Sol–gel-processed Y2O3 films were used as active channel layers for resistive random access memory (RRAM) devices. The fabricated ITO/Y2O3/Ag RRAM devices exhibited the properties of conventional bipolar memory devices. A triethylamine stabilizer with a high vapor pressure and low surface tension was added to realize the local electric field area. During drying and high-temperature post-annealing processes, the large convective flow enhanced the surface elevation, and the increased –OH groups accelerated the hydrolysis reaction and aggregation. These phenomena afforded Y2O3 films with an uneven surface morphology and an increased surface roughness. The increased roughness of the Y2O3 films attributable to the triethylamine stabilizer enhanced the local electrical field, improved device reliability, and achieved successful repetition of the switching properties over an extended period.

Flexible indium tin oxide (ITO)/Y2O3/Ag resistive random access memory (RRAM) devices were successfully fabricated using a thermal-energy-free ultraviolet (UV)/ozone-assisted photochemical annealing process. Using the UV/ozone-assisted photochemical process, the organic residue can be eliminated, and thinner and smother Y2O3 films than those formed using other methods can be fabricated. The flexible UV/ozone-assisted photochemical annealing process-based ITO/Y2O3/Ag RRAM devices exhibited the properties of conventional bipolar RRAM without any forming process. Furthermore, the pure and amorphous-phase Y2O3 films formed via this process showed a decreased leakage current and an increased high-resistance status (HRS) compared with the films formed using other methods. Therefore, RRAM devices can be realized on plastic substrates using a thermal-energy-free UV/ozone-assisted photochemical annealing process. The fabricated devices exhibited a resistive window (ratio of HRS/low-resistance status (LRS)) of >104, with the HRS and LRS values remaining almost the same (i.e., limited deterioration occurred) for 104 s and up to 102 programming/erasing operation cycles.

In this study, we fabricated sol-gel-processed Y2O3-based resistive random-access memory (RRAM) devices. The fabricated Y2O3 RRAM devices exhibited conventional bipolar RRAM device characteristics and did not require the forming process. The long-term stability of the RRAM devices was investigated. The Y2O3 RRAM devices with a 20 nm thick Ag top electrode showed an increase in the low resistance state (LRS) and high resistance state (HRS) and a decrease in the HRS/LRS ratio after 30 days owing to oxidation and corrosion of the Ag electrodes. However, Y2O3 RRAM devices with inert Au-passivated Ag electrodes showed a constant RRAM device performance after 30 days. The 150 nm-thick Au passivation layer successfully suppressed the oxidation and corrosion of the Ag electrode by minimizing the chance of contact between water or oxygen molecules and Ag electrodes. The Au/Ag/Y2O3/ITO RRAM devices exhibited more than 300 switching cycles with a decent resistive window (>103). They maintained constant LRS and HRS resistances for up to 104 s, without significant degradation of nonvolatile memory properties for 30 days while stored in air.

Yttrium oxide (Y2O3) resistive random-access memory (RRAM) devices were fabricated using the sol–gel process on indium tin oxide/glass substrates. These devices exhibited conventional bipolar RRAM characteristics, without requiring a high voltage forming process. The effect of current compliance on the Y2O3 RRAM devices was investigated and revealed that as the set current compliance values increased, the resistance values gradually decreased. Consequently, intermediate resistance values were obtained, which are suitable for multi-level cell (MLC) switching. The fabricated Y2O3 RRAM devices were capable of functioning as an MLC with a capacity of 2 bits in one cell, utilizing three low resistance states and one common high resistance state. The potential of the Y2O3 RRAM for neural networks was further explored through numerical simulations. Hardware neural networks based on the Y2O3 RRAM demonstrated effective digit image classification with a high accuracy rate of approximately 88%, comparable to the ideal software-based classification (~92%). This suggests that the proposed RRAM has the potential to be utilized as a memory component in practical neuromorphic systems.