistry 1.2 Distribution of YREEs i n Natural Water s 1 .2. 1 YREE di str ibution patterns in seawater 1.2.2 YREE distribution patterns in riverwater 1.3 YREE Physical Chemistry 1.3.1 Significance of REE studies in geochemistry 1.3.2 REE complexation with inorganic li gan d s 1.3.3 REE complexation by organic ligands 1.3 4 REE solubility controls in seawater 1 .3.5 REE distribution model s 1.4 Goals of This Re searc h 2. COMPARATIVE CARBONATE COMPLEXATION OF YTTRIUM AND GADOLINIUM AT 25 o C AND 0.7 MOLAL DM .3 IONIC STRENGTH 2.1 Abstract 2 2 Introduction 2.3 Methods and Material s 2.4 Results a nd Di sc us sio n 3. COMPREHENSIVE INVESTIGATION OF YITRIUM AND RARE EARTH ELEMENT COMPLEXATION BY CARBONATE IONS USING ICP MASS SPECTROMETRY 3. 1 Abstract 3.2 Introduction iv VI Vlll 1 4 4 6 6 6 7 9 10 11 1 2 1 3 14 15 18 32 33

PAGE 6

3 3 Experimental procedures 3.3.1 Reagents 3 3 2 TBP extraction procedures 3 3.3 Post solvent exchange procedures 3.3.3.1 Extraction of YREEs from TBP 3.3.3.2 Coprecipitation of YREEs by Fe(ill) hydroxide 3 3.4 Measurement of YREE concentrations 3.3.5 Data processing 3.4 Results and Discussion 3.4.1 REE complexation constants 3.4.2 Comparisons with previous direct measurements 3 5 Summary 4 RARE EARTH AND YTIRIUM PHOSPHATE SOLUBILITIES IN AQUEOUS SOLUTION 4. 1 Abstract 4.2 Introduction 4.3 Materials and Methods 4.3 1 Preparation of materials 4.3 2 X-ray diffraction analysis of rare earth phosphates 4.3 3 Solubility analysis 4.4 Solubility Product Calculations 4.5 Results of YREE Solubility Products 4 6 Phosphate Solubility Product and Coprecipitation Behavior 4 7 Coprecipitation Equilibria 4.8 Conclusions 5. THE INFLUENCE OF PHOSPHATE COPRECIPITATION ON RARE EARTH DISTRIBUTIONS IN NATURAL WATERS 5 1 Abstract 5 2 Introduction 5.3 Analytical Procedures 5 .4 Results and Discussion 5.5 Conclusions 11 35 35 35 38 38 38 39 40 41 41 48 50 53 54 56 56 57 57 60 63 75 77 80 81 82 83 84 93

PAGE 7

6. COMPARATIVE COPRECIPITATION OF PHOSPHATE AND ARSENATE WITH YTTRIUM AND THE RARE EARTHS: THE INFLUENCE OF SOLUTION COMPLEXATION 6.1 Abstract 6.2 Introduction 6.3 Experimental Procedures 6.4 Theory 6.5 Results and Discussion 6.6 Conclusions REFERENCES APPENDICES Appendix 3 1 Efficiency of Rare Earths and Yttrium Coprecipitation by Fe 94 95 96 97 99 110 Ill 117 Hydroxide 118 Appendix 3 2 Rare Earth and Yttrium Recovery During Fe Coprecipitation Procedure Using Pr or Nd as Yield Standard Elements 119 Appendix 3.3 Extraction Efficiency of Rare Earths and Yttrium from 1W Appendix 3.4 Experimenta l Results of Rare Earths and Yttrium Complexation with Carbonate Ions 121 Appendix 5.1 Experimental Data for Figure 5.1 124 Appendix 5.2 Experimental Data for Figure 5.2 125 Appendix 6.1 Experimental Data for Figure 6.1 a 126 Appendix 6.2 Experimental Data for Figure 6.1 b 127 Appendix 6.3 Experimental Data for Figure 6.1 c 128 Appendix 6.4 Experimenta l Data for Figure 6 .ld 129 VITA End page Ill

PAGE 8

LIST OF TABLES Table 1.1 Basic Physicochemical Parameters of Yttrium and Rare Earth Elements and Abundance in Composite Shale 2 Table 1.2 Summary of Rare Earth Element Complexation Constants with Ligands 8 Table 2.1 Yttrium and Gd Carbonate Complexation Results (25C and 1=0.7 molal) 19 Table 2.2 Conditional REE Carbonate Stability Constants (25C and 1=0.7 molal) 22 Table 2 3 Conditional REE Carbonate Stability Constant Comparisons 24 Table 2.4 Ligand Concentrations and REE Complexation Constants in Seawater 25 Table 3.1 Summary of Yttrium and Rare Earth Carbonate Stability Constant Determinations 34 Table 3.2a Conditional Carbonate Stability Constants of Yttrium and Rare Earths Determined at 25C and 1=0. 70.02 molal: coJ3; (M) Results 42 Table 3.2b Conditional Carbonate Stability Constants of Yttrium and Rare Earths Determined at 25C and 1=0.70.02 molal: Results 43 Table 3.3 Summary of Conditional Carbonate Stability Constants of Yttrium and Rare Earth Constants (25c and 1=0 70.02 molal) 44 Table 3.4 Yttrium and REE Carbonate Complexation Constants (25 c and Zero Ionic Strength) 46 Table 4.1 MHPO; and MH2Po;+ Formation Constants (25 C, 1=0.1 molal) 62 Table 4.2a Procedure B Data 66 Table 4.2b Procedure C Data 67 Table 4.3 Solubility Products ( log (M)) for Procedure B Precipitates 70 Table 4.4 Solubility Products for Procedure C Precipitates 71 iv

PAGE 9

Table 4.5 Recommended log Values at 25 C and Zer o Ionic Streng th 74 Table 5.1 Fractionation Factors (A.) and Stoichiometric Ratio Estimate s for REE and Yttrium Coprecipitation with Po; 92 Table 6.1 Fractionation Factors and for Yttrium and Rare Earth Coprecipitation 1 03 Table 6.2 Complexation C o nstant s for Yttrium and Rar e Earth with Oxalat e Appropriate to 25 C and Zero Ionic Strength 107 v

PAGE 10

LIST OF FIGURES Figure 1 1 Rare Earth Concentrations in Shale and in Seawater 5 Figure 2.1 Residu a l s Obtained in the Least Squares Data Analys is 20 Figure 2.2 Yttrium speciation in seawater. 28 Figure 3.1 Yttrium and REE Carbonate Complex a tion Constants at 25 C and 0.70.02 molal Ionic Strength 45 Figure 4.1 Previou s REE and Yttrium Phosph a te Solubility Product Determinati o ns 55 Figure 4.2 Char acte ri s tic Spectra of EuP04 Rh a bdophane 58 Figure 4 3 Rare Earth and Yttrium Solubility Product s at 25 C and Zero Ionic Strength (Pr o cedure A ) 65 Figure 4.4 Compari son of Rare Earth and Yttrium Solubility Product s (Procedures A Band C) 69 Figure 4.5 Recommended Solubility Products at 25C and Zero Ionic Strength Based on Procedure B and C Measurements 72 Figure 4.6 Coprecipitation Behavior of Rare Earth and Yttrium Phosphates 78 Figure 5.1 Yttrium a nd Rare Earth Element Fractionation for a 49-hour Pho s ph a te Coprecipitation Experiment at 25"C 85 Figure 5.2 Yttrium and Rare Earth Element Fractionation for a 24-hour Coprecipitation Experiment at pH= 3 8 86 Figure 5.3 Ytt r ium and Rare Earth Element Speciation in Experimental Solution s a t pH 4.4 and pH 3 8 (25 C 1=0.01 molal) 88 Figure 6 1 Selected Rare Earth Concentrations (La Sm, Lu) as a Function of Time 100 Figure 6 2 Fractionation Patterns Observed in the Coprecipitation Processes 101 vi

PAGE 11

Figure 6 3 Fractionation Factors and from Table 6.1 104 Figure 6.4 Comparison of Phosphate and Arsenate Coprecipitation Pattern s ( A,ii (Phosphate)/ A,ii (Arsenate) 1 06 Figure 6.5 Effect of Solution Complexation on Yttrium Rare Earth Fractionation Factors 109 vii

PAGE 12

FRACTIONATION BEHAVIOR OF THE RARE EARTHS IN NATURAL WATERS: THE ROLE OF CARBONATE COMPLEXATION AND PHOSPHATE COPRECIPITATION by XUEWU LIU An Abstract Of a dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Marine Science University of South Florida December 1997 Major Professor : Robert H. Byrne, Ph.D. viii

PAGE 13

The fractionation behaviors of yttrium and the rare earth elements (YREEs) have been investigated through examination of two important aspects of rare earth geochemi s try: solution complexation and removal from solution by precipitation/coprecipitation. ( 1) YREE complexation with carbonate has been examined through s olvent exchange analysi s u s ing both radiochemical and ICP mass spectrometric procedure s at 0 7 molal ionic s trength and 2s c. ( 2) YREE pho s phate s olubility behavior has been asse ssed in dilute HC104 solution and YREE phosphate coprecipitation behavior has been characterized in the presence and absence of solution complexation. Carbonate stability constants of the r are earth elements show a general increase in the extent of REE complexation with increasing atomic number. Stability constant s determined by ICP-MS analysis are in good agreement with previous direct measurements of s tability constants for the ions y 3+, Ce3+, Eu3+, Gd3+, Tb3+ and Yb3+. My work present s the fir s t direct measurements for carbonate stability constant s of the ions La3+, Pr3+, Nd3+, Sm3 + D y3+, Ho3+, Er3 + Tm3+, and Lu3+. Studies of yttrium carbonate complexation by ICP-MS and radiochemical solvent exchange techniques indicate that the solution complexation of y3+ in seawater most resembles that of Eu 3 + and Tb3+. YREE phosphate solubility products determined using different procedures indicate that phosphate phase s can exert significant controls on REE fractionation patterns in natural waters. Solubility products are strongly dependent on the conditions of solid phas e formation. Fre s h precipitates are much more soluble than slowly formed, well aged, coarse precipitates. The pattern of YREE pho s phate solubility product s i s generally similar to YREE fractionation p a ttern s developed during pho s phate coprecipitation. Phosphate coprecipitation creates solution concentration p a tterns characteristic of input-normalized YREE concentrations in s eawater. Removal of REEs from solution by pho s phate coprec i pitation produces a variety of near-neighbor concentration anomalies (e.g. Eu-Gd-Tb, Dy-Er-Yb and Er-Yb-Lu) which have been reported in high precision analyses of open ocean waters. Nearest neighbor concentration anomalies are very s imilar ix

PAGE 14

for phosphate and arsenate coprecipitation both in the presence and in the absence of solution complexation. Solution complexation influences YREE coprecipitation patterns by strongly enhancing the retention of Yttrium and heavy REEs in solution. Abstract Approved: ______________ Major Professor: Robert H. Byrne, Ph.D. Professor Department of Marine Science Date Approved: ___ __ X

PAGE 15

1. IN'IRODUCTION: AN OVERVIEW OF YT'IRIUM AND RARE EARTH ELEMENT GEOCHEMIS'IRY 1. 1 General Rare Earth Chemistry Rare earth element s (REEs) include the element lanthanum (La) and fourteen elements following La, which are placed separately at the bottom of the Periodi c Table. These fourteen elements following La (Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb and Lu), are commonly called lanthanides and when grouped with La are tenned the rare earth elements (REE). Yttrium (Y), the element above La in Group ill of the Periodic Table, is u s ually con s idered together with the REEs because of its similar ioni c radius and chemical properties. A variety of Y and REE chemical and geochemical characteri s tic s are summarized in table 1 .1. REEs are the products of nucleosynthesis during stellar evolution As a result of th e dependence of nuclide stability on proton number and neutron number, REE concentrations in sedimentary rocks show two distinctive features (Fig. 1 1 a): (I) elements with even atomic numbers are more abundant than those with odd atomic numbers (Oddo-Harkins rule), and (2) ther e i s a systematic decrease in abundance with increasing atomic number. Pm, an odd atomic number REE, is the ultimate example of Oddo-Harkins ruJe. It is esse nti a lly absent from the earth's crust. One important goal of REE research is to investigate the mechanisms of REE fractionations (varia tion s of relative conc entrat ion s) produced in geochemical processes.

PAGE 16

Table 1.1 z Symbol 57 La 58 Ce 59 Pr 60 Nd 61 Pm 6 2 Sm 6 3 Eu 64 Gd 65 Tb 66 Dy 67 Ho 68 Er 69 Tm 70 Yb 71 Lu 39 y Basic Physicochemical Parameter s of Yttrium and Rare Earth Elements and Abundance in Composite Shale Name Electron Ionic Radiia Shale M3+ 0 (A) (!lmollkg) 4f 5s 5p CN=8 CN=9 NASCb PAAS d ---------------------------------lanthanum 2 6 1.160 1.216 295 274 cerium 1 2 6 1.143 1 196 592 571 praseodymium 2 2 6 1 126 1 .179 71.7 63 3 neodymium 3 2 6 1 109 1.163 263 222 promethium 4 2 6 1.093 1 144 s amarium 5 2 6 1 .079 1.132 49.9 37 .2 europium 6 2 6 1.066 1.120 10.6 7.24 g adolinium 7 2 6 1.053 1.107 40.4 29.9 terbium 8 2 6 1.040 1.095 7.74 4 85 dy s pro s ium 9 2 6 1.027 1.083 33.8 27.1 holmium 10 2 6 1.015 1 .072 8 12 6.06 erbium 11 2 6 1.004 1.062 22.4 1 7.3 thu l ium 12 2 6 0 .994 1.052 3.73 2.37 ytterbium 13 2 6 0.985 1.042 20.4 16.2 lutetium 14 2 6 0 .977 1.032 3.49 2.46 yttrium 3d104s24p6 1.019 1.075 404c ashannon (1976). b North American Shale Composite Piper (1974) C Haskin and H as kin (1966) d Post Archean Au s tralian Shale, Taylor and McLennan (1985) 2

PAGE 17

REE fractionations are conveniently examined by comparing REE concentrations in natural environmental samples with those in source materials This comparison, expressed as (sample REE concentration)( source REE concentration)"1 provides input normalized REE concentrations which change in response to the variable reactivities of the REEs in natural waters Most studies of REE fractionations in rocks and minerals normalize REE concentrations to chondrite values (Henderson 1984), while fractionations in natural waters are often quantified through a normalization in which shale concentrations are used to represent the average REE concentrations in the earth's crust (Taylor and McLennan, 1985). Normalization of REEs to shale concentrations accomplishes two things First it removes the naturally occurring saw-tooth distribution in the absolute REE abundances. The absolute REE abundance for a representative (fig 1.1 b) seawater sample (Zhang et al. 1996) has the classic odd-even pattern found in all natural materials. Normalization to North American Shale Composite (N ASC, fig. 1.1 c) or Post Archean Australian Shale (PAAS fig. 1.1d) results in a typical REE fractionation pattern for seawater, including a large Ce depletion and a heavy REE enrichment. Second, since shale normalized concentration patterns represent abundances of a sample relative to continental crust shale patterns serve as process signatures. Any deviation from a flat pattern (identical shale normalized concentrations) is then attributed to chemical fractionation processes in the natural water system. The YREE abundances of two representative shales (NASC and PAAS) are presented in table 1.1. Use of different shales produces slightly different fractionation assessments (Sholkovitz, 1988). 3

PAGE 18

1. 2 Distribution of YREEs in Natural Waters 1. 2. 1 YREE distribution patterns in seawater With the exception of Ce, REEs in seawater generally show increasing concentrations with depth, and deep water concentrations in the Pacific are higher than those in the Atlantic and Indian Oceans. Such profiles suggest that the REEs are scavenged by adsorption onto biogenic particles at shallow depth and are released back into deep waters by redissolution as the particles settle through the water column and partially decompo s e Since deep Pacific waters are older than those of other oceans this process also explains the enrichment of REEs in the deep Pacific relative to the Atlantic. Shale-normalized REE concentrations such as those shown in Figs. 1.1 c and 1.1 d increase substantial l y with increasing atomic number. This general enrichment from La to Lu has been the subject of many investigations. In addition to the general enrichment across the REE s eries, several interesting and important features are often observed in shale normalized REE patterns including obvious negative Ce anomalies (except in anoxic basins), positive Eu anomalies close to hydrothermal vents and positive Gd anomalies in normal seawater. Whereas the Ce and Eu anomalies can be explained by Ce(III)/Ce(IV) and Eu(II)/Eu(III) redox reactions, explanations for Gd anomalies have been based solely on the interplay of solution chemistry( complexation in so luti on) and interactions with particle surface s (Byrne and Kim 1993). Despite the c l ose similarities in their ionic radii Y and Ho are fractionated in seawater and it has been s uggested that the fractionation of Y and Ho in s eawater is due to the differences in their solution complexation or surface reactions (Bau et al., 1996) Conclusive fractionation mechanisms have not been formulated due to lack of relevant thermodynamic data. 4

PAGE 19

b 10' 10" lACe Pr Nd Pm SmEu Gd lb Dy H o Er Tm Yb Lu La Ce Pr Nd PmSm Eu Gd 1b Dy H o Er Tm Yb Lu d IO' IO "' u < < "' < i:l w Ll Ce Pr NdPm Sm E u Gd 1b Dy H o F..t TmYb lu La Ce Pr Nd PmSm Eu Gd lb Dy H o Et T m Yb Lu REE REE Figure 1.1 Rare Earth Concentrations in Shale and in Seawater. (a) Representative rare earth concentrations in shale: PAAS and NASC. ( b) Rare earth concentrations in seawater at different depths (Station SA 7, 14 15.38 S, 154 0.06'E, from Zhang et al. 1996) (c) P AAS normalized rare earth concentrations in seawater. (d) NASC normalized rare earth concentrations in seawater. 5

PAGE 20

1. 2. 2 YREE distribution patterns in riverwater Compared to oceanic processes which influence REE distributions REE chemistries in riverine and estuarine waters are more complex. Rivers and estuaries are highly dynamic systems. Rivers drain through different geological terranes and exhibit l arge ranges in discharge rate, concentration of s u spe nded and colloidal particles and overall chemical composition. Estuaries display large fluctuations in salinity, pH and particle load Relatively speaking, oceanic waters have a much smaller range of composition Major conclusions from studies of rivers and estuaries (Sholkovitz, 1992) are that chemical weathering reactions on the continents leads to progressive differences in the relative abundances of REEs in continental rocks, s u spended riverine particles and solution ( dissolved REEs). Solution and surface chemistry play a m ajor role in establishing the REE composition of freshwater. Both the concentrations and the extent of fractionation of di sso lved REEs in river waters are dependent on pH and the presence of colloidal particles High -pH waters have the lowest REE concentrations and the most fractionated composition A large proportion of the "dissolved" REE inventory in river waters is associated with colloids which undergo large scale coagulation in estuaries This process results in highly fractionated solution phases as REEs are transported via rivers from the continents to the oceans. Therefore, fractionat i on of REEs in oceanic water should be viewed as a continuation of processes which occur in rivers and estuaries. 1. 3 YREE Physical Chemistry 1. 3. 1 Significance of REE studies in geochemistry Due to their coherent group chemistry REEs are considered ideally suitable for studying solution/surface partitioning processes which influence the relative concentrations 6

PAGE 21

of metals in natural water. The coherence of REE chemical behavior stems mainly from similar but monotonically decreasing ionic radii (Table 1 .1) of trivalent REEs between La and Lu. The comparative simplicity of REE chemistry and the existence of multiple Ce and Eu redox states allow REEs to be utilized as probes in many fundamental reactions and processes which act on trace elements in natural waters (Elderfield, 1988, Byrne and Sholkovitz 1996). These fundamental reactions and processes include adsorption to suspended and sinking particles and complexation by inorganic and organic ligands m solution and oxidative processes mediated by microorganisms (Moffett, 1990). The interplay of surface and solution chemistry is a fundamental aspect of trace element cycling in the oceans, and REE investigations provide an insightful means of examining the general nature of this interplay 1. 3. 2 REE complexation with inorganic ligands Natural water contains a complex mixture of anions capable of complexing the REEs. Potentially important inorganic REE complexes include those with co ; so;-, F and CI". Recommended REE stability constants from previous assessments are summarized in table 1.2 Most of the data in table 1 2 are based on LFER as explained in Lee and Byrne (1992, 1993a) REE speciation in seawater was first assessed by Turner et al. ( 1981) and their results were the first to indicate the importance of carbonate complexation in the marine chemistry of the REEs. The REE stability constants in Table 1.2 are based on the experimental data of Cantrell and Byrne (1987a and b) and Lee and Byrne (1993a) and linear free energy relationship assessments of Lee and Byrne (1993a). REE speciation in seawater is strongly dependent on pH The data in table 1.2 indicate that more than 90 % of light REEs and 99 9 % of heavy REEs are complexed by carbonate ions in seawater at pH 8. 7

PAGE 22

Table 1.2 Summary of Rare Earth Element Complexation Constants with Ligands in Seawater REE La 8.06 4.57 7.92 1.95 -9.20 2 57 Ce 8.45 4 85 8.54 1.92 -8.95 2 .76 Pr 8.70 5 .01 8 82 1.95 -8 .81 2.88 Nd 8.88 5.12 9.00 1.97 -8 .70 2 97 Pm Sm 9 .21 5.33 9.28 2.00 -8 50 3.17 Eu 9.32 5.40 9 .31 2 .00 -8.44 3 .24 Gd 9.30 5 27 9 26 1.99 -8.47 3 .21 Tb 9.49 5.38 9.44 1.97 -8.40 3 32 Dy 9.62 5.45 9 57 1.95 -8.35 3 29 Ho 9.69 5.49 9 62 1.92 -8 32 3 27 Er 9.79 5.56 9.80 1.92 -8.27 3 26 Tm 9 92 5.64 9 94 1.92 -8.22 3 26 Yb 10. 04 5.73 10.12 1.91 8 .14 3.30 Lu 10.09 5.74 10.19 1.85 -8 .13 3.29 compiled from Byrne and Sholkovitz (1996). 0.41 0.42 -0.44 -0.48 -0 52 -0 55 -0.57 -0 57 0 59 -0.59 -0 62 -0 63 -0 63 0 66 Note : the constants for carbonate comple x ation are expressed in terms of total carbonate ion concentrations; other constants are expressed in terms of free ion concentrations 8

PAGE 23

The recommended constants for REE complexation with phosphate in table 1.2 are based on the work of Byrne et al. ( 1991 ) and Lee and Byrne ( 1992) The concentration of free phosphate in seawater is generally much smaller than 1x 109 mol kg-1 REE phosphate complexation constants shown in table 1.2 indicate that REE phosphate complexes are relatively minor species in seawater. However in groundwater and riverwater, the complexation by phosphate ions could become important (Lee and Byrne 1992 and 1993b) due to decreased carbonate complexation Based on the data in Table 1 2, REE complexation with sulfate hydroxide, fluoride and chloride is of marginal significance compared to carbonate complexation in normal seawater. REE sulfate concentrations are about the same as concentrations of free M3+ Fluoride and chloride species are 10 to 20% of free M3+ Using a reasonable upper-bound pH for seawater at 25C (pH 8 35), the hydrolysis constant estimate for lutetium indicates that (Byrne and Sholkovitz, 1996) The concentrations of hydrolyzed REE species in seawater are comparable to the concentrations of free REE ions only in relatively warm (e.g. t > 20 C) surface water. 1. 3. 3 REE complexation by organic ligands Even though many trace metals (Cu2+, Zn2+, Fe3+) are heavily influenced by the presence of organic ligands such as fulvic acid, amino acids, etc., it has not been shown that organic ligands in seawater appreciably complex REEs in solution Instead organic ligands are expected to strongly influence the marine behavior of the REEs through the role of organics in REE surface complexation (Byrne and Sholkovitz, 1996; Byrne and Kim 1993) Surface characteristics of particles in natural waters (e.g. functional group chemistry and surface charge) may be determined by adsorbed organics (Davis, 1984; Hunter and Liss 1979) REE uptake behaviors of organisms such as macroalgae (Stanley and Byrne 1990) and diatoms (Bing1er et al. 1989) have shown that organic substrates can produce rare earth abundance patterns similar to shale-normalized REE abundance patterns m 9

PAGE 24

seawater, and adsorptive behaviors can be strongly influenced by solution complexation in seawater. Byrne and Kim (1990) observed that the affinity of REEs for surfaces can be strongly affected by thin organic coatings 1. 3. 4 REE solubility controls in seawater Jonassen et al. (1985) and Choppin (1989) suggested that equilibration of REEs with insoluble salts could be an important aspect of REE marine geochemistry Choppin ( 1986 and 1989) proposed that formation of REE carbonates could set the upper limit of the REE concentrations in seawater. Previous determinations of REE carbonate solubility products suggests that MiC03 ) 3 solubility products are such that log (M) -3 5. 5. Byrne and Sholkovitz ( 1996) concluded that solubility products of this magnitude would support free REE ion concentrations in seawater which are three or more orders of magnitude larger than those observed. Consequently, precipitation of carbonates is not expected in seawater. Previous investigations (Jonassen et al, 1985; Firshing and Brune, 1991; Byrne and Kim, 1993) have shown that REE phosphate solubility products are very low (log about -25 to -23) and phosphate precipitation has been suggested (Jonassen et al., 1985; Byrne and Kim, 1993) as a significant control on REE concentrations in seawater and other natural waters The analysis of Byrne and Kim (1993) indicated that the activity products of REE ions and phosphate ions in seawater at depths below about 300 m are approximately constant and are equal to or greater than previously reported solubility products. Byrne and Kim (1993) noted that REE phosphate precipitation should not involve the formation of pure phases but, rather, the formation of coprecipitates. REE phosphate coprecipitation occurs in a manner which enriches solutions with heavy REEs relative to light REEs, consistent with observed shale-normalized enrichments of the heavy REEs in seawater. 10

PAGE 25

1. 3. 5 REE distribution models REE fractionation modeling of natural waters involves differentiation of the roles of two important processes : solution complexation, and surface complexation by marine particles. Solution complexation has been investigated by many researchers while few works have been directed to quantification of surface complexation behavior. Surface complexation behavior has been modeled by assuming that REE surface complexation exhibits (a) characteristics appropriate to oxide surfaces and thereby metal ion hydrolysis constant behavior ( Elderfield 1988; Erel and Stolper, 1993) or (b) characteristic s appropriate to the behavior of organic ligands (Elderfield, 1988; Byrne and Kim, 1990). Predicted variations in shale normalized concentrations based on different assumptions are similar in producing heavy REE enrichments. Quantitative assessments of REE distributions (Elderfield 1988; Byrne and Kim, 1990, Piepgras and Jacobsen, 1992; Erel and Stolper 1993) have been based on directly measured carbonate stability constants for a few elements and quadratic or LFER estimations for constants not directly measured (Cantrell and Byrne 1987 Lee and Byrne, 1993a). Without direct comparative carbonate complexation measurements for all REE s, considerable uncertainties remain in the models of REE solution complexation. Even though simple solution complexation/surface complexation models are capable of explaining the general enrichment of heavy REEs, it has proven difficult to construct models which closely reproduce all observed distributional features of the YREEs (Byrne and Sholkovitz, 1996). REE fractionation models based on competitive solution/surface comp l exation reasonably depict he avy-REB enrichment in seawater but underestimate the fractionation of La relative to its heavier neighbors. Byrne and Kim ( 1993) noted that REE phosphate coprecipitation also promotes heavy-REB enrichments in solution. Although the importance of REE phosphate precipitation in seawater has not been directly demonstrated, suc h contro l s are consistent with observed total REE concentrations in some environments (Byrne and Kim 1993 ; Johannessen et al., 1995) and are consistent with the general heavy vs light REE fractionation in the oceans. 11

PAGE 26

1. 4 Goals of This Research The objective of this dissertation is to systematically examine two important aspects of REE and yttrium geochemistry: ( 1) inorganic complexation in aqueous solution and (2) YREE removal from solution as insoluble phosphate salts Since it has been proposed that REE geochemistry involves an interplay of these phenomena a detailed quantitative description of solution complexation behavior and solubility behavior is required to further elucidate the fractionation of REEs in natural waters. Previous experimental investigations of REE behaviors involved only a few REEs. This work involves investigation of the solution complexation behavior and solubility (coprecipitation behavior) of yttrium and all the rare earths. In the following chapter (chapter 2), comparative Y and REE complexation beh a vior is investigated by solvent exchange and radioisotope counting techniques. Our results show that even though linear free energy relationships provide good estimates for the general trend of REE complexation across the series, direct experimental determination is essential to unambiguou sly determine the position of yttrium complexation constants in the REE series. In Chapter 3 carbonate complexation of yttrium and all REEs is systematically determined by ICP mass spectrometric techniques This work presents the first direct measurements for the complexation of La3+, P r+, Nd3+, Sm3+, Dy3+, Ho3+ Er3+, Tm3+, Lu3+ by carbonate ion s. Chapter 4 examines the solubility behavior of individual YREE phosphates and shows that REE phosphate solubilities are low enough to set upper limits for YREE concentrations in seawater. Chapter 5 discusses the implications of YREE phosphate coprecipitation in natural waters Chapter 6 then discusses the influence of so lution complexation on coprecipitation behavior. My results demonstrate that YREE coprecipitation in the oceans would support the general YREE pattern observed in seawater (including light REE behavior) and that solution complexation exerts a strong influence on fractionation patterns developed during coprecipitation processes. 12

PAGE 27

2. COMPARATIVE CARBONATE COMPLEXATION OF YT1RIUM AND GADOLINIUM AT 25C AND 0.7 MOLAL DM-3 IONIC S1RENGTH 2.1 Abstract Stability constants for yttrium and gadolinium complexation by carbonate ions have been measured by so lvent extraction procedures at 25C and 0.7 mol dm-3 ionic strength ' The results of six experiments indicate that 1 (Y)=5 71, ' 1 (Gd)=5.65 and log 2 (Gd)= 10.12 where coinCM) = [C03Fr, represents concentrations and When compared with previous complexation results for Eu3+, Gd3+ and Tb3+ obtained using identical procedures, excellent agreement is observed for Gd 3 + complexation, and the solution complexation of y3+ in seawater is seen to closely resemble that of Tb3+. In spite of closely linked solution chemistries of y3+ and Tb3+, the shale normalized concentrations of these elements in seawater should be distinct due to their differing reactivities with organic ligands on particle surfaces. Model calculations indicate that while shale normalized y3+ concentrations in seawater may exceed the shale normalized concentrations of all rare earths, similarities in Y 3 + and Tb3+ solution chemistries may produce substantial coherence in the oceanic distributions of these elements. 13

PAGE 28

2. 2 Introduction The solut ion behaviors of the trivalent rare earths are distinguished principally on the basis of ionic radius, which changes, in small increments, by less than 20% across the fifteen member series (Shannon, 1976). It i s generally expected that the solution chemistries of both the rare earths and yttrium are, as a first approximation, those of h ard charged spheres (Moeller et al, 1965; Wood, 1990) In this case, the chemical and geochemical b e haviors of yttrium would be expected to closely follow the behavior of Ho, which ha s an ionic radius most s imilar to that of yttrium. In contrast to this expectation, the po s ition of Y(III) among the trivalent lanthanide s i s a functi o n not on ly of Y and REE comparative ionic radii but a l so of ligand identity (Moeller, 196 3; 1972 ; Moeller e t al, 1965). It has been suggested ( Borkow ski an d Siekierski 1992 ) that the variable comparative complexation properties of Y and the rare earths are caused by a ligand dependent covalency of REE comp l exa tion (through f orbital delocalization) which is altogether a b se nt in yttrium. In this case it i s predicted that for hard ligands the comparative complexation behavior of Y and the rare earths will approach that expec ted entirely based on e lec tro s tatics, whereby Y assumes a position among th e rare earths based on compara t ive io nic radii In the case of complexation wi th soft li ga nd s, so me extent of charge delocalization is possible with the rare earths and enhanced REE bond strength will result in li g ht er rare earths and yttrium having comparable complexation behaviors. Thus, to the extent that yttrium and th e rare earths are e lectro s tatically simi lar but substantially different in their potential for covalent bonding the comparative chemistries of yttri um and the rare eart h s serve as a useful index of REE-ligand bond covalency. This attribute of comparative Y and REE chemistries i s of special intere s t in natural systems where Y and REE solution chemistry is dominated by hard ion s (CO;-, ro!-) and electrostatic b ehav ior while REE s urfac e chemistry may be dominantly co ntrolled by organic coatings, with 14

PAGE 29

largely unknown functional group chemistries, and some degree of covalency. In order to establish a reference point for assessment of comparative YIREE behavior in seawater we examine, in this work, the comparative complexation properties of yttrium and REEs with co;-, the anion which is expected to dominate YIREE solution complexation in seawater. Our work is the first to directly examine the comparative Mco; and M(C03); formation constant behavior of yttrium and rare earths Th is work performed at the ionic strength of seawater, allows comparative Y and REE oceanic distributions to be viewed in the context of comparative Y and REE sol ution chemistries. 2. 3 Methods and Materials Toward the goal of a direct comparison of our yttrium carbonate comp le xation results with rare earth carbonate complexation results obtained in our previous work (Lee and Byrne, 1993a), the procedures used in our work are essentially identical to those used by Lee and Byrne (1993a). As a further means of relating the present study to previous REE stability constant investigations (e.g. Lee and Byrne, 1993a), it is useful to conduct experiments which have one or more elements in common with previous works. Toward this end, our yttrium complexation investigations included gadolinium in each experimental solution. Our measurements examined the distribution of yss and Gd 153 (carrier free, Isotope Products Laboratories) between an organic phase, tributyl phosphate (TBP), and an aqueous phase (0.70.02 mol dm3 ionic strength) composed principally of NaCl04. Isotopic di s tributions between the phases were examined as a function of aqueous carbonate and bicarbonate ion concentrations. Preparation and purification of our TBP and perchloride solutions is described in detail by Cantrell and Byrne (l987a). Equilibrium of th e two phas es (2 5.0.1 C) was maintained by stirring and vigorous bubbling with N2 15

PAGE 30

(99.998%, Air Products) and C02-N2 mixtures (29.7% C02, Air Products). The equilibration (bubbling period) in our experiments was 20 min., and an additional 20 min. period was allowed for phase separation. Our bubbling procedure provides a finely dispersed mixture of phases. No differences in distribution coefficients have been observed in our work for equilibration periods between 20 minutes and 40 minutes. After measuring the equilibrium pH of the solution phase, replicate 1 cm3 samples of each phase were taken for radiochemical analysis. Measurements of pH on the free hydrogen ion concentration scale (McBryde, 1969, 1971; Byrne and Kester, 1978) were obtained using a Ross Combination electrode (Orion No. 810200) and a Coming Model 130 pH meter. As in previous works (Cantrell and Byrne, 1987 a, b; Lee and Byrne, 1993a), the electrode was calibrated by titrating the unbuffered working solution with strong acid (HC104 ) Three mol dm 3 NaCl was used as our electrode filling solution to prevent KCl precipitation at the electrode liquid junction. Distrihution coefficients in the absence of complexing ligands (DO) and m the presence of co;ions (D) were determined as (1) where [M3+h and [M3'1 indicate total and free (uncomplexed) y3+ and Gd3+ concentrations in the aqueous phase, and represents the total concentration of each isotope in the TBP phase. Distribution coefficients were determined by monitoring, for each isotope, counts above background in 1 cm3 samples of each phase maintained at constant counting geometry. Counting times were sufficient to provide at least 10,000 counts above background for each radioisotope. Our Gamma spectroscopic analyses were obtained using a Canberra Ge(Li) detector (series number 1861) with a Canberra model 90 multichannel analyzer. 16

PAGE 31

Carbonate and bicarbonate ion concentrations in the aqueous medium were adjusted with a titrant solution con sis ting of 0.68 mol dm-3 NaHC03 plus 0.68 mol dm 3 NaCl04 in deionized-di sti lled water. Since the ionic strength of the titrant solution was 1.36 mol dm-3 and the perchloride concentration was 0.68 mol dm-3, titrant additions caused the ionic strength of the solution to vary somewhat while the perchloride ion concentration remained constant. The work of Lundqvist (1982) indicated that at constant perchloride concentration the parameter no (eqn. (1)) is constant. Nonlinear least-squares methods (Byrne and Kester, 1978; and Byrne, 1987a ) were used to estimate Y and Gd carbonate and bicarbonate complexation constants Elemental distribution data were described using the eqn. [)0 D (2) 2 20 where [C03 h=[C03 ]+[NaCOJ, and brackets ([] and [ h) represent concentrations in mol dm3 Carbonate and bicarbonate ion concentrations were determined using the eqns ( Cantrell and Byrne 1987a) (3) where Pco2 is the C02(g) partial is the free hydrogen ion concentration, Ko is the Henry's law constant, and K '1 and are the first and second conditional di ssoc iation constants of carbonic acid a ppropriate to 0.68 mol dm3 NaCl04. The values logK0K.1=-7 .56 and determined previously in Cantrell and Byrne (1987a) were u sed in this work. The parameters co3.B; (M), (M) and Hco3.B; (M) in eqn (2) are cond itional Y and Gd stability constants defined as follows: 17

PAGE 32

and (4) Use of eqn. (2) in nonlinear least-squares analysis involves minimization of the residual sum of squares (RSS) function : RSS= (5) Eqn. (5) provides constancy of weighting to our experimental distribution data while varies by more than three orders of magnitude. 2. 4 Results and Discussion The conditional Y and Gd carbonate and bicarbonate stability constants obtained in this work (25C and 1=0.70.02) are shown in Table 2.1, and residual plots obtained in our six data fits using eqn. (5) are shown in Fig. 2.1. The error limits provided for each experiment in Table 2 1 are based on the quality of each experimental data fit. The error limits provided with the average stability constants are equal to one half the range of the averaged stability constant data Within experimental error, our average c03,B; (Gd) and results agree with the previous results of Lee and Byrne (1993a). The 18

PAGE 33

Tab l e 2.1 Y and Gd Carbonate Complexation Results(25 C and !=0. 7 molal) ( 4.21.38)x 1 os ( 1.82.10)x 1010 50 (6.14 27)x105 (2.58.08)x 1 Ol o 6 (5.59.31)x 10 5 (2.26.06)x10l0 28 (5.44.71)x105 (1.15.09)x10l0 11 ( 4 .8 5 68)x 1 os (2.48.14)x1QIO 40 ( 4.81 55)x 1 os (2 77.11)x10l0 40 Average (5.17.68)x105 (2.18.60)x JOIO 31 (2.90.42)x 1 os ( 1.16.09)x 1010 86 (4.62.48)x105 (1.66.11)x10l0 32 (5.05.23)x105 ( 1.18.05)xi010 45 (4.76. 20)x105 (0.66 0.02)x10l0 22 ( 4.91.33)x 1 os (1.50.06)xiOIO 43 ( 4.71.58)x 1 os (1.71.10)xiOIO 49 Average (4.49.79)x105 (1.31.40)x10i0 46 19

PAGE 34

Figure 2 1 0 100 Y 1 0 Y1 0 Y l :c X Y4 -1-YS .. "' Y6 X + X u 0 .050 :c .. "' ... "' jl., X 0 -c o-P + .. .ll ; 0.000 + u ., = 0 .. 0 "' o ll = X -() .050 -;;; X .. ::r: + -() 100 -5.5 5 -4.5 3.5 log[C01 1 ] T 0.100 .----.---....,----.-----.----....----.----, Gdl Gdl 0 Cdl X CcU + Cc5 "' Cdii 0 "' X A a 4> + + o + 0 0 X 0 f + X A A 0 -()_ 100 ___ ........ ____ ......... ..J) 5 log[C03:t-h Resid u als Obtained in the Least Squares Data Analysis. ( A) Yttrium results (B ) Gadolinium results. 20

PAGE 35

log c03f3; (Gd) result obtained in the present work is 0.03 log units smaller than the log co3f3; (Gd) result of Lee and Byrne (1993a), and our log (Gd) re s ult is 0.03 log unit s larger than the result of Lee and Byrne (1993a) The bicarbonate complexation results obtained in the present work are poorly defined but are comparabl e in magnitude to the results of Lee and Byrne ( 1993). Bas e d on the excellent agreement between our Gd complexation results and those of Lee and B y rne (1993a), the Eu, Gd, and Tb results obtained by Lee a nd Byrne (1993a) can reasonably be integrated with the Gd andY r es ult s obtained in the pre se nt study Table 2.2 provid es a direct compari so n of carbonate and bicarbonate complex a tion results which have been directly obtained for Eu Gd Tb andY, and estimated (Lee and Byrne 1993a) for Dy and Ho The Gd results s hown in Table 2.2 combine the re s ults s hown in Table 2.1 with the six experiments of Lee and Byrne ( 1993a) which were conducted using procedures id en tical to tho se employed in the present study. The Eu results g iven in Table 2.2 repr esent the ave rage of six experiments (Lee and Byrne 1993a) and the Tb res ult s rep rese nt th e average of thr ee experiments (Lee and Byrne 1993a). The Dy and Ho results shown in Table 2.2 are b ase d on our Table 2.2 Gd result s and the linear free energy analysis of Lee and Byrne (1993a). The linear free energy analysis of Lee and Byrne (1993a) indicated that 1(Gd)=0.18, = 0.36. The sta bility constant re s ult s s hown in Table 2.2 are ge nerally in very good agreement with the determinations of others. In Table 2.3, the Eu and Y complexation results shown in Table 2 2 are compared with formation constant result s obtained in a variety of studies under comparable experimental conditions. With the except i on of the works of Spahiu ( 1985 ), all of the s tudies s hown in Table 2.3 were obtained at very low mol dm-3) m etal concentrations. The Y concentrations in the 2 1

PAGE 36

Table 2.2 Conditional REE Carbonate Stability Constant s (25C and 1 = 0.7 molal) Element References Eu 5.81 10.14 Lee and Byrne 1993a Gd 5 67 10.10 Lee and Byrne 1993a; this work Tb 5 .7 9 10.26 Lee and Byrne 1993a Dy 5.85 10.41 Lee and Byrne, 1992 ; this work Ho 5.88 10.46 Lee and Byrne 1992; this work y 5.71 10.34 this work The r es ults s hown for all rare earths, with the exception of Gd are in general agreement with the estimates of Millero (1992). The log 0(Gd) r es ult s of Millero (1992) are based on indirect estimates (Cantrell and Byrne 1987a) rather than direct measurement s, and in a comparative sense (re l ative to Eu and Tb) are larger than those given in this table 22

PAGE 37

work of Spahiu (1985) ranged between 0 05 and 1 mol dm-3. There are no measurements available for comparison with our estimates of Gd Tb, Dy and Ho carbonate complexation constants. The complexation results shown in Table 2.2 indicate that log coi3; (Y) is smaller than log c03,B; (Eu) while log (Y) is larger than log (Tb). Consequently, in a purely chemical sense, the position of yttrium carbonate complexation constants among the series of constants appropriate to the rare earth elements is ambiguous However, in an oceanographic context, questions surrounding the comparative character of REE and Y complexation can be simplified somewhat by considering the comparative extent of Y and REE complexation in seawater as a function of carbonate ion concentration Figures 2A, B C depict the seawater spe ciation of Eu, Gd, Tb, Dy, Ho and Y as a function of lo gM co3 where M co3 is the sum concentration of carbonate ions in seawater in all forms 20 0 M co3 = [C03 ] + [NaC03 ] +[MgC03 ] + [CaC03 ] The total Y and REE concentrations in seawater ([M3+)T) can be resolved into individual species contributions using the following Equation.: [M3+hf[M3+]=1+ciPI(M)[CI-] +FPI(M)[p-] + + ,(M)Mco3 + t(M)MHco 3 (6) Chloride and fluoride complexation stability constants were estimated from the work of Lee and Byrne (1993b). Hydrolysis constants (p; = [MOH2+ ] [H+ ] [M3+ r1 ) and sulfate complexation constants were taken from the results of Byrne et al. (1987) Bicarbonate complexation constant estimates for Eu, Gd, Tb and Y were estimated using the results shown in Table 2.1 plus the results of Lee and Byrne (1993a). Bicarbonate constants for Dy and Ho were estimated using the procedures of Lee and Byrne (1992) 2-5 The free concentrations of Cl, F-and SO 4 were taken as [CI-]= 0 .56, [F-] = 3.5xlO, and 23

PAGE 38

Table 2.3 Conditional REE Carbonate Stability Constant Comparisons Element I Method References Eu 5 82 10.5 1.0 s Lundqvist a ( 1982) Eu 5.86 10.1 0 68 s Cantrell and Byrne (1987a ) Eu 5 .90 9 90 1.0 s Chatt and Raob ( 1989 ) Eu 5.88 10 03 1.0 s Rao and Chatt ( 1991) Eu 5 .80 10.12 0 7 s Lee and Byrne (1993a) Eu 10.1 0 7 sp Thompson and Byrne ( 1987) ---------------------------------------------------------y 5.94 0 7 p Spahiuc ( 1985) y 5.71 10.34 0 7 s this work a recalculated using the carbonic acid and bicarbonate dissociation constants of Cantrell and B y rne ( 1987a). b r e c a lculated by Rao and Chatt ( 1991 ). c calculated using the 0.9 ionic strength results and activity coefficient model of Spahiu ( 1985 ) plus the carbonic acid and bicarbonate dissociation constants of Cantrell and Byrne (1987a). method : s-solvent exchange; sp-spectroscopy; ppotentiometry. 24

PAGE 39

Table 2.4 Ligand Concentration s and REE Complexation Con s tants in Seawater REE 1(M) 1(M) 1 (M) 1 (M) Eu -0.55 2.93 1.9 -8.34 Gd -0.57 2 .90 1.9 -8.54 Tb -0.57 3.01 1.9 -8.44 Dy -0.59 2 98 1.9 8 .54 Ho -0.60 2 .96 1.9 8 .54 y -0.57 3 01 1.9 -8.24 REE 1 (M) 2(M) Eu 5 .36 9 25 1.83 Gd 5 .2 2 9 21 1.86 Tb 5.34 9 37 1.84 Dy 5.40 9 .52 1.74 Ho 5.44 9.57 1.67 y 5.26 9.45 1.49 Ligand concentration s (m o l dm3 S=35) ligand c oncentration cr 0 .56 502 4 0.0095 F -3 .5x I05 Hco; 2xto3 25

PAGE 40

The Cl, F-, and hydrolysis constant estimates used in construction of Figures 2A, B and C are shown in Table 2.4 Carbonate complexation constants consistent with carbonate ion concentrations expressed as total carbonate (Mco3 ) in seawater were calculated using the following eqns. (Cantrell and Byrne, 1987a) : (7) For the sake of simplicity in our Fig 2.2 comparisons, the total bicarbonate ion concentration was taken as approximately constant in seawater (MHco3=2xl0-3) and Mco3 was calculated as logMco3 =log MHco3 +log SWK2 +pH = -11.79 +pH (8) where pH (-log [H+]) is expressed on the free hydrogen ion concentration scale and SWK2 is a bicarbonate ion dissociation constant appropriate to seawater (Byrne et al., 1988). The Fig 2.2A speciation scheme constructed for yttrium using the carbonate complexation constants 1(Y)=5.26 and shows that YC12+, YF 2 + YOH2+ and are minor species in seawater compared to Yeo; and Y(C03);. Very similar results are obtained for Eu, Gd, Tb, Dy and Ho. Fig. 2.2B shows the ratio between total metal and free metal ([M 3+]T/[M3+]) for yttrium, gadolinium and holmium. On the basis of this comparison (Fig. 2 2B), the extent of Y complexation in seawa ter is less than that of holmium and greater than that of gadolinium Fig. 2 .2 C shows a comparison of differences in the extent of Y, Dy Tb and Eu complexation calculated using eqn. (6). The difference term log[M 3+]T/[M3+] -log[Y3+]T/[Y3+ ] plotted as a 26

PAGE 41

function of carbonate ion concentration and pH, indicates that the inorganic complexation of Y in seawater is most closely related to the complexation of Tb. The so lution complexation intensities of yttrium and terbium (Fig. 2.2C) are quite similar over the entire range of conditions encountered in seawater. The generally small temperature dependence for carbonate complexation (Cantrell and Byrne, 1987b; Byrne et al., 1988) and an expected gen eral simi larity in Y and Tb partial molal volumes (Lee and Byrne, 1994) indicates that the solution chemistries of yttrium and terbium should be closely linked throughout the ocean 's volume. In previous assessments of YIREE comparative complexation, yttrium has been likened to Tb (Cantrell and Byrne, 198 7) and Ho (Cantrell, 1988 ; Byrne et al. 1988) Identification of Th as a reasonable Y analog was based (Cantrell and Byrne, 1987a) on strong correlation between REE carbonate and oxalate stability constants, and similarity in Th and Y complexation behavior with oxalate. Identification of Ho as an appropriate REE analog for Y is based (Cantrell, 1988) on the similarity (Shannon, 1976 ) in y3+ and Ho3 + ionic radii. Based upon linear free energy analysis (Lee and B yme, 1993a ), and s imilarity in y3+ and Ho3+ ionic radii, it would be predict e d that yttrium carbonate stability constants should be no larger than those of holmium. Based upon the direct experimental evidence in the pre se nt work, Tb is identified as the REE mo s t similar to Y in carbonate complexation behavior. In view of this identification direct experimental carbonate complexation comparisons s hould be performed in which both Y and Tb (and perhaps neighboring REEs) are present. Through ICP-MS analysis, we are ab l e to conduct such comparisons with Y and a variety of rare earth. Although the solution chemistries of Y and Tb in seawater are expected to be similar, it should be anticipated that, due to differences in their surface chemistries, the oceanic behaviors of th ese metals should be distinct. To the extent that the REE surface complexation propertie s of marine particles are dominated by organics (Byrne and Kim, 27

PAGE 42

pH 0 .00 ' 7 A ll 1.00 =: s: "' .J "" c ] .00 ...... .. . ... .).I l osMcm .. a p H 1.1 '"' H.; .. 1.50 i / Cd s-00 ..!! !.SO '"' -4.20 .... U O .).AI) losMcru pH .. 1A u Olll > uo ?D y > '"' ;,; ::< .0. 1 0 "' .: -0.20 -4.20 ... oo 3 80 .J.IoO losMco, Figure 2.2 Yttrium speciation in seawater. (a) Yttrium Complexation in Seawater as a Function of pH and Mco3 The log[YL]/[Y]T calculations in this figure are based on eqn (6). Mco3 is calculated using eqn (8) with MHC0 3 =2x 10-3 mol dm-3. (b) Comparison of Y Gd and Ho Complexation in Seawater (using eqn. (6)). (c) The Comparative Complexation of Y, Eu, Tb and Dy in Seawater is Depicted Using Yttrium as a Reference Element. The term log([M 3+]T/[M3+]) is calculated u s ing eqn. (6). 28

PAGE 43

1990) and the complexation behavior of Y with organics is generally that of a light rare earth (Lee and Byrne, 1993c) yttrium should not be as effectively scavenged from seawater as is Tb. Some insight into the possible magnitude of the differences in Y and Tb surface complexation can be gained by comparing available organic complexation data for y 3 + and Tb3 + obtained under comparable laboratory conditions ( 0.1 mol dm3 ionic s trength, 20 to 25 C). Calculation of differences in terbium and yttrium complexation constants for Tb andY complexation by 90 organic ligands (Smith and Martell 1975 1976, 1989; Martell and Smith 19 7 4, 1977 1982) demonstrate s that R is greater than or equal to 0.6 for 10% of the ligand s examined while for only 1 % of the available comparisons. Similarly for 20% of the s urveyed ligands while for only 2 % of the formation constant compari s on s Based on the Tb and Y carbonate complexation results reported in this work, and the complexation properties of Tb and Y by organic li gands, terbium should be more strongly scav enged from seawater than yttrium and shale normalized concentrations of yttrium in s eawater should be greater than that of Tb. Field observation s indicate that this expected b e havior is, in f act observed. The work of Shabani et al. ( 1990) provided evidence for shale normalized Y fib ratios which are on the order of 2.5 and the work s of Zhang et al. (1994) and Bau et al. (1995) show substantial shale normalized yttrium enrichments in seawater relative to the heavy rare earths Due to the similarity of yttrium and terbium so lution chemistries, and the simi larity of yttrium and light rare earth s urface chemistries in seawater (Lee and Byrne, 1993c ), it should be anticipated that shale normalized distributions of yttrium should be di s tinct from shale normalized distributions of any particular rare earth. The si mple rare earth s cavenging model of Byrne and Kim (1990) indicated that the shale normalized concentration (([M3+h)sN) of a rare earth or yttrium, can be approximated as: 29

PAGE 44

log([M3+h)sN = log:L, -log(MK) +constant (9) n j The first tenn on the right hand s ide of eqn. (9) is identified with the solution complexation terms shown in eqn (6). The tenn log(MK) refers to the complexation of rare earths and yttrium on particle surfaces. If eqn. (9) is used to predict the relationship between the shale nonnalized solution concentrations of the rare earths and yttrium the following relationship is obtained. (10) With respect to the s urface complexation tenn in eqn. (I 0), the simplest case which can be envisioned is that which invo lv es a suite of organic ligands with constant relative proportions of organic-ligand site-types. In this simple approximation, the tenn log( MK) YK for any particular rare earth becomes a constant. For any heavy rare earth heavier than Sm, the work of Lee and Byrne (1993c) indicated that log(MK) should be greater than 0. YK Surface complexation thereby contributes a constant negative tenn to log([M3+h) SN values [Y3+h for rare earths heavier than Sm. Due to variations in the marine carbonate system, it is clear ( Byrne et al, 1988) that eqn. 10 tenns involving solution chemistry are highly variable The i s most closely a pproximated as zero when M is Tb. In this case, according to our s imple model, the shale normalized distribution s of Y and the heavy rare earths, although consistently distinct du e to organic surface complexation effects, should be most si milar for yttrium-terbium comparisons. As a means of identifying similarities between Y an d rare earth ocean 30

PAGE 45

chemistries, Zhang et al. (1994) examined Y and REE deep water enrichment factor s, Enw, expressed as [M3+ h (3,000 meters)/[M3+ h (surface). Zhang et a1 (1994) found that these enrichment factors plotted against rare earth atomic number show a minimum at Tb, and noted that a similar result is obtained using the data of Piepgras and Jacobsen (1992). Thus, field data lend some support to the simple comparative di st ribution model embodied in eqn. (10) and the contention that the solution chemistry of y3+ should be most closely identified with that of Tb3+ in seawater. 31

PAGE 46

3. COMPREHENSIVE INVESTIGATIO N OF YT1RIUM AND RARE EARTH ELEMENT COMPLEXATION BY CARBONATE IONS USING ICP-MASS SPEC1ROME1RY 3.1 Abstract Carbonate stability constants for yttrium and all rare earth elements have been determined at 2s c and 0.70 molal ionic strength by solvent exchange and ICP-mass spectrometry Measured stability constants for the formatio n of Mco; and M(C03 ) ; from M3+ are in good agreement with previous direct measurements which involved the use of radiochemical techniques and trivalent ions of Y, Ce Eu, Gd, Tb and Yb. Direct ICP-MS mea s urements of Mco; and M(C03 ) ; formation constants are also in general agreement with modeled stability constants for the metals La, Pr, Nd, Sm Dy, Ho, Er, Tm, Lu based on linear free energy relationships. The experimental procedure s developed in this work can be used for assessing the complexation behavior of other geochernically important ligands such as phosphate, sulfate and fluoride. 32

PAGE 47

3. 2 Introduction Examination of rare earth element (REE) complexation behavior i s best achieved through comparative investigations in which all REEs are studied u si ng a consistent set of procedures (Byrne and Sholkovitz 1996). While this situation is typical for REE complexation with organic ligands (cf. Lee and Byrne 1992 Byrne and Li, 1995), it is commonly the case that REE complexation constants for geochemically important inorganic ligands such as carbonate, phosphate and fluoride are obtained from experiments with s imple solution in which REEs are studied one at a time or as a group of no more than a few elements (cf. Lee and Byrne 1993a, b and c; Cantrell and Byrne, 1987a). The inorganic speciat ion of REEs in seawater is dominated by carbonate complexation (Turner et al. 1981 ; Cantrell and Byrne 1987a; Lee and Byrne 1993 a). Previous d e termination s of REE complexation with car bonate ion s have involved only s i x REEs (Table 3.1 ) Among the methods used for assessment of REE carbonate stability constants only solvent exc hange inve s tigation s have provided consistent re s ults for both first ( coJ3; (M)) and second (co (M)) stability constants (Cantrell and Byrne, 1987a) Previous examination ) of REE carbonate complexatio n via radiochemical s olvent exchange technique s have been limited to e l ements readily avai lable in radioactive form (Cantrell and Byrne, 1987a; Lee and Byrne 1993a; Lundqvist 1982) In instances where stability constants have been directly mea s ured for only a few elements linear free-energy relationships (LFER) have been used (Turner et al., 1981; Lee and Byrne 1992 1993a and b; Byrne and Li 1995) to estimate constants for the entire suite of rare earths. It s hould be noted that previous LFER assessments provide a complexation pattern based on rare earth complexation with organic ligands (Lee and Byrne, 1992, 1993a and b ; Byrne and Li 1995) Evaluation of REE 33

PAGE 48

Tabl e 3.1 Summary of Yttri u m and Rare Earth Carbonate Stability Constant Determinations M3+ log coJ3; (M) log solvent exchange other techniques so l vent exchange y 5 71 h 6.02f 10 .34h La 5.67d Ce 5.413 5.26b 6.31 e 9.26 3 9 .37b Pr Nd Pm Sm Eu 5.85 a, 5.81b 7.1 g 10.02 a, 10.14b, 10.03j 5 88i, 5.93i, 5.8ia 10.72i, 10.5i a Gd 5.68b, 5.65h 10 09b, 10 .12h Tb 5.79b 10.27b Dy Ho Er Tm Yb 6.163 6.14b 10.88 3 10.95b Lu a Cantrell and Byrne (1987a) (solvent exchange, 0.7 molal NaC 1 04 ) b Lee and Byrne (1993a) (so l vent exchange, 0.7 molal NaC 1 04 ) c Thompson and Byrne (1987) (spectrophotometry 0.7 molal NaC104 ) d Ciavatta et al. ( 1981) (potentiometry, 3 molar N aClO 4 ) e Ferri et al. (1983) (solubi l ity, 3 molar NaC 1 04 ) f Spahiu (1985) (potentiometry, 3 molarNaC104 ) g Ruzaikina et al. ( 1978) (solubility, 0 ionic strength) h Liu and Byrne (1995) (solvent exchange, 0.7 molal NaCl04 ) i Lundqvist ( 1982) (solvent exchange, 1.0 molar NaC104 ) i,a Lundqvist (1982) recalculated as in Cantrell and Byrne (1987a) j Rao and Chatt ( 1991) (so l vent exchange, 1.0 molar NaC104 ) other techniques 11.1 e 10.6 g 10.1 c 34

PAGE 49

inorganic complexation in natural waters (Lee and Byrne, 1992) is then based, in part, on organic ligand complexation behavior. The interactions of trace metal ions with individual ligands can have unique characteristics whereupon LFER generalizations may not provide reliable predictions (Byrne and Li, 1995). Due to the dominance of REE complexation with carbonate ions in natural waters, direct investigation of carbonate complexation behavior for all rare earths is essential in order to address important questions about the origins of REE abundance patterns in the environment. This work presents the results of direct carbonate complexation measurements for yttrium and the entire suite of rare earth elements (YREEs). The ICP-MS procedures developed in this work can also be used to measure REE stability constants with other geochemically important ligand s such as phosphate fluoride, chloride, sulfate and hydroxide. 3. 3 Experimental procedures 3. 3.1 Reagents Yttrium and REE (YREE) ICP standard s were obtained from SPEX Chemical Tributyl phosphate (TBP) was from Fluka Sodium hydroxide and iron wire (99 97 % ) were obtained from J. T. Baker. Sodium perchloride and sodium bicarbonate were obtained from Aldrich Chemical. All labware was cleaned by washing with Micro solution (Cole-Panner), followed by immersion in 4 M HCI for at least one week and s ubsequent rinsing with Milli-Q water. 3. 3. 2 TBP extraction procedures The tributyl phosphate (TBP) extraction procedures in thi s work closely follow those des cribed in the radiochemical studies of Lee and Byrne (1993a and b), Cantrell and 35

PAGE 50

Byrne (1987a) and Liu and Byrne (1995) with minor modification. Our experiments were conducted using Y and all REEs except Pm Our measurements examined the distributions of YREEs between an organic phase, TBP, and an aqueous phase (0.70.02 molal ionic strength) composed principally of NaCl04. YREE distribution s b etwee n the phases were examined as a function of aqueous carbonate and bicarbonate ion concentrati ons Preparation and purification of TBP and NaC10 4 solutions, described in detail by Cantrell and Byrne (1987a and b) and Lee and Byrne (19 93a and b), can be summarize d as follows: NaCl04 stock solutions (3.4 molal) were allowed to stand at least 24 hours at pH 9 .5, followed by filtration with 0.22 J..Lm polycarbonate filters (Nuclepore Corp.). About 250 cm3 of TBP were mixed with an equal volume of 1 N NaOH in a 500 cm3 separa tory funnel. The mixture was equilibrated by vigorou s ly s haking for 20 minutes. Subsequ e ntly approx imately 20 minute s were allowed for pha se separation and the NaOH solution was discarded. The TBP phase was then added to se veral 30 cm3-capacity Teflon tube s and centrifuged at 10,000 rpm for 5 minute s to elimi n ate trace NaOH. The purified TBP was tran sfer red t o a 500 cm3 se paratory funnel and combined with 250 cm3 of 0.68 molal NaC104 so lution (prepared by dilution of the filtered 3.4 molal NaC104 stock solution). This mixture was equilibrated by vigorously s hakin g for 20 minutes. The TBP phase was then allowed 20 minutes to separate from the NaC104 phase and the aqueous phase was discarded. Another 250 cm3 of 0.68 molal NaCl04 were added to TBP and the procedure was repeated. After separation of TBP from the NaCl04 so lution 200 cm3 of thi s pre equilibrated TBP were mixed in a 500 cm3 Teflon bottle with an equal vo lume of 0.68 molal NaC10 4 solution containing YREEs The Teflon bottle was immersed in a doublewall jacketed beaker which was thermostated at 25.1 oc. Solutions of yttrium plus 13 REEs in 0.68 molal NaCl04 (excl uding Pr or Nd which served as yield standards for monitorin g REE recovery during an Fe coprecipitation procedure ) were prepar ed from individual sta ndard so lution s (SPEX). The initial concentration of eac h YREE in our 3 6

PAGE 51

experimental medium was 50 ppb. Thus initial aqueous YREE concentrations ranged from 6x10 7 molal for Y to 3x107 molal for Lu with total YREE concentration on the order of 5x10 -6 molal. Equilibration of TBP and aqueous solution was effected by vigorously stirring and bubbling with N2 (99.998% Air Products) or C02 N 2 mixtures (30.0% C02, Air Products). The equilibration (bubbling) period in our experiments was 20 minutes, and an additional 20 minutes was allowed for phase separation. Solution pH on the free hydrogen ion concentration scale (McBryde 1969,1971; Byrne and Kester, 1978) was measured using a Ross combination pH electrode (Orion No 810200) and a Coming model 130 pH meter. As in previous works (Lee and Byrne, 1993a and b; Cantrell and Byrne, 1987a) the electrode was calibrated by titrating the unbuffered working solution (0 68 molal NaC104 ) with strong acid (HCl04) Three molal NaCl was used as our electrode's filling solution to prevent precipitation of KCl04 at the electrode's liquid junction. After measurement of the pH of the aqueous phase, samples from the TBP and aqueous phase were taken. Solution pH was then adjusted with a titrant solution consisting of 0 68 molal NaHC03 in 0 68 molal NaCl04 solution. Repetition of this procedure usually resulted in 8 titration points in each experiment within the range pH where hydrolysis and sorption of lanthanide ions have been reported to be insignificant (Lundqvist 1982). After each equilibration and separation of aqueous and organic phases, 10 cm3 of TBP was added directly to a 30 cm3-capacity Teflon centrifuge tube (Nalgene). The tube was then capped and the amount of transferred TBP was weighed. Ten cm3 of aqueous phase was pipetted into a 25 cm3 -capacity plastic bottle and after two hours of further phase separation, 9 cm3 of each aqueous sample was weighed in a 30 cm3 Teflon tube and immediately subjected to extraction procedures by Fe(lll) coprecipitation. 37

PAGE 52

3. 3. 3 Post solvent exchange procedures In contrast to the previous works of Cantrell and Byrne (1987a) and Lee and Byrne ( 1993a and b) which used radiochemical counting techniques sta ble isotopes of YREE s were used in this investigation and the YREE concentrations in each s ample were measured using ICP mass spectrometric techniques. ICP-MS techniques invo l ved YREE analysis in a matrix of 1 % HN03 Consequently new experimental procedure s were developed to extract YREE s from the TBP phase, and also from the aqueou s phase which had high concentrations of ions (Na+ and ClO to ICP-MS analysis 3. 3. 3.1 Extraction of YREEs from TBP YREEs in TBP were extracted into 0.68 molal NaC104 s olutions at high co; concentration (pH 6.9 and 0.30 atm. C02 partial pressure). Ten cm3 of pH 6.9 aqueous solution were comb in ed with each TBP sample in Teflon tubes under 0.30 atrn partial pre ss ure C02 The tubes were capped and the two phases were e quilibrated by vigorously shaking for 5 hours. After allowing 20 minutes for phase separation the aqueou s portion was transferred to a 25 cm3 plastic bottle Next another 10 cm3 of C02-equilibrated NaC104 solution was used for a second extraction. The two extracted so luti ons were then combined. Fifteen cm3 of combined solution were weighed after separation from trace amounts of TBP. Experiments with mixed YREE sta ndard s s howed recoveries of better than 98% for all YREE s. 3. 3. 3. 2 Coprecipitation of YREEs by Fe(Ill) hydroxide YREEs were separated from NaC104 in our aqueous so lution s by coprecipitating the YREEs with Fe(ID) (Elderfie ld and Gre av es, 1982). Fe ( III ) so lution ( 1 mg Fe cm-3 ) was prepared by dissolving 1 g of high-purity iron wire in seve ral cm3 of concentrated nitric acid. The dissolved Fe was then diluted to a total volume of 1000 cm3 Pr or Nd standard solutions (10 or 50 ppb) were added to samp l es which previously ( during phase 38

PAGE 53

equilibration) did not contain one or the other of these metals. These e l ements were used as s tandard s for monitoring YREE recoveries following coprecipitation and subsequent dissolution. Coprecipitation was initiated by adding 0.75 cm3 of 1000 ppm Fe(III) to each sample followed by 2 cm3 of 1M ammonia so lut ion. These samp le s, with pH between 9.5 to 10, were then equilibrated by shaking vigorously for 3 days at room temperature (23C). The precipitates were collected on 0 22 in-line cartridge filters (Coming) attached to 5 cm3 syringes (Fis her Scientific). To each of the filtered so lution s an additional 0.50 cm3 of 1000 ppm Fe(III) was added followed by an additional 3 day equilibration. Subsequently each samp le was filtered through the same filter used to collect the initial coprecipitates The combined coprecipitate was rinsed three times with 15 cm3 of Milli-Q water. Finally 10 cm3 of 1 % HN03 was used to dissolve each coprecipitate in a 12 cm3 plastic tube. Examinations with mixed YREE standards containing all YREEs s howed recoveries for all YREEs greater than 98%. 3. 3. 4 Measurement of YREE concentrations A mixed Y and REE standard solution containing 100 ppb of each element in 1 % HN03 was prepared from the SPEX mixed stan dard solution. Five cm3 of this standard s olution and 1.25 cm3 of s tock iron solution (1000 ppm Fe(III)) were combined with 1 % HN03 to make a solution which contained each YREE at a concentration of 50 ppb. Blank solutions cons i sted of 1 % HN03 with no added YREEs. Fe(III) was added to standard so luti ons and blanks to matrix match standards, blanks and samples One hundred of In and Re mixed solution (prepared from SPEX ICP standard so luti ons) was added to each solution (blanks, standards and samples) to provide an internal ICP measurement standard. A mass-dependent internal standard calibration was performed by interpolating between 115In and 187Re. 89Y was referenced directly to 115ln. YREE concentrations were measured on a Fisons ICP PQS mas s s pectrometer. 39

PAGE 54

3. 3. 5 Data processing ICP-MS output was normalized to the yield standard (either Nd or Pr which was used to monitor Fe coprecipitation and dilution procedures). Subsequently YREE concentrations in TBP and the aqueous phase were calculated based on sample weights and dilutions Distribution coefficients in the absence of complexing ligands (D0 ) and distribution coefficients in the presence of complexing ligands (D) were determined experimentally as D0= [M3+] I [M3+] and D = [M3+] I [M3+] org T org T T (1) where [ ]T and [ ] indicate the total and free (uncomplexed) concentrations of YREEs (M3+) in the aqueous phase, and [M ]T represents the total concentration of YREEs in the TBP. Non-linear least squares methods (Byrne and Kester 1978; Cantrell and Byrne, 1987b, Lee and Byrne, 1993 Liu and Byrne, 1995) were used to estimate REB-carbonate stability constants Our REE distribution data were described using the equation D01D = 1 + .1(M) ]T + .1(M) [HC03 ]T (2) where ]T [NaCO)] and [HCO} ]T = [HCO} ] + and [ ] represents the free molar concentration of each species Carbonate and bicarbonate ion concentrations were determined using the equations (Cantrell and Byrne 1987b) [HCO} ]T= (KJ<1 ) Pco21 [H+] (3) where Pco2 is the C02(g) partial pressure, [H+] is the free hydrogen ion concentration, K0 . is the Henry's law constant, and K 1 and K2 are the first and second conditional dissociation constants of carbonic acid appropriate to 0 68 molal NaCl04 The values pKJ< 1 = 7.56 and pK2 = 9.53, determined previously (Cantrell and Byrne, 1987a), were used in this 40

PAGE 55

0 0 0 work The parameters co3P1(M), co3P2(M), and Hco 3P1(M) in eqn (2) are conditional REB stability constants defined as follows : (4) Use of eqn. (2) in non-linear least squares analysis involved minimization of the residual sum of squares (RSS) function II I. { 1-D; I (1 + co3 /3; (M)([ co;]T );+ co3 (M)([ HC03 /3; (M)([Hco; ]T ); ) } 2 i (5) Bqn. (5) provides equal weighting to each of our experimental data points when the values of Di under our experimental conditions varied by 3 orders of magnitude. 3. 4 Results and Discussion 3. 4.1 REE complexation constants Conditional carbonate stability constants (co3Pn(M)) for yttrium and all REBs except Pm at 25C and 0.70.02 molal ionic strength are summarized in Tables 3.2a and 3.2b. The standard errors produced in each least squares minimization provide a comparative measure of the quality of each data fit. Table 3.3 shows weighted average stability constant results for yttrium and REB carbonate based on the results in Tables 3.2a and 3 .2 b. These data are shown graphically in Fig 3 .1. The trend of Fig 3.1 can be 41

PAGE 56

Table 3 .2a Conditional Carbonate Stability Constants of Yttrium and Rare Earths Determined at 25C and 1=0.70 .02 molal: coJ3; (M) Results M3+ experiment I experiment IT experiment ill y (6.05.75)x105 (5.75.28)x 105 (5.20.36)x105 La ( 1.21.06)x 105 (9 72.22)x 104 ( 1.24.27)x 105 Ce (2.19.11)x105 ( 1.860.22)x 105 (2.34.23)x 105 Pr (3.26.16)x 105 (3 .02.25)x 105 Nd (3.57.21)x105 Pm Sm (5.47.16)x 105 (5.03.32)x 105 (5 .44 0 .11 )x 105 Eu (5.69.I6)x 105 (5.28.33)xi05 (5.68.I9)x1 05 Gd (4.69. 12)xl05 ( 4.1I .29)x I 05 ( 4.49.29)x 105 Tb (5.61.18)x105 (5.06.34)xl05 ( 4 .95.38)x 105 Dy (6.86.I9)xi05 (6.1 0.35)xi05 (6.59.92)x105 Ho (6.93.25)x 105 ( 6.25.42)x I 05 ( 5 .52.50)x 105 Er (7.75.29)x105 (7 .06.49)x 105 (7 .98.31 ))x 105 Tm (9 .39.32)x I 05 (8.21.50)x105 (7 .50. 73 )x 105 Yb ( 1.26.11 )x 106 ( 1 17 .07)x 106 ( 1 .18.25)x 106 Lu ( 1. 1 0. 06) X 1 06 (1.01.06)x106 ( 1 .00.16)x 106 42

PAGE 57

Table 3 2b Conditional Carbonate Stabi l ity Constants of Yttrium and Rare Earths Determined at 25C and 1=0 .70.02 molal : co Results 3 M3+ experiment I experiment IT experiment ill y (l.l5.14)x1010 (1.17.06)x1010 ( 1.42.06)x 101 0 La (3.60.70)xl08 (6.81 .40)x 108 (4.59.70)x108 Ce (1.77.10)x109 (1.98.33)x 109 ( 1.36.25)x 109 Pr (3 .82.24)x109 (3. 19.30)x109 Nd ( 4.48.37)x 109 Pm Sm ( 1 14.30)x 1010 ( 1.08.07)x 1010 (9.74 .17)x109 Eu ( 1.38.03)x 1010 ( 1.31 .07)x 101 0 (1.19.03)x 101 0 Gd (9.70.20)xl09 ( 1.03.06)x 1010 (6.98.40)x 109 Tb ( 1.94 0.04 )x 1010 ( 1 84.08)x 1010 ( 1.52.06)x 1010 Dy (2.51.04 )x 1010 (2.55.09)x 101 0 (1.74 .15)x 1010 Ho (3.20.06)x1010 (3.07.1I )x I 010 (2 70.09)x I 010 Er ( 4.08.08)x I 010 (3.94.I4)x 1010 (2 .88.24)x I 01 0 Tm (5.92.IO)x 1010 (5.6IO.I6)x I 010 (4.68.I5)xi010 Yb (6.67.32)x 1010 (6.04 0.23)x 1010 ( 4. 80.4 7)x 1010 Lu (7 .61.I9)x I 010 (6 .97.2I )x I 010 (5.63.33)x1010 43

PAGE 58

3.3 Summary of Conditional Carbonate Stability Constants of Yttrium and Rare Earths (25"C and 1=0.70.02 molal) ffiE coJ3; (M)S. E. coJ3; (M)S .E. HcoJ3;(M) log co/3; (M) logc0J3;(M) r (5.59.2 1 )x 105 (1.28.04)x 101 0 5 5.75.02 10 11 .01 0.040.0 11 ..a (9 99.20)x 104 (6.01.34)x108 15 5 00.01 8 78.03 0.042.026 ::e (2. 16 .09)x 105 ( 1. 73.09)x 109 23 5.33 02 9.24.02 0.040.0 1 7 ,r (3.19. 1 3)x 105 (3.59.19)xl09 24 5.50 .02 9.56 02 0.035.00 1 .J"d (3.57.21 )x 105 ( 4.48 .37)x I 09 26 5.55.03 9 65.04 0.035 >m ;m (5.42.09)x 105 ( 1.0 1.0 1)x 1010 30 5.73 0.01 10 .01.0 1 0.038.005 (5.63 I 2)x I 05 ( 1.28 0.02)x 101 0 32 5.75.01 10.11.01 0.042 005 }d (4.59.11)x105 (9.16.02)x 109 15 5.66.01 9 96 .01 0 046.0 13 :b (5.41.14)x 105 (1.82.03)x 101 0 25 5.73 .01 10 .26 0.01 0 065 006 )y (6.69 0 16)x105 (2.47 04)x 1010 23 5.83 0 .01 10.39 01 0.054.014 fo (6.55 0.20)x 105 (3.04.05)x 1010 15 10 5.82 0.01 10.48.01 0.078.0 11 (7.59 24)x105 (3.96.07)x 101 0 16 5 88 .01 10 60.01 0.064.0 1 7 (8.85.25)x 1 05 (5.57.07)x 1010 1510 5 95.01 10.75.01 0.078.009 {b ( 1 19 .06)x 106 (6.05 0.17)x 1010 10 6.08 .02 10.78 .01 0 040 005 .. u ( 1.05 .04 )x 106 (7 .05.13)x 1010 1010 6 .0 2 0 .02 10.85.01 0.063.006

PAGE 59

6.2 6.0 5.8 5.6 -ca: 8 u "" ,g 5.4 5.2 5 0 4.8 y La Ce Pr Nd Pm Sm 11.0 10 5 0 10 0 -d' 8 u c.o ,g 9 5 9 0 8.5 y La Ce Pr Nd Pm Sm 0 0 Eu Gd REE Eu Gd REE 0 o Cantrell and Byrne 1987 o Lee and B y rne 1 993 o Liu and Byrn e, 1995 ---this work ThDy H o Er Tm Yb Lu 0 Cantrell and B yrne, 1987 Lee and Byrne, 1993 0 <> Liu and B y rne, 1995 --this work Th Dy Ho Er Tm Yb Lu Figure 3 1 Yttrium and REE Carbonate Comp l exation Constants at 2sc and 0.70.02 molal Ionic Strength. (a) coJ3;(M) Results (b) Results 45

PAGE 60

Table 3.4 Yttrium and REE Carbonate Complexation Constants (25"C and Zero Ionic Strength ) y 7.78 13.16 La 7.03 11.83 Ce 7.36 12 .29 Pr 7.53 12.60 Nd 7.58 12 .7 0 Pm Sm 7 76 13. 06 Eu 7 78 13.16 Gd 7.69 1 3 .01 Tb 7.76 13.3 1 Dy 7.86 13.44 Ho 7.85 13. 53 Er 7 .91 13.65 Tm 7.98 13.80 Yb 8.11 13.83 Lu 8.05 1 3.90 46

PAGE 61

described as a general increase from light to heavy REEs with an obvious negative anomaly for both coJ3; ( Gd) and ( Gd) The HcoJ3; (M) results summarized in Table 3.3 show increasing values between La and Dy and somewhat smaller values for heavier REEs YREE bicarbonate stability constants are relatively poorly defined in our experiments, which were primarily designed to sensitively determine co/3; (M) and (M). The final column of Table 3.3 lists the ratios of second and first stepwise stability constants . 2 ( K 2 I K1 = /32 I (/31 ) ). Our results indicate the REEs have K 2 I K1 values generally within the range 0 .04 to 0 07 Our experimentally determined complexation constants (Table 3 3) at 0.70.02 molal ionic strength can be corrected to zero ionic strength using the following relationship (Ca ntrell and Byrne 1987a) : r [Co2 ] o Mco log co/31 (M)=log coJ/31 (M) + log 3 +log [C0332-]T r M3+ Yeo;(6) ( 7 ) where Y ; are individual ion activity coefficients and the term log [co;] T I [co; ] accounts for car bonate ion pairing with Na+. Using eqns (6) and (7) Cantrell and Byrne (1987a) s howed that formation constants at 0.7 molal ionic strength and zero ionic strength are related as follows: 0 log co3/31 (M)=log co3f31 (M) + 2.03 (8) 0 log co3/32 ( M)=log co3/32 (M) + 3.05 (9) YREE complexation result s at zero ionic strength estimated using eqns. (8) and ( 9 ) are s ummarized in Table 3.4. 47

PAGE 62

3. 4. 2 Comparisons with previous direct measurements Since carbonate complexation is a dominant influence on REE behavior in n atural waters, critical evaluation of comparative REE carbonate complexation constants is essential to interpretation of environmental REE fractionation pattern s Direct measurements of YREE complexation with carbonate ion h ave been obtained only for a few elements ( T a ble 3.1). Con se quently the pre se nt work constitutes the first comprehensive examination of comparative YREE carbona te complexation. Fig 3.1 s how s that the ICP MS complexation results for Y, Ce Eu, Gd Tb and Yb are in good general agreement with previous re s ult s (Liu and Byrne 1995; Cantrell and Byrne, 1987a and Lee and B yrne, 1993a ) obtained using radiochemical techniques. Our coJ3; (Ce) results are nearl y identical to the average d results of Lee and Byrne (1993a) and Cantrell and Byrne (1987a), and is within 0.02 log units of the Cantrell and Byrne (1987a) result. Our log co)3; (Eu) result is approximately 0.1 units lower than the averaged result s of Lee and Byrn e (1993a) and Cantrell and Byrne (1987a), while is within 0.03 units of the average d Lee and Byrne (1993a) and Cantrell and Byrne (1987a) results After small corrections for differences in ionic streng th the log (Eu) results of Rao and Chatt (1991) are within 0 1 lo g units of the results shown in Table 3.3. The corrected (Cantrell and Byrne 1987a) solvent exchange results of Lundqvist (1982) are within 0.07 log units of our log ca )3; (Eu) results and are within 0.4 log units of our log (Eu) result. Our log coJ3 ; (Gd) result is essentially identical to the results of Lee and Byrne ( 1993a ) and Liu and Byrne (1995) while our log c0JJ;(Gd) result is lower than the Lee and Byrne (1993a) and Liu and Byrne (1995) results by about 0.1 log units. Our log coJJ; (Tb) result is 0.06 log units lower than result of Lee and Byrne (1993a) while our result is essentially identical to that of Lee and Byrne ( 1993 a). The Table 3.3 log co3/3; (Y) value is 48

PAGE 63

within 0 .04 log units of the result of Liu and Byrne ( 1995) while the (Y) is lower th a n that of Liu and Byrne (1995) by 0 2 log unit s. Finally, the Yb r es ults
PAGE 64

are larger than the values given in Table 3.3 by approximately 0.6 log units and 0 9 log units. Cantrell and Byrne (1987a) noted that the log co)3; (La) result of Ciavatta et al. ( 1981) may be relatively poorly defined because no more than 6% of the complexed La3+ in the experiments of Ciavatta et al. (1981) was present as LaCO;. The log coJ3; (Ce) results of Ferri et al. (1983), which are much larger than those given in Table 3.3, are also inconsistent with the log coJ3; (Y) results of Spahiu ( 1985). Since the ionic radius and complexation properties of Y3+ clearly place this element between Sm3+ and Ho3+, depending on the nature of the complexing ligand (Liu and Byrne, 1995), it is clear that the YCO ; stability constant should be substantially larger than that of CeCO ; rather than smaller by a factor of two. While the solubility analysis of Ferri et al. (1983) produced log c oJ3; (Ce) estimates larger than our results by nearly an order of magnitude, the solubility analysis of Ruzaikina et al. ( 1978) produced a log (Eu) estimate smaller than that derived in the present work by approximately 0.8 log units As noted by Cantrell and Byrne ( 1987a), the analysis of Ruzaikina et al. ( 1978) appears to have involved inadequate equilibration times and neglect of important mononuclear and polynuclear europium hydrolysis products in their experimental solutions. Solubility analyses are generally complex compared to other procedures used to derive REE stability constants and appear to produce stability constant data of more questionable quality. 3.5 Summary Among the new stability constant results reported in this work, the results presented for yttrium, lanthanum and lutetium are of special interest. The position of Y among the suite of rare earth carbonate stability constants has been assessed previously (Liu and 50

PAGE 65

Byrne 1995) based on direct measurements of (Y) and plus LFER estimates for the stability constants of REEs heavier than Gd. This assessment led to the conclusion that yttrium carbonate stability constants closely resembled tho s e of Tb. In the present work it is seen (Table 3.3) that the closest relationship between yttrium and REE stabi lit y constants is that observed for Y and Eu. The average stability constants ( coJ3; (M) and (M)) for Y and Eu shown in Table 3.3 are identical. Our work shows that while coJ3; (M) for Y and Tb are very similar, log (Tb) is larger than that of lo g (Y) by about 0.15 l og units. In previous assessments of REE stability constants by LFER analysis, formation constants for REEs between Ce and Eu and between Tb and Yb have been assessed (Lee and Byrne, 1993a) by interpolation In contrast, LFER assessments for La and Lu are obtained by extrapolation and can be considered to be somewhat less well defined. Our results indicate that the LFER analysis of Lee and Byrne (1993a) provided a generally reasonable assessment of the comparative formation constant behavior of La and Ce as well as Yb and Lu. In the present work the difference between log coJ3; (Ce) and log co/3; (La) is 0.33 compared to a predicted difference of 0.28 in the work of Lee and Byrne (1993a) The difference, log (Ce)-log (La) obtained in the present work is 0.46 compared to 0 .62 in the work of Lee and Byrne (1993a). Comparison of Yb and Lu stabi lity constants in the present work shows differences, log co)3; (Lu ) log coJ3; (Yb) and log (Lu ) log (Yb ), equal to -0.06 and +0.07. The work of Lee and Byrne ( 1993a) produced corresponding differences equal to +0.0 1 and +0.07. In summary, the excellent overall correspondence between the formation constants of Y, Ce, Eu, Gd, Tb and Yb obtained in the present work and previous formation constant results (Cantrell and Byrne, 1987a ; Lee and Byrne, 1993a; Liu and Byrne, 1 995) obtained through comparatively straightforward radiochemical analysis, indicates that the procedures 51

PAGE 66

employed in our work provide a coherent assessment of comparative YREE carbonate stability constant behavior. As such, interpretations of yttrium and rare earth fractionation in natural waters can now be based entirely on directly measured, rather than estimated, YREE carbonate formation constants. 52

PAGE 67

4. RARE EARTH AND YT'IRIUM PHOSPHATE SOLUBILITIES IN AQUEOUS SOLUTION 4.1 Abstract Rare earth and yttrium phosphate solubility products range over more than one order of magnitude. Minimum solubilities are observed for light rare earths between Ce and Sm. For the elements Ce, Pr, Nd and Sm solubility products (log (M)=log( [M;3+][Po; ])) at zero ionic strength and 25C can be approximated as = -26. 3 0 2 Rare earth phosphate solubility products for well aged, coarse precipitates increase substantially between Sm and Lu, with estimated as 24 7. The solubility product of yttrium is similar to that of Ho (log (Y) = 25.0) and is much higher than those of all light rare earths. The solubility product of La is substantially larger than that of Ce = 0.5) Solubility products are strongly dependent on the conditions of solid phase formation. Fresh precipitates are much more soluble than slowly formed, well aged, coarse precipitates. The pattern of rare earth and yttrium phosphate solubility products is generally similar to the fractionation patterns which are developed during phosphate coprecipitation 53

PAGE 68

4. 2 Introduction Rare earth element phosphates are major REE-bearing minerals in the natural environment (Goldberg et al., 1963; Elderfield and Greaves, 1982), and it has been suggested (Jonassen et al., 1985; Byrne and Kim, 1993) that phosphate precipitation may be an important aspect of REE geochemical behavior. In spite of the possible importance of phosphate phases in the REE geochemical cycle, the systematic features of rare earth element phosphate (REEP04 ) solubility equilibria are poorly understood Reported REEP04 solubility equilibrium products range over more than two orders of magnitude (Fig. 4.1 ). The works of Tananaev et al. (1963, 1967) indicated that La has a higher solubility than Ce and solubility increases between Ce and Gd. The work of Jonassen et al. ( 1985) indicates that solubility decreases between La and Nd, and then increases between Nd and Er. The work of Byrne and Kim (1993) for fresh precipitates showed a generally increasing trend of solubilities between Ce and Yb. Summarizing these results, the general pattern for the whole REE series is an initial trend of decreasing solubilities among the light rare earths followed by an increasing trend for heavy rare earths. The most comprehensive study of REE solubility (Firsching and Brune, 1991) contrasts with the above generalizations. Firsching and Brune ( 1991) obtained a convex solubility pattern with the most soluble phosphates occurring in the middle of the REE series. In order to resolve these conflicting reports for REEP04 solubility behavior in aqueous solution we have examined REEP04 solubility characteristics for REEP04 solid phases using a variety of procedures. Our solubility results for individual rare earths are then examined in the context of comparative REE behavior during the process of REEP04 coprecipitation (Byrne et al., 1996). 54

PAGE 69

-23 0 r--.-.--.-.--.--.--.----,,----.---,,---,---,--,---,--.---,----, -23 .5 24 0 -24 5 =ll--25 0 ell 2 -25 5 -26. 0 -Byrne and Kim 1993 --a-Jonasso n et al. 1985 -26 5 --+-Tananaev et al 1963 1967 -Firsching and Brune 1991 -27. 0 Y La C e Pr Nd Pm Sm Eu Gd Th Dy Ho Er Tm Yb Lu REE Figure 4.1 Previou s REE and Yttrium Phosph ate Solubility Product Determinations (results have been recalculated from the original d ata) 55

PAGE 70

4. 3 Materials and Methods 4. 3.1 Preparation of materials Yttrium and All REEs except Ce and Tb were obtained as oxides (99.9%) from Aldrich. Cerium and Tb (A ldri ch, 99 9%) were obtained as chlorides Solution s were prepared by di ssolv ing each e l eme nt in 0 5 mol dm -3 HC10 4 Subsequently, three procedures were used to prepare Y and REE phosphate so lids In the first preparation (procedure A), each 0 1 mol dm 3 solution of Y or a REE was c ombined with NaH 2P04 (0.1 mol dm -3 ) to produce solutions equimolar in total phosphate and individual REE metal s. The fresh precipitates formed in this process were collected by filtering suspensions with 0.4 filters ( Nuclepore) The light rare earth phosphates (La through Gd) prepared in this manner had a finely dispersed appearance r e lativ e to the heavy rare earths In the second preparatory procedure (procedure B ), rare earth pho sphates were prepared by aging the procedure A precipitates in solution (25 1 C) for two months prior to filtration. Both fresh (procedure A) and aged (proc edure B ) precipitate s were rinsed extensively with deionized water to eliminate excess REEs and phosphate Rinse solutions were monitored spectrophotometrically until phosphate concentration s reached the limit of detection Subsequently each precipitate was rinsed with dilute acid ( pH=3) followed by deionized water washes until the pH of rinse solutions were near neutral. The third preparatory procedure ( procedure C) very closely followed the procedures of Firsching and Brune ( 1991 ). Equimolar amounts of REE and phosphori c acid were ad ded to a 1 dm3 beaker with 5 cm3 of 70% HCl04 to prevent precipitation Deionized 56

PAGE 71

water was added to bring the volume to about 0.9 dm3. Three to five grams of urea (in solid form ) were slowly added with constant stirring. Each solution was heated to hydrolyze urea. This process slowly raised the solut i on pH and brought about precipitation under mild supersaturation. The precipitates were washed with 0.1 mol dm-3 HC104 and with deionized water to eliminate any traces of carbonate 4. 3. 2 X-ray diffraction analysis of rare earth phosphates Rare earth phosphate precipitates from each procedure were examined by Scintag XRD immediately prior to their u s e in solubility analyses The precipitates formed in procedure A were entirely X-ray amorphous for the light rare earths and showed only hints of c rystallinity for the heavy rare earths. These observations are consistent with the physical appearance of the light and heavy rare earth precipitates obtained with this preparatory procedure. The precipitates prepared using procedures B and C were consistent with those of well crystallized solids, and generally conformed to the standard spectra expected for pure REE phosphates with the rhabdophane structure. A typical s pectrum, that o f Eurhabdophane, is shown in Fig. 4.2 4. 3. 3 Solubility analysis All solubility analyses were conducted as dissolution experiments Each experiment was conducted in 0.1 mol dm 3 perchloric acid. One half gram of each REE (and Y) phosphate solid moles) was placed in a series of 150 cm3 plastic bottles with 100 cm3 of perchloric acid. The bottles were stoppered and shaken at 25 1 C for periods up to five months. A minimum of two days was allowed for sample/solution equilibration Each solution was sampled by passing approximately 5 cm3 of each sample through 0 2 J..Lm Millipore filters The phosphate concentration of each filtrate (note that total phosphate equals total metal) was determined spectrophotometrically. Either 0.5 or 1 0 cm3 of each 57

PAGE 72

(A) : E ur o pium Phospha t e Sample 9 0 80 .::70 60 ... 5 50 ... ;> g 4 0 3 0 2 0 1 0 OS 10 1 5 2 0 (B): EuP04 1:1 ) : OS 1 0 1 5 20 (C): EuP04.H20 1:] : I 05 10 15 20 u 2S 30 35 40 IJ : u J I ; 25 30 35 40 : I I : 25 30 35 40 An g l e (2 9 ) Figure 4.2 Characteristic Spectra of EuP04 Rhabdophane. (a) Europium phosphate sample from procedure B. (b) XRD spectrum for EuP04 (c) XRD spectrum for EuP04 H20 58

PAGE 73

filtrate was diluted with deionized water to provide 500 cm3 of total solution. One hundred cm3 of this solution was combined with 1 cm3 of ascorbic acid (20 gm/1 00 cm3 H 20) and then I cm3 of mixed reagent containing ammonium molybdate, sulfuric acid and potassium antimonyl tartarate (Grasshoff et al., 1983). Absorbances of the resulting molybdenum blue phosphate complex were measured at 880 nm on a Varian Cary 2200 spectrophotometer. Solid/ so lution solubility equilibration was very rapid in our experiments. The dis so lved phosphate concentrations measured after elapsed dissolution times as short as 2 4 hours were generally greater than 50% of the final phosphate concentrations ( pro cedure B ) o bserved for dissolution periods between 30 days and 100 days This indicates that dissolution periods of two weeks are more than s ufficient to obtain total phosphate concentrations within 10% of the final equilibrium values. Two factors were responsible for this rapid equilibration. Our solid/solution mass ratios were very large (5x10'3 ) and our solutions were vigorously agitated throughout our experiments. Much longer equilibration periods (Firsching and Brune 1991) are required when solutions are stirred only occasionally. Phy sica l rather than chemical factors appear to exert dominant controls on rare earth element phosphate dissolution kinetics at low pH (pH:=1). The s low changes (bot h increases and decreases) in phosphate concentrations observed subsequent to 24 to 48 hours of equilibration are likely attributable to slow changes in rare earth phosphate surface chemistry (crystallinity and hydration state ) through time The experiments of Firsching and Brune ( 1991 ), in which both total phosphate and total metal concentrations were measured, indicated that REEP04 dissolution was congruent (total dissolved metal= total dissolved phosphate) under the low pH conditions (pH<1.2) in their work. Calculated and measured pH in our work agreed to within .02 units. The solution pH in our experiments never exceeded 1 03 and, following the 59

PAGE 74

observation s of Firsching and Brune ( 1991 ), congruent dis s olution wa s assumed in our solubility product calculations 4. 4 Solubility Product Calculations Solution concentr a tions of total phosphate in each solid/ s olution equilibrium are given a s the following summation : [P01 J r = [H3P04 ] + [H2PO; ] + [H Po;-]+ [Po;-]+ [MH2Po;+] + [MHPO ; ] (1) where brackets ([ ]) denote the concentration of each indicated specie s Solution concentration s of total Y and REEs ([M3+h) are given as (2) Under the conditions of our experiments the sum concentr a tion of HPol-and POt is small % ) compared toH3P04 and H2PO;, and H3P04 con s titutes a pproximately 90 % of the total phosphate. The concentration s of MH2Po; + MHPo; and are small compared to the concentration of uncomplexed metal: [M3+ ] I [M3+Jr ;?: 0.95 and I [M3+] 10 9 U s ing the following equilibrium relation s hips appropriate at 25 C and 0.1 mol dm 3 ionic s trength (Smith and Martell 1976) and K1 = [HPo;-][P01-r'[H+ r K 2 = [H2PO;][HPO i r '[H+r' K 3 = [H3P04 ][H2PO;r'[H+r' 60

PAGE 75

the concentrations of P01 and M3+ in each e x perimental solution are given as: and [P01 ] = [P01 -Jr-[M3+](HP o f3JHPOi-] +H2Po.f3I [H2PO;]) 1 + K1[H+]+ K1K2[H+f + K1K2K3[H+ ]3 (3) (4) Eqns. (3) and(4) were iteratively solved for the concentrations of uncomplexed P01and M3+ using the equilibrium constants (25 C) given in Table 4.1 (Byrne et al., 1991) Subsequently, each metal phosphate solubility product was calculated according to: (5) Under the conditions of our experiments ([M3+]/[M3+]r 0.95 and [H3P04 ] I [P01 Jr=0.90) it should be noted that, for Y and all rare earth elements, eqn. (5) can be well approximated as : influence of alternative choices of phosphate protonation constants on our calculated solubility products Using, for example, the s elected results given in Martell and Smith (1982) at 25 C and 1=0.1 mol dm-3 ( log(K1K2K3)=20.47), calculated logK_ f/M) values would be larger (more positive) by 0 25 for all elements. Solubility products ( Ksp (M)) measured at 25C and 0.1 mol dm-3 were converted to re s ults appropriate at zero ionic strength ( using the following relationship: (6) where r M)+ and r Po]are free ion activity coefficients of M3+ and Po:-. Following the works of Byrne et al. ( 1988, 1991) and Millero and Schreiber ( 1982) these activity 61

PAGE 76

Table 4 1 MHPo; a nd MH2PO i+ Format i on Constants (25C, 1 = 0.1 mola l ) REE y 4. 7 92 2.136 La 4 107 1.787 C e 4 .321 1.924 Pr 4.450 1.998 Nd 4.542 2.049 Pm 4.631 2. 1 03 Sm 4 .7 1 9 2 1 56 Eu 4. 7 75 2 .191 Gd 4.729 2 134 Tb 4 .792 2.136 Dy 4.825 2.126 Ho 4.841 2 1 1 9 Er 4 .881 2 132 Tm 4 927 2 142 Yb 4.993 2 183 Lu 5.002 2 174 p h osphoric acid protonatio n const ants l og K1 = 11. 887 log K2 = 6 892 l og K3 = 1.943 62

PAGE 77

COefficientS are estimated as y M l + = 0.134 and y POt = 0 189 at 0 1 mol dm'3 ionic strength, whereupon log is expressed as: = logK,/M)1.597 (7) 4. 5 Results of YREE Solubility Products The MP04 solubility products (log (M)) obtained for fresh phosphate precipitates (procedure A) are shown graphically in Fig 4.3. The general pattern of solubilities seen in this experiment is : (a) a sharp decrease in between Y and Ce, (b) a slow increase in with increasing atomic number between Ce and Lu and (c) anomalies at Gd and Er. The observed decreases in (Er) through time suggest that this anomaly might have disappeared over a longer period of time The light rare earth precipitates used in this experiment were initially x-ray amorphous and, since no further XRD analysis were conducted to establish precipitate crystallinity solubility product results obtained using procedure A cannot be quantitatively compared with the results obtained in procedures B and C. It should be noted that in contrast to the results obtained using procedures B and C, the phosphate concentrations observed through time in procedure A rose to a maximum at 24 hours and at day 30, had declined to lower values. We attribute the high initial (24 hours) values to the enhanced solubility of the finely dispersed precipitates produced by procedure A. The declining phosphate concentrations in this experiment (subsequent to day 1) indicate that whereas the experiment began as a dis s olution experiment, after 24 hours the experiment proceeded a s a precipitation experiment. For this experiment the final solution concentrations of the heavy rare earths at day 150 came into good agreement with the results obtained using procedure s B and C. For 63

PAGE 78

the light rare earths, which initially showed very poor crystallinity, procedure A solubilities at day 150 (Fig. 4.4) were uniformly larger than the results obtained using coarse crystalline (procedure B and C) materials. The MP04 solubility data obtained using the well crystallized solids formed in procedure B are shown in Table 4.2A. Solubility products derived from these data at days 30 and 100 are shown in Table 4.3. For all elements except Lu very consistent trends in solubility were seen over the final two months of solid/solution equilibration. The general pattern of solubility products observed after 100 days of equilibration (Fig. 4.4) is: (a) a sharp decrease in between Y and Pr, (b) a slow increase in with increasing atomic number between Pr and Lu. (c) an absence of the positive anomalies at Gd and Er seen in Fig. 4.3. MP04 solubility data obtained using the procedure of Firsching and Brune ( 1991) (procedure C) are given in Table 4.2B. Solubility products derived from these data are shown in Table 4.4. The behavior observed for these well crystallized solids was quite consistent over 30 days of equilibration. The final (day 30) pattern of solubilities observed (Fig 4.4) using procedure C is: (a) a sharp decrease in between Y and Ce, (b) nearly constant between Ce and Sm, and (c) generally increasing between Sm and Lu. The solubility products obtained (Fig. 4.4) for the final set of observations in each experimental series are consistent in indicating that decreases sharply from Y to La to Ce. All three curves show generally increasing behavior between Sm and Lu. Between Ce and Sm, curve A exhibits a minimum at Ce, curve B shows a minimum at Pr and curve C shows nearly constant behavior between Ce and Sm. 64

PAGE 79

-24. 0 -24. 5 -25.0 ee-:f -25.5 2 6 0 26.5 --<>--day 1 ---+-day 30 ----+--day 150 Y La Ce Pr Nd P m Sm Eu Gd Tb D y Ho Er Tm Yb Lu REE Figure 4.3 Rare Earth and Yttrium Solubility Product s at 25C and Zero Ionic Strength (Procedure A) 65

PAGE 80

Table 4.2 a Procedure B data Total phosphate (M) [H+ ] Ionic strength Day 1 Day 15 Day30 Day 100 Day 1 D a y 15 Day30 Day 100 D a y 1 D a y 15 Day30 Day 100 0 001409 0 .001445 0 .001482 0 001554 0 0968 0 .0967 0.0966 0.0963 0 .1052 0 1033 0.1033 0 1032 0.000434 0 .000574 0 000607 0.000672 0 0997 0 .0993 0.0992 0.0990 0 1023 0 1027 0 1028 0 1030 0 000211 0.000238 0.000260 0 000297 0 1004 0.1003 0 1002 0.1001 0.1016 0.1017 0 1018 0 1019 0 .000036 0 .000165 0 .000188 0.000210 0 1009 0.1005 0 1004 0 1004 0 .1011 0.1015 0 1016 0.1016 0.000260 0 .000323 0.000354 0 .000389 0 1002 0 1000 0.0999 0.0998 0 1018 0 1020 0 1021 0 1022 0.000217 0.000330 0 000347 0.000388 0.1 004 0.1000 0.1000 0.0998 0.1017 0.1020 0.1020 0.1022 0 .000379 0.000453 0 000487 0.000521 0 .09 9 9 0 .0 9 97 0 0996 0 .0994 0 .1021 0.1024 0 1025 0 1026 0 000443 0 000654 0 000738 0.000841 0 0997 0 .0990 0 0988 0 0985 0 1023 0.1030 0 1032 0 1035 0.000713 0 000 8 53 0 000918 0 000993 0 0989 0.0985 0 0983 0 0980 0 .1031 0 1036 0 1038 0.1040 0 000769 0 000937 0.000992 0 001040 0 0987 0 0982 0.0980 0 .0979 0 1033 0 1038 0.1040 0.1041 0.000905 0 .001073 0.001118 0 .001180 0 .0983 0 .0978 0 0977 0 .0975 0.1037 0 104 2 0 1 044 0 1045 0.000879 0 000998 0 001053 0.001116 0.0984 0 0980 0 .0979 0 .0977 0.1036 0 1040 0 1042 0.1044 0.001100 0.001110 0.001154 0 001215 0 0977 0 .0977 0.0975 0.0974 0.1043 0 1043 0.1045 0.1047 0.001329 0 001211 0 001229 0.001275 0.0970 0.0974 0 0973 0.0972 0.1050 0.1046 0.1047 0 1048 0.001938 0 .002218 0 .002068 0.001366 0 .0952 0 .0944 0.0948 0 0969 0.1068 0 1077 0 1072 0 1051

PAGE 81

Table 4.2b Procedure C Data Total phosphate M [H+] Ionic strength Day2 Day 30 Day2 Day 30 Day2 Day 30 y 0 001321 0.001295 0.0970 0.0971 0.1035 0.1049 La 0 000715 0.000597 0.0989 0 0992 0.1031 0 1028 Ce 0.000477 0.000416 0 0996 0.0998 0.1024 0 1023 Pr 0.000465 0.000411 0 0996 0 0998 0.1024 0.1022 Nd 0.000449 0.000378 0.0997 0.0999 0.1024 0.1021 Pm Sm 0 000461 0.000388 0.0996 0.0998 0.1024 0.1022 Eu 0.000571 0.000482 0.0993 0.0996 0.1027 0.1025 Gd 0 000848 0.000648 0.0985 0 0991 0.1035 0 1029 Tb 0 001132 0 000913 0 0976 0 0983 0.1044 0 10 37 Dy 0.001509 0.001374 0.0965 0.0969 0 1055 0.1051 Ho 0 001651 0.001544 0.0961 0.0964 0.1060 0 1056 Er 0 001607 0.001429 0 0962 0.0967 0 105 8 0.1053 Tm 0.001786 0.001714 0.0957 0.0959 0 1064 0 .1061 Yb 0.002357 0.002304 0 0939 0 0941 0.1081 0 1079 Lu 0.002545 0.002626 0 0934 0.0931 0.1086 0.1089 67

PAGE 82

Comparison of the absolute magnitudes of values rather than trends, reveals the following features: For all three experimental procedures (M) values for Y, Tb, Dy, Ho and Tm are in good agreement. For procedures Band C, the values of Y, La, Nd, Sm, Eu and Tb are in very close accord. For procedures A and B, values for the rare earths between Tb and Lu generally agree within a factor of two. The largest overall discrepancy in the three series of values (Fig. 4.4) is seen in procedure A values between La and Gd. It appears that the light rare earths in procedure A never attained the crystallinity of procedure B and C solids. In contrast, in both physical appearance and XRD characteristics, the procedure A heavy rare earths were comparatively crystalline relative to the light rare earths immediately after formation. This appears to be responsible for the generally good procedure A vs. procedure B agreement for the rare earths between Tb and Lu. In view of the generally good agreement be tween Tb, Dy and Ho for all preparatory procedures, as well as very good agreement for (Y) values using the three preparatory procedures, it is pos sible to speculate that the tendency of these elements to readily form crystalline solids is related to the particular ionic size of these cations In order to provide recommended values for the of well crystallized solids we have averaged the solubility results obtained for the procedure B (day 100) and C (day 30) solids. The uncertainties given with the log estimates in Table 4.5 and Fig. 4 .5 are equal to half the difference between our procedure B ( day 1 00) and procedure C ( day 30) log estimates. Comparison of the Fig. 4.5 results with the results shown in Fig. 4.1 indicates that both sets of results are broadly consistent and, as well, are consistent in a number of important details: The work of Byrne and Kim (1993) showed a trend of Ce, 68

PAGE 83

-24 0 -24.5 -25.0 oe-25.5 ..2 -26 0 -26.5 -27 0 A-<>--day150 B-tday 100 c--day30 Y La Ce Pr Nd Pm S m Eu Gd Tb Dy Ho Er Tm Yb Lu REE Figure 4.4 Comparison of Rare Earth and Yttrium Solubility Products (Procedures A, B and C) 69

PAGE 84

Table 4.3 Solubility Products (log (M)) for Procedure B Precipitates. (Data are corrected to zero ionic strength) REE day 30 day 100 y -24.98 -24.94 La -25.79 -25.70 Ce -26 54 -26.42 Pr -26 82 -26.73 Nd -26 .2 6 -26 .18 Prn Srn -26 .28 -26.18 Eu -25 98 -25 .92 Gd -25.61 -25.50 1b -25.42 -25 .35 Dy -25.35 -25.30 Ho -25 .24 -25.19 Er -25 .29 -25.24 Trn -25.21 -25.16 Yb -25.15 -25.12 Lu -24 .67 -25.05 70

PAGE 85

Table 4.4 So lu bi l ity Products ( log (M)) for Procedure C Precipitates. (Data are corrected to zero ionic strength) REE day 2 day 30 y -25.07 -25.10 La -25.64 -25.80 Ce -26.00 -26. 12 Pr -26.02 -26 13 Nd -26.06 -26 .21 Pm Sm -26.03 -26.19 Eu -25.84 -25 99 Gd -25.49 25.73 Tb -25 .23 -25.42 Dy -24 .96 -25 05 Ho -24.88 -24.94 Er -24.90 -25.0 1 Tm -24.81 -24 .90 Yb -24.57 -24.65 Lu -24.48 -24.44 71

PAGE 86

-24.5 -25. 0 1l -25. 5 -26. 0 -26.5 Y La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb L u REE Figure 4.5 Recommended Solubility Products at 25 C and Zero Ionic Strength Based on Procedure B and C Mea su r eme nts 72

PAGE 87

Eu, Gd and Yb values consistent with the increasing trend for these elements in Fig. 4.5. The relatively high values obtained by Byrne and Kim (1993) are regarded as upper bound values (Byrne and Kim, 1993) becau se they were obtained through an ongoing precipitation process with concomitant formation of fresh (poorly crystalline and finely divided) solids. The results of Jonassen et al. (1985) are in good general agreement with the magnitudes of and shown in Fig. 4.5. Similarly, the results of Tananaev et al. (1963, 1967) are in general agreement with magnitude s of K.?pCLa) and shown in Fig. 4.5 Comparison of Fig. 4.5 with the re s ult of Firsching and Brune ( 1991) exhibits both agreement and differences The values of Fir sc hing and Brune (1991) generally fall between -24.5 and -25. 5 while the range of log values in Fig. 4 5 is -24.7 to -26.5. Our work (Fig. 4.5) is in agreement with the decreasing trend of between Y and La reported by Fir sc hing and Brune (1991). However, the trend of REE values obtained by Firsc hing and Brun e ( 1991 ) contrasts s harply with the initially decreasing trend betw ee n La and the elements between Ce and Sm obtained in our study, as well as the trend of increasing values between Sm and Lu. Examination of the primary solubility data reported in the Firsching and Brune ( 1991) study demonstrates subs tantial uncertaintie s for a number of elements. These reported uncertainties (e.g. 24.78.73) render some of the log trends reported in their work ambiguous. The Y and REE pho s phate solubility product behavior which emerges from this work can be summarized as follows: minimum solubilities are observed in th e neighborhood of Ce and Pr. La so lubilities are s ubstantially larger than those of the heavier near n eighbo r s of La, and the so lubility of LaP04 i s s ub s tantially s maller th an that of YP04 Solubility product s increase for pho sp hates heavier than SmP04 with little 73

PAGE 88

Table 4.5 Recommended log (M) V a lues at 2SOC and Zero Ionic Strength REE log K (M) y -25 0 2 0.08 La -25 7 5 0.05 Ce -26 2 7 0 .15 Pr 26.4 3 0.30 Nd -26 .20 0.02 Pm Sm 26.1 9 0 .01 Eu 25 9 6 0 03 Gd -25.62 0.12 Tb -25 3 9 0.04 Dy -25 1 8 0.13 Ho -25 07 0 .13 Er -25 1 3.11 Tm -25 0 3 0.13 Yb -24 8 9 0 .24 Lu -24 7 5 0.31 74

PAGE 89

consistent evidence for anomalies. For all precipitates (procedures A, B and C) the YP04 solubility product appears to assume a position among REEP04 solubility products generally consistent with the similarity in Y and Ho ionic radii. 4. 6 Phosphate Solubility Product and Coprecipitation Behavior The works of Byrne and Kim (1993) and Byrne et al. (1996) demonstrate that the formation of Y and REE phosphates in the natural environment does not result in discrete phosphate phases for each element. In the natural environment phosphate precipitation involves the formation of phases which incorporate Y and each rare earth (coprecipitation) The fractionation patterns developed in the course of coprecipitation processes can be described (Glynn and Reardon, 1990) by comparing the solubility products of different metals, Mi and Mi aM,apo]aMJaPo](8) are pure endmember solubility product s and the solid phase activities of M ;P04 and M jP04 are denoted as aM,Po. and aM1Po Since the activity coefficients of rare earths in aqueous solution are very similar, eqn. (8) can be rewritten in term s of solution phase concentrations ([M;] and [M)) and solid phase mole fractions XM, and XM1 [M(+]-[MJ+] Y M1P a4 (9) Eqn. (9) indicates that coprecipitation processes will produce rare earth element fractionations (i.e. difference s in relative solid phase vs solution phase concentrations) 75

PAGE 90

unless ( M ) YM Po equals K0 ( M1.) YM PO Furthermore if Y M Po equals Y M .po the sp r 1 4 s p 1 4 4 1 4 metal fractionations produced during coprecipitation processes will closely resemble the pattern of rare earth and yttrium solubility products. Comparison of Fig. 4 5 and a typical fractionation pattern (Fig. 4.6) produced in the coprecipitation study of Byrne et al. (1996) indicateS that, while Y M PO = Y M PO is a USeful general approximation, distinct differenceS I 4 j 4 in log K!(M) patterns and log[M3+] I [M3+]0 patterns are observed. As in the case of REE (and Y) solubility product behavior the comparative solution phase affinities of Y and REEs in a coprecipitation process decrease between Y and Ce and then generally increase between Sm and Lu. The principal contrast between Figs 4 5 and 4.6 is seen in the Fig. 4.6 minimum at Sm compared to the Fig. 4.5 minimum at Pr. Due to r MP04 differences among various rare earths, removal of rare earths from solution exhibits a pattern ( i.e. log[M3+ ] I [M3+]0 ) similar but not identical to the patterns observed for K!CM). Since the data shown in Fig. 4 6 were generated in a coprecipitation experiment, rather than a dissolution experiment, it should be expected that the solid phases formed are fresh and possibly amorphous. As such, it might be most reasonable to compare Fig. 4.6 coprecipitation results with the K!(M) results obtained using fresh precipitates (procedure A, Fig. 4.3 and 4.4). As was the case for comparisons of Figs 4 5 and 4.6 broad general agreement is observed in the shapes of Fig. 4.3 and Fig. 4.6. In contrast to Fig. 4 5 and Fig. 4.6 comparisons however, Fig. 4.3 and Fig. 4.6 exhibit a prominent positive fractionation anomaly at Gd, and in general (Fig. 4.4, Curve A) very similar trends among the heavy rare earths. These comparisons indicate that solid phase activity coefficients, y MPo4 are dependent on the conditions of solid phase formation, and that freshly formed precipitates can give rise to "anomalous" fractionation behavior at Gd and perhaps at Er. 76

PAGE 91

4. 7 Coprecipitation Equilibria Since REE phosphate sol ubility products range over more than an order of magnitude, it is clear that so lubility products appropriate to coprecipitates should vary widely in response to coprecipitate composition A quantitative expression for solid so lution aqueous solution equilibrium (Thorstenson and Plummer, 1977), corresponding to a given phosphate composition, can be written for complex coprecipitates in analogy with the si mple binary saturation constant described by Glynn and Reardon (1990): where XM. indicates a soli d phase mole fraction for metal M; on the copreci pitate surface, I aM;Po. i s a solid phase activity, and K!(M) are solubility products of pure rare earth element phosphates. Eqn. ( 1 0) i s complex in form and shows that the activity behavior of rare earths in solid so lution must be quantified in order to predict the compositional dependenc e of K ; (MP04). Whereas in this work we determined each of the solubility product s, K !(M), in eqn. (10), and the work of Byrne et al. (1996) ex amined the process of REE phosphate coprecipitation, elucidation of the role of solid-solution composition o n the term { a:;o. . . .. } will require direct observation of cop recipitate s u rfaces in aqueou s solution/solid solution equi librium In view of the s ub s t a ntial complexity of soli d solution-aqueous solution equilibria involving REE pho sp hate cop recipitates, Byrne and Kim (1993) approximated eqn. (10) by neglecting the role of the HREEs on the aqueous solution-solid so lution saturation state. Such an approximation follows reasonably from the comparative l y low abundances and s tron g compl exat ion of the HREEs in seawater. In view of the summary sol ubility behavior s hown in Fig. 4 .5, as well as th e comparativ e abundances of the REEs in natural a qu eous solutions (e.g. seawa ter), a reasonable approximation of eqn. ( 10) can be written as 77

PAGE 92

0 00 0 .2 5 0 ......... + '+"' -0 5 0 .., ........ .._, 0. 75 -1.0 0 Y La C e Pr N1 Pm Sm E u Gl 1b D y H o Er Tm Yb Lu Figure 4 6 Coprecipitation Behavior of Rare Earth and Yttrium Phosphates (Byrne e t a l ., 1996). 7 8

PAGE 93

= {( tc ( ( ( rsm }{ a::?o. (11) Since forCe, Pr, Nd and Sm (Fig 4 5 and Table 4 5 ), we can then approximate as K o (MPO ) Io-26.3{aXc, aXp, axNd axSm } sp 4 -CeP04 PrP04 Nd P 04 SmP04 (12) Additional si mplification of this expressi o n requires further understanding of the activ i ty coefficient behavior of REEP04 solid solutions. However, assuming in the simplest case that deviation from ideality is not large compared to uncertainties in individual solubility products, can be written simply in term of solid phase (surface) composition as follows Ko (MPO ) = w -26.3{Xxc. xxp, xxNd xxSm} sp 4 Ce Pr Nd Sm (13) Since, at equilibrium, this solubility product s hould be equal to the io n activity product (lAP), at equilibrium K;,(MP04 ) in eqn. (13) can also be written (Byrne et a l 1996) as (14) where it is assumed that Ce3+, Pr3+, Nd3+ and Sm3+ a queous activity coefficient s are approximately equal. Combining eqns. (13) and (14) the s aturation s tate (.Q) of natural solutions with re s pect to coarse, well aged REEP04 coprecipitates can be approximated as Thi s eqn. further illustrates th a t knowledge of the surface chemical composition of pho s ph a te coprecipitates is a key to modeling the s olubility behavior of the s e s olids. 79

PAGE 94

4. 8 Conclusions Our solubility product analysis for well crystallized solids produced a reasonably consistent trend of values for Y through Ce and Sm through Lu. A minimum in REE so lubilit y products occurs among the light rare earths in the vicinity of Pr. The position of this minimum can vary depending on the conditions of solid phase formation, and thereby, the so lid phase chemistry of REE phosphates. The behavior of Y and REE phosphates provides a general basis for the fractionation patterns developed during REE and Y coprecipitation. A detailed comparison of results and the fractionation patterns log[M3+] I [M3+ ]0 developed during REE coprecipitation indicates that activity coefficient ( y MPo. ) variations in the solid phase exert significant controls on Y and REE coprecipitation behavior. In view of the substantial difference s in solubility behavior developed through different so lid phase form atio n procedures, our results s ugge s t that solubi lity analysis in conjunction with detailed surface c h e mical analysis (e.g. EXAFS, rather than bulk chemical XRD analysis), might further elucidate the role of s urface chemical transformations in the complex and coprecipitation behavior of Y and REEs. 80

PAGE 95

5. THE INFLUENCE OF PHOSPHATE COPRECIPITATION ON RARE EARTH DISTRIBUTIONS IN NATURAL WATERS 5.1 Abstract Although the rare earth elements (REEs) are thought to participate in essentially all of the processes that influence metal distributions in seawater, quantitative descriptions of ocea nic rare earth element (REE) distributions have nearly exclusively involved removal of these elements from solution via adsorption on settling particles Laboratory investigations involving yttrium (Y) and all REEs, except Pm, demonstrate that phosphate coprecipitation of Y and the REEs creates solution concentration patterns characteristic of input-normalized Y and REE concentration s in seawater. Remo va l of the REEs and Y from solution as pho s phate coprecipitates depletes REEs in so lut ion relative to Y, depletes solutions in Sm relative to heavier and lighter elements (relative depletions of the middle REEs), and produces a variety of near-neighbor concentration anomalies (e.g ., Eu-Gd -Tb, Dy-Er-Yb a nd ErYb-Lu) which have been reported in high precision analyses of open ocean waters. 81

PAGE 96

5. 2 Introduction The solubility products of pure REE phosphates are sufficiently small ( o 24.4) (Jonasson et al., 1985) that even picomolar concentrations of REEs can produce saturation and supersaturation. Since the saw-tooth pattern of REE abundances in natural so lution s (Piper, 1974; de Baar et al., 1991) contrasts sharply with the relatively smooth pattern of REE phosphate solubility products (Jonasson et al., 1985 ; Firsching and Brune 1991 ), it follows that pure REE phosphate phases (e.g LaP04 NdP04 etc.) do not form in natural solutions (Byrne and Kim, 1993) Instead, due to close similarities in ionic radii and chemical properties, the entire suite of REEs in the presence of phosphate ions (Pot) should f orm coprecipitates of variable composition (Byrne and Kim, 1993) The process of coprecipitation should preserve the observed sawtooth pattern of REE abundances in natural solutions while reducing overall REE solution concentrations and altering relative co ncentrations. Previous investigation of REE coprecipitation (Byrne and Kim 1993) involved only four elements (Ce, Eu, Gd, Yb), and does not provide a satisf ac tory description of the REE fractionation behavior which accompanies coprecipitation Over the past decade, Thermal Ionization Mass Spectrometry (TIMS) has produced very precise characterizations of the comparative conce ntr ations of ten REEs in seawater. While producing less precise comparative concentration data than TIMS (Klinkhammer et al., 1994), Inductively Coupled Plasma Mass Spectrometry (ICP-MS) allows determinat ion of all fourteen naturally occurring REEs plus yttrium (Y). ICP-MS analyses are thereby especially well suited to laboratory examinations of comparative REE chemical behavior, and have been 82

PAGE 97

used in this work to determine the REE elemental fractionations which accompany precipitation with Po;. The comparative behaviors of metals in coprecipitation processes can be described in terms of proportionality factors ( A;i) between solution concentrations, [M], and relative concentrations, N of metals on coprecipitate surfaces (Doerner and Hoskins 1925): (1) It then follows (Doerner and Hoskins 1925; Byrne and Kim, 1993) that metal ion solution concentrations during coprecipitation exhibit the following relationship: (2) where [0M;] and [0M)denote solution concentrations prior to coprecipitate formation. The form of eqn (2) indicates that unles s A,ii = 1 elements will progressively fractionate during coprecipitation 5. 3 Analytical Procedures Coprecipitation experiments were performed using Y and all REEs, except Pm The initial solutions were sufficiently acidic to be undersaturated with respect to phosphate coprecipitation. The initial total phosphate concentration in our experiments was 2xl0 4 molal and the ionic strength was adjusted to 0 .01 molal with NaN03 Two types of experiments were performed In the first case all initial Y and REE concentrations were identical, [0 M ;3+] = 100 ppb and M ?+] = 1. Ox 1 o-s molal. In the second type of experiment initial total REE concentrations were in proportions identical to those of Y and REEs in mean shale (Piper, 1974) TheY concentration in mean shale was taken as 30 ppm 83

PAGE 98

and L)0M(+]= 8.6xl0 6 molal. REE standard solutions were made up from individual ICP REE sta ndard s (SPEX Chemical, Metuchen NJ). Experimental so lutions were contained in thermo s tated (2 5.0.1 C) Teflon bottles Coprecipitation was induced by incre asing solution pH through addition of dilute NaHC03 After each adjustment, the pH of o ur so lution s wa s generally stable to 0 .01 pH units Solution pH was measured with a glass electrode calibrated on the free hydrogen ion concentration sca le. Sub seq uent to filtration u s ing 0 .22 11m acetate cartridge filters (MSI, W es tborough, MA), so lution concentrations of Y and the REEs were measured with a Fisons PQS ICP mas s spectrometer. 5. 4 Results and Discussion Fig 5.1 s h ows th e re su lt s of ph os phat e coprecipitation involving fourteen REE s an d Y. At the beginning of thi s experiment all metal ions in so luti on h ad approximately identical concentrations. The pattern s s hown i n Fig 5.1, where th e results are expressed as a fraction of the initial so lution concentration of each element exhibit the following features: Sm a nd Eu are strongly deplet e d in sol ut io n r e lative t o their heavier and lighter neighbors Yttrium i s enric hed rel ative to all REES Concentration anomalie s are consistently observed for a variety of e lements r e lative to their immediate neighb ors; Gd and Er are enriched, and Yb is d ep leted relative t o n e ighboring elements. In order to eva luat e the poss ible influenc e of sol ution composi tion on fractionation factors ( A.;j) for Y and REE coprecipitation a dditional coprecipitation experime nt s (Fi g. 5.2) were co nducted in which the initial concentrations of all e lement s were in the sa m e 84

PAGE 99

0 5 0 -1.00 '<:::: ::; -1.50 E -2 .00 4 7 h ours 48 hours 49 h o urs 2 5 0 ---'N d-P-'-m---'Sm_Eu,__ _Gd.___Tb,L----0'y _...H o _E_._r --,1T m-Y-'-b L'-u ---L.___l Figure 5.1 Yttrium and Rare Earth Element Fractionation for a 49-Hour Phosphate Coprecipitation Experiment at 25 oC. Fractionation is depicted as log [M3+] I [Mg+], where [Mg + ] is the initial concentration of each element. The initial solution pH 3 9 (open circles) was increased to 4.4 (solid circles) to increase the rate and extent of coprecipitation. 85

PAGE 100

.25 7 5 9 hrs 27 h r s I.OO L._.,..a..y -LLa -C:f-c---::' Pr-NLd -:P:c"'-m--::-Smc.......,E:f-u --::'Gd:---::lb:'--D='-y__,H'-o Figure 5.2 Yttrium and Rare Earth Element Fractionation for a 24-Hour Coprecipit a tion Experiment at pH = 3.8. With minor exceptions the experimental conditions in this experiment are identical to those used to obtain our Fig 5 1 results. [Mg+] are initial relative REE concentrations in this experiment were in proportions identical to those of mean shale. Y and REE patterns were obtained at elapsed times of 9, 12, 22, 24 and 27 hours The 27-hour coprecipitation results in this work have been used to calculate the fractionation data given in Table 5 1 86

PAGE 101

proportions as those observed in mean shale. Comparison of Figs. 5.1 and 5.2 demonstrates that with the possible exception of Y, fractionation factors are relatively weakly affected by variations in total metal ion concentration ratios. While it i s clear that adsorptive removal of Y and the REEs is an important process in seawa ter, our results indicate that phosphate coprecipitation can also contribute to the disti nctive pattern of input-normalized REE concentrations in the oceans. The patterns shown in Fig. 5.1 are consistent with recent high precision TIMS analyses of open ocean REE distributions (Piepgras and Jacobsen 1992 ; German et al., 1995) and ICP-MS observations of comparative Y and REE behavior. REE pho s phate coprecipitation ( Fig 5 .1) will contribute to (a) observed enrichments of La (Piepgras and Jacobsen 1992, Figs 3 and 10; German et al., 1995, Figs 5, 8 and 9) relative to its immediate neighbor s, ( b) an observed minimum in input-normalized REE concentration s near Sm ( Piep g r as and Jacobsen, 1992, Fig. 10; German et al., 1995, Figs. 5, 8 and 9), (c) Y enrichments ( Sh a bani et al., 1990 ; Zhang et al., 1994) relative to the suite of REE s and (d) concentration anomalies for Gd, Er, and Yb (Piepgras and Jacobsen, 1992, Fig. 10 ; German et al., 1995, F i gs. 5 8 and 9). When comparing our Figs 5.1 and 5.2 results with observed oceanic elemental distributions, it is important to recognize that our experiments depict coprecipitation behavior of Y and the REEs at very low degrees of solution complexation. Under the conditions of experiment 1 (Fig. 5.1, pH 4.4) the fraction of free (uncomplexed) metal, expressed as [M3+]/[M3+]T, was larger than 0.70 for all REEs and Y. In experiment 2 (Fig. 5.2 pH 3.8) the fraction of uncomplexed metal [M3+]/[M3+ h was greater than 0 .84 for all metals. Fig 5.3 provides solution speciation diagrams for Y and the REEs at the highest (pH= 4.4) and lowest (pH= 3.8) degrees of solution complexation in our experiments. In contrast to Fig 5 3 where [M3+]/[M3+h varies between 0.94 and 0.70, m seawater (25C, S = 35) values of [M3+]/[M3+]T typically range (Byrne et al., 1988) from 87

PAGE 102

1.00 0 .80 0.00 f ... 0 40 0.20 0 .60 .... pH3.8 _,,1'04 _,.,, __ ,.,. Figure 5 3 Yttrium and Rare Earth Element Speciation in Experimental Solutions at pH 4.4 (A, cf. Fig. 5 .1) and pH 3.8 (B, cf. Fig 5 .2) (25c, 1=0 .01 molal). Nitrate stability constants for these calculations were obtained from Wood (1990), following the work of Choppin and Straizik ( 1965) Phosphate stability constants were obtained from Byrne et al. (1991). Stability constants were estimated for Y and all REEs using the linear free-energy relationships of Lee and Byrne (1992) and Byrne and Lee (1993) The carbonate and bicarbonate stability constants of Lee and Byrne (1993), in conjunction with the low pH and low concentrations of total in our experiments, indicate that carbonat e and bicarbonate complexation were not significant. 88

PAGE 103

0.1 (La) to 0.01 (Lu) and even lower. Solution complexation increases the affinity of metal s for the so lution phase Thus, in the coprecipitation experiments of Byrne and Kim (1993) fractionation patterns in the presence of carbonate complexation strongly favored the retention of Yb in solution compared to Ce. The experiments of Byrne and Kim (1993) indicate that a high degree of carbonate complexation will alter the Fig. 5.1 and 5 .2 results in a manner which enriches heavy REEs (HREEs) in the solution phase and light REEs (LREEs) in the solid phase Carbonate complexation in seaw ater will alter the s hape s of the curves in Figs. 5.1 and 5 .2 by progressively increa s ing log[M3+ ]/[M3+ ]0 values for HREEs relative to the LREEs The experiments of Byrne and Kim ( 1993) indicated the degree of carbonate complexation in seawater can increase log[M3+]/[M3+]0 values for Yb by more than a factor of two relative to Ce. A high de gree of carbonate complexation will s trongly favor HREE enrichment s relative to the LREEs for both ad sorp tive scavenging (Byrne and Kim, 1990) and phosphate coprecipitation A potentially dramatic chan ge in the fractionation p atterns shown in Figs. 5 1 and 5.2 sho uld be o bs e rved between Y and the LREES. Since the so lution complexation of Y is s imilar to that of Tb (Liu and Byrne, 1995), durin g phosphate coprecipitation under a high degree of carbonate complexation Y s hould be enriched in solution, relative t o the LREES t o a s ub s tantially greater extent than see n in Figs. 5.1 a nd 5.2. Due to th e rel a tively weak comp le xa tion of the LREES, it sho uld be expected that complexation will not have a lar ge effect on comparative fractionations within the LREEs (e.g LaSm) Nev e rthel ess, th e overall effect of progre ss ively s tronger so lution complexation with increasing REE atomic number s hould, to some ext en t decr ease fractionation between La a nd Sm. Since changes in solution comple xat ion constants are relatively s mall for adjacent elements, the fractionation anomalies see n at Er and Yb s hould be weakly affected by complexation However, s ince Gd complexation con stants are as a rule con s i s tently smaller th an those of E u and Tb, the effect of solu ti on 89

PAGE 104

complexation on Gd and its neighbors will reduce the magnitude of the positive fractionation anomaly at Gd seen in Figs. 5 1 and 5.2. The results shown in Figs. 5.1 and 5.2 depict the general nature of Y and REE coprecipitation processes. In addition to assessments of active coprecipitation (Figs. 5.1 and 5.2) it is interesting to consider the nature of the eventual equilibrium that will result from such processes. Although it is unlikely that equilibrium was achieved in any of our experiments it is probable that the degree of supersaturation in the final (27 hour) result shown in Fig. 5.2 is the smalle s t attained in any of our experiments. The data obtained in this case can be used to assess the general character of the solution/coprecipitate equilibrium. Table 5.1 provides log[M3+ ] /[0M3+] results obtained at 27 hours in experiment 2. Column 5 of Table 5.1 exhibits the elemental fractionation factors ().ij) which result when Y is used as a reference element ([MJ+] = [Y3+]). Using the fractionation factors shown in Table 5.1, eqn. (1) can be used to calculate the relative concentrations ( N;) of Y and REEs on coprecipitate surfaces. Subsequently, these relative s urface concentrations can be used to calculate metal ion mole fractions ( XM. ) on the I coprecipitate surface: (3) The results of this calculation, appropriate to the 27-hour coprecipitate of Fig. 5.2, are given in column 6 of Table 5.1. These mole fractions correspond to the mole fractions in the following equilibrium: X Y3+ +X La3+ +X Ce3+ ++X Lu3+ + P03 = MPO Y La Ct Lu 4 4 (4) The solubility product for equilibrium (4) is then estimated as and 90

PAGE 105

(6) The [M?+], XM [Po;] results given in Table 5.1 have been used to calculate I logKs/MP04 ) via eqn. (6) Since [M;3+ ] = [M?+lr in this experiment, we have not accounted for the small effect of solution complexation on our log Ksr (MP04 ) calculation The logKs/MP04)result corresponding to our 27-hour coprecipitate is logKsr(MP04)<21.6. Under the conditions of our experiment (25C, 1=0 .01 molal) the activity coefficients of M?+, Po;ions are approximately 0.4, whereupon the MP04 solubility product at zero ionic strength is estimated as logKs/MP04)<-22.4. This result is high compared to the solubility products of individual pure precipitates (e.g., logKs/GdP04)=:-24). This is not unexpected since our precipitates are very fresh ( 1 day old) and thereby quite active compared to the well-aged solids used in most determinations. More importantly, our result represents an upper bound since equilibrium had not been attained in our experiments. In assessments of the Ks/MP04)of REE coprecipitates it is important to note that K s/MP04)will vary with solution phase and solid phase composition. Although fractionation patterns appear to be generally consistent during coprecipitation processes substantial differences in K s/MP04)should be observed between solids rich in the most solubl e REEs vs. those composed of the least soluble REEs As an example, the comparatively high affinity of Ce, Pr, Nd, Sm and Eu for the coprecipitate s urface results in initial surface enrichments of these elements ( large XM values). As time proceeds I solution concentrations of these elements decrease relatively rapidly resulting in progressively decreasing surface mole fractions. The fractionation behavior shown in Figs. 91

PAGE 106

Table 5.1 Fractionation Factors (A.;) and Stoichiometric Ratio Estimates for REE and Yttrium Coprecipitation with Po:Mi log[MJ0 log[MJ log[MJ / [MJ0 A.i xi y -5.779 -6.100 -0.321 1.000 1.95x10-1 La -5 .862 -6.195 -0.333 1.040 1.63x10 -1 Ce -5.552 -6.090 -0.538 1.677 3.34x10-1 Pr -6.429 -7.045 -0.616 1.922 4.24x10 -2 Nd -5.885 -6.491 -0.606 1.889 1.49x10 -1 Pm Sm -6.602 -7.423 -0.821 2.561 2 .37x 10 -2 Eu -7.275 -8.042 -0.767 2.393 5 32x10 -3 Gd -6.621 -7.250 -0.629 1.962 2 70x10 -2 Tb -7.354 -7.994 -0 640 1.997 4.96x10 -3 Dy -6.760 -7.338 0.578 1.804 2.03x10 -2 Ho -7.377 -7.848 -0.471 1.470 5.1lx10-3 Er -6.950 -7.388 -0.438 1.365 1.37x10 -2 Tm -7.720 -8 182 -0.462 1.442 2.32x10 -3 Yb -6.995 -7.502 -0.507 1 .58 2 1 22x10 -2 Lu -7.660 -8.109 -0.449 1.399 2 67x10-3 Fractionation factors are calculated u s ing Yttrium as a reference element ([M)=[Y3+]). Stoichiometric coefficients, X i (eqns. (3) and (4)) are identified as the surface mole fractions of Yttrium and REEs on the 27-hour old coprecipitate surface. 92

PAGE 107

5. 1 and 5.2 indicates that surfaces rich in Ce, Pr, Nd, Sm and Eu have lower solubility than surfaces rich in heavier and lighter elements. Thus, it is probable that, other factors being equal, K s/MP04 ) increase significantly during the course of coprecipitation when solutions and surfaces become depleted in elements which have larger surface affinities (larger llij factors). 5 5 Conclusions The stoichiometric ratios and fractionation factors ( llij) obtained in this work indicate that environmental phosphate coprecipitation processes have a signature that is distinct from the fractionation patterns expected for adsorptive removal of REEs (Byrne and Kim, 1993 ; Lee and Byrne, 1993) from weakly complexing solutions. In the absence of a high degree of solution complexation phosphate coprecipitation produces middle REE depletions, while for adsorptive scavenging under such conditions this behavior has not been observed and is not generally expected. In the presence of a high degree of solution complexation, as in seawater, phosphate coprecipitation, along with adsorptive scavenging (Byrne and Kim, 1990) will contribute to HREE enrichment s relative to the LREES and also a La enrichment relative to its heavier neighbors. As a consequence of the similar sizes of REE and trivalent actinide ions it should be expected that phosphate coprecipitation may be a significant factor in the mobility of both lanthanides and actinides in ground waters, rivers and lakes. The results obtained in this work demonstrate that conditions prerequisite to coprecipitation involve the contribution of all free trivalent ions to an overall M;P04 activity product. Conditions prerequisite to coprecipitation can be satisfied when conditions required for formation of pure phases are not met. 9 3

PAGE 108

6. COMPARATIVE COPRECIPITATION OF PHOSPHATE AND ARSENATE WITH YTTRIUM AND THE RARE EARTHS: THE INFLUENCE OF SOLUTION COMPLEXATION 6.1 Abstract Coprecipitation of yttrium (Y) and rare earth elements (REEs) with pho s phate and arsenate removes these elements from solution in variable proportions. For both phosphate and arsenate coprecipitation, middle REEs (Sm and Eu) are strongly depleted in solution r e lative to h eavie r and lighter elements. Solution comp l exation by oxalate ( Ox2 ) influences Y and REE removal patterns by strongly enhancing the retention of Y and the heaviest REEs in so lution The extent of this enhancement is well-de sc ribed by a quantitative account of the comparative solution complexation ofY and REEs as M(Oxt and Th e comparative behavior of phosphate and arsenate coprecipitation exhibits both s imilaritie s a nd differenc es. During arsenate coprecipitation the light REEs are retained in so lution, relative to the heavy REEs, to a greater extent than is the case for pho s phate coprec ipitation Notab l e irregularities are observed in the comparative coprecipitation behavior of nearest-neighbor elements (e .g. Eu-GdTb and TmYb Lu). Such irregularitie s are very similar for pho sp hate and arsenate coprecipitation in the absence and in the pre se nce of solution complexation. 94

PAGE 109

6. 2 Introduction Observations of Yttrium (Y) and rare earth element (REE) phosphate solubility products (Liu and Byrne, 1997; Byrne and Kim 1993; Jonassen et al., 1985) indicate that formation of Y and REE coprecipitates may determine the upper bound concentrations of these elements in the ocean and may also strongly influence the relative abundance patterns of Y and REEs in seawater (Byrne et al., 1996; Byrne and Kim 1993). Little work has been previously devoted to examination of coprecipitation on comparative REE abundance s in solution The work of Byrne and Kim (1993) investigated the influence of pho s phate coprecipitation on the removal of Ce, Eu, Gd, Tb and Yb from solution. Their work showed s ub s tantial difference s in the extent of REE removal from solution during coprecipitation and d e monstra ted that removal patterns (fractionation s) are strongly influenced by REE solution complexation. Byrne et al. (1996) inve s tigated the influence of phosphate coprecipitation on fourteen REEs and Y in the absence of solution complexation Their work demonstrated that removal of Y and REE s from solution as pho sp hate coprecipitates depletes Sm in so lution relative to heavier and lighter e lem en t s, and produ ces a variety of nearest neighbor concentration anomalies consistent with concentration anomalie s reported in high precision analyses of open-ocean water. In the present work we have extended the work of Byrne et al. (1996) to include the influence of so lution complexation on comparative Y and REE phosphate coprecipitation. Additionally, in view of the high concentrations of dissolved arsenate in some environmental systems (Maest, et al .. 1992; Anderson and Bruland, 1991 ), we have investigated the comparative behaviors of Y and REE s during coprecipitation with arsenate. This investigation allow s direct examination of comparative phosphate and arsenate coprecipitation with Y and REEs, as 95

PAGE 110

well as a comparative examination of the influence of solution complexation on Y and REE coprecipitation with both phosphate and arsenate. 6. 3 Experimental Procedures The experimental procedures in this work follow tho se de scribed in Byrne et al. ( 1996) with minor modification s. Coprecipitation experiments were performed using Y and all REEs except Pm. Yttrium and REE solutions were made up from individual ICP sta ndards (SPEX Chemical). Sodium phosphate dibasic ( ultra-pure grade), sodium oxalate and sodium borate decahydrate were obtained from J. T. Baker. Sodium arsenate heptahydrate was from Sigma Chemicals. Alllabware was cleaned by washing with Micro so lution (Cole-Parmer), followed by immersion in 4 N HCl for at least one week and s ub sequent rinsing with Milli-Q water Coprecipitation was carried out in 1000 ml-capacity Teflon bottles For experiments not involving oxalate, 500 ml of REE plu s Y solution was mixed with an equal volume of phosphate or arsenate solution The initial Y and REE concentrations after mixing were 100 ppb each ([ 0M?+ ]= 100 ppb ) and the total concentration of REEs and Y ( M?+]) equaled l.Ox 1 o-5 molal. The concentration of phosphate or arsenate was 2x 10-4 molal. For experiments involving oxalate sodium oxalate salt was first dissolved in phosphate or arsenate solution and subsequent to mixing with the REE plus Y solution the final oxalate concentration was 1 0x 1 o-5 molal. After mixing these solutions Teflon bottles were thermostated at 25 .0. 1 C. The total ionic strength of these mixed solutions was on the order of 1xl0-3 molal. The initial solutions were sufficiently acidic (pH 3.3) to keep Y and REEs from precipitating. Solution pH was measured with a Ros s combination electrode calibrated on the free hydrogen ion concentration scale (McBryde, 1969, 1971; Byrne and 96

PAGE 111

Kester 1978) After taking samples for initial concentrations of Y and REEs, the pH was slowly raised by adding 0.2 M borate solution while stirr ing vigorously After each pH adjustment a 5 cm3 sample was taken to check for onset of precipitation. If precipitation was observed from measurement of dissolved Y and REE concentrations samples were taken at regular time interv als. Di sso lved concentrations of Y and REEs were mea s ured with a Fisons PQS ICPMass Spectrometer. No sample processing other than filtration was required prior to analysis. Subsequent to filtration with 0.22 jlm acetate cartridge filters (Coming), 5 cm3 of samp le were combined with 5 cm3 of 100 ppb indium so luti on used as an internal standard. Sample concentrations were mea sured against a mixed Y and REE s tandard so lution ( SPEX Chemical) 6.4 Theory The coprecipitation phenomena ob serve d in this work were quantitative ly de s cribed usmg an equation (Doerner and Hoskins 1925) originally developed to examine the coprecipitation of Ba and Ra sulfate. The results of Doerner and Hoskin s ( 1925) support the ass umption that relative metal concentrations on a coprecipitate surface are directly proportional to relative metal concentrations in sol ution : (1) where (M)s and (M)s are the surface concentrations of metal ions M ; and M j [M/+] and [M]+] are the concentrations of free (uncomplexed) metal ions in sol ution and is a proportionality or fractionation factor for metals M ; and Mj. 97

PAGE 112

If the concen tration s of metals M; and M j on the coprecipitate surface are directly proportional to free metal ion concentra tion s in solutio n as shown in eqn. (1) then, in the presence of solutio n complexation eqn. (1) can be expressed in terms of total metal ion solution concentrations ( [M/+ lr) as follows: (2) (3) Under our experimental conditions, where metal ions are complexed by oxala te, eqn. (3) can be written as Ar, =A.! O+ox.Bt (M)[Ox2]+0x/32(M)[Ox2 ] 2 ) 1 1 11 (1+0x.Bt ( M;)[Ox 2-]+0x/32 (M;)[Ox 2-f)' (4) h .B (M) [MOx + ] d .B (M) [M(Ox)-2 ] w ere an -and brackets [ ] denote the Ox 1 -[M3+][0x2-] Ox 2 [M3+][0x2-f free concentra tions of each indicated species. Beginning with eqns. (1) and (2), following the derivations of Doerner and Hoskins ( 1925) it can be shown that the solution concentrations of metals M; and Mj in a coprecipitation process can be written, in the presence and absence of solut ion complexation, in the fo ll owing forms: (5) and 98

PAGE 113

(6) point in time subsequent to initiation of coprecipitation and [0M?+ lr, [0MJ+ lr, [0M?+] and [ 0MJ+] represent corresponding concentrations at any prior point in time Equations (5) and (6) were used to quantitatively describe the coprecipitation of Y and REE elements during the formation of phosphate and arsenate coprecipitates in the presence and absence of oxalate. 6. 5 Results and Discussion The extent of metal removal from solution during coprecipitation is shown in Fig. 6.1 as a function of time for three rare earths REE concentrations decreased steadily through time, although at variable rates in each experiment. The pH required to initiate coprecipitation in arsenate experiments was higher than that for phosphate due to the greater solubility of REE arsenates Metal solubilization due to oxalate complexation (Figs. 6.1 b and 6.ld) also increased the pH requisite to coprecipitation The comparative behaviors of Y and REEs during coprecipitation are shown in Fig. 6.2. Phosphate and arsenate coprecipitation patterns show both substantial similarities and 99

PAGE 114

pH 4.6-4 5 pH 5 .1-5.0 -0 1 -0 1 -0.2 .:::. -0.3 -.::: ;; -0.4 ::E ..9 -0 .6 -0.7 (la) 0 10 20 30 40 50 60 70 80 20 30 40 50 60 70 80 pH 5 .2 pH 5 8-5.7 pH 6 1-6 0 0 -0. 1 0.2 -0.3 -.::: -0 6 -0 .4 ;; ::E -0 5 bO ..9 -0.8 -0 6 -I Sm (lc) -0 7 (ld) 0 5 I 0 15 20 25 30 0 5 10 15 20 25 30 Time (hours) Time (hours) Figure 6.1 Selected Rare Earth Concentrations (La, Sm, Lu) as a Function of Time (a) rare earth phosphate coprecipitation, (b) rare earth phosphate coprecipitation in the presence of oxalate, (c) rare earth arsenate coprecipitation and (d) rare earth arsenate coprecipitation in the presence of oxalate. 100

PAGE 115

0 0 -0. 1 .f" .., -0.2 ;:::!; 0 :s;: -0 .3 + .... 6 be -0 4 0.5 -0. 2 -0.4 !' ;:::!;0 -0 6 :s;: + -0.8 be -1.0 -1.2 29 (2c) -1.4 L.....l--L.J.....J........L-'-l'-L..L..L...J....J......L.J_J,_LJ Y Ce Nd Sm Gd Dy E r Yb La Pr Pm Eu 1b H o Tm Lu 0.0 -0 2 -0.4 -0.6 -0.8 (2b) Y Ce Nd Sm Gd Dy E r Yb La Pr Pm Eu 1b Ho Tm Lu Figure 6.2 Fractionation Patterns Observed in the Coprecipitation Processes with Times (hours). Fractionation is depicted as log[Mi+ ] 1[0M i + ] and log[Mi+]r 1[0Mi+] r where [0M?+] and [0M?+]r are the initial concentrations of each element. (a) rare earth phosphate coprecipitation, (b) rare earth phosphate coprecipitation in the presence of oxalate, (c) rare earth arsenate coprecipitation, and (d) rare earth arsenate coprecipitation in the presence of oxalate. 101

PAGE 116

significant differences. Minima in phosphate and arsenate coprecipitation patterns are observed at Sm and Eu (Figs. 6.2a and 6.2c) Solution complexation shifts these minima toward lighter elements because heavy REEs are more strongly complexed than light REEs, favoring their retention in solution In the presence of oxalate the minimum in the phosphate coprecipitation pattern is shifted to Ce (Fig. 6.2b) and the arsenate minimum is shifted to Sm (Fig. 6.2d). Inspection of the Fig. 6. 2 results for each of the four coprecipitation experiments shows that coprecipitation patterns developed early in each experiment are generally retained throughout the experiment. For example, higher Er concentrations relative to its nearest neighbors (Ho and Tm) are seen throughout the experimental patterns in Figs. 6.2a and 6.2c, as are elevated Gd concentrations relative to Eu and Tb. The most appropriate means of comparison of the metal fractionation pattern s shown in Fig. 6.2 is examination of fractionation factors and calculated via eqns. (5) and (6). Figs. 6.3a and 6.3b s how phosphate and results and Figs. 6.3c and 6.3d s how arsenate an d results obtained (Table 6.1) using the final four metal fractionation patterns ( log[M( + ] I [0M?+J) in each experiment. Sm is used as the reference element for all calculations ([Sm3+] = [M]+] and [Sm3+]r = [MJ+] r). The error bars given in Table 6.1, and shown graphically in Fig 6.3, represent the standard errors of the calculated and res ult s for the four mea sure ments Comparison of Figs. 6.3a and 6.3c shows that, while results for phosphate and arsenate are similar in general appearance, results for the lighte s t REEs (La, Ce, Pr) relative to the heaviest REEs (Ho Lu) differ distinctly for phosphate and arsenate coprecipitation. These differences are also seen in comparisons of Figs 6.3b and 6.3d. Comparisons of result s with r es ults (Figs. 6.3b v s. 6.3a and 6.3d vs. 6.3c) show that increasing trends in for 10 2

PAGE 117

Table 6.1 Fractionation Factors and for Yttrium and Rare Earth Coprecipitation (calculated using equations (5) and (6)) REE (phosphate) (phosphate) (with oxalate) (arsenate) (arsenate) (no oxalate) (no oxalate) (with oxalate) y 0.352.037 0 .2 16 014 0.310.015 0.216 0.021 La 0.455 0 019 0 .794 0.016 0.171 .006 0.393 0.020 Ce 0 .687 0.013 1.056.014 0.363 006 0.636. 010 Pr 0 .769 0.009 1 054.009 0.503.006 0.713 0 020 Nd 0.770.006 0.961 015 0.558 0.008 0.715. 022 Pm Sm 1 1 Eu 0 .9 56 0 015 0.848 0 10 1.017.007 0.940 0.008 Gd 0. 706 0.014 0.609 0.006 0.738 .009 0.686.011 Tb 0.742 011 0.529.006 0.868. 009 0 .650 0.020 Dy 0 .663 0.018 0.420.007 0.819.012 0.563. 028 Ho 0.530.0 19 0.303 0.007 0.637.015 0.398 0 0 13 Er 0.486 0.011 0.250 0.005 0.586.0 12 0.344.0 14 Tm 0 .500 0 009 0.234 0.004 0.600. 0 13 0.293 0.020 Yb 0.529.0 18 0.238.007 0.681.008 0 322 0.015 Lu 0.473.023 0.203.007 0.581.010 0.265 017 103

PAGE 118

0 0 0.2 0.2 0.4 0.4 A.: IJ 0.6 0.6 0.8 0 0.2 0.4 0.8 (3 a ) ( 3c) Y Ce N d Sm Gd Dy E r Yb La Pr Pm Eu Tb H o Tm Lu 0 8 0 0.8 (3b) (3d) Y Ce N d Sm Gd D y Er Yb La Pr Pm Eu Tb Ho Tm Lu Fi gure 6.3 Fract ion at ion Fac t o r s a nd fro m T ab l e 6. 1 (a) rar e earth phosphate co pr ec ipitation, ( b ) rare eart h phosphate copre cipitation i n the presence of oxa lat e, (c) rar e earth arsenate copre c ipit a tion a nd (d) rare earth arsenate coprecipitation in the presence of oxalate 104

PAGE 119

elements heavier than Sm (Figs. 6.3a and 6.3c) are enhanced in the presence of oxalate (Figs. 6. 3b and 6.3d) None of the curves shown in Fig. 6.3 are smooth. All of the patterns ( ll{ and in Fig. 6 3 show irregularity with increasing atomic number. The most consistent and obvious instances of this irregularity are seen in the patterns of Eu-Gd-Tb and TmYb-Lu As a means of directly comparing the fractionation patterns obtained for phosphate and arsenate, Fig. 6.4 shows ll{ results for phosphate normalized to ll { results for arsenate as well as results similarly compared ( I ). Fig 6 4 shows that arsenate normalization of phosphate f ractionation factors produces curves which are very similar in appearance and are largely devoid of element to element irregularities. The smooth trend observed for normalized ll .. I) data (phosphate/arsenate) shows that observed irregularities in the coprecipitation behavior of adjacent elements are nearly identical in pho s phate and arsenate coprecipitation. The relationships s hown in Fig 6.4 indicate that there are striking similarities underlying the fractionations of Y and REE s during pho s phate and arsenate coprecipitation. The comparative behavior of Eu-Gd-Tb and TmYb-Lu is, for example very s imilar fo r pho s phate and arsenate coprecipitation. It is difficult to rationalize the irregular coprecipitation behavior of Eu Gd Tb and TmYb Lu observed in all experiments directly in terms of crystal radii since the ionic radii of the REE elements decrease monotonically with increas ing atomic number. More appealing explanations for the observed comparative behaviors will likely involve accounts of the irregular extent of REE hydration in solution ( Rizk alla and Choppin, 1991 and 1994) An assessment of the influence of oxalate complexation on phosphate and arsenate coprecipitation is seen in Fig 6 5 The calculated quotients ll{l shown is Fig 6 5 105

PAGE 120

Figure 6.4 Comparison of Phosphate and Arsenate Coprecipitation Patterns ( A.ii (phosphate)/ A.ii (arsenate). 106

PAGE 121

Table 6.2 Complexation Constants for Yttrium and Rare Earth with Oxalate Appropriate to 25 C and Zero Ionic Strength REE y 6.662 10.865 La 5.882 9.405 Ce 6.072 9.835 Pr 6.208 10. 086 Nd 6.303 10.245 Pm Sm 6.479 10 555 Eu 6.532 10.615 Gd 6.500 10 550 1b 6.622 10.825 Dy 6.663 10 .92 4 Ho 6.685 10 974 Er 6.726 11.119 Tm 6.772 11.095 Yb 6.852 11. 375 Lu 6.865 11.437 Complexation constants for Y, Ce, Eu, Tb and Tm are obtained from Martell and Smith (1977). Yb results are from Cantrell and Byrne (1987). Other constants are assessed based on REE linear free energy relationships as characterized by Lee and Byrne (1992, 1993). 107

PAGE 122

provide assessments of Y and REE oxalate complexation in our experiments, relative to Sm: (1 +ax.BI (M;)[0X 2 -]+0x.B2 (M;)[Ox2-f) = Cl+ox.B1(Sm)[Ox 2 -]+0x,B2 (S m)[Ox 2-f). (7) Directly calculated values of I obtained using the Aii results given in Table 6.1, are in good agreement with the results of speciation calculations (right-hand side of eqn. 7) obtained using the oxalate stability constant data given in Table 6.2. Calculated values of the quotient R = (1 +ax.BI (M;)[OX2-]+ ox.B2(M;)[Ox2-f) (1 +ax.BI (Sm)[Ox2-]+0x,B2(Sm)[Ox2-]2 ) (8) shown in Fig. 6 5 are appropriate to the beginning ( Rini riat) and end ( R finat) of eac h experiment. Due to the similar initial concentrations of total metals and oxalate in our experiments (Fig. 6.2), free oxalate concentration varied from 1.7xl06 molal at the inception of coprecipitation and increased to 6.3x 10 6 molal at th e end of each phosphate and arsenate experiment as total metal concentrations decreased. Free oxalate concentration s were determined through iterative calculations using Table 6.2 stability constant data, total oxalate concentrations equal to 1xl05 molal and direct measurements of total dissolved metal concentrations through time (Fig. 6.2). The calculated R values shown in Fig. 6.5 exhibit only minor variation with extent of coprecipitation and are consistent with and x:. results obtained from direct observations ([M3+]/[0M3+]) of metal removal from I) solution during coprecipitation. These results s uggest that Y and REE coprecipitation behavior in complexing media can be reasonably predicted from characterizations (Table 6.1 ) and metal speciation calculations based on the solution concentrations of complexing ligands. 108

PAGE 123

2 5 2 0.5 3 2 5 2 f.:=-1::: 1.5 -=-.....:: 0.5 (Sa) (S b ) experimental -R(calculated ) initial R(calculated) final I [Ox]=6 .2xl0 M ,., ,' .-0 ' [Ox]=l.7x10-6 M ---experimental --R(calculat e d) initial R(calculated) final Y La Ce P r Nd Pm Sm Eu Gd Th D y H o Er Tm Yb Lu Figure 6.5 Effect of Solution Complexation on Yttrium Rare Earth Fractionation Factors. (a) phosphate coprecipitat i on, (b) arsenate coprecipitation. 109

PAGE 124

6. 6 Conclusions Solution chemistry is seen to exert strong controls on the coprecipitation behavior of REE phosphates and arsenates. Strong solution complexation enhances the retention of heavy REEs in solution. The extent of this enhancement, for Y and REE complexation by oxalate, can be quantitatively described in terms of thermodynamic characterizations of Y and REE oxalate stability constants. The comparative behaviors of Y and REE coprecipitation by phosphate and arsenate evidence strong similarities in the fine structure of REE fractionation patterns. Solution chemistry may also play a strong role in thi s instance through the influence of irregular REE hydration behavior on relative metal affinitie s for solution and solids. 110

PAGE 125

REFERENCES Anderson L.C.D. and Bruland K.W. (1991) Biogeochemistry of arsenic in natural waters: The importance of methylated species. Environ. Sci Te c hnol. 25, 420-427. Balistrieri L, Brewer P.G. and Murray J W. (1981) Scavenging residence times of trace metals and surface chemistry of sinking particles in the deep ocean. Deep-Sea Res. 28A, 101-121. Bau M., Dulski P. and Moller P. (1995) Yttrium and Holmium in South Pacific seawater: Vertical distribution and possible fractionation mechanisms. Chernie der Erde. 55, 1-15. Bertram C.J and Elderfield H. ( 1993) The geochemical balances of the rare earth elements and neodymium isotopes in the oceans. Geochirn. Cosrnochirn. Acta 57, 19571986. Borkowski M and Siekierski S. (1992) Factors affecting the position of Y and Actinides(ITI) with respect to lanth anides in the NH4SCN-Adogen-464SCN extraction system. Radiochirn. Acta 56, 31-35. Byrne R.H. and Kester D R. (1978) Ultraviolet spectroscopic study of ferric hydroxide complexation. J. Sol Chern 7 373 383 Byrne R.H. and Kim K.H (1990) Rare earth element scavenging in seawater. Geochirn. Cosrnochirn. Acta. 54, 2645 2656. Byrne R.H and Kim K.H (1993) Rare earth precipitation and coprecipitation behavior : the limiting role of Po;on dissolved rare earth concentrations in seawater. Geochern. Cosrnochirn. Acta 57, 519-526. Byrne R.H. and Lee J.H. (1993b) Comparative yttrium and rare earth element chemistries in seawater Mar. Chern. 44, 121130. Byrne R.H and Li B. (1995) Comparative complexation behavior of the rare earths. Geochirn. Cosrnochirn. Acta 59, 4575-4584. Byrne R.H. and Sholkovitz E.R. (1996) Marine C hemi stry and Geochemistry of the Lanthanides. In Handbook on the Physics and Chemistry of Rare Earths (eds. K.A. Gschneidner and L.Eyring), 487-593 Elsevier. Byrne R H Kump L.R. and Cantrell K.J. (1988) The influence of temperature and pH on trace metal speciation in seawater. Mar. Chern 25 1 63-181 111

PAGE 126

Byrne R.H., Lee J H. and Bingler L.S (1991) Rare earth element complexation by Po; ions in aqueous solution. Geochim. Cosmochim Acta 55, 2729-2735. Byrne R.H., Liu X and Shijf J (1996) The influence of phosphate coprecipitation on rare earth distributions in natural water. Geo c him Cosmochim. Acta 60, 3341-3352. Cantrell K.J (1988) Actinide (III) Carbonate complexation. Pol y hedron 7, 573-574. 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) Temperature dependence of europium carbonate complexation. J Sol. Chem 16, 555-566 Cantrell K.J. and Byrne R.H.(l997) Mixed ligand complexes of the rare earth elements. In preparation. Chatt A. and Rao R.R. (1989) Mat. R es Soc. Symp. Proc. 127 897. as cited in Rao and Chatt (1991). Choppin G.R. and Strazik W F (1965) Complexe s of trivalent lanthanide and actinide ions. I. Outer-sphere ion pairs Inorg Chem 24, 1250-1254. Ciavatta L., Ferri D Grenthe 1., Salvatore F. and Spahiu K. (1981) Studies on metal carbonate equilibria. 3. The lanthanum (III) carbonate complexes in aqueous perchlorate media. Acta Chem. Scand. A35, 403-413. de Baar H.J W. Schijf J., and Byrn e R H. (1991) Solution chemistry of the rare earth elements in seawat er. Eur. J Solid Stat e Inorg. Chem. 28, 357 -373. Doener H .A. and Hoski n s W.M. (1925) Coprecipitation of radium and barium sulfates. J. Am. C h em. Soc. 47, 662-675. Elderfield H. ( 1988) The oceanic c h e mistry of the rare earth elements. Philos. Trans. R So c London A325, 105 126. E lderfield H and Gre aves M.J. (1982) The rare earth elements in seawa ter. Nature 296, 214-219. Elderfield H Upstill-Goddard R. and Sholkovitz E.R. ( 1990) The rare earth e lement s in river s, estuaries, and coastal seas and their s i g nificance to the composition of ocean waters. Geochim. Cosmochim. A c ta 54, 971-991 Erel Y and Stolper E.M. (1993) Modeling of rare -e arth element partitioning between particles and solution in aquatic environments. Ge oc him. Cosmo c him. Acta 57 513-518. Ferri D. Grethe I., Hietanen S. and Salvatore F. (1983) Studies on metal carbonate equilibria. 5 The Cerium (lll) carbonate complexes in aq ueous perchlorate media Acta Chem. S c and A37, 359 365. Firsching F.H. and Brune S .N (1991) Solubility products of the trivalent rar e-earth phosphates. J Chem. Eng. Dat a 36, 93 95. 112

PAGE 127

German C.R., Masuzawa T., Greaves M.J., Elderfield M., and Edwond J.M. (1995) Dis so lved rare earth elements in the Southern Ocean: Ce oxidation and the influence of hydrography. Geochirn. Cosrnochirn. Acta 59, 1551-1558. Glynn P D Reardon E.J., Plummer L.N. and Busenberg E. (1990) Reaction paths and equilibrium end-points in solid-solution systems. Geochirn Cosrnochirn. Acta 54, 267-282. Goldberg E.D. Koide M., Schmitt R.A. and Smith R.H. (1963) Rare earth distributions in the marine environment. J Geophys. Res. 68, 4209-4217 Goldstein S J and Jacob se n S.B (1987) The Nd and Sr isotopic systematics of river water dissolved material: implications for the sources of Nd and Sr in seawater. Chern. Geol. 66 245-272 Goldstein S.J and Jacobsen S.B. (1988a) Rare earth elements in river waters. Earth Plan et. Lett. 89, 35-47. Goldstein S.J. and Jacobsen S.B. (1988b) REE in the Great Whale River estuary, northwest Quebec Earth Planet. Lett 88,241-252. Grasshoff K., Ehrhardt M and Kremling K. (1983) Methods of Seawater Analysis. Verlag Chemie. 419pp. Greaves M.J., Rudnicki M and Elderfield H (1991) Rare earth elements in the Mediterranean Sea and mixing in the Mediterranean outflow. Earth Planet. Sci. Lett. 103, 169-181. Honeyman B D and Sant s chi P.H (1988) Critical review: Metals in aquatic systems. Their scavenging residence times from l aboratory data remain s a challenge. Environ. Sci. Technol. 22, 862. Hunter K.A. ( 1983) The adsorptive properties of sinking particles in the deep ocean. De ep Sea Res. 30, 669. Jonasson R.G., Ban coft G.M and Nesbit H.W. (1985) Solubilitie s of so me hydrous REE pho sp hates with implications for diagenesis and seawater concentrations. Geochirn. Cosrnochirn. Acta 49, 2133-2139. Lee J.H. and Byrne R.H. (1992) Examination of comparative rare earth element complexation behavior u sing linear free energy r e lation s hip. Geochirn. Cosrnochirn. Acta 56, 1127-1137. Lee J.H and Byrne R.H. (1993a) Complexation of trivalent rare earth elements (Ce, Eu, Gd, Tb, Yb) by carbonate ions. Geochirn. Cosrnochirn. Acta 57 295 303. Lee J.H. and Byrne R.H. (1993b) Rare earth element complexation by fluoride ions in aqueous solution. J.Sol. Chern. 22, 751. Lee J.H. and Byrne R.H (1993c) Comparative yttrium and rare earth element chemistries in seawa ter. Mar .Chern 44, 121-130. 113

PAGE 128

Lee J.H. and Byrne R.H. (1994) Pressure dependence of gadolinium carbonate complexation in seawater. Geochim Cosmochim. Acta 58, 4009-4016. Li B. and Byrne R.H. ( 1997) Ionic strength dependence of rare earth-NT A stability constants at 25 c. Aqutic Geochem (in press). Liu X. and Byrne R.H. (1995) Comparative carbonate complexation of yttrium and gadolinium at 25C and 0.7mol dm -3 ionic strength Mar Chem .. 51, 213-221. Liu X. and Byrne R.H. (1997) Rare earth and yttrium phosphate solubilities in aqueous solution. Geochim. Cosmochim Acta 61, 1625-1633. Lundqvist R. (1982) Hydrophilic complexes of the actinides I. Carbonates of trivalent americium and europium. Acta Chem. Scand. A36, 741-750 Maest A.S., Pasilis S.P. Laurence G.M. and Nordstrom D.K. (1992) Redox geochemistry of arsenic and iron in Mono Lake, California, USA. In Y.K. Kharaka and A.S. Maest (Editors) Water-Rock Interaction (Balkema Rotterdam), 507-511. Martell A.E. and Smith R.M. (1974) Critical Stability Constants Vol. 1, Amino Acids. Plenum Press, New York 469 pp. Martell A.E. and Smith R.M (1977) Critical Stability Constants, Vol. 3, Other Organic Ligands. Plenum Press, New York. 495 pp. Martell A.E and Smith R.M. (1982) Critical Stability Constants, Vol. 5, First Supplement. Plenum Press, New York 604 pp Martin J M H<)>gdahl O T. and Philippot J C. (1976) Rare earth element supply to the oceans. J. Geophys. Res 81, 3119-3124. McBryde W A E. (1969) The pH meter as a hydrogen ion concentration probe. Analyst 94, 337-346. McBryde W A.E. (1971) The pH meter as a hydrogen ion concentration probe: A postscript. Analyst 96 739-740. Millero F.J. and Schreiber D.R. (1982) Use of the ion pairing model to estimate activity coefficients of the ionic components of natural waters. Amer. J. Sci. 282, 1508 1540. Millero F.J. (1992) Stability constants for the formation of rare earth inorganic complexes as a function of ionic strength. Geochim. Cosmochim. Acta 56 3123-3132. Moeller T. (1963) The Chemistry of the Lanthanides Reinhold Publishing Corp., New York. Moeller T. ( 1972) Complexes of the Lanthanides. In Lanthanides and A c tinides, Volume 7 (ed. K.W. Bagnall). 275 298. University Park Press. Moeller T ., Martin D.P., Thompson L.C., Fem1s R. Feistel G.R. and Randall W .J. (1965) The coordination chemistry of yttrium and the rare earth metal ions Chem. Rev. 65, 1-50. 114

PAGE 129

Moffett J.W. (1990) Microbially mediated cerium oxidation in seawater. Nature 345, 421-423. Nyffeler U.P., Li Y H and Santschi P.H. (1984) A kinetic approach to describe trace element distribution between particles and solution in natural systems. Geochim. Cosmochim. Acta 48, 1513-1522 Piepgra s D.J. and Jacobsen S.B. (1992) The behavior of rare earth elements in seawater: Precise determination of variations in the North Pacific water column. Geochim. Cosmochim. Acta 56, 1851-1862 Piper D.Z. (1974) Rare earth elements in the sedimentary cycle: A summary Chern. Geol. 14, 285-304. Rao R.R. and Chatt A. ( 1991) Studies on s tability constants of Europium (III) carbonate complexes and application of SIT and ion-pairing Model s. Radiochim. Acta 54, 181-188. Rizkalla E.N. and Choppin G.R. (1991) Hydration and hydrolysis of Lanthanide s. In Gschneider K.A. and Eyring L. (Editors) Handbook on the Phy sics and Chemistry of Rare Earths, Vol. 15. Elsevier. 393-440 Rizkalla E.N. and Choppin G.R. (1994) Lanthanides and Actinides hydration and hydroly s i s. In Gschneider K.A., Eyring L, Choppin G R and Lander G H. ( Editor s) Handbook on the Physics and Chemistry of Rare Earths, Vol. 18: Lanthanides/ Actinides: Chemstry. Elsevier. 529-560. Ru z ikina L.V. Morov LN., Ryabukhim V.A., Ermakov A.N. an d Filimonava V.N. ( 1978) Investi ga tion of the complexing of europium with carbonate ions. Zh. Anal. Khim. 33, 1082 1088. Shabani M B. Akagi T., Shimizu H. and Mashda A. (1990) Determination of trace lanthanides and yttrium in seawater by inductively coupled plasma mass s pectrometry after preconcentration with solvent extraction and back-extraction. Anal. Chern. 62 2709-2714. Shannon R. D ( 1976) Revi se d effective ionic radii and systematic s tudies of interatomic distance s in halides and chalcogenides. Acta Cryst. A32, 751-767. Sholkovitz E.R. ( 1988) Rar e earth elements in th e sediments of the North Atlantic Ocean, Amazon D e lta and East China Sea: R e interpretation of terrigenous input patterns to the oceans. Amer. J. Sci. 288, 236-281. Sholkovitz E.R. ( 1992 ) Chemical evolution of rare earth elements: fractionation between colloidal and so lution phases of filtered water. Earth and P lanet Lett. 114, 77-8 4. Smith R.M. and Martell A.E. ( 1975) Critical Stability Constants, Vol. 2, Amines. Plenum Press New York. 415 pp. Smith R.M. and Mart ell A.E. (19 76) Critical Stability Constants, Vol. 4, Inorganic Complexes. Pl e num Press, New York. 257 pp. 115

PAGE 130

Smith R.M. and Martell A.E. (1989) Criti ca l Stability Constants, Vol. 6, Second Supplement. Pl e num Pr ess, New York 643 pp Spahiu K. ( 1985) studies on metal carbonate equilibria 11. Yttrium (Ill) carbonate complex formation in aqueous perchlorate media of various ionic st rength s. Acta Chern. Scand. A 39, 33-45. Tananaev LV. and Petushkova S.M. (1967) The reaction of gadolinium chloride with orthophosphate ions in aqueous solution at 25 C. Ru ss J. Inorg. Chern. 12, 39-42. Tananaev LV. and Vasil 'ev a V.P. (1963) Lanthanum phosphates. Ru ss J. In org. Chern. 8 555-558. Thompson S.W and Byrne R.H (1987) Indicator ligands in metal complexation stud ies: Role of 4-(2 Pyridylazo) resorcinal in europium carbonate equilibrium investigations. Anal. Chern. 60 19-22 Thorstenson D. C. and Plummer L.N. (1977) Equilibrium criteria for two-component sol ids with fixed composition in an aqueous-phase; example: the magnesium Turner D.R. and Whitfield M (1987 ) An equilibrium model fo r copper in sea and estuarine waters at 25c including complexation with glycine EDTA and NT A. Geochim. Cosmochim. Acta 51, 3231-3239. Turner D.R., Whitfield M. and Dickson A.G. (1981) The equilibrim speciation of dissolved component s in freshwater and seawater at 25 c and 1atm. pressure. Geochim. Cosmochim Acta 45, 855-881 Wood S.A. (1990) The aq ueous geochemistry of the rare-earth elements and yttrium. 1. R e view of available low-temperature data for inorganic complexes and the inorganic REE speciation of natural waters. Chern. Geol. 82, 159-186. Zhang J Amakawa H and Nozaki Y. (1994) The comparative behavior of Yttrium and Lanthanides in the seawa ter of the North Pacific. Geophys. R es L e tt 21 26772680. Zhang J and Nozaki Y. (1996) Rare earth elements and yttrium in seawater: ICP-MS d e termination s in the East Caroline, Coral Sea, and South Fiji Basins of the W es tern South Pacific Ocean. Geochim. Cosmochim Acta 60, 463 1-4644. 116

PAGE 131

APPENDICES 117

PAGE 132

Appendix 3.1 Efficiency of Rare Earths and Yttrium Coprecipitation by Fe Hydroxide REE test I test II 10ppb 20ppb y 10.6 20.5 La 10.5 20 3 Ce 10.2 19.3 Pr 10.0 19.5 Nd 10.1 19. 3 Pm Sm 10.0 19.3 Eu 9.9 19.2 Gd 10.0 19.5 Tb 9.9 19.6 Dy 10.0 19.0 Ho 10.0 19 7 Er 10.2 20 0 Tm 10.3 20.4 Yb 10.4 20.5 Lu 10 3 20.6 118

PAGE 133

Appendix 3.2 Rare Earth and Yttrium Recovery During Fe Coprecipitation Procedure Using Pr or Nd as Yie l d Standard Elements Sample Pr Nd 10 ppb 50ppb 01 10. 1 49.1 A1 9.65 49.6 02 10 3 49.0 A2 9.85 49.2 03 10.2 50. 0 A3 10 1 49.8 04 10.4 49. 5 A4 10.1 50. 0 05 10.1 49. 6 A5 10.0 51.0 06 10.2 49. 5 A6 10 3 48 .2 07 10.2 50.4 A7 10.2 50. 1 08 10. 3 49.4 A8 10.4 49.4 119

PAGE 134

Appendix 3.3 Extraction Efficiency of Rare Earths and Yttrium f r om TBP P h ase REE test I test II 50ppb 50ppb y 50 1 51.0 La 51.7 52.0 Ce 50.8 51.0 P r Nd 51.0 51.3 Pm Sm 50 5 51.0 Eu 51.0 51.3 Gd 50.2 51.2 Tb 51.0 51.2 Dy 50.6 51.0 Ho 50. 6 51.2 Er 51.0 51.4 Tm 50. 2 50. 6 Yb 50.2 50.7 Lu 51.0 50.8 120

PAGE 135

Appendix 3.4 Experimental Results of Rare Earths and Yttrium Complexation with Carbonate Ions: Results of Experiment I (The r esults are s hown as D/D0 as a function of pH) y La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu pH .0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1 .0000 1.0000 1.0000 4.4408 1.9357 0.9692 0.9394 0.9169 0.9000 0.8938 0.9317 0 9066 0 9215 0.9038 0 .8 998 0 .8 842 0.8601 0.8769 5.3180 .5562 0.7917 0.6934 0.6166 0.5120 0 5010 0.5700 0.5156 0.4816 0.4758 0.4439 0.4051 0.3541 0.3777 5.8843 .2 760 0 6083 0.4696 0.38 23 0 .2682 0 .2563 0.3130 0.2591 0.2302 0.2181 0 1953 0 165 4 0 1358 0.1428 6 1176 .1280 0.4109 0 2817 0.2044 0.1301 0.1228 0 1523 0.1175 0.1007 0.0931 0.0817 0.0642 0.0515 0.054 1 6.2950 .0440 0 .2339 0 1342 0.0897 0 0490 0 0449 0 0575 0 0402 0 .032 5 0 0290 0 .0242 0 0181 0 0145 0.0146 6.4911 .0171 0 1 248 0.0620 0 0390 0.0190 0 0169 0.0225 0 0143 0 0112 0.0096 0.0078 0.0057 0.0047 0.0045 6.6399 .0078 0 .068 4 0.0300 0.017 3 0.0076 0.0066 0 0089 0.0052 0.0041 0 0034 0 0027 0.0019 0.0018 0.00 1 6 6 7734

PAGE 136

Appendix 3.4 (Continued) Results of Experiment II y La Ce Pr Nd Sm Eu Gd Th Dy Ho Er Tm Yb Lu pH 0000 1.0000 1.0000 1.0000 1.0000 1.0000 1 0000 1.0000 1 0000 1.0000 1.0000 1.0000 1.0000 1.0000 4.4610 1.9322 0 9310 0 9103 0 8921 0.8815 0 8833 0.8968 0.8977 0.8867 0.8859 0 8780 0 8711 0.8568 0 8574 5.3045 .5802 0.7888 0.7076 0 6121 0.5455 0 5386 0 5907 0.5531 0.5160 0.5097 0.4829 0.4425 0.3930 0.4034 5 8573 .227 8 0 4801 0.3782 0 .2 794 0 2183 0 2104 0 2511 0.214 8 0 1 8 5 2 0 .1781 0 .1587 0 1349 0 1119 0 115 8 6. 1683 .1154 0 3856 0.2569 0 1646 0.1176 0 1114 0 1 371 0.1070 0.0 8 82 0.082 3 0 0711 0 0571 0 0463 0 0470 6 3305 0510 0 2491 0 1443 0 0851 0.0554 0.0508 0.0634 0.0464 0.0368 0.0 33 3 0.0277 0.0215 0.0174 0 0172 6.4742 .0 202 0 1346 0 0702 0 0377 0 0226 0 0201 0.0 2 56 0.0173 0 0132 0.0117 0.0095 0 0070 0 0060 0 0057 6 6280 0098 0.0827 0.0374 0.0192 0.0102 0 0090 0 0115 0.0072 0 0055 0.0047 0 0038 0 0027 0.0025 0 0023 6 7362

PAGE 137

y La Ce Pr 1.0000 1.0000 1.0000 1.0000 0.4036 0.7777 0.6227 0.5201 0.1662 0.4572 0.3209 0.2505 0.0635 0 2687 0 1659 0 1 212 0.0200 0 1488 0.0769 0.0511 0.0052 0 0539 0 0258 0.0154 Nd Appendix 3.4 (Continue d ) Results of Experiment III Sm Eu Gd Th Dy Ho Er Tm Yb Lu pH 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 4 6351 0.3853 0.3745 0.4482 0.3782 0.3560 0.3396 0.3037 0.2752 0.2310 0 2483 5.9756 0.1636 0.1567 0.2017 0.1606 0.1385 0.1325 0.1119 0.0970 0 0759 0.0814 6.2190 0.0663 0.0615 0 08 3 5 0.0615 0.0501 0.0460 0.0373 0 0303 0.0236 0.0246 6.4049 0 0238 0 0212 0.0297 0.0194 0.0156 0.0131 0.0108 0.0083 0 0067 0.0065 6.5807 0.0065 0.0057 0.0088 0.0050 0.0043 0 0031 0.0028 0 0019 0.0018 0 0016 6.7666

PAGE 138

Appendix 5.1 Experimental Data for Figure 5.1 .rs) 0 24 25 25.5 26 27 28 29 30 31 32 109.0 80.88 78.59 74.23 73.79 69.32 65.73 104 .0 67.81 64.79 62.19 58.86 54.50 50.96 105 0 50.72 47 77 43.79 39.8 0 34.76 29.92 106 .0 46.22 43.78 39 .86 35.30 30.00 25.65 106 .0 46.85 44.31 39.96 36.68 3 1.06 26.92 60.93 58.53 57.55 55.48 46 .59 43.26 39.10 37.54 26.56 23.20 20.05 18.48 22.47 19.08 16.32 14.84 23.11 19.50 17.49 15.58 1 06.0 34.98 32 .22 28.30 25.55 20 .3 5 16. 64 1 3.46 11.34 9.38 107.0 39.05 36.17 31.67 28.03 23.11 19.26 16.05 13.4 8 11.45 107 0 51.36 49.54 44 .83 41.94 3 6 .27 30.92 27.93 24.40 21.51 107 .0 50.40 48 68 43 .76 40.77 34.88 29.75 26.75 23.33 20.44 107.5 54.72 52.03 47.84 44.94 39.34 34.40 31.28 27.20 24 .3 0 107.0 65.27 61.74 57.46 55.64 51.04 44.94 43.76 38.52 34.88 107.0 68.16 64.20 61.20 59 .38 54.03 48.36 46.97 41.19 38.73 105. 0 66.46 63.53 59.01 56.38 51.24 45.78 45 .26 39.58 36.86 106.5 62.41 58.47 55.49 52.29 46 .65 41.11 40 .2 6 35.14 31.74 105 0 67.20 64.68 61.53 59.43 53.45 48.30 47 .67 41.48 38.64 8.1 4 10. 1 3 19.47 18.73 22.36 33.49 36. 59 34.55 29.71 3 6.33 33 34 53.19 52.54 35.36 33.59 17.0 1 15.44 13.46 1 2.30 14.4 2 12.93 46 43.71 27.04 11.55 8.97 9.77 47 47.5 43.38 37.71 25 69 20 .07 10.3 5 7.27 8.04 5.30 8.87 5.81 7.32 9.12 18.19 17.23 20 .8 5 3 1 .35 34.35 32.65 28.0 1 34.65 6 .53 5.00 4.25 2 .63 8.24 6 .10 5.45 3.49 17.01 1 3.05 11.98 8 52 15. 94 12.6 3 19.1 3 16. 02 29.64 25.47 32.74 28. 68 30.97 26.88 26.94 22.58 11. 56 14.62 24.18 27.18 25.73 21.30 8.12 10.59 1 8.40 21.19 19.63 15.97 32.86 29.40 27.62 21.21 48 30.63 15.29 5 .08 3.52 3.77 1.77 2. 1 9 5.97 5 .67 7.5 5 1 3.9 1 1 6.37 14.9 1 12.03 16.38 2 1 :

PAGE 139

Appendix 5.2 Experimental data for Figure 5 2 [oM/ + ] [Mj+] (ppb) time(hours) 0 5 6 9 10 11 12 22 24 27 y 152. 33 1 21.41 131.77 109 07 100.39 99 93 99 6 2 82.87 77 .54 72.81 La 1 98 33 164.02 171.75 138 .04 132 88 132.88 127.73 105 .91 97.38 92. 03 Ce 402 .67 283.88 287 .51 2 19.46 204 .96 196 .91 1 88 .45 142 .54 127 24 116 .77 Pr 53.57 37 12 37 .71 2 7 .64 25 .07 23 57 23 .04 16. 29 14. 25 12.96 N:i 193 1 7 131.55 136 .19 98 32 93.11 83.64 82.48 58 .92 51. 77 47 .91 Pm Sm 38 .57 23 .0 3 22 80 14.81 13.65 12.23 11. 88 7 7 1 6.71 5 .82 Eu 8.37 5 26 5.05 3.42 3. 1 0 2 85 2.73 1.83 1.57 1.43 Gl 41.27 26.50 27.49 20 1 4 17. 75 17.04 16.4 3 11.89 1 0 .52 9 .70 Th 7.24 4 75 4 .69 3.41 3 .01 2 86 2 .90 2 .06 1.80 1.6 6 Dy 2 8 83 19. 72 19.84 14.41 13.0 9 12.89 12. 69 9 .11 8 22 7 61 H o 7.07 5 .11 5.42 3 95 3 .90 3.80 3 77 2 .91 2 53 2.39 Er 19.47 14.45 15.07 11.74 11.12 10.86 10 .90 8 55 7 .81 7 .11 Tm 3 3 2 37 2.49 1.95 1.80 1.84 1.80 1.40 1.20 1.14 Yb 17.95 12.21 1 3 03 9.85 9 35 8 99 8 .96 6 84 6.00 5.58 Lu 3.2 2 27 2 42 1.83 1.74 1.77 1.72 1.37 1.24 1.14

PAGE 140

s .) 0 109 108 108 109 109 108 111 108 109 109 109 109 110 109 109 23 94.9 91.9 85.3 84.8 83.5 77. 8 79.4 85.1 84. 1 86.2 88.6 89.5 89.6 88. 1 89.9 23.5 93.6 89.3 84.0 82.4 78.2 74.9 76. 1 8 0.3 80.0 80.9 83.8 84.1 85.3 82.0 83.5 24.5 93.6 87. 2 81.5 80.2 77.3 70.8 74.5 78.3 76.8 78.9 81.9 82 5 82.7 80.7 82.9 Appendix 6. 1 Experimenta l Data for Figure 6.1 a 25.5 26.5 27.5 28.5 29.5 30. 5 3 1.5 32 5 33.5 94.3 91.7 93.0 92.8 93.8 89.7 89 6 88.7 90. 1 87.5 85.7 85.4 86.3 87.0 83.0 82.5 8l.l 80.9 81.6 80.5 79.9 78.3 77.6 75.6 73 1 71.5 70.9 79.0 77.0 77.9 7 5.8 75 1 72.2 70.3 69.0 67.7 78.5 76.4 77.5 74.8 74.0 71.3 70.3 66.8 67.7 72.6 71.6 71.8 69.5 67.2 64.9 62 1 60.5 59.0 74.7 7 4.3 74.0 70.8 69.9 66 6 64.8 63.1 61.9 80.3 78.8 80.1 77.0 77.4 74.1 72.0 70.0 70.0 79.8 78.7 78.7 76.9 76.9 74.3 71.8 70.1 69.6 81.6 79.9 82.3 79.5 78.7 75.8 74.3 73.0 71.6 85.7 85.1 85.2 83.5 83.2 80.2 78.7 77.9 77.7 85.7 86.5 87.3 85. 7 85.3 83.2 8 1.0 80.4 79.2 85.9 86.0 86.4 84.2 85. 3 83.0 80.6 79 1 78.6 85.0 8 4 .8 85.5 83.9 83.8 81.9 78.8 78.0 77.3 86.0 85.6 86.2 84.0 83.9 83.0 80.5 78. 7 78.7 47 53 71 77.4 78.5 73. 1 71.8 7l.l 62.4 59.6 57.0 47.0 56.3 52.9 42.6 56.5 52.9 42.6 46.4 41.9 31.7 50. 1 45.4 33.8 60.4 55.0 44 6 58. 7 54.3 43.1 63.5 58.4 47.6 71.0 65.9 55.3 73.2 68.8 59.2 72.9 68.4 58 .9 71.0 66.4 55 .7 75.3 69.1 60. 1 77 71.5 61.0 44.7 41.0 40.5 30.0 32. 1 43 9 42.1 45. 8 55.4 58.2 57 9 54.7 58.5

PAGE 141

r s. ) 0 107 107 107 108 108 108 110 107 109 109 108 108 108 109 108 23 101 101 101 103 101 23 5 96 8 96.4 94 6 94 6 94.4 101 93 9 102 96 3 101 94.4 102 95.3 102 94 2 102 95. 2 101 95. 5 102 95 8 100 94 9 101 94.2 Appendix 6.2 Experimental Data for Figure 6.1 b 24 5 25 5 26.5 27 5 28 5 29.5 30 5 31.5 90 2 88. 3 87.1 87 6 84. 5 8 2 9 80.4 81. 8 76 1 66 3 60 0 58.4 53. 2 48 8 45 3 43.3 67.4 56 0 49 0 46. 3 40 6 36 3 33. 3 31. 2 67 0 54 5 47.8 45. 6 39. 8 35 8 32 2 31. 0 67. 2 57.0 51.0 48. 8 42 7 38 9 35.8 34.8 63.2 53.4 47 8 44 6 40 2 36 5 33. 9 31.6 68 3 58 7 53.5 51.9 46 0 42.5 39 1 38. 6 74 2 67 7 62 6 61.6 56 5 54 0 50.4 50.0 77 0 70. 9 67 9 66.4 62. 3 58 7 56 3 55 2 79.4 76.0 72.7 73. 1 67 6 65 3 62 7 63.0 84 1 82 1 80 0 80 1 75.8 73 5 71.7 71.6 87. 1 84.2 83. 6 84. 1 80. 5 78 1 76.2 76 5 86. 8 85.5 84.3 84. 0 81.0 79 9 78 3 78.4 86 8 85.6 84 3 85. 1 82.4 79 6 77.2 77.7 87 6 87.3 85. 8 86.8 82. 5 81.7 79.8 80.3 32 5 82.4 42.3 29 5 30. 2 33 0 30 1 37.9 49 0 54.4 63 1 71.6 76 3 77 3 77 6 79.4 33. 5 47 53 80 0 74 0 75.7 39 0 31.5 30 2 27 7 21.0 19. 5 27 6 21.2 19. 6 31.3 24.7 23.0 28. 7 22.8 35. 3 29.9 47 4 42 1 52 7 48 3 60 2 57.3 69 5 68.3 74 0 73.5 75 6 75.2 75 7 76 3 77.7 79.8 21.2 27 9 39 5 46 1 55 1 65. 8 72. 1 73.4 74.1 76 9 71 77 74 6 72 9 26.3 24 2 1 6 8 15.0 17.0 15.4 19.9 18. 0 18.7 24.7 36.5 43.1 52 6 63 6 70 1 72.3 71.2 76.3 17.4 22.9 35. 2 40 9 49 7 61.2 67 6 69 9 69 9 73. 7

PAGE 142

time(hrs.) 0 y 103.7 La 100. 9 Ce 100.7 Pr 103.0 N:l 101.3 Pm Sm Eu Ql Tb Dy Ho Er Tm Yb Lu 102 7 104.3 102 0 101.0 104.3 102. 7 105.3 100 9 103.0 101.3 Appendix 6.3 Experimental Data for Figure 6.lc 83.7 85.5 72.8 66.2 64. 1 1 5 2 4 76 0 76.6 75.1 81.5 82.9 83.2 66 7 65 8 64.0 59. 1 58 0 55.4 55. 5 54.4 51.4 5.5 17.5 74.0 67 5 82.7 78.5 63.2 57. 3 53. 9 47 9 49. 7 43 1 18.5 60.3 73 1 53 0 41.3 37 1 20. 5 58.1 73.4 51.4 40.2 35 6 21 53.4 70.9 46.4 35.1 31.4 45 .2 36.6 34.0 30.0 28.3 21.6 17. 0 15.3 12.1 46 .0 37.5 33.9 29.9 28.4 21.7 16. 8 15.3 11.5 57.2 49.3 46.4 42 1 40 5 32 7 27.5 25.4 21.2 51.6 43.4 40 5 36.3 34 6 26.7 21.6 19. 7 15.8 55 2 47 7 44 1 40 1 38. 2 29 8 24.3 22 7 18. 2 62.6 56 0 52.9 49.4 47 8 38. 7 33.6 31.8 26.7 65.4 58 9 56 9 53. 0 51.9 42 9 36.8 35 8 30.6 61.5 55.9 52 8 49 9 48 8 40.6 34.8 33.3 28.5 58.3 50 3 47.4 44.5 43. 6 36 1 30 3 29 0 24 3 60.4 54.2 51.5 49.2 48.8 41.5 36.0 34.7 29.6 22.5 23.5 25 5 26 26.5 50 1 68.9 44 2 32.5 28.4 47 1 45 9 46 6 43.9 65 7 64.4 64.3 63.8 41.3 38.8 39.1 38.4 29.7 28.1 28.1 27.4 26 0 23.4 23.8 23.2 10.3 8.99 10. 0 8 62 18.8 16.9 13. 9 12.0 15. 9 14. 1 24.1 21.5 27.8 25.3 25 6 23.4 21.5 19.6 26 7 24 7 7.91 7 69 7 62 7.62 15. 1 15. 0 10.7 10.6 1 2.3 12.4 19.4 19.2 22.9 22.3 21.0 20.8 17. 5 17.4 22 1 22.3 7.52 7.32 14.6 10.3 12. 0 18. 8 22.2 20 6 17.1 21.8 28 44 9 63. 6 37 6 26.5 22 3 7.16 6.93 13. 6 9 62 11.2 17. 6 21.0 19.2 16.3 20.7

PAGE 143

Appendix 6.4 Experimental Data for Figure 6.1 d [ oM/+ ] [M/+] (ppb) time(hr s.) 0 3 3.5 4 5.5 7 22 23 24 27 y 103.7 91.3 90.6 88.4 87.5 86.0 82.9 78.7 72.6 71.2 La 100.9 81.6 76.6 76.6 73.3 72. 0 65 6 60.8 56 .6 48.8 Ce 100.7 74.4 67. 0 66.2 62.6 60.6 48.2 44.2 39.9 32.8 Pr 103 .0 73.5 66.3 66.0 60.6 58.9 46.0 41.2 35.8 28.8 Ni 101.3 75.4 68 5 66.0 62.0 59.0 45.2 40.3 35.4 27.9 Pm Sm 102.7 71.4 62.3 59.7 53.2 49.8 32.4 27.4 23.8 18.2 Eu 104.3 74.6 65.3 62 7 57.4 53.0 35.2 30.4 26.6 20.1 Gi 102. 0 82.2 73 .8 72.5 66.0 63.0 46.5 41.8 37.3 30.3 1b 101.0 81.9 75.8 73.6 67.6 65.2 48.8 43 .2 39.1 31.3 Dy 104.3 86.0 80.3 79.1 73.5 70.5 57.1 48.9 45 .2 37.8 Ho 102.7 88.7 8 4.1 84. 5 78.2 76.9 65.5 61.8 56.4 50.8 Er 105 .3 92.4 88. 5 89.5 83.9 80.7 72.3 67.0 63.0 56.7 Tm 100 9 89.8 87.6 89.2 82.4 80. 3 74.3 66.7 65.4 60.5 Yb 103 0 91.6 86.6 87.5 83 2 80.3 72.9 66.4 63.9 58.4 Lu 101.3 90.1 88.3 89.6 84.3 81.3 76.8 70.6 67.8 63.2

PAGE 144

VITA Xuewu Liu received his B.S in marine c hemistry from Ocean University of Qingdao, Qingdao China in 1986. H e s tarted hi s g raduate s tudy in Department of Marine Science of University of South Florida in 1992. His research interest is on the physical chemis try of natural waters. He is currently a member of American Chemical Society and Geochemical Society.


printinsert_linkshareget_appmore_horiz

Download Options

close
No images are available for this item.
Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.