Right-triangular peaks - What's the cause?

[peptidemetdev]

I've tried every reversed phase column I have in my arsenal, and none offered any significant improvement. I did, however, determine that the pore size makes the biggest difference.

On a whim, I decided to try a Hilic column I had on hand (ES Industries Epic HC, 4.6x250mm, 100 angstrom, 5micron, I believe). To my surprise, it gave much better peak shape than anything else! The TFA peak had a plate count of around 6500, and the resolution had increased dramatically

The only problem is that the TFA isn't really retained very long, and I worry about a negative injector peak interfering with results if I can't get the TFA to stick a bit more. Considering that it appears as though the TFA and Acetic acid are retained by something other than reversed phase interaction, could anyone recommend a mobile phase to try on the hilic column (I think it's just a bare silica phase, but I can't find any info about it) that might help retain the TFA a bit? I don't know anything about HILIC phases, and I wouldn't know where to start.

Thanks!

[SIELC_Tech]

You best choice is mixed-mode anion-exchange column. You will retain TFA and have huge capacity compare to your current approach. TFA detection using Primesep B2 and Primesep D columns was presented at one of chromatography Symposiums (Christer Andersson at AstraZeneca). Please provide me with email address and I will send you a reprint on analysis of TFA in drug substances. Also check these links:

http://www.sielc.com/compound_299.html (TFA and other acids with ELSD, you can use phosphoric acid and UV)
http://www.sielc.com/pdf/SIELC_November_2006.pdf (example of ACOH and formic acid with UV detection)

[Uwe Neue]

On a decent C18 column that is compatible with 100% water, you should have gotten good retention for TFA in a phosphoric acid mobile phase. The YMC ODS-A should have worked in 100 % water. Did you try this, or did you stick with your 2% MeOH? A better alternative is of course the modern Atlantis T3, which has been designed to work with 100% water.

[peptidemetdev]

On a decent C18 column that is compatible with 100% water, you should have gotten good retention for TFA in a phosphoric acid mobile phase. The YMC ODS-A should have worked in 100 % water. Did you try this, or did you stick with your 2% MeOH? A better alternative is of course the modern Atlantis T3, which has been designed to work with 100% water.

Sorry I didn't mention it, but I did try a phosphoric acid mobile phase (0.5% since I had some of that on hand, and I knew the column I was using was safe at such a low pH) on a Phenomenex Gemini, same dimensions as the YMC column. I didn't want to risk any other columns at a 100% aqueous because I wasn't sure if any of the other ones would do it. The only result I saw when I tried this was that the elution order changed, but the retention and peak shape did not improve. The Gemini had really poor peak shape and resolution to begin with, though, so I doubt it was the best column to try.

I don't believe that the retention mechanism is a reversed phase one. Columns with smaller pore sizes retained both TFA and Acetate better, with pore sizes of 300 resulting in the compounds not being retained at all. The HILIC column had the best peak shape for TFA and best resolution out of any column I tried, but the TFA was just -barely- retained past the void peak, whereas with the small pore C18 columns I'd see a minute or two between the void peak and the TFA peak. I tried a 5% H2O in ACN with 0.1% phosphoric acid mobile phase on the HILIC column, and that was an absolute mess, so I gave up on it.

I suppose it couldn't hurt to try a single injection or two on an ODS-A column, but YMC's website seems to suggest that only their AQ model is designed for 100% aqueous mobile phases. I'll try that on Monday and let you know.

[JA]

If you really need sensitivity, I would go back to my suggestion of finding a better column. And, if you switch to a smaller bore column (3 mm is my preference), you can slow the flow to 0.6 mL/min. If you are lucky, keep the same injection volume and you will see an increase in sensitivity. You may have to decrease injection volume somewhat if the peak shape is poor.

It could be mentioned, without wanting to derail the thread's main topic, that any improvement in sensitivity by switching to a narrower column [of the same stationary phase] is simply a result of injecting a proportionally larger volume of sample. This was covered in a recent extended discussion relating to column diameters. It seems to me that we don't need to buy in to this 'benefit' unless one is limited with respect to available sample solution.

That is to say that the same 'benefit' can be realised by the original poster (OP) by increasing the injection volume with his current methodology. The provision to "decrease the injection volume if the peak shape is poor" may put the OP right back where he started with the wider column saving himself only some mobile phase consumption. If the switch to one of his narrower columns improved the peak shape and resulting sensitivity it would probably be down to changes in both the physical and chemical properties of the stationary phase.

I've tried every reversed phase column I have in my arsenal, and none offered any significant improvement. I did, however, determine that the pore size makes the biggest difference.

You did this with the same brand of stationary phase?

I didn't want to risk any other columns at a 100% aqueous because I wasn't sure if any of the other ones would do it.

Ordinarily the columns don't break at 100% aqueous, they 'dewet', and can be restored with a high percentage of organic in the mobile phase. You could always check the accompanying column information pamphlet or manufacturer's website.

Columns with smaller pore sizes retained both TFA and Acetate better, with pore sizes of 300 resulting in the compounds not being retained at all. The HILIC column had the best peak shape for TFA and best resolution out of any column I tried, but the TFA was just -barely- retained past the void peak, whereas with the small pore C18 columns I'd see a minute or two between the void peak and the TFA peak. I tried a 5% H2O in ACN with 0.1% phosphoric acid mobile phase on the HILIC column, and that was an absolute mess, so I gave up on it.

This is again determined with columns containing stationary phase of the same chemical identity? Open question: Isn't it otherwise possible to be a result of improved mass transfer from the mobile phase?

I wonder if the retention of your two anions in HILIC mode would be improved by employing a more neutral mobile phase such as phosphate at pH 6.8+.

[Uwe Neue]

As JA has mentioned: columns do not break when you put them inot 100% water. They may just dewet, and you can rewet them with a short slug of acetonitrile. If the column does not work for you, the best option is to work with a column that has been designed for 100% water AND that gives you the best retention, i.e. Atlantis T3.

[peptidemetdev]

]This is again determined with columns containing stationary phase of the same chemical identity? Open question: Isn't it otherwise possible to be a result of improved mass transfer from the mobile phase?

Unfortunately, I do not have any columns that were otherwise identical aside from the pore size. However, I did have multiple different C18 columns and the pattern was clear. TFA/Acetate were not retained on columns with a pore size of 300 angstrom, and they were retained on columns with a pore size between 80 and 120 angstrom. I tried the HILIC column, certain that it wouldn't work, simply because it had a pore size of 100 angstrom. Note that the retention says nothing about the peak shape. The Acetate peak was typically pretty good (if it was resolved from the TFA peak, it generally had a plate count of above 13,000), while the TFA peak frequently tailed, or had that overloaded look to it that I mentioned in the OP.

I wonder if the retention of your two anions in HILIC mode would be improved by employing a more neutral mobile phase such as phosphate at pH 6.8+.

I could certainly try that as well. I'm very curious as to why I see any retention with an almost completely aqueous mobile phase on what is essentially a normal phase column, especially with the Acetic Acid which is retained just as well as it was on the RP columns. Any idea?

[HW Mueller]

The bad retention in the 300A columns could simply be due to a likely lower surface area?
If you use about a neutral pH with a normal phase silica based column you have to fiddle with the ionic strength in order to avoid contributions by exclusion.

[JA]

On the topic of ion exclusion: If I remember correctly, and the information hasn't been superceded, we are informed by manufacturers that peak tailing for bases on a polar embedded phase is significantly reduced due to the 'shielding effect' of a small water layer formed just above the silica surface.

Could we expect that a similar shielding effect will be present in HILIC mode negating any like-charge repulsion?

I will add however, that a few applications I have seen for this approach use a higher buffer concentration (50-100 mM in the final MP) than I would of typically used. Maybe this just seems different for me as I'm usually making 5-20 mM for MS compatibility.

[peptidemetdev]

The bad retention in the 300A columns could simply be due to a likely lower surface area?
If you use about a neutral pH with a normal phase silica based column you have to fiddle with the ionic strength in order to avoid contributions by exclusion.

Actually, I think the higher pH helps to retain the TFA in a significant way. By using a pH 6ish 0.1M ammonium phosphate buffer, the TFA's retention on the HILIC column increased to a k value of around 2, and the plate counts were in the range of 15000 or so. The only problem is that the acetic acid was no longer retained at all. If I could come up with some way to increase the silanol activity without bringing the pH above 3.5, I think I might have something to work with.

I found a Polysulfoethyl-A column in the fridge earlier, and I tried playing around with that. I couldn't get anything to work, though, presumably because the pore size was too large (300A)

I think this project is probably not worth pursuing any more, but I found it all very interesting, and I'd like to thank all of you for your suggestions/explanations!

[HW Mueller]

As already mentioned before, an increase in SiO- could give rise to exclusion of anions, so why would you want to increase "silanol activity"?

[peptidemetdev]

Oh, you're right. I wasn't thinking clearly for a moment there. For some reason, I thought the charges would be complementary, but you're right, they wouldn't be.

So that leaves me to wonder why an increase of pH led to better retention for the TFA. Maybe it was just because I used ammonia instead of sodium as a cation for the buffer? I know flurocarbons have some really bizarre properties, though, so I wouldn't be surprised if there were some sort of weird dipole bonding going on between the silanols and the fluorines on the TFA. Perfluorohexane dissolves oxygen so well that you could actually breathe it in it's liquid form (although I'm sure it wouldn't be very much fun to get it out of your lungs if you wanted to go back to breathing gasses).

[Victor]

What conditions are you using for the HILIC separation?

What is the column? The mobile phase?

[danko]

Hi peptidemetdev,

Try and read SIELC_tech’s post. If you don’t want to experiment with mixed-mode, then just set up an ordinary anion exchange system, as I mentioned earlier, and you will be rewarded.

A bit off topic: Smaller diameter columns (everything else equal) offer higher efficiency and thus higher sensitivity.

Best Regards

[peptidemetdev]

What conditions are you using for the HILIC separation?

What is the column? The mobile phase?

The column is an ES Industries "Epic" HILIC column, 4.6x250, 5µm, 100A (I think). I can't figure out what the stationary phase actually is, but I -believe- it's nothing fancy. Just bare silica, or diol, or something.

The mobile phase that fully retains Acetic Acid and barely retains TFA is 2% MeOH in 0.1M sodium phosphate (pH of around 2-2.5). Switching to 0.1% phosphoric acid resulted in the TFA not being retained, but didn't really effect the acetic acid peak at all. Switching to a 0.1M ammonium phosphate with a pH of around 6 resulted in very good retention with the TFA and no retention of the Acetic Acid. I tried doing phosphoric acid in more of a hilic mode (5% H2O, 0.1% H3PO4), but that didn't work at all.

We used to use a 1mN sulfuric acid mobile phase on a Hamilton X300 IE column for quantification of acetate counter-ion, but we switched to this methodology in order to analyze TFA as well. The method using a C18 column actually seems to be a whole lot more robust (better peak shapes, better resolution, etc.) than trying to TFA and Acetate on the X300. Granted, though, we didn't really explore any method development and just used the method supplied by Hamilton for analysis of acetic acid. My goal was to try and increase the sensitivity on the C18 method (which, actually, is already more sensitive than the ion exclusion method, probably just due to the fact that the flowrate is reduced by 1/4) without purchasing any new column.

In doing some routine work this past week, I discovered that the peak shape for the TFA was actually very good (TP > 10,000) at the lower levels of my concentration curve. My goal was to create a method that would allow analysis of both acetate and TFA salts, and would be capable of detecting residual TFA in an acetate salt at the 100ppm level, but I don't think I'm going to be able to do that. Instead, I should be able to develop a separate method for acetate salts by engineering a special standard curve.

danko: Purchasing a new column (especially one as expensive as a mixed mode column) is, unfortunately, a tough sell right now. We actually have two anion exclusion columns from Hamilton on hand right now, an X100 and an X300. Hamilton doesn't, unfortunately, have any applications on their website that include both acetate and TFA, and it really seems like they have really poor capacity and peak shape (the peaks are symmetrical, but extraordinarily wide).