PHOTOFRAGMENTATION IN FOURIER-TRANSFORM ION-CYCLOTRON RESONANCE - USE OR MISUSE

The noise in a sector or quadrupole mass spectrometer is determined by the source, i.e. the number of ions present in the signal, whereas in Fourier-transform ion cyclotron resonance it is detector limited. In addition, the Fourier transform spreads white noise equally over the whole spectrum. Thus the accuracy for broad-band spectra is generally rather low. However, the precision, i.e. the reproducibility of successive spectra, is astonishingly high. This fact has been used with much success, for example in collisionally activated dissociation in Hadamard spectra. It is shown that photofragmentation spectra are also very reproducible and thus allow-at least in principle-a 'kinetic deconvolution' of a series of such spectra. In such a case, nonlinear regression has to be applied to the data. Since in most kinetics problems several exponential functions are involved, and since it is well known that these exponential functions are notoriously difficult to approximate with a Marquardt-Levenberg algorithm, the validity of the results may be questionable. Three examples (protonated benzene, cycloheptatriene, deuterated toluene) are chosen to demonstrate some of the difficulties and pitfalls of the procedure.