Chromatography Forum

An usual method for quantitation of volatiles by GC is (1) identify by GC-MS (2) take 2nd run on FID for peak area determination for quantitation (3) match peaks of MS with those of FID chromatogram to determine quantity of identified compounds.
On this ground, had anybody quantitated compounds directly by areas of GC-MS chromatograms relative to those of internal standards? Is this a valid method? Are there any papers available for such study on plant volatiles?

There are hundreds of methods based on quantitation by MS, but the ones that gave good results all calibrate the area of the analyte compared to the area of the internal standard vs the quantity of the analyte.

As I understand it, what you are proposing is to compare the area of compound A with the area of internal standard B and then quantitate A as A/B times quantity of B. Then you are in questionable territory no matter what detector you use.

Peter

Some folks make an assumption that the FID trace is effectively a mass trace -- and it comes fairly close for most compounds of interest to someone like a flavorist. The FID essentially counts carbon atoms. And highly oxygenated compounds (like sugars) will never make it down the column. So, you can expect a major portion of the mass of the molecule making it to the detector to be carbon. (Remember, we are looking at the kinds of things in flavors, and aside from H2S and NH3, there is some carbon in just about every molecule that we would want to smell - and some that are of natural origin that we do not want to smell.)

Taking this a step further, the total response on the MS roughly follows that of the FID. So, some will use the Total Ion Chromatogram (Or total ion current or TIC for short) from a full scan acquisition to estimate percent composition in a mixture. And there are rules for adjusting response factors for compound classes. I recall seeing this in some class slides for a flavor chemistry class that were posted on line several years ago. I believe the class was by Dr. Gary Reineccius (a highly respected food and flavor chemist) from the University of Minnesota. And if I can find the slides, I’ll post the link.

Depending on what you are doing, this works quite well enough. If you are using the TIC trace from an MS to estimate that a fragrance mixture has 5% geraniol in it, you are probably good enough. If you use this technique to show that there is 5 ppb of DDT is a soil sample -- you won't have much credibility. (Actually you probably wouldn't find the peak in the TIC anyhow.)

Bottom line - the sophistication of the tool required depends on the task at hand.

There are hundreds of methods based on quantitation by MS, but the ones that gave good results all calibrate the area of the analyte compared to the area of the internal standard vs the quantity of the analyte.

As I understand it, what you are proposing is to compare the area of compound A with the area of internal standard B and then quantitate A as A/B times quantity of B. Then you are in questionable territory no matter what detector you use.

Peter

Isn't what are you saying same as what have I said? For example, if I have added 500 ng internal standard giving peak area of 15000 then quantity of sample giving peak area 30000 would be 1000 ng...? Do you mean this is not a valid method?

Some folks make an assumption that the FID trace is effectively a mass trace -- and it comes fairly close for most compounds of interest to someone like a flavorist. The FID essentially counts carbon atoms. And highly oxygenated compounds (like sugars) will never make it down the column. So, you can expect a major portion of the mass of the molecule making it to the detector to be carbon. (Remember, we are looking at the kinds of things in flavors, and aside from H2S and NH3, there is some carbon in just about every molecule that we would want to smell - and some that are of natural origin that we do not want to smell.)

Taking this a step further, the total response on the MS roughly follows that of the FID. So, some will use the Total Ion Chromatogram (Or total ion current or TIC for short) from a full scan acquisition to estimate percent composition in a mixture. And there are rules for adjusting response factors for compound classes. I recall seeing this in some class slides for a flavor chemistry class that were posted on line several years ago. I believe the class was by Dr. Gary Reineccius (a highly respected food and flavor chemist) from the University of Minnesota. And if I can find the slides, I’ll post the link.

Depending on what you are doing, this works quite well enough. If you are using the TIC trace from an MS to estimate that a fragrance mixture has 5% geraniol in it, you are probably good enough. If you use this technique to show that there is 5 ppb of DDT is a soil sample -- you won't have much credibility. (Actually you probably wouldn't find the peak in the TIC anyhow.)

Bottom line - the sophistication of the tool required depends on the task at hand.

Thanks for the information!
Please send me those slides if you get!

As Don rightly points out the validity (loosely) of a method depends on what you need the results for. If you calculate the quantity of A from the response for a known amount of B you have to assume that the response per unit mass is the same for both compounds, and that they both behave the same through the other steps of the analysis (extraction, injection, adsorption on the column etc). That assumption may or may not be close enough for what you need. Only if you calibrate the two responses against mass of A have you tested the assumption, and proved its validity.

I did an intensive literature review on the selectivity of the FID, and I was surprised to find that even with an homologous series of hydrocarbons there are percent level deviations from uniform response per mole of carbon, and that even moving a double bond or a methyl branch changed the response.

Peter

The MS is not a carbon counter like the FID. First, compounds don't all ionize the same way they burn, which is a good thing in an MS. Second, depending on how you tune versus your neighbor, you are not going to get the same ion distribution as your neighbor.

Best regards.

The thing about approximations is that, when applied properly, they work -- but if you forget they are approximations they will bite you!

And a side note on the FID as a carbon counter: An absolutely wonderful solvent for GD-FID is carbon disulfide. It does not show in the FID trace. Carbon going in disguise! (And yes it does show in the MS... I have to add that for those who are still trying to figure out why carbon disulfide does not show in the FID. )

Yeah you definitely need to be careful quantifying with MS to be sure that you don't let all the different tuning variables distort your results. See it all the time. Unfortunately I'm still a beginner at MS (don't even go for training till April) or I'd try and offer a little more help. I would err on the side of caution though.


Also, I would be really interested in those food chemistry slides--please post them if you can find them!

I've looked for the slides and have not been able to find them. I did note that Gary Reineccius has published a book on food chemistry, which would be a good bet - if you can squeeze a book into your budget.

On tune parameters and calibrations: (And speaking from the GC/MS side - LC/MS is a different animal.)

Tune for a consistant ion rato using your instrument's autotune. Always tune with the same column flow into the ion source and the same ion source temperature. The autotune should take care of the other parameters for you. For most analysis this is fine. There are some folks who will adjust a quadropole a bit manually as well. But, I lean away from this because the methodology becomes operator dependent (not robust across operators). Also, check for background air and water. These result in lowered sensitivity and I've seen water background result in a change in fragmentation in PFTBA. Not major, but I do not like inconsistancy in tune results.

Run check standards periodically and always check your curve after doing anything that can affect your instrument tune.

I hope that helps as you get started.

Thanks all for the comments and the info!
Don...I will keep in mind your tips...Thanx
Ram