Probing the Photoisomerization of CHBr3 and CHI3 in Solution with

Feb 26, 2013 - emergent halogen atom returns to form a C−X−X bonding motif, appear about 9 ps after photolysis for iso-CHBr2−Br and in about 46 ...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/JPCA

Probing the Photoisomerization of CHBr3 and CHI3 in Solution with Transient Vibrational and Electronic Spectroscopy Thomas J. Preston,† Michael A. Shaloski, and F. Fleming Crim* Department of Chemistry, University of WisconsinMadison, Madison, Wisconsin 53706, United States S Supporting Information *

ABSTRACT: Transient infrared absorption spectroscopy monitors condensed-phase photodissociation dynamics of 30 mM CHBr3 and 50 mM CHI3 in liquid CCl4. The experiments have picosecond time resolution and monitor the C−H stretch region of both the parent polyhalomethanes and their photolytically generated isomers. The C−H stretching transitions of these isomers, in which the emergent halogen atom returns to form a C−X−X bonding motif, appear about 9 ps after photolysis for iso-CHBr2−Br and in about 46 ps for iso-CHI2−I. These time scales are consistent with, but differ from, the time evolution of the transient electronic absorption spectra of the same samples, highlighting the subtle differences between monitoring the vibrational and electronic chromophores. The specificity of using vibrational transitions to track condensed-phase reaction dynamics permits reassessment of the transient electronic spectrum of photolysis in neat CHBr3, which has an additional prompt feature near 400 nm. Calculations show that this feature, which arises from a precursor to the isomer, is a charge-transfer transition of a contact pair between the nascent Br fragment and a nearby CHBr3 molecule. Dilution and solvent studies show that transition is independent of the solvent. The iso-CHBr2−Br transition wavelength, however, shifts over the range of 400 to 510 nm depending on the solvent. Time-dependent density functional calculations faithfully reproduce these trends.

1. INTRODUCTION Condensed-phase photolysis can lead to unique products because the high density of the surroundings restricts the translational motion of the dissociating fragments. The subsequent interaction of the fragments can cause them to reassociate,1 and recombination can rearrange the connectivity of the atoms, effecting a solvent-assisted isomerization.2−7 Figure 1 shows a sketch of the potential energy surface for CHBr3, one of the molecules we study here, and its isomer, CHBr2−Br. On the corresponding contour plot, we sketch two reaction pathways in which the solvent inhibits separation of the radical fragments. As the figure illustrates, one of these paths facilitates isomerization in this and other polyhalomethanes.3,5,8−10 Photolysis initiates cleavage of one C−Br bond, but the surrounding molecules prevent complete separation of the CHBr2 and Br fragments. The departing halogen atom can return to bond with the carbon atom, regenerating the parent molecule, or with a bromine atom bound to the carbon, forming iso-CHBr2−Br. In a closely related study, Tarnovsky and co-workers have recently measured the bromoform transient electronic absorption spectrum over a broad range and report on its reactivity with water.11 Many other molecules undergo this solvent-assisted isomerization, forming C−X−X bonding motifs, and these isomers are reactive species in condensed-phase chemistry. Phillips and coworkers first proposed iso-CH2I−I as the methylene transfer agents in cyclopropanation reactions.12,13 Time-resolved transient absorption measurements by Paschen and co-workers show that both iso-CH2I−I and iso-CHI2−I are active in the cyclopropanation of several cycloalkenes.14−16 The electronic absorption spectra of these isomers typically consist of two © 2013 American Chemical Society

broad features that span the near-ultraviolet through nearinfrared wavelength regions, and the larger of these features can have molar extinction coefficients up to 18 000 M −1 cm−1.2,3,7,10,14 This intense absorption and unique bonding structure make the isomerization of the polyhalomethanes ideal targets for exploring ultrafast condensed-phase reaction dynamics. Previous studies determine picosecond formation times using ultraviolet and visible transient absorption4,7,17 and transient resonance Raman spectroscopy,18 both of which exploit the characteristically strong electronic transitions in these molecules. There are several possible chromophores in the condensedphase photolysis of polyhalomethanes, including halomethyl radicals, halogen−solvent complexes, polyhalomethane isomers, and reforming parent molecules. The broad electronic transitions of these species can complicate interpretation of the transient ultraviolet and visible spectra. However, the specificity of vibrational spectroscopy complements the electronic spectroscopy and provides an additional picture of the isomerization. Time-resolved resonance Raman spectroscopy also identifies vibrational frequencies of transient species during isomerization18,19 but necessarily relies on electronic transitions for the resonance enhancement. An alternative approach using transient infrared absorption of the photolysis products is an effective means of increasing understanding of condensed-phase isomerization reactions. We use vibrational spectroscopy to monitor liquid-phase isomerization reactions of the polyhalomethanes CHBr3 and Received: October 30, 2012 Revised: February 25, 2013 Published: February 26, 2013 2899

dx.doi.org/10.1021/jp310737d | J. Phys. Chem. A 2013, 117, 2899−2907

The Journal of Physical Chemistry A

Article

regions of the spectrum, we focus a few nJ of the fundamental 800-nm light into a CaF2 substrate, generating a broad spectrum of light covering about 330 to 1100 nm. For infrared probe light, a continuum-seeded double-pass optical parametric amplifier based on a potassium niobate crystal directly generates tunable light near 3.3 μm to monitor changes in the infrared optical density (OD) of the sample after photolysis. A computer-controlled mechanical translation stage controls the time interval between the two pulses, Δtdelay, as they cross at a small angle in the sample. A gear pump flows the liquid sample between two MgF2 windows separated by a 1-mm polytetrafluoroethylene spacer. We use CHBr3 (Aldrich, 99+ %), CHI3 (Arcos Organics 99+%), and CCl4 (Sigma-Aldrich, Reagent grade 99.9%) as received. Our samples are sufficiently dilute to transmit the 3-μm probe light while providing adequate transient absorption signals. The concentrations, 30 mM CHBr3 and 50 mM CHI3, are the same for both the infrared and ultraviolet/visible probing experiments. We record the intensity of the probe light for every laser pulse on each of the 64 pixels of a mercury−cadmium−telluride array in the infrared measurements, obtaining the spectrum with and without the photolysis pulse present on successive laser pulses. Measuring the infrared intensity on consecutive pulses minimizes spurious transient absorption signals resulting from fluctuations in the laser output. A 0.25-m Ebert monochromator that contains a 300-grooves/mm grating disperses the infrared probe light on the infrared array detector, which sits in the focal plane of the monochromator. The C−H stretching features of thin films of CH2Cl2, CHCl3, and CHBr3 sandwiched between two CaF2 windows serve to calibrate the detector. The resolution of this spectrometer is about 3 nm/ pixel, which is about 2.7 cm−1/pixel in this region of the spectrum. The first 32 elements of the 64-element array respond to the infrared light differently from the last 32 elements. However, all pixels within each 32-element set respond similarly to each other. We compensate for the difference by processing the data from the single detector as two separate, side-by-side arrays. To account for fluctuations in the output of the OPA, we use one pixel in each set of 32 pixels to normalize the response across all pixels. There is typically a sloping baseline through the transient spectrum because of imperfect correlation between the reference wavelength and the remaining 31 elements, and, thus, we fit a linear baseline through the first two and last two pixels of each half of the array. Each pixel of the array has a unique nonzero transient absorption baseline, and we take the average transient spectrum of 9 data points with Δtdelay ≤ −2.5 ps and subtract that value from all transient absorption spectra. This procedure results in highly repeatable transient absorption spectra that have pixel-to-pixel fluctuations near 5 parts in 106 (5 μOD).21 Broadband electronic absorption measurements use a matched pair of Si photodiode arrays as detectors in a pair of modified Czerny-Turner spectrometers that each contain a 600grooves/mm holographic grating.20 This arrangement provides a resolution of 0.7 nm/pixel and about 350 nm of spectral coverage. We divide the continuum probe light into a signal pulse that passes through the sample and a reference pulse that goes around the sample. Sharp holmium absorption features (Holmium wavelength calibration standard, Sigma-Aldrich) provide reliable calibration of both arrays. The reference array monitors fluctuations in the intensity of this probe light. Fast mechanical shutters expose the sample to 100 consecutive

Figure 1. Sketch of the potential energy surface and contour plot of isomerization in CHBr3. Photolysis cleaves a C−Br bond, and the solvent cage restricts free motion along the C−Br bond distance. The system returns to either minimum corresponding to the parent or isomer, transferring energy to the surrounding solvent.

CHI3. Using ultrafast transient infrared absorption spectroscopy, we observe the formation and relaxation dynamics of both the isomers and the parent molecules in liquid carbon tetrachloride (CCl4). Guided by matrix-isolation assignments of vibrational transitions,2,3,5 which first identified the isomeric forms of the polyhalomethanes, we are able to monitor loss of the parent molecules and the ensuing formation of the isomers using their C−H stretching transitions near 3.3 μm. The C−H stretches are particularly suitable for monitoring isomerization reactions and vibrational relaxation because their transition frequencies are well-separated and their widths reflect vibrational excitation in other modes of the molecule. Previous studies assigned the photolysis products of bromoform17 and iodoform10 to their isomers using ultrafast transient electronic absorption spectroscopy. We make transient electronic absorption measurements for these reactions in carbon tetrachloride for comparison with our new transient infrared study. The increased specificity of vibrational spectroscopy allows us to revise our previous assignment of the electronic transition of iso-CHBr2−Br.17

2. EXPERIMENTAL APPROACH The apparatus we use to monitor the evolution of the spectra of the iso-polyhalomethanes with ps time resolution is similar to the one we used recently to study condensed-phase photolysis reaction dynamics.17,20 An ultrafast Ti:sapphire laser operating at 1 kHz with 100-fs pulses centered near 800 nm provides both the photolysis and the probe light for the transient vibrational and electronic absorption spectroscopy. The third harmonic of the laser near 267 nm photolyzes a solution of bromoform (CHBr3) or iodoform (CHI3), initiating isomerization. For probe light spanning the near-ultraviolet and visible 2900

dx.doi.org/10.1021/jp310737d | J. Phys. Chem. A 2013, 117, 2899−2907

The Journal of Physical Chemistry A

Article

photolysis pulses, and we measure the intensity of the probe pulse on both arrays. We then execute this procedure with the photolysis pulse blocked and calculate the resultant transient absorption. Repeating this cycle five times for each Δtdelay and scanning the full series five times at about 70 different time delays provides a sensitivity of about 15 parts in 105 (150 μOD) in the transient electronic spectra. The solvent CCl4 is susceptible to two-photon photolysis, and its transient products appear in the same spectral region as the CHBr3 and CHI3 photoproducts.22 We acquire five scans with only the solvent and use those data to remove those spectral features from the transient spectra with solutes added to the CCl4 solvent. The uncertainties we report encompass both the statistical uncertainty of the fits described below and variations in the reproducibility of different samples on different days. Each of the data sets we present is an average of five or more scans at about 70 different time delays with each data point being the average of about 50 000 transient infrared absorption measurements or 2500 laser pulses for transient ultraviolet/visible measurements. We use the Gaussian 09 suite of programs23 to perform electronic structure calculations on isolated and implicitly solvated parent molecules, isomers, and related structures to obtain vibrational and electronic energy levels and transition strengths. The M06 functional using the Sadlej-pVTZ basis set for all atoms provides harmonic vibrational transition energies and intensities. Time-dependent density functional theory using the M06-2x functional with the same basis set provides the electronic transition energies and oscillator strengths. Previous work has shown that these levels of theory and basis set reliably provide transition energies and strengths for similar isomers.7 One of our calculations includes an explicit Ar atom, in which case we use the similar aug-cc-pVTZ basis set because the Sadlej-pVTZ is unavailable for Ar.

Figure 2. Transient vibrational absorption spectrum in the C−H stretching region after photolysis of 30 mM CHBr3 in CCl4. The two features in the data (closed circles) result from instantaneous loss of parent CHBr3 and growth of isomer iso-CHBr2−Br. Fits to the data (solid lines) provide reliable descriptions of the evolving spectrum, as described in the text.

3. RESULTS Previous experiments assign the infrared spectrum of isoCHBr2−Br entrained in the stabilizing environment of cryogenic matrices,5 and we use that study and our computational results to inform the interpretation of the liquid-phase measurements. Band positions of infrared transitions in matrixisolation measurements typically are within 1−2% of their gasphase values.24,25 There are no published spectra of CHI3 or iso-CHI2−I in cryogenic matrices. Figure 2 shows the transient vibrational spectra, and Figure 3 shows the transient electronic absorption spectra following photodissociation of 30 mM CHBr3 in CCl4. Two prominent vibrational features with opposite signs dominate the infrared spectrum, the negative signal (bleach) near the probe wavenumber ν̃probe = 3025 cm−1 and the absorption near ν̃probe = 3050 cm−1. The negative feature, whose position mirrors the absorption of the C−H stretch of CHBr3, is the only feature in the spectrum at Δtdelay = 1.5 ps. In the following 10 ps, a positive feature belonging to iso-CHBr2−Br rises near ν̃probe = 3045 cm−1, and over the next 1 ns this feature shifts to higher energy and narrows. The bleach, however, grows in magnitude until about 150 ps after photolysis, beyond which it begins to decrease. The 25-cm−1 difference in the C−H stretching frequencies for CHBr3 and iso-CHBr2−Br that we measure at Δtdelay = 1 ns in CCl4 is consistent with the 25-cm−1 difference between the two in cryogenic matrices.5 That same study measures an oscillator strength for the isomer that is approximately four times larger than for the parent.5 At the

Figure 3. Transient electronic absorption spectrum of iso-CHBr2−Br after photolysis of 30 mM CHBr3 in CCl4. The electronic spectrum of the isomer grows over about 50 ps and is reminiscent of the transient vibrational spectrum.

longest time delay we measure, Δtdelay = 1 ns, the area of the feature at ν̃probe = 3052 cm−1 is 1.8 times larger than the area of the bleach at 3025 cm−1. If we assume that one lost parent molecule generates one isomer in the liquid phase, the liquidphase ratio is a lower bound to the true comparison of oscillator strengths because other processes can generate products that we do not measure. The large extinction coefficient of the C−H stretch in the liquid-phase photolysis product is also consistent with isomer formation. We fit the vibrational spectra at each Δtdelay using a sum of Gaussian curves through the transient spectra, 2

G(ν̃) =

c 2 i

∑ Aie−(ν̃− ν̃ ) +G∞

(1)

i=1

νci ,

where their amplitudes, Ai, centers, widths, σi, and a baseline offset, G∞, are parameters of the fit. Restricting |Ai| to be greater than 1 μOD, νci to be within the bounds of the spectrum, σi to be more than 8 cm−1 but less than 65 cm−1, and |G∞| to be less 2901

dx.doi.org/10.1021/jp310737d | J. Phys. Chem. A 2013, 117, 2899−2907

The Journal of Physical Chemistry A

Article

picoseconds cause unreliable fits to eq 1, we restrict our kinetic analysis to time delays longer than 4 ps. We use areas at each time as the transient infrared absorption signal, S(t), and fit the signal to a sum of exponentials

than 100 μOD prevents physically unreasonable fits to the data yet assures flexibility in fitting the transient spectra. If only one feature is present, we set A2 = 0 μOD. We do not intend to imply that a physical model results in Gaussian lineshapes. The function fits the vibrational features well with only a few parameters, as shown by the smooth lines in Figure 2, and the spectral fitting provides ready analysis of band centers, amplitudes, and widths. We use the area of the Gaussian fit to each vibrational feature at each Δtdelay rather than the amplitude of a single ν̃probe to monitor relative populations of species in our probe volume because the off-diagonal anharmonicity changes the vibrational transition frequency for different levels of excitation in the other modes of the molecule,26 which can broaden the feature. The top panel of Figure 4 shows the areas of the bleach as a

n

S(t ) =

∑ Bi e−[(t− t )/τ ] + S∞ 0

i

(2)

i=1

with the fitting parameters of the amplitude Bi and time constant τi associated with each exponential, the baseline offset S∞, and the zero of time t0 dictated by the temporal overlap of the pump and probe. We use n ≤ 3 to fit the data. Table 1 contains the fit parameters for eq 2 for CHBr3, and CHI3 photolysis. The prompt appearance of the bleach in CHBr3 photolysis at ν̃probe = 3025 cm−1 is consistent with rapid photolysis of CHBr3. The temporal evolution differs for the infrared absorption of iso-CHBr2−Br near ν̃probe = 3050 cm−1. Its initial rise, which occurs with time constant τ1 = 8.9 ± 0.8 ps, is similar to the evolution of the broad transient electronic absorption at λprobe = 443 nm, which appears with τ1 = 19 ± 1 ps. The difference between the two time scales likely results from probing two distinct chromophores. The C−H bond, the vibrational chromophore, is spatially separated from the newly forming Br−Br bond, the electronic chromophore.19 It is probably necessary to develop a more nearly complete Br−Br bond in order for the distinct infrared C−H stretching absorption of the isomer to appear at its characteristic wavelength. The broad electronic absorption of iso-CHBr2−Br develops before the photoproducts lose much of their initial energy from photolysis, highlighting the less stringent requirements for electronic excitation. Thus, we use a 2-nm wide region near λprobe = 450 nm to characterize the isomer population and fit that absorption as the signal S(t). The electronic spectra monitor the isomer population but provide no information about the parent. The top panel of Figure 4 shows the area of the C−H stretching transition of iso-CHBr2−Br on the left ordinate and the amplitude of the electronic transition near λprobe = 443 nm on the right ordinate. Both vibrational and electronic spectral features have biexponential kinetics with secondary time constants τ2 = 98 ± 17 ps and τ2 = 124 ± 44 ps, respectively. The 20-ps appearance time for a vibrational feature belonging to iso-CHBr2−Br in dilute CCl4 disagrees with our previous analysis using concentrated CHBr3 samples.17 That previous work relied on transient electronic absorption spectroscopy alone, but the complementary vibrational spectroscopy we use here helps us reassign the transient spectra produced by photolysis of concentrated CHBr3. In the Discussion, we show that in concentrated CHBr3 solutions a charge-transfer transition of the nascent Br atom adjacent to a CHBr3 solvent molecule is responsible for the prompt feature and that, even in neat solutions, the isomer evolves with time scales similar to those observed in dilute solutions.

Figure 4. Time evolution of vibrational (red, left ordinate) and electronic (black, right ordinate) features following photolysis of 30 mM CHBr3 (top) and 50 mM CHI3 (bottom) in CCl4. The vibrational features of the isomers, fit to the extracted area, evolve with a time constant that differs from the electronic features, fit to a 2.5-nm wide spectral window, because of the difference between the electronic and vibrational chromophores. The C−H stretch features of the parent disappear promptly after photolysis, but we do not probe their electronic transitions.

function of time after photolysis of CHBr3 (open circles) near ν̃probe = 3025 cm−1 resulting from the loss of parent CHBr3 and the absorption feature (closed red circles) near ν̃probe = 3050 cm−1 assigned to iso-CHBr2−Br. The solid lines through these data are the multiexponential fits described below. Because nonresonant transient absorption signals during the first few

Table 1. Time Constants and Amplitudes of Multiexponential Fits in CHBr3 and CHI3 Photolysis feature iso-CHBr2−Br (vib.) CHBr3 (vib.) CHBr3 (elec.) iso-CHI2−I (vib.) CHI3 (vib.) CHI3 (elec.)

τ1 (ps) 9.0 5.3 19 46 40 29

± ± ± ± ± ±

0.6 2.6 1.0 11 22 2

τ2 (ps) 98 22 124 445

± ± ± ±

17 7 44 832

307 ± 2000 2902

τ3 (ps) 7.5 × 105 2843

B1/B2 2.5 −0.6 6.2 0.8

B1/B3 1.7 × 10−4 −0.1

8.1 dx.doi.org/10.1021/jp310737d | J. Phys. Chem. A 2013, 117, 2899−2907

The Journal of Physical Chemistry A

Article

The C−H stretching frequency of matrix-isolated iso-CHI2−I is unknown, but the shift to higher energy by about 25-cm−1 for the isomer relative to parent is consistent for iso-CHBr3,5 isoCH2Cl−I,7 and iso-CH2Br−Br.3,27 We measure the time evolution of the previously assigned9 iso-CHI2−I electronic absorption spectrum at λprobe = 453 nm with τ1 = 29 ± 2 ps in CCl4. Wall and co-workers measure an appearance time of 11 ps for iso-CHI2−I in cyclohexane,9 and the difference we report here may result from our increased photolysis energy or the change in solvent. For example, the CCl4 solvent may remove energy from the nascent fragments less efficiently that the cyclohexane solvent. The bottom panel of Figure 4 shows the time evolution of the infrared and visible absorption features for the CHI3 system. The calculated C−H stretch oscillator strength, discussed below, of iso-CHI2−I is about twice as large as that of CHI3, which is consistent with the lower bound of our measured ratio of 1.7 at Δtdelay = 1 ns.

Photoinduced isomerization in iodoform produces transient spectra that are similar to those in bromoform, an observation that supports the similar assignments we make in the two systems. Figure 5 shows the transient vibrational spectra, and

4. DISCUSSION Ultraviolet photolysis of halomethanes with 267-nm light deposits 37 500 cm−1 of energy, and we use the C−H stretching transitions of both parent and isomer molecules to monitor the removal of the excess energy by the solvent. The energy from the ultraviolet photon goes into breaking the carbon−halogen bond, translation of the departing fragments, internal excitation of the polyatomic fragments, and potentially electronic excitation of the halogen atom. Table 2 shows this Table 2. Gas-Phase Energy Partitioning of CHBr3 and CHI3 Photolysis

Figure 5. Transient vibrational absorption spectrum in the C−H stretching region after photolysis of 50 mM CHI3 in CCl4. The two features in the data (closed circles) result from instantaneous loss of parent CHI3 and growth of isomer iso-CHI2−I. Fits to the data (solid lines) provide reliable descriptions of the evolving spectrum, as described in the text.

molecule CHBr3a CHI3

bond energy available (cm−1) energy (cm−1) 21 415 15 889b

15 872 21 398c

translational energy (cm−1)

internal energy (cm−1)

3968 2742

11 904 18 655

a

Reference 21. bReference 23. cUsing soft-fragment impulsive model of ref 22.

energy partitioning for gas-phase photolysis of CHBr3 and includes an estimate for CHI3 photolysis.21,28 Photolysis of gasphase bromoform leaves the CHBr2 fragment with about 12 000 cm−1 of internal energy and deposits about 4000 cm−1 of energy in relative translation, Etrans, of the fragments.28 Data are unavailable for energy partitioning in CHI3 photolysis. We estimate the translational energy using a soft-fragment impulsive model29 ⎛ μ ⎞ Etrans = ⎜⎜ C−I ⎟⎟Eavail ⎝ μCH2I−I ⎠

(3)

from the reduced mass of the C and I atoms, μC−I, the reduced mass of the CHI2 and I, μCHI2−I, and the available energy, Eavail. Cleaving the 191-kJ/mol C−I bond30 with a 267-nm photon provides the radicals with about 3000 cm−1 of translational energy. The internal and translational energy in the fragments after photolysis inhibits prompt isomer formation because halogen−halogen bond strengths in these isomers are only about 3000 to 5300 cm−1.5,7 An ab initio molecular dynamics simulation of condensed-phase CH2I2 photoisomerization in acetonitrile solution shows that an additional source of vibrational excitation in the photoproduct can come from recombination of the photolyzed halogen atom with the

Figure 6. Transient electronic absorption spectrum of iso-CHI2−I after photolysis of 50 mM CHI3 in CCl4. The electronic spectrum of the isomer grows over about 100 ps and is reminiscent of the transient vibrational spectrum.

Figure 6 shows the transient electronic spectra obtained following photolysis of 50 mM CHI3 in CCl4. The transient infrared spectrum has a prompt appearance of the bleach at ν̃probe = 3012 cm−1, and the rise-time τ1 = 46 ± 11 ps of a feature about 23 cm−1 higher in energy at ν̃probe = 3035 cm−1 is reminiscent of CHBr3 photolysis. We assign these transitions to the C−H stretch of parent CHI3 and iso-CHI2−I, respectively. 2903

dx.doi.org/10.1021/jp310737d | J. Phys. Chem. A 2013, 117, 2899−2907

The Journal of Physical Chemistry A

Article

polyatomic fragment.31 With the new halogen−halogen bond, the geometry at the carbon atom changes from planar to tetrahedral, which excites many of the vibrational modes in the nascent iso-CH2I−I, providing an additional source of vibrational excitation.31 The evolution of the vibrational features monitors the removal of this excess energy, which is initially directed along the newly forming X−X bonding coordinate, and subsequent siphoning of the internal energy from the isomer into the solvent. Formation and relaxation times for isomers in solids,7 liquids,8 and supercritical fluids32 are largely independent of the state of matter of the solvent. We expect isomerization to follow similar paths and to have similar time scales in both dilute and concentrated solutions. The spectroscopy, however, changes. A prompt feature, absent in dilute solutions, appears in the electronic spectrum of concentrated CHBr3. We first discuss isomerization dynamics in dilute solutions of CHBr3 and CHI3, which provide us with the most complete picture of isomerization because we have both transient vibrational and electronic absorption spectra. Because the concentrated samples are opaque to the infrared probe light, we discuss the additional electronic absorption feature in CHBr3 within the framework established by the analysis of dilute solutions. 4.1. Isomer Formation and Relaxation in Dilute Solutions. The prompt appearance of the negative feature at ν̃probe = 3027 cm−1 after photolysis of CHBr3 in CCl4, shown in Figure 2, results from immediate loss of the parent. In the gas phase, photolysis creates separated CHBr2 and Br fragments, but we are unable to observe a prompt C−H stretching absorption from the CHBr2 fragment. The weak transition of the radical leaves us blind to its presence although we suspect it is formed initially. We calculate that the oscillator strength of the C−H stretch of the radical is only 30% of that of the parent. Table 3 shows the results of the harmonic frequency

detection limit. Reid and co-workers34 have identified a concerted pathway between parent and isomer that presents an alternative isomerization pathway that involves a gradual evolution from reactants to products. We do not include such a pathway in our model in order to avoid introducing more parameters than warranted by the data, but we cannot rule out its presence as a secondary channel. The two products we observe in the C−H stretching region, iso-CHBr2 −Br and CHBr 3, result from successful and unsuccessful isomerization events, shown by the bifurcating red arrow in Figure 1. As the photolysis products interact with the surrounding solvent, they begin to lose energy and their spectra evolve. Plotting the width of both bleach and absorption features, retrieved from eq 1, as a function of Δtdelay in Figure 7 provides a means of assessing the relaxation

Figure 7. Evolution of C−H stretching transitions of CHBr3 (red) and iso-CHBr2−Br (black). The width of the isomer feature decays quickly during the first 10 ps after photolysis, indicating a prompt removal of energy in the other low-frequency vibrational modes of the molecule. Vibrationally excited CHBr3 has the same transition energy as the parent bleach, complicating the spectral evolution of the bromoform feature. We do not fit the width at time delays less than about 5 ps because the extracted widths are unstable at short times.

Table 3. Harmonic Frequencies of Possible Photoproducts in CHBr3 and CHI3 Systems

species

C−H stretch energy (cm−1)

CHBr3 iso-CHBr2−Br CHBr2 CHBr3−Br CHI3 iso-CHI2−I CHI2 CHI3−I

3130 3155 3156 3121 3112 3137 3179 3116

calculated shift from parent intensity (cm−1) 2.7 22.8 0.8 11.9 5.1 10.6 0.3 11.1

of internal energy in the parent and isomer molecules. Increased widths of the infrared transition are hallmarks of vibrational excitation in other degrees of freedom,26 and without knowledge of the exact modes that couple to the C−H stretches we use their evolving widths to monitor the loss of vibrational excitation from the nascent molecules. The system loses much of its initial kinetic energy during the first collisions with the solvent cage. Rapid kinetic energy loss from photolyzed solutes such as diatomic halogen molecules isolated in cryogenic matrices occurs in a few picoseconds.1 Similar rapid thermalization occurs in liquids,7 and depositing translational energy into the vibrational modes of the nascent isomer with the new halogen−halogen bond31 localizes the photoproducts near either product basin in Figure 1. Because the extracted widths are unstable prior to this initial energy loss and formation of excited molecules, we fit the parent with Δtdelay ≥ 2.5 ps and the isomer with Δtdelay ≥ 5.5 ps. The biexponential decay of the C−H stretching width of isoCHBr2−Br points to a two-tiered relaxation process during which the newly formed solute loses energy to the solvent. The first time scale of about 3 ps, during which the isomer C−H stretching width decreases by half, monitors a fast removal of a large amount of vibrational energy. Strong interactions between the solute and solvent efficiently transfer energy to the solvent, and the large deviations from equilibrium geometries probably contribute to this fast decay. The complicated evolution of the

measured shift from parent (cm−1)

25 26 −9

25

25 67 4

23

calculations described in the Experimental Approach for CHBr3, CHI3, and possible photolysis products for both parent molecules. This harmonic frequency calculation correctly predicts the oscillator strength ratios for the measured symmetric CBr2 stretch, antisymmetric CBr2 stretch, and HCBr deformation modes in gas-phase CHBr2, and each calculated transition frequency differs by only 30 cm−1 from known experimental values.33 The agreement between the known and calculated CHBr2 vibrational transitions demonstrates the fidelity of the calculation. The 12 000 cm−1 of internal energy in the CHBr2 fragment that appears promptly after photolysis 28 can broaden the absorption band, 26 decreasing the already weak signal intensity below our 2904

dx.doi.org/10.1021/jp310737d | J. Phys. Chem. A 2013, 117, 2899−2907

The Journal of Physical Chemistry A

Article

to lie at much shorter wavelengths in a rare-gas matrix than in a solution of CHBr3. Figure 8 shows the evolution of the electronic spectra after photolysis of a 90% CHBr3 solution in CCl4. The Br-CHBr3

infrared bromoform feature, both in Figure 7 and in Figure 2, results from the evolving infrared transition of reforming CHBr3 molecules sitting atop the bleach of CHBr3 lost to isomerization, and we do not try to model this behavior with our fitting routine. A second vibrational-narrowing time scale of about 28 ps in the C−H stretch of iso-CHBr2−Br likely results from a second intermolecular energy transfer process from the partially stabilized isomer to the solvent. This time scale is somewhat faster than energy transfer from CH2I2 to liquid CCl4, for example, in which about 8600 cm−1 of excess energy leaves the molecule in about 58 ps.35 The unique bonding in iso-CHBr2− Br may be responsible for the accelerated intermolecular energy transfer. For example, one source of efficient coupling between the solute and solvent may be extended excursions toward the solvent made by the weakly bound terminal Br, potentially facilitating energy transfer to the solvent. Another possible mechanism is an accidental overlap of the low-frequency Br−Br vibration with a gateway mode of the solvent that efficiently directs energy away from the solute. A third source of increased coupling could be strengthened electronic interactions between the polarizable solvent and the isomer, which has significant ionic character.5,36 We do not determine explicitly the relaxation pathways in this work. Neither the transient vibrational nor electronic spectra evolve appreciably after about 100 ps, indicating that by this time most of the excess energy from photolysis has left the solute. The isomer persists well beyond the time scale of our measurement in roomtemperature CCl4. The overall evolution of the infrared spectra following photolysis of CHI3 is similar to that for CHBr3, but the spectra evolve more slowly in the CHI3 system. We use the same 267nm photolysis wavelength in both molecular systems, but excitation in CHBr3 occurs in the lowest-energy electronic absorption band while in CHI3 it occurs in the third accessible band. The different amounts of electronic, vibrational, and translational excitation in the photoproducts make direct comparison between the isomerization time scales of the two molecular systems difficult. The C−H stretch feature of isoCHI2−I rises about two times more slowly than its electronic transition, but both features persist at Δtdelay = 1 ns. Separation of the two radical fragments is an available pathway,9 and we calculate a 67-cm−1 shift to higher energy for the C−H stretch in the CHI2 radical37 but find that its transition strength decreases by a factor of 20 relative to the parent CHI3. We assign the transition at 3035 cm−1 to the C−H stretch of isoCHI2−I, however, because the vibrational feature grows over many tens of picoseconds as we expect for the isomer. 4.2. Isomerization in Neat CHBr3. The specificity of the transient infrared absorption spectra verifies that the feature near λprobe = 450 nm belongs to the iso-CHBr2−Br species. Transient electronic absorption spectra of concentrated bromoform solutions, however, show a prompt feature appearing near λprobe = 390 nm, which overlaps the rare-gas matrix-isolation spectrum of iso-CHBr2−Br reported by Reid and co-workers.5 We show below that the prompt feature we observe in solution comes from a transition to a charge-transfer state of the nascent Br atom adjacent to an intact CHBr3 molecule. The formation of such a contact pair is much rarer in dilute solutions. The coincidental overlap of the matrixisolation spectrum results from the solvatochromatism of isoCHBr2−Br, which causes the electronic transition of the isomer

Figure 8. Transient electronic absorption spectrum of the CHBr3−Br complex and iso-CHBr2−Br after photolysis of 90% (v/v) CHBr3 in CCl4. The prompt feature near 400 nm results from a transient pairing of the nascent Br atom and a nearby CHBr3 molecule. The decay of this prompt feature gives rise to the isomer feature near 490 nm, whose growth is the same as in dilute CHBr3 solutions.

feature near 390 nm appears within the 0.5-ps time resolution of our experiment, and its decay coincides with the growth of the iso-CHBr2−Br feature near 450 nm. The isosbestic point near 420 nm suggests that the CHBr3−Br complex is a precursor to iso-CHBr2−Br. The residual signal at 390 nm at large Δtdelay results from incomplete separation of the two electronic transitions and measures the high-energy side of the iso-CHBr2−Br feature. A sea of intact CHBr3 molecules surrounds the emerging Br atom in neat CHBr3, and the transient pairing prior to isomerization produces the prompt absorption near 390 nm. Diluting the sample with CCl4 decreases the probability of making the Br−CHBr3 pair. We explore the effect of dilution to verify the assignment. In Figure 9, we normalize the transient signal at all Δtdelay to the isomer signal at Δtdelay = 1 ns (top panel) for CHBr3 concentrations ranging from 100% to 2% (v/v). At Δtdelay = 2 ps, the prompt feature dominates the transient signal (bottom panel), and the normalization scheme highlights the changing signal size of the prompt feature relative to a constant number of isomers. As we expect, the intensity of charge-transfer transition decays linearly with decreasing CHBr3 fraction. Time-dependent density functional calculations, described in the Experimental Approach, provide transition energies and strengths of the Br−CHBr3 and iso-CHBr2−Br species, and the results are consistent with the spectra shown in Figure 9. In the calculations, we optimize Br-CHBr3 in implicit solvents that cover the range of published experimental results. For isoCHBr2−Br we use a similar procedure but also include one explicit solvent molecule. Including explicit solvent moves the calculated transitions by only a few nm. Figure 10 correlates the calculated transitions for the optimized structures of both the contact pair and the isomer. The prompt feature appears near 390 nm in cyclohexane,17 bromoform,17 acetonitrile,38 and carbon tetrachloride, which is consistent with the small range of transition wavelengths near 345 nm that we calculate for the Br−CHBr3 contact pair. This solvent-independent transition is responsible for the decreasingly intense transition that remains at the same wavelength in the bottom panel of Figure 9. The 2905

dx.doi.org/10.1021/jp310737d | J. Phys. Chem. A 2013, 117, 2899−2907

The Journal of Physical Chemistry A

Article

The calculated intensities of the CHBr3−Br and iso-CHBr2− Br electronic transitions are also consistent with the observed trends in the transient spectra. If we assume that each departing Br atom makes this complex in neat CHBr3 but that each Br has only a 15% probability of forming iso-CHBr2−Br,7 then the comparable intensities between the prompt and late features in neat CHBr3 requires the CHBr3−Br transition to be about 0.15 the strength of the iso-CHBr2−Br transition. The calculated 0.18 ratio of oscillator strengths between the two is consistent with the assignment. There are other potential photolysis products that could be responsible for the prompt transition. While the above arguments for the CHBr3−Br transition are strong, they are not definitive. We gather calculated electronic transitions for some possible photoproducts in Supporting Information Table S1. Incompletely relaxed isomer and CHBr2 radical, for example, are present in the sample, but their calculated electronic transitions are inconsistent with the observed trends. The spectrum of the charge-transfer transition of contact pairs between Br and different solvent molecules39 moves by 20 000 cm−1, and their transient signal intensities increase with solvent concentration, both of which are contrary to the observed trends. Tarnovsky and co-workers have also shown that the prompt feature results from a charge-transfer transition between the nascent Br atom and a nearby CHBr3 molecule.11 It is the decreasing CHBr3 concentration that lowers the probability of making the Br−CHBr3 transition en route to isomerization, explaining the trends in the transient electronic spectra in Figure 9 and in acetonitrile,38 cyclohexane,17 and argon.5

Figure 9. Transient electronic absorption spectra of iso-CHBr2−Br and CHBr3−Br in CCl4 for decreasing CHBr3 concentrations, normalized to the maximum signal at Δtdelay = 1 ns. Diluting bromoform in CCl4 shifts the electronic transition of the isomer at Δtdelay = 1 ps to higher energy (top), but at Δtdelay = 2 ps the intensity of the CHBr3−Br electronic transition decreases without appreciably shifting the transition energy (bottom).

5. SUMMARY We report the evolution of the C−H stretching transitions of iso-CHBr2−Br and iso-CHI2−I following photolysis of CHBr3 and CHI3, respectively, in carbon tetrachloride solutions. Photolysis leads to a prompt bleach in the C−H stretching region of the parent molecule. The transient infrared spectra show the appearance of new features with characteristic times ranging from 9 to 46 ps, and these features belong to the isomers. The time evolution of the vibrational isomer spectra after CHI3 photolysis agrees with previously assigned electronic absorption spectra10 but differs for CHBr3 photolysis.17 Comparison of these new vibrational spectra allows reassignment of the iso-CHBr2−Br electronic transition. Electronic structure calculations show that the iso-CHBr2−Br electronic transition shifts about 4000 cm−1 between Ar and CHBr3 solvents, and this behavior explains the discrepancy between the liquid-phase and matrix-isolation spectra. Electronic structure calculations of the CHBr3−Br complex reliably capture the transition energy, intensity, and solvent dependence of the prompt feature in the transient electronic absorption spectra. Measurement of the iso-CHBr2−Br or CHBr3−Br electronic transitions in polarizable cryogenic matrices can complete the picture of this condensed-phase isomerization. The bond specificity of transient vibrational spectroscopy provides additional insight into these photolytically driven isomerization pathways.

Figure 10. Correlation between calculated and observed electronic transitions of iso-CHBr2−Br and Br−CHBr3 in several solvents. The CHBr3−Br transition (blue, open circles) is independent of the solvent. The isomer transition depends strongly on the solvent (dark red, closed circles) and only shifts slightly in energy after including one explicit solvent molecule (light red circles). The line is a guide to the eye.

■

isomer appears near 400 nm in an Ar matrix5 and near 510 nm in acetonitrile solutions,38 correlating well with our calculated transitions at 385 nm in Ar and 469 nm in acetonitrile. The strong solvent dependence causes the large shift in the top panel of Figure 9, likely reflecting the ion-pair character of the isomer described by Reid and co-workers.5

ASSOCIATED CONTENT

S Supporting Information *

Table S1 providing calculated electronic transition wavelengths and intensities of bromoform photolysis products. This material is available free of charge via the Internet at http://pubs.acs.org. 2906

dx.doi.org/10.1021/jp310737d | J. Phys. Chem. A 2013, 117, 2899−2907

The Journal of Physical Chemistry A

■

Article

P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09; Gaussian, Inc.: Wallingford, CT, 2009. (24) Jacox, M. E. Chem. Soc. Rev. 2002, 31, 108−115. (25) Jacox, M. E. Chem. Phys. 1994, 189, 149−170. (26) Baggott, J. E.; Jasinski, J. M.; Zewail, A. H.; Vonputtkamer, K.; Quack, M.; Haas, Y.; Troe, J.; Rabinovitch, B. S.; Heller, E. J.; Wittig, C.; et al. Faraday Discuss. 1983, 341−363. (27) Lugez, C.; Schriver, A.; Schrivermazzuoli, L.; Lasson, E.; Nielsen, C. J. J. Phys. Chem. 1993, 97, 11617−11624. (28) Xu, D. D.; Francisco, J. S.; Huang, J.; Jackson, W. M. J. Chem. Phys. 2002, 117, 2578−2585. (29) Tuck, A. F. J. Chem. Soc., Faraday Trans. 2 1977, 73, 689−708. (30) Furuyama, S.; Golden, D. M.; Benson, S. W. J. Am. Chem. Soc. 1969, 91, 7564−&. (31) Odelius, M.; Kadi, M.; Davidsson, J.; Tarnovsky, A. N. J. Chem. Phys. 2004, 121, 2208−2214. (32) Grimm, C.; Kling, M.; Schroeder, J.; Troe, J.; Zerbs, J. Isr. J. Chem. 2003, 43, 305−317. (33) Jacox, M. E. In NIST Chemistry WebBook; NIST Standard Reference Database Number 69; Linstrom, P. J., Mallard, W. G., Eds.; NIST: Gaithersburg, MD. (34) Kalume, A.; George, L.; Reid, S. A. J. Phys. Chem. Lett. 2010, 1, 3090−3095. (35) Elles, C. G.; Bingemann, D.; Heckscher, M. M.; Crim, F. F. J. Chem. Phys. 2003, 118, 5587−5595. (36) Elles, C. G.; Crim, F. F. Annu. Rev. Phys. Chem. 2006, 57, 273− 302. (37) Zheng, X. M.; Phillips, D. L. Chem. Phys. Lett. 2000, 324, 175− 182. (38) Carrier, S. L. Ph.D. Thesis, University of WisconsinMadison, 2009. (39) George, L.; Kalume, A.; Esselman, B.; McMahon, R. J.; Reid, S. A. J. Phys. Chem. A 2011, 115, 9820−9827.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Present Address †

T.J.P.: School of Chemistry, University of Bristol, Bristol BS8 1TS, United Kingdom Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS A grant from the National Science Foundation (CHE0910917) supports the research described here. The computations used computer resources supported by the National Science Foundation (CHE-0840494). The authors thank Alexander Tarnovsky and Scott Reid for useful discussions.

■

REFERENCES

(1) Apkarian, V. A.; Schwentner, N. Chem. Rev. 1999, 99, 1481− 1514. (2) Maier, G.; Reisenauer, H. P. Angew. Chem., Int. Ed. 1986, 25, 819−822. (3) Maier, G.; Reisenauer, H. P.; Hu, J.; Schaad, L. J.; Hess, B. A. J. Am. Chem. Soc. 1990, 112, 5117−5122. (4) El-Khoury, P. Z.; George, L.; Kalume, A.; Reid, S. A.; Ault, B. S.; Tarnovsky, A. N. J. Chem. Phys. 2010, 132, 124501. (5) George, L.; Kalume, A.; Esselman, B. J.; Wagner, J.; McMahon, R. J.; Reid, S. A. J. Chem. Phys. 2011, 135, 124503. (6) Zheng, X. M.; Kwok, W. M.; Phillips, D. L. J. Phys. Chem. A 2000, 104, 10464−10470. (7) Preston, T. J.; Dutta, M.; Esselman, B. J.; Kalume, A.; George, L.; McMahon, R. J.; Reid, S. A.; Crim, F. F. J. Chem. Phys. 2011, 135, 114503. (8) Tarnovsky, A. N.; Wall, M.; Rasmusson, M.; Pascher, T.; Akesson, E. J. Chin. Chem. Soc. 2000, 47, 769−772. (9) Wall, M.; Tarnovsky, A. N.; Pascher, T.; Sundstrom, V.; Akesson, E. J. Phys. Chem. A 2003, 107, 211−217. (10) El-Khoury, P. Z.; Kwok, W. M.; Guan, X. G.; Ma, C. S.; Phillips, D. L.; Tarnovsky, A. N. ChemPhysChem 2009, 10, 1895−1900. (11) Pal, S. K.; Mereshchenko, A. S.; Butaeva, E. V.; El-Khoury, P. Z.; Tarnovsky, A. N. J. Chem. Phys. 2013, in press. (12) Phillips, D. L.; Fang, W. H.; Zheng, X. M. J. Am. Chem. Soc. 2001, 123, 4197−4203. (13) Phillips, D. L.; Fang, W. H.; Zheng, X.; Li, Y. L.; Wang, D.; Kwok, W. M. Curr. Org. Chem. 2004, 8, 739−755. (14) Tarnovsky, A. N.; Sundstrom, V.; Akesson, E.; Pascher, T. J. Phys. Chem. A 2004, 108, 237−249. (15) Tarnovsky, A. N.; Sundstrom, V.; Akesson, E.; Pascher, T. J. Phys. Chem. A 2005, 109, 954−954. (16) Tarnovsky, A. N.; Pascher, I.; Pascher, T. J. Phys. Chem. A 2007, 111, 11814−11817. (17) Carrier, S. L.; Preston, T. J.; Dutta, M.; Crowther, A. C.; Crim, F. F. J. Phys. Chem. A 2010, 114, 1548−1555. (18) Kwok, W. M.; Ma, C. S.; Phillips, D.; Parker, A. W.; Towrie, M.; Matousek, P.; Phillips, D. L. Chem. Phys. Lett. 2001, 341, 292−298. (19) Zheng, X. M.; Fang, W. H.; Phillips, D. L. J. Chem. Phys. 2000, 113, 10934−10946. (20) Crowther, A. C.; Carrier, S. L.; Preston, T. J.; Crim, F. F. J. Phys. Chem. A 2008, 112, 12081−12089. (21) Preston, T. J. Ph.D. Thesis. University of WisconsinMadison, 2012. (22) Zhang, G.; Thomas, J. K. Photochem. Photobiol. 2006, 82, 158− 162. (23) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. 2907

dx.doi.org/10.1021/jp310737d | J. Phys. Chem. A 2013, 117, 2899−2907