Buffer pH vs analyte pKa about peak tailing

[ym3142]

As promising to Uwe I did the simple test to find if acid dissociation causes peak tailing/split. Here is what I did/found:

luna c18(2) 150x4.6x5u, flow 1.5 , 40% MeOH, 60% buffer, 10ul injection, room temperature;
sample: sodium benzaote 13mg/25ml solvent (solvent=50%water and 50% MeOH);
buffer: 50mM NaOAc, 50mM KH2PO4, pH adjusted to 2.2, 4.7 and 7.0 separatedly.
at pH2.2, tailing 1.2, Rt 7.8, plates 7557;
at pH 4.7 t 1.1, rt 5.0, p 6717;
at pH 7, t 1.3, rt 1.7, p 2712;

it seems that the dissociating and associating is so fast thus the column can only feel its average.

But I did see some times peak can shoulder and/or split due to changing of pH and we can fix it by adjust the pH. Is this just due to insufficient buffer?

Can any one help to further clarify?

[Mark Tracy]

Peaks do split when the analyte pKa is close to the mobile phase pH. This happens when there is inadequate buffering capacity. The sample is at a pH that elutes earlier than the mobile phase and some of it runs ahead until the mobile phase catches up. This is a variation on the "strong solvent" effect. Adjusting the sample pH to the mobile phase pH is one cure.

Hydrogen ion exchange kinetics are limited by mixing and diffusion. The diffusion rate for H+ is extremely fast compared to ordinary molecules.

[ym3142]

Mark,
Thank you for the input.

1,

The sample is at a pH that elutes earlier than the mobile phase and some of it runs ahead until the mobile phase catches up

I am sorry not understanding.

2,

Adjusting the sample pH to the mobile phase pH is one cure.

I did prepared the sample in three pH 2.2, 4.7, and 7. No difference was seen under different preps.

[koen_shimadzu]

As I understand you want to see peak splitting when pH=pKa.
I'm sure that there are pH levels where you will observe this, but they are difficult to find.

The decrease in plate number can possibly due to the decrease in k-value. At the highest pH you're compound elutes almost in the void. Dead volume interferes here your experiments as I expect.

Also the column you're using may be not the best choice. Take an older column which has more silanol activity, spherisorb or something like.
Also with basics the effect will be much more obvious than with acids

Nevertheless I expect the hydrophobic model and secondary interactions will always dominate over possible "ionic formâ€

[Mark Tracy]

The splitting only happens when the mobile phase buffer is too weak to effectively control the sample pH. In your experiments, the mobile phase used a strong buffer at its optimum buffer points.

Immediately after the injection, the sample has not mixed with the mobile phase. If the sample is self-eluting faster than the mobile phase would, you get peak splitting. This is the "strong solvent" effect. Eventually, the analyte falls behind the injection solvent, and its elution is now controlled by the mobile phase.

In order to cause splitting, you need a bad buffer in the mobile phase and a good one in the sample. It is also much more obvious for large sample injections and early peaks. If you tried the experiment at pH 3.4 where neither phosphate nor acetate is an adequate buffer, you should see splitting.

[ym3142]

koen:
thank you for your input.

As I understand you want to see peak splitting when pH=pKa.
I'm sure that there are pH levels where you will observe this, but they are difficult to find.

could you further explain the meaning of "difficult"? why difficult? but there are always 50% acid and 50% its anion when pH = Pka, aren't there?

Also the column you're using may be not the best choice. Take an older column which has more silanol activity, spherisorb or something like.

are you saying that active silanol can make one peak for acid and another peak for the anion(deprotonated acid)?

Nevertheless I expect the hydrophobic model and secondary interactions will always dominate over possible "ionic formâ€

[ym3142]

Mark, thank you again.

so basically you believe acid/base do not split at pH=pKa if the buffer is good and strong. though we also believe there are 50% HB(the acid) and B-(the anion, the depronatd acid) when pH=pka and the two forms elute from column at different speed.

Immediately after the injection, the sample has not mixed with the mobile phase. If the sample is self-eluting faster than the mobile phase would, you get peak splitting. This is the "strong solvent" effect. Eventually, the analyte falls behind the injection solvent, and its elution is now controlled by the mobile phase.

I'd imagine usually the analyte elutes because of the 'pushing" of mobile phase except for some slow diffusion. what kind of force inside a filled column make the analyte run faster than the mobile phase? what kind situation makes this force play a major role?

i will try 3.4 if have chance

thanks again,

[unmgvar]

ym3142,

Mark Tracy is very right, you made your experiment at the good Ph values of your solution when it acts as a buffer.

i would go a little more crazy then just trying ph 3.4.
i would take out the phosphate of the equation and do ph 2 and 7 again.
just for the experiment.

i am myself having a discussion with r&d collegues of mine that have "borrowed" 2 methods and are having robustness issues.
the methods are with an ammonium acetate "buffer" in both cases; the Ph is 7.0
pka's being about 4.8 for acetate and ammonium around 9.3, i told them that they should start there, but then again, regulations, usp...

[danko]

I think the experiment ym3142 has designed is quite OK. The peak splitting he wanted to observe shouldn’t be dependent on the buffer capacity but rather on the pH value and the effect should be expected at pH approximately 4. Here is the rationale: Benzoic acid (pKa 4.21) is completely protonated at pH 2.2 so a reasonable retention and peak shape are to be expected at this pH. At pH 7.8 the benzoic acid is completely deprotonated, so bad retention and peak shape are to be expected. Finally at pH 4.7 benzoic acid is partly acid and partly benzoate, regardless of the buffer capacity. Now if the column had the necessary selectivity it would’ve been able to separate the two species - the deprotonated fraction eluting first and then the protonated. But it’s obviously not the right column to demonstrate the effect on. And not because of silanol deficit but maybe on the contrary. For if the silica is perfectly covered/endcapped, the deprotonated fraction of the analyte would be too polar for the stationary phase and would elute much faster than the protonated species. So I would suggest an endcapped column for the experiment or one of these polar embedded columns which most probably will alter the selectivity. Anyway the silanol activity should not affect the benzoate retention that much because it’s either neutral or negatively charged. The silanol affect as expected mostly positively charged species e.g. amines.

Best Regards

[rhaefe]

we had this discussion before:

http://tinyurl.com/22g2c6

Look at Uwe's response.

[Mark Tracy]

The kinetics of benzoate + proton <==> benzoic acid is on the order of nanoseconds. It is impossible to resolve them by chromatography where the separation mechanism is operating on a time scale about a million times slower. Only mixing effects that operate on a time scale similar to chromatography can cause splitting of acid/base peaks.

If you want to see a real equilibrium effect that causes peak splitting, try fluorescamine derivatives. The isomerization has a time scale of a minute or two.

[koen_shimadzu]

I, unfortunately, never had the chance to study this in detail, but I'm quit sure Mark is right in this.

To fully understand to know what is happening, it's not a matter of just a few experiments.

[HW Mueller]

Why study something that has probably been known for over a hundred years? One just needs a chemistry book.

[ym3142]

Hans,
I agree with you 100%.
Would you mind letting know a books or two(not too many) which explain why an acid and its anion elute as a single peak under some conditons but two peaks under other conditions?
In other words, if the association/disociation speed matters why somtimes the speed is so fast but other times it is not?
I will really Appreciate
and you have a great weekend,

[HW Mueller]

ym3142, First of all one does not need a book (though any book on LC will allude to this) to explain that an extremely fast process (proton exchange in acids) will not be distinguished by a much slower process (chromatography or NMR). One needs a book to learn that protons exchange extremely fast (see Mark´s statement), even among water molecules.

Now where do you get the idea that under some conditions the two entities are separated on a column?