The 2D nature of graphene, its single-molecule sensitivity, low noise and high carrier concentration have generated significant interest in its application in gas sensors. However, due to its inherent non-selectivity and massive p-doping in atmospheric air, its applications in gas detection are often limited to controlled environments such as nitrogen, dry air, or synthetic moist air.
While humidity conditions in synthetic air can be used to achieve controlled hole doping of the graphene channel, this does not adequately reflect the situation in atmospheric air. In addition, atmospheric air contains several gases with concentrations similar to or greater than that of the analytical gas. Such shortcomings of graphene-based sensors impede selective gas detection and identification of molecular species in atmospheric air, which is necessary for applications in environmental monitoring and non-invasive medical diagnosis of disease.
The research team led by Dr. Manoharan Muruganathan (formerly Associate Professor) and Professor Hiroshi Mizuta from Japan Advanced Institute of Science and Technology (JAIST) used the machine learning (ML) models trained on various gas adsorption-induced doping and scattering signals. to achieve both highly sensitive and selective gas detection with a single device.
The performance of the ML models often depends on the input functions. “The conventional graphene-based ML models are limited in their input functions,” says Dr Osazuwa Gabriel Agbonlahor (formerly a postdoctoral researcher). The existing ML models only monitor the changes in the graphene transfer characteristics or resistance/conductivity caused by gas adsorption without modulating these characteristics by applying an external electric field.
Therefore, they lack the distinctive van der Waals (vdW) interaction between gas molecules and graphene, which is unique to individual gas molecules. Therefore, unlike the conventional electronic nose (e-nose) models, we can map the external electric field modulated graphene-gas interaction, which allows for more selective feature extraction for complex gas environments such as atmospheric air.
Our ML models for the identification of atmospheric gases are developed using the graphene sensor functionalized with a porous activated carbon thin film. Eight vdW complex features were used to monitor the effects of the external electric field on the graphene-gas molecule vdW interaction, and consequently mapped the evolution of vdW binding before, during and after the application of the external electric field.
In addition, although the gas detection experiments were performed under different experimental conditions, e.g. gas chamber pressures, gas concentrations, ambient temperature, relative humidity, tuning time and tuning voltage, the models developed proved robust enough to handle these variations in experimental conditions by not exposing the models to these parameters. .
In addition, to test the versatility of the models, they were trained in both atmospheric environments and relatively inert environments commonly used in gas detection, such as nitrogen and dry air. Therefore, a powerful “electronic nose” of atmospheric gas was achieved, which distinguished with 100% accuracy between the four different environments (ammonia in atmospheric air, acetone in atmospheric air, acetone in nitrogen and ammonia in dry air).
The research will be published in the journal Sensors and Actuators B: Chemical.
Osazuwa G. Agbonlahor et al, Machine learning identification of atmospheric gases by mapping the graphene molecule van der Waals complex bond evolution, Sensors and Actuators B: Chemical (2023). DOI: 10.1016/j.snb.2023.133383
Offered by Japan Advanced Institute of Science and Technology
Quote: New machine learning approach selectively identifies one molecule in a billion, with graphene sensors (2023, March 17) Retrieved March 18, 2023 from https://phys.org/news/2023-03-machine-learning-approach-molecule-billion -graphene.html
This document is copyrighted. Other than fair dealing for private study or research, nothing may be reproduced without written permission. The content is provided for informational purposes only.