Acidic compound / Si column tailing

[JA]

A colleague is performing some work on N-hydroxysuccinimide (NHS) and maleimide derivitives 1-4, shown below. While the derivitives will, NHS doesn't want to stick by reverse phase and thus HILIC and normal phase is being looked at.

On a silica column, it is not certain if HILIC is actually "working", as retention is maximised at 100% ACN and decreases on addition of anything above 2% water. With 1% water in the MP sharper peaks are observed than without. Retention can be maintained when the MP consists of 5% IPA in ACN but decreases thereafter. Order of elution in all cases is 1-3-4-2.

On a used silica column (history not known) peak shapes for 1-4 are pretty decent. On a new column, peak shapes for 2 & 3 are poor, showing peak tailing and shouldering/tailing respectively. Addition of small quantities of formic acid to the MP sharpens the peaks - the RT for components 1-3-4 is unaffected while component 2 elutes more quickly. What is the likely cause of tailing for components 2 & 3 given that R contains no basic functionality?

[Bryan Evans]

If you are getting poor retention on a silica column - there's not much
you can do. Looking at those compounds - there's a descent chance they would be retained on an aminopropyl phase.

I recommend trying Unison UK-Amino. You can run under aqueous normal phase conditions - buffers to try would be ammonium formate, ammonium acetate, or acetic acid.

[Einar Ponten]

It seems that you have almost no retention in either RP or HILIC mode. You don't mention the mobile phase details, but you should add buffer and run at a neutral pH. You certainly need water in the mobile phase to build up a water layer on the silica surface.

Before proceeding I would recommend that you consult the following papers:

Retention Behavior of Small Polar Compounds on Polar Stationary Phases in Hydrophilic Interaction Chromatography
Y. Guo, S. Gaiki
J. Chromatogr. A, 1074 (2005) 71-­80

Investigating the Effect of Chromatographic Conditions on Retention of Organic Acids in Hydrophilic Interaction Chromatography Using a Design of Experiment
Y. Guo, S. Srinivasan, S. Gaiki
Chromatographia, 66 (2007) 223-229

Hydrophilic Interaction Chromatography [Review]
P. Hemström, K. Irgum
J. Sep. Sci., 29 (2006) 1784-1821

...and our "A Practical Guide to HILIC" that surely will give you some ideas.

[Mark Tracy]

I'm going to speculate that the new column has ionized silanols due to traces of adsorbed metals or ammonia. 2 & 3 would interact by hydrogen bonding. But since the ionization is uncontrolled, you get bad peak shape. The acid puts the silica surface into a uniform state of protonation, and cleans off any traces of adsorbed ions.

[HW Mueller]

Einar, the Hemström & Irgum article tries to make a case against water partition as the standard mechanism in HILIC. Anyway, since there are considerable differences in chromatography between HILIC columns with different surfaces the water layer would have to be different in these columns. In other words, it is the surface that directs the chromatography irrespective of what interaction mechanism is invoked.

Mark, what´s the new scoop on SiO-? I have not encountered any silica column that doesn´t show the ion exchange typical of SiO- at higher pH, even though the manufacturers claim there are practically no heavy metals left, certainly there was also no NH3 around. Now, are some special reagents required to cause silanol dissociation or do they just modify the dissociation?

JA, if a careful adjustment of pH doesn´t do the trick I am afraid that this is a case for two methods, or a two step (2 dimensional) chromatography.

[Mark Tracy]

All SiO2 phases dissociate, but given the original mobile phase, there is not any means of controlling the degree of ionization. There is no way to know what state the silica is in fresh from the stockroom. I know how hard we work to control metal contamination, but the best we can do is make sure that the columns meet some standard of cleanliness at the time of manufacturing (not at the point of use). That standard is not zero, just the best we can do as mere mortals. In any case, iron and aluminum are known to lower the pKa of silanols adjacent to the metal ion.

[Uwe Neue]

Hans,

Your question was with respect to silanol ionization at higher pH. A silanol on a high-purity silica has a pKa around 7. On a low-purity silica, more acidic silanols exist, with pKa's in the 3 to 4 range. For a hybrid packing like XTerra, the silanol pKa drops into the range of 9 or thereabouts, with few left around 7.

Mark,

I think we need to differentiate between metals that are buried in the matrix of a silica, and metals that sit on the surface. The ones in the matrix are responsible for the acidity of surface silanols, the ones on the surface do not affect this, but are responsible for binding with chelators. The ones in the matrix are a function of hte manufacturing process of the silica. The ones on the surface come from the instrument and the mobile phase. I agree with you that the latter ones are unavoidable, if you use a steel column on a steel instrument.

[JA]

Thanks for the replies.

The columns were both Phenomenex Luna Silica which I was lead to believe was a relatively decent type-B material which shouldn't exhibit the properties of having too many acidic silanols. Besides which, we know the tailing isn't due to uncontrolled ion exchange as the analytes aren't basic.

Can it be said that HILIC isn't working, or is it still possible that analytes interact with the silica support when partitioned into the polar layer covering the surface?

In the absence of HILIC, wouldn't hydrogen bonding be the predominant mechanism of retention for all of the analytes given the abundance of carbonyl groups and electronegative atoms?

To expand on the speculations already made, hydrogen bonding involving analytes which are H-bond donors may exhibit unwanted interactions with a silica surface on which the degree of ionisation was not controlled?

[Uwe Neue]

One column was used, the other one was new. You also observed changes in peak shape as a function of the addition of formic acid.

To me this means that the status of the surface was different between column A and B, and that the nature of this difference has something to do with the ionization of these surface silanols.

I would play with formic acid and ammonium formate until the tailing goes away.