Chromatography Forum

Just a quick question how reliable in terms of specificity when we use MS library like the ones from NIST and Wiley to identify our compund?

Libraries can be reliable - but have to be used with caution. There are three conditions that can lead to improper identification with a library:
1) The compound you are looking for may not be in the library. The library search routing produces a mathematical search score and returns a result based on that - so it can return a match for a compound that is unrelated to your unknown.
2) Isomers and homologs often have similar or identical spectra. Remember that it is detection of the molecular ion and fragments generated from it that gives the spectrum - and sometimes the molecular ion does not survive long enough to give a visible signal and the parts giving differences between two different compounds may be lost in the neutral losses. Thus you need to confirm your identifications with retention time data or other chemical information.
3) The spectrum you are looking at may be contaminated by background from the chromatographic column or by coelution with another compound (or compounds). The software will give you the compound with the best match score. And, from the early days of computing: "Garbage in, garbage out."

Critically review library hits. Does the compound SUGGESTED make sense in terms of the sample and location in the chromatogram? Are the major ions in the library spectrum present in the unknown. Do m+1 and m+2 ratios appear to be similar between library and unknown (while ratios of fragments may change, relative abundances of isotopes do not - unless you are working with compounds like isotopically enriched compounds - in which case you usually know that the compounds are present.)

My car is reliable. I will shortly go down stairs and mumble “To work.â€

Don't know if the OP is still paying attention, but just for posterity I wanted to add few specifics to what Don said in my experience with GC/MS.

1. There are classes of compounds where library matches are unreliable. The worst offenders I know about are the simple alkanes. While a search generally will work well differentiating hexane and heptane, alkanes with carbon numbers in the teens and higher are likely to be wrong, even if the molecular ion registers above the noise. Other noteworthy examples include xylenes and PAH's.

2. High molecular weight compounds fare poorly in library searches in general, even if the molecular ion registers. This is in part due to response factor--the fact that high MW, high retention time compounds have much lower peak heights than lower MW compounds at the same concentration. Then there is the fact that there are more ways to arrange 20 carbon atoms than three carbon atoms.

Having said that, a library match for toluene, for example, isn't likely to be wrong. Personally, I don't trust library matches unless I have supporting information, and I am very uncomfortable when I'm handed a GC/MS report which claims to have identified a compound based on a library search. The library search function is a helpful aid, but it can be, and often is, tragically wrong.

Just my $0.02

Rataan,

To add my free advice (commentary) what about the tune? You take a Standard Spectrum Autotune on an Agilent and butcher it to get it to pass DFTPP. Which tune did the library use?

Remember what they say about free advice and getting what you paid for....

Best regards.