Chromatography Forum

I read somewhere that the sample pH and the mobile phase pH should not be >2 pH units.
If this can not be followed when developing a method, what are the potential problems?

You typically want your MP pH to be 2 or more pH units away from your analyte's pKA. This minimizes peak broadening/splitting due to changes in solubility that can be associated w/ differences in salt/base form partition coefficients.

And a good rule to remember is "to do the best chromatography, disturb the equilibrium of your system as little as possible."; which means that ideally, your diluent should be the same as your mobile phase.

The potential problem is that your chromatography, or at least some part of it, will be at the wrong pH if the sample has a diff. pH than the mobile phase.

I have studied about in a book that if we not vary the Pka units above or below 2 units the peak shape and Rt changes. One of my professor told the reason that the ionisation of the compound will not be 100% if we not vary the Pka 2 units this due to the unequal ionisation of the compound in the buffer.

I suppose some of it comes from the relative strengths of buffers. If you put a small volume of weakly buffered sample into an LC-flow using a concentrated, strongly-buffered eluant, then the sample will have little effect on the pH of the eluent by the time it hits the column, and there shouldn't be much trouble (but see below). If you inject 20uL 100mM-buffered sample into a 10mM buffered eluant flowing at 200uL/min, it stands to reason that the pH of the eluant is going to be severely disturbed over a large amount of the liquid flowing into the column, so the compounds you're trying to separate will experience all sorts of different pH values in their surroundings.

It's also going to depend on the chromatography. I haven't tried it, but I'd expect, for example, that in simple RP chromatography, if the analyte is in the same ionic form at the pH of the sample, and the pH of the running buffer, you'll have less trouble than if its pKa is somewhere between the two.

You should also think about why the sample pH is different. If the pH was chosen to solublise the analyte, dumping it into a buffer at a pH at which it is insoluble might cause all sorts of problems!

But Tom is, of course, right. The less difference between your sample and the eluant at the start of the run, the better.

I think that a correct statement about pH is "In reversed-phase chromatography, pH of the mobile phase needs to be 2 units below pKa for basic compounds and 2 units above pKa for acidic compounds." In mixed-mode chromatography, where you have both reversed-phase and ion-exchange interaction, you can be at the same pH as your pKa and get good peak shape and retention, which comes from both mechanisms. Here are few examples:
http://www.sielc.com/Compound-Ethylglucuronide.html
http://www.sielc.com/Compound-2-4-Dihyd ... -Acid.html
http://www.sielc.com/Application-HPLC-S ... isc-R.html

One should not take this "2 pH away" rule so serious, one can easily imagine or also demonstrate very good chromatography at pH = pKa, etc., especially in larger molecules where van der Waals forces might be equally important as ionic interactions.

Sigh!

DR ran afoul of one version of "Murphy's Law" which goes "No matter what you write, someone will be willing and able to misinterpret it."

There is no rule anywhere that says that you should never have your mobile phase pH within 2 pH units of the pKa. What *is* true is that k' values (and hence, selectivity) change most rapidly as a function of pH when the pH is near the pKa. Therefore, when developing a robust method, it is safer to begin 2 pH units away if possible. If the pH must be near the pKa in order to get good selectivity -- or if you don't know the pKa -- or if you have several analytes with widely varying pKa values, then "you do what you have to do". If you start seeing robustness issues, you know where to look for the cause.

DR summarized the *potential* problems that might occur. That does not say that they *will* occur in any given case.