Wavelet-enhanced TD–GC–MS analysis of molecular pattern alterations in gas samples induced by breath sampling devices

This study presents a computational method to identify volatile organic compound (VOC) artifacts introduced by breath sampling hardware. To exclude endogenous biological variability, ambient air was collected using two sampling devices working in the same experimental conditions: the Mistral end-tidal breath sampler and the ACTI-VOC PLUS pump, a low-emission reference system. VOCs were pre-concentrated on sorbent-packed thermal desorption (TD) tubes and analyzed by TD-gas chromatography-mass spectrometry (TD-GC-MS). Differential chromatograms obtained by subtracting ACTI-VOC signals from Mistral traces were processed using stationary wavelet transform (SWT) to selectively enhance high-frequency features indicative of artifactual emissions. Four new compounds not previously associated with Mistral sampling hardware were consistently detected in Mistral samples and were absent in ACTI-VOC pump controls: 1,3,5-trioxane, 1,3,5,7-tetroxane, (Acetyloxy)acetic acid, and N,N-dimethylformamide. These molecules are indicative of polymer degradation, acetal resin breakdown, and material off-gassing specific to the breath sampler.