Slow falling baseline with FID

[Gopher]

Hello,
While looking for information on a problem I've been having I happened to stumble onto this forum site and figure someone here might be able to help me fix my problem.
I am currently working with a PE autosystem GC equipped with a FID. We don't normally run GC everyday much less every week so the two GC's in the room (the second of which is an autosystemXL) both are hooked up to the same gas cylinders. We use He for carrier and H, and compressed air for the FID. My problem is that when I go to ignite the FID (manually) on the autosystem, it lights and the mV reading jumps to well above the detectors limit. after a few minutes the reading starts to come down but very slowly. it can take upwards of 4 to 5 hours for this GC to become usable, meaning the baseline during a run returns visible peaks and stays above the autozero level. I feel like it is clearly a problem with the FID but what that problem is or how to fix it I just have no idea due to my being a new grad. and only having run GC and not maintained the instrument. Thank you very much for the help.

[Gopher]

I forgot to add that there is a second PE autosystem GC in the building that I don't have much contact with but it appears to also suffer from the same or similar problem. and there are 2 autosystemXL GC's that both have no problems like this.

[Peter Apps]

Do you leave the detector hot when the instrument is on standby ?, if not you are probably getting deposits that slowly cook off when you start up again.

Other possibilities are column bleed, dirty gasses or even inlet contamination, we need more details on how you leave the GC, and what scrubbers you have on your gas lines.

Peter

[AICMM]

I second Peter's emphasis on detector temperature in standby mode and his request for more information. I would also add one thing, house air or cylinder air?

Best regards.

[Gopher]

So I just asked and it looks like there are filters and moisture traps on all the gas lines and they were just replaced recently. we use gas cylinders for all three gases. with regards to the detector temp, we leave it hot all the time at 250C. I'm pretty sure we don't run it any higher due to the column's max use temp. (gas chrom column)

also of note the GC has been sitting unused for over 1 week and I just lit it again and the FID still shoots up but not as high as it has (~50mV as opposed to over 1000mV). usually this GC settles at around 10mV.

I did also just replace the inlet liner no more than a few weeks ago and we have not run it much since.

Thanks!

[Ron]

What column phase and phase thickness are you using? I suspect it is a thick phase column, and may need to be reconditioned.

[Gopher]

It has been a while but it looks like the problem has been resolved. After some time we started having accuracy problems. all of our check standards were running considerably low. after trying and changing everything to do with the samples, we replaced the column. That simple change seems to have fixed both problems. my guess was the column was bad from the very beginning.

Either way thank you for the help and the suggestions. I just figured I would let others know our solution if they end up having similar problems.

[Don_Hilton]

Thanks for posting the resolution.

[dhe0124]

Your wandering FID signal is a common problem. Assuming that the gas flow is proper, here is what I do before a run:
1. Check the column inlet (the column should be about 1/8 inch into the inlet). I usually cut about 1/8 inch to remove any particles in the inlet.
2. I use Vespel/graphite ferrules. They have to be conditioned by heating and cooling at max allowed temp a couple of times. Tighten them before your run, but eventually they shrink from tightening and need to be replaced. Before your run, leak check. I use Snoop, then rinse off.
3. Bake the column (at near maximum allowed) for an hour each day, but an occasional overnight bake at near maximum is best. Typical temps are 240 C. Cool down and check for leaks after the overnight run. I find Snoop more reliable than electronic detectors, but rinse the snoop off.
4. Measure and record all flow rates.
5. The Detector area must be cleaned of any debris.
6. By the way, the air flowrate should be checked occasionally with a flow meter. Low air flows and too fast air flows cause eratic signal.

When all else fails, follow the instructions of the oldest expert: his experience is the most reliable.

[Peter Apps]

Sorry, but Ihave to disagree with a lot of what dhe0124 says.

Downward drifting baselines are certainly not common unless there is something wrong with the system. Upward drift (due most commonly to bleed) is inevitable to some extent in programmed GC.

I cannot imagine any scenario in which doing all this preventive maintenance before every run could be worthwhile.

The 1/8 inch insertion depth for the column is specific to Agilent instruments. Cutting 1/8 inch off the column is pointless - contaminants will penetrate at least 50 cm. You should though cut a cm or so off the end of the column after it has been threaded through the ferrule.

I have never heard of pre-conditioning vespel-graphite ferrules, and I have never experienced their shrinking with time once they have bedded in over the first couple of runs and been re-tightened. I have had VG ferrules in place for months and hundreds of runs with no long-term shrinkage.

Leak checking is good practise, but never ever with Snoop. Get a good leak seeker, or if you want to use liquids use propanol diluted with is own volume of water. NB when using a leak seeker turn the oven fan off. If you use liquids you have to cool the inlet below their boiling point.

Baking columns near their maximum temperature shortens their lifetime. It may be helpful, but it should be invoked only if the samples contain a lot of heavy contaminants. Baking is a lot more effective if the flow through the column is reversed. Paradoxically baking out, especially for long periods can make upward baseline drift even worse by damaging the stationary phase, which then starts to bleed more.

Measuring flow rates is good practise, but is only really valuable if it is done with an external flow meter - the front panel readouts of the GC are linked to its gas controls, and so a fault in a sensor can cause an incorrect flow with the expected indicated flow still displayed on the front panel.

FIDs (especially with capillary columns) last months or years without collecting any debris. The only significant sources of contamination are large volume injections of chlorinated solvents (which make soot), and silylating agents which make silica. It is certinaly not neccessary to clean an FID (which usually implies cooling it, dismantling it, reassembling, heating up and conditioning) before every run, - you would maybe get one analytical run done before it was time for column's overnight bakeout !

Welcome to the forum by the way.

Peter