Environ. Sci. Technol. 2005, 39, 7279-7286
Field Study of Pulsed Air Sparging for Remediation of Petroleum Hydrocarbon Contaminated Soil and Groundwater X I A O M I N Y A N G , * ,† D E N N I S B E C K M A N N , † STEPHANIE FIORENZA,† AND CRAIG NIEDERMEIER‡ BP p.l.c., 28100 Torch Parkway, Warrenville, Illinois 60555, and The RETEC Group Inc., Austin, Texas.
Recent laboratory-scale studies strongly suggested an advantage to operating air-sparging systems in a pulsed mode; however, little definitive field data existed to support the laboratory-scale observations. This study aimed to evaluate the performance of a field-scale pulsed air-sparging system during a short-term pilot test and during longterm system operation. The air-sparging system consisted of 32 sparging points and had been previously operated in a continuous mode for two years before the field study was performed. The field study used instruments with continuous data logging capabilities to monitor the dynamic responses of groundwater and soil vapor parameters to air injection. The optimum pulsing frequency was based on the evidence that the hydrocarbon volatilization and oxygen dissolution rates dramatically dropped after the airsparging system reached steady state. The short-term pilot test results indicated a substantial increase in hydrocarbon volatilization and biodegradation in pulsed operation. On the basis of the results of the pilot test, the airsparging system was set to operate in a pulsed mode at an optimum pulsing frequency. Operation parameters were collected 2, 8, and 12 months after the start of the pulsed operation. The long-term monitoring results showed that the pulsed operation increased the average hydrocarbon removal rate (kg/day) by a factor of up to 3 as compared to the previous continuous operation. The pulsed air sparging has resulted in higher reduction rates of dissolved benzene, toluene, ethylbenzene, and xylenes (BTEX) than were observed during the continuous operation. Among BTEX, benzene’s reduction rate was the highest during the pulsed air-sparging operation.
Introduction In-situ air sparging (IAS) is a remediation technology that involves the injection of air into the saturated subsurface to treat contaminants dissolved in the groundwater and trapped in soil pores in the saturation and capillary zones. IAS has been rapidly adopted as a remedy to treat immiscible-phase source zones and dissolved contaminant plumes and as a barrier to contain dissolved plumes (1-4). IAS also has significant potential for remediating very soluble, but slowly * Corresponding author phone: (630)836-7176; fax: (630)836-7193; e-mail:
[email protected]. † Atlantic Richfield Company (A BP affiliated company). ‡ The RETEC Group Inc. 10.1021/es050084h CCC: $30.25 Published on Web 08/16/2005
2005 American Chemical Society
degrading fuel oxygenates, such as MTBE (5). Reports from field applications indicated that air sparging was very effective at some sites and less effective at other sites (6). A recent field application review found that almost half of the reviewed air-sparging sites achieved a 95% permanent reduction in groundwater contaminant concentrations, and the mass removal rates dramatically dropped over the lifetime of the air-sparging projects (7). Recent laboratory-scale physical model studies have shown significant improvement of contaminant volatilization when air-sparging systems were operated in a pulsed mode (6, 8-10). For instance, pulsed air sparging was evaluated for the remediation of tetrachloroethylene (PCE) present as dense nonaqueous phase liquid (DNAPL) in a two-dimensional laboratory tank (10). The test results demonstrated that pulsed operation accelerated the time-weighted average PCE volatilization by 40-600%, depending on the aggressiveness of the pulsing. Another laboratory-scale study was conducted using a two-dimensional aquifer physical model to quantitatively assess the effects of pulsed air injection on the rate and the extent of petroleum hydrocarbon removal by IAS (6). The study results suggested that pulsed air sparging enhanced the hydrocarbon removal rate by 66%. In another laboratory-scale study, the effect of pulsed aeration on bioreaction rate in a 12-L solid-state bioreactor was investigated (11). The study results demonstrated that the pulsed aeration greatly enhanced oxygen mass transfer and moisture volatilization, resulting in a much higher bioreaction rate than direct aeration. Several mechanisms have been postulated to explain why pulsed operation improves contaminant removal by IAS. Induced groundwater flow and groundwater mixing might be the two dominating mechanisms. As air is introduced into a formation, air displaces groundwater in the largest pores and creates temporary groundwater flow around sparging wells (12). When an air-sparging system achieves steady state, preferential air flow pathways consisting of the largest network of pores are formed in the formation (13), and the induced local groundwater flow ceases (14). The contaminant mass removal and oxygen dissolution rates become limited by contaminant transport from the bulk of the groundwater to the air/water interface at the air flow pathways and oxygen transport in the reverse direction (presumably diffusion dominated) (5). Pulsing the air injection frequently creates non-steady-state conditions and induces groundwater circulation as the air channels form and collapse during each cycle. Chemical mass transfer coefficients in the groundwater are proportional to the groundwater flow velocity (15). The induced groundwater flow created by the pulsed air sparging substantially enhances the contaminant and oxygen mass transfer coefficients in the groundwater (10). Second, contaminants in the immediate vicinity of air channels can be removed within hours to a couple of days of the start of sparging (16), but contaminated groundwater present at a greater distance from the air channels is less treated because of the limited mass transport (i.e., diffusion) in the groundwater. During pulsed air sparging, the lesstreated water flows into the air channels and mixes with the treated water when the air injection is temporarily turned off. This mixing occurs in both microscale in air channels and in macroscale in sparing zones (10). This mixing created by the pulsed operation introduces contaminated water to the vicinity of the air pathways as the channels form and collapse during each cycle (16). The groundwater mixing reduces the degree to which chemical mass transport in VOL. 39, NO. 18, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7279
FIGURE 1. Air-sparging system layout and groundwater monitoring network (a) and stratigraphic cross section from A to A′ (b) of the gas production site in Kalkaska County, Michigan. groundwater governs contaminant removal, resulting in an increase in contaminant mass removal by pulsed air sparging (3). While the recent laboratory-scale studies and theoretical analysis strongly suggested an advantage to operating airsparging systems in a pulsed mode, little definitive field data existed to support the laboratory-scale observations (17, 18). This study aimed to evaluate the performance of a fieldscale pulsed air-sparging system during a short-term pilot test and during long-term system operation. The valuable knowledge gained from the laboratory-scale studies by other researchers was integrated into the design of this field study. The objectives of this field study included (1) measuring the responses of groundwater and soil vapor parameters to air injection; (2) developing a method to determine an optimum pulsing frequency; (3) evaluating the mass of petroleum hydrocarbon volatilized and biodegraded during the pulsed operation and comparing it to the previous continuous operation; and (4) monitoring long-term performance of the pulsed air-sparging system. 7280
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 18, 2005
Material and Methods Site and System Description. The site is located in Kalkaska County, Michigan. An environmental investigation was initiated in 1995 following the decommissioning of a production well at the site. At shallow depths below the site, sediments consist of glacial deposits of medium sands grading into medium-to-coarse sands at a depth of approximately 4.6 m below ground surface (bgs). This medium-to-coarse sand unit extends to a depth of at least 10.7 m bgs (Figure 1b). A shallow aquifer exists at a depth of approximately 4.6-9.1 m bgs within the sand unit. Groundwater within the shallow aquifer flows in a southeast direction with an estimated hydraulic gradient of 0.001 m/meter. Twenty-two shallow groundwater monitoring wells were installed as part of the initial investigation and subsequent investigations. An air-sparging system consisting of 32 sparging wells was installed in 2001 to treat shallow groundwater impacted with BTEX. The sparging wells were screened approximately 4.6 m below groundwater with a screen length of 0.76 m. Part of the sparging system layout is provided in Figure 1a. The sparging system was designed to inject air at
TABLE 1. Monitoring Equipments for Pulsed Air-Sparging Field Study instrument in-situ multiparameter troll
working range and sensitivity
parameters measured
groundwater pressure DO 9000 LTS temperature pressure transducers groundwater pressure multiRAE four-gas monitor hydrocarbon oxygen GasTECH portable gas monitor carbon dioxide
0-30 psig, (0.1% 0-20 mg/L, (0.2 mg/L -5 to 50 °C, (0.1 °C 0-30 psig, (0.1% 0-3000 ppmv, (1 ppmv 0-30%, (0.1% 0-5%, (10% of reading
monitoring locations MW-2, MW-3 MW-5, MW-6, MW-8, MW-14 VMP-1 Deep, VMP-2 Deep, VMP-2 Shallow VMP-1 Deep, VMP-2 Deep, VMP-2 Shallow
FIGURE 2. Responses of groundwater pressure and DO (a) and hydrocarbon concentration (b) in the vadose zone to air injection. 113 L/min per sparging well. Twenty-two groundwater monitoring wells were used to monitor the performance of the sparging system along with five vapor monitoring points (VMPs). Each VMP consisted of a shallow and deep sampling location. The sampling locations of the shallow VMPs are 4 m above the groundwater table, while the sampling locations of the deep VMPs are 0.9 m above the groundwater table. The sparging system was operated in a continuous mode from May 2001 to September 2003. Groundwater analytical data showed that BTEX concentrations at five key groundwater monitoring wells (MW-1, 2, 3, 5, and 8) were orders of magnitude higher than the site clean up criteria: benzene (0.005 mg/L), toluene (0.790 mg/L), ethylbenzene (0.074 mg/ L), and xylenes (0.280 mg/L). A thin layer of light nonaqueous phase liquid (LNAPL) existed in MW-1 and MW-2 (measured thickness of 9 mm in the April 2003 sampling event) during the continuous air sparging. The BTEX concentrations at the five key monitoring wells except MW-3 did not substantially decrease after 2 years of the continuous air-sparging operation. Groundwater and Soil Vapor Parameters Measured. During the pilot test, groundwater pressures and dissolved oxygen (DO) were measured in the selected groundwater monitoring wells (circles in Figure 1a). Table 1 presents a list of the monitoring locations and the field measurement equipment used in the study, including the range and sensitivity of each piece of equipment. Volatile hydrocarbon, carbon dioxide, and oxygen concentrations were measured from selected deep and shallow VMPs (boxes in Figure 1a). Data collected by the monitoring equipment was logged every 30 s during the field study. Selected groundwater and soil vapor parameters were measured during the long-term operation. Design of the Field Study. Initially, 24-h baseline measurements of the soil vapor and groundwater parameters listed in Table 1 were conducted in September 2003 while the air sparging system was operated in the continuous mode. Then, the air sparging was shutdown for 24 h. Continuous measurements of those parameters were conducted until the trapped air within the saturated zone was fully released. After that, the system was restarted and the groundwater and soil vapor parameters were monitored continuously until
FIGURE 3. Groundwater pressure changes after the air injection terminated. the groundwater/air flow reached steady state. Steady-state conditions were determined in the field by analyzing the changes of groundwater pressures and DO and soil vapor hydrocarbon concentrations. After that, a pulsing frequency was selected that would maximize oxygen dissolution and hydrocarbon volatilization. The sparging system was restarted and was set to operate at the chosen pulsing frequency. The groundwater and soil vapor parameters were continuously monitored for 72 h to evaluate the mass of hydrocarbon volatilized and biodegraded and to verify that the optimum pulsing frequency was chosen. The air-sparging system was set to operate in the pulsed mode after the pilot test was completed. Selected operational parameters were measured 2, 8, and 12 months after the start of the pulsed air sparging to evaluate the long-term performance.
Results and Discussion Responses of Groundwater Pressure, DO, and Soil Vapor Hydrocarbon Concentration to Air Injection. Figure 2 shows the responses to the air injection of groundwater pressure and DO in MW-2 and soil vapor hydrocarbon concentration in VMP-1 Deep. The responses in other measured monitoring wells had similar patterns. When air was introduced into the formation, the groundwater pressure increased sharply, reaching a peak in 15 min. The maximum increase in groundwater pressure was 12 cm H2O. The water pressure then tapered off and achieved steady state after approximately 4 h. The trend of the groundwater DO changes was almost VOL. 39, NO. 18, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7281
FIGURE 4. Groundwater fluctuation during pulsed air sparging. identical to the groundwater pressure. DO reached a peak of 2.4 mg/L in the first hour of the air injection and then slowly declined to 0.5 mg/L after 7 h, suggesting that the oxygen mass transport decreased. Similar groundwater pressure and DO response curves were observed in Pixie Newman’s field air-sparging tests (19). There was a 2-h lag of soil vapor hydrocarbon concentration changes after the air injection, but the lag time decreased to half an hour after three pulsing cycles. The initial hydrocarbon concentration at VMP-1 Deep was 20 ppmv before the air injection. It increased to a maximum concentration of 800 ppmv in 3 h after the lag time and then dropped gradually. Groundwater pressure dropped immediately after the termination of air injection (Figure 3). The groundwater pressure decreased 8 cm H2O in MW-2. The collapse of air channels in the formation initially created a groundwater cone of depression around each air-sparging well. Groundwater gradually flowed into the sparging zone and displaced the residual air in the aquifer resulting in an increase of groundwater levels until reaching steady state. Air pressures measured in the soil vapor monitoring points slowly vanished after the air injection, indicating the gradual release of the residual air. Groundwater levels recovered after about 4 h. A pulsing cycle consists of an on period and an off period. A 4-h on period was chosen to maximize the average DO and hydrocarbon concentrations in the on period. The biodegradation and volatilization still occurred to some degree when the residual air in the groundwater was gradually released into the vadose zone after the air injection. A 3-h off period was first chosen for the pilot test, but the test results suggested a longer off period because the groundwater table probably did not completely recover in 3 h. The off period was changed to 4 h after the pilot test. A 4-h-on-4-h-off pulsing cycle was permanently applied to the sparging system in September 2003. Water Pressure Fluctuation during Pulsed Air Sparging. As shown in Figure 4, pulsed air sparging created water pressure fluctuation in the formation. The air flow rate was increased from 113 to 155 L/min temporarily during the pilot test, and a bigger pressure fluctuation was detected. The phenomenon might result from higher air channel volumes at the higher injection rate (2). Pulsing the air injection induces back-and-forth groundwater flow in the formation, which increases the oxygen and hydrocarbon mass transfer rates in the groundwater and promotes groundwater mixing; therefore, hydrocarbon volatilization and degradation should improve during the pulsed air sparging. The Effects of Pulsed Operation on Hydrocarbon Volatilization. Injected air carried the volatilized hydrocarbon into the vadose zone. Air pressure in the deep monitoring point VMP-2 Deep was manually monitored to check whether injected air continuously flowed through the vapor sampling point. Table 2 lists the measured air pressure values during 7282
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 18, 2005
TABLE 2. Air Pressure in VMP-2 Deep during an On Period elapsed time (hh:mm) 0:00 0:02 0:03 0:06 0:08 0:19 0:32 0:42 0:56 4:32
air pressure (cm H2O) 0.00 0.38 0.64 1.14 1.27 1.78 1.91 1.91 1.91 2.03
an on period. After the initial dynamic stage (less than 30 min), the air pressure in VMP-2 Deep remained constant, suggesting continuous and steady air flux in the unsaturated zone. The vadose zone hydrocarbon concentrations just above the water table were monitored to confirm the transfer of hydrocarbons from the saturated zone to the vadose zone and to measure the extent of the hydrocarbon volatilization. As shown in Figure 5, hydrocarbon concentrations were 38 and 20 ppmv at VMP-1 Deep and VMP-2 Deep, respectively, during the continuous air sparging when the baseline measurement was performed in September 2003. Because the hydrocarbon concentration fluctuated during pulsed air sparging, the average hydrocarbon concentration, which was calculated by averaging all the data points in a 72-h monitoring event, was used to indicate the extent of hydrocarbon volatilization. The DO, carbon dioxide concentration, and oxygen concentration were averaged the same way. The average hydrocarbon concentrations increased to 1104 and 738 ppmv at VMP-1 Deep and VMP-2 Deep, respectively, during the pulsed air sparging pilot test in September 2003. The thin layer of LNAPL hydrocarbons, which had existed in MW-1 and MW-2 during the continuous air-sparging operation, was not detected when the groundwater was sampled in October 2003, 45 days after the start of the pulsed air sparging. The average hydrocarbon concentration dropped to 530 ppmv at VMP-1 Deep in November 2003, rose to 919 ppmv in May 2004, but dropped again in September 2004 to 371 ppmv. In November 2003, the vapor monitoring equipment at VMP-2 Deep malfunctioned; in May 2004, the average hydrocarbon concentration was 271 ppmv and then dropped to 30 ppmv in September 2004. Operating the air-sparging system in a pulsing mode significantly boosted the hydrocarbon volatilization rate. The concentrations at those locations dramatically dropped in September 2004 as the mass of remaining hydrocarbons decreased. Effects of Pulsed Air Sparging on Hydrocarbon Biodegradation. The pulsed operation promoted oxygen dis-
FIGURE 5. Soil vapor hydrocarbon concentrations during continuous and pulsed air sparging.
FIGURE 6. Dissolved oxygen at MW-2 during continuous and pulsed air sparging. solution into the groundwater. Figure 6 shows the groundwater DO in MW-2. The DO was only 0.3 mg/L during the continuous air-sparging baseline measurement, presumably because air channels developed, limiting oxygen mass transfer to the groundwater. The pulsed air sparging significantly elevated the DO. During the on periods of pulsed air sparging, DO increased and reached a peak value of 5.8 mg/L in May 2004. The dissolved oxygen dissipated rapidly in the off periods likely as a result of biological activity in the groundwater. Also shown in Figure 6 is the fluctuation of DO in each pulsing cycle. The average DO in MW-2 was 0.95, 1.6, and 2.2 mg/L in September 2003, November 2003, and May 2004, respectively. The steady increase of DO suggested that the pulsed air sparging sustainedly oxygenated groundwater. Carbon dioxide and oxygen concentrations in deep-soil vapor-sampling points were monitored to measure the rate of the hydrocarbon biodegradation. Carbon dioxide and
oxygen concentrations were 0.24% and 21.0%, respectively, in VMP-1 Deep during the continuous air-sparging baseline measurement. In the pulsed operation, carbon dioxide concentrations measured in VMP-1 Deep and VMP-2 Deep increased to 0.96% and 1.42%, respectively, in November 2003 (Figure 7), while the corresponding oxygen concentrations dropped to 18.7% and 18.3%, respectively. The pulsed air sparging resulted in higher dissolved oxygen in the groundwater, therefore stimulating biological activities in the formation. In May 2004, the carbon dioxide concentrations dropped to 0.52%, and oxygen concentrations increased to 20.7%. The site had snow cover until late April in 2004. Biological activity presumably decreased while the temperature was lower in the long winter. Another monitoring event was carried out in September 2004 to address the probable seasonal effects. The carbon dioxide concentrations increased to 0.62%, and oxygen concentrations decreased to 20.4%, VOL. 39, NO. 18, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7283
FIGURE 7. Soil vapor carbon dioxide and oxygen concentrations during continuous and pulsed air sparging. indicating a slight increase of biological activity in the summer 2004. The pulsed air-sparging operation reduced the hydrocarbon mass in the saturated zone, and consequently, the magnitude of hydrocarbon biodegradation decreased. The hydrocarbon mass loss from the site might be the main reason for the depressed CO2 and increased O2 in the vadose zone in 2004. Mass of Hydrocarbon Removed. The mass of the hydrocarbon volatilized and biodegraded in one pulsing cycle is estimated using the following equations.
10-6 m3 60 min 10-6 kg ‚ ‚ (1) L h mg
MVol h Vapor HC ) Qair‚Ton‚C HC ‚ Vapor MBio HC ) Qair‚Ton‚ (CCO2 -
10-6 m3 60 min 10-6 kg 78(C6H6) ‚ ‚ ‚ (2) L h mg 6 × 44(CO2)
FIGURE 8. Mass removal rates of continuous and pulsed air sparging.
Bio where MVol HC and MHC are the mass of hydrocarbon volatilized and biodegraded, respectively, in one pulsing cycle (kg); Qair is the air injection rate (L/min); Ton is the length of the on period; C h Vapor is the average hydrocarbon concentration HC Vapor 3 (mg/m ) in the deep vapor monitoring points; CCo is the 2 average soil vapor carbon dioxide concentration (mg/m3) in Background the deep vapor monitoring points; CCo is the carbon 2 dioxide concentration in the atmosphere (∼630 mg/m3); and the benzene molecular weight (78 g/mol) is used to convert the hydrocarbon concentration from ppmv to mg/m3 using the ideal gas law. The mass of hydrocarbons removed through volatilization and biodegradation was calculated during the on period only. Biodegradation and volatilization during the off period was not included in the removal calculations because of the
difficulty of measuring this in the field. Therefore, this mass estimation may underestimate the mass of hydrocarbon volatilized and biodegraded during the pulsed air sparging. The total hydrocarbon mass removal rate (kg/day) for pulsed air sparging is defined as the hydrocarbon mass removed during the on period divided by the length of a pulsing cycle (including both on and off periods). Only the hydrocarbon and carbon dioxide concentrations in VMP-1 and VMP-2 Deep were measured and then used to calculate the hydrocarbon mass removal rates in this study. So, the calculated values only represent the mass removal rates by air sparging around VMP-1 and VMP-2. Figure 8 depicts the volatilization, biodegradation, and total mass removal rates. Before the pulsed sparging field test was conducted, the total mass removal rate of the continuous sparging was 0.19 kg/day, and the biodegradation
Background CCO )‚ 2
7284
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 18, 2005
FIGURE 9. Benzene concentrations (a) and T + E + X concentrations (b) at key groundwater monitoring wells. Appendix A. Groundwater Quality Data in Key Monitoring Wells T + E + X (mg/L)
benzene (mg/L) sample date
MW-1
MW-2
MW-3
MW-5
10/26/1995 3/21/1996 8/24/1996 3/20/1997 6/22/1997 10/14/1997 4/6/1998 10/31/1998 5/1/1999 11/1/1999 4/30/2000 10/25/2000 10/18/2001 4/13/2002 11/3/2002 4/13/2003 10/18/2003 4/23/2004 10/22/2004
30.000
32.000 23.700 30.000 34.000 24.000 25.000 21.000 26.000
0.700 0.280 0.360 0.150 0.075 0.280 0.110 0.520 0.240 1.400 0.930
0.010
25.000 31.000 31.000 31.000 21.000
0.001 0.001 0.012 0.001 0.001 0.001 0.003 0.002 0.004
MW-8
0.001 0.001 0.001 0.001
MW-1
MW-2
MW-3
MW-5
MW-8
74.600
63.100 43.800 69.400 71.000 55.460 58.600 67.800 78.000
2.93 1.61 0.583 0.322 0.179 0.422 0.35 0.719 0.733 4.571 5.8
7.634
0.006
4.42 3.05 1.754 2.03 0.89 2.937 4.956 4.784 3.67
0.007 0.006 0.001 0.001 0.001 0.012 0.012 0.010 0.037 23.400
110.700 71.600 68.800 84.000 74.200
1.100
0.870 0.420 0.410
4.200 1.600 0.540
0.210 0.043 0.024 0.026 0.001 0.001 0.010
0.015 0.024 0.001 0.001 0.014
0.120 1.400 7.100 5.200 4.800 2.100
rate was 0.18 kg/day per sparging well around VMP-1 and VMP-2; so, 92% of the mass loss was attributable to biodegradation. When pulsed operation began in September 2003, the total hydrocarbon removal rate increased, by 3 times, to 0.58 kg/day per sparging well around VMP-1 and VMP-2 by November 2003. The percentage attributable to biodegradation changed from 53% in September up to 78% in November 2003. The role of biodegradation became relatively less important in the total mass removal in May 2004, and in September 2004, biodegradation accounted for 83% of the total mass removal as the volatilized hydrocarbon mass decreased probably because of the hydrocarbon mass loss from the site. Hydrocarbon mass removal by the air-sparging system significantly increased when the pulsed operation was applied in 2003. The hydrocarbon mass removal rate began to decline in May 2004 as the amount of remaining hydrocarbons decreased, but the total removal rate was still higher than it had been in continuous operation, which is calculated using the 2003 data. It is expected that the remediation time will be lessened as a result of the pulsed approach to air sparging. It is useful to note that the main objective of pulsed air sparing was to create non-steady-state air and groundwater flow in the formation, thereby enhancing chemical mass transport in groundwater and promoting hydrocarbon removal. Additionally, pulsed operation cut the electricity cost by 50%, and these savings on energy cost were just an added benefit of the technique. Groundwater Analytical Data. The groundwater monitoring wells continue to be sampled semiannually to monitor the long-term performance of the air-sparging system and
21.130 25.420 33.790
66.200 94.600 26.950
24.4 5.39 7.88 11.09 3.786 2.183 0.442
8.19 16.35 12.7 21.7 15.18
2.758 32.910 67.400 48.000 48.100 25.980
to determine when site objectives have been met. MW-1 and MW-2 are in the source area (Figure 1a). Benzene concentrations measured in these wells were approximately 34 mg/L before 1998 (Figure 9a). LNAPL was detected in 1999, and benzene was not analyzed when LNAPL was present. LNAPL had been present in the source zone during the continuous air sparging; however, the LNAPL at MW-1 and MW-2 was not detected after 45 days of the pulsed air sparging. Once the LNAPL was removed, the benzene concentrations dropped from 0.87 mg/L in October 2003 to 0.412 mg/L in October 2004 in MW-1 and from 4.2 mg/L to 0.54 mg/L in MW-2 over the same monitoring period. MW-3 and MW-5 are located in the dissolved plume. During the continuous air sparging, the groundwater benzene concentration in MW-3 stabilized at 0.026 mg/L while in MW-5 it slightly increased to 0.024 mg/L. The benzene concentrations at both wells declined to under the detection limit, which was 0.001 mg/L, within 2 months of the pulsed air sparging. They increased to 0.01 mg/L in October 2004 as the groundwater table rose, resulting in the dissolution of the benzene from the vadose zone. MW-8 is at the edge of the sparging system (9 m away from the closest sparging points) and is downgradient from the source zone. The continuous sparging caused the benzene concentration at MW-8 to elevate. However, it stopped increasing in the pulsed operation from September 2003 to April 2004. A likely explanation is that pulsed air sparging created a peripheral zone where hydrocarbon removal occurred over longer times as contaminated water moved in and out of the air contact zone (16). The benzene concentration trend at MW-8 might result from the expansion of the radius of influence under the pulsed air VOL. 39, NO. 18, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7285
sparging. Two additional sparging wells were installed by MW-8 in August 2004. One is located 3.3 m northwest of MW-8, and the other is 10 m southeast of MW-8. The benzene concentration at MW-8 dropped from 4.8 mg/L in April 2004 to 2.1 mg/L in October 2004. Figure 9b shows the total toluene (T), ethylbenzene (E), and xylene (X) concentrations measured in key monitoring wells (E concentration was much lower than T and X concentrations at this site). Figure 9a and 9b indicated that TEX concentrations in these key monitoring wells, in particular the source zone wells MW-1 and MW-2, decreased at a lower rate than benzene concentrations. These observations are supported by the results of Johnson’s air-sparging theoretical analysis which indicated that the chemical property most affecting hydrocarbon volatilization in source zones was solubility (5). The groundwater results at MW-3, a well in the dissolved plume, are an exception to this trend, perhaps because a coarse gravel zone begins at this location. Groundwater monitoring results of other parameters at MW-3 have differed from other wells during this test, for example, the pressure results in Figure 4 and in DO measurements not shown. The results of this field test of pulsed air sparging provided multiple lines of evidence that demonstrated improved performance over continuous operation of air-sparging systems and that supported previous laboratory-scale observations (6, 17). Pulsing air injection increased the dissolved oxygen in the saturated zone and hydrocarbon removal rates from both volatilization and biodegradation mechanisms. The contribution of hydrocarbon mass loss via volatilization to the total hydrocarbon mass removal increased during pulsed sparging. The increased reduction rate for dissolved benzene in the source zone, the most soluble chemical of concern, corresponds to the increased volatilization observed in the vadose zone and supports prior theoretical analysis (5). Because benzene is the risk driver at most petroleum hydrocarbon-contaminated sites, pulsed air sparging could specifically reduce risk faster than continuous air sparging.
Acknowledgments This study was funded by BP Remediation Management Environmental Technology Development Program. We thank Wally Zverina, the RETEC Group, Inc., and Lon Husbands, Kalkaska Float Equipment, Inc., for managing and conducting the field test. We also give special thanks to Kevin Endriss and Kevin Heaton, both BP environmental business managers, for their support of the study.
Literature Cited (1) Suthersan, S. S. In situ air sparging. Remediation Engineering: Design Concepts; CRC press LLC: 1999; Chapter 4. (2) Johnson, P. C.; Johnson R.; Bruce, C. L.; Leeson, A. Advances in in situ air sparging/biosparging. Biorem. J. 2001, 5 (4), 251266. (3) Williams, O. Engineering and Design: In Situ Air Sparging; U.S. Army Corps of Engineers Manual EM 1110-1-4005, 1997; Chapter 6.
7286
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 18, 2005
(4) Marley, M. C.; Bruell, C. J. In Situ Air Sparging: Evaluation of Petroleum Industry Sites and Considerations for Applicability, Design, and Operation; API Publication 4609; American Petroleum Institute: Washington, DC, 1995. (5) Johnson, P. C. An assessment of the contributions of volatilization and biodegradation to in situ air sparging performance. Environ. Sci. Technol. 1998, 32, 276-281. (6) Johnson, P. C.; Das, A.; Bruce, C. L. Effect of flow rate changes and pulsing on the treatment of source zones by in situ air sparging. Environ. Sci. Technol. 1999, 33 (10), 1726-1731. (7) Naval Facilities Engineering Command. Final Air Sparging Guidance Document; NFESC Technical Report TR-2193-ENV, 2001; Appendix A. (8) Reddy, K. R.; Zhou, J.; Kosgi, S. A review of in situ air sparging for the remediation of VOC-contaminated saturated soils and groundwater. Hazard. Waste Hazard. Mater. 1995, 12 (2), 97118. (9) Adam, J. A.; Reddy, K. R. Laboratory study of air sparging TCEcontaminated saturated soils and groundwater. Ground Water Monit. Remed. 1999, 19 (3), 182-190. (10) Heron, G.; Gierke, J. S.; Faulkner, B.; Mravik, S.; Wood, L.; Enfield, C. G. Pulsed air sparging in aquifers contaminated with dense nonaqueous phase liquids. Ground Water Monit. Remed. 2002, 22 (4), 73-82. (11) Yang, X.; Huang T.; Tsao. G. T. Pressure pulsation in solidphase fermentation. Appl. Biochem. Biotechnol. 2002, 98 (100) 599-610. (12) Kueper, B. H.; Frind, E. O. An overview of immiscible fingering in porous media. J. Contam. Hydrol. 1988, 2 (2), 95-110. (13) Ahlfeld, D. P.; Dahmani, A.; Ji, W. A conceptual model of field behavior of air sparging and its implications for application. Ground Water Monit. Rev. 1994, Fall, 132-139. (14) Tomlinson, D. W.; Thomson, N. R.; Johnson, R. L.; Redman, J. D. Air distribution in the Borden aquifer during in situ air sparging. J. Contam. Hydrol. 2003, 67 (1-4), 113-132. (15) Lyman, W. J.; Reehl, W. F.; Rosenblatt, D. H. Handbook of Chemical Property Estimation Methods; American Chemical Society: Washington, DC, 1990; Chapter 17. (16) Johnston, C. D.; Rayner, J. L.; Briegel D. In Situ Air Sparging Just How Efficient is it in Remediating Groundwater Contaminated by Dissolved Petroleum Hydrocarbons. Proceedings of Contaminated Site Remediation Conference, Fremantle, Western Australia, March 1999; pp 697-704. (17) Kirtland, B. C.; Aelion, C. M. Petroleum mass removal from low permeability sediment using air sparging/soil vapor extraction: impact of continuous or pulsed operation. J. Contam. Hydrol. 2000, 41 (3-4), 367-383. (18) Yang, X.; Beckmann, D.; Fiorenza, S.; Niedermeier, C. Pulsed Air Sparging Field Test. Proceedings of NGWA/API conference Petroleum Hydrocarbons and Organic Chemicals in Ground Water: Prevention, Assessment, and Remediation Conference, Baltimore, MD, 2004. (19) Newman, P.; Petersen, S.; Boersma, P.; Huddleston, R. Effects of Sparging on Groundwater Hydraulics and Groundwater Quality. Proceedings of 87th Annual Air and Waste Management Association Conference, Cincinnati, OH, June, 1994.
Received for review January 14, 2005. Revised manuscript received June 23, 2005. Accepted June 27, 2005. ES050084H
