MATRIX ISOLATION STUDY OF THE 193 nm EXCIMER LASER PHOTOCHEMISTRY OF HEXAFLUOROBENZENE

193 nm excimer laser irradiation of Ar/C6F6 samples during deposition onto a cryogenic surface has led to the formation and isolation of a range of products, the dominant being hexafluoro-Dewar benzene. Additional absorptions likely due to the previously unreported hexafluorobenzvalene were observed, along with extensive fragmentation and additional minor products. When either Cl2 or CCl4 was doped into the Ar/C6F6 sample as an electron trap, a number of additional product bands were noted. A few of these were destroyed by subsequent Hg arc irradiation, and at least one is tentatively assigned to the C6F6


INTRODUCTION
The identification of intermediates in photochemical processes is important in un- derstanding the dynamics of excited states of molecules.Since many of these intermediates are very short-lived, transient techniques have been employed on a number of occasions.Alternatively, trapping of initial photochemical products into argon matrices may isolate and preserve these species for spectroscopic study.While this approach has been exploited on many occasions, the advent of the excimer laser has increased the range of accessible species.
In a recent study in this laboratory, the photochemistry of benzene after irradiation with the 193 nm ArF laser line was reported.Both isomerization and fragmentation products were observed, in a ratio that was dependent on the concentration of the precursor in the matrix.Hexafluorobenzene provides an interesting contrast to ben- zene, as a consequence of the stronger C--F bond.Also, while several isomers of C6F6 are well known, much less is known about the isomeric forms of C6F6.Only hexafluoro Dewar benzene has been reported, while very little is known about the hexafluorobenzvalene, prismane or fulvene isomers.In addition, while C6F6 has been the object of several photochemical Sttldies, 3,4 these have not been at 193 nm, where C6F6 is known to absorb strongly.Also, while the gas phase photochemistry of C6F6 has been examined, matrix trapping of intermediates has not been attempted to date.Finally, the photoelectron transitions for hexafluorobenzene occur at 9.8-10.8, 12.5- 13.5 and 16.0 eV.Thus, two-photon ionization to form the radical cation is possible provided a suitable electron trap is present in the matrix.With these considerations in mind, a study was undertaken to examine the argon matrix-isolated products of the 193 nm excimer laser irradiation of Ar/C6F6 samples, with and without an added electron trap.

EXPERIMENTAL SECTION
The experiments in this study were carded out on conventional matrix isolation equipment that has been described. 2,7,8Briefly, gas samples containing C6F (Aldrich) were deposited onto a CsI cold window maintained at 14 K.Some samples were doped with either C12 (Matheson) or CC14 (MCB Reagents).The reagents were purified by repeated freeze-pump-thaw cycles at 77 K prior to sample preparation.Samples were deposited for 20-24 hours before final spectra were recorded from 400 to 4000 cm on a Mattson Cygnus Fourier transform infrared spectrometer at 1 cm resolution.Two arrangements were used for irradiation of the sample by a Lambda Physik EMG 103 MSC excimer laser operhting at 193 nm.In the first (in situ), the sample was deposited, the deposition stopped and the final (pre-irradiation) spectra recorded.The cold window was then rotated by 45 degrees, allowing exposure to the laser beam through a suprasil window.The sample was irradiated for 1-2 hours, after which the cold window was rotated back into the beam of the spectrometer, and additional spectra were recorded.In the second, the cold window was rotated at the beginning of the experiment, and irradiation was concurrent with deposition.In both arrangements, the laser repetition rate was 5 Hz, with a pulse duration of 10 ns and a pulse energy up to 200 mJ.

RESULTS
Prior to irradiation experiments, blank experiments were carried out for each of the parent species.The spectra obtained in these blanks were in very good agreement with literature spectra. 9-1In some of the later experiments, CH4 was observed as a minor impurity arising from the walls of the vacuum system.
Several initial experiments were carried out irradiating Ar/C6F6 and Ar/C6F6/C12 samples in situ.No product absorptions were noted in any of these experiments.Consequently, in situ irradiation was abandoned, and all subsequent experiments were carried with irradiation concurrent with deposition.
A number of experiments were conducted in which samples of Ar/C6F6 were ir- radiated during deposition with the excimer laser.In these experiments, a large number of new infrared absorptions were observed, (see Table 1 and Figures 1 and  2) for typical experiments with concentrations of 1000/1 and 500/1.Over the series of experiments, these bands were observed reproducibly.In a second set of experiments, samples of AI'/C6F6/CC14 were irradiated during deposition, at several different concentration ratios.In these experiments, many new product absorptions were noted, an listed in Table 2.A number of these coincide with product absorptions observed above when C6F was irradiated, although with diminished intensity.Several bands match those observed during the irradiation of blank samples of Ar/CC14, while yet additional bands were new and required the presence of both precursors.Two of these matrix samples were subsequently sub- jected to irradiation from a medium pressure Hg arc lamp.As also noted in Table 2, some of the product bands were reduced in intensity or destroyed by Hg lamp irradiation.
Samples of Ar/C6F6/C12 were irradiated during deposition in a third set of experi- ments, at several different concentration ratios.As in the above experiments, a large number of new infrared absorptions were observed in these experiments, the majority of which corresponded to bands seen in the Ar/C6F6/CC14 experiments.All of the product bands in the Ar/C6F6/C12 system are listed in Table 3.One of these matrices was then subjected to Hg arc irradiation; as also shown in Table 3, several of the product bands were reduced or destroyed by the Hg arc.Infrared spectra, from 1270-1840 cm-, of a sample of Ar/C6F 500 subjected to 193 nm excimer laser irradiation during deposition (upper trace) compared to a spectrum of a similar sample without irradiation (lower trace).

DISCUSSION Product Identification
As is apparent from Tables 1-3, numerous new absorptions were seen following 193 nm excimer laser irradiation with or without added dopant.Previous studies have demonstrated that even with the short time between irradiation and matrix trap- ping there is sufficient time for extensive reaction and rearrangements to occur.As a consequence, radicals are not commonly isolated as they react and recombine rapidly.The literature, then, was searched in an attempt to match the many new absorptions to known fluorocarbons, matching both band position and relative inten- sities to spectra of known compounds.Unfortunately, the infrared spectrum of only two isomers of C6F6 itself are known, the parent D6h benzene-like structure and hexafluoro Dewar benzene?The other possible isomers are unknown although they are potential products in these experiments.Nonetheless, a substantial majority of the new bands could be assigned to known species.As listed in Table 1 for experiments without added dopant, the products include: hexafluoro Dewar benzene,

CC13CF
No change occurred after irradiation with mercury arc lamp.Present in blank experiments but increased in intensity when sample was irradiation during deposition.Overlapped by parent band.Overlapped by product band.decafluorobiphenyl, 2,3 octafluorocyclohexadiene (1, 3 and 1, 4 isomers), 4,5 octa- fluorotoluene 6 and decafluorocyclohexene. 7Weaker bands can be assigned to fragmentation products 8-2 C2F4, C4F2 and CF3CCF.Interestingly, in those experi- ments where CH4 was an impurity in the system, product absorptions due to C6FsH and C6F6fn were observed. 22,23Several additional bands remain.Two of these, at 1599 and 1677 cm-1, fall very near vibrational modes of the hexakis(trifluoromethyl) derivative of benzvalene, 24 and are tentative assigned to hexafluorobenzvalene. Of the remaining bands, most are weak (less than 0.10 absorbance units) and are not readily assigned.They may be due either to absorptions of the other isomers of C6F6, or to additional minor products.In any event, it is clear that extensive reaction and product formation occurs following 193 nm irradiation of these samples.Doping with an electron trap (here, either C12 or CC14) is commonly done in matrix photochemistry experiments to enhance photoionization reactions and cation forma- tion.These products are often detected by photobleaching experiments, where Hg arc irradiation releases electrons from the trap, and neutralization of the cation occurs.Of course, the electron trap may enter into the reaction chemistry; both effects were observed here.
Table 2 lists the products observed following 193 nm laser irradiation of Ar/C6F6/CCI4 samples with varying concentration ratios.Many of the products  observed without added dopant were observed in these experiments as well, although with generally lower yield.In addition, a number of chlorinated products were ob- served, primarily including C6F5C1 and C6FaClz (ortho, meta and para isomers), z2,25 The CC13 radical was seen at 898 cm -, as has been noted in a previous excimer laser irradiation study 26 of CC14 and elsewhere, 27 along with CC13F. 1,28Lesser products derived from this pair of reactants are listed in Table 2.It is noteworthy that several product bands were sensitive to Hg arc irradiation; these will be discussed below.
Experiments involving laser irradiation of Ar]C6F6/C12 samples led to product for- mation as well.Many of the products formed were seen in the above experiments, either during the irradiation of Ar/C6F 6 or Ar/C6F6/CCI4 samples.These are listed in Table 3. Due to CH4 impurity, additional product containing C,H,F and C1 were observed, as listed in the Table .Additional, new products were also seen, and are assigned in Table 3.Finally, several product absorptions were decreased or destroyed by Hg arc irradiation.Some of these were also seen in the CC14-doped experiments, while others were not.
Bands that were destroyed by Hg arc irradiation must be assigned to particularly reactive intermediates, often radical cations when an electron trap is present.Previous studies of the excimer laser irradiation of Ar/CC14 samples observed several such bands, which were assigned in accordance with earlier work to species such as CCI and CI, as well as the CC13 radical. 26,27These bands also observed in this study, and are assigned in Table 2.Additional photosensitive bands common to experiments with both dopants were noted at 479, 1276, 1473 and 1592 cm-.The first two were quite weak before and after Hg arc irradiation and limited conclusions can be reached about the species responsible for these two absorptions.The latter two were quite intense before irradiation, and showed dramatic reduction upon photobleaching with the Hg arc.While interesting, the limited data preclude definitive assignment to the absorbing species.Nonetheless, realistic possibilities should be discussed.
The most likely cationic product in this system is the parent radical cation, C6F , a species seen by emission spectroscopy after 193 nm excimer laser irradiation 29,3 of gaseous and argon matrix samples containing C6F6.The selection rules for emis- sion spectra, however, lead to observation of the totally symmetric modes of the emitting species, while infrared spectra show the antisymmetric modes.Thus, bands observed in emission are not anticipated in the infrared spectrum.However, a salt containing the C6F cation has been prepared 31 (with the AsF anion).The infrared spectrum of this salt shows a strong absorption at 1490 cm-, close to the 1473 cm band observed here.Antisymmetric CF stretching modes of aromatic fluorides are anticipated in this region, so assignment of the 1473 cm band seen here to C6F is likely.The first ionization threshold for C6F is around 10 eV which is readily ac- cessible by absorption of two photons.Multiphoton ionization of CC14 has been seen under identical conditions, 26 so that formation of C6F (and presumably C1-) is quite reasonable.CC14 serves as an effective electron trap by dissociative attachment, lead- ing to CC13 cage-paired with C1-.Hg irradiation may lead to electron detachment, neutralization of cations, and reduction in bands due to the CC13 radical.This reduc- tion was also observed, with the band near 900 cm -1 due to the CC13 radical decreasing significantly upon irradiation.
While assignment of the 1592 cm band to another mode of the C6F cation is possible, the salt spectrum showed no additional intense bands in this region.
Another, more likely, assignment is to non-rotating H20 in the argon matrix.Many researchers have noted that, in general, H20 rotates in solid argon, and gives rise to a well known spectrum. 32However, when certain impurities, including cations, are introduced into the matrix, this rotation is quenched and "non-rotating" H20 is detected 33 near 1592 cm-.It has been observed 34 that when cations in the matrix are photobleached, the local electric field is reduced and the H20 molecules are able to rotate.The band at 1592 cm is reduced, and the normal spectrum of "rotating" H20 is seen.While not definitive, this provides a reasonable explanation for this band, one that is in agreement with previous studies.
A few weak, photosensitive bands remain, at 1337 cm in the CC14 doped experi- ments and at 971,974 cm -1 in the C12 doped experiments.The 971,974 cm doublet was on the low energy shoulder of a very intense band of parent C6F which made reproducible observation of this band difficult.It may also have been in the CC14 experiments, but hidden by the parent band.This position is close and slightly to the red of an intense band at 1020 cm for the C6F cation in salts.Given the dif- ficulty in reproducibly detecting this doublet, such an assignment must be viewed an very tentative.The 1337 cm was in a very congested spectral region, and con- sequently might have escaped detection in the C12 experiments.This band was relatively weak, and thus any assignment would be very speculative.

Comparison with Previous Studies
Infrared multiphoton photochemistry of C6F6 in the gas phase reported C2F4 as the major product. 35These researchers propose initial formation of C6F5o + F, followed by a sequence of steps leading to the observed product.Bryce-Smith and coworkers 36 argue, at least in solution, C--F bond breakage is not the initial step due to the very strong CF bond, and that the solvent plays a significant role in the process.Earlier corona discharge studies 37 of C6F6 samples followed by matrix trapping led to isola- tion of a significant yield of decafluorobiphenyl, C2Fo, which was taken as indication of the initial formation of C6F5.C2F10 was very weakly observed at best in the present experiments despite the fact that the infrared spectrum is well known.This argues that in the present study CmF bond to rupture to form the C6F radical is not a major process.Haller has studied this system extensively, and suggested three competing isomerization processes for C6F6 from higher singlet states.Path (a) involves isomerization to hexafluoro-Dewar benzene from the E2g ($3) or Bu ($2) state of C6F6.Path (b) produces the hexafluorobicyclo (3.1.0)hexenylene biradical from any of the three singlet states S 1, $2 or $3, with little or no activation energy.It is important to note that this biradical is not involved in formation of the Dewar isomer, and is proposed to revert back to C6F6.As such, it is the major competition for isomerization to the Dewar isomer.Path (c) produces hexafluorobenzvalene from the Blu ($2) state with no activation energy required and from B2 (S) with an activation energy required.This mechanism fits many of the observations here, particularly that hexafluoro- Dewar benzene was a major product in these studies, in agreement with path (a).In addition, two bands were observed where hexafluorobenzvalene is anticipated to ab- sorb, as predicted by path (c).Of course, path (b) leads to reformation of the parent species, and would not be directly observable.With the photon flux employed here, additional photochemistry may occur, either from absorption of a photon by the initial products or absorption of a second photon by excited C6F6.This would lead to the more extensive fragmentation that was also observed.Finally, CH4 was an impurity in the present experiments, and products such as C6FsH and C6FsCH3 were detected.This agrees well with a previous study 23 where C6F6 and CH 4 were inten- tionally mixed and irradiated, and the same products (C6Fsn and C6FsCH3) were seen.
In the experiments doped with either CC14 or C12, additional chlorine-containing products were observed.These included C6F5C1, and 0-, m-, and p-C6F4C12.Since the C---C1 bond is readily ruptured by a 193 nm photon, and since extensive fragmentation and rearrangement occurs in the absence of the dopant, it is not surprising that chlorine incorporation into the product species occurs.It is possible that a Figure 2 Infrared spectra, from 1270-1840 cm-, of a sample of Ar/C6F 500 subjected to 193 nm

Table 1
Band Positions (cm-1) and Assignments for the Products of the Excimer-Laser Irradiation of avenumber Figure1Infrared spectra, from 400 to 4000 cm-', of a sample of Ar/C6F 500 subjected to 193 nm excimer laser irradiation during deposition (upper trace) compared,to a spectrum of a similar sample without irradiation (lower trace).

Table 2
Band Positions (cm-) and Assignments for the Products of the Excimer-Laser Irradiation of C6F6 Doped with CC14 in Argon Matrices

Table 3
Band Positions (cm-t) and Assignments for the Products of the Excimer-Laser Irradiation of C6F6 Doped with CI in Argon Matrices

Table 3
Cont'dPresents in blank experiments but increased in intensity when sample was irradiated during deposition.Overlapped by product band.