Xterra column stability at high pH, anyone???

[HW Mueller]

Richard,

your X-terra was pre-used and the Extend was new when you started your stability tests?

Please don´t get me wrong, I have been doing HPLC since ~1970 and know about tricks. Roughly, Waters columns might have been 2% of all columns used.

[Victor]

Richard,

Your results are very interesting but I am not sure exactly what you are saying (and HWM also appears confused).

I think there are 2 separate issues here. The first is whether the column is stable under the high pH conditions. This could be determined by injecting a standard (no horrible matrix compounds) at the beginning of an analysis and then repeating the injection at intervals of hours or days or weeks to see if the peak shape has deteriorated, or the back pressure has increased.

The second issue is whether the column is more prone to contamination by the injection of large numbers of real samples which contain matrix compounds. This might well be intimately connected with the pH issue, or it might not be-for example the different manufacturer columns might also show this variability at pH 7 with one lasting much longer than the other.

Do you have the answers to these questions- I think you probably do, but please let us know.

[Uwe Neue]

Just another comment on the pH stability of XTerra at high pH: I previously gave you the published information on the stability of the packing at pH 11.5. I looked up a few additional data: at pH 9 (which was the pH of this discussion) the stability was 7 times higher than at pH 11.5, all other conditions being equal.

[Richard]

Hi,

Perhaps a bit of background on what I do will help to put things in perspective. I work in a drug discovery environment, we run many hundreds of thousands of analyses annually using generic fast RP LCMS gradient methods (~ 3 min cycle time) with extensive use of MUX technology.

Running at pH10 offers a significant alternative RP LC method to low pH (pH2) RP LC, hence when the Xterra column was launched (that could withstand pH10 mobile phase) we looked to use this to our advantage.

The longevity issue was identified after several column changes on a 4 way MUX system (>20 columns) - we simply could not get the column longevity that we needed. All of the columns were failing at around the same number of injections ~1000. It was this that prompted the longevity study.

For the two tests I used brand new columns in each case.

For the first test I took our standard test mix - a mixture of acid and basic "drug like" compounds (purchased from Aldrich etc) and simply repeatedly analysed them until the column failed. I used the same generic method that I referred to above.

Two things were suggested,

1) The standard samples do not represent what you routinely analyse.
(OK but we use this text mix for low pH applications too and have not identified any issues prior to this).

2) Add a longer 100% organic hold at the end of the gradient.
(OK but the gradient profile is the same for our standard low pH method and this has run several million compounds without a longevity issue)

I took these points on board although I was not convinced.

So the second test comprised "real" samples and a 1 min organic hold was added to the gradient.

Still I generated the same result - column failure after ~1000 injections

You could argue that there is something in the chemistry that happens at high pH that does not happen at low pH - with the standards and samples that I analysed - and this could explain the longevity issue (?). It seems unlikely to me that it is the compounds that are causing the problem as the failure seems to be related to the number of injections rather than the type of compound or any marix effect.

I really did try to get Xterra to work - we get excellent technical support from Waters, but I just couldn't get Xterra to do what I needed hence I switched to another column.

Since then (approximately 18 months) we have run with Extend on the same the method as we tried with Xterra with much better column longevity.

[Uwe Neue]

Richard,

What were your conditions, including temperature?

[HW Mueller]

Wouldn´t one run a series of gradients without injections if one were trying to find out whether the mobile phase pH is degrading the column in this case? (Maybe I misunderstood something in your proceedures).

[Victor]

Richard,

Neither myself, or apparently HWM can understand the details of your argument. This is a pity, because there may be something important here.

If you are testing the high pH stability of a column, you need to take a standard of a few test compounds in mobile phase (no other matrix) and inject it, say once a day until the column no longer gives good performance. You must not inject anything else. The number of sample injections is going to be small. No "real" samples are injected in this investigation. You then should have a result as to how long (in terms of volume of mobile phase or perhaps in terms of time of contact with the mobile phase), that the column is stable. If a column lasts 2 days then the answer is clearly that it is not stable to high pH mobile phase-you needn't perform any other experiments with it. Repeat with the second column. I begin to think you have not performed such experiments.

If the results of the above experiment are reasonable, say each column lasts some weeks, then you can proceed with another test to see how many samples (real samples, plasma extracts or whatever) you can inject before the column is of no use to you. It would be useful to also perform such an experiment at a mobile phase pH where the column was much more stable. It might tell you whether one column is more susceptible to matrix contamination than another-this is a very interesting issue. It could be independent of the pH issue.

These experiments are horribly tedious and of course you probably have no time to perform them. However, all that you can say with confidence is then that for your samples and conditions, column X is better than column Y. This is very useful for you, but of limited value to others.You cannot however say that column X is more stable at high pH conditions than column Y or any details like that.

[Richard]

Hi,

The conditions I use are,

A: 10 mM ammonium bicarbonate adjusted to pH 10 by addition of ammonium hydroxide.

B: MeCN

The gradient runs 5% B to 100% B

Flow rate: 1 ml/min

Temperature: ambient

Column format: 2.1 x 50 5u

Injection 3ul (loop fill) ~ 1 mg/ml soln.

The tests that I have run mirror the conditions that columns will encounter in routine use in my lab. Running pH stability tests without the injection of any compound or matrix will not tell me whether the column will work for my application - this could be the crux of the issue. The more "correct" way of determining pH stability by means of pumping a large volume of mobile phase through between a small number of test solution injections vs running the column under the conditions it will encounter in the lab. Which experiment will be of most value? probably both but... As an analyst I have a rather limited amount of time for method development and tend to rely on the specs provided by manufacturers to make an initial decision on what to use a column for ie for high pH applications Xterra, Extend and Betabasic are obvious columns of choice. So the best way for me to evaluate these columns is to run them head to head (in the way that I have explained) and see what happens.

I have pointed out in previous postings that column lifetime is application specific and explained which column works best for my particular application. I also have pointed out that in order to determine which column works best for your application you need to try if for yourself in your lab etc.

I agree that running at several different pH conditions would be an interesting experiment to do - but it is pretty unlikely that I will get a chance to do that unfortunately. The technologies that are of most use to me are the ones that work straight out the box - so that I can get on with sorting out the next thing on my "to do" list. In this case Extend was the answer - I don't really know why - but as interesting as it is I don't have any more resource to get to the bottom of the longevity issue.

[HW Mueller]

Richard,
"..could not get the column longevity that waters claims" meant to me that you had unequivocal evidence about a false claim. Now, that you could get better results with one column than with another is also interesting, but a different ball game.....OK.

[Basil]

Hi,

I'm also having problems with XTerra RP18 3,5 µ (10 x 4.6)
After 300 injections I've lost 20 % of capacity.

Method:

A: 5 ml of triethylamine in 1 L of water adjusted to pH 10 by addition of formic acid.

B: MeCN

The gradient runs 25% B to 75% B in 25 minutes

Flow rate: 1,2 ml/min

Temperature: 40 ºC

Injection 5ul (loop fill) ~ 1 mg/ml soln in MeCN

Could it be a problem with temperature ?. If I change to ambient then i lose resolution between a main peak and an impurituy. or maybe mobil phase A is to strong for this kind of columns ??

[Richard]

HW,
Sorry about the confusion, it would still be interesting to hear from other users and their experiences though...

[Victor]

Richard,

Ah -so now we have the answer. I agree with HWM. I think users should be as careful in refuting the claims of column manufacturers as column manufacturers should be in making them in the first place. Although your result might still stand-apparently the Xterra column may be less robust to large numbers of injections of your samples under high pH conditions than a rival-there is little else we can deduce from your postings apart from the fact that there might be something for someone else to investigate here.

[Uwe Neue]

To Richard:

For a column without a guard column, a typical column life time under ALL conditions is about 1000 injections. The maximum that I have found for a clean injection was about 2000 injections. Therefore, I do not see a problem that would point to a stability of the packing material under your conditions. As I said before, it does not agree with our findings, and we are typically using the same or very similar operating conditions as you are.

To Basil:

At pH 10, we are generally using ammonium bicarbonate as the buffer. While I do not see a reason why your triethylamine buffer would act any differently, one can never be sure about subtle differences that may influence column lifetime. Also, we typically work with a much lower concentration: you use a 50 mM concentration, we are typically working with 10 mM to 20 mM. You could try to reduce the buffer concentration, or you could try to use a ammonium bicarbonate buffer at the same pH, or both.