Agilent 7890A GC-FID peak tailing

[Yifan Du]

Hi everyone, I am using a new Al2O3/KCl column that was installed just three week ago to measure light hydrocarbons in my gas samples. I inject the gas by a 50 uL syringe under the splitless mode. The temperature program is from 30C to 200C at a rate of 10C/min. The injection temperature is 125C.

In the beginning the measurements were alright, but since last week there was a large tail for the first peak, i.e., the CH4 peak. I have tried several solutions, e.g., changing the septum, changing the liner, changing the O-ring, cutting off several centimeters of the column front and reinstalled the column, baking out the column again, but nothing really works.

I got acceptable measurements very seldomly afterwards, but most of the time the tailing peak is still there. May I ask for some suggestions? Thank you very much

[larkl]

Poor Methane peak shape typically indicates a leak.

Are all the peaks tailing or just the methane? How long did you bake out the column? The alumina columns are VERY sensitive to carrier gas impurities, especially polar (like water). How clean is your carrier gas? One lab that I was in baked all the alumina columns out all night once a week. Although, typically, we see peaks shifting with alumina "contamination".

I've never seen anyone use splitless mode and 50 uL for gases. More typically, I've seen a GSV with a 250uL sample loop, then split mode at maybe 50:1.

[Yifan Du]

Hi Larkl,

Thank you so much for your reply. It is normally only the methane peak tailing. Most of the moisture in my gas samples has been removed by a cooler, but indeed there is the possibility that the treated gas is not totally dry. I baked the column twice during three weeks, but just for 4 hours each time. I will follow your suggestions and bake the column for a longer time at a higher frequency.

I am a fresh user of GC, and was not aware of the difference between splitless and split in the beginning. Many of my measurements had been perfomed at the splitless mode, so did the calibration. Would those measuring results be reliable? Thanks again in advance.

[larkl]

In my experience, splitless is used for very low concentration samples. The solvent band dropped onto the 'head' of the column evaporates, concentrating the solutes. You may be overloading, but with only 50 uL, maybe not. If it calibrated fine, then it's working other than the new problem.

What is your methane concentration? You might try injecting less and see if the peak shape improves. I'd switch the method over to split for a trial run, personally.

If you've baked out several times for 4 hours, the column is likely dry, which is good.

I'm really guessing here, but have you replaced the gold seal in the bottom of the inlet? Early peak shape problems are typically due to a leak.

[Yifan Du]

Hi larkl,

Thanks for the your explanation and suggestions.

The components of my samples are complicated, and the concentrations of them also vary a lot. The methane concentration ranges from around 500 ppm to 2%, but the concentrations of some other hydrocarbons are normally very low, i.e., just a few ppms. I guess considering those trace species, the splitless mode might be better? As you said, maybe it is not the problem as I got fine calibration results, but still, I will give the split mode a try.

And you are right, I have not changed the gold seal in the bottom - this might be a problem. I will try this solution, too

Your answers really help me a lot. Thanks again!