by
lmh » Thu Feb 19, 2015 2:21 pm
I think partition is still a helpful term. It points at the mechanism: the analyte can be "stuck to" the mobile phase or the stationary phase. If it spends, on average, 30% of its time stuck to the mobile phase and 70% to the stationary, it moves at 30% of the linear velocity of the solvent. It may not be partitioning in the strict sense of physically being confined to one of two separate phases, as happens in a two-phase partition in a glass flask, but it is still partitioning in the sense of "spending time associated with one of two phases".
What the physical nature of the association actually looks like, in the column, is a different matter, and probably pretty complex. One can worry too much. Knowing the ways in which it can interact (hydrophobic, ionic etc.) is enough to make some educated conclusions about how to modify a method.
I don't think this is unique to reverse phase, or affected by the fact that the C18 chains are bonded to silica. This doesn't make the material any less stationary. The interaction of a pancake with Teflon in a non-stick frying-pan is unaffected by whether the frying-pan underneath is made of aluminium, copper or steel, and when we toss the pancake we hope for 100% partitioning into air. If the non-stick surface isn't continuous because someone's scratched the pan, then we can get interactions, and a percentage of pancakes will fail to partition correctly! So the partitioning model still holds, it's just that there can be multiple modes of interaction between solid-phase and analyte, and the solid-phase can consist of multiple sorts of surface, all bonded together as one solid unit.