Over 100 mole%??

[fergman]

Hello,

I'm very new to chromatography but am hoping that by the time I get my PhD, I'll be a real pro!

So I'm testing out a Varian CP-4900 Micro GC that we're borrowing from another lab. The columns in it are a porous polymer and a molecular sieve 5A. Right now I'm just trying to verify the calibration done by someone else by injecting room air into the system. My first problem, zero separation, was solved by baking the column out for 8 hours.

However, it will only positively identify N2 and O2 at 90 C and 120 C (at 150 the peaks were too close together, and below 90, they came out too late). Does this seem alright?

Second, the outputs of our method are given in mole %. For O2, I'm getting like 30% and for N2 113%. I checked the analysis of the previous operator, who built the method we're using, for air and he gets the correct proportions of O2 and N2, and also it worked for him at 60 C.

Any thoughts? Let me know what other information I can provide. Thanks!

[AICMM]

fergman,

in opposite order. Sounds like you do not have normalization on in your method and that may be why the results are reporting the way they are. If normalized, they should add up to 100%. What else were you expecting to see from room air injection? Perhaps CO2 on the second channel? What carrier gas are you running on each channel?

90C should be plenty hot enough to get air separated in a hurry so the next thing I would look at is the carrier pressure to see if it is high enough.

Best regards.

[Bruce Hamilton]

I don't know what detectors the instrument has. I assume you expect to see argon/oxygen/nitrogen +?..

If you can't see the individual peaks, we need more details about the instrument setup, gas flows, sample injection, and calibration.

It's quite possible that the detector you have gives different responses for each of the above gases, so you need to calibrate it using at least one standard gas mixture, provided you can see the individual peaks.

Note that normalisation is not usually required if a good calibration is obtained.

If you were only interested in the above, then you can use dried air and pure gases to make mixtures. You can make simple mixtures using a couple of 50-60 mL medical plastic syringes and the plastic luer-fitting medical three-way valves that are used on such syringes. Just remember to wave the mixture around like a flag for several minutes to ensure the gas mixture is homogeneous.

The advantage of such macro standards is that they can show which calibration response factors are incorrect. For accurate work, you should obtain a couple of calibrated gas standard mixtures that represent the upper and lower composition limits of components.

Please keep having fun,

Bruce Hamilton

[fergman]

Thanks for the replies all! At the advice of someone from another lab, I'm running a bakeout for two days because the instrument hasn't really been used since the end of June.

AICMM -

That normalize parameter seemed to do the trick. Now N2 is coming in ~78% and O2 ~22%. But why should I have to normalize my results? In other words, how could the instrument detect more than 100% in the first place? So one detector is set up for H2, N2, and O2 and the other one is set up for CO, CO2, CH4, ethylene, and ethane. All we saw in the carbon channel was a small amount of CO. In terms of carrier gas, it's argon. For carrier pressure, it's set at 30 psi for the carbon channel and 32 psi for the other channel...is this OK?

Bruce -

Right, the detectors are two TCDs. And calibration is definitely something we're going to redo, because it hasn't been calibrated since March and I need to learn how to do it anyway.


I'm still wondering why the N2 and O2 peaks are shifted to the right (come out later) when measured now as compared with when they were calibrated in March.

Thanks again,
=tom

[AICMM]

fergman,

You use normalize when you would be looking at the whole sample for all of the constituents. For example, if you are running CO2 and balance helium in a helium carrier, normalize would not be the best way to go. But when you are looking a complete mixture, then normalization makes sense since it should all add up to 100%. The reason that you probably had high values before normalizing is a slight drift in the TCD calibration.

If you are shooting room air, then on the carbon channel you should not see CO, only in the atmosphere at ppb levels and your TCD most likely won't see that low, especially with argon carrier. It is probably CO2 which should be a very small peak indeed if you are using argon carrier.

Regarding the retention time shift, I would guess that your head pressure is off a little bit low. Which also means your backflush is probably off a little bit as well. That being said, you may be getting water on your sieve.

Best regards.