Assessing Degradation Rates of Chlorinated Ethylenes in Column

variability in rate coefficients on 1 or 2 orders of magnitude. Geometric mean values of surface area normalized rate coefficients (in 10-5 L m-2 h-1)...
0 downloads 0 Views 177KB Size
Download PDF
Environ. Sci. Technol. 2006, 40, 2004-2010

Assessing Degradation Rates of Chlorinated Ethylenes in Column Experiments with Commercial Iron Materials Used in Permeable Reactive Barriers MARKUS EBERT,* RALF KO ¨B E R , ANIKA PARBS, VOLKMAR PLAGENTZ, D I R K S C H A¨ F E R , A N D A N D R E A S D A H M K E Institute of Geosciences, Christian-Albrechts-Universita¨t zu Kiel, Ludewig-Meyn Strasse 10, 24118 Kiel, Germany

Multiple column experiments were performed using two commercial iron materials to evaluate the necessity and usefulness of preliminary investigations in permeable reactive barrier (PRB) design for chlorinated organics. Experiments were performed with contaminated groundwater and involved fresh iron granules or altered iron material excavated from PRBs. The determination of first-order rate coefficients by global nonlinear least-squares fittings indicated a variability in rate coefficients on 1 or 2 orders of magnitude. Geometric mean values of surface area normalized rate coefficients (in 10-5 L m-2 h-1) for fresh gray cast iron and iron sponge, respectively, are: tetrachloroethene (4.5, 2.6), trichloroethene (8.1, 3.3), cis-1,2-dichloroethene (3.1, 2.9), trans-1,2-dichloroethene (9.5, 5.3), 1,1-dichloroethene (4.0, 4.4), and vinyl chloride (1.6, 6.1). The increasing rate coefficients with decreasing grade of chlorination, which characterize degradation at iron sponge are linearly related to diffusion coefficients in water, suggesting diffusion limitation in the degradation process for this particular material, possibly due to a high inner surface. The variability in rate coefficients seems to be too high to use mean rate coefficients from published studies in the design procedure of PRBs, and variabilities cannot be related to groundwater characteristics, water flow through the reactive cells, or secondary corrosion reactions.

Introduction Several constructed permeable reactive barriers (PRBs) use zerovalent iron (ZVI) to treat chlorinated ethylenes (1-4), i.e., tetrachloroethene (PCE), trichloroethene (TCE), and cis1,2-dichloroethene (cDCE). ZVI is capable of degrading these groundwater contaminants by heterogeneous reductive dehalogenation reactions. A successful PRB application needs a careful design of PRB flow through thickness to ensure sufficient residence times because dehalogenation reactions are kinetically controlled and intermediates are produced (e.g., trans-1,2-dichloroethene (tDCE), 1,1-dichloroethene (11DCE), and vinyl chloride (VC)). Material characteristics (5, 6), inorganic groundwater constituents (5, 7-13), or contaminant concentration and distribution (14-18) have proven to effect initial degradation velocity and longevity as well. A detailed description of the heterogeneous redox reactions requires a kinetic rate law based on Langmuir-Hinshelwood-HougenWatson kinetics, including terms for inter- and intra-species * Corresponding author e-mail: [email protected]; phone: +49431-8804609; fax: +49-431-8807606. 2004

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 6, 2006

competition for reactive and nonreactive surface sites (1416, 19), but most often a pseudo first-order rate law with respect to dissolved contaminant concentration is used to describe the degradation behavior of chlorinated hydrocarbons (CHCs) for ZVI. The production and degradation of intermediates is described using fractionation factors, and the normalization of observed rate coefficients (kObs) to the surface area concentration (kSA) permits the comparison of results from various experiments. Here, the BET surface area is commonly used, even if corrosion reactions and secondary mineral precipitations may change the reactive surface area of ZVI over time. Anaerobic corrosion of ZVI yields the production of hydrogen, increasing pH levels, and the precipitation of Fe(II/III)(hydr)oxides. Chloride, sulfate, or bicarbonate accelerate anaerobic corrosion (20). Bicarbonate has been proven to temporarily accelerate CHC degradation, followed by surface passivation in the long term due to carbonate precipitations (7, 9, 21). Other groundwater constituents may also inhibit anaerobic corrosion or degradation reactions, i.e., nitrate (11, 12, 18), chromate (12), silicate (9, 10), or natural organic matter (9, 22). Microbial sulfate reduction observed for field PRBs (3, 4) and column experiments (23, 24) consume hydrogen and produce dissolved sulfide, which primarily precipitates in the form of iron monosulfides (4). Laboratory experiments have shown accelerating (25), but also inhibiting (8), effects of sulfide. In addition, mineral precipitation or gas bubble formation decreases the pore volume in ZVIreactive media and may considerably decrease the hydraulic conductivity of a PRB. Recent reaction and transport models are capable of simulating decreasing reactivity and permeability, both related to iron consumption and mineral precipitation (26), but are not applied to real sites, and terms describing a potential acceleration of degradation velocity are not included. Any guidelines for PRB applications therefore advise preliminary investigations in PRB design using column experiments, and questions are open on reasonability or necessity of these experiments. For cost- and time-saving reasons it may be interesting to use mean degradation rates outlined in the literature and omit the column experiments in the design procedure. This may only be possible if the variability of degradation rates is small using the same type of iron at various geochemical conditions, or if variabilities could be related to any specific groundwater characteristics, water flow through the reactive cell, or secondary corrosion reactions. The authors have performed many column experiments for the design of ZVI PRBs at sites in Europe, and PRBs recently installed in Germany use primarily the two types of ZVI investigated in these experiments. The authors have also investigated material excavated from two PRBs. The experiments have created a series of concentration profiles characterizing the degradation behavior of chlorinated ethylenes. The study aims to assess the variety of degradation rates of various chlorinated ethylenes determined from the multiple column experiments, each of them performed using contaminated groundwater from different sites and commercial iron used in recent PRBs. The collective data demonstrate the range of rate coefficients which could be expected for ZVI PRB applications and are used to evaluate the necessity and usefulness of preliminary investigations in reactive barrier design for chlorinated organics.

Materials and Methods Column Experiments. Grey cast iron from Gotthart Maier AG (Rheinfelden, Germany) used in the experiments has a 10.1021/es051720e CCC: $33.50

 2006 American Chemical Society Published on Web 02/08/2006

columnar and platy shape. The grain-size distribution was 0.3-2 cm or 1-3 cm, and in three columns it was mixed with fine gravel. Iron sponge (ReSponge) is a steelmaking byproduct from ISPAT GmbH (Hamburg, Germany). The direct reduction of ore pellets using the Midrex process yields spheriodal ZVI with a porous inner structure. The grain size is between 0.6 and 1.6 cm, and up to 10% of lump ore is included. ZVI was used as delivered without any pretreatment. Material from coring in the Rheine PRB (3) and Tu ¨ bingen PRB (27) was additionally investigated. Vertical drillings in the up- and downgradient interface regions of the PRBs captured material for the experiments and were performed 3.5 and 4.5 y after installation at the Rheine and Tu ¨ bingen sites, respectively. Material from coring was carefully wet sieved under an argon atmosphere using degassed deionized water and was then packed into columns under water saturation. Here, reference experiments used samples retained from PRB construction (Rheine) or fresh material of equal grain-size distribution (Tu ¨ bingen). The 25 column experiments were all performed using a similar setup (see Supporting Information) and differed in the type of ZVI, flow velocity, groundwater constituents, concentration of the contaminants, and duration (Table SI1).The duration of the experiments was between 103 and 420 days, and between 6 and 10 samples were collected along the flow length (∼1 m) throughout 5-14 sampling campaigns (Table SI-1) and every organic concentration profile was used in data analysis. In total, 15 experiments were performed for the dimensioning of ZVI PRBs and 10 tests used material from the Rheine PRB, the Tu ¨ bingen PRB, or corresponding reference material. All experiments were performed at 20 ( 3 °C using groundwater from contaminated sites. Mean total CHC concentrations were between 18.4 and 1994 µM (Table SI-2). Most groundwaters had a calcium-bicarbonate character, and three showed an impact of calcium sulfate (Figure SI-2). pH was primarily between 7 and 7.7, alkalinities were between 2 and 8.3 mM, and TDS concentrations varied from 8 to 34 mM (Table SI-3). Water used in experiments performed with material from the downgradient section of the PRBs had higher pH levels and smaller calcium, bicarbonate, and sulfate concentrations because the water was taken from wells within the PRBs. Calculation of Rate Coefficients. Pseudo first-order degradation rates were obtained from global nonlinear leastsquares fitting of all CHC concentration profiles from a certain sampling event using the DynaFit (V. 3.28.011) model (28). Residence times were calculated using porosity and water flow; dispersion was neglected. The degradation scheme was adopted from ref 14, but had to be modified because of missing measurements of dichloroacetylene, chloroacetylene, and mass balances of organic carbon generally below 60%. Carbon mass balances might be effected by a fixation in the solid phase or by gas bubble formation (29). For this reason, the model includes a degradation to end-products in addition to the transformation to less chlorinated species (Figure SI1). The pathway to end-products includes the production of ethene, ethane, and other short aliphatic hydrocarbons. The formation of less chlorinated species includes all possible formation reactions, i.e., direct formation by hydrogenolysis or formation via dichloro-elimination and hydrogenation. The total pseudo first-order degradation rate coefficient of a compound is the sum of all coefficients. However, only the concentrations of dissolved chlorinated species, and not the production of any end-product, were used for data fitting. Raw data were filtered and normalized before the leastsquares fitting as described in the Supporting Information.

Results Inorganics. Concentrations of dissolved inorganic constituents changed during flow through the columns typical for ZVI-reactive media, and changes will not be discussed in detail within this study. pH levels generally increased up to more than 10 or 11 in experiments using iron sponge and between 9 and 10 in the outflow of columns using gray cast iron. pH levels were stable or decreased along the flow distance in the experiments with high TCE inflow concentration and gray cast iron, probably due to a higher production of Fe+II and subsequent (hydr)oxide precipitation. All experiments showed a significant loss in alkalinity and, generally, a strong decrease in calcium concentration as well (Figure SI-2). Nitrate was depleted in all experiments, and mole balances between the loss of nitrate and the production of ammonium generally show a decreasing tendency. Some experiments showed a significant loss of sulfate concentration (up to 120 mg/L) in a later stage of the experiment. Both the pronounced loss in sulfate and the decreasing production of ammonium in conjunction with the continuous loss of nitrate indicate microbial activity (23, 30, 31). Overall, water character changed from calcium bicarbonate to sodium chloride or sodium sulfate. Mean TDS concentrations decreased along the flow distance and were between 2 and 22 mM in the outflow of the columns. Degradation of Chlorinated Ethylenes. Figure 1 shows examples of model results for two column experiments. One was performed with gray cast iron (exp. B-GG100) and the other was performed with iron sponge (exp. Anonymous4). TCE was the main contaminant in both experiments, but inflow concentrations differed by a factor of approximately ten. Only the results of two sampling campaigns for each of the experiments are shown in Figure 1. The overall examination of 179, 194, 193, 109, 107, and 101 concentration profiles from all experiments results in a broad statistical spread of pseudo first-order degradation rate coefficients for PCE, TCE, cDCE, tDCE, 11DCE, and VC. Variations are on 2 orders of magnitudes for each species, using kObs or kSA (Figure SI-3). The average values of all calculated rate coefficients indicate most rapid degradation for tDCE and TCE and slowest degradation for cDCE and VC (Table 1). In some cases, calculated kObs was below 6.9 × 10-3 h-1, corresponding to a half-life of more than 100 h. Mean value calculation does not include these results, but their number and proportion are shown in Figure 2 and Figure SI-3. In addition, outliers in the opposite direction (kObs > 10 h-1) indicating half-lives below 0.07 h were excluded in average value calculation and are not shown in the figures, but these are only 1 for 11DCE and tDCE and 2 for TCE. The upper boundary at 100 h half-life was chosen because the decrease in concentration becomes to small for an accurate rate estimation at that slow reaction velocity, and residence times typically between 1 and 2 days in the column experiments. For example, the model fits well to cDCE concentration profiles derived from the experiment B-GG100 (Figure 1) with total rate coefficients corresponding to half-lives of 785 and 5806 h for 21 and 94 PV. This merely indicates that noticeable cDCE degradation cannot be justified. A half-life of a few hundreds of hours more or less will not significantly change the modeled concentration profiles. Here, the calculation of rate coefficients primarily depends on the statistical spread of measured concentrations. A classification of the various materials and a subgrouping of the results from each experiment indicate smaller variations of rate coefficients within most of the individual experiments (Figure 2). The species-specific total range of all rate coefficients for the degradation at fresh or altered ZVI is still on at least 1 order of magnitude. The geometric average values of kSA again show most rapid degradation for tDCE, VOL. 40, NO. 6, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2005

FIGURE 1. Measured concentrations of chlorinated ethylenes in two column experiments using gray cast iron (exp. B-GG100, closed symbols) and iron sponge (exp. Anonymous4, open symbols) from two sampling campaigns at various exchanged pore volumes (PV). The lines indicate the least-squares fit of a pseudo first-order degradation model with total rate coefficients (h-1) as numbered in the figures. (Error bars specify the standard deviation from triplicate samples and are partly covered by the symbols).

TABLE 1. Geometric Average Values of Modeled Pseudo First-Order Rate Coefficient (Observed and Surface Area Normalized), Subgrouped for Experiments Using Various ZVI PCE

TCE

cDCE tDCE 11DCE

kObs [10-2 h-1] all 8.7 16 7.1 grey cast iron, fresh 15 28 11 grey cast iron, altered 8.9 21 2.8 iron sponge, fresh 5.5 6.8 5.4 iron sponge, altered 3.3 9.8 6.9 kSA [10-5 L m-2 h-1] all 3.6 6.7 2.5 grey cast iron, fresh 4.5 8.1 3.1 grey cast iron, altered 5.4 14 1.1 iron sponge, fresh 2.6 3.3 2.6 iron sponge, altered 1.4 4.2 2.9

26 40 33 11

7.9 9.5 9.1 5.3

11 13 9.3 9

4.4 4 10 4.4

VC 6.8 6.8 2.3 13

2.1 1.6 0.61 6.1

followed by TCE, and slowest degradation of cDCE and VC for fresh gray cast iron (Table 1). Interestingly, the mean rate coefficients for the fresh iron sponge loosely indicate increasing degradation rates with decreasing grade of chlorination. Furthermore, the degradation of VC is slower for gray cast iron than for iron sponge by a factor of approximately 3, but the mean degradation rate of the other species is similar (11DCE) or more rapid, up to a factor of approximately 2.5 (TCE). The comparison between calculated kSA for fresh and altered gray cast iron from PRBs indicates a decreasing reactivity for cDCE and VC, but shows slightly 2006

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 6, 2006

increasing rate coefficients for the degradation of other species for altered material. Experiments using altered iron sponge from the Rheine PRB indicate decreasing reactivity only for PCE. Moreover, the degradation of cDCE was more rapid for altered iron sponge than for altered gray cast iron. The results show that cDCE and VC may be the most problematic contaminants for ZVI PRB applications, at least if gray cast iron is used. This is suggested by the decreased mean rate coefficients for altered material and additionally supported by the high number of concentration profiles which correspond to a half-life of more than 100 h. Overall, 50% of the cDCE concentration profiles show this slow degradation velocity in the experiments using material from the PRBs. In addition, 26% and 28% of all cDCE concentration profiles indicate an insignificant degradation in experiments using fresh gray cast iron and iron sponge. Most of these results were obtained from experiments using groundwater from the Bernau site (32), which contained high amounts of TCE (nominal 100 and 300 mg/L). TCE was not completely degraded along the flow distance in these column experiments due to higher concentrations and residence times of less than 1 day. Here, the high background of TCE may explain the small pseudo first-order rate coefficients modeled for cDCE because a suggested competition between TCE and cDCE for reactive sites (14, 16) is not included in the degradation scheme. However, TCE background concentrations were small in experiments using altered material from the PRBs, indicating that other processes inhibit cDCE degradation in these cases. No obvious correlations were

FIGURE 2. Modeled surface area normalized pseudo first-order degradation rate coefficient of CHCs in 25 column experiments using fresh and altered gray cast iron (left) and iron sponge (right). Horizontal solid lines couple results of individual experiments, and vertical lines indicate geometric average value. Bars indicate the number of concentration profiles where the observed half-life was greater than 100 h. able to be obtained from the column experiments between decreased cDCE rate coefficients and corrosion reactions or mineral formation indicated by changing concentrations of dissolved inorganic species. In general, the results support no consistent correlation between the development of CHC degradation rate coefficients and experimental conditions. Most of the experiments show a decreasing tendency in degradation velocity, with highest rates at the beginning. In some cases, fitted rate coefficients indicate quasi steady-

state conditions through the end of the observation period. Other experiments show increasing rate coefficients, indicating highest reactivity through the end of the observation period (Figure SI-4). Moreover, observed reactivity trends were frequently not equal for all chlorinated species in an individual experiment. Therefore it is not surprising that correlations between calculated rate coefficients and exchanged pore volumes are not transferable. A relation of rate coefficients to the load of alkalinity, total inorganic carbon, VOL. 40, NO. 6, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2007

TDS, or oxidation capacity also remains unsuccessful for the comparison of various experiments. However, each individual experiment generally showed a continuous progress of fitted rate coefficients, if increasing or if decreasing reactivity was indicated. The only exceptions are experiments with sulfate reduction in a later stage of the observation period, where discontinuously increasing or decreasing rate coefficients were occasionally observed.

Discussion Pseudo first-order degradation was chosen for data fitting because the structure of the data does not support adequate fittings using models including sorption at reactive or nonreactive sites and intra- or interspecies competition. Such models would likely result in better fittings of the concentration profiles, but due to lack of separate experiments those fittings would not be unique, as the result of overly high degrees of freedom (16, 19). The model used assumes uniform degradation throughout the reaction along the entire flow distance in the column experiments. This means that a variable reactivity which may be caused by corrosion reactions and secondary mineral formation is neglected. However, the pseudo first-order model has correlated well with most of the concentration profiles. A further assumption is a uniform flow through the reactive media, neglecting preferential flow, nonuniform porosity, and changing porosity with time. Mineral precipitation may decrease porosity primarily next to the inflow region of the columns. Concentration profiles of dissolved inorganic species indicated mineral formation primarily located in the first quarter of the flow length in each column experiment. Hydrogen concentration profiles and gas accumulation in downstream gas traps suggested gas bubble formation in the reactive media in some experiments, but gas production was variable in time and space. Both mineral precipitation and gas bubbles decrease porosity in ZVIreactive media and accelerate pore velocity at constant flow conditions. Therefore variations in fitted rates within an individual experiment or between various experiments may be partly due to variable porosities. The magnitude of such variations is linearly related to the change in porosity because of the linear relation between residence time and first-order rate coefficients. The spreading of the rate coefficients was generally on 1 or 2 orders of magnitudes and is likely not explainable by porosity variations alone. Johnson et al. (17) have cumulated previously published first-order rate coefficients for the degradation of CHC at ZVI. Compared to these, the calculated average kSA are smaller for the degradation of the more highly chlorinated species for fresh gray cast iron or iron sponge (Figure SI-5). Other studies have also shown more rapid degradation of TCE and cDCE (6, 15), while comparable or smaller VC rate coefficients have been determined (33, 34). The column experiments presented here indicate degradation rates for TCE higher than those for PCE independent of the material used. A more rapid indirect reduction of TCE by hydrogen (35) or sorption competition (14) may be responsible for that observation. The progressively decreasing pseudo first-order rate coefficients for the degradation of TCE, 11DCE, cDCE, and VC in experiments using gray cast iron follow a tendency frequently observed and was related to decreasing redox potentials or correlated by linear free energy relationships (36-38). The experiments using iron sponge have shown a reverse tendency, i.e., increasing rate coefficients with decreasing grades of chlorination. Calculated mean kSA correlates positively with diffusion coefficients in water for PCE, TCE, DCE, and VC (Figure 3). This implies that diffusion processes may play an important role in CHC degradation at this material, possibly as the result of the high inner porosity of iron sponge, accessible only by diffusion. A higher inner 2008

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 6, 2006

FIGURE 3. Mean kSA (fresh material) vs diffusion coefficients in water at 20 °C (open symbols, gray cast iron; closed symbols, iron sponge). porosity of iron sponge can be assumed from differences in total and effective porosities (0.65 to 0.7 vs 0.41 to 0.5), a difference between calculated geometric surface area and measured BET surface area in 3 orders of magnitude, and from the apparent density of the material (3.4-3.6 g cm-3). cDCE does not follow this relation, and mean rate coefficients would be even smaller if results are included in average calculations which indicate insignificant levels of degradation (i.e., half-life > 100 h). An uncorrelated small degradation velocity of cDCE was also shown in other compilations of degradation rates of CHC. Without a valid explanation for this behavior, the collective results of the column experiments point to cDCE as the contaminant with the highest potential for failing remediation goals, even if national regulations may allow higher limits (i.e., 70 µg/L) for cDCE. Various studies have investigated the effect of groundwater constituents on the degradation behavior of contaminants for ZVI and indicate the inhibition or acceleration of CHC degradation related to the water chemistry. However, most of these studies have investigated the effect of a single constituent using artificial solutions. The column experiments presented here do not support a general dependency (tested by multivariate statistics, e.g., component analysis, cluster analysis, correlation analysis) between degradation rate coefficients or their development throughout experiment run times and the load, concentration, or changes in concentration of any groundwater constituent (Table SI-4, Figure SI6). This suggests that more complex interactions take place in systems using groundwater of more or less complex composition and cumulative or antagonistic effects on CHC degradation are not able to be derived from the plain combination of effects observed in more simple experiments. Implications for PRB Design. Sophisticated hypotheses are required and must be tested by numerical modeling to fully understand the interactive reactions in ZVI PRBs, including terms for inhibition as well as the acceleration of corrosion and degradation and to describe porosities and reactive surface areas variable in time and space, all dependent on (kinetically controlled) corrosion reactions, secondary mineral formation, and material characteristics. A prediction of degradation efficiency and its longevity presents a challenge because such a complete model has yet to be developed, probably because the important processes are not yet identified. Furthermore, at field PRBs complex flow patterns may result in a nonuniform flow through the reactive media and variable residence times (2, 27). Despite these uncertainties, column experiment results have shown to be in good agreement with field results for the Rheine PRB (3) or the Borden PRB (1). Interestingly, field results from the Bernau in situ reaction vessels indicate a small reactivity for cDCE, while TCE was degraded completely at reaction rates comparable to those derived from the column experiments. Reasons for this specific inhibition of cDCE degradation have

not been identified as yet, but the column experiments point to this issue before PRB construction. The characteristic differences in degradation behavior of chlorinated ethylenes have implications for passive in situ groundwater remediation with PRBs using the two types of ZVI investigated. A more rapid degradation of more highly chlorinated ethylenes was observed for gray cast iron, but iron sponge degrades the less highly chlorinated species more rapidly. In addition, inhibition of cDCE degradation was more pronounced when altered gray cast iron was used in column experiments. This suggests that iron sponge may be the better choice for a PRB application if degradation of the less highly chlorinated ethylenes is the limiting factor in PRB design. Furthermore, the effectiveness of the degradation of more highly chlorinated ethylenes is comparable for both materials on a cost basis because the price per volume unit of gray cast iron is approximately twice as high as that for iron sponge. The pseudo first-order rate coefficients determined from the column experiments may have varied by at least 1 order of magnitude, but each specific experiment has indicated a systematic development of degradation rates, and results seem to be transferable to field conditions. This points to complex and interactive, but reproducible, reactions taking place in ZVI PRBs which have not yet been fully understood, and accentuates the need for column experiments as an element of preliminary investigations for PRB applications. In best case the experiments should be carried out until steady state conditions are established, which might require a run time of at least 100-150 days. The high variability of firstorder rate coefficients points to an increased risk for failing remediation goals if mean degradation rates from published studies are used in the design procedure. These types of experiments definitely hold potential for predicting degradation behavior under field conditions, even if flow patterns in a PRB may differ significantly from flow through a laboratory column experiment. Moreover, on a cost-value ratio these experiments are the best available for PRB design to minimize potential failures in passive groundwater remediation using ZVI.

Acknowledgments Research was funded by the German Federal Ministry of Education and Science as a part of the RUBIN Research Network (Contract 02WR0208). I.M.E.S. GmbH, Amtzell, Germany, primarily financed the column experiments for the dimensioning of PRBs. INGAAS GmbH, Berlin, Germany, supported the Bernau preliminary investigations, and two experiments were funded by M&P GmbH, Hannover, Germany.

Supporting Information Available Detailed description of the experiment setup and leastsquares fitting routine. Figures and tables that characterize the contaminated groundwater, illustrate the degradation scheme, show the spreading of all rate coefficients, show examples of the development of pseudo first-order rate coefficients throughout the experiment duration, compare rate coefficients with other studies, and list the correlation matrix between rate coefficients and various experimental state variables. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) O’Hannesin, S. F.; Gillham, R. W. Long-term Performance of an in-situ “iron-wall” for remediation of VOCs. Ground Water 1998, 36, 164-170. (2) Gavaskar, A.; Sass, B.; Gupta, N.; Drescher, E.; Yoon, W.-S.; Sminchak, J.; Hicks, J.; Condit, W. Evaluating the longevity and hydraulic performance of permeable reactive barriers at Department of Defense sites; Battelle: Columbus, OH, 2002.

(3) Ebert, M.; Wegner, M.; Parbs, A.; Plagentz, V.; Scha¨fer, D.; Ko¨ber, R.; Dahmke, A. Prognostizierte und tatsa¨chliche Langzeitstabilita¨t von Fe(0)-Reaktionswa¨nden - Am Beispiel der Reaktionswand am Standort Rheine nach 5-ja¨hriger Betriebszeit. Grundwasser 2003, 3, 157-168. (4) Wilkin, R. T.; Puls, R. W.; Sewell, G. W. Long-term performance of permeable reactive barriers using zerovalent iron: Geochemical and microbiological effects. Ground Water 2003, 41, 493503. (5) Ta´mara, M. L.; Butler, E. C. 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, 1866-1876. (6) Miehr, R.; Tratnyek, P. G.; Bandstra, J. Z.; Scherer, M. M.; Alowitz, M. J.; Bylaska, E. J. Diversity of contaminant reduction reactions by zerovalent iron: Role of the reductate. Environ. Sci. Technol. 2004, 38, 139-147. (7) Agrawal, A.; Ferguson, W. J.; Gardner, B. O.; Christ, J. A.; Bandstra, J. Z.; Tratnyek, P. G. Effects of carbonate species on the kinetics of dechlorination of 1,1,1-trichloroethane by zerovalent iron. Environ. Sci. Technol. 2002, 36, 4326-4333. (8) Dries, J.; Bastiaens, L.; Springael, L.; Diels, L. Kinetics of trichloroethene (TCE) reduction by zerovalent iron: effect of medium composition. In Groundwater Quality: Natural and Enhanced Restoration of Groundwater Pollution (Proceedings of the Groundwater Quality 2001 Conference held at Sheffield, UK, June 2001); Thornton, S. F., Oswald, S. E., Eds.; IAHS Press: Oxfordshire, 2002; Vol. 275, pp 397-402. (9) 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, 1208-1218. (10) Kohn, T.; Kane, S. R.; Fairbrother, D. H.; Roberts, A. L. Investigation of the inhibitory effect of silica on the degradation of 1,1,1-trichloroethane by granular iron. Environ. Sci. Technol. 2003, 37, 5806-5812. (11) Ritter, K.; Odziemkowski, M. S.; Simpgraga, R.; Gillham, R. W.; Irish, D. E. An in situ study of the effect of nitrate on the reduction of trichloroethylene by granular iron. J. Contam. Hydrol. 2003, 65, 121-136. (12) Schlicker, O.; Ebert, M.; Fruth, M.; Weidner, M.; Wu¨st, W.; Dahmke, A. Degradation of TCE with iron: The role of competing chromate and nitrate reduction. Ground Water 2000, 38, 403409. (13) Ko¨ber, R.; Schlicker, O.; Ebert, M.; Dahmke, A. Degradation of chlorinated ethylenes by Fe0: inhibition processes and mineral precipitation. Environ. Geol. 2002, 41, 644-652. (14) Arnold, W. A.; Roberts, A. L. Pathways and kinetics of chlorinated ethylene and chlorinated acetylene reaction with Fe(0) particles. Environ. Sci. Technol. 2000, 34, 1794-1805. (15) Wu ¨ st, W. F.; Ko¨ber, R.; Schlicker, O.; Dahmke, A. Combined zero- and first-order kinetic model of the degradation of TCE and cis-DCE with commercial iron. Environ. Sci. Technol. 1999, 33, 4304-4309. (16) Scha¨fer, D.; Ko¨ber, R.; Dahmke, A. Competing TCE and cis-DCE degradation kinetics by zerovalent iron - experimental results and numerical simulation. J. Contam. Hydrol. 2003, 65, 183202. (17) Johnson, T. L.; Scherer, M. M.; Tratnyek, P. G. Kinetics of halogenated organic compound degradation by iron metal. Environ. Sci. Technol. 1996, 30, 2634-2640. (18) Farrell, J.; Kason, M.; Melitas, N.; Li, T. Investigation of the longterm performance of zerovalent iron for reductive dechlorination of trichloroethylene. Environ. Sci. Technol. 2000, 34, 514-521. (19) Venkatapathy, R.; Bessingpas, D. G.; Canonica, S.; Perlinger, J. A. Kinetics models for trichloroethylene transformation by zerovalent iron. Appl. Catal. B 2002, 37, 139-159. (20) Reardon, E. J. Anaerobic corrosion of granular iron: measurement and interpretation of hydrogen evolution rates. Environ. Sci. Technol. 1995, 29, 9, 2936-2945. (21) Zhang, Y.; Gillham, R. W. Effects of gas generation and precipitates on performance of Fe0 PRBs. Ground Water 2005, 43, 113-121. (22) Tratnyek, P. G.; Scherer, M. M.; Deng, B.; Hu, S. Effects of natural organic matter, anthropogene surfactants, and model quinones on the reduction of contaminants by zerovalent iron. Water Res. 2001, 35, 4435-4443. (23) Gu, B.; Phelps, T. J.; Liang, L.; Dickey, M. J.; Roh, Y.; Kinsall, B. L.; Palumbo, A. V.; Jacobs, G. K. Biochemical dynamics in zerovalent iron columns: implications for permeable reactive barriers. Environ. Sci. Technol. 1999, 33, 2170-2177. VOL. 40, NO. 6, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2009

(24) Ebert, M. Elementares Eisen in permeablen reaktiven Barrieren zur in-situ Grundwassersanierung - Kenntnisstand nach zehn Jahren Technologieentwicklung. Habilitationsschrift, ChristianAlbrechts-Universita¨t, Kiel, Germany, 2004. (25) Sivavec, T. M.; Horney, D. P.; Baghel, S. S. Reductive dechlorination of chlorinated ethenes and ethanes by iron metal and iron sulfide minerals. In Special Symposium of the American Chemical Society, Emerging Technologies in Hazardous Waste Management VII: Atlanta Georgia, 1995; American Chemical Society: Washington, DC, 1995; pp 42-45. (26) Li, L.; Benson, C. H.; Mergener, E. A. Impact of mineral fouling on the hydraulic behavior of continuous-wall permeable reactive barriers. Ground Water 2005, 43, 582-592. (27) Parbs, A.; Ebert, M.; Ko¨ber, R.; Plagentz, V.; Schad, H.; Dahmke, A. Einsatz reaktiver Tracer zur Bewertung der Langzeitstabilita¨t und Reaktivita¨t von Fe(0)-Reaktionswa¨nden. Grundwasser 2003, 3, 146-156. (28) Kuzmic, P. Program DYNAFIT for the analysis of enzyme kinetic data: Application to HIV proteinase. Anal. Biochem. 1996, 237, 260-273. (29) Dahmke, A.; Ebert, M.; Dethlefsen, F.; Ko¨ber, R.; Parbs, A.; Plagentz, V.; Scha¨fer, D.; Silva-Send, N. Bewertung und Weiterentwicklung von Voruntersuchungs-, Monitoring- und Qualita¨tsmanagementansa¨tze fu ¨ r Reaktionswa¨nde - eine vergleichende Labor- und Standortstudie unter besonderer Beru ¨ cksichtigung vollsta¨ndiger, stoffspezifische Bilanzen des Schadstoffumsatzes beim Einsatz reaktiver Wa¨nde.- BMBF Forschungsvorhaben No. 02WR0208; final report; Christian-AlbrechtsUniversita¨t, Kiel, Germany, 2005. (30) Gandhi, S.; Oh, B.-T.; Schnoor, J. L.; Alvarez, P. J. J. Degradation of TCE, Cr(VI), sulfate, and nitrate mixtures by granular iron inflow-through columns under different microbial conditions. Water Res. 2002, 36, 1973-1983.

2010

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 6, 2006

(31) Scherer, M. M.; Richter, S.; Valentine, R. L.; Alvarez, P. J. J. Chemistry and microbiology of permeable reactive barriers for in situ groundwater clean up. Crit. Rev. Environ. Sci. Technol. 2000, 30, 363-411. (32) Hein, P.; Freygang, M.; Vigelan, L. Horizontaler Eisenganulatreaktor zur in-situ-Abreinigung von LHKW-kontaminiertem GW in Bernau bei Berlin. altlasten spektrum 2002, 2, 89-90. (33) Deng, B.; Burris, D. R.; Campbell, T. J. Reduction of vinyl chloride in metallic iron-water systems. Environ. Sci. Technol. 1999, 33, 2651-2656. (34) Su, C.; Puls, R. Kinetics of trichloroethene reduction by zerovalent iron and tin: Pretreatment effect, apparent activation energy, and intermediate products. Environ. Sci. Technol. 1999, 33, 163168. (35) Wang, J.; Farrell, J. Investigating the role of atomic hydrogen on chloroethene reactions with iron using tafel analysis and electrochemical impedance spectroscopy. Environ. Sci. Technol. 2003, 37, 3891-3896. (36) Gillham, R. W.; O’Hannesin, S. F. Enhanced degradation of halogenated aliphatics by zerovalent iron. Ground Water 1994, 32, 958-967. (37) Scherer, M. M.; Balko, B. A.; Gallagher, D. A.; Tratnyek, P. G. Correlation analysis of rate constants for dechlorination by zerovalent iron. Environ. Sci. Technol. 1998, 32, 3026-3033. (38) Burrow, P. D.; Aflatooni, K.; Gallup, G. A. Dechlorination rate constants on iron and the correlation with electron attachment energies. Environ. Sci. Technol. 2000, 34, 3368-3371.

Received for review August 30, 2005. Revised manuscript received January 5, 2006. Accepted January 6, 2006. ES051720E