Infrared and EPR Spectroscopic Studies of 2-C2H2F and 1-C2H2F Radicals Isolated in Solid Argon

Home , Cryochemistry

Journal of Molecular Spectroscopy 205, 269–279 (2001) doi:10.1006/jmsp.2000.8266, available online at http://www.idealibrary.com on

I. U. Goldschleger,* A. V. Akimov,* E. Ya. Misochko,* and C. A. Wight† *Institute of Problems of Chemical Physics of the Russian Academy of Sciences, 142432, Chernogolovka, Moscow region, Russian Federation; and †Department of Chemistry, University of Utah, Salt Lake City, Utah 84112

E-mail: [email protected], [email protected]

Received August 17, 2000; in revised form October 12, 2000

2-fluorovinyl radicals were generated in solid argon by solid-state chemical reactions of mobile F atoms with acetylene and its deuterated analogues. Highly resolved EPR spectra of the stabilized radicals CHFA•CH, CDFA•CD, CHFA•CD, and CDFA•CH were obtained for the first time. The observed spectra were assigned to cis-2-fluorovinyl radical based on excellent

agreement between the measured (a F ϭ 6.50, a ␤H ϭ 3.86, a ␣H ϭ 0.25 mT) hyperfine constants and those calculated using density functional (B3LYP) theory. Analogous experiments carried out using infrared spectroscopy yielded a complete assignment of the vibrational frequencies. An unusual reversible photochemical conversion is observed in which cis-2- fluorovinyl radicals can be partially converted to 1-fluorovinyl radicals by pulsed laser photolysis at 532 nm. Photolysis at 355 nm converts 1-fluorovinyl back to cis-2-fluorovinyl. High-resolution EPR and infrared spectra of 1-fluorovinyl were obtained

for the first time. The measured hyperfine constants (a F ϭ 13.71, a H1 ϭ 4.21, a H2 ϭ 1.16 mT) are in good agreement with calculated values. © 2001 Academic Press Key Words: cryochemistry; EPR; infrared spectroscopy; fluorine; acetylene; 2-fluorovinyl radical; 1-fluorovinyl radical.

INTRODUCTION in the matrix. The most challenging aspect of this method is that reactants are normally closed-shell species; therefore, most Free radicals play an important role in many chemical and of the likely reaction products will also be closed-shell. How- biological processes. Therefore considerable attention has been ever, we have recently explored a modification of the method given to study of their properties (1). The matrix isolation that is based on the high mobility of fluorine atoms in rare gas technique has been a highly successful method for spectro- matrices and on the combination of two complementary spec- scopic investigations of radical products of chemical reactions. troscopic techniques, IR and EPR (3, 4). The ability of fluorine A common variant of this method is to trap gas-phase reaction atoms (e.g., generated by photolysis of F2) to migrate in solid products and freeze them in a matrix consisting of a large argon over long distances (5, 6) allows them to react with excess of argon or neon. The relatively weak guest–host in- closed-shell species to form free radicals at sufficiently high teractions allow spectroscopic characterization of the matrix- concentrations for spectroscopic study. The advantages of us- isolated species under conditions that are similar to the gas ing of two complementary spectroscopic techniques (EPR for phase but at much higher density than can ordinarily be sensitive detection of radicals and IR for characterization of achieved in the gas phase. closed-shell products) were demonstrated earlier for reactions Infrared absorption and EPR spectra of many radicals were of fluorine atoms with methane (3), ethylene (7), nitrogen obtained using this technique and received general acceptance oxide (8), and ammonia (4). (2). However, in some cases, radicals were misidentified or Using this approach, we have shown that complete sets of poorly characterized. In particular, difficulties in their identi- isotropic hyperfine constants (hf) can be determined from fication by IR spectroscopy included (a) the presence of sec- high-resolution EPR spectra of radicals in solid argon. The- ondary gas-phase reaction products; (b) stabilization of prod- oretically calculated isotropic hf constants of radicals and ucts in different structural traps of the solid matrix results in radical-molecular complexes are generally in good agree- multiple infrared bands, or “site splittings”; (c) registration of ment with those measured. The isotropic hf constants low-frequency IR bands ␯ Յ 300 cmϪ1, which are required for change drastically upon even minor distortion from the reliable identification of radical-molecular complexes, struc- equilibrium geometry, whereas vibrational frequencies gen- tural isomers, or conformational isomers, is experimentally erally do not (3). Therefore EPR technique is a more reliable challenging under ordinary conditions. method for identification and clarification of the structure Most of these difficulties can be overcome if the radicals are and geometry of stabilized radicals than the infrared tech- generated by atom–molecular reactions that take place directly nique. The main advantage of using IR is that most of the

269 0022-2852/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved. 270 GOLDSCHLEGER ET AL. molecules can be detected, providing a more comprehensive each other, and the measured hf constants (11, 13) are actually probe of reactants and products. close to those calculated for cis-2-C2H2F. The calculated har- This paper is devoted to the study of fluorovinyl radicals monic vibrational frequencies of both isomers are nearly the formed in the reactions of mobile F atoms with C2H2 mole- same (i.e., the differences in the frequencies are comparable cules. In the gas phase, this reaction leads to the formation of with the accuracy of calculations). two types of products: In this paper, we show how it is possible to form relatively high concentrations of isolated fluorovinyl radicals at 24 K by reaction of thermally diffusing F atoms with matrix-isolated [1] acetylene molecules. We make a definitive assignment of the

spectroscopic properties of cis-2-C2H2F based on an experi- It was determined that the dominant channel is the formation of mental approach (4, 8) that combines measurements made using infrared and EPR spectroscopies. The correlation of the 2-C2H2F radical (9, 10). There are two isomers of this radical, which differ in the location of the F atom with respect to the results is made possible by the discovery of a reversible pho- orbital of unpaired electron: toconversion between cis-2-C2H2F and 1-C2H2F that is observed in both types of experiments.

EXPERIMENTAL

The experimental techniques are similar to those used in our previous studies (3). Samples were prepared by vacuum co-

The first attempt to obtain the EPR spectrum of ␤-C2D2F deposition of gaseous mixtures Ar/C2H2 and Ar/F2 onto a radical by F2 photolysis of solid ternary mixtures Ar/F2/C2D2 at substrate (CsI salt window in IR experiments, or a sapphire rod 4.2 K was made in 1970 by Cochran et al. (11). They observed in the EPR experiments) maintained at 15 K. The composition a poorly resolved anisotropic spectrum, but they were able to of typical samples was Ar:F2:C2H2 ϭ 3000:1:1, though the Ϫ3 determine isotropic hf constants for the F atom (a F ϭ 6.9 mT) relative concentration of reactants were varied from 3 ϫ 10 Ϫ4 and deuterium (a D ϭ 0.6 mT). Based on the known isotropic hf to 2.5 ϫ 10 . The thickness of the samples was approximately constants of both isomers of C2H3 radical (12), they ascribed 100 ␮m. the detected EPR spectrum to trans-␤-C2D2F radical. Better- The EPR experiments were carried out with a homemade resolved spectra of 2-fluorovinyl radical in an SF6 matrix were helium flow cryostat with a movable helium shaft (3). The obtained by Shiotani et al. (13). These authors also assigned temperature stability was ϳ0.1 K over the range 15–40 K. EPR the observed spectra to trans-2-C2H2F radical. Such assign- spectra were recorded using standard 9 GHz spectrometer. ment was based on the results of semiempirical INDO MO Infrared spectra were recorded with a Mattson Model RS/ calculations, which predicted that trans isomer is more stable 10000 FTIR spectrometer (spectral region from 500 to 4000 than the cis isomer. The infrared absorption spectrum of ma- cmϪ1 and spectral resolution 0.5 cmϪ1). The closed-cycle he- trix-isolated fluorovinyl radical was reported by Jacox (14)by lium refrigerator (APD Cryogenics Model DE-202) was used trapping the gas-phase products of the F ϩ C2H2 reaction. to maintain the sample temperature in these experiments. In a Many different species were generated in this experiment, and separate experiment we verified that the light from the ceramic it required exceptional skill and scrupulousness to separate the globar source did not induce any chemical reactions. infrared bands of 2-C2H2F radical from the variety of other Fluorine atoms were generated by F2 photolysis at ␭ ϭ 355 bands. Jacox obtained a reasonable fit of the isotopic data to a nm (third harmonic of Nd:YAG laser in IR experiments) and diagonal valence force field potential using the assumed struc- ␭ ϭ 337 nm (N2 laser in EPR experiments). The average power 2 ture of trans-2-C2H2F radical, and with the support of the did not exceed 10 mW/cm in either type of experiment. To previous EPR study (11), she assigned the observed spectrum distinguish the chemical reactions involving photogenerated F to trans-2-C2H2F. atoms from those of diffusing thermal atoms, photolysis of F2 The first reliable calculations of spectroscopic characteristics molecules was performed at 15 K. Fluorine atoms are able to of fluorovinyl radicals performed with post-Hartree–Fock diffuse in solid argon at temperatures above 20 K. To initiate methods and density functional methods appeared in 1993 reactions of thermal F atoms, we performed annealing of (15). These calculations showed that the energy of the cis photolyzed samples at 24 K. Annealing was carried out by isomer is lower than the trans isomer when the basis set is step-by-step procedure. After completion of photolysis at 15 K, enlarged and differences in zero-point vibrational energy are the sample was annealed 3–5 min at 24 K. Then the tempera- taken into account. The spin density in the 2-C2H2F radical is ture was lowered back to 15 K and the spectrum was recorded. qualitatively different from that of C2H3, thereby casting some This cycle were repeated 10–12 times until reactions were doubt on Cochran’s original assignment. The isotropic hf con- complete. stants of cis and trans isomers of 2-C2H2F strongly differ from UV photolysis at 15 K and subsequent annealing of an Ar/F2

FIG. 1. Infrared spectra of a sample Ar:F2:C2H2 ϭ 3000:1:1 after deposition (trace a). Difference spectrum after exhaustive UV photolysis at 15 K (trace b). Trace c shows difference spectrum of photolyzed sample before and after annealing at 24 K. IR bands of 2-C2H2F are marked with asterisks. All spectra were recorded at 15 K. sample at 24 K lead to the appearance of a sharp absorption at Ϫ1 1490 cm in the infrared spectrum corresponding to FO2 [2] radicals. In the similar EPR studies the spectrum of this radical is also observed. Molecular oxygen is a common impurity in

fluorine gas and is difficult to remove by fractional distillation. After prolonged photolysis, the C2H2 bands are reduced by Therefore oxygen is always present in the samples in small ϳ2%; this corresponds to the fraction of binary complexes in Ϫ Ϫ concentration ϳ10 4–10 5. Since diffusing fluorine atoms re- the samples ϳ3 ϫ 10Ϫ4 (mole ratio) (3, 7). act with O2 molecules to form FO2 radicals (16), we have used Subsequent annealing of the samples at 24 K gives rise to this reaction as an internal standard for characterizing the growth of a new series of eight infrared bands at 643, 698, 785, reaction rate of diffusing F atoms during the annealing cycles. 1070, 1263, 1631, 3026, 3193 cmϪ1 (see Fig. 1c; see also Fig. EPR spectra of freshly prepared Ar:C2H2:F2 samples exhibit no 2 for the deuterated analog). A simultaneous consumption of lines due to paramagnetic species. Very weak infrared bands of bands of isolated C2H2(C2D2) molecules is also observed. The Ϫ1 Ϫ1 CO2 (661.9, 2340.5 cm ) are observed in the IR spectra of sharp band of FO2 radicals at 1490 cm grows in linear deposited samples. proportion to the new product bands and to the decrease of the Ϫ1 736.5 cm C2H2 band, as shown in Fig. 3. This provides convincing evidence that this series corresponds to a primary RESULTS AND DISCUSSION reaction product of diffusing thermal F atoms with isolated

C2H2. In accordance with the gas-phase reaction [1], we might A. Infrared Absorption Spectra of 2-C2H2F Radicals expect that both stereoisomers of 2-C2H2F radical and the ϭ The IR spectrum of Ar:F2:C2H2 3000:1:1 after deposition C2HOHF complex could be formed in the reaction of ther- is shown in Fig. 1a. It contains broad bands of acetylene with mally diffusing F atoms with C2H2. However, growth of the Ϫ maxima at 736.5, 1334.5, and 3288.5 (3302.0) cm 1. Brief UV characteristic absorption bands of HF (4000–3500 cmϪ1) and ϳ Ϫ1 photolysis ( 5 min) at 15 K leads to rapid growth of intense C2H radical (1845 cm )(14) is not observed, signifying that O bands in the regions of C F stretching vibration (1058, 1129, excited 2-C2H2F radicals are efﬁciently stabilized in the argon and 1153 cmϪ1) and HF molecular vibration (3756 and 3900 matrix and do not decay by HF elimination to an observable Ϫ1 cm ) (see Fig. 1b). These bands are attributed to C2H2F2 and extent. Each stereoisomer of 2-C2H2F radical has nine normal a molecular complex of C2HF bound to HF (17, 18). Such vibrations, and their most intense stretching vibrations COF molecular products could be generated in the cage reaction by and CAC have comparable calculated intensities (see Table 1). photolysis of the complex of reactants [C2H2 ϩ F2]: However, in each spectral region a single absorption band is

FIG. 2. Difference spectrum of photolyzed sample (Ar:F2:C2D2 ϭ 3000:1:1; C2H2-D2, 98 atom % D) before and after annealing at 24 K. IR bands of ␤-C2D2F are marked with asterisk. The spectrum was recorded at 15 K.

observed; hence, all of the new bands should be assigned to In samples containing higher F2 concentration (Ar:F2: one isomer of 2-C2H2F. Experiments were carried out in which C2H2 ϭ 3000:3:1) two additional series of product bands samples containing 2-C2H2F radicals were annealed at 30 K for appear upon annealing (see Fig. 4). The growth of these new 72 h, but no changes of the infrared spectra were observed. We bands exhibited a quadratic dependence with respect to the therefore conclude that thermally activated cis–trans isomer- growth of FO2 radicals, as shown in Fig. 3. This behavior is ization does not occur at this temperature. Thus, only one characteristic of secondary F atom addition products. The most stereoisomer of 2-C2H2F radical is formed in the reaction of intense bands of these secondary products are located at 877, Ϫ1 Ϫ1 diffusing F atoms with isolated C2H2. 1153 cm and at 1129, 1714 cm . These bands were earlier

Ϫ1 Ϫ1 Ϫ1 FIG. 3. Growth of (ᮀ) 2-C2H2F radical band (1631 cm ) and consumption of the (ϩ)C2H2 band (736.5 cm ) vs. intensity of the FO2 (1490 cm ) band Ϫ1 during annealing of the photolyzed sample Ar:F2:C2H2 ϭ 3000:1:1. Increasing intensity of IR band of secondary product cis-CHFACHF (1129 cm )inthe sample Ar:F2:C2H2 ϭ 3000:3:1 is marked with ‚. All intensities of IR bands are shown with respect to their maximum value.

TABLE 1 shown in Fig. 5c. At temperatures below 20 K, the spectrum ؊ Calculated Harmonic Frequencies (cm 1) and Isotropic hf exhibits a series of anisotropic bands that are consistent with Constants (mT) of Isomers of 2-Fluorovinyl Radical the spectrum previously reported by Cochran (11). At 30 K, the spectrum consists of eight narrow isotropic lines (the linewidths do not exceed 0.06 mT) having the same integrated intensities. Despite the different widths of the lines, all are

components of hf structure of EPR spectrum of 2-C2H2F rad-

ical. There are three doublet hf splittings in the spectrum: a 1 ϭ

6.50, a 2 ϭ 3.86, and a 3 ϭ 0.25 mT. The hf constants a 1 and

a 2 are close to the literature values for 2-ﬂuorovinyl radical

(11, 13). However, the smallest splittings a 3 were not observed in previous studies. It was determined earlier that isotropic hf constant of the proton of ␣-CH group is nearly zero (20). Thus absence of splitting on ␣-H in the EPR spectra may result from Note. The relative intensities of vibrations are given in parentheses. Calcu- insufﬁcient resolution of the spectra caused by high concen- lations were performed using the B3LYP/6-311ϩϩG(3df,2p) method and tration of the impurities or magnetic interactions with nuclei of basis. a the matrix. To demonstrate the inﬂuence of these effects on the This work. ϭ b IR data are extracted from (14), while EPR are extracted from (13). spectral resolution, we have studied samples Ar:F2:C2H2 83 300:1:1 and Kr:F2:C2H2 ϭ 3000:1:1 (nuclear spin of Kr

equals I ϭ 9/ 2). In the ﬁrst case (Fig. 5a), most of C2H2 assigned to trans- and cis-1,2-diﬂuoroethylene, respectively molecules are present in the sample as dimers (or larger clus-

(19): ters). Therefore 2-C2H2F radicals in this environment are

formed adjacent to C2H2 molecules, which cause line broad- ening due to incomplete averaging of anisotropic magnetic [3] interactions. In the second case (Fig. 5b), strong line broaden- ing results from interaction of unpaired electron with magnetic nuclei of the Kr matrix. Comparison of these spectra (5a and B. EPR Spectrum of 2-C H F Radical 2 2 5b) with the spectrum presented in the Fig. 5c shows that the

The EPR spectrum of a Ar:F2:C2H2 ϭ 3000:1:1 sample after highest spectral resolution is achieved in the argon matrix UV photolysis at 15 K and subsequent annealing at 24 K is (nuclear spin of Ar is equal to zero) under conditions of perfect

FIG. 4. Infrared spectra of secondary products in the Ar:F2:C2H2 ϭ 3000:3:1 sample: (a) after UV photolysis at ␭ ϭ 355 nm and partial annealing; (b) after • • successive visible light photolysis at ␭ ϭ 532 nm and ﬁnal annealing. Spectra were recorded at 15 K. ␤, IR bands of CHFA CH; ␣,CH2A CF; c, cis-CHFACHF; t, trans-CHFACHF; d,CH2ACF2 species.

FIG. 5. EPR spectra of samples after photolysis at 15 K and subsequent annealing at 24 K: (a) Ar:F2:C2H2 ϭ 300:1:1; (b) Kr:F2:C2H2 ϭ 3000:1:1; (c)

Ar:F2:C2H2 ϭ 3000:1:1. Spectra were recorded at 30 K. isolation of stabilized radicals only. Fulﬁllment of this condi- should form CHFA•CD and CDFA•CH radicals. Figure 6 tion made it feasible to observe the smallest splitting a 3 ϭ shows the EPR spectrum of such a sample after low-tempera- 0.25 mT. ture UV photolysis and subsequent annealing at 24 K. There In order to distinguish between the hf constants on H and F are three series of lines in the spectrum. The ﬁrst and dominant nuclei, we have carried out a series of similar experiments with group of 2-C2D2F radical contains six lines: a doublet (a 1 ϭ mixtures containing deuterium-enriched acetylene (75% C2D2, 6.50 mT) of 1:1:1 triplets (a 2 ϭ 0.59 mT). The second series • 22% C2HD, 3% C2H2). Addition of an F atom to C2D2 gives consists of four lines (CHFA CD): doublet (a 1 ϭ 6.50 mT) of • • rise to formation of CDFA CD, whereas addition to C2HD doublets (a 2 ϭ 3.86 mT). There are 12 lines (CDFA CH) in

FIG. 6. (a) EPR spectrum of a sample Ar:F2:(C2D2,(75%) ϩ C2HD (22%) ϩ C2H2 (3%)) ϭ 3000:1:1 after photolysis at 15 K and annealing at 24 K. The spectrum was recorded at 30 K. (b) Simulated spectrum of the sample containing radicals CDFA•CD (78%), CHFA•CD (11%), and CDFA•CH (11%), respectively. The hyperﬁne constants and linewidths are taken from the experimental spectrum (a).

Copyright © 2001 by Academic Press SPECTROSCOPY OF FLUOROVINYL RADICALS 275 the third group (six of them are masked by the more intense lines of the other series) formed from three hf splittings: two ϭ 1 ϭ splittings on nonequivalent nuclei with spin I 2 (a 1 6.50 and a 3 ϭ 0.25 mT) and one splitting on a nuclear with spin

I ϭ 1(a 2 ϭ 0.59 mT). Upon H/D substitution, the hf splittings on proton nuclei should decrease by a factor of

␥H/␥D Ϸ 6.5 times due to the gyromagnetic ratios of the proton and deuteron. However, the hf splitting on 19F should be unchanged. Thus, the unchanged splitting a 1 ϭ 6.50 mT 19 corresponds to hf splitting on F. The doublet splitting a 2 ϭ 3.86 mT in the radicals CHFA•CH and CHFA•CD and the • • triplet splitting a 2 ϭ 0.59 mT in CDFA CD and CDFA CH belong to hf splitting on the proton ␤-H. The remaining doublet • • splitting a 3 ϭ 0.25 mT (CHFA CH and CDFA CH) must then be ascribed to the ␣-H proton. The corresponding triplet A• A• splitting in the CDF CD and CHF CD radicals is less than FIG. 7. Difference IR spectrum of UV photolyzed and annealed sample the linewidth and could not be observed. Thus, in this study we before and after photolysis at ␭ ϭ 532 nm; (a) Ar/F2/C2H2; (b) Ar/F2/C2D2.IR were able to obtain a highly resolved EPR spectrum of bands of 2-C2H2F and 1-C2H2F are marked with ␤ and ␣, respectively. Spectra were recorded at 15 K. 2-C2H2F radical and to determine all its isotropic hf constants. There are several well-known cases in which small radicals and molecules rotate in rare gas matrices almost freely (21, 22). ␥ ␥ ␥ If a radical rotates incoherently, the observed changes in the where e is the gyromagnetic ratio of electron; F, H are the widths and intensities of the lines of EPR spectrum are ex- gyromagnetic ratios of nuclei of proton and ﬂuorine; and plained by incomplete averaging of anisotropic magnetic interactions. Fraenkel and Freed (23) proposed a model of rota- Y ͑␪Ј, ␸Ј͒ ͑k͒ ϭ ͱ ␲ ͳ ⌿ͯ 2k ͯ ⌿ʹ tional diffusion based on the Redﬁeld equations (24). This D i 6 /5 ͑ Ј͒ 3 r i model draws an association between the linewidth of components of the hf structure and the distribution of electron density is the dipole–dipole interaction of the nuclear magnetic mo- in the radical. In the case of a rapidly tumbling radical, the Ϫ ment with the unpaired electron, averaged over its electronic linewidth T 1 is described by 2 wavefunction, ⌿. Based on the observed difference in line- Ϫ1 width, ⌬T 2 Х 0.01 mT at 30 K, we have estimated the Ϫ Ϫ ␶ ϳ T 1 ϭ T 1 ϩ ͑ A ϩ ͸ B m ϩ ͸ C m 2 ϩ ͸ E m m ͒␶ , rotational correlation time from Eqs. [4] and [5] to be R 2 2,0 i i i i ij i j R Ϫ9 i i iϽj 10 s. [4] C. Photoinduced Conversions of 2-C2H2F Radicals

Ϫ1 Exhaustive photolysis of 2-C2H2F radicals at ␭ ϭ 532 nm where T 2,0 incorporates all contributions to linewidth that are causes the intensities of the associated infrared absorption not related to rotation; m i, m j represent the z components of bands to decrease by about half. This change is accompanied nuclear spin quantum number for nuclei i and j; ␶R is the rotational correlation time; and the coefficients A, B, C, E are by growth of a new series of six bands at 798(803), 930, 1133, Ϫ1 determined by anisotropic part of the hf interactions and g 1355, 1654, 2778 cm (see Fig. 7a). In samples containing • tensors. Since the EPR spectrum of CHFA CH radical is 2-C2D2F radicals, the new bands are appeared at 630, 797, Ϫ1 ϭ 1 973, 1138, 1625, 2221 cm (see Fig. 7b). This photoconver- symmetric and nuclear spins of all magnetic nuclei are I 2, only the cross term E m m contributes to the difference in the sion was found to be reversible: subsequent short-time UV ij i j ␭ ϭ linewidths. In the spectrum shown in Fig. 5c, lines with dif- photolysis 355 nm leads to disappearance of new bands and to complete recovery of 2-C2H2F radical bands. Figure 8 ferent signs of m im j product for F and ␣-H nuclei exhibit illustrates the kinetics of photoconversion under successive different widths. Therefore only the term E FHm Fm H contributes visible light and UV photolysis. From this figure, it is evident to the difference in linewidths. The coefficient E FH arises from multiplication of hf interaction tensors (23, 25) that changes in the intensities of infrared bands of 2-C2H2F radical and new series of bands occur with the same photochemical rate. No changes are observed in intensities of others ϭ 8 k 2 infrared bands. 2 2 ͑k͒ ͑Ϫk͒ E ␶ ϭ ប ␥ ␥ ␥ ͸ D D , [5] Such a reversible photoconversion could be explained either FH R 15 e F H F H kϭϪ2 by photoinduced cis–trans isomerization or by a structural

FIG. 8. Variation in IR band intensities of 2-C2H2F and 1-C2H2F radicals upon successive photolysis at ␭ ϭ 532 nm and ␭ ϭ 355 nm. Single exponential ﬁts of experimental data are shown by solid curves. Characteristic times of 532- and 355-nm photolysis are equal to 715 and 12 min, respectively, at a photon ﬂux of 1.1 ϫ 1016 cmϪ2 sϪ1.

isomerization 2-C2H2F 7 1-C2H2F. It can be seen from Fig. 7 complete (as monitored by growth of the FO2 band). Under Ϫ1 that bands in the region of the COF (1133 cm ) and CAC these conditions ﬂuorine atoms reacted not only with C2H2 Ϫ1 (1654 cm ) stretching vibrations in the IR spectrum of the molecules and 2-C2H2F radicals, but also with the new new radical have the greatest intensity, similar to the spectrum 1-C2H2F radicals to form a new series of four bands at 806, Ϫ1 of 2-C2H2F radical. However, in the infrared spectrum of 926, 1293, 1727(1741) cm that are in excellent agreement

2-C2H2F radical, the intensity of the COF stretching band is (apart from minor matrix shifts) with the vibrational frequen- nearly 50% higher than the CAC stretching band, whereas the cies of 1,1-difluoroethylene (19). This result definitively con- new radical exhibits just the opposite behavior. In the COH firms that the new infrared spectrum that appears under 532 nm Ϫ1 stretching region, the only observed band lies at 2778 cm , photolysis of 2-C2H2F radicals corresponds to 1-C2H2F radical. and its intensity is smaller than that of CH␣ vibration in 2-C2H2F radical by a factor of 10. These observations make it TABLE 2 ؊ difficult to assign the new radical species to trans-2-C2H2F and Harmonic Frequencies (cm 1) and Isotropic hf Constants instead favors an assignment to 1-C2H2F. A summary of the (mT) of Isomers of 2-Fluorovinyl-D2 Radical calculated vibrational frequencies and intensities are given in

Table 1 (2-C2H2F), Table 2 (2-C2D2F), and Table 3 (1-C2H2F and 1-C2D2F).

The assignment of the new radical to 1-C2H2F was con- ﬁrmed in experiments where F atoms were allowed to react with it to form 1,1-diﬂuoroethylene

A • ϩ 3 A CH2 CF F CH2 CF2. [6]

To accumulate a sufﬁcient concentration of 1-C2H2F radicals, a partial annealing of a photolyzed sample at 24 K was performed. The temperature was then lowered to 15 K and the ␭ ϭ sample was subjected to a prolonged photolysis at 532 nm Note. The relative intensities of vibrations are given in parentheses. to convert half of the 2-C2H2F radicals to 1-C2H2F. The sample Calculations were performed using the B3LYP/6-311ϩϩG(3df,2p) was then annealed until reactions of diffusing F atoms were method and basis.

TABLE 3 (Calculated Harmonic Frequencies (cm؊1) and Isotropic hf Constants (mT

for 1-Fluorovinyl and 1-Fluorovinyl-D2 Radicals

Note. The relative intensities of vibrations are given in parentheses. Calculations were performed using the B3LYP/6-311ϩϩG(3df,2p) method and basis.

We performed similar EPR experiments and observed the by the F atom at the radical center. Based on H/D isotopic reversible 532 nm photoconversion of 2-C2H2F radicals to substitution we have determined hf splittings of 1-C2H2F:

1-C2H2F radical. The EPR spectrum of the sample after com- ͉A Ќ(F)͉ ϭ 12.04, ͉A ʈ(F)͉ ϭ 17.07, (a F ϭ [2 ⅐ A Ќ(F) ϩ A ʈ(F)]/ pletion of visible light photolysis is shown in Fig. 9. It consists 3 ϭ 13.71), a H1 ϭ 4.21, a H2 ϭ 1.16 mT. Notice that the of three doublet splittings which correspond to the hf structure isotropic hf constant on 19F in this radical is more than twice as ϭ 1 from three nonequivalent magnetic nuclei with spin I 2. large as in 2-C2H2F radical. However, the new spectrum is more complex than EPR spectrum of CHFA•CH radical, since one of the hf splittings has an D. Quantum-Chemical Calculations of Energy and anisotropic component. The fact that this anisotropy is ob- Spectroscopic Characteristics of Fluorovinyl Radicals served in 1-C2H2F and not 2-C2H2F is consistent with the All of calculations were performed using GAUSSIAN 98 relatively large contribution of p electron character introduced suite of codes (26). It was demonstrated earlier that density

• • FIG. 9. (a) EPR spectrum of CH2A CF radicals generated by 532-nm photolysis of CHFA CH radicals. The spectrum was recorded at 30 K. Lines of CHFA•CH radicals are marked with diamonds. (b) Simulated spectrum using the experimentally measured linewidths and hyperﬁne splittings, as indicated.

TABLE 4 Optimized Geometries (Å and degrees) of Isomers of 2-Fluorovinyl Radical

functional (DF) theory calculations are able to accurately re- us to conclude that spectroscopic characteristics of stabilized produce geometric, thermodynamic, and spectroscopic param- radicals formed in the solid-state reaction of mobile F atoms eters of radicals at low computational cost (15, 27). Compar- with C2H2 molecules correspond to those of cis isomer of ison with conventional post-Hartree–Fock methods shows that 2-C2H2F radical. DF results are competitive with QCISD results for geometries Table 5 shows that the calculated energy of the cis isomer is and harmonic frequencies and with coupled cluster computa- lower than that of the trans isomer in all but the lowest level tions of hf interactions. We have used DF method B3LYP with calculation. The energy difference is enhanced when the level 6-311ϩϩG(d,p) and 6-311ϩϩG(3df,2p) basis sets. There of theory is increased and when zero-point vibrational energy are no experimental data of geometries of fluorovinyl radicals. (ZPE) is taken into account. At the highest level of theory the To assess the quality of the geometry predictions we have used difference in the energies of isomers is 1.0 kJ/mol (1.5 kJ/mol ϩϩ results of computations at the QCISD/6-311 G(d,p) level with ZPE). Although such difference is comparable with the of theory. The optimized geometries of stereoisomers of 2-flu- accuracy of energy calculations of isomers, the observed trends orovinyl are compared for the different levels of calculations in clearly indicate that the cis isomer is more stable then trans Table 4. All methods gave similar bond lengths and bond isomer. We have calculated the barrier of cis–trans isomeriza- angles, but the optimized geometry using the B3LYP/6- tion; its height is only 9.0 kJ/mol. Thus, the equilibrium pop- ϩϩ 311 G(3df,2p) method is in better agreement with the ulation of state of trans isomer at low temperatures T Յ 30 K QCISD calculations. is extremely small. This conclusion is supported by our inabil- The next series of calculations was to obtain spectroscopic ity to detect the trans isomer (i.e., its population must be at characteristics of the cis and trans isomers of 2-C H F radicals. 2 2 least 20 times less than the cis isomer). From this limit, we The results of these calculations are compared with the exper- calculate that the enthalpy of the trans isomer must be at least imental results in Table 1, and for deuterated radicals in Table 0.7 kJ/mol higher than cis. 2. Agreement of the measured and calculated frequencies is The calculated spectroscopic features of 1-C H F radical somewhat better for the cis isomer than for trans isomer. 2 2 at B3LYP/6-311ϩϩG(3df,2p) level of theory are presented However, a definitive assignment based only on these frequencies is not possible because the calculations do not account for in Table 3. The calculated hf constants for 1-C2H2F radical effects of anharmonicity or frequency shifts due to the matrix are in excellent agreement with the experiments. Therefore, environment. Table 1 illustrates that even slight changes in we have obtained dependable vibrational frequencies and hf experimental conditions (i.e., Jacox’s results (14) compared constants of 1-C2H2F radical stabilized in solid argon for the with this paper) can result in measurable differences in band first time. positions. Fortunately, a definitive assignment is possible based on a comparison of theoretical and measured isotropic hf constants. TABLE 5 ؊ ؍ ⌬ The calculations predict that the hf constants on F and ␤-H Calculated Difference in Electronic Energy, E0 E(trans) nuclei in cis and trans isomers strongly differ (values of a and E(cis), for 2-Fluorovinyl Radical Isomers with and without Zero- F Point Energy Correction a ␤H differ more than 1.5 and 2 times, respectively). Good agreement of the calculated and measured values is obtained only for the cis isomer. Our previous studies (see (3, 4, 8)) and results of computations (15, 28) show that B3LYP method with appropriate basis set generally reproduces isotropic hf constants with ϳ5% accuracy. Therefore comparison of the calculated and measured hf constants of 2-C2H2F radicals allows

CONCLUSIONS 10. R. L. Williams and F. S. Rowland, J. Phys. Chem. 77, 301–310 (1973). 11. E. L. Cochran, F. J. Adrian, and V. A. Bowers, J. Phys. Chem. 74, In this study we have shown that unambiguous identification 2083–2089 (1970). of the isomeric form of fluorovinyl radicals can be accom- 12. E. L. Cochran, F. J. Adrian, and V. A. Bowers, J. Chem. Phys. 40, 213–219 (1964). plished using a combination of two complementary spectro- 13. M. Shiotani, Y. Nagata, and J. Sohma, J. Phys. Chem. 86, 4131–4137 scopic techniques, EPR and IR, along with quantum chemical (1982). calculations. The EPR technique plays the crucial role in the 14. M. E. Jacox, Chem. Phys. 53, 307–322 (1980). identification, since hyperfine constants are very sensitive to 15. V. Barone, C. Adamo, and N. Russo, Chem. Phys. Lett. 212, 5–11 the geometry distortion of the open-shell species, and they can (1993). 16. E. Ya. Misochko, A. V. Akimov, and C. A. Wight, Chem. Phys. Lett. 274, be precisely predicted by quantum-chemistry calculations. Re- 23–28 1997. liable assignment of the vibrational frequencies of cis-2-C2H2F 17. C. Zetsch, Ph.D. dissertation, Georg-August University, Go¨ttingen, Ger- and 1-C2H2F radicals is made possible by the fact that the same many, 1971. reversible photoconversion of the two species is observed in 18. G. R. Hunt and M. K. Wilson, J. Chem. Phys. 34, 1301 (1961). both the EPR and IR experiments. 19. NIST Mass Spec Data Center, in “NIST Chemistry WebBook, NIST Standard Reference Database Number 69” (W. G. Mallard and P. J. Linstrom, Eds.), February 2000, National Institute of Standards and Tech- ACKNOWLEDGMENTS nology, Gaithersburg, MD. [http://webbook.nist.gov] 20. H. M. McConnel and J. Strathdee, Mol. Phys. 2, 129–138 (1959). This research was supported by the U.S. National Science Foundation 21. M. E. Jacox, J. Mol. Spectrosc. 66, 272–287 (1977). (Grant CHE-9970032) and the Russian Foundation for Basic Research (Grant 22. C. Girardet, L. Abouaf-Marguin, B. Gauthier-Roy, and D. Maillard, Chem. 98-03-33175). Phys. 89, 415–430 (1984); Chem. Phys. 89, 431–443 (1984). 23. J. H. Freed and G. K. Fraenkel, J. Chem. Phys. 39, 326–348 (1963). REFERENCES 24. A. G. Redfield, IBM Res. Dev. 1, 19–29 (1957). 25. J. H. Freed and G. K. Fraenkel, J. Chem. Phys. 40, 1815–1829 (1963). 1. W. E. Jones and E. G. Skolnik, Chem. Rev. 76, 563–591 (1976). 26. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, 2. M. E. Jacox, Rev. Chem. Intermed. 6, 77–120 (1985). J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., R. E. Strat- 3. E. Ya. Misochko, V. A. Benderskii, A. U. Goldschleger, A. V. Akimov, A. V. mann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Benderskii, and C. A. Wight, J. Chem. Phys. 106, 3146–3156 (1997). Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. 4. E. Ya. Misochko, I. U. Goldschleger, A. V. Akimov, and C. A. Wight, Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, Low Temp. Phys. 26, 727–735 (2000). G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. 5. J. Feld, H. Kunttu, and V. A. Apkarian, J. Chem. Phys. 92, 1009–1020 Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B. (1990). Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, 6. E. Ya. Misochko, A. V. Akimov, and C. A. Wight, Chem. Phys. Lett. 293, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. 547–554 (1998). Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill, B. Johnson, 7. E. Ya. Misochko, A. V. Benderskii, and C. A. Wight, J. Phys. Chem. 100, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. 4496–4502 (1996). Replogle, and J. A. Pople, “GAUSSIAN 98, Revision A.3,” Gaussian, 8. E. Ya. Misochko, V. A. Akimov, I. U. Goldschleger, A. I. Boldyrev, and Inc., Pittsburgh, PA, 1998. C. A. Wight, J. Am. Chem. Soc. 121, 405–410 (1999). 27. F. De Proft, J. M. L. Martin, and P. Geerlings, Chem. Phys. Lett. 250, 9. G. A. Kapralova, T. P. Nagornaya, and A. M. Chaikin, Kinet. Katal. 11, 393–397 (1996). 809–820 (1970); Kinet. Katal. 667–670 (1970). 28. V. Barone, J. Chem. Phys. 101, 6834–6838 (1994).