TCD Oxygen Linearlity with HP-MOLSIV Column

[tmalew2]

Does anyone know how sensitive is the TCD to Oxygen when having a
HP-MOLSIV 30m x 0.32mm (ID) x 25um (FILM) column. I am trying to do
calibrations of Oxygen and doesn't seem to be linear. I am injecting
from 25ppm to 2500ppm also have injected 50000ppm.
The sample loop I believe is 250ul in volume. Injecting between 15 to 16psi.

When injecting between 25ppm to 2500ppm I get a curve where the slope gets bigger at higher ppm levels. At 50000 ppm the slope is about 25% higher than at 2500ppm. Between 25ppm and 2500ppm I get a huge difference, bigger than an order of magnitude.

Does the TCD is more sensitive to higher levels of Oxygen or 25ppm is to small or am I using the wrong column for these levels of oxygen.

[chromatographer1]

Based on experience I would believe that it is probably not the detector but the column that is the source of your non-linearity.

Try another column that is not sensitive to oxygen and see if your linearity improves before you touch the detector.

In other words demonstrate that it is the detector that is non-linear with another column. You could use porous polymer or carbon sieve column to separate methane and other gases from the oxygen peak. Nitrogen will not be separated unless you use a carbon sieve.

Good luck and let us know the results, would you?

Rodney George
consultant

[tmalew2]

I am actually trying to detect Oxygen and need to detect oxygen. I am able to detect oxygen but can't get a calibration for it. We only use the TCD with the HP-Molsieve to quantify oxygen and noble gases (but for krypton or neon we don't care how much is there). As for Oxygen we are running oxidation experiments in low ppm levels, and we need to be able to quantify how much oxygen before and after the experiment.
We are just having problems with the Oxygen calibrations, its never linear.

Here is my email tmalew2@uic.edu
I can email someone calibration data for oxygen and show you how the calibration curves changes with increasing ppm levels. I just finished the 50000ppm of O2 calibrations and I got an average of 1.25E-02 ppm/(Area/psi) when I do 25ppm of O2 I get a magnitude smaller. I am assuming its not the TCD but I assume that with the column I have the TCD has a better sensitivity to higher concentration (50,000ppm) than smaller (25ppm).

Since in our experiments we are starting with the max of 2000ppm of O2 how do I calibrate at those levels and lower and get it linear.

I am thinking that I need to get the peaks of O2 bigger at 2000ppm versus what I have now, so maybe use a larger sample loop or do I use a different size column.

[chromatographer1]

I understand what you are doing and the results you are getting.

Please understand that I think that you think that the column is not affecting the linearity of the oxygen response and that the loss of oxygen at low levels you believe to be caused by the detector response. You do not believe it to be a problem with the column.

I believe you make this assumption without a basis in fact. I believe that it is the column that is the problem, not the detector. Do you follow me?

Until you modify the sieve so it does not consume oxygen you will have losses of oxygen at the lower levels. If you were loosing 10 ppm of oxygen with a 25 ppm sample and loosing 10 ppm of oxygen with a 50,000 ppm level sample you would show a non-linear response over the range of your samples. The chemical reduction of the sieve can happen if you are using hydrogen carrier or it if was not conditioned properly before use. Then it will consume oxygen as the gases pass through the sieve.

Oxidize your sieve so that it does not consume oxygen and you WILL get a linear response for oxygen samples. Injecting dilute samples will not do this quickly, but very slowly over a long time. While you might be loosing oxygen at the detector as it reacts with the hot metal in the detector it is not likely.

Good luck and best wishes,

Rodney George
consultant

[GasMan]

I assume that you are using helium as carrier gas and not hydrogen. If you are using hydrogen, then oxygen will react with the hydrogen on hot metal surfaces.

Are you injecting into any other column before the sample goes onto the Mol Sieve column. I have seen problems when oxygen first goes through a Porapak type column if that column has been above 150°C for any length of time. I have also found that the tubing material will have an effect on oxygen levels. I would recommend using nickel tubing for any line that is in contact with your sample.

Gasman

[mattj63]

As I understand it, you calibration is curving upwards (when you have signal on y-axis, and concentration on x-axis). I doubt this is your detector as someone else has already mentioned. TCD is usually pretty consistent and not much to change there. I don't analyze gases but your problem is most likely the chromatography and not the detector.

[MMcI]

Oxygen analysis at low ppm / ppb is often impossible due to the iron content of the tubing or the absorbtion by the column packing.

Use silcosteel transfer tubes to solve the iron problem.

A low water content in mol-sieve will cause losses in oxygen.

Condition the column only to the point where O2 and N2 are separated.

150oC overnight may be enough.

Over-conditioning is evidant when O2 is lost and CO is tailing.

A linear output is possible down below 50ppb using discharge ionization
detector and below 5ppm with TCD.

Porous polymer columns will remove ppm oxygen and are not suitable.

[AICMM]

tmalew2,

Rod is spot on. Just finished a GC set up for 10 ppm fixed gases, balance helium, using DBD-HID mode. Beautiful peaks for all components except oxygen. Oxygen was gone. Now in discussions with the vendor about exchanging columns. In the meantime, purchased a different vendor sieve and oxygen worked beautifully. Problem is probably not the detector but rather the column.

Best regards.