Janet S. Herman

Pyrite oxidation at circumneutral pH

Abstract Previous studies of pyrite oxidation kinetics have concentrated primarily on the reaction at low pH, where Fe(III) has been assumed to be the dominant oxidant. Studies at circumneutral pH, necessitated by effective pH buffering in some pyrite oxidation systems, have often implicitly assumed that the dominant oxidant must be dissolved oxygen (DO), owing to the diminished solubility of Fe(III). In fact, Fe(III)(aq) is an effective pyrite oxidant at circumneutral pH, but the reaction cannot be sustained in the absence of DO. The purpose of this experimental study was to ascertain the relative roles of Fe(III) and DO in pyrite oxidation at circumneutral pH. The rate of pyrite oxidation was first-order with respect to the ratio of surface area to solution volume. Direct determinations of both Fe(II)(aq)> and Fe(III)(aq) demonstrated a dramatic loss of Fe(II) from the solution phase in excess of the loss for which oxidation alone could account. Based on rate data, we have concluded that Fe(II) is adsorbed onto the pyrite surface. Furthermore, Fe(II) is preferred as an adsorbate to Fe(III), which we attribute to both electrostatic and acid-base selectivity. We also found that the rate of pyrite oxidation by either Fe(III)(aq) or DO is reduced in the presence of aqueous Fe(II), which leads us to conclude that, under most natural conditions, neither Fe(III)(aq) nor DO directly attacks the pyrite surface. The present evidence suggests a mechanism for pyrite oxidation that involves adsorbed Fe( II ) giving up electrons to DO and the resulting Fe(III) rapidly accepting electrons from the pyrite. The adsorbed Fe is, thus, cyclically oxidized and reduced, while it acts as a conduit for electrons traveling from pyrite to DO. Oxygen is transferred from the hydration sphere of the adsorbed Fe to pyrite S. The cycle of adsorbed Fe oxidation and reduction and the successive addition of oxygen to pyrite S continues until a stable sulfoxy species dissociates from the surface. Prior work has shown that sulfoxy species of lower oxidation state than sulfate (e.g., thiosulfate or polythionate) may accumulate in solution under some circumstances but not under the conditions of the experiments reported here. In these experiments, the rate of sulfate accumulation in solution is proportional to the rate of pyrite oxidation.

Bacterial transport in porous media: Evaluation of a model using laboratory observations

The factors that control the transport of bacteria through porous media are not well understood. The relative importance of the processes of dispersion, of immobilization of bacterial cells by various mechanisms (deposition), and of subsequent release of these trapped cells (entrainment) in describing transport has not been elucidated experimentally. Moreover, the variability of the phenomenological coefficients used to model these processes, given changes in such primary factors as grain size, organism, and ionic strength of the water, is unknown. We report results of fitting solutions of an advection-dispersion equation, modified to account for deposition and entrainment, to breakthrough curves from packed sand columns using two sizes of sand, two ionic strengths of the carrier solution, and two organisms with different sizes. A solution to the advection-dispersion equation including three processes, that is, dispersion, deposition, and entrainment, provides a match to the data that is superior to that achieved by solutions ignoring one of the processes. Fitted values of the coefficient describing deposition vary in a consistent manner with the control variables (organism, grain size, and ionic strength) and are generally within one order of magnitude of those predicted on the basis of theory.

The influence of mineralogy and solution chemistry on the attachment of bacteria to representative aquifer materials.

Abstract The rate and extent of bacterial attachment to mineral surfaces (chips of quartz, muscovite, limestone, and Fe-hydroxide-coated quartz and muscovite) was investigated by counting the numbers of bacterial cells (Lula-D, an indigenous groundwater organism) associated with each surface over time. The degree of attachment of cells to mineral surfaces was correlated with the sign of the surface charge as estimated from literature values for the isoelectric point; attachment of the negatively charged bacteria was much greater to the positively charged surfaces of limeston, Fe-hydroxide-coated quartz, and Fe-hydroxide-coated muscovite than to the negatively charged surfaces of clean quartz and clean muscovite. Batch experiments determined that the numbers of bacteria attached to the clean muscovite increased with increasing ionic strength of the solution, and the numbers attached to clean quartz were greater at pH 5 than at pH 7. In columns of clean quartz sand under saturated flow conditions, bacteria initially broke through at 1 pore volume but continued to elute for at least 7 pore volumes. Columns of Fe-hydroxide-coated sand retained more of the bacteria added to the columns (99.9% vs 97.4%), and the elution of cells ceased after the primary breakthrough. The results indicate that surface interactions between the mineral grains in an aquifer and the bacterial cells must play an essential role in determining the movement of bacteria through saturated porous media.

Differential dissolution of a Pleistocene reef in the ground-water mixing zone of coastal Yucatan, Mexico

A geochemical explanation is provided for the extensive dissolution observed along the carbonate coast of the Yucatan Peninsula, Mexico. Mixing of fresh ground water with subterranean Caribbean seawater generates a highly reactive geochemical zone that enhances aragonite and calcite dissolution and permits neomorphism of aragonite. Mixing-zone dissolution caused by ground-water discharge is a major geomorphic process in developing caves, coves, and crescent-shaped beaches along the Yucatan coast. Such dissolution has probably been a significant control on permeability and porosity distribution in carbonate rocks in the geologic record.

Spatial distribution of deposited bacteria following Miscible Displacement Experiments in intact cores

Miscible displacement experiments were performed on intact sand columns ranging from 15 to 60 cm in length to determine whether bacterial deposition varies at the centimeter scale within aquifer sediments. A 1-pore-volume pulse of radiolabeled cell suspension was introduced into the columns followed by a 2-pore-volume flush of artificial groundwater. The columns were then drained and dissected along the axis of flow. At ∼1-cm intervals, nine samples were removed for the enumeration of sediment-associated bacteria. Concentrations of sediment-associated (deposited) bacteria varied by up to 2 orders of magnitude in the direction perpendicular to flow demonstrating that bacterial deposition cannot be described mechanistically by a single rate coefficient. Incorporation of a distribution of sediment size and porosity values into Monte Carlo simulations indicates that physical heterogeneities are only partially responsible for the observed variability in deposited bacteria. A simple first-order model (classic filtration theory) adequately described the average spatial distribution of bacteria with depth within the 15-cm column. For the longer columns, however, the average concentration of deposited bacteria did not decrease exponentially with depth. A second-order model, modified to include an influent suspension of bacteria consisting of two subpopulations with separate sticking efficiencies (dual-alpha population), was required to describe the observed decreases of deposited bacteria with depth. A sensitivity analysis was performed with a first-order dual-alpha model to understand the effects of an influent suspension with two subpopulations of bacteria on the decrease of deposited bacteria with flow path length. Numerical simulations show that even for small fractions (0.01) of nonsticky bacteria, the decrease in deposited bacteria may deviate substantially from the exponential decrease expected from colloid-filtration theory. Results from experimental as well as numerical studies demonstrate the importance of column dissections for understanding bacterial deposition in saturated porous media.

Iron reduction in the sediments of a hydrocarbon-contaminated aquifer

Abstract Sediments sampled at a hydrocarbon-contaminated, glacial-outwash, sandy aquifer near Bemidji, Minnesota, were analyzed for sediment-associated Fe with several techniques. Extraction with 0.5 M HCl dissolved poorly crystalline Fe oxides and small amounts of Fe in crystalline Fe oxides, and extracted Fe from phyllosilicates. Use of Ti-citrate-EDTA-bicarbonate results in more complete removal of crystalline Fe oxides. The average HCl-extractable Fe(III) concentration in the sediments closest to the crude-oil contamination (16.2 μmol/g) has been reduced by up to 30% from background values (23.8 μmol/g) as a result of Fe(III) reduction in contaminated anoxic groundwater. Iron(II) concentrations are elevated in sediments within an anoxic plume in the aquifer. Iron(II) values under the oil body (19.2 μmol/g) are as much as 4 times those in the background sediments (4.6 μmol/g), indicating incorporation of reduced Fe in the contaminated sediments. A 70% increase in total extractable Fe at the anoxic/oxic transition zone indicates reoxidation and precipitation of Fe mobilized from sediment in the anoxic plume. Scanning electron microscopy detected authigenic ferroan calcite in the anoxic sediments and confirmed abundant Fe(III) oxyhydroxides at the anoxic/oxic boundary. The redox biogeochemistry of Fe in this system is coupled to contaminant degradation and is important in predicting processes of hydrocarbon degradation.

Effect of surface coatings, grain size, and ionic strength on the maximum attainable coverage of bacteria on sand surfaces.

The injection of bacteria in the subsurface has been identified as a potential method for in situ cleanup of contaminated aquifers. For high bacterial loadings, the presence of previously deposited bacteria can result in decreased deposition rates--a phenomenon known as blocking. Miscible displacement experiments were performed on short sand columns (approximately 5 cm) to determine how bacterial deposition on positively charged metal-oxyhydroxide-coated sands is affected by the presence of previously deposited bacteria. Approximately 8 pore volumes of a radiolabeled bacterial suspension at a concentration of approximately 1 x 10(9) cells ml-1 were introduced into the columns followed by a 2-pore-volume flush of cell-free buffer. It was found that the presence of Al- and Fe-coated sand increased both deposition rates and maximum fractional surface coverage of bacteria on the sediment surfaces. The effect of grain size on maximum bacterial retention capacity, however, was not significant. Decreasing ionic strength from 10(-1) to 10(-2) M KCl resulted in noticeable decreases in sticking efficiency (alpha) and maximum surface coverage (thetamax) for clean silica sand--results consistent with DLVO theory. In columns containing positively charged Al- and Fe-coated sands, however, changes in alpha and thetamax due to decreasing ionic strength were minimal. These findings demonstrate the importance of geochemical controls on the maximum bacterial retention capacity of sands.

Effect of bacterial cell shape on transport of bacteria in porous media

Parameters used to describe the transport of colloids (including bacteria) through porous media either implicitly or explicitly account for colloidal particle size but assume that the particles are spheres of uniform size. Bacteria found in soils and in aquifers exhibit a variety of shapes as well as sizes. We sought to determine if there exists a systematic effect of cell shape on the transport of bacteria in columns packed with clean quartz sand. A pulse of resting cells (14 strains of bacteria isolated from aquifers) suspended in an artificial groundwater was passed through a short column. Properties of the bacteria in the influent pulse were compared with those in the eluent from the columns. Cell shape, as quantified by the ratio of cell width to cell length, affects the transport of bacterial cells through porous media. In addition, the distributions of size and shape of cells in the effluent differed from those in the influent suspension with cells in the effluent being smaller and rounder. Short rods with low water contact angles (a measure of cell-surface hydrophobicity) showed the greatest decrease in cell length during passage through short columns. 21 refs., 3 figs.

Seasonal dynamics of groundwater-lake interactions at Doñana National Park, Spain

Abstract The hydrologic and solute budgets of a lake can be strongly influenced by transient groundwater flow. Several shallow interdunal lakes in southwest Spain are in close hydraulic connection with the shallow ground water. Two permanent lakes and one intermittent lake have chloride concentrations that differ by almost an order of magnitude. A two-dimensional solute-transport model, modified to simulate transient groundwater-lake interaction, suggests that the rising water table during the wet season leads to local flow reversals toward the lakes. Response of the individual lakes, however, varies depending on the lakes position in the regional flow system. The most dilute lake is a flow-through lake during the entire year; the through flow is driven by regional groundwater flow. The other permanent lake, which has a higher solute concentration, undergoes seasonal groundwater flow reversals at its downgradient end, resulting in complex seepage patterns and higher solute concentrations in the ground water near the lake. The solute concentration of the intermittent lake is influenced more strongly by the seasonal wetting and drying cycle than by the regional flow system. Although evaporation is the major process affecting the concentration of conservative solutes in the lakes, geochemical and biochemical reactions influence the concentration of nonconservative solutes. Probable reactions in the lakes include biological uptake of solutes and calcite precipitation; probable reactions as lake water seeps into the aquifer are sulfate reduction and calcite dissolution. Seepage reversals can result in water composition that appears inconsistent with predictions based on head measurements because, under transient flow conditions, the flow direction at any instant may not satisfactorily depict the source of the water. Understanding the dynamic nature of groundwater-lake interaction aids in the interpretation of hydrologic and chemical relations between the lakes and the ground water.

Kinetics of BTX biodegradation and mineralization in batch and column systems

Abstract Flow-through column and liquid batch experiments were performed in the present study in order to evaluate whether the kinetics of biodegradation reactions of organic contaminants for batch conditions were comparable to those measured under solid-to-solution ratios applicable to aquifer or water-saturated soil systems. The biodegradation of benzene, toluene, and xylene was observed under oxic conditions. Steady-state reaction rates were determined for the biodegradation reactions in the flow-through columns and evaluated using a rate law based on the Monod equation for conditions where bacterial growth is negligible. Calculated rate constants ( κ 1 ) for biodegradation, or substrate disappearance, for sole substrate experiments were 1.32 mmol L −1 h −1 for benzene, 1.42 mmol L −1 h −1 for toluene, and 0.833 mmol L −1 h −1 for xylene. Rate constants were determined for batch experiments using a rate law based on the Monod equation that does account for bacterial growth. The maximum specific growth rate, μ max , was found to be similar between batch and column experiments, indicating that there were no mass-transport limitations in the columns and that the solid-to-solution ratio was not a significant factor affecting kinetic parameters. There is considerable variability in rate constants for BTX biodegradation reported in the literature, up to two orders of magnitude for μ max . Rate constants from this study were within the range of published values. For the experiments reported here, rates determined for sole carbon sources could be used to predict the reaction rates of BTX mixtures given some adjustment of cell yields and lag times.