retention time increase

[MARLIN31]

Hi, all,

I am a freshman in GC and my experience is unsufficient to fix the problem I met. We analyze beer for vicynal diketones - diacetyl and 2,3-pentandione. HS, ECD and a packed column (Mss316 Chromosorb101) are used in the analysis. Calibration curves for the compounds are good. But when I run a beer sample, the retention times of VDK increase significantly, e.g. from 3.85 min for diacetyl in the standard solution to 4.12 min for it in the beer sample. When I added the standard solution to beer, incubated it for a couple of days, nothing changed. Two peaks with the maxima at 3.85 and 4.12 min were observed. I checked also if the ethanol content of a sample affects the retention time. Again nothing changed in the retention times. I supposed the yeast biochemistry could affect the results of the beer analysis. Beer from another brewery was analyzed, again the retention times were longer, however a little bit less (4.09min), compared to those for standard solutions. So, I can not say that the problem results from contamination of the column, the effect of some beer constituent or yeast biochemistry. If somebody has any idea what may be the reason for this difference in the retention times of VDK in standard solutions and beer samples, please, let me know.

Thank you,

Alex

[dliebert]

How exactly my similar experience translates to the headspace world, I don't know, but I've seen similar shifts in fish tissue samples with residual lipids in them. Imperfect cleanup leaves some goo in your sample, it makes it onto the column and slows everything down. It's a rarity, and I deal with it by making a special quantitation method with sample-specific retention times. Now, if you're using headspace the goo in your system has to be a volatile goo, eh? So, maybe you can adjust your incubation temperature to prevent it from volatizing, clean-up your beer before analysis to get rid of whatever it is, .... You might run your beer on an MS in scan to try to identify whatever it is. Another way to deal with this sort of matrix effect is to make your calibration standards up in your sample matrix, ie. beer, and to process them just like you do your samples. My first guess would have been the ethanol too, BTW. Good luck.

[Don_Hilton]

What are your headspace conditions (and type of headspace sampler), inlet conditions and GC oven conditions for the analysis -- temperatures and flow rates? What is the solvent in which your standards are made - and run in the headspace vials. And, do you have the same sample volume for the standards as you do for the beer sample? And what are the dimensions of the GC column?

[MARLIN31]

What are your headspace conditions (and type of headspace sampler), inlet conditions and GC oven conditions for the analysis -- temperatures and flow rates? What is the solvent in which your standards are made - and run in the headspace vials. And, do you have the same sample volume for the standards as you do for the beer sample? And what are the dimensions of the GC column?

All conditions were kept the same both for standars and for beer samples. The solvent for standards was 5%(v/v) ethanol. As I previously wrote varying the content of ethanol in the standard samples (from appr. 1 to 50% (v/v)) had no noticeale effect on the retention time of diacetyl. It was around 3.85 min. Inlet temperature - 160 C, oven temperature - 150 C, carrier gas - nitrogen, flow rate - 30ml/min, headspace bath temperature - 50 C, equilibration time - 15 min, transfer line temperature - 110 C, volume of vials - 15 ml, volume of a standard/beer sample - 6ml, column length - 2m, id - 2mm, ECD temperature - 300 C. HS sampler is domestic one and I am not sure that I can translate its type correctly.

[Peter Apps]

Hi, all,

I am a freshman in GC and my experience is unsufficient to fix the problem I met. We analyze beer for vicynal diketones - diacetyl and 2,3-pentandione. HS, ECD and a packed column (Mss316 Chromosorb101) are used in the analysis. Calibration curves for the compounds are good. But when I run a beer sample, the retention times of VDK increase significantly, e.g. from 3.85 min for diacetyl in the standard solution to 4.12 min for it in the beer sample. When I added the standard solution to beer, incubated it for a couple of days, nothing changed. Two peaks with the maxima at 3.85 and 4.12 min were observed. I checked also if the ethanol content of a sample affects the retention time. Again nothing changed in the retention times. I supposed the yeast biochemistry could affect the results of the beer analysis. Beer from another brewery was analyzed, again the retention times were longer, however a little bit less (4.09min), compared to those for standard solutions. So, I can not say that the problem results from contamination of the column, the effect of some beer constituent or yeast biochemistry. If somebody has any idea what may be the reason for this difference in the retention times of VDK in standard solutions and beer samples, please, let me know.

Thank you,

Alex

When I added the standard solution to beer, incubated it for a couple of days, nothing changed. Two peaks with the maxima at 3.85 and 4.12 min were observed.

So if you run standards you get a peak at 3.85 min, if you run beer you get a peak at 4.12 minutes, and if you run been spiked with standard you get two peaks; one at 3.85 min and one at 4.12 min. Is this correct ?

If this is what is happening then the peak at 4.12 min is not diacetyl.

Peter

[MARLIN31]

So if you run standards you get a peak at 3.85 min, if you run beer you get a peak at 4.12 minutes, and if you run been spiked with standard you get two peaks; one at 3.85 min and one at 4.12 min. Is this correct ?

If this is what is happening then the peak at 4.12 min is not diacetyl.

Peter

Hi, Peter,

Thank you for your suggestion. I was thinking of that. But the problem is that there is no peak at around 3.85 min in all my beer samples. Just a smooth line. I analyzed different beer types, "green" beer - no traces of diacetyl at 3.85 min. So, what may happen to diacetyl in my beer samples?

Alex

[Don_Hilton]

spike diacetyl into the beer sample until you see a peak grow into the chromatogram, corresponding to the increasing level of diacetyl added. This will prove the location of the peak. Or, if nothing increases after a healty addition of diacetyl, there is somethign else going on.

[Peter Apps]

So if you run standards you get a peak at 3.85 min, if you run beer you get a peak at 4.12 minutes, and if you run been spiked with standard you get two peaks; one at 3.85 min and one at 4.12 min. Is this correct ?

If this is what is happening then the peak at 4.12 min is not diacetyl.

Peter

Hi, Peter,

Thank you for your suggestion. I was thinking of that. But the problem is that there is no peak at around 3.85 min in all my beer samples. Just a smooth line. I analyzed different beer types, "green" beer - no traces of diacetyl at 3.85 min. So, what may happen to diacetyl in my beer samples?

Alex

If I recall correctly from working on diacetyl and beer off flavours many years ago there should be very little diacetyl in the samples - it produces a flavour fault if the concentration is too high. So the simplest explanation is probably the correct one - you do not see a diacetyl peak in the samples because the diacetyl concentration is below the limit of detection.

You should use Don's suggestion to establish what the limit of detection is.

Peter

[MARLIN31]

spike diacetyl into the beer sample until you see a peak grow into the chromatogram, corresponding to the increasing level of diacetyl added. This will prove the location of the peak. Or, if nothing increases after a healty addition of diacetyl, there is somethign else going on.

Thank you, Don,

I'll try to do it asap. Another thing I'm going to check is to distill a beer sample and then to run the distillate on GC to see if diacetyl is present in it. Simultaneously I'll check the sample by a colorimetric method.

[MARLIN31]

If I recall correctly from working on diacetyl and beer off flavours many years ago there should be very little diacetyl in the samples - it produces a flavour fault if the concentration is too high. So the simplest explanation is probably the correct one - you do not see a diacetyl peak in the samples because the diacetyl concentration is below the limit of detection.

You should use Don's suggestion to establish what the limit of detection is.

Peter

When I prepared the calibration curve the minimal concentration of diacetyl I used was 0.02mg/l (20ppb). At this concentration the diacetyl peak was really small but still noticeable. Usually the concentration of diacetyl in lager beer (what we brew) is 0.04-0.14mg/l. So it seems to be increadible that all the samples I have checked are below 0.02mg/l in diacetyl. Anyway, I'll check Don's suggestion.

[jdezeeuw]

I am surprised that you can work with packed columns for these compounds. When I did "beer" There are a series of compounds that all elute before the ethanol peak and we had to specific capillary columns to get all the separations. Typically carbowax 400 was used, but later on also bonded wax phases were applied. we needed 30m at least to get separations.

If you are using chromosorb 101, you probably will accumulate all the late eluting materials on there. So be prepared for some impact on the chormatography.

jaap de zeeuw, Restek corporation

[MARLIN31]

I am surprised that you can work with packed columns for these compounds. When I did "beer" There are a series of compounds that all elute before the ethanol peak and we had to specific capillary columns to get all the separations. Typically carbowax 400 was used, but later on also bonded wax phases were applied. we needed 30m at least to get separations.

If you are using chromosorb 101, you probably will accumulate all the late eluting materials on there. So be prepared for some impact on the chormatography.

jaap de zeeuw, Restek corporation

For analysis of beer we use different columns and detectors. As for VDK a packed column and ECD is a standard configuration. For other compounds we use capillary columns and other detectors.
As for your suggestion, it is interesting and it could be possible that some beer compounds left in the column could affect the diacetyl retention time but I can not understand how these compounds can discriminate between beer's and standard's diacetyl. These compounds should act in the same way either to diacetyl produced by the yeast or to diacetyl added to the beer sample. But we see two separate peaks on the chromatogram of the beer sample spiked with diacetyl.