Optimized design of carbon nanotube field-effect transistor using Taguchi method for enhanced current ratio performance

• Main Authors: Abdul Hadi, M.F., Hussin, H., Muhamad, M., Wahab, Yasmin Abdul
• Format: Article
• Published: IOP Publishing 2024
• Subjects: QC Physics; TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.um.edu.my/45158/; https://doi.org/10.1088/1402-4896/ad4c1d
• Institution: Universiti Malaya

Presently, the integrated circuit (IC) industry grapples with obstacles in downsizing MOSFET technology further, hindered by its inherent physical constraints. Therefore, the substitution of silicon with carbon nanotubes (CNTs) holds promise for paving a novel path in semiconductor industries, driven by their diminutive dimensions and superior electrical properties. Hence, this project employed SILVACO ATLAS software in conjunction with the Taguchi method to refine a CNTFET design for optimal performance. In this work, response variables consists of on-current (Ion), current ratio (Ion/Ioff) and threshold voltage (Vth) are extracted. In this particular design, the Taguchi method was employed to ascertain the most effective combination of design parameters and materials to achieve optimal CNTFET performance, as assessed by the three key response variables. The design parameter and material that had been chosen were the diameter of carbon nanotube (Dcnt), dielectric material (K) and oxide thickness (tox). Each of the design parameters and material had three different values. For K, the values are 3.9 (SiO2), 25 (ZrO2) and 80 (TiO2). While for Dcnt and tox, the values are 4.0 nm, 6.0 nm, 8.0 nm and 2.0 nm, 4.0 nm, 6.0 nm respectively. According to the Taguchi optimization findings, the ideal combination of parameters comprises a CNT diameter of 4.0 nm, an oxide thickness of 2.0 nm, and the use of TiO2 (80) as the dielectric material. The ANOVA analysis underscores the significance of prioritizing optimization efforts towards the CNT diameter parameter. This is attributed to its substantial contribution, accounting for 93.55% of the variation in the Ion/Ioff value, surpassing the influence of dielectric materials and oxide thickness.