To take the questions out of order
:
For future development, how can i know the right ionic strength i should use to get good peak separation?
You can't. What you have to do is to pick a reasonable value (usually in the 25 mM range) and then change it and see what happens.
If you have peak like the blue one, how can you know that increasing ionic strength will improve the separation?
Again, you can't. Increasing the ionic strength might have made it worse (in which case, decreasing it would probably improve the separation).
In our
Advanced HPLC Method Development course. we spend at least a couple of hours on ion-pair chromatography (and even that manages to only skim the surface) so there's no way I can give a detailed explanation in a Forum post, but here's a grossly oversimplified version:
An ion-pair reagent is essentially a surfactant, with a hydrophobic tail and a (charged) hydrophilic head. The easiest way for me to visualize what happens is that the hydrophobic tail gets embedded in the hydrophobic stationary phase while the charged head stays on the surface in contact with the aqueous mobile phase. In effect, you end up with an ion-exchanger (the difference between this and a conventional "fixed site" ion exchanger is that in the latter, the ion exchange sites are permanently bound to the support, while in the ion-pair system, they are loosely bound by hydrophobicity). The advantage of ion-pair is that you have a "tunable" system: if you add lots of ion-pair reagent (or use a very hydrophobic ion-pair reagent), you will have a system that behaves primarily as an ion-exchanger. If you add only a little bit of ion-pair reagent (or use a not-very-hydrophobic ion pair reagent), it will behave primarily as a reverse-phase system -- and obviously, you can tweak it to be anything in between.
As with all reversed-phase separations, getting reasonable retention is (usually!) not that difficult. Most of the effort goes into selectivity (the relative spacing of the peaks). That is controlled by six parameters:
- solvent strength (% organic for an isocratic separation or %B/minute for a gradient).
- solvent type (the classical Acetonitrile/Methanol/THF triad); usually not that important in ion-pair unless the selectivity problem is with neutral compounds
- temperature
- pH
- buffer and/or additive concentration
- column chemistry.
We can make general predictions (e.g., adding more organic solvent makes everything come off sooner; adding more IP reagent makes ionics come off later and neutrals come off a bit sooner; changing pH can make ionics go either way but usually has little effect on neutrals; increasing the salt concentration makes ionics come off sooner but has little effect on neutrals; etc., etc.) but in any given case it comes down to experimentation: pick a parameter, make a big change, and see what happens.