Chromatography Forum

Hi, I was hoping to get some help selecting calibration gases for a GC I'm currently setting up.

The GC is to be used to test Natural gas samples, where we are looking at the following gases.

Helium, Hydrogen, Nitrogen, Oxygen, Methane, Carbon Dioxide, Ethane, Ethene, Propane, I-Butane, N-Butane, I-Pentane, N-Pentane, Hexane.

With samples generally consisting of 80-95% Methane, and small amounts of the other gases.
However occasionally its possible I'll receive air contaminated, or samples with high CO2 levels, which will cause matrix effect for my other gases.

I know what to do if the air contamination or CO2 contamination become very large, I just need a calibration gas where the balance of my gas is changed from methane to the contaminate.

But how do I handle calibrations when my methane/air or Methane/CO2 level become close to equal.
Is it as easy as mimicing my other calibration gases but change the methane and nitrogen level to be equal (~45% each).

Also if the range of air contamination can be anywhere from 0-99%, how many different contamination level calibration gases will I need?

Are you using FID for the hydrocarbons? If so, the FID is quite linear, so you MIGHT be able to get by with a high and a low standard only for these gases. If you're in a regulatory environment, then your calibration protocols may be more stringent and this won't work.

My first thought when I read your post was that I'd be more worried about the separation and the chromatography with concentrations varying that much. Perhaps you've already crossed that hurdle? Calibrating a method with poor or unreproducible chromatography isn't going to work.

AWatt,

You are using a TCD because of the list of components you want to look at. Hopefully you are using a two channel system because of the hydrogen/fixed gasses you list. The TCD, like the FID, is very linear, so that's the good news.

You have a number of options. 1) Introduce air and use it as a second calibration point with your current standard mix (and leave it at that.) 2) Blend your hydrocarbon mix with a diluent (like air) in a dynamic blender. 3) Purchase another "point" or two to make up your calibration range.

On the other hand, if you have that much air in your sample, are you really going to report values or just say your sample is corrupt?

Best regards,

AICMM

Much of the problem here lies in the non-ideal nature of the gases even though the TCD is a linear detector. To get a very accurate measurement you really need a number of calibration blends. That said, AICMM is correct in posing, do you really need to be all that accurate for poor quality gas samples? Do you not need instead a good measurement of the BTU value of the natural gas?

As the air content goes up the methane content goes down, so depending upon the column selection you pick, the separation of the peak pair stays nearly the same if a correct selection of columns has been picked.

I assume also you will have a separate column train to measure helium and hydrogen using nitrogen carrier. A much larger sample loop to measure these ppm levels of gas would be required.

CO2 is another issue altogether. Much will depend upon the selection of columns you choose for your measurement. What multidimentional column configuration do you plan to use? Reverse column step? Backflush to detector? Why are you not measuring C6-C9+ content?

best wishes,

Rod