Bad Air?

[sassman]

We are performing a method validation on our GC (agilent 6890). When we were about 1/2 way through the response of our main band (ethanol) skyrocketed. It actually went off scale. This also correlated with us changing the air tank. When we changed the air tank again today the response dropped back on scale. What would cause this? How could the air cause your response to increase. We are also using a headspace sampler. When we inject a blank with no ethanol we do not see an ethanol peak.

[chromatographer1]

Ah, a mystery !

New air gives a better response on the FID.

Old air gives a poorer response on the FID.

Which is out of spec ?

My guess is that the new air is not super air, but the old air is bad, contaminated with something that burns (methane, hydrogen?) so the response is lower for ethanol. (different air fuel ratio)

Do you see any change in flame color with the change in air supply?

best wishes,

Rodney George
consultant

[Peter Apps]

If you are using synthetic air i.e. a mixture of nitrogen and oxygen it is just possible that you had a cylinder with the wrong ratios. People have tried enhancing the FID response by changing oxygen concentration, but without significant success so this is not likely. I imagine that there would be considerable interest in a gas mixture that boosted the FID's sensitivity, so you should think about finding out what was in the super air.

Peter

[HW Mueller]

Could it be that totally different air flows are behind that, somehow?

[Peter Apps]

Hello Hans

You are right; changes in flow are the most straightforward explananation, but the 6890 GC has electronic flow control on its flame gasses, so it should be relatively immune to anything happening upstream of the GC, and if it cannot get the flow that it wants it flags an error and stops injections. As with all things electronic this can lead to a false sense of security, and independent checks of the flow rates are a first troubleshooting step.

Peter

[larkl]

I've seen this before. We were burned so many times on one headspace method that we started checking the calibration every time we put a new air 6-pack in service.

[sassman]

Thanks everyone. I was thinking that it could be that the previous air tank had been contaminated with hydrocarbons resulting in decreased response. Changes in flow could also have such an effect, but I wouldn't expect any change as long as the upstream pressure remains the same (right?). Of course the flow controller could go bad, but I still think that bad gas is more likely given that this happened immediately after changing the tank.

[coffee bean]

You haven't mentioned anything about peak width. If your peak width have increased as well and you're on split mode, chances are that your split line is partially blocked. If so, you should replace the split vent filter and maybe clean/replace that copper tubing that goes from the inlet to the split vent filter.

[Peter Apps]

Hi Sassman and larkl

How large was the change in response - a few %, tens of % ? An FID is remarkably robust to changes in air flow, and nobody has ever managed to come up with a gas mixture that significantly improves its sensivity, so your observations are really interesting. What happened to the noise when the signal was increased ?

Has anyone ever seen this with compounds other than ethanol (which has more oxygen in it than most compounds analysed by GC).

Peter

[Ron]

What grade air are you using, and are there any purifiers in the line between the cylinder and the instrument? Is the same supplier supplying all these cylinders?

[tima]

Maybe the tank is not air but pure oxygen??? Timothy