Pressure dependence of gadolinium carbonate complexation in seawater

Geochimica et Cosmochimica

Acta,Vol.58,No.19, pp.

4009-4016,

1994

Copyright 0 1994 Elsevicr Science Ltd

Pergamon

Printed in the USA. All rights reserved 0016-7037/94 $6.00 + .OO

0016-7037(94)00191-X Pressure

dependence of gadoli~um carbonate comp~e~tion in seawater JONGHYEONLEE and ROBERTH. BYRNE

Department

of Marine Science, University of South Florida, St. Petersburg, FL 33701, USA

(Received August 2, 1993; accepted in r~i~ed~o~ April 2 1, 1994)

Abstract-The influence of pressure on Gd carbonate complexation has been investigated at t = 25’C and P = l-600 atm. in synthetic seawater using a tributyl phosphate solvent extraction technique. The pressure dependencies of Gd carbonate stability constants appropriate to seawater ( t = 25°C and S = 35) can be described as follows: I~~~~~(~)

= 5.19-1.30 x 10_3(P-

1);

log g3&(Gd)

N 9.17-1.5 x 10-3(P - I);

log ~~3&(Gd)

= 1.58-3.73 X 10-4(P - l),

where F/$,(Gd) = [GdL,J[Gd3’]-‘[ L]Tn, [ ] and [ IT represent free and total concentrations of indicated species (molal scale), and P is pressure in atmospheres. In seawater with a total alkalinity of 2.4 mmol/kg and total CO, of 2.1 mmollkg, the fraction of total Gd present as free Gd3+ increases by a factor of approximately five or more as pressure increases from t atm. to 500 atm. at 2°C. Gadolinium speciation calculations indicate that changes in [CO:-]T and pH are dominant factors controlling the distribution of Gd species in upper oceanic column (O-1000 m depth). Below 1000 m water depth, the effect of pressure on Gd carbonate complexation exerts a major influence on the distribution of Gd species. INTRODUCIION EXAMINATION OFRAREearth element (REE) behavior in the environment and in the laboratory provides exceptional opportunities to quantify relationships between trace element environmental d~~butions and fundamen~ chemistries (BENDER,1982; ELDERFIELD,1988). While REE distributions have been determined in a very wide variety of environments (DE BAAR et al., 1983,1985a,b, 1988; ELDERFIELD and GREAVES,1982; ELDERFIELDand SHOLKOVITZ,1987; ELDERFIELD et al., 1990; GERMANand ELDERFIELD,1989, 1990,GREAVES et al., 1991;GOLDSTEIN and JACOBSEN, 1988; HOYLEet al., 1984; KLINKHAMMER et al., 1983; PIEFGRAS and JACOBSEN,1992; SHOLKOVITZ and SCHNEIDER,199I ; SHOLKOVITZ et al., 1989), aqueous equilibrium characteristics of the REEs are presently understood only under conditions appropriate to the earth’s surface: the REEs have speciation schemes su~~nti~ly dominated by carbonate complexation (CANTRELL and BYRNE,1987b;BYRNEet al., 1988; LEEand BYRNE,1993) , and while REE carbonate complexation has been examined in terms of temperature ( CANTRELL and BYRNE,1987a), there has been no study documenting the influence of pressure on REE carbonate complexation. Direct rn~su~rne~~ of the influence of pressure on REE equilibria have been obtained only for REE complexation by sulfate ions: LaSO$ (FISHER and DAVIS, 1967) and EuSO$ (HALE and SPEDDING,1972). The partial molal volume changes reported in those studies indicate that REE sulfate stability constants decrease with increasing pressure, up to 40% at 500 atm. In view of the effect of pressure on REEsulfate complexation, we expected that pressure might exert a major influence on REE carbonate complexation equilibria.

in the present work, using a solvent extraction technique, we have examined the carbonate complexation of Gd as a function of pressure in synthetic seawater. Our work constitutes the first experimental measurement of REE carbonate complexation as a function of pressure, and allows assessment of REE ~uilib~um behavior #rou~out the oceanic water column. Our characterization of REE equilibrium chemistry as a function of both temperature and pressure is apparently unique for trace metals in the ocean. THEORY The ~uilib~um pa~tioning of Gd between an organic phase, TBP, and an aqueous phase, synthetic seawater, can be characterized by organic/aqueous distribution coefficients ( CANTRELLand BYRNE,1987b; LEEand BYRNE,1993). At constant temperature, pressure, ionic strength, and perchlorate ion c¢&on, the affinity of tributyl phosphate for rare earth ions f h/I’+) is constant (LUNDQVIST,1982, CANTRELLand BYRNE,1987b). The affinity of rare earth ions for an aqueous phase is dependent on the extent of the aqueous solution complexation of each REE. Thus, under the conditions of our experiments, variations in the distribution of Gd between TBP and aqueous solution is directly dependent on aqueous phase carbonate and bicarbonate ion concentrations (LEE and BYRNE,1993). Distribution coefficients in the absence of carbonate ions (Do) and in the presence of carbonate ions (D) can be expressed as D

=

[GdiiJr and Ed’+ 11:

Do==, T

(1)

where [Gdz&], represents the total concentration of Gd3+ in the TBP phase, and [ Gd3’]t and [ Gd3+J7 represent, respec-

4009

J. H. Lee and R. H. Byrne

4010

tively, the total aqueous Gd concentration in seawater in the absence and presence of carbonate ligands: [Gd3+lT = [Gd’+] + [GdCl’+]

+ [GdSO:] + [Gd(CO,);]

(2)

= [Gd3+] + [GdC12+] + [GdSO:].

(3)

+ [GdHCO:+]

+ [GdCOf]

and [Gd3+];

Equations 2 and 3 can be written in terms of complexation constants T]“fl.(Gd) as [Gd3+lT = [Gd’+]( 1 + g&(Gd)[Cl-]

represent E-ee and total concentrations of indicated species. Note that values of gfi, (Gd) and &#Ir (Gd) are defined in terms of the free chloride and sulfate concentrations, while valuesoff;“co,8l(Gd),&~1(Gd),and%,B2(Gd)aredefined in terms of the total bicarbonate and carbonate concentrations. When pressure is applied to seawater, the relative concentration of each species in Bqns. 2 and 3 varies in response to changing ligand concentrations and changing stability constants ( Eqns. 4,5 ) . The general expression describing the effect of pressure on stability constants (ignoring compressibility terms) can be written (MILLERO,1983) as “L”fl,(Gd) = ~a~(e-(AV;/RT)(P-l)),

+ %,LMGd)PO:-1

+ %,PdGd)[CO:-IT

+ ~,P2(GdNCO:-l~)

(4)

and [Gd3+]$ = [Gd3+]( 1 + gn”pr(Gd)[Cl-] + %,,,B,(Gd)PO:-I), wheret”/3,(Gd)=[GdL.][Gd’+]-‘[L]-”,and[

DO -= D

(6)

+ Kco,B,(Gd)WCO;lT

(5)

]and[

1~

where S,J/3; is a stability constant appropriate to Gd in seawater at 25°C and 1 atm. total pressure, T is absolute temperature, R is 82.057 cm3 atm deg-’ mol-’ , and AV z is the partial molal volume change (cm’/mol) for Gd complexation with a given ligand L in seawater. Using Eqns. 1 and 4-6, the equilibrium partitioning of Gd 3+ between TBP and syntheticseawater,expressedasD’/D= [Gd3’]r/[Gd3’]$,can be characterized by the equation

1 + ~;B~(~-‘“Y~RT”P-“)[C~-]

+ & Bt(e-‘AY~,/RT”P-I))[~O:-] + EC0 Bt(e-‘AVi”llRT”P-‘))[HCO;]~ 1 1 + ~;B~(~-‘“Y”IR~~‘~-“)[c~-] + ~,~t(e-‘“Y~,/R~“P-I))[so~-] -(A~,JR~)(P-~))[co:-]~

+ ~~8:(e-(“Vi~,21R~(P-I))[co~-]:)

1 + 88t(e-‘“V’“IRT”P-“)[~1-]

+ ~,,,B~(~-‘“Y~,IR~(P-‘))[so:-]

(7) ’

where [HCOSIT = [HCOr]

+ [NaHCO!] + [MgHCO;]

(8)

+ [CaHCO;],

and [CO:-]r

= [CO:-]

+ [NaCO;] + [MgCOt]

+ [CaCO~].

(9)

Note that, at 25°C and 1 atm. total pressure, Eqn. 7 can be rewritten as

EL,+ D

Kco~PI[HCO~IT + F~,B~[CWIT + %&mwl:

1 +~~~[C1-]+&3~[SO:-]

f

(10)

= Ko’“K~co,lW+l~

(11)

and [CO:-IT = &““fG”~2&0,/W+1:,

[HCOTIT = [CO:-IT

(CA[H+l,MW+l,

= (CAs”K2)/([H+lT

( 12)

where PCO, is the CO2 partial pressure in the aqueous medium, [H +]r is the total hydrogen ion concentration ([ H+17 = [ H ‘1 + [ HSO;] ), K. is the Henry’s law constant, and ‘wK, and SW K2 are the first and second dissociation constants of carbonic acid appropriate to seawater conditions. The val-

+

2”Kz)

+ 2”K2),

(13) (14)

in which [H+]r at in situ pressure can be calculated from the ratio of carbonate alkalinity (CA ) to total CO2 (CT ) : CA/C-r = (s”‘K,[H+]T + 2”K,““K2)/ ([H+]:

Equation 10 was used to estimate the constants $o,/3f, &$? f , and F&Y: appropriate to seawater at 25°C and 1 atm., using calculated values of s/3 ! , &,j3 t, and distribution data intheform(D”,D,[HC0~]r,[CO~-]r)at250Cand latm. Total bicarbonate and carbonate concentrations at 25 “C and 1 atm. were calculated, using the expressions [HCO;h

ues of sWKl and ‘“K2 were obtained using the algorithms of DICKSON and MILLERO (1987 ) . K. was taken from WEISS (1974). Total bicarbonate and carbonate concentrations in our pressurized solutions at 25“C were calculated using the expressions

+ IWK,[H+lT + sWK,SWK2). (15)

The influence of pressure on the carbonic acid and bicarbonate dissociation constants was assessed using the model and equations of MILLERO ( 1979 ) . Sulfate complexation constants with Na + , Mg2+ , and Ca *+ at 25’C and 1 atm. ( KESTER and PYTKOWICZ, 1968, 1969; PYTKOWICZ and KESTER, 1969) and AV* values for sulfate ion pairs with Na+, Mg2+, and Ca*+ ( MILLERO, 197 1) were used to estimate [SO:-] as a function of pressure at 25°C. The Gd sulfate seawater stability constant at 25°C and 1 atm., &,,@i(Gd), was calculated as 99.6 from the thermodynamic constants of POWELL( 1974) and SMITH and MARTELL (1976). Currently, no experimental data exist for the partial molal volume change of Gd-sulfate complexation in seawater (A V &,)_ HALE and SPEDDING ( 1972) presented AV&, values for Eu3+-SO:- complexation at zero ionic strength ( A V go, = 25.6 cm’/mol). Assuming similar AV”

Gd carbonate complexation in seawater

values for EuSO: and GdSO: complexes and using the molar volume estimates, P”(Gd3+) = -53.9 and P*( SO]-) = 22.9 (MILLERO, 1969, 1977), we calculated ti”(GdSO:) = -5.4 cm3/mol, Using the relationship A~(trans) = p* - 9’ = 1.2 f 0.2 cm3/mol suggested by MILLERO ( 1977) for monovalent cations, the partial molal volume of GdSOt in seawater, v*( GdSOt), can then be estimated as -4.2 cm3/mol. From the values of p*(Gd3’) = -46.2 and r*( SO]-) = 29.7 (MILLERO, 1977), the AV& value appropriate to 35% seawater at 25°C is subsequently calculated as 12.3 cm’/mol. The Gd chloride seawater stabihty constant at 2S’C and 1 atm., ?J#t(Gd), was calculated as 0.3 from the thermodynamic constant of MIRONOV et al. ( 1982). In the absence of A V (: data appropriate to GdCl 2+ formation in seawater, and in view of the minor intluence of GdCl 2+on Gd seawater speciation (BYRNE et al., 1988), we assumed AV& = 0. Using our distribution data (P, Do, D, [HCO;]r, [ CO$-17 ) in Eqn. 7, nonlinear least squares fitting techniques (CANTRELL and BYRNE, 1987b; LEE and BYRNE, 1993) allowed estimation of A V &,,, AV &,, and A V ~~~~~appro-

priate to seawater at 25Y. Use of Eqn. 7 in nonlinear least squares analysis (Marquardt method, SAS Institute inc., 1989) involved minimization of the residual sum of squares n function RSS = T f 1 - (~/~“)~(f~3’]~/[Gd3~f~~~ ), the ratios ([Gd3+]~/~Gd3+]~)i are calculated using Eqn. 7, and (D/DO)i are directly measured distribution coefficients. where

EXPERIMENTAL

SECMON

The solventextractionproceduresused to examine the distribution of ‘53Gd between TBP and synthetic seawater are very similar to those used previously in studies employing 0.7 M NaClO, as the aqueous medium ( CANTRELLand BYRNE,1987a;b; LEE and BYRNE, 1993). Noting that the effect of pressure on dissociation constants of carbonic acid has been directly estimated in seawater ( MILLERO and BERNER,1972, MILLERO,1979), we utilized a synthetic seawater as the aqueous phase in this work. By conducting our inv~tio~ in seawater, Gd carbonate stability constants can be directly expressed in terms of total, rather than free, carbonate ion concentrations. Our synthetic seawater contained perchlorate ions, since ClO; is prerequisite for REE extraction into TBP. The composition of our synthetic seawater (mol/kg HzO) is:

[Na+]r = 0.486 [SO:-],

= 0.028

[MgZ+lT = 0.053 [Cl-],

= 0.032

[Ca’+]r = 0.010 [ClO&

= 0.524,

Our titrant solution consistedof0.34 m NaHCG, (Aldrich Chemical Co., A.C.S. resgent) in synthetic seawater. Thus, the total concentrations of Mg2+. Ca”. SGf. CI-. and ClO; were constant on addition of tit&t ‘solution, while fNa+fT varied between 0.486 and 0.5 10 ma&kg HzO. Total bicarbonate and carbonate wn~ntmtions inoursolutionsvariedwithintheboundsO~fHCO~~<5.6X low2 and 0 5 [CO:-]r < 3.1 X lo4 moI/kg HsO. Procedures for the preparation/purigcation of TBP and aqueous solutions are explained in detail in CANTRELLand BYRNE ( 1987a,b) and LEE and BYRNE (1993); prior to each experiment, stock solutions composed of deionizeddistilled water and NaClO., (Aldrich Chemical Co., anhydrous, 99-l%) were stored at least 24 h at pH 8.5-9.5. These solutions were then filtered with a 0.45 pm polycarbonate filter (Nuclepore Corp.) which had been acid cleaned and rinsed with deionized H20. TBP (Fluka, put&s p.a.) was treated as follows prior to each experiment. Equal volumes of TBP and IM NaOH were shaken vigorously for 5 min in a separatory funnel. The TBP was next centrifuged at 12,000 rpm for 1 min. Subsequently, the TBP was preequilibrated with 0.68 m NaCIO, by combining equal volumes of

4011

0.68 m NaCIOh and TBP and stirring ~gomu~y for 5 min at 25°C. The NaClO, solution was then separated from the TBP and the preequilibrated TBP was subjected to the same treatment with fresh 0.68 m NaClO+ Two types of experiments have been conducted in this work, In the first type, is3Gd distributions between TBP and seawater in a jacketed beaker were examined as a function of bicarbonate and carbonate ion concentrations at 25°C and 1 atm. total pressure. The equilibration procedures used in these experiments closely followed the procedures employed by CANTRELLand BYRNE( 1987a,b) and LEE and BYRNE( I993 ) to investigate REE carbonate comnlexation in NaClO,. In the second type of experiment, we performed TBP/seawater ~uilibmtions at 2S’C and I, 200,400, and 600 atm. in a h&h pressure housing. Our hi~-p~u~ housing is similar in structure to the pressure housing used by ACKER et al. f 1987) to investigate CaCC&(s) dissolution rates in natural seawater. The housing consisted of a 50 cm3-capacity inner cell placed within a stainless steel pressure chamber. The inner ceil consisted of a precision bore glass cylinder with double O-ringed plastic end-windows. This arrangement allows the windows to adjust in response to pressure (ACKER et al., 1987), while isolating the mixture of TBP and seawater within the cylinder. In these experiments, TBP and seawater (plus titrant solution) were equilibrated in a jacketed beaker by stirring and bubbling with a 30% C02-70% N2 &as mixture prior to addition of is3Gd. Following this equilibration, 20 min were allowed for phase separation, and pH measurements were taken to calculate the aqueous phase total alkalinity and total CO2. Since our solutions contained no borate ions, total alkalinity (TA) and carbonate alkalinity (CA) were very nearly identical in our studies. Aqueous and organic phases (50~50) were quickly loaded into the inner glass cell, and 5 - 10 L of %d (Isotope Products Laboratories) was added to the mix&e. The cell was then loaded into the high pressure housing, pressurized, and shaken at 140 cycles/min for approximately 2.5 h. Control experiments showed that Gd distributions between TBP and synthetic seawater reached an asymptotic value after 2 h of shaking and subsequently remained constant for equilibration times up to 24 h. Pressure was applied to the chamber by an Enerpac Model I I-100 stainless steel pump. Pressure measurements were obtained with an Enerpac Model 3 16 SS ( 20000 psi max. and *0.5% accuracy) pressure gauge. Gauge readings were fquently checked during the shaking period, and an average reading was taken for each distribution coefficient measurement. Subsequent to equilibration, the system was depmssurized and 1 mL samples of TBP and seawater were removed for gamma spectroscopic analysis. The pH (--log [ H+b ) of the remaining seawater phase was measured and the seawater was then rebubbled for IO min with the same 30% CC&-70% N2 gas mixture which was used for the previous equilibration at I atm. The pH values measured in rebubbled seawater were consistent (kO.0 1 pH unit) with the pH values measured prior to pressurization. This observation indicates that, in the course of our experiments, CA remained constant throughout the pressurization/depressurization process. However, slightly higher pH values, obtained immediately after depressurization but before rebubbling, indicated a slight loss in C, while transferring the solutions to the pressure chambers Consequently, the Cr values appropriate to our pressurized solutions were calculated from the CA and pH values determined immediately after system depressurization. In both types of experiments, two samples ( 1 mL each) were withdrawn from each phase for gamma spectroscopic analysis in a Bicron 2 X 2 inch well type Nal detector (2MW2Q/2) monitored with a Tracer Northern multichannel analyzer (TN- I7 10). The pH of the aqueous phase ( I atm.) was measured with a Ross combination pH electrode (Orion No. 870200) and a Coming model 130 pH meter. Measurements of pH in our study were obtained on the total hydrogen ion concentration scale ([ H+]r = [H ‘1 + [ HSOi]). RESULTS REE

AND DlSCUSSlON

Carbonate Complexation in Seawater at 2S’C and

I atm The distribution data obtained in our jacketed beaker experiments (Table I ) were used to determine Gd carbonate

4012

J. H. Lee andR. H. Byrne Table 1. Jackebzd beaker dktcihution data obtained in synlktic seawater for Gd CathoMte and bicarbonate complexation analyaia at 25oCand 1 atm. (m&l scale). ____-~---_-----____-____--_.________-”_ Exp PH ._--~.._.__D_..__!~~~---_!-~~~_-O.ooO 3.9941.7880.000 #l 4.7081.7105.9mxm4 3536x10~ 5.3401.5772.56&10-3 6LXhlx107 5.8201.0137.745xlU3 5.929x10-6 6.2850222 2.259x10-~ 5.043~105 #2

3.8491.7l30.000 0.000 4.4291.6923.1~7x10-4 9759x109 1.11ox107 4.956 1.676 l.O6&lV3 1.6gox106 5.546 1.449 4.lz3x1#3 1.497x105 6.021 0.701 123lxl(k* 6.413 0.131 3.O36x102 9.112x105 6.535 OD5t3 4.018x102 1.5%xlOd 6.678 0.020 5.594x10-* 31wJxlW

x3

4.473 1.6911o.ooo 0.000 55478 1.539 3.528x1&3 1231x106 5.794 1287 7.305x103 5.276x106 6.285 0.270 2259x10” 5.043~105 6.564 0.054 4293x10* 1.822xliP --_--------

and bicarbonate stability constants expressed in terms of total CO:- and HCO; concentrations. Table 2 summarizes the stability constant results obtained in synthetic seawater at 25’C and 1 atm. The error limits given with each experimental result represent the standard error and provide a comparative measure of the quality of each data fit. The error limits given with the average stability constant represent one half the range of individual experimental results. Our directly measured Table 2 seawater stability constants, log yo$ i = 5.19 and log %,/3: = 9.17, are in good agreement with the values of log yo$t = 5.27 and log &pi = 9.24 calculated from the co,81j(Gd) values of LEE and BYRNE ( 1993). Thus, our stability constants deter-n&e& in synthetic seawater are in accord with Z&@] predictions (LEE and BYRNE, 1993 ) derived from measurements in 0.7 M NaClO,, . Our bicarbonate complexation results, log &,# f = 1.6, are also in general agreement with previous estimates of CIAVATTA et al. ( 198 I), CANTRELL and BYRNE ( 1987a,b) and LEE and BYRNE ( 1993). The ratio of stepwise stability constants obtained in this study (K,/K, = &~~(~o,~~)-2 = 0.06) is in good agreement with the result ( K2/K1 = 0.06 f 0.02) obtained by LEE and BYRNE (1993)for Gd carbonate complexation in 0.7 M NaC104. Influence of Pressure on the REE Carbonate ~o~p~exation in Seawater The distribution data obtained in our pressure cell experiments are shown in Table 3. For comparison with the stability constant results shown in Table 2, a subset of the Table 3 data, results obtained at I atm. total pressure, were used in Eqn. 10 for determination of To,,/3f , &3p 4, and j&o, (I 1. The stability constant comparison shown in Table 4 demonstrates excellent agreement, and thereby indicates the efficacy of the equilibration procedures used in our pressure cell experiments.

Using our pressure cell dist~bution data (Table 3) and average Gd carbonate and bicarbonate stability constants (&@L and &o,Bf) determined at 1 atm. (Table 2), A v go,, and A v 7~0~)~ values were estimated from Av:co,, Eqn. 7 as follows AP’&,, = 73 + 39 cm3/mol;

(16)

A v Tcop~z= 139 f 53 cm3/mol;

(17)

AJ’&o,

= 21 t 92 cm3/mol.

(18)

Our AV Eo, result indicates that &&(Gd) decreases by approximately one-half for each 2,300 m increase in depth in the ocean. The AV Tco,t, result we obtained indicates that Fo&(Gd) decreases with ocean depth (pressure) substantially more rapidly than “c”o,/?, (Gd). Our AV &,, has no significant bearing on Gd’+ speciation in seawater since GdHCOF is a very minor species relative to carbonate complexes. The uncertainties on our Eqns. 16- 18 best fit estimates are large, in part, due to correlation of the estimated parameters. Thus, a smaller AV &,3,, value implies a larger A V zo, value, etc. A A V Iflcojfzvalue as small as 120 cm”/moI leads to a AV & value equal to 84 cm’/mol, and a A V itcstt value as large as 160 cm3/mol requires a A V zo3 value equal to 62 cm3/mol. In order to further examine the quality of our AV &, estimate, we performed an additional analysis in which Eqn. 7 was truncated by setting :&/3:[CO:-]$ = 0. We then examined the AV &, and AV gco, values provided by least squares analyses in which we successively removed the data points which exhibited the highest degree of solution complexation. Complexation intensity is indicated by the parameter Do/D. Figure 1 shows that as we successively removed our highest Do/D data points, until half of our data remained, AV &, values ranged between 55 and 85 cm3/moi. Our analysis provtded AV &Q results within the range 11 s AV gco, I 49 cm3/mol. Thus, our analyses using a truncated form of Eqn. 7 are in reasonable agreement with the A V zo, and AV &,) results summarized in Eqns. 16 and 18. The range of A V &, and A V Eco, estimates obtained in these analyses using the truncated version of Eqn. 7 is very much smaller than the uncertainties provided with the parameter estimates in Eqns. 16 and 18. Based on the results of our analyses using the truncated version of Eqn. 7, our recommended values and uncertainties in AV & and AV zco, at 25°C can reasonably be expressed as AV&

= (73 1: 17) cm3/mol;

(19)

Table 2. cd carbonate and bicarbonate stability constants (m&J sfaae) detennkd in syntktk seawater at 25oCand 1 attn. in the jacketed beaker exp&menta. Ratios of stepwise StabiRy constants wew calculated as K?/K,= z3j3; cZ3@: )-‘, Exp. #I #2 #3

%3J3f (1.61Si5)x105 (1.46zko.06)x105 (155&.32)x10’

Ave: (154#.OtJ)x1@

k$/K,

%3Sl

jzo3e:

(1.9%0.52)x10* (1.37~.03)x109 (1.1lio.20)x109

75k21 39i7 ozt30

OS@3 0.06 0.05

(1.5oHf.50)x109

38i38

0.06

Gd carbonate complexation in seawater

4013

Table3. Llistribution data for Gd carbonate and bicarbonate complexation obtained in eeawaber at 25oC and l-600 atm. pki values represent in situ values at each presww. Cr and CA represent, respectively, t&al CO? and carbonate alkahnity (molalscale).

presstue

pH

D

C,

labn latm. 1 atm. 1ab.n.

4.326 5555 5.886 6214 6.443

1.462 1.151 I.049 0.480 OS29

o.VUO

0.000

O.ooO

O.OW

8.729x103

1.139xlV3 1.835xlUa 3.289xlVa

4.422xlV3 5.953x103 1286XlV* 2633xlW

4.414x103 5.942xlV3 1.25lx1va 2.616x1@

3.69OxlW 5.29&&F 2.43lxlU5 8.4llxlW

2OOaS.m. 173 atm. 193 atm. 160 atm. 200 atm.

4.616 5.729 5591 6.101 6.445

1.568 o.VUU 1.2VS 1.27VxlW 0.861 2.06lxW 0.564 2.269xlVa 0.163 3.939xW

O.OUO 5.984xlV3 1.173xlW 1.533x102 3264xlU”

O.UOU 5976XlV3 1.17lx1v* 1.528xlVl3.240x10-’

0.0 4.158x106 1.199xW 2.485xlW 1.193x10-4

353 atm. 367 a&n. 380 atm. 357 attn. 357 atm.

4.336 5.613 5.923 6.155 6.376

1.643 1.407 0.898 0.514 0.229

o.OUo 126OxlV2 1.968xlU~ 3459xlVa 5.313xlW

0.000 5649x103 1235xlV~ 2.62lxN-P 4.413~102

OXKM 5.643xlV3 1232xlU2 2609xW 4.382xW

O.VtM 3.404xlW 1.53&&3s 5.957x1@ 1.552x10-4

6tKIatm. 547at.m. 613 atm. 587 atm. 613 atm.

4.4% 6.024 6.123 6.167 6.219

1.329 0.964 0.602 0.622 0.333

o.am 1537xlVa 3.48OxlV”3.9USxlU~ 5.766xlVa

O.VUO l.U95xlV” 2.674xW 3.057~102 4.65Ux1u~

0.00 1.Wlx1va 266lx1v* 3.042xlV2 4.624xlVa

O&IO 1.898xlV’ 6.V5lxlUs 7537x105 1.313x10-4

1

AL’&+

at-m.

= (2 1 -+ 28) cm3/mol.

(20)

In the absence of corroborating evidence for our Eqn. 17 AV ~co1j2 result, and in view of the large uncertainty (+53 cm3/moi) in this estimated parameter, we feel it is advisable to recommend only a lower bound estimate for AV &oat,. The estimate AV tcos)tl r 86 cm3/moI

,-o&(Eu)) (%o,B,(Eu), expressions,

can be related to , Fo,&(Eu)), i?o,,B~(Eu)

seawater constants using the following

5;“co,&(Eu) = ncoJX(Eu)*W;ISWK)~

(22)

%,&(Eu)

=co,B;(Eu).(1Y’tISWIY~)*(~~ISWIU2),

(21)

follows from the A V f-3j1 bounds in Eqn. I7 and is consistent with a AV* for the reaction GdCO; + CO:- ++ Gd(COr); equal to 13 cm3/mol. Influenceof Temperature on REE Complexation in Seawater

where K\ and K; are conditional carbonate dissociation constants appropriate to 0.68 m NaClO., at experimental temperatures. The term K’,ISWK, is nearly equal to one ( K’,/swK, = 0.969 at S = 35 and 25°C) and, following CANTRELL and BYRNE( 1987b), can, therefore,bc assumed approximately constant between 15 and 35’C.

CANTRELL and BYRNE (1987a) examined the temperature dependencies of Eu carbonate stability constants in 0.68 m NaC104 at 15, 25, and 35°C. Their conditional stability constantsin 0.68 m NaC104 (~~~~~(Eu~, ~~~~(Eu),

experWenta. The resultsobtahiedin our jacketedbeakerexperimenta (TableZ)amahownforcompa&m.

SF=uu e

21w,V: (1.39iV.Ol)xl~ -I___ (~*~~)xl~

The ratio

Ki/‘“Kz is calculated as 0.385, 0.348, and 0.324 at 15, 25, and 35’C, respectively, using K2 values of CANTRELLand BYRNE(1987a) and the swK2 values of DICK!SONand MLLERO( 1987), recalculatedand expressedon the free hydrogen ion con~ntration scale. The calcufated values of

Table4 Gdtxbnate and bhrbonate stability constants (molal scale) determined in synthtic seawater at 25oCand 1 atm. in the premurecell

EW -__-

(23)

,_,e: -------__(140#.01)x109 -___l-----_l(1.~~~1~

&3s:

Ka/Kr

9Sl

0.07

38s

0.06

4014

J. H. Lee and R. H. Byrne

- 70

- 66

- 60

55

5ot50 21

20

19

18

17

16

14

13

12

11

10

9

Numberof D&r Points

FIG. 1. Estimated AV & values are shown as a function of the number of data points used in Eqn. 7 when the highest Do/L) points are sequentially removed from the Table 3 dataset and the &&[CO$-1: term in Eqn. 7 is set equal to zero.

&&&,(Eu)and$&,(Eu)at 15,25,and35”CgiveninTable 5 indicate that the influence of temperature on REE carbonate complexation at constant pressure is insignificant. Our Table 5 results indicate that Ml = 0.4 kJ/mol and A& = -5.4 kJ/mol. This is consistent with the general observation ( CHOPPIN,1980; GRANT et al., 1988) that enthalpy changes for REE complexation are usually small. Since enthalpy changes for I:1 aqueous complexes of the REE cations are generally small, it follows that complexation of the REEs in solution is strongly driven by entropic effects (CHOPPIN, 1989). For the REEs, entropies of reaction appear to be strongly driven by the dehydration of the cation. BRI’ITA~N et al. ( 1992) have shown a general correlation between the entropy change of 1:1 complexation and the number of water molecules displaced from the primary coordination sphere of the rare earth ion by the l&and. Using the %J3, (Eu) and dff, values in Table 5, we calculated a reaction entropy for EuCO; formation equal to 109 J mol-’ OK-‘. From the correlation shown in BRITTAIN et al. (1992), this entropy value indicates that two to three water molecules are displaced from the inner Eu3+ coordination sphere upon E&O; forTable 5. Eu carbonate and bicarbonate stability constants calculated from the conditional stability con&ants of Cant&l and Byrne (1987a)using Eqns. 2325. Enthalpy values (AH) appropriate to each constant were cal* using the relationship log ‘;“&(Eu)= -AH/Z.XBRT + constant. ____________________-------_ ______ -- -...._.._--temp. log log 1% &61 (Eu) Z&@u) i%o$JI(W PJ 15 25 35

5.357 5.379 5.369

9.098 9.081 9.035

1.028 1.132 1.463

o&2.1

-5.&l.?

36.8rt12.1

mation. Thus, complexation of Eu(H,O)p by a single CO:- ion adds two to three inner sphere water molecules to the bulk solution. Loss of both inner sphere and outer sphere water molecules during complexation is likely an important aspect of the large partial modal volume changes observed in our study. Although Gd 3+complexation by carbonate ions at constant pressure exhibits only a small temperature dependence, Eqn. 6 expressed in the form log fl”&,(Gd) = log fly@; - &(P-

1)P

(25)

shows explicitly that the influence of pressure on complexation constants is itself temperature dependent. It should be noted, as well, that the term A V ? in Eqn. 25 may also exhibit a significant temperature dependence. The overalf partial molal volume change for a reaction can be represented as a difference between the partial molal volumes of individual products and reactants. Among the individual ion partial molal volumes which influence the reaction A V z values for the equilibria under consideration, only the temperature dependence of carbonate ion partial molal volumes appears to have been directly assessed in seawater. In this case ( MILLERO, 1983), p,-o, decreases from 13.2 cm3/mol to 7.4 cm3/mol between 25 and 0°C. This influence on A V z acts to increase A V &, by 5.8 cm’/mol between 25 and O”C, and would contribute to an i 1.6 cm3/mol increase in AV &o,),. Although the influence of temperature on the partial molal volumes of hydrated rare earth ion is apparently unknown, it can be noted that the influence of temperature on alkaline earth partial molal volumes ( MILLERO, 1983) is substantially smaller than the vco, temperature dependence in seawater. With respect to the contact ion pairs formed between Gd3+ and carbonate ions, MILLERO( 197 1) states that the partial molal volumes of contact ion pairs do not change much between 25 and 0°C. Consequently, it is most probable that,

Gd carbonate complexation in seawater

due to the influence of temperature on &,, partial molal volume changes accompanying the formation of Gd carbonate complexes at the low temperatures characteristic of deep ocean depths will be si~ifi~ntly larger than A V r values at 25’C. Gadolinium Speciation in Seawater BYRNE et al. ( 1988) considered the influence of temperature and pH on various metal speciation schemes in seawater at 1 atm. In the case of Gd, they showed that carbonate complexation accounts for 86-98% of total Gd at pH = 7.6-8.2 and t = 5-25’C. By including the influence of pressure on Gd carbonate complexation, we have constructed Gd speciation schemes appropriate to the natural range of oceanic conditions. sojourn inorganic solution complexation can be expressed as

[Gdhl[Gdl

4015

changes in Gd speciation at depths above 1,000 meters are due to the variations in pH and (consequently) [CO’,- ]7 with depth. Below the depths where pH and [CO:-], show substantial variation, changes in Gd speciation are dominantly due to the influence of pressure on &&(Gd). Pressure is seen to exert a very strong influence on REE aqueous equilibria in the oceans. It should then be expected that the aqueous chemistry of trivalent actinides will exhibit pressure dependencies of a similar magnitude. In view of the very large influence of pressure on the equilibria studied in this work, it would seem advisable to conduct examinations, in general, of the intluence of pressure on important trace metal equilibria in the oceans. Ac~ow~e~e~~-This work was supported by grant No. (XX90 10299 from the Chemical fishy Program of the National Science Foundation. The constructive -criticismof Johnson R. Haas, Frank J. Millero, and David J. Wesolowskiis gratefully acknowledged.

= 1 + Bo(Gd) + %,B,Gd)P:-I, Editorial handling: G. Faure + %,8*wwe-l:~

(26)

where ~~(~) = z&(Gd)[Cl-] + ~,~,(~)[SO~-] + $B,(Gd)[F-] + ““@:(Gd)[H+]-” and “@(Gd) = [Gd(OH)i-“][H’]“[Gd3+]-‘. The influence of temperature on Gd sulfate complexation (m = 4.4 kcal/mol) was adopted from BYRNE et al. ( 1988). The influence of temperature and pressure on the free sulfate ion concentration in seawater, [SO:-], was adopted from KESTERand PYTKOWICZ( 1970) and MIUERO ( 197 I ). In the absence of A V* data appropriate to GdC12+, GdF2+, and Gd( OH):-” formation in seawater, and in view of the minor influence of such species on Gd seawater speciation (BYRNE et al., 1988), the terms g&(Gd)[cI-1, $&(Gd)[F-1, and “8: (Gd) [ H’]-” were assumed to be independent of pressure. The AIII values for GdCl’+, GdF’+, and Gd(OH);f-” formation were adopted from BYRNE et al. ( 1988 ). Table 6 illustrates the effect of temperature and pressure on the distribution of Gd species in seawater with S = 35, C, = 2.1 mmol/kg, and TA = 2.4 mmol/kg. The influence of temperature on Gd speciation is quite small. The extent of Gd3+ complexation in seawater changes by less than 5% as the temperature is decreased from 25 to 2°C at constant pressure. In contrast, as pressure is increased from one to 500 atm. at constant temperature, the extent of Gd3+ solution complexation is seen to decrease by a factor of five or more. Use of our A V r data, and GEOSECS CO2 system data from the Atlantic and Pacific, indicates that in general, substantial Table 6. Effect of temperature and pressure on the distribution of Gd species in seawater with %3!$ CT 2.1 mmol/kg and TA= 2.4 mmol/kg_ -----1_ PH &I --

&m)

25

1 1 500

2 2 -_----

8.043 8.391 8206

1=&r (mM) 0238 0.217 0204

log’

~--rwr~Kw

---_

CaIcuIatedusingtbeAV&~Itsin Eqns. X9-21. l

2.09 2.07 51%

REFERENCES ACKERJ. G.. BYRNE R. H., BEN-YAAKOVS., FEELYR. A., and BlX&R P. R. ( 1987) The effect ofpressure on aragonite dissolution rates in seawater. Geochim. Cosmochim. Acta 51,2 17 1-2 175. BENDERM. L. ( 1982) Trace elements and ocean chemistry. Nature

2%,203-204. BRITTAINH. G., CHOPPING. R., and PARTHELEMY P. P. ( 1992) pH-dependence of the metal ion hydration state in lanthanide complexes of ~lyarnino~ly~~xy~~e ligands. J. Ccwrd.Chem.

26, 143-153. BYRNER. H., KUMPL. R., and CANTRELL K. J. ( 1988) The influence of temperature and pH on trace metal speciation in seawater. Mar. Chem. 25,163-181. CANTRELL K. J. and BYRNER. H. ( 1987a) Temperature dependence of europium carbonate complexation. & Sol. Chem. 16,5X+566. CANTRELLK. J. and BYRNER. H. ( 1987b) Rare earth element complexation by carbonate and oxalate ions. Geochim. Cosmochim.

Acta 51.597-605.

CHOPPIN G. R. ( 1980)

Lanthanide and Actinide Chemistry and Spectroscopy (ed. N. M. EDELSTEIN);ACSSymp. Series No. 131. pp. 173-181. Amer. Chem. Sot. CHOPPING. R. ( 1989) Chemical propertiesof the rareearth elements. In Lonthanide Probesin I&e, Chemical and Earth Sciences:Theoty and Practice (ed. J.-C. G. BONZLIand G. R. CHOPPIN),pp. l4 1. Elsevier.

CIAVATTA L., FERRID., GRENTHEI., SALVATORE F., and SPAHIU

K. ( 198 I) Studies on metal carbonate equilibria. 3. The lanthanum (III) carbonate complexes in aqueous perchlorate media. Acta

Chem. Stand. A35,403-4 13. DE BAAR H. J. W., BACONM. P., and BREWERP. G. ( 1983) Rareearth di~bu~ons with a positive Ce anomaly in the Western North Atlantic Ocean. Nature 301, 324-327. DE BAAR H. J. W., BACON M. P., BREWERP. G., and BRULAND K. W. (1985a) Rare earth elements in the Pacific and Atlantic Oceans. Geochim. Cosmochim. Acta 49, 1943-1959. DE BAARH. J. W., BREWERP. G., and BACONM. P. ( 1985b)Anomalies in rare earth distributions in seawater: Gd and Tb. Geuchim.

Cosmochim. Acfa 49, 196I- 1969. DE BAAR H. J. W., GERMANC. R., ELDERFIELD H., and VANGAANS P. ( 1988 ) Rare earth element distributions in anoxic waters of the Cariaco Trench. Geochim. Cosmochim. Acta 52, I203- 1219. DICKSON A. G. and MILLEROF. J. ( 1987) A comparison of the equilibrium constants for the dissociation of carbonic acid in seawater media. Deep-Sea Res. 34, 1733-1743. ELDERFIELD H. ( 1988) The oceanic chemistry of the a-earth elements. Phil. Trans. Roy. Sot. Land. A325 105-126. ELDERFIELD H. and GREAVESM. J. ( 1982) The rare earth elements in seawater. Nature 2%,2 14-2 19.

4016

f. H. Lee and R. H. Byrne

ELDERFIELD H. and SH~LKOVKZE. R. ( 1987) Rare earth elements in the pore waters of reducing nearshore sediments. E&I Planet. Sci. Lett. 82, 280-288. ELDERF~ELD H., U~ILL-~DDARD R., and SHOLKOVITZ E. R. ( 1990) The rare earth elements in rivers, estuaries, and coastal seas and their significance to the composition of ocean waters. Geochim. Cosmochim. Ac6a 54,97 I-99 I. FISBERF. H. and DAVISD. F. ( 1967) Effect of pressure on the distribution of the (La!Q) + complex ion. J. Phys. Chem. 71,8 19822. GERMANC. R. and ELDERRELDH. ( 1989) Rare earth elements in Saanich Inlet, British Columbia, a seasonally anoxic basin. Geochim. Cosmochim. Acta 53,256 l-257 I. GERMANC. R. and ELDERFIELD H. ( 1990) Ram earth elements in the NW Indian Ocean. Geochim. Cosmochim.Acta 54,1929-1940. GOLDSTEIN S. J. and JACOBSEN S. B. ( 1988) Rare earth elements in river waters Earth Planet. Sci. Lett. 89, 35-47. GRANT P. M., BAISDEN W. F., KJNARDW. F., and TORRESR. A. ( 1988) Enthalpies of formation of the mono~uorolanthanide complexes. Inorg. Chem. 27, I 156-l 158. GREAVESM. J., RUDNICKIM. R., and ELDERFIELD H. ( 1991) Rare earth elements in the Mediterranean Sea and mixing in the Mediterranean outflow. Earth Planet. Sci. Lett. 103, 169- I8 I. HALEC. F. and SPEDDINCF. H. ( 1972) Effect of hinh oressure on the formation of aqueous EuS& at 25’C. J. Ph>s.‘Chem. 76, 2925-2929. HOYLEJ., ELDERF~ELD H., GLEDHILLA., and GREAVESM. ( 1984) The behaviour of the rare earth elements during mixing of river and sea waters. Geochim. Cosmochim. Acla 48, 143- 149. I&STER D. R. and PYTKOWICZR. M. ( 1968) Magnesium sulfate association at 25°C in synthetic seawater. Limnol. Oceanogr. 13, 670-674. KESTERD. R. and PYTKOWICZ R. M. ( 1969) Sodium, magnesium and calcium sulfate ion-pairs in seawater at 25°C. LimnoI. Oceanugr. 14,686-692. KESTERD. R. and PYTKOWICZ R. M. ( 1970) Effect of temperature and pressure on sulfate ion association in seawater. Geochim. Cosmochim. Acta 34, 1039- 105I. KLINKHAMMER G., ELDERFIELD H., and HUDSONA. ( 1983) Rare earth elements in seawater near hydrothermal vents. Nature 305, 185-188. LEEJ. H. and BYRNER. H. ( 1993) Complexation of trivalent rare earth elements (Ce, Eu, Gd, Tb, Yb) by carbonate ions. Geochim. Cosmochim. Acta 57,295-302.

LUNDQVIS~R. ( 1982) Hydrophilic complexes of the actinides. I. Carbonates of trivalent americium and europium. Acfa Chem. Scund. A36,741-750. MILLEROF. J. ( 1969) The partial molal volumes of ions in seawater. Limnol. Oceanogr. 14, 376-385. MILLEROF. J. ( 1971) Effect of pressure on sulfate ion association in seawater. Geochim. Cosmochim. Acta 35, 1089-1098. MILLEROF. J. ( 1977) The use of the specific interaction model to estimate the partial molal volumes of electrolytes in seawater. Geochim. Cosmochim. Acta 41,2 15-223. MILLEROF. J. ( 1979) The the~~yna~~ of the carbonate system in seawater. Ge~him, Cusm~hjm. Acta 43, 1651- 1661. MILLEROF. J. ( 1983) Influence of pressure on chemical processes in the Sea. In Chemical Oceanography (ed. J. P. RILEYand R. CHESTER),Vol. 8, pp. l-88. Academic Press. MILLEROF. J. and BERNERR. A. ( 1972) Effect of pressure on carbonate equilibria in seawater. Geochim. C’osmochim.Actu 36,9298. MIRONOVV. E., AVRAM~NKON. I., KOPERINA. A., BLOKHIN V. V., EIKEM. Yu., and ISAYEVI. D. ( 1982) Thermodynamics of the formation reaction of the monochloride complexes of the rare earth metals in aqueous solutions. Koord. Khim. 8,636-638. PIEPGRASD. J. and JACOBSEN S. B. ( 1992) The behavior of rare earth elements in seawater: Precise determination of variations in the North Pacitic water column. Geochim. Cosmochim. Ada 56, 1851-1862. POWELLH. K..P. ( 1974) Entropy titrations: A reassessment of data for the reaction of the sulphate ion with trivalent lanthanoid ions. 3. Chem. Sot. Dalton Trans., 1108-I 112. PYTKOWICZ R. M. and KESTERD. R. ( 1969) Harned’s rule behavior of NaC1-Na2S0.rsolutions explained by an ion association model. Amer. J. Sci. 267,2 17-229. SAS INSTITUT’E INC. ( 1989) SAS/STAT UserS Guide, Version 6 (4th ed., Vol. 2). SAS Institute Inc. SHOLKOVITZ E. R. and SCHNEIDER D. L. f 1991)Cerium redox cycles and rare earth elements in the %ugassoSea. Geochim. Cosmochim. Acta 55,2737-2743. SHOLKOVITZ E. R., PIEPGRAS D. J., and JACOBSEN S. B. ( 1989) The pore water chemistry of rare earth elements in Buzzards Bay sediments. Geochim. Cbsmochim. Acta 53,2847-2856. SMITHR. M. and MARTELL A. E. I 1976) CriticalSta~~~it~Consfants (Vol. 4); inorganic ComplexesI Plenum Press. ’ WEISSR. ( 1974) Carbon dioxide in water and seawater: the solubility of a non-ideal gas. Mar. Chem. 2,203-2 15.