Spectroscopic Investigation of Zinc-Containing Organometallic

Allan J. McKinley and Emmanuel Karakyriakos , Lon B. Knight, Jr., Robert Babb, and Aaron Williams. The Journal of Physical Chemistry A 2000 104 (16), ...
3 downloads 0 Views 617KB Size
J. Phys. Chem. 1994,98, 10427-10431

10427

Spectroscopic Investigation of Zinc-Containing Organometallic Radicals Prepared Using a Pulsed Electrical Discharge Nozzle Ian M. Povey, Andrew J. Bezant, Gary K. Corlett, and Andrew M. Ellis* Department of Chemistry, University of Leicester, University Road, Leicester LEI 7RH, U.K. Received: June 11, 1994; In Final Form: July 21, 1994@

A pulsed electrical discharge source has been used to obtain jet-cooled laser excitation spectra of organometallic free radicals. High-voltage (600-1200 V) discharge across a Zn(CH3)dinert gas mixture allowed a strong LIF signal of the ZnCH3 radical to be obtained. Extensive fragmentation to form molecules such as C2 and C3 was not observed, indicating that a pulsed discharge source can be a relatively clean source of organometallic intermediates. Interestingly, a weak ZnCH3 spectrum could also be obtained by discharge of Ar/Al(CH3)3 across zinc electrodes at voltages exceeding 1200 V. In this case, the ZnCH3 was formed by reaction between sputtered zinc atoms from the electrodes and methyl radicals produced by fragmentation of aluminum trimethyl. A pulsed discharge across Zn(C2H5)2/Ar produced a new LIF spectrum which was attributed to the zinc monoethyl radical, ZnC2Hs. This is the first report of a spectrum of this radical.

-

Introduction

1 cm

There has been considerable interest in the photodissociation dynamics of group 12 metal alkyls. These volatile compounds can be used to deposit the corresponding metal on a substrate by UV irradiation and are therefore of considerable interest to the semiconductor fabrication industry."-3 In addition, their relative simplicity compared with many metal precursor molecules makes them useful model compounds for understanding organometallic photodissociation processes. The photodissociation of zinc alkyls has been investigated in two recent publication^.^,^ Most recently, Seidler has employed time-resolved IR spectroscopy to monitor products formed by 193 nm photolysis and found that the data is in accord with a simple metal-alkyl bond homolysis mechanism! In a slightly earlier but detailed study of the dynamics, Jackson examined the laser photodissociation of zinc dimethyl, zinc diethyl, and zinc dipropyl at 248 nm.5 In the case of zinc dimethyl, both zinc atoms and zinc monomethyl radicals were detected by laser-induced fluorescence (LIF) spectroscopy. It was deduced that absorption of a 248 nm photon results in fission of one Zn-C bond and formation of highly energized ZnCH3. Unless collisionally stabilized by excess inert gas, ZnCH3 underwent further decomposition to form Zn atoms and CH3. For the higher zinc alkyls, essentially the same photodissociation mechanism was assumed to occur, even though the corresponding zinc monoalkyl radicals could not be detected. The failure to detect either zinc ethyl or zinc propyl was attributed to depopulation of the electronic excited state by rapid internal conversion. If such a mechanism was operating, it would be expected to be much less important in ZnCH3 due to the lower density of vibrational states, thus explaining the LIF observation of ZnCH3. In this paper, we describe a preliminary account of work which shows that the zinc monoethyl radical does have an LIF spectrum. Instead of photolysis, we have employed a pulsed electrical discharge source to fragment zinc dialkyl precursors. Pulsed discharges have recently been employed in several laboratories as a low-cost alternative to laser photolysis as a source of transient molecules.6-8 However, a pulsed discharge

* Author to whom conespondence should be addressed. @

Abstract published in Advance ACS Absrracfs, September 1, 1994.

0022-365419412098- 10427$04.5010

I Teflon flxture

Figure 1. Schematic illustration of pulsed discharge nozzle.

has not so far been exploited as a source of organometallic intermediates. Our source has enabled intense laser excitation spectra of both zinc monomethyl and zinc monoethyl to be recorded. In addition, as will be described, we have successfully employed a discharge to produce organometallic radicals via metal atom sputtering off the surface of the electrodes followed by reaction with organic fragments. This source offers the possibility of preparing a range of new and unusual organometallic intermediates for spectroscopic investigations.

Experimental Section The pulsed electrical discharge source employed was similar to that described by Bondybey et a1.I Briefly, a Teflon fixture was attached to the faceplate of a commercial pulsed valve (General Valve Series 9) and mounted in a vacuum chamber pumped by a Leybold WAU501/D40B Rootslrotary pump combination. Within the fixture, two electrodes were situated approximately 1 mm apart, and when a high, constant voltage was applied across the electrodes, a discharge was initiated by entrance of a short pulse (typically 300 ps duration) of highpressure gas into the fixture channel. Various electrode configurations were employed including disc-shaped and pencil0 1994 American Chemical Society

Povey et al.

10428 J. Phys. Chem., Vol. 98, No. 41, 1994

-

ZnCH,

I

I

23900

-X2A, 0:band

I

23940

I

1

I

23980

Wavenumber 1cm-l Figure 2. Laser excitation spectrum showing the ZnCH3 .&2E,/2-~ZA~ 0; band centered at ca. 23 953 cm-l. The conditions employed were 800 V discharge voltage, 5% zinc dimethyl in helium carrier gas (total pressure 2 bar), and zinc disc electrodes.

shaped electrodes, and all gave virtually the same performance. A typical arrangement is illustrated in Figure 1. Following discharge, the gaseous mixture traversed the fixture and on exit formed a supersonic jet expansion. The supersonic jet was crossed 1-2 cm downstream from the fixture by the beam from a dye laser (Spectra Physics PDL-3 pumped by a Quanta Ray GCR-11 Nd:YAG laser; dye laser linewidth rated as 0.08 cm-'). The unfiltered fluorescence was then collected by anfll.5 lens and detected by a photomultiplier tube (Hammamatsu R562). The output from the PMT was then amplified using a homemade preamplifier, digitized by a CAMAC-based transient digitizer (LeCroy Model 2262), followed by PC storage and manipulation. All zinc dialkyl samples were obtained from Epichem (U.K.) (min purity 99.9%)and were thoroughly degassed on a vacuum line before use. Sample temperatures were controlled using slush baths to achieve the desired vapor pressure of the precursor. Experiments were performed with either helium or argon carrier gas, with precursor concentrations being 5 5 % .

Results and Discussion Zinc Monomethyl. In order to establish the feasibility of using a pulsed electrical discharge to produce organometallic radicals, initial experiments focused on the ZnCH3 radical. LIF excitation and dispersed fluorescence spectra of this molecule have been recorded previously by Robles et al.using photolysis of the zinc dimethyl precursor in a supersonic jet.g The excitation spectrum of the ZnCH3 radical shows intense bands at 23 953 and 24 203 cm-', arising from the AZE1/2-X2A1 and A2E3/2-A2A1 0; transitions, respectively. Structure due to excitation of Zn-C stretching and methyl umbrella vibrations was also observed. The laser excitation spectrum of a discharged Zn(CH3)fie mixture is shown in Figure 2. It consists of a strong band centered at ca. 23 953 cm-' which is only present with a discharge voltage exceeding 600 V (this is the minimum voltage required to initiate the discharge under our conditions). The obvious assignment of this band is to the AZE1/2-g2A1 0; transition of ZnCH3. With a Nd:YAG pumped dye laser, the

spectral region above 24 OOO cm-' could not readily be accessed, and so it was not possible to observe either the A2E3/2-X2A1 0; band nor higher vibronic transitions. However, the absence of any other bands in the range 23 150-24 000 cm-', apart from very weak ZnH features (see below), leaves no doubt that the band shown in Figure 2 arises from the A2E1/2-X2A1 0; transition of the ZnCH3 radical. It is worth commenting on the fact that no other bands, apart from ZnH lines, were observed in the 23 150-24 000 cm-I region. Recent work has indicated that pulsed discharges of the type employed here may be rather destructive sources of transient molecules. For example, C3 has been observed by Baker et al. using LIF spectroscopy after a pulsed discharge through pure CO.lo Indeed, Baker et al. were able to observe rather strong vibrational hot bands from the A1lIu-A1&+ transition of C3 in the spectral region investigated in the present work. However, no evidence for hot bands of C3 was found in our experiments during the discharge of zinc dimethyl nor was the 2-0 Swan band of C2 observed.11 Furthermore, another possible discharge fragment, CH, whose A2A-X211 origin transition falls within the spectral range investigated here, was not detected." The formation of species such as CZ,C3, and CH must involve secondary processes including several fragmentation steps. The energy for each fragmentation step is expected to be produced mainly by collision with metastable rare gas atoms in our experiments, which in turn are produced by electron impact excitation in the discharge. The low precursor concentrations used in our work, as opposed to the experiments by Baker el ~ l . , 'minimize ~ the probability of secondary processes occumng and therefore inhibit the formation of species such as CH, CZ, and C3. Weak ZnH lines arising from the A211-X2E+ 0-0 system were, however, seen in the 23 250-23 700 cm-' region. ZnH could have been formed by direct reaction between zinc and hydrogen fragment atoms or by hydrogen abstraction from an alkyl group by electronically excited zinc atoms. Since ZnH is a prominent secondary photodissociation product when UV photolysis of zinc dimethyl is employed,12 our observations indicate that a discharge of the type used in this work is not necessarily more destructive than photolysis as a source of ZnCH3. The presence of ZnH in the discharge provides a simple means for estimating the rotational cooling achieved during the expansion. In fact, ZnH lines were much stronger in discharges through zinc diethyl (see below) than through zinc dimethyl, and so rotational temperature determinations were more reliable when using the former precursor. Depending on the experimental conditions, simulations of the easily resolved rotational structure have yielded rotational temperatures ranging from about 8 through to 60 K. One would not expect vibrational cooling to be as effective as rotational cooling, but we have nevertheless seen no significant vibrational hot bands of zinc monomethyl during a series of experiments. Zinc Monoethyl: Establishing the Carrier. When a discharge through zinc diethyl was employed instead of zinc dimethyl, a new laser excitation spectrum was observed, as shown in Figure 3. On the high-wavenumber side of the spectrum shown in Figure 3, intense rotational lines of the A2lT-X2Z+ transition of ZnH are clearly indicated. In addition to the narrow ZnH lines, a number of broader features, beginning with the strongest band at 22 515 cm-' and extending out to approximately 23 600 cm-', can be seen. We believe that all of these bands arise from a zinc-containing organometallic, the most likely carrier being zinc monoethyl, ZnC2H5. The evidence for this assignment is considerable, key factors

Zinc-Containing Organometallic Radicals

J. Phys. Chem., Vol. 98, No. 41, 1994 10429 ZnH

1

22500

kn,,z,n-x2c+

,

, 23000

1

24000

23500

Wavenumber/cm-' Figure 3. Laser excitation spectrum obtained following discharge of zinc diethyl/argon mixture (0.1% Zn(Cfi5)z in argon, copper electrodes, 600 V). Rovibronic transitions of ZnH can be readily identified in this spectrum by their characteristically narrow linewidths. As detailed in the text, the remaining, broader bands are all attributed to a single molecule, the zinc monoethyl radical, ZnCzH5. (For a list of band positions, see Table 3).

being as follows: (i) The bands disappear when zinc diethyl is not present. (ii) They are only observed for discharge voltages z 600 V. (iii) Variation of experimental conditions such as discharge voltage (600- 1100 V) and Zn(CzH5)z/Ar ratios yielded no change in relative band intensities, apart from the ZnH bands relative to the remaining bands. This indicates that all non-ZnH bands arise from the same molecule. (iv) Use of other volatile ethyl-containing compounds, such as gallium triethyl or boron triethyl, produced no LIF spectrum in this region. (v) No rotational structure was resolved, thus eliminating small molecules such as CH, Cz, and C3. Since spectroscopic data for these particular molecules are widely available, they can, of course, also be eliminated on the basis of band positions. (vi) The presence of several bands forming no obvious pattern is indicative of a polyatomic molecule of some complexity. The above factors are all consistent with an assignment to a single zinc-containing polyatomic molecule. However, given the observation of an intense ZnCH3 LIF signal on discharge of zinc dimethyl and in view of the reasonable proximity of the bands in Figure 3 relative to the A-2 origin of ZnCH3, the carrier of the spectrum is almost certainly the zinc monoethyl radical. It was mentioned earlier that Jackson attempted to observe zinc monoalkyl radicals by LIF in a photodissociation study of zinc d i a l k y l ~ .ZnCH3 ~ was observed in the photodissociation of Zn(CH3)2, but no evidence for ZnC2H5 was found in the corresponding experiment on Zn(CzH5)z. However, it should be noted that Jackson limited his search to transitions above 23 470 cm-'. As can be seen from Figure 3, the ZnCzH5 bands are mainly rather weak in this region, thus providing a possible explanation for their nonobservation in Jackson's work. Zinc Monoethyl: Comments on the Assignment. Complete assignment of all bands on the basis of the laser excitation spectrum alone is not a practical prospect. The inevitable lower equilibrium symmetry of zinc monoethyl compared with zinc monomethyl adds considerable complexity. The highest feasible point group symmetry for ZnCzH5 is C,, and this has the effect of removing the electronic and vibrational degeneracies found in ZnCH3. Thus, the analogue of the A2E state of ZnCH3 will

TABLE 1: Results from ab Initio Calculations on the X2A1 Electronic State of ZnCH3 vibrational datab geometrical Dammeters' ~~~~

~

vibration

description

metry

sym-

frequency/ cm-'

symm C-H stretch C-H umbrella Zn-C stretch C-H stretch CH3 scissors CH3rock

al a, al e e e

3148 1185(1064) 445 (445) 3245 1546 617(315)

~

m- = 2.052 (1.987) 8, r ~ += 1.091 (1.094) 8, = 109.8' (111.0")

VI

v2 v3 vd vg vg

The geometrical parameters in parentheses are from MP2 calculations by Jamorski and Dargelos (ref 15). The frequencies calculated in this work are harmonic vibrational frequencies. The values in parentheses are fundamental frequencies derived from dispersed fluorescence experiments (ref 9). be resolved into two distinct electronic states in ZnCzH5, A' and A" states, both of which might be observable by LIF. The larger number of vibrational modes and the loss of vibrational degeneracies when compared with ZnCH3 also lead one to expect more vibrational structure in the spectrum of ZnCzH5. To obtain additional information, ab initio calculations have been carried out on both ZnCH3 and ZnCzH5 at the HartreeFock level using GAUSSIAN 92.13 These relatively low-level calculations, performed using a double-5 basis set with added polarization functions on each atom,14 focused on the ground electronic state of each molecule in order to determine equilibrium geometries and vibrational frequencies. The results obtained for the %A1 state of ZnCH3 are summarized in Table 1. Geometry parameters in parentheses in Table 1 are from MP2 calculations by Jamorski and Dargelos using basis sets of comparable quality to our own.l5 Reasonable agreement is obtained between the two sets of parameters. Most relevant to this work is a comparison of the ab initio harmonic vibrational frequencies with those experimental fundamentals that are available (in parentheses in Table 1). The agreement is quite good except for the degenerate CH3 rocking mode, Y6,although it should be noted that the experimental Y6 frequency is only t e n t a t i ~ e .It~ should also be noted that Jamorski and Dargelos have performed extensive CI calculations on ZnCH3 and its low-

Povey et al.

10430 J. Phys. Chem., Vol. 98, No. 41, 1994 TABLE 2: Results from ab Initio Calculations on the X2A' Electronic State of ZnC2H5 vibrational data geometrical parameterso rz.-c = 2.086 8, rc-c = 1.527 rc-w = 1.092 8, rc-H; = 1.093 8, T C - H ~ = 1.090 8, ea-ci. = 114.70 OZ~-C-HI = 105.8' eH;-C-H4 = 107.0" OH~-C-HS = 107.1"

A

vibrationb

description

symmetry

Zn-C-C bend CH3 torsion Zn-C stretch CHdCH3 rock CHdCH3 twist CHzICH, wag C-C stretch CHdCH3 wag CH2/CH3 twist

a' a" a' a" a"

a' a' a' a"

TABLE 3: Band Positions in the Laser Excitation Spectrum of Zinc Monoethyl frequencykmintensityb 22 515 22 740 22 760 22 939 22 951 22 990 23 031 23 141 23 219 23 292 23 410 23 431 23 440 23 469 23 521

frequency/ cm-' 196 231 381 600 960 993 1058 1207 1343

H1 and H2 designate the equivalent hydrogen atoms in the CH2 group. H3, H4, and H5 refer to the methyl hydrogen atoms, H3 being in the Zn-C-C plane. b A nonstandard numbering system for the vibrational modes is employed here for the sake of simplicity. Only those vibrations with harmonic frequencies 1500 cm-I are listed. The description of modes is very approximate and is intended only to convey some indication of the vibrational motion. a

lying electronic states.l 5 Their calculations indicate that the A ~ E - X ~ A transition ' results largely from a zinc-localized 4s 4p transition with no major change in Zn-C bonding character. This assignment is reasonably consistent with the observed excitation spectrum of ZnCH3, which exhibits only a weak progression in the Zn-C stretching vibration, ~ 3 . ~ In Table 2, the results from Hartree-Fock calculations on ZnC2H5 are presented. It has been found from these calculations that ZnCzH5 has C, equilibrium symmetry, which, by analogy with ZnCH3, would give rise to a 2A' electronic ground state. In the case of vibrational frequencies, we have listed only those modes with harmonic frequencies < 1500 cm-', since only these vibrations could conceivably be excited within the scan range covered in the ZnC2Hs excitation spectrum in Figure 3. Even so, there are nine vibrations shown. Of course, not all vibrations need necessarily be excited in the spectrum. To begin with, Franck-Condon arguments essentially preclude excitation of non-totally symmetric vibrations. In addition, by analogy with ZnCH3, it is reasonable to assume that the bands in the excitation spectrum of ZnC2H5 arise from an electronic transition or transitions primarily localized on the zinc atom. If this assumption is valid, one would expect the major vibrational structure to arise from vibrations directly affected by this electronic transition. Thus, for example, significant FranckCondon factors might be expected for excitation of the Zn-C stretching and Zn-C-C bending modes, whereas less FranckCondon activity would be expected for vibrations localized on the ethyl framework. A list of the wavenumbers of bands attributed to ZnC2H5 is given in Table 3. The most intense band, at 22 515 cm-', is assigned to the A-3 0; transition, since no other bands were observed at lower wavenumbers in scans down to 21 000 cm-'. By comparison with the ab initio vibrational data in Table 2, the next reasonably intense band, that at 22 760 cm-', could be attributed to excitation of either the Zn-C-C bend ( V I ) or the methyl torsion ( ~ 2 ) . Given the arguments made earlier, the more likely assignment is to the Zn-C-C bend if it is part of the A-2 system. An alternative assignment is that the 22 760 cm-' band is due to excitation to the B state of ZnC2H5, the other state resulting from the loss of degeneracy when compared with the AZEstate of ZnCH3. This assignment finds favor when the cluster of four bands in the 22 930-23 040 cm-' region are considered. If these bands all arise from the A-2 transition, the implication is that ZnC2H5 has four vibrations with frequencies in the 420-520 cm-' region, all of which are Franck-

-

S

W

m S

m m

m W W W W W W W

m

Positions of band centers. Estimated accuracy, based on comparison with ZnH rovibronic lines, is f l cm-I. Intensity designated as strong (s), medium (m), or weak (w) to assist comparison with Figure 3.

Condon active. While one of these bands could be due to excitation of the Zn-C stretching vibration, ~ 3 in, the A state, the ab initio frequencies in Table 2 are clearly not consistent with them all belonging to a single electronic band system. Although the calculated vibrational frequencies refer to the electronic ground state, whereas the experimental intervals represent vibrational frequencies in the excited state, there is, by analogy with ZnCH3,9 no reason to expect a major change in vibrational frequencies on A-8 electronic excitation. Further support for this suggestion comes from the absence of any long vibrational progressions in the spectral region investigated. Consequently, the likely interpretation of the spectrum in Figure 3 is that it contains contributions from at least two different electronic band systems of ZnC2H5. It seems probable that these include the 2A' and 2A" excited electronic states that correlate with the A2E state of ZnCH3. It will be apparent from the comments made above that a combination of laser excitation data and relatively simple ab initio calculations is insufficient to achieve a firm assignment of even the lowest frequency bands in the excitation spectrum of ZnC2H5. The ab initio calculations on the ground state of ZnC2H5 are helpful but cannot reliably be used to distinguish between closely spaced vibronic bands. Assignment would be greatly assisted by recording dispersed fluorescence spectra for each of the observed excitation transitions. This should allow vibronic bands associated with different electronic states to be distinguished, as well as provide vibrational frequencies for the ground electronic state. Although not currently equipped to perform dispersed fluorescence measurements, we nevertheless aim to perform such experiments in the near future. Sputtering Experiments. Recent work by Bondybey and co-workers has shown that a pulsed discharge source of the type used in the current work can also produce metal atoms by electrode sputtering. For example, a discharge through an inert gas/CFBr3 mixture between copper electrodes has yielded intense LIF spectra of CuBr, while discharge of inert gases across gallium and thallium electrodes has been employed to prepare inert gas van der Waals clusters of these m e t a l ~ . ~ J ~ J ~ During the course of the present work, we attempted to ascertain whether this idea could be extended to prepare organometallic intermediates. The basic aim was to be able to sputter metal atoms off the electrodes, to simultaneously fragment a source of organic ligands, and thus to react the two entities to form detectable amounts of organometallic intennediates. Laser ablation has already been employed by Ellis et al.

Zinc-Containing Organometallic Radicals

J. Phys. Chem., Vol. 98, No. 41, 1994 10431 of volatile precursors. Discharge of a zinc dimethyvinert gas mixture allowed an intense laser-induced fluorescence excitation spectrum of the ZnCH3 radical to be obtained under supersonic jet conditions. Comparable experiments on zinc diethyl yielded a new laser excitation spectrum which was attributed to the ZnC2H5 radical, a molecule whose gas phase spectrum has not previously been recorded. Further work, including dispersed fluorescence measurements, will be needed to identify the specific transitions responsible for the zinc monoethyl spectrum.

23900

23950

24000

Wavenumber/cm-'

Figure 4. Laser excitation spectrum of ZnCH3 produced by sputtering of zinc atoms from zinc electrodes in the presence of Al(CH3)JAr (1.5% aluminum trimethyl in 3 bar of argon, discharge voltage = 1400 V).

with the same overall philosophy,'* but a pulsed discharge equivalent would be an important development because of the much lower cost of the latter source. If successful, a pulsed discharge sputtering source could be used to prepare new organometallic intermediates for gas phase spectroscopic studies which cannot be obtained by simple fragmentation of organometallic precursors. For example, a variety of simple metal carbenes, carbynes, and acetylides could in principle be made by selection of suitable organic precursors and various electrode materials. Our preliminary work has shown that the basic idea is feasible. Using zinc electrodes and a mixture of argon and a methyl precursor, aluminum trimethyl, we have been able to obtain reasonable ZnCH3 LIF signals, an example being shown in Figure 4. To obtain this spectrum, high discharge voltages ('1200 V) are necessary to be able to sputter zinc atoms. Unfortunately, under these vigorous conditions the ZnCH3 signal disappears after, at most, a few minutes of the discharge. This is apparently due to formation of a black film, presumably carbon, over the surface of the electrodes, which does not impede the discharge but does prevent further metal atom sputtering. Confirmation of this was obtained by a discharge study using indium electrodes. When pure argon was used as the discharge gas, an intense atomic indium 2S1/2-2P3/2 laser excitation signal was observed at 22 160 cm-l.I9 This signal could be made to be stable for up to several hours, but on addition of a small amount of a volatile organic or organometallic compound it rapidly disappeared. Once again, a carbon film was formed on the electrodes. A possible solution to the carbon film problem would be to employ a dual-discharge system. We plan to construct a Teflon fixture containing two discharge assemblies, through one of which will flow pure inert gas, while through the other there will be an inert gadorganic precursor mixture. In this manner, it should be possible to obtain clean metal atom sputtering in one channel and organic or organometallic precursor fragmentation in the other. The two gas pulses will then flow into a common channel where reaction will take place followed by expansion into vacuum.

Conclusions A pulsed electrical discharge nozzle has been shown to be suitable for producing organometallic radicals by fragmentation

At high voltages ( 21200 V), metal atom sputtering from the cathode was found to be substantial. This enabled ZnCH3 to be synthesized by direct reaction between sputtered zinc atoms from zinc electrodes and methyl fragments from aluminum trimethyl. This means of synthesizing metal-containing intermediates for spectroscopic studies, first suggested by Schlachta et a1.: appears to be a promising low-cost source of organometallic intermediates for a wide variety of metals.

Acknowledgment. The authors would like to thank Mr. John Weale and Mr. Philip Acton for help in constructing the experimental apparatus, and the SERC's Laser Support Facility for the loan of laser equipment. A.J.B. and G.K.C. are grateful to SERC and the Isle of Man Department of Education, respectively, for the award of research studentships. This work was supported by grants from SERC (GR/H91749) and The Royal Society. References and Notes (1) Bauerle, D. Chemical Processing with Lasers; Springer: Berlin, 1986. (2) Deutsch, T. F.; Ehrlich, D. J.; Osgood, R. M. Appl. Phys. Lett. 1979, 35, 175. (3) Johnson, W. E.; Schlie, L. A. Appl. Phys. Lett. 1982, 40, 798. (4) Seidler, P. F. J . Phys. Chem. 1994, 98, 2095. (5) Jackson, R. L. J. Chem. Phys. 1992, 96, 5939. (6) Bramble, S. K.; Hamilton, P. A. Chem. Phys. Lett. 1990,170, 107. (7) Schlachta, R.; Lask, G.; Tsay, S H.; Bondybey, V. E. Chem. Phys. 1991, 155, 267. (8) Rosser, K. N.; Wang, Q.-Y.; Westem, C. M. J. Chem. Soc., Faraday Trans. 1993, 89, 391. (9) Robles, E. S . J.; Ellis, A. M.; Miller, T. A. Chem. Phys. Lett. 1991, 178, 185. (10) Baker, J.; Bramble, S. K.; Hamilton, P. A Chem. Phys. Lett. 1993, 213, 297. (11) Huber, K. P.; Herzberg, G. Molecular Spectra and Molecular Structure, Vol. 4, Constants of Diatomic Molecules; Van Nostrand: New York, 1979. (12) Robles, E. S. J.; Ellis, A. M.; Miller, T. A. Unpublished data. (13) GAUSSIAN 92; Frisch, M. J.; Trucks, G. W.; Head-Gordon, M.; Gill, P. M. W.; Wong, M. W.; Foresman, J. B.; Johnson, B. G.; Schlegel, H. B.; Robb, M. A,; Replogle, E. S . ; Gomperts, R.; Andres, J. L.; Raghavachari, K.; Binkley, J. S.; Gonzalez, C.; Martin, R. L.; Fox, D. J.; Defrees, D. J.; Baker, J.; Steward, J. J. P.; Pople, J. A. Gaussian Inc.: Pittsburgh, PA, 1992. (14) The double-5 basis sets were taken from Schafer et al. ( J . Chem. Phys. 1992, 97, 2571) and were supplemented with one p polarization function on zinc (exponent 0.16), two d functions on carbon (exponents 0.4 and 1.6), and two p functions on hydrogen (exponents 0.4 and 1.6). (15) Jamorski, C.; Dargelos, A. Chem. Phys. 1992, 164, 191. (16) Stangassinger, A,; Scheuchenpflug, J.; Prinz, T.; Bondybey, V. E. Chem. Phys. Lett. 1993, 209. 372. (17) Stangassinger, A,; Scheuchenpflug, J.; Prinz, T.; Bondybey, V. E. Chem. Phys. 1993, 178, 533. (18) Ellis, A. M.; Robles, E. S. J.; Miller, T. A. J . Chem. Phys. 1991, 94, 1752. (19) Moore, C. E. Atomic Energy Levels; NBS, 1952.