Pilot-Scale Demonstration of Surfactant-Enhanced ... - ACS Publications

24 downloads 1944 Views 318KB Size Report
Feb 12, 2005 - Department of Civil and Environmental Engineering, Tufts. University, 113 Anderson Hall, ... Institute of Technology, Atlanta, Georgia 30332-0512, ... 605 East Robinson Street, Suite 308, Orlando, Florida 32801,. CDM, Inc.
Environ. Sci. Technol. 2005, 39, 1791-1801

Pilot-Scale Demonstration of Surfactant-Enhanced PCE Solubilization at the Bachman Road Site. 2. System Operation and Evaluation C. ANDREW RAMSBURG,† KURT D. PENNELL,‡ L I N D A M . A B R I O L A , * ,† G A R Y D A N I E L S , § CHAD D. DRUMMOND,| MATT GAMACHE,⊥ HSIN-LAN HSU,# ERIK A. PETROVSKIS,X KLAUS M. RATHFELDER,O JODI L. RYDER,# AND THOMAS P. YAVARASKI# Department of Civil and Environmental Engineering, Tufts University, 113 Anderson Hall, Medford, Massachusetts 02155, School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332-0512, Wilcox Associates, 3500 Ludington, Suite 310, Escanaba, Michigan 49829, HSW Engineering, Inc., 605 East Robinson Street, Suite 308, Orlando, Florida 32801, CDM, Inc., One Cambridge Place, 50 Hampshire Street, Cambridge, Massachusetts 02139, Environmental and Water Resources Engineering Program, Univeristy of Michigan, Ann Arbor, Michigan 48109-2125, GeoSyntec Consultants, 8120 Main Street, Dexter, Michigan 48130, and GeoSyntec Consultants, 838 SW First Avenue, Suite 530, Portland, Oregon 97204

A pilot-scale demonstration of surfactant-enhanced aquifer remediation (SEAR) was conducted during the summer of 2000 at the Bachman Road site in Oscoda, MI. Part two of this two-part paper describes results from partitioning and nonpartitioning tracer tests, SEAR operations, and post-treatment monitoring. For this field test, 68 400 L of an aqueous solution of 6% (wt) Tween 80 were injected to recover tetrachloroethene-nonaqueous phase liquid (PCE-DNAPL) from a shallow, unconfined aquifer. Results of a nonreactive tracer test, conducted prior to introducing the surfactant solution, demonstrate target zone sweep and hydraulic control, confirming design-phase model predictions. Partitioning tracer test results suggest PCE-DNAPL saturations of up to 0.74% within the pilotscale treatment zone, consistent with soil core data collected during site characterization. Analyses of effluent samples taken from the extraction well during SEAR operations indicate that a total of 19 L of PCE and 95% of the injected surfactant were recovered. Post-treatment monitoring indicated that PCE concentrations at many locations within * Corresponding author telephone: (617)627-3237; fax: (617)6273819; e-mail: [email protected]. † Tufts University. ‡ Georgia Institute of Technology. § Wilcox Associates. | HSW Engineering, Inc. ⊥ CDM, Inc. # Univeristy of Michigan. X GeoSyntec Consultants, Dexter. O GeoSyntec Consultants, Portland. 10.1021/es049563r CCC: $30.25 Published on Web 02/12/2005

 2005 American Chemical Society

the treated zone were reduced by as much as 2 orders of magnitude from pre-SEAR levels and had not rebounded 450 days after SEAR operations ceased. Pilot-scale costs ($365 900) compare favorably with design-phase cost estimates, with ∼10% of total costs attributable to the intense sampling density and frequency. Results of this pilotscale test indicate that careful design and implementation of SEAR can result in effective DNAPL mass removal and a substantial reduction in aqueous concentrations within the treated source zone under favorable geologic conditions

Introduction The objective of this study was to design, implement, and assess a pilot-scale demonstration of surfactant enhanced aquifer remediation (SEAR) for tetrachloroethene (PCE) dense nonaqueous phase liquid (DNAPL) source-zone mass recovery at the Bachman Road site. Site characterization efforts, laboratory experiments, and mathematical modeling involved in the development of the SEAR pilot-test design are presented in detail in the first part of this two-part series (1) and are briefly summarized below. This paper (part 2) focuses on details of design implementation, test performance, model/data comparisons, test evaluation, and project cost assessment. The Bachman Road site is located in northeast Michigan along the coast of Lake Huron in the town of Oscoda. Preliminary site assessment delineated four distinct groundwater plumes between US 23 and Lake Huron (Figure SI-1 in the Supporting Information) (2). While the release history at this site remains unknown, results from extensive coring and drive point sampling conducted during a 1997 feasibility study funded by the State of Michigan Department of Environmental Quality (MDEQ) suggested that plume B (Figure SI-1 in the Supporting Information) emanated from underneath a building formerly used as a dry-cleaning establishment (1). Analysis of soil cores collected during site characterization efforts revealed that the aquifer was composed of relatively homogeneous, medium- to fine-grained, glacial outwash sands with an organic carbon content of ∼0.02%, and underlain by a clay layer located approximately 7.6 m below ground surface (bgs) (1). Site characterization results suggested that the source-zone underneath the occupied building had relatively low DNAPL saturations without extensive DNAPL pooling. The favorable geology, underlying clay layer, and low suspected DNAPL saturations, along with the relatively shallow water table (varies seasonally between 2.4 and 3.0 m bgs), made the Bachman Road site an attractive SEAR candidate. Tween 80 (polyoxyethylene(20) sorbitan monooleate), a nonionic food-grade surfactant, was selected for use at the Bachman Road site based upon treatability studies conducted with Bachman aquifer material and PCE-DNAPL (3). Economic analyses suggested that treatment of the source zone at the Bachman Road site with an aqueous solution of Tween 80 would be cost-effective in comparison to pump-and-treat, with minimal loss of Tween 80 to the sandy aquifer material (1, 3). Regulatory approval of Tween 80 was based upon its use as a food-grade additive and potential for biodegradation (4-6). Efforts to accommodate physical and fiscal constraints and to minimize injected surfactant solution volume resulted in a unique SEAR design involving flushing cross-gradient to establish hydraulic control (1). Based upon results from the feasibility study, the MDEQ subsequently authorized funding for a team of researchers VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1791

FIGURE 1. Process flow diagram for SEAR treatment at the Bachman Road site.

TABLE 1. Pumping Schedule for SEAR Pilot-Scale Study at the Bachman Road Site well ID

period

S1 surfactant injection (L/min)

S2 surfactant injection (L/min)

S3 surfactant injection (L/min)

W1 water injection (L/min)

W2 water injection (L/min)

W3 water injection (L/min)

EW extraction well (L/min)

1. day -24 to -23 2. day -23 to 0 3. day 0 to 5 4. day 5 to 10 5. day 10 to 12 6. day 12 to 31 7. day 31 to 47 8. day 47+

off 1.9 (water) 1.9 off off off off off

off 1.9 (water) 1.9 1.9 off off off off

off 1.9 (water) 1.9 1.9 off off off off

off 3.8 3.8 3.8 3.8 off off off

off 3.8 3.8 3.8 3.8 off off off

off 3.8 3.8 3.8 3.8 off off off

-19.7 -19.7 -19.7 -19.7 -19.7 -19.7 -28.4 off

(with members from the Great Lakes and Mid-Atlantic Hazardous Substance Research Center, GeoTrans Inc., and MDEQ) to conduct a pilot-scale SEAR demonstration for source zone treatment underneath the northwest corner of the building. This paper begins with a description of the SEAR system which was constructed based upon the design developed in part one (1). Data from a partitioning tracer test provide estimates of DNAPL saturations in selected locations prior to treatment, while results from a conservative, nonreactive tracer test demonstrate hydraulic control and facilitate comparisons to model predictions of the swept zone. Concentrations of PCE and Tween 80 in samples collected during SEAR operations, along with posttreatment monitoring results, are used to assess DNAPL mass reduction and the benefits of mass removal within the treated zone. 1792

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

SEAR System System Engineering. The flushing design developed in the feasibility assessment phase of the project and described in Abriola et al. (1) was implemented in the pilot-scale demonstration. This design incorporated a single extraction well and two sets of three injection wells, one set for delivery of the water required to establish the necessary hydraulic gradient, and the other set for delivery of an aqueous solution of 6% (wt) Tween 80 (Figure SI-2 in the Supporting Information). The SEAR injection, recovery and effluent treatment system is depicted in a process flow diagram in Figure 1, and the pumping schedule is presented in Table 1. A flow rate of 3.8 L/min was maintained in each water injection well (11.4 L/min total) using flow valves (EZ-View, Hedland Inc.) and water pressure in the city water distribution network. Water used to prepare the surfactant solution was stored in

FIGURE 2. Schematic of multilevel sampler locations, with depth (m bgs) denoted next to each point. Angles given represent the angle between the ground surface (horizontal) and axis of multilevel well. a 7500 L tank (Raven Industries) and transferred to mixing tanks using a centrifugal pump (Figure 1). Use of two 9500 L conical bottom polyethylene mixing tanks (Raven Industries) allowed for alternating mixing and injection of the surfactant solution. Approximately 4500 kg of Tween 80 (Uniqema) were delivered and stored on site in 20 55-gal drums. For each 8175 L batch of 6% (wt) Tween 80 solution, a Masterflex model 77601-10 metering pump (Cole-Parmer Instrument Company) was used to transfer approximately 490 L of surfactant into a mixing tank containing 7685 L of water (determined using a Great Plains Industries digital water meter). Dissolution of the surfactant was optimized by positioning the effluent end of the surfactant transfer tubing immediately above the top impeller blade of the model D-3-00 double impeller mixer (Neptune Mixing Company) in each mixing tank. Using this approach, surfactant was completely mixed after approximately 5 h, and mixing was continued until the batch was used as feed for the surfactant injection system (∼14 h). Each batch of surfactant solution was sampled in duplicate for Tween 80 analysis prior to and during delivery to the injection wells. Surfactant solution was delivered at a flow rate of 1.9 L/min to each injection well using three laboratory size piston pumps (model PM6014, Fluid Metering Inc.) equipped with FMI piston heads (model Q2V) and stroke rate controllers (model V200). For the first 5-day period of the demonstration, surfactant solution was injected into three fully screened injection wells (S1-S3) to achieve a total flow rate of 5.7 L/min. Each of these injection wells was equipped with a recirculation system consisting of a 5 hp progressing cavity pump which drew from the bottom of the well and discharged at the top of the well. This recirculation system facilitated in-well mixing and ensured the surfactant solution was evenly distributed over the entire screened depth. During the second 5-day injection period, surfactant solution was delivered into two wells (S2 and S3) screened over the top and bottom 1.2 m of saturated depth, allowing targeted injection of surfactant to suspected higher PCE-contaminated regions of the treatment zone. To facilitate monitoring of the swept zone, five multilevel sample wells were installed within the treatment zone, providing 26 sampling points (Figure 2). Each multilevel well consisted of 5 (or 6 in the case of ML5) 0.95 cm diameter

tubes installed such that the 15 cm screened interval would allow sampling at a specific depth. All multilevel sampling wells extended underneath the building at an angle, facilitating sampling from locations inside the swept zone. Angles of the wells with the ground surface and approximate depth of each multi-level sampling location are shown in Figure 2. The extraction well consisted of 15.2 cm diameter poly(vinyl chloride) (PVC) pipe screened over the entire saturated thickness of the aquifer. Two submersible (one operational, one back-up) groundwater pumps (model 10 SQ 3-15 gpm variable flow rate, Grundfos) were installed in the extraction well and delivered water to the on-site treatment facility at a flow rate of 19.7 L/min. The rate of groundwater extraction was controlled using a globe valve and monitored using a Badger flow meter. Prior to entering the air strippers, a foodgrade silicone antifoaming agent (Trans-10, Trans Chemco, Inc.) was introduced into the liquid stream using a digital peristaltic pump (Masterflex) and subsequently mixed using a 15.2 cm static inline mixer. Rates of addition of the antifoaming agent varied from 2 to 40 mL/min depending on surfactant concentrations present in the air strippers. In addition to a backup generator sized to run all equipment on site, the effluent treatment system also contained two automatic back-up systems which allowed constant groundwater extraction in the case of a treatment system malfunction or shut-down. In the event either air stripper malfunctioned or for periods of scheduled maintenance, a 3800 L tank (Raven Industries) was used to contain diverted effluent water. To ensure against catastrophic failure of the effluent treatment system, any overflow from the smaller effluent storage tank would be automatically diverted into a 75 000 L frac-tank. The frac-tank, however, was not used during the pilot-scale demonstration. Separation of PCE from the effluent aqueous-phase surfactant solution was accomplished using two low-profile air strippers (model EZ-Tray 8.6 SS, QED Environmental Systems, Inc.). In preparation for the pilot-scale demonstration, a laboratory-scale tray stripper system was evaluated for removal of PCE from aqueous solutions of Tween 80 (7). Results indicated that tray strippers were capable of significant PCE separation, but performance predictions must be based upon Henry’s law coefficients obtained for surfactantladen water (7). In the absence of a measured Henry’s law VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1793

coefficient, Kibbey et al. (7) suggest the following correlation: 0

KH )

KH Csurf - CMC 1 + WSR Sorg

(

)

TABLE 2. Selected Properties of Alcohols Typically Used in PITTs

(1) compound

mol wt (g/mol)

Knwa (Lwater/LPCE)

costb ($/L)

2-propanol (IPA)c 2,2-dimethyl-3-pentanol 2,4-dimethyl-3-pentanolc 2-octanolc 2-ethyl-1-hexanol 2,6-dimethyl-2-heptanol 3,5,5-trimethyl-1-hexanolc

60.1 116.2 116.2 130.2 130.2 144.3 144.3

0 30 30 130 140 220 250

9.54 4373.50 156.80 6.20 12.26 73.00 17.33

where, KH is the dimensionless Henry’s law coefficient corrected for solubilization, KH0 is the dimensionless Henry’s Law coefficient in the absence of surfactant, Csurf is the concentration of surfactant [M/L3], CMC is the critical micelle concentration of surfactant [M/L3], Sorg is the aqueous solubility of the organic compound [M/L3], and WSR is the weight solubilization ratio of organic in surfactant solution [M/M]. Design of the air strippers for use at the Bachman Road site was based upon a surfactant concentration of 6 g/L (10× dilution of the influent concentration) in the extraction well. This resulted in a corrected KH (at 25 °C) of approximately 2.7 × 10-2 for KH0 ) 0.724 (7), a Tween 80-PCE WSR of 0.65 (3), a Tween 80 CMC of 13 mg/L (8), and a PCE aqueous solubility of 150 mg/L (9). Consistent with the corrected KH, each air stripper was 2.4 m in height and contained six perforated plates for a liquid flow rate of 19.7 L/min and an air flow rate of 11.9 m3/min. Air flow to each stripper was supplied by a model HP 4C19 blower (Cincinnati Fan and Ventilator Company), with an additional model HP 4C19 blower serving to transfer the off-gas through a knock-down drum and two 635 kg fill vapor phase, activated carbon vessels (TIGG Corporation) connected in series. The treated air was then discharged to the atmosphere under a MDEQ rule 290 exemption. The effluent liquid stream from the air strippers passed through a bag filter and then discharged to the city wastewater treatment facility via a lift station located approximately 50 m from the pilot-scale treatment building. Approval for this discharge was contingent upon submission of 24-h composite effluent samples to the wastewater treatment plant laboratory for BOD analysis. The surfactant-laden effluent constituted a high oxygen demand waste (COD ∼1.8; ref 10) and was therefore assessed an additional discharge fee of $0.02/kg BOD by the operators of the treatment facility (Earth Tech, Inc). System Operation. The pilot-scale SEAR system was constructed over a period of five months, during which time the treatment building was fabricated, components and controllers were installed, and all site improvements were performed. An existing outbuilding on site was renovated for use as a control room for the injection system. The operations and maintenance phase of the project involved flow rate adjustment, sample collection, air stripper and pump maintenance, bag filter replacement, and surfactant mixing. To operate at the desired injection and pumping rates, flow rate adjustments were based upon cumulative averages calculated from readings taken each morning. Multilevel well sampling occurred as frequently as every 4 h. Over the course of the pilot-scale demonstrations, biological growth became significant within the air strippers. Consequently, each air stripper was taken off line approximately weekly (during the surfactant injection phase) and thoroughly cleaned. Additionally, inspection of the submersible groundwater pumps indicated biological growth was significant at the pump intake. Overall, operation of the SEAR system proceeded without incident and at no time did effluent samples exceed discharge limits for chlorinated solvents. The system was taken out of service upon completion of the pilot study.

partitioning interwell tracer tests (PITTs) (e.g., refs 11-16). Assessment of NAPL removal can be made by comparing the volume of NAPL detected by PITTs conducted before and after surfactant flushing (e.g., ref 16). Alternatively, estimates of NAPL saturation obtained from pretreatment PITT results may be compared to the volume of NAPL recovered during active source-zone remediation. Partitioning tracer tests are usually designed with injection and extraction well systems that induce uniform flow in the direction of the natural hydraulic gradient. This allows multiple observation points to be located along particular flow paths to provide insight regarding the spatial distribution of NAPL within the swept zone (11, 12, 17). At the Bachman Road site, however, the unique hydraulic design and physical constraints of sampling beneath an occupied building prohibited locating multiple measurement points along any given flow path. Time and fiscal constraints permitted 2 weeks in which a PITT could be conducted at the Bachman Road site and resulted in the decision to interrogate only a portion of the treatment zone. Selection of alcohols for use at the Bachman Road site was based upon a comparison of NAPL-water partition coefficients (Knw), biodegradability, and approximate cost ($/L alcohol) of several alcohols previously employed in laboratory- and field-scale tests (Table 2). Alcohols which partition strongly (high Knw) were favored to overcome the anticipated low overall saturation in the swept zone (18). Branched alcohols are generally used to minimize potential biodegradation during a tracer test. Three partitioning alcohols (2,4-dimethyl-3-pentanol, 2-octanol, 3,5,5-trimethyl1-hexanol) and one nonpartitioning alcohol (2-propanol) were selected for use in the pilot-scale PITT to provide a range in Knw values (0-250) at reasonable cost (Table 2). In addition, bromide was selected for use in a nonreactive tracer test designed to validate the swept-zone predicted during hydraulic design (1). Simulation of these tracer tests was conducted utilizing MODFLOW (21), MODPATH (22), and MT3DMS (23) with the model grid, hydraulic conductivity field, hydraulic parameters, and model transport properties described by Abriola et al. (1). Selection of an appropriate location for injection of the suite of partitioning tracers was based upon estimated travel times as determined using the Bachman site model generated during the design portion of this study (1). Well S1 was eliminated from consideration because injected fluid from this location did not sweep a substantial portion of the treatment zone. Furthermore travel times for partitioning tracers injected through well S3 were too long for the time frame allotted for the PITT (14 days). Thus, injection of partitioning tracers through well S2 represented a compromise between the desire to interrogate a large portion of the pilot-scale volume and the limitation of a 2-week time frame.

Tracer Test Design

Sampling and Analytical Methods

Detection and characterization of NAPL distribution has been accomplished at both laboratory and field scales using

Sampling. Multi-level wells were sampled at each depth interval in duplicate using a Cole Parmer model 7518-00

1794

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

a

Ref 19.

b

Ref 20. c Selected for use.

peristaltic pump. Following a two-minute purge period, two 40 mL vials containing 1 mL of preservative (40% (wt) NaHSO4) were filled completely to eliminate headspace. Extraction well samples were also obtained periodically from the effluent line prior to entering the on-site treatment facility. Samples of each batch of surfactant solution were collected in a manner similar to that used for the collection of extraction well samples. All samples were preserved on ice and transported to the University of Michigan where they were stored at 4 °C for processing and analysis within 14 days of collection. The BOD load in the aqueous phase effluent from the on-site treatment plant was analyzed by the wastewater treatment plant laboratory using 24-h composite samples (Isco autosampler). Additionally, weekly samples from the treatment effluent were used to monitor for PCE in the aqueous-phase being discharged to the wastewater treatment plant. Vapor-phase effluent monitoring for VOC release was accomplished monthly under a rule 290 exemption, using air samples taken from the vapor phase point of discharge. The vapor-phase effluent was periodically monitored for the presence of PCE using Draeger tube sampling techniques. Analytical. Bromide analysis was performed without sample dilution using a model DX-100 Dionex ion chromatograph (IC) equipped with an Ion Pac AS4A SC 4 mm column and conductivity detector. Use of an aqueous elluent solution of 190 mg/L Na2CO3 and 140 mg/L NaHCO3 resulted in an average retention time of 2.3 min. Calibration samples were prepared using NaBr (Fisher Scientific) and analyzed every 100 samples using a six-point calibration curve (1.36 mg/L quantification limit). Quality control was maintained using a calibration check sample that was analyzed every 40 samples. All samples were analyzed using duplicate injection. Analytical separation of partitioning tracers (2-propanol, 2,4-dimethyl-3-pentanol, 2-octanol and 3,5,5-trimehtyl-1hexanol) was accomplished using a Hewlett-Packard 5890 gas chromatograph (GC) equipped with a Hewlett-Packard 19395A headspace analyzer, DB-WAX column (0.53 mm i.d., 30 m length), and a flame ionization detector (FID). Samples were prepared in duplicate by transferring a 5 mL aliquot into a 10 mL headspace vial, in which samples equilibrated for 120 min at 70 °C prior to analysis. Oven temperature was maintained at 70 °C for the initial 2 min and then increased up to 140 °C using a 25 °C /min temperature ramp. Average retention times of 0.67, 1.87, 3.92, 4.46 min were obtained for 2-propanol, 2,4-dimethyl-3-pentanol, 2-octanol and 3,5,5trimehtyl-1-hexanol, respectively. A six-point calibration curve was prepared using aliquots of each alcohol (obtained in neat form from Aldrich). Calibration standards were analyzed every 100 samples and checked every 12 samples with a single calibration sample. Method quantification limits were determined to be 1.6, 0.08, 0.2, 0.4 mg/L for 2-propanol, 2,4-dimethyl-3-pentanol, 2-octanol and 3,5,5-trimehtyl-1hexanol, respectively Robust and efficient analytical methods were required to process samples (∼4500 not counting duplicates) generated by the intense sampling frequency employed during the SEAR demonstration. All aqueous-phase samples were diluted 1:1 (vol) in 2-propanol for subsequent PCE and Tween 80 analyses. Gas chromatographic (Hewlett-Packard series 6890) separation of PCE was accomplished isothermally (95 °C) using a DB-5 column (J&W Scientific; 0.53 mm i.d., 30 m length). Following separation, the flow was split (1:1) to an electron capture detector (ECD) and FID, allowing analysis of a wide range of PCE concentrations and a lower quantification limit of 0.5 mg/L during the period of surfactant flushing. Average PCE retention time was 1.6 min during a 4 min run. A six-point calibration curve was obtained prior to analysis and checked every 20 samples for quality assurance. Subsequent Tween 80 analysis was performed using a Hewlett-Packard series 1050 high-perfor-

FIGURE 3. Simulated and measured bromide breakthrough curves for the nonreactive, conservative tracer test conducted at the Bachman Road site. mance liquid chromatograph (HPLC), equipped with an Alltech Hypersil ODS (C18) column (7.5 × 4.6 mm, 5 µm pore-size), and Sedere model SEDEX 55 evaporative light scattering detector (ELSD) following the method described by Taylor et al. (24). An elluent flow rate of 1.8 mL/min was initially comprised of 95% glacial acetic acid (ACS certified, Fischer Scientific) and 5% acetonitrile (HPLC grade, Fisher Scientific). This solvent mixture was held constant for 0.3 min, at which time the mobile phase composition was changed to 5% acetic acid and 95% acetonitrile for 0.7 min. The solvent mixture was returned to the initial composition for the remainder of the 1.5 min method. A six-point calibration curve was obtained every 100 samples and was checked every 15 samples for quality assurance. Using this method, average Tween 80 retention time was 1.6 min with a lower quantification limit of 90 mg/L during the period of surfactant flushing. Upon the cessation of surfactant flushing, reduced sample loads could no longer justify dedicating the GC and HPLC instruments to the Bachman Road project. Analysis of posttreatment samples was conducted using methods similar to those used during the pilot-scale test, however, quantification limits shifted over time as the concentration ranges of interest decreased (i.e., additional dilution of high concentration samples was not performed). Resulting PCE quantification limits for samples collected at the end of SEAR operations, and 14, 56, and 270 days subsequent were 5.14, 0.63, 0.02, and 0.02 mg/L, respectively. Quantification limits for Tween 80 in samples collected at the end of SEAR operations, and 14, 56, 270, and 450 days subsequent were 90, 90, 16, 16, and 50 mg/L, respectively. A more thorough analysis for chlorinated hydrocarbons in samples collected 450 days posttreatment was accomplished by the MDEQ laboratory (Lansing, MI) using EPA method 8260, which provided for a PCE quantification limit of 0.005 mg/L.

Results and Discussion Tracer Tests. Prior to the tracer test a steady flow rate through the source area was established for approximately 2 pore vol (see period 2 in Table 1). Subsequently, 8175 L of an aqueous solution containing 1000 mg/L bromide were introduced at a flow rate of 1.9 L/min through wells S1-S3 (Figure 2). Bromide concentrations measured in the extraction well during the tracer test compared favorably with predicted concentrations for the implemented hydraulic design (Figure 3) suggesting that the site model developed in part one (1) effectively simulated the transport properties within the pilotscale test area. Quantifiable concentrations of bromide (i.e., VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1795

FIGURE 4. Simulated sweep of partitioning tracers through the treated zone. >1.4 mg/L or C/C0 > 0.0014) appeared at all monitoring locations except for ML4-B and ML4-C within 10 days of injection. It is hypothesized that locations ML4-B and ML4-C are within a small zone bypassed as a result of flow around a lower-permeability region. Bromide mass balance closure of 85% was achieved by analyzing samples collected over the first 20 days following tracer introduction. Measured concentrations of bromide at this point were tracking with model predictions in which 92% mass balance closure was achieved. These results indicate that effective hydraulic control was obtained for flushing operations at the pilot-scale. Following the injection of the bromide pulse, a total of 2725 L of a solution consisting of 790 mg/L 2-propanol, 855 mg/L 2,4-dimethyl-3-pentanol, 475 mg/L 2-octantol, and 210 mg/L 3,5,5-trimethyl-1-hexanol was introduced through an injection well (S2) at a constant rate of 1.9 L/min. Note that water flooding (1.9 L/min) continued through wells S1 and S3 (see period 2 in Table 1). Model flow path estimates of the tracer swept area from injection at S2 to the extraction well (EW) are shown in Figure 4. Aqueous samples were taken at all monitoring points (Figure 2) at 8-h intervals for 8 days after the start of the tracer injection. Tracer concentration breakthrough curves (BTCs) were analyzed at each multilevel sampling location within the swept area. First temporal moments of the nonreactive and partitioning tracers were calculated and used to determine average DNAPL saturations (SN) present along the flow paths upstream of each sampling location (11, 12, 15, 25). Since a substantial portion of the BTC tail was below analytical quantification limits, the distal portion of each BTC was extrapolated by fitting (nonlinear regression) an exponential function to the final portions of the BTCs (26-28). As a representative example, the extrapolated BTC for sample point ML5-E is shown in Figure SI-3 in the Supporting Information. Calculated DNAPL saturations using the exponential extrapolation technique ranged from 0.15 ( 0.02% to 0.74 ( 0.20% along the measured flow paths (Table 3). The uncertainty reflected in the estimates reported in Table 3 is based upon propagation of both random and systematic errors through the moment analysis (18, 30). In general, a minimum retardation factor of 1.2 is required for adequate separation between the partitioning and conservative tracers (29). Thus, tracers producing retardation factors less than 1.2 were excluded from the determination of average DNAPL saturation along the flow path from the injection well to the measurement location. Because multiple sample points were 1796

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

not located along flow paths, spatial trends in DNAPL saturations could not be delineated. DNAPL saturation values were, however, observed to be highest at the deeper sampling locations, suggesting that the majority of DNAPL mass interrogated by these tracers resided near the bottom of the aquifer (Table 3). Direct comparison of the PITT saturation values to those obtained from soil cores collected during site characterization activities (see ref 1) was limited by (i) the spatially averaged nature of PITT data versus the point measurements obtained from core samples and (ii) the lack of overlap in the volumes interrogated by each of the methods. Fiscal and time constraints prohibited a post-SEAR partitioning tracer test. Surfactant Flush. The surfactant flush was conducted over a period of 10 days and bracketed by periods of water flooding (Table 1). Following the post-treatment water flood, the extraction system continued to produce 19.7 L/min until day 31, at which time the flow rate was increased to 28.4 L/min to maximize capture of injected surfactant and solubilized PCE. Analytical measurements of PCE and Tween 80 concentrations in the extraction well effluent are presented in Figure 5. Note that PCE concentrations observed immediately preceding the initiation of SEAR flushing were approximately 10 mg/L. These concentrations are representative of active flushing conditions within the treatment zone initiated during the tracer tests and are lower than levels measured prior to the start-up of all six injection wells (∼15 mg/L). The vertical lines shown in Figure 5 represent the times corresponding to changes in the influent solution. These changes are denoted using numbers (located along the top of the plot) which correspond to the flushing sequence described in Table 1. Upon the arrival of surfactant in the extraction well, the PCE concentration rose to a value of 44 mg/L consistent with PCE solubilization. Concentrations of PCE in the effluent quickly became transient, suggesting a reduction in source mass. A second peak, however, is evident near day 10, and corresponds to the targeted injection approach used from day 5 until day 10, before the decrease in PCE concentration observed during post-treatment water flooding (period 5). Overall, concentrations in samples from the extraction well indicate that 19 L PCE were removed from within the treatment zone. The deviation between effluent surfactant concentrations (up to ∼10 g/L) and influent values (∼60 g/L) is attributed to dilution, caused by clean water entering the extraction well via its upstream radius of influence, and the hydraulic design which used the

TABLE 3. Calculated NAPL Saturations Based upon Alcohol Retardation during the PITT sampling location

depth (m)

ML2-A

3.4

ML2-B

4.7

ML3-A

3.0

ML3-B

4.3

ML3-C

5.5

ML4-D

6.2

ML4-E

6.8

ML5-C

5.3

ML5-E

6.5

ML5-F

7.2

tracera

relative mass obsd (%)b,c

retardation factorc

NAPL saturation (%)c

DMP OCT TMH DMP OCT TMH DMP OCT TMH DMP OCT TMH DMP OCT TMH DMP OCT TMH DMP OCT TMH DMP OCT TMH DMP OCT TMH DMP OCT TMH

89 ( 1 80 ( 1 83 ( 1 87 ( 1 78 ( 1 73 ( 1 92 ( 1 80 ( 1 84 ( 1 88 ( 1 70 ( 1 55 ( 1 88 ( 1 71 ( 1 59 ( 1 84 ( 2 78 ( 2 70 ( 2 84 ( 1 70 ( 1 63 ( 1 102 ( 1 94 ( 1 86 ( 1 102 ( 1 95 ( 1 91 ( 1 77 ( 1 64 ( 1 62 ( 1

1.03 ( 0.04 1.14 ( 0.04 1.18 ( 0.04 1.03 ( 0.02 1.21 ( 0.03 1.34 ( 0.03 1.02 ( 0.03 1.08 ( 0.03 1.13 ( 0.04 1.01 ( 0.02 1.09 ( 0.02 1.13 ( 0.02 1.02 ( 0.02 1.16 ( 0.02 1.29 ( 0.02 1.02 ( 0.03 1.10 ( 0.03 1.16 ( 0.04 1.04 ( 0.03 1.36 ( 0.03 1.52 ( 0.04 1.09 ( 0.04 1.31 ( 0.04 1.54 ( 0.05 1.20 ( 0.03 1.93 ( 0.06 3.13 ( 0.11 1.06 ( 0.04 1.32 ( 0.04 1.50 ( 0.06

0.10 ( 0.12 0.11 ( 0.03 0.07 ( 0.02 0.10 ( 0.08 0.16 ( 0.03 0.14 ( 0.02 0.07 ( 0.10 0.06 ( 0.02 0.05 ( 0.02 0.05 ( 0.06 0.07 ( 0.02 0.05 ( 0.01 0.07 ( 0.06 0.12 ( 0.02 0.11 ( 0.02 0.06 ( 0.11 0.07 ( 0.03 0.06 ( 0.02 0.15 ( 0.10 0.28 ( 0.04 0.21 ( 0.03 0.29 ( 0.13 0.24 ( 0.04 0.21 ( 0.03 0.68 ( 0.13 0.71 ( 0.08 0.84 ( 0.09 0.20 ( 0.12 0.25 ( 0.04 0.20 ( 0.03

avg NAPL saturation (%)c,d -e 0.15 ( 0.04 -e -e 0.11 ( 0.02 -e 0.24 ( 0.07 0.23 ( 0.05 0.74 ( 0.20 0.22 ( 0.06

a DMP, 2,4-dimethyl-3-pentanol; OCT, 2-octanol; TMH, 3,5,5-trimethyl-1-hexanol. b Zeroth moment normalized by the zeroth moment of 2-propanol. Value ( total uncertainty (i.e., random and systematic). d Average for location based upon tracers with retardation factors greater than 1.20 (29). e Retardation factor less than 1.20 for all tracers at this location, thus average value was not reported. c

FIGURE 5. Concentrations of PCE and Tween 80 in the extraction well. Vertical lines represent a change to the influent solution or total flow rate (1, extraction only; 2, water flood through S1, S2, S3, W1, W2, and W3; 3, fully screened injection of Tween 80 in S1, S2, and S3; 4, partially screened injection of Tween 80 through S2 and S3; 5, water flood through W1, W2, andW3; 6, extraction only; 7, increased rate of extraction; 8, pumping terminated). injection of clean water to produce surfactant sweep across the natural groundwater gradient. Hydraulic control was maintained throughout the test, as 95% of the injected surfactant was recovered in the extraction well. This recovery compares favorably to previous SEAR demonstrations (e.g., ∼94% surfactant recovery reported in ref 17) and consistent with pretest modeling efforts (∼95% reported in ref 1).

Concentrations of PCE in the extraction well effluent were observed to increase more than 4x over initial levels (∼10 mg/L) during surfactant flushing. In contrast, contaminant concentrations increased more than 20× in extraction well data from SEAR demonstrations conducted at CFB Borden and Hill AFB (17, 31, 32). The relatively smaller increases in PCE concentration observed in this SEAR demonstration were VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1797

attributed to low DNAPL saturations present in the treatment zone, and the unique hydraulic design. Previous pilot-scale SEAR tests have been conducted in aquifers where average DNAPL saturations were one to two orders-of-magnitude larger (2-3%), resulting in relatively greater increases in contaminant concentrations in effluent well data (e.g., see refs 17 and 31). It is also likely that the relatively low DNAPL saturations within the treatment zone contributed to the rapid onset of transient behavior observed, which could result from depletion of the DNAPL source mass. In prior studies, the mass of DNAPL present was much greater and allowed for sustained contaminant concentrations during much of the surfactant pulse (17, 31). Concentrations of PCE observed in the effluent near the end of the test were similar to those observed after the start of the extraction well, but prior to the start of the injection wells (15 mg/L). It is important to note that the pilot-scale SEAR flood was designed to sweep only a portion of the contaminated zone (1). Thus, the return of PCE concentrations in the effluent to those values measured prior to treatment may be attributed to the influence of DNAPL source areas located outside the treated zone, but within the extraction well radius of influence (i.e., underneath the building). This hypothesis is further supported by the multilevel sampler concentration analyses presented below, which demonstrate DNAPL solubilization and subsequent reduction in PCE concentrations within the treated zone. The evolution of Tween 80 and PCE concentrations in three, representative multi-level sampling locations within the swept zone is presented in Figure 6. The vertical lines and numbers at the top of plots in Figure 6 represent changes in the influent conditions, and correspond to those used in Figure 5 and Table 1. Data from ML4-D (Figure 6A) represent a location with insignificant PCE concentrations. Concentrations at sampling location ML1-E illustrate a moderate level of solubilization, with PCE concentrations ranging from approximately 15 mg/L up to approximately 85 mg/L during the surfactant pulse (Figure 6B). Monitoring location ML1-E is along one of the longest flow paths generated by the unique hydraulic design used to flush across the natural gradient (see Figure SI-2 in the Supporting Information). The broadening of the surfactant pulse at this location may be attributed to changes in the flow field experienced after water injection into wells W1-W3 was terminated on day 12 (end of period 5 in Figure 6B). The resulting temporal variation in flow paths suggests samples collected from point ML1-E are descriptive of regions within the swept-zone that were not previously monitored (i.e., not on the initial flow path). Surfactantenhanced solubilization within these regions is hypothesized to have contributed to the observed fluctuations in aqueousphase PCE concentrations. Sustained operation of the extraction well (period 6 in Figure 6B) allowed for subsequent capture of surfactant and solubilized PCE from this location. Tween 80 and PCE concentrations measured at ML5-E are shown in Figure 6C, where significant solubilization was observed. Here, concentrations of PCE increased up to ∼5000 mg/L upon arrival of surfactant at this location (∼48× increase over waterflood concentrations). Locally high concentrations such as these represent significant PCE mass removal along this flow path. The high concentrations of Tween 80 achieved at each of these observations points approach injected concentrations (∼60 g/L) indicating effective delivery of surfactant throughout the treatment zone. Post-treatment Monitoring. While fiscal constraints prohibited a post-SEAR partitioning tracer test, sampling of the multi-level locations was conducted 14, 56, 270, and 450 days post-treatment to assess the longer-term impacts of the pilot-scale SEAR flood on concentrations with the treated zone. Monitoring results from the sampling event conducted 270 days following the conclusion of treatment indicate that PCE concentrations in the treatment zone were reduced 1798

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

FIGURE 6. Representative PCE and Tween 80 concentration data measured at multilevel sampling points (A) ML4-D, (B) ML1-E, and (C) ML5-E. Vertical lines represent a change to the influent solution or total flow rate (3, fully screened injection of Tween 80 in S1, S2, and S3; 4, partially screened injection of Tween 80 through S2 and S3; 5, water flood through W1, W2, and W3; 6, extraction only; 7, increased rate of extraction). significantly and had not rebounded (Table 4). One exception to this trend is at sampling point ML1-B where PCE concentrations at 270 days post-treatment increased to 26 mg/L. Concentration data shown in Table 5 for the entire monitoring period, however, indicate that the elevated concentration observed at ML1-B subsequently decreased to 5.2 mg/L. Also shown in Table 5 are concentrations of PCE from ML5-B, ML5-C and ML5-E at 450 days post-treatment. These data demonstrate a 2 order-of-magnitude decrease in concentrations when compared to those measured immediately prior to surfactant injection. The lack of rebound at both 270 and 450 days post-treatment is a reliable indication that flushing with Tween 80 effectively solubilized PCE-DNAPL present within the treated zone. The relatively

TABLE 4. Concentrations Measured at All Sampling Locations at the Start of SEAR and at 270 days Post-treatmenta PCE (mg/L) location ID

start SEAR

Tween 80 (mg/L)

270 days post-SEAR

start SEAR

270 days post-SEAR

ML1-A 3.5 m ML1-B 4.9 m ML1-C 6.0 m ML1-D 6.6 m ML1-E 6.8 m

0.29 0.52 0.25 0.64 4.09

NQ < 0.020 26.2 NQ < 0.020 NQ < 0.020 NQ < 0.020

NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311

NQ < 16.0 NQ < 16.0 NQ < 16.0 NQ < 16.0 205

ML2-A 3.4 m ML2-B 4.7 m ML2-C 5.9 m ML2-D 6.5 m ML2-E 7.3 m

NQ < 0.25 NQ < 0.25 NQ < 0.25 0.43 2.62

0.070 NQ < 0.020 NQ < 0.020 NQ < 0.020 NQ < 0.020

NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311

NQ < 16.0 NQ < 16.0 NQ < 16.0 NQ < 16.0 204

ML3-A 3.0 m ML3-B 4.3 m ML3-C 5.5 m ML3-D 6.6 m ML3-E 6.8 m

NQ < 0.25 1.32 NQ < 0.25 1.15 61.1

2.28 0.152 0.127 0.129 6.31

NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311

NQ < 16.0 NQ < 16.0 NQ < 16.0 NQ < 16.0 9100

ML4-A 2.5 m ML4-B 3.7 m ML4-C 5.0 m ML4-D 6.2 m ML4-E 6.8 m

NQ < 0.25 NQ < 0.25 NQ < 0.25 NQ < 0.25 NQ < 0.25

DRY 0.027 NQ < 0.020 NQ < 0.020 NQ < 0.020

NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311

NQ < 16.0 NQ < 16.0 NQ < 16.0 NQ < 16.0 NQ < 16.0

ML5-A 2.9 m ML5-B 4.1 m ML5-C 5.3 m ML5-D 5.9 m ML5-E 6.5 m ML5-F 7.2 m

0.35 45.3 48.0 0.89 98.8 NQ < 0.25

0.288 0.052 0.100 NQ < 0.020 0.190 NQ < 0.020

NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311

NQ < 16.0 21.9 NQ < 16.0 NQ < 16.0 NQ < 16.0 56.3

a

NQ ) not quantifiable at given limit.

low aqueous-phase concentrations of PCE observed in the deeper sampling locations (samples D, E, and ML5-F) suggest that downward migration of DNAPL free product did not occur within the treated zone (Tables 4 and 5). Migration of DNAPL would be expected to result in elevated concentrations of PCE along the bottom of the aquifer, both during the test (Figure 6C) and at later time. Also shown in Tables 4 and 5 are surfactant concentrations obtained from sampling events occurring up to 450 days following the test. At the conclusion of the test, surfactant concentrations were not detectable at the 90 mg/L level for all sampling locations in the ML1 and ML5 multilevel bundles (Table 5). With the exception of sampling points ML1-E (7.3 m below ground surface) and ML5-F (7.2 m below ground surface), posttreatment surfactant concentrations in bundles ML1 and ML5 did not generally exceed 210 mg/L (∼0.35% of the maximum surfactant concentration observed during the test). At ML5-F surfactant concentrations remained elevated from the end of the SEAR test while at ML1-E, surfactant concentrations increased to 4300 mg/L at 56 days following the pilot-scale flood. At 450 days post-treatment, however, surfactant concentrations had returned to 52.1 mg/L and NQ < 50 mg/L in samples taken from ML1-E and ML5-F, respectively. Thus, elevated surfactant concentrations can be attributed to downstream (natural gradient) migration of uncaptured surfactant solution from a lower permeability zone penetrated by the flow gradients used during SEAR operations. Note that these sampling locations are near the sand-silt-clay transition layer present immediately above the clay layer at the Bachman site (1). Pilot-Scale Costs. SEAR system costs are important components of pilot-scale tests; however, interpretation of cost data must include recognition of the novelty of this innovative technology. For example, this pilot-scale surfactant flood employed a sampling density and frequency that is not

typical of field-scale SEAR implementations to supply the data required for a thorough technology demonstration. The overall costs of this pilot-scale SEAR demonstration are shown in Table 6. Design work included modeling, engineering, contractor selection and award, permitting, site visits, and project meetings. A large portion of the operation and maintenance (O&M) costs consisted of fees (∼$0.02/kg BOD) associated with discharging high BOD (surfactant-laden) water to the municipal wastewater treatment plant. The addition of the BOD surcharge to a regular sewage fee ($0.15/L tap water used on-site) resulted in a total discharge cost of $26 000. Also included in Table 6 are previously reported cost estimates for pilot-scale SEAR using a 4% (wt) Tween 80 solution at the Bachman site (3). Because most of the recoverable equipment was retained on site for future remediation efforts, no equipment costs have been recaptured to date. Thus, actual costs ($365 900) are within 15% of the estimates ($317 300) reported by Ramsburg and Pennell (3) when recoverable equipment cost are excluded. In general, the greater costs incurred resulted from incorporation of a pretreatment PITT test and use of a more intense sampling frequency than originally estimated for the pilotscale demonstration. The overall objective of this study was to demonstrate PCE mass removal from a DNAPL source zone located beneath a former dry cleaning facility by flushing with a nonionic surfactant solution. The removal of 19 L PCE corresponds to a volume average initial saturation in the swept zone of ∼0.04% and is consistent with relatively low NAPL saturations estimated from pretreatment PITT data. While the amount of PCE mass removed at this SEAR site is modest when compared with previous larger-scale solvent disposal areas (e.g., ∼1300 L at Hill AFB, ref 32), posttreatment monitoring data suggest that concentration rebound within the treated zone had not occurred 450 days after treatment. These findings indicate that SEAR may be an effective treatment option for a number of source zones associated with smaller scale dry-cleaning operations. In this pilot-scale study, a total of 68 400 L of a 6% (wt) surfactant solution was flushed through a treatment zone having a pore vol of ∼45 000 L, which is equivalent to ∼1.5 pore vol flood. The use of pore volumes to describe SEAR flooding activities, however, is problematic since this metric is highly dependent upon the treated volume used to determine the dimensional pore volume. For example, had this flood been conducted underneath the entire building, the amount of surfactant required would have remained relatively unchanged (∼4500 kg), but the volume of the treated zone would have increased by nearly 4-fold, thereby decreasing the number of pore volumes of surfactant solution injected (to ∼0.4). This flushing approach was not considered viable at this site, however, because longer treatment times required to transport the surfactant pulse would have greatly increased O&M costs. For this reason implementation of SEAR, particularly at smaller sites, should be optimized for contaminant recovery and cost-effectiveness. The cost of pilot-scale remediation systems are generally skewed toward higher values when compared with similar treatments at full-scale. This typically results from certain threshold costs required for SEAR implementation regardless of scale (i.e., wells, pumps, treatment facilities, etc.). For this test, the overall cost per volume soil was ∼$2100/m3, yet for comparable full-scale operations, estimates have shown this figure drops to ∼$300/m3 (3). It is also important to note that SEAR costs are highly dependent upon site characteristics (NAPL saturation, hydraulic conductivity, site access, etc.). Moreover, costs per volume soil should not be viewed as a cost normalization which allows direct comparison between sites. Comparisons are best made between remediation technologies on a site-specific basis and at comparable scales VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1799

TABLE 5. Detailed Monitoring Results for ML1 and ML5a SEAR location ID

start

ML1-A 3.5 m ML1-B 4.9 m ML1-C 6.0 m ML1-D 6.6 m ML1-E 6.8 m ML5-A 2.9 m ML5-B 4.1 m ML5-C 5.3 m ML5-D 5.9 m ML5-E 6.5 m ML5-F 7.2 m

0.29 0.52 0.25 0.64 4.09 0.35 45.3 48.0 0.89 98.8 NQ < 0.25

ML1-A 3.5 m ML1-B 4.9 m ML1-C 6.0 m ML1-D 6.6 m ML1-E 6.8 m ML5-A 2.9 m ML5-B 4.1 m ML5-C 5.3 m ML5-D 5.9 m ML5-E 6.5 m ML5-F 7.2 m

NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311 NQ < 311

a

post-SEAR end

14 days

56 days

270 days

450 days

PCE Concentrations (mg/L) NQ < 5.14 NQ < 0.63 NQ < 5.14 NQ < 0.63 NQ < 5.14 NQ < 0.63 NQ < 5.14 NQ < 0.63 NQ < 5.14 2.40 NQ < 5.14 NQ < 0.63 NQ < 5.14 NQ < 0.63 NQ < 5.14 NQ < 0.63 NQ < 5.14 NQ < 0.63 NQ < 5.14 NQ < 0.63 NQ < 5.14 14.3

0.295 8.24 0.132 0.136 1.04 0.085 0.457 0.882 0.089 0.937 0.205

NQ < 0.020 26.2 NQ < 0.020 NQ < 0.020 NQ < 0.020 0.288 0.052 0.100 NQ < 0.020 0.190 NQ < 0.020

0.032 5.20 0.063 0.057 0.130 0.022 0.170 0.240 0.020 0.400 0.018

Tween 80 Concentrations (mg/L) NQ < 90 NQ < 90 NQ < 90 NQ < 90 NQ < 90 NQ < 90 NQ < 90 NQ < 90 NQ < 90 122 NQ < 90 122 NQ < 90 104 NQ < 90 NQ < 90 NQ < 90 NQ < 90 NQ < 90 NQ < 90 17 500 15 850

162 105 49.2 85.4 4300 63.7 94.0 54.1 64.9 63.9 151

NQ < 16.0 NQ < 16.0 NQ < 16.0 NQ < 16.0 205 NQ < 16.0 21.9 NQ < 16.0 NQ < 16.0 NQ < 16.0 56.3

NQ < 50.0 NQ < 50.0 NQ < 50.0 NQ < 50.0 52.1 NQ < 50.0 NQ < 50.0 NQ < 50.0 NQ < 50.0 NQ < 50.0 NQ < 50.0

NQ ) not quantifiable at given limit.

TABLE 6. Pilot-Scale Costs type

actual ($)

estimateda ($)

design construction wells equipment PITT surfactant operation and maintenance (O&M) discharge to WWTP utilities on-site office rental in-house analytical on-site labor and expenses

55 900 185 500 25 600 130 500 11 900 17 500 58 800 26 000 3 000 3 800 26 000 65 700

51 300b 168 000 23 500 127 500 na 17 000 43 700 31 200 3 800 na 8 700 54 300c

total

365 900

317 300d

a Ramsburg and Pennell (3). Categories adapted for direct comparison. b Sum of administrative and engineering fees. c Sum of piping and contingencies fees, system maintenance, and carbon disposal. d Excludes equipment recovery and location factor and assumes a negligible time value of money over project duration (∼3 months).

of implementation. Knowledge gained from pilot-scale studies should serve to decrease implementation costs as SEAR technologies are refined. Two key factors affecting the overall cost of SEAR are chemical costs and effluent treatment costs. Chemical costs may be reduced through the development of efficient low-cost surfactant formulations and effective source-zone delineation. In contrast, the cost of treating contaminated, surfactant-laden water is reliant upon advances in separation technologies. While surfactant recycle systems may reduce both chemical and post-recycle treatment costs, economic analyses indicate that savings realized from these systems are often offset by the additional capital expenditure for smaller sites (3). Surfactant recycle systems are envisioned to be more applicable at sites where multiple pore volumes are required (e.g., higher DNAPL saturations), but real-time recycling requires the extraction and separation of significant amounts of surfactant before the cessation of flooding. This 1800

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 6, 2005

may prove difficult at sites such as Bachman Road if fullscale treatment constraints necessitate a hydraulic design that has a swept volume greater than the volume of surfactant solution required for treatment of the DNAPL source.

Acknowledgments The authors express their appreciation to Dr. Peter Adriaens (University of Michigan) for his contributions throughout the Bachman Road project and to Dr. Suresh Rao (Purdue University) for his assistance in selection of appropriate partitioning tracers. Appreciation is also expressed to Duke Engineering & Services, Inc. and the University of Texas for licensing the PITT technology (U.S. Patents 5,905,036 and 6,003,365) for use at the Bachman Road site. Funding for this research was provided by the Michigan Department of Environmental Quality, under Contract Y80011, and by the U.S. Environmental Protection Agency, Great Lakes and MidAtlantic Center for Hazardous Substance Research (GLMACHSRC), under Grant R-825540. The content of this publication does not necessarily represent the views of either agency and has not been subject to agency review. The following individuals provided assistance in data collection and data analysis: Michael Gebhard, Ernie Hahn, and Peter Brink.

Supporting Information Available Information contained in this section includes a site map, particle tracking results for SEAR design, and representative breakthrough curves for the suite of tracers used in the PITT. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Abriola, L. M.; Drummond, C. D.; Hahn, E. J.; Hayes, K. F.; Kibbey, T. C. G.; Lemke, L. D.; Pennell, K. D.; Petrovskis, E. A.; Ramsburg, C. A.; Rathfelder, K. M. Pilot-scale demonstration of surfactantenhanced PCE solubilization at the Bachman Road Site. 1. Site characterization and test design. Environ. Sci. Technol. 2005, 39, 177.

(2) Brown and Root Environmental. Project No. 7M59: Remedial Investigation and Feasibility Study Report. Residential Wellss Bachman Road Site, Oscoda, Michigan; 1996. (3) Ramsburg, C. A.; Pennell, K. D. Experimental and economic assessment of two surfactant formulations for source zone remediation at a former dry cleaning facility. Ground Water Monit. Rem. 2001, 21 (4), 68-82. (4) Schick, M. J. Nonionic Surfactants; Surfactant Science Series 1; Marcel Dekker: New York, 1967; 1085 pp. (5) Swisher, R. D. Surfactant Biodegradation, 2nd ed.; Surfactant Science Series 18; Marcel Dekker: New York, 1987; 1085 pp. (6) Yeh, D. H.; Pennell, K. D.; Pavlostathis, S. G. Toxicity and biodegradability screening of nonionic surfactants using sediment-derived methanogenic consortia. Water Sci. Technol. 1998, 38 (7), 55-62. (7) Kibbey, T. C. G.; Pennell, K. D.; Hayes, K. F. Application of sievetray air strippers to the treatment of surfactant-containing wastewaters. AIChE J. 2001, 47, 1461-1470. (8) Pennell, K. D.; Adinolfi, A. M.; Abriola, L. M.; Diallo, M. S. Solubilization of dodecane, tetrachloroethylene, and 1,2dichlorobenzene in micellar solutions of ethoxylated nonionic surfactants. Environ. Sci. Technol. 1997, 31, 1382-1389. (9) Verschueren, K. Handbook of Environmental Data on Organic Chemicals, 2nd ed.; Van Nostrand Reinhold: New York, 1983. (10) Yeh, D. H.; Pennell K. D.; Pavlostathis S. G. Effect of Tween surfactants on methanogensesis and microbial reductive dechlorination of hexachlorobenze. Environ. Toxicol. Chem. 1999, 18, 1708-1416. (11) Jin, M.; Delshad, M.; Dwarakanath, V.; McKinney, D. C.; Pope, G. A.; Sepehrnoori, K.; Tilburg, C. E.; Jackson, R. E. Partitioning tracer test for detection, estimation and remediation performance assessment of subsurface nonaqueous phase liquids. Water Resour. Res. 1995, 31, 1201-1211. (12) Annable, M. D.; Rao, P. S. C.; Hatfield, K.; Graham, W. D.; Enfield, C. G. Partitioning tracers for measuring residual NAPL: Fieldscale test results. J. Environ. Eng. 1998, June, 498-503. (13) Wilson, R. D.; Mackay, D. M. Direct-detection of residual nonaqueous phase liquid in the saturated zone using SF6 as a partitioning tracer. Environ. Sci. Technol. 1995, 29, 1255-1258. (14) Nelson, N. T.; Brusseau, M. L. Field study of the partitioning tracer method for detection of dense nonaqueous phase liquid in a trichloroethene-contaminated aquifer. Environ. Sci. Technol. 1996, 30, 2859-2863. (15) Sillan, R. K.; Annable, M. D.; Rao, P. S. C.; Dai, D. P.; Hatfield, K.; Graham, W. D. Evaluation of in situ cosolvent flushing dynamics using a network of spatially distributed multilevel samplers. Water Resour. Res. 1998, 34, 2191-2202. (16) Meinardus, H. W.; Dwarakanath, V.; Ewing, J.; Hirasaki, G. J.; Jackson, R. E.; Jin, M.; Ginn, J. S.; Londergan, J. T.; Miller, C. A.; Pope, G. A. Performance assessment of NAPL remediation in heterogeneous alluvium. J. Contam. Hydrol. 2002, 54, 173-193. (17) Brown, C. L.; Delshad, M.; Dwarakanath, V.; Jackson, R. E.; Londergan, J. T.; Meinardus, H. W.; McKinney, D. C.; Oolman, T.; Pope, G. A.; Wade, W. H. Demonstration of surfactant flooding on an alluvial aquifer contaminated with dense nonaqueous phase liquid. In Innovative Subsurface Remediation Field Testing of Physical, Chemical and Characterization Technologies; Brusseau, M. L., Sabatini, D. A., Gierke, J. S., Annable, M. D., Eds.;

(18) (19) (20) (21) (22)

(23)

(24)

(25)

(26) (27) (28) (29)

(30)

(31) (32)

ACS Symposium Series 725; American Chemical Society: Washington, DC, 1999; pp 64-85. Dwarakanath, V.; Deeds, N.; Pope, G. A. Analysis of partitioning interwell tracer tests. Environ. Sci. Technol. 1999, 33, 38293836. Rao, P. S. C. Personal correspondence, 2000. Sigma-Aldrich Corporation, Milwaukee, WI. McDonald, M. G.; Harbaugh, A. W. A Modular Three-Dimensional Finite-Difference Groundwater Flow Model: US Geological Survey: Reston, VA, 1988. Pollock, D. W. User’s guide for MODPATH/MODPATH-PLOT, Version 3: A particle tracking postprocessing package for MODFLOW, the US Geological Survey finite-difference groundwater flow model. Open-File Rep.sU.S. Geol. Surv. 1994, No. 94-464. Zheng, C.; Wang, P. P. MT3DMS: A Modular Three-Dimensional Multispecies Transport Model for Simulation of Advection, Dispersion, and Chemical Reactions of Contaminants in Groundwater Systems; Documentation and User’s Guide. Contract report; SERDP-99-1; 1999. Taylor, T. P.; Pennell, K. D.; Abriola, L. M.; Dane, J. H. Surfactant enhanced recovery from a porous medium containing low permeability lenses 1. Experimental studies. J. Contam. Hydrol. 2001, 48, 325-350. Rao, P. S. C.; Annable, M. D.; Sillan, R. K.; Dai, D.; Hatfield, K.; Graham, W. D.; Wood, A. L.; Enfield, C. G. Field-scale evaluation of in situ cosolvent flushing for enhanced aquifer remediation. Water Resour. Res. 1997, 33, 2673-2686. Cain, R. B.; Johnson, G. R.; McCray, J. E.; Blanford, W. J.; Brusseau, M. L. Partitioning tracer tests for evaluating remediation performance. Ground Water 2000, 38, 752-761. Skopp, J. Estimation of true moments from truncated data. AIChE J. 1984, 30, 151-155. Pope, G. A.; Sepehrnoori, K.; Shiles, G. Numerical simulation of interwell tracers. In Situ 1994, 18, 107-121. Jin, M.; Butler, G. W.; Jackson, R. E.; Mariner, P. E.; Pickens, J. F.; Pope, G. A.; Brown, C. L.; McKinney, D. C. Sensitivity models and design protocol for partitioning tracer tests in alluvial aquifers. Ground Water 1997, 35, 964-972. Jin, M.; Jackson, R. E.; Pope, G. A. The interpretation and error analysis of PITT data. In Treating Dense Nonaqueous-Phase Liquids (DNAPLs): Remediation of Chlorinated and Recalcitrant Compounds; Wickramanayake, G. B., Gavaskar, A. R., Gupta, N., Eds.; Battelle Press: Columbus, OH, 2000; pp 85-92. Fountain, J. C.; Starr, R. C.; Middleton, T.; Beikirch, M.; Taylor, C.; Hodge, D. A controlled field test of surfactant-enhanced aquifer remediation. Ground Water 1996, 34, 910-916. Londergan, J. T.; Meinardus, H. W.; Mariner, P. E.; Jackson, R. E.; Brown, C. L.; Dwarakanath, V.; Pope, G. A.; Ginn, J. S.; Taffinder, S. DNAPL removal from a heterogeneous alluvial aquifer by surfactant-enhanced aquifer remediation. Ground Water Montit. Rem. 2001, 21 (4), 57-67.

Received for review March 19, 2004. Revised manuscript received December 15, 2004. Accepted December 16, 2004. ES049563R

VOL. 39, NO. 6, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1801