HP 5890A Odd chromatogram

[AJKamath]

I could use some helpful suggestions. I'm using a HP 5890a with FID and splitless injection (liquid syringe). The column is a Restek RTX-1 (60m, 0.53mm ID.

At end of last week, GC was acting OK. On Monday, 1st injection of solvent acetone showed nothing. Repeat injection same.

Started troubleshooting:

1. Checked air, He, hydrogen tanks. Replaced septum. Checked parameters: OK. Made sure He flow was registering normal (2.8mL/min) and checked that FID was ON and lighted. Ran test chromatogram to ensure signal reaching integrator (3396 Series II): OK.
2. Turned off instrument and restarted (possible brownout during weekend). Checked parameters: OK. Made sure He flow was registering normal (2.8mL/min) and checked that FID was ON and lighted. Injection still not showing peak.
3. Disconnected outlet of column and dipped end in MeOH to ensure actual He flow: OK.
4. I was using the B detector, so I switched detector PCBs between A & B slots (Yes, I turned off the instrument for the switchout, turned it back on, stabilized, re-checked parameters). Still no peaks. Dropped detector attenuation 5 units. Injection of acetone produced a short, grossly distorted "peak" with plenty of noise at approximately the expected retention time, followed by and merged with a wide blob (~1 min wide), and then a couple of tiny, poorly formed peaks.
5. Switched PCBs back, etc. Still only the ugly little peaks.

I'm thinking that the detector PCB or main PCB may be the problem. I'm helping out a friend in his lab and he doesn't know if the second (A) detector PCB is actually functional. Therefore, switching detector PCBs may not have been a meaningful test.

Thoughts?

[Peter Apps]

Puzzling.

Step one, measure the actual gas flows to the detector using a flow meter. Check for leaks (using a leak seeker) where the column connects to the detector.

Two: check the state of the inlet liner, use a leak seeker to check for leaks leaks anywhere at the inlet end of the plumbing. Measure the actual split purge flow, and the septum purge flow.

I have once in my life come across a column that had cracked - not broken right across, just cracked and leaking, so run the leak seeker over the column.

Dismantle the detector and check that the jet is in place and screwed in tight, not blocked etc, Check for contact with the top electrode, and for frayed cables etc.

If none of that throws up a fault then you may have an electronics problem - the power fluctuations might have cooked a component.

Peter

[AICMM]

AJkamath,

Since you are using a 0.53 you have an opportunity to do a neat little trick. Cool the inlet, stick the liquid syringe with a uL of acetone right into the column, inject your 1 uL and hook the column back up while you are plotting the signal from the FID. If the FID is working and the column is not blocked you will get a whopping peak. If whopping peak, focus on the inlet (maybe splitless/split valve is bad or some other blockage.) If no whopping peak, then focus on column or detector (I would start with the detector, especially the jet.)

Best regards,

AICMM

[larkl]

I'll second what Peter says about a cracked column, I've seen it twice. It wasn't enough to stop all the flow, but it was enough to ruin the chromatography.

It's pretty tedious to leak check the entire column, but you may need to.

[AJKamath]

Thanks for the suggestions. I was out at a plant until 2am, so couldn't try anything until today.

I already checked the detector and inlets for leaks and actually dismantled the detector, cleaned the jet, and reinstalled.

I've dealt with cracked columns before, but they usually manifest as increased retention times and peak broadening or doubling.

First thing is to try a backup column to see if normal peaks register. If so, I can check the problematic column for leaks and actual flow, and then try direct on-column injection as a last resort.

[Peter Apps]

The impact of a leak depends where it is inthe flow stream - near or at the detector you get reduced peak sizes but normal retention, which is what you are seeing.

You still need to measure the flows at the inlet before you change the column.

Peter