Improving the Chemical Selectivity of an Electronic Nose to TNT, DNT and RDX Using Machine Learning.

We used a 16-channel e-nose demonstrator based on micro-capacitive sensors with functionalized surfaces to measure the response of 30 different sensors to the vapours from 11 different substances, including the explosives 1,3,5-trinitro-1,3,5-triazinane (RDX), 1-methyl-2,4-dinitrobenzene (DNT) and 2-methyl-1,3,5-trinitrobenzene (TNT). A classification model was developed using the Random Forest machine-learning algorithm and trained the models on a set of signals, where the concentration and flow of a selected single vapour were varied independently. It is demonstrated that our classification models are successful in recognizing the signal pattern of different sets of substances. An excellent accuracy of 96% was achieved for identifying the explosives from among the other substances. These experiments clearly demonstrate that the silane monolayers used in our sensors as receptor layers are particularly well suited to selecting and recognizing TNT and similar types of explosives from among other substances.

Chemical Selectivity and Sensitivity of a 16-Channel Electronic Nose for Trace Vapour Detection.

Good chemical selectivity of sensors for detecting vapour traces of targeted molecules is vital to reliable detection systems for explosives and other harmful materials. We present the design, construction and measurements of the electronic response of a 16 channel electronic nose based on 16 differential microcapacitors, which were surface-functionalized by different silanes. The e-nose detects less than 1 molecule of TNT out of 10+12 N2 molecules in a carrier gas in 1 s. Differently silanized sensors give different responses to different molecules. Electronic responses are presented for TNT, RDX, DNT, H2S, HCN, FeS, NH3, propane, methanol, acetone, ethanol, methane, toluene and water. We consider the number density of these molecules and find that silane surfaces show extreme affinity for attracting molecules of TNT, DNT and RDX. The probability to bind these molecules and form a surface-adsorbate is typically 10+7 times larger than the probability to bind water molecules, for example. We present a matrix of responses of differently functionalized microcapacitors and we propose that chemical selectivity of multichannel e-nose could be enhanced by using artificial intelligence deep learning methods.

Sensitivity comparison of vapor trace detection of explosives based on chemo-mechanical sensing with optical detection and capacitive sensing with electronic detection.

The article offers a comparison of the sensitivities for vapour trace detection of Trinitrotoluene (TNT) explosives of two different sensor systems: a chemo-mechanical sensor based on chemically modified Atomic Force Microscope (AFM) cantilevers based on Micro Electro Mechanical System (MEMS) technology with optical detection (CMO), and a miniature system based on capacitive detection of chemically functionalized planar capacitors with interdigitated electrodes with a comb-like structure with electronic detection (CE). In both cases (either CMO or CE), the sensor surfaces are chemically functionalized with a layer of APhS (trimethoxyphenylsilane) molecules, which give the strongest sensor response for TNT. The construction and calibration of a vapour generator is also presented. The measurements of the sensor response to TNT are performed under equal conditions for both systems, and the results show that CE system with ultrasensitive electronics is far superior to optical detection using MEMS. Using CMO system, we can detect 300 molecules of TNT in 10(+12) molecules of N2 carrier gas, whereas the CE system can detect three molecules of TNT in 10(+12) molecules of carrier N2.