Ion cyclotron resonance ion trap measurements of energy relaxation in

James D. Faulk, and Robert C. Dunbar. J. Phys. Chem. , 1989, 93 (23), pp 7785–7789. DOI: 10.1021/j100360a012. Publication Date: November 1989...
0 downloads 0 Views 671KB Size
J . Phys. Chem. 1989, 93, 7785-7789

7785

Ion Cyclotron Resonance Ion Trap Measurements of Energy Relaxation in Gas-Phase Ions: Three Techniques Compared for Thiophenol Ion James D. Faulk and Robert C. Dunbar* Chemistry Department, Case Western Reserve University, Cleveland, Ohio 441 06 (Received: February 28, 1989)

Photodissociation measurements in the ICR (ion cyclotron resonance) ion trap were used to find the collisionless relaxation rate of thiophenol ions prepared with 2-3 eV of excess internal energy. Three techniques have been described previously that exploit the kinetic differences between one-photon and two-photon dissociation, and these three methods were applied and compared for this case. The two-light-pulse technique gave good data and a clear interpretation and was considered clearly the best method. The intensity-dependence technique gave good data but a less clear-cut interpretation, while the chopped-laser technique was only marginally useful for this slowly relaxing ion. It was concluded that about 0.3 s is required for excited thiophenol ion to relax from 2.5 to 0.8 eV of internal energy.

Introduction The very slow process of vibrational energy loss from isolated gas-phase molecules is well suited to study using trapped-ion techniques. Alone among convenient laboratory techniques these have the possibility of holding isolated molecules under observation for a long enough time for significant energy loss to occur. The rates of such relaxations are of interest in understanding chemical and physical processes in natural molecule-isolating environments such as near-Earth space and interstellar clouds, but their laboratory study has been slight because of lack of useful methods. Our laboratory has developed several approaches to study slow relaxations of isolated ions in the ion cyclotron resonance (ICR) ion Among these are several techniques exploiting the characteristics of sequential two-photon dissociation at appropriate photodissociating wavelength^.^-^ It is the purpose here to apply three methods based on this theme to measurement of a particularly challenging relaxation system, that of thiophenol ion. The radiative relaxation of this ion is exceptionally slow, and it is of interest to see how well the various approaches succeed in this system and also to establish the rate of radiative relaxation of thiophenol ion with more confidence than any one measurement would provide. The sequential two-photon-dissociation sequence first proposed by Freiser and Beauchamp6 in the trapped-ion environment postulates the successive absorption of two photons, as in the scheme A'

W k3

A'#

hv

dissociationproducts

(1)

The energy of the first photon is rapidly converted to vibrational excitation of the ion, which may then exist in its vibrationally hot state A+# for a long time. The relaxation process k3, which may be collisional or 1R radiative, competes with second-photon absorption to provide the ultimate removal of this A+# species. The two-photon nature of the dissociation is reflected in various aspects of the kinetics,' including light-intensity dependence and time evolution of the A+ abundance, and with suitable experiments these kinetic characteristics can be exploited to characterize the k3 relaxation process. In relaxation experiments based on two-photon dissociation an ion is considered as "relaxed" when its internal energy falls below the one-photon threshold (the energy that, when added to the energy of one laser photon, is sufficient to bring about dissociation). In thiophenol ion, the thermochemistry of the dissociation reaction is known with reasonable confidence8 (assuming the product ion *Author to whom correspondence should be addressed.

0022-3654/89/2093-7785$01.50/0

to have the cyclopentadiene structure, which is the most stable known C5H6+ion and has been found in various experiments to be the product of this reaction9). The dissociation energy is 3.2 eV, and for 514.5-nm photons, the one-photon threshold lies at 0.8 eV. Thus the relaxation process under observation in these experiments is that which takes the ions from their initial internal excitation energy to below 0.8 eV. In the two-pulse and chopped-laser experiments, the initial excitation energy is about 2.5 eV (2.41 eV from the photon plus about 0.1 eV of average thermal energy). In the intensity-dependence experiment, the initial energy is that of the electronimpact-produced ions. This is certainly a broad distribution, and is not known, although some speculation based on the photoelectron spectrum was given in ref 5. As a rough basis for thinking, it can be supposed that the ions initially lying between 0.8 and 3.2 eV are uniformly distributed, so that the average initial energy assignment appropriate for interpretation of the intensity-dependence experiment would be approximately 2.0 eV. As has been discussed previously, the relaxation mechanism and resulting relaxation kinetics are expected to be different for collisional relaxation and IR-radiative relaxation.1° For collisional relaxation by charge transfer with parent neutral, it is reasonable to suppose that a single charge-transfer collision will result in a relaxed parent ion, and the appropriate kinetics for such a case, which lead to exponential time decay of internal energy, have been termed rate-process kinetics. For IR-radiative relaxation, on the other hand, relaxation occurs through a cascade of perhaps 20 successive I R photon emissions, and the resulting energy decay, which exhibits a significant induction time followed by rapid disappearance of excited ions, has been designated cascade kinetics. A previous study of the two-photon dissociation and relaxation of thiophenol ions by the intensity-dependence approach was d e ~ c r i b e d .A ~ collisionless relaxation rate of 3 s-' was reported there. The present results are based on a more extensive and satisfactory data set but give nearly the same result. (1) Dunbar, R. C. J . Am. Chem. SOC.1987, 109, 3215. (2) Huang, F. S.; Dunbar, R. C. J . Am. Chem. SOC.,in press. (Branching ratio of energy-dependent competitive fragmentation in dioxane ion.) (3) Asamoto, B.; Dunbar, R. C. J . Phys. Chem. 1987, 91, 2804. (4) Dunbar, R . C.; Chen, J. H.; So, H. Y.; Asamoto, B. J . Phys. Chem. 1987, 91, 2794. (5) Asamoto, B.; Dunbar, R . C. Chem. Phys. Lett. 1987, 139, 225. (6) Freiser, B. S.; Beauchamp, J. L. Chem. Phys. Lett. 1975, 35, 35. (7) Orlowski, T. E.; Freiser, B. S.; Beauchamp, J. L. Chem. Phys. 1976, 18, 439. Dunbar, R . C.; Fu, E. W. J . Phys. Chem. 1977, 81, 1531. van Velzen, P. N. T.; van der Hart, W. J. Chem. Phys. Lett. 1979, 62, 135. (8) Rosenstock, H. M.; Draxl, K.; Steiner, B. W.; Herron, J. T.J . Chem. Phys. ReJ Data, Suppl. 1977, 6(1). (9) See: Asamoto, B.; Dunbar, R. C. Int. J . Mass Spectrom. Ion Phys. 1988.86, 387, Mabud, Md. A.; Cooks, R. G. Org. Mass Spectrom. 1987, 22, and references in these papers. (10) Dunbar, R. C. J . Phys. Chem. 1983, 87, 3105.

0 1989 American Chemical Society

7786 The Journal of Physical Chemistry, Vol. 93, No. 23, I989

n

1

Faulk and Dunbar

< y;:? -. L -e c z r c r.

8 5ms

--;a;-

aser _ . --,.,-se

+T

3e:oc: -

1

=>-Sf

3e:?Ct

?-lse

Figure I . Pulse sequence for two-pulse, two-photon ion trap photodissociation experiment. The cooling time t between ion formation and the first laser Dulse is chosen long enough for thorough ion thermalization, and the time T between pulses is varied.

-

-

25-s

-7-P-

l

Figure 3. Pulse sequence for intensity-dependence determination of dissociation. Dissociation is compared for two light intensities differing by a factor of 2.08.

initially excited ions to an energy below the one-photon threshold is monitored by an intensity-dependencemeasurement at a variable delay T after the electron beam pulse, as shown in Figure 3. The intensity dependence at a given delay T is actually measured by determining the ratio of dissociation produced by light pulses differing in intensity by a factor of 2.

t

I

I

Figure 2. Pulse sequence for chopped-lasertwo-photon dissociation. The pulse widths are varied along with the pulse rate so that the average power level remains constant.

The Three Methods Two-Pulse Method.3 The pulse sequence for this experiment is shown in Figure 1, After electron-impact formation of the ions, they were allowed to thermalize for 2.5 s before the first laser pulse. The second laser pulse followed after a variable delay time T, and ICR detection of the extent of photodissociation was carried out 20 ms after the second laser pulse. As has been discussed in detail,3 information about energy relaxation derives from observing, as a function of the delay T , the amount of photodissociation which occurs in excess of that expected if the two pulses were widely separated in time. This excess arises because of the presence of not-yet-relaxed ions excited by the first pulse, at the time of arrival of the second pulse. The relaxation rate is most conveniently derived from the data by fitting to a computer-generated simulated curve, in which the variables in the simulation are the relaxation mechanism (rate law, cascade, or other), relaxation rate, and absorption rates for the first and second photons. Chopped-Laser M e ~ h o d . ~ It* 'has ~ been shown that instead of two light pulses, a series of repetitive pulses may be used to illuminate the ions, as diagramed in Figure 2, and relaxation rate information extracted from the dependence of dissociation on the pulse repetition rate. This approach has the advantage of delivering more light to the ions in the course of the ion trapping period. This gives a greater extent of dissociation as compared with the two-pulse method with a similar laser and makes the measurements more precise for a given extent of signal accumulation, leading to a faster experiment. Following electron-impact formation, the ions were allowed to thermalize for 2 s. They were then illuminated with the chopped laser for 4 s, and ICR detection of the extent of dissociation followed the end of the illumination period by tens of milliseconds. Because of concern that the charge-transfer regeneration of parent ions might in some way distort the results (although the chopped-laser method is not expected to be perturbed by such a regeneration), some of the experiments included continuous double-resonance ejection of the m / e 66 daughter ions. In this case the ejecting rf of about 0.4 V/cm was applied a t the daughter ion frequency during the entire illumination period. Intensity-Dependence Method.' This approach depends on the fact that the light-intensity dependences of one-photon and two-photon dissociations are different, being respectively linear and approximately quadratic.' The progressive relaxation of

Experimental Section All work described here was carried out on a homebuilt ICR mass spectrometer previously described?" using phase sensitive bridge detection in the 1.4-T electromagnet. Ion formation was by electron impact at a (nominal) electron energy of 12 eV. Following the appropriate thermalization and laser irradiation periods, ions were excited by the rf excite pulse of 100-ps duration and observed during the 4-ms acquisition period. The extent of photodissociation was monitored by the laser-induced decrease in parent ion intensity. Pressure estimation in these experiments was rather uncertain. The ionization gauge used has been found in calibrations for simpler gases at high pressures to follow the literature correctionsi2 quite well, but this probably does not hold very well for thiophenol at the pressures used here. The best indication of the true thiophenol pressure calibration was derived from the ion-molecule reaction C~HSSH + C6H5'

-+

C ~ H S S H '4- C5H6

(3)

which is believed to occur with near-collisional efficiency. Its rate was measured at various gauge pressures in the (1-10) X lo-* Torr range. Assuming a reaction rate constant of 1 X cm3 molecule-' S - I , the observed rates of the reaction indicated that true thiophenol pressures were of the order of 0.4 times the ionization gauge readings. It was the aim of these experiments to observe collisionless relaxation, but in fact the sample pressures used were probably not quite low enough to remove all collisional component of relaxation, since the radiative relaxation is very slow, while collisional relaxation is very efficient. As described below, this was clearly indicated by the deviation of the relaxation curve in the two-pulse experiment from pure cascade kinetics. The best estimates of pressure suggest a collisional quenching rate of 0.6-0.8 s-l for the two-pulse and chopper experiments and 1 s-I for the intensitydependence experiment (assuming highly efficient quenching by parent neutrals, presumably by resonant charge transfer, at a rate cm3 molecule-' s-l). These values, while very apof 1 X proximate, provide the basis for a collisional correction in assigning the radiative relaxation rates. The light source was the 514.5-nm line from the Coherent Radiation CR12 argon ion laser (upgraded to 1-12 configuration). Laser pulses were produced by a mechanical shutter which was accurate and reproducible to within 2.5 ms. In the two-pulse experiment the continuous laser power was about 1 W, with the beam expanded to a diameter of 3.5 cm at the vacuum window. In the chopped-laser experiment the 4-s irradiation period was gated by the mechanical shutter, while the laser was chopped mechanically by a set of toothed chopper wheels with 9% duty D.; Dunbar, R. C. Reu. Sci. Instrum. 1984, 55, 1 1 16. ( 1 2 ) Bartmess, J. E.; Georgiadis, R. Vacuum 1983, 33, 149. ( 1 1 ) Hays, J.

The Journal of Physical Chemistry, Vol. 93, No. 23, 1989 7787

ICR Ion Trap Measurements of Energy Relaxation I

300

l ..a

I

0

.d .-m V

0

... ... . "-

I

d

9

.

I

200-

VI

. n

......

I 0

I 200

400

600

800

1000

1200

20

I 0

I 500

1000

1500

2000

2500

Delay (ms)

D e l a y Time ( m s )

Figure 5. Intensity-dependence curve showing the dissociation ratio for two light intensities as a function of cooling time between ion formation and dissociation. The solid curve is the computer-modeledcurve assuming exponential ion cooling with a rate of 5.0 s-I.

0

200

400

600

800

1000

1200

Delay T i m e (ms)

Figure 4. Results for two-pulse dissociation of thiophenol ions. The upper dashed curve (--) shows the theoretical cascade-relaxation shape, the lower dashed curve (- - -) shows the theoretical rate-process shape, and the solid curve shows the calculated shape for cascade relaxation mixed with a small component of rate-process collisional relaxation, as discussed in the text. (b) The relaxation data and fitted theoretical curve from (a) are shown, along with two additional fitted curves using relaxation rates respectively 10% lower and 10%higher than the best fit values noted in the text.

factor. The chopper wheels were calibrated, and small corrections were applied to the data assuming square-law intensity dependence of photodissociation. The laser was expanded to a diameter of 2 cm. In the intensity-dependence experiment the low-intensity measurement was made by using a filter that attenuated the laser intensity by a factor measured as 2.08. Results Two-Pulse Method. In Figure 4 is displayed a data set for the dissociation of parent ions as a function of delay time between the two pulses. About 23% of the ions were dissociated at zero delay, falling to about 13% at long delays. In order to extract a relaxation rate constant from these data, the points were fitted to a computer simulation of the two-pulse kinetics. It was not possible to fit the data with a decay curve for pure cascade decay kinetics, which would be appropriate to pure IR-radiative relaxation. Such a curve is plotted as the upper dashed curve in Figure 4a and clearly has the wrong shape. However, it was found that the shape of the curve is extremely sensitive to even a small amount of collisional relaxation (assuming rate-process kinetics for collisional relaxation). The solid fitted curve in Figure 4a is calculated assuming a collisional relaxation rate constant (rate-process kinetics) of 0.8 s-' and a radiative rate constant (cascade kinetics) of 3.3 s-I. (The "rate constant" for radiative relaxation is taken to be the inverse of the time required to relax two-thirds of the ions.) The fit to the data appears to be satisfactory using this kinetic model and these rates. For comparison, the lower dashed curve in Figure 4a is calculated from the pure rate-process model, which would be appropriate for

efficient collisional quenching: This model is not expected to describe the relaxation under these low-pressure conditions and indeed does not give a satisfactory fit to the data. As an indication of the uncertainty in the radiative rate constant determination, Figure 4b shows the fitting curves drawn using values of 3.0, 3.3, and 3.6 s-' (with a corresponding proportional variation in the contribution of 0.8 s-I assigned to collisional relaxation). It appears that all of these fits are reasonable, although the preferred value of 3.3 s-l is superior to the others. Intensity-Dependence Method. In Figure 5 is displayed the change in the intensity-dependence ratio as a function of time following the electron beam pulse. It is clear that right after the beam pulse the intensity dependence is near the linear behavior of one-photon dissociation and that it changes over the course of about a second to the quadratic behavior of two-photon kinetics. The solid curve on the figure shows the result of the detailed simulation, using a relaxation rate of 5 s-l to characterize the decay of the electron-beam-generated population below the one-photon threshold and assuming that 50% of the nascent ions from electron-impact ionization lie above the one-photon threshold. The fit appears quite satisfactory. At the pressure of thiophenol used in this experiment the rate of collisional relaxation is expected to be about 1 s-l. Accordingly, the radiative component of relaxation can be assigned as about 4 s-1. These results are in reasonable accord with the previous measurement by the same t e c h n i q ~ ebut , ~ the present more extensive data set warrants more confidence. The present experiment at 12-eV (as opposed to 17-eV) ionizing energy might be expected to give ions of somewhat lower average initial excitation and correspondingly faster relaxation; the higher initial one-photon fraction (50% vs 40%) and faster relaxation rate (4 s-I vs 3 s-]) found in the present experiment appear to bear this out, although the differences are marginally significant. The overall indication is that the ion population produced at 12 eV does not differ in a major way from that produced at 17 eV. Chopper-Frequency Dependence Method. This relaxation process is at the lower limit of feasibility for the chopped-laser technique, but meaningful results were still possible. Most of the chopper-rate dependence is expected to take place between 1 and 2 Hz, so data were collected at 1, 2, 3, and 18 Hz. As shown in Figure 6 , the drop in dissociation between 1 and 2 Hz is indeed large, but there is little further change between 2 and 18 Hz. Plotted on the figure is the simulated curve assuming a relaxation rate of 2.6 s-l, which gives a reasonable fit to the data. Assuming a collisional relaxation component of 0.6 S-I, the radiative relaxation can thus be assigned as 2.0 s-'. Discussion Comparison ofkfethods. It is interesting to compare these three methods with respect to speed and quality of data acquisition,

7788

The Journal of Physical Chemistry, Vol. 93, No. 23, 1989

040/

0254 0

:

1 5

10

15

20

F r e q u e n c y ( H z)

Figure 6. Chopper-rate dependence of dissociation. The solid curve is the computer-modeled fit assuming exponential ion cooling with relaxation rate constant 2.6 s-'.

susceptibility to artifacts, and interpretability of the results. First we will summarize our experience with the characteristics of the three methods, and then we will discuss their relative performance in the specific study of relaxation processes in thiophenol ion. The two-pulse method, using the gated C W laser to generate the pulses, has been found to require extensive data collection to obtain data sets even with modest signal-to-noise ratios; this reflects the difficulty in obtaining sufficient dissociation with two short, weak pulses. Finite pulse-width limitations limit use of this approach to relaxations slower than perhaps 5 s-'. (These drawbacks will be removed to a large extent with the use of a pulsed laser, which is a very attractive future alternative.) It is also susceptible to a significant artifact due to ion-cloud spreading between the two pulses, unless the laser illumination can be made homogeneous. On the positive side, the relaxation is displayed directly as a function of time, starting from a well-defined excited-ion internal energy level, and interpretation is very straightforward. The chopped-laser method, basically a variant of the two-pulse approach, differs in several important respects: Because the ions are exposed to more light, the intensity problem is alleviated and data are better and more easily obtained; it is easily used to observe fast relaxations; and it is not susceptible to cloud-spreading artifacts. The negative aspect is that the information content is lower: the relaxation process as a function of time is not observed directly and interpretation is thus less clear. The intensity-dependence method is comparable to the two-pulse method in its difficulty of data acquisition and significant scatter and its limitation to slow relaxation processes. It is not very susceptible to cloud-spreading problems. In its present implementation, observing the relaxation of ions initially excited by the electron beam, interpretation is made difficult by the lack of knowledge of the initial distribution of ion energies. If a clean initial-excitation method, such as photoexcitation, could be used, this would be a powerful approach since the relaxation is displayed directly versus time. Two features of thiophenol ion, its slow relaxation and its large photodissociation cross section, both lead to particular advantages for the two-pulse method in observing this relaxation. As has been seen, it was possible to acquire a sufficiently good data set to show quite clearly the time dependence of the relaxation, and a convincing interpretation, including some conclusions about the extent of cascade versus rate-process relaxation, could be carried out. For this ion the chopped-laser technique was unable to capitalize on its advantages of high effective laser intensity, freedom from cloud-spreading problems, and ability to resolve fast relaxations. (For other ions, such as benzene ion and some halogenated benzene ions, these assets have been i n ~ a l u a b l e . ' ~ J ~Accordingly, ) the results from this approach, while concordant with the other

Faulk and Dunbar techniques, were not impressive in this study. The intensity-dependence method gave data as good as or better than the two-pulse data and led to a similar value for the relaxation rate, but the inability to make a definitive interpretation without knowing the ion energy distribution detracted from the value of the results from this technique. Relaxation Mechanism. Radiative energy dissipation from an isolated ion by infrared fluorescence must certainly take place by a series of IR photon emissions, which is the mechanism we have termed cascade re1axati0n.l~ Collisional relaxation may also occur through a cascade mechanism, as is likely to be the case for inefficient quench gases. However, collisional quenching by the parent neutral molecule may well occur by resonant charge transfer. In this case the relaxation would probably occur in a single energy-removal step, giving the situation we have termed rate-process relaxation kinetics.I5 This case may hold even if relaxation is slower than the collision rate, if the charge-transfer event only occurs on a fraction of collisions. Of the three experiments considered here, only the two-pulse technique has the ability to distinguish between these possibilities, since it is the only one that displays the relaxation process versus time for a set of ions prepared with a sharply determined initial excitation. Figure 4a shows the different expected decay functions for pure cascade and rate-process relaxation in the two-pulse experiment and shows the substantial difference in the expected curve shapes. The experimental points do not fit either pure case very well, and as described above, the best fit is to a combination of cascade and rate-process kinetics, as would be expected if both radiative and collisional relaxation were occurring together. Clearly the data scatter is too large to make a very confident assignment of the relative contributions of the two processes, beyond saying that both processes make significant contributions. (The collisional contribution is actually limited by the known collision rate at the pressure used, which sets an upper limit on the contribution expected from rate-process kinetics.) Thiophenol Ion Collisionless Relaxation Rate. Three separate results are reported above for the collisionless relaxation time scale of thiophenol ions, starting with an internal energy approximately in the 2-3-eV region, and relaxing to below 0.8 eV. If we assume as we have done in the past4 that the radiative decrease in ion internal energy can be described by an exponential decay, the first-order rate constant for this cooling process is 3.7 s-I from the two-pulse result, 3.7 s-I from the intensity-dependence result, and 2.3 s-I from the chopped-laser result. There is rough agreement among the three results, and excellent agreement on 3.7 s-l from the two more convincing measurements, giving us confidence that this relaxation process has a time constant of about s. This ion has among the slowest radiative relaxation rate constants of the polyatomic ions that we have studied in this internal-energy regime, comparable to iodobenzene ion (2.9 s-l) and bromobenzene ion (3.6 For a number of ions whose radiative relaxation rates have been measured by trapped-ion techniques, it has been interesting to compare the observed rate with that predicted from the I R intensities of the corresponding neutral molecule (available from 1R absorption intensity measurement^).^^'^ The generalization has been made that the ion radiative rates tend to be somewhat higher than these predictions. The I R absorption intensity data for thiophenol needed to make the neutral-molecule calculation are not readily available, but thiophenol is probably not very different from chlorobenzene, whose calculated neutral-molecule radiative relaxation rate constant of 2.0 S-I was typical of such c o m p o ~ n d s The . ~ radiative rate constant of 3.7 s-l measured here for thiophenol ion is thus somewhat, but not greatly, higher than might be expected from similar neutral-molecule properties. Conclusions It is encouraging to find that all three of these methods could be applied to this relaxation problem to give concordant results.

(13) Ahmed, M. S.; So. H. Y . ;Dunbar, R. C. Chem. Phys. Lett. 1988, 151,

128. (14) So, H. Y . ; Dunbar,

R. C. J . Chem. Phys. 1989. 90, 2192.

(15) Honovich. J . P.; Dunbar, R. C.; Lehman,

89. 2514.

T.J . Phys. Chem. 1985,

J . Phys. Chem. 1989, 93, 7789-7793 The two-pulse method gave the most satisfactory combination of data quality and interpretability of the results and, among the methods for drawing information about relaxation behavior from two-photon kinetics, seems to be best suited to studying slow relaxations. The power-dependence method gave a useful confirmation of the results. The chopped-laser technique, while usable in this system, is more suitable for fast relaxationsI3 or for lowcross-section ions where intensity is a major pr0b1em.I~ To the extent that a pulsed laser can be found to give sufficient fluence in single pulses, two-pulse or intensity-dependence methods will probably be superior to the repetitively pulsed laser method in most respects.

7789

The radiative relaxation of this ion is slower than most previously studied benzene derivative ions. A number of such ions have been found to relax significantly faster than expected on the basis of radiative properties of neutral molecules; the rate found for thiophenol ion was also slightly higher than the expectation based on similar neutrals. Acknowledgment. Appreciation is expressed to the National Science Foundation and to the donors of the Petroleum Research Fund, administered by the American Chemical Society, for support of this research. Registry No. C,H5SH+, 76563-61-6.

Exciplex Formation in the Jet-Cooled van der Waals Cluster of Methyl-Substituted Dienes with Fluorene and 9-Ethylfluorene Michiya Itoh* and Akio Hayashi Faculty of Pharmaceutical Sciences, Kanazawa University, Takara-machi. Kanazawa 920, Japan (Received: October 31, 1988; In Final Form: May 26, 1989)

The exciplex formation in the van der Waals (vdW) cluster of methyl-substituted dienes with fluorene (FR)and 9-ethylfluorene (EFR)was investigated in supersonic expansion. The excitation of vdW complex or cluster bands of 2-truns,4-truns-hexadiene and 2,5-dimethyl-2,4-hexadienewith FR and EFR exhibits resonance and exciplex fluorescences, while that of the cluster of 1,Spentadiene affords only resonance fluorescence. The exciplex fluorescence remarkably increases in intensity accompanied by a spectral red shift with increasing diene pressure in supersonic expansion. The exciplex fluorescence reveals double exponential decay with decay times of 3-6 and 30-60 ns. The longer decay time of the exciplex increases with size of the cluster, though the actual cluster size is not obvious at the present stage. It is likely that the short and long decay times of the exciplex are attributable to the structurally unrelaxed and relaxed exciplex in the cluster, respectively.

Introduction Since our first observation of the exciplex fluorescence from the van der Waals (vdW) complex between 1-cyanonaphthalene (1 -CNN) and triethylamine (TEA) in supersonic free jet,’ several studies on the exciplex formation were reported.” The exciplex formation dynamics from the vdW complex as a weak ground state complex generated in the supersonic free jet are distinguished from the collisional dynamics of the exciplex formation in both the liquid and static vapor phase, which is generally accepted as a consequence of collision between the excited-state and ground-state molecules through an encounter complex formatiom5 In the case of jet-cooled molecules, the transformation of the vdW complex after electronic excitation to the exciplex is a major problem. Saigusa et aL6 have demonstrated the time-resolved fluorescence study of the intermolecular vibration within the vdW complex to the exciplex formation. Recently, we have reported the excimer and exciplex formation from the vdW dimer/heterodimer in the jet-cooled fluorene (FR) and 9-ethylfluorene (EFR).7v8 Further, Anner and Haas9 have reported some vdW complex and/or cluster ( I ) (a) Saigusa, H.; Itoh, M. Chem. Phys. Lett. 1984, 106, 391. (b) Saigusa, H.; Itoh, M. J. Chem. Phys. 1984, 81, 5682. (2) (a) Castella, M.;Prochorow, J.; Tramer, A. J . Chem. Phys. 1984,81, 251 I . (b) Castella, M.; Tramer, A.; Piuzzi, F. Chem. Phys. Left. 1986, 129, 105. 112. (3) (a) Anner, 0.; Haas, Y. J. Phys. Chem. 1986, 90, 4298. (b) Chem. Phys. Lett. 1985, 119. 199. (4) Anner, 0.; Zarura, E.; Haas, Y. Chem. Phys. Lerf. 1987, 137, 121. ( 5 ) (a) Klopffer, W. In Organic Molecular Phofopysics; Birks, J. B., Ed.; Wilev-Interscience: New York, 1973; Vol. 1. (b) Mataga, N. In The Exciplek Gorden, M., Ware, W. R., Eds.; Academic Press:-New York, 1975. (c) Ware, W. R.; Watt, E.; Holmes, J. D. J. Am. Chem. SOC.1974, 96, 7853. (d) Itoh, M.; Mimura, T. Chem. Phys. Left. 1974, 24, 551. (6) Saigusa, H.; Itoh, M.; Baba, M.; Hanazaki, I. J . Chem. Phys. 1987, 86, 2588. (7) Saigusa, H.; Itoh, M. J . Phys. Chem. 1985, 89, 5486. (8) Itoh, M.; Morita, Y. J . Phys. Chem. 1988, 92, 5693.

0022-3654/89/2093-7789$01.50/0

exhibiting exciplex fluorescence in the jet-cooled systems such as anthracene/diethyl ether3 and 9-cyanoanthracene/2,5-dimethyl-2,4-he~adiene.~ On the other hand, Kajimoto and his co-workersI0 have reported the stabilization of the intramolecular charge transfer state of jet-cooled molecules by the solvent cluster in supersonic expansion. This paper presents the cluster and exciplex formations between fluorene (FR) and 9-ethylfluorene (EFR) with substituted dienes such as 2-trans,4-trans-haxadiene(HD) and 2,5-dimethyl-2,4hexadiene (DMHD) in supersonic expansion. In the FR/HD and EFR/HD systems, the resonance and exciplex fluorescence were observed upon excitation of vdW complex and/or cluster bands, while no significant exciplex but only resonance fluorescence was observed in the system with 1,Spentadiene (PD). The exciplex fluorescence increased in intensity accompanied by a spectral red shift, when the diene pressure increased in the carrier gas (He). The exciplex fluorescenceexhibits a double exponential decay with short and long decay times. The latter decay time and magnitude remarkably increase with the size of the cluster, though the actual cluster size is not determined at the present stage. The short and long decay times may be attributed to the structurally (orientationally) unrelaxed and relaxed exciplexes, respectively. The formation of the structurally relaxed exciplex leads to stabilization of the charge transfer state of the exciplex in the cluster of dienes. Experimental Section The pulsed supersonic jet apparatus and procedures are similar to those described An excimer laser pumped dye laser (Lambda Physik EMG 53MSC/FL 3002) was used for laser induced fluorescence (LIF) spectra, whose frequency was doubled (9) Anner, 0.;Haas, Y. J. Am. Chem. SOC.1988, 110, 1416, and references therein. (10) Kobayashi, T.; Kajimoto, 0. J. Chem. Phys. 1987, 86, 11 18, and references therein.

0 1989 American Chemical Society