J. Phys. Chem. 1991, 95, 571-574
571
Photodissociation Dynamics of NH,, NH,D, NHD,, and ND,: Rovibronic Absorption Analysis of the A-X Transition -.I)
Atsushi Nakajima, Kiyokazu Fuke? Keizo Tsukamoto, Yasushi Yoshida, and Koji Kaya* Department of Chemistry, Faculty of Science and Technology, Keio University, 3- 14- I Hiyoshi, Kohoku-ku. Yokohama 223, Japan (Received: May 23, 1990; In Final Form: August 2, 1990) The photodissociation dynamics of A state ammonia molecules has been further investigated for deuterium-substituted derivatives NH2D, NHD2, and ND, measuring the homogeneous line widths on action spectra. The line widths of the 21, vibronic band were determined to be 19, 12, and 1.3 cm-I fwhm for NH2D,NHD2,and ND3, respectively. In consideration of our previously reported value of 26 cm-' for NHp, thcspectral line width decreases in proportion to the number of D atoms. This result shows that the dissociation rate in the A state is addiiive as for the number of H atoms, and it is explained by the fact that the zero-point energies on the potential surface of A state are equally shifted among deuterium-substituted ammonias.
1. Introduction Detailed knowledge about the potential energy surfaces in the electronic excited states is needed in order to understand the ehotodissociation dynamics. In the first electronically excited A'AF state, NH3 predissociates to NH2 (amidogen radical) and HI-5 NH3(g1A,) Lk, NH,(AIA;)
-
NH2(g2BI)+ H(2S)
(1)
but little is known about the mechanism and the dynamics of the ammonia dissociation. The equilibrium geometry of ammonia is planar in the A state, but this A state has a shallow barrier of height ca. 3000 cm-',to repulsive state as one NH bond is stretched to give H NH2(X) at the asymptote. The predissociation rate and the product state distribution are influenced by the vibrational level in Y; and the rotational level in parent molecules.68 Particularly in the predissociation of A state ammonia molecules with Y; 1 2, it has been reported that anharmonic coupling between bending and stretching motions is important. It has also been pointed out that both centrifugal and Coriolis forces may influence the dissociation rate and the eventual product energy disposal and that the relative contribution of the two forces would be dependent upon rotational angular momentum with which the A state ammonia molecules are prepared.2.6.9 Recently, sever$ research groups have studied the homogeneous line width of the A-2 transition to estimate the dissociation rate using various experimental m e t h o d ~ ~ ~ We * ~ have * l ~ previously reported the action spectra of NH2(X) generation in the vicinity : and 21, rovibronic bands of the A-% transition under a of the 0 free jet expansion condition. The advantage of the free jet experiment is to exclude the rotational level dependence because the technique can make the ammonia molecule populate in the lowest possible quantum level. The homogeneous line widths have been estimated as 32 f 2 and 26 f 2 cm-' for Y; = 0 and 1 bands, respectkel I I Then, we examined the homogeneous line widths of the A-$ absorption spectra of all the hydrogen isotope derivatives, Le., NH3, NH2D, NHD2, and ND3, in the 21, rovibronic band in order to know the dependence of isotopic composition. With the number of D atoms in the ammonia, the zero-point energy on potential energy surfaces is lowered by the mass effect and the potential barrier to predissociation increases. The increase in the potential barrier will lead to the decrease in the tunneling rate and will be reflected by the spectral line width, and then we can gain detailed insight into the dissociation dynamics.
+
2. Experimental Section The supersonic jet apparatus used in the present study was similar to that reported previously.I2 Jet-cooled NH,, NH2D, NHD2, and ND3 were photolyzed at 12 mm downstream from 'Present address: Institute for Molecular Science, Mycdaiji, Okazaki 444, Japan.
0022-3654/91/2095-057 1$02.50/0
the nozzle of 400-pm diameter. The light for the photolysis was the second harmonic of a dye laser pumped by a XeCl excimer laser (Lambda Physik FL 3002/LPX 105) in the 213-208-nm region using Bis-MSB dye (Lambda Physik). The second harmonic was generated by a SHG crystal (8-BaB20,), using an autotracking device ( h a d Autotracker 11). The produced NH2, NHD, and ND2 fragments (amidogen radical) were probed by a second dye laser (Lambda Physik FL3002) pumped by the @me XeCl excimer laser, whose wavelength was tuned to the A-X electronic transition of these fragments in the range 600-620 nm using R610 dye (Exciton). The outputs of both the photolysis and probe laser were carefully attenuated and were kept at 5-8 and 2-1 5 pJ/pulse, respectively, in order to avoid power broadening. The line width of the photolysis laser was estimated to 0.35 cm-l by the measurement of the single rotational line width of the NO(A-X) laser-induced fluorescence (LIF) spectrum. The LIF from the fragments in the A state, collected through a lens and a filter (TOSHIBA R-64), was detected by a photomultiplier tube (Hamamatsu R562). The signal from the photomultiplier was fed into a preamplifiq and processed by a boxcar integrator (Stanford Research Systems SR-250). The spectra were not only recorded on a strip chart recorder but also stored in a microcomputer (NEC PC-9801) with a computer interface module (Stanford Research Systems SR-245). The signal intensity was observed to be linearly proportional to both the laser powers of the photolysis light and the probe light. The following timing sequence was typically employed in the measurement: A trigger from a digital pulse generator (Stanford Research Systems DG535) opened the nozzle. When the first 200-fi portion of the pulsed molecular beam came to the photolysis volume, the photolysis laser was fired and the fragments of NH2 or NHD or ND2 were formed. The probe laser was shot with a delay time of 5 ns after the photolysis. The laser-induced fluorescence was detected with a 3-ps gate, which was coincidentally opened with the probe laser pulse. The repetition rate was 10 Hz. (1) Douglas, A. E. Discuss. Furuduy Soc. 1963, 35, 158. (2) Ashfold, M. N. R.;Bennett, C. L.; Dixon, R. N. Chcm. Phys. 1983, 93, 293. (3) Xit, J.; Sha, G.; Zhang, X.; Zhang, C. Chcm. Phys. L i t . 1986,124, 99. (4) Ashfold, M. N. R.: Bennett, C. L.; Dixon. R. N. Faraday Discuss. Chem. Soc. 1986,82, 163. (5) Vaida, V.; McCarthy, M. I.; Engelking, P. C.; Rosmus, P.; Werner, H. J.; Botschwina. P. J . Chem. Phys. 1987,86,6669. (6) Biesner, J.; Schnieder, L.; Schmeer, J.; Ahlers, 0.;Xie, X.: Welge, K. H.; Ashfold, M. N. R.;Dixon, R. N. J . Chcm. Phys. 1988,88, 3607. (7) Biesner, J.; Schnieder, L.; Ahlers, G.; Xie, X.; Welge. K. H.; Ashfold, M. N. R.; Dixon, R. N. J . Chcm. Phys. 1989, 91, 2901. (8) Dixon, R. N. Mol. Phys. 1989, 68, 263. (9) Ziegler, L. D. J . Chem. Phys. 1987,86, 1703. (IO) Ziegler, L. D. J . Chem. Phys. 1985, 82. 664; 1986, 84, 6013. (1 1) Fuke, K.; Yamada, H.; Yoshida, Y.; Kaya, K. J . Chem. Phys. 1988 88, 5238. We have previously reported the line width of NH, in A-2 vibronic transition to 30 i 2 cm-I, but after precise analysis it has been revealed to 26 i2 c d . (12) Fuke. K.; Ozawa, K.; Kaya, K. Chcm. Phys. L r r . 1986, 126, 119.
4
0 1991 American Chemical Society
572 The Journal of Physical Chemistry, Vol. 95, No. 2, 1991
Nakajima et al.
Ro(l)+Ri(O)
Qi(V
211.6
211.5
211.4
WAVE LE NGTH/nm
Figure 3. Action spectrum of NHD2 in the A-g 21, rovibronic transition for monitoring the LIF of ND2 fragment: observed (-) and simulated 1
I
211.1 211.0 WAVELENGTH/nm Figure 1. Action spectrum of ND3 in the A-R 21, rovibronic transition
for monitoring the LIF of ND2 fragment: (a) observed and (b) simulated spectra. In the best fit spectrum, the rotational temperature and the line width were 4 K and I .3 cm-I, respectively.
(- - -) spectra. In the best fit spectrum, the rotational temperature and the line width were 8 K and 12 cm-I, respectively.
TABLE I: Rotational Constants (em-’) for NHzD and NHDz
soecies NH2D NHD2
ground state’ E” C”
A‘
excited state E’
C’
9.3 (5) 6.2 (5)
5.6 (4) 5.0 (3)
3.5 (2) 2.8 (2)
A”
9.68 7.45
6.41 5.34
4.70 3.76
“References 16-18. TABLE II: Line Widths in Various Isotopic Ammonias for the 2: Transition
J’
1
212.2
1
1
.
1
.
1
.
212.0 WAVELE NGTH/nm
1
211.8
Figure 2. Action spectrum of NH2D in the A-2 2; rovibronic transition for monitoring the LIF of N H D fragment: observed (-) and simulated (- - -) spectra. In the best fit spectrum, the rotational temperature and
the line width were 5 K and 19 cm-I, respectively.
3. Results Individual rovibronic levels of NH3(A) show substantial broadening on accouzt of their rapid predissociation. However, the 2l level of ND3(A) predissociates at a sufficiently slow rate that in_dividualrovibronic transitions are resolved in the jet-cooled ND3(A-X) 2; absorption band. The action spectrum is shown in Figure I , which is obtained by scanning the photolysis laser wavelength and monitoring the excitation spectrum for ND2 fragment formation. The ND2 was probed at 16511.1 cm-I, which was tentatively assigned to the R branch of the 0150-0202 vibronic transition on the basis of the reported spectroscopic constants.1°J3 The spectrum can be simulated ,by employing the appropriate spectroscopic constants for both X and A states together with appropriate Honl-London factors and ground-state nuclear spin statistics. The nuclear spin statistics is derived from nuclear spin degeneracy of 10 for AI’, 1 for A i , and 8 for E” in lower inversion level.I4 In the simulation, t h e rotational temperature and line width of ND3 were determined as parameters. The result indicated that a rotational temperature of ND3 was 4 K in the molecular beam and the homogeneous line width was 1.3 f 0.2 cm-l fwhm in the excited state. Figu~es-2and 3 show the action spectra of NHzD and NHD2 in the A-X 2; absorption band, respectively. The spectra were obtained by monitoring the laser-induced fluorescence of NHD and ND2 fragments, respectively, because the dissociation rate of H atom is several times larger than that of the D atom in the (13) Dressler, K.; Ramsay, D. A. Philos. Trans. R. Soc. London 1959.69, 251. (14) Bunker, P. R. Moleculor Symmetry and Spectroscopy; Academic Press: New York, 1979.
species
line width,” cm-’
ratiob
NH, NH2D NHD2 N D3
26 (2) 19 (2) 12 (2) 1.3 (2)
0.73 0.63 0.1 1
“Line widths are fwhm. *Decreasing ratio of line width in an exchange of H atom for D atom.
photolysis of NH2D and NHD2.15 ND2(W) produced in the photolysis of NHD2 was probed at the_sa-me vibronic transition as the photolysis of ND3. The NHD(A-X) transition has never been analzzed in detail; therefore, we measured the LIF spectrum of NHD(X) and the probe laser was set at 16501.0 cm-I, which was tentatively assigned to the R branch of the 0140-020 vibronic transition. The spectra are simulated in order to determine the rotational tempercture and the line width by using rotational constants for the X’AI state of NH2D and NHD2I6I8 and the A’A2’’ state of NH3 and N_D3 and the program BC3.I9 The rotational constants for the A state of NHzD were derived from the known planar equilibrium configuration assuming an N-H(D) bond length of 1.1 A. The rotational constants of NHD2 were derived from almost the same procedure except for slight adjustment of A’to obtain the best fit spectrum. The adjustment is consistent with the v i dependence of rotational constants B‘ of NH3 and ND3,I0which decreases with v i vibrational quantum number. The rotational constants of NH2D and NHD2 are listed in Table I. In the simulation nuclear spin statistia and the mixing of two rotational transition types should be taken into account; c type (0.78)/a type (0.22) for NH2D and c type (0.77)/b type (0.23) for NHD2.I6-l8 Both the rotational temperatures and homogeneous line widths of NH2D and NHD2 in the 2; vibronic transition were determined: 5 K and 19 f 3 cm-l for NH2D, and 8 K and 12 f 3 cm-l for NHD2. The results show the line width decreases with the number of D atoms in ammonia molecules. Nakagawa, T.; Overend, J. J . Mol. Spectrosc. 1974, 50, 333. (16) Cohen, E. A.; Pickett, H.M.J . Mol. Spectrosc. 1982, 93, 8 3 . (17) Coudert, L.; Valentin, A.; Henry, L. J . Mol. Spectrosc. 1986, 120, (15)
185. (18)
Job, V. A.; Kartha, S. B.; Kartha, V. B.; Thakur, K. B. J . Mol.
Spectrosc. 1986, 120, 205.
(19) Koda, S.;Back, R. A. Can. J . Chem. 1977, 55, 1380.
Photodissociation Dynamics of NH3, NH2D, NHD2, and ND3 TABLE IIi: Vibrational Frequencies-Pad Energies from the Potential Minimum of YI) ia tbe First Excited A Stote
species
frcq? cm-' interval 890 (IO) NH3 80 NH2D 810 (IO) 75 NHD2 735 (IO) 75 N D3 660 (IO) Each vibrational frequency is calculated by subtraction of 0 : band transition frequency from 21, transition frequency. u2'
4. Discussion The spectral line width decreases as proportional to the number of D atoms, including our previously reported value of 26 cm-' for NH3 (see Table 11). This result shows that the dissociation rate in the A state is additive as for the number of H atoms. The decreasing ratio of line width in an exchange of H atom for D atom is also shown in Table 11. The predissociation mechanism of ammonia in the state has been known to be the potential tunneling. The narrowing of line width by the increment of the number of D atoms is explained by the following two factors: (i) thejifference of the zero-point energies on the potential surface of A state, which are shifted to the lower energy side in the exchange of H atom for D atom, and (ii) the number of N H bonds for each ammonia molecule. When a vibrational level is energetically lowered by the substitution of a H atom for a D atom, the potential width for tunneling increases, which would result in the decrease of the dissociation rate. Moreover, the line width also depends on the number of bonds that can easily break. For example, in NH2D there are two ways of breaking an N H bond to give H N H D but only one way in NHD2 to give H ND2. Except for ND3, the line widths of the other ammonias correspond to the rate of the dissociation into amidogen radical and H atom. Only the line width of ND3, however, corresponds to the rate of the dissociation into amidogen radical and D atom. The mass difference between H and D atoms will seriously affect the tunneling rate. Indeed the decreasing ratio of the line width (0.1 1) between NHD2 and ND3 is extremely different from the other two (ca. 0.7). In the dissociation of ND3, both the increase in the potential width and the mass difference between H and D atoms can act to decrease the dissociation rate. 4.1. Energy Interval of the v i = 1 Vibrational Level. The rate of dissociation is established by the potential tunneling through a barrier at which the greatest bond displacement from equilibrium is only 0.27 A. For moderate displace-ments from the equilibrium the potential energy surface for the A state can be described as the sum of almost independent contributions from one out-of-plane coordinate and from five in-plane coordinates. As yet it is not feasible to carry out a detailed quantum mechanical treatment with full dimensionality, but a one-dimensional cut along the NH2 H dissociation coordinate can yield semiquantitative informatiom20 As discussed above, the line width is affected by the potential width for tunneling and the mass difference between H and D atoms. On the one-dimensional potential surface along the dissociation coordinate, the energy interval of the Y; = 1 vibrational levels of all the hydrogen isotope derivatives is the summation of two factors; one is the difference of v i vibrational frequency, and the other is energy difference of the zero-point vibration along N-H(D) dissociation coordinate. The former energy interval can be derived by subtraction of 0: band transition frequency from the 21, transition frequency (see Table 111). Evidently as shown in the table, the Y; vibrational frequency varies from 890 to 650 cm-' in the various isotopic compositions at almost regular intervals (ca. 80 cm-I). The latter is estimated to ca. 100 cm-l among deuterium-substituted ammonias in the following procedures: The frequency for the stretching vibration (ul' or u3') of ammonia in the A state has been reported to be 2400-2500 cm-I for NH3 and
A
+
+
+
(20) McCarthy. M. I.; Rosmus, P.; Werner, H.-J.; Botschwina. P.; Vaida, V. J . Chem. Phys. 1987,86, 6693.
The Journal of Physical Chemistry, Vol. 95, No. 2, 1991 573 1900-2000 cm-' for ND3.2' On the potential surface for stretching one N-H(D) bond zero-point energy is half of the frequency of the vibration. Then, the zero-point energy difference of the vibration is ca. 300 cm-' between NH3 and ND3. The interval among individual isotopic ammonias, for example between NH3 and NH2D, is estimated to be ca. 100 cm-'on the assumption of an equal energy interval. Then, the energy interval of Y; = 1 vibrational level Cmong isotopic ammonias can be estimated to ca. 180 cm-I on the A state potential surface; our result shows that the line width decreases to about two-thirds for every 18O-~m-~-reductionof vibrational levels around this energy region on the A potential surface. This ratio is almost consistent with the ratio reported by Dixon:* one-half for every reduction of 230 cm-I. 4.2. The Number of NH Bonds. In a unimolecular dissociation the line width corresponds to the dissociation rate, and the dissociation rate might be expected to increase linearly with the number of N H bonds because of large H / D tunneling ratio. Under this assumption, the line width of NH2D should be at least twice as large as that of NHD2, because in NH2D there are two ways of breaking an N H bond but only one way in NHD2; in addition, the potential width of NH2D for tunneling is smaller than that of NHD2 The measured line width of NH2D, however, is only 1.5 times larger than that of NHD2. This is the same case between NHD2 and NH3. This result shows that the number of N H bonds does not considerably affect the line width though it should be taken into account. By the precise examination, the decreasing ratio of line width between NH3 and NH2D differs from that between NH2D and NHD2 (see Table 11). If only the shift of energy levels would determine the line width, the ratio should be almost the same value. This difference is likely to be reflected by the number of N H bonds. The numbers of the N H bonds of NH3, NH2D, and NHD2 are 3, 2, and 1, respectively. Therefore, the ratio should decrease more efficiently between NH2D and NHD2 than between N H 3 and NH2D, because the decrement of the number of N H bonds is more drastic between NH2D and NHD2 (from two bonds to one bond). Actually, the obtained decreasing ratio (0.63) of line width between NH2D and NHD2 is more drastic than that (0.73) between NH3 and NH2D. 4.3. The Mass Effect for the Potential Tunneling. In ND3, the measured line width corresponds to the rate of the dissociation into the amidogen radical and D atom. Then, the line width is affected by the mass difference between H and D atoms for the potential tunneling. As mentioned above, the energy shift of vibrational level due to an H/D exchange results in the decrease in line width to two-thirds. Namely, when a H atom dissociated in the Y; = 1 level of ND,, the line width would become 8 cm-' in consideration of only the energy level shift, that is, two-thirds of the line width of NHD2 (12 cm-I). However, the line width estimated from the above consideration is still 6 times larger than that of ND3, and this difference corresponds to the mass effect between H and D atoms. In a unimolecular dissociation, the dissociation rate can be represented by a reciprocal of a lifetime. The lifetime, 7 , would be varied by mass effect and potential width on the potential tunneling and is given approximately by22
where Y is the classical vibration frequency and the integration is carried out between the two values of r at which V = E,,,. When comparing two isotopic species XH and XD, the mass difference will affect Y and the factor p1I2in the exponent, both of which will lead to a shorter lifetime for the lighter isotope species. When comparing energetically different quantum levels, potential width through tunneling in the higher energy level will lead to a shorter lifetime. (21) Dixon, R. N. Chem. Phys. Leu. 1988,147, 377. (22) Bell, R. P. The Tunnel Effecr in Chemistry; Chapman and Hall: London, 1980.
J . Phys. Chem. 1991, 95, 514-518
514
The mass difference between H and D atoms seriously affects the tunneling rate, and the amount can be approximately estimated by eq 2. When the vibration between N and H / D atoms is assumed to be a harmonic oscillator, the mass effect for the line width between H and D atoms is estimated by the factor T ( H / r D
rH/rD
(mD/mH)1/2exp[ (mD’/’ - mH’/’) x (“(V-
€LX)l’/’
dr]] ( 3 )
in consideration of both the vibrational frequency and the factor F ~ in/ eq~ 2. Here, mDand mH are the mass weight of D and H atoms, respectively. This factor corresponds to the mass effect
for the potential tunneling between H and D atoms. At the ~ 2 ) = 1 vibrational level of ND, the amount of the factor (3) is about 6, so that the exponent factor amounts to 4.2. This value should be connected with the potential energy function, V(r),and a similar analysis for the 0; and 2; transitions can reveal details of the dissociation mechanism on the potential energy surface. The analysis for the 08 and 2; transitions and an ab initio calculation are in progress in our group.
Acknowledgment. The authors are grateful to Professor S. Iwata for valuable discussion. The financial support from Grant-in-Aid for Scientific Research for Priority Area by the Ministry of Education is greatly acknowledged.
Femtosecond Infrared Spectroscopy: Ultrafast Photochemistry of Iron Carbonyls in Solution Philip A. Anfinrud: Chul-Hee Han, Tianquan Lian, and Robin M. Hochstrasser* Department of Chemistry, University of Pennsylvania, Philadelphia, Pennsylvania I9104-6323 (Received: May 18, 1990)
The photochemistry of cyclopentadienyliron dicarbonyl dimer in cyclohexane has been investigated with a novel femtosecond infrared spectrometer. Visible (580 nm) photolysis proceeds through an excited bound state that persists for about 1 ps. Dissociation occurs through more than one channel and is suggested to generate singly bridged and/or nonbridged dimers and homolytically cleaved species. Solvation of the nascent fragments apparently begins around 5 ps and nears completion around IO ps. This delay in solvation may arise because visible excitation generates photoproducts with sufficient excess energy to inhibit complexation with the solvent at earlier times. Roughly half of the photodissociated dimers recombine geminately within =20 ps, presumably through a singly bridged intermediate. Therefore, singly bridged photoproducts would not participate in photoinitiated bimolecular chemical reactions with diluted species.
Introduction We recently introduced a new method for obtaining IR spectra with femtosccond time reso1ution.l This development has enabled an ultrafast investigation into the photochemistry of [CpFe(CO),], (Cp = q5-C5H5)in cyclohexane where the primary photoproducts and their dynamics have been observed for the first time. Although the photochemistry is complex, the time-resolved IR spectra are sufficiently detailed to illuminate some of the primary photochemical reaction pathways. Coordinatively unsaturated organometallic complexes are thought to be reactive intermediates in homogeneous c a t a l y s i ~ . ~ ? ~ Such species are readily generated upon photolysis of transition-metal carbonyls. Hence, characterization of photochemical reaction pathways for saturated organometallic species is particularly relevant to the understanding of homogeneous catalysis. The photochemistry of dinuclear transition-metal carbonyls has been extensively studied by numerous IR techniques. Near-UV irradiation of [CpFe(CO)z]2 in an organic matrix generated only one photoproduct, which was identified through its IR spectrum as a CO-loss product with three bridging CO bond^.^.^ A time-resolved IR investigation in cyclohexane reported a second intermediate, the homolytically cleaved dimer! It was found that visible excitation provides sufficient energy for homolytic cleavage, but near-UV photolysis is required to form the triply bridged intermediate. While microsecond time-resolved IR spectroscopy has contributed greatly to our understanding of photoinitiated diffusion-controlled chemistry,H this technique is unable to address issues related to the primary events in the photodissociation ‘Present address: Department of Chemistry, Harvard University, 12 Oxford St., Cambridge, MA 02138.
of organometallic compounds. For example, what excited states are involved, what is the quantum efficiency for photodissociation, what are the nascent fragments, how might the solvent affect the partitioning among available reaction pathways, is there any geminate recombination, and what is the mechanism and rate of solvation? This problem was recently addressed in this laboratory by using transient IR spectroscopy with -30-ps time resolution.I0 While these picosecond spectra provided evidence for new intermediates, not even that time resolution was sufficient to characterize the primary processes. The extension of IR spectroscopy into the femtosecond regime has made the current investigation possible. ~
~~~
(1) Anfinrud, P. A.; Han, C.; Hanscn, P. A.; Moore, J. N.; Hochstrasscr, R. M. In Ultrafast Phenomena; Springer: Berlin, 1988; Vol. VI, p 442. (2) Geoffroy, G. L.; Wrighton, M. S. Organometallic Photochemistry; Academic Press: New York, 1978. (3) Parshall, G. W. Homogeneous Catalysts: The Application and Chemistry of Catalysis by Soluble Tramitlon Metal Complexes; Wiley: New York. 1980. - ----, -- -(4) Hooker, R. H.; Mahmoud, K. A.; Rest, A. J. J . Chem. Soc., Chem. Commun. 1983, 1022. ( 5 ) Hepp, A. F.; Blaha, J. P.; Lewis, C.; Wrighton, M. S.Organometallics 1. 17A -lWd - - ., -, - . .. (6) Moore, B. D.; Simpson, M. B.; Poliakoff, M.; Turner, J. J. J . Chem. Soc., Chem. Commun. 1904, 912. ( 7 ) Dixon, A. J.; Healy, M. A.; Hodges, P. M.; Moore, B. D.; Poliakoff, M.; S i m p n , M. B.; Turner, J. J.; West, M. A. J. Chem. Soc.,Fmday Tram. 2 1986,82,2083. (8) Moore, B. D.; Poliakoff, M.; Turner, J. J. J . Am. Chem. Soc. 1986. 108, 1819. (9) Dixon, A. J.; Hcaly, M. A,; Poliakoff, M.; Turner, J. J. J. Chem. Soc.. Chem. Commun. 1986, 994. (10) .Moore, J. N.; Hansen, P. A.; Hochstrasser, R. M. J . Am. Chem. Soc. 1989, 1 1 1 , 4563.
0022-3654191 /2095-0514302.50/0 0 1991 American Chemical Society
