Hayato Nakamura1,Yumeng Zheng1,Kentaro Kinoshita1
Tokyo University of Science1
Hayato Nakamura1,Yumeng Zheng1,Kentaro Kinoshita1
Tokyo University of Science1
Resistive switching (RS) devices of Pt/Nb-doped SrTiO<sub>3</sub> (STO) Schottky junctions have been the subject of considerable interest in terms of both non-volatile and volatile resistance relaxation nature. When used as general non-volatile memory, the continuously tunable write/erase characteristics are advantageous for multibit application. In this sense, resistance state retention characteristics are important and improving resistance volatility is crucial. On the other hand, when used in new computational paradigm for the breakthrough of von Neumann-type computing, controllability of resistance relaxation phenomenon (RRP) is critically important. It has been reported that the RRP after RS in Pt/Nb(0.5 wt%):STO can be used to mimic synaptic plasticity for AI applications [1]. The development of nanoscale devices with this neuromorphic function is the basis for hardware implementation of artificial neural networks. For practical application, details of the resistive relaxation phenomenon is required to be understood. Schottky parameters (SPs) such as barrier height (SBH) and depletion width (<i>W</i><sub>D</sub>) can be determined by combining <i>I-V</i> and <i>C-V</i> measurements. SPs had been estimated only for two extreme states, high-resistance states (HRS) and low-resistance states (LRS) [2], and we reported on the time dependence of SPs during the resistance relaxation last year [3].Since carrier concentration (<i>n)</i> strongly influences SPs, it is important to clarify the time dependence of SPs as a function of <i>n</i>. In this study, we revealed the time dependence of SPs of Nb:STO as a function of Nb doping concentration, for the first time.<br/>Therefore, in this study, we conducted a comparative analysis using devices with varying doping concentrations (0.1, 0.5, 1.0 wt%) to investigate the control of volatility in Pt/Nb:STO junctions. Sequential <i>I-V</i> and <i>C-V</i> measurements were performed on the same junction to extract the time evolution of SPs during the relaxation. Additionally, we investigated the location of electronic traps using the conductance method. "<i>I-V</i> measurement" and "<i>C-V</i> measurement and conductance method" were performed by the two-terminal method using a semiconductor parameter analyzer and impedance analyzer, respectively. Since the resistance value changes logarithmically after RS, the switching unit was employed to quickly switch from the circuit for I-V measurements, which set the junction to HRS or LRS, to the circuit for AC measurements. <br/>We measured the relaxation phenomenon up to 1000 s after RS. After setting the Nb 1.0 wt% device to LRS, SBH was estimated to increase by 0.12 eV from 0.54 to 0.66 eV by <i>I-V</i> measurement and from 0.76 to 0.90 eV by <i>C-V</i> measurement. At the same time, the donor concentration (<i>N</i><sub>D</sub>) decreased from 5.8×10<sup>19</sup> to 4.2×10<sup>19</sup> cm<sup>-3</sup>, and the <i>W</i><sub>D</sub> increased from 20.8 to 26.6 nm. It was observed that all SPs exhibited a linear dependence on the logarithm of time during the resistance relaxation. Furthermore, these trends of temporal changes in SPs were confirmed to be consistent regardless of doping concentrations.<br/>On the other hand, from the results obtained by the conductance method, it was discerned that traps tend to be positioned at deeper locations as the doping concentration decreases. This observation aligned with the 0.2 eV difference in SBH between 0.5 and 1.0 wt% cases. Furthermore, it was found that their respective time dependencies also shift in a direction indicative of traps becoming deeper with time. This result is consistent with the 0.12 eV increase in SBH. Consequently, it is suggested that this effect is attributed to interface-layer traps that become deeper as SBH increases. Therefore, RRP is suggested to occur due to the re-trapping of electrons by defects after they were de-trapped by applying forward bias to set to LRS.<br/><br/>[1] T. F. Tiotto <i>et al., Fron. Neurosci. </i>14, 627276 (2021).<br/>[2] C. Park <i>et al., J. Appl. Phys.</i> 103, 054106 (2008).<br/>[3] H. Nakamura <i>et al.,</i> 2022 MRS Fall Meeting, SF06.06, 3784242.