Environ. Sci. Technol. 2002, 36, 1839-1844
PCB Destruction in Subcritical and Supercritical Water s Evaluation of PCDF Formation and Initial Steps of Degradation Mechanisms ROLAND WEBER,* SHINYA YOSHIDA, AND KEIICHI MIWA Environmental Process Development Department, Ishikawajima-harima Heavy Industries Co., Ltd., 1, Shin-Nakahara-cho, Isogo-ku, Yokohama 235-8501, Japan
The reduction of polychlorinated biphenyl (PCB) emissions to the environment are contemporary issues of global efforts, and possible destruction technologies have to be selected and evaluated for PCB remediation. In this study PCB destruction in subcritical and supercritical water were assessed under oxidative conditions and nonoxidative, alkaline conditions. In both cases PCBs could be destroyed by more than 99%. The formation of polychlorinated dibenzofurans (PCDFs) during PCB destruction was evaluated. Under both oxidative and nonoxidative treatments of sub- and supercritical conditions, the formation of PCDFs was observed. The PCDFs formed in the early stages of PCB destruction resulted in up to 47-fold increase in terms of toxic equivalency (TEQ) compared to the initial PCB mixture. However, the PCDFs were destroyed together with the PCBs under more severe conditions i.e., at higher temperature or prolonged residence time. The mechanism of PCDF formation and the initial step of PCB degradation was evaluated. Our laboratory-scale investigation indicates that PCB destruction under supercritical water conditions is feasible, but because of the PCDF formation potential, in particular the high ratio of toxic 2,3,7,8-substituted congeners, conditions have to be carefully selected.
1. Introduction The release of polychlorinated biphenyls (PCBs) poses a serious threat to public health and the environment (1, 2). Although PCB production has been discontinued, these compounds are still remaining in large quantities in e.g., capacitors, electrical transformers, or storages. Therefore, PCBs are targeted by governments as part of a global treaty on persistent organic pollutants (3). The baseline remediation technology for PCBs is incineration. However, for a safe destruction, temperatures of more than 1100 °C are required (4) which demands state of the art hazardous waste incineration facilities. Therefore, several alternative PCB destruction technologies have been proposed during the last two decades or are already commercially applied including base-catalyzed decomposition (BCD) (5, 6), KOH/PEG methods (7, 8), ultraviolet irradiation (9), catalytic destruction on oxidation catalysts (10-12), or biodegradation (13). However, until to date there is no state of the art technology for a destruction method alternative to hazardous waste incineration. Since PCBs are included in a variety of matrices, different tech* Corresponding author phone: +81-45-759-2149; fax: +81-45759-2149; e-mail:
[email protected]. 10.1021/es0113910 CCC: $22.00 Published on Web 03/13/2002
2002 American Chemical Society
nologies may provide optimum solutions for different types of PCB-waste. Supercritical water oxidation (SCWO) has been studied in the last two decades as an oxidation process for the treatment of hazardous waste (14-16). Above its critical point (TC ) 374 °C, PC ) 218 atm), water has a high solubility for organics and oxygen, so a single phase containing a homogeneous mixture can exist leading to rapid oxidation rates. In laboratory experiments and pilot plants it has been shown that PCBs can be destroyed under sub- and supercritical water conditions (17, 18). In these studies, however, it was not examined or reported if any formation of polychlorinated dibenzofurans (PCDFs) occurs during the PCB destruction. Recently Hatakeda et al. (19) reported the detection of toxicologically nonrelevant monochlorinated dibenzofurans as byproducts during SCWO of 3-monochlorobiphenyl. However, during all thermal treatments of PCBs, the formation of PCDFssespecially the toxic 2,3,7,8substituted congeners (TEQ)shas to be considered as an important parameter and closely evaluated when establishing a treatment method for PCB destruction. It has been shown that the oxidation of PCBs to PCDFs occurs in the presence of air already at temperatures as low as 300 °C (20). Increasing the temperature to 550 °C under short-term pyrolysis conditions, the conversion rates of PCBs to PCDFs were found as high as 25% (21). This covers the temperature range in which sub- and supercritical water (SCW) treatment seems feasible. Therefore one key question for assessment of PCB destruction in SCW (and generally for every PCB destruction technology) is whether the more toxic PCDFs are formed and under which conditions and applications a conversion to PCDFs has no relevance. In this study we investigated the destruction of PCBs during sub- and supercritical water oxidation in the presence of oxygen. In a second experimental series we investigated the potency of sub- and supercritical water to destroy PCBs dissolved in a hydrophobic solvent under nonoxidative, alkaline conditions. For both experimental conditions we evaluated the relevance of PCDF formation. Furthermore, the initial steps of destruction of chlorinated aromatic compounds in sub- and supercritical water were assessed with respect to the degradation mechanism of PCBs and the formation of PCDFs.
2. Experimental Section 2.1. Chemicals. PCBs used in this study included Clophen A 30 and Clophen A 60 mixtures. For the experiments, these PCB mixtures were combined to get a homogeneous homologue distribution of DiCB to HeptaCB. The mixture also contained measurable amounts of MonoCB, OctaCB, and NonaCB. Additionally DecaCB was added (Wellington laboratories, Ontario, Canada) to enable evaluation of the whole range of PCB homologues in the destruction study. The PCDFs, generally present in the ppm levels in commercial PCB mixtures (22), were separated from the PCB mixtures on an alumina column before application, to avoid their interference with PCDFs that may form during the experiments. 2,2′,4,4′,5,5′-HexaCB (PCB #153) and 3,3′,4,4′-TetraCB (PCB #77) (Wellington Laboratories) were used as individual PCB standards for detailed mechanistic studies. 2.2. Experiments. A Microautoclave (44 mL volume) was used as reactor. Hastelalloy C-276 was selected as reactor material because it was shown to have favorable anticorrosion characteristics under supercritical water oxidation conditions VOL. 36, NO. 8, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1839
FIGURE 1. Destruction efficiency of PCB in subcritical and supercritical water under oxidative conditions (oxygen, 15 min, pH 7) and nonoxidative, alkaline conditions (argon, 15 min, 0.1 m Na(OH) pH 13, 0.5 mL of tetradecane). (16, 18). All experiments were performed in batch type mode. For the oxidation experiments 11 mL of water was loaded into the reactor and spiked with 100-1000 µg of PCB mixture. The microreactor was filled with oxygen providing an overstoichometric amount of oxidant even when assuming total oxidation. For nonoxidative conditions, 11 mL of sodium hydroxide solution (0.1 normal) and 0.5 g of tetradecane were filled into the reactor and the reactor was flushed with argon. For the experiments a concentration of 100 µg of PCB dissolved in tetradecane was chosen corresponding to a contamination of 200 ppm. For all experiments using single congeners, 1000 ng of PCB was spiked. The sealed autoclave was heated within 5 min to the respective temperature (measured inside the autoclave wall) and held at this temperature ((3 °C) for the chosen reaction time. The reaction was finally quenched by water spray for rapid cooling. After opening the autoclave, the 13C labeled standards were added. At selected conditions replicates were performed under oxidative and nonoxidative, alkaline treatments. The reproducibility was high (below (10% variability) and satisfying for the purposes of this study (Figure 1). 2.3. Cleanup, Analysis, and Quantification. The cleanup procedures are described elsewhere (23, 24). Analysis was carried out by high-resolution gas chromatography on a HP 6890 gas chromatograph coupled to a HP 5973 mass selective detector (low resolution mass spectrometry) or a Micromass Autospec Ultima (high-resolution mass spectrometry). The quantification of PCBs and PCDDs/PCDFs was carried out by isotope dilution mass spectrometry with 13C-labeled standards. For quantification of TetraCDDs/TetraCDFs to OctaCDDs/OctaCDFs all 17 2,3,7,8-substituted congeners were spiked. For the analysis of MonoCDDs/MonoCDFs to TriCDDs/TriCDFs at least one 13C-labeled congener was spiked for each homologue. The conversion of PCBs to PCDFs (%) in this study are recalculated to the amount of degraded PCBs. Hydroxylated PCBs were quantified using internal native 2,2′,3′,4,4′,5′-hexachloro-3-biphenylol standard spike (25 ng). In preliminary tests, the presence of this standard in the product mixtures was examined for both conditions and found negligible (374 °C) resulted in significant increases of the DE. However the DE at 400 °C (87% ( 6%) was still insufficient for technical application. Destruction rates greater than 99.99% were finally achieved at 450 °C. The destruction efficiency for PCBs in SCWO in the present study are in agreement with the experiments reported in the literature. For example, Hatakeda et al. (19) found insufficient PCB degradation rates in SCW with oxygen at 400 °C. On the other hand, Anitescu and Tavlarides (20) found a high destruction efficiency in the temperature range between 450 and 550 °C. 3.2. Destruction of PCBs by SCW under Nonoxidative, Alkaline Conditions. Nonoxidative conditions were applied to protect the tetradecane from oxidative destruction during PCB degradation. A 0.1 N Na(OH) solution was chosen, which showed a favorable destruction potential in a preliminary test series (25). Under the alkaline, nonoxidative conditions the onset of PCB degradation was observed already around 200 °C. At 350 °C under subcritical conditions 95% of the PCBs were degraded (Figure 1). Under supercritical conditions at 400 °C a destruction efficiency of 99.8% was achieved. Therefore it seems feasible to decontaminate transformer oils, contaminated in the range of several hundred ppm of PCBs, to values below 0.5 ppm already in the lower temperature range of supercritical water under alkaline conditions to meet even the stringent Japanese regulation standards (26). 3.3. PCDF Formation during Degradation of PCBs. PCDFs were formed in all experiments under both oxidative and nonoxidative treatments (Table 1). The transformation yield of total PCBs to total PCDFs under all experimental conditions was below 7% based on the amount of degraded PCBs. PCDDs were not formed under the nonoxidative, alkaline conditions and only to a minor extent under oxidative conditions (
