Photodissociation dynamics of HNCO at 248 nm

9 August 1996

CHEMICAL PHYSICS LETTERS

ELSEVIER

Chemical Physics Letters 258 (1996) 164-170

Photodissociation dynamics of HNCO at 248 nm R.A. Brownsword, T. Laurent, R.K. Vatsa ~, H.-R. Volpp *, J. Wolfrum Physikalisch-Chemisches Institut der Universitfit Heidelberg, Im Neuenheimer Feld 253, D-69120 Heidelberg, Germany

Received 10 May 1996; in fmal form 6 June 1996

Abstract Using the laser photolysis/laser-induced fluorescence (LP/LIF) pump-probe technique, the gas-phase photodissociation dynamics of room temperature HNCO was studied at a photolysis wavelength of 248 nm. H atoms produced via HNCO + h v ~ H(2S)+ NCO(X 2H) were detected by vacuum-UV laser-induced fluorescence (VUV-LIF) at the Lyman-a transition. By means of a calibration method - using H2S photolysis as a source of well defined H atom concentrations the absolute cross section for direct photolytic H atom formation was determined to be tr H = (1.2 + 0.3)× 10 -2~ cm 2 molecule-~. From the H atom Doppler profiles, measured under single-collision conditions, the fraction of the available energy released as product translational energy was determined to be fT(H + NCO) = (0.55 + 0.02). In addition, the second energetically accessible 'spin-forbidden' dissociation channel, HNCO + h v ~ NH(X 3 E - ) + CO(X ~E ÷ ), was investigated. Our results show that at a wavelength of 248 nm, direct H atom formation is the dominant dissociation channel in the HNCO photolysis.

1. Introduction Isocyanic acid (HNCO) plays an important role in a variety of combustion environments ranging from solid propellant combustion [1] to the 'rapid reduction of nitrogen oxides' (RAPRENO x) by cyanuric acid [2], and urea injection [3], which is, in addition to the Thermal De-NO x process [4], one of the after-treatment processes commonly used on stationary combustion systems to control NO x emission. In addition, H N C O is a convenient photolytic source of NCO and NH radicals for laboratory kinetics studies of elementary reactions [5].

" Corresponding author. On sabbatical leave from Chemistry Division, Bhabha Atomic Research Centre, Bombay 400 085, India.

H N C O is a 16-valence-electron molecule with ground state C s symmetry. It is an important molecule o f the class HABC, which exhibits two competing photodissociation channels of the form H + A B C and H A + BC. An energy level diagram for H N C O and its dissociation products is shown in Fig. 1. Energies of the different asymptotes are from [6] and the ordering of the states of H N C O is taken from [7,8]. U V / V U V absorption spectra of H N C O have been measured down to about 120 rim. W o o and Liu [9] studied low-resolution electronic absorption spectra of H N C O and found a series of diffuse bands in the region 2 2 5 - 2 5 7 nm, merging into a continuum at the short-wavelength end. Dixon and Kirby [10] analysed the rotational structure of the absorption spectrum of H N C O between 200 and 280 nm and concluded that the excited state has an equilibrium geometry with an NCO bond angle of

000%2614/96/$12.00 Copyright © 1996 Elsevier Science B.V. All rights reserved. PH S0009-261 4(96)00626-4

R.A. Brownsword et al. / Chemical Physics Letters 258 (1996) 164-170

approximately 120°. The UV photochemistry of HNCO in general is dominated by the following two competing, spin-allowed dissociation channels:

HNCO(X '.~) + h v

-',

NH(a 'A) + CO(X '.Y.,+)

(1) -,

H(2S) + NCO(X2n).

(2)

In the VUV, Okabe [1 l] studied the production of electronically excited N H ( A 3 H , c IH) and NCO(A 2~ +) based on the emission observed in the UV and visible regions and derived a value of D0(H-NCO) ~ 113 kcal/mol. Uno et al. [12] studied the photodissociative excitation processes of HNCO in the wavelength region 107-180 nm. The photodissociation dynamics of HNCO following the 193 nm excitation have been studied in detail using various experimental techniques [6,13-15]. In their photodissociation studies Bradley et al. [16] showed that the threshold for HNCO dissociation lies at wavelengths longer than 240 nm and the quantum yield ~bl of channel (l) increases relative to that of channel (2) with decreasing wavelength. At a wavelength of 206.2 nm, ~bI = th2 = 0.5 has been reported by Woolley and Back [17]. Using 212.6 nm excitation, Yi and Bersohn [18] measured a yield in the H atom channel of ~b2 = (0.13 4- 0.01) and determined the fraction of the available energy channelled into product translational energy to be fT(H + NCO) = 0.58. Recently, the ion imaging technique has been used by Kawasaki et al. to detect CO in the 217 nm photolysis of HNCO [19]. Crim and co-workers [20] investigated the influence of vibrational excitation of the N - H stretching mode on the N C O / N H branching ratio at comparable excitation energies and, based on the S/N ratio in their experiment, estimated this ratio to be about 20 at a photolysis wavelength of 235 nm. It has been shown [6,15] that for photolysis at 193 nm, corresponding to a photon energy of 148 kcal/mol, HNCO is excited to a zg' state from which it dissociates into NH(tA) + CO and H(2S) -INCO(X2H). The production of NH(3E - ) via the dissociation channel,

HNCO(',~) + h~,~ NH(3~-,-) + CO(X'F-,*),

(3)

165

although energetically possible, was not observed in the 193 nm photodissociation of HNCO; this was attributed to the spin-forbidden nature of the process. Dissociation of excited HNCO prepared by chemical reactions, at similar total energies, has also been reported in the reaction of NH(1A)+ CO [7] and H + NCO [21]. Hack and Rathmann [7] concluded that the reaction of NH (IA) + CO proceeds via an HNCO(IA~, l~,) complex, formed without activation barrier, which decomposes mainly into H + NCO, and which also, to a minor extent (about 12%), undergoes intersystem crossing to the 3AZ surface to form NH(3E -) + CO. Dagdigian et al. [21] found that the reaction of H + NCO yields mainly N H ( 3 ~ - ) + C O and attributed the low formation rate of NH(IA) to the fact that the NH(~A)+CO reaction pathway is endothermic. The dissociation energy of HNCO has been a subject of some discussion, and only recently have two studies [6,22] reported an upper limit of D0(H + NCO) ~ ll0.1 kcal/mol. The fact that in this case H

H + NCO (A21£+)

NH (A3rl) + CO ( x l L "+)

(175.2)

(171.9) NH (b l'r+) + CO (X1Z+) (147.5)

150-

NH (alA) + CO (XlL"+) (122.8)

H + NC£ -6

EE 1 0 0 -

(11C NH (X3'r-) + CO (Xl~Z+)

£

(86.8)

t-

¢-

w

50-

0 --

HNCO (X)

Fig. 1. Energy level diagram of the HNCO dissociation processes. The shaded area covers a range of photolysis wavelengths previously used in photodissociation dynamics studies as described in the text.

166

R.4. Brownsword et al. / Chemical Physics Letters 258 (1996) 164-170

atoms could be formed in the 248 nm photodissociation of HNCO (see Fig. 1) motivated the studies to be described in the present Letter. In the following we report on the observation of H atom formation and the measurement of the H atom cross section at this wavelength. We have determined fT(H + NCO) and investigated the influence of the polarization of the photolysis laser on the H atom Doppler profiles. An attempt was also made to detect additional H atoms formed from NH(3E - ) via its reaction with NO in order to determine the quantum yields for the two energetically possible photodissociation channels (2) and (3).

2. Experimental HNCO photodissociation studies were carried out in a flow reactor at mTorr level pressures using the apparatus which has been used previously to investigate the photodissociation dynamics of HNCO at 193 nm and which is depicted and described in detail in Ref. [ 14]. Only a brief summary of the experimental method will be given here. In the present study a KrF excimer laser (Lambda Physik LPX 205i) was used to dissociate HNCO and HaS at room temperature. VUV-probe light for H atom detection (121.567 nm) was generated by resonant third-order sum-difference frequency conversion in a K r - A r mixture. The fundamental laser radiation was obtained from two tunable dye lasers, simultaneously pumped by a XeC1 excimer laser. The generated Lyman-ot light was carefully separated from the fundamental lasers by a monochromator lens and the probe beam was aligned so as to intersect the photolysis beam at right angles in the viewing region of the LIF detector. The delay time between the pump and probe pulses was typically 80-120 ns. The LIF signal was measured through a band-pass filter by a solar blind photomultiplier positioned at right angles to both pump and probe lasers. The VUV-probe beam intensity was monitored after passing through the reaction cell. In order to obtain a satisfactory SIN ratio, each point of the H atom Doppler profiles was averaged typically over 30 laser shots. LIF signal, VUV-probe and the photolysis intensities were recorded with a three-channel boxcar system and transferred to a microcomputer

where the LIF signal was normalized to photolysis and probe laser intensities. During the experiments room temperature HNCO flowed through the reactor at a rate high enough to ensure total renewal of the gas at the intersection region of the pump and probe lasers between successive laser shots. All measurements were carried out at a repetition rate of 6 Hz. The HNCO pressure in the cell was typically 50-100 mTorr; in the calibration measurements HES (UCAR electronic grade) was pumped through the reaction cell at a pressure of typically 20 mTorr. Photolysis laser fluence was kept low to avoid secondary photodissociation of SH radicals. In addition, several test runs were carried out in order to confirm that our result is not affected by side reactions of HNCO photolysis products. First, the pump-probe delay time was varied between 80 and 150 ns. No significant change in the H atom signal was observed, consistent with production of H atoms solely by direct HNCO photolysis. Secondly, in experiments in which the photolysis laser intensity was varied, no non-linear effects on the H atom fluorescence signal could be observed showing that multiphoton processes do not play a role under our experimental conditions. HNCO was prepared as described in our previous studies [14]. The purity was checked by FT-IR spectroscopy and MS analysis. The HNCO sample was stored at liquid nitrogen temperature. During the experiments, the temperature of the sample was maintained around - 50°C. Since the pressure of H2S used in the calibration measurements was markedly lower than that of HNCO, a correction due to the different absorbing conditions at the Lyman-ct wavelength in the test and calibration measurements had to be applied. The absorption cross sections of HNCO and H2S at 121.6 nm were therefore measured in the course of the present studies and found to be O'HNCO(121.6 n m ) = (2.5 _ 0.2) × 10 -17 cm 2 molecule - l and o-H2s(121.6 n m ) = (2.8 + 0.2) X 10 -17 cm 2 molecule- 1, in good agreement with values estimated from the VUV absorbtion spectra of Okabe [11] and Uno et al. [12]. Fig. 2 shows the corresponding Lambert-Beer plots. A small contribution to the LIF signal from H atoms produced directly by the Lyman-ct photolysis of HNCO and H2S was observed. In order to dis-

R.A. Brownsword et al. / Chemical Physics Letters 258 (1996) 164-170 1.5

1

i

i

J

H2S absorption

~c>_ ~¢=¢n1 a = ( 2 . 8 + 0 . 2 ) x t ~

v

c: 0.5

o

°o

1

2

I

i

3

I

i

J

4

I

i

I

J

HNCO absorption 1.5

167

sure of 8 Torr and recording the time evolution of the H atom concentration in order to detect any additional H atoms produced by the reaction NH(X3E - ) + NO ~ H + N20. To remove NO 2 impurities the NO flow was passed through a two-stage acetone trap (maintained at - 8 0 ° C ) and over KOH before entering the reactor. For polarization measurements the photolysis laser beam was linearly polarized (between 90 and 95%) in either a perpendicular or parallel sense relative to the probe beam direction using a 10-plate Brewsterangle suck polarizer, or was used without polarizing elements. The degree of polarization a , defined as the fraction of the total laser intensity polarized in the perpendicular direction, was measured for each of these three configurations with a second identical stack polarizer in a polarizer/analyzer arrangement.

i

_-e,

3. Results and discussion 0.5

3.1. HNCO + h v ~ H + NCO absolute cross section f o r H atom formation

0

1 I

2 I

3 I

4 I

5 I

6 I

n × I [1016 molec cm -2] Fig. 2. Lambert-Beer plots of the absorption of H2S and HNCO at the Lyman-ct transition wavelength 121.6 nm. The absorption path length was 20 c m . The absorption cross sections determined are indicated in the figure.

criminate between this 'Lyman-ot-background' H atom signal and that from H atoms produced in the 248 nm photolysis, an electronically controlled mechanical shutter was inserted into the photolysis beam path. At each point of the H atom Doppler profile, the signal was recorded first with the shutter open, and then with the shutter closed, the difference between the two signals representing the contribution from H atoms generated solely by 248 nm photolysis. A point-by-point subtraction procedure was adopted to obtain a signal free from the ' L y m a n - a background' H atoms described above. The possibility that channel (3) plays a role in 248 nm photolysis was investigated by photolyzing H N C O - N O mixtures in N 2 bath gas at a total pres-

The absolute cross section ~H for H atom formation at the photolysis wavelength of 248 nm was obtained by calibrating the H atom signal SH(HNCO) measured in the HNCO photodissociation against the H atom signal SH(H2S) from well-defined H atom number densities generated by photolyzing H 2S. The absolute cross section for the dissociation channel (2) was determined using the following relationship: orH = TS H(HNCO) O'H2s PH2s/SH (H2S) PHNCO"

(4) Here S H is the integrated area under the respective normalized H atom Doppler profiles. PH2S and PHNCO are the pressures of H2S and HNCO, respectively, tTH~s is the absorption cross section of H2S at 248 nm (O'H: s = 2.6 × 10 -20 cm 2 molecule- t [23]). The factor Y accounts for the different degree of absorption of Lyman-ot radiation by H2S and HNCO, using the previously measured absorption cross sections (see Section 2). Evaluation of the Doppler profiles according to Eq. (4) yielded a value of (r H = (1.2 _ 0.3) × 10 -21 cm 2 molecule- I.

168

R.A. Brownswor d et al. / Chemical Physics Letters 258 (1996) 164-170

3.2. HNCO + hu ~ H + NCO / N H ( S E - ) + CO branching ratio NH(3E

- ) is rapidly converted by NO to H atoms,

with literature values for the branching ratio (]~lH ranging from 0.65 to 0.8 [24-26]. If NH(aE - ) were a significant photolysis product, then in the presence of NO, a clear exponential growth in H atom concentration with time would be observed in addition to the 'background' concentration generated directly by dissociation. The observed kinetic behaviour of the H atoms was fitted using literature values for the NH(35', - ) + NO rate constant and H atom yield [24], and a branching ratio for the dissociation of HNCO to NH(aE - ) of 0.05 + 0.10 at the 90% confidence level was determined. However, as is clear from the quoted errors, the relatively poor S I N ratio of these data does not allow the participation of the NH dissociation channel to be unambiguously identified. Production of H + NCO is therefore concluded to be the dominant channel in the photodissociation of HNCO at 248 nm.

Here the brackets ' ( . . . )' imply an expectation value averaged over all angles. Following Bersohn et al. [28], the average energy released into z-translation can be written as a function of the total translational energy release (ET(lab)), the parameter fl and the degree of the photolysis laser polarization a, thus: ( Ez0ab)(H )) = ( l / a ) ( ET0ab)(H ) ) × [1 + / 3 ( 2 - 3 a ) / 5 ] .

Values of (Ez0ab)(H)) were evaluated from Doppler profiles measured under the three polarization conditions described in Section 2. From a least-

HNCO ,"'-2,

LL i

.

.

.

.

.

.

-3-2-101

3

-3-2-1

0 1 2 3

d)

3.3. HNCO + hv ~ H + NCO (k, ep )-vector correlation and center-of-mass translational energy distribution

VUV-LIF spectra were recorded with unpolarized, horizontally and vertically polarized radiation. The measured H atom Doppler profiles were evaluated in order to determine the fraction of the average available energy, EarI = hv + Eint(HNCO) - D0(HNCO), released as relative product translational energy (ET(cm)) into the H - N C O product channel. Via the linear Doppler shift, ~,- ~'0 = Vz ~'o/c, the observed LIF profiles directly reflect the probability distribution of the velocity component v z of the absorbing atoms along the propagation direction of the probe laser beam, from which the mean translational energy in the z-direction can be determined v i a ( E z ( l a b ) ( H ) ) = 1 / 2 mH(v2). Under the singlecollision conditions of our experiment, however, the H atom velocity distribution will show a spatial anisotropy described by the parameter/3 = 5(P2(kep)), where k and ep are unit vectors in the direction of the H fragment recoil and of the polarization vector of the photolysis beam, respectively [27].

(5)

H=S

IJ. _.% -6-4-20246

-6.4-20246

,"'2.

f)

II

tt/)

"o (1)

E it)

i

.

.

.

.

.

-6-4-20246

.

-6 .4 - 2 0 2 4 6 D o p p l e r shift [cm -1]

Fig. 3. VUV laser-induced fluorescence profiles of H atoms produced in the 248 nm photodissociation of 60 mTorr HNCO with (a) kpr 1 ep and (b) kprtlEp, and of 20 mTorr H2S with (c) kpr _L ep and (d) kpr]lE p. kpr and ep denote the direction of the probe beam propagation vector and of the photodissociation laser polarization vector, respectively. H atom Doppler shapes for the 248 nm photodissociation of H2S, simulated using fl = - 0 . 6 4 [29] for (e) k~. 3_ ep and ( 0 kpr[[£ p, are also shown. The ordinate is in arbitrary units and is not for comparison.

R.A. Brownsword et al./ Chemical Physics Letters 258 (1996) 164-170

squares fit of (Ez0ab)(H)) as a function of a according to Eq. (5), the parameters /3 and (ET(lab)(H)) were determined to be - 0 . 1 3 -I- 0.11 and 3.35 +_ 0.14 kcal/mol, respectively. Doppler profiles of H atoms produced by H2S photolysis at 248 nm were also recorded (see Fig. 3) for which the method of analysis described above gave /3 and (ETcI,b)(H)) values which are in very good agreement with the available literature values o f / 3 = - 0 . 6 4 [29] and (ET(lab)(H)) = 1.08 eV [30]. Finally the average translational energy (ET(cm)) released to the H - N C O products in the center-ofmass system was calculated using the expression: (ET(cm)) = (1 -t- mH//mNCo)(ET(Iab)(H)) - (mJmNCo)(ET<,,b)(HNCO)),

(6)

where (EV(lab)(H)) and (ET0ab)(HNCO)) are the average laboratory translational energies of the H atom

dissociation fragment and the HNCO molecule at T = 300 K, respectively. In principle, it has to be taken into consideration that the measured spectral profile represents a convolution of the laser spectral profile and the Doppler profile of the absorbing H atoms. As described in Ref. [31], the measured probe laser bandwidth of our laser system (Aulaser = 0.4 cm -~) is so small that a back-correction due to the finite probe laser bandwidth was well below the experimental uncertainty. (Ev(cm)) was calculated to be 3.41 + 0.14 kcal/mol. The average internal energy of HNCO is calculated to be 1.14 kcal/mol, using the vibrational frequencies of Ref. [32] and the classical approximation for the average rotational energy. Using the value D 0 ( H - N C O ) = 110.1 kcal/mol reported in [6,22] we determined the fraction fT of the available energy released into the relative translational degree of freedom of the H NCO products to be f T ( H + N C O ) = (Ev(cm)) / ( e . v , ) = (0.55 _+ 0.02). Fig. 3a and b show the H atom Doppler profiles from the HNCO photolysis recorded using different polarizations; no significant difference in the profiles is observed, consistent with an almost isotropic recoil velocity distribution. Although no calculated potential energy surface for the excited state of HNCO exists, information about the upper state potential has been drawn from experimental observations. Dixon and Kirby [10] recorded the UV spec-

169

trum of HNCO in the 200-280 nm region, observing rotational structure at long wavelengths which becomes diffuse below 265 nm and which disappears beyond 244 nm. Crim et al. [20] infer from Dixon and Kirby's results that the excited IA~' surface is bound at excitation energies below 37700 cm - j , predissociative in the N - H coordinate in the region 37700-41000 cm -I and unbound above 41000 cm -~. In the present experiments, the excitation energy is 40300 cm -1, and the almost isotropic recoil velocity distribution observed would be consistent with excitation to a predissociative state with a lifetime of the order of several HNCO ~ rotational periods before decomposition to H + NCO occurs. The measured value of fv(H + NCO) = 0.55, which is significantly higher than the 'prior' statistical value of 0.23 (calculated taking only energy conservation into account [33]), could indicate that the predissociation proceeds via a barrier in the H - N C O exit channel, leading to the observed non-statistical product energy partitioning.

Acknowledgements

The authors would like to thank Professor F.F. Crim and Dr. S.S. Brown for providing details on the method of HNCO synthesis, Professor B. Schramm and Dr. B. Kopff for their help in the HNCO FT-IR analysis. RKV and RAB wish to acknowledge fellowships provided by the DLR Bonn under the Indo-German bilateral agreement (Project No. CHEM-19) and by the HCM programme (Contract No. CHRK-CT 94-0436) of the European Community, respectively. The authors also gratefully acknowledge financial support of the Alexander-vonHumboldt-Stiftung, the Max-Planck-Gesellschaft and the Deutsche Forschungsgemeinschaft.

References [1] T.L. Boggs, Prog. Astronaut. Aerounaut. 90 (1984) 121. [2] D.L. Siebers and J.A. Caton, Combust. Flame 79 (1990) 31. [3] M. Jodal, C. Nielson, T. Hulgaard and K. Dam-Johansen, 23rd Int. Symposium on Combustion (The Combustion Institute, Pittsburgh, PA, 1990) p. 237. [4] J.A. Miller and P. Glarborg, in: Springer series in chemical physics. Gas phase chemical reaction systems: Experiments

170

[5]

[6] [7] [8]

[9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]

R.A. Brownsword et a l . / Chemical Physics Letters 258 (1996) 164-170

and models 100 years after Max Bodenstein, eds. J. Wolfrum, H.-R. Volpp, R. Rannacher and J. Warnatz (Springer, Berlin, 1996) in press. A. Schnck, H.-R. Volpp and J. Wolfrum, Combust. Flame 99 (1994) 491; H. Lillich, A. Schuck, H.-R. Volpp, J. Wolfrum and P.-D. Naik, 25th Int. Symposium on Combustion (The Combustion Institute, Pittsburgh, PA, 1994) p. 993. J. Zhang, M. Dulligan and C. Wittig, J. Phys. Chem. 99 (1995) 7446. W. Hack and K. Rathmann, J. Phys. Chem. 94 (1990) 3636. J.W. Rabalais, J.R. McDonald and S.P. McGlynn, J. Chem. Phys. 51 (1969) 5103; W.H. Fang, X.Z. You and Z. Yin, Chem. Phys. Lett. 238 (1995) 236. S. Woo and T. Liu, J. Chem. Phys. 3 (1935) 544. R.N. Dixon and G.H. Kirby, Trans. Faraday Soc. 64 (1968) 2002. H. Okabe, J. Chem. Phys. 53 (1970) 3507. K. Uno, T. Hikida, A. Hiraya and K. Shobatake, Chem. Phys. Lett. 166 (1990) 475. B. Bohn and F. Stuhl, J. Chem. Phys. 97 (1993) 4891. R.A. Brownsword, T. Laurent, R.K. Vatsa, H.-R. Volpp and J. Wolfrum, Chem. Phys. Lett. 249 (1996) 162. T.A. Spiglanin, R.A. Perry and D.W. Chandler, J. Phys. Chem. 90 (1986) 6184. J.N. Bradley, J.R. Gilbert and P. Svejda, Trans. Faraday Soc. 64 (1968) 911. W.D. Wooley and R.A. Back, Can. J. Chem. 46 (1968) 295. W. Yi and R. Bersohn, Chem. Phys. Lett. 206 (1993) 365. M. Kawasaki, Y. Sato, K. Snto, Y. Matsumi and S.H.S. Wilson, Chem. Phys. Lett. 251 (1996) 67.

[20] S.S. Brown, H.L. Berghout and F.F. Crim, J. Chem. Phys. 102 (1995) 8440. [21] E. Quinofies, J. Chen and P.J. Dagdigian, Chem. Phys. Lett. 174 (1990) 65. [22] B. Russic and J. Berkowitz, J. Chem. Phys. 100 (1994) 4498. [23] G.W. Flynn and R.E. Weston, Ann. Rev. Phys. Chem. 37 (1986) 551. [24] S. Okada, A. Tezaki, A. Miyoshi and H. Matsui, J. Chem. Phys. 101 (1994)9582. [25] R.W. Quandt and J.F. Hershberger, J. Phys. Chem. 99 (1995) 16939. [26] J.L. Durant, Jr., J. Phys. Chem. 98 (1994) 518. [27] R.N. Zare, Mol. Photochem. 4 (1972) 1. [28] G.W. Johnston, S. Satypal, R. Bersohn and B. Katz, J. Chem. Phys. 92 (1990) 206. [29] G.N.A. van Veen, K.A. Mohamed, T. Bailer and A.E. de Vries, Chem. Phys. 74 (1983) 261. [30] G.W. Johnston, B. Katz, K. Tsukiyama and R. Bersohn, J. Phys. Chem. 91 (1987) 5445. [31] S. Koppe, T. Lanrent, P.D. Naik, H.-R. Volpp, J. Wolfrum, T. Arusi-Parpar, 1. Bar and S. Rosenwaks, Chem. Phys. Lett. 214 (1993) 546. [32] A.L.L. East, C.S. Johnson and W.D. Allen, J. Chem. Phys. 98 (I 993) 1299. [33] R.D. Levine and R.B. Bemstein, Molecular reaction dynamics and chemical reactivity (Oxford University Press, Oxford, 1987).