IP: 193.93.193.199 On: Sun, 19 Aug 2018 07:13:16 Copyright: American Scientific Publishers Delivered by Ingenta Copyright © 2018 American Scientific Publishers All rights reserved Printed in the United States of America Article Journal of Nanoscience and Nanotechnology Vol. 18, 984–991, 2018 www.aspbs.com/jnn Mimicking the Synaptic Weights and Human Forgetting Curve Using Hydrothermally Grown Nanostructured CuO Memristor Device T. D. Dongale 1 ∗† , P. S. Pawar 1† , R. S. Tikke 1† , N. B. Mullani 1 , V. B. Patil 1 , A. M. Teli 2 , K. V. Khot 3 4 , S. V. Mohite 2 , A. A. Bagade 2 , V. S. Kumbhar 2 , K. Y. Rajpure 2 , P. N. Bhosale 3 , R. K. Kamat 5 , and P. S. Patil 2 1 Computational Electronics and Nanoscience Research Laboratory, School of Nanoscience and Biotechnology, Shivaji University, Kolhapur 416004, India 2 Department of Physics, Shivaji University, Kolhapur 416004, India 3 Department of Chemistry, Shivaji University, Kolhapur 416004, India 4 Sharad Institute of Technology, College of Engineering, Yadrav 416115, India 5 Department of Electronics, Shivaji University, Kolhapur 416004, India In the present investigation, we have fabricated copper oxide (CuO) thin film memristor by employ- ing a hydrothermal method for neuromorphic application. The X-ray diffraction pattern confirms the films are polycrystalline in nature with the monoclinic crystal structure. The developed devices show analog memory and synaptic property similar to biological neuron. The size dependent synaptic behavior is investigated for as-prepared and annealed CuO memristor. The results suggested that the magnitude of synaptic weights and resistive switching voltages are dependent on the thickness of the active layer. Synaptic weights are improved in the case of the as-prepared device whereas they are inferior for annealed CuO memristor. The rectifying property similar to a biological neuron is observed only for the as-prepared device, which suggested that as-prepared devices have bet- ter computational and learning capabilities than annealed CuO memristor. Moreover, the retention loss of the CuO memristor is in good agreement with the forgetting curve of human memory. The results suggested that hydrothermally grown CuO thin film memristor is a potential candidate for the neuromorphic device development. Keywords: Memristor, Synapse, Neuromorphic Computing, Copper Oxide, Thin Film, Hydrothermal Method. 1. INTRODUCTION It is widely believed that the bottleneck of von Neumann architecture is due to the sequential processing of data and instructions. Moreover, lower processing speed, higher power dissipation, and higher footprint are responsible for inefficient digital computers. 1 On the other hand, the biological brain has remarkable computing and memory capabilities that cannot be emulated by conventional dig- ital computers. There will be huge opportunities in near future if a human can mimic the biological brain by digi- tal means. Owing to this fact, it is obligatory to develop a ∗ Author to whom correspondence should be addressed. † These three authors contributed equally to this work. neuromorphic computing architecture to achieve true par- allel processing and huge computational capabilities simi- lar to biological brain and memristor based neuromorphic computing architecture is one of the solutions. Memristor is a two terminal passive circuit element and its mem- ductance can be controlled by one or more state variables to achieve memory property. 2 The passivity, nonlinearity, and memory property make it favorable and fundamen- tal building block of next-generation memory, logic, and computing applications. 3 The functional element of the biological brain is a neu- ron and its performance decides the computing and mem- ory capabilities of the brain. The memristor is a two terminal element similar to a neuron with intrinsic memory 984 J. Nanosci. Nanotechnol. 2018, Vol. 18, No. 2 1533-4880/2018/18/984/008 doi:10.1166/jnn.2018.14264