Investigating the Photodissociation Dynamics of CF2BrCF2I in CCl4 through Femtosecond Time-Resolved Infrared Spectroscopy

The photodissociation dynamics of CF2BrCF2I in CCl4 at 280 ± 2 K were investigated by probing the C−F stretching mode from 300 fs to 10 μs after excitation at 267 nm using time-resolved infrared spectroscopy. The excitation led to the dissociation of I or Br atoms within 300 fs, producing the CF2BrCF2 or CF2ICF2 radicals, respectively. All nascent CF2ICF2 underwent further dissociation of I, producing CF2CF2 with a time constant of 56 ± 5 ns. All nascent g-CF2BrCF2 isomerized into the more stable a-CF2BrCF2 with a time constant of 47 ± 5 ps. Furthermore, a-CF2BrCF2 underwent a bimolecular reaction with either itself (producing CF2BrCF2Br and CF2CF2) or Br in the CCl4 solution (producing CF2BrCF2Br) at a diffusion-limited rate. The secondary dissociation of Br from a-CF2BrCF2 was significantly slow to compete with the bimolecular reactions. Overall, approximately half of the excited CF2BrCF2I at 267 nm produced CF2BrCF2Br, whereas the other half produced CF2CF2. The excess energies in the nascent radicals were thermalized much faster than the secondary dissociation of I from CF2ICF2 and the observed bimolecular reactions, implying that the secondary reactions proceeded under thermal conditions. This study further demonstrates that structure-sensitive time-resolved infrared spectroscopy can be used to study various reaction dynamics in solution in real time.

Vicinal diiodoethanes, such as CH 2 ICH 2 I, CF 2 ICH 2 I, and CF 2 ICF 2 I, were mainly investigated because the iodoethyl radicals produced upon UV excitation contain weak C−I bonds that readily proceed with secondary dissociation [1][2][3]5,7,8,10,12]. In the case of the photodissociation of CF 2 ICF 2 I in the gas phase, a fraction of the iodoethyl radicals dissociate the secondary C−I bond with a time constant of 26 ± 7 ps to form CF 2 CF 2 and I [2]. Secondary dissociation of haloethyl radicals often occurs because of the weakness of the halogen-carbon bonds; however, these bonds strengthen upon fluorination. The C−I bond is weaker in CH 2 ICH 2 than in CF 2 ICF 2 because the π-bond formed concertedly upon the fission of the C−I bond is stronger in the case of CH 2 ICH 2 [7]. The primary isomerization, and the former is expected to survive sufficiently long to participate in a BR involving the radical. Therefore, comparative reaction dynamics between the two radicals can be probed by observing the entire reaction dynamics of CF 2 BrCF 2 I excited at 267 nm in a CCl 4 solution. Owing to the sensitivity of IR spectroscopy to molecular structure, the conformer-specific reactions of the radicals obtained from the photodissociation of CF 2 BrCF 2 I in solution can be observed in detail, similar to the case of the photodissociation of CF 2 ICF 2 I in solution [8].
In this study, we investigated all the subsequent reactions after the photodissociation of CF 2 BrCF 2 I in a CCl 4 solution at 267 nm through TRIR spectroscopy for up to 10 µs. With the support of quantum-chemical calculations and available spectra, all the transient absorptions in the TRIR spectra of CF 2 BrCF 2 I in CCl 4 were assigned, and their kinetics were determined. Upon excitation, CF 2 BrCF 2 I immediately produces a-CF 2 BrCF 2 , g-CF 2 BrCF 2 , and a-CF 2 ICF 2 [15]. Although a-CF 2 ICF 2 undergoes a simple dissociation of the C−I bond, most of the g-CF 2 BrCF 2 isomerize into a-CF 2 BrCF 2 which undergoes a bimolecular reaction with itself or the dissociated Br atom in the solution. Structure-sensitive TRIR spectroscopy enabled us to determine the real-time reaction dynamics of UV-excited CF 2 BrCF 2 I in CCl 4 at 280 K.

Results and Discussion
The equilibrium electronic spectrum of CF 2 BrCF 2 I dissolved in CCl 4 showed a weak absorption band at 274 nm and a strong band above 200 nm (Figure 1), which was slightly red-shifted from the peaks at 268 and 193 nm in the gas phase [9]. The weak absorption at 274 nm, assigned to the n → σ* transition of the C−I bond, leads to the immediate dissociation of the I atom [1][2][3]12], while the strong absorption above 200 nm, assigned to the n → σ* transition of the C−Br bond, leads to the immediate dissociation of the Br atom [9]. As the wing of this strong absorption band extends up to 300 nm and accounts for 18 ± 3% of the absorption at 267 nm, the photoexcitation of CF 2 BrCF 2 I at 267 nm can lead to the dissociation of not only the I atom but also the Br atom. Specifically, 82 ± 3% of the absorption intensity at 267 nm arises from the weak absorption band at 274 nm; thus, excitation at 267 nm would promote 82 ± 3% (18 ± 3%) of CF 2 BrCF 2 I to the antibonding state of the C−I (C−Br) bond. Consequently, 82 ± 3% (18 ± 3%) of the excited CF 2 BrCF 2 I at 267 nm is expected to dissociate into CF 2 BrCF 2 + I (CF 2 ICF 2 + Br). The contribution from the strong band to the absorption at 267 nm in the gas phase was estimated to be 9 ± 4% [9], which would result in 9 ± 4% CF 2 ICF 2 production upon excitation at 267 nm in the gas phase, which is smaller than that in CCl 4 . tion, and the former is expected to survive sufficiently long to participate in a BR inv the radical. Therefore, comparative reaction dynamics between the two radicals probed by observing the entire reaction dynamics of CF2BrCF2I excited at 267 nm in solution. Owing to the sensitivity of IR spectroscopy to molecular structure, th former-specific reactions of the radicals obtained from the photodissociation of CF2 in solution can be observed in detail, similar to the case of the photodissociat CF2ICF2I in solution [8].
In this study, we investigated all the subsequent reactions after the photodissoc of CF2BrCF2I in a CCl4 solution at 267 nm through TRIR spectroscopy for up to 10 μs the support of quantum-chemical calculations and available spectra, all the transie sorptions in the TRIR spectra of CF2BrCF2I in CCl4 were assigned, and their kinetic determined. Upon excitation, CF2BrCF2I immediately produces a-CF2BrCF2, g-CF2 and a-CF2ICF2 [15]. Although a-CF2ICF2 undergoes a simple dissociation of the C−I most of the g-CF2BrCF2 isomerize into a-CF2BrCF2 which undergoes a bimolecular re with itself or the dissociated Br atom in the solution. Structure-sensitive TRIR spe copy enabled us to determine the real-time reaction dynamics of UV-excited CF2BrC CCl4 at 280 K.

Results and Discussion
The equilibrium electronic spectrum of CF2BrCF2I dissolved in CCl4 showed a absorption band at 274 nm and a strong band above 200 nm (Figure 1), which was s red-shifted from the peaks at 268 and 193 nm in the gas phase [9]. The weak absorp 274 nm, assigned to the n → σ* transition of the C−I bond, leads to the immediate d ation of the I atom [1][2][3]12], while the strong absorption above 200 nm, assigned to → σ* transition of the C−Br bond, leads to the immediate dissociation of the Br ato As the wing of this strong absorption band extends up to 300 nm and accounts for 1 of the absorption at 267 nm, the photoexcitation of CF2BrCF2I at 267 nm can lead dissociation of not only the I atom but also the Br atom. Specifically, 82 ± 3% of t sorption intensity at 267 nm arises from the weak absorption band at 274 nm; thus tation at 267 nm would promote 82 ± 3% (18 ± 3%) of CF2BrCF2I to the antibondin of the C−I (C−Br) bond. Consequently, 82 ± 3% (18 ± 3%) of the excited CF2BrCF2I nm is expected to dissociate into CF2BrCF2 + I (CF2ICF2 + Br). The contribution fro strong band to the absorption at 267 nm in the gas phase was estimated to be 9 ± 4 which would result in 9 ± 4% CF2ICF2 production upon excitation at 267 nm in t phase, which is smaller than that in CCl4.  The TRIR spectra of CF 2 BrCF 2 I in CCl 4 at 280 ± 2 K were measured in the spectral region of 1370−1020 (1020−870) cm −1 that covers all of the C−F stretching modes of the molecule over a broad time span from 0.3 ps to 10 µs (1 ns) after excitation at 267 nm, encompassing the entire excitation-induced reaction. As the photodissociation reaction was found to be complete at 1 µs in previous experiments on similar dihaloalkanes in solution [8,10], the time range of 0.3 ps to 10 µs was expected to be sufficient to observe the fates of all of the intermediates and products induced by the photoexcitation of CF 2 BrCF 2 I in CCl 4 . The absorption bands in the 1300−1100 cm −1 region were congested, while those in the 1020−870 cm −1 region (vide infra) were well resolved. Thus, the TRIR spectra in the 1020−870 cm −1 region were collected up to 1 ns to confirm the band assignment [15] and dynamics of the CF 2 BrCF 2 radical (vide infra), even though this was time-consuming because of the inferior sensitivity of our TRIR spectrometer in such a long-wavelength spectral region.
As shown in Figure 2b, negative absorption bands (bleach) appeared immediately after excitation at the positions of the absorption bands of the equilibrium spectrum of CF 2 BrCF 2 I. The bleach signal arises from the depletion of the population of CF 2 BrCF 2 I in the ground state upon excitation. The four main absorption bands (negative absorption) of CF 2 BrCF 2 I near 1225, 1172, 1115, and 994 cm −1 appeared at 0.3 ps and maintained their magnitude up to 10 µs. This suggests that the photoreaction proceeds within 0.3 ps and none of the nascent photofragments return to the reactant for up to 10 µs. New absorption bands can be categorized into three groups according to their kinetics: (1) bands appearing immediately after excitation, (2) bands growing around 50 ps and decaying near 50 ns, and (3) bands growing at~50 ns and maintaining their amplitudes up to 10 µs, which is the last pump-probe delay time in our experiment. This suggests that the nascent products may undergo secondary reactions to produce intermediates that can further react to generate the final photoproducts. The absorption bands immediately appearing after excitation (group (1)) were assigned to the nascent photoproducts [15] and those maintained up to 10 µs after growing near 50 ns to the final products produced by secondary reactions of the reaction intermediates. Absorption bands near 1278 and 1128 cm −1 grew near 50 ps and decayed around 50 ns; thus, they were assigned to the reaction intermediates that produced the final products through the secondary reactions. The immediately appearing absorption bands near 1260, 943, and 888 cm −1 were initially broad and shifted toward blue with time, suggesting that the nascent photofragments were produced with excess energy that relaxed via the anharmonically-coupled lower-frequency modes (thermal relaxation) [15][16][17][18]. The absorption bands at 1324 and 1175 cm −1 , assigned to CF 2 CF 2 in CCl 4 [8,13,14,19], grew near 50 ns and maintained their intensity up to 10 µs, indicating that CF 2 CF 2 is the final product formed in tens of nanoseconds.
Notably, TRIR spectra beyond 1 µs did not evolve with time, indicating the completion of the photoreaction. Thus, new absorption bands in the TRIR spectra beyond 1 µs should arise from the final products and bleach the signal from the depleted reactant. The TRIR spectra from 1 to 10 µs were averaged to obtain more reliable spectra for the final products. As shown in Figure 3c, the averaged TRIR spectrum overlapped well with the difference spectrum obtained by subtracting the absorption spectrum measured before the pump-probe experiment from that measured after the experiment, confirming that the photoreaction was complete by~1 µs. Apart from the absorption spectrum of CF 2 CF 2 and the inverted absorption spectrum of CF 2 BrCF 2 , the averaged TRIR spectrum contained an additional absorption spectrum, which was assigned to CF 2 BrCF 2 Br based on the reported spectrum of CF 2 BrCF 2 Br in CS 2 (Figure 3a) [20]. Clearly, the photodecomposition of CF 2 BrCF 2 I leads to two products: CF 2 CF 2 and CF 2 BrCF 2 Br. The decomposition of the averaged TRIR spectrum suggests that the photoexcitation of CF 2 BrCF 2 I in CCl 4 at 267 nm produces CF 2 CF 2 (50 ± 3%) and CF 2 BrCF 2 Br (50 ± 3%). Notably, TRIR spectra beyond 1 μs did not evolve with time, indicating the c tion of the photoreaction. Thus, new absorption bands in the TRIR spectra beyon should arise from the final products and bleach the signal from the depleted reacta TRIR spectra from 1 to 10 μs were averaged to obtain more reliable spectra for th products. As shown in Figure 3c, the averaged TRIR spectrum overlapped well w difference spectrum obtained by subtracting the absorption spectrum measured (b) Contour plot of the TRIR spectra of CF 2 BrCF 2 I in CCl 4 at 280 ± 2 K obtained after excitation of CF 2 BrCF 2 I at 267 nm. The data in the spectral regions of 1370−1020 cm −1 and 1020−870 cm −1 were collected from 0.3 ps to 10 µs and from 0.3 ps to 1 ns, respectively. Apparent bands are marked by arrows, and their peak positions are also mentioned. Dotted lines at 50 ps, 50 ns, and 1 µs guide the eye to analyze the apparent kinetics of transient signals. The absorbance difference, ∆A, was obtained by subtracting the absorbance of the sample before excitation from that after excitation. The absorbance is given in optical density (OD), where 1 mOD = 10 −3 OD. (c) Contour plot of the fitted TRIR spectra of CF 2 BrCF 2 I in CCl 4 (see text). averaged TRIR spectrum suggests that the photoexcitation of CF2BrCF2I in CCl4 at 267 nm produces CF2CF2 (50 ± 3%) and CF2BrCF2Br (50 ± 3%). When CF2BrCF2I is excited by UV light, it dissociates I or Br atoms, producing CF2BrCF2 or CF2ICF2 radicals, respectively [4,9,10,15]. As mentioned, because 82 ± 3% (18 ± 3%) of the excited CF2BrCF2I in CCl4 at 267 nm is promoted to the antibonding of the C−I (C−Br) bond, the majority of the produced radicals are expected to be CF2BrCF2. As shown in our previous TRIR spectroscopic experiment on CF2BrCF2I in the time range of 0.3−320 ps, the TRIR spectra consisted of the spectra of three nascent radicals (a-CF2BrCF2, g-CF2BrCF2, and a-CF2ICF2) and the inverted spectrum of CF2BrCF2I for the bleach signal [15]. The final photoproducts were 50 ± 3% CF2CF2 and 50 ± 3% CF2BrCF2Br, which were produced via secondary reactions of the reaction intermediates. Therefore, the TRIR spectra should include the spectra of the reactant (CF2BrCF2I), final products (CF2CF2 and CF2BrCF2Br), and reaction intermediates (a-CF2BrCF2, g-CF2BrCF2, and a-CF2ICF2). The TRIR spectra were globally fitted using the basis spectra of CF2BrCF2I, CF2CF2, CF2BrCF2Br, a-CF2BrCF2, g-CF2BrCF2, and a-CF2ICF2, as shown in the upper panel of Figure  4. The basis spectra of CF2BrCF2I, CF2CF2, and CF2BrCF2Br were obtained from equilibrium FTIR measurements, and those of a-CF2ICF2, a-CF2BrCF2, and g-CF2BrCF2 were obtained from our previous experiments [8,15]. The decomposition of CF2BrCF2I and CF2BrCF2Br into their conformer-specific spectra was not necessary for the fitting because  [20] (magenta line). The spectrum is well described as the sum of the absorption spectra of a-CF 2 BrCF 2 Br (dashed blue line) and g-CF 2 BrCF 2 Br (dashed orange line) using density functional theory (DFT) calculations. (b) Equilibrium FTIR spectrum of CF 2 CF 2 in CCl 4 [8] (crosses). The spectrum is well described by two absorption bands (green line). (c) Averaged spectrum (open circles) of the TRIR spectra of CF 2 BrCF 2 I from 1 to 10 µs overlapped with the equilibrium difference spectrum (gray line) obtained by subtracting the absorption spectrum measured before the pumpprobe experiment from that after the experiment. The averaged spectrum corresponds well with the sum (black line) of the three equilibrium spectra of CF 2 BrCF 2 I (purple line), CF 2 CF 2 (green line), and CF 2 BrCF 2 Br (magenta line). Positions of the bands are given in the figure.
When CF 2 BrCF 2 I is excited by UV light, it dissociates I or Br atoms, producing CF 2 BrCF 2 or CF 2 ICF 2 radicals, respectively [4,9,10,15]. As mentioned, because 82 ± 3% (18 ± 3%) of the excited CF 2 BrCF 2 I in CCl 4 at 267 nm is promoted to the antibonding of the C−I (C−Br) bond, the majority of the produced radicals are expected to be CF 2 BrCF 2 . As shown in our previous TRIR spectroscopic experiment on CF 2 BrCF 2 I in the time range of 0.3−320 ps, the TRIR spectra consisted of the spectra of three nascent radicals (a-CF 2 BrCF 2 , g-CF 2 BrCF 2 , and a-CF 2 ICF 2 ) and the inverted spectrum of CF 2 BrCF 2 I for the bleach signal [15]. The final photoproducts were 50 ± 3% CF 2 CF 2 and 50 ± 3% CF 2 BrCF 2 Br, which were produced via secondary reactions of the reaction intermediates. Therefore, the TRIR spectra should include the spectra of the reactant (CF 2 BrCF 2 I), final products (CF 2 CF 2 and CF 2 BrCF 2 Br), and reaction intermediates (a-CF 2 BrCF 2 , g-CF 2 BrCF 2 , and a-CF 2 ICF 2 ). The TRIR spectra were globally fitted using the basis spectra of CF 2 BrCF 2 I, CF 2 CF 2 , CF 2 BrCF 2 Br, a-CF 2 BrCF 2 , g-CF 2 BrCF 2 , and a-CF 2 ICF 2 , as shown in the upper panel of Figure 4. The basis spectra of CF 2 BrCF 2 I, CF 2 CF 2 , and CF 2 BrCF 2 Br were obtained from equilibrium FTIR measurements, and those of a-CF 2 ICF 2 , a-CF 2 BrCF 2 , and g-CF 2 BrCF 2 were obtained from our previous experiments [8,15]. The decomposition of CF 2 BrCF 2 I and CF 2 BrCF 2 Br into their conformer-specific spectra was not necessary for the fitting because their basis spectra have contributions from both conformers, and these contributions were maintained throughout the experiment. As shown in Figure 2c and the lower panel of Figure 4, the TRIR spectra were well reproduced by the sum of the basis spectra for CF 2 BrCF 2 I, CF 2 CF 2 , CF 2 BrCF 2 Br, a-CF 2 BrCF 2 , g-CF 2 BrCF 2 , and a-CF 2 ICF 2 shown in the upper panel of Figure 4. The time-dependent amplitude changes in the basis spectra were obtained by global fitting of the TRIR spectra. As mentioned, the magnitude of the bleach spectrum did not change throughout the experimental time span, implying that there was no rebinding of the dissociated halogen atom with its counter radical for up to 10 µs. The amplitudes of the remaining five basis spectra revealed rich kinetic information related to the reaction dynamics of the 267-nm-excited CF 2 BrCF 2 I in CCl 4 at 280 K. their basis spectra have contributions from both conformers, and these contributions were maintained throughout the experiment. As shown in Figure 2c and the lower panel of Figure 4, the TRIR spectra were well reproduced by the sum of the basis spectra for CF2BrCF2I, CF2CF2, CF2BrCF2Br, a-CF2BrCF2, g-CF2BrCF2, and a-CF2ICF2 shown in the upper panel of Figure 4. The time-dependent amplitude changes in the basis spectra were obtained by global fitting of the TRIR spectra. As mentioned, the magnitude of the bleach spectrum did not change throughout the experimental time span, implying that there was no rebinding of the dissociated halogen atom with its counter radical for up to 10 μs. The amplitudes of the remaining five basis spectra revealed rich kinetic information related to the reaction dynamics of the 267-nm-excited CF2BrCF2I in CCl4 at 280 K. The amplitude of a basis spectrum (amp) is related to the population of the corresponding compound by ∝ [15,19], where ε and n represent the integrated extinction coefficient and population of the compound, respectively. The integrated The amplitude of a basis spectrum (amp) is related to the population of the corresponding compound by amp ∝ ε × n [15,19], where ε and n represent the integrated extinction coefficient and population of the compound, respectively. The integrated extinction coefficient (ε) of a-CF 2 BrCF 2 , g-CF 2 BrCF 2 , and a-CF 2 ICF 2 were obtained from our previous measurements [8,15], and those of CF 2 BrCF 2 I, CF 2 CF 2 , and CF 2 BrCF 2 Br were determined from equilibrium FTIR measurements. Time-dependent fractional population changes of CF 2 BrCF 2 I, CF 2 CF 2 , CF 2 BrCF 2 Br, a-CF 2 BrCF 2 , g-CF 2 BrCF 2 , and a-CF 2 ICF 2 were derived from time-dependent amplitude changes for the basis spectra of the corresponding compounds using the relation, n ∝ amp/ε . As shown in Figure 5, three nascent radicals (a-CF 2 BrCF 2 , g-CF 2 BrCF 2 , and a-CF 2 ICF 2 ) appeared immediately, implying that the photodissociation of CF 2 BrCF 2 I occurred within 0.3 ps [15]. The decay of g-CF 2 BrCF 2 correlated with the growth of a-CF 2 BrCF 2 because g-CF 2 BrCF 2 isomerizes into a-CF 2 BrCF 2 at a time constant of 47 ± 5 ps [15]. Moreover, the decays of a-CF 2 ICF 2 and g-CF 2 BrCF 2 correlated with the growth of CF 2 CF 2 and CF 2 BrCF 2 Br, implying that these products were formed by the secondary reactions of a-CF 2 ICF 2 and a-CF 2 BrCF 2 . extinction coefficient (ε) of a-CF2BrCF2, g-CF2BrCF2, and a-CF2ICF2 were obtained from our previous measurements [8,15], and those of CF2BrCF2I, CF2CF2, and CF2BrCF2Br were determined from equilibrium FTIR measurements. Time-dependent fractional population changes of CF2BrCF2I, CF2CF2, CF2BrCF2Br, a-CF2BrCF2, g-CF2BrCF2, and a-CF2ICF2 were derived from time-dependent amplitude changes for the basis spectra of the corresponding compounds using the relation, ∝ . As shown in Figure 5, three nascent radicals (a- CF2BrCF2, g-CF2BrCF2, and a-CF2ICF2) appeared immediately, implying that the photodissociation of CF2BrCF2I occurred within 0.3 ps [15]. The decay of g-CF2BrCF2 correlated with the growth of a-CF2BrCF2 because g-CF2BrCF2 isomerizes into a-CF2BrCF2 at a time constant of 47 ± 5 ps [15]. Moreover, the decays of a-CF2ICF2 and g-CF2BrCF2 correlated with the growth of CF2CF2 and CF2BrCF2Br, implying that these products were formed by the secondary reactions of a-CF2ICF2 and a-CF2BrCF2. A kinetic scheme (Scheme 1) was introduced to reproduce the time-dependent fractional population changes, as shown in Figure 5, by optimizing the rate constants of all transitions. As shown in Figure 5, Scheme 1 reproduced the time-dependent fractional population changes. Figure 5. Time-dependent fractional population changes in CF 2 CF 2 (green), CF 2 BrCF 2 Br (magenta), a-CF 2 BrCF 2 (blue), g-CF 2 BrCF 2 (orange), and a-CF 2 ICF 2 (red), which were derived from the timedependent amplitude changes of the basis spectra of the corresponding compounds (see text). Time constants for the changes are also shown. Data (open circles) are reproduced by the kinetic scheme (Scheme 1, see text) introduced to describe the time-dependent fractional population changes by adjusting rate constants.
A kinetic scheme (Scheme 1) was introduced to reproduce the time-dependent fractional population changes, as shown in Figure 5, by optimizing the rate constants of all transitions. As shown in Figure 5, Scheme 1 reproduced the time-dependent fractional population changes.
Scheme 1 indicates that all the excited CF 2 BrCF 2 I at 267 nm dissociated one halogen atom to produce CF 2 ICF 2 or CF 2 BrCF 2 (18 ± 3% a-CF 2 ICF 2 , 33 ± 3% a-CF 2 BrCF 2 , and 49 ± 3% g-CF 2 BrCF 2 ) [15]. Almost all nascent g-CF 2 BrCF 2 isomerizes into a-CF 2 BrCF 2 with a time constant of 47 ± 5 ps [15]. All of the nascent a-CF 2 ICF 2 undergoes secondary dissociation to produce CF 2 CF 2 + I with a time constant of 56 ± 5 ns. The nascent a-CF 2 ICF 2 obtained from CF 2 ICF 2 I in CCl 4 undergoes secondary dissociation with a time constant of 5.5 ns [8]. Although a secondary dissociation time of 56 ns for a-CF 2 ICF 2 is less than that observed previously, it clearly demonstrates that the secondary dissociation of the C−I bond is feasible and can occur on the nanosecond time scale in solution.

Scheme 1.
A kinetic scheme used to fit time-dependent fractional populational changes of various compounds appearing when CF2BrCF2I in CCl4 at 280 K is excited at 267 nm. The rate constants and percentage were obtained by globally fitting the TRIR spectra. Gray, light blue, brown, and purple spheres represent carbon, fluorine, bromine, and iodine atoms, respectively. The quoted uncertainties reflect fitting and estimated experimental error.
Scheme 1 indicates that all the excited CF2BrCF2I at 267 nm dissociated one halogen atom to produce CF2ICF2 or CF2BrCF2 (18 ± 3% a-CF2ICF2, 33 ± 3% a-CF2BrCF2, and 49 ± 3% g-CF2BrCF2) [15]. Almost all nascent g-CF2BrCF2 isomerizes into a-CF2BrCF2 with a time constant of 47 ± 5 ps [15]. All of the nascent a-CF2ICF2 undergoes secondary dissociation to produce CF2CF2 + I with a time constant of 56 ± 5 ns. The nascent a-CF2ICF2 obtained from CF2ICF2I in CCl4 undergoes secondary dissociation with a time constant of 5.5 ns [8]. Although a secondary dissociation time of 56 ns for a-CF2ICF2 is less than that observed previously, it clearly demonstrates that the secondary dissociation of the C−I bond is feasible and can occur on the nanosecond time scale in solution.
The decay of a-CF2BrCF2 was correlated with the growth of the final products, CF2CF2 and CF2BrCF2Br. The possible reactions of a-CF2BrCF2 to produce the final products are as follows.

Scheme 1.
A kinetic scheme used to fit time-dependent fractional populational changes of various compounds appearing when CF 2 BrCF 2 I in CCl 4 at 280 K is excited at 267 nm. The rate constants and percentage were obtained by globally fitting the TRIR spectra. Gray, light blue, brown, and purple spheres represent carbon, fluorine, bromine, and iodine atoms, respectively. The quoted uncertainties reflect fitting and estimated experimental error.
The decay of a-CF 2 BrCF 2 was correlated with the growth of the final products, CF 2 CF 2 and CF 2 BrCF 2 Br. The possible reactions of a-CF 2 BrCF 2 to produce the final products are as follows.
The excess energy in the nascent radicals (a-CF 2 ICF 2 , a-CF 2 BrCF 2 , and g-CF 2 BrCF 2 ) produced during the photodecomposition of CF 2 BrCF 2 I was thermalized with a time con-stant of 15 ± 3 ps [15], which is consistent with the thermalization time constant observed in other reactions in solution [8,19]. The solvent acts as an energy sink for the excess energy of a molecule in solution. All nascent radicals were thermalized with a time constant of 15 ± 3 ps after the photodecomposition of CF 2 BrCF 2 I in CCl 4 , indicating that the observed reactions became thermal within tens of picoseconds after the photodecomposition. Therefore, the radicals should gain the energy required for the secondary bond dissociation that occurs within nanoseconds or longer. The bond dissociation energy is supplied by the solvent, indicating that the solvent becomes an energy source as well as an energy sink. As the solvent can act as an energy sink and an energy source, secondary dissociation at a longer duration than thermalization is a characteristic of the reaction in solution. For the secondary dissociation of the C−X bond, the activation energy should be as high as the C−X bond energy. As the C−Br bond energy of 22.3 ± 2.5 kcal/mol is 15.2 kcal/mol higher than the C−I bond energy of 7.1 ± 2.5 kcal/mol [7], the secondary dissociation of the C−Br bond is expected to take approximately 10 12 times longer than that of the C−I bond, based on the Arrhenius equation for the rate constant at room temperature. Considering that the secondary dissociation time of thermalized a-CF 2 ICF 2 is 56 ns, it is unlikely that the secondary dissociation of Br from thermalized a-CF 2 BrCF 2 will take 130 ns.
The Gibbs free energy of a-CF 2 BrCF 2 Br was calculated to be 0.91 kcal/mol more stable than that of g-CF 2 BrCF 2 Br. Therefore, when thermalized, 72% of CF 2 BrCF 2 Br is in the anti-conformer at 280 ± 2 K, while almost all (97%) the CF 2 BrCF 2 radicals are a-CF 2 BrCF 2 . Although most CF 2 BrCF 2 Br is produced by the BR of a-CF 2 BrCF 2 , the produced CF 2 BrCF 2 Br is in both anti and gauche conformers in the equilibrium distribution. Based on the rotational isomerization time of 47 ps for CF 2 BrCF 2 and the calculated rotational barriers [15], that of CF 2 BrCF 2 Br in CCl 4 at 280 K was estimated [15] to be 12 ns using the calculated Gibbs free energy for the rotational activation. The Gibbs free energy was found to be 6.21 (7.12) kcal/mol for gauche-to-anti (anti-to-gauche) transition by the DFT method using ωB97X-D/aug-cc-pVTZ. As shown in Figure 5, CF 2 BrCF 2 Br was produced with a time constant of 83 ns, which is longer than the estimated rotational isomerization time. Thus, CF 2 BrCF 2 Br was produced in the equilibrium distribution, even though the reactant was mainly the a-CF 2 BrCF 2 radical.

Time-Resolved Mid-IR Spectroscopy
The TRIR spectrometer is based on a commercial Ti:sapphire oscillator/amplifier system (Spitfire Ace, Spectra Physics, Milpitas, CA, USA) that generates 800 nm, 110 fs pulses with a repetition rate of 2 kHz [8,19]. Half of the amplified femtosecond pulses were sent to a third-harmonic generator (THG), and the other half were sent to a homemade linear optical parametric amplifier (OPA). The THG produced 267 nm, 160 fs pulses with 10 µJ of energy. The OPA produced near-IR signal and idler pulses, which were difference frequency mixed in a 1-mm-thick GaSe to generate tunable mid-IR pulses with a duration of 120 fs, 160 cm −1 spectral width, and 1 µJ of energy. The sample was excited by a portion of the 267 nm pulse and probed by a small fraction (~10 nJ) of the mid-IR pulse. The 267 nm excitation pulses were optically delayed up to 1 ns relative to the mid-IR probe pulses by a computer-controlled translation stage (M-415. PD, PI, Karlsruhe, Deutschland). The translation stage was impractical for optical delays beyond 1 ns because of the long physical distances for the stage. Thus, the excitation pulses were replaced with the nanosecond pulses (267 nm, 2.5 ns, and 20 µJ) produced by a commercial nanosecond tunable laser (NT240, EKSPLA, Vilnius, Lithuania) based on an optical parametric oscillator. These nanosecond excitation pulses were synchronized with the femtosecond probe pulses using an electronic digital delay generator (DG535, Stanford Research Systems, Sunnyvale, CA, USA), which produced an optical delay beyond 1 ns. The angle between the linearly polarized probe and the excitation pulses was set as the magic angle (54.7 • ) to obtain an isotropic absorbance proportional to the population of the corresponding compounds free from molecular rotation. The energy of the excitation pulse was reduced to 1-2 µJ to ensure linear absorbance and to minimize thermal lensing, which causes high background absorption on the nanosecond time scale due to solvent heating [21][22][23][24][25]. To measure the spectral region beyond the spectral width (160 cm −1 ) of the probe pulses, the experiment was repeated with probe pulses centered at 1300, 1230, 1160, 1080, or 950 cm −1 , and the data were combined to obtain a broad spectrum. The probe pulses were passed through the sample and were detected by a 1 × 128-pixel MCT array detector (MCT-8-128, InfraRed Associates, Stuart, FL, USA) for wavenumbers > 1000 cm −1 or a 1 × 64-pixel MCT array detector (MCT-16-64, InfraRed Associates, Stuart, FL, USA) for wavenumbers < 1000 cm −1 in a 320 mm monochromator (HR320, Horiba, Miami, FL, USA) with a grating of 100, 75, or 50 lines/mm. The 1 × 64-pixel MCT array detector was custom-made to enhance spectral responsivity at longer wavelengths (<1000 cm −1 ). The resulting spectra have a spectral resolution of 0.9-1.2 cm −1 for the spectral regions > 1000 cm −1 and 2.7 cm −1 for spectral regions < 1000 cm −1 . The instrument response function was determined by the transient absorption of the Si wafer and was 0.2 ps in the subnanosecond experiment (0.3 ps-1 ns) and 2.5 ns in the nanosecond experiment (1 ns-10 µs). The spectra were collected at pump-probe delay times ≥ 0.3 ps to avoid signal complications due to pumpprobe overlap [26][27][28].

Sample Preparation
CF 2 BrCF 2 I was purchased from Alfa Aesar, and CCl 4 was obtained from Sigma-Aldrich and used without further purification. A flowing sample cell consisting of two 2-mm-thick BaF 2 windows and a 100-µm-thick Teflon spacer was connected to a peristaltic tubing pump (Masterflex L/S, Cole-Parmer, Vernon Hills, IL, USA), refreshed with 50 mM CF 2 BrCF 2 I in CCl 4 fast enough for each probe pulse at a repetition rate of 2 kHz. Concentration and path length were set to obtain the maximum transient absorption signal in the spectral region of interest. A sufficiently large volume of the sample was used to maintain the decrease in the sample concentration as <2% due to the photoreaction of CF 2 BrCF 2 I by 267 nm excitation. The temperature of the sample was maintained at 280 ± 2 K by passing the coolant through the sample mount block. The low temperature was beneficial in minimizing the noise in the TRIR spectra caused by the temperature gradient in the probe spot of the sample induced by the pump energy (i.e., thermal lensing) appearing on the nano-to microsecond time scale [21][22][23][24][25]. To ensure the integrity of the samples, equilibrium UV-Vis and FT-IR spectra were collected before and after the TRIR experiments.

Computational Details
All quantum-chemical calculations were performed using the Gaussian 09 software (Revision D.01). The DFT method with the ωB97X-D (long-range corrected hybrid density functionals with damped atom-atom dispersion corrections) functional was used for molecular geometry optimization and the energies of the optimized compounds. ωB97X-D has been reported to yield satisfactory accuracy in thermochemistry, kinetics, and noncovalent interactions [29,30]. The aug-cc-pVTZ basis set was used for the C, F, and Br atoms, and the aug-cc-pVTZ-PP basis set was used for the I atom. Solvent effects were incorporated into the polarizable continuum model using the integral equation formalism variant.

Conclusions
In this study, the photodissociation dynamics of CF 2 BrCF 2 I in a CCl 4 solution were determined over a broad time span, encompassing the entire reaction, after excitation at 267 nm using TRIR spectroscopy. UV excitation led to ultrafast dissociation of I or Br atoms, resulting in CF 2 BrCF 2 and CF 2 ICF 2 radicals, respectively, which further proceeded with conformer-specific reactions. CF 2 BrCF 2 was the major photofragment with both conformers of a-CF 2 BrCF 2 (33 ± 3%) and g-CF 2 BrCF 2 (49 ± 3%), while a-CF 2 ICF 2 (18 ± 3%) was the minor one. The less stable g-CF 2 BrCF 2 was isomerized with a time constant of 47 ± 5 ps into a-CF 2 BrCF 2 . a-CF 2 BrCF 2 further reacted bimolecularly with itself at a diffusion-limited rate to produce CF 2 BrCF 2 Br and CF 2 CF 2 , or with the dissociated Br atom to form CF 2 BrCF 2 Br. However, no secondary dissociation of C−Br in a-CF 2 BrCF 2 was observed. In contrast, a-CF 2 ICF 2 underwent simple secondary dissociation of the I atom with a time constant of 56 ± 5 ns to produce CF 2 CF 2 + I. The final products contained 50 ± 3% CF 2 BrCF 2 Br and 50 ± 3% CF 2 CF 2 . Structure-sensitive TRIR spectroscopy allowed us to determine the entire reaction dynamics of photoexcited CF 2 BrCF 2 I in a CCl 4 solution, including the conformer-specific dynamics of the CF 2 BrCF 2 radicals.