[curiouscat]

How good / reliable are these automatic libraries that come with GC-MS's these days? I'm not talking about very complex or exotic molecules, liquid hydrocarbons and oxygenates in the C6-C14 range is what most of my targets will be.

Does one always need a human interpretation or have the algorithms become good enough to run on their own now? Reason I ask is that we don't have an in-house GCMS at the moment but do have access to an instrument at another related site. But they don't have a chemist trained / willing to interpret a GC-MS.

As an aside, how do the automatic algorithms work? Is it just based on a m/z match or something more sophisticated.

The application goal is to get a more thorough understanding of what by-products we are forming in our process to then enable some sort of by-product reduction strategy.

[Archer]

I find the algorithms are OK, it is my understanding that they look at m/z and relative intensity. There is no substitution for human input (I.e. checking the retention time of a reference martial.

I look at such a range of substances, some times I can do a better job than the library, other times I wouldn't have a hope of predicting the EI fragmentation.

[WW]

It depends.  If you are working with a very clean sample and have low baseline, you are probably ok.  However as you do as I do and inject extractions from soils/sediments/plants, they are often unreliable.  There are a lot of "ghost" m/z peaks in a lot of the spectra, and this often causes the software to misidentify the compound.

Never blindly trust the software, use it as a guide.  it is no substitute for comparing the R.T. and spectra of a standard