Oxygen Analyzers Vs Gas Chromatography Results

[arnaeem]

I am looking forward to an expert opinion on the following matter:

We are representing a (P.S.A. Carbon Molecular Sieve based) "industrial" Nitrogen gas generators manufacturing company. We've recently commissioned an N2 generator designed to deliver N2 gas with 10 PPM Oxygen contents. If someone is aware of Pressure Swing Adsorption, then the N2 gas being produced (by this technique) is a mixture of itself, O2, Argon and may be some other traces in it (we don't know as our main concern is to monitor the concentration of Oxygen). Furthermore, our generator has a built-in (online) Oxygen analyzer in it.

Our customer used to buy N2 gas cylinders in the past and the gas company claims it to be 99.999% Nitrogen with a couple of O2 PPM. The cylinder gas is very pure indeed as it's being produced from liquid Nitrogen!! The problem creates when the customer collects a sample from gas cylinder and run a test on their Shimadzu GC (which they purchased a few months ago). Their GC detects the Oxygen contents as 68 PPM and if we allow the same cylinder gas to run through our integral analyzer, it also reads it around 66 PPM. It means our analyzer is working fine? On the other side, when our generator is in running / production mode, the same integral (built-in) O2 analyzer shows the O2 contents as 6 PPM. The customer then collects the gas sample from our generator in the same container and under the same ambient conditions, but their GC always detects between 1.5% - 1.7% Oxygen contents (which should actually be around 10 PPM)??

As I have already stated, it's a brand new system and we have verified that our O2 sensor + analyzer are calibrated and sensing low / high contents of Oxygen in different N2 gas streams.

I am not an expert in GC's at all ..... Is there a possibility of "wrong" method being applied on their GC (column), which is not able to detect / separate Oxygen from other traces and hence depicting a misleading composition?

Will really appreciate to have some valid arguments / opinions.

Regards,

ASAD RAZA NAEEM
asad.arne@yahoo.com

[chromatographer1]

As you have noted the problem lies in their GC analysis.

This can be due to the collecting the sample, the injecting of the sample, or the identification of 'oxygen'.

My guess would be one of the first two. BUT consider: Did they empty their cylinder before they filled it with low O2 content nitrogen? If Oxygen is in the cylinder then they contaminate the nitrogen with oxygen. Sound possible?

Do they get low numbers when they test certified gas samples?

Without the details any answer is only a guess.

Gook luck,

Rod

[JI2002]

Sample handling is the key. Given 210,000 ppm of oxygen in the room air, the samples from the cylinder/nitrogen generator can be easily contaminated with oxygen from the point of sample collection to the point of injection to GC. No matter how well the lines, syringes, loops... are flushed, possibility of sample contamination can not be eliminated.

[arnaeem]

@ JI2002 and Rod

Thanks to both of you!

The sample collection method is the same for both Cylinder and our N2 gas generator.

The sample is collected in a small 316L SS pipe of dia around 3 inches and 6 inches in length. The customer has fabricated this sampling pipe in shape of a "sampling bulb" with ball valves at inlet and outlet.

The N2 gas from a high pressure cylinder is regulated down to 5 barg and flown across the sampling assembly for 3-4 minutes. The outlet valve is closed followed by the inlet valve. Hence, they trap a sample gas within this assembly and take to their lab in order to test on a GC. There is a specific sampling port with Silicon seal. They use a special needle and syringe to tap out further samples for the GC. There's apparently no leakage during collection of a sample in the assembly.

The inlet/outlet valves are opened in order to release any previous sample inside!!

The same ambient conditions and method are followed in case of our Nitrogen gas generator! Everything looks the same apparently?

[chromatographer1]

"
The N2 gas from a high pressure cylinder is regulated down to 5 barg and flown across the sampling assembly for 3-4 minutes. The outlet valve is closed followed by the inlet valve. Hence, they trap a sample gas within this assembly and take to their lab in order to test on a GC. There is a specific sampling port with Silicon seal. They use a special needle and syringe to tap out further samples for the GC. There's apparently no leakage during collection of a sample in the assembly.
"

Very easy for human error to enter into the situation here, especially in the sampling from the cylinder, also known as a 'bomb' in industry circles. It doesn't take much of an error with a syringe to leak 1% air in the syringe sample, or to incompletely flush the 'bomb' with the previous sample of gas.

A sampling valve with enough sample gas to fully purge the supply line, the sample loop, and the vent line is the PROPER hardware to perform gas analysis.

There is an old expression that applies here sadly: GIGO

garbage in, garbage out.

best wishes,

Rod

[arnaeem]

Our customer is using the following instrument:

Shimadzu GC-2010
Shimadzu, Molecular Sieve 5A, 3.0m x 3.0mm packed
TCD detector
TCD Temp 220 °C, Column 70 °C, 95 °C, 10 °C/ min, Injector temp 180°C, Flow rate 50 ml/min
Atmospheric air is used as standard

From the Chromatograms, it seems that the GC is only able to detect N2 and O2 gases. I have three different chromatograms i.e. atmospheric air, Gas Cylinder and the Gas from our N2 generator. How can I share with you? Please let me to have your email address.

[chromatographer1]

Use chromatographer1 at either yahoo or aol

I will review the 'grams when I receive them.

Rod

[arnaeem]

Is this chromatographer1@yahoo.com or chromatographer140@yahoo.com?

[arnaeem]

@ ROD

Email sent at chromatographer1@yahoo.com and aol.com both.

Regards,

ASAD

[chromatographer1]

ASAD

There are reasons for my not typing my email address in a public forum.

Are you aware of them?

Rod

[arnaeem]

I'm sorry as I'm really not aware of them ...