ASCII.jp The University of Tokyo simultaneously detects trace gas components contained in exhaled breath by machine learning

A research team at the University of Tokyo has developed a sensor that uses an ion gel and multiple electrodes, and succeeded in simultaneously detecting hydrogen, ammonia, and ethanol in ppm units present in a mixed gas such as exhaled breath. Conventionally, it has been difficult to detect a plurality of trace gas components in a mixed gas at the same time.

A research team at the University of Tokyo has developed a sensor that uses an ion gel and multiple electrodes, and succeeded in simultaneously detecting hydrogen, ammonia, and ethanol in ppm units present in a mixed gas such as exhaled breath. Conventionally, it has been difficult to detect a plurality of trace gas components in a mixed gas at the same time. The research team focused on the diversity of the interface between the ion gel and the electrode. We have developed a sensor with a structure in which four types of electrodes, gold, chromium, platinum, and rhodium, are in contact with the ion gel. Of the four electrodes, select and use two electrodes with an electrical switch. A voltage is applied to one electrode and the other is connected to the gate of the MOS transistor. When a gas is blown, the ion gel occludes it and reacts on the electrodes to change the potential. The time dependence of this sensor response was input, and the hydrogen, ammonia, and ethanol concentrations in the sprayed mixed gas were used as outputs to train the neural network, and the respective gas concentrations in the mixed gas were estimated. As a result, we succeeded in identifying the concentration of hydrogen, ammonia, and ethanol in ppm units contained in human exhaled breath in the mixed gas. If the sensor is miniaturized and mounted on a mobile terminal or the like, it will be possible to easily judge the health condition by reading the change in the trace gas contained in the exhaled breath. The research results were posted online on March 17th in "ACS Sensors".

(Sasada)

[Read this article at MIT Technology Review]

Display format: PC ⁄ smartphone