Determining cause of random peak

Hi MassMan , the two most likely candidates would be the rotor seal of the Autosampler injection valve, or a buildup of contamination on the column itself.
If you have a zero dead volume union (ZDV), you could replace the column with a ZDV and then do a series of injections of your mobile phase and see if any peaks come through after injection. If no peaks then suspect column as the source of the ghost peak.

If you have a spare new-ish or known good LC column alternative then put that on the system and check with injections of Mobile Phase, ACN or MeOH (if compatible with the column used). If you are still getting the same peak then it probably else where in the system (Needle seat, capillaries, rotor seal of injection valve, any column switching valves). If no peak, suspect the Bio-Rad column is contaminated.

Conversely you can disassemble and sonicate your Rheodyne valve rotor seal (or replace seal if required), then re-assemble, run a pressure test to makes sure no leaks and then re-test with blank runs and injections of mobile phase or MeOH or ACN to see if the ghost peak is still present.

I hope that gets you somewhere. Please let me know how you get on.

Regards,

Damian

Thanks Damian

I removed column and added a zero union fitting and peak showed up right after injection. Removed Rheodyne valve seal and sonicated that and head in Meoh for 20 minutes. Still have peak. Bypassed all column heater valves and peak still exist. Bypassed needle, seat and injection pump valve, still have peak. Replaced needle and seat with new and rebuilt metering valve, peak still there. Replaced rheodyne valve seal with another good used one, still a peak. If I inject 1ul I get the same peak area as a 10ul injection but if I inject 20ul it does double. Any other advise would be helpfull.

A little more data. If I cycle the bypass valve no peak. If I do a needle rinse 100s I see the peak. Rinse is DI water and the lines have all been rinsed with Meoh. I turned off the needle rinse on the method earlier as I thought it might be the problem with no change but something is being brought into the equation that the needle rinse procedure does.

Thanks

Chris

Hi Chris,

You have certainly done an extensive effort at troubleshooting this issue.

Ok so to summarize it looks like the peak only occurs which the injection valve is actuated correct? If you just run your gradient without injection and with a ZDV you get no signs of a peak? If so, then one of the only candidates remains the rotor seal of the injection valve. If sonication hasn't helped it may need to be replaced.

Agilent Autosamplesr are a flow-through design which means in mainpass mode the needle, needle seat and sample loop are included in the flow path, except in bypass mode. So, if those elements were contaminated you should see you contamination peak even without injection.

Point of clarification - "I removed column and added a zero-union fitting and peak showed up right after injection" - are you sure this is the same peak? Was this peak reproducible over multiple injections?

 I would connect the pump straight to the detector, run a normal gradient and confirm no peaks. If no peaks then the pump and the detector can be eliminated from t-shooting. Then connect the autosampler into the flow path (Pump_>ALS->Detector) and run a gradient again (no injection) and confirm no peaks. Then repeat with MP, ACN or MeOH injections to see if a peak occurs. If it does then it really has to be to be the ports of the rotor seal that are introduced into the flow path when the valve switched to bypass mode during injection, and that is literally the only point of difference in these two flow path comparisons. Or it is coming from your MP, think Glassware, pippettes etc. that you use to prepare your mobile phase, not just the solvent itself.

If that doesn't shed some light on the issue then it's time for an engineer visit I think Chris. Based on the info you have already provided I'm out of ideas other than rotor seal replacement, which is an expensive part.

Regards,

Damian