An unexpected co-crystal with a variable degree of order: 1:1 rac-1,2-cyclohexanediol/triphenylphosphine oxide.

A 1:1 co-crystal of rac-trans-1,2-C(6)H(10)(OH)(2) and (C(6)H(5))(3)PO has been found that is unusual because there are no strong interactions between the two kinds of molecules, which are segregated into layers. Furthermore, neither pure rac-1,2-cyclohexanediol (CHD) nor pure triphenylphosphine oxide (TPPO) has any obvious packing problem that would make the formation of inclusion complexes likely. The TPPO layers are very much like those found in two of the four known polymorphs of pure TPPO. The hydrogen-bonded ribbons of CHD are similar to those found in other vic-diol crystals. The co-crystals are triclinic (space group P\overline 1), but the deviations from monoclinic symmetry (space group C2/c) are small. The magnitudes of those deviations depend on the solvent from which the crystal is grown; the deviations are largest for crystals grown from acetone, smallest for crystals grown from toluene, and intermediate for crystals grown from ethanol. The deviations arise from incomplete enantiomeric disorder of the R,R and S,S diols; this disorder is not required by symmetry in either space group, but occupancy factors are nearly 0.50 when the structure is refined as monoclinic. When the structure is refined as triclinic the deviations of the occupancy factors from 0.50 mirror the deviations from monoclinic symmetry because information about the partial R,R/S,S ordering is transmitted from one diol layer to the next through the very pseudosymmetric TPPO layer. Analyses suggest individual CHD layers are at least mostly ordered. The degree of order seems to be established at the time the crystal is grown and is unlikely to change with heating or cooling. Thermal data suggest the existence of the co-crystal is a consequence of kinetic rather than thermodynamic factors.