Active Electrode Dependence of HfO_<i>x</i>-Based Flexible RRAM With a Stack Engineered Thermal Enhancement Layer

The statistical analysis and resistive switching (RS) nature of the HfOx-based flexible resistive RAM (RRAM) with a chalcogenide stack layer and germanium selenium telluride (GST) with different top electrodes (Al, Ag, and In) have been investigated. The Ag and Al electrodes exhibit a self-rectifying gradual interfacial switching, while indium shows strong filamentary nature. The GST, known for its outstanding thermal confinement, drastically enhances switching performance of the RRAM by excellent thermal confinement and oxygen modulation. In Ag-based devices, the analog switching of 1.5-<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu $ </tex-math></inline-formula>W power is achieved. The current conduction mechanism shows Fowler-Nordheim (FN) tunneling with 0.0324-eV barrier height in Ag devices resulted in a high current increase in low set voltage (~0.3 V). The device also exhibits high temperature (up to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$85~^{\circ }$ </tex-math></inline-formula>C) stability, prolonged data retention capabilities (up to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$10^{{4}}$ </tex-math></inline-formula> s), and 3000 cycles dc endurance. It also sustains a stable memory window under mechanical flexing (up to 5 mm). The device is also well-adapted for neuromorphic computing applications by demonstrating exceptional synaptic learning characteristics, such as paired-pulse facilitation (PPF), long-term potentiation (LTP), paired-pulse depression (PPD), and long-term depression (LTD).