Ar or N2 in TCD

[christiancarioca]

Can anyone explain to me the following fact?

I am analyzing the same pure methane stream under the same conditions (carrier flow, block temperature, volume injected), using N2 or Ar as carrier gas.

I read that Ar should give a better sensitivity, being other variables equal, because the difference between thermal conductivy of methane and Ar is larger than with N2.

Actually, the methane peak area measured under N2 is higher (more sensitive) that with Ar.

Why does the theory not match with the experiment?

I'd appreciate you comments.

[chromatographer1]

Conductivity Values @ 49°C from CRC Handbook of Chemistry and Physics editor Weist

Hydrogen 471.1
Helium 376.1
Nitrogen 65.7
Argon 45.5
Oxygen 68.2
Methane 89.3

yet @ 93°C the value for methane is 106.6.

Weist does not list the conductivities for Argon or Nitrogen at 93°C.

What temperature are you running your oven and your detector?

Interesting issue. Are your peak widths and peak shapes the same?

Rod

[R. E. Kaiser]

There is a very sensitive effect of the temperature of the HCD sensing element. Change in small steps the voltage / current of the HCD and you will see strong sensitivity changes of the methane signal.
This happens with carrier gas N2 as well as with Ar.

REK

[chromatographer1]

Generally Christian, when experiment does not meet the theoretical, the problem is usually with the experiment not being conducted as one assumes it is. That is, something is wrong or is different than the conditions are supposed to be.

Occasionally, the theory is wrong (the literature).

What you are seeing is due to the integration being different from the carrier gas affecting the peak shape or the difference in conductivity at whatever temperature you are using favors the nitrogen as the more sensitive gas to measure methane.

Or the literature could be wrong or the values are different at your temperature (although we don't know your working temperature).

Of course, the obvious question are: are the carrier gases you are using pure when they arrive at the inlet of your GC? Is the integration correct from your DCS?

Desiring to solve the mystery.....

best wishes,

Rod

[rekuci]

This may not fully explain it, but I did a few calculations using an empirical formula describing the thermal conductivity of gas mixtures, eg from Stewart, Lightfoot, and Bird (the authors of the classic text on chemical engineering transport phenomena).

To give an example, consider 80% Ar or N2 with 20% CH4 passing through the TCD. Rough calculation (using estimated thermal conductivities and viscosities intermediate between the values given in the CRC for 300K and 400K) gives the following:

CH4/Ar mix: ~25 mW/m.K vs. ~20 mW/m.K for argon
CH4/N2 mix: ~32 mW/m.K vs. ~29 mW/m.K for nitrogen

The differences are now only 5 mW/m.K for argon and 3 mW/m.K for nitrogen, much less than the differences between the pure gases of approx ~25 mW/m.K for argon vs. CH4, and ~16 mW/m.K for nitrogen vs. CH4. Also, I'm not accounting for pressure or any other impurities that may be in one of the carriers, or any detector anomalies. If the nature of the sample plug is different, for example say you have a neater sample plug of 30% CH4 in Ar, but 60% CH4 in N2 (evidenced by a wider peak with Ar), then you're looking at the following comparisons:

CH4/Ar mix: ~27 mW/m.K vs. ~20 mW/m.K for argon
CH4/N2 mix: ~39 mW/m.K vs. ~29 mW/m.K for nitrogen

Now suddenly you have a larger sensitivity with nitrogen carrier for methane. I'm not sure what could cause this exactly, but say nitrogen interacts and competes for adsorption sites with the carrier, this would give a wider elution. You'd think this would happen with N2 and not Ar, but who knows. Long response, but I thought it was an interesting problem...at any rate you can come up with scenarios that narrow the conductivity gap or even reverse it.