Enhanced analog synaptic behavior of SiNx/a-Si bilayer memristors through Ge implantation

Abstract

Conductive bridging random access memory (CBRAM) has been considered to be a promising emerging device for artificial synapses in neuromorphic computing systems. Good analog synaptic behaviors, such as linear and symmetric synapse updates, are desirable to provide high learning accuracy. Although numerous efforts have been made to develop analog CBRAM for years, the stochastic and abrupt formation of conductive filaments hinders its adoption. In this study, we propose a novel approach to enhance the synaptic behavior of a SiNx/a-Si bilayer memristor through Ge implantation. The SiNx and a-Si layers serve as switching and internal current limiting layers, respectively. Ge implantation induces structural defects in the bulk and surface regions of the a-Si layer, enabling spatially uniform Ag migration and nanocluster formation in the upper SiNx layer and increasing the conductance of the a-Si layer. As a result, the analog synaptic behavior of the SiNx/a-Si bilayer memristor, such as the nonlinearity, on/off ratio, and retention time, is remarkably improved. An artificial neural network simulation shows that the neuromorphic system with the implanted SiNx/a-Si memristor provides a 91.3% learning accuracy mainly due to the improved linearity.

Cluster-type analogue memristor by engineering redox dynamics for high-performance neuromorphic computing

“…1a). To our best knowledge, the two main features of analogue devices-the linear switching performance and the large dynamic range of Ti 4.8% :a-Si device-display a clear deviation from the previously reported Ag 9,10,15,17,18,[33][34][35][36][37][38][39][40][41][42][43] -and Cu 44,45 -based analogue CBRAM (Fig. 2c).…”

Emulating the short-term plasticity of a biological synapse with a ruthenium complex-based organic mixed ionic–electronic conductor

“…LTspice (17.0.22.0) was used to simulate the STDP learning rule by employing an OMIEC memristor as a STM fitted with a volatile memory model (VM model) and an electrochemical metallization (ECM) device as a long-term memory (LTM) fitted with a Yakopcic model. 20,[42][43][44] These two devices are connected in a series circuit to conduct a SPICE simulation for STDP curves by applying two consecutive pulses. Each pulse is composed of a short-duration pulse with a high amplitude (500 ms with 14 V for a presynaptic spike or 500 ms with 18 V for a postsynaptic spike) and a long-duration pulse with a low amplitude (200 ms and 6 V).…”

“…Many research groups implemented LTP by using memristors such as Ag/Si, SiN x /a-Si, TiN/TaO x , TiO x /Al 2 O 3 , and HfO x /AlO x based devices. [19][20][21][22][23] However, a relatively small number of STP-based memristors have been developed, although their volatile characteristics can be used in spatio-temporal or reservoir computing in recurrent neural networks. 24,25 STP memristors are also potentially applicable to a synaptic memory, which comprises both long-term and short-term partitions, to improve the learning efficiency.…”

Ion-Movement-Based Synaptic Device for Brain-Inspired Computing

“…Kim et al reported improved conductance linearity and dynamic range in a Ag/SiN x /a-Si/p ++ Si-based ECM device through Ge implantation [ 36 ]. Ge implantation induces structural defects in the bulk and surface regions of the a-Si layer, which enables spatially uniform Ag migration and nanocluster formation in the upper SiN x layer, leading to enhanced synaptic behavior.…”

“…Recently, synaptic devices based on various types of resistive switching devices have been utilized for brain-inspired computing [ 25 , 26 , 27 , 28 , 29 , 30 ]. Brain-inspired computing based on synaptic devices has achieved particular progress from ion-movement-based resistive switching mechanisms such as cation-movement-based filaments [ 31 , 32 , 33 , 34 , 35 , 36 ], anion-movement-based filaments [ 37 , 38 , 39 , 40 , 41 , 42 , 43 ], cation-movement-based ferroelectric polarization reversal [ 44 , 45 , 46 , 47 , 48 , 49 ], and ion-movement-based electrochemical electrolytes [ 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 ]. These ion-movement-based resistive switching devices can show a gradual change in conductance and nonvolatile characteristics, which have not been implemented in Mott-insulator-based resistive switching devices.…”