Comparative Discrimination Spectral Detection Method for the Identification of Vapors Using Overlapping Broad Spectral Filters

Abstract

We present a comparative discrimination spectral detection approach for the identification of chemical vapors using broad spectral filters. We applied the method to flowing vapors of as-received and non-interacting mixtures for the detection of the volatile components of a target chemical in the presence of interferents. The method is based on measurements of the overall spectral signature of the vapors, where the interferent spectrum largely overlaps the target spectrum. In this work we outline the construction of a set of abstract configuration-space vectors, generated by the broadband spectral components from sampled chemical vapors, and the subsequent vector-space operations between them, which enable the detection of a target chemical by comparative discrimination from interferents. The method was applied to the C-H vibrational band from 2500 to 3500 cm(-1), where there is large spectral signal overlap between the chosen target chemical and two interferents. Our results show clear detection and distinction of the target vapors without ambiguity.

Filter selection criteria for the discrimination of strongly overlapping chemical spectra

“…Though these results illustrate the interaction between the pairs of chemical discrimination vectors they do not directly provide information about sensing ability of these chemicals using various filter sets. To confirm the previous angle and standard deviation findings that support our initial filter selection, we conducted CDSD analysis [4] using an injection experiment of interferent (acetone, hexanes) onto our target chemical (fuel oil). For these experiments, ideal sensing would result in two distinct realms of detection.…”

“…We know from previous work that the ability to discriminate between chemicals is directly related to the separation of the vectors formed by the interaction between each individual chemical and the selected filter set. 3,4 Importantly, the separation (i.e. the angle between any 2 vectors and the subsequent volume formed by the separation of all 3 vectors) between the vectors is a key factor in whether discrimination between chemicals will occur.…”

“…To clearly demonstrate sensing with this 3-filter approach, we developed a model, comparative discrimination spectral detection (CDSD) which uses the relationship between the chemical discrimination vectors as a means to distinguish between closely matched chemical. 4 For any optical filter based sensing approach, the initial selection of filters to be used is of paramount importance. These methods discriminate between chemicals based on variances in the interaction between the chosen filters and the target chemicals specific absorbance bands.…”

Discrimination Between Explosive Materials and Isomers Using a Human Color Vision-Inspired Sensing Method

“…In order to determine whether the human color vision sensing approach would be capable of discriminating between the explosives RDX, PETN, and HMX, we first calculated discrimination results using methods similar to those described previously. 21,26,27 Briefly, the integrated product of the average spectral values obtained from the FT-IR spectrum for each sample and each of the three optical filter spectra were calculated for each surface concentration of RDX along with the HMX and PETN samples. The integrated products of the three filters were then normalized such that the vector formed by these coordinates would have a unit length of one.…”

“…Comparative discrimination spectral detection is a vector-based analysis technique used to classify optically filtered absorption data. Described in full detail elsewhere, 21,22 it operates by defining unique training vectors in an n -dimensional space (for these studies n = 3, as there are three optical filter channels) based on the combined optical response of provided training data with each optical filter. All vectors are normalized such that any measurement is bounded to a maximum value of unity.…”