IV. NONELECTROMAGNETIC CONTRIBUTIONS TO SERS As successful as the em theory has been, there exists a number of observations that do not seem to be easily explicable in terms of it, which include, for example, the following. (1) The molecules CO and N2, whose Raman cross sections are almost identical, produce SERS spectra of rather different intensities (Moskovits and DiLella, 1982b).When coadsorbed in equal numbers onto cold-condensed silver films, CO produces a broad Raman peak shifted by some 28 cm- from that of unbonded CO, while N2 pro-duces a narrow Raman peak, shifted by only 2 cm-',whose intensity is about fifty times lower than that of CO. Suggestions that CO perhaps displaces N2 from the surface were countered with experiments in which N2 and CO were adsorbed independently with results similar to those in the coadsorption experiments. Likewise, postulating different surface bonding orientations for CO and N2, the former being "end-on" bonded, while the latter is sidewise bonded, cannot account for the full fiftyfold discrepancy. On the basis of the Raman depolarization ratios of the two molecules, at most a factor of 10 may be obtained for the ratio of the Raman intensities resulting from the adoption of the two orientations. Molecules such as methane and ethane produce SERS spectra some one-hundred-fold less intense than do ethylene or benzene, even though the ordinary Raman spectra of all of these molecules contain bands of equivalent strength. (2) In its simplest forms the em theory predicted that the SERS spectrum of a molecule should be a uniformly enhanced version of the ordinary Raman spectrum (McCall et al., 1980). Observations show this not to be true. This discrepancy is dramatically shown in the SERS spectrum of C6F6 (Fig. 24), which shows the C-F symmetric "breathing" vibration to be the dominant feature, while the same vibration is among the weakest in the ordinary Raman spectrum of the same molecule. Moreover, the SERS spectrum of many molecules (Erdheim et al., 1980; Dornhaus et al., 1980; Moskovits and DiLella, 1980,1982a,1982b) displays normally forbid-den features, often with intensities comparable to the most intense of the allowed transitions. (3)In SERS studies involving roughened electrodes immersed in aqueous electrolytes or aqueous metal colloids one does not observe a strong SERS spectrum of water, which presumably surrounds the SERS-active features in abundance. (4) It has also been observed that the SERS excitation spectrum obtained from aqueous colloidal silver does not correspond well with the absorption spectrum of the same colloid (Creighton et al., 1979; Mabuchi et al., 1982; Sii-man et al., 1983). The latter usually consists of a strong peak at about 390 nm with weaker features to the red of this. The 390-nm peak is associated with spherical colloidal particles, while the red peak or peaks are commonly believed to come about from aggregated colloid whose optical properties were discussed either in terms of ellipsoidal particles or by considering coupling between two or among several spheres. Although it is clear from the work of Aravind et al. (1981) that dipolar coupling is in-sufficient to account for the electrostatic fields about spheres spaced so closely together, dipolar coupling, often couched in the language of effective medium theories such as the Maxwell-Garnett theory does produce a red-shift plasmon peak whose exact position depends on the density of (hence average distance between) the colloidal spheres. 请翻译这段英文

已完成理解「IV. NONELECTROMAGNETIC CONTRIBUTIONS TO SERS As successful as the em theory has been, there exists a number of observations that do not seem to be easily explicable in terms of it, which include, for example, the following. (1) The molecules CO and N2, whose Raman cross sections are almost identical, produce SERS spectra of rather different intensities (Moskovits and DiLella, 1982b).When coadsorbed in equal numbers onto cold-condensed silver films, CO produces a broad Raman peak shifted by some 28 cm- from that of unbonded CO, while N2 pro-duces a narrow Raman peak, shifted by only 2 cm-',whose intensity is about fifty times lower than that of CO. Suggestions that CO perhaps displaces N2 from the surface were countered with experiments in which N2 and CO were adsorbed independently with results similar to those in the coadsorption experiments. Likewise, postulating different surface bonding orientations for CO and N2, the former being "end-on" bonded, while the latter is sidewise bonded, cannot account for the full fiftyfold discrepancy. On the basis of the Raman depolarization ratios of the two molecules, at most a factor of 10 may be obtained for the ratio of the Raman intensities resulting from the adoption of the two orientations. Molecules such as methane and ethane produce SERS spectra some one-hundred-fold less intense than do ethylene or benzene, even though the ordinary Raman spectra of all of these molecules contain bands of equivalent strength. (2) In its simplest forms the em theory predicted that the SERS spectrum of a molecule should be a uniformly enhanced version of the ordinary Raman spectrum (McCall et al., 1980). Observations show this not to be true. This discrepancy is dramatically shown in the SERS spectrum of C6F6 (Fig. 24), which shows the C-F symmetric "breathing" vibration to be the dominant feature, while the same vibration is among the weakest in the ordinary Raman spectrum of the same molecule. Moreover, the SERS spectrum of many molecules (Erdheim et al., 1980; Dornhaus et al., 1980; Moskovits and DiLella, 1980,1982a,1982b) displays normally forbid-den features, often with intensities comparable to the most intense of the allowed transitions. (3)In SERS studies involving roughened electrodes immersed in aqueous electrolytes or aqueous metal colloids one does not observe a strong SERS spectrum of water, which presumably surrounds the SERS-active features in abundance. (4) It has also been observed that the SERS excitation spectrum obtained from aqueous colloidal silver does not correspond well with the absorption spectrum of the same colloid (Creighton et al., 1979; Mabuchi et al., 1982; Sii-man et al., 1983). The latter usually consists of a strong peak at about 390 nm with weaker features to the red of this. The 390-nm peak is associated with spherical colloidal particles, while the red peak or peaks are commonly believed to come about from aggregated colloid whose optical properties were discussed either in terms of ellipsoidal particles or by considering coupling between two or among several spheres. Although it is clear from the work of Aravind et al. (1981) that dipolar coupling is in-sufficient to account for the electrostatic fields about spheres spaced so closely together, dipolar coupling, often couched in the language of effective medium theories such as the Maxwell-Garnett theory does produce a red-shift plasmon peak whose exact position depends on the density of (hence average distance between) the colloidal spheres. 请翻译这段英文」