On column degradation

[JA]

Without expert knowledge at our workplace I'm turning to the forum with the following puzzle. In another department at work, a colleague of mine has been looking at the HPLC analysis of the final product shown in the attached scheme.



Without having performed any of the work myself, and without seeing the actual chromatograms, the information is currently limited thus:

Using a water/acetonitrile isocratic method, a reasonable purity (>95%) is indicated on a Zorbax StableBond C18 column. The same mobile phase, when used on an Ascentis Express C18, displays additional early-eluting peaks which reduce the apparent purity by approximately 10%. Gradient runs on the Zorbax give results consistent with the isocratic data, while such a 'neutral' gradient has not been utilised on the Ascentis Express.

On the Ascentis Express column, the usage of acidic (H2O/MeCN/TFA) or basic (amm bicarb/MeCN) mobile phases appear to promote degradation and lead to a lower apparent purity than observed in any of the preceding work.

Said colleague reasoned that the Ascentis Express possesses more acidic (Si-OH) and basic (Si-O-) sites in comparison to the Zorbax StableBond support, particularly since production of the latter employs the use of bulky endcapping reagents to prevent C18 ligand hydrolysis. A second colleague then assigned Ascentis Express to be the more acidic stationary phase.

Is there an expectation that an acid labile compound would display a lower apparent purity on a stationary phase possessing a larger degree of surface silanols? Are silanols acidic in the sense they can donate a proton to a particular analyte being chromatographed? Do we even know which of the two columns would have more surface silanols and do we know which is likely to have more higher energy sites?

I appreciate that the information provided doesn't present a complete comparison of the two columns (I might be able to dig more up if it will assist) but I'd like to hear your thoughts.

[Kostas Petritis]

Probably the caffeine/phenol Tanaka's test will be able to indicate which one has more residual silanols in the surface. Having said that your collegues sound like organic synthesis people on denial... oh look this column shows lower purity so there must be something wrong with this column. On a similar note, organic synthesis people do not like using ELSD for purity analysis because it also shows compounds with non-chromophores thus decreasing the purity...

[HW Mueller]

Also, degradation on a column produces wide, ugly peaks, unless the degradation occurs and completes immediatly at the start of the column.

Silanol activity can be controled via mobile phase pH.

[danko]

As I see it, there is a lot of acid (TFA) in the case of the “Ascentis Express columnâ€

[tom jupille]

I'll repeat HW Mueller's response:

degradation on a column produces wide, ugly peaks, unless the degradation occurs and completes immediately at the start of the column

Re the relative silanol activity of the two columns, you can look up their selectivity properties in the PQRI database on the USP web site:
http://www.usp.org/USPNF/columnsDB.html
Here's the comparison (A is a measure of hydrogen-bonding acidity; B is a measure of hydrogen-bonding basicity).

Zorbax StableBond 80A C18: A = 0.264, B = -0.001
Ascentis Express C18: A = 0.023, B = -0.052

That's not surprising considering that the StableBond technology is now 15 or 20 years old, versus a couple of years for the Ascentis Express (newer technology columns in general tend to be of higher purity and have fewer "active" silanols).

It looks to me like those impurity peaks are real. Which, by the way, supports the argument that you should never rely on a single HPLC method for purity!

[JA]

Thank you for the replies so far.

Kostas:
Your take on the organic synthesis colleagues is unfortunately quite accurate, but it's the way it is when you work in a QC/Analytical Development department. Historically we have had a similar experience when upgrading our NMR spectrometer to a significantly more sensitive high field magnet: "where did all these extra peaks come from?.."

Hans:
You triggered my memory for the oft-cited proline rotamer interconversion. In my younger days I undertook a short study of the degradation by hydrolysis of 'activated' expoxides on a couple of stationary phases; Spherisorb ODS2 and a HiChrom phase (RPB, I think). A ring-opened degradant was seen only on Spherisorb, although I cannot recall what the peak shape looked like: nice and sharp or broad/bridging.

danko:
I might not have been clear in my original message; the comments regarding use of acidic (TFA) or basic (amm bicarb) mobile phases was only to indicate the analyte appeared to dislike conditions away from pH neutral. The underlying issue, and reason behind my request for comments, was the observation of dissimilar results on the two columns in the absence of additives, i.e. water/acetonitrile only.

Tom:
A brilliant link! Embarrisingly, I have played with that tool a good while ago, even read a 2002 "Column Watch" article in LC/GC America guest authored by Snyder & Dolan introducing the selectivity function Fs, and still forgot about it
We will continue to reason our side, now armed with more numbers.

What I'm still unclear about is whether the accessibility of surface silanols, denoted by the A-term, is equivalent to the "acidity" of the stationary phase in the sense of it being a hydrogen ion donor? If we believe that degradation of the analyte is promoted by hydrogen- or hydroxide- ion is it expected when there is only an association between the silanol hydrogen and the analyte?

[tom jupille]

What I'm still unclear about is whether the accessibility of surface silanols, denoted by the A-term, is equivalent to the "acidity" of the stationary phase in the sense of it being a hydrogen ion donor? If we believe that degradation of the analyte is promoted by hydrogen- or hydroxide- ion is it expected when there is only an association between the silanol hydrogen and the analyte?

I'm not enough of an organic chemist to comment on the degradation mechanism (yes, that has to do with a memorable ether fire in my first semester of graduate school ), but:
the "hydrophobic subtraction" model was developed by correlating differences in relative retention with known parameters of probes (analytes). A column that retains basic compounds more strongly (relative to neutrals) is presumably acidic and has a high A value. A column that retains acidic compounds more strongly (again, relative to neutrals) is presumably basic and has a high B value. The math is more complex than this, but you get the general idea. Interpreting what the parameters mean in other types of interactions (i.e., other than chromatographic retention) is speculative at best.

[JA]

Here I come from another angle. We are told that amino columns are used for sugar analysis as the alkaline environment within the pores promotes rapid anomer interconversion resulting in a single peak. I'll take that to mean the amino functions mop up H+ from the mobile phase.

Conversely then, I ponder the possibility of a locally acidic environment on stationary phases comprising a greater degree of low-pKa silanols.

Following from your earlier post Tom, would you be kind enough to set me straight on the A and C-terms from the hydrophobic subtraction model? I'm resistant to equating increasing values of the A-term with column acidity as it has been described and depicted as increased retention of bases due to hydrogen bond association with an analyte. I picture the silanol hydrogen remaining as Si-OH. Increasing values of the C-term for cation exchange give me the impression that this is where a proton can be abstracted, Si-OH ⇌ Si-O- + H+ shifted to the right. Isn't C therefore representative of 'column acidity'?

I completed tabulating the data for the columns in question, plus a few others we use, for information. The largest difference between the two appears in the C-term at neutral pH.

[HW Mueller]

Isn´t the anomer interconversion acid catalyzed, like eterification, with base it hydrolyses (here, opens the ring)? If correct that would mean the acidic mobile phase equilibrates the anomers so fast (faster than the chrom.) that they show as one peak. Also, one usually does HPLC with the column/mobile phase in equilibrium, there should not be any mop up.

Others are better qualified to answer the question regarding the C term. It is curious, though, that results at pH 7 are remembered, but applied at pH 2 where they probably have no bearing.

[JA]

Apparently the equilibration of glucose anomers by mutarotation takes hours at RT and, while catalysed by both acid and base, the process occurs more quickly in dilute alkali (pH 10) by an order of ~5000 times*. I won't shame myself by posting my sketched mechanisms but I think both proceed via a ring-opened intermediate which has free rotation.

Hans:
With regard to the chromatographic process occuring with the column in equilibrium with the mobile phase, I recall reading that things are stagnant within the pores.

I was confused by your comment regarding pH 7 and pH 2, could you please elaborate?

*Glycoscience: Chemistry and Chemical Biology, 2001, ISBN: 3-540-67764-X

[HW Mueller]

The last first: It seems that comparisons at a certain pH, such as this, are often used in aguments at other pH where it doesn´t hold. For instance, I doubt that there are many SiO- around in any column at pH 2.

Do you have a ref. on this base catalysis? I have trouble picturing a mechanism.

There was a discussion not too long ago that surprised me as it seems clear that there is no bulk movement of mobile phase into the pores. Certainly there is diffusion into them, so that they are in equilibrium with the rest of the column. The environment, especially near the surface of the pores must be different than in the rest of the mobile phase, but mopping up anything would require strong disequilibrium. I don´t see from where the energy for that would come.

[HW Mueller]

OK, I found a mechanism proposal for basic anomer conversion in Metzler, Biochemistry. The open chain variation is deemed to pick up a proton to be able to reclose. What I said above regarding a problem with closing up in base holds probably only for very strong bases.

[tom jupille]

JA - your interpretation matches mine (whether it's correct or not is another question! ). Either way, though, the StableBond is more "acidic" than the Ascentis Express, and those "extra" peaks are far more likely to be real than artifacts from on-column degradation.

[JA]

I'm a bit stuck now with the comment indicating the mobile phase inside the pores is in equilibrium with the bulk solution travelling down the column - do we mean that there is no possibility to have a different pH inside and outside? Experimentally, isn't the glucose analysis evidence enough?

For kicks here are the sketched mechanisms I referred to earlier:

[HW Mueller]

The surface should certainly be different than the bulk mobile phase, but everything is very close to equilibrium. adsorption of an analyte will probably change things again, but that is also an equilibrium. There could also be some buffering when some acid is injected, as an example, but mopping up in the sense that H+ will be permanently removed??
Did I miss the evidence for anomer interconversion by the column?