- Email: [email protected]
Using polymer mats to biodegrade atrazine in groundwater: laboratory column experiments
Journal of Contaminant Hydrology 54 (2002) 195 – 213 www.elsevier.com/locate/jconhyd
Using polymer mats to biodegrade atrazine in groundwater: laboratory column experiments B.M. Patterson a,*, P.D. Franzmann a, G.B. Davis a, J. Elbers b, L.R. Zappia a a
CSIRO Land and Water, Private Bag No. 5, Wembley, WA 6913, Australia b Paderborn University, Germany
Received 9 August 2000; received in revised form 11 July 2001; accepted 19 July 2001
Abstract Large-scale column experiments were undertaken to evaluate the potential of in situ polymer mats to deliver oxygen into groundwater to induce biodegradation of the pesticides atrazine, terbutryn and fenamiphos contaminating groundwater in Perth, Western Australia. The polymer mats, composed of woven silicone (dimethylsiloxane) tubes and purged with air, were installed in 2-m-long flowthrough soil columns. The polymer mats proved efficient in delivering dissolved oxygen to anaerobic groundwater. Dissolved oxygen concentrations increased from < 0.2 mg l 1 to approximately 4 mg l 1. Degradation rates of atrazine in oxygenated groundwater were relatively high with a zero-order rate of 240 – 380 mg l 1 or a first-order half-life of 0.35 days. Amendment with an additional carbon source showed no significant improvement in biodegradation rates, suggesting that organic carbon was not limiting biodegradation. Atrazine degradation rates estimated in the column experiments were similar to rates determined in laboratory culture experiments, using pure cultures of atrazinemineralising bacteria. No significant degradation of terbutryn or fenamiphos was observed under the experimental conditions within the time frames of the study. Results from these experiments indicate that remediation of atrazine in a contaminated aquifer may be achievable by delivery of oxygen using an in situ polymer mat system. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Atrazine; Biodegradation; Polymer; Oxygen delivery
*
Corresponding author. Tel.: +61-8-9333-6276; fax: +61-8-9333-6211. E-mail address: [email protected] (B.M. Patterson).
0169-7722/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 7 7 2 2 ( 0 1 ) 0 0 1 7 8 - 4
196
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
1. Introduction Contamination of anoxic groundwater by atrazine (2-chloro-4-ethylamino-6-isopropylamino-1,3,5-triazine), terbutryn (2-tert-butylamino-4-ethylamino-6-methylthio-1,3,5triazine) and fenamiphos [1-(methylethyl)-ethyl-3-methyl-4-(methylthio) phenylphosphoramidate] from infiltration of pesticide-laden washdown water has been reported in Dianella, a suburb in Perth, Western Australia (Appleyard, 1995). Such contamination is of great concern in Perth because 35% of domestic drinking water supply is obtained from groundwater, and a similar quantity is used directly on gardens. During initial ‘‘pump-and-treat’’ remediation over 4 months, atrazine concentrations increased from approximately 800 to 2000 mg l 1 after pumping commenced (possibly as a result of plume breakthrough) and then decreased to approximately 600 mg l 1 when pumping was halted. Fenamiphos concentrations remained relatively constant before and during pumping at approximately 50 mg l 1; however, concentrations increased to approximately 100 mg l 1 once pumping was halted (Appleyard, 1995). The ‘‘pumpand-treat’’ remediation suggested that short-term strategies might not be suitable for this site. Both atrazine (Nair and Schnoor, 1992; Blume et al., 1993; Blumhorst and Weber, 1994) and fenamiphos (Ou et al., 1994; Kookana et al., 1997; Patterson et al., 1998) have been previously reported to undergo biodegradation. However, this research has focussed on the fate of these compounds in contaminated shallow soil profiles rather than groundwater. Rates of degradation of these compounds vary significantly and seem to be dependent on individual site conditions, but generally decrease with depth through the vadose zone, largely due to lower total organic carbon levels at depth (Franzmann et al., 1998). There have been limited reports of atrazine and/or fenamiphos remediation in groundwater, while some studies have reported atrazine to be recalcitrant in aquifer material (Agertved et al., 1992). Herrling et al. (1994) reported bioremediation of a groundwater plume of triazine pesticides (including atrazine) using an in-well aeration technology. Mirgain et al. (1995) observed in laboratory experiments that there was a threshold concentration of 200 mg l 1 below which atrazine degradation could not be induced. Bacteria capable of complete mineralisation of atrazine have recently been isolated from the Dianella site (Tilbury, 1998). The bacteria were found to degrade atrazine through an enzymatic mediated hydrolytic dechlorination pathway to hydroxyatrazine, with further degradation to CO2 and H2O. Microcosm experiments by Franzmann et al. (2000) indicated that the half-life of atrazine mineralisation in unamended anoxic aquifer material from the Dianella site was greater than 20 years. However, amendment of microcosms with an atrazine-degrading culture (collected from the Dianella site), and a source of oxygen and carbon, decreased the half-life of atrazine mineralisation to approximately 20 days. In the microcosms that received considerable biomass of atrazine-degrading aerobic cells ( f 105 cells g 1), some atrazine mineralisation still occurred in the absence of oxygen, which is consistent with the findings that the enzymatic steps in the mineralisation pathway are hydrolytic and do not require oxygen (De Souza et al., 1998). Since the organic carbon content of the aquifer material from the Dianella site has been measured at
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
197
0.7 g kg 1, the addition of a carbon source may not be required. These microcosm experiments indicated that oxygen and atrazine-degrading bacteria are an essential requirement for atrazine remediation at the Dianella site. Oxygen delivery for remediation may normally involve air injection (air sparging), addition of peroxide, or addition of aerated water (Davis et al., 1998). These oxygen delivery techniques have been used successfully to remediate petroleum-contaminated groundwater (Johnston et al., 1998). Oxygen delivery has also been achieved using a more passive oxygen diffusion system made of sealed silicone (dimethylsiloxane) polymer tubing installed in a borehole and pressurised with oxygen (Gibson et al., 1998). Yamagiwa et al. (1998) used a similar oxygen supply method in a bioreactor for simultaneous organic carbon removal and nitrification. Adaptation of this polymer tubing technique may well be suited as a longer term, semipassive strategy to remediate contaminated groundwaters, using natural groundwater flow to deliver contaminated groundwater to the polymer mats engineered as an in situ biologically active barrier. The advantage of using diffusion of oxygen across a membrane to deliver oxygen to the subsurface is that the location of delivery of oxygen may be more reproducible and less dynamic and disturbing than other oxygen delivery techniques (e.g. air sparging). Delivery in this way is thought to enhance the formation of a more stable atrazine-degrading biofilm on the surface of the polymer mat (Yamagiwa et al., 1998), since it should be able to achieve simultaneous and more consistent delivery of oxygen from polymer membrane/biofilm interface to the aqueous/biofilm interface and contaminant from the aqueous/biofilm interface to the polymer/biofilm interface. This biologically active polymer-mat barrier concept was tested in laboratory largescale (150 mm diameter, 2.0 m long) column experiments to investigate the potential for biodegradation of atrazine, terbutryn and fenamiphos in groundwater for a variety of groundwater conditions. The experiments were conducted under anaerobic and aerobic groundwater conditions, and with and without an additional carbon source. In particular, oxygen delivery to anoxic groundwater was tested using air purging of silicone polymer tubing woven to form a polymer mat, relying on sustained oxygen diffusion through the polymer tubing wall. Large-scale flow-through column experiments were used in preference to microcosm experiments, since column experiments are more representative of field conditions. By using large-scale columns, similar hydrogeological conditions, such as flow velocity, and soil porosity can be set up, side-wall effects can be reduced, and amendments and remediation options can also be better tested.
2. Materials and methods 2.1. Experimental set up and operation Three aluminium columns (2.0 m long, 150 mm diameter) were constructed with fine stainless steel mesh placed at both ends, 10 cm from the top and bottom of the column to prevent soil deposition in the feed and effluent lines. A polymer ‘mat’ was installed 46 cm from the base of each column. Seven water sampling ports were also positioned along the
198
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
length of each column, at 4 (port A), 37 (port B), 57 (port C), 77 (port D), 107 (port E), 137 (port F) and 182 cm (port G) from the base of the column. The first port was set below the stainless steel mesh (below the soil) and was used to monitor influent aqueous concentrations to the column. A set of four solution delivery ports was positioned around the circumference of the column at a distance of 48 cm from the base of the column. These were used to inject a bromide tracer to estimate soil porosity. A schematic of a column is shown in Fig. 1.
Fig. 1. Schematic of the flow-through column.
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
199
The oxygen delivery mats consisted of a 5.0-m length of silicone tubing (2.0 mm i.d., 3.0 mm o.d.) with a fine stainless steel spring inserted into the centre of the polymer tubing to provide support and eliminate the possibility of the tubing twisting or collapsing. The polymer tubing was then woven through a stainless steel support frame consisting of two 14-cm diameter mesh disks. The two disks were 3 cm apart to form a bilayered polymer mat (see insert in Fig. 1). The mats were installed in the column with an inlet and outlet port at each end of the polymer tube that attached to the side of the column. Oxygen was delivered by flushing air through the inner gas space of the polymer tubing via the inlet port on the side of the column. The effluent air exiting the inner gas space of the polymer tubing via an outlet port on the side of the column was released into the atmosphere. The columns were filled with low organic-content (0.02% w/w) leached Bassendean sand, collected from the Dianella site, approximately 300 m down hydraulic gradient from the source of contamination. At this location, the groundwater was anoxic, but not contaminated with atrazine, terbutryn or fenamiphos. To maintain anoxic conditions, the soil was collected from 3.0 m below the water table and contact between the sediment/ groundwater and air was avoided by first filling the columns with groundwater under an atmosphere of nitrogen. A bailer was then used to obtain saturated sediment from below the water table. The bailer containing sediment was repeatedly filled then opened inside the column displacing excess groundwater and gradually filling the column. The columns were sealed on site. The columns were operated in saturated up-flow mode, under a hydraulic head that varied from 1.0 to 1.5 m. Gravity volumetric flow through each column was restricted to a constant 706 ml day 1 (0.54 cm h 1; assuming a soil porosity of 0.31) using a peristaltic pump on the effluent line. This velocity is similar to groundwater velocity found in the shallow aquifer beneath the Dianella site. This construction was used to avoid contact between plastic tubing used in peristaltic pumps and column influent. Experiments were run at room temperature, which was controlled coarsely at 22 C, similar to groundwater temperatures (15 –20 C). Fresh groundwater for the columns was collected fortnightly from a domestic bore within the plume and known to have been contaminated by atrazine for over 5 years. The groundwater was stored under an atmosphere of nitrogen. Groundwater chemistry data are shown in Table 1. Atrazine concentrations were over an order of magnitude greater than terbutryn or fenamiphos. Atrazine-degrading bacteria had previously been detected in groundwater collected from this bore (Tilbury, 1998). To maintain uniform concentrations of atrazine, terbutryn and fenamiphos in the influent groundwater throughout the experiment, the groundwater was spiked with a concentrated aqueous solution of the three pesticides to give final concentrations of approximately 500 mg l 1 of each. Selected physical properties of the three pesticides are given in Table 2. Also, nitrogen gas was continuously injected into the gas space above the reservoir of spiked groundwater at a rate of 5 ml min 1 to maintain anaerobic groundwater during the experiment. Preliminary microcosm experiments, using atrazine-degrading bacteria isolated from groundwater at the Dianella site, indicated that an added organic carbon source improves the mineralisation rate of atrazine (see Franzmann et al., 2000). To assess if low concentrations of naturally occurring total organic carbon present in the groundwater
200
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
Table 1 Groundwater chemistry data Parameter
Unit
Concentration
pH EC Cl N – NO3 SO24 P – PO4 Fe Mn Total organic carbon Dissolved oxygen Atrazine Terbutryn Fenamiphos
– mS m 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1
5.8 69.8 130 3.3 92 0.01 0.3 < 0.02 8.0 < 0.2 440 15 20
(8.0 mg l 1) were limiting degradation, groundwater for one column (column 1) was amended with an additional carbon source. Sodium acetate, a nonfermentable carbon source, was added to the groundwater used in column 1 to give a final concentration of 15 mg l 1. Also, the microcosm experiments indicated that while an electron acceptor was not required for atrazine degradation (degradation was via hydrolytic dechlorination; De Souza et al., 1998), enhanced growth of the bacteria occurred under aerobic conditions (see Franzmann et al., 2000). To assess the effect of oxygenated groundwater conditions on the degradation of the pesticides, the groundwater in column 1 (with an additional carbon source) and column 2 (without an additional carbon source) were purged with air through the internal gas volume of the polymer tubing mats. For a short period, air purging was halted and nitrogen purging was carried out to determine if physical removal (gas stripping via permeation through the polymer) of the pesticides was contributing to mass removal of the pesticides from the groundwater.
Table 2 Physical properties of the three pesticides investigated Compound
CAS no.
Aqueous solubility (mg l 1)
Octanol – water partitioning coefficient (Kow)
Henry’s Law constant (kPa m3 mol 1)
Atrazine Terbutryn Fenamiphos
1912-24-9 886-50-0 22224-92-6
33a 22c 330d
479b 5500c 1780d
2.48e 7c 3.0e 6c 5.8e 6d
a b c d
Weber (1994). Lopez-Avila et al. (1986). Ciba-Geigy data (1989). Miles data.
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
201
Column 3, containing a nonpurged capped polymer mat, was used as a control column to assess degradation of the pesticides under anoxic conditions. 2.2. Air purging rates through polymer tubing The rate of air purging through the internal gas volume of the polymer tubing mat to maintain near 21% oxygen concentrations at the exit of the polymer tubing was determined theoretically using a model of the transport of oxygen through the polymer tubing wall. The model was formulated to represent the experimental conditions during purging of the polymer mats. In brief, air was purged through the internal air space of the polymer tube. The influent air had a concentration of 21% oxygen at the tube inlet and was monitored at the exit of the polymer tubing. The air purging leads to redistribution of the oxygen concentration in the inner gas space, polymer and external (gaseous or aqueous phase), via diffusion of oxygen radially from the internal gas space through the polymer and into the external phase. A model was required to determine the optimum purging rates for various tubing lengths. A rate of air purging could then be selected to allow an excess supply of oxygen for the full length of the tubing. The model accounts for diffusion of oxygen from the inner gas volume of a tube through the wall of a cylindrical polymer tube with an inner radius = b and outer radius = a. The model also accounts for the longitudinal advective transport (in the z direction) of oxygen within the internal gas space, to the end of the polymer tube by an advective air stream (Fig. 2). Assuming radial diffusion through the polymer tube wall, a zero oxygen concentration in the phase external to the polymer tube, rapid radial equilibration within the inner gas space, and advective air flow inside the polymer tube, the equation governing the concentration of the oxygen in the gas inner space is (van Genuchten and Alves, 1982): @Cg @ 2 Cg @Cg ¼ Dz Fp vg @t @z2 @z
ð1Þ
subject to the following initial and boundary conditions: Cg ¼ Co
at z ¼ 0
ð2Þ
@Cg ¼0 @z
as z ! 1
ð3Þ
where Cg is the concentration of oxygen in the inner gas space of the polymer tubing, Fp is the mass flux of oxygen from the inner gas space through the polymer tube wall (at r = b) per unit volume of the inner gas space, Dz is the diffusion coefficient of oxygen in the gas space, z is the length dimension along the polymer tubing from the inlet to the outlet, and vg is the velocity of the gas stream through the polymer tube. Boundary condition (2) states that the oxygen concentration entering the tubing (at z = 0) is Co. Boundary condition (3) states that the oxygen concentration within the polymer tubing at location z is constant when z is very large (approaching infinity).
202
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
Fig. 2. Model geometry of gas-purged polymer tubing. Cg is the concentration of oxygen in the inner gas space of the polymer tubing, Cp is the concentration of oxygen in the polymer tubing, Kpg is the polymer – gas partitioning coefficient, b is the inner radius of the polymer tube, a is the outer radius of the polymer tube, and z is the longitudinal direction.
Assuming steady-state conditions, that is, @C/@t = 0, and assuming diffusion in the z direction is negligible, Eq. (1) simplifies to vg
@Cg ¼ Fp @z
ð4Þ
The mass flux term Fp in Eq. (4) can be written as Fp ¼
@Cp 2 Dp b @r
at r ¼ b
ð5Þ
where Dp is the diffusion coefficient through the polymer, r is the radius of the polymer tubing from the centre of the tube, and Cp is the concentration of oxygen within the polymer tubing wall. To determine Cp(r), we can write the equation describing steady-state conditions for the radial transport of oxygen in the polymer as (Crank, 1975): @Cp @ r ¼0 for b < r < a ð6Þ @r @r subject to the following boundary conditions: Cp ¼ Kpg Cg ðzÞ
at r ¼ b
ð7Þ
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
Cp ¼ 0
at r ¼ a
203
ð8Þ
where Cp is the oxygen concentration in the polymer tubing and Kpg is the polymer –gas partitioning coefficient. Boundary condition (7) assumes a step change in the oxygen concentration at the polymer/gas boundary. Boundary condition (8) assumes the oxygen concentration at the outer edge of the polymer is zero. This condition is taken to provide an indication of the flux/flow rate required to balance a potential maximum consumption rate of oxygen during atrazine biodegradation. Integration of Eq. (6) with respect to r gives r
@Cp ¼ A1 @r
ð9Þ
and a further integration of Eq. (9) with respect to r gives Cp ðrÞ ¼ A1 ln r þ A2
ð10Þ
where A1 and A2 are constants determined from the boundary conditions at r = a and r = b. Substituting boundary conditions (7) and (8) into Eq. (10) and eliminating A2 [since we only need A1 to fulfill Eq. (9)] gives A1 ¼
Kpg Cg ln b a
ð11Þ
Substituting Eq. (11) into Eqs. (9) and (5) gives Fp ¼
2Dp Kpg Cg b2 ln b a
at r ¼ b
ð12Þ
Eq. (4) then becomes @Cg 2Dp Kpg Cg ¼ @z vg b2 ln b a
ð13Þ
which is now a first-order differential equation in z that can be solved for Cg(z), since the other parameters are constant. Solving Eq. (13) subject to Cg = Co at z = 0 gives z2Dp Kpg
b Cg ðzÞ 2 ¼ e vg b lnð a Þ Co
ð14Þ
The gas velocity through the polymer tube (vg) is calculated as vg ¼
fr Ac
ð15Þ
204
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
where fr is the gas flow rate through the polymer tubing and Ac is the cross-sectional area within the tubing. Eq. (14) can be used to determine the minimum flow rate through the inner gas space of the polymer tubing to allow effectively uniform supply of oxygen for the full length of the tubing, by taking the effluent oxygen concentration (Cg at z = L) to be 90% of the influent oxygen concentration (Co) and using the following parameters: an inner polymer radius (b) of 0.10 cm, outer polymer radius (a) of 0.15 cm, a length of polymer tubing (L) 500 cm long, a diffusion coefficient (Dp) for oxygen through the polymer (dimethylsiloxane) tubing is 1.8 10 5 cm2 s 1 (Charati and Stern, 1998), and a Kpg value of 0.33 for oxygen. The Kpg value is based on a polymer fractional free volume of 0.33 (Charati and Stern, 1998) and assuming no oxygen sorption to the polymer (Schomburg, 1990), that is, the only oxygen is in the free volume. Using the above parameters gives a minimum flow rate through the inner space of the polymer tubing of 14 cm s 1 (26 ml min 1). For the majority of the experimental periods, air was purged through the polymer mats in each column at a rate of 40 ml min 1. Note, again that using boundary condition (8) makes this minimum flow rate estimate in fact an overestimate of the flow rate required to deliver oxygen into the external aqueous phase outside the polymer tubing.
3. Degradation rates and porosity estimation Estimated zero-order degradation rates (zd) of the pesticides were calculated using Eq. (16), from the difference in organic concentrations between sampling ports and the time required for groundwater to move from one port to the other zd ¼
Cu Cd t
ð16Þ
where Cu is the concentration of the pesticide in the groundwater collected from the port up-gradient of the oxygen delivery mat, Cd is the concentration of the pesticide in the groundwater collected from the port down-gradient of the oxygen delivery mat, and t is the estimated time for the groundwater to move from the up-gradient to the down-gradient port. First-order degradation rates (expressed as half-lives) were estimated by plotting the pesticide concentration data for the monitoring ports (ports B, C and D) against the time taken for water to migrate from the polymer mat to the monitoring port, and fitting an exponential curve to the experimental data. The time in days (t) required for groundwater to move from the up-gradient to the down-gradient port was calculated from the flow rate of groundwater through the column using Eq. (17)
t¼
pR2 hh Fr
ð17Þ
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
205
where R is the radius of the soil column (cm), h is the distance from the up-gradient to the down-gradient port, h is the soil porosity (dimensionless), and Fr is the flow rate of groundwater through the column (cm3 day 1). The effective soil porosity (h) was estimated from a bromide tracer experiment conducted during the course of the experiment. Bromide was used as a conservative tracer. A 500 ml solution of 250 mg l 1 NaBr was injected into each column through the four concentric solution delivery ports located 48 cm from the base of the column. Bromide concentrations were monitored in the highest port (port G) and the soil porosity was estimated from the time of breakthrough using Eq. (18): h¼
Fr tb pR2 hb
ð18Þ
where tb is the time of 50% mass breakthrough of the bromide after bromide injection into the column and hb is the distance between the solution injection ports and the monitoring port.
4. Analytical methods During the experiment, 5-ml water samples were collected from the columns in a glass syringe. A microextraction technique was used to extract the pesticides from water samples. This method was similar to that described by Patterson et al. (1993). The water sample was then spiked with 10 ml of a surrogate internal standard solution (1.0 g l 1 4,4V-dibromobiphenyl in methanol). Two milliliters of diethyl ether was then added and the pesticides were extracted from the water sample by shaking for 2 min. The diethyl ether layer was then transferred to a 2-ml crimp top autosampler vial, and 1 ml of the extract was then analysed by gas chromatography– mass spectrometry (GC-MS). Gas chromatographic analyses of the pesticides were performed on a Varian 3400 gas chromatograph with a Saturn II Mass Spectrometer. Analytical conditions of the GC-MS were as follows. A 25 m 0.22 mm i.d. BPX70 capillary column with a film thickness of 0.25 mm was used. A 1-ml extract was injected into a septum programmable injector (SPI) held at 80 C for 0.5 min, followed by temperature programming to 290 C at 100 C min 1 and held at this temperature for 22 min. The column was held at 80 C for 2.6 min followed by temperature programming at 25 C min 1 to 290 C and held at this temperature for 14 min. The mass spectrometer and transfer line were held at 220 and 260 C, respectively. The mass spectrometer was operated in full scan mode with a range from 50 to 350 m/z. Water samples were analysed for dissolved oxygen using an Oxi 320 dissolved oxygen meter. The probe of the dissolved oxygen meter was incorporated within a low-volume (2 ml) flow cell, to reduce the volume of water required for analysis. Approximately 10– 20 ml of water from a sampling port was allowed to flow through the flow cell before the flow was ceased and the dissolved oxygen concentration recorded. Water samples were also analysed for atrazine-degrading bacteria using the technique described by Jayachandran et al. (1998).
206
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
Bromide was analysed by HPLC with a Waters IC-Pak Anion (4.6 mm 5 cm) using UV detection at 200 nm. The eluent (a solution of 10 mM NaCl and 5 mM NaH2PO4 H2O, buffered to pH 6.2 with Na2HPO4) was pumped through the column at a flow rate of 1.0 ml min 1. The volume injected into the HPLC was 15 ml of the filtered groundwater.
5. Results and discussion 5.1. General trends Ratios of groundwater concentrations before the oxygen delivery mat (port B) and immediately after the mat (port C) are plotted for atrazine, terbutryn and fenamiphos for the three columns in Fig. 3. For terbutryn and fenamiphos, no significant differences in concentration ratios were observed between the control column (column 3) and columns 1 and 2, indicating that the addition of oxygen or oxygen and acetate did not enhance the removal of terbutryn or fenamiphos under these experimental conditions and time frames. Concentration ratios of terbutryn for all columns increased from approximately 0.6 to 1.0 during the course of the experiment. This phenomenon was most likely due to sorption of terbutryn to the polymer tubing during the experiment. Towards the end of the experiment, equilibrium partitioning may have been established between the polymer tubing and water phase, resulting in ratios close to 1.0. Also, ratios of terbutryn concentrations at port B to concentrations at port A (the two ports up-gradient of the mat) show consistent values of around 1.0 throughout the experiment for all three columns (data not shown), suggesting little sorption between sampling ports when polymer mats were not present. Strong partitioning of selected organic compounds to silicone has previously been reported (Patterson et al., 2000). The degree of partitioning was shown to be related to the organic compound octanol –water partitioning coefficient (Kow). Terbutryn has a relatively high Kow value of 5500, which would suggest strong partitioning to the silicone tubing mat. In addition, the Kow values for atrazine and fenamiphos are significantly lower than terbutryn (see Table 2), so changes in concentration ratios for these compounds due to sorption onto the polymer mat would be less likely. Rapid decreases in atrazine concentrations were observed in both columns where air was purged through the gas delivery mats (columns 1 and 2), and only in the region close to the polymer mats. No atrazine decrease was observed in the non-purged column (column 3). Removal of atrazine from the water in columns 1 and 2 may have resulted from either biological degradation or physical removal by diffusion of atrazine through the polymer tube and then volatilising into the purged gas inside the tubing. Physical removal due to volatilisation seemed unlikely due to the low Henry’s Law constant of atrazine (2.48 10 7 kPa m3 mol 1). Assuming equilibrium between the purged gas and the groundwater, based on the Henry’s Law constant of atrazine, the total mass flux of atrazine
Fig. 3. Plots of ratios of atrazine, terbutryn and fenamiphos concentrations before the oxygen delivery mat (port B) and immediately after the mat (port C) for columns 1, 2 and 3.
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
207
208
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
that could be removed through purging would be significantly less than 1% of the mass flux of atrazine in the groundwater flowing past the mat. Note, even though there is a polymer between the water and gas phases, at equilibrium, the partitioning is still governed by the Henry’s Law constant (Patterson et al., 2000). Also, since all three pesticides had similar Henry’s Law constants, preferential atrazine removal by volatilisation seems unlikely. Hydroxyatrazine concentrations in groundwater down-gradient of the polymer mats were measured during the period of oxygen delivery, using the HPLC method of Vermeulen et al. (1982). No hydroxyatrazine was detected ( < 1 mg l 1) in the groundwater, indicating that either hydroxyatrazine was not formed or hydroxyatrazine degradation rates were rapid enough that accumulation of hydroxyatrazine did not occur. To confirm that the removal of atrazine was not due to physical stripping, purging of the mats with air was halted and the mats were then purged with nitrogen gas. Once nitrogen purging commenced, dissolved oxygen concentrations decreased from approximately 4 mg l 1 to less than 0.2 mg l 1 at port C for both columns. There was also a corresponding increase in atrazine concentrations relative to (influent) concentrations at port B. These changes in dissolved oxygen and atrazine ratios are shown for column 2 in Fig. 4. These data demonstrate that atrazine removal was through oxygen-mediated biotic or abiotic processes. Microcosm experiments (Franzmann et al., 2000) have shown no atrazine mineralisation in soils that have been taken from this aquifer and sterilised. In addition, atrazine-degrading bacteria have been detected in these columns; therefore,
Fig. 4. Changes in dissolved oxygen (port C) and atrazine ratios in column 2 before and after air purging of the polymer mat halted and nitrogen purging commenced.
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
209
biodegradation by oxygen-requiring microorganisms seems the most plausible mechanism for atrazine removal. Once oxygen delivery was halted, it is assumed that the microorganisms stopped metabolising the hydrolytic enzymes required for atrazine degradation, and the enzymes already present lost their activity over a period of 1 – 3 days. 5.2. Soil porosity Soil porosities of each column were estimated from their bromide breakthrough curve (see Fig. 5) and using Eq. (18). While bromide peak heights/areas varied between columns, estimated breakthrough times were similar, especially for columns 2 and 3. Estimated soil porosities for all three columns were similar with an average porosity of 0.33 F 0.02 (error = standard deviation). These porosities were similar to the estimated field porosity of 0.37 F 0.02, determined from recovered intact cores. 5.3. Biodegradation rates Since degradation of atrazine occurred rapidly and only in the region close to the polymer mats, it was difficult to determine if degradation was zero order or first order; therefore, degradation rates of the pesticides were calculated as both zero-order and firstorder degradation rates. First-order rates are expressed as half-lives. Zero-order biodegradation rates were calculated from the difference in concentrations of the pesticides in groundwater collected from ports B and C [up-gradient and down-
Fig. 5. Bromide breakthrough at port G, 182 cm from the base of the column, after bromide injection into solution delivery ports at 48 cm from the base of the column.
210
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
gradient of the oxygen delivery mat using Eq. (16)]. The times required for groundwater to migrate from port B to C were estimated from Eq. (17) using the soil porosity and the volumetric flow rate through the column of 706 ml day 1. The estimated travel times between ports B and C were 1.5 days for columns 1 and 3, and 1.6 days for column 2. Degradation rates between ports B and C were estimated during stable periods of air purging of the mats and nitrogen gas purging of the mats. For the air purging, degradation rates were estimated from average concentrations of the pesticides between the 5th and 22nd June 1998 (n = 7). During this time, concentrations were stable for both ports B and C. During nitrogen gas purging, degradation rates were estimated from average concentrations between the 26th June and 13th July 1998 (n = 5). Again, during this time, concentrations were stable for both ports B and C. Zero-order biodegradation rates estimated for both these periods are given in Table 3. When the groundwater was oxygenated, there was a rapid decrease in atrazine concentrations with degradation rates of 380 and 240 mg l 1 day 1 for columns 1 and 2. These rates are possibly an underestimate, since degradation may be negligible in the time groundwater takes to migrate between port B and the oxygen delivery mat. Also, the degradation may only occur in a narrow zone at the mat or down-gradient from the mat. While degradation rates were different between the columns with and without acetate, this difference is most likely due to differences in the initial atrazine concentrations at port B in columns 1 and 2. The average concentrations were 600 F 70 and 370 F 50 mg l 1 (errors are the standard deviation). Estimated half-lives, based on average concentrations between the 5th and 22nd June 1998 for ports B, C and D, are given in Table 4. Atrazine half-lives for oxygenated conditions were 0.35 F 0.04 and 0.35 F 0.07 days for columns 1 and 2. These data would suggest that the naturally occurring dissolved organic carbon in the groundwater (8 mg l 1) might be sufficient for bacteria to degrade atrazine. These degradation rates compare favourably with a half-life of 0.26 F 0.2 days, estimated from culture experiments using citrate as a carbon source (Tilbury, 1998).
5.4. Oxygen delivery Based on Eq. (14) with an air-flow rate of 40 ml min 1 and Cg/Co = 0.9, 10% (1.12 mg min 1) of the total oxygen influx through the inner space of tubing should diffuse through
Table 3 Estimated zero-order degradation rates for atrazine, terbutryn and fenamiphos Zero-order degradation rate (mg l 1 day 1) Atrazine
Oxygen + acetate Oxygen only Control, no oxygen
Terbutryn
Fenamiphos
Oxygen purge
Nitrogen purge
No purge
Oxygen purge
Nitrogen purge
No purge
Oxygen purge
Nitrogen purge
No purge
380 240 –
< 10 < 10 –
– – < 10
< 10 < 10 –
< 10 < 10 –
– – < 10
< 10 < 10 –
< 10 < 10 –
– – < 10
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
211
Table 4 Estimated half-lives for atrazine, terbutryn and fenamiphos Half-life (days) Atrazine Oxygen purge Oxygen + acetate 0.35 F 0.04 Oxygen only 0.35 F 0.07 Control, no oxygen – Culturea 0.26 F 0.02
Terbutryn
Fenamiphos
Nitrogen No Oxygen Nitrogen No Oxygen Nitrogen No purge purge purge purge purge purge purge purge > 20 > 20 – –
– – > 20 –
> 20 > 20 – –
> 20 > 20 – –
– – > 20 –
> 20 > 20 – –
> 20 > 20 – –
– – > 20 –
Errors are standard errors. a Tilbury (1998).
the polymer and be instantaneously consumed at the outer surface of the polymer tubing, or transported past the mat with groundwater flow. If all the oxygen that diffuses through the polymer is transported into the groundwater flowing past the polymer mat, the mass flux rate of oxygen (1.12 mg min 1) should be sufficient to saturate the anoxic groundwater flowing past the mat (4.9 10 4 l min 1) with oxygen to give an oxygen concentration in the groundwater of approximately 8.7 mg l 1. Oxygen concentrations measured in the groundwater at port C (11 cm down-gradient from the polymer mats) in columns 1 and 2 were approximately 4 mg l 1. The lower oxygen concentration suggests that there has been oxygen consumption in the groundwater between the contact time with the polymer mat and the sampling time at port C, a time difference of approximately 20 h. As significant atrazine degradation occurred within this zone, some oxygen consumption would appear likely. Another possible explanation for the lower oxygen concentrations may be that the delivery of oxygen to groundwater was less efficient than predicted. Possible reasons could be mass transfer limitations at the polymer – water interface, or insufficient polymer – water surface area resulting in a mixture of oxygen-saturated groundwater (groundwater in contact with the polymer) and anoxic groundwater (groundwater that did not make contact with the polymer). Correspondence in atrazine half-lives between microcosm and column experiments suggests oxygen did not limit degradation.
6. Conclusions The column experiments have demonstrated that bacteria in the pesticide-contaminated groundwater from the Dianella site were capable of degrading atrazine under aerobic conditions. Degradation rates of atrazine were relatively high with a zero-order rate of 240– 380 mg l 1 day 1 or half-life of 0.35 days. Amendment with an additional carbon source showed no significant improvement in biodegradation rates, suggesting that organic carbon was not limiting biodegradation. Atrazine degradation rates estimated in the column experiments were similar to rates determined in laboratory culture experiments.
212
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
No significant degradation of terbutryn or fenamiphos was observed under the experimental conditions and time frames studied. An alternative remediation strategy may be needed for these compounds. The polymer tubes constructed into mats and purged with air provided an efficient method to deliver dissolved oxygen to anaerobic groundwater. Dissolved oxygen concentrations increased from < 0.2 mg l 1 to approximately 4 mg l 1. The use of polymer tube mats may provide a low-infrastructure, long-term field technique for the delivery of oxygen to groundwater. In comparison to alternative oxygen delivery techniques, such as air sparging or peroxide amendment, this technique may be more cost-effective and reliable, although this aspect has not been assessed here. Results from these experiments indicate that remediation of atrazine at the Dianella site may be achieved through oxygenation of the anaerobic contaminated zone of the aquifer. No significant degradation of terbutryn or fenamiphos was observed under the experimental conditions within the time frames of the study. While bacteria capable of degrading atrazine have been isolated from the central portion of the pesticide plume, such bacteria have not been isolated from groundwater at the leading edge of the pesticide plume. Therefore, if remediation is desirable at the leading edge of the pesticide plume, augmentation of this zone of contaminated groundwater with atrazine-degrading bacteria may also be required.
Acknowledgements The authors thank David Briegel, Terry Power, Nathan Innes and Dean Coad for their technical assistance. Thanks also to Dr. C. Barber and Dr. P. Clement for critically reviewing the manuscript. This work was partly funded by the Water and Rivers Commission of Western Australia through the Centre for Groundwater Studies.
References Agertved, J., Rugge, K., Baker, J.K., 1992. Transformation of the herbicide MCPP and atrazine under natural aquifer conditions. Ground Water 30 (4), 500 – 506. Appleyard, S.J., 1995. Investigation of ground water contamination by fenamiphos and atrazine in a residential area: source and distribution of contaminants. Ground Water Monit. Rev. 15 (4), 110 – 113. Blume, E., Bischoff, M., Moorman, T., Turco, R.F., 1993. Assimilative capacity of subsurface soils for atrazine. Agron. Abstr. 83, 241. Blumhorst, M.R., Weber, J.B., 1994. Chemical versus microbial degradation of cyanazine and atrazine in soils. Pestic. Sci. 42, 79 – 84. Charati, S.G., Stern, S.A., 1998. Diffusion of gases in silicone polymers: molecular dynamics simulations. Macromolecules 31, 5529 – 5535. Ciba-Geigy data, 1989. Crank, J., 1975. The Mathematics of Diffusion, 2nd edn. Clarendon Press, Oxford. Davis, G.B., Johnston, C.D., Patterson, B.M., Barber, C., Bennett, M., 1998. Estimation of biodegradation rates using respiration tests during in situ bioremediation of weathered diesel NAPL. Ground Water Monit. Rem. 18 (2), 123 – 132. De Souza, M.L., Wackett, L.P., Sadowsky, M.J., 1998. The atzABC genes encoding atrazine catabolism are
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
213
located on a self-transmissible plasmid in Pseudomonas sp. strain ADP. Appl. Environ. Microbiol. 64, 2323 – 2326. Franzmann, P.D., Zappia, L., Patterson, B.M., Rayner, J.R., Davis, G.B., 1998. Mineralization of low concentrations of organic compounds and microbial biomass in vadose zone soils from Perth Coastal Plain. Aust. J. Soil Res. 36, 921 – 939. Franzmann, P.D., Zappia, L.R., Tilbury, A.L., Patterson, B.M., Davis, G.B., Mandelbaum, R.T., 2000. Bioaugmentation of atrazine and fenamiphos impacted groundwater: laboratory evaluation. Bioremediation J. 4 (3), 237 – 248. Gibson, T.L., Abdul, A.S., Chalmer, P.D., 1998. Enhancement of in situ bioremediation of BTEX-contaminated ground water by oxygen diffusion from silicone tubing. Ground Water Monit. Rem. 18 (1), 93 – 104. Herrling, B., Stamm, J., Alesi, E.J., Bott-Breuning, G., Diekmann, S., 1994. In situ bioremediation of groundwater containing hydrocarbons, pesticides, or nitrate using vertical circulation flows (UVB/GZB technique). In: Hinchee, R.E. (Ed.), Air Sparging for Site Remediation. Lewis Publishers, Boca Raton, pp. 56 – 80. Jayachandran, K., Stolpe, N.B., Moorman, T.B., Shea, P.J., 1998. Application of 14C-most-probable-number technique to enumerate atrazine-degrading microorganisms in soil. Soil Biol. Biochem. 30 (4), 523 – 529. Johnston, C.D., Rayner, J.L., Patterson, B.M., Davis, G.B., 1998. The contribution of volatilisation and biodegradation during air sparging of dissolved BTEX-contaminated groundwater. J. Contam. Hydrol. 33, 377 – 404. Kookana, R.S., Phang, C., Aylmore, L.A.G., 1997. Transformation and degradation of fenamiphos nematicide and its metabolites in soils. Aust. J. Soil Res. 35, 753 – 761. Lopez-Avila, V., Hirata, P., Kraska, S., Flanagan, M., Taylor Jr., J.H., Hern, S.C., Melancon, S., Pollard, J., 1986. Movement of selected pesticides and herbicides through columns of sandy loam. In: Garner, W.Y., Honeycutt, R.C., Nigg, H.N. (Eds.), Evaluation of Pesticides in Ground Water. American Chemical Society, Washington, DC, ACS Symposium Series 315, 311 – 328. Miles data, Agricultural Chemicals Division, 8400 Hawthorn Road, Kansas City, MO 64120. Mirgain, I., Green, G., Monteil, H., 1995. Biodegradation of the herbicide atrazine in groundwater under laboratory conditions. Environ. Technol. 16, 967 – 976. Nair, D.R., Schnoor, J.L., 1992. Effect of two electron acceptors on atrazine mineralization rates in soil. Environ. Sci. Technol. 26, 2298 – 2300. Ou, L.-T., Thomas, J.E., Dickson, D.W., 1994. Degradation of fenamiphos in soil with a history of continuous fenamiphos application. Soil Sci. Soc. Am. J. 58, 1139 – 1147. Patterson, B.M., Power, T.R., Barber, C., 1993. Comparison of two integrated methods for the collection and analysis of volatile organic compounds in groundwater. Ground Water Monit. Rem. 13 (3), 118 – 123. Patterson, B.M., Franzmann, P.D., Rayner, J., Davis, G.B., 1998. Mobility and biodegradation of low-levels of organic compounds in the vadose zone of the Swan coastal plain. Conference paper for the IAH International Groundwater Conference, Melbourne, Australia, February 1998, pp. 533 – 538. Patterson, B.M., Davis, G.B., McKinley, A.J., 2000. Volatile organic compounds in groundwater, probes for the analysis of. In: Meyers, R.A., et al., (Ed.), Encyclopedia of Analytical Chemistry: Instrumentation and Application. Wiley, Chichester, pp. 3515 – 3526. Schomburg, G., 1990. Gas Chromatography: A Practical Course. VCH Publishers, New York, p. 298. Tilbury, A., 1998. Hons. Thesis, Department of Chemistry, The University of Western Australia. van Genuchten, M.T., Alves, W.J., 1982. Analytical solutions of the one-dimensional convective – dispersive solute transport equation. U.S. Dep. Agric., Tech. Bull. No. 1661, 151 pp. Vermeulen, N.M.J., Apostolides, Z., Potgieter, D.J.J., Nel, P.C., Smit, N.S.H., 1982. Separation of atrazine and some of its degradation products by high performance liquid chromatography. J. Chromatogr. 240, 247 – 253. Weber, J.B., 1994. Properties and behavior of pesticides in soil. In: Honeycutt, R.C., Schabacker, D.J. (Eds.), Mechanisms of Pesticide Movement into Ground Water. Lewis Publishers, Boca Raton, pp. 15 – 41. Yamagiwa, K., Yoshida, M., Ito, A., Ohkawa, A., 1998. A new oxygen supply method for simultaneous organic carbon removal and nitrification by a one-stage biofilm process. Water Sci. Technol. 37 (4 – 5), 117 – 124.
Using polymer mats to biodegrade atrazine in groundwater: laboratory column experiments B.M. Patterson a,*, P.D. Franzmann a, G.B. Davis a, J. Elbers b, L.R. Zappia a a
CSIRO Land and Water, Private Bag No. 5, Wembley, WA 6913, Australia b Paderborn University, Germany
Received 9 August 2000; received in revised form 11 July 2001; accepted 19 July 2001
Abstract Large-scale column experiments were undertaken to evaluate the potential of in situ polymer mats to deliver oxygen into groundwater to induce biodegradation of the pesticides atrazine, terbutryn and fenamiphos contaminating groundwater in Perth, Western Australia. The polymer mats, composed of woven silicone (dimethylsiloxane) tubes and purged with air, were installed in 2-m-long flowthrough soil columns. The polymer mats proved efficient in delivering dissolved oxygen to anaerobic groundwater. Dissolved oxygen concentrations increased from < 0.2 mg l 1 to approximately 4 mg l 1. Degradation rates of atrazine in oxygenated groundwater were relatively high with a zero-order rate of 240 – 380 mg l 1 or a first-order half-life of 0.35 days. Amendment with an additional carbon source showed no significant improvement in biodegradation rates, suggesting that organic carbon was not limiting biodegradation. Atrazine degradation rates estimated in the column experiments were similar to rates determined in laboratory culture experiments, using pure cultures of atrazinemineralising bacteria. No significant degradation of terbutryn or fenamiphos was observed under the experimental conditions within the time frames of the study. Results from these experiments indicate that remediation of atrazine in a contaminated aquifer may be achievable by delivery of oxygen using an in situ polymer mat system. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Atrazine; Biodegradation; Polymer; Oxygen delivery
*
Corresponding author. Tel.: +61-8-9333-6276; fax: +61-8-9333-6211. E-mail address: [email protected] (B.M. Patterson).
0169-7722/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 7 7 2 2 ( 0 1 ) 0 0 1 7 8 - 4
196
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
1. Introduction Contamination of anoxic groundwater by atrazine (2-chloro-4-ethylamino-6-isopropylamino-1,3,5-triazine), terbutryn (2-tert-butylamino-4-ethylamino-6-methylthio-1,3,5triazine) and fenamiphos [1-(methylethyl)-ethyl-3-methyl-4-(methylthio) phenylphosphoramidate] from infiltration of pesticide-laden washdown water has been reported in Dianella, a suburb in Perth, Western Australia (Appleyard, 1995). Such contamination is of great concern in Perth because 35% of domestic drinking water supply is obtained from groundwater, and a similar quantity is used directly on gardens. During initial ‘‘pump-and-treat’’ remediation over 4 months, atrazine concentrations increased from approximately 800 to 2000 mg l 1 after pumping commenced (possibly as a result of plume breakthrough) and then decreased to approximately 600 mg l 1 when pumping was halted. Fenamiphos concentrations remained relatively constant before and during pumping at approximately 50 mg l 1; however, concentrations increased to approximately 100 mg l 1 once pumping was halted (Appleyard, 1995). The ‘‘pumpand-treat’’ remediation suggested that short-term strategies might not be suitable for this site. Both atrazine (Nair and Schnoor, 1992; Blume et al., 1993; Blumhorst and Weber, 1994) and fenamiphos (Ou et al., 1994; Kookana et al., 1997; Patterson et al., 1998) have been previously reported to undergo biodegradation. However, this research has focussed on the fate of these compounds in contaminated shallow soil profiles rather than groundwater. Rates of degradation of these compounds vary significantly and seem to be dependent on individual site conditions, but generally decrease with depth through the vadose zone, largely due to lower total organic carbon levels at depth (Franzmann et al., 1998). There have been limited reports of atrazine and/or fenamiphos remediation in groundwater, while some studies have reported atrazine to be recalcitrant in aquifer material (Agertved et al., 1992). Herrling et al. (1994) reported bioremediation of a groundwater plume of triazine pesticides (including atrazine) using an in-well aeration technology. Mirgain et al. (1995) observed in laboratory experiments that there was a threshold concentration of 200 mg l 1 below which atrazine degradation could not be induced. Bacteria capable of complete mineralisation of atrazine have recently been isolated from the Dianella site (Tilbury, 1998). The bacteria were found to degrade atrazine through an enzymatic mediated hydrolytic dechlorination pathway to hydroxyatrazine, with further degradation to CO2 and H2O. Microcosm experiments by Franzmann et al. (2000) indicated that the half-life of atrazine mineralisation in unamended anoxic aquifer material from the Dianella site was greater than 20 years. However, amendment of microcosms with an atrazine-degrading culture (collected from the Dianella site), and a source of oxygen and carbon, decreased the half-life of atrazine mineralisation to approximately 20 days. In the microcosms that received considerable biomass of atrazine-degrading aerobic cells ( f 105 cells g 1), some atrazine mineralisation still occurred in the absence of oxygen, which is consistent with the findings that the enzymatic steps in the mineralisation pathway are hydrolytic and do not require oxygen (De Souza et al., 1998). Since the organic carbon content of the aquifer material from the Dianella site has been measured at
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
197
0.7 g kg 1, the addition of a carbon source may not be required. These microcosm experiments indicated that oxygen and atrazine-degrading bacteria are an essential requirement for atrazine remediation at the Dianella site. Oxygen delivery for remediation may normally involve air injection (air sparging), addition of peroxide, or addition of aerated water (Davis et al., 1998). These oxygen delivery techniques have been used successfully to remediate petroleum-contaminated groundwater (Johnston et al., 1998). Oxygen delivery has also been achieved using a more passive oxygen diffusion system made of sealed silicone (dimethylsiloxane) polymer tubing installed in a borehole and pressurised with oxygen (Gibson et al., 1998). Yamagiwa et al. (1998) used a similar oxygen supply method in a bioreactor for simultaneous organic carbon removal and nitrification. Adaptation of this polymer tubing technique may well be suited as a longer term, semipassive strategy to remediate contaminated groundwaters, using natural groundwater flow to deliver contaminated groundwater to the polymer mats engineered as an in situ biologically active barrier. The advantage of using diffusion of oxygen across a membrane to deliver oxygen to the subsurface is that the location of delivery of oxygen may be more reproducible and less dynamic and disturbing than other oxygen delivery techniques (e.g. air sparging). Delivery in this way is thought to enhance the formation of a more stable atrazine-degrading biofilm on the surface of the polymer mat (Yamagiwa et al., 1998), since it should be able to achieve simultaneous and more consistent delivery of oxygen from polymer membrane/biofilm interface to the aqueous/biofilm interface and contaminant from the aqueous/biofilm interface to the polymer/biofilm interface. This biologically active polymer-mat barrier concept was tested in laboratory largescale (150 mm diameter, 2.0 m long) column experiments to investigate the potential for biodegradation of atrazine, terbutryn and fenamiphos in groundwater for a variety of groundwater conditions. The experiments were conducted under anaerobic and aerobic groundwater conditions, and with and without an additional carbon source. In particular, oxygen delivery to anoxic groundwater was tested using air purging of silicone polymer tubing woven to form a polymer mat, relying on sustained oxygen diffusion through the polymer tubing wall. Large-scale flow-through column experiments were used in preference to microcosm experiments, since column experiments are more representative of field conditions. By using large-scale columns, similar hydrogeological conditions, such as flow velocity, and soil porosity can be set up, side-wall effects can be reduced, and amendments and remediation options can also be better tested.
2. Materials and methods 2.1. Experimental set up and operation Three aluminium columns (2.0 m long, 150 mm diameter) were constructed with fine stainless steel mesh placed at both ends, 10 cm from the top and bottom of the column to prevent soil deposition in the feed and effluent lines. A polymer ‘mat’ was installed 46 cm from the base of each column. Seven water sampling ports were also positioned along the
198
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
length of each column, at 4 (port A), 37 (port B), 57 (port C), 77 (port D), 107 (port E), 137 (port F) and 182 cm (port G) from the base of the column. The first port was set below the stainless steel mesh (below the soil) and was used to monitor influent aqueous concentrations to the column. A set of four solution delivery ports was positioned around the circumference of the column at a distance of 48 cm from the base of the column. These were used to inject a bromide tracer to estimate soil porosity. A schematic of a column is shown in Fig. 1.
Fig. 1. Schematic of the flow-through column.
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
199
The oxygen delivery mats consisted of a 5.0-m length of silicone tubing (2.0 mm i.d., 3.0 mm o.d.) with a fine stainless steel spring inserted into the centre of the polymer tubing to provide support and eliminate the possibility of the tubing twisting or collapsing. The polymer tubing was then woven through a stainless steel support frame consisting of two 14-cm diameter mesh disks. The two disks were 3 cm apart to form a bilayered polymer mat (see insert in Fig. 1). The mats were installed in the column with an inlet and outlet port at each end of the polymer tube that attached to the side of the column. Oxygen was delivered by flushing air through the inner gas space of the polymer tubing via the inlet port on the side of the column. The effluent air exiting the inner gas space of the polymer tubing via an outlet port on the side of the column was released into the atmosphere. The columns were filled with low organic-content (0.02% w/w) leached Bassendean sand, collected from the Dianella site, approximately 300 m down hydraulic gradient from the source of contamination. At this location, the groundwater was anoxic, but not contaminated with atrazine, terbutryn or fenamiphos. To maintain anoxic conditions, the soil was collected from 3.0 m below the water table and contact between the sediment/ groundwater and air was avoided by first filling the columns with groundwater under an atmosphere of nitrogen. A bailer was then used to obtain saturated sediment from below the water table. The bailer containing sediment was repeatedly filled then opened inside the column displacing excess groundwater and gradually filling the column. The columns were sealed on site. The columns were operated in saturated up-flow mode, under a hydraulic head that varied from 1.0 to 1.5 m. Gravity volumetric flow through each column was restricted to a constant 706 ml day 1 (0.54 cm h 1; assuming a soil porosity of 0.31) using a peristaltic pump on the effluent line. This velocity is similar to groundwater velocity found in the shallow aquifer beneath the Dianella site. This construction was used to avoid contact between plastic tubing used in peristaltic pumps and column influent. Experiments were run at room temperature, which was controlled coarsely at 22 C, similar to groundwater temperatures (15 –20 C). Fresh groundwater for the columns was collected fortnightly from a domestic bore within the plume and known to have been contaminated by atrazine for over 5 years. The groundwater was stored under an atmosphere of nitrogen. Groundwater chemistry data are shown in Table 1. Atrazine concentrations were over an order of magnitude greater than terbutryn or fenamiphos. Atrazine-degrading bacteria had previously been detected in groundwater collected from this bore (Tilbury, 1998). To maintain uniform concentrations of atrazine, terbutryn and fenamiphos in the influent groundwater throughout the experiment, the groundwater was spiked with a concentrated aqueous solution of the three pesticides to give final concentrations of approximately 500 mg l 1 of each. Selected physical properties of the three pesticides are given in Table 2. Also, nitrogen gas was continuously injected into the gas space above the reservoir of spiked groundwater at a rate of 5 ml min 1 to maintain anaerobic groundwater during the experiment. Preliminary microcosm experiments, using atrazine-degrading bacteria isolated from groundwater at the Dianella site, indicated that an added organic carbon source improves the mineralisation rate of atrazine (see Franzmann et al., 2000). To assess if low concentrations of naturally occurring total organic carbon present in the groundwater
200
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
Table 1 Groundwater chemistry data Parameter
Unit
Concentration
pH EC Cl N – NO3 SO24 P – PO4 Fe Mn Total organic carbon Dissolved oxygen Atrazine Terbutryn Fenamiphos
– mS m 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1 mg l 1
5.8 69.8 130 3.3 92 0.01 0.3 < 0.02 8.0 < 0.2 440 15 20
(8.0 mg l 1) were limiting degradation, groundwater for one column (column 1) was amended with an additional carbon source. Sodium acetate, a nonfermentable carbon source, was added to the groundwater used in column 1 to give a final concentration of 15 mg l 1. Also, the microcosm experiments indicated that while an electron acceptor was not required for atrazine degradation (degradation was via hydrolytic dechlorination; De Souza et al., 1998), enhanced growth of the bacteria occurred under aerobic conditions (see Franzmann et al., 2000). To assess the effect of oxygenated groundwater conditions on the degradation of the pesticides, the groundwater in column 1 (with an additional carbon source) and column 2 (without an additional carbon source) were purged with air through the internal gas volume of the polymer tubing mats. For a short period, air purging was halted and nitrogen purging was carried out to determine if physical removal (gas stripping via permeation through the polymer) of the pesticides was contributing to mass removal of the pesticides from the groundwater.
Table 2 Physical properties of the three pesticides investigated Compound
CAS no.
Aqueous solubility (mg l 1)
Octanol – water partitioning coefficient (Kow)
Henry’s Law constant (kPa m3 mol 1)
Atrazine Terbutryn Fenamiphos
1912-24-9 886-50-0 22224-92-6
33a 22c 330d
479b 5500c 1780d
2.48e 7c 3.0e 6c 5.8e 6d
a b c d
Weber (1994). Lopez-Avila et al. (1986). Ciba-Geigy data (1989). Miles data.
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
201
Column 3, containing a nonpurged capped polymer mat, was used as a control column to assess degradation of the pesticides under anoxic conditions. 2.2. Air purging rates through polymer tubing The rate of air purging through the internal gas volume of the polymer tubing mat to maintain near 21% oxygen concentrations at the exit of the polymer tubing was determined theoretically using a model of the transport of oxygen through the polymer tubing wall. The model was formulated to represent the experimental conditions during purging of the polymer mats. In brief, air was purged through the internal air space of the polymer tube. The influent air had a concentration of 21% oxygen at the tube inlet and was monitored at the exit of the polymer tubing. The air purging leads to redistribution of the oxygen concentration in the inner gas space, polymer and external (gaseous or aqueous phase), via diffusion of oxygen radially from the internal gas space through the polymer and into the external phase. A model was required to determine the optimum purging rates for various tubing lengths. A rate of air purging could then be selected to allow an excess supply of oxygen for the full length of the tubing. The model accounts for diffusion of oxygen from the inner gas volume of a tube through the wall of a cylindrical polymer tube with an inner radius = b and outer radius = a. The model also accounts for the longitudinal advective transport (in the z direction) of oxygen within the internal gas space, to the end of the polymer tube by an advective air stream (Fig. 2). Assuming radial diffusion through the polymer tube wall, a zero oxygen concentration in the phase external to the polymer tube, rapid radial equilibration within the inner gas space, and advective air flow inside the polymer tube, the equation governing the concentration of the oxygen in the gas inner space is (van Genuchten and Alves, 1982): @Cg @ 2 Cg @Cg ¼ Dz Fp vg @t @z2 @z
ð1Þ
subject to the following initial and boundary conditions: Cg ¼ Co
at z ¼ 0
ð2Þ
@Cg ¼0 @z
as z ! 1
ð3Þ
where Cg is the concentration of oxygen in the inner gas space of the polymer tubing, Fp is the mass flux of oxygen from the inner gas space through the polymer tube wall (at r = b) per unit volume of the inner gas space, Dz is the diffusion coefficient of oxygen in the gas space, z is the length dimension along the polymer tubing from the inlet to the outlet, and vg is the velocity of the gas stream through the polymer tube. Boundary condition (2) states that the oxygen concentration entering the tubing (at z = 0) is Co. Boundary condition (3) states that the oxygen concentration within the polymer tubing at location z is constant when z is very large (approaching infinity).
202
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
Fig. 2. Model geometry of gas-purged polymer tubing. Cg is the concentration of oxygen in the inner gas space of the polymer tubing, Cp is the concentration of oxygen in the polymer tubing, Kpg is the polymer – gas partitioning coefficient, b is the inner radius of the polymer tube, a is the outer radius of the polymer tube, and z is the longitudinal direction.
Assuming steady-state conditions, that is, @C/@t = 0, and assuming diffusion in the z direction is negligible, Eq. (1) simplifies to vg
@Cg ¼ Fp @z
ð4Þ
The mass flux term Fp in Eq. (4) can be written as Fp ¼
@Cp 2 Dp b @r
at r ¼ b
ð5Þ
where Dp is the diffusion coefficient through the polymer, r is the radius of the polymer tubing from the centre of the tube, and Cp is the concentration of oxygen within the polymer tubing wall. To determine Cp(r), we can write the equation describing steady-state conditions for the radial transport of oxygen in the polymer as (Crank, 1975): @Cp @ r ¼0 for b < r < a ð6Þ @r @r subject to the following boundary conditions: Cp ¼ Kpg Cg ðzÞ
at r ¼ b
ð7Þ
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
Cp ¼ 0
at r ¼ a
203
ð8Þ
where Cp is the oxygen concentration in the polymer tubing and Kpg is the polymer –gas partitioning coefficient. Boundary condition (7) assumes a step change in the oxygen concentration at the polymer/gas boundary. Boundary condition (8) assumes the oxygen concentration at the outer edge of the polymer is zero. This condition is taken to provide an indication of the flux/flow rate required to balance a potential maximum consumption rate of oxygen during atrazine biodegradation. Integration of Eq. (6) with respect to r gives r
@Cp ¼ A1 @r
ð9Þ
and a further integration of Eq. (9) with respect to r gives Cp ðrÞ ¼ A1 ln r þ A2
ð10Þ
where A1 and A2 are constants determined from the boundary conditions at r = a and r = b. Substituting boundary conditions (7) and (8) into Eq. (10) and eliminating A2 [since we only need A1 to fulfill Eq. (9)] gives A1 ¼
Kpg Cg ln b a
ð11Þ
Substituting Eq. (11) into Eqs. (9) and (5) gives Fp ¼
2Dp Kpg Cg b2 ln b a
at r ¼ b
ð12Þ
Eq. (4) then becomes @Cg 2Dp Kpg Cg ¼ @z vg b2 ln b a
ð13Þ
which is now a first-order differential equation in z that can be solved for Cg(z), since the other parameters are constant. Solving Eq. (13) subject to Cg = Co at z = 0 gives z2Dp Kpg
b Cg ðzÞ 2 ¼ e vg b lnð a Þ Co
ð14Þ
The gas velocity through the polymer tube (vg) is calculated as vg ¼
fr Ac
ð15Þ
204
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
where fr is the gas flow rate through the polymer tubing and Ac is the cross-sectional area within the tubing. Eq. (14) can be used to determine the minimum flow rate through the inner gas space of the polymer tubing to allow effectively uniform supply of oxygen for the full length of the tubing, by taking the effluent oxygen concentration (Cg at z = L) to be 90% of the influent oxygen concentration (Co) and using the following parameters: an inner polymer radius (b) of 0.10 cm, outer polymer radius (a) of 0.15 cm, a length of polymer tubing (L) 500 cm long, a diffusion coefficient (Dp) for oxygen through the polymer (dimethylsiloxane) tubing is 1.8 10 5 cm2 s 1 (Charati and Stern, 1998), and a Kpg value of 0.33 for oxygen. The Kpg value is based on a polymer fractional free volume of 0.33 (Charati and Stern, 1998) and assuming no oxygen sorption to the polymer (Schomburg, 1990), that is, the only oxygen is in the free volume. Using the above parameters gives a minimum flow rate through the inner space of the polymer tubing of 14 cm s 1 (26 ml min 1). For the majority of the experimental periods, air was purged through the polymer mats in each column at a rate of 40 ml min 1. Note, again that using boundary condition (8) makes this minimum flow rate estimate in fact an overestimate of the flow rate required to deliver oxygen into the external aqueous phase outside the polymer tubing.
3. Degradation rates and porosity estimation Estimated zero-order degradation rates (zd) of the pesticides were calculated using Eq. (16), from the difference in organic concentrations between sampling ports and the time required for groundwater to move from one port to the other zd ¼
Cu Cd t
ð16Þ
where Cu is the concentration of the pesticide in the groundwater collected from the port up-gradient of the oxygen delivery mat, Cd is the concentration of the pesticide in the groundwater collected from the port down-gradient of the oxygen delivery mat, and t is the estimated time for the groundwater to move from the up-gradient to the down-gradient port. First-order degradation rates (expressed as half-lives) were estimated by plotting the pesticide concentration data for the monitoring ports (ports B, C and D) against the time taken for water to migrate from the polymer mat to the monitoring port, and fitting an exponential curve to the experimental data. The time in days (t) required for groundwater to move from the up-gradient to the down-gradient port was calculated from the flow rate of groundwater through the column using Eq. (17)
t¼
pR2 hh Fr
ð17Þ
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
205
where R is the radius of the soil column (cm), h is the distance from the up-gradient to the down-gradient port, h is the soil porosity (dimensionless), and Fr is the flow rate of groundwater through the column (cm3 day 1). The effective soil porosity (h) was estimated from a bromide tracer experiment conducted during the course of the experiment. Bromide was used as a conservative tracer. A 500 ml solution of 250 mg l 1 NaBr was injected into each column through the four concentric solution delivery ports located 48 cm from the base of the column. Bromide concentrations were monitored in the highest port (port G) and the soil porosity was estimated from the time of breakthrough using Eq. (18): h¼
Fr tb pR2 hb
ð18Þ
where tb is the time of 50% mass breakthrough of the bromide after bromide injection into the column and hb is the distance between the solution injection ports and the monitoring port.
4. Analytical methods During the experiment, 5-ml water samples were collected from the columns in a glass syringe. A microextraction technique was used to extract the pesticides from water samples. This method was similar to that described by Patterson et al. (1993). The water sample was then spiked with 10 ml of a surrogate internal standard solution (1.0 g l 1 4,4V-dibromobiphenyl in methanol). Two milliliters of diethyl ether was then added and the pesticides were extracted from the water sample by shaking for 2 min. The diethyl ether layer was then transferred to a 2-ml crimp top autosampler vial, and 1 ml of the extract was then analysed by gas chromatography– mass spectrometry (GC-MS). Gas chromatographic analyses of the pesticides were performed on a Varian 3400 gas chromatograph with a Saturn II Mass Spectrometer. Analytical conditions of the GC-MS were as follows. A 25 m 0.22 mm i.d. BPX70 capillary column with a film thickness of 0.25 mm was used. A 1-ml extract was injected into a septum programmable injector (SPI) held at 80 C for 0.5 min, followed by temperature programming to 290 C at 100 C min 1 and held at this temperature for 22 min. The column was held at 80 C for 2.6 min followed by temperature programming at 25 C min 1 to 290 C and held at this temperature for 14 min. The mass spectrometer and transfer line were held at 220 and 260 C, respectively. The mass spectrometer was operated in full scan mode with a range from 50 to 350 m/z. Water samples were analysed for dissolved oxygen using an Oxi 320 dissolved oxygen meter. The probe of the dissolved oxygen meter was incorporated within a low-volume (2 ml) flow cell, to reduce the volume of water required for analysis. Approximately 10– 20 ml of water from a sampling port was allowed to flow through the flow cell before the flow was ceased and the dissolved oxygen concentration recorded. Water samples were also analysed for atrazine-degrading bacteria using the technique described by Jayachandran et al. (1998).
206
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
Bromide was analysed by HPLC with a Waters IC-Pak Anion (4.6 mm 5 cm) using UV detection at 200 nm. The eluent (a solution of 10 mM NaCl and 5 mM NaH2PO4 H2O, buffered to pH 6.2 with Na2HPO4) was pumped through the column at a flow rate of 1.0 ml min 1. The volume injected into the HPLC was 15 ml of the filtered groundwater.
5. Results and discussion 5.1. General trends Ratios of groundwater concentrations before the oxygen delivery mat (port B) and immediately after the mat (port C) are plotted for atrazine, terbutryn and fenamiphos for the three columns in Fig. 3. For terbutryn and fenamiphos, no significant differences in concentration ratios were observed between the control column (column 3) and columns 1 and 2, indicating that the addition of oxygen or oxygen and acetate did not enhance the removal of terbutryn or fenamiphos under these experimental conditions and time frames. Concentration ratios of terbutryn for all columns increased from approximately 0.6 to 1.0 during the course of the experiment. This phenomenon was most likely due to sorption of terbutryn to the polymer tubing during the experiment. Towards the end of the experiment, equilibrium partitioning may have been established between the polymer tubing and water phase, resulting in ratios close to 1.0. Also, ratios of terbutryn concentrations at port B to concentrations at port A (the two ports up-gradient of the mat) show consistent values of around 1.0 throughout the experiment for all three columns (data not shown), suggesting little sorption between sampling ports when polymer mats were not present. Strong partitioning of selected organic compounds to silicone has previously been reported (Patterson et al., 2000). The degree of partitioning was shown to be related to the organic compound octanol –water partitioning coefficient (Kow). Terbutryn has a relatively high Kow value of 5500, which would suggest strong partitioning to the silicone tubing mat. In addition, the Kow values for atrazine and fenamiphos are significantly lower than terbutryn (see Table 2), so changes in concentration ratios for these compounds due to sorption onto the polymer mat would be less likely. Rapid decreases in atrazine concentrations were observed in both columns where air was purged through the gas delivery mats (columns 1 and 2), and only in the region close to the polymer mats. No atrazine decrease was observed in the non-purged column (column 3). Removal of atrazine from the water in columns 1 and 2 may have resulted from either biological degradation or physical removal by diffusion of atrazine through the polymer tube and then volatilising into the purged gas inside the tubing. Physical removal due to volatilisation seemed unlikely due to the low Henry’s Law constant of atrazine (2.48 10 7 kPa m3 mol 1). Assuming equilibrium between the purged gas and the groundwater, based on the Henry’s Law constant of atrazine, the total mass flux of atrazine
Fig. 3. Plots of ratios of atrazine, terbutryn and fenamiphos concentrations before the oxygen delivery mat (port B) and immediately after the mat (port C) for columns 1, 2 and 3.
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
207
208
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
that could be removed through purging would be significantly less than 1% of the mass flux of atrazine in the groundwater flowing past the mat. Note, even though there is a polymer between the water and gas phases, at equilibrium, the partitioning is still governed by the Henry’s Law constant (Patterson et al., 2000). Also, since all three pesticides had similar Henry’s Law constants, preferential atrazine removal by volatilisation seems unlikely. Hydroxyatrazine concentrations in groundwater down-gradient of the polymer mats were measured during the period of oxygen delivery, using the HPLC method of Vermeulen et al. (1982). No hydroxyatrazine was detected ( < 1 mg l 1) in the groundwater, indicating that either hydroxyatrazine was not formed or hydroxyatrazine degradation rates were rapid enough that accumulation of hydroxyatrazine did not occur. To confirm that the removal of atrazine was not due to physical stripping, purging of the mats with air was halted and the mats were then purged with nitrogen gas. Once nitrogen purging commenced, dissolved oxygen concentrations decreased from approximately 4 mg l 1 to less than 0.2 mg l 1 at port C for both columns. There was also a corresponding increase in atrazine concentrations relative to (influent) concentrations at port B. These changes in dissolved oxygen and atrazine ratios are shown for column 2 in Fig. 4. These data demonstrate that atrazine removal was through oxygen-mediated biotic or abiotic processes. Microcosm experiments (Franzmann et al., 2000) have shown no atrazine mineralisation in soils that have been taken from this aquifer and sterilised. In addition, atrazine-degrading bacteria have been detected in these columns; therefore,
Fig. 4. Changes in dissolved oxygen (port C) and atrazine ratios in column 2 before and after air purging of the polymer mat halted and nitrogen purging commenced.
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
209
biodegradation by oxygen-requiring microorganisms seems the most plausible mechanism for atrazine removal. Once oxygen delivery was halted, it is assumed that the microorganisms stopped metabolising the hydrolytic enzymes required for atrazine degradation, and the enzymes already present lost their activity over a period of 1 – 3 days. 5.2. Soil porosity Soil porosities of each column were estimated from their bromide breakthrough curve (see Fig. 5) and using Eq. (18). While bromide peak heights/areas varied between columns, estimated breakthrough times were similar, especially for columns 2 and 3. Estimated soil porosities for all three columns were similar with an average porosity of 0.33 F 0.02 (error = standard deviation). These porosities were similar to the estimated field porosity of 0.37 F 0.02, determined from recovered intact cores. 5.3. Biodegradation rates Since degradation of atrazine occurred rapidly and only in the region close to the polymer mats, it was difficult to determine if degradation was zero order or first order; therefore, degradation rates of the pesticides were calculated as both zero-order and firstorder degradation rates. First-order rates are expressed as half-lives. Zero-order biodegradation rates were calculated from the difference in concentrations of the pesticides in groundwater collected from ports B and C [up-gradient and down-
Fig. 5. Bromide breakthrough at port G, 182 cm from the base of the column, after bromide injection into solution delivery ports at 48 cm from the base of the column.
210
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
gradient of the oxygen delivery mat using Eq. (16)]. The times required for groundwater to migrate from port B to C were estimated from Eq. (17) using the soil porosity and the volumetric flow rate through the column of 706 ml day 1. The estimated travel times between ports B and C were 1.5 days for columns 1 and 3, and 1.6 days for column 2. Degradation rates between ports B and C were estimated during stable periods of air purging of the mats and nitrogen gas purging of the mats. For the air purging, degradation rates were estimated from average concentrations of the pesticides between the 5th and 22nd June 1998 (n = 7). During this time, concentrations were stable for both ports B and C. During nitrogen gas purging, degradation rates were estimated from average concentrations between the 26th June and 13th July 1998 (n = 5). Again, during this time, concentrations were stable for both ports B and C. Zero-order biodegradation rates estimated for both these periods are given in Table 3. When the groundwater was oxygenated, there was a rapid decrease in atrazine concentrations with degradation rates of 380 and 240 mg l 1 day 1 for columns 1 and 2. These rates are possibly an underestimate, since degradation may be negligible in the time groundwater takes to migrate between port B and the oxygen delivery mat. Also, the degradation may only occur in a narrow zone at the mat or down-gradient from the mat. While degradation rates were different between the columns with and without acetate, this difference is most likely due to differences in the initial atrazine concentrations at port B in columns 1 and 2. The average concentrations were 600 F 70 and 370 F 50 mg l 1 (errors are the standard deviation). Estimated half-lives, based on average concentrations between the 5th and 22nd June 1998 for ports B, C and D, are given in Table 4. Atrazine half-lives for oxygenated conditions were 0.35 F 0.04 and 0.35 F 0.07 days for columns 1 and 2. These data would suggest that the naturally occurring dissolved organic carbon in the groundwater (8 mg l 1) might be sufficient for bacteria to degrade atrazine. These degradation rates compare favourably with a half-life of 0.26 F 0.2 days, estimated from culture experiments using citrate as a carbon source (Tilbury, 1998).
5.4. Oxygen delivery Based on Eq. (14) with an air-flow rate of 40 ml min 1 and Cg/Co = 0.9, 10% (1.12 mg min 1) of the total oxygen influx through the inner space of tubing should diffuse through
Table 3 Estimated zero-order degradation rates for atrazine, terbutryn and fenamiphos Zero-order degradation rate (mg l 1 day 1) Atrazine
Oxygen + acetate Oxygen only Control, no oxygen
Terbutryn
Fenamiphos
Oxygen purge
Nitrogen purge
No purge
Oxygen purge
Nitrogen purge
No purge
Oxygen purge
Nitrogen purge
No purge
380 240 –
< 10 < 10 –
– – < 10
< 10 < 10 –
< 10 < 10 –
– – < 10
< 10 < 10 –
< 10 < 10 –
– – < 10
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
211
Table 4 Estimated half-lives for atrazine, terbutryn and fenamiphos Half-life (days) Atrazine Oxygen purge Oxygen + acetate 0.35 F 0.04 Oxygen only 0.35 F 0.07 Control, no oxygen – Culturea 0.26 F 0.02
Terbutryn
Fenamiphos
Nitrogen No Oxygen Nitrogen No Oxygen Nitrogen No purge purge purge purge purge purge purge purge > 20 > 20 – –
– – > 20 –
> 20 > 20 – –
> 20 > 20 – –
– – > 20 –
> 20 > 20 – –
> 20 > 20 – –
– – > 20 –
Errors are standard errors. a Tilbury (1998).
the polymer and be instantaneously consumed at the outer surface of the polymer tubing, or transported past the mat with groundwater flow. If all the oxygen that diffuses through the polymer is transported into the groundwater flowing past the polymer mat, the mass flux rate of oxygen (1.12 mg min 1) should be sufficient to saturate the anoxic groundwater flowing past the mat (4.9 10 4 l min 1) with oxygen to give an oxygen concentration in the groundwater of approximately 8.7 mg l 1. Oxygen concentrations measured in the groundwater at port C (11 cm down-gradient from the polymer mats) in columns 1 and 2 were approximately 4 mg l 1. The lower oxygen concentration suggests that there has been oxygen consumption in the groundwater between the contact time with the polymer mat and the sampling time at port C, a time difference of approximately 20 h. As significant atrazine degradation occurred within this zone, some oxygen consumption would appear likely. Another possible explanation for the lower oxygen concentrations may be that the delivery of oxygen to groundwater was less efficient than predicted. Possible reasons could be mass transfer limitations at the polymer – water interface, or insufficient polymer – water surface area resulting in a mixture of oxygen-saturated groundwater (groundwater in contact with the polymer) and anoxic groundwater (groundwater that did not make contact with the polymer). Correspondence in atrazine half-lives between microcosm and column experiments suggests oxygen did not limit degradation.
6. Conclusions The column experiments have demonstrated that bacteria in the pesticide-contaminated groundwater from the Dianella site were capable of degrading atrazine under aerobic conditions. Degradation rates of atrazine were relatively high with a zero-order rate of 240– 380 mg l 1 day 1 or half-life of 0.35 days. Amendment with an additional carbon source showed no significant improvement in biodegradation rates, suggesting that organic carbon was not limiting biodegradation. Atrazine degradation rates estimated in the column experiments were similar to rates determined in laboratory culture experiments.
212
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
No significant degradation of terbutryn or fenamiphos was observed under the experimental conditions and time frames studied. An alternative remediation strategy may be needed for these compounds. The polymer tubes constructed into mats and purged with air provided an efficient method to deliver dissolved oxygen to anaerobic groundwater. Dissolved oxygen concentrations increased from < 0.2 mg l 1 to approximately 4 mg l 1. The use of polymer tube mats may provide a low-infrastructure, long-term field technique for the delivery of oxygen to groundwater. In comparison to alternative oxygen delivery techniques, such as air sparging or peroxide amendment, this technique may be more cost-effective and reliable, although this aspect has not been assessed here. Results from these experiments indicate that remediation of atrazine at the Dianella site may be achieved through oxygenation of the anaerobic contaminated zone of the aquifer. No significant degradation of terbutryn or fenamiphos was observed under the experimental conditions within the time frames of the study. While bacteria capable of degrading atrazine have been isolated from the central portion of the pesticide plume, such bacteria have not been isolated from groundwater at the leading edge of the pesticide plume. Therefore, if remediation is desirable at the leading edge of the pesticide plume, augmentation of this zone of contaminated groundwater with atrazine-degrading bacteria may also be required.
Acknowledgements The authors thank David Briegel, Terry Power, Nathan Innes and Dean Coad for their technical assistance. Thanks also to Dr. C. Barber and Dr. P. Clement for critically reviewing the manuscript. This work was partly funded by the Water and Rivers Commission of Western Australia through the Centre for Groundwater Studies.
References Agertved, J., Rugge, K., Baker, J.K., 1992. Transformation of the herbicide MCPP and atrazine under natural aquifer conditions. Ground Water 30 (4), 500 – 506. Appleyard, S.J., 1995. Investigation of ground water contamination by fenamiphos and atrazine in a residential area: source and distribution of contaminants. Ground Water Monit. Rev. 15 (4), 110 – 113. Blume, E., Bischoff, M., Moorman, T., Turco, R.F., 1993. Assimilative capacity of subsurface soils for atrazine. Agron. Abstr. 83, 241. Blumhorst, M.R., Weber, J.B., 1994. Chemical versus microbial degradation of cyanazine and atrazine in soils. Pestic. Sci. 42, 79 – 84. Charati, S.G., Stern, S.A., 1998. Diffusion of gases in silicone polymers: molecular dynamics simulations. Macromolecules 31, 5529 – 5535. Ciba-Geigy data, 1989. Crank, J., 1975. The Mathematics of Diffusion, 2nd edn. Clarendon Press, Oxford. Davis, G.B., Johnston, C.D., Patterson, B.M., Barber, C., Bennett, M., 1998. Estimation of biodegradation rates using respiration tests during in situ bioremediation of weathered diesel NAPL. Ground Water Monit. Rem. 18 (2), 123 – 132. De Souza, M.L., Wackett, L.P., Sadowsky, M.J., 1998. The atzABC genes encoding atrazine catabolism are
B.M. Patterson et al. / Journal of Contaminant Hydrology 54 (2002) 195–213
213
located on a self-transmissible plasmid in Pseudomonas sp. strain ADP. Appl. Environ. Microbiol. 64, 2323 – 2326. Franzmann, P.D., Zappia, L., Patterson, B.M., Rayner, J.R., Davis, G.B., 1998. Mineralization of low concentrations of organic compounds and microbial biomass in vadose zone soils from Perth Coastal Plain. Aust. J. Soil Res. 36, 921 – 939. Franzmann, P.D., Zappia, L.R., Tilbury, A.L., Patterson, B.M., Davis, G.B., Mandelbaum, R.T., 2000. Bioaugmentation of atrazine and fenamiphos impacted groundwater: laboratory evaluation. Bioremediation J. 4 (3), 237 – 248. Gibson, T.L., Abdul, A.S., Chalmer, P.D., 1998. Enhancement of in situ bioremediation of BTEX-contaminated ground water by oxygen diffusion from silicone tubing. Ground Water Monit. Rem. 18 (1), 93 – 104. Herrling, B., Stamm, J., Alesi, E.J., Bott-Breuning, G., Diekmann, S., 1994. In situ bioremediation of groundwater containing hydrocarbons, pesticides, or nitrate using vertical circulation flows (UVB/GZB technique). In: Hinchee, R.E. (Ed.), Air Sparging for Site Remediation. Lewis Publishers, Boca Raton, pp. 56 – 80. Jayachandran, K., Stolpe, N.B., Moorman, T.B., Shea, P.J., 1998. Application of 14C-most-probable-number technique to enumerate atrazine-degrading microorganisms in soil. Soil Biol. Biochem. 30 (4), 523 – 529. Johnston, C.D., Rayner, J.L., Patterson, B.M., Davis, G.B., 1998. The contribution of volatilisation and biodegradation during air sparging of dissolved BTEX-contaminated groundwater. J. Contam. Hydrol. 33, 377 – 404. Kookana, R.S., Phang, C., Aylmore, L.A.G., 1997. Transformation and degradation of fenamiphos nematicide and its metabolites in soils. Aust. J. Soil Res. 35, 753 – 761. Lopez-Avila, V., Hirata, P., Kraska, S., Flanagan, M., Taylor Jr., J.H., Hern, S.C., Melancon, S., Pollard, J., 1986. Movement of selected pesticides and herbicides through columns of sandy loam. In: Garner, W.Y., Honeycutt, R.C., Nigg, H.N. (Eds.), Evaluation of Pesticides in Ground Water. American Chemical Society, Washington, DC, ACS Symposium Series 315, 311 – 328. Miles data, Agricultural Chemicals Division, 8400 Hawthorn Road, Kansas City, MO 64120. Mirgain, I., Green, G., Monteil, H., 1995. Biodegradation of the herbicide atrazine in groundwater under laboratory conditions. Environ. Technol. 16, 967 – 976. Nair, D.R., Schnoor, J.L., 1992. Effect of two electron acceptors on atrazine mineralization rates in soil. Environ. Sci. Technol. 26, 2298 – 2300. Ou, L.-T., Thomas, J.E., Dickson, D.W., 1994. Degradation of fenamiphos in soil with a history of continuous fenamiphos application. Soil Sci. Soc. Am. J. 58, 1139 – 1147. Patterson, B.M., Power, T.R., Barber, C., 1993. Comparison of two integrated methods for the collection and analysis of volatile organic compounds in groundwater. Ground Water Monit. Rem. 13 (3), 118 – 123. Patterson, B.M., Franzmann, P.D., Rayner, J., Davis, G.B., 1998. Mobility and biodegradation of low-levels of organic compounds in the vadose zone of the Swan coastal plain. Conference paper for the IAH International Groundwater Conference, Melbourne, Australia, February 1998, pp. 533 – 538. Patterson, B.M., Davis, G.B., McKinley, A.J., 2000. Volatile organic compounds in groundwater, probes for the analysis of. In: Meyers, R.A., et al., (Ed.), Encyclopedia of Analytical Chemistry: Instrumentation and Application. Wiley, Chichester, pp. 3515 – 3526. Schomburg, G., 1990. Gas Chromatography: A Practical Course. VCH Publishers, New York, p. 298. Tilbury, A., 1998. Hons. Thesis, Department of Chemistry, The University of Western Australia. van Genuchten, M.T., Alves, W.J., 1982. Analytical solutions of the one-dimensional convective – dispersive solute transport equation. U.S. Dep. Agric., Tech. Bull. No. 1661, 151 pp. Vermeulen, N.M.J., Apostolides, Z., Potgieter, D.J.J., Nel, P.C., Smit, N.S.H., 1982. Separation of atrazine and some of its degradation products by high performance liquid chromatography. J. Chromatogr. 240, 247 – 253. Weber, J.B., 1994. Properties and behavior of pesticides in soil. In: Honeycutt, R.C., Schabacker, D.J. (Eds.), Mechanisms of Pesticide Movement into Ground Water. Lewis Publishers, Boca Raton, pp. 15 – 41. Yamagiwa, K., Yoshida, M., Ito, A., Ohkawa, A., 1998. A new oxygen supply method for simultaneous organic carbon removal and nitrification by a one-stage biofilm process. Water Sci. Technol. 37 (4 – 5), 117 – 124.