«THERMODYNAMIC MODELING OF METAL ADSORPTION AND MINERAL SOLUBILITY IN GEOCHEMICAL SYSTEMS A Dissertation Submitted to the Graduate School of the ...»
We observed an incorporation effect that is dramatically larger than that predicted by ideal solid-solution behavior. For example, the 6511 ppm Np soddyite corresponds to a XSOddyite value of 0.9908. Ideal solid-solution behavior, then, would correspond to an aqueous U02+2 decrease of only 0.004 log molality units, essentially an unchanged concentration relative to the value measured for the experiments involving the 24 ppm Np solid. In actuality, the 6511 ppm Np soddyite experiment exhibited a UCV2 molality decrease of 0.36 log molality units (Figure 5.1B), an effect that is approximately two orders of magnitude greater than that expected for ideal solidsolution behavior. Values of Tsoddyite increasingly depart from ideal behavior (where rsoddyite = 1) as XSOddyite decreases (Figure 5.2). With the limited data available, it is difficult to reliably extrapolate the relationship that is depicted in Figure 5.2 to determine the solubility behavior of solids with higher levels of Np incorporation. However, the effect of Np incorporation into the soddyite structure is dramatic even for the low levels of Np incorporation documented in this study, and Figure 5.2 suggests that the effects are likely to increase with increasing Np incorporation.
5.4 Conclusions The experimental results reported here indicate that neptunyl substitution for uranyl in soddyite is much more complex and exerts a much larger effect on mineral solubilities than the ideal Np4+ substitution for U4+ observed by Rai et al. (2004). Several mechanisms may explain the deviation from ideal solid solution behavior observed in soddyites containing Np(V). Np(V)C2+ exhibits markedly different bond strengths than are present within U(VI)C2+2, and due to the charge imbalances that accompany NpC2+ substitution for U02+2, a co-substitution (most likely with Na"1) must occur. These complicating factors make the substitution of NpC2+ for UC2+2 in uranyl phases much more complex than the simple (and ideal) one-for-one incorporation that is possible with
Figure 5.2: The solid phase activity coefficient (rSOddyite) as a function of the mole fraction of uranyl soddyite (Xso,uyue)Np substitution for U inuraninite.
The more complex substitution is likely to occur for any Np(V)C2+ substitution for U(VI)02+2 in uranyl compounds in general, so the large Np(V) incorporation effect on the solubility of uranyl minerals is likely to significantly affect the solubility of the wide range of uranyl phases that incorporate even fairly low concentrations of Np. Our results suggest that incorporation of Np(V) into secondary uranyl phases can not only limit and control the mobility of Np in repository systems, but may decrease the mobility of U as well through non-ideal solidsolution effects.
This dissertation provides new insight into the adsorption of monovalent cations onto bacterial cell walls, the adsorption of cadmium onto mixtures of soil components, the determination of biomass C in soils, and the release of Np and U from a uranyl silicate mineral, soddyite. In Chapter 2,1 propose an alternative method to account for the effects of ionic strength on the adsorption of divalent and trivalent metal cations onto bacterial surfaces. The effect of ionic strength on the adsorption of metals has typically been accounted for by varying the extent and strength of the electric field surrounding the bacteria. My results indicate that ionic strength can also be accounted for by calculating equilibrium constants for complexes between monovalent cations in solution and the discrete organic functional groups present on the bacterial surface. In this way, higher-charged metals in solution and monovalent cations compete directly for adsorption to the bacterial surface functional groups. I test this direct competition model by fitting existing data sets of Cd adsorption to Bacillus subtilis in the presence of monovalent cation competition. The results show that the apparent equilibrium constants for Cd-surface complexes is higher than was previously calculated when monovalent cation competition was not included. The non-electrostatic, direct competition approach I develop here is simpler than the electric field approach because it does not entail the use of arbitraryfittingparameters, and is therefore easier to apply to complex geochemical systems.
The universal equilibrium constant I propose for complexes between monovalent metals and Site 2 of B. subtilis (log K = 1.9 ± 0.3) is fairly well constrained by my work.
However, to more rigorously test the accuracy of this value, monovalent cation adsorption experiments should be performed at different cation:/?, subtilis concentration ratios. Experiments containing a lower cation:bacteria ratio are likely to give a greater range of metal uptake than the low percentage (0 - 20%) of metal uptake observed in our experiments. Models of data at different catiombacteria ratios also may indicate if adsorption is occurring only on Site 2 of the bacteria, as I conclude from experiments in Chapter 2, or if other sites need to be invoked to describe the observed adsorption behavior. Additionally, Yee and Fein (2001) demonstrate that the adsorption behavior of Cd across a wide range of pH values is nearly identical for seven Gram-positive and Gram-negative species. Repeating monovalent cation adsorption experiments using other bacterial species will reveal if the adsorption of these cations is a similarly universal phenomenon.
I test validity of the fumigation-extraction method (Vance et al., 1987; Tate et al.,
1988) for determining biomass C in soils in Chapter 3. The method is based on the premise that soils can be exposed to chloroform gas for 24 hours to lyse bacterial cells in the soil, and that the chloroform can be fully recovered by evacuation after that time. I test this assumption by performing the fumigation-extraction process on individual components of soils, including kaolinite, montmorillonite, quartz sand, humic acid, and bacteria. Chloroform substantially adsorbs to both of the clays, enhancing the carbon pool in extraction solutions of these materials. Since biomass C is measured only as the total organic carbon in the extracts, this adsorbed chloroform that enters into the extraction solutions is likely to be misinterpreted as biomass, causing artificially high biomass C measurements. Thus, in soils containing a significant clay fraction, the fumigation-extraction method may be invalid, or require a correction factor to be usable.
A comprehensive study to determine which types of natural soils adsorb chloroform during the fumigation-extraction process is necessary. Among the factors that may influence chloroform vapor adsorption to a soil are the clay content, as discussed here in Chapter 3, soil moisture (Chen and Dural, 2002), and the presence of adsorbed bacteria and biofilms on clay surfaces. A universal correction factor to adjust experimental data for the adsorption of chloroform to soils is unlikely to be successful because of the complexity and variety of soil compositions found in nature. However, it.
may be possible to salvage the fumigation-extraction method for an individual soil by determining the fraction of carbon in its extract that is due to bacterial lysis, and the fraction due to chloroform. Methods including gas chromatography - mass spectrometry (GC-MS), or UV-visible spectrometry methods can quantify chloroform in solutions. If the chloroform concentration in extracts is determined using one of these methods, it could be subtracted from the total organic carbon measurement in a sample to give the amount of carbon attributed to the soil sample. In the fumigation-extraction method as currently used, two pools of carbon are determined: that due to cell lysis, which may include chloroform artifacts, and the background carbon in the soil prior to fumigation. Using the correction procedure I propose above, carbon would be divided into three discrete pools: adsorbed chloroform, that due to the lysis of bacteria, and background carbon extracted from unfumigated soil samples. Thus, a true measure of biomass C may be feasible, despite the adsorption of chloroform vapor.
In Chapter 4,1 rigorously test the component additivity (CA) approach in mixtures of B. subtilis cells, HFO, kaolinite, and dissolved acetate. Cd is adsorbed to each component individually, and equilibrium constants for Cd-sorbent surface complexes are calculated using existing surface complexation models for each sorbent. I conclude that Cd adsorption to two- and three-component mixtures of the geosorbents is well predicted by combining the constants calculated from the one-component systems.
In experiments including dissolved acetate as a representative organic acid, however, the • CA approach does not work well even in simple one-sorbent systems. Invoking only aqueous Cd-acetate complexes causes the predictive model to underestimate the observed extent of adsorption at all pH levels. To obtain a better fit, it is necessary to calculate equilibrium constants for the complexation of a cadmium acetate, Cd(CH3COO)+, to each sorbent. Accounting for ternary complexation among solid surfaces, metals, and dissolved organic acids is likely important in describing the distribution and transport of heavy metals. Careful consideration of ternary complexation is likely to be important in applying the CA approach to realistic geologic systems, where metal-organic aqueous complexes may be the dominant species adsorbing to soil or aquifer components.
Our data require models that include adsorption of the Cd(CH3COO)+ complex to each sorbent in order to account for the observed adsorption behaviors. More detailed evidence regarding the types and coordination environments of ternary metal-organicsurface complexes forming in our experiments could be provided by a spectroscopic technique such as extended adsorption x-ray fine structure (EXAFS) or Fourier transform - infrared spectroscopy (FT-IR). Future studies should determine the impacts of more complex organic ligands on metal adsorption, including fulvic and humic acids that will be present in natural systems. The ultimate application of the CA approach is in predicting metal distribution in soils. The physical components that make up soils can be divided into a few, broadly defined categories, or many detailed categories. The challenge is in defining these component categories well, so that surface complexation modeling provides accurate predictions of metal distributions.
The solubility effect of Np(V) incorporation into the uranyl silicate phase, soddyite, is discussed in Chapter 5. My results show that a relatively minor substitution of less than 1 atomic % NpC2+ for UC«22+ in the crystal structure results in a dramatic decrease in phase solubility, approximately 100 times more than would be expected by invoking ideal solid solution chemistry. This non-ideal behavior has important implications for nuclear repositories or Np contaminated groundwaters in oxidizing conditions where soddyite may be present or may form. Specifically, the incorporation of Np(V) into soddyite and the subsequent decrease in phase solubility may dramatically limit the mobility of both Np and U in these systems.
The exact structural mechanisms of Np incorporation into soddyite are not known with certainty. Due to the low concentrations of Np incorporated into the soddyite phases synthesized in my studies, elucidating the coordination of Np(V) in the crystal structure is difficult. If a phase with a higher Np concentration in the solid can be synthesized, EXAFS may reveal more about the binding environment of Np in soddyite.
Our studies cover only a small fraction of the theoretical solid solution between uranyl soddyite and a potential end member containing 100% neptunyl cation in the uranyl sites. I propose that the concentration of Np released into solution from Np-soddyite phases can be predicted by a simple linear distribution coefficient, Kd, approach at the modest levels of Np incorporation detailed here. However, it is unclear how far into the solid solution this linear relationship extends. The deviation of Np-soddyite from ideal behavior, as quantified by the soddyite activity coefficient rSOddyite, is directly related to the mole fraction of uranyl soddyite in a phase. Because of the rapid decrease in rSOddyite from pure soddyite, where rSOddyite = 1, to 1% Np-incorporated soddyite, where rSOddyite= 0.44, this relationship cannot remain linear much further into the solid solution. Another empirical fit, such as a Freundlich or Langmuir isotherm, may substitute for the linear Kd approach that is valid at low levels of Np incorporation. Additionally, solubility studies on soddyites with more than the 1% Np substitution in my study will provide a better understanding of the observed decrease in solubility. Future work should also include solubility studies of other Np-incorporated uranyl phases, such as uranophane and Nacompregnacite, that may incorporate significant amounts of Np.
The characterization of heavy metal and radionuclide transport in geologic settings requires knowledge of the amount and reactivity of each component in the system. My dissertation contributes to the understanding of these topics by improving existing thermodynamic models for metal adsorption and mineral solubility, and revealing a critical shortcoming in a widely used method of determining biomass in soils. As our understanding of individual mechanisms, such as those described here, increases, we will be able to model more accurately the distribution of metals in geologic systems of increasing complexity.
Ali, M. A., Dzombak, D. A. (1996) Effects of simple organic acids on sorption of Cu and Ca2+ on goethite. Geochim. Cosmochim. Acta 60:291-304.
Ams, D. A., Fein, J. B., Dong, H., Maurice, P. A. (2004) Experimental measurements of the adsorption of Bacillus subtilis and Pseudomonas mendocina onto Fe-oxyhydroxidecoated and uncoated quartz grains. Geomicrobiol. J. 21:511-519.
Baes, C. F., Jr., Mesmer, R. E. (1976) The Hydrolysis of Cations. Wiley, New York.
Bakken, L. R. (1985) Separation and Purification of Bacteria from Soil. Applied and Environmental Microbiology 49:1482-1487.