Robustness in AEX, mobile phase composition

[MK]

I'm doing robustness on an anion exchange method with following mobile phases (a salt gradient from 100% A to 25% B).

A = 300mM NaCl – 50mM HEPES – pH 7.5
B = 1500 mM NaCl – 50mM HEPES – pH 7.5

I've thought of doing pH 7.5+-0.2. ICH says that influence of variations in mobile phase composition should also be tested. Should I do it for Hepes buffer, for NaCl, for both or forget it? I've done only RP before and there it is just the organic portion...

thanks
MK

[tom jupille]

You have buffer in both the A and the B reservoirs. Assuming you're allowing a 0.2 pH unit window, the paranoid approach would be the following runs:

A = 7.5, B = 7.5 (control)
A = 7.3, B = 7.3
A = 7.3, B = 7.7
A = 7.7, B = 7.7
A = 7.7, B = 7.3

In principle it's probably better to adjust the pH before you add the NaCl (unless you calibrate your pH electrode in a high salt solution), but whichever way you do it, make sure to write up the documentation in such a way that there is no ambiguity about the procedure.

[MK]

Thank you Tom,

what i was going to do was just:

A = 7.3, B = 7.3 and
A = 7.7, B = 7.7

cause mixing solutions with pH's 7.3 and 7.7 will produce something between these values and what i've read about robustness you just need to do the both ends of the (small but delirable) range then assuming that the effect is linear? Or were you thinking of the slight pH gradient formed in 7.3&7.7 and 7.7&7.3 combinations?

Do I need to vary buffer concentration or can I just say that it is there in excess just to keep the pH constant?

What about NaCl? It is there to elute and could then be "compared" to organic modifier in RP methods.

Many thanks again
MK

[tom jupille]

Hey, just because I'm paranoid doesn't mean they're not out to get me!

Seriously, though:

Yes, if you have a different pH in the A and the B reservoirs, then you are effectively running a pH gradient superimposed on the salt gradient. Because the column packing itself can "titrate" the mobile phase, non-linear effects are quite possible (this is one of the reasons we tell people to avoid pH gradients if possible). You're dealing with a narrow range, so I wouldn't really expect a problem, but the point to robustness evaluation and validation is to demonstrate that no problem exists.

As long as you are not overloading the system or "challenging" it with an injection of wildly different pH, buffer concentration shouldn't be an issue. However, now that you mention it, it wouldn't hurt to do a couple of extra runs with 40mM and 60 mM just to be sure (again, better to demonstrate that no problem exists).

And yes, the NaCl plays the same role in ion exchange that the organic solvent does in reversed-phase.