1998_annable_hill tracers

1998_annable_hill tracers - PARTITIONING TRACERS FOR...

Info icon This preview shows pages 1–6. Sign up to view the full content.

View Full Document Right Arrow Icon
Image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 2
Image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 4
Image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 6
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: PARTITIONING TRACERS FOR MEASURING RESIDUAL NAPL: FIELD-SCALE TEST RESULTS By Michael D. Annable,l Associate Member, ASCE, P. S. C. Rao,2 Kirk Hatfield,3 Associate Member, ASCE, Wendy D. Graham,4 A. L. Wood,‘ and C. G. Enfield‘ ABSTRACT: The difficult task of locating and quantifying nonaqueous phase liquids (NAPLs) present in the vadose and saturated zones has prompted the development of innovative, nondestructive characterization tech- niques. The use of the interwell partitioning tracer’s (IWPT) test, in which tracers that partition into the NAPL phase are displaced through the aquifer, is an attractive alternative to traditional coring and analysis. The first field test of IWPT was conducted in a hydraulically isolated test cell (3.5 by 4.3 m) to quantify the total amount of a complex NAPL (a mixture of JP-4 jet fuel and chlorinated solvents) trapped within a 1.5-m smear zone in a shallow, unconfined sand and gravel aquifer at Hill Air Force Base (AFB), Utah. Tracer breakthrough curves (BTCs) were measured in three extraction wells (EWs) following a tracer pulse (0.1 pore volume) introduction through four injection wells (IWs). The measured retardation of the partitioning tracer (2,2-dimethyl-3-pentanol) relative to the nonreactive tracer (bromide) was used to quantify the NAPL present. The EW data were used to estimate an average NAPL saturation of 4.6—5.4% within the test cell. NAPL saturations estimated by using measured concentrations in soil cores of two significant compounds present in the NAPL were 3.0 and 4.6%. INTRODUCTION Quantifying the total amount, composition, and spatial dis- tribution of nonaqueous phase liquids (NAPLs) trapped in po- rous media at hazardous waste sites is of paramount impor- tance in determining environmental impacts and in selecting appropriate remediation alternatives. NAPLs, both lighter and denser than water, can be present in the vadose and saturated zones and can act as long-term sources contributing to exten- sive ground-water contamination. Locating, quantifying, and delineating such NAPL source zones has been a difficult task facing contaminant hydrologists. The use of partitioning trac- ers presents an attractive alternative to the traditional intrusive techniques that require soil coring and analysis (Pope et al. 1994a; Jin et al. 1995; Wilson and Mackay 1995). By selecting tracers that partition into the NAPL phase with predictable or measurable relationships, it is possible to determine the NAPL quantity present in the tracer flow field. Laboratory-scale batch and column tests have shown the potential use of long-chain alcohols as partitioning tracers for quantifying NAPL content in aquifers and of perfluorocarbons for accomplishing the same in the vadose zone [see Jin et al. (1995)]. In this paper, results from the first field test of partitioning tracers to quantify residual NAPLs in the saturated zone at an NAPL-contaminated site are reported. The site is located at Hill Air Force Base (AFB) in Utah. The use of an interwell partitioning tracer’s (IWPT) test at the site was motivated by the need to quantify the initial volume of residual NAPL pres- ent in an isolation test cell. The test cell was installed for the purpose of evaluating the use of in-situ flushing with cosol- lAsst. Prof., Dept. of Envir. Engrg. Sci., Univ. of Fla, Gainesville, FL 32611. E-mail: [email protected]fl.edu. 2Grad. Res. Prof, Soil and Water Sci. Dept. Univ. of Fla., Gainesville. FL. 3Assoc. Prof., Dept. of Civ. Engrg., Univ. of Fla., Gainesville, FL. ‘Assoc. Prof.. Dept. of Agric. and Biol. Engrg.. Univ. of Fla, Gaines- ville, FL. ’Res. Soil Sci., Nat. Risk Mgmt. Res. Lab., U.S. Envir. Protection Agency, Ada, OK 74820. 6Sr. Res. Envir. Sci., Nat. Risk Mgrnt. Res. Lab., U.S. Envir. Protection Agency. Ada, OK. Note. Associate Editor: Hilary I. Inyang. Discussion open until No- vember 1. 1998. To extend the closing date one month, a written request must be filed with the ASCE Manager of Journals. The manuscript for this paper was submitted for review and possible publication on April/ May, 1996. This paper is part of the Journal of Environmenml Engi- neering, Vol. 124. No.6, June, 1998. ©ASCE, ISSN 0733-9372/98/0006- 0498—0503/$8.00 + $.50 per page. Paper No. 17661. 498 / JOURNAL OF ENVIRONMENTAL ENGINEERING /JUNE 1998 vents for remediation of a sand-gravel-cobble aquifer contam- inated with a complex (mixed waste) NAPL. The objectives of the partitioning tracer test were to demonstrate the method of application for quantifying both the residual NAPL volume and distribution within the test cell. The results were compared to estimates of NAPL volume based on the soil core data col- lected. These results were then used for simulation and opti- mization of the cosolvent flushing experiment design. PARTITIONING TRACERS: THEORETICAL BASIS Jin et a1. (1995) present in detail the experimental and the- oretical basis for the use of partitioning tracers. A brief sum— mary is provided here. The theory behind partitioning tracers for quantification of NAPL saturation and volume is based on well-characterized tracer partitioning relationships between the aqueous and NAPL phases, the simplest approach being a lin— ear relationship: C KNw=C—:: (1) where CN = concentration of the tracer in the NAPL phase; C, = tracer concentration in the aqueous phase; and KNW = NAPL- water partition coefficient. Partitioning refers to the distribu- tion of tracer between the NAPL and water, and it can be predicted using the UNIFAC-solubility method [see Wang et al. (1996)]. This partitioning relationship can be introduced into the advection-dispersion equation describing solute trans- port in porous media. In a steady-state flow field with phase partitioning according to (l), the effect is to delay or retard the partitioning tracer transport rate. This is manifested by a partitioning tracer breakthrough curve (BTC) for which the delayed average travel time is a function of the average NAPL saturation, SN, within the tracer flow field (Pope et al. 1994a; Jin et a]. 1995): t R: 1 + ——K””S” = _p (1 — SN) t. (2) where R = retardation factor for the partitioning tracer, which is calculated as the ratio of average travel times for the par- titioning tracer pulse (tp) and the nonpartitioning tracer pulse (tn) introduced and displaced simultaneously during steady flow. Sorption onto solid particles is assumed insignificant. This relationship forms the basis for determining unknown NAPL volumes present in an aquifer and requires the esti- mation of average travel times (t, and 1,.) using first temporal moments of the BTCs for partitioning and nonpartitioning tracers measured in effluent from each extraction well (EW). The effective pore volume for each EW can be determined using the nonreactive tracer average travel time, t,,_,: VeJ = Qt [mi (3) where Q, = pumping rate from EW 1'. The NAPL saturation, SW, for the swept volume of each EW, can be determined using (Jin et a1. 1995) (th _ mi) [th _ tn.i(KNw T 1)] The total NAPL volume (VNJ) in the swept zone of each well is then given by SNJ = (4) _ SNJVtJ _ 1 — Sm The total for all EWs can then be summed (i = 1, ..., k). Tracers should be selected to provide adequate retardation to give reliable estimates of NAPL saturation but not so large that the tracer test duration is unreasonable. Jin (1995) rec- ommends a range of 1.2 < R < 4. It is advisable to estimate SN values based on breakthrough data for multiple partitioning tracers. Vm (5) FIELD EXPERIMENT Motivation for Study The initial plan for characterization of the test cell was pri- marily by soil core samples. Within the test cell, seven screened wells were installed for injection and extraction of fluids, and 12 multilevel samplers (MLSs) were installed (Fig. 1) to monitor the progression of tracers and contaminant con- centration changes within the flow domain during cosolvent flushing. The MLSs were constructed using stainless-steel tub- ing terminating at different depths with a stainless-steel filter and were installed with a 25-t cone penetrometer truck (CPT) using an expendable drive tip that was left ahead of the sam- pler bundle when the rod was removed. The initial plan called Sheet Pile Extends 3.0 in into Clay Layer <I> [W3 Injection Well 93 EW3 Extraction Well 0 1 Multilevel Sampler (MLS) 0 1 MLS With Piezometer FIG. 1. Schematic Diagram of Test Cell Installed at HIII AFB, Which Was Used for Conducting Partitioning Tracers Test for CPT collection of soil samples to minimize media distur- bance during installation of the MLSs. However, because of large cobbles, CP’I‘ soil—sample collection was not feasible. Soil samples taken during the installation of injection wells (IWs), EWs, and MLSs with piezometers provided a substan- tial set of soil cores (102 samples in the NAPL smear zone obtained during well installation by hollow-stem auger) avail- able for preremediation site characterization of NAPL distri- bution. The lack of CPT core samples in a portion of the test cell, in part, stressed the need for an integrated nonintrusive partitioning tracer test that provided a measure of the NAPL content on the basis of a much larger sample volume. Site History and Characteristics Operable Unit 1 (OU-l) at Hill AFB has a number of con- taminant sources located across the site. The remediation test cell is located hydraulically downgradient of two chemical dis— posal pits which were predominantly used to dispose of avi- ation fuels (JP-4) and chlorinated solvents used during the 19405 and 19505. Upgradient of the cell is a former fire-train- ing area, which may have contributed unextinguished fuels and combustion by-products to the site. The resulting NAPL is lighter than water and exists as a plume covering several hec- tares delineated by up to 0.15 m of free product measured in wells. Test Cell Installation and Instrumentation An isolation test cell was constructed at the OU-l site using the Waterloo sealable sheet pile barrier system (Starr et al. 1992, 1993) (Fig. l). The cell measures 3.5 by 4.3 m at the ground surface, and the sheet piles extend to a depth of 9.1- m below ground surface (BGS). The confining unit, consisting primarily of clay with fine sand stringers, is encountered at ~6.1-m BGS. The cell joints were sealed with a special grout (a mixture of fly ash, concrete, and other proprietary constit- uents; supplied by the Waterloo group) prior to installation of wells and MLSs. Prior to the cell installation, wells drilled for locating the cell indicated that the water table was encountered approxi- mately 5.8-m BGS. Results from field monitoring of soil cores using a photoionization detector and observing dark black col- oration of the soil indicated that the NAPL was present in a smear zone between 4.6- and 6.1-m BGS. The highest levels of contamination were present in the bottom portion of this zone near the existing water table. The vertical extent of the NAPL smear zone is consistent with the historical records of seasonal water-table variation (1.5 m) at this site. Based on this information, MLSs were installed in the test cell in a uni- form grid pattern with vertical positions from 4.6- to 6.4-m BGS at a uniform spacing of 0.38-m. Four injection wells (IWl—4) and three extraction wells (EW1—3) were installed at opposite ends of the test cell. These wells were constructed with 3.0-m-Iong, 5.1-cm stainless-steel slotted screen with the bottom placed at 6.1-m BGS. Following the well installation, the cell was hydraulically tested by rais- ing the water level in the cell from approximately 5.2- to 4.6- m BGS. Based on the measured water-level rise induced by the injection of 1,740 L of water, the effective porosity, (1),, was estimated to be 0.20. Tracer Experiments A preliminary test with two nonsorbing tracers (bromide and pentafluorobenzoate) was conducted to evaluate the hydrody- namic properties of the test cell. Results of the tracer test pro- vided an estimated pore volume of 4.37 m3 based on bromide and 4.11 1113 based on pentafluorobenzoate. The results of that JOURNAL OF ENVIRONMENTAL ENGINEERING /JUNE 1998 / 499 TABLE 1. Tracer Partltlon Coefflclents (Km) Measured In Batch Tests Uslng NAPL from HIII AFB Test Cell and Tracer Con- centratlons Used In Fleld Study Tracer injection concentration (mg/L) (3) Bromide 273 Ethanol 1330 n-Pentanol 989 n-Hexanol 945 2,2-Dimethyl-3-pentanol 878 n-Heptanol n-Octanol tracer test, including tracer arrival times for the MLSs, were used to design the partitioning tracers test. The partitioning tracers were chosen based on the KM, values measured in batch and column tests conducted using the NAPL from the OU-l test site (Pope et al. 1994b). Table 1 lists the alcohol tracers considered and the measured KM, values. Other desirable tracer characteristics include ( 1) nonhazardous; (2) nontoxic; (3) nondegrading; (4) low volatility; (5) reasonable cost and avail- ability; and (6) easily quantifiable, especially in the presence of numerous NAPL constituents. The tracers selected for this study were bromide (nonreactive tracer applied as KBr), eth- anol, n-pcntanol, n-hexanol, and 2,2-dimethyl-3-pentanol. The partitioning tracer test was conducted over an eight—day period in October 1994, during which time water samples were collected at frequent intervals from the EWs and MLSs. The sampling schedule was based on data from the preliminary tracer test and the partitioning characteristics of the tracers, as well as the results from one-dimensional and three-dimen- sional simulation models [simulations were performed at the University of Florida using a three-dimensional numerical code under development and at the University of Texas at Aus- tin using UTCHEM (Pope et al. 1994b)]. An ~O.l pore vol- ume pulse of a mixture of the tracers was introduced over a 3.3-h period after steady flow had been established for ~3 pore volumes; tap water from the Hill AFB was used for this study. The original source of tap-water on the base is from a deeper aquifer that has a similar total dissolved solids concen- tration compared to the surficial aquifer. Flow was maintained using three dedicated pumps for the EWs and a single pump with a four-head peristaltic cartage with independent lines to each of the IWs. Steady, uniform flow was maintained by continually monitoring and adjusting flow rates throughout the test. Tracer solution was mechani- cally mixed and delivered uniformly to the four IWs for the same duration. The water-table level in the cell was maintained at 4.57 I 0.03-m BGS. The average flow rate during the tracer study was 3.2 L/min, equivalent to an average Darcy velocity of ~0.03 m/h, or an average hydraulic residence time in the test cell of ~1 (1. Based on the measured hydraulic gradient of 0.05 within the cell, an average saturated hydraulic con- ductivity of 17 m/d (2.0 X 10" m/s) was calculated. Analytlcal Methods Water samples were collected from the EW lines immedi- ately after the pumps. These were cooled and shipped in quan- tity by overnight freight to the University of Florida for anal- ysis. At the University of Florida, samples were stored in a refrigerator or cold room at 5°C until analyzed. Alcohol tracers were analyzed by direct injection onto a gas chromatograph (J&W RTX-624 capillary column, 0.53-mm i.d., 70-m long) equipped with a flame-ionization detector. Bromide concentra- tions in the samples were analyzed using ion chromatography SOD/JOURNAL OF ENVIRONMENTAL ENGINEERING /JUNE 1998 (Dionex Ionpac AS4—SC analytical column, 4-mm id, 0.25- m long, conductivity detector). RESULTS AND ANALYSIS Interpretation of EW Data The results presented here focus on the samples collected from the three EWs. The data were used to estimate the av- erage NAPL saturation and the total NAPL volume within the swept zones for each of the three EWs. Data from the MLSs were used to estimate the spatial distribution of NAPL within the test cell. Fig. 2 shows the BTCs for all tracers used for EW3. Retar- dation of the partitioning tracers compared to bromide is taken as evidence for the presence of residual NAPL. BTCs mea- sured in laboratory columns using uncontaminated soils from the site show very low retardation for all of the tracers used in the field study. The data for the partitioning tracers n-pen- tanol, n-hexanol, and 2,2-dimethyl-3-pentanol show increasing retardation proportional to KM, for each tracer as expected from (2). Loss of some of the alcohols (ethanol, n-pentanol, n-hexanol) based on zeroth moments, is likely due to bio- degradation during the study. The field data for bromide and 2,2-dimethyl-3-pentanol, which had the best mass recovery (102 and 92%, respectively) and temporal separation, were first analyzed by calculating zeroth and first temporal moments for quantifying mass re- covery, well swept volume, and average NAPL saturation. The BTCs are presented in Fig. 3 and the results are presented in Table 2. The calculated residual saturation, SN = 0.046, rep- resents an estimated NAPL volume of 0.239 m3 within the test cell. Comparison of SN estimates based on data for the indi- vidual EWs indicates a general trend of increasing NAPL con- tents from the EWl side to the EW3 side of the cell. This is consistent with the location of the suspected source of NAPL (i.e., the CPDs) relative to the cell. The total swept volume of 4.96 m3 yields a cell-average mobile water content of 0.21, which compares well with the value of 0.2 that was estimated based on hydraulic testing of the cell. Note also that of the total swept volume, approximately 41% is captured by one well (EWl) with the other wells at approximately 29% each. These data suggest that the EWl side of the test cell contains less permeable regions, when compared to the EW3 side. The average NAPL saturation based on the entire cell swept volume, SN = 0046—0054, is somewhat lower than an NAPL saturation of about 0.067 measured using partitioning tracers in composite core material taken from the cell (Johnson 1996). The composite sample was a combination of soil cores from the zone of highest contamination and was used to evaluate 0.2 4» n-pen ano + 2,2-dimethyl-3-pentanol + n-hexanol 0.15 V \ + bromide o 2 0 1 0 0.05 0 0 1000 2000 3000 4000 5000 6000 Effluent Volume (L) FIG. 2. EW3 BTCa Measured for Nonsorblng Tracers (Bromlde and Ethanol) and Partltlonlng Tracers (n-Pentanol. n-Hexanol. and 2,2-DImethyI-3-Pentanol) durlng Fleld Study Conducted In Test Cell at Hlll AFB 2,2-Dimethyl-3—pentanol 0 2 4 6 8 10 0 2 4 6 8 1 0 Effluent Volume (L)*1000 FIG. 3. Ew BTCs Measured for Nonsorblng Tracer (Bromide) and Partltlonlng 'n'ecer (2,2—Dlmethyl-s-Pentanol) durlng Fleld Study Conducted In Test Cell at HIII AFB cosolvent flushing. The difference between the field and the composite is likely due to the fact that the core samples used were taken from the zone with the heaviest NAPL contami- nation and, therefore, may not be representative of the entire cell. Extrapolation of Truncated Data The tracer test was continued until ~7 pore volumes of water had been displaced through the test cell. As evident in Fig. 2, the last samples collected still have concentrations of tracers (l—ppm range) that might contribute to the calculated moments. Truncated BTCs produce moments that underesti- mate the NAPL present. To improve the estimated NAPL sat- uration in the test cell an exponential extrapolation of the tracer data was performed. The tail of each BTCs (last 5—10 data points) was fit by a log-linear regression, which was then TABLE 2. Mess Recovery, Well Swept Volume, and NAPL Sat- uratlon tor EWs Br swept volume‘ (m3) (5) recovery recovery Note: Br = bromide and DMP = 2,2-dimethyl—3-pentanol. ‘Log-linear extrapolation. used to extrapolate the data to very low concentrations so that temporal moments calculated converged to a constant value. The extrapolated results are also presented in Table 2. Com- paring these results with results based on no...
View Full Document

{[ snackBarMessage ]}

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern