Chromatography Forum

Hello all,

I am experiencing a lack of sensitivity from my Agilent 7890 FID detector. I am measuring ambient air for CH4. My standards range from 0.5 to 3.0 ppm, the linear peak ht only varies by maybe 1-2 pA and therefore reproducibility is horrible. I see much more sensitivity in my 5890s and a similar instrument (7890) shows maybe 5 pA difference in peak ht from 1 to 1.4 ppm CH4 samples.

My flow is approx 23, H2 30, air 300, N2 makeup 1, oven temp 80-90, detector t = 300.

I tried to inject manually or automatically and the results are the same. A higher range of stds 8-60 does not yield large differences in peak hts either.

What can I do to improve this?

Thanks much!

I should add my flow rate is 23 at a epc set pressure of 25. A similar setup in another lab shows a flow of 53 at a pressure of 20? Maybe they selected a wrong carrier type, mine being N2.

First step is to check all the flows using an independent flow meter, not the onboard gas controls of the GC.

Peter

Actual flows as follows

Carrier = 24
carrier plus makeup = 25
carrier plus h2 = 55
carrier plus air = 333

They look okay.

What to try next?

Thanks

How are you generating your standards, aqnd how are you injecting them ?

What do the peaks look like ?, flat tops ? Please post a chromatogram - instructions are in a sticky at the top of the page.

Peak height does not change much, what about area ?. Please post the heights and areas for a set of standards.

Peter

kahmark,

How much are you shooting and how are you shooting it? According to my rough calculations, 2 ppm at 1 cc is about 1.5 ng on column (splitless) so you are pretty low amount for the typical FID.

Best regards.

Peter,

The peak hts are approx 15 pA with areas in the 20-30+ range. The peak shape is perfect. Are stds in the CH4 range of 0.6 - 3.5 ppm doable. I have another GC (5890) that we run the same suite of stds and get obtainable results, I see a nice curve with one ml injections. Also, we have changed the FID signal board and all GC parts are new.

I do notice at injection the total flow at the injector goes to zero for a bit which means the EPC is sensing not to add pressure. So I slowed the speed down so i see flow always, this did not help.

I'll try to post a screen shot of my range.

THANKS!
Kevin

Peter,

I can't understand how to insert an image, could you make it more plain for me. Or, I can email a screenshot to you.

Thank you

I doubt that my instructions would be any clearer than this : viewtopic.php?t=12659

Total flow dropping to zero when you inject probably means a substantial leak, try a new septum (if you are using a syringe) and check for leaks if you are using a valve.

Please review the posts so far - there are unanswered questions that will help people help you.

Peter

Standards are generated from a multi-std cylinder into 6 ml glass vials with Labco septa. We use a gerstel MPS2 to inject one ml of sample into a purged packed inlet with a septum purge of 0.5 (low).

Our std range is 0.7 to 3.6 ppm CH4. This number also accounts for a two ml dilution with N2 to give the vial enough pressure to draw 2 - 1ml samples. We flush the 6 ml vials prior to injecting the appropriate amount of standard into the vial. We then add 2 mls of overpressure.

The standards are then loaded into the custom autosampler and one ml is drawn and injected into the GC at a rate of 100ul/sec (ten sec inj).

Peak hts and areas do not change much at all over the range of stds.

I also run N2O from the same vials onto a back column and see discernable area/peak ht differences. I've have a split after the FID injection that takes 2ml/min flow and diverts it to a Licor gas analyzer which shows a difference over a std range in the same vial as the CH4 std.

This led to Agilent checking the FID signal board and replacing it with no success.

Screenshot to come.

Again I see no significant difference in area or peak ht for our std range. I run a similar setup minus the dilution on a 5890 GC, curves look good w/ r2 of .994

Thanks,
Kevin

Here is the chromatogram, the top plot is the FID plot the bottom is the ECD plot (two columns for the ECD) As you can see each methane looks to be nearly the same peak ht. These stds are 2 to 4.7 ppm CH4.

Hi Kevin

Puzzling

No matter what flow and pressure disturbances there may be, there should still be a concentration difference in what goes through the detector, and that should give a change in peak area.

An independent detector in parallel with the FID does detect a difference in methane concentration - am I right ?

Are you absolutely sure that the peak you are looking at on the FID is actually methane ? If it was some other component of the matrix gas its concentration would not change.

What column are you using ? - there are some gas separation experts on this forum who could probably advise whether the elution time of your peak is appropriate to methane.

Peter

Hi Peter,

Yes puzzling.

I can add higher amounts of CH4 say 25 -50 ppm and see a response. There are no other peaks on the FID so unless CH4 is getting masked or lost completely I'd say that peak is methane. Did you see the chromatogram? What I didn't have connected at the time of the chromatogram was the Licor measurement.

The independent detector measures differences in CO2, and our ECD in parallel measures N2O. Those show adequate results.

COLUMN- Restek PP-Q, 3m, 80/100 mesh, 22ml/min flow (approx). I use these same columns in four other GCs with results.

I've now had several Agilent and a Gerstel engineer in to look at this with no progress.

I've even doubled the injection amount with no result.

Today is another day!

Kevin