Response Factor of ECD, FID and TOF-MS

[Peter Apps]

So take him up on his offer . Take two different batches of the same product, and spike one of them with some chlorophenols at single ug/kg levels (this is a very common off-odour problem). Let him see what he can find.

Please post the results here.

Peter

[zupahfly]

AHAHAH

Well actually that guy made us two samples as a favour to himself i suppose in order to sell us a service. That's why i'm here tryng to have a "free of interest" opinion.

Moreover, I dont think he will run anything else for free...

But i got you point.

[Don_Hilton]

I do GCxGC-TOFMS - and it is a great technique. I also have a wrench in the tool box in the garage and it is a great tool. For nuts and bolts, the wrench works great - but when I need to cut something...

Consider your samples -- both the target of your analysis and the matrix. It is fairly easy to spike an essential oil with some other compound and then find the difference. So, that demonstrates that you may be able to find the off odor if it is a contaminant and you have a sample of that same lot, uncontaminated, for comparison. And this assumes that your contaminant is at a sufficient concentration you can detect it - consideration of the target. (On my GCxGC-GOTMS instrument I expect to get useful spectral information for identification above about 1 ppb - and that is for something like a polychloroanisole.) Now for consideration of the matrix: If you have two different lots of an essential oil, you immediately include all the natural variation that occurs in plant material. If you are looking to compare a retained sample lot of an essential oil with a sample from the same lot that has spoiled, you encounter all of the chemical changes that have occurred in that material - and expect many that have nothing to do with the fragrance. In these cases where you are not analyzing the same oil, (just spiked and unspiked) a lot of things change - most are unrelated to sensory quality.

To determine significant differences between a sample that is bad and one that is good, you need to do something like run many good samples to measure the natural variation of all the components. Then, you run the bad sample and see what falls outside of natural variation. Understand the training set must be sufficiently broad to encompass the natural changes. For example, you can run menthol samples from all over the world - and you get them from, say a 2010 or 2011 crop. Now you run a menthol with an off note -- if you are working with a sample from the 2009 crop, expect to find many compounds outside of the expected range.

And speaking of menthol, a source of an off note is the presence of enantiomers, which you will not be able to identify by GC/MS - the spectra look the same. The peak that should have a minty smell will just smell like the mint is "dirty." No wax or methyl silicone type column will separate them. (Here is where you drag a chiral column in.)

I've run essential oils by GCxGC-TOFMS, and it is easy to get over 30,000 peaks from such a run. Using the statistical techniques I described, you may be able to reduce the number of peaks of interest significantly. There is a chance, you will see a chemical name from a library hit that you can recognize - and can purchase - and has the off note that you are hunting. But, if I have to go with chances, I'd buy a few lottery tickets before betting on this one. (The second lottery ticket is the backup plan.)

The most accurate detector of the off note is the nose. Keep the nose in the picture as you go through the discovery process. Is there a place for GCxGC-TOFMS with essential oils? Yes. It is great for pesticide analysis with no sample prep or other target analysis for multiple comonents at low concentration. It may be good for forensic fingerprinting of oils. And those are just what pop into my head. Consider the analytes of intest and the matrix - how will you extract the information you need from the technique? Now pick the tool.

(I did shorten a piece of wood with a wrench once - used the wrench as a hammer and beat on the end of the wood until I could fit it into what ever it was that I needed to prop open. But that is a story for another time...)

[zupahfly]

Hi there,

Perhaps i should be more clear about this. I'm responsible for quality control/R&D at a cork transformation company. We perform everyday sensorial analysis trhough a group os panelists who sniff cork samples tha stay in a water soak 24h (this is a ISO procedure and has to be like that). What is happening is that sometimes off odours are detected among those macerations. These off odous are usually like musty, mouldy, earthy, mushroom, green pepper and so on... We currently have GC-ECD to control chloroanisoles, but only. Since this is being a growing problem, we want to eliminate them and our first approach is to identify them so we cand develop techniques that will clean our products. We are talking here of acting compounds at 1 ppt levels...

So my job now is to find the best, more reliable (and cheap if possible) way to identify these compounds so we can somehow latter find a solution to eliminate them. We've been across several techniques, but one of them was GCxGC TOFMS which seem very informative, but the correlation with the nose seemed a bit weak, because it is always based on statistics. Nevertheless, the thing is that i'm not a GC expert and thats why i'm asking for some help here...

You guys have been terrific! But i'm still not sure which path to follow... Lets say, money its not a problem, Where should i go?

[Peter Apps]

Really this depends on what in house analytical capability you have. If by good fortune the off-odours you get are due to known compounds that other people have identified (and there is a lot of literature on cork taints and wine off-flavours) then you might be able to harness the GCxGC-TOFMS to do targetted analyses. I would recommend this as a first step anyway to avoid re-inventing the wheel.

If you have new compounds, that nobody has identified before you are going to need the full GC - GC sniffing - MS set up I described before, but you will also need high res MS to get the structure, if the off-odours are at ng/l you are never likely to get enough of them to do NMR. This is highly specialized work, and unless you have a really good lab, with all the hardware and people who know how to use it I would seriously consider contracting it out.

Peter

[GOM]

Hi,

You may well have considered these but whenever I had a sample with malodour descriptors of musty, mouldy, earthy, mushroom then it usually had one or more of the following

Geosmin,
2-methyl isoborneol,
2-isopropyl-3-methoxypridine,
2-isobutyl-3-methoxypridine,
octen-3-ol,
3-octanone

Regards

Ralph