Despite many advances in recent years, crystal engineering remains a risky venture. A successful outcome requires manipulation of the noncovalent bonding and properties such as size, shape, repulsion, attraction, polarity, and chirality. In this Account, we describe the interplay of crystal engineering and synthetic organic chemistry required to develop the family of helical tubuland diol hosts, the members of which exhibit a wide range of tube dimensions and inclusion properties. Certain alicyclic dialcohols crystallize with a hydrogen-bonded network structure, termed the helical tubuland lattice, in space group P3121 (or its enantiomorph P3221). Double helices of diol molecules surround parallel tubes that contain guest molecules, which are included on the basis of size and shape rather than functional group. The crystal structure of (diol)3·(chloroacetic acid)1.2 is illustrative. These chiral helical tubulate lattice inclusion compounds are formed when the racemic host diol is allowed to crystallize from solution. Complete enantiomer separation occurs during this process, producing a 1:1 mixture of pure (+)- and pure (?)-crystals (a conglomerate). The challenge of creating this family of compounds required the development of much synthetic chemistry, in particular new pathways to alicyclic ring systems with specific substitution patterns. It was also necessary to understand and control the supramolecular properties of the diol molecules. What makes the original compound tick, and why did it behave in this remarkable manner, when most of its structural neighbors crystallize totally differently? The synthesis of new helical tubuland diols requires not just preparation of a new molecular structure but also a transplant of the original unchanged hydrogen-bonding supramolecular synthon. Synthesis of the specific crystal space group is necessary. This was achieved by defining structural characteristics, termed molecular determinants, which are essential for the helical tubuland structure to occur. If these requirements were met, then the target molecule had a high probability of success. This investigation has close conceptual parallels with the search for pharmacophore properties of bioactive molecules. In both situations, parts of a molecule with little or no chemical reactivity may actually play vital supramolecular roles. The review illustrates how crystal engineering is based on specific supramolecular properties that can be uncovered and then exploited by synthetic chemists.
School of Chemistry, The University of New South Wales, UNSW Sydney, NSW 2052, Australia