Full text loading...
I-V characteristics for the Ru/HfO2/TiOx/Ru device structure. (a) Initial I-V characteristics before any bias was applied. (b) I-V switching characteristics. Curve 1 shows the forming step, where a compliance current was used to prevent a hard breakdown. Curves 2 and 3 show the reset process, where devices start to show a gradual reset around −1.5 V on curve 2. Curves 4 and 5 show the setting of the device from a well defined HRS state with a compliance level of 5 mA.
Conduction mechanism fittings using temperature dependent measurements for VRS, LRS, and HRS. Data is shown by shapes for a wide temperature range, while conduction mechanism fitting is shown with a solid line for smaller temperature ranges. (a) VRS fitting for F-P emission between 250 K and 365 K and TAT fitting between 125 K and 160 K. (b) LRS fitting for Ohmic conduction between 80 K and 365 K. (c) HRS fitting for F-P emission between 190 K and 365 K and TAT fitting between 80 K and 97.6 K. I-V fittings for conduction mechanisms in different temperature regions are included in the insets for VRS, LRS, and HRS.
Multiple LRS states were achieved by altering the compliance level. The resistance of the resulting linear I-V characteristics was extracted and the resistance vs. compliance level is shown. The area of the filament was then calculated based off several assumptions, including a single Hf filament with a constant cross sectional area stretching from the B.E. to the TiOx layer. The area vs. compliance current is shown on the right axis.
Multiple HRS states were achieved by altering the reset voltage. The resistance vs. reset voltage is shown, where the resistance was taken at −0.5 V after the device has been reset. The change in resistance is believed to be due to an increase in the thickness (treox) of the re-oxidized HfO2 layer. treox vs. reset voltage is shown on the right axis.
Triangular pulse waveform capturing the gradual reset. (a) Transient waveform of both the voltage applied and the current measured. Voltage was applied to the top electrode, while the current was measured from the bottom electrode. Device initially started to reset at 4.9 ms and continued to reset close to 7.1 ms. (b) The total energy and the resistance were extracted and are shown versus time. The reset process continued to occur until the energy began to saturate and the final HRS state was achieved. (c) Resistance vs. time shown during the reset where a decrease in resistance is observed initially before the device resets fully. (d) Current vs. voltage, illustrating the gradual reset from the transient waveform.
Article metrics loading...