Seasonal variations in the atmospheric distribution of a reactive chlorine compound, tetrachloroethene (CCl2=CCl2)

Tropospheric mixing ratios of CCl2=CCl2 were measured at remote surface locations in the Pacific between 71°N and 47°S during September and December of 1989, and March and June of 1990. The observed gradient of decreasing concentrations from the northern to southern hemisphere, and very low concentrations in the southern hemisphere throughout the year, indicates a predominant input from the northern hemisphere. Our seasonal measurements in the northern hemisphere showed maximum CCl2=CCl2 concentrations occurring in the late winter and minimum concentrations occurring in the late summer. This distinct seasonal variation is strongly coupled to the atmospheric abundance of hydroxyl radical, the only important species responsible for CCl2=CCl2 removal. Using the estimated global CCl2=CCl2 emissions the lifetime is calculated to be about 5.4 months which is in good agreement with the 4.0 month estimate obtained from the inverse ratio of its measured hydroxyl reaction rate constant compared with that of methylchloroform (CH3CCl3).


Introduction
The latitudinal distribution and atmospheric survival of reactive chlorinated molecules has recently attained regulatory policy significance because of concern over the effects on stratospheric ozone anticipated from the hydrochlorofluorocarbons (HCFCs) now being developed and used as substitutes for the fully-halogenated chlorofluorocarbons (CFCs) [WMO- NASA, 1990]. We have measured the atmospheric concentration of the reactive gas tetrachloroethylene, CC12=CC12, in all four seasons during 1989-1990 at remote surface locations covering the latitude range from 71øN to 47øS. Tetrachloroethene is man-made with no confirmed natural sources and is primarily used as a dry cleaning and degreasing solvent [Most, 1989]. This similarity in usage to other anthropogenic chlorocarbon compounds, especially CH3CC13 and CC1F2CC12F (CFC-113), makes its geographical emission pattern very much like those of the CFCs and CFC substitutes.
The double bond in its molecular structure makes it susceptible to hydroxyl radical addition reaction, which ultimately results in its atmospheric removal [Itoh et al., 1994]. While the CC12=CC12 emissions are likely to be distributed more or less uniformly throughout the year, the HO source strength is quite variable with both season and latitude, and transport between the northern and southern hemispheres also has a significant seasonal dependence. With the majority of the CC12=CC12 release occurring northward of 30øN latitude and reaction with HO radicals as the only significant sink, the result-Copyright 1995 by the American Geophysical Union.

Paper number 95GL01069
0094-8534/95/95 GL-01069503.00 ing distribution of this compound is a sensitive test of the interplay between atmospheric transport and chemical destruction. The year-round measurements of CC12=CC12 also furnish information applicable to the atmospheric behavior of proposed short-lived HCFC and HFC substitutes [WMO 1990].

Experimental
Air samples were collected in four periods, each approximately two weeks in length, centered in mid-September 1989, early December 1989, plus mid-March and mid-June in 1990. For each collection period, approximately 50 to 60 surface-level air samples were introduced at ambient pressure into preevacuated 2-liter stainless steel canisters in remote Pacific locations, covering latitudes from Barrow, Alaska (71øN) to southern New Zealand (47øS). Within 3 weeks of collection, samples were returned to the University of California Irvine laboratory for gas chromatographic (GC) analysis on separate aliquots for halocarbons  and other trace gases [Blake et al., 1994]. The halocarbon analyses in 1989-1990 were performed by cryogenically preconcentrating a 26.3 ml STP air sample aliquot at -165øC on an 1/8" stainless steel loop packed with 80/120 mesh glass beads. The loop was warmed and flushed with ultra pure nitrogen gas onto a 60m x 0.75 mm I.D. glass SPB-1 capillary column (Supelco).
Three aliquots of each air sample were analyzed, with measurements of a reference air sample of known mixing ratios bracketing each triplicate analysis. The linearity of the ECD response for various halocarbons, including CC12=CC12, has been rigorously studied. The CC12=CC12 concentrations for background samples fall well within the linear working range of its calibration curve .
Mixing ratios for the halocarbons were calibrated by diluting a pure halocarbon mixture to the part per trillion by volume (pptv) level using a 4-step volumetric dilution method and a glass vacuum line. The absolute dilution error was determined from pressure and volume measurement uncertainties. The overall calibration error (1•) for CC12=CC12 is estimated to be 8% at about 2 pptv and 3% at about 40 pptv. Using this dilution method, our calibration scales for CFC-12 and CFC-11 are in agreement within 2% with the absolute values in other measurement programs [NASA, 1994]. The only uncertainty not included in the error assessment is surface adsorption which for CC12=CC12 is believed to be quite small on a glass surface. One torr of water vapor was added to preserve standard mixtures in the glass bulb and minimize any adsorption problems during analysis.
A second method was also employed using CFC-12 in the reference air as the basis for an indirect measurement of the dilution factor. In this method, the initial partial pressures in the mixture for CFC-12 and CC12=CC12 were carefully measured. This mixture was subsequently diluted to the parts per trillion level. This method depends on the accuracy of CFC-12 in the reference air, the analytical precision for CFC-12 measurements and the accuracy for measuring the initial partial pressures of pure halocarbons. Both methods were used several times and they agree within 3% for CC12--CC12.

Results and Discussion
Latitudinal Mixing ratios in the southern hemisphere were from 6 pptv near the equator and decreased to about 2 pptv south of the tropics (Figure 3). By contrast, after excluding several significantly elevated samples for which the concentrations of other halocarbons also suggested recent urban augmentation, the CC12=CC12 measurements in the September 1989 showed minimum mixing ratios for the northern hemisphere of less than 10 pptv. The southern hemispheric mixing ratios were about 1-2 pptv, not much different from those of the March 1990 trip. Overall, our reported mixing ratios fall in the lower end of published values, but are in good agreement with the measurements made by Koppmann et al. [1993] in August/September, 1989 and the measurements made by Rudolph et al. [1984] in spring 1983. Our data also agree with the summer and winter measurements made in Hokkaido, Japan between 42 ø-45øN [Makide et al., 1987].
After removing the significantly elevated samples shown in Figures 1-4 If both the CC12=CC12 atmospheric burden and its emissions are known, the atmospheric lifetime of CC12=CC12 can be calculated, and therefore by inference the average HO concentration can be estimated, as has been done for CH3CC13 [Makide and Rowland, 1981;Prinn et al, 1992]. Using CC12=CC12 in some respects could be more beneficial than using CH3CC13, because the interplay with HO is more sensitive for CC12=CC12 than for These observations provide two additional insights into the problem of ozone depletion by chlorine released in the stratosphere from organochlorine compounds. First, the transport of such molecules from the troposphere to the stratosphere takes place almost entirely through the tropical tropopause. Because the typical mixing ratios in the tropics in all seasons are around 5 pptv of CC12=CC12 or 20 pptv of C1, the fraction of the current 4,000 pptv of tropospheric organochlorine attributable to CC12=CC12 as a source is about 0.5%. The delivery of only 20 pptv C1 to the stratosphere from yearly atmospheric emissions of about 400 kilotons contrasts strikingly with the delivery of more than 1,000 pptv C1 from CC12F2 (2 C1 atoms x 500 pptv mixing ratio) from the accumulation from yearly emissions over the 1970' s and the 1980's of about 400 kilotons/year. The second point is a corollary of the first. When chlorocarbon molecules have effective tropospheric sinks, they do not reach the tropical tropopause in large concentration. Furthermore, the close comparison for CC12=CC12 between average lifetime calculated versus CH3CC13 and the observed average lifetime from actual atmospheric burden lends strong support for the use of the reaction rate constant ratio versus CH3CC13 procedure for molecules not yet introduced into the atmosphere. This is particularly relevant for lifetimes calculated in this manner for HCFCs, HFCs, and other hydrogen-containing molecules proposed as technological substitutes for the CFC's being replaced under the terms of the Montreal Protocol.