Unknown peak causing decreased analyte response

[AVgt]

Hi,
I hope that someone might be able to come up with an idea on how to proceed with troubleshooting of a method.
I am running an ion-pairing method on a Dionex Ultimate 3000.
Mobile phase components: Octanesulphonic acid, citric acid, phosphate buffer, EDTA, 10 % MeoH.
Electrochemical detection.
Column: Kinetex, C18
My analytes are among others Dopamin and serotonin.
I have been running the method a few months now, but lately, an unknown peak at 2.8 min has emerged. It comes and goes. However, it is not present, when I inject mobile phase or 10 % MeOH. But it pops up periodically in my calibration standards and samples with a varying response. But when it is present in the chromatogram with a high response, it lowers the responses of all the analytes accordingly. I thought at first that it could be an autosampler problem, and changed syringe and needle, no effect. I have also prepared new mobile phases, new standards solutions, cleaned the electrochemical cell and changed the column. No effect.
It is difficult to find out what is going on, since the peak is periodic.
Any suggestions?
Thank you in advance.
Best regards,
Anne

[Gerhard Kratz]

Do you operate at constant column temperature? Flow is constant?

[AVgt]

Yes. The flow is 0.8 ml/min and the column temperature 30C.

[tom jupille]

Intermittent problems like this are the *absolutely* most difficult to troubleshoot!

How often does the problem occur?
When you see the extra peak, can you re-inject the same sample and see it again? Or re-run the same sample sequence and see it at the same point?
How wide is the extra peak compared to the other peaks in the same chromatogram (if it is much wider than the other peaks, then it may be a contaminant from a previous injection finally coming out)?

[AVgt]

I agree, it is challeging
The unknown is at least present in each sequence (around 15-20 samples), mostly several times.
I cannot reproduce the peak from a given sample, when I re-inject it.
I think it is somewhat broader than the other peaks, but not alarmingly much. So maybe you are right about it being a contaminant from a previous sample coming out. Actually, I moved the column and mobile phase to a similar HPLC system, and did not see the peak. So perhaps it is building up over time. However, it is difficult to get rid off, since I do not know what it is.

But thank you for your questions. Nice to hear an opinion from some one from the outside.

[tom jupille]

Ok, so a couple of possibilities here:

1. The fact that the problem occurs only on a particular system points to it being a hardware issue, but you have already changed the most likely components. Maybe something like the needle wash solution? Or dust in the autosampler that gets picked up occasionally (there was some discussion of that kind of problem last spring, but not on a Dionex system: viewtopic.php?f=1&t=26464)?

2. Since your method is isocratic, and the problem peak is somewhat broader than other peaks, that suggests that if it's a late-eluter, it would have to come from the preceding sample (if it came from an even earlier sample, the width would be *much* broader). A test for that would be (when you see the unknown peak), to re-inject the *previous* sample and then wait for twice the usual run time.