Rationale: Ion mobility spectrometry (IMS) is a powerful analytical tool that has been widely applied in many fields. However, the limited structural resolution of IMS results in peak overlapping in the analysis of samples with similar structures. We propose a novel method, improved particle swarm optimization (IPSO), for the separation of IMS overlapping peaks.
Methods: This method, which prevents local optimization, is used to identify the peak model coefficients of IMS. Moreover, we use the half-peak width characteristics of IMS to determine the particle position range, which eliminates impossible combinations of single peaks and reduces the difficulty of identification of coefficients.
Results: During a comparison in performance between IPSO and the genetic algorithm (GA), the results show that the maximum separation error of IPSO is only 1.45%, while the error of the GA is up to 17.43%. Moreover, the time consumed by IPSO is 95% less than that of the GA, and IPSO has a greater robustness under the same separation error conditions.
Conclusions: The method proposed provides accurate analytical results in separating overlapping IMS peaks even in cases of severe overlaps, which greatly enhances the structural resolution of IMS.
© 2020 John Wiley & Sons, Ltd.