Hydrous Ferric Oxide Incorporated Diatomite for Remediation of

Among them, Fe0 has been extensively utilized as a reactive medium in PRBs .... (1 g), pH (6.8), empty bed contact time (EBCT) (3.3 min), arsenic (10 ...
0 downloads 0 Views 135KB Size
Download PDF
Environ. Sci. Technol. 2007, 41, 3322-3328

Hydrous Ferric Oxide Incorporated Diatomite for Remediation of Arsenic Contaminated Groundwater M I N J A N G , * ,† S O O - H O N G M I N , † JAE KWANG PARK,† AND ERIC J. TLACHAC‡ Department of Civil and Environmental Engineering, University of WisconsinsMadison, 1415 Engineering Drive, Madison, Wisconsin 53706, and Natural Resource Technology, Incorporated, 23713 West Paul Road, Suite D, Pewaukee, Wisconsin 53072

Two reactive media [zerovalent iron (ZVI, Fisher Fe0) and amorphous hydrous ferric oxide (HFO)-incorporated porous, naturally occurring aluminum silicate diatomite [designated as Fe (25%)-diatomite]], were tested for batch kinetic, pH-controlled differential column batch reactors (DCBRs), in small- and large-scale column tests (about 50 and 900 mL of bed volume) with groundwater from a hazardous waste site containing high concentrations of arsenic (both organic and inorganic species), as well as other toxic or carcinogenic volatile and semivolatile organic compounds (VOC/SVOCs). Granular activated carbon (GAC) was also included as a reactive media since a permeable reactive barrier (PRB) at the subject site would need to address the hazardous VOC/SVOC contamination as well as arsenic. The groundwater contained an extremely high arsenic concentration (341 mg L-1) and the results of ion chromatography and inductively coupled plasma mass spectrometry (IC-ICP-MS) analysis showed that the dominant arsenic species were arsenite (45.1%) and monomethyl arsenic acid (MMAA, 22.7%), while dimethyl arsenic acid (DMAA) and arsenate were only 2.4 and 1.3%, respectively. Based on these proportions of arsenic species and the initial As-to-Fe molar ratio (0.15 molAs molFe-1), batch kinetic tests revealed that the sorption density (0.076 molAs molFe-1) for Fe (25%)-diatomite seems to be less than the expected value (0.086 molAs molFe-1) calculated from the sorption density data reported by Lafferty and Loeppert (Environ. Sci. Technol. 2005, 39, 21202127), implying that natural organic matters (NOMs) might play a significant role in reducing arsenic removal efficiency. The results of pH-controlled DCBR tests using different synthetic species of arsenic solution showed that the humic acid inhibited the MMAA removal of Fe (25%)diatomite more than arsenite. The mixed system of GAC and Fe (25%)-diatomite increased the arsenic sorption speed to more than that of either individual media alone. This increase might be deduced by the fact that the addition of GAC could enhance arsenic removal performance of Fe (25%)-diatomite through removing comparably high portions * Corresponding author phone: +82-2-3702-6592; fax: +82-23702-6609; e-mail: [email protected]; current address: Soil Remediation Team, Korea Mine Reclamation Corporation, Coal Center, 80-6 Susong-dong, Jongno-gu, Seoul, 110-727, Korea, Republic of. † University of Wisconsin--Madison. ‡ Natural Resource Technology, Incorporated. 3322

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 9, 2007

of NOMs. Small- and large-scale column studies demonstrated that the empty bed contact time (EBCT) significantly affected sorpton capacities at breakthrough (C ) 0.5 C0) for the Fe0/sand (50/50, w/w) mixture, but not for GAC preloaded Fe (25%)-diatomite. In the large-scale column tests with actual groundwater conditions, the GAC preloaded Fe (25%)diatomite effectively reduced arsenic to below 50 µg L-1 for 44 days; additionally, most species of VOC/SVOCs were also simultaneously attenuated to levels below detection.

Introduction Human activities and natural phenomena have caused releases of arsenic into groundwater and surface water, creating potentially serious environmental problems for humans and other living organisms. Since arsenic contamination is a health risk for many countries around the world, there is an urgent demand for a highly effective, reliable, and economical technique for the removal of arsenic from groundwater. Permeable reactive barrier (PRB) technology may be a practical and economical alternative to conventional groundwater remediation systems for treating arsenic contaminated groundwater. PRBs involve the placement or formation of a reactive treatment zone in the path of a dissolved contaminant plume with the objective of passively removing target contaminants or altering them by physical, chemical, and/ or biological processes to reduce their toxicity and/or mobility in the subsurface. PRBs have been successfully applied to many organic compounds such as chlorinated ethenes, including trichloroethylene and tetrachloroethylene, and inorganic contaminants such as chromium, uranium, arsenic, and other dissolved metals in groundwater (1-5). PRBs have become increasingly popular because the operation and maintenance costs are significantly less expensive than the traditional “pump and treat” method. Many materials have been used as a reactive medium in pilot-scale PRBs, including recycled foundry waste, zerovalent iron (Fe0), activated alumina, and ferric oxide. Among them, Fe0 has been extensively utilized as a reactive medium in PRBs because it has been shown to be effective in removing anionic metals, such as Cr(VI), Cr(III), As(V), and As(III), as well as halogenated organic compounds by means of coprecipitation and reductive dehalogenation (6-8). Several forms of Fe0 with good structural and hydrodynamic properties are also commercially available. However, Fe0 can potentially elevate pH and Fe(II) concentration in groundwater due to a corrosion reaction with dissolved oxygen. This reaction will change groundwater conditions, resulting in potentially undesirable chemical reactions and a reduction in oxidation potential. In addition, due to relatively slow adsorption kinetics and competition effects of other anions such as silicate or phosphate, it is necessary to design a PRB with an extra thickness of Fe0 materials (9). To overcome these drawbacks of Fe0, an effective and economical adsorptive medium has been developed in our previous study for the removal of arsenic from water (10). Previous research has shown that amorphous hydrous ferric oxide (HFO) incorporated into porous, naturally occurring aluminum silicate diatomite [designated as Fe (25%)-diatomite] was more efficient than the conventional medium, AAFS-50, for arsenite and arsenate removal, and that the increase in efficiency was particularly pronounced with respect to arsenite (10). Diatomite (or diatomaceous earth) is a lightweight sedimentary rock composed principally 10.1021/es062359e CCC: $37.00

 2007 American Chemical Society Published on Web 03/31/2007

of silica microfossils of aquatic unicellular algae having a variety of pore structures with up to 80-90% voids. Diatomite has not only been approved as a food-grade material by the Food and Drug Administration (FDA), but it is also stable in the liquid phase since it originated and is produced from a sea or lake. These properties will result in good stability in a saturated environment, such as a PRB, as well as environmental friendliness. Meanwhile, amorphous HFO has been widely studied as a promising adsorptive material for removing both arsenate and arsenite (11) due to its high selectivity for arsenic species. An incipient wetness impregnation method using a vortexing device was developed to disperse and incorporate an HFO precursor homogeneously on the pore surfaces of diatomite. Since a large volume of precursor solution is not needed or wasted, this technique is simple, economical, and environmentally friendly. In this study, two reactive media [Fe0 and Fe(25%)diatomite] were tested with groundwater from a hazardous waste site in northern Wisconsin containing high concentrations of both organic and inorganic species of arsenic, as well as other toxic or carcinogenic VOC/SVOCs. Granular activated carbon (GAC) was also included in this study as a reactive media since a PRB at the subject site would need to address the hazardous VOC/SVOC contaminations as well as arsenic. Adsorption with GAC is a common, proven technology for removal of VOC/SVOC compounds from water. The VOC/SVOC removal was characterized in the largescale column tests as described below to determine the effect of simultaneous arsenic adsorption with both media. In an attempt to discover another purpose for the use of GAC, we tried to find some synergetic effects of GAC on the arsenic removal performance of Fe (25%)-diatomite by removing natural organic matters (NOMs) in groundwater, which have ubiquitous presence in natural aquatic systems and significant effects on the arsenic mobility due to their redox and complexation capabilities. Since NOMs have negatively charged functional groups such as carboxylic, phenolic, quinone, amino, sulfhydryl, nitroso, and hydroxyl functional groups (12) at a neutral pH, they could give not only significant inhibitory effects on the arsenic removal performance of Fe (25%)-diatomite, but also desorption effects that are comparative to phosphate (13). Fisher Fe0, Fe (25%)-diatomite, and GAC were first subjected to kinetics tests to determine arsenic adsorption rates and capacities with actual groundwater and synthetic solutions. Fisher Fe0 showed not only the highest arsenic removal speeds, but also the highest arsenic sorption capacities among the other commercialized Fe0 products (7). Then, the Fe0/sand (50/50, w/w) mixture and GAC preloaded Fe (25%)-diatomite were tested in smallscale column tests with a fast flow rate. Finally, these two reactive media were tested in large-scale columns to determine breakthroughs with actual groundwater flow conditions at the subject site. All testing utilized groundwater from the subject site. The objectives of this study were (1) to investigate the arsenic removal efficiences of Fe (25%)-diatomite compared with Fe0, (2) to evaluate the enhancement of arsenic removal performance with the addition of GAC, and (3) to find out the applicabilities of dual media [GAC preloaded Fe (25%)-diatomite] for actual high concentrations of organic and inorganic arsenic, as well as VOC/SVOC contaminated groundwaters through different sizes of column tests.

Materials and Methods Sample Collection and Characterization. The bulk groundwater sample was collected from a monitoring well at the subject site. This particular location was selected based upon analytical data from previous characterization work. The bulk sample was collected utilizing low-flow methods (14) to minimize aeration of the groundwater during sample collection. This was done to minimize changes in arsenic

speciation and loss of VOC/SVOC compounds which can occur during sample aeration. The bulk sample was filtered during collection with a 0.45-µm inline disposable groundwater filter to remove suspended solids that could interfere with laboratory chemical analysis and testing. Following filtration, the bulk sample was preserved in a refrigerator (4 °C) during long-term storage and handling, then pumped directly into 19 L Tedlar and/or Teflon bags to minimize aeration for column tests. The arsenite oxidation processes and microbial arsenic transformations considered for these procedures were negligible. The total arsenic concentrations of groundwater or synthetic solutions were analyzed with an inductively coupled plasma atomic emission spectrometer (ICP-AES, Jobin Yyon Inc., Edison, NJ) for samples having arsenic concentrations larger than 1 mg L-1. An atomic absorption spectrophotometer (AAS, Varian AA-975) and a GTA-95 graphite tube atomizer with a programmable sample dispenser (Palo Alto, California) were also used for samples having arsenic concentrations lower than 1 mg L-1. In this case, a 50 mg L-1 nickel solution was used as a matrix modifier. The detection limits of ICP-AES and AAS-graphite were 0.033 and 0.0006 µmol L-1 for arsenic, respectively. Ion chromatography and inductively coupled plasma mass spectrometry (IC-ICPMS) were used to quantify arsenic species in groundwater samples. The chromatographic separation of arsenic species was achieved using gradient elution with dilute sodium hydroxide on a high-capacity anion exchange column. A DX500 ion chromatograph (Dionex) and an Elan 6000 ICP-MS (Perkin-Elmer) were used for arsenic speciation analysis. The groundwater sample also contained 2-3 mg L-1 volatile organic compounds (VOCs) and 3-4 mg L-1 semi-volatile organic compounds (SVOCs). VOC/SVOC compounds were identified and quantified by gas chromatography and mass spectrometry (GC-S) following USEPA methods SW8468260B (VOCs) and 8270B (SVOCs). Analytical testing for chemical composition and VOC/SVOC compounds of the groundwater was conducted at TestAmerica Analytical Testing Corporation (Watertown, WI). Preparation of Fe (25%)diatomite is described in the Supporting Information. Batch Kinetic Study. A groundwater sample of 300 mL was utilized for each kinetic of Fe (25%)-diatomite, Fe0, GAC, or mixed media [GAC and Fe (25%)-diatomite]. These samples were purged with nitrogen gas for 24 h to remove the VOC/ SVOC compounds prior to kinetics. Except Fe0 (half volume), masses of media were predetermined to have the same volume (Table S3 in the Supporting Information). Most cases of the column studies showed the mixture of Fe0 with sand at the condition of 50/50 (w/w) to avoid a significant clogging effect. An aliquot of 1.5 mL of suspension was withdrawn at 10-60 min intervals and centrifuged at a speed of 10 000 rpm for 5 min. A 1-mL sample of the suspension was diluted with 19 mL of 1% HNO3 solution prepared with deionized water. These samples were then analyzed for total arsenic utilizing an ICP-AES (Jobin Yyon Inc., Edison, NJ). pH-Controlled Differential Column Batch Reactors (DCBRs). The pH-controlled DCBR tests were conducted not only to estimate the removal kinetics of different arsenic species on Fe (25%)-diatomite, but also to find the NOM addition effect on arsenite and MMAA that were the dominant arsenic species of groundwater. Figure S1 in the Supporting Information shows the schematics of pH-controlled DCBRs. The most important reason to use pH-controlled DCBR is to control the pH condition of feeding arsenic solution because pH is one of the most significant parameters in affecting arsenic species and active surface sites of metal oxide (10). For the evaluation of the effect of NOM on the arsenic removal for Fe (25%)-diatomite, humic acid (sodium salt, SigmaAldrich) was used as a representative of NOM species in groundwater (15, 16). Arsenite standard solution for ICP VOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3323

TABLE 1. Small-Scale and Large-Scale Column Tests Setup and Operational Conditions column ID

flowrate (mL h-1)

SLVa (m d-1)

EBCTb (day)

A B

6.46 6.76

0.88 0.92

0.3 0.310

small-scale column tests Fe0/sand (50/50, w/w) 4 g of GAC preloaded 10 g of Fe (25%)-diatomite

A B

7.33 7.79

0.088 0.094

5.1 4.7

large-scale column tests Fe0/sand (50/50, w/w) 40 g of GAC preloaded 200 g of Fe (25%)-diatomite

media setup

bed volume (mL)

total mass of media

46.9 50.2

90 14

886.9 937.1

1,800 240

a SLV ) Surface loading velocity. b EBCT is the time required for the liquid in an adsorption bed to pass through the column assuming that all liquid passes through at the same velocity. It is equal to the volume of the empty bed divided by the flow rate.

(arsenic trioxide, Sigma-Aldrich), sodium arsenate heptahydrate (Na2HAsO4‚7H2O), monosodium acid methane arsonate (Na2HAsNaO3, Chem Service), and cacodylic acid (C2H7AsO2, Sigma-Aldrich) were used to prepare 1000 mg L-1 stock solutions of arsenite, arsenate, MMAA, and DMAA, respectively. The pH-controlled DCBR tests were run with the following conditions: media mass (1 g), pH (6.8), empty bed contact time (EBCT) (3.3 min), arsenic (10 mg as As L-1), and humic acid (100 mg L-1, 50.3-65.3 mg C L-1). For arsenic kinetic results, a pseudo second-order kinetic equation was found to fit well for many chemisorption processes using heterogeneous materials (17). Therefore, all of the kinetic data from our experiments were fit with a pseudo secondorder kinetic model in estimating the rate constants, initial sorption rates, and adsorption capacities for total arsenic. Ho et al. described the pseudo second-order kinetic rate equation (18, 19). The pseudo-second-order kinetic model can be solved with the following equations. The kinetic rate equation is expressed as follows (17-20):

dqt ) k2(qeq - q)2 dt

(1)

where qeq is the sorption capacity at equilibrium, and q is the solid-phase loading of arsenic. The k2 (g mmol-1 min-1) is the pseudo-second-order rate constant for the kinetic model. By integrating eq 1 with the boundary conditions of q ) 0 (at t ) 0) and q ) qt (at t ) t), the following linear equation can be obtained:

t 1 1 ) + t q v0 qeq

(2)

v0 ) k2 ‚ qeq2

(3)

In this equation, v0 (mmol g-1 min-1) is the initial sorption rate. Therefore, the v0 and qeq values of kinetic tests can be determined experimentally by plotting t versus t/qt. Column Studies. The 15-mm diameter and 150-mm height glass columns (Bio-Rad Laboratories, Hercules, CA) were used for the small-scale column experiments. The setup of media and operational parameters is shown in Table 1. For column B, 4 g of GAC was preloaded to 10 g of Fe (25%)diatomite. The total mass of Fe0/sand (50/50, w/w, 90 g) in column A was 6.4 times higher than that of GAC preloaded Fe (25%)-diatomite. The groundwater sample was purged with N2 gas for 24 h to remove VOC/SVOC compounds. After about 200 mL of distilled water were passed through the columns, the groundwater sample was pumped into the bottom of the columns at a flowrate of approximately 6 mL hr-1. Effluent samples were collected and diluted with 1% HNO3 solution prior to total arsenic analysis as described in the previous section. The large-scale column tests were conducted for about 3 months at a flowrate of 0.088-0.094 m day-1 which 3324

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 9, 2007

approximated groundwater flow conditions at the subject site. The VOC/SVOC removal was characterized during largescale columin testing, which also allowed for analysis of the effect of VOC/SVOC compounds on arsenic adsorption. Table 1 summarizes the experimental setup of the large-scale column tests. Predetermined amounts of media were packed into 50-mm diameter by 500-mm high glass columns (Ace Glass). In column B, GAC (40 g) was preloaded to Fe (25%)diatomite (200 g). The total mass of GAC preloaded Fe (25%)diatomite (240 g) was 7.5 times less than the Fe0/sand mixture (1800 g, column A) at the similar bed volume. Bags, connections, and transfer tubing made of Teflon were used to prevent aeration and/or VOC/SVOC loss during column operation. Effluent samples were collected at 2-3 day intervals, and diluted with 1% HNO3 solution prior to total arsenic analysis and VOC/SVOC analysis using the analytical methods described in the previous section.

Results and Discussion Sample Characteristics and Arsenic Speciations. Table S1 shows the chemical analysis of the groundwater sample using various analysis methods (21-23). The sample had very high arsenic concentrations of 341 mg L-1 (4.55 mmol L-1) that were 6820 times higher than 50 µg L-1. The extremely high arsenic concentration represents arsenic contamination caused by human activities such as arsenic-based herbicide or pesticide production at the subject site. Meanwhile, naturally occurring arsenic has been found in the groundwater of most aquifers of Wisconsin. Arsenic contamination is especially prevalent in the sedimentary bedrock of northeastern Wisconsin (24). The primary cause for this arsenic contamination is the oxidation of sulfide mineralized zones containing arsenic. About 3.5% of private drinking water wells in Outagamie and Winnebago counties in Wisconsin were reported to have an arsenic concentration higher than 50 µg L-1. Along with arsenic, the groundwater contains high concentrations of other anions such as sulfate (1600 mg L-1 or 16.7 mmol L-1) and chloride (300 mg L-1 or 8.5 mmol L-1), as well as cations such as sodium (440 mg L-1 or 19.1 mmol L-1), calcium (690 mg L-1 or 17.3 mmol L-1), and magnesium (49 mg L-1 or 2.0 mmol L-1). Since typical groundwater contains much lower concentrations, it was thought that various salts (CaSO4, NaCl, MgCl2, etc), in addition to arsenic, leaked into the groundwater from the hazardous waste sites. Compared to typical groundwater (150 mg C L-1), the groundwater contains a high TOC concentration (72 mg C L-1), representing potentially high NOMs or organic matters derived from the hazardous waste site. Table S2 in the Supporting Information shows the retention time, formulas, dissociation constants, concentrations, and percentages of known arsenic species (arsenite, arsenate, MMAA, and DMAA) and several unknown arsenic species detected by IC-ICP-MS analysis for the groundwater sample. Dominant known arsenic species detected were

arsenite (143 880 µg L-1, 45.1%) and MMAA (72 350 µg L-1, 22.7%), but DMAA and arsenate were only 7600 (2.4%) and 4010 µg L-1 (1.3%), respectively. Since arsenite is highly toxic and less removable for most metal oxides in subsurface environments, there is an urgent need to remediate groundwater in this area with an effective method. Methylated arsenic species have been known to have less toxicity than inorganic arsenic; however, they have not only higher mobilities in subsurface environments, but also possibilities for the demethylation process. The total arsenic concentration of all species determined by IC-ICP-MS was about 94% of the total arsenic determined by ICP-AES, showing a reliable measurement of arsenic analysis. Along with methyl- and demethylation of arsenic species, the unknown species might be formed through chemical or microbial transformations of known arsenic species. In the previous study (25), the major portion of arsenic found within the groundwater at the site was in the organic form even though large variations of speciation existed depending on specific sites. Accordingly, in order to set an efficient strategy of remediation, it is critical to speciate arsenic for each contamination site because the predominant organic forms of arsenic (MMAA or DMAA) and inorganic arsenite could not be easily adsorbed in minerals and are transported in the dissolved form due to their high acid dissociation constants (26). Batch Kinetic Study (Adsorption Characteristics). Kinetic tests were conducted to determine several adsorption parameters such as initial sorption rates, pseudo secondorder rate constants, and equilibrium sorption capacities. Table S3 in the Supporting Information shows the kinetic data obtained from the pseudo second-order kinetic model for the sorption of total arsenic on Fe (25%)-diatomite, Fe0, and GAC, respectively. All data in Figure S2 show reliable fitting conditions since determination coefficients were above 0.99 (Table S3). Fe (25%)-diatomite (20.5 mg g-1) exhibited about 3 or 6 times higher arsenic sorption capacities than Fe0 (6.6 mg g-1) or GAC (3.8 mg g-1), respectively. The initial sorption rates (0.24 mg g-1 min-1) of total arsenic removal for Fe (25%)-diatomite were also 1.2-1.8 times greater than those of Fe0 (0.19 mg g-1 min-1) and GAC (0.11 mg g-1 min-1). The sorption capability of GAC might be due to organic arsenic contents that have hydrophobic characteristics. Fe0 showed 0.0049 molAs molFe-1 of sorption density. Other literature showed the range of 0.00035-0.0056 molAs molFe-1 of arsenic sorption capacities for Fe0 in batch and column tests, in which higher sorption capacities were shown for higher arsenic concentrations and arsenite species (6, 7, 9, 27-29). Surface precipitation and/or adsorption by ferrous and/or ferric oxyhydroxides continuously produced by corrosion of Fe0 have been known to be the main mechanisms of inorganic arsenic removal for Fe0 (7, 29). Fe (25%)-diatomite had 0.076 molAs molFe-1 of sorption density at 0.15 of Asto-Fe molar ratio. In other studies using HFO, Raven et al. (30) showed sorption maxima of about 0.6 (or 0.58) and 0.25 (or 0.16) molAs molFe-1 were achieved for arsenite and arsenate at pH 4.6 (or pH 9.2), representing a higher sorption capacity for arsenite than arsenate at higher arsenic concentrations. Dixit and Hering (11) showed 0.31 and 0.24 molAs molFe-1 for arsenite and arsenate at 0.03-0.3 of As-to-Fe molar ratio, respectively. Lafferty and Loeppert (31) also reported a higher sorption maxima of arsenite for HFO than other arsenic species. The calculated sorption maxima of arsenite, arsenate, MMAA, and DMAA on ferrihydrite were 0.2, 0.105, 0.094, and 0.075 molAs molFe-1 with 0-0.22 of As-to-Fe molar ratio at pH 7.0. Thus, based on the above observations, the sorption densities of arsenite for HFO were close to the sorption maxima at 0.97 of determination coefficients (R 2)(Table S4). Fe (25%)-diatomite showed similar initial sorption rates (0.015-0.02 mg g-1 min-1) and qeq (9.310 mg g-1) for arsenate, arsenite, and MMAA, while much less for DMAA that had 0.005 mg g-1 min-1 (v0) and 2.7 mg g-1 (qeq). Fe (25%)-diatomite had about 100% of arsenic sorption densities close to sorption maxima (0.037 molAs molFe-1) for arsenate, arsenite, and MMAA, while only 25% (0.01 molAs molFe-1) for DMAA. These sorption behaviors are quite similar to those reported by Lafferty and Loeppert (31), in which they studied the sorption behaviors of inorganic and methyl-arsenic compounds with ferrihydrite and goethite. Although spectroscopic analysis is needed to find out the surface complexes’ geometry, the plausible reason for the similar adsorption trends for arsenite, arsenate, and MMAA is the fact that they have two hydroxyl groups which could complex with the surface of HFO. The lower capacity for DMAA might not only be due to the additional methyl group of molecular geometry instead of oxygens, but also VOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3325

FIGURE 2. Kinetic results using the groundwater sample, batch test using the single medium of GAC (11.3 g L-1) or Fe (25%)-diatomite (8.7 g L-1), and the mixed media of GAC (11.3 g L-1) and Fe (25%)diatomite (8.7 g L-1) because of the electron donating characteristics of the methyl groups that could weaken the Fe-O-As bond (31). In the presence of humic acid, concentrations of sorbed arsenite and MMAA decreased, even though the decrease trends were different. The adsorption trend of arsenite in the presence of humic acid was initially similar to that of arsenite without humic acid for about 3 h, but gradually increased and equilibrated at about 10 h with about 50% of sorbed arsenic concentration without humic acid. In the case of MMAA, humic acid significantly inhibited the MMAA adsorption with Fe (25%)-diatomite. Only about 10% of the adsorption capacity of MMAA was obtained with humic acid. Bowell (32) also showed that the methyl-arsenic adsorption capacities of several ferric oxides (goethite, hematite, or lepidocrocite) in the presence of fulvic acid were much less than those of inorganic arsenic at a neutral pH, representing that the competition effect of organic acids is greater toward MMAA than arsenite. This higher competition for MMAA might be due to the fact that the hydrophobic characteristics of MMAA could create a similar sorption mechanism with humic acid for the sorption sites of HFO, and negatively charged MMAA species could be more easily complexed by metals and deprotonated functional groups within humic acid. More scientific investigations are needed for the competition effect between organic acids and inorganic/ organic arsenic species. Batch Kinetic Study [GAC Addition on Arsenic Removal Performance of Fe (25%)-Diatomite]. Additional kinetic studies were conducted to determine the synergetic effect of the addition of GAC to Fe (25%)-diatomite in groundwater. Figure 2 shows the kinetic results of GAC and Fe (25%)diatomite individually and mixed together for groundwater. The mixed system exhibited an initial sorption rate of approximately 0.577 mg g-1 min-1, which is about 2.5 times higher than that of the individual system consisting of Fe (25%)-diatomite. Thus, these increases of initial sorption rates might be the result of the addition of GAC, deducing that the addition of GAC could enhance the arsenic removal performance of Fe (25%)-diatomite by removing comparably high portions of NOMs that can compete for or block the sorption sites of Fe (25%)-diatomite (32). Column Studies. Figure 3 A shows the results of the smallscale column test, which compared column A [Fe0/sand mixture (50/50, w/w)] and column B (GAC preloaded Fe (25%)-diatomite). Column B demonstrated better arsenic removal capacity than column A. The column containing Fe (25%)-diatomite could reduce the arsenic concentration in the effluent to less than 1 mg L-1 from the influent arsenic concentration of 341 mg L-1 for 7.5 BV, while column A did 3326

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 9, 2007

FIGURE 3. (A) Small-scale column tests, column A: Fe0/sand mixture (50/50, w/w) (90 g) and column B: GAC (4 g) preloaded Fe (25%)diatomite (10 g), and (B) large-scale column tests, column A: Fe0/ sand (50/50, w/w) (1800 g) and column B: GAC (40 g) preloaded Fe (25%)-diatomite (200 g). not reduce the arsenic concentration to less than 1 mg L-1 after 0.8 BV. Arsenic concentrations in the effluent from column A increased sharply to 170 mg L-1 until approximately 8 BVs had passed through the column, then continued to increase steadily after that point, but not as quickly. The breakthroughs (C ) 0.5 C0) of column A and B were 11.3 and 18 BV, at which arsenic adsorption capacities (or sorption densities) for column A and B were 2.6 mg (g of Fe0)-1 (or 0.002 molAs molFe-1) and 26.8 mg [g of Fe (25%)-diatomite]-1 (or 0.099 molAs molFe-1), respectively. Figure 3B shows arsenic removal observed for about 3 months during the large-scale column tests, which were conducted at conditions which approximated anticipated PRB conditions in the field (e.g., flowrate, anaerobic environment, presence of VOC/SVOC compounds). Column B [GAC preloaded Fe (25%)-diatomite] showed greater arsenic removal capabilities than column A (Fe0/sand mixture) did for groundwater, with a reduction in effluent arsenic concentrations to less than 1 mg L-1 for 44 days (8.9 L or 9.5 BV), obtained by ICP-AES analysis. The large-scale Fe0/sand mixture did not reduce the arsenic concentration in the column effluent below 1 mg L-1, even though the EBCT (5.1 d) of large-scale column tests was 16 times longer than that of small-scale column tests (0.3 d). Based on the breakthrough of 1 mg L-1, this result shows 42 L kg-1 of the normalized volume of treated groundwater to media mass, which is about 8% higher than the results from the small-scale column tests (39 L kg-1 for column B, Figure 3A). Effluent samples of large-

scale column B were again analyzed with AAS-graphite which had higher sensitivity than ICP-AES. AAS analysis shows arsenic was treated to less than 50 µg L-1 for 8.4 L (9 BV), even though arsenic concentrations were initially higher than 50 µg L-1 due to an unstable initial operational condition. Large-scale column B showed 20.8 mg [g of Fe (25%)diatomite]-1 of sorption capacity (or 0.078 molAs molFe-1), while column A had 4.4 mg (g of Fe0)-1 (or 0.0033 molAs molFe-1) at breakthrough (C ) 0.5 C0). Thus, compared to small-scale column tests, large-scale column A showed about a 1.7 times higher sorption capacity or density, while largescale column B had about 78% of arsenic sorption capacity. Different results of the normalized volume of treated groundwater to media mass and sorption capacities at breakthrough might be mainly a result of different flowrates and media compositions. However, based on those observations, results between the small- and large-scale column tests for the GAC preloaded Fe (25%)-diatomite indicated that arsenic removal performance is independent of flowrate in this range during the test period, compared to the Fe0/sand mixture that showed flowrate-dependent characteristics on arsenic removals. The increase of arsenic removal performance in the large-scale column test could be explained by the fact that the corrosion and hydrolysis process related to Fe0 is the rate-limiting factor of arsenic removal since the arsenic removal step by HFO could be very fast, as shown in the column containing Fe (25%)-diatomite (9). Except for some VOC/SVOC compounds that were at high concentrations in influents, the GAC preloaded Fe (25%)diatomite removed most species below the detection levels at both 7 and 15.6 BV (Table S5). However, the Fe0/sand mixture provided limited attenuation of halogenated VOCs, BETX (benzene, ethylbenzene, toluene, and xylenes), and naphthalene, but did not affect the concentrations of acetone, phenol, or 4-methyl 2-pentanone (MIBK). In the case of GAC preloaded Fe (25%)-diatomite, trichloroethene, cis-1,2dichloroethene, and vinyl chloride were observed to be attenuated during the large-scale column test. Chlorobenzene, 1,2-dichlorobenzene, 1,4-dichlorobenzene, and methylene chloride were attenuated significantly (Table S5). Degradation of halogenated VOCs, particularly chlorinated ethenes, by corrosion reactions catalyzed by Fe0 has been well-documented (33-36). However, degradation with Fe0 has been either not demonstrated or not evaluated in previous studies (8) for these halogenated VOCs. In addition, limited attenuation of BETX and naphthalene was observed in the Fe0/sand column. No relative decrease in concentration was observed for phenol, acetone, or MIBK either. In fact, the concentration of acetone in the effluent from the column of the Fe0/sand mixture was actually higher than the influent concentration during column operation. Implications for PRB Application. Large-scale column studies showed that GAC preloaded Fe (25%)-diatomite is promising in the PRB application because high arsenic concentrations were effectively reduced to below 50 µg L-1, and most species of hazardous VOC/SVOCs were also simultaneously attenuated below the detection levels with the actual groundwater condition. The independency of EBCT (or flowrate) in terms of sorption capacities at breakthrough (C ) 0.5 C0) or normalized volume of treated groundwater to media mass at 1 mg L-1 breakthrough in treating arsenic must be a beneficial characteristic because this system can be effectively applied to groundwater having a wide range of flowrate. Along with high sorption speeds, the GAC preloaded Fe (25%)-diatomite could be easier to handle and install for PRB applications since their bulk density is 6-8 times lighter than the Fe0/sand mixture. During the 3-month operation of large-scale column tests, Fe (25%)-diatomite did not break off, indicating good hydraulic stabilities. With spectroscopic analysis, more abiotic studies are underway

to find the competition effects of NOMs and removal mechanisms of different arsenic species with different ratios of GAC and Fe (25%)-diatomite.

Acknowledgments The present study was supported by the owner/operator of the subject site, who wished to remain anonymous in this publication. The authors also acknowledge the work of Dr. Dirk Wallschlaeger at Trent University in Peterborough, Ontario, Canada for the identification and quantification of organic and inorganic species of arsenic in the groundwater samples used in this study.

Supporting Information Available Methodology of Fe (25%)-diatomite preparation, schematic of pH-controlled DCBR (Figure S1), characteristics of the groundwater sample (Table S1), arsenic speciation results (Table S2), conditions and results of batch kinetics tests (Table S3 and Figure S1), results of the pH-controlled DCBR tests (Table S4), and the results of VOC/SVOC analysis for influents and effluents of each column in large-scale column tests (Table S5). This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) 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 (6), 2004-2010. (2) Wilkin, R. T.; Su, C.; Ford, R. G.; Paul, C. J. Chromium-removal processes during groundwater remediation by a zerovalent iron permeable reactive barrier. Environ. Sci. Technol. 2005, 39, 4599-4605. (3) Morrison, S. J.; Mushovic, P. S.; Niesen, P. L. Early breakthrough of molybdenum and uranium in a permeable reactive barrier. Environ. Sci. Technol. 2006, 40 (6), 2018-2024. (4) Weisener, C. G.; Sale, K. S.; Smyth, D. J. A.; Blowes, D. W. Field column study using zerovalent iron for mercury removal from contaminated groundwater. Environ. Sci. Technol. 2005, 39 (16), 6306-6312. (5) Szecsody, J. E.; Fruchter, J. S.; Williams, M. D.; Vermeul, V. R.; Sklarew, D. In situ chemical reduction of aquifer sediments: Enhancement of reactive iron phases and TCE dechlorination. Environ. Sci. Technol. 2004, 38 (17), 4656-4663. (6) Nikolaidis, N. P.; Dobbs, G. M.; Lackovic, J. A. Arsenic Removal by Zero-valent Iron: Field Laboratory and Modeling Studies. Water Res. 2003, 37, 1417-1425. (7) Su, C.; Puls, R. W. Arsenate and Arsenite Removal by Zerovalent Iron: Kinetics, Redox Transformation, and Implications for In Situ Groundwater Remediation. Environ. Sci. Technol. 2001, 35, 1487-1492. (8) USEPA. Permeable Reactive Barrier Technologies for Contaminant Remediation; EPA/600/R-98/125; Office of Research and Development, Office of Solid Waste and Emergency Response, USEPA: Washington, DC, 1998. (9) Su, C.; Puls, R. W. In Situ Remediation of Arsenic in Simulated Groundwater Using Zerovalent Iron: Laboratory Column Tests on Combined Effects of Phosphate and Silicate. Environ. Sci. Technol. 2003, 37, 2582-2587. (10) Jang, M.; Min, S. H.; Kim, T. H.; Park, J. K. Removal of Arsenite and Arsenate Using Metal Oxide Incorporated NaturallyOccurring Porous Diatomite. Environ. Sci. Technol. 2006, 40 (5), 1636-1643. (11) Dixit, S.; Hering, J. G. Comparison of arsenic (V) and arsenic (III) sorption onto iron oxide minerals: implications for arsenic mobility. Environ. Sci. Technol. 2003, 37, 4182-4189. (12) Redman, A. D.; Macalady, D. L.; Ahmann, D. Natural organic matter affects arsenic speciation and sorption onto hematite. Environ. Sci. Technol. 2002, 36 (13), 2889-2896. (13) Ghosh, A.; Saez, A. E.; Ela, W. Effect of pH, competitive anions and NOM on the leaching of arsenic from solid residuals. Sci. Total Environ. 2006, 363, (1-3), 46-59. (14) USEPA. Low-flow (minimal drawdown) ground-water sampling procedures; Office of Research and Development, Office of Solid Waste and Emergency Response, USEPA: Washington, DC, 1996. VOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3327

(15) Saada, A.; Breeze, D.; Crouzet, C.; Cornu, S.; Baranger, P. Adsorption of arsenic (V) on kaolinite and on kaolinite-humic acid complexes: role of humic acid nitrogen groups. Chemosphere 2003, 51, 757-763. (16) Ko, I.; Kim, J. Y.; Kim, K. W. Arsenic speciation and sorption kinetics in As-hematite-humic acid system. Colloids Surf., A 2004, 234, 43-50. (17) Reddad, Z.; Gerente, C.; Andres, Y.; Le Cloirec, P. Adsorption of several metal ions onto a low-cost biosorbent: kinetic and equilibrium studies. Environ. Sci. Technol. 2002, 36 (9), 20672073. (18) Ho, Y. S.; McKay, G. A comparison of chemisorption kinetic models applied to pollutant removal on various sorbents. Trans. IChemE 1998, 76B, 332-340. (19) Ho, Y. S.; McKay, G. Sorption of dye from aqueous solution by peat. Chem. Eng. J. 1998, 70, 115-124. (20) Ho, Y. S.; McKay, G. Kinetic models for the sorption of dye from aqueous solution by wood. Trans. IChemE 1998, 76B, 183-191. (21) USEPA. EPA methods for chemical analysis of water and wastes, method 310.1; USEPA: Washington, DC, 1983; p 16. (22) USEPA. Determination of inorganic anions by ion chromatography: SW846, method 9056; Office of Solid Waste and Emergency Response, USEPA: Washington, DC, 1994; p 16. (23) USEPA. Inductively coupled plasma-atomic emission spectrometry: SW846, method 6010B; Office of Solid Waste and Emergency Response, USEPA: Washington, DC, 1996; p 25. (24) Burkel, R. S.; Stoll, R. C. Naturally occurring arsenic in sandstone aquifer water supply wells of Northeastern Wisconsin. Ground Water Monit. Rem. 1999, 19 (2), 114-121. (25) URS. Summary of Findings: 1974-2000 TYCO Suppression Systems-ANSUL (EPA ID#WID 006 125 215); URS Corporation: Milwaukee, WI, 2001; p 38. (26) Vogels, C. M.; Johnson, M. D. Arsenic Remediation in Drinking Waters Using Ferrate and Ferrous Ions; Department of Chemistry and Biochemistry, New Mexico State University: Las Cruces, NM, 1998; p 21.

3328

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 9, 2007

(27) Lien, H.-L.; Wilkin, R. T. High-level arsenite removal from groundwater by zero-valent iron. Chemosphere 2005, 59, 377386. (28) Lackovic, J. A.; Nikolaidis, N. P.; Dobbs, G. M. Inorganic arsenic removal by zero-valent iron. Environ. Eng. Sci. 2000, 17, 29-39. (29) Cheng, Z.; Van geen, A.; Louis, R.; Nikolaidis, N.; Bailey, R. Removal of methylated arsenic in groundwater with iron filings. Environ. Sci. Technol. 2005, 39, 7662-7666. (30) Raven, K. P.; Jain, A.; Loeppert, R. H. Arsenite and Arsenate Adsorption on Ferrihydrite: Kinetics, Equilibrium, and Adsorption Envelopes. Environ. Sci. Technol. 1998, 32 (3), 344-349. (31) Lafferty, B. J.; Loeppert, R. H. Methyl arsenic adsorption and desorption behavior on iron oxides. Environ. Sci. Technol. 2005, 39, 2120-2127. (32) Bowell, R. J. Sorption of arsenic by iron oxides and oxyhydroxides in soils. Appl. Geochem. 1994, 9, 279-286. (33) Gillham, R. W.; O’Hannesin, S. F. Enhanced degradationof halogenated aliphatics by zero-valent iron. Ground Water 1994, 32 (6), 958-967. (34) Johnson, T. L.; Scherer, M. M.; Tratnyek, P. G. Kinetics of halogenated organic compound degradation by iron metal. Environ. Sci. Technol. 1996, 30 (8), 2634-2640. (35) Matheson, L.; Tratnyek, P. Reductive dehalogenation of chlorinated methanes by iron metal. Environ. Sci. Technol. 1994, 28, 2045-2053. (36) Roberts, A. L.; Totten, L. A.; Arnold, W. A.; Burris, D. R.; Campbell, T. J. Reductive Elimination of Chlorinated Ethylenes by ZeroValent Metals. Environ. Sci. Technol. 1996, 30 (8), 2654-2659. (37) Burrows, G. J.; Turner, E. E. A new type of compound containing arsenic. J. Chem. Soc. 1920, 117, 1373-1380. (38) Burrows, G. J.; Turner, E. E. Experiments on the production of compounds containing arsenic as a center of optical activity. J. Chem. Soc. 1921, 119, 426-437.

Received for review October 3, 2006. Revised manuscript received February 26, 2007. Accepted February 28, 2007. ES062359E