Dechlorination of Polychlorinated Biphenyls in ... - ACS Publications

Nov 11, 2003 - JOSEPH SILVERMAN, AND. MOHAMAD AL-SHEIKHLY*. Department of Materials and Nuclear Engineering,. University of Maryland, College ...
0 downloads 0 Views 91KB Size
Environ. Sci. Technol. 2003, 37, 5773-5777

Dechlorination of Polychlorinated Biphenyls in Industrial Transformer Oil by Radiolytic and Photolytic Methods CYNTHIA G. JONES,† JOSEPH SILVERMAN, AND MOHAMAD AL-SHEIKHLY* Department of Materials and Nuclear Engineering, University of Maryland, College Park, Maryland 20742 PEDATSUR NETA AND DIANNE L. POSTER Chemical Science and Technology Laboratory, National Institute of Standards and Technology, Gaithersburg, Maryland 20899

Used electrical transformer oils containing low or high concentrations of polychlorinated biphenyls (PCBs) were treated using electron, γ, and ultraviolet radiation, and the conditions for complete dechlorination were developed. Dechlorination was determined by analysis of the inorganic chloride formed and the concentrations of remaining PCBs.Transformer oil containing ≈95 µg g-1 PCB (≈3.5 mmol L-1 Cl) is completely dechlorinated by irradiation with 600 kGy after the addition of 10% triethylamine (TEA). Transformer oil containing >800 000 µg g-1 PCB (17.7 mol L-1 Cl) requires an additional solvent to prevent solidification. When this oil is diluted with 2-propanol (2-PrOH) and TEA (v/v/v, 1/79/20), complete dechlorination is achieved with a dose of 2500 kGy. Ultraviolet photolysis of the same oil/2-PrOH/TEA solutions led to 90% dechlorination after exposure for 120 h in our experimental setup. Such yields were obtained by radiolysis with a dose of 2000 kGy (300 h in our Gammacell). Replacing TEA with KOH in 2-PrOH solutions greatly increases the yield of dechlorination in both the radiolytic and the photolytic experiments, demonstrating that a chain reaction plays a role in both of these treatment methods and suggesting that both methods deserve further consideration for large-scale application.

esolv- + ArCl f Ar• + Cl-

(1)

However, the oil contain various aromatic hydrocarbons, such as biphenyl (Ph2), that can react with the electrons to form radical anions:

esolv- + Ph2 f Ph2•-

(2)

PCBs are dechlorinated also by electron transfer from the aromatic radical anions:

Ph2•- + ArCl f Ph2 + Ar• + Cl-

Introduction Although the use of polychlorinated biphenyls (PCBs) has been tightly restricted in the United States since 1977, PCBs still remain in use in a variety of industrial and commercial applications when circumstances permit. The concern for environmental contamination of PCB-laden transformer oil today is that of overheating of electrical components, such as electrical transformer or capacitor oil containing PCBs that can produce emissions of vapors or expansion of the oil, leading to small spills, leaks, or airborne releases into the environment. In addition, accidental spills are a problem as well as possibly illegal disposal. Under current regulations, equipment that contains PCBs must be specifically identified. * Corresponding author telephone: (301)405-5214; fax: (301)3149467; e-mail: [email protected]. † Permanent address: U.S. Nuclear Regulatory Commission, Mail Stop O16-C1, Washington, DC 20555. 10.1021/es030412i CCC: $25.00 Published on Web 11/11/2003

Transformers and capacitors, the largest reservoirs of PCBs still in use today, are included in this category. It is estimated that about 125 million transformers containing PCBs were in use as of 1999, based on required user information data compiled by the United States Environmental Protection Agency (EPA) Office of Pollution Prevention and Toxics (1). Although there were, on average, about 25 000 high-level concentration (>500 µg g-1) PCB transformers disposed of per year from 1990 to 1994 (2), the current U.S. inventory has remained essentially constant since 1988. This is primarily due to three factors. First, there are only four incinerator disposal facilities for high-concentration PCB-laden oils left in the United States. Second, the disposal costs are continuing to increase as access to facilities becomes more limited. Last, there are no regulatory requirements or incentives for early (i.e., before end-of-life) PCB disposal. As a national strategy, the EPA seeks a 90% reduction nationally of the high-level PCBs (i.e., >500 µg g-1) used in electrical equipment by the year 2006 (3). Radiation processing and treatment of materials is a mature industry, which has been successfully used in the degradation of many chemical toxins and unwanted byproducts (4-8). In the present study, we examine the radiolytic degradation of PCBs in two electrical transformer oils containing ≈95 µg g-1 and >800 000 µg g-1 PCB, which were donated by the Constellation Energy Group (CEG) (9). While earlier studies examined radiolysis of oil containing low concentrations of PCBs (10), this is the first study that is aimed at achieving the complete (>99.9%) dechlorination of used transformer oil containing high concentrations of PCBs and that compares radiolytic with photolytic dechlorination methods. In previous studies, we demonstrated the complete radiolytic degradation of 2,2′,6,6′-tetrachlorobiphenyl (PCB54) in transformer oil (the PCB congener was added to clean transformer oil in the laboratory) and discussed the mechanism for this process under various conditions (11, 12). PCBs (ArCl) are dechlorinated by reacting with solvated electrons:

 2003 American Chemical Society

(3)

These reactions explain why PCBs can be dechlorinated despite the presence of aromatic hydrocarbons in the oil and despite the formation of biphenyl as a radiolysis product that reacts rapidly with solvated electrons. In a recent paper, we reported preliminary results on the radiolytic dechlorination of PCBs in a commercial contaminated oil (13). The complete results, demonstrating quantitative dechlorination of a PCBladen oil, are presented in this paper along with a comparison of the radiolytic yields obtained using various additives. Moreover, we compare the radiolytic with a photolytic treatment method and suggest the latter as a viable alternative.

Experimental Section Materials. Samples of PCB-laden transformer oils containing ≈95 µg g-1 and >800 000 µg g-1 PCBs were donated by CEG. These two oils were selected to be representative of typical VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5773

transformer oils: (i) at concentrations that have been typically treated with ionizing irradiation in previous studies (≈95 µg g-1) and (ii) at extremely high concentrations that would currently require special treatment (i.e., incineration) for disposal (14). Gas chromatographic analysis of the low- and high-level PCB-laden transformer oils closely matched the PCB congener patterns of Aroclors 1260 and 1016, respectively, which are typical of transformer oils (15). Because the composition of the high level PCB-laden transformer oil was given as >800 000 µg g-1, 1 mL of the sample (F ) 1.342 ( 0.002 g mL-1) was mixed with 24 mL of clean Diala AX transformer oil (F ) 0.885 ( 0.002 g mL-1) and analyzed at the Materials and Chemistry Laboratory (Oak Ridge, TN) for the total chlorine content. The sample was found to contain Cl at a concentration of 27.8 ( 0.6 mg g-1. After correcting for the dilution factor and for the densities of the sample and the clean oil, the concentration of Cl in the original sample is 467 ( 10 mg g-1 or 17.7 ( 0.5 mol L-1. Since Aroclor 1016 contains total Cl at a concentration of ≈410 mg g-1 (16), the oil is likely a mixture of Aroclors, such as 1016 and possibly 1242, which has a similar PCB congener pattern (16). Irradiation. The transformer oils were irradiated in the presence of various additives (2-PrOH, TEA, KOH). For γ-radiolysis and electron beam irradiation with doses 4000 kGy, a container with a coldfinger (16) was used to control the temperature and to prevent sample loss by boiling or evaporation. All solutions were in equilibrium with atmospheric pressure and temperature before irradiation and were sealed to prevent rapid diffusion of air. Oxygen present in solution is consumed in the early stages of irradiation. γ irradiations were carried out with either the NIST Gammacell 220 60Co source, with a dose rate of 6.40 kGy h-1, or at the University of Maryland Gamma Laboratory, with a dose rate of 6.84 kGy h-1. Electron beam irradiations were carried out with a Varian linear accelerator at the University of Maryland, using 3-µs pulses of 7 MeV electrons. To prevent overheating, the solutions were irradiated with sequences of 10 000 pulses at a time, followed by a 2-min cooling period. Dosimetry of the electron beam with radiochromic film dosimeters (17) showed an absorbed dose difference of about 10% between the front and the back of the 5-mL cylindrical irradiation vial. Fricke dosimetry (air-saturated 10-3 mol L-1 FeSO4 in 0.4 mol L-1 H2SO4), taking G(Fe3+) ) 1.54 µmol J-1, gave an absorbed dose of 10 Gy/pulse. For the photolysis experiments, solutions (5-mL aliquots) were placed in Pyrex (cutoff ≈ 300 nm) or quartz (cutoff < 200 nm) cylindrical cells and irradiated with a 300-W UV Xenon arc lamp (ILC Technology LX 300 UV) with continuous stirring. The output of the lamp is nearly flat throughout the visible range, and it drops in the UV range to about 50% at 300 nm and to about 10% at 220 nm. The samples were placed 10 cm away from the front of the lamp, and no condensing lenses were used. Some of the samples were deoxygenated by bubbling with N2 before irradiation, while others were irradiated in sealed vials without bubbling. Analysis. After irradiation, the chloride ions were extracted from the organic medium into aqueous NaOH solutions, diluted to measurable concentrations, and analyzed by ion chromatography using a Dionex DX-500 apparatus (ED40 electrochemical detector, GP40 gradient pump, 25-µL injection loop) with a Dionex IonPac AS11 column (4 mm) and NaOH eluent (gradient method increase from 0.2 to 15 mmol L-1 over 15 min). Five-point calibration curves were generated before and periodically throughout the analyses period to verify the stability of the system. Duplicate or sometimes triplicate samples were prepared and analyzed for each sample. The combined standard uncertainties in the measurements of chloride ions concentrations by ion chroma5774

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003

FIGURE 1. Radiolytic dechlorination of PCBs in transformer oil containing ≈95 µg g-1 PCB and mixed with TEA (9/1, v/v), irradiated with γ (O) or electron beam (b). The dotted line indicates the level of complete dechlorination. The solid line is a fit to both sets of data. tography and in the reported radiolytic yields (G values) are estimated to be (5%. The concentrations of PCBs in the irradiated and unirradiated transformer oils were determined by capillary gas chromatography (GC) with electron capture detection (ECD) using a 5% phenyl methylpolysiloxane column (60 m × 0.25 mm; 0.25 µm film). Prior to GC-ECD analysis, gravimetrically prepared aliquots of ≈0.5 g from each oil sample (three for each dose) were processed through silica solid-phase extraction (SPE) cartridges (Waters Corp.) using hexane as the mobile phase. The collected eluants were reduced to ≈0.5 mL under nitrogen and fractionated on a semipreparative aminopropylsilane liquid chromatographic column to further isolate the PCBs from the transformer oil using hexane as the mobile phase. Collected fractions were concentrated to ≈0.5 mL under nitrogen, processed through aminopropyl SPE cartridges using hexane, and concentrated for duplicate GC-ECD analyses. Prior to processing, gravimetrically prepared aliquots of 2,2′,4,5′,6-pentachlorobiphenyl (PCB-103) and octachloronaphthalene (OCN) were added to each of the ≈95 µg g-1 PCBs in oil samples for use as internal standards. Similarly, hexachlorobenzene (HCB) and mirex (hexachloropentadiene dimer, C10Cl12) were added to each of the >800 000 µg g-1 PCBs in oil samples for the same purpose. Prior to irradiation, the oils were examined for the presence of PCB-103 as well as PCB congeners that may have possibly coeluted with the internal standards under the stated GC conditions. In addition, calibration standards of Aroclors 1260 and 1016 [NIST Standard Reference Materials (SRMs) 3080 (Aroclor 1260 in Transformer Oil) and 3075 (Aroclor 1016 in Transformer Oil)] were gravimetrically prepared (in triplicate) and processed alongside the oil samples to generate PCB response factors relative to the internal standards. Results are based on the areas of the dominant GC peaks and the two internal standard peaks. For example, for Aroclor 1260 the areas of 43 peaks were used. This approach is similar to EPA Method 505 (18), although with a larger number of peaks (19).

Results and Discussion Radiolysis Experiments. As in our earlier experiments (11), PCB-laden oils were irradiated in the presence of TEA. For the transformer oil containing ≈95 µg g-1 PCBs, addition of 10% (v/v) TEA was found to permit complete radiolytic dechlorination. The formation of inorganic chloride during both electron and γ irradiation (Figure 1) shows that essentially all (>99%) of the initial organic chlorine was converted to inorganic chloride ions after irradiation with

FIGURE 2. GC-ECD chromatograms (same scale, x (time) and y (ECD response), for all) of (A) Aroclor 1260 in transformer oil (as SRM 3080); (B) the unirradiated oil containing ≈95 µg g-1 PCB and 10% TEA; and (C) the same oil mixture irradiated with a dose of 600 kGy. Internal standards, added after irradiation, are indicated. absorbed doses of 600 kGy. The measured initial radiolytic yield for the dechlorination of this oil mixture is 0.02 µmol J-1. GC-ECD analysis of the unirradiated oil (Figure 2) confirmed the presence of Aroclor 1260 and indicated a PCB concentration of 91.6 ( 3.8 µg g-1, which is very close to CEG’s original estimate of 95 µg g-1. GC-ECD analysis of the irradiated (600 kGy) oil (Figure 2) led to an estimated 93% dechlorination of these oil samples, as compared to the ion chromatography results of 99%. The level of dechlorination as determined by GC-ECD represents a final calculated PCB concentration of 6.1 ( 1.0 µg g-1, well below the 50 µg g-1 threshold to declare this transformer oil as uncontaminated, non-PCB waste. In an attempt to achieve full radiolytic dechlorination of the concentrated PCB oil (>800 000 µg g-1), we mixed equal volumes of this oil with TEA and irradiated the mixture with an electron beam or in a γ source. Measurements of inorganic chloride production show efficient radiolytic dechlorination with an initial G value of 0.5 µmol J-1 for both types of irradiation. At doses near 4000 kGy, the mixture solidified due to massive formation of solid triethylammonium chloride. At this point the sample underwent about 16% dechlorination. Full dechlorination could not be reached because further irradiation of the solid mixture was ineffective for dechlorination. To achieve full radiolytic dechlorination it was necessary to dilute the PCB further in order to prevent solidification. Subsequent experiments were carried out with 2-PrOH solutions containing only 1% of the >800 000 µg g-1 PCB oil and 20% TEA (Figure 3a). In this case, essentially all of the original organic chlorine in the PCBs was converted to inorganic chloride by electron or γ irradiation at doses around 2500 kGy. There is 800 000 µg g-1 for the undiluted PCB oil. GC-ECD analysis (Figure 4) and chloride ion analysis (Figure 3a) of the samples irradiated with a dose of 2000 kGy indicate a radiolytic dechlorination of 99% and 93%, respectively, yielding a PCB concentration of 97.8 ( 6.6 µg g-1 remaining after irradiation. For the 3000 kGy irradiation, both analytical measurements show 100 (

FIGURE 3. Radiolytic dechlorination of PCBs in transformer oil containing >800 000 µg g-1 PCB diluted 1:100 with 2-PrOH solutions (a) containing 20 vol % TEA and (b) containing 1 mol L-1 KOH, irradiated with γ (b) or electron beam (O). The dotted lines indicate the level of complete dechlorination.

FIGURE 4. GC-ECD chromatograms (same scale, x (time) and y (ECD response), for all) of 2-PrOH solutions containing 20% TEA and 1% PCB-laden transformer oil (>800 000 µg g-1 PCB) (A) unirradiated; (B) γ irradiated with a dose of 3000 kGy; and (C) UV irradiated for 220 h. Internal standards, added after irradiation, are indicated. 1% dechlorination, resulting in a final calculated PCB concentration of e16.9 µg g-1, again well below the 50 µg g-1 threshold to declare this transformer oil as uncontaminated, non-PCB waste. Comparison of TEA and KOH as Additives. It is known that high concentrations of KOH in 2-PrOH lead to very high radiolytic dechlorination yields due to a chain reaction (2025). To compare the effects of KOH and TEA, we carried out radiolysis experiments using KOH (1 mol L-1) instead of TEA (20%, 1.4 mol L-1). As expected, very high radiolytic yields were achieved with KOH. The dose required for complete VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5775

FIGURE 5. Photolytic dechlorination of PCBs in transformer oil containing >800 000 µg g-1 PCB, diluted 1:100 with 2-PrOH solutions that contain various additives. (a) No additive (O); 20 vol % TEA [(in Pyrex cell initially under air (b); in quartz cell, deoxygenated (4)]. (b) 1 mol L-1 KOH [(in Pyrex, deoxygenated (2) or initially under air (b); in quartz, deoxygenated (4)]; additional experimental points (2, b) at 4 h show complete dechlorination. dechlorination of a 2-PrOH solution containing the same concentration of PCB (1% of the >800 000 µg g-1 sample, i.e., 0.177 mol L-1 Cl) was 50 kGy (Figure 3b) (i.e., 50 times smaller than the dose required in the oil/2-PrOH/TEA (1/79/20) experiment). Thus, KOH has a significant advantage over TEA as an additive when the PCBs are dissolved in 2-PrOH because of the chain reaction but does not have this advantage in MeOH or t-BuOH solutions, where the chain reaction is negligible (23). Photolysis Experiments. UV photolysis was also conducted to compare the applicability of this method to that of ionizing radiation for the dechlorination of PCB-laden transformer oil. 2-PrOH solutions containing 20% TEA and 1% PCB-laden oil (>800 000 µg g-1) were photolyzed with a Xenon lamp. The formation of inorganic chloride as a function of photolysis time (Figure 5a) indicates efficient dechlorination. A photolyzed sample that showed 92% dechlorination based on chloride yield indicated 95% dechlorination based on GC-ECD analysis of the oils (Figure 4C). Thus, UV irradiation can be as effective in dechlorinating PCBs as ionizing radiation. We assume that the dechlorination reactions are similar to those occurring under radiolysis, except that the initiation step is not the electron formed from the solvent but rather an electron transfer from a reductive quencher, such as TEA, to an excited aromatic molecule, either ArH or ArCl. Further details of the mechanism can be examined by utilizing pure compounds and specific wavelengths. In the present study, however, our aim is to demonstrate the feasibility of the photolytic process with commercial samples and standard equipment. With the photochemical setup and the γ source used in this study, the exposure time required to achieve complete dechlorination by UV photolysis was lower than that required for γ 5776

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003

radiolysis. This comparison was made with identical mixtures of PCB-laden oil with 2-PrOH and TEA. It is possible, however, that the conditions for photochemical treatment can be further optimized. Photolysis of PCBs in 2-PrOH solutions not containing TEA also led to dechlorination. The efficiency of dechlorination during the first several hours was about half that obtained in the presence of TEA. However, while complete dechlorination was achieved with TEA after extended photolysis, dechlorination in the absence of TEA stopped at a level of ≈25%. In contrast, photolysis of PCBs in 2-PrOH solutions in the presence of 1 mol L-1 KOH led to much faster dechlorination than in the presence of TEA (Figure 5b). Under our experimental conditions, the initial photochemical efficiency was an order of magnitude higher and the time required to achieve practically complete dechlorination was nearly 2 orders of magnitude lower with KOH than with TEA. Clearly, a chain reaction develops in the photochemical system as well. These photolysis experiments were carried out in glass containers sealed from the atmosphere. When the solutions were deoxygenated by bubbling with N2 prior to photolysis, the dechlorination reaction began immediately upon photolysis, and the yield of Cl- increased linearly with time of photolysis and then leveled off. When the solutions were not deoxygenated but were sealed, there was a lag time in the beginning of the dechlorination (Figure 5b) corresponding to the time required to consume the O2 present in solution. The results shown in Figure 5, panels a (b,O) and b (b,2), are for photolysis in Pyrex containers. Pyrex absorbs most of the light at wavelengths below 300 nm. The aromatic compounds in the oil, which are excited in these experiments, generally have stronger absorptions below 300 nm. The highpressure Xenon lamp used in this work has considerable light output at wavelengths below 300 nm, which is wasted when the samples are photolyzed in Pyrex containers. By using quartz containers, we can utilize the energy of the lower wavelengths. Indeed, we found that the initial photochemical yields in deoxygenated solutions were increased by a factor of 4-6, both in the TEA and the KOH solutions, and that the time required for complete dechlorination was decreased by using quartz instead of Pyrex irradiation cells. Thus, with the same oil/2-PrOH/TEA (1/79/20) mixtures, complete dechlorination is achieved after 120 h of photolysis under our experimental conditions and by radiolysis with 2000 kGy or 300 h in our γ source. This comparison clearly favors photolysis when considering the cost of the equipment and the associated safety requirements. The photolytic efficiency can be improved by a factor of 4-6 by replacing the Pyrex cell with a quartz cell (or by using a windowless configuration) to permit utilization of wavelengths below 300 nm, which are abundant in high-pressure xenon lamps (with quartz windows) and more effective for the excitation of aromatic compounds. Photolysis in 2-PrOH in the absence of TEA gives a lower yield of dechlorination, but this yield is sufficient to initiate a chain reaction in the presence of high KOH concentrations, leading to very high photochemical efficiencies. By comparing the initial dechlorination yields in 2-PrOH solutions in the presence and absence of KOH, we estimate a chain length of 450, similar to values reported in PCB/2-PrOH/KOH solutions treated with ionizing radiation (21-24). To compare the photolytic with the radiolytic efficiency, we take the nominal output of our xenon lamp and assume that all the energy is absorbed by the solution (although in practice only a fraction is collected and absorbed). The total UV output (from 200 to 390 nm) is reported by the manufacturer to be between 7 and 9 W. From the observed initial yield in a deoxygenated oil/2-PrOH/TEA mixture photolyzed in a quartz cell, we estimate the photochemical

efficiency as 8 × 10-9 mol J-1. This is approximately 30 times less than the observed radiolytic efficiency. This factor probably can be gained by the lower cost of equipment and operation of UV sources as compared with electron beam and γ sources. Thus, at least for dechlorination of PCBs in 2-PrOH solutions, photolytic treatment may be advantageous to radiolytic treatment. Before any such treatment method is applied, one must consider and examine whether any harmful products are formed as a side effect of this treatment. The balance of the chain reaction involving 2-PrOH and KOH, whether initiated by radiolysis or photolysis, is

ArCl + (CH3)2CHOH + KOH f ArH + (CH3)2CO + H2O + KCl (4) ArCl represents all PCB congeners that undergo stepwise dechlorination until they are converted to biphenyl. The chain reaction involves two radicals as the chain carriers, namely, Ar• and (CH3)2C4 O-, and is terminated by radical-radical reactions. Combination of two Ar• radicals leads to formation of tetraphenyls or chlorinated tetraphenyls and then to longer polyphenyls, all of which are eventually dechlorinated. Combination of Ar• with (CH3)2C4 O- may form adducts (such as R-hydroxyisopropylbiphenyls). Thus, while most of the PCBs are converted to biphenyl, some PCBs form various polyphenyls and hydroxyalkyl derivatives. These products are much less toxic than PCBs and can be more readily and safely treated for disposal or separated and recycled. The use of TEA instead of KOH is not recommended because (a) the efficiency of both the photolytic and radiolytic methods is lower with TEA, (b) TEA is volatile and toxic, and (c) radicalradical reactions involving TEA may form undesirable amine side products, such as R-(diethylamino)ethylbiphenyls. Since the treatment involves reducing conditions, except for the small amount of oxygen that is consumed in the beginning of the process, these amines are not expected to form any harmful nitrosamines. Nevertheless, such assumptions must be experimentally verified before scale-up. TEA may be replaced with another reductive quencher that is more benign, but we have not found any such quencher that is a sufficiently strong base to replace KOH [because it is necessary to ionize (CH3)2C4 OH into (CH3)2C4 O- to achieve a chain reaction]. Scale-up of either the radiolytic or the photolytic method from a small sealed vial to commercial quantities requires that the material be circulated in front of the energy source either by stirring or in a close-loop flow under anaerobic conditions, although the initial amount of oxygen present in the mixture does not necessarily have to be removed. While treatment with γ radiation or electron beam can be performed in closed metal containers, treatment with UV light may involve circulation of the liquid within a quartz tube surrounded by a bank of UV lamps. The initial cost of the UV lamps setup is expected to be much less than that of an electron accelerator or a γ source setup. On the other hand, the latter setups may run for years with minimal additional cost while the UV lamps may have to be replaced relatively frequently. Radiolytic or photolytic treatments of PCBs in a mixture with 2-PrOH and KOH are not too different from the basecatalyzed decomposition or the polyglycoxide methods in terms of the treatment materials but may be advantageous in terms of energy consumption because they involve a chain reaction that is merely initiated by the irradiation. PCBs can be dechlorinated also by reaction with strongly reducing metals and by catalytic hydrogenation. These methods do not involve a strong base but use a highly reactive metal that must be rereduced for recycling or specific catalysts and high pressures of hydrogen, all of which involve costly materials

and equipments. Detailed cost analysis and comparison must be performed before any recommendation can be made for a practical treatment method.

Acknowledgments We thank the National Science Foundation for support of previous studies (Grants BES-9800192 and BES-9320339) by the authors in this area, the U.S. Nuclear Regulatory Commission for support of C.G.J. during the course of this work, and Drs. Robert Huie and Jan Grodkowski for helpful discussions. The mention of commercial equipment or material does not imply recognition or endorsement by the National Institute of Standards and Technology, nor does it imply that the material or equipment identified is necessarily the best available for the purpose.

Literature Cited (1) U.S. Environmental Protection Agency. Management of Polychlorinated Biphenyls in the United States; Office of Pollution Prevention and Toxics: Revised January 30, 1997 (http:// irptc.unep.ch/pops/indxhtms/cspcb01.html). (2) U.S. Environmental Protection Agency. PCB workgroup meeting overheadssdisposal data (from 1990 to 1994 annual reports); March 23, 1988 (http://www.epa.gov/grtlakes/bns/stakeholder98/ pcbovers.htm). (3) U.S. Environmental Protection Agency, Office of Science and Technology. February 2, 1999 (http://www.epa.gov/ostwater/ fish/pcb99.html). (4) Cooper, W.; Cadavid, E.; Nickelsen, M.; Lin, K.; Kurucz, C.; Waite, T. J. Am. Water Works Assoc. 1993, 85 (9). (5) Exner, J. H. Detoxification of Hazardous Waste; Ann Arbor Science Publishers: Ann Arbor, MI, 1982. (6) Gray, K. A.; Hilarides, R. J. Radiat. Phys. Chem. 1995, 46, 10811084. (7) Hilarides, R. J.; Gray, K. A.; Guzzetta, J.; Cortellucci, N.; Sommer, C. Environ. Sci. Technol. 1994, 28, 2249-2258. (8) Matthews, S. M.; Boegel, A. J.; Loftis, J. A.; Caufield, R. A.; Mincher, B. J.; Meikrantz, D. H.; Murphy, R. J. Radiat. Phys. Chem. 1993, 42, 689-693. (9) Constellation Energy Group (formerly Baltimore Gas and Electric), 7609 Energy Parkway, Suite 101, Baltimore, MD 21226. (10) Arbon, R. E.; Mincher, B. J.; Knighton, W. B. Environ. Sci. Technol. 1996, 30, 1866-1871. (11) Chaychian, M.; Silverman, J.; Al-Sheikhly, M.; Poster, D. L.; Neta, P. Environ. Sci. Technol. 1999, 33, 2461-2464. (12) Schmelling, D. C.; Poster, D. L.; Chaychian, M.; Neta, P.; Silverman, J.; Al-Sheikhly, M. Environ. Sci. Technol. 1998, 32, 270-275. (13) Chaychian, M.; Jones, C.; Poster, D.; Silverman, J.; Neta, P.; Huie, R.; Al-Sheikhly, M. Radiat. Phys. Chem. 2002, 65, 473-478. (14) U.S. Environmental Protection Agency. PCB Sources and Regulations; Revised October 4, 1999 (http://www.epa.gov/ bnsdocs/pcbsrce/pbcsrce.html). (15) Erickson, M. D. Analytical Chemistry of PCBs, 2nd ed.; CRC: New York, 1997. (16) Jones, C. G. Ph.D. Dissertation, University of Maryland, 2001. (17) Far West Technology, Inc., 330-D South Kellogg, Goleta, CA 93117. (18) Methods for the Determination of Organic Compounds in Drinking Water; EPA-600/4-88/039; U.S. Environmental Protection Agency, Office of Research and Development: Cincinnati, 1988 (revised July 1991); p 109. (19) Poster, D. L.; Schantz, M. M.; Leigh, S. D.; Wise, S. A.; Standard Reference Materials (SRMs) for the calibration and validation of analytical methods for PCBs (as Aroclor Mixtures). J. Res. NIST (submitted for publication). (20) Sherman, W. V. J. Phys. Chem. 1968, 72, 2287-2288. (21) Sawai, T.; Shinozaki, Y. Chem. Lett. 1972, 865. (22) Sawai, T.; Shimokawa, T.; Shinozaki, Y. Bull. Chem. Soc. Jpn. 1974, 47, 1889-1893. (23) Sawai, T.; Shimokawa, T.; Sawai, T.; Hosoda, I.; Kondo, M. J. Nucl. Sci. Technol. 1975, 12, 502-507. (24) Singh, A.; Kremers, W.; Smalley, P.; Bennett, G. S. Radiat. Phys. Chem. 1985, 25, 11-19. (25) Mincher, B. J.; Curry, R. C.; Brey, R. Environ. Sci. Technol. 2000, 34, 3452-3455.

Received for review March 28, 2003. Revised manuscript received July 8, 2003. Accepted September 19, 2003. ES030412I VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5777