Environ. Sci. Technol. 2003, 37, 2525-2533
Inoculation of a DNAPL Source Zone To Initiate Reductive Dechlorination of PCE DAVID T. ADAMSON, JAMES M. MCDADE, AND JOSEPH B. HUGHES* Department of Civil and Environmental Engineering, Rice University, 6100 Main Street, Houston, Texas 77005-1892
The ability to inoculate a PCE-NAPL source zone with no prior dechlorinating activity was examined using a near fieldscale simulated aquifer. A known mass of PCE was added to establish a source zone, and the groundwater was depleted of oxygen using acetate and lactate prior to culture addition. An active and stable dechlorinating culture was used as an inoculum, and dechlorination activity was observed within 2 weeks following culture transfer. PCE reduction to TCE and cis-DCE was observed initially, and the formation of these compounds was accelerated by the addition of a long-term source of hydrogen (Hydrogen Releasing Compound). cis-DCE was the predominant chlorinated ethene present in the effluent after 225 days of operation, and production of VC and ethene lagged the formation of TCE and cis-DCE. However, dechlorination extent continued to improve over time, and VC eventually became a major product, suggesting that reinoculation was unnecessary. The detection of Dehalococcoides species in the source culture and in the simulated aquifer postinoculation indicated that the metabolic capability to dechlorinate beyond cis-DCE (t ) 86 days and t ) 245 days) was present. Elevated levels of TCE and cis-DCE were present in the source zone, but neither VC nor ethene were detected in the vicinity of NAPL. The results of this research indicated that adding dechlorinating cultures may be useful in the application of source zone bioremediation but that dechlorination beyond cis-DCE may be limited to regions downgradient of the source zone.
Introduction The emergence of anaerobic bioremediation as a technically practical means of treating chlorinated solvent-contaminated aquifers has been fostered by an increased understanding of the metabolic capabilities of dehalorespiring organisms that obtain energy through the dechlorination of these pollutants (1-3). For these organisms, the presence of chlorinated ethenes represents an opportunity for growth, and rapid rates of dechlorination can be achieved if conditions for growth are appropriate (4-10). At many locations, however, chlorinated ethene dechlorination is not observed due to improper aquifer conditions (i.e. oxidation-reduction potential), the absence of dehalorespiring strains, or both, and considerable interest exists in the ability to induce and sustain dehalorespiration to achieve remedial objectives (11-14). * Corresponding author phone: (713)348-5903; fax: (713)348 5203; e-mail:
[email protected]. 10.1021/es020236y CCC: $25.00 Published on Web 04/17/2003
2003 American Chemical Society
Bioaugmentation (defined as the introduction of microorganisms capable of transforming a target compound in a contaminated media) is one approach that has been increasingly investigated in an effort to improve the ability of existing dechlorinating populations present in situ. To date, several bench- and field-scale studies have shown that dechlorination rate and extent can be improved after bioaugmentation into porous media where low levels of dechlorination activity are often present (15-23). Despite these findings, the efficacy of bioaugmentation remains controversial as the activity of introduced strains is difficult to distinguish from those already present and the perception that similar increases in activity could be achieved through conventional stimulation techniques in many cases. In other instances, when no dechlorination activity is present, the addition of dechlorinating mixed cultures represents an inoculation of dehalorespiring strains, and the impact on dechlorination potential is more readily discernible. In either case, concerns exist that the metabolic advantage possessed by the cultures added may not be sufficient to overcome the stresses associated with the introduction to a non-native environment (24, 25). Periodic supplemental additions of cultures have been necessary to maintain activity in some cases. Because dechlorination requires contact between dissolved compounds and microorganisms, anaerobic bioremediation stimulated through culture addition has been traditionally discussed in terms of treating (and preventing further migration of) the dissolved chlorinated solvent plume. In most cases, continued plume development is the result of the slow dissolution of chlorinated ethenes from a nonaqueous phase liquid (NAPL) source. Any treatment scheme designed to reduce the persistence of the contamination source must consider the limiting effect of the slow dissolution from NAPLs (26, 27). Recent studies strongly suggest the potential to biochemically alter the longevity and fate of NAPLs in the subsurface (28-30) by developing rapid dechlorination in close proximity to the NAPL zone. These studies suggest that concerns associated with toxicity of chlorinated ethenes at high concentrations should not rule out the potential for bioremediation in NAPL source areas and that the high aqueous-phase concentrations observed in the vicinity of the NAPL may be exploited by dehalorespiring cultures. If colonization of a source-zone region by these organisms occurs, then the rate of PCE dissolution from the NAPL may be increased. Additionally, metabolic conversion of PCE to trichloroethene (TCE) and cis-dichloroethene (cis-DCE) increases dissolution extent as they are less hydrophobic than PCE (28, 31, 32). Despite the promise of these processes, no detailed evaluations of bioaugmentation or inoculation in source-zones have been conducted to date. The objective of this study was to evaluate the ability to inoculate a nondechlorinating porous medium containing a pure PCE NAPL at a near field-scale under controlled conditions. In this experiment, an active and stable lab-scale dechlorinating culture (9, 28, 29, 33) was used as an inoculum to develop large volumes of culture needed to inoculate a previously verified field-scale simulated aquifer system (34) known as the Experimental Controlled Release System (ECRS). Prior to culture addition, a known mass of pure PCE had been added to the simulated aquifer. Results show that the introduced culture rapidly began to dechlorinate PCE even in regions containing NAPL. Moreover, activity (measured as both rate and extent of dechlorination) improved over the course of the monitoring period, suggesting that the VOL. 37, NO. 11, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2525
culture was stable and that the need for reaugmentation in these conditions was minimal.
Materials and Methods Chemicals. The following chemicals were obtained in liquid form: tetrachloroethene (99+%, Acros), sodium-DL-lactate (60% solution in water, Acros), sodium hydroxide (1-N, Fisher Scientific), Hydrogen Releasing Compound (glycerol polylactate, Regenesis). The following chemicals were obtained in crystal form: sodium acetate (Fisher-Scientific) and sodium bicarbonate (Fisher Scientific). Analytical. Chlorinated ethenes concentrations in experimental samples were determined using headspace analysis. Samples (100 µL) were injected directly into a gas chromatograph (GC) (Hewlett-Packard 5890) equipped with a flame ionization detector (FID) and a packed column (6 ft × 1/8 in. OD) containing 60/80 Carbopack B/1% SP1000 (Supelco). The operating parameters of the GC have been previously described by Carr and Hughes (9). Standards were prepared by adding PCE, TCE, and cis-DCE dissolved in methanol, and VC, ethene, and methane gases, all at known volumes, to a serum bottle (70 mL) containing deionized water (50 mL). Volatile fatty acids (acetate and propionate) were analyzed by first filtering aqueous samples (2.7 mL) through a syringe filter (0.2 µm) into a screw-cap vial (10 mL). To this sample 0.3 M oxalic acid (0.3 mL) was added to yield a final concentration of 0.03 M oxalic acid. If not analyzed immediately, the sample was stored at 4 °C until analysis. Samples were analyzed using a GC (Hewlett-Packard 5890) equipped with a flame ionization detector (FID) that contained a glass packed column (2 m × 2 mm i.d.) containing 80/120 Carbopack B-DA*/4% Carbowax 20 M (Supelco). Using a syringe, a liquid sample (1 µL) was injected directly into the column. The operating parameters for the GC were as follows: oven temperature was 175 °C, detector temperature was 200 °C, and the injector temperature was 200 °C. The flow rate for N2 (carrier gas) was 24 mL/min; air and H2 were used as detector makeup gases. Chemical oxygen demand (COD) was measured using the closed reflux colorimetric method in Standard Methods for the Examination of Water and Wastewater (35). Because lactate was supplied to the system in the form of a polymer for the majority of the experiment, lactate was not measured directly. The input from HRC addition was defined in terms of total combined lactate, which included free aqueous lactate and polymerized lactate that had not undergone hydrolysis. Concentrations of total combined lactate were calculated by subtracting the measured acetate and propionate concentrations from the COD results (mg/L). The conversions for milligrams of acetate and propionate to milligrams of COD were as follows: 1.09 mg COD/mg acetate and 1.53 mg COD/ mg propionate. Dissolved oxygen (DO, dissolved oxygen probe YSI) and pH (Denver Instruments) were measured using aqueous samples. Culture Development. A bioreactor was constructed to collect and maintain a large volume of a rapidly dechlorinating, methanogenic mixed culture that could be used for inoculation. The vessel was a high-density polyethylene tank (112.5 L) equipped with ports for injection of nutrients, pH control, PCE addition, liquid sampling, recycling the contents of the bioreactor, and headspace analysis. The vessel was filled approximately 2/3 full with tap water (75 L) and then connected to the effluent of a laboratory column previously enriched for rapid and complete PCE dechlorination that has sustained activity for over 9 years (33). The remaining 1/3 of the bioreactor (37.5 L) was allowed to fill with the effluent from this column over a period of several days. Subsequently, the system was operated as a fed-batch reactor for 210 days with routine additions of PCE and lactate. 2526
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 11, 2003
FIGURE 1. Plan view of ECRS tank and injection points. Chlorinated ethene distributions were determined by headspace analysis using a gas sampling port. Electron donor was added when COD analysis demonstrated that previous additions had been consumed. The bioreactor was also monitored for pH, and sodium hydroxide (1 N) was added as needed for pH adjustment. After 210 days of operation, the biomass concentration was measured as 12 mg of volatile suspended solids per liter. Inoculation Study. To evaluate the efficacy of inoculation, a large-scale aquifer simulator previously described by Reeves et al. (34) was obtained. The system (depicted in Figure 1) is comprised of a metal tank (5.49 m long × 2.13 m wide × 1.83 m high) open to the atmosphere and packed with sand obtained from a quarry in New Caney, TX. This material was characterized as silty sand, light brown with a permeability of 1.01 × 10-1 cm/s, a pH of 6.15, organic carbon < 0.03%, a grain density of 2.66 g/cm3, and a porosity of 0.35 following packing. The tank was fitted with both influent and effluent lines, and a gravel layer (0.43 m wide) was placed at each end of the tank to improve vertical and horizontal distribution of flow in and out of the system. Water flow was controlled using an influent pump (Milton Roy LMI pump C131-27, Pump and Controls, Inc.) at a flow rate of 24 L/h, corresponding to a velocity of 0.5 m/day and a travel time of 9 days from the influent line to the effluent line. The source water for the ECRS was from the Rice University potable water supply, consisting of a mix of surface water and groundwater. Analysis of dissolved ionic compounds is routinely carried out as part of overall water quality assessments of the supply. These reports indicated low levels of sulfate (44 mg/L), nitrate (0.38 mg/L), and total iron (0.03 mg/L) were typically present in the influent. Multiple internal sampling or injection points were installed using 1/4 inch stainless steel tubing installed during the packing of the system. Initially, the flow of water in the ECRS system was carried out in recycle. Activated carbon canisters (CANSORB XP polyethylene drums containing TIGG 5D virgin liquidphase activated carbon with total surface area of 1050 m2/g) were installed in the effluent line to remove any chlorinated solvents present before reintroduction into the influent of the tank. Later in the operation of the system, once-through flow was used, and the activated carbon units treated the waste flow prior to discharge. ECRS operation was conducted for 317 days. To initiate the study (day 0), neat PCE (1 L total) was added in equal volumes (333 mL) through three sample lines approximately 3 m down gradient from the inlet of the ECRS tank (see Figure 1). PCE delivery was accomplished via syringes under minimal pressure. To deplete dissolved oxygen in the system, sodium acetate (197 g) was introduced to the ECRS influent on day 16. A solution of acetate was mixed in an external reservoir (220 L) filled with tap water and pumped into the ECRS tank
over the course of 12 h. Recycle in the ECRS tank and the lack of significant sorption of acetate by the activated carbon allowed for contacting and mixing of acetate throughout the system (bulk concentration of 0.7 mM). To further depress DO and to provide a fermentable substrate, lactate was added to the ECRS tank. Sodium lactate (1.5 L, 60% solution in water) was injected using syringes to 14 sample lines (100 mL/line) both upstream and near the PCE injection location (see Figure 1). This level of lactate addition would achieve a concentration of 3.0 mM if completely mixed and without degradation. Each line was flushed with deionized water (100 mL) following lactate addition. Immediately prior to inoculation of the ECRS, carboys (18.75 L) were purged with N2 to evacuate oxygen and then filled with culture from the larger bioreactor under a N2 sparge to provide positive pressure and maintain anaerobic conditions during the transfer. The contents from these smaller carboys were transferred under a nitrogen purge into the ECRS through sample lines on days 32 and 33. Additions (18.75 L each for a total of 112.5 L) were carried out using gravity flow (300-700 mL/minute) in three lines in the same location of PCE addition and three sample lines approximately 1.22 m upgradient of PCE addition (see Figure 1). At the time of transfer, the biomass concentration in the bioreactor was 12 mg/L, resulting in an input of 1350 mg of biomass (measured as volatile suspended solids) to the ECRS tank. Following inoculation, lactate was added to the ECRS tank as an electron donor for dechlorination every 4 days using the same procedure described for the depletion of residual oxygen. For these lactate additions, sodium lactate was diluted (1:1 v:v) with deionized water and injected (83 mL/line) through sample lines (lines 3, 9, and 17) in the inoculation zone (see Figure 1). To provide higher and continuous levels of electron donor, HRC (78.8 L total) was injected into the ECRS tank on day 64 using a direct push geoprobe method. Seven locations upstream of PCE addition were chosen for HRC addition (11.25 L each). The first four points (see Figure 1) of HRC addition were perpendicular to flow approximately 1.3 m upstream of PCE addition. The next three were placed just downgradient (0.46 m) of the first HRC injection areas. Prior to injection, HRC was heated using a water bath (60°F-80 °F) to produce a liquid that could be pumped easily. A pH decrease from near neutral to almost 5.0 was observed after the addition of HRC due to the hydrolysis of lactic acid esters, their fermentation, and the limited ability of the activated carbon to adsorb organic acids. To address this, the system was changed to a once-through mode on day 93. Subsequent operation used tap water buffered with sodium bicarbonate (3 kg per 220 L) as the influent. An immediate rise in pH was observed, and the effluent pH eventually stabilized at approximately 6.3 by day 130. The mean pH for the final 187 days of monitoring was 6.31 ((0.22). Monitoring of the ECRS was carried out daily for the majority of the duration of the experiment. Effluent concentrations of chlorinated ethenes, methane, dissolved oxygen, volatile fatty acids, and COD were analyzed, along with temperature, pH, and flowrate. Periodically, samples were taken from interior lines to measure many of these same parameters at various locations in the ECRS tank. Microcosms. To evaluate dechlorination rate and extent, microcosm studies were carried out during several phases of this study. In all cases, microcosms used bottles (70 mL) filled with an aqueous sample (50 mL unless otherwise stated) sealed with Teflon coated septa and aluminum crimps. The incubation of microcosms was carried out at 22 °C (unless otherwise stated) and chlorinated ethenes were analyzed using headspace analysis. Initial microcosm studies focused on the ability to enrich dechlorination activity using aquifer sand obtained from the
ECRS system. All tools, microcosm bottles (70 mL), septa, aluminum crimps, were autoclaved (121 °C) for 1 h. After cooling, bottles were amended with ECRS tank sand (50 g), sealed, and purged with ultrapure nitrogen, and a solution (20 mL) of filter-sterilized deionized water containing lactate (1.1 mM) and PCE (6 µM) was added via a syringe. Duplicate sand microcosms were sacrificed every 14 days for analysis over a period of 17 weeks. To assay dechlorination in the bioreactor used as a source culture for inoculation, microcosm studies were conducted periodically using a 48-h incubation period. Sealed bottles were sparged with ultrapure nitrogen gas to create anaerobic conditions and an aliquot (50 mL) from the bioreactor was added. The microcosms were sparged with H2/CO2 gas mix (80%/20% mix, respectively) to remove background chlorinated ethenes and to ensure an adequate supply of electron donor. Bottles were then spiked with PCE ((14-145 µM) initial concentration) and maintained on a stir plate at room temperature. Similarly, microcosm studies were conducted using samples obtained from the ECRS effluent line. These microcosms were identical in construction to those used to assay the lab culture. For these microcosms, ECRS effluent was collected and transferred anaerobically to microcosm bottles. Microcosms were spiked with concentrations of PCE to coincide with that measured directly from the ECRS effluent (24-96 µM). Several sets of microcosms were conducted to evaluate how electron donor availability and temperature may have influenced dechlorination activity. The evaluation of the influence of electron donor availability consisted of microcosms amended with lactate (2.59 and 5.03 mM) or H2/CO2 (with unamended microcosms serving as controls). To evaluate the influence of temperature, separate microcosms were incubated at 12, 22, and 35 °C. ECRS Community DNA Analysis. Denaturing gradient gel electrophoresis (DGGE) analysis and the Gene-Trac Dehalococcoides assay were conducted to provide an assessment of microbial community characteristics of the source culture, bioreactor culture, and the ECRS after inoculation. These tests were performed by the SiREM Laboratory (Guelph, Ontario) following procedures described elsewhere (36, 37). Aliquots (liquid, groundwater, or soil samples) were collected and transferred to appropriately sized plastic sampling containers. Samples were stored at 4 °C prior to shipment. The Gene-Trac assay has been designed to detect for the presence of organisms belonging to the Dehalococcoides group, members of which are the only identified organisms capable of dechlorinating PCE or TCE to ethene (8, 37, 38). It employs polymerase chain reaction (PCR) analysis using primer sets specific to DNA sequences in the 16S rRNA gene of this group (detection limit of 200300 organisms per liter of water). DGGE, another PCR-based test, was conducted using universal bacterial primers and a 8.0% 0.5X TAE acrylamide gel with a 20-55% gradient of urea/formamide at a constant 60 °C. Band separation and sequencing provided an indication of microbial diversity (39).
Results Sand Microcosm Results. To determine the biological activity of indigenous bacteria in the ECRS sand, microcosms were constructed using ECRS tank sand spiked with PCE and lactate. Headspace samples analyzed by GC determined that no dechlorination activity was stimulated in the microcosm bottles after 17 weeks of incubation in anaerobic conditions. Bioreactor Microcosm Results. Microcosms were performed to give a semiquantitative analysis of reductive dechlorination activity within the bioreactor. Complete reduction of PCE (14-145 µM) to less chlorinated ethenes was typically observed within 24 h of incubation. Prior to day 61 of bioreactor operation, PCE was reduced to TCE and VOL. 37, NO. 11, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2527
FIGURE 3. Chlorinated ethene concentrations in ECRS effluent expressed as mole fractions. Vertical dashed line at day 32 represents inoculation of ECRS. Vertical dashed line at day 64 represents addition of HRC. Vertical dashed line at day 93 represents switch from recycle mode to once-through mode of ECRS operation.
FIGURE 4. Methane concentration in ECRS effluent. Vertical dashed line at day 32 represents inoculation of ECRS. Vertical dashed line at day 64 represents addition of HRC. Vertical dashed line at day 93 represents switch from recycle mode to once-through mode of ECRS operation.
FIGURE 2. Chlorinated ethene concentrations (µmol/L) in ECRS effluent: (a) PCE, (b) TCE and cis-DCE, and (c) VC and ethene. Vertical dashed line at day 32 represents inoculation of ECRS. Vertical dashed line at day 64 represents addition of HRC. Vertical dashed line at day 93 represents switch from recycle mode to once-through mode of ECRS operation. cis-DCE after 48 h of incubation. Subsequent microcosms conducted through day 210 of bioreactor operation demonstrated more complete dechlorination, and the major transformation products after 48 h were VC (>85%) and ethene (2-15%). ECRS Monitoring. The results of PCE, TCE, and cis-DCE concentrations in effluent samples are displayed in Figure 2a,b, with concentrations expressed as µmol/L. Prior to culture addition, PCE concentration stabilized at 59.9 (( 20.3) µmol/L, and no dechlorination products were observed. After inoculation and before HRC addition, TCE was observed as the sole dechlorination product in effluent samples. Following HRC addition, TCE concentrations increased significantly (p e 1.0 × 10-10), and cis-DCE began to appear in the effluent. PCE concentrations were reasonably stable (56.5 ( 10.7 µmol/ L) until HRC was added and then decreased significantly (p e 1.9 × 10-10) to a mean concentration of 27.1 (( 13.8) µmol/ L. After the system was switched to a once-through operation, another statistically significant (p e 2.6 × 10-5) decrease in PCE concentration was observed (mean concentration of 11.6 ((11.5) µmol/L for day 95 through day 317). A gradual decrease in PCE concentration was noted throughout the remainder of the monitoring period, with levels never exceeding 10 µmol/L after day 160. 2528
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 11, 2003
At day 80 a decrease in the cis-DCE concentration was observed, which coincided with a decrease in the pH to 5. Once the system was taken off recycle mode and the pH began to rise, the cis-DCE concentrations also increased (Figure 2b), reaching a peak of 40 µmol/L on day 182. Following this point, cDCE concentrations gradually declined while VC began to increase. The first detection of substantial concentrations of VC (0.39 µmol/L) in the ECRS effluent was on day 192 and continued to be present in the effluent for the remainder of the experiment with a peak concentration (11.6 µmol/L) observed on day 266 (Figure 2c). Ethene was first observed when the system was taken off recycle mode and pH increased in the ECRS tank, with concentrations increasing to a peak of ∼1 µmol/L. This increase was transient, although concentrations remained between 0.1 and 0.4 µmol/L once VC began to be detected in the effluent. Over the course of the entire monitoring period, 78% of the added PCE was accounted for in the effluent, as determined by the summation of daily molar concentrations of PCE and metabolites for the given pumping rate. PCE (prior to and following inoculation) contributed to 44% of this total, with the remainder consisting of TCE (22%), cis-DCE (29%), VC (5%), and ethene (
