In the present work we studied the geometric structure of tetrahedrane (I), azatetrahedrane (II), diazatetrahedrane (III), triazatetrahedrane (IV), and tetraazatetrahedrane (V) (see Fig. i). The data available from the literature on the structure of these compounds, with the exception of the parental hydrocarbon [5], are basically results obtained by the MINDO/3 method and a nonempirical method in a minimal basis. However, the MINDO/3 method is known to be unsuitable for calculations of compounds with NN bonds, and the latter method is known to be unsuitable for the determination of the structure of stressed compounds. We carried out the optimization of the geometry of molecules I-V by a nonempirical method in the 3-21G basis [6], whose accuracy in the determination of bond lengths in organic molecules is comparable to experimental accuracy. The GAUSSIAN-80 system of programsadapted for the ES1052 computer was used. In addition, semiempirical calculations were performed by the MNDO method, as well as with the use of a variant of the MINDO/3 method with corrected values for the two-center parameters aNN and BNN, which has demonstrated its superiority over the ordinary versionof the method on an extensive group of nitrogen-containing compounds (MINDO/3M [7]). The data from the calculations of the bond lengths in the molecules studied are presented in Table I~ The data from [i, 2] (MINDO/3), [4, 8, 9] (MNDO), and [2] (STO-3G) were used in part in compiling this table. In the literature there are data from the calculation of bond lengths in double-zeta basis: The results for I in [5] differ only slightly from the values which we obtained, but the data for V in [3], which were obtained with the use of a pseudopotential, deviate significantly from the results presented. A possible reason for this disparity is the greater structural lability of the molecules of the tetrahedrane series, as well as the necessity of using gradient methods of optimization in carrying OUt nonempiri -~ cal structural calculations (see, for example, [i0]).
Read full abstract