Passivation-Dependent Device-Level Thermal Robustness of Indium Gallium Oxide Thin-Film Transistors under Dynamic...

As conventional silicon-based dynamic random-access memory (DRAM) approaches its physical and electrical scaling limits, oxide semiconductors have emerged as promising channel materials for next-generation memory devices because of their low off-state current and process compatibility. Among them, indium gallium oxide (IGO) thin-film transistors (TFTs) have attracted considerable attention for DRAM-related applications; however, their implementation remains challenging because the device stack must withstand high-thermal-budget back-end-of-line processing, including annealing at 600 °C in a N2 ambient which can promote oxygen-deficient conditions in oxide layers. Here, we systematically investigate the effects of passivation-layer selection on the electrical performance and thermal reliability of IGO TFTs by comparing Al2O3-, HfO2-, and SiO2-passivated devices before and after 600 °C N2 annealing. Among the passivation layers examined, Al2O3 uniquely preserved stable transistor operation after the high-temperature process, whereas the HfO2- and SiO2-passivated devices exhibited severe degradation of switching behavior and became electrically unextractable after 10 min annealing. In contrast, the Al2O3-passivated TFTs retained excellent electrical characteristics, including a field-effect mobility of 94.08 cm2V-1s-1 and a subthreshold swing of 84 mVdec-1 after 10 min annealing. Moreover, the positive-bias temperature stress stability was markedly improved, with the threshold-voltage shift at 100 °C decreasing from 63 to 20 mV after extended annealing. The superior reliability of Al2O3 layer may be attributed to its strong Al-O bonding and high film density, which provide enhanced chemical and structural stability during high-temperature N2 annealing. These findings suggest that Al2O3 passivation is an effective strategy for enabling thermally robust and electrically reliable IGO TFTs for advanced DRAM applications.