Unraveling the importance of fabrication parameters of copper oxide-based resistive switching memory devices by machine learning techniques

Abstract

In the present study, various statistical and machine learning (ML) techniques were used to understand how device fabrication parameters affect the performance of copper oxide-based resistive switching (RS) devices. In the present case, the data was collected from copper oxide RS devices-based research articles, published between 2008 to 2022. Initially, different patterns present in the data were analyzed by statistical techniques. Then, the classification and regression tree algorithm (CART) and decision tree (DT) ML algorithms were implemented to get the device fabrication guidelines for the continuous and categorical features of copper oxide-based RS devices, respectively. In the next step, the random forest algorithm was found to be suitable for the prediction of continuous-type features as compared to a linear model and artificial neural network (ANN). Moreover, the DT algorithm predicts the performance of categorical-type features very well. The feature importance score was calculated for each continuous and categorical feature by the gradient boosting (GB) algorithm. Finally, the suggested ML guidelines were employed to fabricate the copper oxide-based RS device and demonstrated its non-volatile memory properties. The results of ML algorithms and experimental devices are in good agreement with each other, suggesting the importance of ML techniques for understanding and optimizing memory devices.

Resistive switching and role of interfaces in memristive devices based on amorphous NbOx grown by anodic oxidation

“…Indeed, anodic oxidation is largely exploited in the fabrication of RS devices based on TMO such as HfO x , TaO x , TiO x , and also NbO x . [30][31][32][33][34][35][36][37][38][39][40] However, in the literature, few studies have been conducted on the anodic oxidation of niobium as a resistive switching layer. 30,[35][36][37]41 In this work, we report on a comprehensive investigation of the resistive switching mechanism in amorphous NbO x realized by anodic oxidation by combining the analysis of the resistive switching properties of Nb/NbO x /Au devices together with a detailed chemical, structural and morphological analysis of the NbO x layer and its interfaces.…”

“…Indeed, anodic oxidation is largely exploited in the fabrication of RS devices based on TMO such as HfO x , TaO x , TiO x , and also NbO x . 30–40 However, in the literature, few studies have been conducted on the anodic oxidation of niobium as a resistive switching layer. 30,35–37,41…”

Electrochemical anodic oxidation assisted fabrication of memristors

“…With the advantages of simple architecture, high-density scalability, low power consumption, good compatibility with the CMOS technique, and multi-level storage [28][29][30][31][32], ReRAMs are key candidates for the next generation non-volatile memories (NVMs) [33,34]. For NVM applications, memristors require excellent endurance and retention, good yield, on/off ratio and linearity [35,36]. Apart from NVMs, memristors are also applied to logic operations [37], bio-inspired neuromorphic computations [38][39][40][41][42], and data encryption [43,44].…”

Intensive harmonized synapses with amorphous Cu2O-based memristors using ultrafine Cu nanoparticle sublayers formed via atomically controlled electrochemical pulse deposition

“…Since it is difficult to secure the reproducibility of each CF that is randomly distributed in a probabilistic manner, it is vulnerable to long-term operation and noise signals, causing problems in storing the delicately adjusted resistance states. [29][30][31][32] The aforementioned filamentary 1, 2, and interfacial types still have severe weak points to become an ideal synapse device, which is strongly connected with the filament formation mechanism (Scheme 1d). 15,18,[22][23][24][25][26] Therefore, an unprecedented conductive filament structure is needed to secure good retention, fast driving speed, and ideal RS behavior.…”

“…Since it is difficult to secure the reproducibility of each CF that is randomly distributed in a probabilistic manner, it is vulnerable to long-term operation and noise signals, causing problems in storing the delicately adjusted resistance states. 29–32…”