Page 1 of 1

GC-FID Question

Posted: Wed Oct 09, 2013 9:19 pm
by Chenosuke
Hi,

I work at an alcoholic beverage company and one of my tasks is to quantify ethyl acetate (among other compounds) in our products. The method we use is pretty straightforward in that there is minimal sample prep - just add an aliquot of internal standard to the sample and go.

Now, the issue I have here is this: I noticed recently that there is a fairly uniform decrease in peak response for all compounds depending on the alcoholic strength of the sample. The trend is: the higher the alcoholic strength (%alc/vol), the higher the peak response. We have products and ingredients that span a range of strengths from 40-96%alc/vol, but I ran a quick experiment recently which has me questioning the method a bit. I ran a product at 80% strength, then diluted it with distilled water to 40%. The peak areas for all compounds dropped by 35-40%. This, fortunately was also true for the internal standard, so it compensated for the loss when calculating the final concentrations. Also, I should point out that all retention times remained repeatable to within 0.002 minutes.

The question I have is, what exactly does water do to my system to reduce the response of every single compound in my run? I am now preparing more samples to run to see if the amount of reduction is repeatable and also I will test at different strengths to see if it is a linear reduction as the water content of my sample increases. The reduction was observed on both GC systems with identical methods loaded.

I am running 2 Agilent 6890 GCs with J&W DB-624 (60m x 0.25mm ID) columns installed, identical auto-injectors, identical FIDs.

Carrier gas is H2 @ 50cm/sec
Inlet temp = 250 deg
Injection volume = 0.5 microlitre (to reduce chance of backflash considering water's expansion volume)
Split Ratio = 10:1
FID temp = 250 deg

I think I've posted all the relevant information about my setup, but if you need any more, feel free to ask.

Any ideas about this would be greatly appreciated, thanks for your time.

Chen.

*edit*In my rambling, I forgot to specify one thing in my question. The reduction in peak response that I have observed so far does not appear proportional to the dilution. Meaning, the reduction I saw in response based on a dilution from 80% to 40%, was different from what was expected from a dilution from 80% to 60% and from 80% to 70%, which suggests that there is another factor here at play other than simply the dilution effect.

Re: GC-FID Question

Posted: Thu Oct 10, 2013 7:06 am
by Peter Apps
My first guess is that it has something to do with the different expansion volumes affecting inlet pressure and split ratio.

If the rationale behind diluting high strength products is to be able to use a single calibration for all products, then why not do the "dilution" with alcohol ? which will give a much more tractable matrix.

Peter

Re: GC-FID Question

Posted: Thu Oct 10, 2013 4:28 pm
by Bigbear
Your assay would work better if you had access to a headspace unit. That being said what liner do you use ( 4 mm I hope) , what is the inlet pressure?

Re: GC-FID Question

Posted: Thu Oct 10, 2013 7:12 pm
by Don_Hilton
I don't know how large the change in response is, but I'll toss an idea out for you: What happens if you let the needle stay in the inlet just a bit longer after the injection completes - or if you increase the inlet temperature a bit? I am wondering if this could be related to heating of the needle and evaporation of the volume in the needle being a function of the alcohol/water mix as the needle warms in that brief bit of time the needle is in the inlet...

Re: GC-FID Question

Posted: Thu Oct 10, 2013 9:15 pm
by Chenosuke
Thanks for all of your responses.

@ Peter: You know, I have thought about "diluting" with alcohol in the past, but somehow, it never crossed my mind when thinking about this. Now that you mention different expansion volumes and split ratio, I think I will try running the samples splitless and see what happens. I'll definitely revisit the "dilution with alcohol" idea though.

@Bigbear: I wish I had access to a headspace sampler (my boss isn't very receptive to the idea of buying one). Right now, I try to keep the samples as cool as I can and leave not so much headspace in the vials. For this application, I use 78.5 x 4mm split liners. Inlet pressure is usually somewhere around 21.7psi.

@Don: I could try extending the needle's dwell time. I'm not so crazy about increasing inlet temp though. I read somewhere that furfural (one of our analytes) decomposes when heated above 250 degrees. That would be bad. Also, if the needle trick doesn't work, chances are increasing inlet temp won't do it either.

Thanks for the ideas!

Chen.