Pyrite Oxidation in Unsaturated Aquifer Sediments. Reaction

With few exceptions the argon content at the end of the experiment was ...... of sulphate on groundwaters of Triassic carbonate aquifers, Upper Silesi...
0 downloads 0 Views 96KB Size
Download PDF
Environ. Sci. Technol. 2001, 35, 4074-4079

Pyrite Oxidation in Unsaturated Aquifer Sediments. Reaction Stoichiometry and Rate of Oxidation MARTIN SØGAARD ANDERSEN,* FLEMMING LARSEN, AND DIEKE POSTMA Groundwater Research Centre, Environment & Resources, Technical University of Denmark, Building 204, DK 2800 Kgs Lyngby, Denmark

The oxidation of pyrite (FeS2) contained in unsaturated aquifer sediment was studied by sediment incubation in gas impermeable polymer laminate bags. Reaction progress was followed over a period of nearly 2 months by monitoring the gas composition within the laminate bag. The gas phase in the incubation bags became depleted in O2 and enriched in CO2 and N2 and was interpreted as due to pyrite oxidation in combination with calcite dissolution. Sediment incubation provides a new method to estimate low rates of pyrite oxidation in unsaturated zone aquifer sediments. Oxidation rates of up to 9.4‚10-10 mol FeS2/g‚s are measured, and the rates are only weakly correlated with the sediment pyrite content. The reactivity of pyrite, including the inhibition by FeOOH layers formed on its surface, apparently has a major effect on the rate of oxidation. The code PHREEQC 2.0 was used to calculate the reaction stoichiometry and partitioning of gases between the solution and the gas phase. Pyrite oxidation with concurrent calcite dissolution was found to be consistent with the experimental data while organic carbon oxidation was not. The reaction involves changes in the total volume of the gas phase. The reaction scheme predicts the volume of O2 gas consumed to be larger than of CO2 produced. In addition the solubility of CO2 in water is about 30 times larger than of O2 causing a further decrease in total gas volume. The change in total gas volume therefore also depends on the gas/water volume ratio and the lower the ratio the more pronounced the loss of volume will be. Under field conditions the change in total volume may amount up to 20% in the absence of calcite and over 10% in the presence of calcite. Such changes in gas volume during the oxidation of pyrite are expected to result in pressure gradients causing advective transport of gaseous oxygen.

Introduction The oxidation of pyrite (FeS2) is a process of major environmental impact. It is the cause of acid mine drainage, the development of acid sulfate soils, and of aquifer contamination (1, 2). The problems with pyrite oxidation arise when atmospheric oxygen gains access to sediment layers containing pyrite. In aquifers, groundwater abstraction often causes dewatering of sediments, and the resulting oxidation of pyrite produces a high groundwater sulfate concentration, a low pH, and enhanced heavy metal (Ni, Co, As, Zn) contents * Corresponding author phone: (+45) 45 25 22 57; fax: (+45) 45 88 59 35; e-mail: [email protected]. 4074

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 20, 2001

(3, 4). In association with drained sulfide mines and tailings the effects are even more severe (5-7). The processes involved in pyrite oxidation in unsaturated sediments or mine tailings are complex. One important aspect is the transport of gaseous O2 to the site of pyrite oxidation. In mine tailings O2 may become depleted from the soil gas at less than a meter below the surface and diffusive transport of O2 becomes important (8). In some aquifers the unsaturated zone has restricted connection to the atmosphere, and advective gas transport driven by atmospheric pressure variations may also contribute to O2 transport (9). In addition the consumption of O2 by pyrite oxidation may cause a loss in total gas volume and induce O2 transport along a pressure gradient (10), but such changes in gas volume during pyrite oxidation have not been studied well. Also reaction kinetics may be a limiting factor for pyrite oxidation in the field (1, 2). Pyrite oxidation has been studied extensively in the laboratory with focus on both the inorganic reaction kinetics (11-13) and bacterial catalysis (14, 15). Few methods are available to estimate field rates of pyrite oxidation. In mine tailings with a steep O2 gradient in the soil air the diffusive flux of O2 passing through the tailing surface can be estimated using Fick’s first law (8). The flux of O2 into the tailing may also be derived from O2 depletion in a headspace chamber and Ficks second law (16). Others have used the accumulation of sulfate over time as a measure for the rate of pyrite oxidation (17, 18). These methods measure the cumulative effect of oxygen transport and the rate of pyrite oxidation. With the sulfate accumulation method, there are also problems with heterogeneous pyrite oxidation and sulfate leaching from the sediment. In aquifer sediments the concentration of pyrite is generally much lower than in mine tailings, and the above methods become difficult to apply. This paper presents a new approach toward estimating the rate of pyrite oxidation in aquifer sediments containing only small amounts of pyrite. Unamended aquifer sediment is incubated in gas-impermeable polymer laminate bags, and the gas composition is monitored over time. The change in gas composition in the headspace over time is used to determine the rate of pyrite oxidation and to evaluate the reaction stoichiometry. PHREEQC modeling is used to verify the reaction scheme and to calculate the partitioning of gases between the gas and aqueous phase and the resulting volume effects during the oxidation of pyrite.

Materials and Methods Sediments. The sediment used in the incubation experiments comes from the Beder well field, near Aarhus, Denmark (4, 9, 19). The well field is situated in a sandy aquifer below a clayey till cap. Groundwater abstraction has lowered the water table by more than 10 m causing pyrite oxidation in the sediment below the till cap. Sediment samples for incubation were taken from the dewatered sandy sediment within the zone of pyrite oxidation. Therefore all sediment samples have been subject to partial pyrite oxidation under field conditions before incubation in the laboratory. Sediments were collected using a piston corer with 5 cm diameter PVC core liners. Cores were sectioned, sealed, and frozen for storage. Incubation Techniques. The sediment were incubated in bags consisting of polymer laminate. The polymer laminate used was produced by Danisco Flexible Inc. for storage of food stuffs. Going from inside to outside the laminate consisted of the following layers: 72 µm low-density polyethylene, 16 µm polyamide, 15 µm aluminum foil, and 16 µm polyamide. The layers are glued together. The aluminum foil acts as a gas diffusion barrier, and the polyamide provides 10.1021/es0105919 CCC: $20.00

 2001 American Chemical Society Published on Web 09/08/2001

TABLE 1. Sediment Properties and Experimental Incubation Conditions and Rate of Pyrite Oxidation Expressed in Mol of Pyrite Oxidized per Gram of Pyrite per Second sample

pyrite, mmol/kg

Org-C, mmol/kg

Carb-C, mmol/kg

gas vol ,mL

gas/water ratio

sed. wt, g

rate, mol/g‚s

Hb5 5,8 Rb11 11,1 Rb11 6,4 Rb13 13,8

1.59 7.18 6.42 2.80

144. 17.8 19.3 29.9

490. 406. 452 423

21.3 33.8 40.0 38.2

3.1 17.4 17.2 10.5

82.79 99.68 103.4 105.3

3.52‚10-10 0.75‚10-10 0.14‚10-10 1.02‚10-10

tear strength. The inside cover of polyethylene allows the laminate to be welded together by heat. A commercial ElwisPack PV50242 plastics welding apparatus was used for making and sealing the sediment incubation bags. Sediment Incubation. About 100 g of sediment, retaining its natural moisture content, was placed in each laminate bag, and the bag was then sealed by welding. Using needles inserted into the bag, the atmospheric air within the bag was replaced by a gas mixture, consisting of 8% O2 and 92% N2, by flushing its volume three times. The final gas volume within the bag was adjusted to about 40 mL. Subsequently the reaction bags were incubated at in situ groundwater temperatures of 8 °C. The sediment volume within each bag was calculated from the weight of the sediment and a solid-phase density of 2.65 g/mL, equal to that of quartz. The water volume was determined by weighing and drying a subsample at 105 °C for 24 h. The total volume of the bag was determined by water displacement at 20 °C. The volume of the empty bag was determined independently. The gas volume in the bag is then obtained as the difference. The pyrite content of the sediments was determined using the Cr(II)-reduction method (20-22). The sediment carbon content was determined by combustion and IR detection of CO2 using a LECO CS-225. The carbonate content was determined as the loss in carbon content when the sample is acidified. Monitoring the Gas-Phase Composition. Gas samples were taken from the laminate bags using gastight syringes and needles. A piece of low diffusive adhesive tape was partially mounted on the laminate bag. The sampling needle was inserted through the laminate at a low angle, and the tape was immediately pressed over the needle to provide a seal. The syringe volume was flushed three times with the gas within the bag, and then a 0.2 mL sample was taken for analysis. The needle was gently withdrawn, and the protective tape was pressed over the hole. Subsequently the sampling spot was welded together using a pair of flat tongs heated to 165 °C. The whole sampling procedure was carried out in duplicate. The gas composition was determined using a ML GC 82 gas chromatograph with a Shintzu C.B3A integrator. Helium was used as carrier gas, and the GC was operated at 60 °C. A Porepac column was used to separate CO2 from O2, Ar, and N2. A Molsi column did separate O2+Ar from N2. Thus two injections of 0.1 mL each were done for each sample to obtain the full gas composition. The argon content of the gas in the incubation bag was measured only at the start and end of each experiment. For this purpose a Haye Sep A column was mounted in the GC and cooled to a temperature in the range +2 and -4 °C. The precision of GC measurements was determined by replicate analysis of air, giving a standard deviation of 0.22% for O2+Ar and 0.73% for N2. A 5% CO2 standard in N2 gave a precision of 0.15% for CO2. Several tests were carried out to check the methodology employed. Incubation bags without sediment were monitored for 3 months and showed no change in gas composition. Since the gas composition within in the incubation bags differs from the atmosphere, there is apparently no leakage. The gas in the incubation bags is also free of argon, while atmospheric air contains about 0.93% argon. With few

exceptions the argon content at the end of the experiment was always less than 0.1%, the detection limit of the method. The general conclusion is therefore that the incubation bags are tight and can be sampled without contamination.

Results and Discussion Reaction Stoichiometry. Table 1 lists the composition of some of the aquifer sediments used for incubation. The pyrite content of the sediment generally ranges from 1 to 10 mmol/ kg although a few samples had a higher pyrite content. The CaCO3 content is in the range of 406-490 mmol/kg and occurs as finely dispersed CaCO3. The organic carbon content is between 18 and 30 mmol/kg except for one sample with 144 mmol/kg. The natural water content of the sediment samples ranges from 1 to 17% in correspondence with their origin in the unsaturated zone. Figure 1 shows examples of the development in gas composition during the sediment incubation. The general trend is a decrease in O2 and an increase in CO2 and N2. However, there is a major difference in the extent of O2 consumption. In sample Hb5 5,8 the O2 becomes almost depleted after about 1000 h, while in Rb11 6,4 the O2 only has decreased from the initial value of 8.5% to 7% after 1400 h. The decrease in O2 is presumably due to pyrite oxidation, while the increase in CO2 suggests concurrent CaCO3 dissolution. The overall reactions are

FeS2 + 15/4O2 + 7/2H2O f Fe(OH)3 + 2SO42- + 4H+

(1)

followed by

2CaCO3 + 4H+ f 2Ca2+ + 2CO2 + 2H2O

(2)

The volume of CO2 gas produced by reaction 2 is less than the volume of O2 gas consumed by reaction 1. The gas mixture used in our experiments contained 8% O2 and following the reaction scheme 8% × 8/15 ) 4.27% CO2 should be produced. The laminate bags are a system of constant total pressure, and the total volume of gas in contact with oxidizing pyrite should therefore decrease by 8 - 4.27 ) 3.73%. Since N2 behaves chemically conservative and is quite insoluble in water, the relative contribution of N2 to the gas-phase composition must increase when the sum of O2 and CO2 decrease, as is observed experimentally (Figure 1). Treating N2 as an inert and insoluble gas, the loss in volume of gas phase is given by

∆Vloss ) Vo(1 - N2(o)/N2(t))

(mL)

(3)

Here (o) and (t) refer to times zero and t. The calculated change in volume is shown on the right-hand axis in Figure 1. It varies from a negligible volume change for exp. Rb11 6,4 to a change in volume of nearly 4.96% in exp. Hb5 5,8. A volume change of 4.96% exceeds the 3.73% predicted by reactions 1 and 2, and therefore additional processes must be involved. CO2 is about 30 times more soluble in water than O2. When O2 is consumed and CO2 is produced, the partitioning of gases between the water and gas phase must therefore change. Due to the difference in gas solubility the VOL. 35, NO. 20, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4075

FIGURE 1. The evolution in gas-phase composition during pyrite oxidation in unsaturated aquifer sediment incubated in gas impermeable polymer bags. Symbols indicate experimental data, the solid line represents a data fit for O2, and the dotted lines for CO2 and N2 are modeled by PHREEQC. The volume loss is calculated from the N2 increase using eq 3. change in gas-phase volume also depends on the initial gas/ water volume ratio in the sediments (Table 1) which varies between 3.1 and 17.4. The reaction scheme of 1+2 assumes the bicarbonate ion concentration to remain constant. However, the increase in CO2 resulting from reaction 2 may cause additional dissolution of CaCO3 and the formation of HCO3-.

CaCO3 + CO2 + H2O f Ca2+ + 2HCO3-

(4)

The importance of reaction 4 as compared to reaction 2 depends on the relative volumes of the gas and water phases and at a low gas/water ratio bicarbonate formation gains importance. The depletion of O2 and the increase in CO2 could in principle also be due to the oxidation of organic matter in the sediment, as given by the general equation:

CH2O + O2 f CO2 + H2O

(5)

Here organic matter represented in simplified form as a carbohydrate and carbohydrates are by far the most important substrates for organic matter degradation in sediments (23). Again the CO2 produced may interact with CaCO3 through reaction 4 and form bicarbonate. PHREEQC Modeling. The chemical reactions involved in the oxidation of pyrite coupled to calcite dissolution and the associated partitioning between water and gas phases are complex. To quantify the processes occurring in the incuba4076

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 20, 2001

tion bags, the development in gas and water phase composition was calculated with the code PHREEQC 2.0. (24). PHREEQC may operate with separate gas and water phases, and reaction products and reagents are partitioned among these phases. The input data file of PHREEQC specifies the absolute amounts of the water and gas volumes present in the incubation bags. The initial gas-phase composition in the incubation bag was used as starting point for the simulations. To check the stoichiometry of reactions 1 + 2, the gas and water phases were first brought in equilibrium with goethite and calcite. Next, increments of pyrite were added in amounts corresponding to reaction 1 to match the experimental O2 data and are given by the solid lines in Figure 1. The changes in the N2 and CO2 contents of the gas phase were calculated with PHREEQC and are given as broken lines in Figure 1. The PHREEQC results, based on pyrite oxidation in combination with calcite dissolution, compare well with the experimental data. In experiment Rb11 11,1 PHREEQC predicts, however, more CO2 in the gas phase than was experimentally observed. A comparison of the initial pyrite content of the sediment (Table 1) with the amount of pyrite consumed during incubation according to the model, ranges from 10 to 2.6% pyrite consumed in samples Hb5 5,8 and Rb11 11,1, respectively, the samples showing most O2 depletion. In samples Rb13 13,8 and Rb11 6,4 only 4 and 1% of the available pyrite is consumed. PHREEQC also calculates a decrease in gas volume as reflected in the increase of N2.

of inorganic carbon between gas and water phases at a gas/ water volume ratio of 3.1 and 30. Figure 3a shows the depletion of micromoles of O2 from the gas phase versus the accumulation of micromoles of CO2 in the gas phase. A low gas/water volume ratio gives a relatively low increase of CO2 in the gas phase. When the gas/water ratio increases, more micromoles of CO2 end up in the gas phase, and the model runs with a gas/water ratio of 20 and 30 yield a stoichiometric ratio of O2 consumption over CO2 increase in the gas phase that is close to 2 as is predicted by reactions 1 and 2. At a low gas/water ratio a larger part of the produced CO2 remains in solution due to the high solubility of CO2 in water and more bicarbonate formation (reaction 4). In combination with the stoichiometry of reactions 1+2 this causes the total gas volume to decrease and thereby N2 to increase as illustrated in Figure 3b. For a gas/water ratio of 3.1 the volume is reduced by 4.96%, and for a gas/water ratio of 30 the volume is reduced by only 0.37%. By increasing the gas/water ratio, the total gas volume increases, and the number of micromoles of CO2 produced by CaCO3 dissolution contribute less to the total gas volume. Therefore the increase in CO2% is largest for a low gas/water ratio and decreases with higher gas/water ratios (Figure 3c). The CO2 content will through reaction 4 influence the bicarbonate content of the water. For gas/water volume ratios of 20 and 30 the change in CO2 during the reaction is small (Figure 3c), and the calculated alkalinity is near constant (Figure 3d). The overall stoichiometry of reactions 1 and 2 added remains then valid. However for a gas/water ratio of 3.1 the alkalinity increases with 1.90 mmolc/L, and the overall stoichiometry becomes FIGURE 2. PHREEQC modeling results comparing the effect of pyrite oxidation and organic carbon oxidation on the CO2 and the total gas volume. Experiment Hb5 5,8 is used as starting point, and the experimental data are included as filled squares.

2.37CaCO3 + 4H+ f 2.37Ca2+ + 0.74HCO3- + 1.63CO2 + 1.64H2O (6)

PHREEQC was also used to test the assumption that it is oxidation of pyrite rather than oxidation of carbon that takes place in the incubation bags. Instead of pyrite, increments of zerovalent carbon were added to match the experimental O2 depletion, following reaction 5, in otherwise similar model runs. Figure 2 shows the PHREEQC predicted development in gas-phase CO2 and total gas volume together with the experimental data. Reaction 5 predicts for carbon oxidation that the amount of O2 consumed equals the CO2 produced. For pyrite oxidation the CO2 production is roughly half the O2 consumption. Therefore the PHREEQC calculations show a larger increase in CO2 for carbon oxidation than for pyrite oxidation. The modeled line for pyrite oxidation is in good agreement with the experimental data. The total volume, calculated by PHREEQC, decreases for both pyrite and carbon oxidation. For pyrite oxidation the decrease in volume is the combined effect of reactions 1+2, the higher solubility of CO2 compared to O2 and the dissolution of CaCO3 (reaction 4). For carbon oxidation reaction 5 predicts no change in total volume. Here it is the difference in solubility of O2 and CO2 in combination with bicarbonate formation following reaction 4 that causes the volume increase. The effect of the gas/water volume ratio on the gas-phase composition was explored in a series of PHREEQC model runs with pyrite oxidation. Experiment Hb5 5,8 was selected as point of reference for this exercise since the gas/water volume ratio of 3.1 is lower than in most other experiments, and therefore it serves as a good endmember. In the various model runs the volume of gas is varied, while the volume of water is kept constant. The calculations were performed for the gas/water volume ratio of 3.1 in Hb5 5,8 and for hypothetical ratios of 5, 10, 20, and 30. The results are summarized in Figure 3 and Table 2 showing the partitioning

Rate of Pyrite Oxidation. The initial rate of pyrite oxidation in the aquifer sediment can be estimated from the oxygen depletion curve in the incubation experiments (Figure 1). The initial rate of pyrite oxidation is given by the stoichiometry of reaction 1, dmFeS2/dt ) (4/15)‚dmO2/dt. The initial O2 concentration in the different reaction bags was almost constant, and the initial rates are valid for O2 ) 8% ( 0.5%. Figure 4 shows the initial rate of pyrite oxidation against the initial pyrite content of the sediment. Results for a wider range of Beder aquifer sediments are plotted and show a large variation in the rate of pyrite oxidation with values going up to about 0.05 µmol FeS2/h. The rate of pyrite oxidation seems to be almost unrelated to the amount of pyrite present. Therefore the reactivity and specific surface area of the pyrite present in the sediment must have a dominant effect on the reaction rate. Pyrite grains were isolated from the sediment and inspection by scanning electron microscope (SEM) revealed a wide variation of habits of pyrite, ranging from framboids with a high specific surface area to detrital fragments of pyrite crystals with a low specific surface area (25). In addition, SEM showed the pyrite grains to be coated with a variable iron oxide layer, which is consistent with the partial oxidation of pyrite in the field prior to incubation in the polymer bags. The formation of iron oxide is known to have a strongly impeding effect on the rate of pyrite oxidation (11). The variability in pyrite crystal habits in combination with iron oxide coating probably explains the large range in reactivity. Comparison of the rate of pyrite oxidation in the aquifer sand with results of most laboratory studies is difficult because they are mostly reported with respect to the pyrite surface area, which for pyrite in the aquifer sand is hard to determine. However, if the rates of pyrite oxidation measured at pH 6 and 7 in the laboratory on pure pyrite by Moses and Herman VOL. 35, NO. 20, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4077

FIGURE 3. The effect of the gas/water (G/W) ratio on reaction stoichiometry and partitioning between gas and aqueous phase, as modeled by PHREEQC 2.0. Sample Hb5 5,8 with a gas/water ratio of 3.1 is used as starting point, and the experimental data are included as filled squares. For comparison, PHREEQC results for hypothetical gas/water ratios of 5, 10, 20, and 30 are given.

TABLE 2. Calculated Partitioning of Inorganic Carbon between the Gaseous and Aqueous Phases at Gas/Water Volume Ratios of 3.1 and 30 CO2 gas CO2 aq HCO3-

G/W 3.1

G/W 30

40% 12% 48%

78% 2.5% 19.5%

(13) are recalculated to 8 °C, using the temperature correction of (26), the results are in the range 5.5-8.7‚10-12 mol FeS2/ g‚s. Likewise the study of Williamson and Rimstidt (12) yields a value of 6.0‚10-11 mol FeS2/g‚s. Rate estimates for the Beder aquifer are in the range 4.0‚10-12-9.4‚10-10 mol FeS2/g‚s being confidently within the same range. Field Implications. The determination of the rate of pyrite oxidation by using the sediment incubation method may provide a good estimate of the in situ rate of pyrite oxidation. It allows a independent assessment of the effects of gas transport and pyrite oxidation kinetics on the rate of pyrite oxidation in the field. The variation in the rate of pyrite oxidation in the field is high and apparently the complex result of the amount of pyrite present, its mode of occurrence, and its previous exposure to oxidation. Particular caution is therefore warranted when applying rate laws for pyrite oxidation obtained using pure pyrite in the laboratory to field situations. The initial O2 content within our incubation bags was 8%, while the atmosphere contains 20% O2. The changes in gas volume during the oxidation of pyrite in an unsaturated zone 4078

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 20, 2001

FIGURE 4. The initial rate of pyrite oxidation measured in incubation bags versus the initial content of pyrite in the aquifer sediment. in contact with the atmosphere will therefore be 2.5 times higher than the ones we have found in the laboratory. In the absence of CaCO3 the change in volume may amount up to 20% and over 10% when CaCO3 is present. A decrease in total pressure of this magnitude must induce advective transport of air. Methods that interpret the O2 gradients in mine tailings as due to diffusion only (8, 16, 18) will therefore

underestimate the actual flux of O2 going into the sediment. Increases of N2 in the gas phase of mine tailings have in fact been measured, and advective transport of air toward the site of pyrite oxidation is considered to be a significant process (10).

Acknowledgments Special thanks are due Peter Togeskov of Danisco Flexible Inc. for advice concerning the properties of different polymer laminates and for providing these materials as well.

Literature Cited (1) Environmental Geochemistry of Sulfide Oxidation; Alpers, C. N., Blowes, D. W., Eds.; ACS Symposium Series 550; American Chemical Society: Washington, DC, 1994. (2) Evangelou, V. P.; Zhang, Y. L. Critical Rev. Environ. Sci. Technol. 1995, 25, 141-199. (3) Kinniburgh, D. G.; Gale, I. N.; Smedley, P. L.; Darling, W. G.; West, J. M.; Kimblin, R. T.; Parker, A.; Rae, J. E.; Aldous, P. J.; O’Shea, M. J. Appl. Geochem. 1994, 9, 175-195. (4) Larsen, F.; Postma, D. Environ. Sci. Technol. 1997, 31, 25892595. (5) Stollenwerk, K. G. Appl. Geochem. 1994, 9, 353-369. (6) Morin, K. A.; Cherry, J. A. Chem. Geol. 1986, 56, 117-134. (7) Cravotta, C. A. In Environmental Geochemistry of Sulfide Oxidation; Alpers, C. N., Blowes, D. W., Eds.; ACS Symposium Series 550; American Chemical Society: Washington, DC, 1994; pp 345-364. (8) Elberling, B.; Nicholson, R. V.; David, D. J. Nordic Hydrol. 1993, 24, 323-338. (9) Elberling, B.; Larsen, F.; Christensen, S.; Postma, D. Water Resour. Res. 1998, 34, 2855-2862. (10) van Berk, W.; Wisotzky, F. Environ. Geol. 1995, 192-196. (11) Nicholson, R. V.; Gillham, R. W.; Reardon, E. J. Geochim. Cosmochim. Acta 1990, 54, 395-402.

(12) Williamson, M. A.; Rimstidt, J. D. Geochim. Cosmochim. Acta 1994, 58, 5443-5454. (13) Moses, C. O.; Herman, J. S. Geochim. Cosmochim. Acta 1991, 55, 471-482. (14) Nordstrom, D. K.; Southham, G. Rev. Mineral. 1997, 35, 361390. (15) Bond, P. L.; Druschel, G. K.; Banfield, J. F. Appl. Environ. Microbiol. 2000, 66, 4962-4971. (16) Elberling, B.; Nicholson, R. V.; Reardon, E. J.; Tibble, P. Can. Geotech. J. 1994, 31, 375-385. (17) Fennemore, G. G.; Neller, W. C.; Davis, A. Environ. Sci. Technol. 1998, 2680-2687. (18) Elberling, B.; Nicholson, R. V. Water Resour. Res. 1996, 32, 17731784. (19) Larsen, F. Ph.D. Dissertation, Technical University of Denmark, 1996. (20) Zhabina, N. N.; Volkov, I. I. Environ. Biogeochem. Geomicrobiol. Proc. Int. Symp. 1977, 735-745. (21) Postma, D.; Boesen, C.; Kristiansen, H.; Larsen, F. Water Resour. Res. 1991, 27, 2027-2045. (22) Heron, G.; Bjerg, P. L.; Graversen, P.; Ludvigsen, L.; Christensen, T. H. J. Contam. Hydrol. 1998, 29, 301-317. (23) Conrad, R. FEMS Microbiol. Ecol. 1999, 28, 193-202. (24) Parkhurst, D. L.; Appelo, C. A. J. User’s guide to PHREEQC. (Version 2); U.S. Geol. Surv. Water Resour. Inv. 99-4259; 1999. (25) Andersen, M. S. M. Sc. Dissertation, Technical University of Denmark, 1997. (26) Nicholson, R. V.; Gillham, R. W.; Reardon, E. J. Geochim. Cosmochim. Acta 1988, 52, 1077-1085.

Received for review January 31, 2001. Revised manuscript received July 23, 2001. Accepted July 25, 2001. ES0105919

VOL. 35, NO. 20, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4079