The photoelectron (PE) spectra of silatranes (1), 2,8,9-trioxa-5-aza-l-silabicyclo [3.3.3] undecanes, have been analyzed and partially reinvestigated. In the PE spectra of 1 the bands in the 8.2-8.9 eV region, which were previously ascribed to the lone electron pair (LEP) orbital of the nitrogen atom or the SiN bond, are due to the presence of impurities, the products of incomplete hydrolysis. Interpretation of the bands in the PE spectra of pure compounds was carded out by analysis of orbital-correlation diagrams and results of MNDO and AM 1 calculations. The first ionization potential of 1 is associated with a MO localized to a great extent on the axial X--S i--N fragment. The increase with shortening of the Si--N distance is related to the corresponding increase in the degree of noncoplanarity of the nitrogen atom when its LEP is involved in the hypervalent X--Si--N bond.KEY WORDS: Photoelectron spectra; silatranes; orbital-correlation diagram.
~TRODUCTIONSilatranes, XSi (OCHRCH2)3N (1), hold a special position among organic derivatives of pentacoordinate silicon (and its subgroup elements). It is the investigation of their physicochemical properties that has been revealed the most general signs of expansion of the silicon coordination number [1][2][3]. Nevertheless, the main question concerning the orbital structure of 1 remains in dispute. Thus, an empirical assignment of the PE bands of 1-methyl-, 1-hydro-, and 1-ethoxysilatranes turns out to be ambiguous and leads to alternative versions for orbital interactions in such molecules [4]. Subsequent studies of the PE spectra of 1-methylsilatrane, its 3,7,10-trimethylderivative and 1-methyl-2-carbasilatrane as well as a number of model 2-aklanolamines and their O-silylated analogues have made the question of the orbital structure of 1 even more complicated. Initially, the first maximum (8.5 eV) in the PE spectrum of 'Institute of Organic Chemistry, Siberian Branch, Russian Academy of Sciences, Irkutsk, Russia. 2 Institute of Organic Chemistry, Siberian Branch, Russian Academy of Sciences, 664033 Irkutsk, Russia.
1891-methylsilatrane was assigned to the presence of an impurity [5], but subsequently to the nitrogen electron pair of the silatrane molecule in an unusual (and unknown) conformation [6]. Recently the PE spectra of a large series of silatranes were interpreted to a great extent on the basis of ab initio STO-3G calculations of 1-methyl-, 1-fluoro-, 1-chloro-, 1-vinyl-, and 1-chloromethylsilatrane [7]. In this study, however, no account was taken of the experimentally confirmed significant substituent effect on the Si--N bond length of 1-substituted silatranes, nor the pronounced difference of its values in the crystalline and gaseous phase [8,9]. These calculations were carded out without geometry optimization using the same SiN distance for all molecules, which was taken from the X-ray study of 1-methylsilatrane. Therefore, the suggested assignment of the bands in the PE spectra of 1 [7] does not seem to be convincing. This is expressed, in particular...