«THERMODYNAMIC MODELING OF METAL ADSORPTION AND MINERAL SOLUBILITY IN GEOCHEMICAL SYSTEMS A Dissertation Submitted to the Graduate School of the ...»
In Chapter 5,1 report on experiments that elucidate the effect of Np0 2 + incorporation on the solubility of soddyite. I synthesized soddyite samples in the presence of various aqueous Np(V) concentrations and measured the release of U, Np, and Si into solution as a function of time under controlled pH conditions. These data, through the mass action equation for the Np-soddyite dissolution reaction, were used to determine the solid phase activity coefficient for each Np-incorporated phase. The values of these activity coefficients indicate whether Np incorporation follows ideal solid-solution, and the determination of the coefficients as a function of Np content of the solid phase enables extrapolation of the results to solid phases with more extensive Np incorporation. I also use the data to calculate distribution coefficients that describe the extent of Np release from the Np-bearing soddyite phases. Therefore, the results of this study enable quantitative models of both Np and U mobility under repository conditions.
2.1. Introduction Surface and ground waters typically contain a range of metal ions that compete for adsorption sites on surfaces of soil and aquifer components. Although the binding of environmentally important, divalent and trivalent metals onto soil components has been studied (e.g., Beveridge and Murray 1976; Beveridge 1989; Ledin et al. 1997; Yee and Fein 2001; Covelo et al. 2007), the binding behavior of monovalent cations onto bacterial surface functional groups is not known. Monovalent cations represent a major component of the total dissolved ions in many natural waters. Additionally, concentrated solutions of monovalent salts such as NaCl or NaC104 typically are used to buffer ionic strength in metal-bacteria adsorption experiments.
The adsorption of monovalent cations onto mineral and bacterial surfaces has been accounted for indirectly through construction of electric double or triple layer models, which ascribe the association of the cations with the surface to electrostatic interactions with the surface electric field (Stumm et al. 1970; Davis et al. 1978).
Implicit in electrostatic surface complexation models of experimental adsorption data is the assumption that monovalent cations do not compete with higher-charged metals for adsorption onto specific bacterial surface functional groups. Typically, the adsorption of multi-valent cations to minerals and bacteria decreases with increasing ionic strength (e.g., Daughney and Fein 1998; Gu and Evans, 2008). Electrostatic multi-layer models ascribe these ionic strength effects to the contraction of the surface electric field of the sorbent due to non-specific outer-sphere electrostatic attraction of monovalent counterions to the electric field of the sorbent (e.g., Davis and Kent 1990; Koretsky 2000).
However, electrostatic models require the optimization of a number of parameters from experimental data such as surface electric field capacitance values. Direct modelindependent determination of these parameters is impossible, and application of these models to complex real systems is problematic (Davis et al. 1998).
An alternative approach to accounting for ionic strength effects on multi-valent cation adsorption is to ascribe the adsorption behavior to direct competition between the electrolyte monovalent cations and the less abundant multivalent cations for specific surface sites. This approach obviates the need for determining electrostatic modeling parameters by instead including electrostatic effects in the apparent metal-surface equilibrium constants (Davis and Kent 1990). Equilibrium constants determined using this approach have the advantage of being independent of ionic strength and pH.
Although the binding between surface functional groups and monovalent cations is likely to be weak, the concentration of the background electrolyte in metal adsorption experiments is often several orders of magnitude greater than that of the metal of interest. Thus, it is possible that monovalent metals significantly reduce the adsorption of higher charged metals via specific adsorption onto sites at the bacterial surface.
In this paper, we report the results of experiments conducted to determine the extent of adsorption of three monovalent cations, Li+, Rb+, and Na+, to the Grampositive soil bacterial species Bacillus subtilis as a function of ionic strength and pH.
Adsorption data are modeled using a non-electrostatic surface complexation model (NEM) approach, and discrete metal-bacteria binding constants are determined for each monovalent cation. These constants are then used to determine the competitive effect of monovalent cations on the adsorption of Cd at bacterial surface sites.
2.2 Methods 2.2.1 Bacteria growth and preparation The Gram-positive soil bacterium Bacillus subtilis was initially cultured on agar slants made of 0.5% yeast extract and trypticase soy agar. Cells from the slant were transferred to 3 ml of growth medium consisting of trypticase soy broth (TSB) and 0.5% yeast extract and allowed to grow for 24 hours at 32°C. After the growth period, these bacteria were transferred to 21 of identical broth and allowed to grow for another 24 hours at 32°C, reaching stationary phase. Bacteria were harvested by centrifuging the broth at 9,000 g for 10 minutes to pellet the bacteria. After decanting the broth, the bacteria were washed four times in NaC104 electrolyte solutions of the same ionic strength as the target ionic strength for an individual adsorption experiment, between 10" * and 10"' M. After each wash, the bacteria were centrifuged at 8100 g for 5 minutes to pellet the bacteria and the electrolyte was discarded. The cells were then resuspended in fresh electrolyte using a Vortex and stir rod. The bacteria were transferred to a weighed centrifuge tube after the final wash, and centrifuged two times for 30 minutes at 8100 g, • decanting the remaining supernatant each time. The weight of the resulting wet bacterial pellet has been determined to be 8 times the dry weight (Borrok et al. 2005). The growth and washing procedure renders the bacteria alive, but metabolically inactive (Borrok et al. 2007).
2.2.2 Lt andRb+ adsorption experiments Batch metal-bacteria adsorption experiments were performed with the monovalent cations Li+ and Rb+ (separately) in the presence of a NaC104 electrolyte, as a function of NaC104 concentration and pH. A perchlorate electrolyte solution was chosen because this anion does not complex strongly with metal cations. Because Na and K are present in biological cells and leach into solution to some extent, it is impossible to conduct adsorption experiments with these elements due to mass balance difficulties. Instead, we conducted Li+ and Rb+ adsorption measurements because these elements are not present appreciably in cells, and therefore rigorous constraints on their mass balances could be imposed on the experimental systems.
To verify that Li+ or Rb+ did not enter into electrolyte solutions from within the cells, control experiments were performed by suspending 20 g l"1 wet mass B. subtilis cells in 10"1,10"2, and 10"3 molal NaC104 electrolyte solutions. Each ionic strength bacterial suspension was separated into a set of test tubes, and the pH of each tube was adjusted using small volumes of HC1 or NaOH to a value between 2 and 10. The tubes were placed on a rotary shaker to equilibrate for 2 h, after which the steady-state pH of each system was measured. The systems were then centrifuged at 8100 g for 10 min to separate the bacteria from the solution, and the resulting supernatants filtered through 0.45 \xm nylon membranes and acidified with 25 [xL of 2.0 M HC1 per 10 ml of solution.
Metal concentrations in all solutions were immediately analyzed using Inductively Coupled Plasma - Optical Emission Spectrometry (ICP-OES). In all control experiments, the measured concentrations of Li+ and Rb+ were below the detection limits of the ICP-OES.
Isotherm Li+ and Rb+ adsorption experiments were performed as a function of ionic strength at a range of fixed pH values between pH 3 and 9. Each experiment was initiated by suspending 20 g l"1 wet mass B. subtilis in a NaC104 electrolyte solution with a concentration between 10"' and 10"3 M. The pH of each suspension was adjusted as needed with small volumes of NaOH or HC1, and the systems were allowed to equilibrate on a rotary shaker for 2 h. This process of pH adjustment was repeated until each system was within ±0.1 units of the target pH. The amount of acid or base added was recorded in order to calculate the final experimental ionic strength value. After each suspension reached steady-state pH conditions, 10 ml aliquots of the bacterial suspension were removed from each system, placed into polypropylene test tubes, spiked to afinalconcentration of 2.34 x 10"5 M Li+ or Rb+ using 1000 mg l"1 Li or Rb stock solutions prepared from LiC104 or RbC104 salts, and allowed to react for another 2 h.
The preparation of these LiC104 and RbC104 stock solutions and all additions of the solutions to experiments were performed gravimetrically. After reaching steady-state, the pH of experimental systems with Li+ and Rb+ were measured. ThefinalpH levels for the Li+ experiments were 2.99±0.03, 4.99±0.03, 6.95±0.07, and 9.19±0.09, and for the Rb+ experiments, 2.98+0.03, 5.18±0.10, 6.98±0.04, and 9.02±0.04. Henceforth, for convenience we refer to these experiments as pH 3, 5, 7, and 9, respectively. The test tubes were centrifuged, filtered, and acidified in the same manner as the control experiments (described above), and metal concentrations in the solutions were immediately analyzed using ICP-OES. We found that the signal strength of the ICP-OES varied strongly with solution ionic strength and composition (data not shown). To control for this effect, we centrifuged andfilteredthe extra Li- and Rb-free bacterial suspension that was not used in each Li+ and Rb+ adsorption experiment, and made Li or Rb calibration standards for ICP-OES analysis using the resulting supernatants. In this way, each experimental system had its own set of calibration standards made in a background matrix that was identical to that of the experimental samples. Analytical uncertainties associated with the ICP-OES analysis, as determined by repeat analyses of calibration standards, were less than ±3% in all cases. The amount of metal adsorbed in each experimental system was determined by subtracting the measured metal concentration remaining in solution from the initial known metal concentration in each experiment.
2.2.3 Cd adsorption experiments Batch Cd metal adsorption experiments were performed between pH 2.0 and 5.5 ' in the presence of Li-, Na-, and KCIO4 electrolytes to determine if the type of monovalent cation in the buffering electrolyte affected Cd adsorption behavior.
Experiments were initiated by suspending 20 g 1"' washed B. subtilis cells in a 0.1 M Li-, Na-, or KCIO4 solution containing 8.90 x 10"5 M Cd from a 1000 mg l"1 Cd stock solution. The Cd stock solution was prepared from a Cd(ClC»4)2 salt. The experimental pH adjustments, equilibration time, centrifugation, filtering, and acidification for each system was identical to the procedures used for Li+ and Rb+ batch adsorption experiments described previously. Concentrations of Cd remaining in the filtered supernatants were analyzed using ICP-OES. Cd standards were prepared gravimetrically from a 1000 mg l"1 Cd stock solution made from a Cd(N032 salt, diluted to desired concentrations using the same perchlorate salt matrix as the experimental systems being analyzed. The Cd signal strength reported by the ICP-OES did not vary significantly with solution ionic strength, and uncertainties were within ±3% for these experiments.
2.3 Results The measured extents of Li+ and Rb+ adsorption onto B. subtilis are shown as a function of ionic strength at various pH values in Figures 2.1 and 2.2, respectively. The addition of NaOH to achieve the pH 9 conditions, or HC1 to achieve pH 3 and 5, substantially increases the ionic strength of the background electrolyte in the experiments that had an initial ionic strength of 10"3 M. For this reason, each experiment is plotted using the actual ionic strength of the experiment, which is the sum of the ionic strength contribution from the NaC104 electrolyte and that from the acid or base additions.
In both the Li+ and Rb+ experiments, no significant adsorption is observed at any ionic strength in experiments conducted at pH 3, whereas significant adsorption is observed in the pH 5, 7, and 9 experiments. The pH 5, 7, and 9 experimental results indicate that the extent of Li and Rb adsorption increases significantly with decreasing ionic strength of the experimental electrolyte. As a function of pH, the amount of metal adsorbed to the bacteria in the pH 5, 7, and 9 experiments does not change significantly at any ionic strength level. This suggests that at pH 5 and above, Li and Rb adsorption behavior is more dependent on ionic strength than pH under our experimental conditions.
-2.5 -2 -1.5 -0.5 Ionic Strength (log M) Figure 2.1: (A) Lithium adsorption to B. subtilis as a function of ionic strength and pH.
Initial experimental conditions were 20 g l"1 B. subtilis cells and 2.34 x 10"5 M Li. The pH 5 model curve represents the best-fit model that accounts for Li adsorption onto Site 2 only. Curves for the pH 7 and pH 9 models show the extent of adsorption that would be predicted using the KNQ and Ku values determined from modeling the pH 5 data, and assuming no additional adsorption of Li onto Sites 3 or 4. (B) Best-fit model for Li adsorption to Site 2 of B. subtilis at pH 5 (solid curve). Dashed curves are models resulting from a ± 0.2 variation in the best-fitting log stability constant value of Ku
Figure 2.3 shows the results of the experiments of Cd adsorption onto B.
subtilis cells. Cd adsorption increases as a function of increasing pH from approximately 10% at pH 2 to nearly 80% at pH 5.5. The extent of adsorption does not vary significantly as a function of the type of monovalent salt used to buffer the ionic strength of the experimental system, suggesting that the three monovalent cations studied here compete with Cd for adsorption to bacterial surface functional groups to a similar degree.
In general, the extent of adsorption of the monovalent cations studied here was much less than that observed for divalent cations under similar experimental conditions.