Knowledge Bank

[FIGURE] The trimethylenemethane (TMM) derivatives 2-methylene-1,3 cyclopentadienyl (I) and 2-isopropylidene-1,3 cyclopentadienyl (II) have been studied using \emph{ab initio} wavefunctions constructed from a (9s5p/4s) / [4s3p/2s] Gaussian basis set. The states of interest are the planar $\mathrm{<mrow\; data-mjx-texclass="ORD">3</mrow>}B2,1A1$, and $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}B2$; and the perpendicular, or orthogonal $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}B1$ and $\mathrm{<mrow\; data-mjx-texclass="ORD">3</mrow>}B1$. Optimized singled configuration wavefunctions were used for all states except the $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}A1$ state for which an optimized two-configuration wavefunction was used. These derivatives are quite similar to TMM itself - $\mathrm{<mrow\; data-mjx-texclass="ORD">3</mrow>}B2$ is a highly delocalized structure; $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}B2$, $\mathrm{<mrow\; data-mjx-texclass="ORD">3</mrow>}B1$ and $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}B1$ resemble allyl plus methyl, while $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}A1$ may be described as ethylene plus two methyls. Comparison of compounds I, II, geometry-optimized TMM, and TMM fragments with the I and II geometries shows that the energy levels of TMM are altered both by the geometric discortion and the alkyl substitution achieved by the incorporation of TMM in a cyclopentyl ring. In the case of compound I, the lowest singlet is $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}A1$, which is 9.1 Kcal/mole above $\mathrm{<mrow\; data-mjx-texclass="ORD">3</mrow>}B2$. $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}B1$ is 5.2 Kcal/mole higher. When the same geometries are used for compound II, $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}A1$ is 10.9 Kcal/mole above $\mathrm{<mrow\; data-mjx-texclass="ORD">3</mrow>}B2$, and $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}B1$ is 0.9 Kcal higher yet. Thus in this compound $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}B1$ and $\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}A1$ are much closer in energy than in compound I. These results are in marked contrast to conclusions drawn from experimental studies which have been interpreted as indicating a single-triplet splitting of a much smaller magnitude.