Bacterial communities were detected in deep crystalline rock aquifers within the Columbia River Basalt Group (CRB). CRB ground waters contained up to 60 μM dissolved H 2 and autotrophic microorganisms outnumbered heterotrophs. Stable carbon isotope measurements implied that autotrophic methanogenesis dominated this ecosystem and was coupled to the depletion of dissolved inorganic carbon. In laboratory experiments, H 2 , a potential energy source for bacteria, was produced by reactions between crushed basalt and anaerobic water. Microcosms containing only crushed basalt and ground water supported microbial growth. These results suggest that the CRB contains a lithoautotrophic microbial ecosystem that is independent of photosynthetic primary production.

The sorption of Cs ϩ was investigated over a large concentration range (10 Ϫ9 -10 Ϫ2 mol/L) on subsurface sediments from a United States nuclear materials site (Hanford) where high-level nuclear wastes (HLW) have been accidentally released to the vadose zone. The sediment sorbs large amounts of radiocesium, but expedited migration has been observed when HLW (a NaNO 3 brine) is the carrier. Cs ϩ sorption was measured on homoionic sediments (Na ϩ , K ϩ , Ca 2ϩ ) with electrolyte concentrations ranging from 0.01 to 1.0 mol/L. In Na ϩ electrolyte, concentrations were extended to near saturation with NaNO 3(s) (7.0 mol/L). The sediment contained nonexpansible (biotite, muscovite) and expansible (vermiculite, smectite) phyllosilicates. The sorption data were interpreted according to the frayed edge-planar site conceptual model. A fourparameter, two-site (high-and low-affinity) numeric ion exchange model was effective in describing the sorption data. The high-affinity sites were ascribed to wedge zones on the micas where particle edges have partially expanded due to the removal of interlayer cations during weathering, and the low-affinity ones to planar sites on the expansible clays. The electrolyte cations competed with Cs ϩ for both high-and low-affinity sites according to the trend K ϩ ϾϾ Na ϩ Ն Ca 2ϩ . At high salt concentration, Cs ϩ adsorption occurred only on high-affinity sites. Na ϩ was an effective competitor for the high-affinity sites at high salt concentrations.In select experiments, silver-thiourea (AgTU) was used as a blocking agent to further isolate and characterize the high-affinity sites, but the method was found to be problematic. Mica particles were handpicked from the sediment, contacted with Cs ϩ (aq) , and analyzed by electron microprobe to identify phases and features important to Cs ϩ sorption. The microprobe study implied that biotite was the primary contributor of high-affinity sites because of its weathered periphery. The poly-phase sediment exhibited close similarity in ion selectivity to illite, which has been well studied, although its proportion of high-affinity sites relative to the cation exchange capacity (CEC) was lower than that of illite. Important insights are provided on how Na ϩ in HLW and indigenous K ϩ displaced from the sediments may act to expedite the migration of strongly sorbing Cs ϩ in subsurface environments.

Abstract--Adsorption of uranyl to SWy-1 montmorillonite was evaluated experimentally and results were modeled to identify likely surface complexation reactions responsible for removal ofuranyl from solution. Uranyl was contacted with SWy-1 montmorillonite in a NaC104 electrolyte solution at three ionic strengths (I = 0.001, 0.01, 0.1), at pH 4 to 8.5, in a N2~) atmosphere. At low ionic strength, adsorption decreased from 95% at pH 4 to 75% at pH 6.8. At higher ionic strength, adsorption increased with pH from initial values less than 75%; adsorption edges for all ionic strengths coalesced above a pH of 7. A site-binding model was applied that treated SWy-1 as an aggregate of fixed-charge sites and edge sites analogous to gibbsite and silica. The concentration of fixed-charge sites was estimated as the cation exchange capacity, and non-preference exchange was assumed in calculating the contribution of fixed-charge sites to total uranyl adsorption. The concentration of edge sites was estimated by image analysis of transmission electron photomicrographs. Adsorption constants for uranyl binding to gibbsite and silica were determined by fitting to experimental data, and these adsorption constants were then used to simulate SWy-1 adsorption results. The best simulations were obtained with an ionization model in which A1OH2 + was the dominant aluminol surface species throughout the experimental range in pH. The pH-dependent aqueous speciation of uranyl was an important factor determining the magnitude of uranyl adsorption. At low ionic strength and low pH, adsorption by fixed-charge sites was predominant. The decrease in adsorption with increasing pH was caused by the formation of monovalent aqueous uranyl species, which were weakly bound to fixed-charge sites. At higher ionic strengths, competition with Na + decreased the adsorption of UOz 2+ to fixed-charge sites. At higher pH, the most significant adsorption reactions were the binding of UP22+ to A1OH and of(UO2)3(OH)5 + tO SiOH edge sites. Near-saturation of A1OH sites by UOz 1+ allowed significant contributions of SiOH sites to uranyl adsorption.

The subsurface behaviour of 99 Tc, a contaminant resulting from nuclear fuels reprocessing, is dependent on its valence (e.g., IV or VII). Abiotic reduction of soluble Tc(VII) by Fe(II) (aq) in pH 6-8 solutions was investigated under strictly anoxic conditions using an oxygen trap (<7.5 · 10 À9 atm O 2 ). The reduction kinetics were strongly pH dependent. Complete and rapid reduction of Tc(VII) to a precipitated Fe/Tc(IV) form was observed when 11 lmol/L of Tc(VII) was reacted with 0.4 mmol/L Fe(II) at pH 7.0 and 8.0, while no significant reduction was observed over 1 month at pH 6.0. Experiments conducted at pH 7.0 with Fe(II) (aq) = 0.05-0.8 mmol/L further revealed that Tc(VII) reduction was a combination of homogeneous and heterogeneous reaction. Heterogeneous reduction predominated after approximately 0.01 mmol/L of Fe(II) was oxidized. The heterogeneous reaction was more rapid, and was catalyzed by Fe(II) that adsorbed to the Fe/Tc(IV) redox product. Wet chemical and Fe-X-ray absorption near edge spectroscopy measurements (XANES) showed that Fe(II) and Fe(III) were present in the Fe/Tc(IV) redox products after reaction termination. 57 Fe-Mö ssbauer, extended X-ray adsorption fine structure (EXAFS), and transmission electron microscopy (TEM) measurements revealed that the Fe/Tc(IV) solid phase was poorly ordered and dominated by Fe(II)-containing ferrihydrite with minor magnetite. Tc(IV) exhibited homogeneous spatial distribution within the precipitates. According to Tc-EXAFS measurements and structural modeling, its molecular environment was consistent with an octahedral Tc(IV) dimer bound in bidentate edge-sharing mode to octahedral Fe(III) associated with surface or vacancy sites in ferrihydrite. The precipitate maintained Tc(IV) aq concentrations that were slightly below those in equilibrium with amorphous Tc(IV)O 2 AEnH 2 O (s) . The oxidation rate of sorbed Tc(IV) in the Fe/Tc precipitate was considerably slower than Tc(IV)O 2 AEnH 2 O (s) as a result of its intraparticle/intragrain residence. Precipitates of this nature may form in anoxic sediments or groundwaters, and the intraparticle residence of sorbed/precipitated Tc(IV) may limit 99 Tc remobilization upon the return of oxidizing conditions.

The potential for reduction of 99 TcO 4 Ϫ (aq) to poorly soluble 99 TcO 2 · nH 2 O (s) by biogenic sediment-associated Fe(II) was investigated with three Fe(III)-oxide containing subsurface materials and the dissimilatory metal-reducing subsurface bacterium Shewanella putrefaciens CN32. Two of the subsurface materials from the U.S. Department of Energy's Hanford and Oak Ridge sites contained significant amounts of Mn(III,IV) oxides and net bioreduction of Fe(III) to Fe(II) was not observed until essentially all of the hydroxylamine HCl-extractable Mn was reduced. In anoxic, unreduced sediment or where Mn oxide bioreduction was incomplete, exogenous biogenic TcO 2 · nH 2 O (s) was slowly oxidized over a period of weeks. Subsurface materials that were bioreduced to varying degrees and then pasteurized to eliminate biological activity, reduced TcO 4 Ϫ (aq) at rates that generally increased with increasing concentrations of 0.5 N HClextractable Fe(II). Two of the sediments showed a common relationship between extractable Fe(II) concentration (in mM) and the first-order reduction rate (in h Ϫ1 ), whereas the third demonstrated a markedly different trend. A combination of chemical extractions and 57 Fe Mössbauer spectroscopy were used to characterize the Fe(III) and Fe(II) phases. There was little evidence of the formation of secondary Fe(II) biominerals as a result of bioreduction, suggesting that the reactive forms of Fe(II) were predominantly surface complexes of different forms. The reduction rates of Tc(VII)O 4Ϫ were slowest in the sediment that contained plentiful layer silicates (illite, vermiculite, and smectite), suggesting that Fe(II) sorption complexes on these phases were least reactive toward pertechnetate. These results suggest that the in situ microbial reduction of sedimentassociated Fe(III), either naturally or via redox manipulation, may be effective at immobilizing TcO 4 Ϫ (aq) associated with groundwater contaminant plumes.