Library search mismatches

[borislaw]

Dear All,

I am new to mass spectrometry, and I have a question:

Why ion abundances in mass spectra obtained on different instruments (with the same conditions) are not always matching library search? As far as I understand, once I have calibrated my system (i.e. have proper ion ratios for "standard") I should obtain relative abundances very similar to those in library, otherwise it's useless, no? Does it matter what kind of instrument was used to build the library (quad, ion trap)?

[AICMM]

borislaw,

Which "standard"? PFTBA, DFTPP, BFB? Which MS; quad, ion trap, sector? Which tune (at least on Agilent); STUNE, Atune, Max Sens tune? I don't think it is as standardized as you might be expecting. I suspect that many of the Wiley spectra are a result of environmental analysis under BFB or DFTPP but I don't know that that is the same of the NIST data base.

Sorry.

Best regards.

[Don_Hilton]

Because the ion sources of various instruments are not identical, identical conditions are not identical! While the simple explanation of MS is an electron of some fixed energy ionizing a molecule and causing it to fragment, there are a number of other factors. Even the filament temperature in the mass spec has an effect. And given that filaments age, you can not get identical conditions twice in the same mass spectrometer - if you want to get really fussy about it. (And, this is why one solution to an instrument that will not pass tune is to switch to the other filament.)

When we tune a mass spectrometer, we make conditions such that the fragmentation will be the same for the particular molecule used to test the tune (like PFTBA). And for the work that most of us do, this works quite well enough. When I run the same compound in the same mass spectrometer tuned to fragment the tuning compound the same way - I get reproducible results. And, the results for that compound look quite similar to those in mass spectral libraries.

Spectra in libraries are from various instruments with ion sources run at various temperatures, gas flow rates (or without gas flow for insertion probe work), differing filament temperatures, differing geometry and such. Also, that 70 volt electron energy: It depends on how the electron is measured. But it all works remarkably well, actually.

One of the really important things is that the fragmentation pathways remain the same. So, major fragments will always be present in electron impact spectra - even if ratios change slightly. (And yes, the ion trap does look different from the others. But off in that land of traps there remains consistancy.)

[borislaw]

Thanks for explanation. Just to clarify: how "slight" those ratios might be/allowable? Should one examine only "fingerprint" match, but not ion ratios? The instrument in use is ion-trap, calibration gas is PFTBA, analysis performed in EI positive mode and results are compared to Wiley library.

[Don_Hilton]

The trap has a reputation for a couple of things: 1) sensitivity under specific types of conditions and 2) spectra that look different from other types of EI instruments.

I've known one fellow who used ion trap, but would take his samples over to another instrument (it was either a quad or TOF - I don't remember) for obtaining spectra for identification.

I'm mostly a food, flavor and fragrance guy, so I can't offer any more because most all of my friends stay away from traps -- because of the difference in spectra. The folks who I've known who have used traps have used them for analysis of target analytes in environmental work.

[AICMM]

Borislaw,

If I am not mistaken (and it has been a while) the EPA gives you a 20% window around the expected ion ratio relative to the parent ion. So, for example, if 98 is within +/-20% of its expected ratio to 177, then it is considered a confirming factor in identifying the component. So, in short, not a trememdously tight ion window for identification.

Best regards.

[Ron]

In many cases ion traps give different fragmentation patterns (ion ratios) than quadrupoles, and external ionization traps give different fragmentation patterns than internal ionization traps. Most library mass spectra are collected using quadrupole instruments, so in many cases the library spectra will look different than the ones you collect.

[tangaloomaflyer]

You may also need to subtract any background, although if you are using an ion-trap i would expect this is not too much of a problem