Electroforming-Free Y₂O₃Memristive Crossbar Array with Low Variability

Abstract

Transition metal oxides play a very important role to develop the memristive crossbar array for nonvolatile memory for storage and logic operations. However, the development of a high-density memristive crossbar array for complex applications is restricted due to low device yield and high device-to-device (D2D) and cycle-to-cycle (C2C) variability in device switching voltages. Here, we report the fabrication of a stable, highly scalable, reproducible, Y2O3-based memristive crossbar array of (15 × 12) on silicon by utilizing a dual ion beam sputtering system. The fabricated crossbar array exhibits the intrinsic nonlinear characteristics of the memristive element by displaying a high endurance (∼7 × 105 cycles), high current ratio (>200), good retention (∼1.5 × 105 s), high device yield, low device-to-device (D2D) (0.25), and cycle-to-cycle (C2C) (0.608) variability in the SET/RESET voltages of the memristive device, which can be further suitable for analog computation and logic operations.

Enhancing memristor fundamentals through instrumental characterization and understanding reliability issues

“…Memristors typically exhibit undesirable variations in resistance values and SET and RESET voltages from one device to device (also called D2D or spatial variation) and from one cycle to cycle (also called C2C or temporal variation). 165,166 From the perspective of physical mechanism, the cycle-to-cycle variation originates from the stochastic formation and rupture of conductive filaments. 43 The device-to-device variation stems from the limitations of fabrication techniques, which produce nonideal film morphology and homogeneity, resulting in the different electrical characteristics between memristors.…”

Experimental Validation of Switching Dependence of Nanoscale Y2O3 Memristors on Electrode Symmetry via Physical Electrothermal Modeling

“…Figure 1b depicts the field emission scanning electron microscopy (FESEM), which reveals a continuous SL of Y 2 O 3 with compact grains, as reported previously. 10 Furthermore, Figure 1c depicts the cross-sectional view of the high-resolution transmission electron microscopy (HR-TEM), wherein all the deposited layers are distinct to each other having clear interfaces in the fabricated structure. The HRTEM analysis also confirms the absence of interfacial oxide between the Si and Y 2 O 3 insulating layer, which is beneficial for uniform SL deposition 10 to form an abrupt interface.…”

“…10 Furthermore, Figure 1c depicts the cross-sectional view of the high-resolution transmission electron microscopy (HR-TEM), wherein all the deposited layers are distinct to each other having clear interfaces in the fabricated structure. The HRTEM analysis also confirms the absence of interfacial oxide between the Si and Y 2 O 3 insulating layer, which is beneficial for uniform SL deposition 10 to form an abrupt interface. Figure 1d shows the atomic force microscopy (AFM) analysis, in which the surface root mean roughness of the SL is below 2 nm.…”

“…Recently, our research group has successfully fabricated Y 2 O 3 -based memristive crossbar arrays on 3 inch-diameter Si substrates with record-high endurance and retention properties and device production yield along with low device-to-device (D2D) and ultralow cycle-to-cycle (C2C) variability parameters. , To further extend the commercial viability of the Y 2 O 3 -based memristor, it is critical to thoroughly investigate the role of electrode materials in device switching. Therefore, in this work, for the first time in the literature, we thoroughly investigate the various simulated characteristics of the Y 2 O 3 memristor with symmetric and asymmetric electrode combinations via physical electrothermal modeling with the corresponding experimental validation.…”

“…Also, the symmetric electrode-based simulated results are thoroughly experimentally validated. 10,11 The presented simulation study is based on the quantitative thermodynamic numerical model of the memristor, which is further correlated with the free energy of the utilized materials. Here, memristors are designed and modeled in COMSOL Multiphysics, 9 which helps to solve the partial differential equations by utilizing the finite element method (FEM).…”