Changing retention times during LCMS runs

[Jasmine]

Hi,

I have been doing an analysis using LCMS. I am running isocratically but using two solvents mixing at constant speed throughout the run. Solvent A is 10mM Ammonium Trifluoroacetate and the solvent B is Methanol. I run isocratically at 80%A/20%B. What I noticed during my runs are that a few injections seems to have changed the retention time, like the peaks all come out a minute or two later than previous injections and then later the retention times will changed back to original again. Then after 10 or twenty more injections, the same delayed retention times happens again for a couple of injections followed by a series of injection with the normal original retention times.

Why does this happen? I am running isocratically and I have allowed the system enough time to fully equilibrate. Perhaps anyone can explain why this is happening and advice me on it.

Thanks!

[MG]

That sounds like it could be temperature fluctuations in your lab, unless your column is temperature-controlled.

[Victor]

I think I can offer a possible answer to your problem.

However, first of all, stick a pH electrode in your buffer A (no need to add the organic solvent) and tell us the result.

[HW Mueller]

Your problem will probably disappear if you premix a whole batch of your mobile phase and use only one good pump.
Victor, careful with where you stick your electrodes, they are "dirt slingers". It´s better to take out maybe 2 x 2mL of the solution to be measured, place into two separate test tubes (carbonate polymer), and measure one after the other.

[Victor]

Good point HWM excellent advice.

And please do not forget to read the post on perchlorate!

[Jasmine]

The pH of my mobile A is around 5

[Klaus]

Could it be a pump problem?
Mixing MeOH and water on the low pressure side of the pump often leads to degasing. The bubbles in your pump would lower the flow and increase the retention time.
Did you check if the void peaks shift together with the main peak? Did you check the recorded pressure curves for changes?

[Victor]

Jasmine- I was not sure exactly how you made your solution A, but it appears this is just a solution of ammonium trifluoroacetate. It has no buffer capacity at this pH and I suspect your instability is due to fluctuations in the mobile phase pH caused by injection of your samples, or other factors. Is it essential that you work at this pH? If so I would recommend changing your additive-ammonium acetate adjusted to pH 5 might work better.

[Jasmine]

First of all, thank you all for sending replies to my query.

Klaus, I cannot see void peaks in my LCMS chromatograms but I did notice that all my main peaks got delayed. I'm sorry but my system do not have the pressure curve monitoring function. However, you and the others may have a point about either my pump or the solvent degassing during the runs. I will try to mix both buffer and organic modifier in one bottle using one solvent line and see if the situation improves.

Victor, please tell me the optimum pH range for ammonium trifluoroacetate so that I can a diiferent pH. I cannot use ammonium acetate buffer because it produce slight interfering peaks in my runs.

[Victor]

Jasmine,

As a rule of thumb you can use buffer solutions within +/- 1 unit of their pKa.

The amonium ion has pKa of about 9.3 and can be used between 8.3 and 10.3.

Trifluoroacetate has a pKa of 0.3 which would suggest it does not have much use in HPLC. However, it is generally used as a 0.1% solution with a pH of about 2.2. THe idea is that any strong acid has a reasonable buffer capacity because it has large reservoir of H+ ions which resist pH change. (the same sort of argument applies for a solution of a strong base) . There have been arguments on this board about how you define a buffer. Forget these- 0.1% TFA resists pH change-that's good enough. However, this solution may be too acid for your application-it may change the selectivity of your separation. You could also try ammonium formate at about pH 4.7 if ammonium acetate cannot be used. This is at the limit of its buffering range (formate pKa = 3.75), however, it is a better buffer than trifluoracetate at this pH.

[Victor]

Jasmine,

I might add that HW Mueller and Klaus may together have solved your problem. It is not ALWAYS necessary to have good buffering in HPLC. It may be possible to use mobile phases such as yours-it all depends on your samples and how reproducibly you can make up this rather undesirable mobile phase.