We've been having a discussion around the water cooler and have hit a brickwall, thought y'all could help. It's about buffers, organic solvents and pH/pKas. Some people think that when you add IPA to your buffer (lets say 10%) and the pH changes that everything is valid and you can use the pH reading from the meter. The rest of us are saying that the meter doesn't read it "exactly" right and that not only do you shift the Ka of your buffer/water, but also your analyte Ka should shift. Also that the shift in Ka for water is not always going to be proportional to the shift you find in the Ka for the analyte. What I guess I'm asking is that if my buffer is pH 9.0, I add 10 % IPA and the pH meter reads 8.3, is this a valid reading or do I need to also take into account that the pKa for my analyte will shift in this new organic buffer.
Thanks for any input

Unfortunately, the real situation is extremely complex and there is no simple answer. You are right, however, that changes in the solvent dialectic constant will in fact shift the pKa of buffer species, of your analyte species and even the properties of your pH electrode. This topic has been discussed on numerous occasions in this forum and I recommend that you search the archives if you're interested in a detailed discussion. But, the common consensus you will find in previous discussions is that one should measure pH only in aqueous solutions since there are no suitable reference materials for calibrating a pH meter in a nonaqueous (or partially nonaqueous) solvent system.

Thanks Chris, I think your info makes my point. But a quick follow up (That wasn't really clear in previous threads): The shift of pKa of my buffer will not be an equivalent shift of the pKa of the analyte (especially if they are structurally different e.g. ammonium acetate vs citrate or NADP)?

You are correct. One cannot assume that environmental factors (regardless of whether the parameter is dielectric constant, temperature or the nature of the solvent) has an identical effect on the pKa of all analytes. Unfortunately, as I mentioned, there isn't any tabulated data on the effect of dielectric constant on pKa but there is tabulated data on the effect of temperature on pKa. Still, this data illustrates the point nicely: the pKa for both acetic and formic acid reach a maximum at 25°C, butyric acid is highest at 0°C while both the first and the second pKa for citric acid are highest at 50°C although the third pKa for citric acid is highest at 10°C. Similar sorts of analyte specific dependence on solvent composition can also be expected.

I agree with Chris. In most practical cases for the practical chromatographer, the best way of measuring the pH is in water, where you have defined buffer pK values, defined reference solutions etc. Only thorough theoretical work should be done by measuring pH in the presence of an organic solvent. But those that need to do this kind of stuff also know how to do it right.

The general rule of the shift in the pK of buffers and analytes is as follows: the pKa of acids increases with the addition of an organic solvent, and the pKa of bases decreases when you add an organic solvent. This general rule is useful, if you create a chromatographic condition for an analysis, and you suddenly find that the retention of one of your analytes is more wiggly than you would like to see.

This is why we tend to set MP pH at least 1.5 pH units away from the main analyte pKa and we differentiate between 'pH' and 'apparent pH' for cases where a pH adjustment is made in an otherwise final solution (containing the organic).

Adjusting pH of MP w/ organic in it isn't going to get you an accurate number, but as long as you're not camping on the pKas of analytes, consistency is maintained and reproducibility is possible.

DR: It is not a question that you do not get an accurate number when you measure the pH in the presence of an organic solvent. The real question is the meaning of the number. Unless you have a calibration, you do not know. A phosphate buffer at pH 7 is a perfect buffer. You add organic solvent and measure again, you may have a value of 8. It means nothing from the standpoint of reproducibility of chromatography, it is still a buffer of optimum buffer capacity.

I have a mission in life to stamp out the notion that one can not, or should not, make pH measurements in partially aqueous solvents. To answer the first post, and some of the others....

1. the glass electrode responds nearly ideally in water, alcohols, acetic acid, acetonitrile. Hence, stuffing the electrode into any of these solutions or mixtures thereof, will yield a meaningful pH. In some cases, especially when the alcohol content is not too great (say less than 50%) the pH reading will correspond closely to a measurement of hydrogen ion activity. I have recently been involved on a project where thousands of these measurments have been made in a solvent with only 5% water.

2. If one wants to compare pH among partially aqueous solutions where the water/organic ratio remains nearly the same and some acid or base component changes, the ONLY reliable measurement that can be made is in the solvent mixture, and NOT the aqueous component before mixing.

3. Pka's and pKb's change depending on the solvent properties, with dielectric constant being the dominant factor in most cases of interest to chromatographers. (eg. It gets very hard to separate a proton form an already charged dihydogen phosphate when the dielectric gets smaller, so this species gets very basic at low dielectric. )

4. One can not presume either the pH of the buffer nor the pK of the analyte from an aqueous measurement before addition of organic (trends can be estimated). Extrapolating aqueous pH readings to mixed solvents can, and has, led to incorrect interpretations of data, the most famous of which was Kirkland's erroneous conclusions that mineral acid buffers dissolved columns faster than amine buffers. If you want to do equilibrium calculations you must have pk and pH data in the solvent composition of interest.

5. There are NIST pH calibration buffer recipes for methanol/water solvents. There are data for pKa and Pkb for common acids and bases in methanol, ethanol, acetic acid and acetonitrile.

Conclusion: to answer some questions one must make measurements in the solvent of interest. In other cases, and most practical LC situations, a measurement in water before dilution with organic will suffice. The trick is knowing the difference, which isn't easy.

Uwe - Agreed, the number is going to be different and we need to be aware of that. My point was that as long as the methods specify when the pH is to be measured and adjusted (either before of after addition of organic component), everyone will do it consistently for a given method and things should go well. The purpose of staying more thatn 1 pH unit away from pKas of main analytes is to allow for a little movement of pH with addition of organics and still be 'safely' away from a critical pH. (a pragmatic approach)

Bill - thanks for your input. Knowing the difference isn't easy, but being certain that everyone preparing a given phase consistently is helpful (and not too difficult).

Bill,

Where can we find pH calibration buffer recipes for methanol/water solvents and data for pKa and Pkb for common acids and bases in methanol, ethanol, acetic acid and acetonitrile? Are there corresponding pH calibration buffer recipes for acetonitrile? Is this only available from NIST? Can you supply a link or does access to this data require a subscription?

Chris

Bill - I have a mission in life to prevent people from complicating life without reason. :-)

In order to get reproducible chromatography of ionic compounds, the first choice is to work in a pH range far away from the pK of the compounds. However, this is not always a good place to be, since one does not have many tools for the manipulation of the selectivity of a separation. Thus one needs to consider intermediate pH ranges, where some of the analytes may be only partially ionized. In order to get reproducible chromatography under these conditions, it is best to work directly at or close to the pK of the buffer. The pK-values of buffers in water are known to every chemist ... acetate pK = 4.75 etc... This is simple, I can measure the pH of my buffer in water, and I know where I am. In the moment I add the organic solvent the pH is shifting, and the pK of the buffer is shifting. Without going to the literature and reading endless tables, how do I know to which w/s or s/s pH I shall adjust my acetate buffer in 47% acetonitrile to get the maximum buffer capacity?
I do not know this value, and I refuse to learn it, because I can survive perfectly by knowing that my acetate buffer has the maximum buffer capacity at pH 4.75 at room temperature in water.
Please don't misunderstand me: If I need to, I'll measure the w/s pH, and I even have book chapters in the works that have tables of these things.
But I do like the KISS principle....

Uwe,

I hope you read to the bottom of my previous note because we don't disagree.

I survived in a large corporate LC lab, and I don't think that I ever measured, or needed to measure, an eluant pH, aqueous or mixed solvent. As you noted, rarely does one have the fundamental data to proceed rigorously. Optimum pH was arrived at empirically by mixing acid and base components, the weights and volumes of components were noted and for ever more the buffers were prepared by weighing and pipeting according to the recipe established in the scouting work. As I have passionately stated before, it is greatly more precise to prepare buffers by weighing and pipeting things. After all, that is how NIST buffers are prepared.

If I encountered a target pH from a lit. reference it was often not clear how the buffer was in fact prepared. That, plus considering the effort of finding and calibrating a pH meter it was faster to reestablish optimum empirically and proceed as usual.

But, if one needs to understand equilibria in the mixed solvent, that is where the hydrogen ion activity must be measured. People need to recognize that such mixed solvent measurements can be made.

Chris, the pK data and buffer recipes I have are in original papers and 3 good books. I do not know if it has been tabulated in a NIST report or not. Contact me off the Forum and I can provide references. Most of this work was done by only a few authors so it is easy to search out. The best work was done by Bates of NIST and an English person, whose name I can't remember how to spell, in the 1950's. There are lots of data for methanol, reasonable data for acetonitrile and less useful LC relevant data for acetic acid because the thrust of this work was to understand weak base titrations.

The correct specification of a buffer preparation is a common problem. There was a previous thread on this topic.