How efficient is TLC in completely separating each individual component of an unknown mixture? Would you need to first have a bit of an idea of whats in the mystery mixture in order to find suitable absorbents?

ive only had experience with silica plates, but i know there are others around (alumina seems to be the second option).

These seem to do well for most things, but remember you can also fine tune them with the solvent choice. Trial and error is usually the name of the game. Most TLC uses are limited to finding out if a reaction is complete (the reactant spot goes away, it done), little real information can be found with them due to thier size and general non-selectability.

TLC plates are naturally limited with choice, so most of the time you have to make do. Once you get into more elaborate chromatography setups (GC or LC) you can buy columns made from anything you like, from simple silica to human serum.

But yes, the more you know about the mixture you are trying to seperate the easier it will be to get it working nicely.

yes...you need a solvent that will be selective for the things you are trying to seperate