Apparent pH and gradient pH

[wanda50]

Part 1
Many of our methods have mobile phases pH'ed (e.g., 3.5 +/- 0.1) after the organic component has been added. Those developers call it "apparent pH".

Part 2
Another quirk some developers have is to put 0.1% TFA in the aq ACN MP A while not putting any TFA in MP B. With our compounds, this definitely adds a pH component to the reverse phase separation, but I'm not sure how they actually determined the impact. I learned to keep every separation as simple as possible, and this just appears to me as unpredicatable.

Since I am revisiting these methods, I have to have my arguements ready for any method changes that I propose to management. I have never tried to argue with those folks, but those practices make the hair on the back on my neck stand up.

If anyone has comments, I would really appreciate it.

[mbicking]

1. This is generally not recommended because, among other things, a) the definition of pH changes when you are no longer in all water, b) the equilibria change, and c) your pH electrode is very unhappy, unstable, and maybe unreliable.

There have been several LCGC articles on this topic, and maybe someone can point you to the links. Or, search the archives. In practice, you will have a more reproducible measurement if you measure pH in water only. Yes, the apparent pH will be different, but this approach is likely to more reproducible from batch to batch. The key in most labs in consistency and reproducibility.

2. Others in this forum may have other opinions, but what you are describing is not always unusual. I will add an acid to the mobile phase simply to lower the pH. This will ensure that acids and bases are in the desired ionization state: acids neutral, bases positive charge. This is not a "buffer" in the traditional sense, but most people don't really care. They just want a low pH to make sure the ionization form is correct.

Certainly you could put the acid in both mobile phase components, and this will prevent some baseline problems during gradients at low wavelengths. Yes, apparent pH will change during the gradient (see 1. above), but again, if the solution is sufficiently acidic, that's all you need. However, I have had good general success only putting the acid in the water. I do not have reproducibility problems with these systems.

If your peak shape is good, retention times are reproducible, and baselines are OK, then I would save your energy for other battles.

Let us know how it goes.

[tom jupille]

Merlin posted while I was composing this, so it's a bit repetitive:

Consistency is more important than correctness. Since HPLC methods are developed empirically anyway, so long as the mobile phase is prepared exactly the same way every time, you should get consistent results.

When you get away from 100% aqueous solutions, the questions "what is pH?", and "how do you measure it?" rear their ugly heads. The pKa of a buffer will change with the addition of organic solvent (which will shift the equilibrium in favor of the undissociated form), so that if you adjust pH in the aqueous and then add organic, the hydronium ion concentration in the final solution will have shifted. Of course, if you try to measure the pH in the final solution (aqueous + organic), you have to take account of the fact that electrode response will also shift in the presence of organic solvent, so you should calibrate the electrode in the same aqueous+organic mixture. But you can't accurately calibrate the electrode unless you know the pKa of the standard buffer in that same mixture. And so on . . .

About 5 years ago, Bill Tindall wrote a series of three articles in LC/GC dealing with pH measurements for HPLC purposes. You can get to them here. Bill's suggestion (which is what we recommend in our Advanced HPLC Method Development course) is the following: when you document a method, ignore the pH and specify the mobile phase composition entirely by weight and/or volume (e.g. dissolve a grams of acid and b grams of base in c mL of deionized water and add d mL of acetonitrile). This has the additional advantage that weighing is much more accurate and precise than pH measurement.

When you deal with gradients and things like TFA, the situation becomes even more complex. Not only does the pKa of the buffer shift dynamically during the gradient, but TFA is a (weak) ion-pairing reagent, and the surface concentration of TFA will decrease during the gradient. Again, consistency is more important than correctness, so once you have established what works, it is imperative to do it the same way every time. In addition to preparing the mobile phases by weight, you should be careful to use the same equilibration time before every gradient.

[Uwe Neue]

To Part 1: for multiple reasons, some already mentioned, you are better off creating a buffer and measuring pH in water. However, if you have a method that works and is reproducible with the pH measured in the presence of the organic solvent, then there is absolutely nothing wrong with it. Let us take your pH 3.5 buffer measured in the presence of an orgainc solvent. This could be a phosphate buffer very close to its pKa, since the pKa of acidic buffers is increasing when you add organic. If this is indeed the case, you will get reproducible results and you should stay happy and leave the method alone.

To Part 2: I also prefer for multiple reasons to add TFA to both the organic component and the aqueous component of the gradient. However, if retention is reproducible, I would not bother trying to fix it. If the analyte concentration is low, their pKa is a mile apart from the pKa of TFA and the form in which they are injected is in the same form as they would elute due to the presence of TFA, you may get reproducible results even if the buffer is weak. If this is the case, fold your hands over your belly and twiddle your thumbs.

[DR]

My $0.02 on part 2 - Thinking in terms of partitioning and the function of TFA (as an ion pairing agent), there is a possibility of trouble that is proportional to the retention time of your last peak of interest, compounded by its tendancy to tail. In other words, once your peaks of interest are in solution, the TFA concentration at the column head is no longer as important. If you have some tailing later eluters, you may benefit from having TFA in your B phase.

[wanda50]

Thanks SO much for all your help. If I have crazy typos, I have gloves on since the lab is about 65F.

Apparent pH was more to me an intellectual exercise since the results are consistent, and have been in our QC labs for 15 years. I really liked Tom's scenario where you have to pH in the same composition. as everyone said - consistency is the key, but the whole thing irks me.

TFA in only MP A has been a pet peeve of mine for years for all the reason stated. When I develop, I use TFA in both phases. e.g., To reduce the baseline rise especially at 220 nm, I use 1 mL/L in MP A and 0.9 ml/L for MP B. Not a difficult thing, and keeps the pH about the same as well as ion paring impact. I have never figured out whether if it is a true lack of understanding of the role of TFA or laziness for the developers. I believe the answer is yes to both. Again - consistency and if it isn't broke, don't fix it.

You guys are so eloquent compared to me that I really appreciate your time.