Environ. Sci. Technol. 2001, 35, 406-410
Electrochemical Treatment of 2,4,6-Trinitrotoluene and Related Compounds JAMES D. RODGERS AND NIGEL J. BUNCE* School of Engineering and Department of Chemistry and Biochemistry, University of Guelph, Guelph, Ontario, Canada N1G 2W1
This work involves electrolysis of nitrotoluene congeners, which are persistent pollutants that enter the environment as a consequence of their manufacture and use as explosives. Reduction to aminotoluenes occurred with high current efficiency at a variety of cathodes, at potentials -0.5 to -1 V vs SCE. The products were formed in high chemical yield and with excellent mass balance. Preliminary experiments were also carried out to find methods of removing the electrolysis products from solution by oxidative oligomerization. The most satisfactory method was partial reoxidation at a Ti/IrO2 anode, suggesting an overall remediation technology in which reduction is followed by reoxidation of the spent catholyte in the anode compartment of the same electrolytic cell.
Introduction The large scale manufacture and use of nitroaromatic compounds such as 2,4,6-trinitrotoluene (TNT), 1,3,5-trinitrobenzenene, dinitrotoluenes (2,4-DNT, 2,6-DNT), tetryl, and picric acid as explosives has led to significant contamination of soils (1) and groundwater. The U.S. ceased production of TNT in the mid-1980s, but environmental contamination has continued because of the open detonation and burning of explosives at army depots and artillery ranges and demilitarization at ordnance disposal sites (2). Nitroaromatic explosives are toxic: for example, the 14-day LC50 values toward the guppy (Poecilia reticulzata) are 12.5 mg L-1 for 2,4-DNT and 18 mg L-1 for 2,6-DNT (3). TNT urinary metabolites are mutagenic, but the major documented effects on exposed humans are hemolysis, hepatotoxicity, and changes in hepatic enzyme levels (4). Nitrotoluenes are recalcitrant toward chemical or biological oxidation and hydrolysis (5). Biological treatment of nitrotoluenes is limited by their toxicity at high concentrations to microorganisms and sometimes produces recalcitrant or toxic byproducts (6). Chemical treatments may be energy intensive (e.g., rotary kiln incineration), be ineffective at low concentrations, or may cause other environmental problems such as NOx emissions. Electrolysis, the use of electrical energy to drive an otherwise unfavorable chemical reaction, is a developing technology for environmental remediation. Recent examples of its use include treatment of chromium (7), cyanides (8), ethanol (9), EDTA (10), dye wastes (11), methanol (12), and landfill leachates (13). Attractions of electrochemical technology include the low costs of electricity compared with * Corresponding author phone: (519)824-4120 ext. 3962; fax: (519)766-1499; e-mail:
[email protected]. 406
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 2, 2001
chemicals and the “green” issue of using only electrons as reagents (14). Electrochemical technologies also offer the prospect of relatively low capital cost, modular design, operation at atmospheric pressure and temperature, and the possibility of higher energy efficiency than thermal treatment or photolysis. However, a significant drawback is that parasitic reactions, such as electrolysis of water, often compete with electrolysis of the contaminant(s) and lower energy efficiency. The electrochemical reduction of nitroaromatic compounds has long been known (15). Nitrobenzene undergoes a four-electron reduction to phenylhydroxylamine in aqueous solution at pH > 4 and a six-electron reduction to aniline at lower pH. Reduction takes place in several steps involving sequential electrochemical and chemical reactions (16-18). An important consideration in the present context is that reduction occurs at low negative potentials, at which parasitic electrolysis of water is minimized.
Experimental Section 2,4-Dinitrotoluene (2,4-DNT), 2,6-dinitrotoluene (2,6-DNT), and their reduction products 2-amino-4-nitrotoluene (2A4NT), 4-amino-2-nitrotoluene (4A2NT), 2,4-diaminotoluene (2,4DAT), and 2,6-diaminotoluene (2,6-DAT) were obtained from Aldrich; 2,4,6-trinitrotoluene, 2-amino-4,6-dinitrotoluene (2ADNT), 4-amino-2,6-dinitrotoluene (4-ADNT), 2,4-diamino6-nitrotoluene (2,4-DANT), and 2,6-diamino-4-nitrotoluene (2,6-DANT) were obtained from Chemservice. 2,4,6-Triaminotoluene (TAT) was a gift from Dr. J. Hawari of the National Research Council of Canada. Purities of all compounds were verified by HPLC to be g 98%. Horseradish peroxidases (HRP: Sigma) consisted of a Type I (EC 1.11.1.7, RZ 1.3) and a Type II (EC 1.11.1.7, RZ 1.9) with activities of 220 and 240 units mg-1, respectively. One unit of activity is defined as the number of µmol of hydrogen peroxide converted per minute at pH 7.4 at 25 °C. Voltammetry. Reduction potentials were determined in N2-purged solutions at a dropping mercury electrode using a Radiometer polarographic analyzer with a platinum bulb as a counter electrode and a saturated calomel reference electrode. Conditions for differential pulse polarography (DPP) were as follows: purge time 1000 s, initial potential ) 0 mV, final potential ) -1200 mV, step duration ) 0.4 s, step amplitude ) 1 mV, pulse duration ) 40 ms, and pulse amplitude ) 25 mV. Electrolysis. Electrolyses were carried out amperostatically or potentiostatically, controlled by a Pine RDE4 potentiostat with flow-through, sandwich type cells constructed from Plexiglas pieces with dimensions of 90 × 15 × 45 mm. The center of each Plexiglas piece was bored out to provide 2.5 mL anode and cathode compartments, which were separated by a Nafion 417 cation exchange membrane. Anolyte and catholyte solutions were delivered from 60 mL glass reservoirs and were pumped at 7 mL min-1 upward through the cell (to facilitate removal of gas bubbles) using a peristaltic pump. Connection to the pump head was by means of LFL-Tygon; Teflon tubing was used for connections between the pump and the reservoirs. A three-electrode system was used, with electrode potentials measured vs external Ag/AgCl. Anodes (counter electrodes) were either Pt foil or IrO2/Ti (10 cm2). Cathodes (10 cm2) were constructed from solid Pb, Ni-plated Ni wire (0.5 mm diameter), Raney nickel, Pt, or reticulated carbon (RTC). Pb and Ni were obtained from Aldrich, RTC from ERG Materials. The Ni-plated cathode was prepared by electrodepositing a 1.4 M NiSO4/NiCl2 solution for 2 h at 15 mA cm-2 onto Ni wire. The Raney Ni cathode was prepared by 10.1021/es001465s CCC: $20.00
2001 American Chemical Society Published on Web 12/09/2000
TABLE 1. Comparison of Nitroaromatic Potential Peak Values at a DME under Different Conditions (vs SCE) substrate
Hetmana (23)
Pearsonb (22)
this workc
TNT 2,4-DNT 2,6-DNT
-0.35, -0.55, -0.9 V -0.68, -0.88 V N/A
-0.10, -0.20, -0.31 V -0.18, -0.30 V -0.24, -0.40 V
-0.11, -0.23, -0.34 V -0.10, - 0.19 V -0.10, -0.18 V
a 30 mL of pyridine, 7 mL of 1 M KNO , 35 mL of 2 M NH NO , 28 mL of H O, pH 7.6. 3 3 3 2 - HCl, pH 2.5. c 0.05 M phosphate at pH 2.
moistening a 4:1 mixture of Raney Ni with nickel powder and compressing the mixture at 10 000 psi followed by leaching with 3 M NaOH for 1 h at 70-80 °C, cf. VelinPrikidanovics and Lessard (19). The latter authors used a similar procedure except that they leached for 8 h. The supporting electrolyte was 0.05-0.1 M disodium hydrogen phosphate adjusted to pH 2. Oxidation experiments employed an IrO2/Ti anode and Pb cathode. Chromatography. Organic compounds were analyzed by HPLC, consisting of a Waters model 501 pump (flow rate 0.5-1.0 mL/min), Rheodyne injector containing a 20 µL sample loop, and Waters model 440 UV absorbance detector operated at 254 nm. The detector output was processed with Peaksimple software. Separation was achieved on a reverse phase C18 (25 cm × 4.6 mm) stainless steel column, using 40:60 to 70:30 (v/v) MeOH:sodium phosphate buffer (pH 2) as the mobile phase. All solvents used were Fisher Scientific HPLC grade. A Hewlett-Packard Model 5890 Series II gas chromatograph with a Hewlett-Packard Model 5971 mass selective detector was used to identify volatile products. Separation was achieved with a 30 m DB5MS column (i.d. 0.25 mm) using He as carrier gas. Injector and detector temperatures were 225 °C and 250 °C, respectively. The oven temperature program was 60 °C, ramped at 7 °C/min for 15 min, at 15 °C/min for 9 min, and then held constant at 300 °C for 6 min. Sample introduction was by solid-phase microextraction (SPME) using a 100 µm poly(dimethylsiloxane)-coated fiber, which was conditioned at 200 °C prior to use (20). Extraction consisted of immersing the fiber for 30 min in a stirred 2 mL aliquot of the aqueous sample, which was adjusted to pH 6-8 with KOH to protect the fiber and to deprotonate products containing NH2 groups. Nonvolatile electrolysis products were identified using either probe mass spectrometry with electrospray ionization or by LC/MS, with separation effected on a 10 cm C18 column (mobile phase 40% H2O: 60% CH3CN, flow rate 0.7 mL/ min). In these cases, electrolysis employed ammonium acetate as the supporting electrolyte. Samples were run in APCI [(M + 1) and (M - 1)] mode and then subjected to MS/MS for positive identification. Peroxidase Oxidation. Enzymatic oxidation employed a 25 mL batch reactor containing 10 mL of HRP solution (0.2 mg/mL: ∼50 units/mL) prepared in 0.05 M sodium phosphate buffer (pH 7) and 1 mL of the aminotoluene derivative (36 mM for 2,6-DAT and 2,4-DAT, 5 mM for TNT reduction products). To this solution was added H2O2 (1 mL of a 36 mM solution for DAT, 5 mM for TNT reduction products) prepared in the same buffer. Aliquots were taken at specific time intervals, filtered through a 0.45 µm filter, quenched by the addition of a few drops of concentrated sulfuric acid, and analyzed by HPLC for reactant loss.
Results and Discussion Differential pulse polarography was used to determine the reduction potentials of the target nitro compounds under our experimental conditions (pH 2, chosen because explosives wastes are generally acidic due to the method of preparation). Each nitrotoluene exhibited the same number of polarographic peaks as the number of nitro groups in the
b
5% ethanol, 95% 0.01 M potassium hydrogen phthalate
FIGURE 1. Reduction of 2 mM 2,6-DNT as a function of electrode material at current density 10 mA cm-2.
TABLE 2. Current Efficiency (%) for the Reduction of a 2.5 mM 2,4-DNT Solution over 1 h of Electrolysis at a Ni Plated Ni Cathode current, mA cm2
efficiency, %
current, mA cm2
efficiency, %
1.3 6.3 12.5
69.7 45.1 23.1
18.8 25.0
16.2 13.0
molecule, and since each peak was shown to be associated with reduction by ∼6 electrons, we concluded that each nitro group was reduced completely before the reduction of another nitro group commenced. Peak potentials for the reduction of nitroaromatic compounds are strongly influenced by the experimental conditions, especially pH, because reduction is a proton-consuming reaction. This is shown by the comparative data in Table 1.
ArNO2 + 7H+ + 6e- f ArNH3+ + 2H2O 2,6-Dinitrotoluene was the first target for electrolysis, because the symmetry of the molecule limits the number of reduction products. Figure 1 shows the results of electrolysis of 2,6-DNT at different cathodes using a current density of 10 mA cm-2. The reactivity order was Ni-plated Ni wire > Pb > Raney Ni > reticulated carbon (RTC) > Pt. The data for Pt and RTC are plotted on a linear scale (left-hand ordinate); all others are fitted logarithmically, consistent with pseudofirst-order kinetics (right-hand ordinate). The poor performance of Pt is consistent with the lack of response in voltammetric experiments and is consistent with previous studies (21). In the case of Raney Ni, which functions as a reductant by means of electrocatalytic hydrogenation, its inability to reduce DNT may be because the cell potential was insufficiently negative, even though aromatic compounds reduce more easily than water (22). Current efficiency is the ratio of electrons used in the desired transformation of the analyte to the total number of electrons passed. Allowing for the number of electrons (six) to reduce NO2 to NH2, current efficiency for 2,6-DNT ranged from
