Bench-Scale Investigation of Permanganate Natural Oxidant Demand

A large fraction of the organic carbon resisted oxidation over the 21-week duration .... experiments were conducted using a low and high darcy flux (â...
0 downloads 0 Views 169KB Size
Environ. Sci. Technol. 2005, 39, 2835-2840

Bench-Scale Investigation of Permanganate Natural Oxidant Demand Kinetics K E V I N G . M U M F O R D , †,§ N E I L R . T H O M S O N , * ,† A N D RICHELLE M. ALLEN-KING‡ Department of Civil Engineering, University of Waterloo, Waterloo, Ontario, Canada N2L 3G1, and Department of Geology, University at Buffalo, State University of New York, Buffalo, New York 14260

A vital design parameter for any in situ chemical oxidation system using permanganate (MnO4-) is the natural oxidant demand (NOD), a concept that represents the consumption of MnO4- by the naturally present reduced species in the aquifer solids. The data suggest that the NOD of the aquifer material from Canadian Forces Base Borden used in our study is controlled by a fast or instantaneous reaction captured by the column experiments, and a slower reaction as demonstrated by both column and batch test data. These two reaction rates may be the result of the reaction of MnO4- with at least two different reduced species exhibiting widely different rates of permanganate consumption (fast rate >7 g of MnO4- as KMnO4/kg/day and slow rate of ∼0.005 g/kg/day), or a physically/chemically rate-limited single species. The slow NOD reaction prevented fulfillment of the ultimate NOD during the days- to months-long batch experiments and allowed significant early MnO4- breakthrough (>98%) during transport in the column experiments. A large fraction of the organic carbon resisted oxidation over the 21week duration of the batch experiments. This result demonstrates that NOD estimated from total organic carbon measurements can significantly overpredict the NOD value required in the design of an in situ chemical oxidation application.

Introduction In situ chemical oxidation technologies involve the application of a chemical oxidant to the contaminated subsurface to treat target contaminants, such as dense nonaqueous phase liquids (DNAPLs). Frequently used aqueous oxidants are hydrogen peroxide, Fenton’s reagent, persulfate ion, and permanganate ion (MnO4-). Permanganate has received considerable attention because it is more persistent in situ than the other reagents (1). When a permanganate solution is injected into the subsurface, a portion is consumed by the reaction with a variety of nontarget reduced species associated with the aquifer material (e.g., organic carbon and reduced forms of iron, manganese, and sulfur) (2-6). This * Corresponding author phone: (519)888-4567 ext. 2111; fax: (519)888-6197; e-mail: [email protected]. † University of Waterloo. ‡ University at Buffalo, SUNY. § Current address: Department of Civil Engineering, McMaster University, Hamilton, ON, Canada, L8S 4L7. 10.1021/es049307e CCC: $30.25 Published on Web 03/01/2005

 2005 American Chemical Society

consumption, which is expressed as the mass of oxidant consumed per mass of dry solids (3, 7), is referred to as natural oxidant demand (NOD) with single-value estimates usually reported (2, 3, 5). Carbonaceous matter (non-carbonate carbon material including combustion products as well as organic matter) can exist in a variety of forms in the subsurface environment (e.g., 8, 9), some of which are more resistant to oxidation than others. The possibility of multiple reduced inorganic species as well as a variety of forms of carbonaceous matter creates an extremely heterogeneous environment within which NOD reactions occur. However, the extent to which different species comprising aquifer materials contribute to NOD is not well understood. In addition to the chemical heterogeneity imposed by the presence of various reactant forms, a fraction of the reduced species associated with aquifer solids may be physically less accessible due to incorporation into mineral matrices or aggregates (e.g. 10, 11). When significant, the above reactions can negatively affect the performance of an in situ chemical oxidation application by increasing the oxidant dosing requirement and concomitantly increasing treatment cost (12). Thus, consideration of the NOD is an important design criterion at potential in situ chemical oxidation treatment sites. The existing conceptual model for NOD assumes that the reaction of oxidant with any reduced species present in the aquifer material is extremely rapid (13) and represents an instantaneous sink for injected oxidant. This conceptual model, suggested by Barcelona and Holm (4), treats NOD as a single value applicable throughout an entire oxidant application. The implication of this instantaneous sink conceptual model is that no injected oxidant can leave a region of an aquifer prior to the satisfaction of the entire NOD within that region. An alternative conceptual model is to treat the NOD as a number of parallel reactions between the oxidant and reduced aquifer solid species that occur at different rates and with different stoichiometries, as illustrated for an idealized pore space containing residual DNAPL and various reduced aquifer solid species in Figure 1. The single NOD value used in the existing conceptual model represents the ultimate NOD, the maximum possible oxidant mass that could be consumed. The existing conceptual model assumes that the reaction between MnO4- and reduced aquifer solid species is rapid relative to the rates of dissolved DNAPL reactions and mass transport processes; thus permanganate transport is possible only after consumption of the ultimate NOD. In the alternative conceptual model, the transport and consumption of aqueous oxidant is not strictly a function of the ultimate NOD and contaminant demand, but is a function of transport processes and reaction rates associated with all dissolved DNAPL components and reduced aquifer solid species. The primary objective of this study was to demonstrate and quantify the kinetic nature of NOD processes with aquifer material. Column experiments were used to estimate the NOD during oxidant transport through aquifer material, and batch tests were used to generate supporting data to further our understanding of the relations between aquifer material properties and NOD temporal variation. A focus in this study was to test the importance of organic matter as a primary reductant in a typical low organic carbon content aquifer material. All experiments were performed using permanganate as the oxidant and uncontaminated material from the well-studied Canadian Forces Base Borden (Borden) aquifer (e.g., 14-17) as a representative low carbon content aquifer solid with attributes typical of many aquifers within which persistent DNAPLs reside (18, 19). Typical of many VOL. 39, NO. 8, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2835

FIGURE 1. Pore-scale conceptual model for NOD showing the possibility of reaction with reduced aquifer solid species, reaction with dissolved DNAPL species, and transport of unreacted MnO4-. Reduced species include: (a) pyrite (dark brown cubic grains) and organic matter coatings (orange) on grains, (b) reduced minerals such as magnetite (narrow lath) and reduced carbonaceous matter such as charcoal, (c) dissolved and nonaqueous phase oil (DNAPL, gray), and (d) organic matter and disseminated pyrite entrained within carbonate grains and lithic fragments. Nonreactive solids shown include quartz (clear), feldspar (pink), and pure calcite (green). sedimentary aquifers, the Borden aquifer contains a variety of forms of carbonaceous matter and reduced minerals (20, 21) that may react with chemical oxidants.

Experiment Methods Aquifer Material. The experiments were conducted using random subsamples of dried, homogenized material from saturated zone core samples collected in the sand pit area. Additional details concerning aquifer material collection and preparation are described in the Supporting Information. The organic carbon (OC) content of the aquifer solids is 0.020%, consistent with prior measurements of depthintegrated samples from the sand pit area of the Borden aquifer (21). Throughout this manuscript, we use OC to designate the carbonaceous matter, determined by hightemperature combustion following acidification to remove carbonates (22). X-ray diffraction analyses confirmed that the minerals present in this aquifer sample are consistent with previously reported observations (21). Aqueous KMnO4 Preparation and Analysis. All KMnO4 solutions were prepared by adding analytical grade KMnO4 (EM Science) to Milli-Q water and boiling for ∼1 h. The cooled solution was filtered (0.45-µm glass fiber, Pall Corporation) and standardized by titration into a sulfuric acid and sodium oxalate solution (23). KMnO4 concentrations were determined by spectrophotometry (Milton Roy Co., Spectronic 20D) at 525 nm with a method detection limit of 1.3 mg/L. The spectrophotometer was calibrated prior to batch test sampling events and prior to each column experiment. 2836

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 8, 2005

Column Experiments. Four aquifer sand column experiments (Col-1 through Col-4) and a control column experiment were performed to observe temporal variation in NOD during MnO4- transport. The column experiments were conducted using a low and high darcy flux (∼47 and ∼50 cm/day) in combination with a low and high influent KMnO4 concentration (∼2.2 and ∼4.5 g/L). The approximately 13.5 cm long by 2.5 cm inside diameter sand columns were dry packed in ∼1 cm compacted lifts and contained glass beads to act as flow distributors at either end. The control column was packed exclusively with glass beads, and mass balance calculations indicated no evidence of MnO4- consumption due to contact with the experimental apparatus. All experiments were performed at 18-20 °C. Each column was purged with CO2 and saturated with deionized water prior to oxidant flushing (1 or 2 flushes per column). Oxidant flushing was followed by deionized water flushing. A conservative tracer (70-100 mg/L Br-) test was performed before the initial and after the final water flush. The influent reservoir was sampled at least once during each tracer test and oxidant flush. Additional column experiment setup and details, including specific injection rate and KMnO4 concentration, tracer solution analysis method, and physical dimensions, are provided in the Supporting Information. During the tracer tests, column effluent samples were collected every 4 or 15 min at high or low flow rates, respectively, and analyzed by ion chromatography (Dionex). During the oxidant and deionized water flushes, the column effluent was analyzed continuously for KMnO4 concentration using a flow through cell (Fisher Scientific, 14-385-926A) and a data recorder (Sciemetric Instruments Inc., model LLSYS). Batch Experiments. Three groups of batch tests (BT-1 through BT-3) were conducted to investigate the relationship between selected aquifer material properties and long-term NOD variation. Subsamples of bulk aquifer material were used in BT-1 to directly complement the column experiments. Because organic carbon content in the Borden aquifer is correlated to grain size (21), three grain size fractions were used in BT-2 to explore the relationship between NOD and OC content. Subsamples of crushed (98% of the influent concentration) was observed. The shape of these breakthrough curves is typical of those from Col-1, -2, and -3. The permanganate mass that exited each column was determined by piecewise integration of the permanganate breakthrough curves. Mass balance calculations indicated no evidence of MnO4- consumption during the second oxidant flush in Col-1 through Col-4 (net mass loss ) 0.00

FIGURE 2. Col-4 breakthrough curves showing data at early time with the time line of the start of the second permanganate flush shifted to coincide with the start of the first permanganate flush and the tracer tests. g, stdev ) 0.01 g). Despite these high permanganate mass recoveries, we observed a slight difference between the Brbreakthrough curve and the shifted permanganate breakthrough curve, suggesting a very slow reaction resulting in minimal permanganate consumption during the second flushing period. It is unclear whether this reaction is due to the catalytic decomposition of MnO4- resulting from contact with mineral surfaces or MnO2 (produced from the reaction with reduced aquifer material species) (24) or the reaction with reduced aquifer species. NOD values for Col-1 through Col-4 were 0.26 (37% relative standard deviation (rsd)), 0.63 (23% rsd), 0.37 (57% rsd), and 0.15 (34% rsd) g/kg, respectively, with an average of 0.35 g/kg (all expressed as mass of KMnO4 per mass of dry aquifer material). The large relative standard deviation in these results, calculated using propagation of error techniques (25), is due to small variations in the injection rate and the small differences between the MnO4- mass injected and recovered. Because the injection rate was well controlled (rsd ) 2.7% at flow rates of 0.16-0.53 mL/ min), it was accepted that large relative errors were a consequence of tests conducted with low demand aquifer material. It is expected that similar testing of higher demand aquifer material would result in less relative error. There is a distinct delay between the Br- breakthrough curve and first permanganate flush breakthrough curve (Figure 2). The permanganate breakthrough curve also demonstrated a slow approach to greater than 98% breakthrough (C/Co > 0.98) (Figure 2). The appearance of permanganate as a sharp front after the tracer breakthrough is consistent with a fast reaction (KMnO4 consumption rate >7.2 g/ kg/day) and suggests that breakthrough was possible only after complete or near complete consumption of one or more physically accessible reduced aquifer solid species. Despite minor variations between permanganate results from different column experiments (Figure 3), we observed that the columns injected with 2.1-2.2 g/L KMnO4 solution required the injection of a greater volume of oxidant solution to achieve initial breakthrough as compared to those injected with ∼4.5 g/L KMnO4 (Figure 3). We interpret this result to indicate the critical mass of MnO4- required for consumption of the “fast reacting” reduced aquifer solid species. Assuming piston-like flow conditions, the average permanganate mass consumed that would yield the initial delay or retardation of the MnO4- breakthrough curve was estimated to be 0.10 g of KMnO4/kg (8.4% rsd). Surprisingly, this fast demand was relatively consistent across all four columns and represents between 16% and 60% of the total NOD realized in each of the column experiments. The deviation of the late-time or tail portion of each of the permanganate breakthrough curves as compared to the tracer breakthrough

FIGURE 3. Early portion of the permanganate breakthrough curves from Col-1 through Col-4 and a typical Br- tracer breakthrough curve. Note the similarity between breakthrough curves from columns injected with similar concentrations but at different darcy fluxes (between Col-1 and Col-2, and between Col-3 and Col-4).

FIGURE 4. Average NOD results (symbols) from batch tests with the different size fractions (BT-2) and the bulk aquifer material (BT-1). Error bars represent the standard deviation based on triplicate reactors. curves suggests the presence of an ongoing slow reaction of MnO4- (KMnO4 consumption rate ∼0.04 g/kg-day) with reduced aquifer solid species (Figures 2 and 3). Batch Experiments. The NOD for each of the batch tests was quantified by the change in KMnO4 mass relative to the control divided by the mass of aquifer material. NOD increased rapidly prior to the first sampling episode and more slowly thereafter (Figure 4). The bulk material had a NOD of ∼1.2 g/kg following 21 weeks of exposure. It is apparent that the NOD has yet to reach a constant or maximum value by the end of the experiment (21 weeks) in all except the finegrained reactors, and therefore these data provide only a minimum bound on the ultimate NOD. The general trend of a declining rate of increasing NOD with time observed in the batch experiments is similar to the results presented by Chambers et al. (26), who also demonstrated that NOD was increasing at the end of their 14-day batch experiments. The much lower NOD observed in the column experiments (average of 0.35 g/kg) compared to the batch experiments (>1.2 g/kg) demonstrates that significant transport of MnO4- in Borden aquifer material occurs when far less than complete consumption of the reduced aquifer solid species has occurred. These batch experiments suggest that slow oxidation of reduced species in the aquifer solids is likely continuing at the termination of each of the column experiments at a rate less than detectable by the column design and analytical methods employed. In these column experiments, it is not possible to resolve permanganate mass loss equivalent to bulk material > fine fraction (Figure 4). It has previously been noted (14) that, because the OC content is correlated to the fraction of carbonate grains, it is likely that the majority of the OC is incorporated in the larger carbonate grains. The relationship between NOD and particle size is likely attributable to the different OC concentrations present in the various size fractions (Table 1) and suggests that OC is a principal reactant of NOD. The theoretical prediction of MnO4-:OC stoichiometry for the complete oxidation of a simple carbohydrate (27) by permanganate as given by

3CH2O + 4MnO4- f 3CO2 + 4MnO2(s) + H2O + 4OH(1) is 13.2 (wt/wt). Oxidation of more oxidized forms of organic matter (e.g., organic acids) or incomplete oxidation of carbohydrate to organic acid (e.g., 27) results in stoichiometric ratios as low as one-half of that cited above, while oxidation of more reduced forms of organic matter produces somewhat greater ratios. The MnO4-:OC stoichiometry observed in these batch experiments for the bulk material, medium-coarse, and coarse fractions (9 ( 6 to 15 ( 8, Table 1) is comparable to the values for complete oxidation of a simple carbohydrate. The relatively large uncertainties in our results arise from the reasonable measurement uncertainty associated with distinguishing OC loss in low OC aquifer material. Nonetheless, these results are consistent with the hypothesis that the primary NOD in the aquifer material is asserted by the OC. Although OC contents decreased in each of the test reactors, 60-90% of the initial OC remained after 14 weeks of exposure. It is likely that the ultimate NOD is reached prior to consumption of the total OC. The carbonaceous matter in the Borden aquifer material is quite heterogeneous, including a variety of more condensed, and probably recalcitrant forms, such as intertinite (black carbon) and high reflectance vitrinite (28, Louigi 2002, unpublished data, personal communication). Incomplete wet oxidation of some forms of carbonaceous matter by persulfate, a stronger oxidant than permanganate (Eo of 2.0 V as compared to 1.7 2838

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 8, 2005

FIGURE 5. Best fit lines to the average NOD data presented in Figure 4 using the empirical model NOD(t)/NODmax ) a ln t + b (a and b are fitting parameters) with NODmax estimated from the initial OC content and the stoichiometry given by eq 1. The coefficient of determination is indicated for each fit. V for permanganate), is a well-recognized limitation of this technique for OC analysis (28). The implication is that significant overestimation of the actual in situ natural OCderived permanganate demand is likely if complete oxidation of all OC is assumed. Rate-Limiting Processes and Conceptual Models. The BT-3 batch test (using crushed (7.2 g/kg/day) observed in the column data due, in part, to the temporal sampling program used in this batch test investigation. The data presented here suggest the presence of either two (or a continuum of) reactions with distinct intrinsic rates (fast or instantaneous, and slow) or/ and a chemically fast reaction that is physically rate limited. Pertinent examples of both of these processes exist in the literature. Previous reports have coupled oxidation reaction rate to the chemistry of various aquifer material species. For example, the reaction rates of phenols and phenolic functional groups in organic matter are fast as compared to those of more aromatic structures (24, 30). The observed oxidation kinetics of Fe(II) present in aquifer materials has been successfully modeled using three rates (instantaneous, rapid, and slow) representing reactions between pseudo-species of iron and a probe chemical, hexavalent chromium (31). Finally, Ball and Roberts (11) observed that the sorption rate of probe hydrophobic contaminants in Borden aquifer material is consistent with intragranular diffusion to reactive sites within porous grains (14), and more recent studies have confirmed the presence of heterogeneous carbonaceous matter in this aquifer material (32). The heterogeneous nature of both the carbonaceous matter and the iron species present within Borden aquifer material coupled with physical-rate limitations provides ample opportunities for a continuum of NOD reaction rates to be observed. Future research efforts are required to determine mechanisms controlling permanganate consumption in this aquifer material.

Supporting Information Available

Implications for In Situ Chemical Oxidation Applications. The Borden aquifer is representative (i.e., low OC content and concentration of reduced inorganic species capable of being oxidized by permanganate) of many aquifers in which chlorinated solvents persist and in situ chemical oxidation applications are considered. Although aspects of this aquifer’s properties may not be similar to all sites where a permanganate-based in situ chemical oxidation application is considered for source zone treatment, a number of findings from our work may be generally relevant to other field sites. Clearly, consideration must be directed toward understanding the kinetics of the reaction between permanganate, or other oxidants, and the aquifer material rather than assuming the reduction capacity of the aquifer material to be an instantaneous sink of the injected oxidant. Simple estimates of NOD based on the OC content will lead to a significant overprediction of the NOD magnitude in circumstances where the majority of the OC content is resistant to oxidation over the time scale of several months. This approach undervalues the proportion of injected oxidant available for reaction with the target contaminants. NOD data generated from column experiments will underestimate the maximum NOD but yield a value that may be more representative of in situ application conditions. Long-term (>several months) batch experiments are required to predict ultimate NOD values. Finally, all bench-scale determinations of NOD should be conducted at concentrations and test durations that are representative of the proposed application conditions.

K.G.M. was supported by a Natural Sciences and Engineering Research Council (NSERC) of Canada Postgraduate Scholarship, and an Ontario Graduate Scholarship. Additional funding was provided by a NSERC Discovery Grant (N.R.T.). Organic carbon analyses were conducted by S. Chatten of the University of Waterloo Earth Science Geochemistry Lab. This manuscript also benefited from petrographic analyses by B. Louigi, LAOP, Tu ¨ bingen, Germany.

Additional information on experimental methods, column details and operating conditions, results from BT-3, and the empirical model best-fit parameters. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) USEPA Field Applications of In Situ Remediation Technologies: Chemical Oxidation; EPA542-R-98-008, 1998. (2) Hood, E. Permanganate flushing of DNAPL source zones: Experimental and numerical investigation. Ph.D. Thesis, University of Waterloo, Department of Civil Engineering, Waterloo, ON, 2000. (3) Siegrist, R. L.; Urynowicz, M. A.; West, O. R.; Crimi, M. L.; Lowe, K. S. Principles and Practices of In Situ Chemical Oxidation Using Permanganate; Battelle Press: Columbus, OH, 2001. (4) Barcelona, M. J.; Holm, T. R. Oxidation-reduction capacities of aquifer solids. Environ. Sci. Technol. 1991, 25, 1565-1572. (5) Drescher, E.; Gavaskar, A. R.; Sass, B. M.; Cumming, L. J.; Drescher, M. J.; Williamson, T. K. J. Batch and column testing to evaluate chemical oxidation of DNAPL source zones. In Proc. from the 1st International Conf. on Remediation of Chlorinated and Recalcitrant Compounds; Battelle: Monterey, CA, 1998. (6) Li, X. D.; Schwartz, F. W. Efficiency problems related to permanganate oxidation schemes. In Proc. from the 2nd International Conf. on Remediation of Chlorinated and Recalcitrant Compounds; Battelle: Monterey, CA, 2000. (7) MacKinnon, L. K.; Thomson, N. R. Laboratory-scale in situ chemical oxidation of a perchloroethylene pool using permanganate. J. Contam. Hydrol. 2002, 56, 49-74. (8) Allen-King, R. M.; Grathwohl, P.; Ball, W. P. New modeling paradigms for the sorption of hydrophobic organic chemicals to heterogeneous carbonaceous matter in soils, sediments, and rocks. Adv. Water Res. 2002, 25, 985-1016. (9) Gustafsson, O.; Bucheli, T. D.; Kukulska, Z.; Andersson, M.; Largeau, C.; Rouzaud, J. N.; Reddy, C. M.; Eglinton, T. I. Evaluation of a protocol for the quantification of black carbon in sediments. Global Biogeochem. Cycles 2001, 15, 881-890. (10) Holmen, B. A.; Gschwend, P. M. Estimating sorption rates of hydrophobic organic compounds in iron oxide- and aluminosilicate clay-coated aquifer sands. Environ. Sci. Technol. 1997, 31, 105-113. (11) Ball, W. P.; Roberts, P. V. Long-term sorption of halogenated organic-chemicals by aquifer material. 2. Intraparticle diffusion. Environ. Sci. Technol. 1991, 25, 1237-1249. (12) Mumford, K.; Lamarche, C.; Thomson, N. R. Natural oxidant demand of aquifer materials using the push-pull technique with permanganate. ASCE J. Envir. Eng. 2004, 130 (10), 11391146. doi: 10.1061/ASCE0733-93722004130:101139. (13) Zhang, H.; Schwartz, F. W. Simulating the in situ oxidative treatment of chlorinated ethylenes by potassium permanganate. Water Resour. Res. 2000, 36, 3031-3042. (14) Ball, W. P.; Roberts, P. V. Long-term sorption of halogenated organic-chemicals by aquifer material. 1. Equilibrium. Environ. Sci. Technol. 1991, 25, 1223-1237. (15) Allen-King, R. M.; Halket, R. M.; Gaylord, D. R.; Robin, M. J. L. Characterizing the heterogeneity and correlation of perchloroethene sorption and hydraulic conductivity using a faciesbased approach. Water Resour. Res. 1998, 34, 385-396. (16) Mackay, D. M.; Freyberg, D. L.; Roberts, P. V.; Cherry, J. A. A natural gradient experiment on solute transport in a sand aquifer. 1. Approach and overview of plume movement. Water Resour. Res. 1986, 22, 2017-2029. (17) Mackay, D. M.; Ball, W. P.; Durant, M. G. Variability of aquifer sorption properties in a field experiment on groundwater transport of organic solutes: Methods and preliminary results. J. Contam. Hydrol. 1986, 1, 119-132. VOL. 39, NO. 8, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2839

(18) Jackson, R. E. The migration, dissolution, and fate of chlorinated solvents in the urbanized alluvial valleys of the southwestern USA. Hydrogeol. J. 1998, 6, 144-155. (19) Rivett, M. O.; Feenstra, S.; Cherry, J. A. A controlled field experiment on groundwater contamination by a multicomponent DNAPL: creation of the emplaced-source and overview of dissolved plume development. J. Contam. Hydrol. 2001, 49, 111-149. (20) Ran, Y.; Xiao, H.; Huang, W. L.; Peng, P. A.; Liu, D. H.; Fu, J. M.; Sheng, G. Y. Kerogen in aquifer material and its strong sorption for nonionic organic pollutants. J. Environ. Qual. 2003, 32, 17011709. (21) Ball, W. P.; Buehler, C. H.; Harmon, T. C.; MacKay, D. M.; Roberts, P. V. Characterization of a sandy aquifer material at the grain scale. J. Contam. Hydrol. 1990, 5, 253-295. (22) Churcher, P. L.; Dickhout, R. D. Analysis of ancient sediments for total orgnaic carbon - Some new ideas. J. Geochem. Explor. 1987, 29, 235-246. (23) Franson, M. A. H. Standard Methods for the Analysis of Water and Wastewater, 20th ed.; American Public Health Association: Washington, DC, 1998. (24) Stewart, R. Oxidation by permanganate. In Oxidation in Organic Chemistry; Wilberg, K. B., Ed.; Academic Press: New York, 1965. (25) Miller, J. C.; Miller, J. N. Statistics for Analytical Chemistry, 3rd ed.; Ellis Horwood Limited: New York, 1993. (26) Chambers, J.; Leavitt, A.; Walti, C.; Schreier, C. G.; Melby, J. Treatability study-fate of chromium during oxidation of chlorinated solvents. Proc. from the 2nd International Conf. on

2840

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 8, 2005

(27) (28)

(29) (30)

(31)

(32)

Remediation of Chlorinated and Recalcitrant Compounds; Battelle: Monterey, CA, 2000. Stumm, W.; Morgan, J. J. Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters; John Wiley & Sons: New York, 1996. Powell, R. M.; Bledsoe, B. E.; Curtis, G. P.; Johnson, R. L. Interlaboratory methods comparison for the total organiccarbon analysis of aquifer materials. Environ. Sci. Technol. 1989, 23, 1246-1249. Grathwohl, P. Diffusion in Natural Porous Media; Kluwer Academic Press: Boston, MA, 1998. Matthiessen, A. Evaluating the redox capacity and the redox potential of humic acids by redox titrations. In Humic Substances in the Global Environment and Implications on Human Health; Senesi, N., Miano, T. M., Eds.; Elsevier Science: Amsterdam, 1994. Hwang, I.; Batchelor, B.; Schlautman, M. A.; Wang, R. Effects of ferrous ion and molecular oxygen on chromium(VI) redox kinetics in the presence of aquifer solids. J. Hazard. Mater. 2002, B92. Taylor, K. M.; Allen-King, R. M.; Gaylord, D. R. Characterization of geochemical factors controlling PCE sorption to carbonaceous matter by lithofacies. AGU Fall Meeting; American Geophysical Union: San Francisco, CA, 2004.

Received for review May 7, 2004. Revised manuscript received December 21, 2004. Accepted January 19, 2005. ES049307E