HPLC Shrinking Chromatogram

[basidiomycete]

I need help with my HPLC. I am using a Waters Alliance HPLC system to look for free amino acid content in tomato fruits. I’ve been working on this project for about 4 years now and only recently have I had problems. I noticed that the chromatogram started to shrink, it used to span 15 to 45min, but now the whole chromatogram is held in a space of 40 to 45min. It’s really strange, and it doesn’t seem to matter what I do, I’ve purged the pumps and the injector multiple times, I’ve changed the column twice, I’ve looked it up and down for leaks and I’ve gone through several batches of eluent. I have no idea what’s wrong because I haven’t done anything differently. I use exactly the same method now that I used 4 years ago. Any help would be greatly appreciated.

[Kostas Petritis]

Have you checked the flow rate of your mobile phase B (strong solvent) under pressure (i.e. is the flow rate at 100% B coming out from the column the correct one)?

[basidiomycete]

The flow rate (1mL/min is what it should be) at 100% mobile phase B (60% acetonitrile 40% water) is the correct flow rate. The prssure really dosn't change much (it hovers at just below 1500PSI).

[Kostas Petritis]

Can you describe your gradient? Did you try to change the method lately (i.e. did you make sure that the method was not changed in the software)? Are you injecting standards? Did the change from not shrink to shrink happened suddently or over time?

[basidiomycete]

The method starts out with 100% phosphate buffer (eluent A) and gradually works up to 67%A33%B over a 37min period, then eluent B reaches 100% 4min later and over a one minute period goes back to 100%A which runs for about 20min. The method lasts a total of 64 minutes. I haven't changed the method ever and I've checked the software multiple times to make sure the method is correct and that it's using the correct method. I have used glutamate as a standard, it used to have a retention time of 26min but now it peaks at 45min. I've watched other amino acids move too, like glutamine which used to peak at 27min (consistently for 4 years) but now peaks at 45min. The shrinking was gradual (over a period of a month). One profile of a tomato fruit extract I have from late June shows an otherwise readable chromatogram condensed into a box between 35 and 45min. Exactly the same fruit sample now produces a chromatogram boxed into a space between 40 and 45min.

[danko]

Basidiomycete

It could be something else, but I have a strong feeling that you are dealing with a faulty GPV.
If I were you, I’d perform a gradient test and thus make sure that the GPV is functioning adequately.
A quicker test (although not as evident as an actual gradient test) could be connecting mobile phase A and B to channels C and D.

Best Regards

[lamond1973]

I am inclinded to agree with Danko. We used to use an Alliance with our LCMS but we started to find we had problems with graident elutions. Sometimes the gradient would freeze part way through as the Alliance started to wash the injection loop. Other times we'd see peaks changing retention times during the same run. A Waters Engineer changed the valve Danko mentioned and we still found we had problems. Alliance systems have low pressure mixing which can be inefficient. I'd recommend using a binary system with two pump heads for efficient mixing when using gradients - we have reconfigured our system with a binary pump and we now get consistent chromatography.

[lmh]

Sorry, this may be a really obvious point to most people, but I'll mention it just in case, as an extra to lamond's comments. Forgive me if I'm "teaching grandmothers to suck eggs"

Quaternary low-pressure mixing pumps are very good IF they're pumping against a back-pressure. They absolutely rely on the backpressure to close the valves in the gradient proportioning valve. If you do a pump-test with no backpressure regulator, or you run a method with a flow that is too low, or a column that is short and generates little backpressure, you will get the wrong gradient.

If you have a method that is at an edgily low pressure, it will work fine just after the system has been serviced, but drift off a few months later if any tiny amounts of dirt accumulate in the GPV.

Binary high-pressure systems don't tend to suffer from these issues, but you're paying a lot more, and you can't mix three components in a gradient.

[basidiomycete]

Thanks Kostas Petritis, danko, lamond1973 and lmh for the suggestions and advice. I think the problem is solved. There is definitely something wrong with the gradient. Thanks everyone.