Destruction of Methyl Bromide Sorbed to Activated Carbon by

Avenue, Parlier, California 93648, United States. Environ. Sci. Technol. , 2015, 49 (7), pp 4515–4521. DOI: 10.1021/es505709c. Publication Date ...
0 downloads 4 Views 366KB Size
Download PDF
Article pubs.acs.org/est

Destruction of Methyl Bromide Sorbed to Activated Carbon by Thiosulfate or Electrolysis Yu Yang,†,⊥ Yuanqing Li,‡,⊥ Spencer S. Walse,§ and William A. Mitch*,‡ †

Department of Civil and Environmental Engineering, University of Nevada, Reno, Nevada 89557, United States Department of Civil and Environmental Engineering, Stanford University, Jerry Yang and Akiko Yamazaki Energy and Environment Building, 473 Via Ortega, Stanford, California 94305, United States § San Joaquin Valley Agricultural Sciences Center, USDA-ARS, 9611 South Riverbend Avenue, Parlier, California 93648, United States ‡

S Supporting Information *

ABSTRACT: Methyl bromide (CH3Br) is widely used as a fumigant for postharvest and quarantine applications for agricultural products at port facilities due to the short treatment period required, but it is vented from fumigation chambers to the atmosphere after its use. Due to the potential contributions of CH3Br to stratospheric ozone depletion, technologies for the capture and degradation of the CH3Br are needed to enable its continued use. Although granular activated carbon (GAC) has been used for CH3Br capture and thiosulfate has been used for destruction of CH3Br in aqueous solution, this research explored techniques for direct destruction of CH3Br sorbed to GAC. Submerging the GAC in an aqueous thiosulfate solution achieved debromination of CH3Br while sorbed to the GAC, but it required molar concentrations of thiosulfate because of the high CH3Br loading and produced substantial concentrations of methyl thiosulfate. Submergence of the GAC in water and use of the GAC as the cathode of an electrolysis unit also debrominated sorbed CH3Br. The reaction appeared to involve a one-electron transfer, producing methyl radicals that incorporated into the GAC. Destruction rates increased with decreasing applied voltage down to ∼−1.2 V vs the standard hydrogen electrode. Cycling experiments conducted at −0.77 V indicated that >80% debromination of CH3Br was achieved over ∼30 h with ∼100% Coulombic efficiency. Sorptive capacity and degradation efficiency were maintained over at least 3 cycles. Capture of CH3Br fumes from fumigation chambers onto GAC, and electrolytic destruction of the sorbed CH3Br could mitigate the negative impacts of CH3Br usage pending the development of suitable replacement fumigants.

■

INTRODUCTION The fumigant methyl bromide (CH3Br) has been widely used to control agricultural pests in soils and commodities;1,2 however, concerns over the potential for CH3Br emissions to contribute to depletion of the stratospheric ozone layer led to the international regulation of its use via the Montreal Protocol in 19873 and in the United States via the Clean Air Act.4 Although the lifetime of CH3Br in the atmosphere is relatively short at ∼2 years, its ozone depletion potential is considered ∼60% that of chlorofluorocarbon-11 (CFC-11).1 Methyl bromide emissions from natural sources, particularly the ocean, exceed all anthropogenic sources.1 Transitioning from the agricultural use of CH3Br has been challenging.2 Many proposed replacements cannot match the relatively short treatment time required for CH3Br pesticidal efficacy.1 Rapid, efficacious treatment is particularly important for postharvest quarantine preshipment (QPS) applications, which are very often constrained by shipping or port logistics. Accordingly, key postharvest QPS uses of CH3Br were exempted from the phaseout as outlined in the Montreal Protocol.2 For most postharvest applications, CH3Br currently is vented to the atmosphere from fumigation chambers. Pending the © 2015 American Chemical Society

identification of suitable replacement fumigants, capture and destruction of CH3Br may enable its continued use for these applications without contributing to ozone depletion. Methyl bromide vented from fumigation containers can be captured on granular activated carbon (GAC),5 but disposal of the GAC as hazardous waste would be expensive. Previous research has noted that CH3Br and other fumigants react rapidly with reduced sulfur species in aqueous solution via nucleophilic substitution reactions.5−9 For example, CH3Br was degraded by the hydrogen sulfide produced by biological sulfate reduction in anaerobic sediments.6 Ammonium thiosulfate rapidly destroyed CH3Br via a nucleophilic substitution reaction with thiosulfate in aqueous solution, indicating the potential for addition of this fertilizer to fumigated soils to reduce CH3Br emissions:7 CH3Br + S2 O32 − → CH3S2O3− + Br − Received: Revised: Accepted: Published: 4515

(1)

November 22, 2014 March 5, 2015 March 19, 2015 March 19, 2015 DOI: 10.1021/es505709c Environ. Sci. Technol. 2015, 49, 4515−4521

Article

Environmental Science & Technology

vials by spiking through the septa to achieve various concentrations. Over 24 h, samples were sacrificed for analysis of residual CH3Br by immediate extraction. GAC and associated water were transferred to filter paper contained within a glass funnel and were washed with 20 mL of deionized water. The ∼20 mL filtrate, used as a rough estimate of CH3Br in the pore spaces, and GAC were extracted separately by shaking each with 10 mL of ethyl acetate for 2 min. Electrolytic Degradation of CH3Br. Electrolysis was conducted using a CH-600D potentiostat (CH Instruments, Austin, TX, USA). The borosilicate glass electrochemical cells were constructed similarly to those in a previous study.15 Briefly, cathode and anode chambers (∼50 mL internal volume each) were separated by a cation exchange membrane (Ultrex CMI-7000, Membranes International, USA). In the cathode chamber, a Ag/AgCl (1 M KCl) reference electrode (CHI111, porous Teflon tip, CH Instruments, Austin, TX, USA) was placed within 0.5 cm of the working electrode, and a platinum wire (0.127 mm thick, Alfa Aesar, Ward Hill, MA, USA) was used as a counter electrode in the anode chamber. For the working electrode, CH3Br was sorbed to GAC by spiking 5 μmol of CH3Br into 10 mL of deionized water in the presence of 8 g of GAC in headspace-free vials and shaking the vials on a horizontal shaker overnight. After overnight equilibration, the GAC was rinsed with 20 mL of deionized water and transferred into ∼1.5 cm × 6 cm cylinders constructed from sheet graphite. The mass of CH3Br sorbed to the GAC postequilibration was measured as described above to mark the initial concentration of CH3Br on the GAC prior to electrolysis. Generally, ≥90% of the initial CH3Br mass added was sorbed to the GAC after equilibration, and no significant mass of CH3Br occurred in the aqueous phase. A copper wire was connected to the top of the graphite sheet tube in the headspace above the water. Both cathode and anode chambers were filled with deionized water buffered at pH 7 with 100 mM phosphate buffer. The potentiostat was set to apply to the working electrode voltages ranging from −345 to −1795 mV vs the standard hydrogen electrode (S.H.E.). At various times up to 24 h, the reaction was halted, and the CH3Br in the aqueous phase supernatant and the residual CH3Br sorbed to the GAC or occurring in the aqueous phase within pore spaces were immediately and separately extracted into ethyl acetate. The CH3Br in the supernatant was measured by extracting 10 mL of supernatant with 10 mL of ethyl acetate; the CH3Br in the GAC pore spaces and sorbed to the GAC were analyzed as described above for the thiosulfate reactions. The CH3Br in the supernatant and in the filtrate from rinsing the GAC particles was always negligible, indicating that desorption from the GAC to the aqueous phase was unimportant. Cycling Experiments. Cycling experiments were conducted to compare the efficacy of thiosulfate and electrolysis to degrade high levels of CH3Br sorbed to GAC over three sorption−degradation cycles. Pure gas-phase CH3Br was routed from a cylinder for 3 min into a 25 mL glass vial containing preweighed GAC particles via a syringe through a Teflon-lined septum; a needle through the septum permitted the gas to exit. The GAC particles were then treated with either thiosulfate or electrolysis. For the reaction with thiosulfate, two 2 g aliquots of the CH3Br-spiked GAC particles were extracted with ethyl acetate as described above for analysis of the initial CH3 Br concentration. An additional 20 g of the spiked GAC particles were soaked in 30 mL of either 2 or 4.5 M sodium thiosulfate

This reaction has also been applied for CH3Br capture and destruction from postharvest fumigation chambers.10 In this scheme, CH3Br fumes are captured on GAC. The GAC bed is heated to revolatilize the CH3Br, which then is purged through a separate tank containing a thiosulfate solution to destroy the CH3Br. The GAC can be reused to capture the fumes from the next fumigation treatment. The revolatilization of CH3Br from the GAC, and the separate thiosulfate treatment bath entails the implicit assumption that sorption of CH3Br to the GAC sequesters CH3Br from destruction by thiosulfate in the aqueous phase.7 The first objective of this work was to determine whether CH3Br can be degraded by thiosulfate while still sorbed to GAC, a finding that would enable capture and degradation of CH3Br within a single GAC treatment unit. Previous research has indicated that compounds may be destroyed while still sorbed to black carbons. Hydrolysis of CH3Br sorbed to activated carbon has been reported at ∼3 h half-lives at 80 °C.5 In previous work, we demonstrated that sorption of the explosives RDX and nitroglycerin to black carbons promoted their degradation by hydrogen sulfide even at room temperature via either nucleophilic substitution (RDX) or reduction (nitroglycerin) pathways.11−14 Because of the high concentration of CH3Br that would be sorbed to the GAC under conditions relevant to postharvest fumigation chambers, substantial concentrations of thiosulfate would be required for treatment. A second objective was to evaluate whether CH3Br captured on GAC could be reductively destroyed by converting the GAC to a cathode within an electrolysis cell, in effect substituting electrons from the electric grid for thiosulfate as the reagents responsible for CH3Br degradation. The results of the electrolysis experiments lay the groundwork for a simpler treatment scheme for the capture and destruction of CH3Br from fumigation chambers.

■

MATERIALS AND METHODS Materials. Alfa Aesar (Ward Hill, MA) sheet graphite (0.13 mm thickness, catalog number 43078) was used as received. Fisher (Pittsburgh, PA) GAC (6−14 mesh) was rinsed three times with deionized water and then oven-dried overnight prior to use. The Supporting Information describes other reagents, standards preparation, and analytical methods. Thiosulfate-Mediated Degradation of CH3Br. In the absence of GAC, CH3Br (1.7 μmol) was spiked through Teflon-lined septa into headspace-free vials containing 5 mL of deionized water buffered at pH 7 with 10 mM phosphate buffer and containing various concentrations of thiosulfate. Over 24 h, samples were sacrificed for analysis of residual CH3Br by immediate extraction; samples were transferred into 10 mL vials containing 3 mL of ethyl acetate and extracted by shaking for 2 min. Ethyl acetate extracts were analyzed by gas chromatography with electron capture detection, as described below. In the presence of GAC, CH3Br (1 μmol) was spiked through Teflon-lined septa into headspace-free vials containing 3 mL of deionized water buffered at pH 7 with 10 mM phosphate buffer and 2.4 g GAC. To equilibrate the CH3Br with the GAC, samples were shaken on a horizontal shaker at 100 rpm for 24 h. Under these conditions, there was negligible supernatant; nearly all of the water occurred in pores or pore spaces between GAC particles. Analysis of the water and GAC after 24 h equilibration (see below) indicated that >99% of the CH3Br was sorbed to the GAC. Thiosulfate was added to the 4516

DOI: 10.1021/es505709c Environ. Sci. Technol. 2015, 49, 4515−4521

Environmental Science & Technology

■

Article

RESULTS AND DISCUSSION Thiosulfate-Mediated Degradation of CH3Br. Initial experiments were conducted to evaluate destruction of 1.7 μmol (333 μM) of CH3Br by 0, 3, and 30 mM sodium thiosulfate in a 5 mL aqueous phase at pH 7 in the absence of GAC. To enable comparison with loss rates observed when CH3Br was sorbed to GAC, loss of the total mass (M) of CH3Br was monitored over time. No significant loss of CH3Br was observed over 24 h in the absence of thiosulfate (Figure 1A). In the presence of thiosulfate, CH3Br loss followed

solutions buffered at pH 7 with 100 mM phosphate buffer. After 15 h, the GAC particles were filtered using Whatman 2 μm pore size filter papers and rinsed with 20 mL of deionized water. Two 2 g aliquots of the GAC particles were extracted by ethyl acetate for analysis of residual CH3Br. The remaining GAC particles were dried at 40 ◦C for 2 h in an oven. The dried GAC was loaded with gas-phase CH3Br as described above to initiate the second of three cycles. Two 2 g aliquots were extracted at the beginning and end of each cycle to measure CH3Br. For the electrolysis treatment, 2 g of the CH3Br-spiked GAC particles were split into four 0.5 g aliquots. The four aliquots were transferred into ∼1.5 cm × 2 cm cylinders fashioned from sheet graphite and were submerged in deionized water buffered at pH 7 with 100 mM phosphate buffer. One aliquot was analyzed immediately to determine the initial CH 3 Br concentration. A second aliquot served as the cathode in the electrolysis cell described above, applying −0.77 V versus S.H.E. to the cathode. After 30 h, the electrolysis was halted together with the other two nonelectrolyzed control aliquots to complete the first cycle. Residual CH3Br and other products were measured. To ensure that the mass of GAC subject to electrolysis was consistent throughout the cycles and that the extraction procedures involved in sample analysis did not alter the activity of the GAC during subsequent cycles, separate sets of aliquots were set up in the same fashion to evaluate the second and third cycles without intermediate analyses. For example, to evaluate the second cycle, the GAC particles were oven-dried overnight at 120 °C after the first cycle without analysis and then reloaded with gaseous CH3Br to initiate the second cycle; analyses were conducted only after the second cycle. Because of the high volatility of CH3Br, there were concerns about the loss of CH3Br into the headspace of the cathode cell during electrolysis or during sample processing (e.g., filtration of GAC particles prior to extraction for analysis of sorbed CH3Br). Several lines of evidence indicate that these losses were negligible. First, CH3Br sorbed strongly to the GAC. As indicated below, after pre-equilibration of GAC with CH3Br in water, >99% of CH3Br generally was sorbed to the GAC, and negligible concentrations were in the aqueous phase. The strong sorption would hinder CH3Br loss to the headspace of the cathode cells. Second, at the high CH3Br loadings onto GAC from the gas phase used to initiate the cycling experiments, the CH3Br concentrations measured by direct extraction of the GAC particles with ethyl acetate (prior to exposure to water) matched those estimated from gravimetric analysis of the GAC (100.3% ± 6.3%). After exposure of the same GAC loaded with CH3Br to water, filtration, and extraction, as described above, the recovery of CH3Br was ∼62.5% (±2.1%). That this represented an extraction efficiency from the wet carbon rather than CH3Br losses during processing was demonstrated when a subsequent extraction of the same carbon retrieved an additional 20% (±0.7%) of the CH3Br initially sorbed to the GAC, such that ∼83% overall recovery was attained from the wet carbon via the two extractions. Extraction of the aqueous phase yielded an additional 2.0% (±0.7%), indicating that the desorption from GAC to aqueous phase was negligible. Third, as indicated below, the aqueous bromide concentration measured after the electrolysis cycling experiments indicated a 98% molar yield from the initial CH3Br, indicating that the fate of CH3Br was reaction, not losses by volatilization.

Figure 1. (A) Degradation by 3 mM thiosulfate of the mass (M) of 1.7 μmol of CH3Br in 5 mL deionized water without GAC (aqueous) or of 1.0 μmol of CH3Br in 3 mL deionized water with 2.4 g GAC. Solutions buffered at pH 7 with 10 mM phosphate. Controls = without thiosulfate. Error bars represent the standard deviation of experimental triplicates. (B) Pseudo-first-order observed degradation rate constants of total CH3Br mass as a function of applied thiosulfate concentrations with and without GAC. Error bars represent the standard error of the regression.

pseudo-first-order kinetics (e.g., Figure 1A for 3 mM thiosulfate). The slope of the plot of the pseudo-first-order loss rate constants (k obs = k[S 2 O 3 2− ]) vs thiosulfate concentrations (Figure 1B) provided a second-order rate constant (k) of 2.2 × 10−2 M−1 s−1 (±0.1 M−1 s−1 standard error of the regression), similar to the 2.1 × 10−2 M−1 s−1 rate constant determined by Wang et al.7 using ammonium thiosulfate. When these experiments were repeated in the presence of 2.4 g of GAC pre-equilibrated with 1.0 μmol of CH3Br, such that >99% of the CH3Br was sorbed to the GAC surface, no significant decay of CH3Br was observed in the absence of thiosulfate (Figure 1A). In the presence of thiosulfate, the total mass of CH3Br (equivalent to that recovered from the GAC surface) decayed in a pseudo-first-order fashion initially, but the 4517

DOI: 10.1021/es505709c Environ. Sci. Technol. 2015, 49, 4515−4521

Article

Environmental Science & Technology rate declined after ∼4 h (e.g., Figure 1A for 3 mM thiosulfate). Pseudo-first-order loss rate constants were calculated using data from the first 4 h. For thiosulfate concentrations ≥3 mM, kobs did not vary (Figure 1B). The results indicate that GAC can mediate the destruction of CH3Br sorbed to the GAC surface by thiosulfate over relatively short time scales (half-lives ∼ 2 h); however, the lack of increase in kobs with increasing thiosulfate concentration suggests that thiosulfate is not involved in the rate-limiting step. To further evaluate how the thiosulfate reacts with GACsorbed CH3Br, GAC (2.4 g) was treated for 24 h with 3 mL of 3 mM thiosulfate in 10 mM phosphate-buffered deionized water at pH 7. After the aqueous phase was decanted, the GAC was rinsed with 20 mL of deionized water, 3 mL of 10 mM phosphate-buffered deionized water at pH 7 was added to the GAC, and 1 μmol of CH3Br was injected. The vials were capped and slowly mixed by rotation for 24 h. Analysis of the residual CH3Br on the GAC indicated that 61.3% (±6.9% standard deviation of experimental triplicates) of the CH3Br was degraded by the thiosulfate pretreated GAC, but no significant loss of CH3Br was observed in controls in which no pretreatment of the GAC occurred. Reduced sulfur species, including thiosulfate, are potent reductants and nucleophiles.16 To differentiate these reaction pathways, an electrochemical cell was constructed as described previously.12 Briefly, anode and cathode cells consisted of 24 mL glass vials, each containing 20 mL of 10 mM phosphatebuffered deionized water at pH 7 and 0.1 g sheet graphite electrodes. The sheet graphite electrodes extended into the headspace above the solution, where they were connected by copper wire. A salt bridge constructed of Teflon tubing filled with an agarose gel containing 1 M potassium chloride completed the electrical circuit between the electrode solutions. The anode solution contained 3 mM thiosulfate, and the cathode solution contained 1 μmol of CH3Br. After 24 h, only 21.3% (±2.3% standard deviation for experimental duplicates) of the CH3Br was recovered from the cathode cell, whereas 90.7% (±4.2%) was recovered from controls lacking thiosulfate in the anode cell. Similar results were found for higher CH3Br loadings. Anode and cathode cells (50 mL each) containing 10 mM phosphate-buffered deionized water at pH 7 were separated by a cation exchange membrane but connected by a copper wire. GAC exposed to gas-phase methyl bromide (35% by weight loading) was packed into a 1.5 cm × 4 cm sheet graphite cylinder in the cathode chamber. A sheet graphite electrode (0.1 g, 0.5 cm × 4 cm) was placed in the anode chamber containing 2 M thiosulfate. After 24 h, only 17.5% of the CH3Br was recovered from the cathode cell, whereas 107% was recovered from controls lacking thiosulfate. The physical separation of the thiosulfate and CH3Br rules out that the reaction involved a nucleophilic substitution between aqueous thiosulfate and aqueous or sorbed CH3Br. Together, these results suggest that thiosulfate degrades CH3Br sorbed to GAC at least partly via a reduction reaction. The lack of dependence of kobs on increasing thiosulfate concentration (Figure 1B) and the finding that graphite pretreated with thiosulfate can mediate CH3Br destruction suggests that degradation is mediated by a reactive intermediate on the carbon surface rather than directly by thiosulfate. The nature of the reactive intermediate is unclear; however, an organosulfur intermediate is possible because Xu et al.13 detected the occurrence of sulfur on the surface of sheet graphite pretreated with hydrogen sulfide.

Electrolytic Degradation of CH3Br. To avoid the need for the thiosulfate reagent, we evaluated whether CH3Br sorbed to GAC could be degraded by reduction via electrons supplied from the electric grid. Electrolysis of CH3Br presorbed to GAC as the cathode was conducted at constant voltage. The reaction was halted at different times for measurement of residual CH3Br on the GAC. For controls in which the copper wire did not connect the electrodes, no degradation was observed over 24 h. When the wire was connected, the degradation of sorbed CH3Br followed pseudo-first-order kinetics (e.g., Figure 2 for

Figure 2. Pseudo-first-order degradation of the mass (M) of CH3Br presorbed to GAC in 100 mM phosphate-buffered deionized water at pH 7 under an electrolysis voltage of −1295 mV vs S.H.E. The average relative standard deviation of experimental replicates was 17%.

1295 mV vs S.H.E. applied to the cathode). As discussed further with regard to the cycling experiments, reduction at the cathode of CH3Br may occur by either 1 (eq 2) or 2 electron transfers (eq 3) coupled with oxidation of water at the anode (reverse of eq 4): CH3Br + e− ↔ CH3• + Br − +

−

(2)

CH3Br + H + 2e ↔ CH4 + Br O2 + 4H+ + 4e− ↔ 2H 2O

−

(3) (4)

Figure 3 provides the observed pseudo-first-order degradation rate constants for CH3Br sorbed to GAC as a function of the applied electrolysis voltage. The kobs increased to a

Figure 3. Pseudo-first-order kobs of CH3Br presorbed to GAC under electrolysis voltages ranging from −345 to −1795 mV vs S.H.E. Error bars represent the standard error of the regression. 4518

DOI: 10.1021/es505709c Environ. Sci. Technol. 2015, 49, 4515−4521

Article

Environmental Science & Technology maximum of ∼0.11 h−1 as the voltage applied to the cathode was reduced to −1295 mV vs S.H.E. and remained at this value at −1795 mV vs S.H.E. The plateau in kobs likely reflects the onset of significant competition for electrons from proton reduction (eq 5) at this voltage. Although the standard reduction potential for reduction of protons to hydrogen gas at pH 7 is −0.41 V vs S.H.E., carbon electrodes often exhibit a significant overpotential for this reaction.17 2H+ + 2e− ↔ H 2(g)

deviation of experimental duplicates) molar yield of bromide and 47% (±0.4%) of methyl thiosulfate compared with CH3Br degraded. Bromide and methyl thiosulfate are the expected products of a nucleophilic substitution pathway (eq 1). The fate of the other 53% of methyl groups associated with degraded CH3Br is unclear. No methanol was detected in the supernatant, although controls indicated that methanol would not sorb significantly to the GAC. To address the potential for formation of methanethiol as a product of methyl thiosulfate, the experiment was repeated for 24 h, but no methanethiol was detected in either the GAC or aqueous phases; for an initial CH3Br loading on the GAC of 0.25 mol and a detection limit of 0.8 μmol in the methylene chloride extract, any methanethiol formation would have been readily detected. Bromide is also the expected product of reduction pathways (eqs 2 and 3). The electrochemical cell experiments suggested that reductive degradation pathways also were operative (see above). Attempts to evaluate the potential for formation of methane via a two-electron reduction (eq 3) were unsuccessful because of the difficulty of measuring methane sorbed to GAC. Aliquots of GAC and associated aqueous solutions were transferred after the 15 h thiosulfate contact period from headspace-free vials to larger vials with headspace; analysis of the headspace using a gastight syringe did not detect methane. Control experiments indicated that methane would volatilize readily from aqueous solution. However, when GAC was treated with a stream of 99.7% pure methane, methane sorbed strongly to the GAC (10% loading by weight). The GAC with sorbed methane was transferred into either open or sealed vials and then heated at 120 °C overnight. The weight of the GAC particles in the open vials returned to levels prior to methane exposure, indicating that desorption of methane occurs at high temperature with sufficient dilution gas. However, for the sealed vials at 120 °C, we detected no methane in the headspace of the vials and no significant decrease in weight of the GAC particles. Previous research indicated a strong vacuum is needed to recover methane from GAC,18 but the loss in methane resulting from generation of this vacuum would preclude quantification of methane. Alternatively, a one-electron reduction of CH3Br could produce methyl radicals (CH3•; eq 2). Combination of methyl radicals with the GAC or with each other (as a result of the high concentration of CH3Br on the GAC surface) is a possibility. Incorporation of methyl radicals into the GAC would result in an increase in GAC weight that would not decrease upon heating at 120 °C overnight in open containers. Thus, measurement of a permanent increase in GAC weight might distinguish between the one-electron or two-electron reduction pathways. Although this technique would be useful for electrolysis (see below), it would not be definitive for thiosulfate because an increase in weight might also result from precipitation of the thiosulfate in GAC porewaters during heating. For the electrolytic treatment, the initial weight loading of CH3Br and degradation efficiency achieved over a 30 h treatment cycle remained high over all three cycles (Table 2). Product characterization after the first cycle indicated 98% (±2% standard deviation of experimental duplicates) molar yield of bromide compared to CH3Br loss. The Coulombic efficiency for CH3Br reduction (ε), the percentage of electrons emitted by the potentiostat that resulted in CH3Br reduction, was calculated as

(5)

Cycling Experiments. To compare the efficiency of thiosulfate-mediated and electrolytic degradation of CH3Br sorbed to GAC under conditions relevant to field-scale operations, GAC was loaded to ∼15−30% by weight by passing a stream of pure CH3Br gas over the GAC for ∼3 min. Note that these loadings are more variable and likely less than those could have been achieved by prolonged passage of CH3Br over the GAC; prolonged treatment with CH3Br would have rapidly depleted the CH3Br cylinder. The GAC was then treated with either 2 or 4.5 M aqueous thiosulfate solutions or electrolysis at −0.77 V vs S.H.E. in solutions buffered at pH 7 with 100 mM phosphate buffer. After treatment, aliquots of the GAC were extracted to measure residual CH3Br, and other aliquots were oven-dried and reloaded with gas-phase CH3Br for the second of three treatment cycles. Control experiments conducted with CH3Br sorbed to the GAC submerged in 100 mM phosphate buffer at pH 7 in the absence of thiosulfate or electrolysis indicated no significant loss of CH3Br over 30 h. For treatment with 4.5 M thiosulfate, the initial CH3Br loading declined rapidly from 15% to 0.1% by weight over the three cycles (Table 1). For an initial CH3Br loading of 15% Table 1. Results of Cycling Experiments for Thiosulfate Treatment cycle 4.5 M thiosulfate 1 2 3 2 M thiosulfate 1 2 3 a b

initial loading (weight %)

removal (%)

14.9 (2.4)a 4.4 (0.7) 0.1 (0.04)

77 (8.4) 100b 71.9 (11.9)

29 (0.2) 29 (0.4) 20 (0.6)

93 (0.5) 97 (0.2) 95 (0.03)

Parentheses provide standard deviation of experimental duplicates. No residual CH3Br detected in either duplicate.

(cycle 1), the 4.5 M thiosulfate solution represented a molar ratio of aqueous thiosulfate to sorbed CH3Br of ∼4:1. At this concentration, which is close to the solubility of sodium thiosulfate, precipitation of clear crystals was observed on the carbon surface. Over the 15 h exposure to thiosulfate during each cycle, the degradation of sorbed CH3Br was high over the three cycles (Table 1). Together, the results suggested that precipitation of sodium thiosulfate resulted in blockage of CH3Br sorption sites, but not loss of the ability of GAC to mediate degradation of sorbed CH3Br. When the cycling experiment was repeated with 2 M thiosulfate, no precipitate was observed. The initial weight loading of CH3Br and the degradation efficiency over the 15 h thiosulfate contact period remained high over all three cycles (Table 1). Product analysis conducted after the first cycle for the 2 M thiosulfate treatment indicated 90% (±0.5% standard 4519

DOI: 10.1021/es505709c Environ. Sci. Technol. 2015, 49, 4515−4521

Article

Environmental Science & Technology

during subsequent cycles. Further research is needed to understand the nature of the carbon products formed during electrolysis. Environmental Relevance. The capture and destruction of CH3Br fumes vented from fumigation chambers onto GAC would mitigate the impacts of the continued use of CH3Br, pending identification of suitable replacement fumigants by preventing emissions to the atmosphere and the associated contribution of these emissions to stratospheric ozone depletion. Our results suggest that CH3Br fumes captured on GAC can be destroyed by submerging the GAC bed in an aqueous thiosulfate solution while the CH3Br is still sorbed to GAC. Such a scheme would obviate the need to heat the GAC to drive the CH3Br back into the gas phase and to purge the vapors through a separate reactor unit containing a thiosulfate bath. Following direct destruction of the GAC-sorbed CH3Br, the GAC bed could be drained, dried, and returned to service to capture fumes from another fumigation chamber. Although the observed pseudo-first-order loss rate of CH3Br did not increase with thiosulfate concentration, >90% CH3Br destruction was attained within 15 h during the cycling experiments (Table 1). This time scale is short enough that continuous use of fumigation chambers could be attained if several GAC beds attached to a single fumigation chamber were cycled through capture, destruction, and regeneration cycles. Although feasible, thiosulfate treatment of CH3Br-saturated GAC has several significant drawbacks. Because of the high loading of CH3Br on the GAC under conditions relevant to capture of CH3Br fumes from fumigation chambers, molar concentrations of thiosulfate would be needed for CH3Br destruction. At the highest CH3Br loadings on GAC, stoichiometric addition of thiosulfate may approach the solubility of thiosulfate. At these levels, thiosulfate precipitation in the pores may deplete the sorptive capacity of the GAC such that the GAC could not be reused over many cycles. In addition to the cost associated with the molar concentrations of thiosulfate, near molar concentrations of methyl thiosulfate would be formed. The toxicity of methyl thiosulfate is unclear, but its evaluation would be necessary to understand the implications for disposal at high concentrations. A preferable alternative may be the reductive debromination of CH3Br captured on GAC by conversion of the GAC to a cathode within an electrolysis cell. Reducing the voltage to −0.77 V vs S.H.E. permitted destruction of sorbed CH3Br by 82−88% over 30 h. Although this time scale is longer than that associated with thiosulfate treatment, cycling of GAC beds through capture, treatment, and regeneration cycles should be feasible. Cycling experiments conducted when the GAC was loaded at the high concentrations anticipated for CH3Br capture at fumigation facilities indicated that the GAC maintained its performance in terms of CH3Br sorption capacity, degradation efficiency, and Coulombic efficiency over at least three cycles. By supplying the active reagent from the electric grid, this scheme avoids the need to purchase, transport, and handle large quantities of thiosulfate and avoids methyl thiosulfate formation. Disposal of bromide, a primary waste product of the electrolysis process, to seawater may be possible at port facilities. Instead of methane production, the fate of the methyl group appears to be incorporation into the GAC or production of alternative hydrocarbons. Although methane production would have enabled its capture for energy generation, the lack of significant methane production may avoid the potential for explosion risks.

Table 2. Results of Cycling Experiments for Electrolysis Treatmenta cycle

initial loading (weight %)

removal (%)

Coulombic efficiencyb (%)

weight increasec (%)

1 2 3

38 (3)d 31 42 (6)

84 (0.8) 88 82 (10)

103 (2) 92 116 (16)

5.6 (1.2) 4.8 (0.6) 6.3

Electrolysis conducted at pH 7 and −1.0 V vs S.H.E. bCoulombic efficiency calculated assuming a one-electron transfer. cCompared with initial GAC weight. dParentheses provide standard deviation of experimental duplicates. a

ε=

nFMCH 3Br Q

(6)

where n is the number of electrons transferred, F is the Faraday constant (96 485 C mol−1e), M is the moles of CH3Br degraded, and Q is the cumulative coulombs of charge transferred from the potentiostat. Assuming a two-electron transfer (eq 3), the calculated Coulombic efficiency generally was ∼200%. Such a high efficiency would require that, in addition to reducing CH3Br to methane (eq 2), the electrolysis alters the GAC surface to promote alternative degradation pathways. For example, generation of nucleophilic groups on the GAC surface may promote the hydrolysis of CH3Br to methanol (eq 7), raising the apparent Coulombic efficiency above 100%. CH3Br + H 2O ↔ CH3OH + Br −

(7)

However, no methanol or methane were detected. Because of the difficulty associated with analysis of methane sorbed to GAC, an additional electrolysis experiment was conducted under the same conditions, except that the sheet graphite cylinder contained no GAC. CH3Br was injected to the aqueous phase to achieve a concentration of 8 mM. After 24 h, no methane was observed in the headspace, despite 91% loss of CH3Br. Alternatively, CH3Br reduction via electrolysis forms methyl radicals (eq 2) that incorporate into GAC. The Coulombic efficiencies calculated assuming a one-electron transfer (eq 2) were ∼100% for all three cycles (Table 2). After each cycle, the GAC particles were dried overnight at 120 °C and weighed. No increase in weight was observed for controls in which GAC loaded with CH3Br was exposed to phosphate buffer over 30 h without electrolysis. However, an increase in weight was observed for aliquots treated by electrolysis (Table 2). For the first cycle, the 5.6% (±1.2%) increase in weight was comparable to the 3.2% expected if all of the methyl groups associated with the CH3Br that degraded during the electrolysis were to become incorporated into the GAC. Previous research has indicated that formation of methyl radicals via plasma pyrolysis of methane can lead to the production of aromatic groups, particularly in the presence of graphite electrodes.19 No further increase in weight was observed during subsequent cycles. However, significant generation of methane during these latter cycles is unlikely because it would necessitate Coulombic efficiencies >100%. The results suggest that CH3Br electrolysis proceeds predominantly by one-electron reduction (eq 2) rather than via methane generation (eq 3). One possibility is that methyl radical generation results predominantly in incorporation into GAC during the first cycle, but production of alternative hydrocarbons that are not covalently bound to the GAC (and therefore subject to volatilization upon heating) 4520

DOI: 10.1021/es505709c Environ. Sci. Technol. 2015, 49, 4515−4521

Article

Environmental Science & Technology

(9) Zheng, W.; Yates, S. R.; Papiernik, S. K.; Guo, M.; Gan, J. Dechlorination of chloropicrin and 1,3-dichloropropene by hydrogen sulfide species: Redox and nucleophilic substitution reactions. J. Agric. Food Chem. 2006, 54, 2280−2287. (10) Nordiko Quarantine Systems. Submission to ERMA on the reassessment of methyl bromide; February, 2010; http://www.nordiko. com.au/fileadmin/pdf_reports/Website_ERMA_Submission_Feb_ 2010.pdf (accessed September 3, 2013). (11) Kemper, J. M.; Ammar, E.; Mitch, W. A. Abiotic degradation of RDX in the presence of hydrogen sulfide and black carbon. Environ. Sci. Technol. 2008, 42 (6), 2118−2123. (12) Xu, W.; Dana, K. E.; Mitch, W. A. Black-carbon mediated destruction of nitroglycerin and RDX by hydrogen sulfide: Relevance to in situ remediation. Environ. Sci. Technol. 2010, 44, 6409−6415. (13) Xu, W.; Pignatello, J. J.; Mitch, W. A. The role of black carbon electrical conductivity in mediating hexahydro-1,3,5-trinitro-1,3,5triazine (RDX) transformation on carbon surfaces by sulfides. Environ. Sci. Technol. 2013, 47, 7129−7136. (14) Xu, W.; Pignatello, J. J.; Mitch, W. A. Reduction of Nitroaromatics Sorbed to Black Carbon by Direct Reaction with Sorbed Sulfides. Environ. Sci. Technol. 2015, DOI: 10.1021/es5045198. (15) Radjenovic, J.; Farre, M. J.; Mu, Y.; Gernjak, W.; Keller, J. Reductive electrochemical remediation of emerging and regulated disinfection byproducts. Water Res. 2012, 46, 1705−1714. (16) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. Environmental Organic Chemistry, 2nd ed.; John Wiley & Sons: Hoboken, 2003. (17) Kinoshita, K. Carbon: Electrochemical and Physicochemical Properties; Wiley: New York, 1988. (18) Matranga, K. R.; Myers, A. L.; Glandt, E. D. Storage of natural gas by adsorption on activated carbon. Chem. Eng. Sci. 1992, 47, 1569−1579. (19) Fincke, J. R.; Anderson, R. P.; Hyde, T. A.; Detering, B. A. Plasma pyrolysis of methane to hydrogen and carbon black. Ind. Eng. Chem. Res. 2002, 41, 1425−1435.

Although the electrolysis system is promising, additional research is needed. In addition to evaluating designs for GACbased electrolysis units, our laboratory is determining the effect of GAC properties on both CH3Br sorptive capacity and electrolytic degradation efficiency to determine whether electrolytic degradation can be achieved over shorter time scales. The fate of the methyl group in CH3Br needs further evaluation. If the production of methyl radicals leads to the production of conjugated aromatic regions, the high conductivity of these regions might enhance destruction rates in subsequent electrolytic cycles. However, if nonreactive carbon deposits form, performance may decline over subsequent cycles as a result of blockage of sorption and reaction sites.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional materials and methods. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel: +1 650 725-9298. Fax: +1 650 723-7058. E-mail: [email protected]. Author Contributions ⊥

Y.Y. and Y.L. contributed equally to this work

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was funded by the USDA Agricultural Research Service as well as the USDA Foreign Agricultural Service and the California Dried Plum Board under the Technical Assistance for Specialty Crops program, agreement no. 201019. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer.

■

REFERENCES

(1) Ristaino, J. B.; Thomas, W. Agriculture, methyl bromide, and the ozone hole: Can we fill the gaps? Plant Dis. 1997, 81, 964−977. (2) Johnson, J. A.; Walse, S. S.; Gerik, J. S. Status of alternatives for methyl bromide in the United States. Outlooks Pest Manage. 2012, 23, 53−58. (3) United Nations Environment Programme - Ozone Secretariat. The Montreal Protocol on Substances that Deplete the Ozone Layer; http://ozone.unep.org/new_site/en/montreal_protocol.php (accessed August 28, 2013). (4) United States Environmental Protection Agency. The Phaseout of Methyl Bromide; http://www.epa.gov/ozone/mbr/ (accessed August 28, 2013). (5) Gan, J.; Megonnell, N. E.; Yates, S. R. Adsorption and catalytic decomposition of methyl bromide and methyl iodide on activated carbons. Atmos. Environ. 2001, 35, 941−947. (6) Oremland, R. S.; Miller, J. G.; Strohmaier, F. E. Degradation of methyl bromide in anaerobic sediments. Environ. Sci. Technol. 1994, 28, 514−520. (7) Wang, Q.; Gan, J.; Papiernik, S. K.; Yates, S. R. Transformation and detoxification of halogenated fumigants by ammonium thiosulfate. Environ. Sci. Technol. 2000, 34, 3717−3721. (8) Bondarenko, S.; Zheng, W.; Yates, S. R.; Gan, J. Dehalogenation of halogenated fumigants by polysulfide salts. J. Agric. Food Chem. 2006, 54, 5503−5508. 4521

DOI: 10.1021/es505709c Environ. Sci. Technol. 2015, 49, 4515−4521