Advances of RRAM Devices: Resistive Switching Mechanisms, Materials and Bionic Synaptic Application

Abstract

Resistive random access memory (RRAM) devices are receiving increasing extensive attention due to their enhanced properties such as fast operation speed, simple device structure, low power consumption, good scalability potential and so on, and are currently considered to be one of the next-generation alternatives to traditional memory. In this review, an overview of RRAM devices is demonstrated in terms of thin film materials investigation on electrode and function layer, switching mechanisms and artificial intelligence applications. Compared with the well-developed application of inorganic thin film materials (oxides, solid electrolyte and two-dimensional (2D) materials) in RRAM devices, organic thin film materials (biological and polymer materials) application is considered to be the candidate with significant potential. The performance of RRAM devices is closely related to the investigation of switching mechanisms in this review, including thermal-chemical mechanism (TCM), valance change mechanism (VCM) and electrochemical metallization (ECM). Finally, the bionic synaptic application of RRAM devices is under intensive consideration, its main characteristics such as potentiation/depression response, short-/long-term plasticity (STP/LTP), transition from short-term memory to long-term memory (STM to LTM) and spike-time-dependent plasticity (STDP) reveal the great potential of RRAM devices in the field of neuromorphic application.

Keywords: 2D materials; RRAM; artificial intelligence; bionic synaptic application; switching mechanisms; thin film.

Figure 1

Illustration of sandwich structure for RRAM devices.

Figure 2

(a) Unipolar and (b) bipolar Modes for RRAM devices.

Figure 3

Switching mechanism of unipolar (a,b) and bipolar (c,d) AlO_x-based RRAM devices, reproduced from [39], with permission from Springer Nature, 2020.

Figure 4

Schematic of discontinuous and continuous states of mini-band DBs in the middle of silicon oxide band gap.

Figure 5

Illustration of multilevel RESET process of Ag/SiO₂/Pt RRAM device.

Figure 6

Various thin film materials of RS medium for RRAM devices. (a) Silk protein, reproduced from [65], with permission from John Wiley and Sons, 2016. (b) PVK-C₆₀, reproduced from [72], with permission from American Chemical Society, 2007. (c) Albumen, reproduced from [67], with permission from Springer Nature, 2015. (d) PVA, reproduced from [71], with permission from John Wiley and Sons, 2011. (e) AlO_x, reproduced from [19], with permission from Elsevier, 2020. (f) CsPbBr₃, reproduced from [73], with permission from John Wiley and Sons, 2019. (g) Ge₂Sb₂Tr₅, reproduced from [38], with permission from John Wiley and Sons, 2019. (h) GO, reproduced from [74], with permission from AIP Publishing, 2013.

Figure 7

RS performance, including (a) bipolar I-V characteristic, (b) resistance distribution, (c) endurance and (d) retention performance, of Ni/solution-processed AlO_x/Pt RRAM devices annealed at different temperatures, reproduced from [9], with permission from MDPI (Basel, Switzerland), 2019.

Figure 8

MoS₂-rGO hybrid with (a) TEM and (b) HRTEM. (c) Bipolar I-V curves and (d) endurance properties of Ti/MoS₂-rGO/ITO RRAM device, reproduced from [26], with permission from Elsevier, 2019.

Figure 9

(a) Bipolar RS performance and (b) AES depth profiles of Hf/Ta₂O₅/Pt RRAM devices, reproduced from [129], with permission from AIP Publishing, 2013.

Figure 10

(a) Computing systems based on traditional Von Neumann architecture, the memory address of program instruction and the memory address of data point to different physical locations in the same memory device. (b) Computing systems based on neuromorphic architecture with integration of a single synaptic device into each unit.

Figure 11

Structure of synapse in neural network.

Figure 12

(a) Potentiation and depression response of Ag/SiO_x:Ag/TiO_x/p⁺⁺-Si device with repeated voltage sweeps. (b) Conductance modulation and (c) PPF of Ag/SiO_x:Ag/TiO_x/p⁺⁺-Si device by repeating consecutive pulses. (d) Repeated STP response with the model fitting of TiO_x-based RRAM device, reproduced from [158], with permission from Springer Nature, 2020. (e) Synaptic facilitation response to consecutive pulses of the device with h-BN, reproduced from [159], with permission from Elsevier, 2020.

Figure 13

(a) Gradual modulation for conductance with long-term potentiation/depression response of Ag/HZO/ITO/PET RRAM device. Retention performance of Ag/HZO/ITO/PET RRAM device in (b) long-term potentiation process by consecutive positive pulses and (c) long-term depression process by consecutive negative pulses, reproduced from [79], with permission from Springer Nature, 2019. (d) Synaptic weights modulation of Au/PEDOT:PSS/ITO RRAM device by 10 consecutive pulses. (e) Long-term potentiation/depression under 600 consecutive pulses in one operation and (f) five operations of long-term potentiation/depression for Au/PEDOT:PSS/ITO RRAM device, reproduced from [163], with permission from MDPI (Basel, Switzerland), 2018.

Figure 14

(a) Illustration of Cu atom dynamics of Cu/a-Si/Pt device during the transition from STM to LTM. (b) Relationship between normalized current response and retention time when the transferring process from STP to LTP occurred in Ag/SiO_x:Ag/TiO_x/p⁺⁺-Si device. (c) Synaptic weight response to changes of pulse amplitude and relaxation time τ, reproduced from [155], with permission from Springer Nature, 2020.

Figure 15

(a) Implementing programming pulses and (b) STDP behavior of Ag/SiO_x:Ag/TiO_x/p⁺⁺-Si RRAM device, reproduced from [155], with permission from Springer Nature, 2020. (c) Applied pre/post-spikes with sequences and (d) STDP characteristics of TaN/HfO₂/Al₂O₃/ITO RRAM device, reproduced from [20], with permission from Elsevier, 2020. (e) STDP results with pulse interval modulation of pre- and post-synaptic spiking for PEDOT:PSS-based RRAM device, reproduced from [163], with permission from MDPI (Basel, Switzerland), 2018. (f) Experimental and fitting results of STDP behaviors for Ag/SrTiO₃/RGO/FTO RRAM device, reproduced from [166], with permission from Elsevier, 2018.