1604
J . Phys. Chem. 1990, 94, 1604-1607
Mechanisms of the Reaction of Unsaturated Organic Halides with Small Gas-Phase Vanadium Clusters Li Song, John E. Freitas, and M. A. El-Sayed* Department of Chemistry and Biochemistry, University of California. Los Angeles, Los Angeles, California 90024- I569 (Received: June 21, 1989; In Final Form: September 14, 1989)
Small gas-phase vanadium clusters synthesized in a molecular beam system by a laser vaporization technique are reacted with I-bromopropene, 3-bromopropene, 3-chloropropene, and propene in a fast-flow reactor, and the analysis of the products was made by laser ionization-time-of-flight mass spectrometry. The results observed can be explained in terms of three different mechanisms for the reaction of halopropene: halogen abstraction, halogen substitution followed by molecular hydrogen evaporation, and molecular addition followed by evaporation of molecular hydrogen and hydrogen halide. The addition reaction followed by molecular hydrogen evaporation can account for the reaction involving propene. The dependence of the relative importance of the three different mechanisms on the vanadium cluster size and the structure of the halopropene is discussed.
Introduction The reactivity of gas-phase metal clusters as a function of cluster size has been the subject of many recent studies.]-” Successful combination of laser vaporization, supersonic expansion, and a fast-flow reactor yields rich information on the relative reactivity and the stability of neutral gas-phase metal clusters. By studying the branching ratios of two possible pathways in the reaction of niobium clusters with some reagent (Le., BrCN), it was possible to examine the dependence of stereospecificity of the reaction on cluster sizeGZ1The loss of the stereospecificity on large clusters provides some insight into the mechanisms of such reactions.
(1) Morse, M. D.; Geusic, M. E.; Health, J. R.; Smalley, R. E. J . Chem. Phys. 1985, 83, 2293. (2) Hamrick, Y.; Taylor, S.; Lemire, G. W.; Fu,Z.-W.; Shui, J.-C.; Morse, M. D. J. Chem. Phys. 1988,88, 4095. (3) Geusic, M. E.; Morse, M. D.; Smalley, R. E. J . Chem. Phys. 1985,82, 590. (4) Zakin, M. R.; Cox, D. M.; Whetten, R. L.; Trevor, D. J.; Kaldor, A. Chem. Phys. Lett. 1987, 135, 223. ( 5 ) Alford, J. M.; Weiss, F. D.; Laaksonen, R. T.; Smalley, R. E. J . Phys. Chem. 1986, 90, 4480. (6) Elkind, J. L.; Weiss, F. D.; Alford, J. M.; Laaksonen, R. T.; Smalley, R. E. J. Chem. Phys. 1988,88, 5215. (7) Whetten, R. L.; Zakin, M. R.; Cox, D. M.; Trevor, D. J.; Kaldor, A. J . Chem. Phys. 1986,85, 1697. (8) Richtsmeier, S. C.; Parks, E. K.; Liu, K.; Pobo, L. G.; Riley, S. J. J. Chem. Phys. 1985,82, 3659. (9) Parks, E. K.; Liu, K.; Richtsmeier, S.C.; Pobo, L. G.; Riley, S. J. J. Chem. Phys. 1985, 82, 5470. (IO) Liu, K.; Parks, E. K.; Richtsmeier, S. C.; Pobo, L. G.; Riley, S. J. J. Chem. Phys. 1985, 83, 2882. (1 1) Parks, E. K.; Nieman, G. C.; Pobo, L. G.; Riley, S. J. J. Phys. Chem. 1987, 91, 267 1. (121 St. Pierre. R. J.; El-Saved. M. A. J . Phvs. Chem. 1987. 91, 763. (13) St. Pierre, R. J.; Chronister, E. L.; El-Sa&, M. A. J. Phys. Chem. 1981, 91, 5228. (14) Zakin, M. R.; Cox, D. M.; Kaldor, A. J. Phys. Chem. 1987,91,5224. (15) Parks, E. K.; Weiller, B. H.; Bechthold, P. S.;Hoffman, W. F.; Nieman, G. C.;Pobo, L. G.; Riley, S. J. J . Chem. Phys. 1988, 88, 1622. (16) Hoffman 111, W. F.; Parks, E. K.; Riley, S. J. J . Chem. Phys. 1989, 90, 1526. (17) Zhang, Q.L.; OBrien, S. C.; Heath, J. R.; Liu, Y.; Curl, R. F.; Kroto, H. W.; Smalley, R. E. J . Phys. Chem. 1986, 90, 525. (18) Zakin, M. R.; Brickman, R. 0.; Cox, D. M.; Kaldor, A. J . Chem. Phys. 1988, 88, 3555. (19) Kaldor, A.; Cox, D. M.; Zakin, M. R. A h . Chem. Phys. 1988, 70, 21 I . (20) Cox, D. M.; Reichmann, K. C.; Trevor, D. J.; Kaldor, A. J . Chem. Phys. 1988.88, 11 1. (21) Song, Li; Eychmuller, Alexander; El-Sayed, M. A. J. Phys. Chem. 1988, 92, 1005.
0022-3654/90/2094-1604$02.50/0
It was reported earlier that a r-bond in the reagent molecule is often necessary to initiate an addition type r e a ~ t i o n . ’ ~ - ’For ~*~~ some of the organic compounds studied, this addition product can undergo varying degrees of dehydrogenation, which depends on the relative stabilities of the bare metal cluster, the reagent molecule, and the final product. Most recently, the reactions of niobium clusters with some organic bromidesz3 and carbon dioxide24were studied in detail. In the reaction of Nb, with carbon dioxide, a strong oxygen abstraction product was observed on small clusters. As the cluster size increases, molecular C02 addition product was observed to become dominant and the oxygen abstraction efficiency dropped. This was explained by proposing that exothermic oxygen abstraction takes place for all clusters. For small clusters, C O evaporation takes place in order to cool the cluster product, but as the cluster size increases, evaporation becomes less probable.24 In the study of the reaction of Nb, with some saturated and unsaturated organic bromides, it was found that only a bromine abstraction product was observed for the saturated bromides, while both bromine abstraction and addition-dehydrogenation reaction products were observed for the unsaturated bromides. The lack of correlation between the bromine abstraction and additiondehydrogenation reaction leads23to the conclusion that they were formed by two different types of collisions. In this present study, the reaction between vanadium clusters and propene and a few of its haloderivatives was examined. The possible mechanisms accounting for the observed results are outlined, and the change in the relative importance of each is explained in terms of the structures of the reactants and the size of the cluster. Experimental Section The experimental details have been described e l s e ~ h e r e . ~ ~ ? ~ ~ Briefly, vanadium clusters were synthesized by laser vaporization (6 mJ per pulse of 355-nm laser light from a Quanta-Ray Model DCR-1A running a t 10 Hz) of a solid vanadium rod in a pulse of high-purity helium (99.999%, Spectra Gases). The metal atom plume formed upon each laser shot was entrained and quenched in the helium pulse where condensation-nucleation results in cluster formation. The mixture of vanadium clusters with helium carrier gas then was expanded into a fast-flow reactor where a pulse of reactant seeded in high-purity helium was injected. The (22) St. Pierre, R. J.; Chronister, E. L.; Song,Li;El-Sayed, M. A. J. Phys. Chem. 1987, 91, 4648. (23) Song, Li; El-Sayed, M. A. Chem. Phys. Lett. 1988,152, 281. (24) Song,Li; Eychmuller, Alexander; St. Pierre, R. J.; El-Sayed, M. A. J . Phys. Chem. 1989, 93, 2485. (25) Geusic, M. E.; Morse, M. D.; OBrien, S. C.; Smalley, R. E. Rev.Sci. Instrum. 1985, 56, 2123.
0 1990 American Chemical Society
Reaction of Organic Halides with Vanadium Clusters
The Journal of Physical Chemistry, Vof. 94, No. 4, 1990 1605 -
t
1-Bromopropene
(c)
(c) 3-Chloropropene
I
t t-
(b) 3-Bromopropene
t (
cn
z
'II
(b) 3-Bromopropene
I
W
t-
z H
II
(a) 1-Bromopropene
(0)
Propene 0
I
0
XX: X:
0
70
140
210 280 350 420 490 130
MASS ( A T O M I C UNITS) Figure 1. Comparison of the reaction product distribution (V,R and V,X) of vanadium clusters (V,) and different halopropenes: (a) Mass spectrum after reacting with 1-bromopropene. VC3H3has almost vanished while VBr is relatively strong under similar conditions to that in (b). All products were weaker than in (b). The species of interest are labeled V,, V,R, and V,Br for the bare clusters, the alkyl abstraction products, and the bromine abstraction products, respectively. (b) Mass spectrum after reacting with 3-bromopropene. For x = 1, strong VC3H3 was observed. For x = 2, both V&H3 and V&H were observed. For x = 3, both V3C3and V3C3Hwere observed. As x increases, the resolution was lost. (c) Mass spectrum after reacting with 3-chloropropene. For x = 1, both VC3H3and VCI were observed and resolved. No attempt was made to analyze the other products quantitativelydue to the overlap of V,CpH and V,CI.
mixture of the vanadium clusters and their reaction products was ionized with an unfocused 193-nm (6.42-eV) ArF excimer laser beam (Lambda Physik, EMG 101), and the ions were detected by a 1.7-m time-of-flight mass spectrometer.
Results Parts a, b, and c of Figure 1 show the mass spectra of vanadium clusters and their products after reacting with 1-bromopropene, 3-bromopropene, and 3-chloropropene, respectively. All spectra in Figure 1 were taken at a fluence of approximately 600 pJ/cm2. From these spectra, V,X (X = C1, Br) were observed for all cluster sizes and for all the organic halides used. Figure 1b shows a very strong peak of VC3H3. VBr is also relatively strong under our experimental conditions. In case of the vanadium dimer, both V2C3H3and V2C3H are observed, indicating increased dehydrogenation on the dimer over that on the monomer. On clusters with x L 3, the hydrocarbon products observed are V,C3H and V,C3, suggesting increased ability for the dehydrogenation process. The first strong peak after each V, is V,O which was formed during the vaporization process. The peak in front of each V,O is V,C. This is basically due to the carbon impurity or contamination in our system since it was observed even without any reactant present. At x L 5, a relatively strong VxC2peak is observed, suggesting the loss of CH3 and H, after halogen substitution or the loss of CH4 and H X after molecular addition. Also, peaks of V,COz and V,03 begin to appear. These products could result from impurities such as O2or C 0 2 in the inlet line of the reactor, or
167
204
241
278
315
M A S S (ATOMIC UNITS) Figure 2. Expanded parts of mass spectra resulting from the reaction of V, with propene (a), 3-bromopropene (b), and I-bromopropene (c). Notice in the small cluster region that V,C3 and V,C3H (or VxC3H2)are
well-resolved. The separation between one mass unit and two mass units is easy to distinguish. The reaction of V, with propene produces hydrocarbon-V, products with an even number of hydrogens for x 2 2 resulting from addition followed by molecular hydrogen evaporation. As x increases, the efficiency of dehydrogenation increases rapidly, producing only vanadium carbides V,C3 for x 2 4. The mass peaks labeled 0 are vanadium oxides V,O.
the former could be formed from further elimination of a C2 from Vx02C3. Relatively strong V,C2 peaks were observed for the reaction of V, with all the other organic compounds for x L 5 . Figure l a shows that the peak intensity of VC3H3has almost vanished when 1-bromopropene was used, while VBr is much stronger than VC3H3. All other alkyl abstraction (or bromine substitution) products of the form V,C3H have almost disappeared as well. A relatively strong V,C3 signal is observed for x 1 3. Figure I C shows the resulting spectrum of vanadium clusters after reacting with 3-chloropropene. The hydrocarbons combined with V, are similar to those resulting from the reaction with 3-bromopropene. VC3H3is much stronger than that observed for the reaction with 1-bromopropene (in Figure la) but relatively weaker than that from the reaction with 3-bromopropene (in Figure lb). On larger clusters ( x 1 4), the mass peaks for the products V,Cl, V,C, and V,C3H begin to overlap with each other. The reaction of V, with propene (a full spectrum of which is not shown here) gives products of the form V,C3Hy (y = 0, 2, 4,6) for x L 2; Le., hydrocarbon products with zero or an even number of hydrogen atoms are observed. For x 1 4, the y = 0 product dominates, with weak partial dehydrogenation products distributed on the right side of the total dehydrogenation product. N o product peak was observed for the monomer. Parts a, b, and c of Figure 2 show an expanded part of the mass spectrum resulting from the reaction of V, with propene, 3bromopropene, and 1-bromopropene, respectively. In these expanded spectra, it is easier to compare the products of the forms V,C3, V,C3H, and V,C3Hz. Notice in Figure 2b that the reaction with 3-bromopropene produces V3C3and V3C3H while that with propene produces V3C3,V3C3HZ,and V,C3H4. Comparing spectra
1606 The Journal of Physical Chemistry, Vol. 94, No. 4, 1990
Song et al. clusters have a smaller hydrogen content. The results of the reactions with the halopropene can best be classified in terms of the following three mechanisms: 1. halogen abstraction: V,
+ C3H5X
--+
V,X
+ (C3H5)
2. halogen substitution followed by the nH2 evaporation: V,
+ C3H5X
-X
[V,C,HS]
+
VXC3Hs-2, + nH2
3. complex formation (addition reaction) followed by HX and nHz evaporation: V, 1
2
3
4
+ C3H5X
-+
[V,C,HSX]
-t
VxC3H4-2,, + nH2
+ HX
Only mechanism of the type 3 above is expected to take place for the reaction with C3H6: 4. vx + C3H.5 ---* [VxC3H6] VxC3H,+2, + nH2 +
LN ( Laser lntnesity )
Figure 3. Ionization laser power dependence of V, and VC3H3. Logarithm of the ion signal intensity normalized to the peak intensity of the vanadium monomer signal is plotted versus the logarithm of the laser intensity. From this plot, both V, and VC3H3are found to be ionized by a single-photon process. Of all the other reaction products and bare clusters, only VBr and VCI were found to require two photons for their
ionization. b and c of Figure 2 shows that while the reaction with 3bromopropene produces V3C3and V3C3H,only V3C3is produced for the reaction with 1-bromopropene. The ionization laser power dependence of V2 and VC3H3 is shown in Figure 3. In this power dependence study, the laser fluence was changed from 200 to 1700 pJ/cm2, almost over an order of magnitude. To eliminate fluctuations from spectrum to spectrum in the power dependence study, the intensity of each peak of interest was normalized to the peak intensity of the vanadium monomer. It is known that the vanadium monomer has to be ionized by a two-photon process at 193 Therefore, a plot of the logarithm of the normalized peak intensity versus the logarithm of the laser fluence will give a straight line with a slope of n - 2 . (n is the number of photons required for the ionization process of the cluster products.) If a slope of -1 is obtained, the ionization is achieved by one photon. From Figure 3, it can be seen that both V, and VC3HI are ionized by a single-photon process. The above observation for the ionization potential of V2 is consistent with the one reported by Cox et aL2' From the power dependence of the other products, it was concluded that both VBr+ and VCI+ are observed as a result of the absorption of two photons (having a slope of zero in a logarithm-logarithm plot) while the rest of the products require only one photon at 193 nm.
Possible Mechanisms and Discussion The main important results of the reaction with the halopropene can be summarized as follows: (a) The formation of V,Br or V,CI is observed for all clusters. (b) V,C3 is observed for x I 3. (c) For the reaction with 1-bromopropene, no strong hydrogen-containing product peaks are observed while 3-bromopropene gives V,C3H, where y is 0, 1, or 3 with y = 0 and 1 dominating for x I 4. (d) In the reaction with propene, V,C3 and V,C3H, are observed for x L 3 withy = 0, 2 , 4, and 6. (e) In the reaction with all of the propene and halopropenes, the products of larger (26) (a) Moore, C. E. Analyses of Optical Spectra; NSRDS-NBS34; Office of Standard Reference Data, National Bureau of Standards: Washington, DC, 1970. (b) Herzbcrg, G. J. Mol. Spectrosc. 1970, 33, 147. (c) Gaydon, A. G. Dissociation Energies and Spectro of Diatomic Molecules, 3rd ed.; Chapman and Hall: London, 1968. (d) Ferguson, K.C.; Okafo, E. N.; Whittle, E . J . Chem. Soc., Foraday Trans. I 1973, 69, 295. (e) Morrison, R. T.; Boyd, R. N. Organic Chemistry, 2nd ed.; Allyn and Bacon: Boston, 1966. (27) Cox, D. M.; Whetten, R. L.; Zakin, M. R.; Trevor, D.J.; Reichman, K. C.; Kaldor, A. Advonces in Laser Science-I; Stwalley, W. C., Lapp, M., Eds.; American Institute of Physics: New York, 1986.
A distinction between mechanisms 2 and 3 for the reactions with halopropenes can be made from the number of the hydrogen atoms on the cluster products. Mechanism 2 produces an odd number of hydrogen atoms while mechanism 3 produces an even number or zero hydrogen atoms. Mechanism 4 produces, as observed, an even number or zero hydrogen atoms on the cluster product. The halogen abstraction reaction (mechanism 1) seems to occur for all the clusters with little sensitivity to cluster size. However, both mechanisms 2 and 3 seem to be sensitive not only to the cluster size but also to the position of the halogen with respect to the C=C double bond. The formation of VxC3 (via mechanism 3 from halopropene and 4 from propene) seems to dominate for 1-bromopropene and propene. These occur as follows:
-
V3C3 + 3H2 AH, (1) V, + 1-BrC3H5 V3C3 + 2H2 + HBr AHz (2) Reaction 2 is about 20 kcal/mo126b*cmore exothermic than reaction 1. This might explain the fact that the observed total dehydrogenation in reaction 2 is more complete than that in reaction 1. 3-Bromopropene and 3-chloropropene react with V, via both mechanisms 2 and 3 for x L 3 but via only mechanism 2 for smaller clusters. This is understandable since for larger clusters the probability of having an addition (resulting from a sticky type collision) becomes larger. Therefore, both V,C3 (resulting from mechanism 3) and V,C3H (resulting from mechanism 2 ) are observed for x L 3. From the above, it seems that, in the reaction with V, for x 2 3, mechanism 3 is dominant for 1-bromopropene, but mechanisms 2 and 3 are competitive for the reaction with 3-bromopropene. This might be explained by the presence of the resonance form in 1-bromopropene, but not in 3-bromopropene (this is due to the presence of resonance forms involving the bromine-carbon bond in 1-bromopropene, which are absent in 3-bromopropene): V3
+ C3H6
+
Br+=CH-C-H--CH3
-
Br-CH=CH-CH3
Such a dipolar form greatly reduces the barrier for the addition (complexation) reaction required for mechanism 3 in the reaction with 1-bromopropene. In addition to this, the resonance in CH2=CH-CHz' makes the bond energy of C-Br in 3-bromopropene much less than that of 1-bromopropene (47 vs 70 kcal/mo126d~e).These factors could account for the observed dominance of mechanism 3 in the reaction of V, with l-bromopropene and a competition between mechanisms 2 and 3 in the reaction of V, with 3-bromopropene. The observation of the halogen substitution of 3-chloropropene might also suggest the importance of the carbon-halogen bond energy in determining the product distribution. The C-CI bond energy is about 60 kcal/mol.26e Thus, according to the above argument, the probability of halogen substitution of 3-chloropropene should be between that of 1-bromo- and 3-bromopropene. This can be seen from Figure la-c, where VC3H3is the strongest
J. Phys. Chem. 1990, 94, 1607-1611 for the reaction of vanadium with 3-bromopropene, medium intense for 3-chloropropene, and the weakest for 1-bromopropene. Fragmentation of the vanadium clusters upon reaction with 3-bromopropene is also possible. This mechanism is similar to mechanism 3 except that a VBr molecule is evaporated from the cluster product: V, C3H5Br [V,C3H5Br] Vx-1C3H5-2n+ nH, + VBr
+
-
-
The power dependence of VC3H3is very interesting. From the literature,26awe know that the ionization potential of V is 6.74 eV. After formation of VC3H3, the ionization potential was lowered by at least 0.32 eV. This indicates that, in the product VC3H3,C3H3is somewhat of an electron donor, making vanadium easier to ionize. This is similar to the effect of a benzene molecule on the ionization potential of a niobium atom.I3 Since our mass spectrometer only measures the mass peak width and intensity, it is hard to deduce any structural information about VC3H3. However, it is known that C3H3+(cyclopropenyl cation) is a relatively stable aromatic species.% Therefore, VC3H3+could be formed as a result of its special stability. There is no doubt that further investigations are needed in order to determine the structure of VC3H3. In our previous report,23we studied the reaction of niobium clusters with a few halopropenes. Although M,Br (M = Nb, V) was observed for both metal clusters and for all cluster sizes, no halogen substitution product was observed for the reaction with Nb, (mechanism 2). Instead, molecular addition followed by dehydrogenation to give product of the form Nb,C3HBr was observed. Since Nb and V are in the same column of the periodic table, one would expect similar reactions for both metal clusters. (28) Kemp, D. S.;Vellaccio, F. Organic Chemistry; Worth Publisher Inc.: New York, 1980.
1607
The similarity was observed for the reaction with carbon dioxide and oxygen isotopically labeled carbon dioxide.29 However, the products of reactions of Nb, and V, with halopropenes seem to be formed from different reaction mechanisms. This difference could be a result of the larger barrier for the substitution reaction with Nb, (due to its larger size), a smaller barrier for the addition reaction to Nb,, or simply a difference in the temperature of clusters produced from the two metals.
Conclusions In this paper we present evidence for three different possible mechanisms responsible for our observations on the reaction with vanadium clusters V,: (1) halogen abstraction, (2) halogen substitution reaction followed by molecular hydrogen evaporation, and (3) addition type reaction, a result of sticky type collisions, followed by evaporation of molecular H2 or HX. The addition reaction followed by molecular H2 evaporation could account for the products of the reaction with propene. For x = 1, both halogen abstraction reaction and halogen substitution reaction occur for 3-halopropene. For x I 3, the reaction is dominated by both mechanisms 2 and 3. Due to the structural difference, the relative importance of mechanisms 2 and 3 is different for 1-bromo- and 3-bromopropene. The extent of the dehydrogenation following halide substitution is found to increase with x. The reaction of niobium clusters with the same halopropenes does not seem to undergo halogen substitution reaction. Acknowledgment. We thank the Office of Naval Research for financial support. Registry No. V, 7440-62-2;C3H6, 115-07-1; BrCH=CHCH3, 59014-7; BrCH2CH=CH2, 106-95-6;3-chloropropene, 107-05-1. (29) Song, Li; El-Sayed, M. A. Unpublished results.
Activation of Carbon Monoxide on Nickel-Aluminum Alloy Surfaces and by Interstitial Trapping In a Nickel Matrix. Structure and Electronic Factors from Molecular Orbital Theory Alfred B. Anderson* and S.-F. Jen Chemistry Department, Case Western Reserve University, Cleveland, Ohio 44106 (Received: June 26, 1989)
An atom superposition and electron delocalization molecular orbital study of an isolated CO molecule bound to a cluster model of the Ni( 111) surface shows the site preference order is 3-fold > 2-fold > 1-fold and the vibrational frequencies match well those observed by Campuzano and Greenler. CO is attracted to a coadsorbed AI atom, forming a nonlinear AI-OC-Nix complex with low vibrational frequency, a result of 5 0 and K donation to empty AI 3p orbitals which allows increased K* back donation. This gives an electronic explanation for the activation of CO on Ni-AI alloy surfaces observed by Rao et al. and Yates and co-workers. When trapped in Ni matrices CO again exhibits a large decrease in vibrational frequency and this is shown to stem not from a Ni,-CO-Ni, structure, which favors u interactions, but rather from an interstitial defect structure where Ni atoms contact both C and 0 ends and back donate to the empty K* orbitals.
Introduction It has recently been found that on AI-promoted Ni, adsorbed CO has a greatly weakened stretching frequency.l.2 The complex structure AI,-OC-Ni, was proposedZ based on the facts that CO binds perpendicular to clean Ni surfaces through the C atom and that AI is oxophilic. However, C O interacts only by weak phy(1) Rao, C. N. R.; Rajumon, M. K.; Prabhakaran, K.; Hegde, M. S.; Kamath, P. V. Chem. Phys. Lett. 1986, 129, 130. (2) Chen, J. G.; Crowell, J. E.; Ng, L.; Basu, P.; Yates, J. T., Jr. J. Phys. Chem. 1988, 92, 2514.
0022-3654/90/2094- 1607$02.50/0
sisorption bonding to and (1 11)4 A1 surfaces. The explanation for this weak absorption lies in the closed-shell repulsive interaction between the occupied CO 5a orbital and occupied AI u orbitals, which counteracts the effect of AI back-donation bonding to the empty CO K* orbital^.^ When AI atoms are bound (3) Ryberg, R. Phys. Rev. B: Condens. Matter 1988, 37, 2488. (4) Chiang, T.-C.; Kaindl, G.; Eastman, D. E. Solid Stare Commun. 1980, 36, 25. ( 5 ) Persson, B. N. J.; Muller, J. E. Surf. Sci. 1986, 271, 219. Bagus, P. S.;Nelin, C. J.; Bauschlicher, C. W., Jr. Phys. Reu. B Condens. Matter 1983, 28, 5423.
0 1990 American Chemical Society