If you show an organic chemist the structures of two stereoisomers, she’ll probably have no issue characterizing them as conformational or configurational isomers. Structures differing in the arrangement of atoms about one tetrahedral stereocenter are naturally configurational isomers; structures differing only with respect to rotation about a single bond are naturally conformational isomers. Baby stuff.
The criterion for making these judgments appears to be whether the isomers are related by one or more rotations about formally single bonds (conformational) or any other sort of topological change (configurational). The terms “configuration” and “conformation” only have meaning, then, when another isomer is considered. Put another way, a single structure can be both a configurational and conformational isomer simultaneously depending on context. That’s not so bad, although the situation causes issues for the novice, to whom the relevant “virtual isomer” may not immediately spring to mind.
The real issue comes to the fore when one considers why the hell “rotation about formally single bonds” is the criterion for conformational isomerism. This is one of those theoretical facts that you’ll tell a student, and afterwards they might just stare back at you blankly (if they aren’t of the disposition to accept theoretical facts at face value). They’ll wonder what’s so special about single-bond rotations.
Upon reflection, you’ll realize that there’s nothing special about bond rotations at all, except that they’re easy at room temperature. Falling further down the rabbit hole, you may conclude that any structures related by an easy process at room temperature ought to be called conformational isomers. This includes, for example, amines containing stereogenic nitrogens related by amine inversion (which occurs with lightning speed at room temperature).
The new definition admits more isomers as conformers and removes the illusory “specialness” of single-bond rotations, but it creates additional problems! All of a sudden, the isomeric relationship between two structures depends on how much energy is available in the surroundings: if the activation energy of interconversion of the isomers is less than the ambient energy (proportional to the ambient temperature), the isomers are conformers; if not, they are configurational isomers. It seems we can transform conformers into configurational isomers simply by lowering the ambient temperature! That doesn’t seem right—whether two isomers are configurational or conformational ought to depend only on their structures, right? What activation barrier corresponds to an “easy” process? Would an easy process on Venus be the same as an easy process here on Earth? Yikes…
The only way to sharply divide conformation and configuration under this new definition is to establish a threshold for activation energies of interconversion, above which isomers are considered configurational. I’ve heard that IUPAC has defined the threshold as 19 kcal/mol (80 kJ/mol). This energy corresponds to a rate of about once per second at room temperature, assuming a frequency factor of 1014 per second. IUPAC’s threshold seems sensible at room temperature, but what about –78 °C or refluxing toluene? To the chemist working at these temperatures, the definition still seems arbitrary.
There really are no general solutions to the ambiguities associated with conformation and configuration aside from setting an arbitrary cutoff point or using the “single-bond rotations only” definition. However, the concept of residual stereoisomerism works well for practical purposes and allows one to throw conformation and configuration out the window. Residual stereoisomers are all the isolable species present in a particular sample under a given set of conditions. 1,2-Dichloroethane consists of one residual stereoisomer at room temperature, but cool the stuff down to 50 K and three (the two enantiomeric gauche structures and one anti structure) emerge. The single isomer at room temperature is best conceived as a “smeared” average of the conformers, weighted by their Boltzmann populations, but it’s tough to argue that the conformers have independent existence at room temperature. Cooling them down makes them isolable from one another, at which point it makes more sense to think about them independently. Whether we call them conformers or configurational isomers at 50 K is irrelevant.
Will conformation and configuration ever be purged from the organic chemistry literature? Probably not, as these terms are central to the discipline (despite their ambiguities). Still, residual stereoisomerism is a valuable concept for the practicing chemist, to whom the isolability of a structure is often most important. Eliel aptly calls the condition for the existence of a residual stereoisomer the isolability criterion. Conformers by the second definition fall on the “not isolable” side of the criterion, whereas configurational isomers fall on the “isolable” side.
As an aside, the “single-bond rotations only” definition of conformation does have some value from a topological standpoint, when the student first encounters the idea that groups can rotate about single bonds. In practice however, conformation connotes flexibility and configuration connotes rigidity, and these gut feelings grow stronger with additional studies. It’s worth re-evaluating the rigor of one’s instincts from time to time!