Self-assembly crystallisation

In examining a crystalline structure as revealed by diffraction experiments it is all too easy to view the crystal as a static entity and focus on what may be broadly termed attractive intermolecular interactions (dipole-dipole, hydrogen bonds, van der Waals etc., as detailed in Section 1.8) and neglect the actual mechanism by which a crystal is formed, i.e. the mechanism by which these interactions act to assemble the crystal from a non-equilibrium state in a super-saturated solution. However, it is very often nucleation phenomena that are ultimately responsible for the observed crystal structure and hence we were careful to draw a distinction between solution self-assembly and crystallisation at the beginning of this chapter. For example paracetamol, when crystallised from acetone solution gives the stable monoclinic crystal form I, but crystallisation from a molten sample in the absence of solvent... 

In this context, some comments on protein crystallisation can be made. The process of crystallisation can be viewed as one of self-assembly of the quaternary structure, although the constituent units now have a well-defined arrangement in space, in contrast to their less rigid shape in liquid crystalline mesophases. Indeed, twisted structures are very commonly found in globular protein crystals, which are reminiscent of the hyperbolic forms of micro- and mesoporous zeolites, described in Chapter 2. 

For the class of compounds discussed here, the self-assembly process always takes place in solution, where the components have sufficient mobility. For the characterization of the product the method of choice is X-ray crystallography, which requires the isolation of a crystalline solid. However, a total reliance on crystal structure determination poses several problems. The structural information available from single crystal diffraction is very complete, but it is quite often impossible to grow suitable crystals, and even when it is possible, disordered solvent and counter ions can give considerable problems in the resolution and refinement of the structure. A more fundamental question is whether the crystallization process, itself a form of self-assembly, has not resulted in a structural change. As an example we may quote the complex [Cu2(mimpy)2] which exists as isolated double helical units in solution, but which crystallises in columns with strong stacking interactions between individual units [5]. 

Students learn in their first year of studies that crystallisation is a means of purification, and the crystallisation may well have resulted in the isolation of one product from a solution which in fact contained several complexes. Two final problems are associated with the lability of most systems studied rapid rearrangement to give the least soluble product is always possible, even when it is only a minor species in solution, and even if one can be sure that the structure in solution is indeed identical with that in the crystal, it is still necessary to know if the complex in solution is in dynamic equilibrium with traces of other complexes. To understand completely the self-assembly process fundamental to this area of supramolecular chemistry we must study the reactions in solution as well as the crystal structures of the products. We have chosen to present here a number of examples from our own work, selected to illustrate the methods available for the study of self-assembly in solution, and the influence of the factors in Table 1 on the process. 

Key words membrane contactors, hydrophobic membranes, interfaces, self-assembly, colloidal crystals, phase separation, water desalination, membrane crystallisation, membrane distillation. 

Siripitayananon J, Wangsoub S, Olley RH, Mitchell GR (2004) The use of a low-molar-mass self-assembled tranplale to direct the crystallisation of poly (epsilon-caprolactone). Macromol Rapid Commun 25 1365-1370... 

---