Low-cost desorption unit coupled with a gold nanoparticles gas sensors array for the analysis of volatile organic compounds emitted from the exhaled breath (gastric cancer and control samples)

• Published in: Microelectronic engineering 2021-01, Vol.237, p.111483, Article 111483
• Main Authors: Durán-Acevedo, Cristhian Manuel, Cáceres-Tarazona, Juan Martín
• Format: Article
• Language: English
• Subjects: Cancer; Classification; Data processing; Desorption; Desorption unit; Discriminant analysis; Equipment costs; Exhaled breath; Function analysis; Gas sensors; Gold; Low cost; Nanoparticles; Nanoparticles gas sensors; Neural networks; Nitrogen; PARC; Principal components analysis; Selectivity; Sensor arrays; Tenax tubes; Tubes; VOCs; Volatile compounds; Volatile organic compounds
• ISSN: 0167-9317; 1873-5568
• DOI: 10.1016/j.mee.2020.111483

This paper describes the development and validation of a low-cost desorption unit (DU) equipment for the desorption of pre-concentrated volatile compounds in adsorption tubes ORBO™-420 Tenax. The system initially heats the tubes to a temperature of 250C degrees, releasing the trapped VOCs and then drags them by compressed nitrogen flow to a gold nanoparticles gas sensor array. In each of the experimental tests, the equipment was able to ensure the same desorption conditions in the set of measurements carried out, preventing the sample from being contaminated by external agents or from being diluted during extraction. On the other hand, good repetitiveness and selectivity were observed in the sensors responses. To validate the functioning of the equipment, a set of previously acquired samples of exhaled breath from patients confirmed with gastric cancer and control patients (i.e, with gastritis and ulcer) were analyzed. The results were compared with a previous study and it was shown that the desorption unit was able to extract the volatile compounds emitted from exhaled breath and improving the classification. The samples were analyzed through data processing techniques such as: Principal Component Analysis (PCA), Discriminant Function Analysis (DFA), Probabilistic Neural Network (PNN) and Supporting Vector Machines (SVM), for the analysis and the classification of the data obtained, where 98.31% of the total variance was reached through the PCA analysis and 100% success rate was obtained in the classification by using SVM. [Display omitted] •Low cost desorption unit to be used with instruments such as a multisensory system.•Development of new technology for exhaled breath analysis and other applications.•Portable equipment for sampling systems (e.g, Tenax Tubes).