Identifying volatile organic compounds by determining their diffusion and surface adsorption parameters in microfluidic channels

Abstract

In artificial olfaction systems, target gas classification generally takes place in high-dimensional feature spaces with physicochemically indefinable principal axes. This makes practically impossible to determine the position of a new target gas in the feature space before performing cumbersome training tests. Here, for the first time, we demonstrate the feasibility and advantages of forming a feature space with physicochemically meaningful principal axes in conjunction with an olfaction system. The diffusion progress rate of a single component gas in an air-filled microfluidic channel depends on a number of its physicochemical parameters, the combination of which is the unique property of the gas and can be utilized for its recognition. These parameters are determined via fitting of the model-based predictions to the experimental diffusion rates in two different cross-section capillaries. The system utilized comprises two cylindrical channels with respective diameters of 1000 and 50μm. In each test, the open ends of both channels are suddenly exposed to the analyte contaminated atmosphere. The temporal profiles of the diffusion rates are recorded by continuous resistance measurements on the chemoresistive sensors spliced to the channels’ opposite ends. Fitting diffusion equation solutions to the experimental profiles related to the 1mm channel results in the analyte diffusivity, in air. Similar results from the micro-channel, however, fit the solutions of a modified diffusion equation which accounts for the channel wall adsorption of the analyte molecules and results in another parameter which is related to the analyte–channel wall interactions. Relating these parameters to the analyte affords its recognition in a physically meaningful feature space.