Chromatography Forum

Hi Everyone,
I have been tasked with calculating the detectivity of a FID detector.

Can anyone please advise me on the steps i need to do this?

Thanks.

You have been given an impossible task! The detectability of the detector varies with the susceptability of the molecule. For instance you won't see any pesticides (at the required level of ppt) at all with an FID so we use an ECD (Electron Capture Detector) instead.

Hi,
Thanks for the response. Impossible it does seem at the minute.

Starting point is a FID test mix with the target compound being hexadecane so at least i have a starting point!

There are a number of schools of thought on this. Your limit of detection will be the lowest concentration where you can reliably say that the signal your detector gives you is above the noise. The one that i see most is that limit of detection is defined as the amount of analyte that gives a signal that is 3 times larger than the noise in the baseline (I usually take that noise at the retention time for the analyte).

This topic comes up frequently. Use the search box above to see if you can find some threads that might help you.

Thank you for your recommendations. will definitely search for this.

Do you want to calculate it or measure it ?

Assuming the latter, it is not a trivial task. With a GC inlet and column feeding the detector you will never know how much sample you are feeding to the detector. You need to feed sample directly to the detector, ideally from a diffusion standard that you weigh at intervals to measure how much sample it si feeding to the detector - you can control its temperature with the GC oven and use the GC pneumatics to flush gas form the diffusion standard to the detector.

Or you can report the manufacturer's claim. It is hardly ever the detectivity that is the limiting factor in an anaysis.

Peter

There is usually a section in the instrument manual for the detector that gives some parameters of absolute limits of detection as determined by the vendor. You can perhaps back out of those estimates an idea of how to calculate it for a given column and flow.

Hi guys,
sorry was offline for the bank holiday weekend (in UK). I am working with the vendor to try and lower their specified detectivity.

I have a few ideas of how I will approach this now so just going to try a few different methods to see what works best (and gives the most reproducibile results)

Thanks for your suggestions.