baseline noise and spikes

[dlntx9]

Just started working with a "new" 5890, and was not having any issues for the first day of running standards, but now i'm only seeing the solvent peak, and am getting baseline noise as well as spikes. The gas hasn't changed, and as far as I can tell none of the flow rates were adjusted.

So far I have trimmed the column, changed the septa, cleaned the jet/fid, but after that everything looks essentially the same. The CH3OH blanks essentially look the same as the standards that I have been running.

Any ideas on what else I can do to troubleshoot this.

in the picture below the solvent peak came out at about 1 minute, but on that run I didnt set the software to start recording until 2 minutes.

[Bigbear]

More information would help you getting answers.

Column/ flow/ injection parameters / compounds and oven ramp would all be helpful

[dlntx9]

I'm using helium as a carrier gas, and no makeup gas.

The carrier gas flow rate measured at teh detector with an electronic flow meter is 2 ml/min
hydrogen is at 35 ml/min and air is at 400 ml/min

suptum purge is set at 2.0 ml/min
head pressure is 10 psi
the split/splitless inlet flow measurement is 26 ml/min

Autosampler Injection
Injection volume: 1uL (split ratio 22:1)
Injector Temp: 280C

COLUMN:
RESTEK: RXI-35 SIL MS 0.25mmID x 15 meter 0.25µmdf, (#13820)

OVEN:
T1 at 220oC for 1 minute
Rate 20oC /min
T2 at 280oC for 1.5minute
Run time = 5.5 minutes

DETECTOR:
FID at 320C

[dlntx9]

I have the oven set at 300 and have been monitoring the baseline. still seeing noise and strange spikes.


Edit, I disconnected the column from the detector, plugged the inlet to the detector and re-lit it. The noise is still showing up.

Last night I pulled the detector out and sonicated each piece in methanol, is their a better way of cleaning it?

[Peter Apps]

The top chromatogram shows what looks like chromatographic peaks, probably form column bleed due to the high temperature of the oven.

The second chromatogram is 99% certain electronic noise, possibly a poor connection somewhere, a bad component or noisy mains supply. How does it look with the flame not lit ?

Peter

[dlntx9]

When the flame is not lit the baseline is flat with no noise, which made me think it was most likely not electrical noise.

[Peter Apps]

When the flame is not lit the baseline is flat with no noise, which made me think it was most likely not electrical noise.

There is always noise - are you sure that you are looking at the trace with the same vertical scale as the one you posted ?

Peter

[dlntx9]

I shouldn't have said no noise, but it was much cleaner when not lit vs lit. With no column attached lit I saw a lot of noise, and spikes. When not lit I saw no spikes and little noise. When monitoring lit vs unlit I was using the same vertical scale.

[MCAbraha]

but it was much cleaner when not lit vs lit. With no column attached lit I saw a lot of noise, and spikes. When not lit I saw no spikes and little noise. When monitoring lit vs unlit I was using the same vertical scale.

Seems to me if you have spikes/noise without a column, but flame lit, that you have some sort of impurity somewhere in your fuel gas lines (h2 or compressed air) which is ionized via the flame and causing spikes/noise in the baseline current. I don't have as much FID experience as ECDs though so this should be confirmed, but I have worked with them before.

I agree with whom said it was likely electrical noise. I know we had a Varian GC or two, fitted with ECDs, that would have electrical spikes in the baseline that would occur when the autosampler was moving and was tied to a faulty ground connection..

Hope this helps..

[aidnai]

Interesting problem. It does LOOK like electronic noise. I had very similar appearing problem that turned out to be from the mains -- an audio equipment grade power filter cleaned up my baseline and improved SNR a little better than 10x. I believe the culprit was a switching power supply with a low power factor, plugged into the same circuit of course (powering an FTIR).

However, your troubleshooting steps indicate contaminated air or hydrogen. If you can, switch tanks and see if that helps. Also the lines/regulators may be contaminated so do some extra purging before forming a conclusion.

Good luck!

[James_Ball]

Have you tried without column but with makeup gas?

5890 FIDs normally use a total flow of about 30ml/min column plus makeup gas.

[lynoguchi]

I found some of these spikes after a maintenance in the FID. I found that the problem is the grey gasket in the FID base. When this gasket get old, it lost the capacity to isolate the FID from GC and you see these spikes when the instrument is turned on.
Another thing that could be the problem is the JET connection in FID, check if the JET is not loose.

Regards.

[la_donz27]

Hi;

For the top chromatogram problems (contaminant peaks), look at the split vent line (take off, clean or replace) and do a thorough inlet maintenance, including cleaning the injector insides with a Q-Tip dipped in appropriate solvent. Column bleed will only manifest itself as a baseline rise, as columns don't produce peaks on their own. Ghost peaks have to be coming from somewhere upstream of the column.

For the bottom electronic noise issues, have you tried a thorough FID detector maintenance? I use a brass brush to scrub the jet, and make sure the spring is making good contact. Sometimes cleaning and just re-seating everything makes a world of difference on an FID having spiking issues. Also make sure the column is installed correctly by inserting all the way in then backing off just a millimeter or so as you tighten the nut. If it's crushed in there, it could crack and cause these issues.

Good luck, hope this helps.

[dlntx9]

Thanks for all the suggestions. The FID was very dirty, and I decided to just replace it rather than clean it. The plastic parts weren't looking very good, and the metal parts were rusty. Right now I'm just waiting on the parts from restek to show up. Once the new jet and fid are installed I'll post an update.