J. Phys. Chem. 1981, 85,2042-2046
2042
Surface-Enhanced Raman Spectra of Some Butenes Adsorbed on Silver Daniel P. DiLella and Martin Moskovits’ Lash Miller Chemical Laboratories and Erindale College, University of Toronto, Toronto, Ontarlo MSS 1A 1. Canada (Recelvd: February 3, 198 1)
The surface-enhanced Raman scattering (SERS)spectra of cis-Zbutene,trans-2-butene,isobutylene, and 1-butene adsorbed on a silver film vapor-deposited onto a low-temperature substrate are reported. The enhancement factors for the various vibrational modes observed in the surface-enhancedspectra are most easily understood if one assumes that the ad-moleculesare ?r-bondedto the surface through their double bonds. The SERS spectrum originating from the first monolayer is shown to be at least 100 times more intense than that due to the second monolayer, although some enhancement is present even in the Raman spectrum of layers not bonded to the metal. The special enhancement enjoyed by the first monolayer is attributed to a charge-coupling mechanism discussed in a previous publication. Introduction Surface-enhanced Raman spectroscopy, wherein unusually intense Raman scattering is observed emanating from molecules adsorbed on some metal surfaces, has been the subject of a large number of recent publications, too many to list. Early developments in this field are reviewed by Furtak and Reyez,’ while some more recent developments are contained in the proceedings of a recent Raman conference.2 Despite this intense activity several controversial issues are still extant. In the following paper we present the results of a Raman study of four butenes condensed on vapor-deposited silver metal. Experimental Section The apparatus and the method have been described previ~usly.~Briefly, silver wire wound on a tantalum support was electrically heated in vacuo ( lo* torr). The vapor was deposited on a polished aluminum substrate cooled to 11 K by means of an Air Products Displex closed-cycle refrigerator. A layer of gas was immediately deposited over the fresh silver mirror. Gases were obtained from the Matheson Co. (CP grade) and used without further purification. The cooled aluminum substrate was so placed that only the upper half of it intercepted the metal vapor. The whole substrate received the same dose of gas, however, so that the lower (silver-free) area could be used as a reference. The Raman spectra of the polycrystalline butenes were measured from thick films deposited on the silver-free substrate. Spectra were excited with either the 488.0- or 514.5-nm line of a Control Model 554A argon ion laser. The scattered light was collected at 90° and dispersed with a Spex Model 14018 double monochrometer. Conventional photon counting was used. N
Results The spectra shown in Figure 1, A and C, were obtained from, respectively, the silver-covered and the silver-free portions of a substrate on which -200 monolayers of trans-2-butene were deposited. Figure 1B is the Raman spectrum of a thick (>5000 nm) layer of polycrystalline trans-2-butene deposited over the silver-free portion of the substrate. Comparison of Figure 1, A and C, shows that none of the Raman intensity in Figure 1A arises from the normal Raman scattering by the 200-monolayer film cov(1) T. E. Furtak and J. Reyez, Surf. Sci., 93, 351 (1980). (2) W. F. Murphy, Ed., “Proceedings of the 7th International Raman Conference”, North-Holland Publishing Co., Amsterdam, 1980. (3) D. P. DiLella, A. Gohin, R. H. Lipson, P. McBreen, and M. Moskovita, J. Chem. Phys., 73, 4282 (1980). 0022-3654181l2085-2042$01.25/0
ering the substrate. Figures 2-4 are spectra analogous to Figure 1, A and B, but with cis-2-butene, 1-butene, and isobutylene as the adsorbates. Similar experiments using methane, ethane, or propane as adsorbate did not produce significant SERS spectra. The broad bands seen occasionally in the 1300-1600-~m-~ region of the spectra are similar to those first reported by Otto.4 The intensity of these bands relative to the SERS spectra varied from sample to sample and from location to location on the same sample. There seemed to be no correlation between the intensity of these bands and the intensity of the SERS spectra. Cooney et al. have attributed these bands to carbonaceous deposits on the metal surface; this assignment was supported by the work of Tsang et aL6 Discussion The spectra of parts A of Figures 1-4 are clearly too intense to be the result of normal Raman scattering by so thin an adsorbate layer (