Master Thesis: Optimization of Ag-based Volatile Diffusive Memristors for Neuromorphic Computing Applications

 

Brain inspired neural networks are emerging as potential alternatives of the traditional Von-Neumann Architecture, mainly due to their unique structure of combining the memory and processor together. One example are Spiking Neural Networks (SNNs), where information is processed through the spiking of neurons. CMOS-based integrate-and-fire neuron is typically used for processing the incoming signal. However, such neuron, composed of several transistors and a capacitor, is area- and energy-inefficient. Recently, a compact two-terminal threshold switch (TS)-based neuron has been exploited as replacement of the complex circuitry. Neurons play the active element role in the circuit to execute current summation, integration and firing. As frequent integrating and firing events are involved, the power consumption of neurons needs to be low enough to ensure an energy efficient computing element. Therefore it is important to enhance the volatility behavior and improve stability, variability and cycling endurance of the TS-based neurons.

Tasks:

• Fabrication of micro-structured ECM cells using technologically promising materials such as HfO2, TiO2, SiO2, or mixed as electrolyte and Ag/AgTe metal as top electrode. The different oxides are to be grown using atomic layer deposition (ALD) technique.
• Electrical characterization of the fabricated devices. The current-voltage measurements should be carried out in continuous and pulse mode as a function of temperature and atmosphere.
• The target is to maximize the volatility behavior by applying techniques such as doping and/or annealing as well as performance enhancement in general. These includes switching stability, variability and cycling endurance.

Applicants profile:
You are studying electrical engineering, physics or materials science in the master’s program with knowledge in the field of semiconductors or information technology and are interested in experimental and interdisciplinary work. You have basic knowledge of fabrication and measurement technology, data analysis and programming skills and are ready to familiarize yourself with new methods.

Contact:
M. Sc. Solomon Amsalu Chekol Tel: +49 2461 616288 ; E-mail: s.chekol@fz-juelich.de
Dr. Susanne Hoffmann-Eifert Tel: +49 2461 616505; E-mail: su.hoffmann@fz-juelich.de

Forschungszentrum Jülich GmbH
Peter Grünberg Institut
Electronic Materials (PGI-7)
JARA-FIT Institut Green IT

This advertisement as pdf file:   Master Thesis: Optimization of Ag-based Volatile Diffusive Memristors for Neuromorphic Computing Applications (PDF, 1 MB)