C. E. Turick, P. C. McKinsey, M. A. Phifer,
F. C. Sappington, and M. R. Millings
Westinghouse Savannah River Company
Aiken, SC 29808

An acidic/metals/sulfate groundwater contaminant plume emanates from a 12.5-acre sedimentation basin, due to the contaminated runoff the basin receives from an adjacent 8.9-acre coal pile. The most geochemically important metals present in the plume include iron and aluminum. Additionally elevated concentrations of beryllium, cadmium, chromium, copper, mercury, nickel, and zinc are also present. The plume is located within a 50-foot thick water table aquifer, which consists of a series of interbedded sand, silt, and clay layers with saturated hydraulic conductivities ranging from 1E-3 to 1E-7 cm/s. The free surface of the water table ranges from at grade to 15 feet below grade. The soils in the upper portion of the aquifer are generally at the lower end of the hydraulic conductivity range, whereas the soils in the lower portion generally contain more sand and are at the higher end of the hydraulic conductivity range. Higher contaminant concentrations are generally present in the upper portion of the aquifer.

In situ remediation of this groundwater plume through sulfate reduction promoted by organic injection has been evaluated through laboratory microcosm testing. Such remediation involves the oxidation of an organic substrate by sulfate-reducing bacteria (SRB) for energy and growth, the use of sulfate as an electron acceptor resulting in the production of sulfide, and the subsequent in situ precipitation of metal sulfides.

Three carbon sources, soybean oil, sodium lactate and Hydrogen Release Compound (HRC) were evaluated in microcosms for their effectiveness in stimulating SRB activity. Increased population densities of SRB occurred as a result of the addition of soybean oil to non-contaminated sediments and groundwater as well as sediments and groundwater receiving pH amendments. Hydrogen sulfide concentrations also increased in microcosms demonstrating increased SRB densities. Sodium lactate was shown to serve as a carbon and energy source for growth of SRB from D-area sediments, however an inhibitory response was observed at lactate concentrations greater than or equal to 6.3 g/L. Addition of HRC, a slow release lactate-containing polymer, resulted in a decrease in pH and microbial activity in microcosms and was therefore removed from consideration as a carbon source for bioremediation at this site.

Soybean oil and sodium lactate have been selected as the organic substrates for injection during the field demonstration, based upon the results of the laboratory microcosm testing. The soybean oil is intended to provide a long-term, slow release, carbon source for the SRB, and the sodium lactate is intended to provide a short-term, immediately available carbon source.

The Laboratory Bacteria Population and Organic Selection Testing, was designed to answer questions that pertain to the sulfate reduction remediation of the pH/metals/sulfate groundwater plume emanating from the D-Area Coal Pile Runoff Basin (DCPRB). The laboratory work was designed to help direct the subsequent DIW-1 Field Organic Application, which will be a pilot-scale field application of sulfate reduction remediation to a portion of this pH/metals/sulfate groundwater plume. The following are the questions that the Laboratory Bacteria Population and Organic Selection Testing, was designed to answer:

1. Are sulfate reducing bacteria (SRB) present and if so in what concentration in the vicinity of DIW-1?
2. Are SRBs associated with the groundwater in the vicinity of DIW-1? (This addresses the issue of transport of sulfate reducers within the groundwater and potentially impacts the selection of the carbon source (i.e. miscible versus immiscible))
3. Do indications exist that other bacteria populations, necessary to facilitate sulfate reduction, are present in the vicinity of DIW-1?
4. What is the best carbon source (i.e. out of lactate, HRC, and soybean oil) to promote sulfate reduction for an extended time period in the vicinity of DIW-1?
5. Can sulfate reduction occur efficiently under the current pH conditions in the vicinity of DIW-1? Does increasing the pH of the groundwater cause sulfate reduction to occur more efficiently?
6. Is sulfate reduction prematurely terminated, while sufficient carbon source and sulfate are still present? (If so this would be an indication that some micronutrient limitation may exist)

Groundwater samples were collected from monitoring wells DCB-19A, DCB-19B, and DCB-8, near the DCPRB. Soil cores were taken from locations adjacent to the monitoring wells and within the same elevation as the monitor well screens. Groundwater from monitoring wells DCB-19A, DCB-19B, and DCB-8 was collected in 5 liter, sterile, plastic bottles and all headspace in the bottles was eliminated to minimize exposure to oxygen. Groundwater field measurements and soil cores were taken as outlined in the Test Plan (Phifer et al. 2001). Both the groundwater samples and soil cores were refrigerated for 24-48 hours prior to actual use.

4.1 Microbial Population Analysis Methods

Microbial analyses of the groundwater and soil core samples were conducted in order to address the first three questions dealing with microbial populations. Microbial analyses and associated analytical methods are summarized below. Details of the microbial analyses and associated analytical methods were detailed in Phifer et al. 2001.

AODC is a microscopic method for direct bacterial cell counting of bacteria in environmental samples. This test was performed on both the groundwater and the soil cores as previously described (Phifer et al. 2001).

These determinations were conducted with the spread plate method using two types of media, Tryptic Soy Agar (TSA) and R2A. TSA is nutrient rich and conducive for bacteria, which require high nutrient concentrations. R2A is a minimal medium designed for growth of bacteria, which require low nutrient concentrations. Plate counts were conducted under aerobic (AHPC) and anaerobic (AnHPC) conditions. These tests were performed on the groundwater and the soil cores. In order to perform these tests on the soil cores, the bacteria were desorbed from the soil using a phosphate buffer solution. Serial dilutions were performed as necessary with both the groundwater and the phosphate buffer solutions to facilitate quantification of the bacterial counts.

In summary Aerobic Heterotrophic Plate Counts (AHPC) were conducted using the following four media types: TSA pH 7.0, TSA pH 4.0, R2A pH 7.0, and R2A pH 4.0. Anaerobic Heterotrophic Plate Counts (AnHPC) were conducted using the following four media types: TSA pH 7.0, TSA pH 4.0, R2A pH 7.0, and R2A pH 4.0.

SRB were quantified with a 3 tube most probable number sulfate reducing assay (MPN-SRA) using media specifically designed for growth of a variety of SRB. This test was performed on both the groundwater and the soil cores. In order to perform these tests on the soil cores, the bacteria were desorbed from the soil using a phosphate buffer solution. Serial dilutions of the groundwater and phosphate buffer solutions were added to test tubes, in triplicate and incubated anaerobically for 8 weeks. The quantity of SRB was determined with the MPN-SRA based on the number of positive test tubes at each dilution. These values were then calculated statistically to determine the mean and 95% confidence limit of SRB per sample. Microtiter plates were also used for MPN determination in this study in order to compare results with the test tube method.

Sulfate reducing plate counts (SRPC) were also conducted in order to quantify the SRB. Groundwater and soil samples were treated as above (see section 3.1.3) and dilutions were spread onto sterile plates.

In conjunction with anaerobic microcosm testing, the groundwater from monitoring wells DCB-19A, DCB-19B, and DCB-8 and groundwater from DCB-19A, which was brought into equilibrium with limestone, were analyzed for the parameters in Table 1. DCB-19A groundwater, which was contacted with a mixture of 7 kg of limestone and 5 liters groundwater for 24 hours, was considered in equilibrium with limestone. These analyses provided baseline data for comparison to the final results from the anaerobic microcosm testing.

Anaerobic microcosms were prepared using groundwater and corresponding soil cores as outlined in Table 2 to test organic substrate amendments. To ensure anaerobic conditions, samples were handled in an anaerobic glove box. The oxidation-reduction potential (ORP) of the glove box was monitored while samples were processed using a methylene blue indicator. The indicator is a color strip [blue for oxidized (71mV) and white for reduced (-49mV)] that was calibrated with an ORP probe. The color stayed white throughout the inoculation of test tubes, indicating that Eh values were less than or equal to -49mV.

The microcosms are 200-ml airtight glass bottles with a mini-nert top. 100 ml of the appropriate groundwater and 50 grams of corresponding soil were placed in each microcosm. The organic substrate amendments tested included lactate, soybean oil, and Regenisis Hydrogen Release Compound (HRC). The quantities of organic substrate (as outlined in Table 2) were placed in the appropriate microcosm. Controls, which did not contain an organic substrate amendment, were included in the testing. A total of 48 microcosms were prepared and tested.

These microcosms were sampled and analyzed over time for the parameters listed in Table 1 to monitor the enhancement of sulfate reducers. Liquid aliquots from the microcosms were analyzed for metabolic products of the organic substrate amendments that influence SRB growth as well as H₂S, the product of sulfate reduction. Organic breakdown products include acetate, butyrate, propionate, and valerate. Acetate, butyrate, and propionate are potential SRB carbon sources that could arise from vegetable oil and lactate degradation. These tests were performed with liquid aliquots from the microcosms.

Differences between the microcosms with and without pH adjustment were analyzed to determine the efficacy of pH treatment of the subsurface. The pH of the microcosms was also analyzed over time to determine if microbial activity plays a significant role in pH adjustment in these samples.

At the conclusion of the microcosm testing, selected microcosms were sampled and analyzed for the parameters listed in Table 1, based upon microcosm microbial performance and microcosm sample availability.

Additional details concerning the microcosm testing are provided in Phifer et al. 2001.

The affect of various concentrations of lactate on the growth of SRB was examined in order to determine if inhibitory concentrations of lactate might exist in the microcosms. Lactate concentrations ranging from 0 – 2.5% (percent as mls of 60% sodium lactate per 100 mls of solution) were added to a minimal salt solution for SRB. Inocula (1% vol/vol) was from test tubes that were positive for SRB from D-Area groundwater/sediments. Growth was monitored over several weeks and positive growth was determined as production of a black color and precipitate in the media.

General characteristics of the groundwater (Table 3 field parameters) and groundwater and sediment microbial populations were determined during this study. A preliminary, broad assessment was necessary to determine the feasibility and technical approach for bioremediation to occur at the DCPRB. The Table 3 field parameters show that the contaminated wells (DCB-19A and 19B) have higher oxidation-reduction potential (ORP), dissolved oxygen (DO), and conductivity values but lower pH, compared to the uncontaminated background well, DCB-8. The higher ORP and DO values in the contaminated water are indicative of lower microbial activity than compared to the background well. DIW-1 is immediately downgradient of DCB-19A and 19B. On average DIW-1 has a pH and conductivity between that of the background well and DCB-19A and 19B, and an ORP and DO that is lower than that at the other three locations.

The reason that groundwater from DCB-19A and DCB-19B rather than groundwater from DIW-1 was utilized in the laboratory testing was because of the three organics to be tested (HRC, soybean oil, and sodium lactate), one of them (i.e. HRC) could not be injected through DIW-1. The HRC would have had to be applied as columns between the DCB-19 cluster and DIW-1. Therefore testing was conducted with groundwater from DCB-19A and DCB-19B, which represents the worst case condition that any of the organics would be exposed to, so that all three organics were tested under the same conditions. Therefore any organic which performed reasonably well with this worst case groundwater would be expected to perform better in DIW-1, since DIW-1 groundwater conditions are closer to optimum for SRB growth. See Table 3 for a comparison of DCB-19A and DCB-19B groundwater to that of DIW-1 and the background well DCB-8.

Direct microbial counts, utilizing acridine orange as the stain, resulted in fluorescent interference from colloidal particles in the samples. Some colloidal particles fluoresced at similar wavelengths to that of the acridine orange stained bacteria resulting in false positives. Consequently results from these direct counts were not used in this portion of the study.

The aerobic heterotrophic and anaerobic heterotrophic plate counts (AHPC and AnHPC) provided information on the bacterial population as a whole and demonstrated a trend of decreased microbial density/activity as a function of contamination (Figures 1a-1c). Aerobic microbial plate counts from near the uncontaminated control well DCB-8 did not vary significantly between media type or pH (Figure 1a). Anaerobic plate counts were below detectable levels from sediment of DCB-19B and groundwater from both DCB-19A and B (Figures 1b and 1c). Appendix A provides the actual data associated with each of the figures in tabular format.

SRB were detected with the test tube method of the most probable number – sulfate reducing assay (MPN-SRA) technique in groundwater of DCB-8 and sediment of DCB-19A (Figures 1a and 1b). Growth was slower in microtiter plate method of the MPN-SRA and some cross contamination resulted during incubation of the microtiter plates. Additionally SRB counts were significantly lower with sulfate reducing plate counts (SRPC) than that estimated with the MPN-SRA. Because of these results, only the test tube - MPN-SRA method was employed throughout the rest of the study. Use of the microtiter plates and SRPC was discontinued.

For sulfate reduction to proceed in pure culture, anaerobic conditions within a pH range of 5.5-8.0 are optimal. However previous work has demonstrated SRB activity in sediments with pH values as low as 2 (Ehrlich, 1996, Fauque, 1995, and Tuttle et al. 1969). The reason for this is that the SRB are believed to be growing in small "pockets" of elevated pH throughout the subsurface. The elevation of pH in these pockets is likely a result of non-SRB activity. The presence of aerobic heterotrophs provides the potential for establishing anaerobic conditions by utilization of dissolved oxygen during growth. The three sites characterized demonstrated aerobic heterotrophic activity in groundwater, sediment or both at pH 4 and 7. The presence of SRB from these subsurface samples offers potential for sulfate reduction to proceed if stimulated with organic carbon.

As outlined in section 4.2 groundwater from monitoring wells DCB-19A, DCB-19B, and DCB-8 and groundwater from DCB-19A, which was brought into equilibrium with limestone, were analyzed for the parameters in Table 1 in order to provided baseline data for comparison to the final results from the anaerobic microcosm testing. Table 4 provides the results of this baseline analysis.

Notes to Table 4:
ND = None detected

1. This is DCB-19A groundwater, which has been brought into equilibrium with limestone.
2. The ORP was not analyzed in the laboratory; see the Table 3 field results.
3. Data obtained for DCB-19A prior to limestone treatment; limestone treatment was not anticipated to significantly affect these parameters.

Microcosms were constructed, as outline in section 4.2, in order to simulate nutrient additions to the subsurface. Periodic monitoring of the microcosms was accomplished and provided information related to microbial activity in relation to specific treatments. The microcosms were analyzed after 2 and 4 months of incubation for the parameters listed in Table 1 and as modified as shown in Table 5. Colloidal particles from the microcosms were autofluorescent at similar wavelengths to that of the acridine orange stained bacteria and interfered with AODC direct counts by producing false positives and general difficulty in microbial enumeration. Acridine orange is an epifluorescent stain. This problem was overcome by using the epifluorescent stain 4’6-diamidino-2-phenylindole (DAPI) instead of acridine orange. Fluorescent interference was not a problem with DAPI because it fluoresces at a different wavelength than the sediment particles. Aerobic Heterotrophic Plate Counts (AHPC) and Anaerobic Heterotrophic Plate Counts (AnHPC) were not conducted on the microcosms due to the time, materials and space required. To perform these plate counts 6 dilutions, in duplicate, from each of the 48 microcosms would be required, resulting in the need for an inordinate amount of labor and anaerobic incubation space, making it impractical. Sufficient data was obtained with direct counts with DAPI in conjunction with volatile fatty acid (VFA) analyses in lieu of AHPC and AnHPC. Sulfate Reducing Plate Counts (SRPC) were dropped and replaced with the Most Probable Number – Sulfate Reduction Assay (MPN-SRA) because of better microbial growth in the MPN-SRA liquid media than on the SRPC solid media (plates), and the inordinate amount of labor and anaerobic incubation space associated with the plate counts.

Microbial activity was evident by 2 months incubation, but activity levels were low and therefore only direct DAPI microbial counts, MPN-SRBs and pH were monitored after 2 months of incubation. After 4 months of incubation a complete analysis as outlined in Table 5 was conducted.

pH values (Figure 2) increased with the addition of soybean oil and lactate. DCB-19A microcosms treated with limestone demonstrated greater pH values than DCB-19A microcosms without limestone treatment. pH values did not change significantly from month 2 to month 4 of incubation. HRC treatments resulted in decreased pH that also remained unchanged throughout incubation. Due to consistently low pH and minimal bacterial activity in the HRC treated samples, HRC treated microcosms were not analyzed further. Appendix A provides the actual data associated with each of the figures in tabular format.

Direct microscopic enumeration of the microcosms provided general information about the microbial population as a whole (Figure 3). The initial direct count results are not available due to the problems associated with the acridine orange stain initially used as discussed in section 5.1. Overall microbial density increased in soybean oil and lactate amended microcosms after 2 months incubation. Microbial activity was minimal in HRC amendments. By 4 months incubation, a significant increase in microbial density was detected in lactate amended microcosms. In particular both pH treated and untreated DCB-19A samples demonstrated increased microbial density. All of the samples examined throughout the study demonstrated significant variability among samples. One possible explanation for this is microbial growth may occur in pockets in these sediments. Often microbial cells were detected in clusters that were often associated with sediment particles. This high degree of heterogeneity would increase the variability of the data.

Volatile fatty acid (VFA) determinations were conducted after 4 months incubation in order to assess microbial activity in the microcosms (Figure 4). The VFAs included acetate, propionate, formate, isobutyrate, butyrate, isovalerate, valerate, and isocaproate. As carbon sources are broken down, ultimately to CO₂, various breakdown products are indicative of microbial activity. Carbon sources broken down by one portion of the microbial population become available carbon sources for other microbes. VFAs typically indicate the breakdown of more complex carbon sources. In the case of soybean oil, its partial breakdown would provide less complex carbon sources required by SRB. VFAs were at background levels (initial VFA concentrations) for the unamended samples, as expected, since these microcosms did not receive additional carbon. Soybean oil-amended samples had increased VFA concentrations relative to controls, indicating that microbial activity was underway in the microcosms. The uncontaminated DCB-8 samples demonstrated considerably higher VFA concentrations than the other samples treated with soybean oil. VFA production was also evident in lactate amended microcosms with higher acetate concentrations than those of the soybean oil-treated contaminated sediments. While VFAs indicate microbial activity, the build-up of VFA concentration over 1000 mg/l is often indicative of imbalanced conditions indicating that VFA utilizing microbes are inactive or not present. Since SRB use carbon sources such as acetate, proprionate, lactate, etc., the presence or absence of SRB will play a role in the utilization or accumulation of VFAs in the subsurface. VFA concentrations by month 4 did not indicate imbalanced conditions.

SRB were detected in unamended microcosms from DCB-8 and DCB-19A with and without pH adjustment (Figure 5) and the population density of SRB from DCB-8 was significantly greater than that from DCB-19A. This was expected due to the impact of contaminants on DCB-19A. In fact, the initial SRB density from DCB-19A was just above our detection limit. SRB were not detected in unamended microcosms from the most contaminated site, DCB-19-B. In addition, SRB were not detected in any other treatments of DCB-19B microcosms. Relative to initial SRB densities, soybean oil additions resulted in increased SRB population densities after 2 and 4 months incubation in DCB-8 and DCB-19A (pH adjusted) microcosms. Although lactate is considered the carbon source most SRB utilize, no SRB were detected in lactate amended microcosms.

Hydrogen sulfide concentrations increased over background levels with soybean oil amended microcosms from DCB-8 and the pH treated DCB-19A (Figure 6). These data correspond with the amended microcosms with the highest SRB densities (Figure 5) and indicates active SRB populations in these microcosms. Although SRB densities increased and hydrogen sulfide was detected, the SRB activity was too low to decrease, significantly, the sulfate concentrations in the microcosms (Figure 7). Sodium lactate amendments demonstrated increased sulfate concentrations (Figure 7); possibly as a result of desorption phenomena from the high sodium concentrations. The lack of hydrogen sulfide and decreased sulfate concentrations in the lactate-amended microcosms indicate that SRB activity is minimal at best. This is corroborated by the absence of SRB in MPNs from the lactate amendments. The increase in microbial density from the direct microscopic counts and elevated VFAs from lactate amendments along with the absence of SRB indicate that microbial activity in these samples is fermentative.

The low growth rates of SRB were not unexpected due to the presence of potential inhibitory metals in the samples. For instance Cu and Al concentrations in DCB-19A and DCB 19B were much higher than those detected in DCB-8 and pH treated DCB-19A water (Table 4). One of the microcosms produced negative results in the lower dilutions but positive results in higher dilutions. One explanation for such results is the gradual dilution of inhibitory compounds to the point that inhibition is minimized in the higher dilutions. If compounds inhibitory to SRB activity are present in these sediments and/or groundwater, they were not detected in all MPN-SRA studies and therefore are not likely distributed homogeneously throughout the aquifer.

Considering SRB were detected in DCB-8 and DCB-19A microcosms, the lack of SRB in lactate amendments indicates inhibition of SRB activity, due to contaminants or lactate concentrations or both. The degree of lactate utilization was low in all lactate amended, microcosms (Figure 8). Assuming that SRB are inhibited by lactate concentrations, their activity may be restored over time if lactate is degraded down to non-inhibitory levels. In order to assess the role of lactate on SRB activity we conducted the following supplemental experiment.

Although lactate is a common carbon source for SRB and high lactate concentrations have been used in some recent bioremediation efforts (Martin et al., 2001) we examined the potential for sodium lactate to inhibit SRB growth. Sodium lactate concentrations ranging from 0.125 – 0.5% (percent as mls of 60% sodium lactate per 100 mls) demonstrated sulfate reduction within 7 days (Table 6). Sodium lactate concentrations o f 1.0 and 1.5% additionally demonstrated sulfate reduction within 60 days. However no sulfate reduction as detected with sodium lactate concentrations of 2.0 and 2.5% (Table 6). Since the lactate concentrations used in the microcosms were in the higher range, inhibition may have resulted. Whether the inhibition is due to sodium concentration, substrate overload, or a chemical breakdown of lactate to another inhibitor is unknown at present. If SRB were inhibited by sodium lactate, this could explain the increase in acetate concentrations in the lactate amended, microcosms (Figure 4) and the absence of hydrogen sulfide (Figure 6). Appendix A provides the actual data associated with each of the figures in tabular format.

Physical, chemical and biological parameters of the subsurface of D-Area were assessed for bioremediation potential. The goal of bioremediation at D-Area is immobilization of soluble metals by in-situ hydrogen sulfide production. This approach requires an existing population of SRB. In addition, a mixed microbial population of sufficient size is required to support the growth and activity of SRB. Conditions that support SRB activity include anaerobic conditions and a pH to near neutrality. Based on the results of our tests we have determined the following:

1. SRB are present in the vicinity of DIW-1 ranging in population size from 1-30 cells/ml of water or g of sediment.
2. SRB are associated with the groundwater near DIW-1 and indicates microbial transport in the groundwater.
3. The mixed microbial community is capable of both anaerobic and aerobic growth with rich or minimal nutrient media. This indicates a robust population capable of establishing anaerobic conditions in the D-Area subsurface.

Although the microbial population density appears low, the microbes are capable of growth in conditions ranging at least from pH 4-7 at subsurface temperatures. Aerobic bacteria will decrease oxygen concentrations in the subsurface and create anaerobic conditions. The presence of anaerobic heterotrophic bacteria indicates that nutrient breakdown is possible, thus ensuring nutrient requirements of SRB will be met. Since samples near DCB-19A and DCB-19B represent the worst conditions in the subsurface at D-Area and microbial growth was detected there; there is a high probability that microbial growth, including sulfate reduction will occur with the addition of the proper carbon and energy sources.

The potential for microbial growth was examined in this study by the addition of either lactate, HRC or soybean oil to sealed, anaerobic microcosms containing groundwater and sediment from D-Area. Sodium lactate and the lactate containing commercial compound, HRC, were chosen because lactate is regarded as a universal carbon and energy source for SRB. The addition of lactate into the subsurface should therefore elicit a rapid increase in SRB activity and growth due to its high level of solubility in water. Soybean oil and HRC were regarded as a "slow-release" nutrients that could be injected in high volume periodically and gradually dissolve into the groundwater due to its low solubility in water. This study examined the feasibility of using these carbon sources for bioremediation at D-area and the following conclusions were reached:

1. Both lactate and HRC demonstrated inhibitory effects on SRB activity. HRC decreased the pH of the microcosms and appeared to have a negative effect on microbial growth in general. Sodium lactate inhibited SRB activity but was capable of stimulating non-SRB microbial activity. In addition, the pH increased upon addition of sodium lactate. Inhibitory conditions resulting from lactate were concentration dependent with no inhibition evident with less than 1% of a 60% sodium lactate solution. Soybean oil was capable of stimulating the microbial population as a whole including SRB. Hydrogen sulfide was detected in soybean oil amended microcosms. Lactate does serve as a carbon and energy source to D-Area SRB but may have inhibitory effects above 1% concentrations. Soybean oil will serve as an effective carbon and energy source to the mixed population, including SRB.
2. SRB were detected in microcosms under typical subsurface pH conditions. However, pH amended microcosms demonstrated increased SRB density and activity indicating that pH adjustment would increase the rate of SRB growth and hence hydrogen sulfide production. It should be noted that the average pH of groundwater in DIW-1 is between that of the most contaminated groundwater tested (i.e. DCB-19A and DCB-19B) and that of the background well, DCB-8, and it is essentially the same as that of the DCB-19A groundwater in equilibrium with limestone.
3. Overall bacterial growth rates were slow for contaminated microcosms. This may be due to one or more of the following conditions:
1. Low initial bacterial density,
2. Growth inhibition from contaminants, and/or
3. Micronutrient limitations.
4. In general, data from microcosms are useful in developing an understanding of biogeochemical interactions. Since microcosms are closed systems, their results may differ to some degree to open systems encountered in fieldwork. For example water flow in an open system will dilute added nutrients with time. In addition, while closed systems may encounter trace nutrient limitations; open systems are more likely to have a continuous input of trace nutrients. Factors such as these are important to consider when interpreting microcosm data.

Based upon the results presented above the following organic selection recommendations are made:

• The HRC should be dropped from further consideration due to its production of a pH decrease (i.e. it is producing more unfavorable conditions for SRB growth) and its resultant negative impact upon both the anaerobic heterotroph and SRB populations.
• The use of soybean oil should proceed to the field stage due to its generally positive impact upon anaerobic hetertrophs, SRB, and pH.
• The use of sodium lactate should proceed to the field stage due to its generally positive impact upon anaerobic hetertrophs and pH. Although for the most part no change in SRB population was detected with use of the sodium lactate, sodium lactate is a well-known substrate for culturing SRB populations. In fact 5 ml/L 60% sodium lactate syrup (3.15 g/L lactate) was used successfully in our culture media for our most probable number sulfate reducing assay. In addition sodium lactate concentrations (vol/vol) up to 0.5% (as a 60% sodium lactate syrup) demonstrated growth of SRB isolated from D-area. It is believed that the direct use of the higher 60% sodium lactate concentration in the microcosms resulted in overdosing the SRB with sodium lactate, which actually acted as an inhibitor at those high concentrations. Therefore the sodium lactate initially injected in the field will either be diluted to reduce its concentration, injected at a slow enough rate to allow sufficient mixing with the groundwater, or both to achieve a concentration around 3.15 g/L lactate. We know this concentration is not inhibitory to our SRB in D-Area, based on our laboratory results. The concentrations injected may be increased over time as it is demonstrated that the SRB can handle the previous concentrations.
• Both soybean oil and sodium lactate should be injected into the South wing wall of DIW-1 and only soybean oil should be injected into the North wing wall. This will allow determination of the effectiveness of soybean oil use alone, determination of the added benefit of sodium lactate use to that of soybean oil, and estimation of the effectiveness or degree of inhibition of sodium lactate use.
• DCB-8 groundwater, at least in part, should be utilized as dilution water for sodium lactate injection since it has higher naturally occurring SRB populations than groundwater around DIW-1. DCB-8 groundwater use should accelerate the SRB population growth where injected in DIW-1.
• In addition to carbon addition, other nutrient amendments should be considered because the low microbial growth rates detected in microcosms are likely due to lack of nutrients such as phosphate and ammonium.

The following are monitoring recommendations for the field application of the selected organics and additional treatment options that should be considered:

• The pH should be carefully monitored during field application of the selected organics.
• Groundwater from selected wells near DIW-1 and selected DIW-1 monitoring points should be monitored for VFAs, SRB and inorganics such as hydrogen sulfide, phosphate, and ammonium during field application of the selected organics.
• Adjustment of the pH to near neutrality in the field would increase the rate of SRB activity in the subsurface.

1. Ehrlich, H.L. 1996. In. Geomicrobiology. pp.312-367. Marcel Decker Inc. N.Y., N.Y.
2. Fauque, G.D. 1995. Ecology of sulfate reducing bacteria. pp. 217-235. In. Sulfate-reducing bacteria. (ed) L.L.Barton. Plenum Press. N.Y, N.Y.
3. Martin, J.P., K.S. Sorenson, Jr., and L.N. Peterson. 2001. Favoring efficient in situ TCE dechlorination through amendment injection strategy. In. Anaerobic degradation of chlorinated solvents. (eds) V.S. Magar, D.E. Fennell, J.J. Morse, B.C. Alleman and A. Leeson. Battelle Press. Columbus, OH.
4. Phifer, M.A., C.E. Turick, P. McKinsey, F.C. Sappington, and M.R. Millings. 2001. D-Area coal pile runoff basin sulfate reduction remediation treatability study plan. WSRC-RP-2001-00923.
5. Tuttle, J.H., P.R. Dugan and C.I. Randles. 1969. Microbial sulfate reduction and its potential utility as an acid mine water pollution abatement procedure. Applied Microbiology. 17:297-302.

Heterotrophic populations were detected from each site. Sulfate reducing bacteria (SRB) were detected from DCB-8 and DCB-19-A. The most contaminated site, DCB-19B, had the lowest microbial density with no detectable SRB. The abbreviations in the graphs stand for: Aer = aerobic; An = anaerobic; Gw = groundwater; Sed = sediment. R2A is a low-nutrient growth medium and TSA is a nutrient rich growth medium.

The initial pH values are those of the groundwater alone prior to use in the microcosms. The no amendments, pH values are those of the microcosms (i.e. groundwater and soil) without organic amendments. Soybean oil and lactate amended microcosms demonstrated an increase in pH relative to both the initial and unamended samples. Limestone treatment (pH-DCB-19A) also increased the pH. HRC amendments had decreased pH. pH values did not change throughout an additional 2 months incubation.

Overall microbial density increased in soybean oil and lactate amended microcosms after 2 months incubation. Microbial activity was minimal in HRC amendments. By 4 months incubation, a significant increase in microbial density was detected in lactate amended microcosms.

VFAs increased after 4 months incubation, in carbon amended microcosms as a result of carbon source breakdown, relative to those with no additional carbon.

DCB-8 and DCB-19A microcosms demonstrated indigenous SRB populations with increased SRB density by 4 months incubation with soybean oil. The DCB-8 microcosm, treated with soybean oil, had SRB populations above the upper limit of quantification (indicated as (>) on graph). SRB were not detected from any amendments with DCB-19B microcosms or from any microcosms with lactate or HRC.

Hydrogen sulfide production occurred from soybean oil amended DCB-8 and DCB-19A (pH treated) microcosms. Hydrogen sulfide concentrations from other microcosms were not significantly above background groundwater levels (initial). Initial hydrogen sulfide concentrations were not determined for DCB-19A, pH treatments.

Sulfate did not demonstrate significant decreases as a function of amendment additions. Increased sulfate concentrations from lactate amendments may be a result of elevated Na concentration.

Figure 8. Lactate Utilization in Microcosms

Percent lactate utilized after 4 months incubation in microcosms was minimal, however no significant differences were evident between treatments.

Table 3. Direct Microbial Counts