Methyl-torsion-facilitated internal energy delocalization following electronic excitation in m -fluorotoluene: Can meta and para substitution be directly compared?

Coupling between vibrations, and between vibrations and torsions—a generalization of intramolecular vibrational redistribution (IVR)—provides routes to internal energy delocalization, which can stabilize molecules following photoexcitation. Following earlier work on p -fluorotoluene ( p FT), this study focuses on m -fluorotoluene ( m FT) as probed via the S 1 ↔ S 0 electronic transitions and the D 0 + ← S 1 ionization, using two-dimensional laser-induced fluorescence and zero-electron-kinetic energy spectroscopy, respectively. Wavenumbers are reported for a number of vibrations in the S 0 , S 1 , and D 0 + states and found to compare well to those calculated. In addition, features are seen in the m FT spectra, not commented on in previous studies, which can be assigned to transitions involving vibration–torsion (“vibtor”) levels. Comparisons to the previous work on both m -difluorobenzene and m FT are also made, and some earlier assignments are revised. At lower wavenumbers, well-defined interactions between vibrational and vibtor levels are deduced—termed “restricted IVR,” while at higher wavenumbers, such interactions evolve into more-complicated interactions, moving toward the “statistical IVR” regime. It is then concluded that a comparison between m FT and p FT is less straightforward than implied in earlier studies.


I. INTRODUCTION
Anharmonic coupling in molecules leads to delocalization, and so dispersal, of internal energy within a molecule-an important aspect to enhanced photostability. [1][2][3][4] Which of the vibrations in a molecule can couple, and to what extent, depends on a number of factors, but having the same symmetry and being close in energy are two key considerations, 1 along with the relative motions of the atoms in those vibrations. Such coupling leads to actual vibrational motions having mixed character, and this can be discerned through an analysis of the vibrational activity in electronic and photoelectron spectra, where vibrational eigenfunctions of one electronic state are projected onto those of another. To first order, vibrational fundamentals do not couple anharmonically, but sometimes activity in fundamentals other than that excited can be seen in experimental spectra. This can be caused by changes in geometry that lead to significant Franck-Condon factors (FCFs), or as a result of Duschinsky rotations. 5 Methylation in biomolecules has been invoked as a key factor in photoinduced carcinogenicity, 6 and consequently, an understanding of the role of methyl groups in the modification of a molecule's photobehavior is of key importance. Parmenter and coworkers have published numerous studies on the effect of methylation on intramolecular vibrational redistribution (IVR), with the most pertinent study here being that of Timbers et al. 7 In that work, it was concluded that m-fluorotoluene (mFT) undergoes intramolecular vibrational redistribution (IVR) more than an order of magnitude faster than p-fluorotoluene (pFT) and, hence, that the location of substituents is important in internal energy delocalization.
Recently, we published resonance-enhanced multiphoton ionization (REMPI) and zero-electron-kinetic-energy (ZEKE) studies of the low-wavenumber regions of mFT 8 and m-chlorotoluene (mClT), 9 which mainly focused on torsions and some vibration-torsion (vibtor) levels. The mFT study complemented the two-dimensional laser-induced fluorescence (2D-LIF) study of Stewart et al., 10 who examined the first 350 cm −1 of the S 1 ← S 0 transition and the first 550 cm −1 of the S 1 → S 0 transition via a mixture of LIF, dispersed fluorescence (DF), and 2D-LIF. In these studies, the spectra were assigned in terms of transitions involving torsional, vibtor, and vibrational levels in the S 0 and S 1 states. Earlier, Ito and co-workers [11][12][13][14] reported laser-induced fluorescence (LIF), DF, REMPI, and ZEKE spectra of the low-wavenumber region of mFT. Very recently, we compared the activity of three Duschinsky-mixed vibrations for mFT and mClT, using a combination of 2D-LIF and ZEKE spectroscopy, finding that the spectra were exquisitely sensitive to small changes in molecular mass and electronic structure. 15 Previously, we have studied the pFT molecule using REMPI, ZEKE, and 2D-LIF spectroscopy, showing how IVR evolves through the restricted to statistical regimes. It was seen that this is not a smooth evolution and that the methyl group plays a key role in this. We were able to identify key interactions and energy delocalization routes. [16][17][18][19][20][21][22][23] In the present work, we extend our studies to the S 1 ← S 0 spectrum of mFT in the range 0 cm −1 -1350 cm −1 above the origin, using 2D-LIF, REMPI, and ZEKE spectroscopy. This encompasses a number of fundamental, overtone, combination, and vibtor levels. As with pFT, it is found that there is again an evolution from interactions between small numbers of levels, termed restricted IVR, at low wavenumber to widespread interactions, approaching statistical IVR, at higher wavenumbers. We identify specific interactions and derive wavenumbers for a number of vibrations across the S 0 , S 1 , and D 0 + electronic states. Over several years, our group has been examining the spectroscopy of fluoro-substituted and methyl-substituted benzene molecules. In doing so, we have examined the vibrational labeling of these molecules, putting forward general schemes for the ring-based vibrations of monosubstituted benzene molecules 24 and for each of the three isomeric classes of disubstituted benzene molecules. [25][26][27] These schemes were developed since neither Wilson 28 /Varsányi 29 nor Mulliken 30 /Herzberg 31 notations are appropriate. Indeed, separate schemes were required for these four classes of substituted benzene molecules, since the forms of the vibrations differ substantially between them. This implies that the comparison of vibrational activity between classes of substituted benzene molecules poses difficulties; however, the schemes do allow for direct comparison within each isomeric class. These labeling schemes are based on monofluorobenzene and difluorobenzene molecules, with consistent labels across both symmetric and asymmetric substitutions in the latter cases. They have allowed vibrational activity to be compared across spectra of p-difluorobenzene (pDFB), p-chlorofluorobenzene (pClFB), pFT, and p-xylene (pXyl)-see Refs. 17 and 32-although these ideas flow through our recent work across these molecules.
In the present case, a comparison will be made between the previously published detailed laser fluorescence study of m-difluorobenzene (mDFB) 33 and mFT, for which we expect the vibrational activity to be similar if the same vibrational numbering scheme is used-as will be seen, this expectation is largely borne out. Of course, the spectrum of mFT will be complicated by the contributions from torsion and vibtor levels, as noted above.

II. EXPERIMENTAL
The REMPI/ZEKE 34 and 2D-LIF 18 apparatuses are the same as those employed recently. In both experiments for mFT, a freejet expansion of mFT (Sigma-Aldrich, 99% purity) in 2 bar Ar was employed.
For the 2D-LIF spectra, the free-jet expansion was intersected at X/D ∼ 20 by the frequency-doubled output of a single dye laser (Sirah Cobra-Stretch), operating with Coumarin 503 and pumped with the third harmonic of a Surelite III Nd:YAG laser. The fluorescence was collected, collimated, and focused onto the entrance slits of a 1.5 m Czerny-Turner spectrometer (Sciencetech 9150) operating in single-pass mode, dispersed by a 3600 groove/mm grating, allowing ∼300 cm −1 windows of the dispersed fluorescence to be collected by a CCD camera (Andor iStar DH334T). At a fixed grating angle of the spectrometer, the excitation laser was scanned, and at each excitation wavenumber, the camera image was accumulated for 2000 laser shots. This allowed a plot to be produced of fluorescence intensity vs both the excitation laser wavenumber and the wavenumber of the emitted and dispersed fluorescence, termed a 2D-LIF spectrum. 35,36 Band positions for the 2D-LIF spectra are given for the estimated band center.
For the REMPI and ZEKE spectra of mFT, the focused, frequency-doubled outputs of two dye lasers (Sirah Cobra-Stretch) were overlapped spatially and temporally and passed through a vacuum chamber coaxially and counterpropagating, where they intersected the free-jet expansion. The excitation laser was operated with Coumarin 503 and pumped with the third harmonic (355 nm) of a Surelite III Nd:YAG laser, while the ionization laser was operated with Pyrromethene 597 and pumped with the second harmonic (532 nm) of a Surelite I Nd:YAG laser. The jet expansion passed between two biased electrical grids located in the extraction region of a time-of-flight mass spectrometer, which was employed in the REMPI experiments. These grids were also used in the ZEKE experiments by application of pulsed voltages, giving typical fields of ∼10 V cm −1 , after a delay of up to 2 μs; this delay was minimized while avoiding the introduction of excess noise from the prompt electron signal. The resulting ZEKE bands had widths of ∼5 cm −1 -7 cm −1 . Electron and ion signals were recorded on separate sets of microchannel plates. Band positions for REMPI and ZEKE bands are given for the maximum, and ZEKE spectra were generally obtained when exciting through the intermediate band maximum.
For the REMPI spectrum of mDFB, a free-jet expansion of mDFB (Sigma-Aldrich, 99% purity) in 5 bar Ar was employed.

III. RESULTS AND ASSIGNMENTS
A. Nomenclature and labeling 1. Vibrational and torsional labeling We shall employ the Di labels 27 for the mFT vibrations, as used in Refs. 8-10; this Cs point group labeling scheme is based on the vibrations of the mDFB molecule. As such, we shall transcribe the Wilson/Varsányi labels in Ref. 11  Since we shall also be referring to the methyl torsional motion for mFT, for which use of the G 6 molecular symmetry group (MSG) is appropriate, we shall employ those symmetry labels throughout. The torsional levels will be labeled via their m quantum number, 8,10 and the correspondence between the Cs point group labels and the G 6 MSG labels is given in Table II. To calculate the overall symmetry of a vibtor level, it is necessary to use the corresponding G 6 label for the vibration and then find the direct product with the symmetry of the torsion (Table II), noting that a C 3v point group direct product table can be used, since the G 6 MSG and the C 3v point group are isomorphic. The torsional levels in mFT are labeled with the signed quantum number m (m = 0, 1, 2, . . .). The m = 0 level is singly degenerate, while levels with |m| ≠ 3n (n = 1, 2, . . .) are doubly degenerate, consisting of +/− pairs, and levels with m = 3n form linear combinations of the +/− pairs that split in energy under the influence of the torsional potential; these are labeled m = 3(+) and m = 3(−). 8 Under the free-jet expansion conditions employed here, almost all molecules are expected to be cooled to their zero-point vibrational level, and thus essentially, all S 1 ← S 0 excitations are expected to originate from this level. In contrast, owing to nuclear-spin and rotational symmetry, the mFT molecules can be in one of the m = 0 or m = 1 torsional levels, with approximately equal population in each; 38 residual population in the m = 2 level is also seen. [8][9][10] The available experimental vibrational wavenumbers for mFT are presented in Table III The level of theory employed is given in the footnotes of Tables I and III, and Gaussian 16 39 was used for all of these calculations.

Coupling and transitions
In the usual way, vibrational transitions will be indicated by the number, i, of the Di vibration, followed by a superscript/subscript specifying the number of quanta in the upper/lower states, respectively. When required, torsional transitions will be indicated by m   followed by its superscripted value, and vibtor transitions will be indicated by a combination of the vibrational and torsional transition labels. When designating transitions, we shall generally omit the initial level, since it will be evident from either the jet-cooled conditions or the specified intermediate level.
As has become common usage, we will generally refer to a level using the notation of a transition, with the level indicated by the specified quantum numbers, with superscripts indicating levels in the S 1 state and subscripts indicating levels in the S 0 state. Since we will also be referring to transitions and levels involving the ARTICLE scitation.org/journal/adv ground state cation, D 0 + , we shall indicate those as superscripts, but with a single, additional, preceding superscripted "+" sign. Relative wavenumbers of the levels will be given with respect to the relevant zero-point vibrational level with m = 0, in each electronic state.
For cases where the geometry and the torsional potential are both similar in the S 1 and D 0 + states, the most intense transition is usually expected to be that for which no changes in the torsional and/or vibrational quantum numbers occur, designated as Δm = 0, Δv = 0, or Δ(v, m) = 0 transitions, as appropriate. However, as will be seen (and as reported in Refs. 8, 9, 14, and 15), the Δm = 0 and Δ(v, m) = 0 transitions are almost always not the most intense bands in the ZEKE spectra for mFT, indicative of a significant change in the torsional potential upon ionization. The intensities of low-wavenumber features in the S 1 ↔ S 0 transitions have been discussed in Ref. 10, and reference will be made to that work when appropriate.
If two levels are close in wavenumber and have the same overall symmetry, then (except between vibrational fundamentals, to first order) interactions can occur, with the simplest example being the anharmonic interaction between two vibrational levels-the classic Fermi resonance. 40 Such couplings are only expected to be significant for small changes in the vibrational quantum number, Δv ≈ 3, and for levels that lie close in energy. 41 For molecules that contain a hindered internal rotor, such as mFT and pFT, and if vibration-torsion coupling occurs, then interactions can also involve torsional or "vibtor" levels. This is expected to be significant only for changes in the torsional quantum number, Δm, of ±3 or ±6 in descending order of likely strength. 42 The result of such interactions is the formation of eigenstates with mixed character. Often, the resulting eigenstates will be referred to by the dominant contribution, with the context implying if an admixture is present.
B. An overview of the REMPI spectrum In Fig. 1, we show the S 1 ← S 0 REMPI spectrum of mFT over the range 0 cm −1 -1350 cm −1 . It may be seen to be rich in structure, some of which has been assigned previously. 8,[10][11][12][13][14][15] We note the good agreement with the appearance of the LIF spectrum in Ref. 11, which covers a similar range. Wavenumbers of some of the S 1 vibrations are in dispute. 10 We highlight that the 0 cm −1 -350 cm −1 region has been discussed in depth in Refs. 8 and 10, and the 400 cm −1 -480 cm −1 region was the focus in Ref. 15. Also in Fig. 1, we present the corresponding REMPI spectrum of mDFB, which compares very well with the LIF spectrum presented in Ref. 33.

ARTICLE scitation.org/journal/adv
Overall, ignoring the torsional transitions for mFT, the correspondence between the main activity in both REMPI spectra in Fig. 1 is striking and adds confidence to the assignments for mFT discussed later. A number of values and assignments in Table III are from the present work and will be discussed in Subsections III C-III G, where we break the discussion up into five main regions. Through these five regions, it will be seen that the coupling evolves from being absent, through well-defined coupling between a small number of levels at low wavenumber, into widespread coupling, approaching statistical IVR at higher wavenumbers.
In some of the figures, vertically integrated traces of the 2D-LIF spectra are presented. Each of these looks very similar to the corresponding section of the REMPI spectrum, confirming that the fluorescence collected is representative of the absorption spectrum in that region.
C. 2D-LIF and ZEKE spectra via the origin m = 0 and m = 1 levels In Fig. 2, we show the 2D-LIF spectrum recorded when exciting through the pure torsional m 0 and m 1 excitations, and many assignments are also shown. The 0 cm −1 -550 cm −1 region of the emission spectrum has been assigned and discussed in depth by Stewart et al., 10 and we concurred with those assignments in our ZEKE study. 8 In Ref. 10, only the 0 cm −1 -65 cm −1 region of the 2D-LIF spectrum via the m 0 and m 1 bands was presented, although DF spectra were presented up to 550 cm −1 , with full assignments. These will prove useful in assigning the 2D-LIF spectra in the present work, when we excite at higher wavenumbers.
The assignment of these vibrational bands allows S 0 vibrational wavenumbers to be established, which are included in Table III, some of which were reported by Stewart et al. 10 These values are close to those established by IR and Raman spectroscopy 37 and those discussed in Ref. 27, providing further confirmation of the assignments. Other features can be identified as vibtor levels associated with these vibrational transitions, and indeed, subject to sensitivity, we expect to see the main pattern of vibtor and related transitions that are observed for the origin, for each vibrational transition.
In Fig. 3, we show the ZEKE spectra recorded via both m 0 and m 1 . The regions up to ∼850 cm −1 were assigned in Ref. 8; in the present work, the spectra are extended up to ∼1850 cm −1 and show more vibrational bands with their associated vibtor structure. As will be seen, these help in the assignment of other ZEKE spectra presented later. The vibrational wavenumbers arising from these spectra are also included in Table III.
The integrated 2D-LIF trace covering this region of the S 1 ← S 0 excitation is shown at the top of Fig. 4. The three most intense features in this region are Y 1 and the overlapped X 1 and 18 1 bands. The S 1 D X and D Y vibrations are highly mixed forms of the S 0 D 19 and D 20 vibrations, and these assignments, alongside the Duschinsky mixing that gives rise to the mixed character, have been discussed in Ref. 15. In Fig. 4, we also show an extended 2D-LIF spectrum recorded across this region. It can be seen that numerous features can be identified in the 2D-LIF image that are not obviously associated with discrete features in the integrated 2D-LIF (nor REMPI) spectrum; this highlights one aspect of the extra information that 2D-LIF spectra provide. The assignments of many of the bands are straightforward, combining the knowledge of the torsional levels from Refs. 8 and 10 and the vibrational levels from Ref. 15, and these are indicated in the figure, with numerous features arising from Δ(v, m) = 0 transitions. However, there were also numerous cases where the assignment was less obvious, and some of these will be discussed in the following.
We see a 2D-LIF feature at (428, 675) cm −1 , whose assignment is to the Δ(v, m) = 0 band (28 1  Overall, this behavior is reminiscent of the fact that the (29 1 30 1 m 1 , 29 1 30 1 m 1 ) band was also observed to be very much weaker than (29 1 30 1 m 0 , 29 1 30 1 m 0 ). 10 The reason for the weakness of the transition involving m = 1 was not entirely clear, but the assigned (29 1 30 1 m 1 , 29 1 30 1 m 1 ) band was suggested as being significantly shifted from its expected excitation position, 10 and it was hypothesized that there could be a 29 1 30 1 m 1 . . .m 7 interaction, although this would be a (Δv = 2, Δm = 6) interaction, and so perhaps this is not expected to be strong. In general terms, although we do expect more widespread vibtor coupling involving e symmetry torsions, there is no evidence of such coupling for the 28 1 29 1 m 1 level in the 2D-LIF spectrum, and so the absence of the m = 1 component is somewhat puzzling.
In Fig. 5(a), we show the ZEKE spectrum recorded when exciting at 427 cm −1 , which shows a strong Δm = 3 band, + 28 1 29 1 m 3(+) , in line with expectations when exciting through totally symmetric vibrations; 8 there are also the expected weaker + 28 1 29 1 m 0 and + 28 1 29 1 m 6(+) bands. There is no clear evidence for the ZEKE bands expected to arise from ionization from 28 1 29 1 m 1 , confirming that the major contribution to the REMPI spectrum at this wavenumber is indeed 28 1 We also see evidence in the 2D-LIF spectrum (

ARTICLE
scitation.org/journal/adv there being 27 1 m 3(−) activity also at the Y 1 m 0 excitation position, and the "reverse" activity with the 20 1 m 0 band extending to higher excitation wavenumbers. The width of the bands, and their limited activity, makes it difficult to ascertain whether shifts from expected band positions have occurred, which would be another signature of such an interaction.
Owing to the energetic closeness of the 27 1 m 3(−) and 28 1 29 1 m 0 bands in the S 1 state (Fig. 4), it may be expected that when recording ZEKE spectra when exciting at 0 0 + 427 cm −1 , we might also observe bands arising from excitation of 27 1 m 3(−) . The weak feature at 816 cm −1 in the ZEKE spectrum when exciting here, shown in Fig. 5(a), is at about the expected wavenumber for the Δm = 3 band, + 27 1 m 6(−) , to which it is tentatively assigned. This is in line with the increased propensity of Δm = 3 transitions seen in the ZEKE spectra. 8 Overall, therefore, the ZEKE spectrum recorded when exciting at 427 cm −1 is consistent with the assignments shown in the 2D-LIF spectrum in Fig. 4 but is not definitive. In addition, even though the 27 1 m 3(−) and 28 1 29 1 m 0 transitions overlap to some extent, no evidence of an interaction is seen.
A reasonable assignment for the weak 2D-LIF features in Fig. 4 that appear at an excitation wavenumber of 435 cm −1 can be proposed. This is close to the expected excitation wavenumber for 29 2 m 3(+) , and a ZEKE spectrum recorded at this position, see Fig. 5(b), shows bands that are assignable to + 29 2 m 0 and + 29 2 m 6(+) . In line with observations in our previous paper, 8 the + 29 2 m 3(+) band is expected to be weak and is not discernible in the spectrum.
The prominent feature at (514, 880) cm −1 in Fig. 4 is straightforwardly assignable as (28 2 m 1 , 28 2 m 1 ), based on the appearance of the corresponding m = 2 component at (515, 899) cm −1 . This is consistent with the observation of 28 2 in the case of mDFB. 33 Accompanying the prominent band is a weaker feature at (522, 879) cm −1 , which can be assigned as (28 2 m 0 , 28 2 m 0 ). It is initially surprising that this feature is so weak, as usually the m = 0 and m = 1 components for a vibrational band have comparable intensities, with the m = 1 band being slightly more intense (see Fig. 2). A clue as to the interpretation of these anomalous relative intensities comes from the ZEKE spectra, presented in Fig. 6, recorded at positions corresponding to the maxima of each of these features. The spectrum when exciting at 515 cm −1 is seen to be more complicated than the spectrum recorded at 520 cm −1 . With insight from the appearance of the ZEKE spectrum via m 1 [see Fig. 3(a)], 8 which demonstrates prominent + m 2 , + m 4 , and + m 5 vibtor bands, and those via m 0 levels, which show prominent + m 3(+) and + m 6(+) vibtor bands [ Fig. 3(b)], the bands arising from both the 28 2 m 1 and 28 2 m 0 levels can be straightforwardly identified. Two aspects of the spectra then become clear: first, there are bands arising from 28 2 m 0 in both spectra, and second, there are other bands in the spectra. The + 19 1 m 1 and + 20 1 m 1 bands in the 515 cm −1 ZEKE spectrum, and the + 21 1 29 1 m 3(−) band in the 520 cm −1 spectrum, arise from Franck-Condon (FC)-like activity; however, there are some other relatively intense bands that appear in both sets of spectra. Our favored assignment of these is to + 28 1 29 1 m x vibtor transitions, and their assignment suggests their source is 28 1 29 1 m 3(+) , with the + 28 1 29 1 m 3(+) / + 28 2 m 0 band being clearly seen in Fig. 6(a), but not immediately in Fig. 6(b), where its position is indicated.
The expected position of 28 1 29 1 m 3(+) in S 1 is around 522 cm −1 , and this has the same symmetry, a 1 , as the 28 2 m 0 level, which is also expected to be close to this position; consequently, these two levels can interact. Furthermore, they differ by Δv = 2 and Δm = 3, making the suggested interaction sensible. We thus suggest a ARTICLE scitation.org/journal/adv 28 2 m 0 . . .28 1 29 1 m 3(+) interaction that gives rise to two eigenstates of mixed composition. The relative intensities of the ZEKE bands suggest that the 28 2 m 0 . . .28 1 29 1 m 3(+) state (note that the leading term implies the dominant contribution to the eigenstate) gives rise to the higher, ∼520 cm −1 , feature, while the 28 1 29 1 m 3(+) . . .28 2 m 0 eigenstate gives the lower feature; its transition is overlapped with that of 28 2 m 1 at ∼515 cm −1 , explaining the more-complicated structure in the ZEKE spectrum recorded at that position, Fig. 6(a). This assignment is also consistent with the appearance of the 2D-LIF spectrum (Fig. 4), where the 28 1 29 1 m 3(+) emission band has intensity across the 513 cm −1 -523 cm −1 region, consistent with the proposed interaction. We note the coincidence in the wavenumber of 17 1 and 28 1 29 1 m 3(+) , which makes the interpretation of the 2D-LIF spectrum less straightforward initially, but we accept the 28 1  The region of the REMPI spectrum of mFT between 670 cm −1 and 750 cm −1 can be seen to consist of three main features (Fig. 1), with the lowest wavenumber of these excitation bands being assigned to 17 1 (designated 1 1 by Okuyama et al. 11 ). The other two features were assigned to 12 1 and 13 1 (designated 9b 1 and 18b 1 by Okuyama et al. 11 ), but these two assignments have been questioned by Stewart et al. 10 When comparing to the mDFB spectrum (see Fig. 1), it can be seen that similar activity is seen, with Graham and Kable 33 having assigned these four main features to 26 1 28 1 , 17 1 , 27 2 and an overlapped feature consisting of 24 1 28 1 and 25 1 28 1 . In a future paper, 43 we shall examine the two higher-wavenumber features in this region of the mFT spectrum in more detail, but for the purposes of the present work, we concentrate on the lowest wavenumber feature.
The integrated 2D-LIF trace is shown at the top of Fig. 7(c), which closely resembles the corresponding section of the REMPI spectrum. This excitation feature is assigned to the two m components of the 17 1 transition, in agreement with the assignment of Okuyama et al. 11 and consistent with the mDFB spectrum 33 (Fig. 1). Both the 2D-LIF spectrum [ Fig. 7(c)] and the ZEKE spectra [ Fig. 7(b)] support the 17 1 m 1 and 17 1 m 0 assignments; however, there are significant additional features in both spectra. Furthermore, although, when the 2D-LIF spectrum is examined, the strongest emission does indeed correspond to the two m components of 17 1 , it is clear that the 17 1 m 0 emission band is more intense

ARTICLE
scitation.org/journal/adv than 17 1 m 1 -also reflected in the associated vibtor levels. The 15 1 bands are notably intense-cf. the 960 cm −1 feature (Sec. III F). Despite these observations, the Duschinsky matrices [ Fig. 7(a)] indicate that there is almost no mixing between these modes during excitation or ionization. (In the matrices, the depth of shading of D 15 implies that there is some predicted mode mixing of D 15 between electronic states, but this is composed of several minor contributions from vibrational modes other than D 17 and so is not discussed further here.) In contrast, via the origin, the m 1 band is more intense than m 0 (see Fig. 2), and this is the case with each vibrational emission band. We interpret this as an indication that the 17 1 m 1 (e symmetry) S 1 level emits to more S 0 levels than 17 1 m 0 (a 1 symmetry); furthermore, the ZEKE spectrum recorded via 17 1 m 1 shows more bands than that recorded via 17 1 m 0 . Taken together, it is concluded that the 17 1 m 1 level is likely interacting with other e symmetry levels.
In the 0 cm −1 -550 cm −1 emission region (not shown) of the 2D-LIF spectrum recorded via 17 1 m 0,1 , essentially all of the features seen via the origin (see Fig. 2 and Ref. 10) can also be seen. In the region shown in Fig. 7(c), emission to a number of vibrational levels and their associated vibtor levels can be seen, and a selection of assignments is shown.
The largely discrete nature of the emission spectrum suggests interactions will be between a small number of levels, and we initially considered an interaction with an m = 2 level for the most efficient coupling with 17 1 m 1 (Δm = 3, recalling that the m quantum number is signed) and with Δv ≤ 3. This led to the assignment of the emission band at (682, 784) cm −1 to 18 1 29 1 m 2 , which is supported by the observation of the weaker 19 1 29 1 m 2 emission band at (682, 767) cm −1 . There are other emission bands that can be assigned when exciting at 682 cm −1 -these can be associated either with activity from 18 1 29 1 m 2 or with further interactions between e symmetry levels in S 1 . Although it is not possible to be definitive, the latter is supported by the fact that a number of levels are at a wavenumber position in S 1 that suggests that they could interact with 18 1 29 1 m 2 , each related by ≤ 3 changes in v and/or m. As such, we conclude that a number of concurrent stepwise interactions are occurring, which link 17 1 m 1 to a number of e symmetry levels, explaining the significantly lower 17 1 m 1 intensity compared to that of 17 1 m 0 . For example, 18 1 29 1 m 2 can interact with 21 1 29 2 m 2 via a Δv = 3 interaction, and the latter can then undergo a further Δv = 3 interaction with 29 3 30 1 m 2 . That multiple interactions are likely to be occurring is supported by the significant amount of underlying activity that can be seen in the 2D-LIF and ZEKE spectra in Fig. 7.
In general, for solely anharmonic vibrational coupling, each m level of a particular vibration would behave similarly. However, in the case under discussion, the coupling is with a vibtor level that can only interact with one of the m levels of 17 1 , and generally, there is a greater likelihood of coupling between e symmetry levels than for a 1 (or a 2 )-see Sec. IV. Another possibility for an m-specific interaction with 17 1 m 1 would be with the m = 1 level of an a 2 symmetry vibrational energy level, as the vibtor symmetry would then be e in both cases. However, the coupling mechanism could not simply be anharmonicity but would have to result from a breakdown in the separability of vibrational and torsional motions. It is also possible for m-specific interactions to occur with other vibtor levels of a 1 symmetry, but we do not see evidence of such interactions here.
The ZEKE spectrum via 17 1 m 1 , presented in Fig. 7(b)(i), is consistent, but not definitive, with regard to the suggested couplings: for example, although we do see + 18 1 m 2 , we do not see a clear band for + 18 1 29 1 m 2 , which would be expected at 805 cm −1 and, if present, would be overlapped by the + 17 1 m 2 band. Furthermore, the expected Δm = 3 ZEKE band would be + 18 1 29 1 m 5 , anticipated at 952 cm −1 , which would overlap the + 17 1 m 5 band, and could contribute to the higher-than-expected intensity for this feature. In addition, bands associated with ionization from 21 2 m 4 are all expected in positions that overlap with other bands. However, the greater complexity of the m = 1 ZEKE spectrum compared to that of m = 0 is consistent with m-specific coupling. Part of the difficulty in reaching definitive conclusions from the ZEKE spectra is that the significant change in the magnitude of the torsional barrier, coupled with the change in phase upon ionization, leads to activity arising from a number of m levels in the ZEKE spectrum for a specific m intermediate level. 8 For a limited amount of vibrational activity, this is actually a good assignment tool, as distinct patterns of bands can be identified for each FC active vibration. However, when interactions in S 1 have occurred, particularly those involving vibtor levels, the resulting increase in the number of bands leads to difficulties in reaching a definitive assignment. In both ZEKE spectra in Fig. 7(b), we have given suggested assignments to most of the intense bands at lower wavenumbers, and a number of these also appear in combination with + 18 1 at higher wavenumbers (not indicated). Also present in those spectra are bands that are assigned to combinations, and these appear to arise from FC-like activity. The relevant region of 2D-LIF and three ZEKE spectra recorded across this region are shown in Fig. 8. Okuyama et al. 11 assigned the main feature to 15 1 (denoted 12 1 in their work), and we concur with this assignment of the main excitation bands to 15 1 m 0,1 .
The integrated 2D-LIF spectrum is shown at the top of Fig. 8(b), which closely resembles this region of the REMPI spectrum. The 2D-LIF spectrum consists of a number of well-defined bands, falling into two main columns of activity, centered at excitation wavenumbers of 960 cm −1 and 964 cm −1 ; a weaker column of activity is seen at excitation wavenumbers close to 970 cm −1 . Above about 1000 cm −1 in emission, there is a less well-defined structure extending across the spectrum, suggesting that this emission originates from coupled levels, which will be discussed below.
The band intensities in the 2D-LIF spectrum are not as expected, with the 17 1 mx emission bands being significantly more intense than the 15 1 mx emission bands; the assignment is clear, however, since the 17 1 bands were straightforwardly assigned above in Sec. III E, where we also commented that the 15 1 emission bands were unexpectedly intense when exciting 17 1  The region between 0 0 + 967 and 0 0 + 975 cm −1 has been enhanced by a factor of six, as indicated, as this structure is weak compared to the main activity. Selected assignments are shown-see the text for further discussion.
and Fig. 8(a)(ii)], which show the expected strong vibtor bands associated with + 15 1 . We thus conclude that there are non-diagonal FCFs associated with emission from 15 1 , which must be related to geometry changes, since we do not see evidence of a Duschinsky rotation between these vibrations in the Duschinsky rotation matrix [see Fig. 7(a)], and indeed, the ZEKE spectra [ Fig. 8(a)] do not exhibit + 17 1 m x bands; neither do we see + 15 1 m x bands when exciting via 17 1 m 0,1 [see Fig. 7(b)]. The 2D-LIF spectrum [ Fig. 8(b)] also shows significant torsional bands, together with vibtor bands associated with the main emissions. These are largely as expected, and their assignments are straightforwardly obtained both by the 2D-LIF spectrum obtained via m 0 and m 1 (Fig. 2) and by comparison with the work of Stewart et al., 10 as well as the wavenumbers of other vibrations, obtained in the present work (see Table III). When exciting at 970 cm −1 , the strongest 2D-LIF band is at (970, 1402) cm −1 . The assignment of this band to the Δ(v, m) = 0 band (18 1 28 2 m 0 , 18 1 28 2 m 0 ) is relatively straightforward, fitting the expected wavenumbers in both the S 0 and S 1 states, and also being consistent with the 18 1 28 2 band seen for mDFB (Fig. 1). Furthermore, vibtor bands associated with + 18 1 28 2 m 0 are seen in the ZEKE spectrum recorded when exciting via the intermediate band maximum 0 0 + 971 cm −1 , although it is noted that + 18 1 28 2 m 3(+) is not the most intense band in the spectrum, as would be expected. We note that the strongest emissions seen when exciting across 957 cm −1 -966 cm −1 all extend to higher excitation wavenumbers, consistent with either coincidental FC activity or an interaction between one or both 15 1 m 0,1 levels and a level at 970 cm −1 . Since the profile of the 18 1 28 2 emission band is strongest for the m 0 component, we suggest there is a 15 1 . . .18 1 28 2 interaction for both m components, but for the m = 1 levels, further interactions cause a dissipation of the emission intensity across numerous transitions. In contrast, the interaction with 18 1 28 2 m 0 is weaker and less profligate, and so the emission band is more pronounced. At this excitation wavenumber, we can also anticipate possible activity from other levels, including 17 1 21 1 m 0,1 and the vibtor levels 18 1 19 1 m 3(+) and 18 1 20 1 m 3(+) . Relatively weak, but clearly observable bands at the correct wavenumbers for the 17 1 21 1 m 0,1 emissions can be seen in the 2D-LIF spectra; moreover, bands arising from + 17 1 21 1 m x vibtor levels can also be seen in the ZEKE spectrum in Fig. 8(a)(iii). In addition, + 18 2 m x ZEKE bands are seen, but these are thought to arise from FC activity, since these are also seen in other spectra when exciting fundamentals. In summary, it seems clear that interactions are occurring, and the evidence is that this predominantly involves the m = 1 components and involves widespread ARTICLE scitation.org/journal/adv coupling; the coupling with the m = 0 component is less definitive and is at best restricted in nature. The main activity comes from 15 1 m 0,1 , but there is clear activity from 18 1 28 2 m 0,1 and persuasive evidence for involvement of 17 1 21 1 m 0,1 ; however, whether these levels are interacting significantly or not is less clear, but if they are, then the stronger interaction might be expected to be between 15 1 and 17 1 21 1 , which is Δv = 3, while the other interaction would be Δv = 4. The ZEKE spectrum recorded via 0 0 + 971 cm −1 [ Fig. 8(a)(iii)] is rich in structure, and its assignment is challenging. We highlight that there are ZEKE bands at the correct position for 17 1 21 1 m 0 activity, for example, the intense + 17 1 21 1 m 3(+) band, but the corresponding activity expected for 17 1 21 1 m 1 is not seen, in line with comments in the previous paragraph. We note a strong series of bands at 1026 cm −1 , 1212 cm −1 , and 1338 cm −1 that appear to be the m = 0, 3(+), and 6(+) components associated with + 18 2 , which are indicated in the figure. Although possible assignments could be put forward for other bands in this spectrum, we generally refrain from doing so, since these are not definitive. For example, in cases where an interaction can be suggested, such as 19 1  Looking at the ZEKE spectra via 15 1 m x [ Fig. 8(a)(i) and Fig. 8(a)(ii)], there are a series of bands labeled "Q 1 m x " and we show the Δv = 0 band as well as the corresponding vibtor structure. Despite these bands being well-resolved and prominent in both ZEKE spectra, there is no evidence for corresponding activity in the 2D-LIF spectrum. Although it is difficult to determine the identity of "Q," it may be associated with a level that is in Fermi resonance with + 15 1 in the cation (and so each corresponding pair of vibtor levels is also interacting); one promising candidate is Q = + 25 1 29 1 . This could also simply arise from FC-like activity, of course, and we noted the appearance of + 25 1 29 1 m x bands when exciting via 17 1 m 1 [ Fig. 7(b)].
In summary, at the very least, the 2D-LIF and ZEKE spectra suggest that there are likely numerous interactions occurring with the 15 1 level, supported by the appearance of many bands alongside those of + 15 1 m x in the ZEKE spectrum recorded at 0 0 + 971 cm −1 . We also see clear evidence for 18 1 Tables I and III (for mDFB, the level of theory used for the cation was the same as that used for mFT). The motions are distinctive, and hence, assigning each vibration from the mode diagram is straightforward.

ARTICLE scitation.org/journal/adv
1200 cm −1 , and the integrated spectrum at the top of the spectrum closely resembles this region of the REMPI spectrum. The main emission band for mFT, when exciting across 1254 cm −1 -1262 cm −1 , is at 1271 cm −1 . Comparing this value with the liquid-phase IR/Raman values suggests an assignment of the emission to 11 1 , and this would be in line with the calculated wavenumbers. Note that Okuyama et al. 11 assigned a value of 1267 cm −1 to an S 0 vibration, which they labeled ν 14 in Varsányi 29 notation and would correspond to mDFB mode ν 25 in Mulliken notation 33 and, hence, D 11 here; however, it was shown in previous work that Varsányi modes ν 3 and ν 14 have got confused over the years and, further, that these labels do not describe the motions of the atoms in disubstituted benzene molecules. 24,25 With these caveats, the present assignment and that of Ref. 11 are in agreement. In Fig. 10, we show the calculated motions of D 10 and D 11 for mDFB and mFT in each of the three electronic states, showing that their motions are distinctive and, hence, assignment of each from the calculations is unambiguous.
A comparison of the REMPI spectrum of mFT with that of mDFB also suggests that the excitation at 1267 cm −1 should be assigned as 11 1 ; however, this was assigned as 10 1 (denoted 6 1 ) in the fluorescence study. 33 In mDFB, under C 2v point group symmetry, the D 10 vibration is of a 1 symmetry, while that of D 11 is b 2 ; thus, the 10 1 transition would be symmetry allowed. However, other transitions involving b 2 symmetry vibrations were assigned in Ref. 33, with 21 1 , 19 1 , and 20 1 transitions being notable; these are likely to be vibronically induced. These are also all seen in mFT 15 (where they all become symmetry allowed), but also the 18 1 transition is moderately intense in mFT but is absent in the mDFB spectrum (see Fig. 1 and Ref. 33), even though D 18 is totally symmetric in both molecules. Hence, there is no prima facie reason not to assign the mDFB transition at 1267 cm −1 to 11 1 , which would bring consistency with the mFT assignment. We highlight that Table I shows that the calculated values for D 10 and D 11 in the S 0 state of mDFB are too close to be discriminant (but their motions and so identities are clear-see Fig. 10), and with either assignment, there is a 40 cm −1 difference between the calculated value and the experimental value. Additionally, we are particularly cautious regarding the calculated S 1 values, which we have found to be often less reliable than those for the S 0 and D 0 + states. 23,44 Further evidence is gleaned from related symmetrically substituted molecules: in Ref. 27, vibrational wavenumbers are presented for five such molecules. Excluding mDFB, the wavenumber for the D 10 vibration lies below that of D 11 for both the experimental and calculated values for all of the other molecules. For mDFB, as noted, the calculated values are only a few cm −1 apart, but the experimental values, as assigned, are clearly reversed compared to those of the other molecules. Given the variation in these wavenumbers with mass-and given that the corresponding values for m-xylene and resorcinol are consistent with each other, but the reverse of the previously assigned values for mDFB-we suggest that the D 10 and D 11 assignments need to be reversed as well, and this has been done in Table I. This is then consistent with the mFT results obtained herein. Consequently, as with 18 1 , it appears that 10 1 simply is not active in mDFB, despite being totally symmetric, while we conclude that 11 1 must be vibronically active. Furthermore, we note that Graham and Kable 33 have commented that previous assignments of the b 2 symmetry vibrations of mDFB are questionable, noting that the assignment of 11 1 in S 1 to a value of 1608 cm −1 does not seem to be correct, and indeed, this would not agree well with the calculated value in Table I. In summary, the most intense 2D-LIF band for mFT in Fig. 9(c) at (1260, 1271) cm −1 is assigned as (11 1 m 0 , 11 1 m 0 ) and is significantly more intense than the corresponding m = 1 band. The ZEKE spectra [ Fig. 9(b)] are consistent with this assignment, with the main bands being assigned as the expected vibtor levels via the two m components. Several other fundamentals are also seen, and, where the sensitivity allows, the expected associated vibtor structure is seen.
With the assignment of the excitation to 11 1 , the observed structure in the ZEKE spectrum allows a vibrational wavenumber of 1275 cm −1 to be obtained for + 11 1 . (We note that, unhelpfully, the experimental value for this vibration falls between the calculated values for + 11 1 and + 10 1 , and so this cannot be used as further evidence for this assignment, which is largely based on the 2D-LIF spectrum-see Table III.) The ZEKE spectra have a significant underlying unstructured signal, which is akin to the broad background in the 2D-LIF spectrum and again is consistent with significant IVR occurring. Another progression of vibtor levels is also seen in both ZEKE spectra, consistent with a vibration with the wavenumber 1330 cm −1 , which can be plausibly assigned to + 16 1 20 1 , which could be arising from a Δv = 3 interaction; the activity in + 10 1 likely arises from FC activity, since the Duschinsky matrices, Fig. 9(a), show that D 10 and D 11 are not significantly mixed upon ionization. If the + 16 1 20 1 m x assignments are correct and an interaction is indeed occurring, then this suggests a value for D 16 in S 1 of ∼840 cm −1 , which is in fair agreement with the calculated value. The interaction would be expected for both m components, which is consistent with the ZEKE spectrum. The 16 1 20 1 bands, expected at ∼1370 cm −1 , are in a region of the 2D-LIF spectrum that consists of unstructured emission, effectively ruling out the possibility of definitive identification. This emission, together with the unstructured background in the ZEKE spectra, suggests significant interactions are occurring, but we cannot provide unambiguous assignments for all of the bands nor identify the likely myriad of interacting levels in the spectrum.

IV. FURTHER DISCUSSION
In the above, we have looked at the assignment of a selection of bands across the lowest ∼1350 cm −1 of the S 1 state of mFT. Clearly, IVR cannot occur for the origin and the very lowest levels, but as discussed in Ref. 10, even below 350 cm −1 , there are interactions occurring between vibrations, torsions, and vibtor levels. Here, we have extended the examination of levels, where we see limited interactions are present for levels below 950 cm −1 , but significant IVR occurs above this, moving toward the statistical (dissipative) IVR regime; the latter is demonstrated by the presence of a largely unstructured underlying background in both the 2D-LIF and ZEKE spectra recorded at ∼960 cm −1 and ∼1260 cm −1 . On top of this underlying background, there are numerous well-resolved bands, showing that in the present experiments, some energy remains localized to particular vibrations, while some is dissipated through a range of motions.
We noted above that Timbers et al. 7 have compared the behavior of mFT and pFT, concluding that at about 1200 cm −1 , the ARTICLE scitation.org/journal/adv rate of IVR was an order of magnitude faster for mFT than for pFT, based upon quenching experiments. We have studied pFT in a range of internal energies, and we have found that below 1000 cm −1 , limited IVR occurs involving both anharmonic and vibration-torsional coupling. [16][17][18]20,21 At ∼1015 cm −1 , we found that coupling occurred involving two largely separate overtone levels, providing two routes for energy delocalization in pFT, 23 while in the region 1190 cm −1 -1240 cm −1 , there was more widespread IVR, but two levels less than 40 cm −1 apart behaved significantly differently. 17,22 In both of the latter, there were still structured bands on top of a broad background, suggesting at least some energy remains localized. In the case of pXyl, however, at these energies, most structure was lost in the ZEKE spectra recorded, suggesting almost complete delocalization of energy; these observations were discussed in terms of symmetry and the density of states (DOS). 17 It was concluded that although the DOS buildup is critical in providing pathways to widespread IVR, this is determined largely by the presence of one or more methyl groups, rather than the symmetry per se. On top of this, the DOS buildup is not smooth, and so at lower internal energies, serendipity can play a large role in determining whether a particular vibration is located in a "clump" of levels; even then, there needs to be a means of efficient coupling to these. Such coupling will clearly depend on symmetry, but also on the motions involved; such considerations lead to the "Tier Model" of IVR, whereby coupling between particular levels is efficient and facilitates pathways to coupling with a wide range of "bath states." 1 Here, we make further comparison between mFT and pFT. In Fig. 11, we show their DOS plots for totally symmetric vibrations, and also when including torsions and vibtor levels. It is clear that for mFT, there are more totally symmetric vibrations, as all in-plane vibrations are totally symmetric, while for pFT, these split into a 1 ′ (a 1 in C 2v ) and a 1 ′′ (b 2 in C 2v ) subgroups-this is evident in Fig. 11.   FIG. 11. Density of states (DOS) plots for mFT and pFT, using calculated vibrational wavenumbers from Table III and Ref. 23. In plots (a) and (b), only vibrational levels are included, while in plots (c) and (d), vibrations and vibtor and torsional levels are included. In both plots, we indicate levels that are accessible from the m = 0 and m = 1 levels, which are those that are the most populated in the free-jet expansion (a 1 + e for mFT and a 1 ′ + e ′′ for pFT), together with all levels. Note that we do not consider rotational levels in this work.

ARTICLE scitation.org/journal/adv
Notably, the buildup in the DOS is more erratic for pFT than for mFT. This difference is clearer once vibtor levels are included, where again an approximate doubling of the available energy levels is seen for mFT compared to pFT, for levels accessible from m = 0 and m = 1. Furthermore, it can be seen that the buildup of levels is generally more continuous once vibtor levels are included compared to only the vibrations, which is somewhat erratic, particularly for the totally symmetric levels in pFT.
With regard to previous IVR experiments on pFT, there has been some uncertainty regarding the vibrations excited, which has been discussed. 45 For pFT, there are two main fundamentals at 1196 cm −1 and 1232 cm −1 , but there are other levels nearby, as examined in depth in our recent work. 22 In Ref. 7, it seems the latter level is excited, which is the 5 1 vibration, mainly corresponding to an in-phase stretch of the C-F and C-CH 3 bonds, with the former motion dominating. 25 As we have discussed in the present work, for mFT, the 1260 cm −1 transition is assigned as the vibronically induced 11 1 , which is largely a ring-based distortion. As such, the vibrational motion is quite different for the two molecules, making the comparison less straightforward. Indeed, the motion of D 11 (see Fig. 10) in mFT will involve the adjacent C-H bonds interacting with the methyl group more strongly than for the 5 1 vibration in pFT, which would be one explanation for the increased IVR as a result of vibration-torsional coupling.
In the experiments by Parmenter's group, 7 reliance is placed upon collisional quenching with O 2 . The idea is that the excited electronic state is vibrationally excited following laser excitation and that there is time dependence for the IVR to occur. In addition, the higher the partial pressure of O 2 , the more rapid the quenching, and the less the time molecule had to undergo IVR. However, this can only occur with levels that are excited coherently within the width of the laser pulse, which will be a few cm −1 for a nanosecond pulse (not stated in Ref. 7, but the laser system mentioned suggests this was the case). In Ref. 7, a fit is made to the data to determine k IVR , with electronic and vibrational collision quenching, k V , included. Various assumptions were made in determining k IVR , with the end conclusion being that this was roughly an order of magnitude larger than that for pFT. A discussion of the possible rationalization of this observation was then made, including the DOS of the coupled vibrational levels, the effect of the methyl rotor not being on a principal axis, and the magnitude of the torsional barrier.
With regard to the DOS, we note that there are two factors that increase this in mFT relative to pFT, both related to the reduction in molecular (point group) symmetry from G 12 (C 2v ) for pFT to G 6 (Cs) for mFT. For the vibrations, using molecular symmetry group labels, mFT will have a greater number of totally symmetric vibrations as both the a 1 ′ and a 1 ′′ symmetry vibrations in pFT have the same symmetry (a 1 ) in mFT; Fig. 11 indicates that the number of a 1 vibrations in mFT is comparable to the number of a 1 ′ + a 1 ′′ symmetry vibrations in pFT but with a smoother buildup for mFT. Furthermore, considering molecular symmetry, the number of e torsional levels in mFT is about the same as the number of e ′ + e ′′ torsional levels in pFT. Taken together, see Fig. 11, it can be seen that the total number of vibrational + vibtor levels is about the same in pFT and mFT. However, considering only the states that have the same symmetries as the m = 0 and m = 1 levels (the ones with the dominant populations in the free-jet expansion), then there are about twice as many levels for mFT as for pFT.
The aforementioned DOS does not include rotational levels, and it was argued in Ref. 7 that coupling with rotational levels would be more significant in mFT than pFT; if so, then it may be that there is an effect from the use of room temperature and high pressure conditions in Ref. 7, where rotational effects would be expected to be more significant than they would be in jet-cooled, gas-phase studies. 45 It was also commented in that work 7 that at the internal energies employed, pFT will couple to 10-50 levels, while mFT will couple to essentially an infinite number. The DOS plots in Fig. 11 do not support this latter comment and, further, the 2D-LIF spectra do not either, where the structure is seen, albeit on a background, for both pFT (Ref. 19) and mFT (present work); this is in contrast to the fluorescence spectrum for mFT reported in Ref. 7, where no structure is evident when the quencher is absent.
We concur with the comments in Ref. 7 that the significantly higher barrier in mFT is expected to produce larger torsion-vibration coupling terms. One rationale for the higher barrier in mFT compared to pFT is in terms of hyperconjugation: in pFT, hyperconjugation is not expected to be a large effect and weaker van der Waals type interactions are thus expected to dominate, explaining the lower torsional barrier.
As noted above, there has been some ambiguity in the levels employed for IVR studies on pFT, 45 which is pertinent as the vibrational motion is expected to be critical in the observed coupling. For pFT, at 1200 cm −1 , a rather different time-dependent behavior has been observed for the two main levels, 46,47 which was discussed in terms of a rotation-dependent vibration-torsion interaction that occurred specifically for one of the m levels. Rotational dephasing was concluded to be a time-dependent effect only and is not seen in frequency-resolved experiments. 19 In summary, direct comparison between different isomers of substituted benzene molecules is difficult because of the different forms of the vibrations, even if the observed activity seems quite similar. Furthermore, the conditions used in an experiment are expected have a strong bearing on the results, 45 and so caution is strongly advised when making general deductions from a single experiment. Having the ability to resolve vibrational, vibtor, and torsional structures in a spectrum does seem to give the ability to identify explicit coupling channels, but only when the coupling is reasonably limited. Once the coupling becomes widespread, however, this advantage is lost in frequency-resolved experiments if the resolution is not sufficient to resolve all features, or if the spectrum becomes too complicated to assign definitively. Time-resolved photoelectron spectroscopy experiments with picosecond pulses can, however, still be useful in picking out zero-order state contributions in such circumstances, as long as the frequency resolution can be maintained at tens of cm −1 . 46,47

V. CONCLUSIONS
We have presented 2D-LIF and ZEKE spectra obtained when exciting through selected levels up to 1350 cm −1 in the lower wavenumber region of the S 1 ← S 0 excitation in mFT. We have assigned the majority of the main features observed, but there are many weak features and also broad unstructured backgrounds in some cases. The assigned features confirm that there is widespread vibtor coupling occurring in this molecule, as well as some anharmonic vibrational coupling; these become more prevalent to ARTICLE scitation.org/journal/adv higher internal energies and are more common for the e symmetry torsional levels than for the a 1 symmetry levels. Explicit couplings can be identified in some cases, while, in others, only potential couplings have been suggested, based upon some of the observed 2D-LIF bands. When there are numerous couplings, the ZEKE spectra become very difficult to assign, owing to the number of bands arising, also causing each to have a lower intensity, and these are located on a rising unstructured background.
Comparing pFT and mFT, we agree with many, but not all, of the ideas expressed in Ref. 7, but we have highlighted that both the number of vibrations and also torsions, and so vibtor, levels are responsible for the stark increase in the DOS; of course, all of these levels will have an associated set of rotational levels, and hence in room temperature studies, this difference will be exacerbated compared to experiments employing a free-jet expansion. We have emphasized that it is difficult to compare these molecules directly, with the particular vibration excited at ∼1250 cm −1 being different for the two molecules. Moreover, relying on Wilson/Varsányi labels to identify a vibration can be misleading, since the motions of the atoms for a particular vibrational wavenumber can be very different for meta and para substituted molecules as discussed in Ref. 27. Last, the buildup in vibrational levels is somewhat erratic at these low wavenumbers, and this suggests that at low energies, notwithstanding the more-rapid buildup in the DOS for mFT, the rapidity with which IVR efficiency can increase is restricted since coupling elements will still depend on Δv being small.
Our conclusion is that it seems clear that mFT undergoes more rapid IVR than pFT, but ascertaining the precise reasons for this, and quantifying them, is far from straightforward.