Sciencemadness Discussion Board

GC/MS vs. NMR for unknown sample anaylsis

crazyboy - 26-12-2010 at 16:29

I have several compounds whose identity I would like to confirm unequivocally. I would also like to have at least a rough idea of the relative abundance and identity of any impurities present. I have strong suspicions as to the identity of these compounds so I would know where to look in the spectral library.

What analytical tool should I use GC/MS or proton NMR? I am more comfortable anaylzing data from the MS and the GC gives a readout of the realtive abundance of all the products present. However the GC/MS takes about 40 minutes per sample and I would like to test one or two samples.

Proton NMR takes a fraction of the time (about 3 minutes) but I am completely unable to interpret the readout.

What do you guys think?

TheOrbit - 26-12-2010 at 17:20

in most cases we use combination of both, as MS rarely give M+ and fragmentation is not always that easy and NMR is very useful in multi-protons containing cpds

aonomus - 26-12-2010 at 17:29

A few questions:

- Where do these samples come from, reactions, complete unknowns?
- How pure are these compounds?
- Are they reaction mixtures in solvent, or isolated solids/liquids?
- What is the sample matrix made up of?
- How much sample is available?

To generalize, the preferred method for unknown analysis would probably be something along the lines of LC/MS or GC/MS, though depending on the sample matrix it might require something like SPE/SPME to clean it up, or at least something to get rid of inorganics that might can stuck on a GC column (whereas inorganics on an LC system would probably just get eluted out at the solvent front).

If you are limited *only* to GC/MS and 1H-NMR, I would side with using GC/MS and try to base your analysis of the peak masses off of any known information such as related structures, etc. If you can use both positive and negative ionization modes (if your MS supports it).

1H-NMR can only do so much, it can help to confirm a structure in mind or confirm the addition or change of substrate, but identifying something unknown completely from scratch would be pretty tough unless it was very very simple and ideal (like textbook ideal) with regards to any upfield/downfield shifts of peaks (ie: none).

If you can, run the sample on an HPLC just to see the predominance of polar vs nonpolar compounds. Perhaps if you can collect the appropriate fractions and separate by HPLC, NMR might be more useful at that point.

DDTea - 26-12-2010 at 17:42

I'm going to jump on the bandwagon in favor of GC/MS in this case. The big advantage of MS in structure analysis is that it can tell you the molar mass of your unknown. Combined with sample introduction via GC, it tells you the mass of all the compounds present, provided that your separation is good.

Here's the way I'd do this analysis: GC/MS as a starting to point to, first of all, see how many compounds are present in your samples and to get a rough idea of what they are (especially if you're relying on a spectral library). That should also give you your mass peak and be good starting point. Next, if it's available to you, HPLC to separate the sample into its components. After that, run FT-IR to determine which functional groups are present (this takes even less time than NMR). Following that, use 1-H NMR as a confirmation.

Running NMR on an impure sample will probably lead to problems, especially if there are hydrogen bonds or proton transfers occurring--your peaks might broaden to the point that they're unusable, especially if you're not comfortable interpreting spectra.

The goal in spectroscopy is always to "overdetermine" the structure. Hunches are good, but they're only a starting point. It's always better to be thorough. When I took a course in organic spectroscopy, this is the method we used (combined with, at times with chiroptical methods to determine the stereochemistry).

crazyboy - 26-12-2010 at 18:35

Quote: Originally posted by aonomus  
A few questions:

- Where do these samples come from, reactions, complete unknowns?
- How pure are these compounds?
- Are they reaction mixtures in solvent, or isolated solids/liquids?
- What is the sample matrix made up of?
- How much sample is available?


They are products of reactions and compounds which are labeled but whose identity is difficult to verify.

They seem to be quite pure but that is half of what I would like to find out.

They are all isolated compounds both solids and liquids.

No matrix they are isolated.

Most of the samples I have 5-10g of one I have ~100mg.

Well it looks like I'll be using GC/MS that is what I prefer to do however I will need to be careful not to interfere with the other researchers. Maybe I'll run the samples during lunch or early in the morning...