GC-MS detection of furaneol

[rambiochem]

I am analysing fruit solvent extracts by GC-MS. In one of the samples, I found furaneol (one of the most important flavor compounds) before few days. But after that furaneol is not getting detected in the same sample. We tried re-extraction, that too didn't work. All other conditions are exactly the same. We do see a peak in TIC at the corresponding time, but in the MS spectra, the peak of m/z 128 (of furaneol) is replaced by that of m/z 140. Rest MS pattern appears similar. Can anybody help?

[Don_Hilton]

Whatever you are looking at is a larger molecule that furaneol. The fact that the new spectrum shows a peak at 140 - this is a difference of 12 from the molecuar weight of furaneol - it will not be the simple addition of a carbon atom. It would take examination of the full spectrum for me to be able to begin to guess what is going on. My inclination is that this compound is not a furaneol derivative resulting from standing in air or by reaction with solvent.

Since it looks like the furaneol is going away, what solvents are you using? Solvents like pentane and methylene chloride are nice because they will not react with common flavoring materials. Others like alcohols can form esters, ethers, acetals, and ketals. This can really confuse the analysis. Again, more information would help, including the steps in your preparation and storage conditions. And, in the second preparation, did you use the same batch of fruit and how was it stored. And if there was a nice strawberry note to the fruit in the first prep, but not the second, then there is also a clue. (And if you were looking in pineapple for this - so much for that idea.)

Has a library search given you a clue as to the identity of the new compound? And what MS libraries do you have available?

And, what column are you using? One possibility is to look at a list of flavoring compounds ordered by retention index (flavornet.org has a nice one - and free!) and see if that gives any ideas.

It also crosses my mind to take a look for the ion trace for m/z 128 to see if it is indeed totally going away or if there is this other compound coming in and swamping out the signal for the furaneol.

[ramkulkarni]

Whatever you are looking at is a larger molecule that furaneol. The fact that the new spectrum shows a peak at 140 - this is a difference of 12 from the molecuar weight of furaneol - it will not be the simple addition of a carbon atom. It would take examination of the full spectrum for me to be able to begin to guess what is going on. My inclination is that this compound is not a furaneol derivative resulting from standing in air or by reaction with solvent.

Since it looks like the furaneol is going away, what solvents are you using? Solvents like pentane and methylene chloride are nice because they will not react with common flavoring materials. Others like alcohols can form esters, ethers, acetals, and ketals. This can really confuse the analysis. Again, more information would help, including the steps in your preparation and storage conditions. And, in the second preparation, did you use the same batch of fruit and how was it stored. And if there was a nice strawberry note to the fruit in the first prep, but not the second, then there is also a clue. (And if you were looking in pineapple for this - so much for that idea.)

Has a library search given you a clue as to the identity of the new compound? And what MS libraries do you have available?

And, what column are you using? One possibility is to look at a list of flavoring compounds ordered by retention index (flavornet.org has a nice one - and free!) and see if that gives any ideas.

It also crosses my mind to take a look for the ion trace for m/z 128 to see if it is indeed totally going away or if there is this other compound coming in and swamping out the signal for the furaneol.

Thank you very much for the reply and for the information.
Here are the details...
Solvent: DCM
MS libraries available: (1) Wiley Registry of Mass Spectral Data, 8th Edition (2) NIST98
Column used: SLB5MS (Sigma-Aldrich)
The clues given by library search have very low significance (~300 out of 1000) so they do not point at any of the compounds confidently!
For extracting the flavor compounds, we crush the tissue (mango) in liquid nitrogen, put the powder into the solvent, keep shaking for 1hr, decant the supernatant, dry over sodium sulphate, concentrate on rotary evaporator and store the final extracts in glass inserts at -20 C. All the tissues used before and now have been stored at -80 C.
GC program:
40 C for 5 min
10 C/min upto 220C
220 C for 5 min
Carrier helium @ 1ml/min
Injector at 200 C, mass source at 180 C
I have sent both of the spectra to your e-mail address since they can not be attached here (am I right?)
Thanking you in anticipation
Ram

[Don_Hilton]

This will take more than a few minutes, but as I get a chance I'll look at it. I'll put down my thoughts (with the right to change my oppinion as I come back and look more) but I will have to come back to this - bed time here and I need to work in the morning...

Just because this compund coelutes with furaneol on this column, I will assume that the molcuar weight is not too much higher. So 140 is a good possibility. The presence of stuff at m-1 and m-3 would indicate that loss of a hydrogen radical leads to a stable ion and m-3, I ocasionally see with some very stable aromatic systems. Either that or the moecule is heavier and we are looking at fragments from multiple decomposition pathways.

Examiation of m/z 140 and 141 in the figure (I'll see if I can add that detail to this post later) Shows a ratio m/z141 to 140 of 8.0% There is no m/z 142 signal showing. This indicates that it is not lilely that there are many oxygen atoms in the molecule. (It is possible to loose the signal for a single oxygen atom if the spectrum is weak) And the ratio would lead you to look for something with 7 carbon atoms, maybe 8.

link to picture of potential molecular ion here

If there are only seven carbon atoms in this molecule, there are another 56 mass units to account for. And, that is not going to be hydrogen. Assuming that m/z 140 is the molecuar ion, then if nitrogen is present, we have to have an even number of them. And the parts don't seem to be adding up. And if this is C10H10, molecules with long alkyl prtions tend to fragment, leading to relatively weak molecular ions and no m-3.

So the question is: is this a deconvoluted mass spectrum? And if not, have you tried deconvolution software on the chromatogram to "clean up" the chromatogram? Some data systems have their own deconvolution software. If yours doesn't, you should be able to use AMDIS and there is proabably a copy on the disk that came with your NIST library.

We'll see what else comes to mind -- later...

[Peter Apps]

Hi Ram

It would be a big help if you could post the section of your TICs that includes the furaneol and the mystery peak. Instructions for posting graphics are in a sticky at the top of the LC page. Can you also look in detail at the spectrum across the peak on the two chromatograms to see if the peak is pure or not ?

Peter

[ramkulkarni]

Hi Ram

It would be a big help if you could post the section of your TICs that includes the furaneol and the mystery peak. Instructions for posting graphics are in a sticky at the top of the LC page. Can you also look in detail at the spectrum across the peak on the two chromatograms to see if the peak is pure or not ?

Peter

Hi Peter
Below is a part of TIC having furaneol.
Both peaks at that RT are pure. Before the spectra from all parts of the peak (although the peak was little flat) showed very high similarity with furaneol and now I don't know what it is but the spectra is similar throughout the peak.

[Peter Apps]

Thanks Ram, the pictures help a lot.

I am pretty sure that on the "before" chromatogram you have two peaks that are almost completely overlapping - hence the uniform spectrum across the peak.

On the "now" chromatogram the peak is a much better shape (which is not to say that it is not a mixture !).

Can you do single ion traces for m/z 128 and 140 for both chromatograms and post them for us ?

There are some other changes and shifts between the two traces - are then any very large peaks eluting close to this section of the chromatogram ?, and do they change ?

Is there any possibility of running this with two-dimensional GC, or on a different column ?

Peter

[rambiochem]

Thanks for the reply Peter....
Earlier elution in "NOW" is, I suppose, because we had to cut the column by few cm at the MS end before taking "NOW" runs.
Below are the ion traces...
Please refer to the earlier TIC for corresonding times.
m/z 128 is totally absent at the expected time in "NOW" indicating that furaneol is absent? Is it so?
To add to the mystery, a new peak has emerged for trace 128 "NOW" at a different RI (12.04 min). Its MS spectra which doesn't show significant similarity on library search is also attached below...

[Don_Hilton]

Looking at the chromatographic traces, the furaneol is gone. There is another compund that elutes close to the position where the furaneol was.

The source of the new peak or peaks is not clear. It (they) may not be related to the furaneol. If your sample is oidized, these are fragmetns of who knows what.

From your initial posting, I assume that you want help in figuring out why the furaneol peak went away.

I don't think that there is enough information in what we are looking at here to tell you what happened to the sample between before and after. An exhasutive review of the chromatogram may show some identifiable peaks that would give some clues. We'd have to have the data files (and the time) to look at these. I've not been able to get back to trying to get some information out of the after peak spectrum. (Life gets in the way of having fun.)

Am I missing what you wanted to know?

[Peter Apps]

Adding to what Done says:

You might have had furaneol before, but if it was there it was mixed with something else, which means that any library search is highly suspect. Whatever it was has subsequently disappeared.

Judging from the baseline noise the new peak at 128 is a lot smaller than the original one, which may be why you do not get a good library match.

Do you have a standard of furaneol that you can inject to establish its retention time ?

Peter

[rambiochem]

Thanks Don and Peter for the opinions.
Don...
Yah it seems that furaneol has gone away form this extract; but now we are trying re-extracting the same tissue, but this too is not working...
If we consider that furaneol is undergoing some kind of chemical modifications upon storage, why it is not seen in the the fresh extract is the only question that remains unanswered now!
Peter...
Forgot to answer your previous question...
Neither we have facility for 2D GC, nor do we have any other column...
We don't have std. furaneol also..
I was behind this compound since it is one of the imp flavor compounds.
anyway now i think i will have to lose this compound from the list of abt 50 mango volatiles

[WK]

Try using a new liner or replacing the wool.
WK

[rambiochem]

Ok...I will try this also...
But how do you think this would help?

[WK]

Furaneol,maltol,ethyl maltol will all be difficult by GC.
I have determined them in simple flavours by HPLC.
WK

[rambiochem]

That's a grt info
Thanks!