P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 13:53 Style file version Nov. 19th, 1999 C 2002) Journal of Solution Chemistry, Vol. 31, No. 2, February 2002 (° Solubility of Siderite (FeCO3 ) in NaCl Solutions ∗ Carlos A. R. Silva,1 Xuewu Liu,2 and F. J. Millero2 Received June 21, 2001; revised December 3, 2001 The solubility of siderite (FeCO3 ) at 25◦ C under constant CO2 partial pressure [ p(CO2 )] was determined in NaCl solutions as a function of ionic strength. The dissolution of FeCO3 (s) for the reaction FeCO3 (s) + 2H+ = Fe2+ + CO2 (g) + H2 O ∗ K so = [Fe2+ ] pCO2 /[H+ ]2 has been determined as a function of pH = − log[H+ ]. From these values we have determined the equilibrium constant for the stoichiometric solubility to FeCO3 (s) in NaCl £ ∗ = [Fe2+ ] CO2− K sp 3 ¤ These values have been fitted to the equation ∗ ] = −10.9 + 2.518 I 0.5 − 0.657 I log[K sp o ) − 10.9 in water is with a standard error of s = 0.15. The extrapolated value of log(K sp in good agreement with data in the literature (−10.8 to −11.2) determined in solutions of different composition and ionic strength. The measured values of the activity coefficient, γ T (Fe2+ ) γ T (CO2− 3 ), have been used to estimate the stability constant for the formation of the FeCO3 ion pair, K ∗ (FeCO3 ). The values of K ∗ (FeCO3 ) have been fitted to the equation (s = 0.09) log[K ∗ (FeCO3 )] = 6.3 − 2.3135 I 0.5 + 0.7091 I The value of log[K o (FeCO3 )] in water found in this study (6.3 ± 0.2) is slightly higher than the value found from extrapolations in 1.0 m NaClO4 solutions (5.9 ± 0.2). These differences are related to the model used to determine the activity coefficients of the Fe(II) and carbonate species in the two solutions. KEY WORDS: Siderite; solubility; activity coefficients; sodium chloride; Pitzer equations. 1 Departamento de Oceanografia e Limnologia, Centro de Biociências, Universidade Federal do Rio Grande do Norte, CP 1202, 59075-970, Brasil. 2 Rosenstiel School of Marine and Atmospheric Science, University of Miami, 4600 Rickenbacker Causeway, Miami, Florida 33149-1098; e-mail: [email protected] 97 C 2002 Plenum Publishing Corporation 0095-9782/02/0200-0097/0 ° P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 13:53 98 Style file version Nov. 19th, 1999 Silva, Liu, and Millero 1. INTRODUCTION Among the most chemically reactive and common minerals in the aquatic systems, carbonate minerals play an important role in the cycling of the chemical elements. Under anoxic conditions, the solubility of ferrous iron (Fe2+ ) is frequently controlled by the siderite or ferrous carbonate (FeCO3 )(1) by the equilibrium FeCO3 (s) = Fe2+ + CO2− 3 (1) A number of workers(2–8) have determined the stoichiometric solubility product of FeCO3 (s) at low ionic strengths (≤1 m). If the activity of FeCO3 (s) is assumed to be 1.0, the stoichiometric solubility product is defined by £ ¤ ¡ ¢ ∗ o K sp = [Fe2+ ] CO2− γ T (Fe2+ ) γ T CO2− = K sp (2) 3 3 o where K sp is the infinite-dilution solubility in water and γ T (i) and [i] are the total activity coefficient and concentration of species i. The subscript T is used to denote the total activity coefficient (which accounts for all of the ionic interaction in the solution).(9) Ideally one would like to be able to calculate the solubility product o and estimates of the activity coefficient of Fe2+ and CO2− from values of K sp 3 . 2− 2+ and CO can form comSince Fe can form complexes with OH− and CO2− 3 3 plexes with Mg2+ , Ca2+ , and Fe2+ , one has to account for the formation of these complexes(9) in natural waters. At the present time, Pitzer(10) computer codes are available(9) that can be used to determine the activity coefficients of Fe2+ and CO2− 3 in dilute solutions of most natural waters made up of the major seasalts [Na+ , − 2− − − − Mg2+ , Ca2+ , K+ , Sr2+ , Cl− , SO2− 4 , HCO3 , Br , CO3 , B(OH)4 , F , CO2 , and 2− 2+ B(OH)3 ]. The data for examining the interactions of Fe and CO3 in concentrated solutions are presently not available. In this study, the solubility of FeCO3 (s) was measured at 25◦ C in NaCl solutions as a function of ionic strength. These results have been used to examine the interactions of Fe2+ and CO2− 3 at high ionic strengths. 2. EXPERIMENTAL 2.1. Preparation of Solid FeCO3 Ferrous carbonate was prepared using deoxygenated Milli-Q water (Millipore) and a mixture of ferrous sulfate heptahydrate (reagent grade) and sodium bicarbonate (reagent grade) according to the methods of Greenberg and Tomson.(8) The prepared seed crystal was heated in a water bath at 50◦ C for 10 days. As shown by Bruno et al.,(7) preparations of FeCO3 (s) by these methods provides high-purity crystalline siderite. P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 Solubility of Siderite 13:53 Style file version Nov. 19th, 1999 99 2.2. Working Vessel The measurements were made in a Plexiglas vessel designed for temperaturecontrolled reactions.(11) Four holes were made in the top of the vessel for connecting a combination electrode (8102 ORION), a bubbler, titration tube, and a tube used to take out the sample. The measurements were made at 25◦ C using a Neslab water bath set with a Guiline Platinum resistance thermometer. The measurements were made in reagent-grade NaCl as a function of the molal ionic strength. The NaCl solutions (0.1, 0.7, 1.4, 2.0, 2.5, 4.0, and 5.5 m) were prepared with deoxygenated Milli-Q water and sodium chloride (reagent grade) in a glove box devoid of oxygen. Approximately 9 cm3 of a siderite solution was introduced with a syringe under a stream of CO2 (research grade) into the working vessel. The solutions were bubbled with 5% CO2 gas (CO2 /N2 —Anaerobic Biological Atmosphere) at a total pressure of 1 atm. The CO2 /N2 gas was passed over four oxygen scavenger traps (purple vanadium chloride) and a 0.1 µm Vacu-Guard Whatman filter before it was passed through the test solution. The solution was stirred to suspend the solid particles in the solution. Measurements at high pH were unsuccessful due to the oxidation of Fe(II) to Fe(III) over time. 2.3. Titration and Total Iron Ferrous Analysis The solubility measurements were carried out by titrating the initial solutions [NaCl + FeCO3 (s)] with bicarbonate (2.0 m NaHCO3 ) over the pH range of 5.76 to 7.77. After each addition, the solutions were equilibrated at a constant temperature while bubbling with CO2 (g). The equilibrium was assumed when the potential of the glass electrode remained within 0.1 mV for 24 h.(7) The inner electrolyte solution of the combination pH electrode was filled with 3.0 M NaCl to reduce the liquid junction potential. The emf of the electrode system was related to the molal concentration of the proton [H+ ] by E = E ∗ − (RT /F) ln[H+ ] (3) where E ∗ is the apparent standard potential in the NaCl media. The values of E ∗ were determined by titrating each NaCl solution with 0.2489 M HCl. A syringe of 10 cm3 and an Acro-disc PF syringe filter (0.8 µm/0.2µm) were used to collect filtered solutions under vacuum conditions. The filtered solutions were immediately transferred to a glass vial containing 0.1 M HCl for weighing and analysis of total dissolved iron using the Ferrozine method.(12) 3. RESULTS AND DISCUSSION The solubility of FeCO3 (s) in NaCl solutions was measured by monitoring the ferrous iron concentrations [Fe2+ ] as a function of hydrogen concentrations at constant CO2 (g) partial pressure. The effects of pH on the concentration of P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 13:53 Style file version Nov. 19th, 1999 100 Silva, Liu, and Millero Table I. Values of log [Fe(II)] in NaCl Solutions as a Function of pH for Equilibrated Solutions of FeCO3 (s) at Constant pH m (mol-kg−1 ) pH log[Fe(II)] m (mol-kg−1 ) pH log[Fe(II)] 0.1 5.76 5.83 6.00 6.23 6.33 −2.897 −3.122 −3.360 −3.833 −4.297 2.5 6.39 6.50 6.73 6.95 −2.884 −3.302 −3.639 −4.020 4.0 0.7 5.86 5.95 6.33 6.39 6.50 −2.859 −3.048 −3.833 −4.009 −4.189 6.50 6.58 6.74 6.91 7.18 −2.981 −3.063 −3.323 −3.645 −4.228 5.5 1.4 5.97 6.15 6.39 −2.254 −2.652 −3.034 6.73 6.99 7.25 7.39 7.77 2.0 6.22 6.35 6.49 6.97 7.13 −2.830 −2.958 −3.240 −4.171 −4.570 −3.027 −3.379 −3.850 −4.228 −5.024 Fe2+ in NaCl at different ionic strengths are given in Table I and shown in Fig. 1. The slope of each curve was nearly −2 (−2.0 ± 0.1) and is independent of ionic strength, in accord with the work of Bruno et al.(7) In the range 5.76 ≤ pH ≤ 7.77, the [Fe2+ ] released during the dissolution of FeCO3 (s) is a result of the reaction FeCO3 (s) + 2H+ = Fe2+ + CO2 (g) + H2 O (4) ∗ The constant K so is defined as ∗ K so = [Fe2+ ] p(CO2 )/[H+ ]2 (5) ∗ K so where p(CO2 ) is the partial pressure of CO2 (in bars). The values of have been determined from plots of log { p(CO2 )/[H+ ]2 } vs. log[Fe2+ ] and the results are summarized in Table II and shown as a function of I 0.5 in Fig. 2. These results have been fitted to the linear equation (σ = 0.15) ∗ log(K so ) = 6.95 + 0.988 I 0.5 (6) ∗ The interpolated value of logK so (7.9 ± 0.15) in 1.0 m NaCl is in reasonable agreement with the value found by Bruno et al.(7) in 1.0 M NaClO4 (7.65 ± 0.1). P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 13:53 Style file version Nov. 19th, 1999 Solubility of Siderite 101 Fig. 1. Solubility of FeCO3 (s) as a function of log(Fe2+ ) and pH in NaCl solutions at 25◦ C and p(CO2 ) = 0.05 bar. ∗ The measured values of K so can be used to determine the dissolution of FeCO3 (s) given by FeCO3 (s) = Fe2+ + CO2− 3 ∗ The stoichiometric solubility product K sp for this equilibrium is given by £ ¤ ∗ K sp = [Fe2+ ] CO2− 3 ∗ ] and Product Table II. The Solubility Constant log[K so ∗ ] for FeCO (s) in NaCl Solutions log[K sp 3 Molality (mol-kg−1 ) ∗ ) bar log(K so ∗) log(K sp 0.1 0.7 1.4 2.0 2.5 4.0 5.5 −7.26 ± 0.12 −7.52 ± 0.04 −8.39 ± 0.05 −8.41 ± 0.06 −8.52 ± 0.09 −8.81 ± 0.06 −9.25 ± 0.08 −10.22 ± 0.12 −9.53 ± 0.04 −8.58 ± 0.05 −8.61 ± 0.06 −8.55 ± 0.09 −8.56 ± 0.06 −8.61 ± 0.08 (7) (8) P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 13:53 102 Style file version Nov. 19th, 1999 Silva, Liu, and Millero ∗ ) as a function of the square root of Fig. 2. Values of log(K so ionic strength I . where [Fe2+ ] was obtained from experimental analysis and [CO2− 3 ] was calculated from the values of p(CO2 ) and pH. The CO2 equilibrium is controlled by the following equilibria(13,14) CO2 (g) = CO2 (aq); + CO2 (aq) + H2 O = H + HCO− 3; 2− + HCO− 3 = H + CO3 ; K 0∗ = [CO2 ]/ p(CO2 ) K 1∗ + [HCO− 3 ]/[CO2 ] = [H ] £ ¤ − K 2∗ = [H+ ] CO2− 3 /[HCO3 ] (9) (10) (11) From these stoichiometric dissociation constants the concentrations of [CO2− 3 ] may be determined from £ 2− ¤ CO3 = K 0∗ K 1∗ K 2∗ p(CO2 )/[H+ ]2 (12) ∗ can be defined as and K sp ∗ ∗ K sp = K 0∗ K 1∗ K 2∗ K so (13) The stoichiometric values of K 0∗ , K 1∗ , and K 2∗ in NaCl solutions have been determined from the Pitzer(10) equations of Millero and Roy,(14) which are based ∗ determined on the measurements of Thurmond and Millero.(15) The values of pK sp ∗ from Eq. (13) are given in Table II. The values of log K sp have been extrapolated to infinite dilution using the equation © ¡ ¢ª ∗ − log γ F (Fe2+ )γ F CO2− (14) Y = log K sp 3 (9) where γ F (Fe2+ ) γ F (CO2− 3 ) is the trace activity coefficient product of the free ions (see Table IV). The value of Y should be constant if the trace or free activity coefficients accounted for all of the ionic interactions in the solution. As discussed later, this is not the case due to the formation of the ionic complex FeCO3 . P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 13:53 Style file version Nov. 19th, 1999 Solubility of Siderite 103 ∗ ) − log{γ (Fe2+ ) γ (CO2− )] Fig. 3. Values of Y = log(K sp F F 3 as a function of molality m. A plot of Y versus the molal ionic strength (I ) in Fig. 3 gives a value of log o ∗ = −10.9 ± 0.2. The experimental values of log K sp are plotted versus I 0.5 in K sp Fig. 4 and have been fitted to the equation (s = 0.15) ∗ ) = −10.9 + 2.518 I 0.5 − 0.657 I log(K sp (15) ∗ at 1 m (−9.0 ± 0.2) determined from Eq. (15) is higher than The value of log K sp ∗ the value (log K sp = −9.55 ± 0.2) determined from the measurements of Bruno et al.(7) in 1 m NaClO4 . If the carbonate constants in 1 m NaCl (K 0∗ = 10−1.57 , K 1∗ = ∗ = −9.3 ± 0.2 is 10−5.94 , and K 2∗ = 10−9.50 )(14) are used, a value of log K sp found for 1 m NaClO4 , which is in better agreement with our value. Since the ∗ ) as a function of the square root Fig. 4. Values of log(K sp of ionic strength I . P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 13:53 Style file version Nov. 19th, 1999 104 Silva, Liu, and Millero Table III. Comparisons of the Solubility Product of FeCO3 (s) in Water at 25◦ C ◦C Solution Ionic strength (mol-kg−1 ) ∗) log (K sp 25 25 25 25 25 30 43 50 NaCl Water Water NaClO4 NaClO4 Water Water NaClO4 0.1 to 5.5 ∼0.01 ∼0.01 0.1 1.0 ∼0.01 ∼0.01 1.0 −10.9 ± 0.2 −10.8 −11.0 −10.2 −10.8 −10.7 −10.9 −11.2 Ref. This study a b c d e a f a Greenberg and Tomson, 1992 (Ref. 8). 1991 (Ref. 6). c Singer and Stumm, 1970 (Ref. 4). d Bruno et al., 1992 (Ref. 7). e Latimer, 1952 (Ref. 3). f Reiterer et al., 1981 (Ref. 5). b Braun, carbonate constants in NaCl and NaClO4 solutions are similar,(16) this is a reasonable calculation. As shown in Table III, our value for the solubility product extrapolated to ino ) = −10.9) is in good agreement with literature data.(3,5–8) finite dilution (log(K sp o Part of the differences in logK sp is related to the different methods used to extrapolate the results to infinite dilution. To make a reliable extrapolation to pure water, it is necessary to have a reliable model to estimate the activity coefficients. The activity coefficients for free ions in NaCl and NaClO4 solutions can be estimated from Pitzer’s equations.(9,14,16) These models consider the changes in the interactions of Fe2+ and CO2− 3 with NaCl as well as the formations of complexes (i.e., FeCO3 o ).(17) The measured solubility constants for between Fe2+ and CO2+ 3 the dissolution of FeCO3 (s) is related to the activity coefficients by ¡ ¢ ∗ o = K sp γ T (Fe2+ ) γ T CO2− (16) K sp 3 The measured activity coefficient products determined from ¡ ¢ ∗ o /K sp = K sp γ T (Fe2+ ) γ T CO2− 3 (17) are given in Table IV and shown as a function of the concentration of NaCl in Fig. 5. The values of the total activity product can be compared to the trace activity coef2+ and CO2− ficients, γF (Fe2+ ) γF (CO2− 3 ), determined in NaCl devoid of Fe 3 . The (9) values determined from the Pitzer equation are given in Table IV and shown as a function of the concentration of NaCl in Fig. 5. The trace activity coefficients are higher than the measured values. This is due to the formation of FeCO3 complexes P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 13:53 Style file version Nov. 19th, 1999 Solubility of Siderite 105 Table IV. Values of the Total Activity Product, γ T γ T = γ T (Fe2+ ) γ T (CO2− 3 ), the Stability Constant 0, for the Formation of the FeCO3 Complex, logK ∗ (FeCO3 ), and the Solubility Product in Water, logK sp at 25◦ C Measured I γT γT γ F (Fe2+ ) γ F (CO2− 3 ) Calc. γT γT Calc. Log K ∗ (FeCO3 ) Calc. ∗) log (K sp 0.0 0.1 0.7 1.4 2.0 2.5 4.0 5.5 1.0000 0.2089 0.0427 0.0048 0.0051 0.0045 0.0046 0.0051 1.0000 0.3896 0.2550 0.2656 0.3019 0.3474 0.5871 1.0833 1.0000 0.3333 0.1227 0.0727 0.0532 0.0433 0.0288 0.0240 1.0000 0.0323 0.0115 0.0049 0.0047 0.0044 0.0046 0.0055 6.30 5.41 4.80 4.59 4.51 4.48 4.53 4.71 — — — −10.89 −10.94 −10.91 −10.90 −10.87 Mean: −10.9 ± 0.03 that lower the activity coefficients. The values of the total activity coefficients are related to the free activity coefficients by ¡ ¢ ¡ ¢ = αFe αCO3 γ F (Fe2+ ) γ F CO2− (18) γ T (Fe2+ ) γ T CO2− 3 3 The fraction of the free Fe2+ and CO2− 3 can be determined from ¡ £ ¤ ¢−1 αFe = [Fe2+ ]F /[Fe2+ ]T = 1 + K ∗ (FeCO3 ) CO2− 3 F £ ¤ £ 2− ¤ ∗ 2+ −1 αCO3 = CO2− 3 F / CO3 T = (1 + K (FeCO3 ) [Fe ]F ) Fig. 5. Measured and calculated values of the activity coefficient product, γ T (Fe2+ ) γ T (CO2− 3 ), as a function of molality at 25◦ C. The line is calculated using log K o (FeCO3 = 5.9). (19) (20) P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 13:53 106 Style file version Nov. 19th, 1999 Silva, Liu, and Millero The values of K ∗ (FeCO3 ) are the stoichiometric constants for the formation of the FeCO3 ion pair Fe2+ + CO2− 3 = FeCO3 (21) 2+ 2+ The terms [Fe2+ ]F and [CO2− 3 ]F are the concentration of free Fe (=[Fe ]T αFe ) 2− 2− and free CO3 (= [CO3 ]T αCO3 ), respectively. The stoichiometric association constant K ∗ (FeCO3 ) is related to the thermodynamic value, K o (FeCO3 ) by © ¡ ¢ª (22) K ∗ (FeCO3 ) = K o (FeCO3 ) γ F (FeCO3 )/γ F (Fe2+ ) γ F CO2− 3 Since the total concentrations of [Fe2+ ]T = [CO2− 3 ]T when FeCO3 (s) dissolves, the fraction of free ions are also equal [Fe2+ ]F = [CO2− 3 ]F and can be estimated from combining Eqs. (19, 20, and 22). This gives the quadratic equation £ ¤ ∗ 0.5 (23) [Fe2+ ]F = CO2− 3 F = {−1 + [1 + 4K (FeCO3 )] }/2K ∗ (FeCO3 ) The values of K ∗ (FeCO3 ) needed to determined [Fe2+ ]F = [CO2− 3 ]F depends on the value selected for the thermodynamic value (K o (FeCO3 ). The literature values for log [K o (FeCO3 ) vary from 5.5(7,18) to 5.7.(19) The latter value was redetermined from the solubility measurements of FeCO3 (s) of Bruno et al.(7) in NaClO4 solutions by King.(19) Bruno et al.(7) determined the constant for the reaction Fe2+ + CO2 (g) + H2 O = FeCO3 + 2H+ , log β11 = −12.98 ± 0.16 (24) The value of K ∗ (FeCO3 ) at 1 m can be determined from K ∗ (FeCO3 ) = β11 /K 0∗ K 1∗ K 2∗ = 10−12.9 /(10−7.62 10−9.52 ) = 104.24 (25) CO2− 3 (0.0952) in 1.0 m NaClO4 The free activity coefficients of Fe2+ (0.254) and gives a value of log [K o (FeCO3 ) m o )] = 4.24 + 1.62 = 5.9 ± 0.2, which is slightly higher than the value determined by King.(19) By adjusting the values of K o (FeCO3 ), one can determine γ T (Fe2+ ) γ T (CO2− 3 ) [Eq. (18)] and compare them with the values determined from the measured val∗ . A value of log K o (FeCO3 ) = 6.3 ± 0.2 gives the best fit of γ T (Fe2+ ) ues of K sp 2− γ T (CO3 ) (see Table IV and Fig. 5). This value of K o (FeCO3 ) is higher than the values determined by King(19) from the measurements made in 1.0 m NaClO4 by Bruno et al.(7) The value of log K o (FeCO3 ) = 6.6 determined by Mattigod and Sposito(20) are the only literature values as large as found in this study. It should be pointed out that the value of log K o (FeCO3 ) = 5.7 does not give a reliable fit of our measured values of γ T (Fe2+ ) γ T (CO2− 3 ) (see Fig. 5). The values of log K ∗ (FeCO3 ) calculated from Eq. (22) are given in Table IV. These results are shown in Fig. 6 and have been fitted to the equation (s = 0.1) log (K ∗ (FeCO3 ) = 6.3 − 2.3135 I 0.5 + 0.7091 I (26) It should be pointed out that we have assumed that the activity coefficient of the ion pair is γ F (FeCO3 ) = 1.0 in all of these calculations. P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 13:53 Solubility of Siderite Style file version Nov. 19th, 1999 107 Fig. 6. Values of log K ∗ (FeCO3 ) for the formation of the FeCO3 complex in NaCl as a function of the square root of ionic strength I . o One can also determine the value of K sp from the calculated values of 2− 2+ γ T (Fe ) γ T (CO3 ) by rearranging Eq. (16). The values above 0.7 m, calculated in o = −10.9 ± 0.03 this manner, are given in Table IV. The average value of log K sp is in excellent agreement with the extrapolated value. In summary, our solubility measurements of FeCO3 (s) in NaCl solutions over a wide range of ionic strength and pH < 7.2 can be adequately represented using o and K o (FeCO3 ) and a Pitzer model. The resultant infinite dilution constants for K sp values of both thermodynamic constants depend on the model used to determine the activity coefficients of Fe(II) and carbonate species. The small differences in our values in NaCl and literature results in NaClO4 are related to differences in the activity coefficients in the solutions. Since most natural waters contain NaCl as the major component, our results may prove more useful in examining the behavior of Fe2+ in anoxic waters containing carbonate. ACKNOWLEDGMENTS The Oceanographic Section of the National Science Foundation supported this research. Carlos A. R. Silva wishes to acknowledge the support of the Conselho Nacional de Desenvolvimento Cientı́fico e Tecnológico (CNPq – Processo 200992/97), sponsor of his post-doctoral stay at the Rosenstiel School. REFERENCES 1. P. Wersin, L. Charlet, R. Karthein, and W. Stumm, Geochim. Cosmochim. Acta 53, 2787 (1989). 2. H. Smith, J. Amer. Chem. Soc. 40, 879 (1918). 3. W. E. Latimer, The Oxidation States of the Elements and their Potentials in Aqueous Solutions, 2nd edn. (Prentice-Hall, New York, 1952). P1: GDX/GVM/LCR/GMF P2: GCR Journal of Solution Chemistry [josc] pp439-josl-369369 April 9, 2002 13:53 Style file version Nov. 19th, 1999 108 Silva, Liu, and Millero 4. 5. 6. 7. 8. 9. 10. P. C. Singer and W. Stumm, J. Amer. Water Works Assoc. 62, 198 (1970). F. Reiterer, W. Johannes, and H. Gamsjäger, Mikrochim. Acta 1, 63 (1981). R. D. Braun, Talanta 38, 205 (1991). J. Bruno, W. Stumm, P. Wersin, and F. Brandberg, Geochim. Cosmochim. Acta 56, 1139 (1992). J. Greenberg and M. Tomson, Appl. Geochem. 7, 185 (1992). F. J. Millero and D. Pierrot, Aquatic Geochem. 4, 153 (1998). K. S. Pitzer, Activity Coefficients in Electrolyte Solutions, Vol. I, 2nd edn., (CRC Press, Boca Raton, FL 1991), pp. 75–153. F. J. Millero, S. Hubinger, M. Fernandez, and S. Garnet, Environ. Sci. Technol. 21, 179 (1987). L. C. Stookey, Anal. Chem 42, 779 (1970). F. J. Millero, Geochim. Cosmochim. Acta 59, 661 (1995). F. J. Millero and R. N. Roy, Croatica Chem. Acta 70, 1 (1997). V. Thurmond and F. J. Millero, J. Solution Chem. 11, 447 (1982). F. J. Millero, Geochim. Cosmochim. Acta 56, 3123 (1992). F. J. Millero, W. Yao, and J. Aicher, Marine Chem. 50, 21 (1995). F. J. Millero and D. J. Hawke, Marine Chem. 40, 19 (1992). D. W. King, Environ. Sci. Technol. 32, 2997 (1998). S. V. Mattigod and G. Sposito, Soil. Sci. Amer. J. 41, 1092 (1977). 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.
© Copyright 2026 Paperzz