GC-FID with constant continuous peaks

[Korffski]

Hello,

I am using a 5890 series II GC with FID to analyze toluene and TCE with manual head-space injections. The column is an hp-5ms 30x.32x.25.

Sporadically there will be constant continuous small peaks, almost like the normal noise is amplified. Each peak is 150-200 area counts and my sample response is approximately 100,000 area counts. It comes and goes for no apparent reason. But now it has been doing it constantly for about 2 hours which is a first.

What should I do?

Thanks,
Korffski

[Peter Apps]

It will help if you post a picture of the noise - there are instructions in a sticky on the LC forum.

Peter

[Korffski]

Thank you for the direction on how to upload a picture. Saved me a lot of time.

I happened to catch it going back to normal.

[Peter Apps]

My guess is a fluctuating gas flow, probably one of the FID gasses. Check the flow readouts on the GC, and also use an independent flow meter.

Peter

[aidnai]

The pattern of cycling does look like a gas flow. If that turns out to be a dead end, you could consider electrical noise sources as well. My TCD picks up CDMA type cell phones, and though that is not as periodic it's conceivable a pump, compressor, etc turning on and off could cause an issue.

[Korffski]

I tested both air and hydrogen just after the regulators and they did not fluctuate. Should I test next to the GC as well?

What would I do about electronic noise?

Sorry for all the questions our department analytical chemist just left and I am just an engineering student.

[Peter Apps]

You need to measure actual flows downstream of the detector.

Do you have a flow meter - what kind ? Do you have a way to connect the flow meter to the outlet of the detector ? Somewhere in the users manual there will be a procedure that you need to follow.

Turn off all the gasses to the detector (except the carrier gas). Turn them on one at a time and measure the flows - you are looking for short-term fluctuations.

Most GCs will have a readout of detector gas flow on their front panels or software - check that, but if there is a problem with the flow control then the readout may not show it.

Peter

[rb6banjo]

It has been a while since I saw this on one of my systems but as I recall, I took the detector apart and cleaned the flame jet. I assembled it again and the problem disappeared. You might try that. It's a fairly easy thing to do.

[CE Instruments]

It has been a while since I saw this on one of my systems but as I recall, I took the detector apart and cleaned the flame jet. I assembled it again and the problem disappeared. You might try that. It's a fairly easy thing to do.

This would be my first port of call. You can check for electrical noise by just turning off the flame and gases and monitoring the baseline.
Can we assume this did not occur after a change of column , connection , septa ? All can give traces like this and should be the first thing to question if prior to whatever you did the GC worked fine

[erdno]

Dear Korffski!

I saw this last time when my Jet was obturated. It is possible when u overtight the screw and ferrule go inside the Jet. So your flow will fluctuate since it can not flow properly.

Can i ask why do u have so high negative signal? I have never seen it before using FID.

Erdno

[Korffski]

Just cleaned the detector and flame jet by following manual instructions. The cleaning process turned the methanol brown so it was filthy. Did not fix the issue.

No electronic noise with flame and gas off.

Have not changed a column or connection but I did replace the septa and inlet liner recently. Very close to the time I started seeing the issue, but the other people I asked did not think this could cause my issue. Thoughts?

Erdno,

I will check what you suggested. The negative signal was due to having the zero offset shut off, when on it is closer to a baseline of 15.

Peter,

I am going to check the flow next. I have an old aligent ADM 1000 flow meter and the FID flow adapter (took me forever to find in our messy lab).

[Korffski]

Peter or anyone,

How much fluctuation in flow is considered acceptable? With just the carrier gas the flow seemed to slowly bounce from 2.84-2.88 mL/min and the addition of the FID gases did not seem to change it at all.

Thanks,
Korffski

[Peter Apps]

I doubt that such a change in carrier flow (if it is even real and not just measurement variation) would cause the baseline ripple - unless you have a contaminant in the carrier gas that gives a background signal. So don't worry about it for now.

Over what time period are you measuring flows - how fast does the flow meter react ? The baseline wobble has a period of about 2.5 s, if the meter is averaging over longer than that then the ripple will not show up.

What are the measured flows for hydrogen, air and make-up at the outlet of the detector.

The wobble being intermittent adds a problem. You need to watch the baseline until it starts wobbling, then shut down the gasses, start them up one by one and do the measurements, the put the detector back on to make sure that the wobble is still happening. NB if turning any one gas on and off makes the wobble go away it is a pretty strong sign that the gas flow is the problem.

Check that the pressures in the lines going into the GC are all within the ranges they should be.

Peter

[fahim.khattak]

me too was facing the same problem. there is simple solution to this problem that is thermal cleaning, just increase your FID temperature. suppose if you are operating FID at 220 degree celsius, change it 280 or 300 degree celsius with continuous flow of carrier gas for about 120 to 180 minute. mine 100 percent sure your problem will solve.

[Peter Apps]

me too was facing the same problem. there is simple solution to this problem that is thermal cleaning, just increase your FID temperature. suppose if you are operating FID at 220 degree celsius, change it 280 or 300 degree celsius with continuous flow of carrier gas for about 120 to 180 minute. mine 100 percent sure your problem will solve.

Has to be worth a try Korffski (as long as the column will stand it). If you do try it please post whether it works.

Peter