Environ. Sci. Technol. 2003, 37, 2394-2399
Influence of Surface Composition on the Kinetics of Alachlor Reduction by Iron Pyrite DANIEL L. CARLSON,† M O L L Y M . M C G U I R E , †,‡ A . L Y N N R O B E R T S , * ,† A N D D. HOWARD FAIRBROTHER‡ Department of Geography and Environmental Engineering and Department of Chemistry, Johns Hopkins University, 3400 North Charles Street, Baltimore, Maryland 21218-2686
Even though naturally occurring iron sulfide minerals have previously been shown capable of promoting reductive dehalogenations, the role of surface composition has not been fully investigated. Some researchers have proposed that sulfur species represent redox-active moieties on iron pyrite surfaces. Results from this study indicate that neither the stoichiometric (100) pyrite surface, nor monosulfide defects, play direct roles in the observed reduction of the herbicide alachlor [2-chloro-2′,6′-diethyl-N-(methoxymethyl)acetanilide]. Pyrite surfaces were initially characterized by X-ray photoelectron spectroscopy (XPS); the samples were then transferred to a liquid cell coupled to the ultrahigh vacuum chamber. Aliquots were periodically removed from the liquid cell to monitor the appearance of the reductive dechlorination product, 2′,6′-diethyl-N(methoxymethyl)acetanilide. In experiments with unaltered pyrite (100) surfaces, rates of reaction decreased over time, even though no change in surface composition could be discerned via XPS. In contrast, ion-bombarded surfaces, which are dominated by monosulfide species, exhibit an initial induction period of low reactivity during which the highly defective surface is oxidized by water. Only after the monosulfide defects undergo oxidation does the rate of alachlor reduction increase, precluding a direct role for these defects as alachlor reductants.
Introduction Chloroacetamides, including alachlor [2-chloro-2′,6′-diethylN-(methoxymethyl)acetanilide], represent an important class of herbicides that are widely encountered in surface and groundwater (1). Their ubiquity and toxicity have led to the establishment of drinking water standards for alachlor as well as the inclusion of chloroacetamide degradation products on EPA’s Contaminant Candidate List of substances that may be regulated in the future (2). As many as 20 different degradation products of alachlor have been observed in groundwater; six of these, including the reductive dechlorination product deschloroalachlor [2′,6′-diethyl-N-(methoxymethyl)acetanilide], have been positively identified by comparison with authentic reference materials (3). Many chloroacetamide degradates are recognized to result from photolysis (4) or biodegradation (5). Apart from studies * Corresponding author phone: (410)516-4327; fax: (410)516-8996; e-mail:
[email protected]. † Department of Geography and Environmental Engineering. ‡ Department of Chemistry. 2394
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 11, 2003
of the ethane sulfonic acid and oxanilic acid derivatives (68), however, our current understanding of the environmental degradation of alachlor is somewhat fragmentary. Thermochemical reactions of alachlor with inorganic species in particular have received relatively little attention. Recent studies have shown that alachlor is labile to SN2 displacement by reduced sulfur nucleophiles including thiosulfate (9), bisulfide, and polysulfides (10, 11). Alachlor has also been found to react with granular iron particles (12) and with microbially reduced smectite (13) via reductive dechlorination. The reactive entity in the latter case has been hypothesized to comprise structural iron components of the clay. Iron sulfide minerals, which may possess both reactive sulfur and reactive iron sites, are common minor constituents of sediments and aquifers. The prevalence of these minerals as well as their reactivity has led several groups of researchers to investigate their ability to reduce C1 and C2 organohalides (14-20). Prior work by Kriegman-King and Reinhard focused on reactions of CCl4 with iron pyrite (14). X-ray photoelectron spectroscopic (XPS) analyses of pyrite samples exposed to aqueous solutions of CCl4 showed the ratio of sulfur to iron(II) in the near-surface region was in excess of the stoichiometric 2:1 (14). The depletion of iron at the surface as well as the observation of CS2 as a product led these researchers to propose the reactive moiety to be a surface sulfur species. Butler and Hayes also examined the surface species responsible for reducing C1 and C2 organohalides, using mackinawite (FeS) as a model iron sulfide (17-20). They hypothesized the rate-determining step to be an initial oneelectron transfer. This step does not involve a proton, and, therefore, the rate should be independent of pH; nonetheless, an increase in reaction rate with pH was observed in the range of 7.1-9.5 (17). This was interpreted to signify that two different pH-dependent reactive surface species were involved. The speciation of surface sulfide was postulated to be dominated by tFeSH throughout the pH range investigated, while iron undergoes deprotonation from tFeOH2+ to tFeOH with increasing pH (17). Therefore, these researchers hypothesized surface Fe to be directly involved in the rate determining step. The substantially greater reactivity they computed for tFeOH relative to tFeOH2+ was attributed to differences in electron density (17). Although the observed dependence of reaction rates on pH did not preclude a role for other ionizable surface acid/base pairs (such as tFeSH2+ and tFeSH or tFeSH and tFeS-), the inhibitory or rateenhancing influences of various added solutes (2,2′-bipyridine, 1,10-phenanthrolene, cysteine, methionine) provided additional evidence that surface iron species, rather than surface sulfur species, served as reactive moieties (17). Other researchers (16) have also suggested that iron species are responsible for reductive dechlorination of trichloroethene by another iron sulfide mineral, troilite (FeS). XPS has been widely used to investigate mechanisms through which iron sulfides react, including reductive dechlorination reactions (14, 16), because of its ability to differentiate among surface species with different oxidation states. The use of XPS to identify surface reactive sites on pyrite has recently been explored by Guevremont et al. (21) in investigations of pyrite oxidation by H2O and O2. Their techniques relied on XPS surface characterization of the pyrite, in combination with He+ ion bombardment of the pyrite surface to introduce monosulfide sites. Studies employed a reaction vessel attached to the ultrahigh vacuum (UHV) surface analysis chamber to avoid exposure of the 10.1021/es0262028 CCC: $25.00
2003 American Chemical Society Published on Web 04/22/2003
SCHEME 1
sample to the ambient atmosphere during the course of reaction and XPS analysis. These researchers found that the stoichiometric pyrite surface only reacted in the presence of both H2O and O2. The monosulfide sites introduced by ion bombardment, however, could be oxidized by H2O alone. Studies using XPS, temperature programmed desorption, and photoemission of adsorbed xenon have reported evidence that H2O dissociates on monosulfide sites (21, 22). Together, these results suggest monosulfide sites may be more reactive than the stoichiometric pyrite surface. In this research, we explore the effect of surface composition on the reductive dechlorination (Scheme 1) of alachlor by iron pyrite, FeS2. Experiments were carried out using modifications of some techniques successfully employed in previous work by others (21). Pyrite samples were cut from as-grown crystals, and some samples were subsequently altered by Ar+ bombardment. Samples were exposed to a deoxygenated aqueous solution of alachlor in a reaction vessel attached to the UHV chamber, and the reaction was followed by measuring the appearance of the product, deschloroalachlor, using gas chromatographic/mass spectrometric (GC/MS) analysis. The current study explores the effects of the differences between native pyrite and modified pyrite possessing monosulfide sites on the reductive dechlorination of alachlor.
Experimental Section Chemicals. Alachlor [2-chloro-2′,6′-diethyl-N-(methoxymethyl)acetanilide; 99%], metolachlor [2-chloro-2′-ethyl-6′methyl-N-(1-methyl-2-methoxyethyl)acetanilide; 99%], and propachlor [2-chloro-N-isopropylacetanilide; 99%] were obtained from Chem Service (West Chester, PA). Pyrite crystals, originating from the Navaju ´ n deposit near Logron ˜ o, Spain, were obtained from Ward’s Scientific (Rochester, NY). Deschloroalachlor was synthesized by dissolving approximately 100 mg of alachlor in 100 mL of a deoxygenated 50:50 water:acetone mix. Approximately 2 g of acid-washed granular iron (Fisher Scientific electrolytic iron powder, 100 mesh) was added to the reaction mixture. The suspension was rotated in a serum bottle on a jar mill in the dark until the reaction of alachlor was complete (approximately two weeks), as verified by periodic GC/MS analysis of the reaction mixture. The solution was then acidified and extracted with n-hexane; the product was recovered by evaporating the hexane. The EI mass spectrum conformed to previously published reports (3, 12). Purity was greater than 99% by GC/MS. The 1H NMR spectrum was determined on a UNITYplus 400 MHz NMR in CDCl3. The spectrum lacked the singlet for alachlor assigned in the literature (23) to C(O)CH2Cl at 3.71 ppm. Instead, a singlet at 1.73 ppm was observed, produced by the three equivalent hydrogen atoms in deschloroalachlor. All other chemical shifts were equivalent to those reported for alachlor. Sample Preparation, Modification, and Analysis. Pyrite (100) faces cut from as-grown crystals were cleaned by ultrasonic treatment in methanol, followed by acid-washing in deoxygenated 1 N HCl overnight. Samples were then transferred to an ultrahigh vacuum (UHV) chamber, in which selected samples were sputtered using argon ion bombardment for various times to introduce controlled amounts of
FIGURE 1. Schematic of experimental system. surface monosulfide defects. Both ion-bombarded and natural (cleaned but not sputtered) samples were characterized via X-ray photoelectron spectroscopy (XPS). The samples were then transferred to a liquid cell (under positive nitrogen pressure) coupled to the UHV chamber via a chamber of intermediate pressure (Figure 1), so as to avoid exposure to the ambient atmosphere. The UHV chamber had a base pressure of 4 × 10-10 Torr. Spectra were obtained with a Physical Electronics (Φ) 10-360 Precision Energy Analyzer using the Mg KR line (1253.6 eV) of a Φ 04-500 Dual Anode X-ray source operating at 15 kV and 300 W. All spectra shown were obtained with a pass energy of 44.75 eV and a resolution of 0.125 eV/step. Binding energies were corrected using the S 2p3/2 peak from the bulk pyrite disulfide species at 162.3 eV (24). A Φ 04-300 ion gun (2 keV, 25 mA) was employed for the pyrite surface modifications using an Ar pressure within the gun of 15 mPa. The gun was oriented approximately 45° to the sample surface and focused to a spot slightly larger than the area of the sample surface (∼1 × 1 cm2). The sulfur spectra were fit to Voight peak shapes (consisting of 2p3/2 - 2p1/2 doublet pairs with a spacing of 1.2 eV) using a commercial software package (Wavemetrics). Binding energies were compared to a recently published database of species found on pyrite surfaces before and after oxidation (24). Experimental Setup. The liquid cell (Figure 1) was kept in the dark under a positive pressure of high-purity nitrogen (99.995%) saturated with water. The solution in the cell consisted of 20-30 mL of distilled, deionized water (MilliQ Plus UV system, Millipore Corp.), deoxygenated by sparging with N2 and spiked with a saturated aqueous solution of alachlor. The resulting solution had an alachlor concentration of ∼25 µM and a background deschloroalachlor concentration of ∼0.5 nM. No pH buffer was employed in these experiments. Changes in pH over the course of the reactions were negligible, as the pH remained near 7.0 in all experiments. The solution was mixed with a magnetic stirbar. Temperature was not controlled and varied between 25 and 27 °C. Analysis of Aqueous Solution. Reaction kinetics were followed by periodically removing 2-mL aliquots of the reaction solution and determining the concentration of the reductive dechlorination product, deschloroalachlor. The volume withdrawn was replaced with fresh alachlor solution kept under a nitrogen atmosphere; the resulting dilution was accounted for in the subsequent kinetic analysis of the deschloroalachlor data. Aqueous samples were extracted with 0.5 mL of n-hexane (ultra resianalyzed grade; J. T. Baker) containing propachlor and metolachlor as internal standards. The hexane extracts were analyzed using a Thermo Finnigan Trace quadrupole GC/MS equipped with an on-column VOL. 37, NO. 11, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2395
FIGURE 2. XPS spectra of the Fe 2p and S 2p region of natural and Ar+ ion beam modified pyrite with corresponding fits to the data. The natural surface is dominated by disulfide (S22-, shown as dotted lines) and Fe(II). Ar+ ion sputtering produces monosulfide species (FeS, shown as solid lines).
FIGURE 3. XPS spectra of the Fe 2p region and S 2p showing the evolution of the surface composition of a modified pyrite sample (sputtered with an Ar+ beam) following exposure to water. Monosulfide sites are consumed within the first few hours, and Fe oxides are produced.
injector and a 30 m J&W DB-5, 0.25 mm i.d. × 0.25 µm fusedsilica capillary column. Selected-ion monitoring (using the sum of ions m/z ) 202, 220, and 235 for quantification) was used in electron impact (EI) mode with an electron energy of 70 eV. The software package Scientist (Micromath) was used to model the appearance of deschloroalachlor over time.
Results and Discussion Surface Characterization. The S(2p) and Fe(2p) XPS spectra of a natural acid-washed pyrite (100) surface are shown in Figure 2. The spectral envelope as well as the major peaks in the S(2p) and Fe(2p) regions at 162 and 707 eV, respectively, are characteristic of stoichiometric FeS2 (24). In addition, a small amount of spectral intensity is observed at higher binding energies in the S(2p) region. This has been observed previously in XPS studies of pyrite and attributed to polysulfide (Sn2-) species and/or final-state effects in the photoemission process as well as inadequate background subtraction (25). Trace amounts of silicates were observed in some samples, in addition to a small amount of carbon and oxygen attributed to surface contaminants. Sputtering the pyrite sample with Ar+ ions resulted in a broadening of the Fe(2p) region to higher binding energies (Figure 2), previously ascribed to a combination of structural disordering at the surface and the production of Fe3+ species (21). Simultaneous changes in the S(2p) region included an increase in spectral intensity at lower binding energies consistent with the production of monosulfide (FeS) species (21, 26). The concentration of FeS species was found to increase as a function of sputtering time up to 55 min. No additional changes were observed in the XPS spectra for longer sputtering times, indicating that the concentration of surface FeS species had reached a limiting value. Surfaces sputtered for 55 min were dominated by monosulfide species (FeS). This is confirmed by elemental peak area ratios, which show that sputtering reduced the sulfur/ iron ratio, removing sulfur. Separate studies indicated that the monosulfide defects disappeared after exposure to anoxic water over a period of several hours (Figure 3) as previously reported (21), resulting in a predominantly iron oxide surface. This process further reduced the sulfur/iron ratio to less than unity and resulted in incorporation of oxygen into the surface. The initial monosulfide species are believed to undergo oxidation to the metastable intermediate S2O32- (21), which would be removed from the surface by dissolution. Another possible oxidation product of the initial monosulfides is elemental sulfur, which would not be detected via XPS because of the evaporative loss of elemental sulfur under vacuum (27). Reaction of Alachlor with Natural Pyrite Surfaces. The parent compound, alachlor, reacted at a rate that was too 2396
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 11, 2003
FIGURE 4. Appearance of deschloroalachlor during reaction with natural pyrite surfaces from two different crystals; error bars indicate 95% confidence intervals. Note the different time axes used for the different samples, as well as differences in both [Y]0 and kobs. Lines indicate a model fit (eq 4) assuming a limited number of reactive sites that are consumed in a 1:1 stoichiometry during reaction with alachlor. slow to quantify by monitoring its disappearance (
