抄録
The precision or relative standard deviation (RSD) of measurements in a sector type mass spectrometer equipped with an electron ionization system is examined experimentally and theoretically. The observed RSD is obtained by the usual replication and the theoretical RSD is predicted from the probabilistic properties of the baseline noise by the uncertainty theory called the FUMI theory (Function of Mutual Information). By comparing the experimental and theoretical RSD, this paper identifies two kinds of error sources: (1) for small peaks, the major cause of the measurement uncertainty is the baseline noise originating mainly from the detection unit; (2) for large peaks, the measurement error is subject to ion/particles processes such as the fluctuation of gaseous sample concentration, electron ionization efficiency, and ion beam flux and there is a lower limit of the measurement RSD (≈3%). In case 1, the measurement RSD can successfully be predicted by the theory. In case 2, the observed RSD is even larger than the theoretical RSD, and the repeated measurements are necessary to estimate the measurement uncertainty. The peak corresponding to the limit of detection (RSD=33%) is demonstrated.
