Buffer pH vs analyte pKa about peak tailing

[ym3142]

Hans,
Thanks for your reply.
Now I know that you are the one who does not believe these two species can be separated in a single injection, though, which has been being experienced by so many other chemists here.

Thanks anyway,

[danko]

Mr. Mueller,

I’m sorry to inform you, but you’re beginning to contradict your self - first you state: “One just needs a chemistry bookâ€

[ym3142]

Danko,

You are the man!!!

[Uwe Neue]

I have never seen two peaks originating from a proton exchange. I have seen strange peak shapes, when the amount of injected analyte in one form overwhelmed the buffering capacity of the mobile phase buffer.

I have seen secondary equilibria that can give two peaks or peaks with shoulders that depend on pH. A typical example is sugars or ...pril drugs. There you have a slow conversion of two structurally different forms of the parent compound, and the speed of conversion is pH dependent.

We need to differentiate such things from the protonation equilibrium.

[JA]

We've certainly observed peak splitting when injecting a sample which had a pH mismatch with the mobile phase. It is my interpretation that this is because one form of the molecule, either ionised or neutral, becomes favoured in the plug of injection solvent and will have an RT distinct from that of the other form which is favoured at the mobile phase pH (or the average, if an equilibrium is set up). What confuses me about Mark's statement that "splitting only happens when the mobile phase buffer is too weak to effectively control the sample pH" is whether the mobile phase buffer capacity is actually the critical parameter - does this mean if we took the case of an ionisable sample dissolved in a strong acid and injected into an eluent stream comprised of an equally strong base we would not get peak splitting? In the same way that the strong solvent affect causes peaks to run ahead, isn't the critical process that of the sample solvent taking time to mix properly with the mobile phase which leads to peak splitting?

Most of this focusses on the properties of the sample solvent, however it is later mentioned that if the experiment were repeated at a pH where neither phosphate nor acetate is an adequate buffer then splitting should be observed. It's known that injecting ionisables into an unbuffered mobile phase can give distored peaks, and I now have the strong feeling I have missed the explanation as to why - could someone point it out for me. I can't comprehend it when we consider that compounds can be run with good peak shape when the pH is buffered at or close to the pKa as evidenced by ym3142's data and my own observations. If the species are rapidly interconverting in both the buffered and unbuffered mobile phase then where is the distinction?

Is it possible to be definite that peaks split in an unbuffered mobile phase, or simply that peaks will be misshapen?

[ym3142]

To myself, and to Uwe, Danko, Hans, Mark and all other participants here.

I believe we have the answer. Danko is the major contributor.

As for benzoic acid, no peak splitting/shouder should be seen when mobile phase pH<2.2 or >6.2(as i remember the acid pka is 4.2)

when the pka is from 2.2 to 6.2 or especially around 4.2 and when there is enough buffer then no splitting/shoulder either there will no splitting/shoulder because ;

when the pka is from 2.2 to 6.2 or especially around 4.2 and when there is not eoungh buffer then splitting/shouder happens becuase the fast dissociation/association is restrained. why restrained? because there is not enough proton in the environment locally(as Danko said).
As a result, some anion elutes faster than other neutral acid.

any one has some other ideas/suggestions.

I am happy.

[ym3142]

sorry , i tried to say:

when the pka is from 2.2 to 6.2 or especially around 4.2 and when there is enough buffer then no splitting/shoulder either because of the fast dissociation/association ;

[Uwe Neue]

JA is right. You can get truely split peaks, if your sample is injected in a buffer-type of environment at one pH and the mobile phase buffer is at another pH. His example is right on, and this is actually a not uncommon problem in dissolution testing. The analyte is dissolved in a rather strong acid at low pH, and you can get two peaks, if your mobile phase is in a rather weak buffer at a drastically different pH, where the analyte would be predominantly in a different form than at the acidic pH used in dissolution testing. This is equivalent to the injection of a sample dissolved in DMSO, where a large fraction of the analyte never gets out of the DMSO and is properly retained on the RP column. In this case, you also get two peaks: the analyte comigrating with the DMSO peak and the sampel properly retained by RPLC.

But this does not mean that you can separate the ionic form from the non-ionic form of the analyte. You have a situation where the analyte is not in equilibrium with the mobile phase, because it is migrating in an environment that is different from the mobile phase. You get two peaks that each depend on the local environment

To go back to Danko:

If you see a split at the peak top, or other strange peak shapes, they are in the case of simple ionizable analytes NOT due to a separation of the ionic from the non-ionic form, but due to differences in the mobile phase environment of your sample for at least some period of time during the migration process. If you get strange peak shapes, including double peaks, this does not mean that you have separated the ionic form from the nonionic form of the same substance. It only means that there was a mismatch between the injection solvent and the mobile phase. I guarantee you that I can generate a double peak for benzene, and I do not think that I will be able to make you believe that I managed to separate the benzene with the double bond on the right side from the benzene with the double bond on the left side...

And finally to YM3142: Your experiments are in agreement with expectation, including the low plate count at low retention, which is due to extra-column bandspreading.

The world of chromatography is completely intact...

[ym3142]

Uwe:

But this does not mean that you can separate the ionic form from the non-ionic form of the analyte

i agree. Becuase no one can collect one peak as ionic form and another peak as nonionic form since the dissociation/association never stops. But at every moment anion always elutes faster than the neutral causing spitting/shoulder. but the process is dynamic. So do you agree wih this?

It only means that there was a mismatch between the injection solvent and the mobile phase.

i know strong solvent effect. so is this "mismatch" meaning the same as "strong solvent effect" or more, such as weak solvent; buffer/non buffer solvent; diffferent buffer concentration solvent or others?
Does a weak peak also cause bad peak shape some times?
Do you have an example of 10 ul injection of solution to give a bad peak shape bcuase the diluent doesnot contain buffer(same strength though)?

I guarantee you that I can generate a double peak for benzene

do you have ways other than using strong solvent effect to
generate a double peak for benzene in a normal column(not dead) and normal setup(such as good connections)?


To any others,
do you have good experiment or results to help me to clarify this?

[HW Mueller]

Danko, please read my statement again, you will see that there is not even a hint of a contradiction. Now I don´t want to go into the contradictions in some other statements here.

Danko and ym3142: Just go into a chemistry library, start by picking out any general chemistry text, then get yourself any general organic text, maybe next a text on physical organic chemistry, then one on physical chemistry. Finally a book on chemical equilibrium would be helpful.
You will have learned that H2O is a protic solvent, that is, any water molecule readily gives up a proton. How long do you think it will take a carboxylate anion to yank a H+ from the multitude of H2O around it? Did you really think that its going to sit around and wait until a H3O+ comes around (even though that would also be way too fast for resolution via chromatography)?

[danko]

ym3142 wrote:

do you have good experiment or results to help me to clarify this?

I think an anion exchange setup/analisys would be nice.
F. ex. Use the same mobile phases as A-eluents. Get a portion of each and dissolve NaCl e. g. 0.2 - 0.3 mol/L and use it as B-eluent. Then run a very shallow gradient on an anion exchange column from 0% B and until you se a peak.
Maybe the middle buffer (pH 4.7) should be adjusted t0 pH 4.2 – not because 4.7 is that bad but I think the anaytes pKa is the best pH value for this particula experiment.
And by the way, in order to avoid the “strong solventâ€

[danko]

Hi Hans,

Nobody asked for a reference on a physical chemistry book, but as ym314 wrote:

which explain why an acid and its anion elute as a single peak under some conditions but two peaks under other conditions?

By the way, I didn’t realise that water was a stronger acid than benzoic acid

Best Regards

[Peter Apps]

This topic is on the way to becoming a replay of the never ending saga of the minimum weight !

What can we all agree on ?,

that in straightforward RP LC an anion will move faster through the column than its corresponding unionised molecule,

that a weak acid will be predominantly ionised or not depending on the pH of the solution in which it finds itself,

that the kinetics of protonation are far faster than the kinetics of chromatography

by changing the pH of the mobile phase we can make a molecule move faster by ionising it, and slower by de-ionising it.

as long as the composition of the mobile phase is uniform across a peak, the migration speed of all the molecules in the peak will be equally affected by the mobile phase pH. Under these conditions it is not possible to separate ionised from unionised.

Conversely, if the composition of the mobile phase changes across the peak, the molecules in one part of the peak will move at a different rate to those in another part. In the extreme this could lead to two peaks, as long as there is no exchange of molecules between the peaks and the composition difference can be maintained throughtout the peak's migration down the column.

In practise, differences in mobile phase composition between different parts of a peak are generated by injections in "strong solvents" or at pHs that ionise the analyte. These differences are not maintained along the column because the sample plug mixes with the mobile phase on either side of it. The result is shouldered peaks or, rarely, incompletely separated peaks. These distorted peaks might lead to the belief that if only the conditions were optimised they would resolve completely and ions would have been separated from molecules by chromatography.

The only other circumstances which generate mobile phase differences across the width of a peak are very steep gradients. These could be stable along the column, and I would guess that they might lead to two resolved peaks at the end. I want to stress though that this is not a chemical separation in any meaningful sense of the word because which of the two peaks a particular ion/molecuale ended up in would depend entirely on the random process of which end of the sample plug it was in when the sample was injected to the column.

When chemistry fights with physics, physics always wins !

Peter

[ym3142]

some new results under same condition as those in my initial post except that the Buffer is 50mM sodium acetate; 35% MeOH,
PH 2.17 Rt 10.2 tf 1.2, pl 8143
Ph 3.36 Rt 8.9, tf 1.2, pl 7923
Ph 6.55 Rt 1.7, tf 1.2, pl 3070

I need take a bit time to think about all these and will be back.
Thanks for participation.

By the way,
Hans,
You got your point. Since it is in the water there should be enough protons and there should be no "deficiency" issue.

[HW Mueller]

It seems that Peter and ym3142 gave adequate closing statements.

(Except: The reason physics wins is that chemistry is a part of physics).