Energy and symmetry ( C 3 ␯ nomenclature ͒ of the normal modes of 

Contexts in source publication

Context 1

... As in Paper 1, an important aspect is the use of a single experimental method at a number of wavelengths across the A -band absorption spectrum ͑ 340–220 nm ͒ . This allows the identification of trends in energy disposal as the dissociation energy is varied and comparison of the measured trends with those predicted in a recent state-of-the-art molecular dynamics studies employing time-dependent quantum mechanical calculations with five active vibrational modes by Hammerich et al. 5 and trajectory calculations on six-dimensional ͑ 6D ͒ and full nine-dimensional ͑ 9D ͒ ab initio potential energy surfaces by Amatatsu et al. 6 These studies can be compared to previous studies that employed a pseudotriatomic model for methyl iodide. 7–9 Most of the present knowledge of methyl iodide photo- dynamics is reviewed in the theoretical studies 5,6,8 mentioned in Sec. I and in a recent review article by Kinsey and co-workers. The main highlights and questions remaining are summarized in the following. Absorption is dominated by a parallel transition to the 3 Q 0 state 3 ͑ Fig. 1 ͒ , which correlates adiabatically with the products CH 3 ϩ I * where I * refers to iodine atoms in the spin–orbit excited 2 P 1/2 state. Perpendicular transitions to the 1 Q 1 and 3 Q 1 states, which correlate with ground-state 2 P 3/2 iodine atoms ͑ I ͒ , are also allowed optically, but are much weaker ͑ Ͻ 2% of the total absorptivity ͒ . The experimental observation of a large yield of I atoms with an angular distribution characteristic of a parallel transition implies that curve crossing takes place ͑ i.e., crossing at the seam of the 3 Q 0 Ϫ 1 Q 1 conical intersection ͒ . Off-axis nuclear motion which presumably arises from the zero-point motion of the degenerate ͑ e ͒ vibrational modes of CH 3 I couples the two surfaces via the kinetic energy operator. 5,11 The balance between the parallel and perpendicular character of this ͑ I ͒ channel is a direct probe of the contributions from the 3 Q 0 and 1 Q 1 states, respectively. Photodissociation on the 3 Q 0 surface is extremely rapid and axial, and the CH 3 I retains C 3 ␯ symmetry during the initial stages of dissociation. 6 The rapid traverse from a bound molecule to separated fragments, reflecting the steep walls of the repulsive electronic states, is confirmed by direct measurement of an appearance time of ϳ 150 fs for the fragments using ultrafast pump–probe laser techniques. 12 For the parallel X → 3 Q 0 transition, the transition dipole points along the C–I bond, thus molecules with bond axes parallel to the E field direction of the linearly polarized light beam are primarily excited and dissociated along the laser polarization direction. The fragment angular distribution is given by I( ␪ ) ϭ ␴ /4 ␲ ͓ 1 ϩ ␤ P 2 (cos ␪ ) ͔ where ␴ is the total cross section, ␪ the angle between E and the fragment velocity vector v ͑ both in the laboratory frame ͒ , and P 2 (cos ␪ ) the second order Legendre polynomial. Anisotropy is characterized by ␤ ( Ϫ 1 р ␤ р 2) which for methyl iodide approaches the maximal value possible ␤ Х 2 for the I * channel. Another sign of the impulsive character of the photodissociation is that most of the excess energy in the system ͑ the photon energy being ϳ 2–4 eV higher than the C–I bond energy ͒ appears as translational recoil energy. Time-of-flight TOF measurements of the translational energy distributions of nascent I and CH 3 fragments began more than 25 years ago. 13 Bond energies can be extracted from TOF data using energy balance arguments; the upper limit of the C–I bond energy is the photon energy minus the total kinetic energy of the fastest moving fragments, i.e., the fragments with the least amount of internal energy, assuming that the internal energy of the parent CH 3 I molecule can be neglected. Analysis using this assumption yielded a bond energy of 2.19 eV, a value which is still quoted in recent literature. 14 This too-low value provided the zero internal energy point for the TOF translational energy distributions, resulting in a large overestimation of the internal energy content. This, of course, posed a rather unfair challenge to theoretical models designed for fitting these distributions. 7 In 1988, a ͑ unpublished ͒ high resolution TOF measurement 15 of A -band dissociation at 248 nm pointed out the importance of internal energy in CH 3 I. A subsequent study 16 at 193 nm of photodissociation via the B band yielded a bond energy of 2.38 Ϯ 0.03 eV, in much better accord with predictions from thermodynamic data. 17 The present study reveals more details on the contribution of vibrationally excited CH 3 I to the fragment energy distributions and our analysis yields a slightly higher bond energy of 2.41 Ϯ 0.02 eV. Trends in vibrational energy disposal can be predicted by comparing the geometric structure and normal vibrational modes of methyl iodide and the methyl radical. The following arguments are adapted from an article by Barry and Gorry. 18 Figure 2 shows the energies and nuclear motions for normal modes in C 3 ␯ symmetry for the two molecules. Three CH I modes, ␯ ( a ), ␯ ( e ), and ␯ ( e ) appear in the CH radical as 1 ( a 1 ), 3 ( e ), and 4 ( e ), respectively, with essentially the same relative motion and energy. The energy of the umbrella mode ␯ 2 ( a 1 ) of CH 3 I ͑ 1254 cm Ϫ 1 ͒ drops by half in the equivalent ␯ 2 ( a 1 ) mode of the CH 3 radical ͑ 606 cm Ϫ 1 ͒ on removal of the I atom. Two CH 3 I modes disappear, ␯ 3 ( a 1 ), the C–I stretch, which couples directly into the dissociation coordinate, and ␯ 6 ( e ), the degenerate methyl rock. Based on this very qualitative picture alone, ͑ ignoring, e.g., the topological features of the potential energy surfaces ͒ one predicts that vibrational energy in the CH 3 I modes, ␯ 1 ( a 1 ), ␯ 4 ( e ), and ␯ 5 ( e ), would appear as vibrational energy in the CH 3 product while energy in ␯ 3 ( a 1 ), the C–I stretch, would appear as increased translational energy. As Barry and Gorry pointed out, ␯ 6 ( e ), the degenerate methyl rock, should be an important coupling mode for promoting curve crossing, and ␯ 6 internal energy could appear in the methyl fragment as transverse recoil in the form of increased out-of-plane rotation or as excess translational energy. Geometric arguments can also be used to predict excitation of vibrational energy in the CH 3 product. As a free molecule, the methyl radical is planar. During breaking of the C–I bond, the pyramidal geometry of the methyl group in the parent molecule must flatten towards planarity, which should result in the excitation of ␯ 2 , the umbrella mode, in the CH 3 fragment. In this study we show that vibrational excitation is indeed dominated by umbrella mode vibration but, especially upon photolysis at the high energy side of the absorption spectrum, there is also significant excitation of ␯ 1 , the CH 3 C–H symmetric stretch. In fact, not only are both ( a 1 ) vibrational modes in CH 3 excited, tentative evidence is also presented for excitation of ␯ 4 , the degenerate ͑ e ͒ methyl rocking motion. Information on the product quantum state distributions for ␯ 2 and ␯ 1 is presented here for dissociation at several wavelengths across the A band. Previous studies using resonance enhanced multiphoton ionization ͑ REMPI ͒ , 19–21 infrared emission, 22,23 and Raman spectroscopy 24 have provided qualitative information on the ␯ 2 vibrational energy distribution following dissociation of methyl iodide, mainly at 266 nm. A slight presence, no more than a few percent, of molecules excited with one quantum of ␯ 1 , the C–H stretch, has been indicated in the REMPI studies. A rather beautiful picture of rotational energy disposal in methyl iodide dissociation has developed over the last de- cade. REMPI and Raman studies have shown that rotation around the CH 3 I C 3 symmetry axis ͑ given by the rotational quantum number K ͒ is conserved in the CH 3 radical, as is the ratio of ortho-para molecules, which is demanded by nuclear spin conservation. CH 3 I rotation perpendicular to the C 3 axis is coupled into a translational energy release. The conservation of K results in a strong ͑ E , v , J ͒ vector correlation, with E the light electric field direction, v the recoil direction, and J the rotational angular momentum of the CH 3 radical, and thus a strong alignment of the methyl radicals. This CH 3 ( J , K )-dependent angular momentum alignment has been quantified in previous REMPI studies. 19,20 During the dissociation process an extra kick of a few quanta into tumbling motion ͑ rotation perpendicular to the C 3 axis, denoted by quantum number N ͒ of the methyl radical occurs, increas- ing the total J value of each K manifold of CH 3 . More detailed information on K conservation has been gained from studies of the photodissociation of fully ͑ J,K,M ͒ state- selected CH 3 I using hexapole fields. 25 Internal energy disposal depends also on which dissociation channel (I * or I ͒ is excited. Methyl radicals produced in conjunction with I atoms are reported to be significantly more excited, both rotationally 26 and vibrationally. 20 Qualitatively, following the above-mentioned normal mode arguments, zero-point motion of ␯ 6 should induce extra curve crossing and enhanced tumbling motion in the resulting CH 3 ͑ ϩ I ͒ products. An abrupt change of the methyl group from pyramidal to planar at the seam of the 3 Q 0 Ϫ 1 Q 1 conical intersection is found from the trajectory studies of Amatatsu et al. ; 6 this change causes a higher umbrella mode excitation in this channel. For the I * channel the trajectory studies have indicated that the relaxation of CH 3 from pyramidal to planar geometry is gradual, resulting in a much lower excitation of the umbrella mode, in agreement with experiment. No significant excitation of the other methyl radical vibrational modes in either the I or I * channel is predicted by theory, including the five-dimensional ͑ 5D ͒ studies of ...

Context 2

... the lower laser powers used at 286.81 nm ͑ϳ 500 ␮ J/pulse ͒ resonance enhanced ionization of selected states is indicated. From the fact that these features cannot be discerned in the I atom images it is clear that the total activity in this mode must be small. These states cannot account for more than ϳ 10% of the total population, considering the good quality of the ␯ 2 fits to the I * channel. It is also interesting that two quanta are observed, suggesting a significant amount of excitation in this mode. More study is needed in this spectral region to quantify these observations. As noted previously, the contribution from vibrationally excited CH 3 I was responsible in the early TOF spectra for the long-standing miscalibration of the kinetic energy release curves and a too-low C–I bond energy. In all of the kinetic energy curves shown in Figs. 5–8 and in the high resolution TOF study at 248 nm, an extra peak for ␯ ϭ 0 in both I and I * channels is seen at higher kinetic energy release. This is indicative of molecules excited in a CH 3 I vibrational mode which couples into the dissociation coordinate, most likely ␯ 3 ͑ C–I stretch ͒ of CH 3 I. At room temperature, the Boltzmann population of the lower energy CH 3 I vibrational modes is ␯ 3 ( a 1 ) ϳ 8% and the doubly degenerate ␯ 6 ( e ) ϳ 5%. Vibrational cooling in a supersonic expansion is known to be much less efficient than rotational cooling, especially for higher energy vibrations. In order to avoid a signal from clusters, the early stages of the molecular beam expansion is probed, which contains more vibrationally excited molecules. The CH 3 I hot-band component seen in the I * channel CH 3 kinetic energy curves at 266 nm is centered ϳ 1200 cm Ϫ 1 higher, with twice the linewidth and ϳ 12% of the area of the cold ␯ ϭ 0 peak. More than one quantum of the ␯ 3 vibration ͑ 528 cm Ϫ 1 ͒ contributes thus to the hot band, and CH 3 from (CH 3 I) 2 clusters could also pos- sibly underly the contributions from vibrationally excited CH 3 I. At shorter dissociation wavelengths it is not possible to separate these effects. At longer dissociation wavelengths the hot-molecule peaks become distinct, however, which allows identification of the vibrational mode and its angular distribution. Two raw images taken at different times in the molecu- lar beam expansion are shown in Fig. 10 for dissociation at 305 nm and detection of CH 3 ( ␯ ϭ 0). Very early in the expansion, in the first 5 ␮ s of the ϳ 200 ␮ s long pulse, internal energy relaxation is poor and the raw image, Fig. 10 ͑ b ͒ , shows extra rings at a higher total energy than E a v l * for the I * dissociation channel. These extra rings have the same angular distribution, ␤ ϳ 1.9, as the ␯ ϭ 0 peak. No signal is seen in the middle of this ‘‘warm’’ image. Around 20 ␮ s later in the expansion the extra ring above ␯ ϭ 0 disappears, Fig. 10 ͑ a ͒ , and a signal grows in the middle of the image which is indicative of clusters. Improved cooling in this case decreases the hot-molecule contribution. Also shown in Figs. 10 ͑ a ͒ and 10 ͑ b ͒ are the corresponding kinetic energy distributions for a vertical slice of each raw image. In the warm beam two peaks spaced at energies corresponding to an extra 530 Ϯ 25 cm Ϫ 1 kinetic energy release are seen above the ␯ ϭ 0 peak. These correspond very well with the spacing of the CH 3 I umbrella mode vibrations. While these CH 3 I ␯ 3 peaks are also seen at shorter dissociation wavelengths for a warm beam, they stand out at the longer wavelengths because of their enhanced Franck–Condon factors compared to CH 3 I in the ( ␯ ϭ 0) state, and because the extra ‘‘kick’’ from the C–I vibration is a relatively larger fraction of the total kinetic energy release. Careful inspection of the raw images also shows a signal at total kinetic energies higher than E a v l , the kinetic energy of zero internal energy fragments formed in the I atom dissociation channel. A portion of a raw image for detection of CH 3 ( ␯ ϭ 0) at 310 nm dissociation is shown in Fig. 11. As seen in this warm image, the signal for the I channel is broader at ␪ ϭ 90° than at ␪ ϭ 0°. Angular integration of the image in 15° wide segments, shown in the middle part of Fig. 11, reveals that the segment at 90° peaks at higher kinetic energies than the segment at 0°. The width of the 90° peak decreases at colder beam conditions, but does not decrease all the way down to the width at 0°, which implies that complete cooling of the beam is not possible. A two- Lorentzian fit to the angular distributions for the I channel ( ␯ ϭ 0) peak suggests that two contributions are present, one from cold CH 3 I yielding ␯ ϭ 0 CH 3 with a beta parameter of ␤ ϭ 1.7, and the second a channel ϳ 900 cm Ϫ 1 higher in energy ͑ separation of the peak maxima ͒ with a slightly negative beta parameter of ␤ ϳ Ϫ 0.2. An extra energy release of ϳ 900 cm Ϫ 1 corresponds very well with the dissociation of CH 3 I excited in ␯ 6 ϭ 1 ͑ 885 cm Ϫ 1 ; see Fig. 2 ͒ , the degenerate CH 3 rocking mode. This peak-to-peak spacing also implies that the ␯ 6 vibrational energy is fully converted into translational energy release. A persistent contribution of ␯ 6 ϭ 1 to the I channel at long wavelengths that peaks at ␪ ϭ 90 ° will slightly affect the results of the previous analysis in Paper 1 of decomposition of the CH 3 I A band. In Paper 1 beta parameters for the I and I * channels were measured from I and I * atom images, which include contributions from the ␯ 2 and ␯ 1 ϭ 1, ␯ 2 manifolds. The hot CH 3 I contributions appear predominantly in the ( ␯ ϭ 0) peak, which for the I channel is a minor ͑ Ͻ 20% ͒ component. Still, the hot CH 3 I lowers the total beta value for the I channel somewhat, which will mainly result in an overestimation of the contribution of the 3 Q state. In Paper 1 the I, I * atom images were taken under ‘‘colder’’ molecular beam conditions compared to those of the raw images shown in Figs. 10 ͑ b ͒ and 11. Information on the rotational energy distributions of the CH 3 fragments can be obtained from analysis of the CH 3 0 0 0 band profile. While a thorough analysis was not carried out, a few trends can be noted. Figure 12 shows portions of the O and P branches of this transition for several different dissociation wavelengths. These spectra were obtained by first successively storing vertical profiles through the image while scanning the detection laser wavelength, thus building up a two-dimensional wavelength–speed distribution image. Inte- gration of these successive profiles at the proper speed yields a signal from the I or I * channel. Only a signal for methyl radicals formed in the I * channel is shown, except for when the dissociation laser was set at 305 nm, where the I channel signal is strong. Overlap of the P (4) 0 0 0 band head by the Q branch of the 1 1 1 transition is indicated schematically by an extra shaded component in the spectra. As shown previously, the 1 1 1 contribution is strongest in the I channel but it also appears in the I * channel at short dissociation wavelengths. Once the 1 1 1 component is discounted, we observe no large difference in rotational excitation between the I and I * channels for dissociation at 305 nm. The I channel appears to be slightly warmer judging by the higher population of P (6). Comparison of the I * channels shows that dissociation at 266 and 274 nm appears to produce the least amount of rotational excitation, again, judging by the small P (6) and large P (2) peaks, while the spectrum at 240 nm appears to be roughly as warm as the 305 nm ͑ I ͒ spectrum. A quantitative analysis requires higher quality spectra of the entire profile. The higher R and S band heads are disturbed, however, by CH ϩ 3 signals from dissociation of CH 3 I ϩ produced by (2 ϩ 1) REMPI of Rydberg states. Previous CH 3 rotational analysis following dissociation at 266 nm shows that a simple Boltzmann distribution is not sufficient to reproduce the rotational population distributions. At the least, separate temperatures for the J and K manifolds are necessary along with a population sum which conserves the CH 3 I ortho-para ratio in CH 3 . More appropriately, a line-by-line analysis is needed to extract each ͑ J,K ͒ state population. For these reasons a quantitative analysis was not attempted. Qualitatively, it can be estimated that the J manifold ‘‘temperature’’ for dissociation at 240 nm is less than 50 K warmer than for dissociation at 266 nm. No striking difference in rotational energy disposal is thus evident as the dissociation energy is increased. A crude means of comparing the total amount of vibrational excitation, and also the amount of rotational excitation for the ␯ ϭ 0 radicals, as a function of the dissociation wavelength is to compare the peak widths ͑ FWHMs ͒ for each kinetic energy curve. Figure 13 displays data for the I atoms in terms of the total kinetic energy release, i.e., corrected for the mass partition factors, the CH 3 ( ␯ ϭ 0) peak widths, trends in the total rotational energy content can be noted. As seen in Fig. 13, the CH 3 ( ␯ ϭ 0) I * peak width at 266 nm dissociation ͑ not corrected for hot-molecule contributions as in Fig. 4 ͒ is slightly less than the corresponding peak widths at 240 and 300 nm. This is in agreement with the rotational energy distributions probed by the REMPI profiles shown in Fig. 12. The CH 3 ( ␯ ϭ 0) I channel at 266 nm shows a higher degree of rotational excitation than the I * channel, in accord with previous studies. Also in agreement with the trends shown in Fig. 12 is the nearly equivalent width of the CH 3 ( ␯ ϭ 0) peaks for I and I * at 300 nm. Peak widths from the ␯ 1 ϭ 0 components of the iodine atom images are also displayed in Fig. 13. A smooth increase in peak width for the I * curve is seen as the dissociation wavelength is decreased. For the I curve the width appears to decrease slightly at 240 compared to 266 nm, but the contribution from the ␯ 1 ...

Context 3

... later ͒ , especially for the I channel. Slightly less umbrella mode excitation is seen at the longer photolysis wavelength of 286 nm, Fig. 7. In this case, the I * atom curve is less broad than at 266.1 nm ͑ϳ 175 compared to ϳ 190 meV FWHM total KE ͒ thus the CH 3 ( ␯ ϭ 0) peak occupies a correspondingly larger fraction of the curve. For the I channel the ␯ 1 ϭ 1 component becomes more distinct than at 266 nm, suggesting also less excitation of the higher ( ␯ 2 Ͼ 3) umbrella modes. For off-resonant detection of CH 3 at 286.81 nm, extra rings appear on the outer side of the image ͓ Fig. 3 ͑ b ͔͒ with ϳ 1400 cm Ϫ 1 energy spacing. These features can be tentatively ascribed to activity of the methyl ␯ 4 vibration, which will be discussed in Sec. IV D. At 305 nm, Fig. 8, the signal strength is much lower but again the umbrella mode excitation in the I channel appears to have decreased even more than at 286 nm since the I atom curve peaks at almost the same kinetic energy as the CH 3 ( ␯ ϭ 0) curve. For the I * channel, the I * atom peak width is slightly less than that at 286 nm but a definite con- clusion is not possible due to the low signal strength. In Figs. 6–8 the height of the ␯ ϭ 0 peak has been scaled to suggest its fractional contribution for both the I and I * channels. These fractions, which can be estimated from the curves shown, are only very qualitative estimates. Evidence for excitation of ( 1 1, 2 0) was noted in previous REMPI studies, but up to now the total ␯ 1 ϭ 1 activity ( ␯ 1 ϭ 1, all ␯ 2 states ͒ has not been quantified. The high resolution TOF study 15 at 248 nm found significant ␯ 1 ϭ 1, ␯ 2 activity and will be compared with the present results here in this section. ␯ 1 ϭ 1 activity is obvious for shorter dissociation wavelengths ͑ Fig. 6 ͒ in both the I and I * channels. At longer wavelengths, the presence of ( ␯ 1 ϭ 1, ␯ 2 ϭ 0) is still seen in CH 3 detection on the P (4)0 0 0 ϩ Q 1 1 1 line, even for photolysis at 305 nm ͑ Fig. 8 ͒ . To estimate the ␯ 1 ϭ 1 activity, the separate I and I * atom curves could be fit with two Lorentzian line shapes. For the weakest signals ͑ at longer dissociation wavelengths ͒ the ␯ 1 ϭ 1 component is very weak, causing a large uncertainty. Figure 9 shows the results of this analysis. Plotted is the fraction of ␯ 1 ϭ 1, ␯ 2 activity for the I and I * channels as a function of the energy available. Data for the I * channel are plotted at an energy which is lowered by the internal energy of the excited iodine atom. Data for the I and I * channels overlap reasonably well when plotted in this manner. Also shown in Fig. 9 are data from the TOF results 15 at 248 nm. Reasonable agreement is found between the velocity mapping and the TOF measurements. As seen in Fig. 9, the ␯ 1 ϭ 1 activity increases with available energy, reaching ͑ϳ 30% ͒ at the highest energy for the I channel. The I * channel is much less excited, which could be the result of the lower amount of available excess energy. The total amount of ␯ 1 ϭ 1 activity, obtained by combining the fractional excitation per channel with the channel quantum yields reported in Paper 1, is ϳ 10% at 240 nm, 7% at 266 nm, and ϳ 4% at 300 nm. The 1 1 1 REMPI curves shown in the lower panels of Figs. 5, 6, and 8 probe only the ␯ 2 ϭ 0 component of the ␯ 1 ϭ 1 curve. Most striking about these curves is their con- tinued presence even at very long dissociation wavelengths where the contribution of the full ␯ 1 ϭ 1, v 2 peak is no longer visible in the I atom curve. This indicates that the ␯ 2 population shifts towards ␯ 1 ϭ 1, ␯ 2 ϭ 0 at lower dissociation energies. Most of the CH 3 detection described in this study was with (2 ϩ 1) REMPI via the 3 p z Rydberg state around 333 nm. (2 ϩ 1) REMPI detection of CH 3 is also possible via the 4 p z Rydberg state around 286 nm. CH 3 I itself absorbs in this region, making a single-laser experiment with higher sensitivity possible. Figure 7 shows iodine atom and CH 3 ( ␯ ϭ 0) kinetic energy curves from velocity images taken in the 286 nm wavelength region. Very similar results are found for detection at the Q branch via either the 3 p z or 4 p z Rydberg states. On scanning through the lower energy region of the Q branch of the 4 p z state ͓ between the various P ( K ) and O ( K ) lines and detecting at higher sensitivity ͔ new features appear in the kinetic energy release corresponding to the I dissociation channel in the CH 3 images. A typical kinetic energy curve obtained from a single-laser image taken at 286.81 nm is shown in the lower part of Fig. 7. As expected, CH 3 ( ␯ ϭ 0) is ionized in this wavelength region which is within the rotational envelope of the 0 0 0 band. Surprisingly, two extra peaks, spaced by 1400 Ϯ 50 cm Ϫ 1 , are seen at lower kinetic energy release ͑ higher internal energy ͒ , below the ␯ ϭ 0 peak in the I dissociation channel, which corresponds well with the energy expected for CH 3 excited in ␯ 4 ϭ 1 and 2, ( ␯ 4 ϭ 1396 cm Ϫ 1 ; see Fig. 2 ͒ . It is unlikely that the first peak corresponds to the umbrella mode, since ␯ 2 ϭ 2 is at 1288 cm Ϫ 1 . The second peak falls around the same energy as ␯ 2 ϭ 5 (2790 cm Ϫ 1 ) but resonance enhanced ionization of the ␯ 2 ϭ 5 state ͑ or of ␯ 2 ϭ 2) at 286.81 nm is not expected. A definite assignment of REMPI lines in this region to 4 1 1 or 4 2 2 is not possible due to the much stronger 0 0 0 lines. It is difficult to estimate the total amount of molecules populating the two 1400 cm Ϫ 1 states seen in Fig. 7. The probability of selectively observing a population of much less than 1% of the total population with a good signal to noise ratio seems unlikely. With laser pulse energies maxi- mized ͑ϳ 8 mJ/pulse, nonresonant ionization 33 of all CH 3 states is observed, producing an image similar to the sum of an I and an I * image. With the lower laser powers used at 286.81 nm ͑ϳ 500 ␮ J/pulse ͒ resonance enhanced ionization of selected states is indicated. From the fact that these features cannot be discerned in the I atom images it is clear that the total activity in this mode must be small. These states cannot account for more than ϳ 10% of the total population, considering the good quality of the ␯ 2 fits to the I * channel. It is also interesting that two quanta are observed, suggesting a significant amount of excitation in this mode. More study is needed in this spectral region to quantify these observations. As noted previously, the contribution from vibrationally excited CH 3 I was responsible in the early TOF spectra for the long-standing miscalibration of the kinetic energy release curves and a too-low C–I bond energy. In all of the kinetic energy curves shown in Figs. 5–8 and in the high resolution TOF study at 248 nm, an extra peak for ␯ ϭ 0 in both I and I * channels is seen at higher kinetic energy release. This is indicative of molecules excited in a CH 3 I vibrational mode which couples into the dissociation coordinate, most likely ␯ 3 ͑ C–I stretch ͒ of CH 3 I. At room temperature, the Boltzmann population of the lower energy CH 3 I vibrational modes is ␯ 3 ( a 1 ) ϳ 8% and the doubly degenerate ␯ 6 ( e ) ϳ 5%. Vibrational cooling in a supersonic expansion is known to be much less efficient than rotational cooling, especially for higher energy vibrations. In order to avoid a signal from clusters, the early stages of the molecular beam expansion is probed, which contains more vibrationally excited molecules. The CH 3 I hot-band component seen in the I * channel CH 3 kinetic energy curves at 266 nm is centered ϳ 1200 cm Ϫ 1 higher, with twice the linewidth and ϳ 12% of the area of the cold ␯ ϭ 0 peak. More than one quantum of the ␯ 3 vibration ͑ 528 cm Ϫ 1 ͒ contributes thus to the hot band, and CH 3 from (CH 3 I) 2 clusters could also pos- sibly underly the contributions from vibrationally excited CH 3 I. At shorter dissociation wavelengths it is not possible to separate these effects. At longer dissociation wavelengths the hot-molecule peaks become distinct, however, which allows identification of the vibrational mode and its angular distribution. Two raw images taken at different times in the molecu- lar beam expansion are shown in Fig. 10 for dissociation at 305 nm and detection of CH 3 ( ␯ ϭ 0). Very early in the expansion, in the first 5 ␮ s of the ϳ 200 ␮ s long pulse, internal energy relaxation is poor and the raw image, Fig. 10 ͑ b ͒ , shows extra rings at a higher total energy than E a v l * for the I * dissociation channel. These extra rings have the same angular distribution, ␤ ϳ 1.9, as the ␯ ϭ 0 peak. No signal is seen in the middle of this ‘‘warm’’ image. Around 20 ␮ s later in the expansion the extra ring above ␯ ϭ 0 disappears, Fig. 10 ͑ a ͒ , and a signal grows in the middle of the image which is indicative of clusters. Improved cooling in this case decreases the hot-molecule contribution. Also shown in Figs. 10 ͑ a ͒ and 10 ͑ b ͒ are the corresponding kinetic energy distributions for a vertical slice of each raw image. In the warm beam two peaks spaced at energies corresponding to an extra 530 Ϯ 25 cm Ϫ 1 kinetic energy release are seen above the ␯ ϭ 0 peak. These correspond very well with the spacing of the CH 3 I umbrella mode vibrations. While these CH 3 I ␯ 3 peaks are also seen at shorter dissociation wavelengths for a warm beam, they stand out at the longer wavelengths because of their enhanced Franck–Condon factors compared to CH 3 I in the ( ␯ ϭ 0) state, and because the extra ‘‘kick’’ from the C–I vibration is a relatively larger fraction of the total kinetic energy release. Careful inspection of the raw images also shows a signal at total kinetic energies higher than E a v l , the kinetic energy of zero internal energy fragments formed in the I atom dissociation channel. A portion of a raw image for detection of CH 3 ( ␯ ϭ 0) at 310 nm dissociation is shown in Fig. 11. As seen in this warm image, the signal for the I channel is broader at ␪ ϭ 90° than at ...