Abstract
Neuromorphic computing offers parallel data processing and low energy consumption and can be useful to replace conventional von Neumann computing. Memristors are two-terminal devices with varying conductance that can be used as synaptic arrays in hardware-based neuromorphic devices. In this research, we extensively investigate the analog symmetric multi-level switching characteristics of zinc tin oxide (ZTO)-based memristor devices for neuromorphic systems. A ZTO semiconductor layer is introduced between a complementary metal-oxide-semiconductor (CMOS) compatible Ni top electrode and a highly doped poly-Si bottom electrode. A variety of bio-realistic synaptic features are demonstrated, including long-term potentiation (LTP), long-term depression (LTD), and spike timing-dependent plasticity (STDP). The Ni/ZTO/Si device in which the adjustment of the number of states in conductance is realized by applying different pulse schemes is highly suitable for hardware-based neuromorphic applications. We evaluate the pattern recognition accuracy by implementing a system-level neural network simulation with ZTO-based memristor synapses. The density of states (DOS) and charge density plots reveal that oxygen vacancies in ZTO assist in generating resistive switching in the Ni/ZTO/Si device. The proposed ZTO-based memristor composed of metal-insulator-semiconductor (MIS) structure is expected to contribute to future neuromorphic applications through further studies.
Highlights
The recent emergence of artificial intelligence (AI), big data, and the Internet of Things (IoT) has defined a new paradigm of digital system alternation that has dramatically increased data processing complexity in terms of represented power, size, the number of gates, the amount of memory, and types of environment [1]
The zinc tin oxide (ZTO) supercell was modeled with the following atomic ratio of Zn 4% and Sn 1%, which is very close to our experimental sample ratio, so it contains a total number of 64 atoms (Zn = 24, Sn = 6 and O = 34)
The ionic positions, cell volume and lattice parameters of the system were fully relaxed with the conjugate gradient (CG) method until Hellmann Feynman forces became smaller than 0.02 eV/Å while the energy convergence criteria was met at 1×10−5 eV [50]
Summary
The recent emergence of artificial intelligence (AI), big data, and the Internet of Things (IoT) has defined a new paradigm of digital system alternation that has dramatically increased data processing complexity in terms of represented power, size, the number of gates, the amount of memory, and types of environment [1]. Used in a conventional computing system can execute logic processing and arithmetic operations, but it is susceptible to problems due to scaling, power consumption, and device heating [2], [3]. To overcome the problems affecting these conventional computing systems, most studies have concentrated on finding new types of computing architectures, such as a neuromorphic or in-memory computing systems [4], [5]. Neuromorphic systems mimic neurons and synapses of the human brain, in which a variety of arithmetic, logic, learning, and memory activities are conducted using a low amount of. Neuromorphic engineering requires control of the conductance in an analog manner, spike time-dependent plasticity (STDP), and a CMOS compatible with neuron circuits
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
