RESUMO
Molecular mechanics (MM) with MMFF94 and MMX force fields and ab initio (RHF/6-31G*,RHF/6-311G**, and B3LYP/6-311G**) calculations are used with lanthanide-induced shift (LIS) to investigate the conformations of N-methyl-2-pyrrolidone 1, N-methyl-2-piperidone 2, ε-caprolactam 3, γ-valerolactam (1,5-dimethyl-2-pyrrolidone) 4, 2-azetidinone 5, 4-methyl azetidinone 6, 4-phenyl azetidinone 7, and N-methyl-4-phenyl azetidinone 8. The Yb(fod)3 paramagnetic induced shifts of all the 1 H and 13 C nuclei are measured and the corresponding diamagnetic complexation shifts obtained by the addition of Lu(fod)3 . The complexation model (two-, three-, or four-site) used depends on the relative rates of the processes involved. The amide inversion is the same order as that of the 5- and 6-membered lactam rings and much faster than the lanthanide complexation and the inversion of the 7-membered ring. Both MM and ab initio calculations give an envelope conformation for 1 with C-4 out of the ring plane in agreement with the LIS analysis. For the piperidone ring of 2, the half-chair is calculated as the most stable form. The LIS analysis confirms this but cannot exclude a small amount (<2%) of the boat conformation. For 3, the LIS analysis gives a minimum for 90:10% chair to boat conformation, and 4 exists in two envelope conformations with the C5 -Me ps-eq and ps-ax in an eq/ax ratio of 94:6%. In 2-azetidinone 5, the ab initio calculations gave both ring and nitrogen planar, but the MMFF94 calculations give a butterfly ring and pyramidal nitrogen. The LIS analysis for 5 gave good agreement (Rcryst 0.46%) for the MMFF94 geometry with endo NH but the planar ab initio geometries worse agreement (Rcryst = 1.1%). For 4-methyl-2-azetidinone 6, the MMFF94 geometry gave good agreement (Rcryst 0.96%) with two butterfly conformations with axial and equatorial methyl groups in 1:1 ratio. All the planar geometries gave worse agreement (Rcryst >1.5%). In 4-phenyl azetidinone 7, the MMFF94 geometry with 60% of the axial conformer gave Rcryst 1.2% but the other geometries Rcryst >1.5%. In contrast the N-methyl-4-phenyl-2-azetidinone 8 gave good agreement for all the geometries. The butterfly conformation gave Rcryst 1.1% for 80% of the axial conformer and the planar geometries Rcryst 0.98%. The LIS results confirm the ab initio and MM optimised geometries, but the conformer energies at times differ from the calculated values. They also differ considerably from the corresponding values for the lactones studied previously, and possible reasons for this are discussed.
RESUMO
A refined Lanthanide-Induced-Shift Analysis (LISA) is used with molecular mechanics and ab initio calculations to investigate the conformations of benzamide (1), N-methylbenzamide (2), N,N-dimethylbenzamide (3) and the conformational equilibria of 2-fluoro (4), 2-chloro (5) and N-methyl-2-methoxy benzamide (6). The amino group in 1 is planar in the crystal but is calculated to be pyramidal with the CO/phenyl torsional angle (ω) of 20-25°. The LISA analysis gave acceptable agreement factors (Rcryst ≤ 1%) for the ab initio geometries when ω was decreased to 0°, the other geometries were not as good. In 2, the N-methyl is coplanar with the carbonyl group in all the geometries. Good agreement was obtained for the RHF geometries, with ω 25°, the other geometries were only acceptable with increased values of ω. In 3, good agreement for the RHF and PCModel geometries was found when ω was changed from the calculated values of 40° (RHF) and 90° (PCModel) to ca. 60°, the X-ray and B3LYP geometries were not as good. The two substituted compounds 4, 5 and 6 are interconverting between the cis (O,X) and trans (O,X) conformers. The more stable trans conformer is planar in 4 and 6 but the cis form non-planar. Both the cis and trans conformers of 5 are non-planar. There is an additional degree of freedom in 6 due to the 2-methoxy group, which can be either planar or orthogonal to the phenyl ring in both conformers. The conformer ratios were obtained from the LISA analysis to give Ecis-Etrans in 4 > 2.3 kcal/mol (CDCl3 ) and 1.7 kcal/mol (CD3 CN), in 5 0.0 kcal/mol (CD3 CN) and in 6 > 2.5 kcal/mol (CDCl3 ) and 2.0 kcal/mol (CD3 CN). These values were used with the observed versus calculated (1) H shifts to determine the conformer ratios and energies in DMSO solvent to give Ecis-Etrans 1.1, -0.1 and 1.8 kcal/mol for (4), (5) and (6). Comparison of the observed versus calculated conformer energies show that both the MM and ab initio calculations overestimate the NH..F hydrogen bond in (4) by ca. 2 kcal/mol.
