Environ. Sci. Technol. 2000, 34, 2810-2816
Slurry-Phase Biological Treatment of 2,4-Dinitrotoluene and 2,6-Dinitrotoluene: Role of Bioaugmentation and Effects of High Dinitrotoluene Concentrations CHUNLONG ZHANG,† J O S E P H B . H U G H E S , * ,† SHIRLEY F. NISHINO,‡ AND JIM C. SPAIN‡ Department of Environmental Science and Engineering, Rice University, 6100 Main Street, Houston, Texas 77005, and Air Force Research LaboratorysMLQR, Tyndall Air Force Base, Tyndall AFB, Florida 32403
A pilot-scale study was conducted to evaluate the use of aerobic slurry reactors to treat soils that were highly contaminated with 2,4-dinitrotoluene (2,4-DNT) and 2,6dinitrotoluene (2,6-DNT). Contaminated soils were obtained from Volunteer Army Ammunition Plant (VAAP; Chattanooga, TN) and Badger Army Ammunition Plant (BAAP; Baraboo, WI). Concentrations of 2,4-DNT and 2,6-DNT were 19 000 and 1380 mg/kg in VAAP soil and 8900 and 480 mg/kg in BAAP soil. Soils were homogenized and subjected to a soil washing process; the resulting soil slurry was subsequently fed to an Eimco bioreactor (70-L) operated in a draw-and-fill mode. Degradation of either isomer required augmentation with a DNT-mineralizing culture. Stable performance and essentially complete degradation of 2,4DNT (within ∼2 days) was demonstrated for both soils at slurry concentration (sum of aqueous, sorbed, and crystalline phases) exceeding 11 000 µM. Incomplete degradation of 2,6-DNT was observed after inoculation, and low-level degradation activity could not be sustained without repeated bioaugmentation. Changing reactor operation to maintain low slurry-phase concentrations of 2,4DNTsthrough continuous feeding or by reducing the volume of soil slurry fed during draw-and-fillsimproved the ability to sustain 2,6-DNT degradation activity. Complementary studies conducted in shake flasks demonstrated that the high concentrations of 2,4-DNT resulted in an inhibition of 2,6-DNT degradation. The impact of 2,4-DNT on 2,6-DNT degradation required a dualstage approach to achieve complete treatment of both contaminants. Operating two reactors in series, where 2,4DNT was degraded in the first reactor and 2,6-DNT was degraded in the second reactor, allowed for stable drawand-fill operation. High nitrite concentrations resulting from 2,4-DNT degradation in the first reactor had no apparent impact on subsequent 2,6-DNT degradation.
* Corresponding author phone: (713)348-5903; fax: (713)348-5203; e-mail:
[email protected]. † Rice University. ‡ Tyndall Air Force Base. 2810
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 13, 2000
Introduction Dinitrotoluenes (DNT) are intermediates in the manufacture of 2,4,6-trinitrotoluene (TNT) and precursors of toluene diisocyanate used for the manufacture of polyurethane foams. Improper handling and disposal has resulted in extensive DNT contamination of soil and groundwater (1-3). 2,4-DNT and 2,6-DNT are common contaminants at facilities involved in the nitration of toluene. Both 2,4-DNT and 2,6-DNT are listed as priority pollutants; therefore, the remediation of contaminated media at former and current production facilities is required. To provide a better understanding of the fate of DNT in the environment and to investigate the potential for bioremediation processes as a possible treatment option, research has been conducted on the metabolism of DNT under both aerobic and anaerobic conditions. The aryl nitro groups of both DNT isomers are subject to cometabolic reduction under anaerobic conditions (4-7). Reductive metabolism yields contaminant transformation but does not yield mineralization. Under aerobic conditions, 2,4-DNT and 2,6-DNT can serve as sources of carbon and energy for microbial growth. Oxidative metabolism leads to mineralization of the aromatic ring, with the release of the nitro groups as nitrite (8, 9). Aerobic mineralization provides the basis for bioremediation strategies to treat DNT-contaminated soils and groundwater in reactors augmented with DNT-degrading bacteria. In studies using aerobic DNT-degrading mixed cultures, both 2,4-DNT and 2,6-DNT were degraded in an aerobic fluidizedbed biofilm reactor treating contaminated water (10, 11). More recently, a bench-scale study demonstrated the ability to degrade both isomers of DNT in aerobic slurry reactors fed soil from a former TNT manufacturing plant at Hessisch Lichtenau, Germany (12). The present study was conducted to determine the potential for the use of slurry-phase bioremediation processes for the treatment of soils highly contaminated with both DNT isomers at the pilot scale. The study focused on the ability to sustain degradation activity in the presence of potentially inhibitory levels of DNT or nitrite (produced during DNT metabolism). Highly contaminated soils were obtained from two former army ammunition plants and subjected to aerobic biological treatment in pilot-scale slurry reactors. Results demonstrate that inoculation with DNT-degrading cultures was required to initiate rapid degradation of either isomer and that 2,4-DNT inhibited the sustained biodegradation of 2,6-DNT following inoculation. A strategy using a sequential treatment of contaminated soils was designed to overcome the inhibitory effects of 2,4-DNT and to achieve sustained degradation of 2,6-DNT.
Materials and Methods Bacterial Culture. A mixed culture containing the 2,4-DNTdegrading strain Burkholderia cepacia JS872 and the 2,6DNT-degrading strains B. cepacia JS850 and Hydrogenophaga palleronii JS863 was grown in 18-L batches in nitrogen-free minimal medium (13) supplemented with 2,4-DNT (500 µM-1 mM) and 2,6-DNT (250-500 µM) as sole source of carbon, nitrogen, and energy. The culture was incubated in a Biostat C reactor (B. Braun) at 30 °C, with stirring at 400 rpm and sparged with air at 15 L/min. DNT was added daily with no attempt to use aseptic technique. Cells were harvested by filtration on a 0.45-µm Pellicon cassette filter (Millipore), washed once with phosphate buffer (20 mM, pH 7.2), and suspended in 1 L of phosphate buffer before use in experiments. 10.1021/es000878q CCC: $19.00
2000 American Chemical Society Published on Web 06/06/2000
TABLE 1. Characteristics of Test Soilsa soil
moisture (%)
bulk density (g/mL)
pH
VAAP
3.07
1.40
4.60
BAAP
1.62
1.69
9.35
a
2,4-DNT (mg/kg)
2,6-DNT (mg/kg)
18540A 10890B 8940
1380A 870B 480
A and B denote two batches of VAAP soil.
Test Soils and Slurry Preparation. Contaminated soils were collected from the Volunteer Army Ammunition Plant (VAAP; Chattanooga, TN) and the Badger Army Ammunition Plant (BAAP; Baraboo, WI). Soils were air-dried. Then gravel and large debris were removed, followed by repeated sieving and tumbling until all soils passed through a 20-mesh sieve. Soils were then subjected to a homogenization procedure, including manual mixing followed by the use of a sample splitter (Gilson Screen Co., Malinta, OH) to obtain a homogeneous material for use in biodegradation studies. Both soils were heavily contaminated with 2,4-DNT and 2,6DNT. Table 1 presents results of soil characterization following homogenization. Because large amounts of sand interfered with reactor operation, soils were treated by soil washing before use. Soil washing was performed in a 14-L cylinder with an upward jet flow of warm (60 °C) tap water to separate DNT-associated fines from the clean sand. The resulting soil slurry was used in bioreactors, and the sand was discarded. A total of 60 L of water was generally used to wash one batch of soil; however, the ratio of water to soil varied depending on the desired DNT loading for a given feeding cycle. Because the two soils varied in their DNT concentration as well as their sand content, the notation used throughout this paper to describe the amount of material treated in the slurry reactors is described as the nominal solids loading rate, expressed as percent (w/v). The term corresponds to the mass of soil introduced into the soil washing apparatus (kg) for a given feed cycle divided by the reactor volume (70 L). Eimco Slurry Reactors. Two identical 75-L Eimco Biolift slurry reactors (model B75LA, Tekno Associates, Salt Lake City, UT) were used in this study. The reactors were equipped with agitation, aeration, and temperature control. Temperature was maintained at 30 °C, and pH was maintained in the range of 6.75-7.25 using a pH controller and 12.5 N NaOH. The Eimco reactors were identical to previously described systems (14), except that the impeller-mixing device was not used. The slurry reactor uses an airlift, bottom rakes, and diffuser tubes to mix the soil slurry and aerate to the saturated dissolved oxygen level. The diffuser tubes mounted on the rake arms supply the oxygen to the slurry and aid in maintaining a suspension of smaller particles. The rakes and the central airlift continuously resuspended particles of all sizes. Three sampling ports were located along the side wall of the reactor at 2, 15, and 40 cm from the bottom of the reactors. Routine samples were taken from the top sampling port of the reactor side wall. Analytical. To determine DNT concentrations in experimental systems, slurry samples were transferred to a beaker and mixed vigorously with a magnetic stir plate. Subsamples were withdrawn through a large-bore pipet tip for subsequent extraction and HPLC analysis using a Hewlett-Packard series 1050 HPLC equipped with a UV detector as previously described (12). The extraction procedure allowed for the determination of the aqueous-phase concentration, the solidphase concentration (centrifugation of slurry samples, followed by the extraction of solids with acetonitrile), or the overall slurry concentration. The mass of 2,4-DNT in the reactors often exceeded its aqueous solubility, and crystalline DNT was often present at the beginning of the cycles.
Therefore, the results of HPLC analysis are presented as the slurry-phase concentration, which corresponds to the total amount of DNT per liter of slurry. Reactor Operation. Initially, 60 L of soil slurry was added to a reactor along with NaH2PO4 and Na2HPO4 to yield a total phosphate concentration of 20 mM (pH 7). The final volume of slurry in the reactor was brought to 70 L with tap water. After temperature equilibration and pH adjustment, the slurry was inoculated with the mixed bacterial culture. After start-up, reactors were operated in several modes to assess DNT degradation extent and the ability to sustain DNT degradation activity. Initially, reactors were operated in a draw-and-fill mode. In routine draw-and-fill operation, the reactor contents were drained, and 7 L of slurry was reserved for re-inoculation. Soil slurry (60 L) was added at the desired nominal solids loading rate along with phosphate buffer and the inoculum from the previous feed cycle and brought to 70 L total volume with tap water. After degradation was complete, the draw-and-fill cycle was repeated. At high DNT loading rates (nominal solids loading rates of 20% and 30% for VAAP soil and 40% for BAAP soil), inorganic nutrients including CaCl2‚2H2O (10 mg/L), sodium citrate (2.35 mg/L), FeCl3 (2.16 mg/L), and MgSO4‚7H2O (20 mg/L) were added. Following the initial studies, reactor operation was varied to assess impacts on the ability to sustain 2,6-DNT degradation activity, which was problematic during draw-and-fill operation. First, one reactor was operated in a continuousfeed periodic waste mode. Soil slurry was continuously pumped into the reactor at a flow rate of 5-8 mL/min, which was approximately equivalent to the solids loading of drawand-fill mode. When the slurry volume accumulated to 70 L, 15 L of slurry was wasted. Subsequently, the same reactor was operated in an intermittent draw-and-fill mode by increasing the draw-and-fill feeding frequency and decreasing the amount of soil slurry per feed (7.5 L for the first five cycles, and 10 L for the remaining three cycles). Prior to each intermittent feeding, the same amount of slurry was removed from the reactor to maintain the total slurry volume of 70 L. Finally, the reactors were returned to the draw-and-fill mode used initially. In these studies, the reactors were operated as described previously, except that reactors were operated in series. For sequential reactor operation, the first reactor was charged with slurry. After reaching the desired level of 2,4-DNT, its contents was transferred to the second reactor for further treatment. Shake Flask Experiments. Shake flask studies (2-L Erlenmeyer flasks) were conducted to investigate the factors that control the sustainability of 2,6-DNT degradation in slurry systems. A stock of slurry sample containing 45 µM 2,6-DNT and 6 µM 2,4-DNT was obtained from a slurry reactor operated in the sequential mode. From this stock, 500-mL subsamples were added to flasks containing either 2,6-DNT (to reach slurry concentrations of approximately 0, 100, 200, and 500 µM) or 2,4-DNT (to reach slurry concentrations of approximately 0, 50, 500, 1000, 2500, and 5000 µM), which had been coated on the inside of the flask with an acetone solution that was subsequently evaporated. The concentrations selected were within the range of concentrations tested in the slurry reactor. The flasks were then incubated on a rotary shaker (New Brunswick Scientific, Edison, NJ) at a constant temperature of 30 °C. The pH of the slurry was adjusted manually at least twice a day to maintain neutral pH. Similar shake flask tests were also performed to investigate the extent of biodegradation after the extended aeration of slurry samples containing residual 2,4-DNT and 2,6-DNT. In these studies, reactor effluent was placed in 1-L Erlenmeyer flasks, and the slurry-phase concentrations of both isomers were monitored for 10-12 days at 30 °C. VOL. 34, NO. 13, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2811
FIGURE 1. Draw-and-fill operation. Slurry-phase concentrations of 2,4-DNT (b) and 2,6-DNT (O) in a reactor fed VAAP soil slurry at 5% nominal loading rate.
FIGURE 2. Draw-and-fill operation. Slurry-phase concentrations of 2,4-DNT (b) and 2,6-DNT (O) in a reactor fed VAAP soil slurry at 20% and 30% nominal loading rates.
Results Slurry Reactor Start-Up. VAAP and BAAP soil were fed to slurry reactors (10% solids loading rate) to investigate the ability of indigenous soil microorganisms to rapidly degrade 2,4-DNT and 2,6-DNT. Both reactors were monitored for DNT disappearance over a 3-day period. No degradation of either isomer was observed (data not shown). After inoculation with the DNT-degrading mixed culture, rapid disappearance of 2,4-DNT and 2,6-DNT began immediately. Related studies indicate that 2,4-DNT degradation by indigenous bacteria could be expected to begin after 3 weeks. The reactors were inoculated to reduce potentially long acclimation times. Draw-and-Fill Operation. Figure 1 presents the concentrations of 2,4-DNT and 2,6-DNT throughout 13 feeding cycles of continuous operation with VAAP soil at 5% nominal solids loading rate. The degradation of 2,4-DNT proceeded rapidly, and complete disappearance was noted in approximately 45 h after inoculation. Nearly complete degradation of 2,4-DNT was observed during each feeding cycle (approximately 2-day periods); however, the degradation of 2,6-DNT was slow and incomplete (∼40% depending on the feeding cycle) even though the initial 2,6-DNT concentration was approximately 10 times less than that of 2,4-DNT. There were marked differences in the initial DNT concentrations after the sixth feeding cycle because a different batch of VAAP soil was used for subsequent testing. One additional alteration in operation involved the buffer used. The slurry was initially supplemented with 20 mM phosphate. Phosphate concentrations 2812
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 13, 2000
were changed to 1 mM after 435 h of operation. No changes in the rate and extent of degradation were noted, and 1 mM phosphate was used throughout the remaining studies. The VAAP nominal solids loading rates were increased to determine the maximum loading rates achievable (Figure 2). 2,4-DNT was degraded nearly completely in the slurry reactor with a nominal solids loading rate of 20%. When the reactor was fed VAAP soil at a 30% nominal solids loading rate (17 000 µM 2,4-DNT, 1500 µM 2,6-DNT), complete degradation of 2,4-DNT was no longer observed. The loss of activity may have resulted from an unidentified nutritional limitation, or perhaps it was due to nitrite accumulation. When a new batch of slurry was added at 20% loading rate, the system recovered. It is also clear that 2,6-DNT degradation was minimal at the high loading rates (Figure 2). Figure 3 presents the results of draw-and-fill feeding of BAAP soil slurry at five nominal solids loading rates: 5%, 10%, 20%, 30%, and 40%. As observed with VAAP soil, 2,4DNT degradation activity could be sustained (7 weeks) in draw-and-fill operation. Degradation of 2,4-DNT was nearly complete in the slurry reactor with BAAP soil at a nominal solids loading rate of up to 40%. This loading rate corresponded to an initial 2,4-DNT concentration of 11 230 µM. The degradation of 2,6-DNT was negligible, especially at higher loading rates, which is consistent with the results obtained with VAAP soil (Figures 1 and 2). Single Reactor Fed Slurry in Continuous and Intermittent Modes. Reactor performance was evaluated under conditions of continuous feed and then intermittent feed
FIGURE 3. Draw-and-fill operation. Slurry-phase concentrations of 2,4-DNT (b) and 2,6-DNT (O) in a reactor fed BAAP soil slurry at 5, 10, 20, 30, and 40% nominal solids loading rates.
FIGURE 4. Single reactor fed soil slurry in continuous and intermittent modes. Slurry-phase concentrations of 2,4-DNT (b) and 2,6-DNT (O) in a reactor fed BAAP soil slurry. Also shown is the amount of 2,4-DNT (9) and 2,6-DNT (0) fed to the reactor during continuous feed mode. (Figure 4) to evaluate how feeding strategies influence degradation activity. As in the draw-and-fill mode (Figures 1-3), nearly complete degradation of 2,4-DNT was established after 2 days when the BAAP soil slurry was continuously fed to the reactor. The 2,4-DNT accumulated in the reactor to a concentration of 770 µM and then decreased to below the detection limit. A marked contrast to draw-and-fill operation was the degradation of 2,6-DNT during the continuous feed mode. Approximately 90% of the 2,6-DNT was degraded during the continuous feed period. Both 2,4-DNT and 2,6-DNT were degraded during the rapid feeding low volume draw-and-fill operation. The slurry concentrations of 2,6-DNT decreased to as low as 1.8-7.1 µM at the end of intermittent feedings during the last five cycles (Figure 4). The initial DNT concentrations after feeding were significantly less than during draw-and-fill operation. For example, the slurry concentrations of 2,4-DNT were always below 1000 µM. As shown in Figure 4, 2,6-DNT degradation activity did not occur during certain feed cycles. Re-inoculation (t ) 310, 401, and 645 h) with active 2,6DNT-degrading bacteria rapidly restored activity in the reactor. 2,6-DNT Degradation in Shake Flask Studies. A series of shake flask experiments were conducted to assess possible impacts of 2,6-DNT or 2,4-DNT concentration on 2,6-DNT degradation. The effects of 2,6-DNT concentration on its own biodegradation are shown in Figure 5. Complete 2,6-DNT
FIGURE 5. Effects of 2,6-DNT concentration on its own biodegradation in shake flasks. Initial 2,6-DNT concentrations (spiked) were 0 (b), 0.1 (9), 0.2 (2), and 0.5 mM (1). degradation occurred in all cases, including concentrations as high as 650 µM. The 2,4-DNT concentrations in the suspension were low (∼6 µM) throughout. The effects of 2,4-DNT on 2,6-DNT degradation were determined in an experiment with a series of flasks containing VOL. 34, NO. 13, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2813
FIGURE 6. Effects of 2,4-DNT concentration on 2,6-DNT biodegradation in shake flasks. Slurry-phase concentrations of 2,6-DNT (A) and 2,4-DNT (B). Initial 2,4-DNT concentrations (spiked) were 0 (O), 0.05 (9), 0.5 (2), 1 (1), 2.5 (b), and 5 mM ((). the same concentration of 2,6-DNT and various concentrations of 2,4-DNT (Figure 6). At a concentration of 50 µM, 2,4-DNT did not affect the degradation of 2,6-DNT. 2,4-DNT concentrations of 500 and 1000 µM inhibited 2,6-DNT degradation, but after 2,4-DNT was degraded, 2,6-DNT degradation began. At a 2,4-DNT concentration of 2500 µM, inhibition appeared to be irreversible, as 2,6-DNT was not degraded even after 2,4-DNT was completely degraded. The two highest 2,4-DNT concentrations tested exceeded the solubility limit of 2,4-DNT, thus similarly high aqueous-phase concentrations were reached in both systems (Figure 6B). Sequential Reactor Fed Effluent in Draw-and-Fill Mode. In previous draw-and-fill operation (Figures 1-3), 2,6-DNT degradation was either incomplete or absent. Attempts to enhance 2,6-DNT degradation through extended aeration after 2,4-DNT had disappeared (the first six feedings in Figure 3) did little to increase 2,6-DNT degradation and negatively impacted 2,4-DNT degradation in subsequent feed cycles. The extended lag periods are evident upon comparison of the lag phases of the first six feedings with those of the next three feedings (Figure 3). On the basis of the results of shake flask tests, continuous feed studies, and rapid feeding low volume draw-and-fill operation, it was apparent that high 2,4-DNT concentrations were influencing the ability to maintain active 2,6-DNT degradation in the routine draw-and-fill mode. To develop rapid and sustainable 2,6-DNT degradation in slurry reactors, a sequential reactor system was investigated where the first reactor was operated to achieve 2,4-DNT degradation and the second reactor was fed the effluent from the first reactor for the purpose of 2,6-DNT degradation. The second slurry reactor was initially filled with a 50% dilution of the effluent containing residual 2,6-DNT in BAAP soil and inoculated with an active 2,6-DNT-degrading culture. The 2,6-DNT disappeared after an acclimation period of approximately 4 days. When 2,6-DNT degradation was established, the feed volume of 2,6-DNT slurry was gradually increased until undiluted effluent from the first reactor was used. The results (Figure 7A) clearly demonstrated the sustained degradation 2814
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 13, 2000
FIGURE 7. Slurry-phase concentrations of 2,4-DNT (b) and 2,6-DNT (O) in reactors fed (A) BAAP and (B) VAAP slurry from the first reactor. The numbers in the figures denote the volume (L) of slurry remaining from the previous cycle, the volume (L) of reactor effluent taken from the first reactor in series, and the volume (L) of tap water added. of 2,6-DNT in soil from BAAP at a concentration range of 50-275 µM. The reactor was supplemented with additional BLKN nutrients after a prolonged lag phase at t ) 296 h, and the nutrient supplement was added throughout the remaining cycles. For VAAP soil (Figure 7B), reactor operation started with an initial 2,6-DNT concentration of approximately 100 µM. The 2,6-DNT concentrations were increased continuously over eight draw-and-fill cycles by increasing the feed volume of 2,4-DNT-depleted effluent from the first reactor. 2,6-DNT was degraded effectively at a concentration range of 100300 µM. There was a brief lag phase when 2,6-DNT concentration reached 300 µM, but 2,6-DNT degradation subsequently resumed. Residual Concentrations and Partitioning after Extended Aeration. The residual DNT concentrations after treatment using sequential reactor operation varied slightly from one feeding cycle to another. Residual slurry-phase 2,4DNT or 2,6-DNT concentrations were occasionally as high as 25 µM for both soils. It may have been possible to reduce this residual through extended aeration in slurry reactors; however, extended aeration led to long lag phases in subsequent feeding cycles, and in some case re-inoculation was required to restore activity. Therefore, shake flask tests were used to evaluate the minimum residual DNT concentrations achievable after an extended treatment. Extended aeration led to further degradation of both isomers. The aqueous-phase effluent concentrations (BAAP soil) after 9 days of aeration were 1.3 and 0.29 µM for 2,4DNT and 2,6-DNT (Table 2), which are below the EPA treatment standards (40 CFR, Section 268.48) of 1.76 and 3.02 µM for 2,4-DNT and 2,6-DNT, respectively. Extended aeration was also performed with an effluent from the first reactor in series when operated at 30% and 40% nominal solids loading of BAAP soil and 20% nominal solids loading of VAAP soil. 2,4-DNT was degraded during the extended
TABLE 2. Residual Concentration and Partitioning in Aqueous and Solid Phases after Extended Aeration (BAAP Soil) single reactor treatment
concn % total
Kd (L/kg)b
2,4-DNT 2,6-DNT 2,4-DNT 2,6-DNT 2,4-DNT 2,6-DNT
aqueous (µM)
solida (mg/kg)
6.0 149 38 51
36 535 62 49 46 28
dual reactor treatment aqueous (µM) 1.3 0.29 48 53 151 125
solida (mg/kg) 25 4.7 52 47
a Solid phase: extractable with acetonitrile. b K ) C /C , where C d s w s is the solid-phase concentration (mg/kg) and Cw is the aqueous-phase concentration (mg/L).
aeration, but the concentration of 2,6-DNT remained constant, confirming the lack of 2,6-DNT-degrading activity in the first reactor. Table 2 also presents the solid-phase concentration and the partitioning characteristics of DNT in the samples taken at the end of extended aeration. Solid-phase concentrations were significantly reduced after the dual reactor treatment. This is particularly evident for residual 2,6-DNT, i.e., 535 mg/kg for single reactor treatment vs 4.7 mg/kg for dual reactor treatment. The partitioning coefficients (Kd) of residual 2,4-DNT and 2,6-DNT were higher after dual reactor treatment, implying that residual DNT was not readily desorbed at low solid-phase concentrations. The percentage of residual DNT (based on mass) was almost equally distributed between aqueous phase and solid phase for both single and dual reactor configurations.
Discussion Results from this pilot-scale study demonstrate that high concentrations of 2,4-DNT can be rapidly degraded in aerobic slurry reactors. Previous studies have demonstrated the ability to treat lower concentrations of 2,4-DNT present in either groundwater or soils (10-12). In this study, concentrations of 2,4-DNT as high as 9750 µM were reduced to near detection limits in 42 h for VAAP soil slurries (Figure 2). In BAAP soil slurries, 2,4-DNT concentrations were reduced from as high as 10 840 µM to below detection limits in 46 h (Figure 3). It is clear that rapid and sustained degradation of high concentrations of 2,4-DNT can be achieved in slurry reactors. Sustained growth and stoichiometric release of nitrite indicate that the DNT was mineralized rather than transformed. The initial lack of DNT degradation in the reactors suggested an inadequate population of appropriate DNTdegrading bacteria in both contaminated soils. Shake flask studies (data not shown) testing the need for inoculation with both soils demonstrated that 2,4-DNT but not 2,6-DNT began to disappear after several weeks of incubation under conditions simulating slurry reactor operation. The apparent absence of rapid 2,4-DNT degradation activity in these soils prior to inoculation is consistent with the observed persistence of 2,4-DNT at contaminated sites. High concentrations of 2,4-DNT are found in near-surface soils at manufacturing facilities despite the fact that 2,4-DNT can serve as a growth substrate and its degradation rates can be rapid under appropriate conditions. Previous studies have demonstrated that 2,4-DNT degradation can be inhibited by high nitrite concentrations. For example, Nishino et al. (15) reported nitrite toxicity to 2,4DNT-degrading strains at concentrations ranging from 10 000 to 20 000 µM. Assuming a stoichiometric release of nitrite from DNT, toxicity to 2,4-DNT strains would therefore be expected after the degradation of 5000-10 000 µM DNT. In
our studies, rapid and complete degradation of 10 000 µM DNT was achieved for both soils (Figures 2 and 3). The average observed molar ratio of nitrite release to 2,4-DNT degraded in slurry reactors was 1.7:1 (data not shown), and nitrite concentrations exceeded 20 000 µM in some feeding cycles (nitrite carryover occurred from retentate during draw-andfill operation adding to observed levels in any given feed cycle). Nitrite levels in the reactor reached 35 000 µM in the 30% VAAP feeding cycle and may have been the cause of decreased degradation activity. Rapid degradation of 2,4DNT was re-established without inoculation in the subsequent feed cycle. Inoculation with 2,6-DNT-degrading bacteria was necessary but not sufficient to establish the sustainable and rapid degradation of 2,6-DNT. Complete and sustained degradation of both DNT isomers has previously been demonstrated in bench-scale draw-and-fill slurry reactors fed contaminated soils (12). Studies conducted to assess the potential treatment of dissolved-phase DNT contamination have also demonstrated that the degradation of both isomers can be achieved in a single bioreactor (10, 11). Degradation of 2,6-DNT was problematic, however, in a pilot-scale fluidized-bed bioreactor operated at VAAP to treat contaminated groundwater (16). Several factors could have contributed to the lack of rapid and sustained 2,6-DNT degradation during the initial stages of slurry reactor treatment of BAAP and VAAP soils reported in this study. The most important difference between our results and previous slurry reactor studies (12) appears to be the slurry-phase DNT concentration after feeding in a drawand-fill mode. In this study, the slurry concentration of 2,4DNT after feeding ranged from a low of 2000 (5% nominal loading using BAAP soil) to 17 000 µM (30% nominal loading using VAAP soil). The highest concentration of 2,4-DNT fed in the studies reported by Nishino et al. (12) was approximately 1000 µM. Concentrations of 2,6-DNT also differed. In the present study, 2,6-DNT concentrations ranged from approximately 80 to 1500 µM, as compared to a range of approximately 300-500 µM in those reported by Nishino et al. (12). When slurry reactors were fed BAAP soils in a continuous-feed mode, low slurry-phase concentrations of both 2,4-DNT and 2,6-DNT were maintained and 2,6-DNT was degraded. Similarly, decreasing the initial concentrations of both isomers through reduced volume draw-and-fill operation allowed for improved 2,6-DNT degradationsas compared to initial draw-and-fill operationsalthough inoculation was required to sustain operation to multiple feed cycles. On the basis of our results from shake flask study and the requirement for continued re-inoculation of slurry reactors to restore even minimal 2,6-DNT degradation, it appears that 2,4-DNT at high concentrations inhibits 2,6-DNT degradation. The effects on viability and populations of 2,6DNT-degrading bacteria were not determined. Little is known regarding the regulation of 2,6-DNT degradation; therefore, it is impossible to rule out an irreversible inhibition mechanism as the loss of activity. Further studies would be required to determine precisely why 2,4-DNT so strongly influences the ability to express and sustain 2,6-DNT degradation activity. Degradation of 2,6-DNT was slower than 2,4-DNT under conditions when activity could be sustained, including a sequential reactor configuration and dilution approaches. Possible explanations for this include nitrite inhibition and slow growth kinetics of 2,6-DNT degraders. Results from shake flask studies suggest that inhibition from 2,6-DNT on its own degradation would not be expected since the initial concentration of 2,6-DNT (311 µM) was below the inhibitory levels. The potential inhibition of 2,6-DNT by nitrite was not observed in the second reactor where nitrite concentrations VOL. 34, NO. 13, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2815
were ∼35 000 µM. However, previous studies have shown that the growth rates of 2,6-DNT-degrading bacteria were slower than growth rates observed for 2,4-DNT-degrading bacteria (17). Presumably, the slower growth rates coupled with decreasing inoculum volume were the reason for the increase in treatment times in the final feed cycles. Such apparent differences in the bacterial growth rates during sequential treatment can introduce added complexity to the application of slurry reactor treatment. For complete treatment to occur, all material must pass through both reactors. If the residence time for treatment in each reactor is not the same, additional capacity for the slower treatment step must be included in scale-up (18). This study demonstrates the ability to treat soils contaminated with high concentrations of 2,4-DNT and 2,6DNT through a process of soil washing followed by oxidative biodegradation in sequential slurry reactors. It is evident that spatially separating the biodegradation processes of 2,4DNT and 2,6-DNT degradation is required where high concentrations of 2,4-DNT inhibit the degradation of 2,6DNT. The ability to treat highly contaminated materials without extensive dilution allows for smaller and less costly reactor systems. In addition, it reduces the volume of residuals requiring further treatment (i.e., dewatering and nitrite removal). Since the studies were done at a pilot scale, it was possible to conduct a preliminary cost estimate for full-scale treatment using a similar approach to the one outlined by Dupont et al. (19). The estimates suggest that the cost of sequential slurry reactor treatment would range from $100 to $200 per cubic yard. The cost is competitive with thermal treatment, indicating that detailed evaluations of this treatment approach are warranted (18, 20). Successful operation of bioreactors will be facilitated by an initial inoculation with 2,4-DNT- and 2,6-DNT-degrading strains, and the ability to sustain rapid 2,6-DNT degradation will control the efficacy of the processes.
Acknowledgments This work was supported by the Strategic Environmental Research and Development Program (SERDP) through the Applied Research Associates, Inc., Albuquerque, NM, under F08637-9B-C-6002, and by the Defense Special Weapon Agency (DSWA 01-97-0020). We would like to thank the U.S. Army Waterways Experiment Station (WES) for providing the Eimco reactors and Gunter Brox of Tekno Associates for providing the updated Eimco reactor manual.
2816
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 13, 2000
Literature Cited (1) Feltes, J.; Levsen, K. J. High Resolut. Chromatogr. 1989, 12, 613616. (2) Bradley, P. M.; Chapelle, F. H.; Landmeyer, J. E.; Schumacher, J. G. Ground Water 1997, 35, 12-17. (3) Levsen, K.; Muβmann, P.; Berger-Preiβ, E.; Preiβ, A.; Volmer, D.; Wu ¨ nsch, G. Acta Hydrochim. Hydrobiol. 1993, 21, 153-166. (4) Noguera, D. R.; Freedman, D. L. Appl. Environ. Microbiol. 1996, 62, 2257-2263. (5) Noguera, D. R.; Freedman, D. L. Water Environ. Res. 1997, 260268. (6) Cheng, J.; Kanjo, Y.; Suidan, M. T.; Venosa, A. D. Water Res. 1996, 30, 307-314. (7) Hughes, J. B.; Wang, C. Y.; Zhang, C. Environ. Sci. Technol. 1999, 33, 1065-1070. (8) Spanggord, R. L.; Spain, J. C.; Nishino, S. F.; Mortelmans, K. E. Appl. Environ. Microbiol. 1991, 57, 3200-3205. (9) Haigler, B. E.; Nishino, S. F.; Spain, J. C. J. Bacteriol. 1994, 176, 3433-3437. (10) Lendenmann, U.; Spain, J. C.; Smets, B. F. Environ. Sci. Technol. 1998, 32, 82-87. (11) Smet, B. F.; Riefler, R. G.; Lendenmann, U.; Spain, J. C. Biotechnol. Bioeng. 1999, 63, 642-653. (12) Nishino, S. F.; Spain, J. C.; Lenke, H. Knackmuss, H.-J. Environ. Sci. Technol. 1999, 33, 1060-1064. (13) Bruhn, C.; Lenke, H.; Knackmuss, H.-J. Appl. Environ. Microbiol. 1987, 53, 208-210. (14) Lewis, R. F. Air Waste 1993, 43, 503-508. (15) Nishino, S. F.; Spain, J. C.; He, Z. In Biodegradation of Nitroaromatic Compounds and Explosives; Spain, J. C., Knackmuss, H.-J., Hughes, J. B., Eds.; CRC Press: Boca Raton, FL, Vol. 2, in press. (16) Spain, J. C.; Nishino, S. F.; Greene, M. R.; Forbort, J. E.; Nogalski, N. A.; Unterman, R. In Bioremediation of Nitroaromatic and Haloaromatic Compounds; Alleman, B. C., Leeson, A., Eds.; Battelle Press: Columbus, OH, 1999. (17) Bausum, H. T.; Michell, W. R.; Major, M. A. J. Environ. Sci. Health A 1992, 27, 663-695. (18) Zhang, C.; Hughes, J. B.; Nishino, S. F.; Spain, J. C. Biodegradation of 2,4-Dinitrotoluene and 2,6-Dinitrotoluene in a Pilot-Scale Aerobic Slurry Reactor System; Final Report to the U.S. Air Force Research Laboratory on Grant F08637-9B-C-6002; Tyndall AFB, FL, 1999. (19) Dupont, R. R.; Bruell, C. J.; Downey, D. C.; Huling, S. G.; Marley, M. C.; Norris, R. D.; Pivertz, B. In Bioremediation: Design & Application; Anderson, W. C., Ed.; American Academy of Environmental Engineers: Annapolis, MD, 1997. (20) Christoulatos, C.; Koutsospyros, A. In Biological Treatment of Hazardous Wastes; Lewandowski, G. A., DeFilippi, L. J., Eds.; John Wiley & Sons: New York, 1998.
Received for review January 7, 2000. Revised manuscript received April 12, 2000. Accepted April 25, 2000. ES000878Q
