On column degradation

[JA]

Maybe it was my choice of words which leads to this confusion. Not at any point did I suggest H+ were permanently removed from the mobile phase. I guess I didn't specify the degree of 'mopping up' - it was meant to indicate there are less than there should be.

My line of thought was that whatever equilibrium there is has been shifted by the basic nature of the amino ligand. Ok there's an equilibrium of ab/desorption of H+ from the surface, but it is shifted towards the conjugate acid NH3+ leaving a deficiency in the mobile phase within the pores. This solution is thus rendered alkaline. Sugars (glucose) entering the pores now rapidly interconvert and the two anomers are indistinguished by the stat phase.

Effectively then, yes, anomer interconversion is caused by the column.

Something like that.

[Uwe Neue]

A little note to clarify things: the concentration of amine on the surface is about 4 micromol/m^2. WIth a surface area of 350 m^2/g, this gives about 1.4 mmol/g. With a pore volume of the packing of about 1 mL/g or a bit less, this makes to about 1.5 mmol/mL, or about 1.5 mol/L of base.

With other words: the base in the pores is highly concentrated, and you have to work hard to neutralize the base...

[JA]

Is it therefore reasonable to say that the pH in the pores is somewhat similar to a 1.5 M solution of propylamine?

Is the pore volume significant? I mean, would the analyte experience an alkaline environment immediately upon entering the pore, or only as it approaches the surface? Saying that, I don't have an idea of the scale of a typical organic molecule with respect to the pore dimensions... how much room there is to 'rattle around in'.

[HW Mueller]

With a mobile phase of pH = 7 the stuff is already neutralized. A little bit of acid in the mobile phase and one has only NH3+? Or is this sugar stuff done as a nonequilibrium chromatography?
Anyway, what are the conditions of the LC that gives a single peak for the two anomers (pH of sample, pH of mobile phase, equilibrized?)? Why is it assumed to react only on the column?

[Uwe Neue]

Even neutral pH gives 2 peaks on a neutral column (say a diol column), and you definitely get 2 peaks under acidic mobile phase conditions. Only basic columns give a single peak.
The anomer reaction is a reaction that is happening all the time (mutarotation). The speed of the mutarotation is what counts. Under alkaline conditions it is fast enough to give a single peak.

The environment in the pores is definitely different from the external mobile phase, even on a partially neutralized amino column (due to silanols in the pores).

On a brandnew amino column, the pH in the pores is what you will get with a 1.5 molar solution of propylamine. (I asked an ant to put a pH meter into the pores, and send me a lab report; once I have it, I will publish it here.)

[JA]

Hans:
due to my possible stupidity, can you please explain how "the stuff is already neutralized"? The concentration of H+ in the mobile phase is orders of magnitude less than 1.5 molar. Due to the stagnant nature of the eluent in the pores the bulk of it isn't replenished very frequently / at all?

A typical amino- / carbohydrate-column application might include a mobile phase of 25:75 v/v water/acetonitrile with the sample dissolved in water or a mixture with acetonitrile.

Without having performed the work I'm going to speculate that you won't get a single peak in neutral (and perhaps acidic) mobile phases on other polar columns such as silica, cyano, diol, or in HILIC. Must be the reason the amino is chosen..?

edit: took so long to write and post, didn't see Uwe's contribution above.

edit2:
Uwe, now you have mentioned partial neutralisation of the amino functions by residual silanols.. can you provide a similar estimate of their concentration for me? This is where I have been ultimately trying to go with regard to an "acidic C18" (Resolve, Spherisorb ODS-1/2, Zorbax StableBond, etc.)

The joke about the ant and his pH electrode.. isn't this application just another example of a chemical probe. An indirect measurement if you will. They find accepted use elsewhere - we calibrate the difference between the apparent (measured by regular thermocouple) and actual sample temperature in our NMR spectrometer by measuring the distance between two signals in the spectrum of ethylene glycol.

[Uwe Neue]

JA, I don't know where you want to go with the silanol concentration in the pores. On an unendcapped C18, it is of the same order of magnitude as the amine on the amino column discussed above. Endcapping however, can hide the silanols quite effectively.

On a standard amino column, you also get a hydrolysis of the bonded phase, until there are enough silanols around to stabilize the pH and prevent the bonded phase from further dissolution.

[Bryan Evans]

I'll offer some "practical" comments:

- The different results from the 2 different phases makes sense.
One is more acidic than the other - so perhaps the bases were
retained more on the more "acidic" column (and co-eluted with the active).
The more "inert" phase had less retention for the bases, so they
did not co-elute with the active - resulting in different values for
purity (just a guess).

They are 2 different bonding chemistries - the fact that purity had
2 different results is good. (they may have had similar results under
acidic conditions). Here's a good reference:

Quick check on peak homogeneity
More Practical Problem Solving in HPLC, pg. 82-83
Stavros Kromidas

You can see it below as well (Cadenza CL-C18) is partially endcapped
and shows greater retention for levamisol than the fully endcapped
ODS phases (under neutral mobile phase conditions).

http://www.silvertonesciences.com/files/TI212E.pdf

- Yes, NH2 offers the best peak shape for NP of reducing sugars.
Below are maltose oligosaccharides on ODS and Unison UK-Amino
(doublet peaks occur on the ODS phase due to recognition of anomers):

http://www.silvertonesciences.com/files/TI347E.pdf

[JA]

Bryan:
thanks for the chromatograms and comments. Our group had made suggestions similar to yours about the increased retention for the additional early-eluting peaks observed on the Ascentis Express column but not on the Zorbax StableBond. We didn't get as far as investigating peak purity or additional alternate conditions before our QC/QA/Production/Safety head (yes that's right, all four) allowed the product out of the door.

Uwe:
I wasn't about to follow up with criticism of Waters' columns I was merely dropping their names as phases with a high degree of cation exchange capacity (which I have been equating to acidity). Here is where I was going with the silanol concentration...

We've discussed how amino- / carbohydrate-columns create an alkali environment in the pores based upon their ligand density. There's quite a lot of base in there. I was hoping for either expert opinion or analytical proof, perhaps via use of a chemical probe, of whether a C18 column with high cation exchange capacity (examples in my previous message) actually exhibits an acidic environment within the pores.

Going right back to the compound in my original message, let's say if it is stirred up in a mildly acidic solution and chromatographed in water/acetonitrile on a modern high purity, neutral C18, we see a decrease in apparent purity - the compound is acid labile.

If a fresh solution of the compound is prepared an analysed on an "acidic C18", can it display additional peaks as a result of degradation? This is what our chemists reasoned - the Ascentis Express caused the product to go off and is thus the more acidic column.
The chemists basically leave me like this..

[HW Mueller]

JA, if reaction takes place on the column, producing sharp (normal) peaks, then everything must have happened already at the start of the column (reaction is much faster than the chrom.).
Again, evidence has been mentioned in another chain that there is no bulk movement of liquid into the pores, but you must have diffusion, certainly nothing stagnant.

Generally, could it be that amino columns are so inefficient that they just don´t separate the anomers? Again, what are the pH´s of mobile phases involved?

Uwe, your ants are ions you inject?
Very roughly it may be permissible to assume a pKa of 9 for the amines, in spite of the influence of silanes + ?, and possibly a Donnan effect (don´t see the latter readily). If I am not off completely this would mean that at a pH of about 9 one has as many -NH2 as -NH3+. If the pH changes to 7 it means that the H+ concentration changed a 100 times, that should be enough to reduce the -NH2 to maybe 1%, etc.
Maybe this is a parallel situation: A 4.6?x250mm Atlantis Silica column on a mobile phase of 0. 01 M H2SO4 (pH between 2 and 1) will pass Cl- near to. A 0.00001 M NaOH phase (pH near 9) gives strong exclusion of Cl-. If I understood you correctly than this column has many more silanols than an amino column has -NH2.

[Uwe Neue]

JA, I did not see it as a potential attack on Waters. Resolve C18 is a non-endcapped column with plenty of silanols on the surface.

But this discussion triggered some thoughts for fixing a problem with an acid-labile analyte: there is a Spherisorb ODSB around which has a basic surface. It could be useful for preventing degradation of an acid-labile analyte.

However, I do not agree that an on-column degradation will get more peaks. If you have on-column degradation AND there is still a peak for the parent analyte, you will get lines connecting the retention of the degradant peaks and the parent peak (fronting or tailing). If there would be nothing of the parent peak around any more, and you only get the impurity peaks, then a on-column degradation would be believable, but this is not what you are describing.
Consequently, I agree that the second column just provides more resolution for all the stuff in your sample than the first column.

HW: the same amino column separates the anomers at acidic pH, but gives you 1 peak at alkaline pH. Has nothing to do with efficiency...

[HW Mueller]

OK, it is beginning to make sense now. The mobile phase has to be basic. That could mean that the basic mobile phase is just right for doing a fast interconversion, while the acidic one is at much less of an optimum. If, however, this basic phase gave two peaks on a column without amines than the evidence would be strong for (very fast) on-column reaction.
I checked some more into this anomer interconversion (this time in Hine, Physical Organic Chemistry). There is an interesting morsel: 0.001M
2-hydroxypyridine catalyzes the glucose conversion 7000 times faster than a 0.001M pyridine and 0.001M phenol solution, even though
2-hydroxypyridine has 1/10000 the basicity of pyridine and 1/100 the acidity of phenol. I wouldn´t be surprised anymore, then, that the amino column with basic mp would have the right amount, proportion, and position of SiO- and NH3+, that it would indeed have a strong edge over purely homogeneous solutions.

In regard to the original question it can be reiterated: In absence of further evidence the resolution explanation is the better. A rapid enough catalysis by one of several similar type of columns should be rare.

[Tom Waeghe]

I have worked with the Zorbax SB phases over the years under many different circumstances and what you're seeing in an unbuffered environment at "neutral" pH (pH 5-7 water plus increase in apparent pH due to ACN/water mixture) is failure of some of those basic compounds to elute. I've seen that many times and have also seen on-column degradation (hydrolysis) of labile compounds like sulfonyl chlorides. The SB column is intentionally not endcapped whereas the Ascentis Express C18 definitely is. That fact alone is enough to see big differences in elution profiles in an unbuffered environment such as ACN/water.

An experiment to try would be ACN/citrate or ACN/ammonium acetate buffer(10-20 mM) at pH 5.75 to 6.00 with both the Zorbax SB and the Ascentis Express column. The just-mentioned conditions are "not the best" for stability (ix-nay on phosphate and carbonate buffers at mid-pH for non-endcapped columns) for the SB column so keep the temperature below 35C and minimize the runs under those conditions for that column.