Grey-Iron Foundry Slags As Reactive Media for Removing Trichloroethylene from Groundwater

Transcription

1 Environ. Sci. Technol. 2009, 43, Grey-Iron Foundry Slags As Reactive Media for Removing Trichloroethylene from Groundwater DANIEL B. COPE Geosyntec Consultants, th Street, Suite 400, Oakland, California CRAIG H. BENSON* Department of Civil & Environmental Engineering, University of Wisconsin, 1415 Engineering Drive, Madison, Wisconsin Received May 16, Revised manuscript received October 13, Accepted October 17, A feasibility study was conducted using slags from six greyiron foundries to evaluate their potential as reactive media for permeable reactive barriers (PRBs) to remove aqueous trichloroethylene (TCE) from groundwater. Batch tests indicated that the slags exhibit varying degrees of reactivity ranging from nonreactive to reactivity comparable to that obtained with commercially available granular zerovalent iron on a surfacearea-normalized basis. TCE removal follows pseudo-firstorderkinetics, andproduceslesser-chlorinatedethenebyproducts (e.g., 1,1-DCE, cis-dce). Greater reactivity was obtained with the slags having the highest iron content and the lowest reactivity was obtained with the slag having the lowest iron content, suggesting that iron is a primary reductant in the slags. Batch tests on the two most reactive slags indicated that the rate coefficients are linearly related to surface area over the range tested, and are sensitive to initial TCE concentration. Column studies showed that reactivity is lower under flow-through conditions than anticipated based on batch tests. Calculations indicate a 2-m-thick slag PRB can degrade TCE to less than mg/l for influent concentrations less than 2 mg/l at seepage velocities below 0.1 m/d. Introduction Permeable reactive barriers (PRBs) containing zerovalent iron (ZVI) are used for in situ treatment of groundwater containing trichloroethylene (TCE) and a variety of other organic contaminants (1-5). Increasing world demand for iron and interest in sustainable treatment materials has created an incentive to identify alternative low-cost reactive media for PRBs from unconventional sources, including byproducts from metal forming and processing operations (6-8). One such material is slag generated from iron casting operations, which is usually discarded as waste. Iron slag is derived from molten impurities that accumulate on the surface of liquid iron. These impurities are periodically tapped from the furnace, and cooled by air or water quenching to form slag. Iron slags consist primarily of aluminum and silicon oxides derived from impurities in the iron source, as well as oxides of calcium and magnesium from limestone or dolomite added as fluxing agents (9, 10). * Corresponding author benson@engr.wisc.edu. Slags also contain iron and manganese oxides, sulfur (from coke added as fuel), alkalis, and other trace elements (11, 12). Particles of iron metal also exist within the solidified slag matrix. When crushed to form construction materials, foundry slags have the appearance and texture of coarse sand or pea gravel, and exhibit no change in volume or surface area when hydrated. This study evaluated the feasibility of using grey-iron foundry slag, a byproduct of the iron casting industry that has been landfilled traditionally, as a reactive medium to remove TCE from water. Batch and column tests were conducted with TCE solutions prepared with DI water to evaluate reactivity of the slags and to confirm that TCE was being removed. Identifying mechanisms controlling the reactions, defining all reaction byproducts, and studying the impacts of ambient groundwater chemistry were beyond the scope of this feasibility study. Experimental Section Grey-Iron Foundry Slags. Six slags from grey-iron foundries (designated A-F) were evaluated. These slags were selected because of their range of total iron contents so that the influence of iron content on reactivity could be examined. Production methods and the iron source at each foundry can be found in Section S1 of the Supporting Information (SI). The slags were obtained in pieces ranging in size from 50 to 500 mm and crushed to pass the No. 4 sieve (4.75 mm). Properties of the slags are in Table 1 and in the SI (S1). Leaching of contaminants of concern is an issue when industrial byproducts are used in applications where they may impact groundwater. Metz and Benson (7) and Eberhardt and Benson (13) have evaluated leaching of contaminants from crushed grey-iron slags for conditions relevant to PRBs. They show that a variety of trace elements leach from greyiron slags, but in most cases concentrations fall below maximum contaminant levels. Moreover, they illustrate that trace-element concentrations in leachate from grey-iron slags are lower than those eluted from typical ZVI used for PRBs. Control Materials. Zero-valent iron obtained from Peerless Metal Powders and Abrasives, Co. (Detroit, MI) was used as a reactive control material for comparative batch tests. Uniformly graded quartz sand (Iota Quartz 10TA6, mm, Unimin Corporation, New Canaan, CT) was used as a nonreactive control material. Metal and organic constituents in the sand were removed prior to testing using the method in Ma et al. (12). Characteristics of the ZVI and sand are in Table 1. Batch Tests. Batch tests were conducted to evaluate reactivity using TCE solutions prepared in DI water. DI water was used in lieu of groundwater for this feasibility assessment to minimize confounding factors affecting reactivity. In some cases, however, the reactivity of iron-based PRB media can be affected by the presence of common constituents in groundwater (14-16). Thus, site-specific assessments would be needed if slag was considered for an actual PRB application. Samples for batch testing were prepared by adding reactive material and TCE solution to 40-mL amber borosilicate glass vials. The vials were fitted with polypropylene screw caps equipped with polytetrafluoroethylene-faced silicone septa. Three replicate vials and one control vial (only TCE solution) were prepared for each reaction time. Vials were rotated at 30 RPM at 23 ( 2 C for 3, 6, 12, 24, 48, or 96 h. Supernatant from the vials was analyzed for ph, electrical conductivity (EC), redox potential (Eh), and concentrations of TCE and dechlorination byproducts. Control tests using sand showed /es801359d CCC: $ American Chemical Society VOL. 43, NO. 1, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY Published on Web 11/26/2008

2 TABLE 1. Properties of Grey-Iron Foundry Slags and Control Materials particle size characteristics a material D 10 (mm) D 30 (mm) D 60 (mm) USCS classification specificgravity, G s b hydraulic conductivity c (m/s) organic carbon content d (%) iron content e specificsurface area f (%) (m 2 /g) A SW B SW C SP D SW E SW F SP SAND SP 2.65 NM NM NM NM ZVI SP a Determined according to ASTM D422. b Specific gravity determined according to ASTM D c Determined according to ASTM D2434. d Percent by weight; average of four samples obtained by dry combustion method. e Determined by X-ray fluorescence spectroscopy. f Determined by multipoint krypton BET surface area analysis; D 10, particle diameter at which 10% of material is finer by weight; D 30, particle diameter at which 30% of material is finer by weight; D 60, particle diameter at which 60% of material is finer by weight; USCS, Unified Soil Classification System; G s, specific gravity of solids; NM ) not measured. FIGURE 1. Dechlorination of TCE by slags and ZVI during initial batch tests. Solid lines represent best-fit of eq 1. Error bars show standard error (may be hidden by symbols). Letters in legend represent slags. minimal loss of TCE (see SI S7). Additional details regarding the batch test methods are described in the SI (S7). Details on methods used to prepare the TCE solution are in the SI (S3). A first-order model with instantaneous linear sorption was fit to the TCE concentrations using a nonlinear iterative optimization algorithm to determine the bulk first-order removal coefficient (k obs ) and the slag-phase partition coefficient (K pi )(17, 18): C 0 k C aq (t) ) 1 +F m K pi exp( obs t 1 +F m K pi) (1) where C aq ) TCE concentration at time t, C 0 ) initial TCE concentration, and F m ) the mass of reactive material to solution volume. Surface-area-normalized rate coefficients (k SA ) were calculated as the ratio of k obs to the surface area to solution volume ratio (F a ). Column Tests. Slag was placed in each column by filling and gently withdrawing a tremie tube, and then packed by FIGURE 2. Concentrations of 1,1-DCE and cis-dce during initial batch tests. Error bars show standard error (may be hidden by symbols). tapping the sides of the column until settlement ceased. Weight-volume relationships were used to compute the total porosity. Influent solution was pumped upward at a constant rate using a peristaltic pump. Sampling was conducted using ports at the ends of the column and at five equi-spaced locations along the column length. Samples were collected ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 1, 2009

3 TABLE 2. Experimental Conditions, Partition Coefficients, and Rate Coefficients for Batch Tests on Slag and ZVI material G m a (kg/l) G a b (m 2 /L) C 0 c (mg/l) *K pi d (L/kg) *k obs e (10-3 hr -1 ) *k SA f (10-4 L/m 2 -hr) R 2 Slag A ( ( ( Slag B ( ( ( Slag C ( ( ( Slag D ( ( ( Slag E ( ( ( Slag F ( ( ( ZVI ( ( ( a Solid mass-solution volume ratio. b Surface area-solution volume ratio. c Initial TCE concentration. d Linear partition coefficient. e Bulk pseudo-first-order removal rate coefficient. f Surface-area-normalized pseudo-first-order rate coefficient. * Reported with standard error from regression. TABLE 3. Experimental Conditions, Partition Coefficients, and Rate Coefficients for Batch Tests with Varying Surface Area and Concentration using Slags A and C slag and test C 0 a (mg/l) G m b (kg/l) G a c (m 2 /L) *K pi d (L/kg) *k obs e (10-2 hr -1 ) *k SA f (10-4 L/m 2 -hr) R 2 A ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( C ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( a Initial TCE concentration. b Solid mass-solution volume ratio. c Surface area-solution volume ratio. d Solid phase partition coefficient. e Bulk pseudo-first-order rate coefficient. f Surface area-normalized pseudo-first-order rate coefficient. * Reported with standard error from regression. FIGURE 3. Relationship between bulk rate coefficient (k obs ) and surface area-solution volume ratio (G a ) for the (a) A and (b) C slags. Solid lines represent linear regressions of k obs on G a. Error bars show standard error (may be hidden by symbols). from top to bottom using a gastight syringe equipped with a two-way valve. As with the batch tests, all column tests were conducted with solutions prepared with DI water. Additional details of the column test apparatus are in the SI (S8). A control test conducted using sand indicated minimal losses within the column (see SI, S8). A bromide tracer study was conducted in each column prior to permeation with TCE solution. Bromide concentrations were determined using the colorimetric technique described in Lepore (19) (SI S6). The one-dimensional advection-dispersion-reaction equation (ADRE) was fit to Br - concentrations in the column effluent using the CXTFIT software package (20). Fitting with CXTFIT yielded the seepage velocity (v s ) and dispersion coefficient (D). Effective porosity (n e ) was calculated as the ratio of the specific discharge (q), to v s, and the longitudinal dispersivity was VOL. 43, NO. 1, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 171

4 TABLE 4. Experimental Conditions, Transport Parameters, and Rate Coefficients for Column Tests column test C 0 a (mg/l) G m b (kg/l) G a c (m 2 /L) G d d (Mg/m 3 ) v s e (m/d) r L f (m) *K pi g (kg/l) *k obs h ( 10-2 hr -1 ) *k SA i ( 10-4 L/m 2 -hr) *R 2 Slag A, Low C o ( ( ( Slag A, High C o ( ( ( (0.81 ( 0.05) (5.24 ( 0.00) (0.74 ( 0.00) (0.878) Slag C, Low C o ( ( ( Slag C, High C o ( ( ( a Average influent TCE concentration. b Solid mass to solution volume ratio. c Surface area to solution volume ratio. d Dry density. e Seepage velocity. f Longitudinal dispersivity. g Solid-phase partition coefficient. h Pseudo-first-order rate coefficient. i Surface-area-normalized pseudo-first-order rate coefficient. * Model parameters in parentheses were obtained by fitting ADRE to TCE effluent data. effluent. Surface-area-normalized rate coefficients were computed as k obs /F a for the column. Analytical Methods. Aqueous concentrations of chlorinated ethenes were measured using a Shimadzu (Columbia, MD) GC-2010 gas chromatograph (GC) equipped with a flame ionization detector (FID) and a 30 m 0.53 mm i.d. SPB-624 capillary column with a 3.00-µm film thickness (Supelco Corporation, Bellefonte, PA). The FID combustion gases were air (400 ml/min) and hydrogen (60 ml/min). Nitrogen was used as the carrier gas (0.20 m/s). The GC was connected to an OI Analytical (College Station, TX) model 4560 purgeand-trap sample concentrator equipped with a Supelco (Bellefonte, PA) VOCARB 4000 trap. The following method detection limits (MDLs) were obtained: TCE, 9 µg/l; 1,1- DCE, 30 µg/l; cis-dce, 30 µg/l; trans-dce, 32 µg/l. Reliable detection of vinyl chloride was not possible using the analytical methods employed. Additional details on sample preparation and analytical procedures are provided in the Supporting Information (S3, S5). FIGURE 4. Effect of iron content on (a) bulk first-order rate coefficient (k obs ) and (b) surface-area-normalized first-order rate coefficient (k SA ) from batch tests. Dashed lines indicate trend only. Error bars show standard error (may be hidden by symbols). C o ) 20 mg/l. computed as D/v s. Molecular diffusion was ignored because the flow rates employed ensured an advection-dominated system. Bulk rate coefficients (k obs ) for TCE removal were determined by fitting eq 1 to steady-state TCE concentration profiles (i.e., a stationary distribution of pore water concentrations within the column) assuming that the seepage velocity was steady along the length of the column. Rate coefficients determined in this manner were comparable to k obs obtained from a direct fit of the ADRE to TCE effluent data for those columns where TCE was detected in the column Results and Discussion Relative Reactivity of Slags. Batch tests were conducted on all six slags and ZVI under similar experimental conditions to evaluate relative reactivity and to examine the relationship between reactivity and iron content. The target initial TCE concentration was 20 mg/l and F m was 0.25 kg/l. The initial concentration was selected from within the range of TCE concentrations observed at field-scale PRB installations (2, 3, 18, 21-23) and F m was selected to ensure that the full particle size distribution of each slag was represented in a vial. All batch tests were conducted in triplicate. The standard error of concentrations from the triplicate tests is shown with the error bars in the graphs depicting results from the batch tests. Error bars that are not visible fall within the periphery of the symbol. TCE concentrations decreased with time for all materials (Figure 1), with the most rapid removal of TCE with the ZVI control. Slags A and C removed TCE more slowly than ZVI, but faster than the other slags. Concentrations of 1,1-DCE and cis-dce increased (Figure 2) as TCE concentrations decreased (Figure 1) for the slags and ZVI, and higher cis- DCE and 1,1-DCE concentrations were associated with greater reductions in TCE concentration. Other byproducts may have been generated (e.g., vinyl chloride, ethene, acetylenes, etc.), but were not included in the analysis. The 1,1-DCE and cis-dce concentration records suggest that slags degrade TCE to lesser-chlorinated byproducts in a manner similar to ZVI (24, 25). No trans-dce was observed above the MDL during any of the tests. The absence of trans-dce and predominance of cis-dce may indicate that reductive dechlorination of TCE and trans-dce occurs via the β-elimination pathway observed in ZVI systems (4, 24). However, this pathway could not be confirmed because analyses for ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 1, 2009

5 FIGURE 5. TCE and cis-dce profiles for (a) Slag A, Low C o (C 0 ) 1.5 mg/l, v s ) 0.12 m/d), (b) Slag A, High C o (C 0 ) 14 mg/l, v s ) 0.14 m/d), (c) Slag C, Low C o (C 0 ) 1.5 mg/l, v s ) 0.12 m/d), and (d) Slag C, High C o (C 0 ) 14 mg/l, v s ) 0.12 m/d). Solid lines are fit of eq 1. Error bars show standard error (may be hidden by symbols). chloroacetylene and acetylene, byproducts generated through via the β-elimination pathway, were not included in the analysis. Bulk rate coefficients (k obs ) obtained by fitting eq 1 to the batch test data for all of the slags are summarized in Table 2 along with the experimental conditions. The good fit of eq 1 to the TCE concentrations for ZVI and Slags A and C (R 2 > 0.977) suggests that removal of TCE for these materials is pseudo-first-order, as has been observed for ZVI (e.g., 24-26). Poorer fits were obtained for the less reactive slags due to the lower removal of TCE. Instantaneous partitioning also plays a small role in the removal process, as illustrated by the small partition coefficients (K pi ) summarized in Table 3. The low partition coefficients are consistent with the low organic carbon content of the slags (Table 1), which are formed under extreme temperatures during which combustion of nearly all organic matter occurs. Effect of Surface Area. Batch tests were conducted with Slags A and C to examine the effect of F a on k obs. Tests were performed at target initial TCE concentrations of 2 and 20 mg/l and F m ) 0.25, 0.33, and 0.55 kg/l (F a ) 50, 68, and 108 m 2 /L for Slag A; F a ) 300, 420, and 658 m 2 /L for Slag C). Rate and partition coefficients computed for each of the batch tests are summarized in Table 3. TCE was removed more rapidly at higher F a for both slags. For batch tests conducted at F m higher than 0.25 kg/l, maximum concentrations of 1,1-DCE fell below the MDL for both slags. An approximately linear relationship was obtained between k obs and F a for each slag at high and low initial TCE concentrations, indicating that normalization by surface area is acceptable within the range of surface areas that were evaluated. Estimates of k SA obtained by regressing k obs on F a are shown in Figure 3. The k SA for Slag A are larger than the k SA for Slag C by a factor of 4, even though the two slags exhibit comparable k obs at equivalent F m (Table 4). The lower k SA of Slag C may indicate that this slag has a lower density of reactive surfaces or lower reactivity of active reductants distributed over the total surface area of the slag (26) due to differences in iron content or species in the slags (Table 1). The effect of iron content on k obs and k SA of all of the slags is in Figure 4. The highest rate coefficients were obtained with the highest iron content (Figure 4), and the lowest rate coefficient was obtained with the slag having the lowest iron content. This suggests that reactivity on a bulk and surface- VOL. 43, NO. 1, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 173

6 FIGURE 6. Required thickness of a PRB containing Slag C to meet the MCL for TCE (0.005 mg/l) in effluent for different combinations of influent concentration and seepage velocity. area-normalized basis is influenced by the abundance of iron in the slag and that iron is a reductant involved in the dechlorination of TCE using slags. However, the absence of a clear trend relating reactivity and iron content suggests that some of the iron in slag may be in species that cannot contribute to reduction of TCE or is unavailable for reaction (e.g., iron trapped within the interior of a slag particle that is not in contact with the solution). Effect of Initial TCE Concentration. The k SA from batch tests obtained using low and high concentration differ by more than a factor of 2 for Slags A and C (Table 3 and Figure 3). Others have observed similar effects of concentration with ZVI (4, 27, 28). For example, Arnold and Roberts (4) report an order of magnitude decrease in k obs for an increase in initial TCE concentration from 0.5 to 23 mg/l in batch tests conducted with ZVI. The dependence of rate coefficient on initial concentration may be caused by competition for reactive surface sites, as observed in ZVI systems at high TCE concentrations (4, 28). Flow-Through Behavior. Column tests were conducted using target influent TCE concentrations of 2 and 20 mg/l. A target v s of 0.1 m/d was chosen to represent typical conditions observed in the field (30). Experimental conditions for the column tests are summarized in Table 4. Magnitude of the influent concentration (C o ) is depicted in Table 4 as Low C o and High C o. In each column, TCE and byproduct concentrations attained a steady-state profile, suggesting a constant rate of reaction (Figure 5). Breakthrough of TCE above the MDL was only observed in effluent from Slag A with high influent TCE concentration. The average steady-state concentration was 0.14 mg/l (breakthrough data in Supporting Information S10). cis-dce was only observed above the MDL in effluent from Slag A at average steady-state concentrations of mg/l (low C o ) and 0.77 mg/l (high C o ). Steady-state concentration profiles of TCE and cis-dce are shown in Figure 5. The data in Figure 5 correspond to triplicate samples collected from the effluent ports, with the error bars corresponding to the standard error. Error bars that are not visible fall within the periphery of the symbol. TCE concentrations decrease approximately exponentially along the length of each column, suggesting dechlorination by pseudo-first-order kinetics, as observed in the batch tests. Concentrations of cis-dce increase as TCE concentrations decrease in each column for x/l < 0.5, indicating cis-dce is generated as a dechlorination byproduct of TCE. Maximum cis-dce concentrations in the columns containing Slag A are greater than those in the columns containing Slag C, as also observed in the batch tests. Concentrations of cis-dce fall below the detection limit by x/l ) 1 in columns containing Slag C, while cis-dce concentrations diminish slightly in the downstream portions of the columns containing Slag A. The propensity for Slag A to generate more cis-dce than Slag C may be caused by differences in the composition and the distribution of reductants in the slags, which may influence reactivity toward specific contaminants or promote different degradation pathways (26). Ebert et al. (23) observed differing relative proportions of TCE degradation byproducts in column tests conducted under similar conditions using ZVI from different sources. The TCE concentration measurements obtained from the first (x/l ) 0.17) and third (x/l ) 0.5) sample ports from Slag C with low TCE influent concentration (Figure 5c) deviate from the trend in concentrations measured at the other sample ports. A mass of corroded iron approximately 30 mm in diameter consisting of several iron pieces was identified in slag removed from the column between x/l ) 0.17 and This iron mass may have caused the anomalous concentrations at the first and third sampling ports. For example, the less permeable iron mass may have caused local increases in seepage velocity, which would reduce contact times. Additionally, the presence of a larger mass of iron may have resulted in a local reduction in iron surface area, reducing reactivity. Comparison of rate coefficients from the batch and column tests (Tables 2-4) indicates that k SA for the column tests are approximately 1 order of magnitude lower than k SA from the batch tests. This difference in the k SA may be due to a transition from reaction-rate-limited kinetics in batch systems to mass-transport-limited kinetics in columns, as observed in ZVI systems (e.g., ref 26). Cwiertny and Roberts (31) report a monotonic decrease in k SA with increasing F a for polyhalogenated alkanes (e.g., 1,1,1-trichloroethane) in batch tests with ZVI, and Gotpagar et al. (32) indicate that k obs approaches a maximum at higher F a. Wust et al. (27) suggest that diffusion into micropores in ZVI results in lower rate coefficients in column tests than in batch tests. Practical Implications This study suggests that iron foundry slags with higher iron content (>30%) have adequate reactivity to remove TCE from water, particularly for applications where velocities and TCE concentrations are lower. As an example, the thickness of a PRB containing Slag C required to achieve TCE effluent concentrations below the USEPA MCL (5 µg/l) was computed with eq 1 using the bulk rate coefficient obtained from Slag C with high influent TCE concentration. Seepage velocities between 0.01and 1.0 m/d were selected to bracket typical field conditions (30). Dispersion was ignored and K pi was set to 0. Results of these computations are shown in Figure 6 as contours of thickness required to achieve the MCL as a function of influent TCE concentration and groundwater seepage velocity. For a typical field condition (C o ) 2 mg/l, v s ) 0.1 m/d), a 2-m-thick PRB is required. Thinner PRBs would be practical for sites with lower seepage velocities or lower influent concentrations. The wide range of rate coefficients obtained for the slags (Tables 2-4) illustrates that slag-specific testing would be required when considering grey-iron slag as a reactive medium for a PRB. Batch testing could be used to identify candidate slags with suitable reactivity. However, the findings from this study indicate that the reactivity obtained from ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 1, 2009

7 batch tests may not represent the reactivity under flowthrough conditions, particularly when F a in the flow-through condition is much higher. Rate coefficients for flow-through conditions obtained by linearly scaling rate coefficients from batch tests most likely would overestimate rate coefficients realized in a flow-through scenario, resulting in a PRB that is too thin. Column testing generally will yield rate coefficients that are more representative of flow-through conditions, and can also be used to evaluate temporal interactions with sitespecific groundwater that may affect reactivity and the longterm effectiveness of slag as a reactive medium. Although the findings of this study indicate that slags can be used to remove TCE from groundwater, additional study is needed to confirm the degradation pathways and to determine the presence of other byproducts some of which could be contaminants of concern. Exploration of the reaction mechanisms might also lead to insights into factors that could inhibit or enhance reactivity. Acknowledgments Support for D.B.C. was provided by the Roy F. Weston Distinguished Graduate Fellowship and the National Defense Science and Engineering Graduate Fellowship. Support for C.H.B. was provided by his Wisconsin Distinguished Professorship and the Recycled Materials Resource Center at the University of Wisconsin-Madison. Supporting Information Available Additional information on the properties of the slags, the analytical methods, the batch and column test procedures, and data from the column tests. This material is available free of charge via the Internet at Literature Cited (1) Orth, W.; Gillham, R. Dechlorination of Trichloroethene in Aqueous Solution using Fe 0. Environ. Sci. Technol. 1996, 30 (1), (2) Puls, R.; Blowes, D.; Gillham, R. Long-term Performance Monitoring for a Permeable Reactive Barrier at the US Coast Guard Support Center, Elizabeth City, North Carolina. J. Hazard. Mater. 1999, 68 (1-2), (3) Vogan, J.; Focht, E.; Clark, J.; Graham, W. Performance Evaluation of a Permeable Reactive Barrier for Remediation of Dissolved Chlorinated Solvents in Groundwater. J. Hazard. Mater. 1999, 68 (1-2), (4) Arnold, W.; Roberts, A. Pathways and Kinetics of Chlorinated Ethenes and Chlorinated Acetylene Reaction with Fe(0) Particles. Environ. Sci. Technol. 2000, 34 (9), (5) Field, J.; Reed, R.; Istok, J.; Semprini, L.; Bennet, P.; Buscheck, T. Trichlorofluoroethene: A Reactive Tracer for Evaluating Reductive Dechlorination in Large-Diameter Permeable Columns. Ground Water Monit. Rem. 2005, 25 (2), (6) Wilkens, J.; Shoemaker, S.; Egler, A.; Sinha, R.; Bazela, W.; Bain, J. Arsenic Removal from Groundwater using a PRB of BOF Slag at the DuPont East Chicago (IN) Site; Presented at Remediation Technologies Development Forum Permeable Reactive Barriers Meeting, October 15-16, 2003, Niagara Falls, NY. (7) Metz, S.; Benson, C. Iron Foundry Slags as Permeable Reactive Barrier Materials for Removing Arsenic from Groundwater; Geotechnical Special Publication No. 174; ASCE: Reston, VA, 2007; pp (8) Kang, W., H.; Hwang, I.; Park, J. Y. Dechlorination of Trichloroethylene by a Steel Converter Slag Amended with Fe(II). Chemosphere 2006, 62, (9) Lewis, D. Properties and Uses of Iron and Steel Slags; In Conference Proceedings, Symposium on Slag; National Institute for Transport and Road Research: Pretoria, South Africa, (10) Sharma, H.; Reddy, K. Geoenvironmental Engineering; John Wiley and Sons: Hoboken, NJ, 2004; 968 pp. (11) Proctor, D.; Fehling, K.; Shay, E.; Wittenborn, J.; Green, J.; Avent, C.; Bigham, C.; Connolly, M.; Lee, B.; Shepker, T.; Zak, M. Physical and Chemical Characteristics of Blast Furnace, Basic Oxygen Furnace and Electric Arc Furnace Steel Industry Slags. Environ. Sci. Technol. 2000, 34 (8), (12) Ma, X.; Benson, C.; McKenzie, D.; Aiken, J.; Pedersen, J. Adsorption of Pathogenic Prion Protein to Quartz Sand. Environ. Sci. Technol. 2007, 41 (7), (13) Eberhardt, M.; Benson, C. Trace Element Leaching from Grey- Iron Foundry Slags; Geo Engineering Report No ; University of Wisconsin: Madison, WI, (14) Farrell, J.; Kason, M.; Melitas, N.; Lie, T. Investigation of the Long-Term Performance of Zero-Valent Iron for Reductive Dechlorination of Trichloroethylene. Environ. Sci. Technol. 2000, 34 (3), (15) Tamara, M.; Butler, E. Effects of Iron Purity and Groundwater Characteristics on Rates and Products in the Degradation of Carbon Tetrachloride by Iron Metal. Environ. Sci. Technol. 2004, 38 (6), (16) Klausen, J.; Vikesland, P. J.; Kohn, T.; Burris, D. R.; Ball, W. P.; Roberts, A. L. Longevity of Granular Iron in Groundwater Treatment Processes: Solution Composition Effects on Reduction of Organohalides and Nitroaromatic Compounds. Environ. Sci. Technol. 2003, 37, (17) Eykholt, G.; Li, L. Fate and Transport of Species in a Linear Reaction Network with Different Retardation Coefficients. J. Contam Hydrogeol 2000, 46, (18) Lee, T.; Benson, C.; Eykholt, G. Waste Green Sands as Reactive Media for Groundwater Contaminated with Trichloroethylene (TCE). J. Hazard. Mater. 2004, B109, (19) Lepore, B. Measuring and modeling of water and solute flow in macroporous silt loam soils, PhD thesis, Univ. of Wisconsin- Madison, Madison, WI, 178 p. (20) Toride, N.; Leij, J.; van Genuchten, M. The CXTFIT Code for Estimating Transport Parameters from Laboratory or Field Tracer Experiments, Version 2.0; Research Report No. 137; U.S. Salinity Laboratory, USDA, ARS: Riverside, CA, (21) McMahon, P.; Denney, K.; Sandstrom, M. Hydraulic and Geochemical Performance of a Permeable Reactive Barrier Containing Zero-Valent Iron, Denver Federal Center. Ground Water 1999, 37 (3), (22) Yabusaki, S.; Cantrell, K.; Sass, B.; Steefel, C. Multicomponent Reactive Transport in an In Situ Zero-Valent Iron Cell. Environ. Sci. Technol. 2001, 35, (23) Ebert, M.; Kober, R.; Parbs, A.; Plagentz, V.; Schafer, D.; Dahmke, A. Assessing Degradation Rates of Chlorinated Ethylenes in Column Experiments with Commercial Iron Materials Used in Permeable Reactive Barriers. Environ. Sci. Technol. 2006, 40, (24) Gillham, R.; O Hannesin, S. Enhanced Degradation of Halogenated Aliphatics by Zero-Valent Iron. Ground Water 1994, 32 (66), (25) Roberts, A.; Totten, L.; Arnold, W.; Burris, D.; Campbell, T. Reductive Elimination of Chlorinated Ethylenes by Zero-Valent Metals. Environ. Sci. Technol. 1996, 30 (8), (26) Johnson, T.; Scherer, M.; Tratnyek, P. Kinetics of Halogenated Organic Compound Degradation by Iron Metal. Environ. Sci. Technol. 1996, 30 (8), (27) Wust, W.; Kober, R.; Schlicker, O.; Dahmke, A. Combined Zeroand First-Order Kinetic Model of the Degradation of TCE and cis-dce with Commercial Iron. Environ. Sci. Technol. 1999, 33, (28) Grant, G.; Kueper, B. The Influence of High Initial Concentration Aqueous-Phase TCE on the Performance of Iron Wall Systems. J. Contam. Hydrol. 2004, 74, (29) Fetter, C. Contaminant Hydrogeology, 2nd Ed.; Prentice Hall: Upper Saddle River, NJ, 1999; 500 pp. (30) Li, L.; Benson, C.; Lawson, E. Impact of Mineral Fouling on Hydraulic Behavior of Permeable Reactive Barriers. Ground Water 2005, 43 (4), (31) Cwiertny, D.; Roberts, A. On the Nonlinear Relationship between k obs and Reductant Mass Loading in Iron Batch Systems. Environ. Sci. Technol. 2005, 39 (22), (32) Gotpagar, J.; Grulke, E.; Tsang, T.; Bhattacharyya, D. Reductive Dehalogenation of Trichloroethylene Using Zero-Valent Iron. Environ. Prog. 1997, 16 (2), ES801359D VOL. 43, NO. 1, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 175