Vadose Zone Journal 2:148-153 (2003)
© 2003 Soil Science Society of America
SPECIAL SUBMISSIONS: Contaminant Characterization, Transport, and Remediation in Complex Multiphase Systems
Application of Gas-Phase Partitioning Tracer Tests to Characterize Immiscible-Liquid Contamination in the Vadose Zone Beneath a Fuel Depot
Mark L. Brusseau*,a,
Kali M. Bronsonb,
Stephanie Rossb,
Nicole T. Nelsonb and
Tyson D. Carlsonb
a Soil, Water and Environmental Science Department and Hydrology and Water Resources Department, University of Arizona, Tucson, AZ 85721
b Hydrology and Water Resources Department, University of Arizona, Shantz Bldg. 38, Rm. 429, University of Arizona, Tucson, AZ 85721
* Corresponding author (brusseau{at}ag.arizona.edu)
Received 16 October 2002.
 |
ABSTRACT
|
|---|
Gas-phase partitioning tracer tests were conducted at a fuel depot to evaluate the utility of the partitioning tracer method for characterizing organic immiscible-liquid contamination in the vadose zone. One test was conducted within the boundaries of an existing fuel dispensing island and former underground fuel tank facility. Core sampling indicated that the vadose zone at this location is heavily contaminated throughout its vertical extent by fuel-related hydrocarbons. A tracer test was conducted after 30 mo of operation of a soil vapor extraction (SVE) system. Analysis of hydrocarbon concentrations in the SVE effluent indicates that approximately 355 000 L of hydrocarbons were recovered during the 30-mo operation period. Comparing this value to the initial volume of hydrocarbons present, estimated to be approximately 454 000 L based on core data, produces an estimate of 99 000 (64 000–134 000) L of hydrocarbons remaining within the area influenced by the SVE system. Extrapolation of the tracer test results (Sn = 0.44%) to the SVE-impacted zone produces an estimate of 107 000 (59 000–155 000) L of hydrocarbon present. The two values and associated approximate uncertainty ranges are relatively similar. The second test was conducted approximately 200 m from the former tank facility. Analysis of borehole material collected during well drilling at this location indicates minimal hydrocarbon contamination of the vadose zone, except in the vicinity of the water table (approximately 30 m below ground surface [bgs]), on which floating free product is found. Analysis of the tracer test results produced a hydrocarbon saturation value of 0.37%, which is significantly smaller than the initial value of approximately 1.9% (i.e., before SVE operation) estimated based on core data for the former tank facility location. The lower level of vadose-zone contamination suggests that the source of immiscible-liquid contamination found at the second location may be related to lateral migration of floating free product along the water table from the up-gradient tank facility, rather than vertical migration from the surface above. However, additional studies would be required to more fully evaluate this hypothesis. The results of this work illustrate the utility of the gas-phase partitioning tracer method for characterizing immiscible-liquid contamination in the vadose zone.
Abbreviations: BCF, bromochlorodifluoromethane • bgs, below ground surface • PVC, polyvinyl chloride • SVE, soil vapor extraction
|
INTRODUCTION
|
|---|
THE IMPACT OF ORGANIC, immiscible-liquid contamination, often referred to as "nonaqueous phase liquid", on the characterization and cleanup of contaminated sites is now well recognized. Immiscible liquids act as long-term sources of both vapor-phase and groundwater contamination. Clearly, an accurate assessment of the risk posed by contamination at a particular site cannot be made without knowledge of the occurrence and distribution of immiscible liquids in the vadose zone. Furthermore, it has been clearly demonstrated that groundwater remediation programs will not be successful unless source zones are controlled, including those in the vadose zone. Thus, accurate risk assessments and successful subsurface-remediation efforts are dependent on an accurate characterization of immiscible-liquid contamination in the vadose zone.
The use of traditional methods for immiscible-liquid characterization, including core, soil-gas, and pore-water sampling, is limited for large field sites because of the need for what is usually a cost-prohibitive amount of sampling to account for subsurface heterogeneity. The partitioning tracer method is an alternative method for larger-scale characterization of immiscible-liquid contamination in the subsurface. Partitioning tracers have been used since the 1970s in the petroleum industry to determine residual oil saturations (Tang, 1995). More recently, the use of partitioning tracer tests to characterize immiscible-liquid contamination in environmental systems has been demonstrated in laboratory and field studies for both saturated (e.g., Jin et al., 1995; Wilson and Mackay, 1995; Nelson and Brusseau, 1996; Annable et al., 1998; Brown et al., 1999; Cain et al., 2000; Meinardus et al., 2002) and vadose zone (Simon et al., 1998; Deeds et al., 1999; Mariner et al., 1999; Whitley et al., 1999) systems. While the effectiveness of the partitioning tracer method has been evaluated in several studies for saturated-zone applications, such evaluations are more limited for vadose zone applications. The purpose of the field tests reported herein was to evaluate the use of the gas-phase partitioning tracer method for characterizing immiscible-liquid contamination in the vadose zone beneath a fuel depot.
|
BACKGROUND
|
|---|
Partitioning Tracer Method
The conceptual and theoretical basis for describing the transport of tracers that partition to immobile, immiscible fluids has been presented in detail elsewhere (e.g., Brusseau, 1992; Jin et al., 1995; Brusseau et al., 2003). The procedure for estimating immiscible-liquid saturation using data collected with a partitioning tracer test involves calculation of a retardation factor, R, for the partitioning tracer, which is usually done by a comparative moment analysis with a nonreactive tracer. The retardation factor is defined as the quotient of the travel times of the partitioning and nonreactive tracers. This R determined from the tracer test is equated to the mass-balance definition of R, given by
 | [1] |
where Kng is the immiscible-liquid–gas partition coefficient, Kd is the soil–water partition (sorption) coefficient, Kh is gas–water partition (Henry's) coefficient, 
b is bulk density of the porous medium, 
w is volumetric water content, g is volumetric gas content, and n is volumetric immiscible-liquid content. When the partitioning tracer is minimally retained by the soil and water phases, Eq. [1] can be simplified to
 | [2] |
where saturations (Si = volume phase i/volume of pore space) have been used in place of volumetric fluid contents, and where Sw + Sg + Sn = 1.
Inspection of Eq. [2] reveals that once a value for R is determined from the tracer test, values for Sn can be obtained with knowledge of only the immiscible-liquid–gas partition coefficient and the water saturation. Partition coefficients can be measured in laboratory experiments and can be estimated in some cases. If the partitioning tracer is retained by water, or is sorbed by the soil, additional parameters are required (see Eq. [1]). When interpreting the results of a partitioning tracer test, it is important to recognize that the saturation values obtained from the test are "large-scale" values, representing an averaging across the measured domain. The sensitivity of the partitioning tracer method for measuring immiscible-liquid saturation is a function of the area of influence of the tracer test (swept volume), the amount of immiscible liquid in that swept volume, the immiscible-liquid–gas partition coefficient of the tracer, and constraining factors (rate-limited mass transfer, heterogeneity, mass loss, analytical limitations).
The partitioning tracer method is implemented using relatively standard tracer test techniques. By virtue of sampling a much larger volume of the subsurface than point-sampling methods, the partitioning tracer method generally has a greater chance of detecting immiscible-liquid saturation. Another advantage of this method is the absence of significant depth of measurement limitations, unlike coring, which generally becomes economically and technically constrained for deeper systems. This is a distinct advantage for applications in deep vadose zones, which are common in the western USA.
Site Background
The City of Tucson fuel depot, the site at which the tests were conducted, began operations in 1974, providing fuel for city vehicles (see Fig. 1). The site originally contained 23 underground tanks to store and dispense fuels, including regular and unleaded gasoline and diesel fuel. In June of 1989, fuel was observed to be leaking from the tanks at the northern end of the pumping station. All underground storage tanks were removed by 1992 (HydroGeoChem, Inc., 1994). In February 1996, a SVE system was installed to remediate the contaminated soil. The extracted gas passes through a thermal oxidizing unit for treatment.
The vadose zone at the site primarily consists of low permeability sandy silts to silty clays, with a few interspersed zones of silty sands. The vadose zone extends to approximately 30 m bgs, where a shallow saturated zone is encountered. The hydraulic gradient for the shallow zone is north-northwest across the site (see Fig. 1). Floating free product is observed in several wells emplaced in the shallow saturated zone.
Two sets of core-sampling studies (HydroGeoChem, Inc., 1994) were conducted in an area within which are the existing fuel dispensing island and former underground fuel tank facility (approximate sampled area encompasses the lower left-hand block in Fig. 2). The first set comprised 33 borings drilled to between approximately 12 and 18 m bgs, while the second set comprised 28 borings drilled to 30 or more m bgs. For both sets, samples were collected at 1.5- to 3-m intervals, for a total of approximately 600 samples. Based on the core-sampling results, the vadose zone at the former fuel tank facility location was heavily contaminated throughout its vertical extent by fuel-related hydrocarbons, and the areal extent of the contaminated zone was initially approximately 5000 m2. Concentrations of total petroleum hydrocarbons ranged from approximately 10 to approximately 50 000 mg kg-1, with an average value of 2992 mg kg-1. Given these high concentrations and the properties of the porous media, it is assumed that essentially all of the contaminant mass is present as immiscible liquid. Using the mean hydrocarbon concentration, a porosity of 0.3, and a hydrocarbon density of 0.8 kg L-1, a volume of approximately 454 000 L of hydrocarbons is estimated as initially being present at the site within the zone of influence of the SVE system (i.e., a depth interval of approximately 17 m). This translates to an equivalent hydrocarbon saturation of approximately 1.9%, which is consistent with other large-scale values reported for field systems (e.g., Ostendorf et al., 1991; Poulsen and Kueper, 1992; Cohen and Mercer, 1993; Deeds et al., 1999).
|
MATERIALS AND METHODS
|
|---|
Two gas-phase partitioning tracer tests were conducted at the site (see Fig. 1 and 2). One test was conducted within the boundaries of the existing fuel dispensing island and former underground fuel tank facility. As noted above, extensive hydrocarbon contamination occurred throughout the vadose zone at this location, and a SVE system had been in operation for 30 mo before the test. The second test was conducted approximately 200 m from the former tank facility. Analysis of borehole material collected during well drilling at this location indicated minimal (trace or nondetect) hydrocarbon contamination of the vadose zone as measured by a portable photoionization detector, except in the vicinity of the shallow saturated-zone water table, on which floating free product is found. No remedial actions had been conducted at the second location before the tracer test.
Sulfur hexafluoride (SF6) was used as the nonreactive, nonpartitioning tracer and bromochlorodifluoromethane (BCF) was used as the partitioning tracer. Note that while SF6 has been used as an immiscible-liquid partitioning tracer for saturated-zone applications (Wilson and Mackay, 1995; Nelson and Brusseau, 1996), it serves as a nonpartitioning tracer for gas-phase vadose zone applications (e.g., Mariner et al., 1999). The nondimensional Henry's coefficients for SF6 and BCF are 132 (Olschewski et al., 1995) and 2 (measured), respectively. The nondimensional immiscible-liquid–gas partition coefficient for BCF was measured to be 62, based on the results of a series of batch experiments conducted using immiscible liquid collected from the site. The magnitude of the Henry's coefficient for BCF suggests that partitioning to soil water may contribute to BCF retention. Simple dimensional analysis (with Kng = 62 and Kh = 2) shows that the water saturation would have to be approximately 120 times the immiscible-liquid saturation for water retention to contribute roughly equally to tracer retention. This would equate to a water saturation of about 20% for this case, which is a reasonable field value. However, analysis of the tracer-test results to be discussed below indicated that partitioning to soil water did not significantly influence tracer transport. This is not unexpected given the limited precipitation and SVE operations before the tracer tests.
For the first test, conducted in the vicinity of the former tank facility, a large screened interval was used for the extraction well (E-1), extending from 9 to 24 m bgs, to interrogate a large portion of the vadose zone. The injection (I-1) and extraction wells are 9.3 m apart. The well casing is polyvinyl chloride (PVC) schedule 40 pipe, 3.8 cm in diameter. An air compressor was used to supply ambient air to the injection well. A flow meter was connected to the system near the injection well to measure the injection flow rate. A gas cylinder containing the gas tracer mixture was connected to the system with a 1-cm-diam. stainless-steel tube, and the flow was controlled by a flow meter and a ball valve. An anemometer (TSI 8355 Velocicalc Velocity Meter, TCI, Inc., Shoreview, MN) was attached to the extraction well through an access point at the top of the PVC riser pipe.
A flow field was created using ambient airflow rates of approximately 0.3 and 0.6 std-m3 min-1 for injection and extraction, respectively. The flow rates were measured every 5 min. Pressures were measured and recorded every 5 min with pressure transducers, installed in the injection and extraction wells, connected to a data logger (Campbell CR10X, Campbell Scientific, Logan, UT) and laptop computer. Steady-state flow conditions were indicated when pressure variances were minimal and flow rates were stable. The tracer mixture was then injected for 15 min, during which time approximately 4 m3 of tracer gas were injected into the subsurface. Injection concentrations of 40 ppmv were used for both BCF and SF6. The SVE system was not in operation during the tracer test.
Samples were collected from ports attached to the injection and extraction wells to monitor injection and effluent concentrations, respectively. The sampling ports consisted of septum injector nuts (Valco Instruments Co. Inc., Houston, TX) connected to each well. Gas samples were collected by withdrawing samples through the septum into a needle-tipped syringe. The gas samples were then injected from the syringe into evacuated 80-mL aerosol canisters (Tracer Research, Tucson, AZ). Samples were analyzed for SF6 and BCF using a gas chromatograph equipped with an electron capture detector (Shimadzu GC-17, Shimadzu Scientific, Kyoto, Japan), coupled with a headspace autosampler (Tekmar 7000, Tekmar, Cincinnati, OH). The quantifiable detection limits for BCF and SF6 were 0.05 and 0.04 µg L-1, respectively.
The methods used for the second tracer test, conducted approximately 200 m from the former tank facility, were similar to those used for the first test. The injection (I-2) and extraction (E-2) wells are approximately 23 m apart. In this case, the test was conducted in a narrow interval, isolated between 24 and 27 m bgs with the use of packers, near the bottom of the vadose zone. The duration of the tracer injection was approximately 26 min, during which time approximately 7.3 m3 of tracer gas were injected into the subsurface.
The tracer breakthrough curves were analyzed by calculating the zeroth and first temporal moments to quantify mass recovery, travel time, and retardation (e.g., Jin et al., 1995; Cain et al., 2000). The travel time values were calculated using log-linear extrapolation of the elution tails.
|
RESULTS
|
|---|
Pressures and flow rates were stable throughout the course of both tests. The breakthrough curves for SF6 and BCF are shown in Fig. 3 and 4 for Locations 1 and 2, respectively. Mass recoveries were 44/30% and 35/22% for SF6 and BCF for Tests 1 and 2, respectively. Retardation factors of 1.3 and 1.2 were calculated for the partitioning tracer, BCF, for Tests 1 and 2, respectively. These values translate to immiscible-liquid saturations of 0.44 and 0.37% for Locations 1 and 2, assuming Sw = 0.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 3. Results for Tracer Test 1, conducted within the boundaries of a former fuel tank facility. Breakthrough curves for SF6 and BCF.
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 4. Results for Tracer Test 2, conducted approximately 200 m from the former fuel tank facility. Breakthrough curves for SF6 and BCF. Lines are used in place of data symbols to help visualize the dual-peak behavior exhibited by the data.
|
|
Extrapolation of the saturation value obtained for the first tracer test to the contaminated zone influenced by the SVE system produces an estimate of 107 000 L of hydrocarbon present in the SVE-impacted zone. This estimate is influenced by several sources of uncertainty, including uncertainty associated with the Sn value itself (due primarily to uncertainty in R and Kng values), with its representativeness of regions outside the tracer-swept zone, and with the size of the contaminated zone. For example, Meinardus et al. (2002) reported uncertainly values for Sn ranging from 11 to 67%. For this study, using an estimated uncertainty value of 45%, the volume of hydrocarbons present is estimated to range from approximately 59 000 to 155 000 L. Recognizing this uncertainty, the estimated volume can be compared with a value obtained independently from analysis of soil core and SVE effluent data. As of the date of the partitioning tracer test, approximately 355 000 L of hydrocarbons had been removed from the vadose zone by the SVE system (Fluor Daniel, G.T.I., Tempe, AZ, personal communication, 1998). Comparing the mass-removed value to the initial volume estimate of 454 000 L, based on the core data, an estimated 99 000 L of hydrocarbons remained in the SVE-impacted area of the vadose zone at the time of the tracer test. A range of 64 000 to 134 000 L is estimated based on an approximated uncertainty value of 35%, which incorporates uncertainties in SVE effluent and core analyses as well as the size of the domain. The two values and associated approximate uncertainty ranges are relatively similar, suggesting that the gas-phase partitioning tracer method produced a reasonable measure of in situ hydrocarbon saturation.
Both the SF6 and BCF breakthrough curves obtained from the tracer test conducted at Location 2 exhibit dual-peak behavior (see Fig. 4), indicating the presence of distinct flowpaths of differing permeability. Inspection of Fig. 4 reveals that the first BCF peak is retarded and exhibits reduced concentrations compared with the SF6 peak. Conversely, the second BCF peak is essentially coincident with the SF6 peak, exhibiting minimal retention. This differential retardation behavior observed for BCF suggests that the petroleum-hydrocarbon saturation is nonuniformly distributed within the swept zone. In addition, the minimal retardation associated with the second BCF peak indicates there was negligible retention of that tracer mass, either by petroleum-hydrocarbon saturation or by soil water. The latter observation suggests that the impact of soil water on BCF retention was indeed not significant for the tracer tests.
Analysis of the results of the second tracer test produced a hydrocarbon saturation value of 0.37%, which is significantly smaller than the initial value of approximately 1.9% (i.e., before SVE operation) estimated for the former storage tank area, based on core data, located approximately 200 m away. The lower level of contamination is consistent with the analyses of borehole material collected during well drilling at the second location. No remedial actions had been conducted at the second location before the tracer test. Thus, the lower level of vadose zone contamination suggests that the source of immiscible-liquid contamination (including floating free product) found at this location may be related to lateral migration of free product along the shallow saturated-zone water table from the up-gradient tank facility location (see Fig. 1), rather than vertical migration from the surface above. However, additional studies would be required to more fully evaluate this hypothesis.
|
CONCLUSIONS
|
|---|
On the basis of the results presented above, we conclude that the gas-phase partitioning tracer method appears to have been successful in detecting and measuring petroleum hydrocarbon contamination in the vadose zone beneath a fuel depot in Tucson, AZ. The results of this work illustrate the utility of the gas-phase partitioning tracer method for characterizing immiscible-liquid contamination in the vadose zone. By virtue of sampling a much larger volume of the subsurface than is done in point-sampling methods, the partitioning tracer method generally has a greater chance of detecting immiscible-liquid saturation. Another attribute of this method is the absence of depth limitations, which is a distinct advantage for deep vadose zones, such as are common in Arizona and other western states of the USA, wherein the application of point-sampling methods would generally be cost prohibitive.
|
ACKNOWLEDGMENTS
|
|---|
This research was supported by the City of Tucson Office of Environmental Management, by the U.S. Environmental Protection Agency, and by the NIEHS Superfund BRP. We thank the City of Tucson (Karen Masbruch and Richard Byrd), Fluor Daniel (Curtis Wright and Jim Goetz), and University of Arizona personnel (Brent Cain, and others) for their assistance.
|
REFERENCES
|
|---|
- Annable, M.D., P.S.C. Rao, K. Hatfield, W.D. Graham, A.L. Wood, and K.G. Enfield. 1998. Partitioning tracers for measuring residual NAPL: Field-scale test results. J. Environ. Eng. 124:498–503.
- Brown, C.L., R.E. Jackson, M. Delshad, and G.A. Pope. 1999. Demonstration of sufactant flooding of an alluvial aquifer contaminated with dense nonaqueous phase liquid. p. 64–85. In M.L. Brusseau et al. (ed.) Innovative subsurface remediation: Field testing of physical, chemical, and characterization technologies. American Chemical Society, Washington, DC.
- Brusseau, M.L. 1992. Rate-limited mass transfer and transport of organic solutes in porous media that contain immobile immiscible organic liquid. Water Resour. Res. 28:33–35.
- Brusseau, M.L., N.T. Nelson, and M.S. Costanza-Robinson. 2003. Partitioning tracer tests for characterizing immiscible fluid saturations and interfacial areas in the vadose zone. Available at www.vadosezonejournal.org. Vadose Zone J. 2:138–147 (in this issue).[Abstract/Free Full Text]
- Cain, R.B., G.R. Johnson, J.E. McCray, W.J. Blanford, and M.L. Brusseau. 2000. Partitioning tracer tests for evaluating remediation performance. Ground Water 38:752–761.[ISI]
- Cohen, R.M., and J.W. Mercer. 1993. DNAPL site evaluation. CRC Press, Boca Raton, FL.
- Deeds, N.E., G.A. Pope, and D.C. Mckinney. 1999. Vadose zone characterization at a contaminated field site using partitioning interwell tracer technology. Environ. Sci. Technol. 33:2745–2751.
- HydroGeoChem, Inc. 1994. Corrective action plan for soil and groundwater. P prepared for COT-OEM, 16 Dec. 1994. Thomas O. Price Service Center, Tucson, AZ.
- Jin, M., M. Delshad, V. Dwarakanath, D.C. McKinney, G.A. Pope, K. Sepehrnoori, and C.E. Tilburg. 1995. Partitioning tracer test for detection, estimation, and remediation performance assessment of subsurface nonaqueous phase liquids. Water Resour. Res. 31:1201–1211.
- Mariner, P.E., M. Jin, J.E. Studer, and G.A. Pope. 1999. The first vadose zone partitioning interwell tracer test for nonaqueous phase liquid and water residual. Environ. Sci. Technol. 33:2825–2828.
- Meinardus, H.W., V. Dwarakanath, J. Ewing, G.J. Hirasaki, R.E. Jackson, M. Jin, J.S. Ginn, J.T. Londergan, C.A. Miller, and G.A. Pope. 2002. Performance assessment of NAPL remediation in heterogeneous alluvium. J. Contam. Hydrol. 54:173–193.[ISI][Medline]
- Nelson, N.T., and M.L. Brusseau. 1996. Field study of the partitioning tracer method for detection of dense nonaqueous phase liquid in a trichloroethene-contaminated aquifer. Environ. Sci. Technol. 30:2859–2863.
- Olschewski, A., U. Fischer, M. Hofer, and R. Schulin. 1995. Sulfur hexafluoride as a gas tracer in soil venting operations. Environ. Sci. Technol. 29:264–266.
- Ostendorf, D.W., L.E. Leach, E.S. Hinlein, and Y. Cie. 1991. Field sampling of residual aviation gasoline in sandy soil. Ground Water Monitor. Remed. 11:107–120.
- Poulsen, M.M., and B.H. Kueper. 1992. A field experiment to study the behavior of tetrachloroethylene in unsaturated porous media. Environ. Sci. Technol. 26:889–895.
- Simon, M., M.L. Brusseau, R. Golding, and P.J. Cagnetta. 1998. In Proc. Inter. Conf. Remediation of Chlorinated and Recalcitrant Compounds. Monterey, CA. 18–21 May. Battelle, Columbus, OH.
- Tang, J.S. 1995. Partitioning tracers and in-situ fluid-saturation measurements. SPE Form. Eval.10(March):33–39.
- Whitley, B.A., D.C. McKinney, G.A. Pope, B.A. Rouse, and N.E. Deeds. 1999. Contaminated vadose zone characterization using partitioning gas tracers. J. Environ. Eng. 125:574–582.
- Wilson, R.D., and D.M. Mackay. 1995. Direct detection of residual nonaqueous phase liquid in the saturated zone using SF6 as a partitioning tracer. Environ. Sci. Technol. 29:1255–1258.
This article has been cited by other articles:
|
|
|
|
 
M. L. Brusseau, M. L. Brusseau, N. T. Nelson, and M. S. Costanza-Robinson
Partitioning Tracer Tests for Characterizing Immiscible-Fluid Saturations and Interfacial Areas in the Vadose Zone
Vadose Zone J.,
May 1, 2003;
2(2):
138 - 147.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
