TFA at 215nm [August 26, 2004]

[admin]

By WK on Thursday, August 26, 2004 - 08:34 am:

I am using 0.1%TFA in HPLC water and 0.1%TFA in HPLC MeCN with a gradient 10%B to 40%B in 4mins and a C18 BDS column/215nm.
I have an inevitable ramp up in the baseline while the composition changes (I assume slight different concs of TFA are causing this).
I measured the UV abs at 215nm today of each solvent and worked out a dilution of the 0.1%TFA in MeCN with MeCN to get the same abs as the 0.1%TFA in water (9parts to 1part).
I will try this experiment tomorrow - has anyone tried this before? Did it work for you?

-------------------------------------------------------------------------------------------------------
By BT on Thursday, August 26, 2004 - 10:17 am:

Your ramp is due to the shift from water to acetonitrile and the differences in their UV absorbances. Most solvents have greater absorbance than water, so pretty much every gradient will do this, especially at lower wavelengths. To confirm, run your gradient with just MeCN and water. You'll probably see the same baseline.

-------------------------------------------------------------------------------------------------------
By A on Thursday, August 26, 2004 - 10:27 am:

A lot of people doing this type of gradient try to balance the TFA to get less of a rise in the baseline. Often it is 0.1% TFA in water against 0.085% TFA in the ACN. This can help.

-------------------------------------------------------------------------------------------------------
By WK on Friday, August 27, 2004 - 02:50 am:

Thank you BT and M.
Update:
My 9:1 dilution still had a baseline rise - not so bad (maybe 75%) compared to the undiluted rise.
And the change in TFA conc in MeCN had a small effect on retention.
I dont want to go back to phosphate - I will be forever unclogging/replacing capillaries!!
My next experiment will be to try and get the baseline flatter and then change the ramp to resolve my close eluting pairs.
Watch this space!

-------------------------------------------------------------------------------------------------------
By BT on Friday, August 27, 2004 - 07:40 am:

Are you using the increase in absorbance from the MeCN and the decreasing absorbance of the TFA (due to lowered concentration in the B mobile phase) to end up with a totally flat baseline? I might have to try that with a different buffer that is giving me problems.

-------------------------------------------------------------------------------------------------------
By WK on Friday, August 27, 2004 - 08:05 am:

Yes,BT.
I know that it probably won't work - but its worth a try! Something else in the separation will give - perhaps the buffering.
One of my peaks comes out on the "hill" and its not good integration.
I lose sensitivity at wavelengths above 215nm.

Let me know how you get on...

-------------------------------------------------------------------------------------------------------
By Chris Pohl on Friday, August 27, 2004 - 08:13 am:

WK

The reason why your first attempt did not completely eliminate the problem is that there are two components to the drift: the difference in the absorbence of your starting and ending eluent and an added portion coming from the stationary phase as the increasing solvent decreases the amount of TFA adsorbed on stationary phase. As you've already seen, the first part is straightforward to correct for but in my experience it's quite difficult to completely eliminate the second effect because of the details of the gradient start and end. Usually, if you take the approach of making a compensating decrease in the absorbance of the strong eluent you can get the beginning and ending level to match but then you usually have a curved baseline with a high point in the middle of your gradient.

-------------------------------------------------------------------------------------------------------
By WK on Friday, August 27, 2004 - 08:24 am:

Thanks Chris,
Changing the subject a little:
The TFA adsorbed onto the stationary phase - does it get removed upon neat MeCN washing or does it build up over time to saturate the column?
This is an interesting thread.

Next week I will try to see if I can improve so that my peak comes out on a flatish piece of baseline.

-------------------------------------------------------------------------------------------------------
By Chris Pohl on Friday, August 27, 2004 - 08:38 am:

WK

Washing with neat MeCN should remove the adsorbed TFA quantitatively from a reversed phase column with the exception of any TFA which is interacting "nonspecifically" with polar sites in the underlying support material (residual silanols, metal contamination, trace residuals from samples, etc.).

[HW Mueller]

Chris, do you have some ref(s) to share on a study of these strong TFA interactions? I am still trying to find out why my TSK-Gel Super SW 3000 column changed permanently after using TFA.

[Chris Pohl]

Sorry, I can't provide you any references on secondary retention mechanisms for TFA. This is just an observation.

[HW Mueller]

Thanks, Chris, good to know there is an observation behind this rather than just a feeling. I have seen this kind of thing with other substances, not sure yet what gives so it´s better not to relate details. But... your observation.... sure am curious whether its MS?

[Victor]

HW-your observation is interesting. I also managed to destroy a PLRP (polystyrene-DVB) RP column by using HFBA with it. I just assumed the stuff was retained irreversibly-but this is obviously not a silica -based phase!

[Newman768]

to HW Mueller and Chris Pohl:

I'm doing some analysis with TFA (based on ion pairing mechanism with a fluorinated phase) and only one (of 4) columns has survived more than some weeks. All columns were flushed by default (and very effective) with ACN at the end of the sequence to purge the TFA away.
The interesting stuff is: the only column which is surviving (since 2 or 3 years) was packed from another vendor.
Conclusion: same packing material packed from different vendors don't keep the same time!?!

Which makes me wonder if the degeneration (or changing) of a column could be related first to the (in my case stainless steel) column itself and only second to the stationary phase.
One (theoretical) reason I can think of: If TFA is reacting with the capillares of the lc system (it just needs very small inclusions of Mangansulide in stainless steel to let this steel corrode) and taking Chris posting from August 27, 2004 - 08:38 am into account the adsorbtion of TFA on the stationary phase should rise to a higher level and damages the column.

My question to both of you:
Is my theory too theoretical or does it makes sense in some way?

[Chris Pohl]

HW Mueller,

The experimental observation leading to the conclusion of possible secondary retention mechanisms was based on the washout profile using UV detection. If you bring a reversed phase column into equilibrium with a mixture containing TFA, passing a high content solvent mobile phase through the column (preferably using a separate system completely free of any TFA) should produce a abruptly declining trailing edge as a solvent elutes the TFA from the stationary phase. The presence of a persistent "tail" with a protracted washout under such conditions is indicative of a secondary retention mechanism.

Newman768

I doubt that TFA is all that corrosive as carboxylic acids are generally not very good chelating agents. The low pKa of TFA indicates that it is capable of attacking surfaces that are already undergoing corrosion but most likely it is not the corrosive agent itself. Chloride is the most commonly used reagent in HPLC which has significant corrosion potential depending upon the condition of the equipment, the pH of previous solutions passed through the equipment as well as the pH of chloride solutions being utilized.

[HW Mueller]

Chris, sounds logical if you saw that without other substances in there somewhere and if air didn´t get in at a mobile phase change.

Newman, I agree with Chris here. In my case I thought that the H+ (due to TFA) damaged the column, but then it should not have been partially reversible. Maybe it happened in your case, the pH being different when different columns were used, or the stat phase having been improved?

Victor, does your column also have a secret (proprietary) surface modification? Did you ever contact the vendor on this?

[Victor]

HWM- no, I don't think the column I was using (which comes from Polymer laboratories) has any strange modifications. Perhaps I should contact the vendor about it.

However, I haven't had much problem with using TFA in the mobile phase with this sort of column or with silica-based columns. Obviously, TFA is a strong acid and leads to lower pH solutions than some other additives. This may be detrimental to the end-capping of silica-based ODS columns. However, HWM I do not share your general bad experiences with this additive.

[Newman768]

to Chris:

then I'm asking myself, what else is the reason for the observed rapid column degeneration? With one or two columns I wasn't even able to get the expected rt and seperation of the compound of interest from the matrix.
I don't want to believe that it's only related to different vendors. This assumption would mean that the company which's producing this column material isn't able to pack a hplc column.
Or are fluorintaed phases generelly less stable under acitic conditions than a typical C18?
I need a rough method for the analysis and if the trick with fluorinated phases and TFA won't work, I have a problem. The compound(s) itself is too polar to be analysed under ordinary rp conditions.

[HW Mueller]

Newman,
what pH range does the manufacturer suggest?

[WK]

Sorry for breaking up the thread!!
Are you still out there BT?
I have sorted my problem out by reducing the concentration of TFA in the acetonitrile to approx. 0.07%v/v. The baseline is much lower at the end of the ramp and my peak of interest elutes on the top of the ramp - the integration is OK there. The decrease in buffering through the ramp is OK for the other early components of interest.
How did you get on with your other buffer?
Regards
WK

PS: The rest of this thread is very interesting.

[Newman768]

to HW Mueller:

the pH ange is 2 to 7.5
and pH of mobile phase is around 2.2.
Concentration of TFA is 0,1% or 0,05%, nothing unusual and the vendor is posting such conditions in the brochure.

[Chris Pohl]

Newman768,

First a question: when you referred to columns from two sources using the same packing material

same packing material packed from different vendors don't keep the same time

are you literally referring to media from a single source packed by two different vendors or do you mean two vendors supplying perfluorinated media which is nominally the same based on product description?

If it is the case of two sources both of whom are purchasing the same packing material from the same original synthesis source, the discrepancy could be batch related. But if you simply mean that two vendors are supplying nominally identical media, it's not surprising to find that there is a significant difference in stability from vendor to vendor. This is equally true with standard C18 bonded phases.

As to why you might see exceptionally poor stability under the conditions you describe relative to the stability normally observed for C18, there are three major factors to consider: first, the perfluorinated ligand typically used for making bonded phases is not as large (in terms of the length of the carbon backbone) as with C18. Shorter chains normally correlate to lower stability, all other things being equal. Second, the bonding density is typically lower in the case of perfluorinated bonded phases due to steric factors. Lower bonding density normally correlates to lower stability as well. And finally, there is a matter of stationary phase pH. Even if you have chosen conditions to produce a mobile phase pH which is > the minimum allowed, it is reasonable to expect that your stationary phase pH under the conditions you describe will actually be < what is allowed. The reason for this stems from the fact that the TFA concentration on the surface of the stationary phase can easily be significantly higher than the concentration in the mobile phase. This means that the local pH in the stationary phase can easily fall below an acceptable threshold. Furthermore, this phenomenon should be exacerbated by using perfluorinated media since TFA has a higher affinity for such materials than for C18.

So, all other things being equal, you should expect your perfluorinated phase to be less stable under the conditions you describe, leaving you with three options: raise the mobile phase pH (which should provide a corresponding increase in stationary phase pH), try a different retention mechanism or switch to a C18 phase. If the stability differences are really just vendor related, then obviously you should stick with the one that gives the best stability since likely they have optimized bonding conditions in such a way as to improve stability.

[Newman768]

Chris,

Fluofix (the phase I'm talking about) is made by Neos. The very unstable columns were bought from Venditor B in '03 and '04 and the "everlasting" Fluofix were bought from Venditor A in '00.
With the B columns I didn't get baseline seperation or anything that has something in common with batch-to-batch reproducibility.
Seems to me that between '00 and '03/'04 Neos has changed the column chemistry a little. Or do you have an other explanation?
Guess, I choose the easiest way and adjust pH.