United Kingdom INFRARED LASER PHOTOCHEMISTRY OF TRANS-1 , 2-DICHLOROETHYLENE . EVIDENCE FOR A CI ATOM CHAIN REACTION

The TEA-CO2 laser induced reaction of trans-l,2-dichloroethylene (TDCE) was investigated at 925 cm-1. The laser radiation was focused to yield beam waist fluences of approximately 125 J cm-2. The major reaction productwas c/s-l, 2-dichloroethylene (CDCE), with a few per cent ofchloroacetylene, and minor amounts of acetylene, dichloroacetylene, chloroethylene and an unidentified C4 compound also being formed. The reaction of pure TDCE was studied as a function of number of laser pulses and total pressure (0.02 to 5 torr). Some experiments were also done with added ethane (20%) and propane (2%). Evidence was obtained that the formation ofthe cis isomer occurs via two mechanisms, (1) a unimolecular isomerization, and (2) a CI atom chain reaction. The results are consistent with laser induced decomposition of TDCE occurring through the three lowest energy channels: unimolecular structural isomerization (57.4 kcal/mol); molecular HCI elimination (69 kcal/mol); and C-CI bond scission (89 kcal/mol).


INTRODUCTION
We have been interested in experimental characterization of reaction temperature in laser powered pyrolysis. An approach to this problem that appears promising is the measurement of the temperature dependent equilibrium composition in isomerization reactions. We selected trans-l,2-dichloroethylene (TDCE) as a candidate for such a study since investigations of its multiple photon photochemistry reported the dominant reaction to be isomerization to c/s-l,2-dichloroethylene (CDCE), with only minor amounts of fragmentation occurring. -4 Furthermore, the equilibrium cis to trans ratio is calculated to decrease from 2.46 to 1.07 as temperature increases from 300 K to 1500 K using statistically derived thermodynamic functions. 5 During the course of this work we discovered that the isomerization occurs by two mechanisms, a unimolecular transformation, and a C1 atom catalyzed chain reaction.

EXPERIMENTAL
Samples of TDCE and CDCE were obtained from the Aldrich Chemical Company. Their purity was checked by gas chromatography. TDCE contained small amounts J, R. GUCKERT AND R. W. CARR of CDCE (1.06%), 1,1-dichloroethylene (0.37%) and acetone (0.59%). Both samples were degassed and used without further purification.
Cylindrical reaction vessels were constructed from 2.5 cm diameter Pyrex tubing, with planar NaC1 windows attached by epoxy cement at each end, and a glass-Teflon vacuum valve at the mid-point. Their length was either 29.3 cm or 12.8 cm. The cells were evacuated to 10torr before each experiment. Pressures were measured with either an ionization gauge or a capacitance manometer.
A Lumonics 102 TEA CO laser, tuned to the P(40) line at 925 cm -, was used for all photolyses. The laser pulse had a FWHM of 160 ns as measured with a photon drag detector. The radiation was focused at the center of the 29.3 cm reactor by a 40 cm focal length Ge lens in most experiments. A 1.54 cm diameter circular aperture, placed before the lens, selected a very nearly uniform radial intensity distribution from the laser beam. The laser pulse energies, reported as fluences (J/cm), were calculated for a point inside the front window based on the area of the aperture. The actual fluence inside the front window of the reactor was greater than this due to the reduced area of the partially focused beam at that point. The diameter at the beam waist was approximately 1 mm, producing fluences up to 125 J-cm-. Laser pulse energy was measured by a Scientech calorimeter and power meter connected to a homebuilt integrator. 6 Reaction product mixtures were analysed by FID gas chromatography, using a 3 ft Porapak Q column operated at 115C. For some analyses the column length was increased to 6 ft. For quantitation of the reaction products the response of the FID was assumed proportional to number of carbon atoms. Positive identification of the products was made by GC-MS.

RESULTS
The infrared spectra of TDCE and CDCE are shown in Figure 1. While CDCE does not have any absorption features between 900-1100 cm-1, TDCE has a band centered at 898 cm -1 which extends into the CO2 laser frequency range. This band, containing well defined rotational lines, corresponds to an out-of-plane wagging motion of the CHC1 groups. 5'7 The major product from multiple photon infrared excitation of TDCE at 925 cmis its isomer CDCE. Several decomposition products are also formed, the most abundant (up to 12%) being chloroacetylene. The other products, acetylene, dichloroacetylene, chloroethylene and an unidentified compound are formed in small amounts (approx. 1% or less). Evidence suggests that the unidentified compound was a C4 species, probably diacetylene. The acetone impurity decreased upon irradiation, but the influence of this reaction was small because of its trace amount. The quantity of the other impurity, 1,1-dichloroethylene, did not change significantly during photolysis.
The per cent conversions of TDCE as a function of the number of laser pulses at 5 torr and at 0.05 tort are shown in Figure 2. At each pressure the conversions reach a constant limiting value that is less than 100%. The 5 tort data are nearly at their  limiting value by 3000 to 4000 pulses, and the 0.05 torr conversions reached 68% after 8000 pulses, compared with 66% after 5000 pulses. Also, the limiting 5 torr conversion is slightly less than the limiting 0.05 torr conversion.
The yields of CDCE and chloroacetylene from the 0.05 and 5 torr experiments are plotted vs number of laser pulses in Figure 3. The 5 torr yield of chloroacetylene is significantly larger than the 0.05 torr yield over the entire range of pulses investigated, while for CDCE the yields cross between 2000 and 3000 pulses. The 5 torr yield of CDCE decreases slightly after about 2000 to 3000 pulses, while chloroacetylene continues to increase. On the other hand, the 0.05 torr yields of CDCE and chloroacetylene remain constant, within experimental error, above 5000 torr since the 8000 pulse yields of these species are 61.5% and 5.6%, respectively.
Spots the size of the laser beam appeared on the reaction vessel windows in the 5 torr experiments if the reactor was not evacuated for a sufficient period of time. When spots appeared, the windows were replaced. Several experiments were conducted to determine if heterogeneous reactions were occurring on the windows, as reported in a previous study of the MPD of TDCE. 3 The short reaction vessel was placed between the laser output coupler and the beam waist so that the fluence incident on its rear window was comparable to the influence inside the front windows of the longer cell used for routine experiments (about 4-5 J/cm2). In several experiments at 5 torr, including an 8500 pulse run, neither reaction products nor window spots were detected. However, at 0.05 torr, CDCE was produced at 1% per 1000 pulses. We conclude that the contribution of surface reactions to the observed volatile products is negligible under our conditions. Additional experiments were performed with pure CDCE. Although it does not have a readily discernible fundamental absorption band within the CO2 laser frequency range, CDCE has been reported to give small yields of TDCE upon laser irradiation at 935 cm-. In the present 925 cm -1 experiments, the yield of TDCE after 3000 pulses was 1% at 5 torr and 3.5% at 0.05 torr. Thus, MPD of CDCE is minor compared to that of TDCE. The effect of pressure was investigated by photolyzing TDCE over the range 0.02 torr to 5 torr. All of these experiments were conducted at 0.5 J/cm 2 and 5000 pulses since both the conversion and the yield of CDCE approached limiting values at these conditions. The results are in Table 1. The data, reported as area %, does not contain the acetone and 1,1-dichloroethylene impurities in TDCE since these were only accurately measurable at sample pressures above 3 torr due to longer GC retention times than the other trace species reported, C2H2 and C2H5C1, and even then would make negligible changes in reported percentages. The area per cent of CDCE and chloroacetylene in 5000 pulse experiments are plotted vs pressure in Figures 4 and 5. Some experiments were done with mixtures of 2% propane in TDCE. The small amount of propane should have very little effect on the excitation of TDCE. Table 2 shows that the yield of CDCE is reduced by about one third over a wide pressure range with the addition of propane. Furthermore, analysis of the irradiated mixture on the longer GC column revealed that propane underwent reaction. At 0.05 torr about 50% of the propane is consumed, which increases to about 75% at 5 torr. Propylene is formed in these experiments, although its amount cannot compensate for the loss of propane. Direct photodecomposition of propane was shown not to occur by irradiation of pure propane. Similar results were obtained from the irradiation of an 80% TDCE/20% ethane mixture. The results in Table 3 show that after correcting for the initial ethane concentration the yield of CDCE at both 0.05 and 5 torr is reduced by more than a factor of two with ethane present. While ethane at this pressure may quench hot TDCE or other intermediates leading to CDCE, it is unlikely that this alone can account for the loss of CDCE since quenching should be virtually absent at 0.05 torr, and large reductions in CDCE yield are observed at this pressure. Also, the  In another experiment the fluence was attenuated with CaF2 fiats. The beam was focused in the center of the short cell, producing a fluence of approximately 40 J/cm2, 68% lower than in the preceding experiments, at the beam waist. Yet the addition of 2% propane resulted in a 70% reduction of the CDCE yield (5 torr, 2 x 10 3 pulses), indicating that the scavengeable intermediate is still present at the lower fluence. t-C2H2C12 c-C2H2C12 In scheme 1, asterisk represents a highly vibrationally excited species.

DISCUSSION
Since infrared multiple photon processes are frequently dominated by the lowest energy channel, scheme (1) is expected to occur. However, the formation of decomposition products suggest that some of the higher energy channels in Table 4 may also be accessed at the experimental conditions of this work. An indication that reaction (1) by itself cannot account for the isomerization comes from the data in Figure 4, which show that CDCE and chloroacetylene, which together account for more than about 95% of the products, make up at the most about 70% of the reaction mixture. Furthermore, the conversions in Figure 2 show that the reaction cannot be driven to completion, even at 0.05 torr. This is not the expected result if the isomerization occurs by unimolecular transformation of laser pumped TDCE since CDCE does not photolyze appreciably, and in a unimolecular isomerization TDCE should be able to be essentially completely converted at pressures low enough that collisions are not significant during the laser pulse. At 0.05 torr the collision interval is about 1.4/ see.
The results from the experiments with added propane and ethane provide additional evidence that cis-trans isomerization occurs at least in part by another reaction path. The reduction of the CDCE yield, the consumption of propane and the production of propylene strongly suggest that propane acts to suppress a portion of the overall reaction involving thermal reactions of a reactive intermediate that causes isomerization of TDCE. The species responsible is most probably atomic C1, which may originate directly by C--C1 bond scission.
If laser excited TDCE undergoes C---CI bond cleavage, the CI atoms can promote reversible cis-trans isomerization, providing an explanation for the inability to drive the reaction to completion. Steel s suggest a CI atom chain reaction in the thermal isomerization of TDCE, while Ayscough et al. 9 and Knox and Riddick found evidence from photochlorination studies that isomerization proceeds via a chemically activated CI atom adduct. Furthermore, C--CI bond scission has been observed in MPD of several halocarbons, including trichloreothylene, xx The following reactions should then be considered. k- C2H2C13" --c-C2H2C12 + Cl (2) k-t k The addition of C1 occurs readily. The 298 K rate constants are reported to be 6.64 x 10 -11 cm 3 molecule -1 s -(C1 + TDCE) and 9.96 x 10cm 3 molecules -1 (CI + CDCE). 7 Isomerization will occur at the pressures of this work since the chemically activated trichloroethyl radical is only 50% collisionally stabilized at 150 torr, and furthermore, Wai and Rowland have reported the lifetime for rotation about the C--C bond to be < 10s. TM Finally, reactions 3 and 4 are competitive with addition of C1 to TDCE and CDCE even at room temperature, where their rate constants have the values 5.7 x 10cm 3 moleculesand 1.6 x 10cm 3 molecule -1 s -, respectively. 9 Cl + C2H6 HCI + C2H5 (3) CI + C3Hs HCI + C3H7 (4) It is also possible to envision H atom catalyzed isomerization. Production of H atoms by the cleavage of C--H bonds during MPD may occur, but is less likely than C---CI rupture due to the higher energy required. Furthermore, attack of either CDCE of TDCE by addition of H to the double bond would be followed by elimination of C1 in preference to H, leading to the formation of C2H3C1. Small amounts of this substance are formed, and provide evidence that H atoms may be present. However, the yield of C2H3CI is not reduced, but rather increased when ethane or propane are added. Thus any H atoms present are not scavenged by these species, consistent with the higher activation energy for abstraction of H from the C--H bonds by H atoms (7-9 kcal/mol) than by Cl atoms (approx. 0.1 kcal/mol). The reduction of CDCE yields upon addition of ethane or propane can be best explained ifisomerization occurs via the C1 atom chain of scheme 2, and Cl atoms are scavenged by ethane and propane.
In addition to C---C1 photodissodation, C1 atoms may be formed by HC1 elimination from hot TDCE followed by secondary photolysis of the product C2HC1. HCI elimination is energetically favoured over C--C1 cleavage. Elimination of HC1 has been reported in the MPD of chloroethylene 4 and trichloroethylene. 12 The presence of C1 atoms offers an explanation for the inability to drive the multiphoton reaction of TDCE to completion at low pressures. The finite cis/trans ratios could be due to competition between the depletion of TDCE by reaction scheme 1 and its regeneration by C1 catalyzed isomerization. Also, the amount of C1 is probably proportional to C2HC1 which is either produced concurrently with C1, or is a precursor of C1 by secondary photolysis. Thus, a decrease in the C2HC1 yield should be accompanied by an increase in the CDCE yield because of a reduction in the radical catalyzed isomerization. This trend is observed in Figure 4 at pressures less than 0.1 torr, where thermal effects should be absent.
Above 0.2 torr the CDCE/TDCE ratio decreases with increasing pressure. Since the isomerization is exothermic, this trend would be predicted if the system were thermalized by collisional redistribution of deposited vibrational energy. The ratio CDCE/TDCE 1.46 at 5 torr corresponds to a temperature of 610K if chemical equilbrium is reached. However, at this temperature the value of the overall thermal unimolecular rate constant is 6 x 10 -8 sec-, 13'4 which is much too low to account for the observed isomerization yield since the reaction time is only expected to be a few/z sec. 2 If the transient temperature were enough higher to produce thermal reaction rates permitting equilibrium to be approached in this short time, CDCE/ TDCE would be smaller than observed. Thus, at 5 torr the system must still not be totally thermalized.
While a thorough investigation of the formation of minor products was not done, observed trends in their yields permit some tentative, qualitative conclusions to be drawn.
The presence of CHC1 is most easily explained by elimination of HC1 from vibrationally excited TDCE and CDCE. It may also arise by H elimination from C2H2C1 radicals, either thermally or via secondary photolysis. A possible route to C2HC1 consisting of H abstraction from either TDCE or CDCE by C1, followed by C1 elimination from the C2HC1 is unaffected by the C1 atom scavengers C2H6 and C3H8. The initial decrease of C2HC1 with increasing pressure may be explained if collisional deactivation is responsible for a decrease in the relative importance of the higher energy HC1 elimination channel relative to the cis-trans isomerization channel at pressures below 0.1 torr. Around 0.1 torr collisions between two TDCE* molecules become possible, since the lengths of the mean free path (ca. 0.2 mm) and the radius of the beam at the focal point (ca. 0.5 mm) are comparable. If these collisions promote further up-pumping of one of the colliders, they will enable the reaction to proceed through the higher energy reaction channel, providing an explanation for the increase in C2HCI yield with increasing pressure.
While acetylene formation may be explained by elimination of C1 from C2H2C1 radicals and possibly also by C12 elimination from hot TDCE and CDCE, the increase of acetylene yield in the presence of the C1 atom scavengers provides an argument that a sequence consisting of H abstraction from TDCE or CDCE by atomic chlorine, followed by C1 elimination, may occur. Also, the increase of C2H3C1 in the presence of propane and ethane suggests that this species may be formed via H abstraction by C2H2C1 radicals. There is insufficient information on C2C12 except to say that it may arise from H2 elimination from either TDCE or CDCE.
An estimate of the importance of reaction channel three, CC1 dissociation, can be made by estimating chain lengths for the C1 atom catalyzed isomerization. If chain termination in pure TDCE is assumed to occur by H atom transfer from TDCE to C1, reaction 5; C1 + C2H2C12 HC1 + C2HC12 (5) the chain length is given by kt/k5. Data reported in Cillien et al. 21 can be used to estimate ks 6 x 10 -13 cm 3 molecule -1 s -1 at 298K, from which the chain length is approximately 100. If the dominant termination step in the presence of ethane or propane is reaction 3 or 4, respectively, then the chain lengths are about 6 or 18 at 298K. Applying these estimates of chain length to the low pressure data and assuming that TDCE is formed by only the two routes of schemes I and 2 it can be concluded that C--C1 bond scission is only 1% relative to laser selective unimolecular isomerization. Also, if C2HC1 is assumed to arise solely by HC1 elimination from TDCE at low pressures, then its yields can be used to estimate that the HC1 elimination channel is about 10-12% relative to laser selective unimolecular isomerization.