Dynamic background noise removal from overlapping GC-MS peaks via an entropy minimization algorithm

Abstract

A major challenge for accurate GC-MS analysis is to achieve excellent chromatographic separation and a high signal-to-noise ratio. In fact, real-world samples are often contaminated with impurities that manifest as intense background noise in GC-MS data. Herein, a dynamic background noise removal system is developed based on an entropy minimization algorithm to extract pure mass spectra of straight-chain n-alkane components in a sample of hydrogenated jet fuel. This system is capable of extracting pure component spectra from overlapping peaks that are heavily superimposed by background noise without a priori information. The performance of the system surpasses that of the background subtraction method applied in Agilent ChemStation software. The robustness of the system to accurately analyse real-world samples is expected to simplify and revolutionise the application of GC-MS for chemical analysis.