«THERMODYNAMIC MODELING OF METAL ADSORPTION AND MINERAL SOLUBILITY IN GEOCHEMICAL SYSTEMS A Dissertation Submitted to the Graduate School of the ...»
The inclusion of Na+ competition in the models significantly affects the calculated speciation of the cell wall functional groups. The differences in the speciation of Site 2 in models with and without the Na-Site 2 reaction are depicted in Figures 2.3B and 2.3C, respectively. In models that consider Na-Site 2 complexation, the Na-Site 2 complex represents the dominant surface species with increasing pH as the protonated site deprotonates. In addition, the stability of the Na-Site 2 complex is such that deprotonation occurs at lower pH than would be predicted from the model that does not consider Na-Site 2 complexation. Under the conditions of these calculations, when NaSite 2 complexation is included, the model predicts that half of the Site 2 sites are protonated at approximately pH 3.8, whereas in models that neglect the Na-Site 2 complexation this point occurs at pH 4.8. This same shift in speciation of the protonated site would occur if we modeled the reactions using electrostatic surface models. The NaSite 2 complexes represent the equivalent of Na+ counter-ions that would be predicted to accumulate in the diffuse layer of electrostatic models, with the only difference being the location of the Na+ cations relative to the cell wall functional group sites.
2.5 Conclusions Our results demonstrate that monovalent cations adsorb to bacterial cell walls under some conditions, and that the adsorption can be modeled with site-specific adsorption reactions. With this approach, the ionic strength effect on metal adsorption onto bacteria can be modeled as a competition for available binding sites between the metal of interest and the monovalent cations of the background electrolyte. The stability constant values for Li-, Rb-, and Na-bacterial surface complexes are reasonably close to ' each other, and reasonable estimates of monovalent adsorption behavior can be achieved by assuming a universal stability constant for all monovalent-bacterial surface complexes. Although monovalent cations adsorb much more weakly than do divalent and trivalent cations, in systems where the concentration of monovalent cations is much higher than that of the higher-charged metals that are present, the monovalent cations can effectively compete with other cations for available sites and diminish the extent of adsorption of those other cations. The approach of accounting for the ionic strength effect on adsorption by modeling it as competition for specific binding sites yields adsorption models that are easier to apply to complex geologic systems.
3.1 Introduction Soils are complex and variable mixtures of inorganic and organic components and free and adsorbed water. The inorganic fraction typically consists of minerals such as silica, feldspars, clays, and micas; the organic portion includes humic substances and microorganisms. These components represent a mixture of polar and nonpolar adsorbents that can strongly affect the distribution, speciation, and bioavailability of metals and organic pollutants in soils (e.g., Steffan and Akgerman, 1998). In order to develop quantitative adsorption models of contaminant distributions in soil systems, the amount of each component must be precisely determined.
Sequential extraction approaches (e.g., Tessier et al., 1979; Li et al., 1995) can be used to estimate the abundances of broad categories of minerals (carbonates, silicates, etc.) relative to organic components in soil. Furthermore, mineral identities and abundances in a soil sample can be reasonably estimated using x-ray diffractometry approaches (e.g., Brindley, 1984; Bish and Post, 1993). However, the approaches that have been developed to estimate biological cell mass or numbers in soils lack the precision that is necessary for the estimates to be useful in quantitative geochemical speciation and transport models. Commonly used methods of biomass determination include direct counting of bacteria using an optical microscope, and the use of adsorbent fluorescent dyes for spectrophotometric measurement of cell numbers. Bacterial counts based on optical microscopy or spectroscopy typically yield estimates of cell numbers with uncertainties of an order of magnitude or more, and fluorescent dyes can illuminate particles that are not bacteria leading to falsely high counts (Faegri et al., 1977;
Pogalazova et al., 1996). Lindahl and Bakken (1995) showed that physical dispersion methods to separate cells from soil, such as ultrasonication, blender, drill-gun, and shaking approaches, can damage cells. The percentage, viability, and purity of bacterial cells separated from soil also depend strongly on the dispersion technique and whether a surfactant is used in the separation procedure (Bakken, 1985).
Jenkinson and Powlson (1976a, b) introduced a biocidal fumigation method to determine the cell biomass-C in soils. Their work demonstrates that chloroform (CHCI3) fumigation effectively lyses cells in 24 hours. In this procedure, after the fumigation period, the fumigated soil sample and an un-fumigated control soil are placed in an incubator. The C from organisms that were killed and lysed during the fumigation process is readily mineralized to CO2, so that the difference in CO2 gas evolution between fumigated and un-fumigated samples is a measure of the biomass-C (Smith et al., 1995). Vance et al. (1987) and Tate et al. (1988) introduced a similar method that involves a fumigation step, but in this procedure organic C is extracted from fumigated and un-fumigated samples using a 0.5 M K2SO4 solution instead of an incubation approach. The K2SO4 solutions are then analyzed for total organic carbon (TOC), and biomass-C is calculated as the difference in TOC in the extracted solutions from the fumigated and the un-fumigated samples. The fumigation method is based on the assumption that the enhanced amount of organic C extracted from a sample relative to a control is due entirely from cell lysis caused by chloroform fumigation. Vance et al.
(1987) found an empirical linear relationship between biomass-C and organic C released by fumigation-extraction. Clearly, if chloroform sorbs onto any of the soil components and is extracted using the K2SO4 solution, then the subsequent measurements of enhanced TOC would in part be caused by the presence of chloroform and not only by lysis of biological organisms.
Although the fumigation method is commonly used, and is often considered the preferred method for biomass determination in soil (e.g., Franzluebbers, 1999), control experiments for the approach have never been performed, and so the accuracy of the procedure remains untested. Specifically, in applying the fumigation methods, one implicitly assumes that the introduced chloroform can be completely evacuated from the soil after the 24 h fumigation period. Because chloroform itself contains organic C and could add to the extractable C pool, this assumption must be tested with control experiments involving each of the major components of soils. Haney et al. (1999, 2001) questioned the acceptability of the fumigation-extraction method to determine biomassC by showing that the amount of C extracted using a 0.5 M K.2SO4 solution can vary significantly as a function of pH.
Chloroform is volatile under the experimental conditions, and volatile organic
carbon (VOC) vapors can sorb substantially to clays (e.g., Guo et al., 1998). Generally, montmorillonite, a 2:1 clay mineral, can sorb 200-300 mg VOC per g clay, and the 1:1
kaolinite has a sorption capacity of lA to Vi of that (Thibaud-Erkey et al., 1995). There is evidence that suggests that chloroform can sorb to soils both from aqueous solution (Dural and Peng, 1995) and from the atmosphere (Farrell and Reinhard, 1994; ThibaudErkey et al., 1995; Yeo et al., 1997; Chen and Dural, 2002), and therefore may show up as biomass C upon extraction with K2SO4.
In this paper, we test the validity of the fumigation-extraction method by performing fumigation control experiments with individual soil components, including humic acid, sand, clays, and bacteria. If the fumigation-extraction approach is valid, then we should observe no difference in TOC in the extracted solutions from the fumigated and un-fumigated samples, except for the bacterial samples, where cell lysis should enhance the extracted TOC in the fumigated samples. However, if chloroform sorbs onto any of the surfaces to a significant extent, then we would observe enhanced TOC in extracted solutions relative to the un-fumigated controls even though no biomass is present in those samples. Our control experiments will determine whether chloroform sorbs onto common soil components. If chloroform sorption does occur onto some soil components, then our experiments will constrain the types of soil for which the fumigation-extraction method can yield accurate biomass C analyses.
3.2 Materials and Methods Two clays, a silica sand, a humic acid, and a pure strain of bacteria were used as control materials for testing the fumigation-extraction method. The average grain diameter, with la uncertainties, of the silica sand (Accusand 40/50) was determined using scanning electron microscopy, and was found to be 469±89 \xm (n = 84). Schroth et al. (1996) analyzed a suite of Accusand grades and found the average diameter of the 40/50 grains to be 359±10 um (n = 4) using sieve analyses. The clays that were used in this study included a kaolinite (KGa-lb) and a Na-rich montmorillonite (SWy-2), both of which were obtained from the Source Clays Repository and have been characterized extensively. Dogan et al. (2006) found the mean BET surface areas of KGa-lb and SWy-2 to be 13.1 m2 g'1 and 22.7 m2 g"1, respectively. Cerato and Lutenegger (2002) estimated specific surface areas of 15 m2 g"1 for KGa-lb and 637 m2 g"1 for SWy-2 using the ethylene glycol monoethyl ether (EGME) method. Also available are chemical analyses (Mermut and Cano, 2001), infrared analyses (Madejova and Komadel, 2001) and powder x-ray diffraction analyses (Chipera and Bish, 2001) of both clays. We conducted fumigation tests using dried forms of both the kaolinite and montmorillonite samples, and we also conducted a test using a wetted montmorillonite powder in order to test whether wetness of the clay affects the results. In this case, the SWy-2 montmorillonite sample was wetted by soaking the clay in excess water for 1 h in 250 mL centrifuge tubes. After soaking, the tubes were centrifuged for 10 min at 7500 rpm three times and the water decanted each time. The amount of water retained by the clay after centrifugation was calculated by determining the weight difference between wet and dry SWy-2 samples, measuring the mass of the wet and dry clay samples before and after drying each in an oven at 105°C for 24 h. The dry SWy-2 sample contained
5.0±0.2% water by weight and the wet SWy-2 sample contained 82.8±1.3% water by weight. The humic acid experiments used a commercial humic acid, which has been characterized by Malcolm and MacCarthy (1986), obtained from the Aldrich Chemical Company. The silica sand, humic acid, and clays were used in experiments without washing or other modification. Bacillus subtilis, a gram-positive soil bacterial species, was grown from pure culture slants. Cells from the slant were transferred to 3 mL of trypticase soy broth (TSB) with 0.5% yeast extract and allowed to grow for 24 hours at 32°C. These cells were transferred to 1 L of the same media and allowed to grow for 24 hours at the same temperature. Cells were then harvested in early stationary phase by centrifugation and washed three times in test tubes in a 0.1 M NaC104 solution.
The fumigation procedure that we used closely followed the methods of Jenkinson and Powlson (1976a, b) and Vance et al. (1987). Experiments were conducted as a function of the mass of a soil component, with six experiments at different masses performed for each component. We used 0-50 g kaolinite, 0-50 g dry montmorillonite, 0-80 g wet montmorillonite, 0-50 g silica sand, 0-200 mg humic acid, and 0-200 mg bacteria. For each soil component separately, a known mass of material was spread evenly into two glass Petri dishes: one dish was fumigated with chloroform gas for 24 hours, and the other was an un-fumigated control. The fumigation samples were placed in a glass dessicator lined with moist filter paper. A small beaker that contained approximately 25 mL of ethanol-free chloroform and a few boiling chips was placed in the center of the dessicator. Once sealed, the dessicator was evacuated for 2 min, causing the chloroform to boil, thereby exposing the samples to chloroform vapor.
After 24 hours in the dark, the beaker of chloroform was removed from the dessicator, and the dessicator was evacuated 8 times to remove chloroform vapor from the samples.
Evacuations were performed by connecting the dessicator to a vacuum for 3 min, then sealing the dessicator from the vacuum, and finally slowly opening it to the atmosphere to refill with air. No noticeable chloroform odor existed after this procedure. We performed kinetics experiments, measuring the amount of organic C remaining on samples of dry S Wy-2 montmorillonite as a function of evacuation time. The samples exhibited extracted organic C concentrations that were independent of evacuation time, so all subsequent experiments were conducted with eight 3 min evacuations.