Loss of selectivity in Xbridge columns at hight pH

[vikmnilu]

Hi to everybody!

Recently, we purchased a Xbridge C-8 4.6*150 mm 3.5µM particle size. I found it almost perfect for my separation purposes, which are pyrene and its metabolites produced by some different organisms.

At the beginning, the separation was great, I only found two peaks that I was not happy with and these days I have been trying to separate them better so I could colect them with the fraction collector in order to measure them later on in Scintillation counter (we also use radiolabeled pyrene as a tracer).

Well, what it was a decent, but not enough good, separation of these two peaks has converted in only one peak in the last runs, and I seriuosly do not know why.

The conditions I am using are

Mobile phases
A=Ammonium acetate 10mM +Triethylamine (TEA) 0.1% pH=9.5
B=Acetonitrile
I have with different gradients and it does not work in any of them.

I am using Agilent 1100 series and we have changed the autosampler recently, but that I think should not influence.

History of alst runs: Used high ph (9-9.5 throughly in all the runs since we bought it, however, not so many runs). Few days ago, I run the same samples at ph=5 with another Zorbax C8 XDB column coupled to this one.
Then I switched to ph=9 again and I have been using ph=9-9.5 since then.

At Waters they told me that if you switch from basic to acidic or neutral pH and then back to basic you loose some selectivity, but if I understood correctly that was only if ion pairs are used... and TEA is!

So I may have answered my question... but Any idea of how to return to the original selectivity? I regret so much to have returned to pH 5 for Only a day!!!! (7 hours run approx!), but hei, if someone know something about this or has experienced something similar, I would appreciate an answer.... it is a 3 months old column and I basically want to base my third paper in that separations...

Well, Thanks in advance!!!


Victor

[Mattias]

You should be able change pH, and still keep the original selectivity when you return to the first conditions.

I have had a lot of discussions with Waters regarding a change of selectivity of a similar column, but in an RP-Shield version. Waters have admitted that the column contains some residues that work as ion-exchangers. The C18 part of the column is stable up to very high pH values, but these ion-exchange groups are not. I am running at pH 7, but I still see that I loose retention of some peaks as the column gets older. The other peaks are unaffected.

It has probably nothing to do with your problem. You can easily test it by lowering the mobile phase pH from pH 9.5 to e.g. pH 9.0. It your selectivity returns, you may have the same problem.

The quickest and probably cheapest solution is to order a new column...

[HW Mueller]

Maybe you are just injecting too much gunk?

[danko]

Hi Guys,

I can recomend very stable columns - even at very high pH (as well as very low).
Drop me a mail if you'd like to know more about it.

Best Regards

[vikmnilu]

Hi there and thanks very much for the answers!

To be honest, buying another column is not an option, at least yet. Sad but true. With the 23% taxes in Finland the column price went up to 700 and siomething when I expected something like 500 euros.

I have been using in most of the cases guard column, it is not Waters, unfortunately this would had raised the bill quite much. It is an agilent XDB C-18 (yes, different phase) thant as far as I know bears until pH 9 (I hope this is not a real problem).

I think that, still with the guard column, I may have injected too mucg gunk, as HW Mueller suggest, but I have no idea... I am these days going to try to regenerate the column and see if the selectivity, at least partially returns.

Thanks for the effort again.

Victor

[rwang]

Since you go that far using Zorbax XDB-C18 for guard column, you might be better off using Gemini C18/C18 NX (at least they will tolerate higher pH).

Changes of selectivity after high pH exposure also happens to other columns, not just Xbridge. If you can afford a new column, buy another Xbridge. If you can re-develop the method and want to jump between different pH values, buy an XSelect, as Waters claim its CSH columns will not change selectivity after exposure to different mobile phases. Please note XSelect columns have a narrower pH range (1-11 for C18) than XBridge.

More information can be found in this white paper:

http://www.waters.com/webassets/cms/lib ... 3469en.pdf

Disclaimer: I don't work for Waters.

[danko]

Why (the mechanism behind) would temporary pH change (fx. form 2 to 7 and later back to 2) induce irreversible selectivity change, unless we're talking about extreme pH values - i.e. column material degradation? Also, what would be the “save” pH alteration that does not cause the assumed selectivity change?

Best Regards

[vikmnilu]

Hei and thank you all for your replies.
Danki, I have no idea why that happens, but it does happen and certified by Waters, although they told me that only if you use these ion pairs...but I have no idea of other information regarding to that. Wierd and annoying stuff anyway.

The point is that we do not have money to buy another column. I tried to regenerate the column last week with a through consecutive running of organic solvents (ACN->Isopropyl alcohol->Dichloromethane-> Hexane and back, for two days in total. I have not tried yet, but I hope is that I injected too much dirt, so I can get the separation again.

Fingers crossed!

[ym3142]

this may not be related to your current problem: Why you have TEA in your mobile phase? One pKa for ammonium acetate is 9.25, right? It will provide the buffer capacity.

[cristogalan]

Hi to everybody!

vikmnilu, how do you solve your sample? we are working with BEH Acquity columns from waters and we had problems with some columns because we lost selectivity and efficiency too. the Waters technician asked us about how we solved the samples and the problem was that our gradient start with a high portion of aqueous mobile phase and we solved the samples with 100% organic phase, usually Acetonitrile. So, now we are improving the sample solution in order to avoid this problems. Perhaps you have the same problem with your Xbridge column.

More over, why do you use TEA in your mobile phase? remember that TEA modify your column and if you use the column for another aplication without TEA and after you buy a new column and try to reproduce your conditions,without TEA, perhaps the selectivity will be totally different, because the TEA's effect.

Best regards

[kplc]

(FYI) Currently using an XBridge C-18 4.6*150 mm 3.5µm particle size and the following mobile phase:
(A) 0.1% NH4OH + 0.55mL HCl per liter (pH9.5);
(B) acetonitrile.

Using the above method, if I install a brand new column and inject a comparison sample- then imp X (~0.6 area%) elutes at RRT 1.08. If I then stop and wash that column at low pH (0.1% H3PO4) then at high pH (0.1% NH4OH), then repeat the injection (pH 9.5 mobile phase)- imp X will elute at RRT 1.2.

As a general result of this wash/conditioning, separation is gained between RRT 1.00 and RRT 1.35. Yet the impurity peaks before RRT 1.00 and after RRT 1.35 elute in exactly the same position as they did when the column is brand new (~ 40 peaks total in sample). This also holds true for the columns purchased in 2009, 2010 and 2011. They all may be used interchangeably with unremarkable variation in RRT values per sample.

We inadvertently discovered this as our XBridge columns had to be initially used to switch back and forth between low pH and high pH methods. Another unexpected benefit was that since the wash procedure was initiated, blanks are very clean.

Not what we planned but we were fortunately able to work with it.

[vikmnilu]

Hi again
thanks for the interest shown and the replies.

ym3142 and cristogalan: I use TEA after a lot of tries without different methods, i found out that, for this special samples (as I said) some aquatic worm exposed to pyrene biotransforms it into several metabolites (I am able to spot up to 4 in FLD, but may be more at lower concentrations). Two of them eluted together, resulting in a horrible peak shape with a huge tailing with the normal amm Acetate 10 mM ph5+ ACN. After using TEA (and pH over 9) in Amm Acetate eluent, this problem dissappears and two better resolved peaks appear. I was not able to achieve this with any other conditions, e.g. TEA and Ph 5, not TEA and pH 9-9.5, etc. I guessed that the metabolites are interfering too much with the silica phase of the column (i used at that time a Zorbax XDB C8) and that provoked the tailing. My guess with TEA is that it would avoid these extra interactions and allow the metabolites to elute alone and better defined Retention times. That is why I bought the Xbridge for better pH conditions.

kplc: I also use Methanol as sample solvent, I have thought about what you said and some people told me about it. Do we really think it is a critical step? There are thousands of articles published where the solvent is organic and the gradient starts at high aqueous mobile phase, and I have been running these samples for quite a long time with Methanol as a sample solvent... I may want to try that, will probably do it for the third article (PhD student that needs 4 to graduate), but I do not think that it will change things critically, as I saw the separation happening with MeOh as solvent. Any other opinions about this?

Also, kplc, thanks for the description of what happened with these columns. At waters they said that the labs with money what they usually do is having two identical columns (Xbridge), to use one with high pH and the other with low pH, because indeed the change happens.
Well, i did not know that it ould appear that fast and it is a pity indeed for such good column.

I have not yet tried the run again after the washing (see above) of the column, but will do soon, also will try the solvent change to the initial mobile phase % and I will post the results here.

Thank you all, it is a pleasure to belong and discuss in this forum!!

[ym3142]

Dear Vikmnilu

Both

Two of them eluted together, resulting in a horrible peak shape with a huge tailing with the normal amm Acetate 10 mM ph5+ ACN. After using TEA (and pH over 9) in Amm Acetate eluent, this problem dissappears and two better resolved peaks appear. I was not able to achieve this with any other conditions, e.g. TEA and Ph 5, not TEA and pH 9-9.5, etc.

And

As a general result of this wash/conditioning, separation is gained between RRT 1.00 and RRT 1.35. Yet the impurity peaks before RRT 1.00 and after RRT 1.35 elute in exactly the same position as they did when the column is brand new (~ 40 peaks total in sample). This also holds true for the columns purchased in 2009, 2010 and 2011. They all may be used interchangeably with unremarkable variation in RRT values per sample.

indicates that your mobile phase system did not have sufficient buffer capacity. One reason may be due to the impurities between 1 to 1.35 min have pka's around 9.5. If you can increase buffer concentration or change pH to increase the buffer capacity you may get the same separation as NH4OH, HCl, TEA at pH 9.5.

By the way, not many people are using HCl, which is considered corrosive the steel tube.

[vikmnilu]

Dear all,

my worst fears have been confirmed, and, despite washing the Xbridge over a weekend with organic solvents, I have run some samples and the selectivity has not returned to its original "value". This is, the two peaks elute together, although the shape is great.

So, any ideas of what can I do now?? How to improve the separation again, how to separate these two peaks?
Maybe It is better if I clarify what I have been running, because it may be a bit confusing now:
Suming up:

Old conditions: Zorbax XDB C8 150mm long 4.6 mm id. Particle size 5µm. Amm Acetate 10 mM ph 5 (A) & ACN (B).
-Sample 1. (Organism 1 exposed to pyrene and extracted, final solvent MeOH). Resolution is horrible.
-Sample 2. (organism 2 exposed to pyrene and extracted, final solvent MeOH). Resolution is ok.

New conditions: Xbridge C8 150* 4.6 mm; 3.5 particle size. Amm acetate 10 mM; TEA 0.1% pH 9.5 (A) & ACN (B).

-Sample 1. (Organism 1 exposed to pyrene and extracted, final solvent MeOH). Resolutioon is very good.

-Sample 2. (organism 2 exposed to pyrene and extracted, final solvent MeOH). Resolution was good at the beginning, although could be better: two peaks, (lets called them x & z) elute very close, although they are baselined separated, or almost. I try to separate them more, with a few (meaning really a few, maybe 4 hours) runs with pH 5. After these runs, conditions are not ok, peaks x & z elute together and old selectivity is lost. The peaks shape is great though, meaning that the peaks x & z really elute together.

After the loss of selectivity, still samples 1 are reasonably ok.

Well, this is a real pity and I am now regretting to have tried to get the peaks x & z more separated, I guess it is because I changed back to pH 5 that this has happened.

Ok, time to continue now.


P.S.: I still need to change the final solvent of the samples to the same starting conditions as HPLC run (80% aqueous phase- 20 % ACN), to see if there is any difference, but I do not think so.

Thanks in advance

Victor

[ym3142]

victor,

what did you mean by -"Sample 1. (Organism 1 exposed to pyrene and extracted, final solvent MeOH). Resolutioon is very good."

did you mean that resolution for x and z is more than 2?

If so, preparation of sample 2 need to modify.

Other wise, do you know if x and/or z is ionizable? if so, try 1) different pH, 2) ion pair reagent.

aromatic system? if so, try Phenyl, or PFP or diphenyl phase.

did you try MeOH.

Did you try THF, if you have to.

if none of above works, try other C18's.

if x and z are structural very similar try old tech silica gel column(type A).

Good luck