Non-target analysis of vapor mixtures using silicon nanowire array sampling and thermal desorption

This work presents and evaluates an algorithmic approach to deconvolving the elution profiles of chemical components of vapor mixtures that have been sampled and desorbed from a novel preconcentrator based on highly ordered silicon nanowire arrays. The arrays provide a medium for both preconcentration and partial chromatographic resolution, which is then further leveraged with multichannel detection. Here, mixtures of nitro aromatic vapors are sampled and then thermally desorbed from the device, at which point they are detected by a conventional mass selective detector. The overlapping elution profiles observed from the array are sequentially extracted using a chemometric analysis approach based on evolving factor analysis and multivariate curve resolution by alternating least squares, enabling qualitative and quantitative analysis of individual components without target analyte libraries or complete chromatographic separation. This work examines the analytical capabilities conferred to multichannel detection by silicon nanowire array pre-concentration and partial separation and discusses the technique's limitations, illustrated by both experimental and simulated data.