HPLC Silica column problems

[rekuci]

I'm trying to quantify a few nucleoside analogues and their enzyme reaction products (ribose-like and a purine), in PBS pH 8.0. The enzymes (mine anyway, not sure about the other user) are in picomolar quantity and are chloroform extracted as part of the sample preparation.

A method has already been developed in the lab for doing this, but I think previously a different column was being used (reverse phase). There were many problems with this column so a new one was ordered by the other user of the instrument. It is a Waters Spherisorb 5 um Silica column. Not sure if this is ideal for this application, but it's too late now, and I just started using the instrument a couple days ago.

There are already serious problems with the new column. Retention times have quickly reduced and peak resolution is now terrible, after less than a dozen injections. The mobile phase is 70% water/30% acetonitrile. It has been flushed for many hours with water and cleaned once without flipping the column around with a acetonitrile > acetone > hexanes > acetone > acetonitrile gradient. None of this has helped.

Are the analytes possibly adsorbing too strongly in the column? Can anyone suggest an appropriate cleaning or regeneration method? And is this mobile phase appropriate for this type of column? Thanks in advance for any suggestions.

[Uwe Neue]

I am surprised about your mobile phase. Why did you select this composition?

On this column, the retention mechamism could be ion-exchange in your mobile phase. If this is the case, you need to control the pH by adding a buffer to get reproducible results. I woudl start with 20 mM phosphate buffer pH 7, and then one can improve the sepration by chanign one or the other of the mobiel phase paraemters.

[rekuci]

It was what the other user selected, and for what reasons I don't know. I have no idea how to go about selecting a mobile phase for this system, but it sounds like it should be something more nonpolar, such as hexane/methanol? It is highly possible that the only reason for this mobile phase selection by the other user was that it was what she used in the reverse-phase column. In which case it wouldn't make sense to keep it the same.

I should put the disclaimer that this is my first experience with HPLC, all of two days ago! I'm trying to cram on the technical aspects but so far haven't found any online literature on mobile phase selection for polar columns. Next step is to contact the company of course...one of the test chromatograms that came with the column does use this mobile phase, but the second test chromatogram uses nonpolar mix.

20 mM PBS is probably a good idea, right now it's only 10 mM in the sample.

[agrp87]

silica is the normal phase, acetonitrile is primarily used in reversed phase. Use Methanol or chloroform instead of acetonitrile. Ask an organic chemist what they use on their flash columns. I would probably start out with 75:25 MeOH:CCl3.

You want non-polar for reversed phase and polar for normal phase.

[Uwe Neue]

This is complicated....

I think that for compounds that are as polar as your analytes, multiple retention mechanisms can be made to work, and the silica column may be actually a good choice.

I suggest to do the following: use a pH 7 phosphate buffer of about 10 to 20 mM concentration. Start with your mobile phase (30% acetonitrile) and inject. See where the peaks are eluting. Then go to 50% acetonitrile and to 70% acetonitrile. Tell me what you see (retention changes and resolution changes), and we will take it from there.

If you know, it will be good if you could report what the difficulties were with the reversed-phase system that had been used before.

[Uwe Neue]

PS: I should have mentioned that you can contact me at the email address below if your are uncomfortable posting the details of your results on public board.

[rekuci]

Thanks for the suggestion.

The other user says this column and mobile phase worked well in the past and these elution problems are more recent. There might be some kind of contamination in the sample matrices.

Any comments on Supelco's "silica column regeneration solution" containing glacial acetic acid, dimethoxypropane, and methylene chloride? Would this be appropriate for a silica column used in reverse-phase mode? The 10 minute flush they claim is required would be far faster than a solvent gradient cleaning!

[Uwe Neue]

I do not see why you would need a column cleaning on the silica column. I think that you will need controlled conditiosn to get retention. Under your mobile phase condition, I do not expect anything to be retained except by ion-exchange. Thus you need to work out a mobile phase that gives you a controlled ion-exchange.

Try the mobile phases that I gave you earlier and let me know what is happening.

[rekuci]

The reason that so much effort has been spent with cleaning is that supposedly this column worked for this analysis previously (not done by me), the components eluted with up to 5 minutes between. But I was told that the peaks get closer together and elute faster the more injections you run, so after a dozen or so injections, a water flush and eventually a full solvent cleaning was necessary. But it doesn't seem right to me - if the conditions were appropriate, shouldn't retention time not change?

And then what's weird is that this new column is having these problems right from the start. We supposedly have guard columns on order (it's hard for me to keep up, since I'm only a grad student in this lab part-time).

With this additional background, does this sound more like contamination now?

[Uwe Neue]

You managed to confuse me utterly.

I understood that you started off with some different reversed-phase column, and that you could not get this column to work.

You also said that you are currently using a Waters Spherisorb silica column. This is NOT a reversed-phase column. You may get your separation to work on a silica column, but not with the mobile phase that you are currently using.

Now, tell me what the old column was that you could not get to work, and what the mobile phase was that was used on the old column, when the retention changed.

[rekuci]

Sorry to confuse, here is what I gather the history is:
- Reverse-phase column didn't work well for this analysis (column #1)
- Silica column (Waters Spherisorb) worked well with the water/acetonitrile mobile phase for many months but always with decreasing retention times with each injection, needed flushing and cleaning every dozen injections or so (column #2)
- Above column became unusable so purchased a new identical silica column, but it is already having problems from the start with too low retention, and flushing and cleaning are now not working. (column #3)

The other user calls this using a normal phase column in 'reverse phase mode', to use a polar solvent with a polar column.

[Uwe Neue]

OK. Now I understand better.

There are several issues with using a silica column in your mobile phase with ionizable or ionic compound. The first one is that you need pH control to get reproducible results. The reason for this is that you are using the silica as an ion-exchanger. To get controlled retention, you need to control the ionic state of the silica. It also requires that the silica is in a form that permits ion-exchange. You should be able regenerate it by washing it briefly with an acid, for example a 0.1 molar phosphoric acid solution.

If you indeed use 70% aqueous and 30% organic, I do not think that hydrophilic interaction contributes to your retention mechanism. However, please verify that this is indeed 70% aqueous and not 70% organic. This would change this completely.

[rekuci]

Ok, I guess an acid wash makes sense...IIRC, silica has an isoelectric point of pH 2, so if anything is ionically sticking to it, you'd have to acid wash it. The pI's of 2 zwitterionic molecules present in the sample mix are 5.65 and 6.05, so at least at the sample pH of 8.0, both silica and the molecules should have a net negative charge and should repel each other. But I doubt the 5 µL of sample volume will have an effect on total pH in the column.

Should the HPLC grade water used for the mobile phase be buffered? It isn't right now...which doesn't make sense to me when thinking about this more.

[Uwe Neue]

Let us first wash the material with acid, then run your normal mobile phase. I expect an increase in retention as you have seen on the older column.

If this looks good, and the retention changes to quickly to shorter retention, I would recommend that we start playing with a buffer to control retention. Since you are telling me that two of your compounds have a pI around 6, it might be worthwhile to use an acetate buffer at pH 4.5. At what wavelength are you detecting?

[rekuci]

I'm going into the lab tomorrow and I'm going to try an acid wash, but in case we don't have phosphoric acid (quite likely), acetic acid should also be ok? The silica column regeneration solutions are largely acetic acid.

One of the compounds I'm detecting is unstable at alkaline pH (it actually degrades to one of the other compounds I'm trying to identify), but hopefully should be ok over the course of just a run through the column. So the idea is to have the pH at slightly less than the pI so that the silica is (-) and the analytes are (+), increasing retention times? It's no secret what I'm trying to analyze...I'm looking for S-adenosyl-L-homocysteine and methylthioadenosine and their enzyme reaction products, S-ribosyl-L-homocysteine, adenine, and methylthioribose. It's not necessary for everything to be eluted, but the more the better because I'm trying to determine the enzyme kinetics.

Detection is at 210 nm and 260 nm. S-ribosylblah absorbs at 210 but not 260, which is useful in identifying this compound (which is not commercially available, so no standard). Acetic acid has a UV cutoff of 210 nm, what does that mean exactly and does that mean that 210 nm detection is a problem?