Faraday Discuss. Chem. SOC., 1982, 73, 137-151 Spectroscopy and Photophysics of Organic Clusters BY DONALD H. LEVY, CHRISTOPHER A. HAYNAM AND DONALD V. BRUMRAUGH * James Franck Institute and Department of Chemistry, University of Chicago, Chicago, Illinois 60637, U.S.A. Received 1 1 th December, 198 1 A supersonic free jet expansion has been used to prepare small clusters of aromatic organic com- pounds, and their structure and energy-transfer properties have been studied by laser-induced fluores- cence. In the first, the two monomer units were side by side in a planar configuration. In the second, the two molecules were in a T-shaped configuration. In the T-shaped geometry the two rings are not equivalent, and spectra resulting from excitation of one or the other ring could be distinguished.A mixed-gas expansion of tetrazine and benzene in helium produced a mixed tetrazine-benzene dimer, a trimer consisting of two tetrazines and one benzene, and bands that were attributed to larger clusters. The benzene-tetrazine dimer had a stacked, parallel-plate geometry, while the T2B trimer had a benzene stacked parallel over a planar tetrazine dimer. Dimers of dimethyl tetrazine were found to be more tightly bound in the excited electronic state. Excitation of the 6 d ring mode in the dimethyl tetrazine dimer led to intra- molecular vibrational relaxation into the cluster modes. Dissociation was observed following excit- ation of the 6aX transition of the tetrazine-benzene dimer. The dimer of s-tetrazine was found to have two observable conformations. In recent years the technique of laser-induced fluorescence spectroscopy has emerged as a powerful tool for the study of Van der Waals molecules.’ While it lacks the extreme resolution of electric or magnetic resonance methods, it is very sensitive and relatively easy to apply.Moreover, analysis of the electronic spectrum is the only way to obtain information about electronically excited states of Van der Waals molecules. In addition to examining the structure of electronically excited states, laser-induced fluorescence has also allowed the study of Van der Waals photochemistry and photophysics. Until the studies described in this paper, all work in our laboratory was directed toward Van der Waals molecules composed of a single chemically bound molecule, the substrate, surrounded by one or more rare-gas atoms or first-row diatomic mole- cules Van der Waals bound to the substrate.The substrate was the chromophore and absorbed in the visible or non-vacuum U.V. where there were easily available excitation lasers. Also the molecule had to fluoresce with some finite quantum yield so as to be detectable by fluorescence techniques. The Van der Waals atoms or molecules had much larger electronic excitation energies, and only their ground electronic states seem to be important in describing the structure and dynamics of these species. We have now begun the study of small clusters of aromatic organic molecules where there is potentially more than one chromophore in the molecule. Prior to this work there were no spectroscopic data on the structure of such clusters.More- over, there were no data on how energy transfer either within a monomer unit or from * Present address: Eastman Kodak Company, Kodak Park, Rochester, New York 14650, U.S.A.138 ORGANIC CLUSTERS one monomer to another was influenced by the weak bonds of the cluster. Therefore this appears to be a fruitful area of study. This paper is a report of laser-induced fluorescence studies of clusters containing s-tetrazine or substituted tetrazines as the chromophore. A preliminary report of work on benzene clusters has been published.' The clusters are formed in a supersonic free jet and are small polymers of a single organic monomer or mixed polymers of more than one organic monomer, e.g. tetrazine and benzene.In favourable cases we are able to resolve rotational structure in the fluorescence excitation spectrum and in this way probe the structure of the cluster. Analysis of the dispersed fluorescence spectrum following excitation to a well-defined excited-state level provides both struc- tural information about the ground electronic state as well as dynamical information about the redistribution of energy between the time of excitation and the time of emission. EXPERIMENTAL The experimental apparatus is similar to that which we have used to study the spectroscopy of rare-gas-substrate Van der Wads molec~les.~ The clusters are prepared in a mixed-gas superonic expansion consisting of a small amount of the molecule of interest seeded into a carrier gas, usually helium.When two monomeric units are to be combined in a cluster, seeded gas mixtures of the individual monomers in helium are prepared, and the two seeded mixtures are then mixed with each other prior to expansion through the nozzle. Mixtures are prepared by passing the carrier gas over a solid or liquid sample of the seed molecule, and the concentration of the seed in the mixed gas is controlled by varying the temperature of the mixing chamber thus varying the vapour pressure of the seed. In this way we can independently control the total pressure and the individual concentrations of the seed molecules. Typical operating conditions would be seed gas in helium at a total pressure of 5 atm expanded through a 100 pm diameter nozzle. It should be understood that syn- thesis of a particular cluster may require expansion conditions which are quite different from these typical values.Fluorescence was excited by an argon-ion laser pumped tunable dye laser. Spectra were taken using three different dye-laser configurations depending on the required resolution and scan length. Long-range survey scans were made using a low-resolution laser containing only a birefringent filter as a tuning element. These spectra had a resolution of ca. 1 cm-I and had resolved vibrational structure but no resolved rotational structure. They could, however, cover a range of several hundred cm-'. High-resolution spectra were taken with a single-mode laser which could be hopped from mode to mode but could not be scanned smoothly between modes. In this case the effective resolution was the mode spacing which was on the order of 0.01 cm-', and the length of a single scan was ca.4 cm-'. In most cases these conditions allowed us to scan a complete vibrational band while resolving rotational structure. Finally, in a few cases, we were able to use a single-mode ring laser which could be scanned between modes. This laser had an effective resolution of ca. 0.003 cm-' but could only scan a 1 cm-' interval and thus required several individual scans to cover a vibrational band. Fluorescence excitation spectra were taken by scanning the dye-laser wavelength and collecting as much fluorescent light as possible at all frequencies. Light was collected by an achromatic camera lens and focused to an image on a slit which could be used to spatially select the fluorescent light.The Doppler width of the spectrum could be reduced by narrow- ing the slit, and the width of the slit was chosen in each case to match the Doppler width to the resolution of the laser. The fluorescence was detected by a cooled photomultiplier behind the slit operating in a photon-counting mode. Dispersed fluorescence spectra were taken by tuning the laser to an absorption frequency of the cluster, keeping the laser frequency fixed, and then dispersing the fluorescent light with a 1 m monochromator with a dispersion of 4 8, mm-'. The fluorescent light was collected and imaged onto the monochromator slit by a camera lens and periscope.D . H . LEVY, C . A . HAYNAM AND D . V . BRUMBAUGH 139 Most components of the experimental apparatus have been described in more detail elsew here.3 RESULTS AND DISCUSSION A .TETRAZINE Before describing the spectra of tetrazine clusters, it is helpful to examine the fluorescence excitation spectrum of the monomer shown in fig. 1. Tetrazine itself is a near symmetric oblate top ( K = 0.74), the near symmetric-top axis being the out- of-plane axis. The visible transition is a n * t n transition that is out-of-plane polarized, and therefore the transition moment is parallel to the symmetric-top axis. The ground and excited electronic states have almost the same geometry and rotational constants, and this produces the typical parallel band structure seen in fig. 1 . The He <,,-> I I I I , L . , I I 1 I " " . " " ' -98 -84 -70 -56 -42 -28 -14 Yo 14 28 42 56 70 frequency/GHz FIG.1 .-High-resolution fluorescence excitation spectrum of tetrazine, He-tetrazine and He2- tetrazine. The bottom trace shows the composite spectrum of the three molecules. band has a strong central Q branch consisting of many overlapped lines and evenly spaced P and R branches going off to the low- and high-frequency sides of the Q branch. The slight asymmetry of the molecule produces a slight splitting of the individual K components of a given J transition, but the dominant feature of the140 ORGANIC CLUSTERS spectrum is the regular spacing between adjacent P- and R-branch lines. To a first approximation this spacing is given by the average rotational constant B = i(l3 + C) given by the moments of inertia about the in-plane axes.Also shown in fig. 1 are the fluorescence excitation spectra of two Van der Waals molecules, He-tetrazine and He,-tetrazine.' The helium atoms sit above and below the plane of the tetrazine ring and their additional mass contributes to the moments of inertia about the in-plane axes and thus reduces B. However, the Van der Waals molecules are still near symmetric tops, the ground and excited electronic state geometries are still similar, and the transition is still parallel polarized. Therefore the overall appearance of the spectrum is the same, only the scale is changed due to the smaller rotational constant. B. TETRAZINE DIMERS In fig. 2 we show the low-resolution fluorescence excitation spectrum of a mixture 1 D I I I I I I I I 18300 18200 18100 18000 v/cm-' FIG.2.-Low-resolution fluorescence excitation spectrum of tetrazine in helium at (a) T, = 0 "C and (6) T, = 24 "C, where T, is the temperature of the tetrazine sample container. Increasing T, produces a higher concentration of tetrazine. Features assigned to the tetrazine dimer are marked D. The upper and middle ( x 30) traces are normalized to produce the same intensity at the strongest feature. The lower ( x 1) trace is identical to the middle trace but is taken at 1/30 the sensitivity to keep all features on scale. of tetrazine in helium. The quantity T, is the temperature of the mixing chamber containing the tetrazine sample, and therefore the lower trace is taken at a higher concentration of tetrazine, all other experimental parameters being held constant. Both spectra have been renormalized to have the same intensity in the off-scale tetrazine monomer origin band at 18 128 cm-l, and we note the three features marked D increase in intensity relative to the monomer origin as the concentration is raised.We assign these three features to the tetrazine dimer. The high-resolution fluorescence excitation spectra of the two dimer bands that are red-shifted with respect to the monomer origin are shown in fig. 3. The overallD . H . LEVY, C. A . HAYNAM A N D D. V . BRUMBAUGH 141 pattern of the rotational structure is characteristic of a perpendicular transition (transition moment perpendicular to the symmetric-top axis) of a near symmetric top with similar ground- and excited-state geometries. The overall pattern of these spectra rule out certain geometries.In the first place, the overall extent of the spectrum would tend to rule out a trimer or higher polymer as being responsible for these bands. Moreover, a stacked structure with two parallel plates on top of each vlcm-' FIG. 3.-High-resolution fluorescence excitation spectra of the perpendicular tetrazine dimer bands. The geometries responsible for the two spectra are shown in the figure. (a) vo = 18 089 cm-', (b) vo = 18 103 cm-'. other would be a near symmetric top with the common out-of-plane axis being the symmetry axis, and this would produce a parallel-type transition. This would be true regardless of the interplate separation, and therefore the dimer producing the spectra in fig. 3 cannot have a stacked parallel plate geometry. The spacing between the large Q-branch peaks in fig.3 is largely determined by the rotational constant A - B. The spacing of the Q branches in the 18 103 cm-l band is surprisingly large for a molecule as big as the dimer, and this greatly restricts the possible geometries that are compatible with the spectrum. The only geometries that reproduce the large Q-branch splittings require an A-inertial axis that lies in the planes of both monomer units passing through or near the centre of both monomers. Moreoever, the extensive structure in the four central Q-branch features involving K = 1 levels is produced by the slight asymmetry and disappears in a true symmetric top. A geometry that has two perpendicular rings (fig. 4) with the A-axis lying in the plane of both rings is too symmetric to produce the observed K = 1 splitting6 The geometry that gives the best fit to the spectrum has the two rings lying in the same plane with a separation of 5.6 8, between their centres.The spectrum is rela- tively insensitive to the rotation of the two rings about their own out-of-plane axes, and we are still trying to determine how well we can measure these geometric para- meters. Note that we see only a single transition arising from the side-by-side planar geometry, which implies that the orientation of the rings about their out-of-plane axes produces an overall geometry in which the two rings are symmetrically equiva-142 ORGANIC CLUSTERS c h -1 VO v/cm - 1 FIG. 4.-Synthesized high-resolution spectrum of the tetrazine dimer in a twisted geometry. This geometry is obtained from the planar geometry shown in the lower trace of fig.3 by rotating one ring 90" about a line passing through the centres of the two rings. lent.* One such geometry that would produce the observed spectrum and that would have two equivalent rings is shown in fig. 5. This structure allows partial hydrogen bonding between the hydrogens on one ring with the nitrogens on the other ring. Such a structure not only reproduces the spectrum and preserves the equivalence of the rings, but it explains the forces that hold the dimer in a planar configuration. The band centred at 18 089 cm-l is also a perpendicular transition but the smaller spacing between the Q branches indicates that it is produced by quite a different geometry.The best fit that we have been able to obtain is produced by a T-shaped geometry with the centres of the two rings separated by 4.37 A. This geometry has been proposed for the ground state of the benzene dime^-,^ and it is the geometry of nearest neighbours in the tetrazine crystal * where the centre of the nearest-neighbour rings is 4.40 A. In a T-shaped configuration the two rings are not symmetrically equivalent, and therefore each electronic transition should produce two bands corresponding to excitation of the two inequivalent rings. A T-shaped dimer of tetrazine is a near prolate symmetric top with the symmetry axis being the line drawn between the centre of the two rings. Since the z*+n transition is polarized out-of-plane, excitation of the ring that is parallel to the symmetry axis (the upright of the T) will produce a perpendicularly polarized rotational structure (AK = & l), whereas excitation of the ring that is perpendicular to the symmetry axis (the crosspiece of the T) will produce a parallel polarized rotational structure (AK = 0).The band at 18 278 cm-' is parallel polarized (fig. 6 ) and its rotational structure leads to the same inter-ring * The helium Van der Waals molecule formed by binding a single helium atom above the plane of one ring makes the two rings unequivalent. In this species we observe two transitions separated by 0.2 cm-'.D . H. LEVY, C . A. HAYNAM AND D. V. BRUMBAUGH 143 I I I I ~ I I I I ~ I I I I ~ I I ~ - 1 UO 1 v/cm-' FIG. 5.-Synthesized high-resolution spectrum of the tetrazine dimer in the planar geometry shown in the lower trace of fig.3. -0.2 -0.1 Z'O 0.1 0.2 0.3 v,'cm-' FIG. 6.-High-resolution fluorescence excitation spectrum of the parallel component of the T-shaped tetrazine dimer. vo == 18 278 cm-', R = 4.369 -+ 0.005. distance as that of the 18 089 cm-I perpendicular band. We therefore assign the 18 278 cm-' and 18 089 cm-' bands as the parallel and perpendicular components arising from a T-shaped geometry. c. MIXED TETRAZINE-BENZENE CLUSTERS In fig. 7 we show the fluorescence excitation spectrum that is produced when a mixture of tetrazine and benzene seeded into helium carrier gas is expanded in a144 M D ( a ) M ORGANIC CLUSTERS M n I D M, /"0 :M . /xlo supersonic free jet. For comparison, the spectrum of tetrazine in helium is reproduced in the lower trace.In the spectral region shown in the figure, three new bands (marked with asterisks) appear upon the addition of benzene, and these are assigned to mixed clusters of tetrazine and benzene. The band at 18 260 cm-l, shown in fig. 8, has rotational structure associated with I I I I 1 - 2 -1 L'o 1 v/cm - FIG. &--(a) High-resolution fluorescence excitation spectrum of the stacked parallel plate mixed dimer of tetrazine and benzene. (b) Also shown is the spectrum of the He-tetrazine-benzene Van der Waals molecule. vo = 18 260 cm-'.D . H . LEVY, C . A . HAYNAM A N D D . V . BRUMBAUGH 145 a parallel-type transition, and we beliebe that it is produced by a dimer containing one molecule of benzene and one of tetrazine.The geometry that best fits this band is a structure with the two rings stacked and parallel to each other. A second blue-shifted band can be seen in fig. 8 which also has a parallel rotational structure. We assign this to a Van der Waals molecule consisting of a helium atom bound to the exposed tetrazine face of the tetrazine-benzene dimer. There are several bands to the red of the 18 260 cm-l band that are due to higher clusters of benzene and tetrazine. Two of these are marked in fig. 7, but there are other weaker bands that are observable at higher sensitivity and under different expansion conditions. We are in the process of analysing their rotational structure when this is possible, but the overall width of the rotational profiles suggests that all of these bands are due to clusters larger than the dimer. For example, the rotational structure of the 18 190 cm-I band shown in fig. 9 is due to a B-type transition and "0 1 2 vlcrn-' 3 FIG.9.-(a) High-resolution fluorescence excitation spectrum of the B-type transition produced by the mixed (tetra~ine)~-benzene trimer. The geometry responsible for this band has the two tetrazine rings planar and side-by-side, as in the lower trace of fig. 3, with the benzene stacked above one of the tetrazinesand parallel to the tetrazine ring. Also shown are the spectra of two helium Van der Waals molecules of this trimer : (b) He-(tetra~ine)~-benzene and (c) He2-(tetrazine)2-benzene. vo = 18 190cm-'. can be reproduced assuming a trimer consisting of two tetrazine molecules side by side in a plane with a benzene monomer stacked above the plane of one of the tetrazine rings and parallel to that ring.Also seen in this figure are two bands assigned to Van der Waals molecules containing one and two helium atoms bound to the (tetrazine),-benzene trimer. D . MILD EXCIMERS-DIMETHYL TETRAZINE DIMERS All of the cluster spectra described above have the common characteristic that in the fluorescence excitation spectrum the bands assigned to the clusters are only slightly shifted from those of the monomer. The monomer-cluster shift is a measure of the difference in binding energy of the cluster in its ground and excited electronic state, a red shift corresponding to a deeper well in the excited electronic state. There- fore the small spectra1 shift implies that the potential surface describing the cluster146 ORGANIC CLUSTERS binding does not change very much upon electronic excitation. The fact that the dominant spectral features are single vibronic bands rather than long, intense vibra- tional progressions supports this interpretation that the potential surface is relatively insensitive to the electronic state. A small spectral shift with no extensive activity in the low-frequency vibrational modes has been the rule for rare-gas-substrate Van der Waals molecules as well as for the molecular clusters described above.The fluorescence excitation spectrum of dimethyl tetrazine (DMT) is shown in fig. 10. Once again, the temperatures that label the spectra are the temperatures of x1 XI0 { l I l , ] , 1 , 1 , 1 , 1 , 1 17500 17000 v/cm - FIG, 10.-Low-resolution fluorescence excitation spectrum of dimethyl tetrazine (DMT) in helium.Temperatures noted to the right of each trace are the temperatures of the sample container holding the DMT. Higher temperatures correspond to higher concentrations of DMT in the gas mixture. Features assigned to the DMT monomer are marked M. Other features are assigned to the DMT dimer. The lower three traces have been normalized to a common intensity for the strongest feature. the DMT reservoir, and therefore higher temperatures correspond to higher con- centrations of the seed gas. The lower three spectra are normalized to a common intensity for the DMT dimer origin band at 17 496 cm-', and features that grow in relation to the monomer as the concentration is raised are assigned to the DMT dimer.The spectrum of the DMT dimer is qualitatively different from that of the tetrazine dimer. The largest wavelength feature of the DMT dimer spectrum is red shifted by 548 cm-l with respect to the monomer origin, and a fairly harmonic progression of five members with we = 79 cm-l is observed. The red shift indicates that the excited electronic state has a significantly larger binding energy than the ground electronic state, and this would be expected to produce vibrational activity in the dimer stretching mode, as is observed. The ability of aromatic hydrocarbons to form excimers in solution is well-known,1° and there is some similarity between the DMT gas-phase dimer spectrum and the spectra of aromatic hydrocarbon excimers.Excimers are more tightly bound in their excited states by several thousand cm-l, and the fluorescence is very strongly red shifted from the absorption. The influence of the excimer state has been invoked to interpret the supersonic-jet spectrum of the benzene dimer,2 although the initialD . H . L E V Y , C. A . HAYNAM AND D. V . RRUMBAUGH 147 absorption in that molecule has been assigned to a less tightly bound T-shaped dimer. The supersonic-jet spectrum of the DMT dimer is similar to that of a hydrocarbon excimer in that it is red shifted, but the shift is different in magnitude from that expected from a hydrocarbon excimer. The change in binding energy of the DMT dimer upon electronic excitation is about an order of magnitude less than that ex- pected for a hydrocarbon excimer, and for this reason we have called these states mild excimers.We do not know of any observation of an 7c"t-n excimer in solution, and the spectral resolution available in solution is probably too low to allow the observation of a mild excimer. The only other mild excimer that we have observed is a mixed cluster of DMT and tetrazine. The fluorescence excitation spectrum of this molecule is shown in fig. 1 I . * 1 17300 17200 17100 vlcrn-' FIG. 11 .-Low-resolution fluorescence excitation spectrum of (a) tetrazine, (b) dimethyl tetrazine and (c) a mixture of tetrazine and dimethyl tetrazine. The spectrum of the DMT-tetrazine cluster has a very long vibrational progression with an intensity that peaks in the middle of the progression and dies off to both the red and the blue.One member of the progression is overlapped by the DMT dimer feature near 17 100 crn-l, and at least two weaker members of the progression can be seen to the red of 17 100 cm-l. Because of the intensity fall-off, we cannot be sure that the longest wavelength member of the progression that we observe is the origin of the transition, but if it is the origin, the excited electronic state is more tightly bound than the ground state by ca. 450 crn-'. E. RELAXATION AND PHOTOCHEMISTRY When two monomer units are formed into a dimer, the six translational and six rotational degrees of freedom of the separated monomers become three translational and three rotational degrees of freedom of the cluster plus six new vibrational modes which we will call cluster modes.Because the cluster binding is so weak, the cluster modes have low frequencies, and therefore even at moderate excitation energies the148 ORGANIC CLUSTERS density of vibrational states can be large. We should expect to see intramolecular relaxation effects produced by the large density of vibrational states, and ultimately the flow of energy into the stretching mode should dissociate the cluster. These effects have been observed in both chemically bound molecules ''J~ and Van der Waals m~lecules,'~ and we have recently started to study such effects in molecular clusters. In fig. 12 we can see the effect of a large density of states on the fluorescence I 0 I I I 4 1 1 2 vlcm-' I FIG.12.-High-resolution fluorescence excitation spectrum of (a) the origin and (b) 6aA bands of the dimethyl-tetrazine dimer. The instrumental resolution was the same for both spectra. (c) The assignment of the (DMT),-He Van der Waals molecule is uncertain. excitation spectrum of the DMT dimer. The lower trace shows the rotational structure of the origin band of the dimer. The spectrum, although complex and at this time unanalysed, is sharp and reasonably well resolved. The upper trace shows the rotational envelope of the 6a; transition taken with the same instrumental resolu- tion. The band is clearly broadened, and we observe only a band contour with no resolved single lines. The V6n mode in the DMT is an in-plane elongation of the ring and has a frequency of 517 cm-' in both the monomer and the dimer.Because of the high density of cluster mode states, even at the relatively small excitation energy of 517 cm-', there are a large number of isoenergetic states that can mix with the 6a' level. These states can be described by a wavefunction that has all of the monomer modes in their zero-point levels (or perhaps has some of the low-frequency monomer modes partially excited) but has varying degrees of excitation in the six cluster modes. The mixing of the 6a' level with the various isoenergetic cluster levels is often referred to as intramolecular vibrational re1axati0n.l~ It should be mentioned that whether or not there is a time-dependent relaxation depends on whether the excitation of the collection of isoenergetic states is coherent or incoherent.The effect of a large density of states may be clearly seen in fig. 13, which shows the dispersed fluorescence spectra of the DMT monomer (lower trace) and the DMT dimer. Both of these spectra are taken by tuning the exciting laser to the 6a; absorp- tion frequency, keeping the excitation frequency fixed, and dispersing the emitted light, The feature marked with an asterisk in both spectra is at the laser frequency and is a combination of resonance fluorescence and scattered laser light, The monomer emission spectrum, while complex, consists of well-resolved singleD . H . LEVY, C. A . HAYNAM A N D D. V . BRUMBAUGH 149 * 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I I I I I 18 000 17500 17000 16500 vlcm-' FIG. 13.-Dispersed fluorescence spectra obtained by exciting (a) the dimethyl tetrazine (DMT) mono- mer 6aJ transition and (6) the DMT dimer 6a; transition.The excitation frequencies are marked *. features. On the other hand, the dimer spectrum has some weak sharp features just to the red of the exciting frequency but is dominated by a few broad features several hundred cm-I wide. When relaxation occurs, a large number of cluster states are populated and the observed emission spectrum is the superposition of the emission spectra from each of these states. If the potential surface did not change at all upon electronic excitation, only Av = 0 emission transitions would be allowed, they would all occur at the same frequency, and the emission spectrum would be narrow. Be- cause there is a significant change in the frequency of the stretching mode and pre- sumably some small change in the other five cluster modes, the emission spectra from the several states are not identical and the composite spectrum is broad.The coarse structure in the dimer spectrum is due to the activity of the V6n mode leading to a 6a: progression in the emission spectrum. The appearance of the weak sharp features could arise in either of two ways. If the initial excitation were coherent, the initial state would be equivalent to the single 6af level and would therefore produce a sharp emission spectrum. In time this coherent superposition would dephase into the collection of relaxed states and would produce a broad emission spectrum. If the dephasing time were short compared with the fluorescence lifetime of a few ns, the emission would be broad; if it were long, the emission would be sharp.If the dephasing time was similar to the fluorescence life- time, the spectrum would have both broad and sharp features and the relative intensi- ties would be a measure of the ratio of the two lifetimes. A second, and probably more likely, explanation of the appearance of both sharp and broad features would presume that the excitation was incoherent. In this case the individual excited-state wavefunctions would be linear combinations of the 6a' state with no excitation in the cluster modes and the 6a0 level with excitation in the cluster modes. Therefore each individual state would have some transition prob- ability for emission to a few de-excited (or weakly excited) ground-state cluster levels (sharp spectrum) and some transition probability for emission to many excited ground- state cluster levels (broad spectrum).The ratio of sharp to broad intensity in the150 ORGANIC CLUSTERS emission spectrum would be a measure of the fraction of the 6a' basis function that appeared in each of the excited eigenstates. If this is the mechanism involved here, the appearance of sharp structure requires that the number of cluster levels not be too large so that after mixing each eigenstate still has some significant fraction of the 6a' state. This second mechanism is now thought to be responsible for the appearance of sharp and broad structure in the emission spectra of jet-cooled alkyl benzenes.12 A final phenomenon which we should consider is the possibility of photochemistry, the transfer of enough energy from the initially excited ring vibration to the stretching cluster mode to allow the cluster bond to break.In the spectra just described, the 6ai excitation of the DMT dimer, this process is energetically forbidden. In fig. 10 we can see a dimer feature at 17 465 cm-', just to the red of the monomer origin. This is the 6ai band of the dimer and has built on it the stretching progression of the dimer. Because the 6ai feature is to the red of the monomer, we know that the excitation energy of the 6a' level above the zero-point level of the excited state is less than the difference in binding energy between the ground and excited electronic states. There- fore the 6a' vibrational energy is certainly less than the excited-state binding energy itself, and even if all of this energy flows into the stretching coordinate it will be in- sufficient to break the bond. We have looked for photochemistry in the DMT dimer by exciting up to 6a3, but all emission seems to come from the bound dimer, not from the monomer fragment that would be produced by a photochemical reaction.This failure to break the bond could be an energetic constraint if 6a3 contained less than the binding energy. Even if 6a3 were above the energy threshold, the presence of the other non-dissociating modes could lengthen the dissociation lifetime so that it was much longer than the fluorescence lifetime. Very recently we have looked for photochemistry in the tetrazine-benzene dimer.In this case v,, is a higher frequency (703 cm-' as opposed to 517 cm-l in DMT), and since it does not form even a mild excimer, the binding energy is probably lower. We find that excitation to 6a' produces only relaxation but that excitation to 6a2 pro- duces photochemistry. We are currently trying to determine if the lack of photo- chemistry at 6a' is due to insufficient energy or whether it is due to an insufficiently rapid rate. This material is based upon work supported by the National Science Foundation under grant CHE-7825555, and by the donors of The Petroleum Research Fund ad- ministered by the American Chemical Society. C. A. H. was supported by the Fannie and John Hertz Foundation. D. H. Levy, Annu. Rev. Phys. Chem., 1980, 31, 197; in Photoselective Chemistry, Adv. Chem., Phys., ed. J. Jortner, R. D. Levine and S. A. Rice (Wiley-Interscience, New York, 1981) vol. 47, part I, pp. 323-362. P. R. R. Langridge-Smith, D. V. Brumbaugh, C. A. Haynam and D. H. Levy, J. Phys. Chem., 1981, 85,3742. W. Sharfin, K. E. Johnson, L. Wharton, and D. H. Levy, J. Chern. Phys., 1979,71, 1292; R. E. Smalley, D. H. Levy and L. Wharton, J. Chem. Phys., 1976, 64, 3266. R. E. Smalley, L. Wharton, D. H. Levy and D. W. Chandler, J. Mol. Spectrosc., 1977, 66, 375. R. E. Smalley, L. Wharton, D. H. Levy and D. W. Chandler, J. Chern. Phys., 1978,68,2487. All synthetic spectra were produced using a spectral simulation program originally written by L. Pierce, Notre Dame University. ' K. C . Janda, J. C. Hemminger, J. S. Winn, S. E. Novick, S. J. Harris and W. Klemperer, J. Chem. Phys., 1975,63,1419; J. M. Steed, T. A. Dixon and W. Klemperer, J. Chem, Phys., 1979 70 4940. D. V. Brumbaugh, C . A. Haynam and D. H. Levy, J. Chem. Phys., 1980,73, 5380. * F. Bertinotti, G . Giacomello and A. M. Liquori, Acta Crystallogr., 1956, 9, 510.D . H . LEVY, C. A . HAYNAM AND D. V . BRUMBAUGH 151 lo J. B. Birks, Rep. Prhgr. Phys., 1975, 38, 903. l1 J. B. Hopkins, D. E. Powers and R. E. Smalley, J . Chem. Phys., 1979,71,3886; 1980,72,2905; 1980, 72, 5039; J. B. Hopkins, D. E. Powers, S. Mukamel and R. E. Smalley, J. Chem. Phys., 1980,72, 5049; J. B. Hopkins, D. E. Powers and R. E. Smalley, J. Chem.Phys., 1980, 73, 683. l2 P. S. H, Fitch, C. A. Haynam and D. H. Levy, J. Chem. Phys., 1981,74, 6612. l3 K. E. Johnson, W, Sharfin and D. H. Levy, J. Chem. Phys., 1981, 74, 163; J. E. Kenny, T. D. Russell and D. H. Levy,’J. Chem. Phys., 1980,73,3607; J. E. Kenny, K. E. Johnson, W. Sharfin and D. H. Levy, J . Chem. Phys., 1980, 72, 1109.