I know that alcohols effect the pH value (mv) at a certain concentration. For example: If you add methanol 70% to your 25 mM phosphate buffer and then pH your pH is off. I'm looking for the chemical reason or chemical explaination as to why this happensand. What concentration of alcohol does this begin at?

probably what you are asking is why alcohols affect the reading on a pH meter. The answer involves both electrochemistry, and solution thermodynamics. If you want to read fundamentals get Bates, Determination of pH or Galster, pH Measurement. I will try to summarize and you can dig in these or possibly other books if you want details.

First, pH is not a fundamental measurement of anything, contrary to what is commonly taught. So, a pH measurement does not measure H+ concentration, and in most cases neither does it measure H+ activity. In most cases what a pH measurement tells you is a comparitive measurement between a standard(what you calibrated with) and your sample. If the number is bigger the sample is more basic. How much more basic, well you can't say unless a lot about the standard and sample are the same.

The measurement of pH is in essance a measurement of the voltage generated by a battery whose cells are the glass membrane of the electrode and the reference electrode. Another voltage is generated in your case and that is the voltage generated at the porus frit, or similar device, that provides contact between the reference and glass electrodes. This is called the junction potential. When you go from a dilute specially buffered aqueous solution (the standard) to alcohol/ water this junction potential changes an unpredictable and unmeasurable amount and shows up as a pH change. So, this is the first problem.

In addition, the activity coefficients of everything change when alcohol is added. And finally, the standard state of the hydrogen reference defined as zero under all conditions, in fact will be different in alcohol vs water. Bottom line, you can't readily interpret the difference in what the pH meter says when you stick it in dilute buffered standard and then in alcohol.

A proper use of pH would be to determine that a separation in 75% methanol worked best when the pH meter read 5. Don't worry what 5 means because you can't know. Forever more adjust to 5 and the separation will continue to be fine. Providing you tell someone exactly how you calibrated, someone else could adjust to a reading of 5 and they too would be able to repeat your work. If you want to predict what pH you will need in 50 % methanol or 40% acetonitrile, well that can not be calculated.

In summary, pH is a comparitive measurement that approaches some fundamental meaning only when comparing a reading in identical solutions. Under most circumstances you can not derive a proton concentration or activity from the measurement.

pH measurement should be done on the aqueous component only before mixing. Do not make pH measurements on water-alcohol mixtures. It just doesn't make sense.

Hmmm, interesting discussion on pH. Actaully though, pH correlates very well to H ion concentration and activity if properly calibrated, but only in aqueous solutions. When speaking of organic matrices, acidity and basicity take on different meanings totally unrelated to "pH". I suggest a literature search on "Sorenson" to understand pH fully.

comment to anonymous : the approach you suggest to adjust pH in aqueous phase and then mix is a reasonable one. However, according to the NIST definition of pH( see Bates ref), there is nothing unacceptable about a measurement in alcohol. In fact, pH scales and standards have been developed for methanol/water mixtures. pH measurements are properly,and routinely, made in many nonaqueous solutions, for example acetic acid (ref Bruckenstein and Kolthoff, J. Am. chem. Soc. 78, 10(1956).

reply to chromgod: The Sorenson definition of pH, made in 1909, is pH = -log [H+] ref. Biochem. Z. 21, 131(1909). This definition of pH, while conceptually simple, ignores junction potential and activity coefficients. Worse, it causes chemists to want to give a pH measurement more fundamental meaning than justified. the modern definition of pH, and standards to support the definition, was largely developed by Bates at what was called the Bureau of Standards in the late 40's early 50's. If you want to "understand pH fully" I would suggest either of the two books referenced earlier, or Bates, Part I section B chapter 10 of the Treatise on Analytical Chemistry
,editor Kolthoff/Elving, or Bates and Guggenheim, Pure Appl. chem. 1, 163 (1960).

By the way, the internationally accepted definition of pH is based on measurement with a glass electrode or any other cell reversible to H+ and it is:

pH(unknown) = pH(standard) + (EMF unknown - Emf standard)/ RTln10

nothing is specified about the standard, so one can define any standard one wishes. Under very special conditions this pH approaches -log H+ activity, but in almost all cases pH is only a comparitive or relative measurement.

By the way, the British have a different pH scale than the US. (Usually it is the Frence that do this sort of thing) so a measurement using British standards will yield a different value than the rest of the world (not by a whole lot, but it is an issue in the second decimal place).

There is probably no area of chemistry more misunderstood than pH measurement.

A few comments: there are a ton of recent publications on the subject of pH and pH measurements in reversed-phase mobile phases by two groups of workers from the University of Barcelona, Roses and Bosch, and Barbosa et al. These publications deal with the shift of pH and pK of buffers and samples in the presence of organic solvents in the mobile phase.
From the standpoint of chromatography, the only things that counts are reproducibility and repeatability. For both, the simplest thing to do is to measure the pH in water. The reproducibility of a chromatographic method depends how far away the mobile phase pH is from the pKa of the buffer and the pKa of the analytes. The pKa's of buffers are known in water. The repeatability depends on the calibration of the pH meter. All pH meters that I know of are calibrated in water. The definition of pH in the presence of an organic solvent is of interest to those who would like to understand the details of the retention mechanisms of ionizable compounds in RPLC. But for the average practitioner, this is nothing but an unnecessary complication.

Uwe,
What journal is the work by Roses and Bosch, and Barbosa et al in? Is it published in English?
Thanks.

ewe,
I do like corresponding with someone with a name. I wish less people would choose anonymous. Anonymous #1"s question was "why" the pH reading changes from water to alcohol and that querry has probably been covered in more detail than they preferred. It takes a 2 hour lecture to do the topic justice.

Now on the subject you raised of making a reproducible buffer for chromatography ........
If one is trying to reproduce work and all that was provided was "0.05M phosphate buffer at pH 2.1" then a pH meter might be needed the first time. However, it is more reproducible to prepare a buffer from weighed reagents than to adjust to some pH with a variable component. Think of the errors involved in the quoted approach. Two measurement errors are involved in establishing the slope and intercept for meter calibration and a third error is associated with the buffer measurement. What is important in the chromatography is concentration ( or activity) so the error that counts is the antilog of the sum of the measurement errors. Alternatively, anybody can weigh to 1 part in 10,000. Which is why the buffers used to prepare pH standards are prepared by weighing. As an alternative to the description quoted above, authors of procecdures should provide weights of components. For example, " a 0.05 M phosphate buffer of approximate pH 2.1 was prepared by adding xx g of sodium dihydrogen phosphate and xx g of phosphoric acid to 1 L of water." Such an approach would lead to less ambiguity in concentration when preparing buffers like phosphate( you know the phosphate concentration if one component isn't variable) and improved reproducibility of the resulting pH.

Good point, Bill. However, the pH of a buffer is the major source of the variability. The ionic concentration plays such a small role in RPLC that one can nearly ignore it.

To Anomymous:
The Roses Bosch papers can be found in Anal. Chem, J. Chromatogr. and Anal. Chim. Acta. A fairly recent one in Anal. Chem. 72 (2000), 5163

Measuring the pH of your buffers with a meter which is calibrated with commercial standards prevents a drift to lab or even person specific pH. A scale (balance) may be more accurate than a pH meter, but this doesnīt do you any good if your substance has changed (picked up water, etc.).
If one does an experiment in which one wants to try different specified pH its tough to do it by weighing. Thatīs why we use sort of a combination of what Bill and Uwe suggested. For instance: a phosphate buffer, pH 7.4, is prepared by dissolving the desired mols (weighed) of NaH2PO4 and Na2HPO4 SEPARATLY in water, determining their pH as a check, then mixing them until the pH of 7.4 is obtained. This way the concentration stays correct, no matter how often you need to adjust. This is important to us as the ionic strength is nearly as important as pH in SEC of proteins.
Now it would interest me to see some information on whether organic modifiers not only influence the activity of H+, but interfere with the measurementīs chemistry/physics as well. One can imagine that organics interfere differently with different types of electrodes and especially with semiconductor electrodes. Any info?

R.J.M.Vervoort and co-workers published a review paper about RP systems for the analysis of pharmaceuticals. In chapter 4 (eluent composition, only 3 pages) the authors present a quite concise and useful summary of the "pH problem" from the point of view of day-to-day chromatoraphy practice: J. Chromatogr. A, 897 (2000) 1-22

You asked if organic modifiers(solvents) "interfer" with the pH measurement. It may be best to phrase the answer in terms of how an organic solvent may affect the measurement. Other than affecting activity, the effect will be on the reference electrode. During a pH measurement there is a counter flow of pos and neg ions across the junctiion. The junction is the frit, porus plug, or what ever that keeps liquid from running out of the reference electrode. If the flux of +/- ions is not perfectly balanced, and it never will be, a potential is developed in the junction called the junction potential. Zwitter ions, colloids, polyelectrolytes, solvents, can have a great effect on the junction potential. this potential gets summed along with the potential developed by the glass electrode and its reference and appears on the meter. So, 0.059 V of junction potential will alter the observed pH by 1 unit.

The solvent can also precipitate stuff in the junction and have an additional effect on junction potential.

To minimize these problems, measurments in nonaqueous media are usually done with double junction ref. electrodes, and the high surface area glass joint kind work particularly well. As an example, for a measurement in acetic acid solvent, one might fill the outer chamber with LiCl dissolved in acetic acid.

I sense that some of the people who have responded feel there is something not quite right about a pH measurement in a nonaqueous environment. Outside the chromatography lab these measurements are common and routine. One must appreciate that pH has been defined as an operation, not a fundamental property of matter. The international definition of pH says, in essence, you stick the elecrode into what ever you want and the pH of the system is defined as what appears on the meter. In this context it is not relevant to worry about what effect the solvent may have on the glass sensing membrane. What ever the effect is, if any, it is part of the measurement and therefore can not be considered an interference. Unfortunately most analytical courses teach, incorrectly, that pH is - log H or -log aH, so people want to be able to interpret their pH measurements in these terms. Only under conditions of very low ionic strength simple inorganic salts in a mid range of pH, will a pH measurement have any relationship to -log H or aH. Fortunately, almost nobody ever needs to use their meter to measure -log H. What you normally need to do is to determine that today's pH in some process is the same as yesterdays, and that is a comparison. It is inappropriate to worry about what the meter reading means beyond a comparison of two solutions.
PS a big interference with semiconductor electrodes is light! We tried using one upside down to measure small volumns and light from a window drove it nuts. We had reliability problems with electrodes and finally junked it.

Thanks Bill for confiming my imaginary. I have been trying to take the "believe" out of chromatography, thusly, it would have been nice to see a study (figures) on this. In the meanwhile we will continue to optimize pH ourselves while adapting a literature method. It probably will not help too much if authors would disclose the type of electrode and meter they used.

We also junked the semiconductor electrodes.

Bill:
pH measurements are not made in a vacuum. They have a purpose. In HPLC, the purpose is to get the same retention times all the time. The retention of ionizable compounds may wiggle around with the pH, if the pH is unstable. Therefore an important element of the pH measurement is to know how close or far away you are from the pKa of the buffer that you are using. The pK values of buffers are known and can be found easily in water. As you add organic, these values shift, as does the measurement of the pH. Therefore, in the context of reproducibility and control over chromatographic retention, there is only one way to do this both correctly and simply: measure the pH in water!

ewe,
from your response I can see that I have not been clear about non- and partially aqueous pH measurements. I have been discussing the issue globally, in response to my understanding of the questions and comments, and some may have interpreted my discussion only in the context of preparing a LC buffer. My time as an electrochemist exceeds my time as a chromatographer, and I needed to be reminded that this is an LC Forum. So, to summarize( I hope more clearly):

1. If the choice is between adjusting an aqueous solution to a given pH, then mixing with solvent, or adjusting the partially aqueous mixture, of course the former is prefered. I didn't mean to imply differently.

2. When possible, it is more accurate, more precise and quicker to prepare the aqueous buffer component by weighing components.

3. When writing up methods or papers, one should provide weights of buffer components as well as the target pH of the buffer.

4. And finally, the point where I have created the confusion.........outside the world of preparing LC buffers, there are needs to measure the pH of partially aqueous and nonaqueous solutions, for example in the manufacture of chemicals. There is nothing inappropriate in making a pH measurement in a nonaqueous solution, so long as the purpose of the measurement is comparitive. Nonaqueous pH measurements are more difficult to do than aqueous measurements, so you don't do them when you can avoid it(which takes us back to 1.)

Now, how about diffusion of organic solvent into the electrode (mine certainly can do this), how long does it take to diffuse out again (any figures, anybody?)? On practical terms: someone determines the pH of his aqu/org mobile phase, but then decides to measure his separate aqu. part for publication purposes (or whatever). His electrode is full of organics, the reported aqu. pH is in reality one of an unknown mixture.
Therefore my conclusion: For LC do all pH measurements only in pure aqu. solutions with an electrode only used in aqu. media. Donīt insert your electrode into the aqu. solution, take a portion of it out and measure in a test tube (there was something on the latter in an LC GC, not too long ago).

HW,
I don't understand what you mean by organic diffusing into your electrode. I have never seen an arrangement where this was possible. A pH reading is in reality done with 2 electrodes, though in many cases the two are contained in one body. H+ is sensed at a glass electrode which is a thin glass membrane with something like HCl on one side and the world on the outside. There is no likelyhood that any HPLC solvent will diffuse into the glass membrane. So, as soon as the glass electrode is removed from something like methanol, the methanol evaporates and thats that.


The other electrode is the reference. It is protected from the outside world by a junction, that annoying thing I have discussed earlier. References are always rigged so flow goes from the reference to the outside world, so there is no chance for solvent to enter the reference and muck it up. Alternatively, the filling solution is geled which provides a diffusion barrier in and out. What little solvent might be at the junction surface, perhaps 5 uL, would soon evaporate, or diluted to oblivion when placed into the next mixture. Of course it is a good idea to wash the electrodes between any measurement.

The only other place solvent might enter the picture is the polymer or glass housing for the electrodes. Either is resistant to attack by solvents, and easily dried/washed.

So, to put things in LC terms, the problem of "carryover" would seem insignificant. In reproducing someone elses work, a greater worry would be the inappropriate way electrodes are calibrated by some. And the readout of many modern meters are so highly damped it is hard to get a precise calibration even if you try to do it right. This is why I prefer someone to tell me the weights they used, which there is no proplem reliably reproducing.

We need a new topic. How about the height to which instruments can be stacked? I once saw 2 complete HP 1100 systems and several detectors in one vertical high rise.

Do you know or do you assume that there is no diffusion across the junctions (there are all kinds of junctions).
Incidentally, gels are used to study diffusion without bulk movement.
If one keeps it relatively simple, by staying in aqu. solution, one does not need to be so much affraid of pH measurement as to forgo it.

i personally don't like the gel references because there isn't bulk flow out. Of course, they don't drizzle reference salts into the sample. But they are more prone to mucking up. Just about any problem one ever has with a pH electrode is probably some problem at the junction, so a nice drizzely one helps to clean itself. The diddly amount of organic that may diffuse into the gel junction is unlikely to cause a problem unless one left it in the organic solvent a long time. If one intends to do pH measurements in solvents for a long time, a double junction should be used anyway. The end!! If someone wants more pH info, hire me to do a class or read an above referenced book.

There is also some interesting modern information (FREE) on electrodes in http://www.weissresearch.com/FQA.htm, which also gives a feeling for why some of us try to keep it simple.