Study of a Palladium (Pd)/Aluminum-Doped Zinc Oxide (AZO) Hydrogen Sensor and the Kalman Algorithm for Internet-of-Things (IoT) Application

• Published in: IEEE transactions on electron devices Vol. 67; no. 10; pp. 4405 - 4412
• Main Authors: Chen, Wei-Cheng, Niu, Jing-Shiuan, Liu, I-Ping, Chi, Cheng-Yu, Cheng, Shiou-Ying, Lin, Kun-Wei, Liu, Wen-Chau
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.10.2020; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Aluminum; Aluminum-doped zinc oxide (AZO); Catalytic activity; Chemical sensors; Computer simulation; Hydrogen; hydrogen sensor; Internet of Things; Internet of Things (IoT); Kalman algorithm; Kalman filters; Metals; Operating temperature; Palladium; Radio frequency; Response time; RF sputtering; Sensors; Sputtering; Thin films; Zinc oxide; Zinc oxides
• ISSN: 0018-9383; 1557-9646
• DOI: 10.1109/TED.2020.3018084

A palladium (Pd) thin film is decorated on a radio frequency (RF) sputtered aluminum-doped zinc oxide (AZO) thin film to produce a hydrogen sensor. Due to the catalytic activity of the Pd metal, the studied thin film-based device shows remarkably enhanced hydrogen-sensing characteristics. Experimentally, a very high sensing response of <inline-formula> <tex-math notation="LaTeX">1.12\times 10 ^{4} </tex-math></inline-formula> with a response time of 23 s is obtained under 1% H 2 /air gas at 300 °C. Furthermore, even under an extremely low concentration of 40-ppb H 2 /air, a sensing response of 0.17 is acquired. The optimal operating temperature of the studied device is 300 °C. A hypothesis is used to interpret the related hydrogen-sensing mechanism of the studied device. A thermodynamic analysis is employed to study the surface coverage of hydrogen molecules on the device's surface. Furthermore, for the application in wireless transmission of the Internet of Things (IoT), an interesting Kalman algorithm is used to reduce redundant data, save hardware costs, and reduce network congestion. The simulated results show that 93.9% of the redundant data can be removed. The studied device exhibits advantages of a simple structure, easy fabrication, low cost, a widespread sensing range of hydrogen concentration, a very high sensing response, and an extremely low detecting limit, as well as being suitable for IoT application.