| 摘要 |
To determine the elemental composition of a sample by mass spectrometry, an experimental spectrum consisting of data counts observed at different values of mass-to-charge ratio is first established. The method then comprises the steps of a. generating a list of possible constituents of the sample, each said constituent having a known isotope ratio pattern; b. assigning a prior probability distribution to the said list of possible constituents; c. defining a set of trial compositions corresponding to the results of steps a and b; d. using the isotope ratio patterns together with the responsivity of the spectrometer to define a trial mass spectrum corresponding to each said trial composition; e. defining a posterior probability distribution by using probability calculus to compare the trial mass spectra and the experimental mass spectrum; f. generating a number of trial sets of possible compositions of the sample with their associated likelihood, e.g. In Gibbs sampling, which trial sets are consistent with the said posterior probability distribution; and g. using the trial sets and associated likelihoods to obtain mean values for the amount of each said constituent of the sample, together with corresponding associated uncertainties. |
