Premium
Uncertainty contribution of derivatization in gas chromatography/mass spectrometric analysis
Author(s) -
Vilbaste Martin,
Tammekivi Eliise,
Leito Ivo
Publication year - 2020
Publication title -
rapid communications in mass spectrometry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.528
H-Index - 136
eISSN - 1097-0231
pISSN - 0951-4198
DOI - 10.1002/rcm.8704
Subject(s) - derivatization , bstfa , chemistry , gas chromatography–mass spectrometry , chromatography , gas chromatography , measurement uncertainty , trimethylsilyl , monte carlo method , mass spectrometry , uncertainty quantification , uncertainty analysis , statistics , organic chemistry , mathematics
Rationale The purpose of the current work is to realistically assess the uncertainty contribution in gas chromatography/mass spectrometry (GC/MS) analysis originating from less‐than‐ideal derivatization efficiency. Methods As the exemplary analytical method a two‐step derivatization method with KOH and BSTFA ( N , O ‐bis(trimethylsilyl)trifluoroacetamide), applied for the analysis of fatty acid triglycerides (using real measurement data), was selected. The derivatization efficiencies were in the range 0.89–1.04. In this study, two approaches for bottom‐up uncertainty evaluation were compared: the traditional GUM approach and the Monte Carlo method (MCM). Both were used with and without taking correlation between input quantities into account. Results The most reliable uncertainty estimates were in the range 0.07–0.08 (expanded uncertainties at 95% coverage probability). A strong negative correlation was found between the slope and intercept of the calibration graph ( r = −0.71) and it markedly influenced the uncertainty estimate of derivatization efficiency. The MCM was found to give somewhat higher uncertainty estimates, which are considered more realistic. Conclusions Derivatization directly affects the analysis result. Thus, in the case of this exemplary analysis, just derivatization alone (i.e. if all other uncertainty sources are neglected) causes relative expanded uncertainty around 8%, being thus an important and in some cases the dominant uncertainty contributor.