Molecular Control of Solid-State Reactivity and Biradical Formation from Crystalline Ketones

Abstract

A growing number of solid-state reactions with high selectivities and specificities have been reported in recent years.1 As the potential of highly selective transformations where solvents can be omitted is increasingly appreciated, it has become desirable to search for reliable methods to carry out reactions in the crystalline solid state.2 In this Communication, we analyze some of the factors that affect the photochemical reactivity of crystalline ketones capable of forming biradicals and radical pairs.3 Our interest in biradicals comes from the fact that they occur in a large number of organic reactions, and strategies for their generation may help develop protocols where selectivity and specificity may be achieved by means of crystal control. The main challenge in the preparation and use of biradicals in solid-state reactions comes from the fact that biradical formation requires cleavage of a σ bond under conditions where separation of the radical termini is essentially impossible (Scheme 1). It is known that photochemical excitation (step 1) to dissociative surfaces in solution and in the gas phase leads to separation of atoms A and B (step 2) to form extended biradicals that explore bimolecular reactions and unimolecular rearrangements.4 In contrast, the two radical centers in the solid state are held close together by their rigid environment, and the two termini are likely to collapse to remake the bond that was originally broken (step 3). For biradical products to form in the solid state, chemical reactions that are comparable to or faster than step 3 will be required. We suggest that proper conditions may be facilitated by molecular design. For instance, in the case of ketone precursors, we propose that dissociative triplet state surfaces and irreversible radical-radical reactions may be formulated with some certainty from known substituent effects.4 A triplet surface will slow down step 3, and a rapid decarbonylation reaction may make up for the lack of biradical separation in a sequence that should lead to products from dialkyl biradicals (Schemes 2 and 3).5 To test our hypothesis, we analyzed the solid-state reactivity of several known 2-phenyland 2,6-diphenylcyclohexanones with 2-hydroxy, 2-methyl, 2,6-dihydroxy, or 2,6-dimethyl substituents. The selection of compounds 1-6 (Scheme 2) was based on the expected photochemical effects of their R-substituents and on their abilities to form good crystals. We have elucidated all their X-ray structures.6 The positions of their R-substituents (axial/equatorial) are represented in the scheme. The conformation about the phenyl groups is such that good benzylic stabilization is expected. If the orientation of the ring is maintained though the reaction, the p-orbitals of the aromatic systems should be kept in close alignment with the benzylic radical p-orbital (see Supporting Information). Several reports on the solution photochemistry of most of the compounds in Scheme 2 are available.7 Although all the compounds in the scheme have high and very similar reaction yields in solution (Φ > 0.6), variations in their solid-state reactivity are very large. Crystals of 1 are photostable. Relative yields of solid-state reaction for the other compounds are shown in parentheses in Scheme 2. These were determined in side-