HPLC Ghost peak

[Consumer Products Guy]

Agilent 1100 with other vendor's conductivity detector. Hooked up the conductivity detector, get a ghost peak before one of the peaks of interest. Also have very similar set up on 1050 system, where I don't see this ghost peak. Co-worker made new mobile phase, new standard mix, swapped the two columns (same part number), not related to that. I felt it was autosampler-related so I rigged up a manual injector on the 1100 and that did NOT show the ghost peak, so isolates the issue as 1100 injector issue. Since I had the parts, I swapped in a new rotor seal, cleaned the rotor parts, no effect. I also then replaced the metering piston seal as that hadn't been replaced in years, no effect. The rotor stator face holes show one to appear to be contaminated with a residue, but ultrasonic mixing with methanol didn't help that either, and I didn't have a spare one ($212). Injecting pure water did show the ghost peak, and didn't really matter in size whether I injected 5.0 ul or 0.1 ul, so it appears that just switching to the "bypass" mode causes it. I'm letting it flush overnight in the bypass mode as mobile phase is very dilute acid only. I may also swap out the needle seat, as maybe that's contaminated, and I have a new one of those. I understand that mobile phase continually flushes out the autosampler tubing including the needle ineterior during the run (mainpass) so I'm ruling all tubing out. Any other thoughts, as I've swapped mobile phase column, etc. to the 1050, and no ghost peak there. Oh, yes, the multi-channel mixing valve on both pumps has been bypassed with an adapter, so they are both functioning as isocratic-only pumps currently, so solvent mixing from other reservoirs is eliminated. Any suggestions from a second pair of eyes? Thanks?

[HW Mueller]

Is the 1100 co-injecting air?

[Consumer Products Guy]

HB - I've been working with HPLC for 25 years, and have to admit I don't know what you mean by "co-injecting air". I have standard solution in the autosampler vial, so no air gets in that way; I assume for what you suggest that I would need to have a connection somewhere where air could get sucked in, correct? Anyway, after thinking about this overnight, thought maybe a contaminated needle seat could be the source of a regular ghost peak with each injection, visible with conductivity detector but not picked up when we had used the HPLC with UV detector. I replaced the needle seat this morning and the problem appears to have been solved, see below (ghost peak at about 2 minutes, before two peaks of interest).

[tom jupille]

An air bubble would make sense with UV detection (especially at low wavelength), but not with conductivity.

Since replacing the needle seal fixed the problem, I'd be inclined to suspect that contamination (perhaps from some long-ago buffer?) was indeed the culprit.

[Consumer Products Guy]

I'd be inclined to suspect that contamination (perhaps from some long-ago buffer?) was indeed the culprit.

That's my thinkin' right now. And, yes, that's a well-used column, need to clean out or use my brand-new one, which gives sharper peaks. Those are organic acids.

[HW Mueller]

Good to know hat a conductivity detector is not influenced by air (I was thinking of dissolved air). I thought that the applied potential might be enough to reduce some O2, or that viscosity effects may show.