Development of an Activated Carbon-Based Electrode for the Capture

Agricultural Research Service, San Joaquin Valley Agricultural Sciences Center, USDA, 9611 South Riverbend ... Publication Date (Web): September 9, 20...
0 downloads 0 Views 6MB Size
Download PDF
Article pubs.acs.org/est

Development of an Activated Carbon-Based Electrode for the Capture and Rapid Electrolytic Reductive Debromination of Methyl Bromide from Postharvest Fumigations Yuanqing Li,† Chong Liu,‡ Yi Cui,‡ Spencer S. Walse,§ Ryan Olver,∥ David Zilberman,∥ and William A. Mitch*,† †

Department of Civil and Environmental Engineering, Stanford University, 473 Via Ortega, Stanford, California 94305, United States Department of Materials Science and Engineering, Stanford University, McCullough Building, Stanford, California 94305, United States § Agricultural Research Service, San Joaquin Valley Agricultural Sciences Center, USDA, 9611 South Riverbend Avenue, Parlier, California 93648-9757, United States ∥ Department of Agricultural and Resource Economics, Giannini Hall, University of California at Berkeley, Berkeley, California 94720, United States ‡

S Supporting Information *

ABSTRACT: Due to concerns surrounding its ozone depletion potential, there is a need for technologies to capture and destroy methyl bromide (CH3Br) emissions from postharvest fumigations applied to control agricultural pests. Previously, we described a system in which CH3Br fumes vented from fumigation chambers could be captured by granular activated carbon (GAC). The GAC was converted to a cathode by submergence in a high ionic strength solution and connection to the electrical grid, resulting in reductive debromination of the sorbed CH3Br. The GAC bed was drained and dried for reuse to capture and destroy CH3Br fumes from the next fumigation. However, the loose GAC particles and slow kinetics of this primitive electrode necessitated improvements. Here, we report the development of a cathode containing a thin layer of small GAC particles coating carbon cloth as a current distributor. Combining the high sorption potential of GAC for CH3Br with the conductivity of the carbon cloth current distributor, the cathode significantly lowered the total cell resistance and achieved 96% reductive debromination of CH3Br sorbed at 30% by weight to the GAC within 15 h at −1 V applied potential vs standard hydrogen electrode, a time scale and efficiency suitable for postharvest fumigations. The cathode exhibited stable performance over 50 CH3Br capture and destruction cycles. Initial cost estimates indicate that this technique could treat CH3Br fumes at ∼$5/kg, roughly one-third of the cost of current alternatives.

■

captured by sorption to granular activated carbon (GAC),6,7 but the exhausted GAC would require disposal as a hazardous waste. Reduced sulfur species (e.g., thiosulfate) can degrade aqueous or GAC-adsorbed CH3Br, but this treatment requires the purchase of large quantities of reagent and can produce undesirable byproducts (e.g., methyl thiosulfate).8−11 As an alternative, electrochemical reduction has been demonstrated to achieve dehalogenation of an array of halogenated organic compounds.12−14 We demonstrated recently the potential for capture of vented CH3Br fumes by GAC, followed by submergence of the GAC bed in water and conversion of the

INTRODUCTION

Methyl bromide (CH3Br) has been widely used as a fumigant to control agricultural pests both for preharvest soil treatments and postharvest commodity treatments.1 However, the use of this effective fumigant has been restricted internationally via the Montreal Protocol in 1987 and in the United States via the Clean Air Act2 to limit its contribution to stratospheric ozone depletion.3 While the goal of these regulations has been to phase out CH3Br use, exemptions have continued for postharvest quarantine preshipment (QPS) applications due, in large part, to the slower kinetics featured by potential alternatives,4,5 resulting in difficulties meeting QPS-related logistical and regulatory requirements. Currently, CH3Br effluent from most postharvest chamber fumigations is vented to the atmosphere. To mitigate its contribution to ozone depletion, CH3Br fumes could be © 2016 American Chemical Society

Received: Revised: Accepted: Published: 11200

July 13, 2016 September 4, 2016 September 9, 2016 September 9, 2016 DOI: 10.1021/acs.est.6b03489 Environ. Sci. Technol. 2016, 50, 11200−11208

Article

Environmental Science & Technology

Preparation of Carbon-Based Working Electrodes. The GAC/graphite cathode was prepared as described in our previous study.15 Briefly, 1.0 g of GAC particles in a 25 mL glass vial were loaded with CH3Br by passing gas-phase CH3Br through a Teflon-lined septum into the vial using a syringe needle; a second needle permitted the gas to exit. After CH3Br sorbed to the GAC from the gas phase, the loading of CH3Br on the GAC was determined gravimetrically. The loose CH3Brloaded GAC particles were then transferred into a ∼1.5 × 2 cm cylinder fashioned from sheet graphite to aid the distribution of current to the GAC particles during electrolysis. For this study, the loose GAC particles were loaded with only CH3Br and treated by electrolysis one time; repeated cycles of loading and electrolysis were evaluated in our previous study.12 To enhance the conductivity of the cathode, cathodes consisting of GAC or activated charcoal bound by PVDF to carbon cloth as a current distributor were fabricated using the method described by Xie et al.18 Briefly, Fisher GAC particles were ground into various size ranges using a planetary ball mill (Nanjing T-Bota Scietech Instruments & Equipment Co., Ltd., Nanjing, China) at 400 rpm for 8−10 h. Inks were made by mixing these particles with PVDF in NMP solvent and stirring overnight. For the cathode used for most experiments, Fisher GAC particles were mixed with PVDF in a 92:8% ratio by weight. This mixture in turn was mixed with NMP in a 1:5 ratio by weight. A piece of carbon cloth (∼7 × 9.5 cm) was rolled into a cylindrical shape so that it could fit in the electrochemical chamber after its construction. A titanium wire was attached to the cathode to deliver the current. The cathode was coated by the GAC-containing ink and then dried in a vacuum for 2 days to remove the NMP solvent. This process was repeated again to reach a maximum GAC mass loading on the carbon cloth of ∼1.00 g or 15 mg GAC/cm2. The cathode was loaded with CH3Br from the gas phase in the same fashion as described for the GAC/graphite cathode, and the CH3Br loading was determined gravimetrically before the cathode was placed in the electrochemical cell. Note that this cathode construction procedure was conducted only once for each GAC/carbon cloth-based cathode before any exposure to CH3Br. Thereafter, the cathodes were reused over several cycles of gas-phase CH3Br capture and electrolytic destruction. Between each use, the cathode was extracted with ethyl acetate to measure residual CH3Br, rinsed three times with deionized water, and oven-dried for 2 h at 105 °C to prepare it to capture the next batch of gasphase CH3Br, unless otherwise noted. Electrolytic Degradation of CH3Br. Electrolysis was conducted as described in our previous study.15 Briefly, the electrochemical cell consisted of two 150 mL glass chambers connected by a 4 cm diameter glass tube. Each chamber featured a port capped with a PTFE-lined septum to enable sampling during the experiment. Cathodic and anodic chambers were separated by a cation exchange membrane (Ultrex CMI-7000, Membranes International, Ringwood, NJ) and filled with deionized water buffered at pH 7 with 100 mM phosphate buffer unless otherwise specified. A Ag/AgCl (1 M KCl) reference electrode (CHI111, porous Teflon tip, CH Instruments, Austin, TX) was placed within 0.5 cm of the working cathode, while a platinum wire counter electrode (CH Instruments) was used in the anodic chamber. The working electrode was transferred into the cathodic chamber immediately after CH3Br treatment and connected to a CH-600D potentiostat (CH Instruments) via the titanium wire. Both chambers were capped with minimal headspace (∼3 mL) and

GAC into a cathode within an electrolysis cell to reductively debrominate the sorbed CH3Br.15 The GAC bed could be drained, dried, and then reused to capture CH3Br from the next fumigation cycle. When an electric potential of −0.77 V vs standard hydrogen electrode (SHE) was applied to the GAC cathode, >80% degradation of CH3Br was achieved over ∼30 h with quantitative yield of bromide. We also demonstrated that a GAC-based cathode was able to capture and reductively dehalogenate a range of chlorinated and brominated disinfection byproducts (including chloropicrin) in reclaimed wastewater,16 suggesting that this scheme could also apply to the capture and destruction of alternative halogenated fumigants. While this previous work demonstrated proof of concept, the electrodes were not suitable for full-scale application. The cathode consisted of GAC particles wrapped in sheet graphite, while the anode was a platinum wire. In addition, the ∼30 h time scale of electrolytic degradation is too slow relative to the ∼24 h interval between consecutive CH 3 Br chamber fumigations typically encountered at ports. Hypothesizing that the slow kinetics were attributable predominantly to the resistance between the GAC particles, a primary goal of the current study was to develop and test a new electrode in which a thin layer of GAC particles is bound to a conductive material to distribute the current. Although alternative conductive carbon-based materials (e.g., graphite) have been used to construct electrodes for other purposes, these materials feature low specific surface areas.17 The high specific surface area of GAC is needed to efficiently capture and sorb CH3Br due to its high volatility. In addition to evaluating a range of materials and designs for constructing the electrode, we assessed the effect of reaction conditions on CH3Br debromination rates. Moreover, we tested alternative designs for the anode because of the cost of the platinum anode used in our earlier study. Finally, we assessed the ability of the GAC-based electrode to maintain performance over multiple cycles.

■

MATERIALS AND METHODS Materials. Alfa Aesar (Ward Hill, MA) sheet graphite (0.13 mm thickness, catalog number 43078) and carbon felt (3.18 mm thickness, catalog 50346147), Fuel Cell Earth (Stoneham, MA) carbon cloth (0.38 mm thickness, 45 × 45 yarns/in, carbon content 99%), Fisher (Pittsburgh, PA) binder-free glass microfiber filters and HPLC grade ethyl acetate, and SigmaAldrich (St. Louis, MO) polyvinylidene fluoride (PVDF), Nmethyl-2-pyrrolidone (NMP), and 1,2-dibromopropane (DBP) were used as received. Deionized water (18 MΩ) was produced with a Millipore Elix 10/Gradient A10 water purification system. Other chemicals were all reagent grade. Fisher GAC (6−14 mesh), Cabot Norit (Alpharetta, GA) Hydrodarco 3000 GAC (30−60 mesh), and Sigma-Aldrich activated charcoal particles (100 mesh) were rinsed three times with deionized water and oven-dried overnight before use or further treatment. A 1 lb cylinder of compressed CH3Br, Meth-O-Gas 100 (>99.5% purity), was obtained from Cardinal Professional Products (Woodland, CA). Methyl bromide standards were generated by purging gas phase CH3Br into a glass vial placed in an ice bath to form liquid CH3Br. Then, an aliquot of liquid CH3Br was transferred to a 1.5 mL glass vial and weighed to determine the moles of CH3Br. A liquid stock solution formed by diluting the CH3Br with ethyl acetate was stable for ∼6 months when stored at −20 °C and was used to construct CH3Br standard curves. 11201

DOI: 10.1021/acs.est.6b03489 Environ. Sci. Technol. 2016, 50, 11200−11208

Article

Environmental Science & Technology

Figure 1. Comparison of electrolytic degradation of CH3Br by the GAC/graphite (graphite) and GAC/carbon cloth (CC) cathodes at −1 V at pH 7 and 100 mM phosphate buffer in terms of (A) current and (B) total cell voltage/potential and resistance. (C) SEM image of the GAC/carbon cloth cathode (15 mg GAC/cm2). (D) CH3Br decay and bromide formation over time for electrolysis using the GAC/carbon cloth cathode. Total bromine = [CH3Br] + [Br−].

to measure the loss of CH3Br over various times. Open circuit control experiments were conducted, but no loss of CH3Br or production of Br− were observed when no voltage was applied. Analytical Methods. CH3Br in ethyl acetate extracts was analyzed by gas chromatography (Varian 3800 system) with electron capture detection at 280 °C using an Agilent GCGASPRO column (30 m × 0.32 mm). The column temperature program was 70 °C for 1 min, ramping to 180 °C at 25 °C/min, and then ramping to 250 °C at 40 °C/min and holding for 1 min. 1,2-Dibromopropane injected into the ethyl acetate was used as an internal standard. Bromide ion was analyzed using a Dionex ICS-1000 ion chromatography system equipped with an Ion Pac-AS 11-HC column using a 50 mM NaOH/deioned water (50:50) eluent at a 1 mL/min flow rate with a conductivity detector. Particle sizes of GACs were determined by scanning electron microscopy (SEM). Carbon surface areas were determined by BET analysis using N2 sorption.

stirred after the experiment was started. The potentiostat was operated in bulk electrolysis mode with a constant voltage of −1.0 V vs SHE applied to the working cathode unless otherwise specified. Hereafter, all voltages are referenced to SHE. The total cell voltage/potential between the cathode and anode was measured by a multimeter. The cathodic chamber was sampled for bromide at various times during the electrolysis by withdrawing ∼0.6 mL liquid via the sampling port using a glass syringe. To measure residual CH3Br, the electrolysis was halted. The CH3Br in the supernatant was measured by extracting 10 mL of supernatant with 3 mL of ethyl acetate for 2 min. The residual CH3Br on the carbons was extracted using ethyl acetate for 5 min; initial experiments demonstrated that this extraction time was sufficient for CH3Br extraction. For the GAC/graphite cathode, the GAC particles were removed from the sheet graphite tube, filtered onto glass fiber filters, and then immediately extracted using ethyl acetate at a ratio of 0.5 g GAC/10 mL of ethyl acetate. For the carbon cloth electrode, the entire electrode was submerged in 35 mL of ethyl acetate for extraction. Because this extraction was sacrificial, separate experiments were needed 11202

DOI: 10.1021/acs.est.6b03489 Environ. Sci. Technol. 2016, 50, 11200−11208

Article

Environmental Science & Technology

■

RESULTS AND DISCUSSION

The Coulombic efficiency (CE), a measure of the percentage of electrons supplied that participate in the desired reaction, was evaluated using eq 1:

GAC/Graphite vs GAC/Carbon Cloth-Based Cathodes. The performance of the GAC/graphite cathode was similar to that observed in our previous study.15 The CH3Br loading on the GAC particles ranged from ∼30−40% by weight after purging with gaseous CH3Br, and >80% degradation was attained after 30 h of electrolysis at −1 V with nearly 100% molar yield of bromide relative to CH3Br destroyed. The current dropped to ∼8 mA within 5 min and remained constant thereafter (Figure 1A). Measuring the resistance associated with the junction between the GAC and the current distributor can be tricky, particularly for the GAC/carbon cloth electrode, where a microprobe would be necessary to ensure contact of the probe with only the small GAC particles. Instead, we compared the effects of different cathode designs on the total cell resistance. For the GAC/graphite cathode, the total cell resistance was calculated to be 391 Ω by dividing the 4.3 V total cell voltage/potential measured at the start of the electrolysis by the current measured by the potentiostat (Figure 1B). Over the first 2 h of electrolysis, the current declined slightly more than the total cell voltage/potential, leading to a small increase in resistance (∼450 Ω). Hypothesizing that the slow kinetics for CH3Br destruction with the GAC/graphite cathode resulted primarily from high resistance between the loose GAC particles, we constructed a cathode in which a thin layer of crushed Fisher GAC particles was bound by PVDF to carbon cloth as a conductive current distributor. While crushing the GAC particles might have altered their electrocatalytic activity, we posited that a thin layer of GAC particles directly bound to the carbon cloth would minimize any resistance between GAC particles. Compared to sheet graphite (0.007 m2/g), carbon cloth has a higher specific surface area (1.61 m2/g), enabling higher GAC loadings. Initial tests demonstrated negligible sorption of CH3Br by sheet graphite or carbon cloth alone, indicating the need to incorporate GAC with its even higher specific surface area (1088 m2/g for Fisher GAC19) to effectively capture CH3Br. PVDF was used as a binding agent because it is poorly soluble in water and features low chemical reactivity. SEM images of the GAC/carbon cloth electrode indicates a submonolayer coating of GAC particles on the carbon cloth with GAC particle diameters of ∼0.1−10 μm (Figure 1C), far smaller than the 2− 5 mm diameter GAC particles in the GAC/graphite cathode. The GAC/carbon cloth electrode could achieve ∼30% CH3Br loading by weight relative to the GAC content of the electrode. The slightly lower sorption capacity compared to the GAC/ graphite electrode may result from the partial blockage of GAC pores by the PVDF binder. The electrolytic debromination of CH3Br with the GAC/ carbon cloth cathode exhibited biphasic kinetics (Supporting Information, Figure S1). Approximately 80% removal occurred over the first 2 h (pseudo-first order degradation rate (kobs) = 0.77 h−1), but then the degradation rate slowed (kobs = 0.13 h−1), achieving 96% overall removal after 15 h. The apparent bromide molar yield was only ∼70% relative to CH3Br loss over the first 2 h but approached 100% after 10 h (Figure 1D). This incomplete yield over the first 2 h is an apparent yield based on aqueous phase bromide measurements and may result from poor mass transfer of bromide from the GAC, which is further evaluated below.

CE =

nFNCH3Br Q

(1)

where n is the number of electrons transferred per reaction, F is the Faraday constant (96485 C mol−1e), NCH3Br is the moles of CH3Br degraded, and Q is the cumulative Coulombs of charge transferred. For the GAC/carbon cloth electrode, the Coulombic efficiency was ∼80% initially (Supporting Information, Figure S2). This efficiency was calculated assuming a 1 electron transfer (eq 2), similar to our previous study.16 CH3Br + e− → CH3• + Br

(2)

During the early stages of the electrolysis, the ∼30% loading of CH3Br by weight on the GAC would be equivalent to molar concentrations if the CH3Br were dissolved in the aqueous phase. Thus, sorbed CH3Br can outcompete other reagents (e.g., bubbles were observed without CH3Br, likely due to proton reduction to hydrogen gas) for the consumption of electrons. The Coulombic efficiency declines as the electrolysis continues because the CH3Br is depleted. After ∼2 h, the Coulombic efficiency was ∼45%. The initial electrolytic current of the GAC/carbon cloth electrode (∼100 mA) was an order of magnitude greater than that of the GAC/graphite electrode (Figure 1A). With a total cell voltage/potential of 13 V, the initial resistance of the GAC/ carbon cloth electrode was 138 Ω (Figure 1B). The ∼1000-fold smaller GAC particle diameters and close binding of each GAC particle to the conductive carbon cloth current distributor likely explains this lower resistance, in turn promoting a higher current and faster CH3Br destruction for the GAC/carbon cloth electrode. The high initial current could also be associated with a capacitive current for initial charging of the electrodes; there was no significant difference in current profiles with and without CH3Br. However, the capacitive current likely was short-lived (e.g., the rapid decline in current observed over the first minute for the GAC/graphite electrode). The high Coulombic efficiency (∼80% over the first 1 h for the GAC/ carbon cloth electrode) demonstrates that current resulted in reaction; in the absence of CH3Br, current likely resulted in reduction of other targets (e.g., protons or dissolved oxygen). The current declined with destruction of CH3Br and eventually converged to the current observed for the GAC/graphite cathode after ∼8 h (Figure 1A). The residual current likely is associated with proton reduction, for which carbon electrodes feature a high overpotential.17 After 2 h, the current had decreased, but the total cell voltage/potential had declined to 7.4 V such that the resistance increased to ∼210 Ω, which is still much lower than the resistance of the GAC/graphite cathode. Variations in Electrode Designs. We evaluated the effect of five aspects of the electrode design with respect to CH3Br destruction efficiency. First, we compared carbon cloth-based cathodes constructed using Fisher GAC to cathodes constructed using Norit GAC and Sigma-Aldrich activated charcoal. These materials featured a range of specific surface areas (1088 m2/g for Fisher GAC, 650 m2/g for Norit GAC, and 600 m2/g for the activated charcoal)16 and conductivities (0.92 S/mm for Fisher GAC and 73 S/mm for the activated charcoal, not available for the Norit GAC16). The carbons were 11203

DOI: 10.1021/acs.est.6b03489 Environ. Sci. Technol. 2016, 50, 11200−11208

Article

Environmental Science & Technology

Figure 2. (A) SEM image of a GAC/carbon cloth electrode fabricated with Fisher GAC (100−250 μm particle diameters). (B) CH3Br loss and (C) bromide yield over 2 h of electrolysis at −1 V and pH 7 with 100 mM phosphate with electrodes featuring Fisher GAC particle diameters of