(OCS),, and (OCS) - American Chemical Society

Infrared Photodlssoclatlon of Ar*OCS, (OCS),, and (OCS)3 In Pulsed Molecular Beams: Spectroscopy and Dynamics. Mark A. Hoffbauer,+ Kopln Llu, Clayton ...
1 downloads 0 Views 966KB Size
J. PhyS. Chem. 1983, 87,2096-2102

2096

TABLE V : Best Least-Squares Rotational Constants of He,-Methyl-s-tetrazine Derived from Fitting t h e Observed Torsionless Rotational Lines in t h e Fluorescence Excitation Spectrum

parameter Arrame; em-'

B , cmC,cm

'

K

RH~-MT,'

ground electronic state

excited electronic state

0.1029 0.0658 0.0645 -0.933 3.25 0.03

0.1059 0.0653 0.0631 -0.898 3.24

*

a Distance between t h e helium a t o m and t h e center of the

T h e parameters used in the fit methyl-s-tetrazine ring, were t h e ground-state bond length (3.25 t 0.03 A ) and t h e difference (excited - ground) in bond length (-0.008 t 0.03 A ) .

3.25 f 0.03 8,is similar to the 3.32-A distance observed in He2-s-tetrazine.z6 Conclusion The fluorescence excitation spectrum of methyl-s-tet-

razine has been observed for the first time. The well-resolved rotational structure in the free jet cooled electronic spectrum has been studied, and both torsionless (m = 0) and torsional (m # 0) lines have been identified. Torsionless spectra have been computed and used to obtain rotational constants and information about geometric parameters for methyl-s-tetrazine. Some m f 0 levels are observed to be present regardless of expansion conditions and may limit the application of this technique in more complex systems. The spectrum of the Hez-methyl-stetrazine complex was analyzed and yielded a ring-to-helium distance of 3.25 A.

Acknowledgment. We thank Dan Russell and Young Park for help in synthesizing methyl-s-tetrazine, We also thank Prof. C. J. Seliskar for helpful discussions. This work was supported by the National Science Foundation under Grant CHE-7825555. C.A.H. was supported by the Fannie and John Hertz Foundation. Registry No. MT,67131-36-6; He, 7440-59-7; He2, 12184-98-4.

Infrared Photodlssoclatlon of Ar*OCS, (OCS),, and (OCS)3 In Pulsed Molecular Beams: Spectroscopy and Dynamics Mark A. Hoffbauer,+ Kopln Llu, Clayton F. Glese, and W. Ronald Gentry' Chemical Dynamics Laboratoty, Unlverslty of Minnesota, Minneapolis, Minnesota 55455 (Received December 22, 7982)

Molecular beam experiments have been performed to study the infrared single-photon dissociation of the Ar.OCS, (0CS)2,and (OCS)3van der Waals clusters. C02 lasers were used to excite the first overtone of the OCS v2 bending vibration at frequencies around 1045 cm-'. Measurements include (a) the laser fluence dependence of the photodissociation yield at fixed frequency, (b) the laser frequency dependence of the dissociation yield at constant fluence, and (c) the speed and angle distributions of OCS products from the dissociation of (0CS)2. Single homogeneous absorption peaks were observed for APOCS and (OCS)2,with line widths of 1.0 and 3.7 cm-l, respectively, Corresponding to uncertainty principle lifetimes of 5 and 1.5 ps. The spectrum contains two peaks, with widths of about 6 cm-' (0.8 ps). The absorption intensities per OCS molecule in the clusters are much smaller than that of isolated OCS, reflecting interactions which modify the Fermi resonance which is largely responsible for the 2v2 oscillator strength. The OCS product velocity vector distribution from (OCS)2 photodissociation is isotropic, with only about 1%of the available energy appearing in translation. The data provide further examples in which the range of infrared photodissociation line widths for polyatomic van der Waals clusters is remarkably small. We tentatively conclude that the lifetimes corresponding to these spectral widths are not the vibrational predissociation lifetimes but instead are the lifetimes for relaxation of the initial nonstationary coherent states by anharmonic coupling within the van der Waals complexes.

Introduction The study of van der Waals (vdW) molecule structure and dynamics was one of the last scientific interests of Willis Flygare, and one to which he brought the full measure of originality, enthusiasm, and insight which characterized his scientific career.l It is highly fitting that this commemorative issue contain several papers dealing with this area of research, and we are pleased to be able to make our own contribution. van der Waals molecules provide a uniquely fertile ground for investigating weak intermolecular interactions. In the vdW molecule A-B, where A and B are normal covalently bound singlet molecules, the internal vibrational Present address: Chemistry Division, Naval Research Laboratory, Washington, DC 20375.

0022-3654/83/2087-2096$0 1.50/0

motions of A and B are typically only weakly perturbed by the interactions which bind A and B together in the cluster. However, if A and B are relatively small molecules, then in most cases even a single quantum of excitation in one of the lowest-energy vibrational modes of A or B is more than sufficient to dissociate the complex. An important conceptual question arises as to how one should think about the flow of energy from an internal vibration of one of the constituent covalent molecules into those coordinates which describe motions of the vdW bond, i.e., motions of A with respect to B. After all, the coordinates which represent vibrational and librational motions of A (1)T.J. Balle and W. H. Flygare, Reu. Sci. Instrum., 62, 33 (1981); T.J. Balle, E.J. Campbell, M. R. Keenan, and W. H. Flygare, J. Chem.

Phys., 72,922 (1980).

@ 1983 American Chemical Society

I R Photodissociation of Ar-OCS, (OCS),, and (OCS),

and B with respect to each other in the vdW molecule become translational and rotational coordinates of the separating A and B molecules once the vdW “bond” is broken. If one views the vdW motions as vibrations and librations, then the energy flow should be considered to be intramolecular vibrational-to-vibrational(V-V) energy transfer within the vdW molecule. On the other hand, if one thinks of the vdW motions in the excited molecule as translations and rotations, then the coupling process is more like vibrational-to-rotational-and-translational (VR,T) energy transfer in a low-energy bimolecular collision. Simple theoretical models such as the “energy gap” model of Beswick and Jortner2 and the “momentum gap” model of Ewing3 emphasize the V-T and V-R,T aspects of the dissociation process. Nevertheless, the transition in the character of the vdW motions from that best described as vibrational to that best described as translational must be a gradual one, and the time scale for that transition is important to the interpretation of experimental data on such processes. Experimental studies of the infrared photodissociation of vdW clusters have recently been carried out in several laboratories, including our ~ w n . ~In’ most ~ cases the data are limited to measurements of the dissociation yield as a function of photon frequency, which we will call, for convenience, the “dissociation spectrum”. For molecules which dissociate rapidly and with unit probability upon absorption of a single photon, the dissociation spectrum is identical with the single-photon absorption spectrum for the cluster. In a few cases, the dissociation product speed and angle distributions have also been measured, giving direct information on the final energy d i s p o s i t i ~ n . ~In J~ every case in which the homogeneous line width for the dissociation spectrum has been obtained, the line widths have been very large-typically 1-20 cm-l full-width at half-maximum (fwhm), corresponding to uncertainty principle lifetimes 7 in the range of 0.3 to 5 ps. Although the experimental line width data seem clear enough, the interpretation of these data is perhaps not. The question is whether the lifetime derived from the line width for photodissociation is the lifetime of the state created upon photoabsorption with respect to intramolecular V-V energy transfer and dephasing, or the lifetime of the vdW molecule with respect to dissociation. This question was raised and discussed extensively in our previous report dealing with the infrared photodissociation of isotopic ethylene dimer^.^ It is, of course, closely related to the

The Journal of Physical Chemistty, Vol. 87, No. 12, 1983 2097

0

0.1

I

I

I

I

0.2

0.3

0.4

0.5

Fluence (J/cm2) Flgure 1. Measured laser fluence dependences for photodlssociation of Ar-OCS, (OCS),, and (OCS), at 1045.1 cm-’. Note that the ordinate is logarithmic. The soiM lines are ieast-squares fits to the data.

question which we posed above, since the dissociation lifetime is essentially the time required for the vdW coordinates to become translational and rotational in character, while it is clearly possible, a t least in principle, for the decay of the transition dipole moment due to anharmonic V-V coupling within the vdW complex to be considerably faster than the dissociation rate. At the present time, it appears that the best way to address these questions experimentally is to investigate the infrared photodissociation phenomenon for a wide variety of vdW clusters, in order to learn how the results depend on the vdW bond energy, the structure and complexity of the cluster, and the nature of the excited state which is prepared initially. In this report, we present our resulta on one of the simplest systems yet studied, Ar-OCS, and two related systems, (0CS)2and (OCS)3. Since OCS is a linear triatomic molecule, it has three vibrational modes, the v1 “symmetric” stretch at 859 cm-’, the doubly degenerate v2 bend at 527 cm-l, and the v3 “asymmetric” stretch at 2062 cm-l. In these experiments we used a COP laser to excite the first overtone of the bending mode (2v2) at frequencies between 1025 and 1055 cm-l.

(13)M. P.Casassa, D. S. Bomse, J. L. Beauchamp, and K. C. Janda, J. Chem. Phys., 72,6805 (1980). (14)T. E. Gough, R. E. Miller, and G. Scoles, J. Chem. Phys., 69,1588

Experimental Section The apparatus and experimental techniques used in these studies were similar to those described previously for our experiments on isotopic ethylene dimersa4 The OCS vdW clusters were generated in a pulsed molecular beam by the expansion into vacuum of a 2% OCS-98% Ar mixture at a total source pressure of about 7 atm. The characterization of the source conditions and molecular beam composition was carried out by observing changes in the mass spectrum of the primary beam species as a function of the gas mixture and pressure. Typically, the molecular beam velocity spread Aulv was about 9% fwhm, corresponding to a translational temperature of 2 K. Meerta et al.16 measured the OCS rotational temperature from the expansion under similar conditions of a 5% OCS-95% Ar mixture, and obtained a value of about 3 K, consistent with the expectation that the rotational and translational temperatures should be about the same. Of the various ionic species in the mass spectrum characteristic of the APOCS, (OCSl2, and (OCSI3clusters, the peaks due to the parent ions ArOCS+, (0CS)2+,and (OCS)3+,respectively, were the most intense. Therefore these peaks were used for the attenuation measurements.

(15) M. A. Hoffbauer, W. R. Gentry, and C. F. Giese in ‘Laser Induced Processes in Molecules”,Vol. 6,K. Kompa and S. D. Smith, Ed.,Springer Series in Chemical Physics, Springer, Berlin, 1978.

(16) W. L. Meerta, G. ter Horst, M. J. L. J. Reinartz,and A. Dymanus, J . Chem. Phys., 62,341 (1975).

(2)J. A. Beawick and J. Jortner, Adu. C h m . Phys., 47,363-506 (1981). (3)G. E. Ewing, J. Chem. Phys., 71,3143 (1979). (4) M. A. Hoffbauer, K. Liu, C. F. Giese, and W. R. Gentry, J. Chem. Phys., in press. (5)T. E.Gough and R. E. Miller, Chem. Phys. Lett., 87,280(1982). (6)M. F.Vernon, D. J. Krajnvich, H. S. Kwok, J. M. Lisy, R. Y. Shen, and Y. T. Lee, J. Chem. Phys., 77,47 (1982). (7)T. E. Gough, R. E. Miller, and G. Scoles, J. Phys. Chem., 86,4041 11981). ‘ ( S i J. Geraedte, S. Setiadi, S. Stolte, and J. buss, Chem. Phys. Lett., 78,277 (1981). (9) J. M . Lisx A. Tramer, M. F. Vernon, and Y. T. Lee, J. Chem. Phys., 75,4733 (1981). (10)M. F. Vernon, J. M. Lisy, H. S. Kwok, D. J. Krajnvich, A. ” m e r , Y.R. Shen, and Y. T. Lee, J. Phys. Chem., 86,3327 (1981). (11)M. P. Caaaaaa, D. S. Bomse, and K. C. Janda, J.Phys. Chem., 86, 2623 (1981). (12)M. P.Caeassa, D. S. Bomse, and K. C. Janda, J. Chem. Phys., 74,

.----,-

6044 - - - - (1981).

1197A\. ~ - --,. .

2098

Hoffbauer et ai.

The Journal of Physical Chemistry, Vol. 87, No. 12, 1983

Results Our experiments include measurements of (a) the laser fluence dependence of the photodissociation yield at fixed laser frequency, (b) the frequency dependence of the photodissociation yield at constant laser fluence, Le., the "photodissociation spectrum", (c) "hole-burning" effects in which one laser irradiates the clusters at some fixed frequency while a second laser is tuned over the absorption band, and (d) the speed and angle distributions of OCS produced from the photodissociation of (0CS)2. Photodissociation Spectra. The laser fluence dependence of the photodissociation yield is shown in Figure 1. In this experiment a nearly coaxial laser beam-molecular beam geometry was used in order to achieve uniform laser fluence over the entire molecular beam pulse sampled by the detector. For a simple kinetic scheme in which the ground state is connected by a single-photon absorption to the excited state, which in turn is strongly and irreversibly coupled to the continuum, we have the Beer's law expression -In (f,) = a(v)F where f, is the fraction of dimer remaining undissociated, &) is the frequency-dependent absorption cross section, and F is the laser fluence, measured in number of photons per unit area. For each of these clusters, a linear dependence of In (f,) on fluence is observed at a frequency of 1045.1 cm-', near the 2u2 absorption of OCS. These data were collected by tuning the mass spectrometer successively to each cluster ion peak in the same molecular beam. Since the molecular beam and laser beam conditions were identical for all three clusters, the relative slopes in this plot reflect accurately the relative absorption cross sections of the various clusters at this frequency. The variation of the absorption cross sections with frequency is given in Figure 2 for each of the clusters. For convenience, both the single-laser spectra and the twc-laser spectra are shown in this figure. The solid lines are nonlinear least-squares fits of Lorentzian line shapes to the single-laser spectra obtained with a constant laser fluence of -0.2 J/cm2. The spectrum of Ar-OCS reveals that only two C 0 2 laser lines produce large attenuations of the cluster signal, and the fit yields a peak location of 1046.0 cm-' and a width of 1.0 cm-' fwhm. The (0CSI2spectrum is appreciably broader, with a peak location of 1044.7 cm-' and a width of 3.7 cm-l fwhm. A reasonable fit to the (OCS)3spectrum requires the s u m of two Lorentzians with peak locations of 1044.0 and 1035.8 cm-' and widths of 6.2 and 6.3 cm-' fwhm, respectively. As expected for weakly bound vdW molecules, the peak absorption frequencies are close to that of the infrared-active 2u2 mode absorption of OCS, with small shifts similar to those observed in the liquid and solid infrared spectra. The previously determined frequencies for the 2u2 mode of OCS are 1047.04 cm-' in the gas phase,17 1045.38 cm-' in a liquid argon solution,la and 1042.7 cm-' in the solid phase.lg No significant attenuation of the clusters was observed at frequencies outside of the range shown. The two-laser spectra were recorded in order to examine the homogeneity of the absorption lines. In these experiments, one laser is tuned to a conatant frequency corresponding to one of the points in the absorption spectrum, and its fluence is adjusted to give 10-30% dissociation of ~~

(17)J. S.Wells, F. R. Peterson, and A. G. Maki, Appl. Opt., 18,3567 (1979). (la) A. C.Jeannotte and J. Overend, Spectrochim. Acta, Sect. A , 36, 1213 (1979). (19)F.D.Verderame and E. R. Nixon, J. Chem. Phys., 44,43(1966).

2 [ , ,, 01025

;-,,L,, j 1035

1045

1055

FREQUENCY (cm-')

Figure 2. Photodissociation spectra of APOCS, (OCS)*, and (OCS),. The solid line is a Lorentzian fit to the data points (0)measured with a single laser operated at a fluence of 0.2 Jlcm'. The two-laser data (X) are displaced upward for clarlty, and the point corresponding to the fixed-frequency laser (0.1 J/cm') is circled. The dashed line is the same as the solid line, displaced upward for comparison with the two-laser data.

the clusters. The additional attenuation of the cluster signal is then measured as the second laser is scanned over the entire absorption band. If the absorption line were inhomogeneously broadened, the second laser would cause relatively little additonal dissociation near the frequency of the first laser, since the first laser has already depleted that population. The result of subtracting the signals with the second laser off and on would be a spike in the twolaser photodissociation spectrum near the frequency of the first laser. To obtain the two-laser spectra, the fixedfrequency laser was kept at a constant fluence of -0.10 J/cm2 and a fixed frequency of 1046.9 cm-' as indicated by the circled symbol in the figure. The second laser was then scanned over the absorption band at a constant fluence of -0.2 J/cm2. The two-laser spectra in Figure 2 are displaced upward in the plots in order to make them distinguishable from the single-laser spectra. The dashed lines are the same functions as the best fits to the single-laser spectra, displaced upward by the same amounts. Within experimental uncertainty the single-laser and two-laser spectra are identical, indicating that any inhomogeneous contributions to the observed line widths are small compared to the homogeneous line widths. Neither the absorption peak locations, widths, nor cross sections were found to vary significantly with molecular beam source conditions or laser fluence. In Table I, we summarize the spectral data for these OCS clusters. The peak frequencies, frequency shifts relative to isolated OCS, lifetimes 7 calculated from the homogeneous line widths, and transition dipole moments ( P ) are ~ given for each absorption peak. The error limits

IF? Photodissociation of Ar-OCS, (OCS),, and (OCS),

The Journal of Physical Chemistry, Vol. 87, No. 12, 1983 2099

TABLE I: Spectral Data for OCS Clusters

(OCS), 1 0 - 4 ( ~) , , e

cluster u o , cm-' Ar.OCS 1046.0 i 0.2 (OCS),

1044.7 i

(OCS), 1044.0 i 1035.8 2

Avo,'

-1.0 0.1 - 2 . 3 0.2 -3.0 0.5 -11.2

' A v o = uo(ckster) - v0(-).

cm-' i i. i

i 7

7,O

0.2 0.1 0.2 0.5

PS

D2

5.3 i 1.5 2.0 ?: 1.45 f 0.1 6.7 i 0.86 f 0.1 14.4 f 6.5 i 0.84 t 0.2

1045.1 c n i

0.8 1.0 3.2 2.3

= 1/[2?ic(fwhm)] cm-' vo=

Reference 17 gives the following values for OCS: 1047.04 cm-'

( I J ) , ~ ,= ~

14.9 X

D2.

/

/ IO"

- --

100 m/sec

given are the statistical uncertainties of Lorentzian fits to at least three independent measurements of each spectrum. The relative values of ( P ) for ~ the various clusters are determined by normalizing each spectrum to the cross section values determined for all three clusters at one laser frequency under identical molecular beam and laser conditions. The absolute values of ( P ) depend ~ on an absolute measurement of the laser fluence actually experienced by those clusters sampled by the mass spectrometer. The fluence measurement might be in error by as much as 30%. Even allowing for this uncertainty and the additional uncertainty in fitting a peak shape to the sparse data set, the absorption intensity is extraordinarily small for Ar. OCS-roughly an order of magnitude less than that for isolated gas phase OCS. The (OCS)zintensity is only about 45% of the monomer intensity, while the absorption intensity of (integrated over both peaks) is only slightly larger than that of the monomer. Thus, while the OCS vibrational frequency within the cluster is not very sensitive to interactions with the other cluster species, both the line width and the absorption intensity are extremely sensitive to these interactions. A "local mode" picture of the clusters, in which each OCS molecule absorbs independently, is clearly not appropriate for these cases. Product Velocity Distributions. While the relationship between the absorption line width and the vibrational predissociation dynamics remains questionable, the dissociation product speed and angle distributions unambiguously give direct information on the dissociation process. The product translational energy distribution is sensitive to the presence or absence of a potential energy barrier in the dissociation coordinates and to the partitioning of the product energy between internal and translational coordinates. The product angle distribution with respect to the axis of laser polarization reflects both the distribution of transition dipole moments for the absorbing clusters and the coupling of rotational and orbital angular momentum in the separating molecules. OCS product velocity distributions were measured by scanning repeatedly over a set of predetermined laboratory scattering angles, accumulating time of flight (TOF) spectra at each angle. In order to obtain the greatest possible sensitivity to anisotropy in the product angle distribution in the center of mass (CM) coordinate system, we used two COz lasers, with both laser beams oriented perpendicular to the primary molecular beam axis, but with one laser polarization parallel to, and the other perpendicular to, the molecular beam axis. The data were acquired by alternating the laser polarizations on successive shots. The results for the photodissociation of OCS dimers at 1045.1 cm-l are displayed in Figure 3 as contour maps of the probability for scattering into the Cartesian element d3v' in velocity space. Under the molecular beam source conditions used in these experiments, a significant fraction of Ar-OCSwas also present in the molecular beam along with (OCS)z. The monomer product speed distribution therefore contains

X.90"

O

'

7 5432

/

I05

0"

\50% DIMER B E A M PROFILE

2 5 0 % DIMER B E A M PROFILE Flgwe 3. Contour maps representingthe speed and angle distributions of the OCS monomer fragments from the photodissociation of (OCS),. The (OCS), primary beam fwhm speed and angle distributions are also shown and the laser polarizations are Indicated by the vertical and horizontal arrows superimposed on these distributions. The contour lines were visually drawn through the data points from individual TOF distributions, which are shown for every other contour. The laser fluences were both 0.2 J/cm2.

some OCS resulting from the photodissociation of ArsOCS as well as (OCS)2. Based on the relative dissociation cross sections for ArSOCS and (OCS)z and the relative mass spectrometer signals for these two clusters in the primary beam, we would estimate that