SEMI - EMPIRICAL DFT - BASED ATOMIC SCALE MODELING OF GATE-STACK ARMCHAIR GRAPHENE NANORIBBON FET FOR PERFORMANCE ENHANCEMENT AND SENSOR APPLICATION

Abstract

The continuous downscaling of conventional MOSFETs has reached a point where short channel effects, leakage, and static power dissipation increasingly limit further miniaturization, and motivate researchers to find new material for channel and device architectures that can overcome these limitations. Among the emerging materials, armchair graphene nanoribbons (AGNRs) offer tunable bandgaps, atomically smooth edges, and quasi-ballistic transport, making them well suited for replacing the silicon channel in next generation field-effect transistors. The present thesis develops an atomistic-scale investigation of a gate-stack armchair graphene nanoribbon field-effect transistor (GS AGNRFET) using Quantum ATK simulation tool and systematically examines dielectric engineering, channel-width engineering, computational-model reliability, and edge passivation chemistry for the dual for ultra-low-power logic and gas sensing applications. The present study begins with a density functional theory-based analysis of the GS AGNRFET, in which the gate-stack architecture reduces the off-state current by nearly 73% and improve the switching ratio (SR) by a factor of approximately 213 over the conventional GNRFET. A temperature-dependent analysis from 250 K to 350 K reveals reduction by 48% in on-state current, degradation of SR by 98%, increase in subthreshold swing by 56%, which confirms that the proposed device is suitable for low-temperature regimes such as cryogenic electronics and biomedical instrumentation. Further, edge passivation of AGNR with boron (13th group element) reduces the bandgap and raise the density of states near the Fermi level and reduces the off-state current by ~7.8 × 105 times, enhances the on-state viii current by ~139 times, and improving the SR from ~7.1 × 103 to ~1.45 × 108 relative to the pristine AGNRFET. After this, channel-width engineering is subsequently examined for AGNRs belonging to the 3p + 1 family (N = 4, 7, 10, 13) using a Slater–Koster DFT framework. The study shows that the GS – AGNR (N = 4) FET exhibits the largest bandgap of approximately 1.98 eV, suppressed density of states near the Fermi level, and superior gate control. Also, GS – AGNRFET shows improved transfer chatterati’s over all other configurations, which is further confirmed through transmission spectra analysis, projected density of states analysis and electrostatic potential profiles. Further, to analyze the reliability of the findings across simulation methodologies, the Extended Hückel (EH) model with self-consistent iteration within the NEGF formalism is then employed. The comparison between the AGNR (N = 4) and AGNR (N = 7) as a channel material in device configurations using EH model shows that widening the channel from N = 4 to N = 7 increases the on-state current by 481 times, lowers the off-state current by 99.92%, reduces the threshold voltage by 27.59%, and improves DIBL by 16%. After that, the study of edge passivation using the remaining Icosagens elements ( group 13th elements namely Al, Ga, and In ) indicates that passivation of heavier Icosagens atoms on AGNR (N = 7) shows metallic bandgap and therefore unsuitable for semiconducting use, whereas the same elements substantially enhance the performance of the AGNR (N = 4) channel. The aluminium passivated GS – AGNRFET (Al-AGNRFET) shows improved results in terms of on-state current (rise from 1.06 µA to 18.4 µA), reduction of the off-state current, and improvement in SR. Also, transconductance, device efficiency, transmission probability, and transfer characteristics analysis highlights Al – AGNRFET and Ga – passivated GS – AGNRFET are suitable candidates for high–speed logic, low–power memory devices, quantum computing in space exploration, and neuromorphic computing applications. Lasty, the proposed Al – AGNRFET is analysed as a nanoscale gas sensor for the detection of toxic gases like NO2 and NH3 using a two-step methodology that combines Linear Cobination of Atomic Orbitals calculator for bulk configurations analysis and EH model for simulation at the device level. The pristine AGNR exhibits chemisorption with physically unrealistic recovery times, while the Al-passivated ribbon shows stable physisorption with recovery times of 5.4 × 10−4 s for NO2 and 1.1 × 10−10 s for NH3. The sensitivity values of 736% for NO₂ and 116% for NH₃ in Al – AGNR, which are 3.36 and 6.60 times higher than the P AGNR, respectively. At device level, the opposite threshold voltage shifts of −0.10 V and ix +0.15 V under NH3 and NO2 exposure, along with an on-current-based selectivity coefficient of 12.12, confirm that the proposed Al - AGNRFET device is suitbale to detect and distinct between the two gases. Overall, the proposed GS-AGNRFET device is optimised through gate-stack architecture, channel - width tuning, computational model reliability and Icosagens edge passivation, and shows high switching ratio, reduced off-state current, reliable performance at low temperature and selectively detect NO2 and NH3 gases. Owing to these improvements, the proposed device is a suitable candidate for use in low-power logic, toxic gas sensing, bio- medical applications and energy- efficient nanoelectronics areas.