Electronic Instabilities Leading to Electroformation of Binary Metal Oxide‐based Resistive Switches

Abstract

Oxide‐based resistive switching devices are a leading contender for the next generation memories. Before use, each device has to go through a conditioning process called electroformation which has been suggested to be initiated by the accumulation of oxygen vacancies. Here, experimental evidence is presented which shows that both Ta2O5‐x‐ and TiO2‐x‐based crossbar devices, exhibit characteristic electronic instability leading to a reversible constriction of the current flow to a narrow filament prior to permanent change. Thus, it is asserted, electroformation is initiated through purely electronic and reversible events, to be followed later by structural changes in the material, like oxygen vacancy redistribution. Furthermore, the electronic instability responsible for electroformation also gives rise to negative differential resistance (NDR) and that this characteristics appears to involve two distinct mechanisms: a thermal one in which Joule heating causes resistance to decrease as current increases and a second electronic mechanism that appears not to require Joule heating for NDR. Using a combination of thermometry and thermal modeling, a self‐consistent temperature and filament radius as a function of power are found for the 5 μm cross‐bar devices. In the thermal NDR regime, the filament appears to be ∼500 nm in diameter and has a peak temperature of ∼300 °C, while in the adiabatic regime, the estimated filament diameter is much smaller (<50 nm) and the maximum temperature inside it exceeds 800 °C.

Transient phases during fast crystallization of organic thin films from solution

“…17 This was confirmed experimentally in our TiO x and TaO x devices. 11 The filament, thus formed is volatile 16 as the I-V curves are fully reversible up to this point. While the temperature may increase during the initial thermal NDR beyond that of uniform heating, 11 the filament size at this stage is large and the temperature rise is limited to a range where it does not cause permanent changes in the device characteristics.…”

“…11 The filament, thus formed is volatile 16 as the I-V curves are fully reversible up to this point. While the temperature may increase during the initial thermal NDR beyond that of uniform heating, 11 the filament size at this stage is large and the temperature rise is limited to a range where it does not cause permanent changes in the device characteristics. 11 Typically, the threshold voltage is defined as the voltage at which the device undergoes an abrupt transition from the unformed high resistance state to the conducting state which could either be volatile or not.…”

“…While the temperature may increase during the initial thermal NDR beyond that of uniform heating, 11 the filament size at this stage is large and the temperature rise is limited to a range where it does not cause permanent changes in the device characteristics. 11 Typically, the threshold voltage is defined as the voltage at which the device undergoes an abrupt transition from the unformed high resistance state to the conducting state which could either be volatile or not. Frequently, this voltage corresponds to the highest voltage on the I-V curve at which the device differential resistance becomes zero.…”

“…10 Clear experimental evidence of these two processes in TiO x and TaO x devices has been provided in our earlier publication. 11 However, the nature of threshold switching in oxides remains uncertain. The early suggestions favored an electronic process [11][12][13] but the process can also be interpreted as a result of secondary phase nucleation 14 possibly involving vacancies and reduction of the functional oxide.…”

“…11 However, the nature of threshold switching in oxides remains uncertain. The early suggestions favored an electronic process [11][12][13] but the process can also be interpreted as a result of secondary phase nucleation 14 possibly involving vacancies and reduction of the functional oxide. In this work, we attempt to determine the lowest device temperature at which the threshold switching can be initiated.…”