Colloid-facilitated transport of cesium in variably saturated Hanford sediments

Radioactive 137Cs has leaked from underground waste tanks into the vadose zone at the Hanford Reservation in south-central Washington State. There is concern that 137Cs, currently located in the vadose zone, can reach the groundwater. In this study, we investigated whether, and to what extent, colloidal particles can facilitate the transport of 137Cs at Hanford. We used colloidal materials isolated from Hanford sediments. Transport experiments were conducted under variably saturated, steady-state flow conditions in repacked, 20 cm long Hanford sediment columns, with effective water saturations ranging from 0.2 to 1.0. Cesium, pre-associated with colloids, was stripped off during transport through the sediments. The higher the flow rates, the less Cs was stripped off, indicating in part that Cs desorption from carrying colloids was a residence-time-dependent process. Depending on the flow rate, up to 70% of the initially sorbed Cs desorbed from colloidal carriers and was captured in the stationary sediments. Less Cs was stripped off colloids under unsaturated than under saturated flow conditions at similar flow rates. This phenomenon was likely due to the reduced availability of sorption sites for Cs on the sediments as the water content decreased and water flow was divided between mobile and immobile regions.     

In situ mobilization of colloids and transport of cesium in Hanford sediments

Radioactive waste, accumulated during Pu production, has leaked into the subsurface from underground storage tanks at the U.S. Department of Energy's Hanford site. The leaking solutions contained 137Cs and were of high ionic strength. Such a tank leak was simulated experimentally in steady-state flow experiments with packed Hanford sediments. The initial leak was simulated by a 1 M NaNO3 solution, followed by a decrease of ionic strength to 1 mM NaNO3. Cesium breakthrough curves were determined in both 1 M and 1 mM NaNO3 background. Colloidal particles were mobilized during the change of ionic strength. Mobilized colloids consisted mainly of quartz, mica, illite, kaolinite, and chlorite. Electrophoretic mobilities of colloids in the eluent solution were -3(microm/s)(V/cm) and increased to less negative values during later stages of mobilization. Mobilized colloids carried a fraction of the cesium along. While transport of cesium in 1 M NaNO3 background was much faster than in 1 mM NaNO3, cesium attached to colloids moved almost unretarded through the sediments. Cesium attached to mobilized colloids was likely associated with high affinity sorption sites on micas and illites.     

In situ colloid mobilization in Hanford sediments under unsaturated transient flow conditions: effect of irrigation pattern

Colloid transport may facilitate off-site transport of radioactive wastes at the Hanford site, Washington State. In this study, column experiments were conducted to examine the effect of irrigation schedule on releases of in situ colloids from two Hanford sediments during saturated and unsaturated transientflow and its dependence on solution ionic strength, irrigation rate, and sediment texture. Results show that transient flow mobilized more colloids than steady-state flow. The number of short-term hydrological pulses was more important than total irrigation volume for increasing the amount of mobilized colloids. This effect increased with decreasing ionic strength. At an irrigation rate equal to 5% of the saturated hydraulic conductivity, a transient multipulse flow in 100 mM NaNO3 was equivalent to a 50-fold reduction of ionic strength (from 100 mM to 2 mM) with a single-pulse flow in terms of their positive effects on colloid mobilization. Irrigation rate was more important for the initial release of colloids. In addition to water velocity, mechanical straining of colloids was partly responsible for the smaller colloid mobilization in the fine than in the coarse sands, although the fine sand contained much larger concentrations of colloids than the coarse sand.     

Transport of strontium and cesium in simulated hanford tank waste leachate through quartz sand under saturated and unsaturated flow

We investigated the effects of water saturation and secondary precipitate formation on Sr and Cs transport through quartz sand columns under saturated and unsaturated flow. Column experiments were conducted at effective water saturation ranging from 0.2 to 1.0 under steady-state flow using either 0.1 M NaNO(3) or simulated tank waste leachate (STWL; 1 M NaNO(3) and 1 M NaOH) mimicking Hanford (Washington, USA) tank waste. In 0.1 M NaNO(3) columns, Sr transported like a conservative tracer, whereas Cs was retarded relative to Sr. The transport of Sr and Cs in the 0.1 M NaNO(3) columns under all water saturations could be described with the equilibrium convection-dispersion equation (CDE). In STWL columns, Sr mobility was significantly reduced compared to the 0.1 M NaNO(3) column, because Sr was incorporated into or sorbed to neo-formed secondary precipitates. Strontium sequestration by precipitates was confirmed by additional batch and electron micrograph analyses. In contrast(,) the transport of Cs was less affected by the STWL; retardation of Cs in STWL columns was similar to that found in 0.1 M NaNO(3) columns. Analysis of STWL column data revealed that both Sr and Cs breakthrough curves showed nonideal behavior that suggest nonequilibrium conditions, although nonlinear geochemical behavior cannot be ruled out.     

Changes in uranium speciation through a depth sequence of contaminated Hanford sediments

The disposal of basic sodium aluminate and acidic U(VI)-Cu(ll) wastes in the now-dry North and South 300 A Process Ponds atthe Hanford site resulted in a groundwater plume of U(VI). To gain insight into the geochemical processes that occurred during waste disposal and those affecting the current and future fate and transport of this uranium plume, the solid-phase speciation of uranium in a depth sequence of sediments from the base of the North Process Pond through the vadose zone to groundwater was investigated using standard chemical and mineralogical analyses, electron and X-ray microprobe measurements, and X-ray absorption fine structure spectroscopy. Near-surface sediments contained uranium coprecipitated with calcite, which formed due to overneutralization of the waste ponds with base (NaOH). At intermediate depths in the vadose zone, metatorbernite [Cu(UO2PO4)2 x 8H2O] precipitated, likely during pond operations. Uranium occurred predominantly sorbed onto phyllosilicates in the deeper vadose zone and groundwater; sorbed uranium was also an important component at intermediate depths. Since the calcite-bearing pond sediments have been removed in remediation efforts, uranium fate and transport will be controlled primarily by desorption of the sorbed uranium and dissolution of metatorbernite.     

Colloid-facilitated Cs transport through water-saturated Hanford sediment and Ottawa sand

In this study, a series of saturated column experiments were conducted to investigate effects of colloids on Cs transport in two types of porous media (Hanford sediment characteristic of 2:1 clay minerals and silica Ottawa sand). The colloids used were obtained by reacting Hanford sediment with simulated tank waste solutions. Because of the highly nonlinear nature of Cs sorption found in batch experiments, we used two different concentrations of Cs (7.5 x 10(-5) M and 1.4 x 10(-8) M) for the transport experiments. The presence of colloids facilitated the transport of Cs through both Hanford sediment and Ottawa sand via association of Cs with mobile colloidal particles. Due to the nonlinearity of the Cs sorption, the colloid-facilitated Cs transportwas more pronounced atthe low Cs concentration (1.4 x 10(-8) M) than at the high concentration (7.5 x 10(-5) M) when expressed relative to the inflow Cs concentration. In the absence of colloids, no Cs moved through the 10-cm long columns during the experiment within about 20 pore volumes, exceptfor the high Cs concentration in the Ottawa sand where a complete Cs breakthrough was obtained. Also, it was found that colloid-associated Cs could be partially stripped off from colloids during the transport. The stripping effect was controlled by both Cs concentration and sorption capacity of the transport matrix.     

Distribution and retention of 137Cs in sediments at the Hanford Site, Washington

137Cesium and other contaminants have leaked from single-shell storage tanks (SSTs) into coarse-textured, relatively unweathered unconsolidated sediments. Contaminated sediments were retrieved from beneath a leaky SST to investigate the distribution of adsorbed 137Cs+ across different sediment size fractions. All fractions contained mica (biotite, muscovite, vermiculatized biotite), quartz, and plagioclase along with smectite and kaolinite in the clay-size fraction. A phosphor-plate autoradiograph method was used to identify particular sediment particles responsible for retaining 137Cs+. The Cs-bearing particles were found to be individual mica flakes or agglomerated smectite, mica, quartz, and plagioclase. Of these, only the micaceous component was capable of sorbing Cs+ strongly. Sorbed 137Cs+ could not be significantly removed from sediments by leaching with dithionite citrate buffer or KOH, but a fraction of the sorbed 137Cs+ (5-22%) was desorbable with solutions containing an excess of Rb+. The small amount of 137Cs+ that might be mobilized by migrating fluids in the future would likely sorb to nearby micaceous clasts in downgradient sediments.     

Effect of saline waste solution infiltration rates on uranium retention and spatial distribution in Hanford sediments

The accidental overfilling of waste liquid from tank BX-102 at the Hanford Site in 1951 put about 10 t of U(VI) into the vadose zone. In order to understand the dominant geochemical reactions and transport processes that occurred during the initial infiltration and to help understand current spatial distribution, we simulated the waste liquid spilling event in laboratory sediment columns using synthesized metal waste solution. We found that, as the plume propagated through sediments, pH decreased greatly (as much as 4 units) at the moving plume front. Infiltration flow rates strongly affect U behavior. Slower flow rates resulted in higher sediment-associated U concentrations, and higher flow rates (> or =5 cm/day) permitted practically unretarded U transport. Therefore, given the very high Ksat of most of Hanford formation, the low permeability zones within the sediment could have been most important in retaining high concentrations of U during initial release into the vadose zone. Massive amount of colloids, including U-colloids, formed at the plume fronts. Total U concentrations (aqueous and colloid) within plume fronts exceeded the source concentration by up to 5-fold. Uranium colloid formation and accumulation at the neutralized plume front could be one mechanism responsible for highly heterogeneous U distribution observed in the contaminated Hanford vadose zone.     

Adsorption of natural organic matter to air-water interfaces during transport through unsaturated porous media

To better understand how interactions with the air phase influence the movement of natural organic matter (NOM) through the vadose zone, we measured the transport of soil-humic acid (SHA) through laboratory columns packed with partially saturated sand. Our results demonstrate that sorptive reactions at air-water interfaces reduce SHA mobility and that the affinity of SHA for the air phase increases as the porewater pH declines from 8 to 3.9. SHA desorption from air-water interfaces is negligible for conditions of constant pH, but release of bound SHA occurs in response to perturbations in porewater pH. We analyzed the effluent samples collected from our laboratory columns using high-performance size-exclusion chromatography. The results of this analysis demonstrate that the SHA did not fractionate appreciably during transport through the columns, suggesting that the various components of the SHA pool (as distinguished on the basis of molecular weight) express an equal affinity for the air-water interfaces over the range of pH conditions tested. A mathematical model incorporating irreversible, second-order rate laws to simulate adsorption at air-water and solid-water interfaces closely describes the SHA breakthrough data. The mass-transfer parameters that govern this model vary in a discernible fashion with changes in porewater pH, and the parameter trends are consistent with published theories for SHA adsorption.     

Tracking sources of unsaturated zone and groundwater nitrate contamination using nitrogen and oxygen stable isotopes at the Hanford site, Washington

The nitrogen and oxygen isotopic compositions of nitrate in pore water extracts from unsaturated zone (UZ) core samples and groundwater samples indicate at least four potential sources of nitrate in groundwaters at the U.S. DOE Hanford Site in south-central Washington. Natural sources of nitrate identified include microbially produced nitrate from the soil column (delta15N of 4 - 8 per thousand, delta18O of -9 to 2 per thousand) and nitrate in buried caliche layers (delta15N of 0-8 per thousand, delta 18O of -6to 42 per thousand). Isotopically distinctindustrial sources of nitrate include nitric acid in low-level disposal waters (delta15N approximately per thousand, delta 18O approximately 23%o) per thousandnd co-contaminant nitrate in high-level radioactive waste from plutonium processing (6'5delta1of 8-33 % o, per thousand18delta oO -9 to 7%0). per thousandThe isotopic compositions of nitrate from 97 groundwater wells with concentrations up to 1290 mg/L NO3- have been analyzed. Stable isotope analyses from this study site, which has natural and industrial nitrate sources, provide a tool to distinguish nitrate sources in an unconfined aquiferwhere concentrations alone do not. These data indicate that the most common sources of high nitrate concentrations in groundwater at Hanford are nitric acid and natural nitrate flushed out of the UZ during disposal of low-level wastewater. Nitrate associated with high-level radioactive UZ contamination does not appear to be a major source of groundwater nitrate at this time.     

Influence of oxidation states on plutonium mobility during long-term transport through an unsaturated subsurface environment

Lysimeter and laboratory studies were conducted to identify the controlling chemical processes influencing Pu(IV) mobility through the vadose zone. A 52-L lysimeter containing sediment from the Savannah River Site, South Carolina and solid PuIV(NO3)4 was left exposed to natural wetting and drying cycles for 11 years before the lysimeter sediment was sampled. Pu had traveled 10 cm, with >95% of the Pu remaining within 1.25 cm of the source. Laboratory studies showed that the sediment quickly reduced Pu(V) to Pu(IV) (the pseudo-first-order reduction rate constant, Kobs, was 0.11 h(-1)). Of particular interest was that this same sediment could be induced to release very low concentrations of sorbed Pu under oxidizing conditions, presumably by oxidation of sorbed Pu(IV) to the more mobile Pu(V) species. Transport modeling supported the postulation that Pu oxidation occurred in the lysimeter sediment; the inclusion of an oxidation term in the model produced simulations that capture the Pu depth profile data. By not including the oxidation process in the model, Pu mobility was grossly underestimated by a factor of 3.5. It is concluded that both oxidation and reduction mechanisms can play an important role in Pu transportthrough the vadose zone and should be considered when evaluating disposal of Pu-bearing wastes.     

Spectroscopic and diffraction study of uranium speciation in contaminated vadose zone sediments from the Hanford site, Washington state

Contamination of vadose zone sediments under tank BX-102 at the Hanford site, Washington, resulted from the accidental release of 7-8 metric tons of uranium dissolved in caustic aqueous sludge in 1951. We have applied synchrotron-based X-ray spectroscopic and diffraction techniques to characterize the speciation of uranium in samples of these contaminated sediments. UIII-edge X-ray absorption fine structure (XAFS) spectroscopic studies demonstrate that uranium occurs predominantly as a uranium(VI) silicate from the uranophane group of minerals. XAFS cannot distinguish between the members of this mineral group due to the near identical local coordination environments of uranium in these phases. However, these phases differ crystallographically, and can be distinguished using X-ray diffraction (XRD) methods. As the concentration of uranium was too low for conventional XRD to detect these phases, X-ray microdiffraction (microXRD) was used to collect diffraction patterns on approximately 20 microm diameter areas of localized high uranium concentration found using microscanning X-ray fluorescence (microSXRF). Only sodium boltwoodite, Na(UO2)(SiO3OH) x 1.5H20, was observed; no other uranophane group minerals were present. Sodium boltwoodite formation has effectively sequestered uranium in these sediments under the current geochemical and hydrologic conditions. Attempts to remediate the uranium contamination will likely face significant difficulties because of the speciation and distribution of uranium in the sediments.     

Residual waste from Hanford tanks 241-C-203 and 241-C-204. 2. Contaminant release model

Release of U and 99Tc from residual sludge in Hanford waste tanks 241-C-203 and 241-C-204 atthe U.S. Department of Energy's (DOE) Hanford Site in southeastern Washington state was quantified by water-leaching, selective extractions, empirical solubility measurements, and thermodynamic modeling. A contaminant release model was developed based on these experimental results and solid-phase characterization results presented elsewhere. Uranium release was determined to be controlled by two phases and occurred in three stages. In the first stage, U release is controlled by the solubility of tejkaite, which is suppressed by high concentrations of sodium released from the dissolution of NaNO3 in the residual sludges. Equilibrium solubility calculations indicate the U released during this stage will have a maximum concentration of 0.021 M. When all the NaNO3 has dissolved from the sludge, the solubility of the remaining cejkaite will increase to 0.28 M. After cejkaite has completely dissolved, the majority of the remaining U is in the form of poorly crystalline Na2U2O7 [or clarkeite Na[(UO2)O(OH)](H20)0-1]. In contact with Hanford groundwater this phase is not stable, and becquerelite becomes the U solubility controlling phase, with a calculated equilibrium concentration of 1.2 x 10(-4) M. For Tc, a significant fraction of its concentration in the residual sludge was determined to be relatively insoluble (20 wt % for C-203 and 80 wt % for C-204). Because of the low concentrations of Tc in these sludge materials, the characterization studies did not identify any discrete Tc solids phases. Release of the soluble fraction of Tc was found to occur concomitantly with NO3-. It was postulated that a NaNO3-NaTcO4 solid solution could be responsible for this behavior. The Tc release concentrations for the soluble fraction were estimated to be 2.4 x 10-6 M for C-203 and 2.7 x 10(-5) M for C-204. Selective extraction results indicated that the recalcitrant fraction of Tc was associated with Fe oxides. Release of the recalcitrant fraction of Tc was assumed to be controlled by dissolution of Fe oxide in the form of ferrihydrite. Based on this assumption and measured values for the ratio of recalcitrant Tc to total Fe in each bulk sludge, the release concentration of the recalcitrant fraction of Tc was calculated to be 3.9 x 10(-12) M for C-203 and 10.0 x 10(-12) M for C-204.     

In-mouth mechanisms leading to flavor release and perception

During eating, foods are submitted to two main oral processes-chewing, including biting and crushing with teeth, and progressive impregnation by saliva resulting in the formation of a cohesive bolus and swallowing of the bolus. Texture influences the chewing behavior, including mastication and salivation, and in turn, these parameters influence texture perception and bolus formation. During this complex mouth process, flavor compounds are progressively released from the food matrix. This phenomenon is mainly dependent on the food texture, the composition and in-mouth breakdown, and on saliva impregnation and activity, but an individual's anatomical and physiological aspects characteristics should also be taken into account. This article reviews the knowledge and progresses on in-mouth processes leading to food breakdown and flavor release and affecting perception. Relationships between food texture and composition, food breakdown, oral physiology, and flavor release are developed and discussed. This review includes not only the mechanical aspects of oral physiology but also the biological aspects such as the influence of saliva composition, activity, and regulation on flavor perception. In vitro and in silico approaches are also described.     

Effect of temperature on Cs+ sorption and desorption in subsurface sediments at the Hanford Site,U.S.A

The effects of temperature on Cs+ sorption and desorption were investigated in subsurface sediments from the U.S. Department of Energy Hanford Site. The site has been contaminated at several locations by the accidental leakage of high-level nuclear waste (HLW) containing 137Cs+. The high temperature of the self-boiling, leaked HLW fluid and the continuous decay of various radionuclides carried by the waste supernatant have resulted in elevated vadose temperatures (currently up to 72 degrees C) below the Hanford S-SX tank farm that have dissipated slowly from the time of leakage (1970). The effect of temperature on Cs+ sorption was evaluated through batch binary Cs(+)-Na+ exchange experiments on pristine sediments, while Cs+ desorption was studied in column experiments using 137Cs(+)-contaminated sediments. Cs+ adsorption generally decreased with increasing temperature, with a more apparent decrease at low aqueous Cs+ concentration (10(-10)-10(-6) mol/L). Cs+ desorption from the contaminated sediments increased with increasing temperature. The results indicated that the free energy of Na(+)-Cs+ exchange on the Hanford sediment had a significant enthalpy component that was estimated to be -17.87 (+/- 2.01) and -4.82 (+/- 0.44) kJ/mol (at 298 degrees C) for the high- and low-affinity exchange sites, respectively. Both Cs+ adsorption and desorption at elevated temperature could be well simulated by a two-site ion exchange model, with the conditional exchange constants corrected by the exchange enthalpy effect. The effect of temperature on Cs+ desorption kinetics was also evaluated using a stop-flow technique. The kinetics of desorption of the exchangeable pool (which was less than the total adsorbed concentration) were found to be rapid under the conditions studied.     

Kinetic desorption and sorption of U(VI) during reactive transport in a contaminated Hanford sediment

Column experiments were conducted to investigate U(VI) desorption and sorption kinetics in a sand-textured, U(VI)-contaminated (22.7 micromol kg(-1)) capillary fringe sediment from the U.S. Department of Energy (DOE) Hanford site. Saturated column experiments were performed under mildly alkaline conditions representative of the Hanford site where uranyl-carbonate and calcium-uranyl-carbonate complexes dominate aqueous speciation. A U(VI)-free solution was used to study contaminant U(VI) desorption in columns where different flow rates were applied. Sorbed, contaminant U(VI) was partially labile (11.8%), and extended leaching times and water volumes were required for complete desorption of the labile fraction. Uranium-(VI) sorption was studied after the desorption of labile, contaminant U(VI) using different U(VI) concentrations in the leaching solution. Strong kinetic effects were observed for both U(VI) sorption and desorption, with half-life ranging from 8.5 to 48.5 h for sorption and from 39.3 to 150 h for desorption. Although U(VI) is semi-mobile in mildly alkaline, subsurface environments, we observed substantial U(VI) adsorption, significant retardation during transport, and atypical breakthrough curves with extended tailing. A distributed rate model was applied to describe the effluent data and to allow comparisons between the desorption rate of contaminant U(VI) with the rate of shortterm U(VI) sorption. Desorption was the slower process. We speculate that the kinetic behavior results from transport or chemical phenomena within the phyllosilicate-dominated fine fraction present in the sediment. Our results suggest that U(VI) release and transport in the vadose zone and aquifer system from which the sediment was obtained are kinetically controlled.     

Heavy metal immobilization through phosphate and thermal treatment of dredged sediments

Disposal of dredged sediments is expensive and poses a major challenge for harbor dredging projects. Therefore beneficial reuse of these sediments as construction material is highly desirable assuming contaminants such as heavy metals are immobilized and organics are mineralized. In this research, the effect of the addition of 2.5% phosphate, followed by thermal treatment at 700 degrees C, was investigated for metal contaminants in dredged sediments. Specifically, Zn speciation was evaluated, using X-ray absorption spectroscopy (XAS), by applying principal component analysis (PCA), target transformation (TT), and linear combination fit (LCF) to identify the main phases and their combination from an array of reference compounds. In dredged sediments, Zn was present as smithsonite (67%) and adsorbed to hydrous manganese oxides (18%) and hydrous iron oxides (15%). Phosphate addition resulted in precipitation of hopeite (22%), while calcination induced formation of spinels, gahnite (44%), and franklinite (34%). Although calcination was previously used to agglomerate phosphate phases by sintering, we found that it formed sparingly soluble Zn phases. Results from the U.S. EPA toxicity characteristic leaching procedure (TCLP) confirmed both phosphate addition and calcination reduced leachability of heavy metals with the combined treatment achieving up to an 89% reduction.     

Gas-phase and transpiration-driven mechanisms for volatilization through wetland macrophytes

Natural and constructed wetlands have gained attention as potential tools for remediation of shallow sediments and groundwater contaminated with volatile organic compounds (VOCs). Wetland macrophytes are known to enhance rates of contaminant removal via volatilization, but the magnitude of different volatilization mechanisms, and the relationship between volatilization rates and contaminant physiochemical properties, remain poorly understood. Greenhouse mesocosm experiments using the volatile tracer sulfur hexafluoride were conducted to determine the relative magnitudes of gas-phase and transpiration-driven volatilization mechanisms. A numerical model for vegetation-mediated volatilization was developed, calibrated with tracer measurements, and used to predict plant-mediated volatilization of common VOCs as well as quantify the contribution of different volatilization pathways. Model simulations agree with conclusions from previous work that transpiration is the main driver for volatilization of VOCs, but also demonstrate that vapor-phase transport in wetland plants is significant, and can represent up to 50% of the total flux for compounds with greater volatility like vinyl chloride.     

Transport of silica colloids through unsaturated porous media: experimental results and model comparisons

We present results on the migration of silica colloids through laboratory columns packed with partially saturated quartz sand. The transport of the silica colloids responds to changes in the steady-state volumetric moisture content (theta) and for low theta depends on the wetting history of the sand pack prior to colloid injection. A mathematical model that incorporates a first-order rate law to simulate film straining and a second-order rate law to simulate partitioning at air-water interfaces closely describes colloid transport and mass transfer over the range of experimental conditions tested. The mass-transfer parameters of the model are sensitive to changes in both the level of water saturation and the flow rate. A semiempirical expression, based on a modification of film-straining theory, accounts for the observed variation in the first-order rate coefficient with changes in theta and average porewater velocity. Our work indicates that the presence of the air phase substantially influences porewater concentrations of mineral colloids in water-unsaturated media and that the kinetics of particle removal attributed to air-water boundaries reflects the contribution of multiple mass-transfer mechanisms.     

Virus retention and transport in chemically heterogeneous porous media under saturated and unsaturated flow conditions

Retention and transport of colloids and microorganisms are complex processes, especially in the vadose zone due to the more complicated water flow regime and additional interfacial reactions involved. In this study, we examined the retention and transport behavior of two bacteriophages, MS-2 and phiX174, in homogeneous and chemically heterogeneous media under variably saturated conditions. Column experiments with glass beads (treated to have either hydrophilic or hydrophobic surface properties) were conducted using a phosphate-buffered saline solution at different pore water ionic strengths ranging from 0.025 to 0.163 M. In columns packed with 100% hydrophilic glass beads, retention of the viruses increased with decreasing water content and increasing ionic strength, a result similar to those reported in the literature. However, greater retention of both MS-2 and phiX174 was observed in saturated columns than in unsaturated columns packed with a 1:1 mixture of hydrophilic and hydrophobic glass beads, especially at high ionic strengths. This result contradicts the common belief that viruses (and colloids in general) are subject to greater removal in unsaturated media. Our study suggests that while the mechanisms controlling colloid interfacial interactions (i.e., attachment on solid-water and air-water interfaces and film straining) on the pore scale are relevant, nonuniform wetting conditions due to heterogeneous grain surface hydrophobicity can strongly influence water flow and phase interconnection. Under these conditions, hydrodynamic effects on the mesopore scale will dominate pore-scale interfacial reactions in controlling the extent of colloid retention and movement in unsaturated media.