Chromatography Forum

Hi there,
I am trying to figure out optimal conditions for most symmetrical peaks in my GC-application. I am trying to analyze a gas mixture consisting of N2(~65%), H2(~15%), CO(~10%), CO2(~10%) and have got the a sampling valve with sampleloop (250µL), which is heated up to 150°C.

Now, which temperature should I use best for the inlet (split-mode 40:1, carriergas: He, and right now I use 50°C)? What effect has the inlet temperature?
I am using a tandem capillary column (molsieve 5A and porabond Q are connected via Y-pieces) and a TCD at 160°C (is that to low aswell?). But the peaks are very wavey.

Thank you

Hi, I do not have experience with the gas analysis or the type of columns You are using.

However inlet temperature is always set in such range that the sample and solvent would evaporate instantly but no too high that thermal decomposition of the sample would occur. So in in Your case it is almost irrelevant as the analytes are gases. For pratical reasons and to prevent unexpected effects (condensation etc.) generally inlet tempereature is set higher as the highest columnt temperature used during analysis. Also in Your case probably would be good idea to have inlet temperature similar to gas sampling valve temperature.

Regarding peak shape - inlet parameters could in some cases influence peak shape of liquid samples. However it should not make any difference for gas samples unless some flow disturbances are created.

Main factors which influence peak shape are column properties, carrier flow and column temperature. Since You can not change column (I presume)- set inlet temperature in a range of 150C, test column flow and temperature changes to improve peak shape.

In any case, I would put Inlet temperature of inlet higher, then temperature of incoming sample to prevent condensation of impurities.

T(Valve)<T(inlet)<T(oven)<T(detector)

It might help if you could post a picture of the chromatogram. Your desccription of the peaks being "wavey" is not entirely clear as to what is going on - but might be important information.

buxbaum,

At 250 uL and split 40:1 with dual column flow into a TCD I would be surprised if you are seeing much in the way of peaks at all.... I might even think you are just looking at baseline fluctuation. Not being snotty here, it's just that TCD is not that sensitive a detector and it is a flow sensitive detector so you might not be working in a range it can analyze. I can guarantee you are not seeing the hydrogen (being that you are using He carrier under these conditions.)

If I were in your shoes I would look at several things. First, cut the split way down, say 4:1 or so. Second, think about connecting the columns directly to the valve. Third, think about a bigger sample loop.

Since all of your components are gases anyway it is not absolutely critical what your inlet port is at. Having said that, it would be a good idea to at least keep it over 100 to keep the water from condensing out.

Best regards,

AICMM

Thanks for all the replies so far.
@AICMM:
I can't go down that low with the spillter ratio, because then I would "overload" the capillary column with all the other components, I think. But I think you are right, it is very hard to quantifiy the very small hydrogen peak, next to the giant nitrogen peak. Do you have experience with detecting hydrogen in helium? Would a µ-TCD instead of a regular TCD help, since I have a capillary column? I am thinking about using a second column with argon or nitrogen carrier gas, just to separate the hydrogen and analyze it in argon/nitrogen with TCD.

" I am thinking about using a second column with argon or nitrogen carrier gas, just to separate the hydrogen and analyze it in argon/nitrogen with TCD. "

This is a very good idea. You do not want to TRY to measure hydrogen with a helium carrier. Read the archives on this topic for details. Best gas is nitrogen.

Rod

buxbaum,

I have to apply my standard disclaimer here. You see, I make a helium ionization detector. Having made the disclaimer, a helium ionization detector is ideal for this application (be it mine or the competitions.) It will see hydrogen no problem (discrete positive peak.) It will see all of the components you are currently looking at and it will work quite well at the concentrations you are talking about with split injections (although you might end up cutting the loop size down.)

Rather than adding a second TCD, might I suggest you replace your current TCD with an HID and carry on? They aren't that hard to run (if done right), they can use FID electronics if you have that handy, they only require one carrier gas, you only have to have one injection valve (not two as with 2 TCD's), etc., etc., etc. (to borrow from the King and I.)

Feel free to write me at aicmm at flash.net for further discussion.

Best regards,

AICMM