Optimization of Gd(0.2)Ca(0.8)MnO(3)-based capacitive memristors

Abstract

After the time of miniaturizing transistors used in integrated circuits comes to an end, another computational method is likely to answer the ever-growing need for more computational power. One of the possible options for evolving past the traditional von Neumann architecture is neuromorphic, human brain-inspired, computing, based on the memory and data transfer properties of synapses and neurons. These two main components of the human brain can be artificially fabricated with different methods and materials. Memristors, short for memory resistors, have been proven to work both as synapses and neurons, making them an ideal component. In this thesis I studied and optimized the fabrication process of Gd(0.2)Ca(0.8)MnO(3) (GCMO) thin film capacitive memristive devices. The fabrication methods tested in this thesis were pulsed laser deposition, photolithography with etching, and electron beam deposition. Five iterations of samples with varying combinations of the fabrication methods for the three layered capacitive structure were studied. The characterization of the fabricated samples was done mainly with electrical transport measurements using ArC ONE (Memristor characterisation platform), with additional surface and material quality checks done with atomic force microscopy, scanning electron microscopy, and energy-dispersive spectroscopy. The measurement results showed some memristive properties, for example, accessible high resistance and low resistance states, for the final fabricated device. As the difference in resistance value between these two states was found to be much smaller than what has been measured for a planar GCMO memristor, some optimization is still left to do on the fabrication process of the capacitive structure. Also, the device-to-device variation was high. Manganite oxide memristors have been widely studied in relation to the possibility of working as a synapse or a neuron. For example, for PCMO, both planar and capacitive memristors have been researched. As no measurement results for GCMO-based memristors with capacitive structures have been published previously, it was crucial to find out whether capacitive GCMO possesses the same properties as its planar counterpart. A new, capacitive structure could advance the usability of GCMO as a component for future neuromorphic applications, and deepen the knowledge of the memristive properties of the material.