United States Environmental Protection Agency Environmental Sciences Research Laboratory Research Triangle Park NC 27711 Research and Development EPA-600/S3-83-001 Mar. 1983 &EPA Project Summary Photochemical Reactivity of Perchloroethylene Basil Dimitriades, Bruce W. Gay, Jr., Robert R. Arnts, and Robert L.Seila Perchloroethylene (PCE), a solvent used in dry cleaning, has been suspected of contributing significantly to photo- chemical ozone/oxidant (O3/Ox) prob- lems in urban atmospheres. Past evi- dence, however, was neither complete nor consistent. To interpret more con- clusively the past evidence, and to understand further the role of PCE in the O3/OX problem, a smog chamber testing program was conducted. The program's objectives were (a) to gener- ate additional evidence on the mecha- nism of the PCE reaction in smog chamber atmospheres, and (b) to ex- trapolate the smog chamber findings regarding PCE reactivity to the real atmosphere. Results showed that (a) in smog chambers, PCE reacts and forms O3/Ox following a Cl-instigated photo- oxidation mechanism rather than the OH-initiated mechanism accepted in current smog chemistry, and (b) in the real atmosphere neither the Cl-insti- gated nor the OH-instigated photooxi- dations of PCE can generate substantial concentrations of Os/Ox. In fact, PCE contributes less to the ambient O3/0» problem than ethane. This Project Summary was developed by EPA's Environmental Sciences Re- search Laboratory, Research Triangle Park, NC, to announce key findings of the research project that is fully docu- mented in a separate report of the same title (see Project Report ordering infor- mation at back). Introduction The approach adopted by the U.S. Environmental Protection Agency (EPA) to reducing photochemical ozone (O3) and other oxidants (Ox) has been based on unilateral control of the volatile organic compound (VOC) precursors. However, since the individual VOCs differ widely in 03- or 0,-forming potential, EPA accepted the concept of discriminate control of organic emissions, and from time to time produced lists of VOCs that could be exempted from control by virtue of their negligibly low photochemical reactivity. Perchloroethylene (PCE), a VOC used as a solvent and emitted in significant amounts from dry-cleaning operations, was first thought to be unreactive and was exempted from control. That early judgment was based on Los Angeles Air Pollution Control District (LAAPCD) smog chamber studies of solvent reactivities in the 1960s, which are now considered to be in error. In the years following the LAAPCD studies, there were several other PCE reactivity studies upon which EPA judged that PCE's photochemical reactiv- ity is not sufficiently low to justify total exemption from control. EPA's latest judgment on PCE reactivity was based mainly on findings from EPA and EPA-sponsored studies in the mid- 1970s. Other studies during that period produced results that agreed with the EPA policy on PCE in some respects, disagreed in others, or were not compa- rable. More recently, there have been additional studies, some directly ad- dressed to PCE's Os-forming reactivity and some to various mechanisms of PCE photooxidation. The new evidence ap- peared again to partly agree and partly disagree with the latest EPA judgment on PCE reactivity. However, it became also apparent that most of the conflicting evidence could be reconciled, which led ------- to reexamining the question of PCE reactivity in the light of the reconciled evidence. We concluded that more exper- imental testing was needed to completely reconcile conflicting evidence and to support a more definitive and reliable judgment regarding the reactivity of PCE. This report describes our latest exper- imental study. We also reanalyze and reinterpret all evidence now available on the atmospheric chemistry of PCE, and offer a new judgment of PCE's photo- chemical reactivity with respect to the ambient oxidant problem. To facilitate an understanding of the rationale behind the objectives and design of our new study, and the reanalysis and interpretation of the experimental evidence, we discuss the concept of reactivity, the past data and other relevant literature on PCE reactivity, and the current specific issues and information gaps. The Concept of Reactivity The reactivity aspect of primary interest in this report is the ability of the VOC to participate in atmospheric reactions and form Oa/Ox. The prevalent method for measuring such reactivity has been the smog chamber method, in which the test VOC is exposed in a smog chamber to reactant mixture and radiation conditions similar to those in real urban atmos- pheres, and the yield in 03/0X is meas- ured directly. In the past three decades there have been numerous smog chamber studies of VOC reactivity, and O3/Ox-yield reactivity data are now available for a large number of VOCs. There are problems, however, in using the existing data base for this particular reactivity index. These prob- lems forced researchers and regulators to (a) turn to simple, two-class reactivity classification schemes ("reactive," "un- reactive") in preference to individual VOC reactivity ranking schemes, and (b) con- sider other less direct but more usable reactivity indices. The reactivity index used here in addition to the "O3/Oxyield" was the "VOC consumption rate" reac- tivity. The significance of the latter index lies mainly in the fact that a VOC cannot produce 63/0, unless it participates— and, hence, is consumed—in the atmos- pheric reaction. The VOC consumption rate, therefore, offers the most reliable basis for recognizing the totally unreac- tive VOCs. Thus, VOCs whose consump- tion rates are negligibly small are classi- fied as unreactive. By extension, it was also assumed, at first, that VOCs mani- festing significant consumption rates in smog chambers are reactive. This latter assumption, however, has been question- ed by some recent studies, and, as discussed later, by this study also. Analysis and Interpretation of Existing Evidence on Perchloroethylene Reactivity Existing evidence considered here con- sists mainly of data on the O3/0, yield and PCE consumption rate reactivities obtained in the laboratory. Since crude measures of consumption rate reactivity can also be derived from ambient con- centration and emission rate data, these were included in this analysis. Each piece of reactivity data was critiqued to the extent allowed by the reported informa- tion, and labeled as supporting a "negli- gibly reactive" or "reactive" classifica- tion for PCE. "Negligibly reactive"was defined here as equal or less reactive than ethane (the latter organic taken by the authors to be a "boundary" species separating the reactive VOCs from the unreactive ones). Existing data on Oa/Ox yield and con- sumption rate reactivities were extracted from 13 formal and informal reports on laboratory studies. Estimates of consump- tion rate reactivity of PCE were also derived from PCE emission and aeromet- ric data. Examination of those data revealed considerable inconsistencies. Thus, of the 13 studies, six supported high or border- line O3/Ox yield reactivity, six supported negligible 03/0xyield reactivity, and all of the studies supported high but widely varying consumption rate reactivities. The aerometric data supported negligible reactivity. In searching for clues to the causes of these inconsistencies, the smog chamber studies' results were compared with reactivities deduced from smog chemistry considerations. To explain, currently ac- cepted smog chemistry (derived from atmospheric chemistry studies of hydro- carbons and aldehydes) explains VOC reactivity in terms of the VOC's ability to participate in a chemical process, a key step of which invariably is the initial reaction of the VOC with OH radicals. Reactions subsequent to the OH attack also have a role in the overall 03-forming process, but obviously, only if the initial reaction with OH occurs at a significant rate. Therefore, the value of the rate constant for the PCE + OH reaction can provide a useful first check on the reli- ability of the existing laboratory data. Such a value is known, and based on that and on the OH concentration in the atmosphere, the rate of PCE consumption in ambient air was computed to be 0.36%/h. Such a rate is 1 to 1 Vi orders of magnitude lower than the rates observed in smog chambers, which means either that the performance of the smog cham- bers was not consistent with smog chem- istry predictions or that current smog chemistry, for some reason, is not appli- cable in the PCE case. To further evaluate these two alternate explanations, con- sumption rates were computed from k0H data for several VOCs (hydrocarbons and halocarbons), and compared with those observed in the smog chambers. Results showed generally good agreement be- tween observed and predicted consump- tion rate reactivities for hydrocarbons but extremely poor agreement for halocar- bons. The halocarbon data disagreement is invariably in the direction of higher observed than computed reactivities. Therefore, the inconsistencies of the existing PCE reactivity data base cannot be attributed directly to smog chamber- related factors alone. The more logical conclusion is that current smog chemistry cannot be applied to polychlorinated ethylenes, or, more specifically, the re- action with OH is not the key step in the halocarbon consumption process. Some species, more potent than OH, must be responsible for the rapid PCE consump- tion in smog chambers. This raises four questions: 1. What is the chemical species that causes rapid PCE reaction in smog chambers? 2. What is the chemistry of the process following that species attack? Does it result in O3/OX production? 3. Does this chemistry explain the in- consistencies among the various studies data? 4. Is this species and associated chem- istry operative in the real atmos- phere? With respect to the first and second questions, past basic and smog chamber studies suggested that the PCE consump- tion observed in smog chambers is caused by a Cl-instigated chain photooxidation process which, apparently, is also capable of generating O3. The evidence available, however, on the origin of the Cl atoms in such systems, and on the effects of experimental factors on the PCE reaction with Cl, was lacking. In the absence of such evidence, the support of the Cl- instigated mechanism was weak, and the question of whether this mechanism can explain the inconsistent results of past ------- studies could not be answered. Further- more, there was no evidence that the Cl- instigated photooxi.dation mechanism, presumably operative in smog chamber systems, is necessarily operative in the real atmosphere also. These deficiencies of the past evidence led these investi- gators to conduct a new PCE study, and to reexamine the question of PCE reactivity in the light of the new experimental evidence. Experimental Procedures and Results of New Perchloroethylene Reactivity Study The analysis of existing evidence on PCE reactivity indicated that this experi- mental effort's objectives should be to (a) confirm that in the smog chamber meas- urements of PCE reactivity the operative chemistry in PCE degradation is Cl- rather than OH-instigated photochemistry, (b) understand such Cl chemistry well enough to reconcile some seemingly conflicting evidence, and (c) extrapolate the laboratory findings on PCE reactivity to the real atmosphere. To achieve the first objective, tests were conducted in which PCE was re- acted with Cl atoms, and tests in which PCE was reacted both in the presence and the absence of Cl scavengers. Also, direct photolysis of PCE and of some of its photooxidation products was investigated to determine the source of the Cl atoms that instigate the rapid disappearance of PCE observed in some smog chambers. To achieve the other objectives, smog chamber irradiation tests were conducted to determine the effects on PCE con- sumption of the following factors: initial PCE concentration, presence/concentra- tion of organic co-reactant, radiation spectrum, chamber wall material, and chamber use before the test. The irradiation tests used the absorp- tion cell of a long path Fourier Transform Infrared Spectrophotometer (FTIR) and several bags made of Teflon as smog chambers. Irradiation in both cases was provided by fluorescent lamps simulating natural sunlight. Thirty-seven smog chamber tests were conducted. Results showed that injection of CI2 accelerated strongly the PCE re- action, and that presence of a Cl-scaveng- ing co-reactant inhibited the PCE reaction drastically. These results indicated that PCE consumption was instigated by Cl atoms. To identify the sources of Cl atoms, tests were conducted to determine whether PCE or its reaction products photolyze to Cl atoms. Results suggested that direct photolysis of PCE is the most likely source of Cl atoms, although other sources (e.g., the PCE + OH reaction) could not be completely ruled out. In an effort to reconcile some incon- sistencies in the data from past studies, tests were included to study some exper- imental factors that were known or suspected to have significant effects on the PCE reaction and to differ in the various investigations. Most important among such factors were: radiation spec- trum, chamber wall, and initial reactant concentration. Of these factors, radiation spectrum probably varied the most with investigation, especially within the 2800- 3300A wavelength region. Several tests, therefore, were conducted in which this factor was varied either by changing the lamp composition or by using glass to screen out selected wavelengths. Results showed that the PCE reaction was ex- tremely sensitive to radiation within the 2800-3300A wavelength range. Two smog chamber wall factors were investigated: one related to the radiation transmitted by the window material, and the other related to the scavenging of radicals—especially of Cl atoms—by the inside wall surface. As already discussed, the radiation transmission factor in this study was a strong one. In regards to the radical scavenging role of walls, tests in this study were merely suggestive of such an effect. Of the initial reactant concentration factors in photochemical reactivity stud- ies, those relevant here are the concen- tration of the organic reactant, the con- centration of NO,, and the interaction of these two factors or the organic-to-NOx ratio. In the case of PCE, the data from this study indicated that the NO, does not influence the PCE reaction rate appre- ciably. The PCE concentration, however, was found to have a strong effect. The final objective in this experimental effort was to obtain evidence that would permit extrapolation of the laboratory findings on PCE reactivity to the real atmosphere. The requisite evidence con- sidered the effects of various reactant and radiation conditions on PCE reactivity, especially the radiation intensity and spectrum, the presence of co-reactant organics, and the PCE concentration. Data on the co-reactant, PCE concentra- tion, and radiation factors have been obtained in this study, including data from tests in which PCE-air mixtures were irradiated in parallel, with the laboratory radiation system and with natural sunlight. Results from the latter, natural sunlight irradiation tests showed the PCE rates to be lower than those observed in the artificial sunlight tests when all lamps were on, and comparable to those observed when the 2800-3300A component was reduced. Thus, reactivity data taken with Teflon film smog cham- bers and near-UV-rich radiation would tend to be erroneously high. Discussion The evidence obtained in this and the previous studies on the effects of the initial reactant concentrations, CI2, and organic co-reactants upon PCE activity is consistent with a Cl-instigated chain photooxidation mechanism analogous to the OH-initiated mechanism accepted in current smog chemistry: CI-CCI3CCI2 CCI3CCI2 + 02 - CCI3CCI202 CCI3CCI202 + NO - CCI3CCI20 NO + 0 Io2 I 03 CCI3CCI20 - CCI3C(O)CI + Cl -COCIz+CCIs 02 COCI2+CIO In the absence of NO, main reaction products should beCCI3CCI(O) and COCI2. In the presence of moderate concentra- tions of NO, 03 also should form through photolysis of NO2, as well as PAN-type products arising from the CCI3CO radical through reactions similar to those in smog chemistry for hydrocarbons/aldehydes. Finally, as with all VOC/NOX systems, excess NO should suppress production of 03, PAN and other oxidants. Of the various conceivable origins of Cl atoms, direct photolysis of PCE appears to be the one most consistent with existing data, although other sources (e.g., re- action of PCE with OH) cannot be completely ruled out. The above mecha- nism seems to provide reasonable expla- nations of the wide diversity of PCE reactivity results obtained in the various ------- studies. Thus, the higher consumption rate reactivity results observed in most studies are probably due to the higher intensities of 2800-3300A radiation used in those studies. The high 03/0X yield reactivity results are probably due to the same factor and also to optimum PCE-to- NOX ratio conditions. The final and most important question addressed was whether this Cl-instigated photooxidation chemistry observed in smog chambers is effective in producing O3 in the real atmosphere also. An answer was provided by the evidence from this study related to the effects of the PCE concentration and organic co-reactant factors on PCE reactivity. Based on that evidence, the presence of much higher concentrations of Cl-scavengingco-react- ants in the real atmosphere would cause the PCE to have an extremely low reactiv- ity, lower than that of ethane. Conclusions and Recommendations Currently accepted smog chemistry, with OH attack as the key initial reaction step, does not explain the rapid disappear- ance of PCE observed in some smog chambers. A Cl-instigated photooxidation mechanism appears to be more consist- ent with the smog chamber data available. This should be verified through quantita- tive comparisons of existing smog cham- ber data with mechanistic predictions using modeling techniques. Based on the proposed Cl-instigated photooxidation mechanism, it is derived that in the real atmosphere the PCE consumption rate is extremely small because relatively high concentrations of Cl-scavenging organics are present in ambient air. In fact, the PCE consumption rate, and, hence, its 03/0xyield reactivity also, are estimated to be lower than those of ethane. The EPA authors Basil Dimitriades (also the EPA Project Officer, see below), Bruce W. Gay, Jr., Robert R. Arms, and Robert L. Seila are with the Environmental Sciences Research Laboratory, Research Triangle Park, NC 27711. The complete report, entitled "Photochemical Reactivity of Perchloroethylene," (Order No. PB 83-163 014; Cost: $8.50, subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at: Environmental Sciences Research Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC 27711 frU.S. GOVERNMENT PRINTING OFFICE: 1983-659-017/7009 United States Environmental Protection Agency Center for Environmental Research Information Cincinnati OH 45268 Postage and Fees Paid Environmental Protection Agency EPA 335 Official Business Penalty for Private Use $300 IL ------- |