Recognition of odors effusing from living leaves using a SAW sensor system

抄録

An odor sensing system based on a SAW device has been developed for the measurement of odors from herb, which is a common name of the grass effusing relatively strong odors, and recognition of eight kinds of herb odors was performed using the system. Four kinds of lipid films were used as odor sensitive film in the study. The lipids of the films are (1) lecithin from eggs, (2) sphingomyelin, (3) L-α-phosphatidyl ethanolamine solution 70%+phosphonicacid dioleyl ester 30% and (4) cholesterol 70%+sphingomyelin 30%. The lipids were applied to the SAW resonator surface by spin coating, at speed 3000 r. p. m. for 20 seconds, from a lipid solution of 20μl. A neural network algorithm was applied to identify the odors. In the recognition process, patterns consisting of maximum frequency changes of the sensors and differential frequency changes of the sensors between adsorption and desorption of the odors were taken as inputs of the network, and the patterns were normalized by a specific normalization method, so that the recognition of the odors was improved and done at rate of 99%.