Dear All,
I have a question about a compound's PKa.Will you concern about this parameter when you design a method for a acid or basic compound, especiall for those with more groups? If you do, where this parameter can be search for?This is a question presented itself when I read HPLC book of L.R. Snyder.
Thank you for your reply.
Michelle

In principle, I am not concerned about the pKa of a compound. It is good to know, but there is no fundamental reason to worry about it. If you know it, you know in which pH range you need to play to optimize a method. If your optimized method has a pH that is close to the pK of your analyte(s), you need to make sure that the pH is measured well and adjusted well in order to get reproducible results. For this it would be good to know the pK of your analytes. But when you do a method rebustness study, you will find out very quickly, if a compound pK is clsoe to the pH chosen.

What's wrong with a titration?

Try Lange's Handbook of Chemistry. This reference has pKa's for hundreds of compounds and provides a basis for estimation through analogy in the case where your compound is not listed.

You almost don’t need to worry about pKa value if you do just regular reverse phase chromatography. If you do mixed mode separation when compounds resolved based on their hydrophobicity and their pKa value, it is equally important to know the pKa as to know compounds hydrophobicity.
Example below shows separation of 5 compounds based o their hydrophobicity and pKa value vs. similar separation based on hydrophobicity only.
http://allsep.com/makeChr.php?chr=Chr_023

ACD Labs offers software which calculates pKa value of ionizable compounds quite accurately (see link below)
http://www.acdlabs.com/products/phys_chem_lab/logd/suite.html

Scifinder (http://www.cas.org/SCIFINDER/scicover2.html) now offers ACD labs pKa calculator on line for free if you subscribe Scifinder service.

Many thanks for so much helpful information.

I agree with Uwe that knowledge of a compound's exact pKa is not essential to develop a method. It may be useful to know if the compound is an unusually strong acid (for example a sulfonic acid) and hence especially low pH is called for, or perhapos one will be driven to ion pairing or ion chromatography. But for a typical organic acid a bit of phosporic acid is sufficiently low pH to protonate it and gain retention.

Why titration may not particularly useful? For more information read the great number of papers by Roses, or LC/GC last Nov, Dec,& Jan. To summarize, the pKa that is important is the pKa in the mobile phase during elution, which can differ substantially from the pKa in water(presuming the compound elutes with some organic modifier present). To get the pKa in the mobile phase the true thermodynamic pH of the mobile phase will need to be measured during the titration. This measurement will therefore involve calibrating the electrode in this solvent. Buffers for typical organic modified mobile phases (methanol, acetonitrile, etc) are now available to make these measurements, but most will choose to simply lower the pH to the limit of the column and get on with the analysis. If there is no retention then try something else.

Sure, the activity of a protolyte is seldom know or easily determined in other solvents or mixtures than aqueous solution. Still, titrations are possible in most solutions.

An experienced will figure out a suitable pH for the mobile phase anyway, without experiments.

Still, a lot of problems arise from the fact that several people are not considering the dissociation of the analyte.

is it advisable to keep the compound in the ionized state or in the unionized state? Accordingly we can modulate the pH of the mobile phase.

It's usually preferred to operate with analytes that are in their neutral form when using a reversed phase column. This generally gives better retention, better peak shape and better reproducibility. The major exception to this would be basic solutes where the pH necessary to produce the neutral form of the analyte might result in significantly shorter column life, depending upon the basicity of the analyte. Of course, the other exception to this is in cases where you need to adjust selectivity in order to improve the resolution of two closely eluting analytes. Increasing the ionization of one of the poorly resolved components can be an effective tool in adjusting resolution. Of course, in the case of ion exchange or ion pair chromatography the preference is to operate with analytes in the ionized form (with analyte charge opposite that of the stationary phase).

I would like to ask also about Luna C18(2) column (Phenomenex). This column is so hydrophopic, that when I analyse some weak acids (pKa about 2-3, e.g. 3,5-dinitrobenzoic acid), they are retarded strongly even without any modifying acid added.
Surprisingly, diminishing pH (by adding HCOOH for example) is diminishing the retention factor!!! Usually it should be in opposite way. At lower pH the analytes should be less ionized and more retarded on the column.
Can you tell me what makes this column so exeptional?

There are several possible explanations depending upon your analytical system. If you are using a 100% aqueous mobile phase, formic acid can act as a mobile phase solvent and result in reduced retention for some analytes. On the other hand, carboxylic acids form strongly hydrogen bonded dimers under low dielectric constant conditions (i.e. in the stationary phase). Formic acid could conceivably be forming mixed dimers with your analytes. The mixed dimers would most certainly have lower retention than analyte dimers. Finally, this effect could also be a function of formic acid's effect on your buffer system. What buffer are you using?

The eluent is just 60% ACN in water. No buffer.
Well... I just seek a reason, why this column behaves in this interesting way, because all publications I know, predict the opposite behavior at least for "old era" C18 phases!!!

What is your analyte?

Sorry, I see now that your previous post suggested that your analytes were substituted benzoic acids. Actually, if you're compounds are analogues of 3,5-dinitrobenzoic acid then the expected effect of formic acid depends on just how much formic acid you have added. Since dinitrobenzoic acids are pretty strong acids (the most weakly acidic members of the dinitrobenzoic acid family are still 10 times stronger acids than formic acid), your observation may be related to the fact that you've chosen an acid which isn't particularly good suitable for suppressing the ionization of your analytes. Probably, formic acid is masking a secondary retention mechanism rather than suppressing ionization.

But please concentrate on column! The column makes a difference. When I use older era columns like Lichrospher the effect is as expected (and described in a bulk of papers - no surpise!) and your digression about formic acid simply doesn't explain it.
I also used for example bnzoic acid, 2,6-dihydroxybenzoic acid, so the pKa is covered in the range between 1.5-4.
In eluent I used also ion-pair known agents like TFA. The results are very similar.

If you know that the problem is observed only on some columns under identical conditions, this would seem to point to a secondary retention mechanism. Since I'm not privy to the specifics of the Luna bonding chemistry I can't be of much help as to possible explanations with out more information. Have you observed the same phenomenon any other columns? If so, which ones?