Related Experiment Video
Updated: Apr 1, 2026
08:07
Assembly and Characterization of Biomolecular Memristors Consisting of Ion Channel-doped Lipid Membranes
Published on: March 9, 2019
8.4K
PdNeuRAM: forming-free, multi-bit Pd/HfO2 ReRAM for energy-efficient neuromorphic computing
Erbing Hua1, Theofilos Spyrou2, Majid Ahmadi3,4
1Department of Quantum and Computer Engineering, Delft University of Technology, Delft, The Netherlands. e.hua@tudelft.nl.
Communications Engineering
|March 30, 2026
Summary
New memristor devices eliminate the need for high-voltage electroforming, enabling energy-efficient computing. These palladium-hafnia (Pd/HfO2) devices reduce power consumption for neuromorphic applications.
Area of Science:
- Materials Science
- Electrical Engineering
- Computer Science
Background:
- Memristor technology promises energy-efficient computing but faces challenges like resistance drift, variability, and electroforming.
- Filamentary resistive random-access memory (RRAM) typically requires high-voltage electroforming, increasing power and reducing endurance.
Purpose of the Study:
- To develop forming-free memristive devices based on HfO2.
- To investigate the mechanism behind forming-free resistive switching.
- To demonstrate the potential for energy-efficient neuromorphic computing.
Main Methods:
- Fabrication of HfO2-based memristive devices with Pd and Ti electrodes.
- Combined electrical and materials characterization (e.g., XPS, TEM).
- Analysis of oxygen vacancy formation and migration barriers.
Main Results:
- Developed Pd/HfO2 forming-free memristive devices (PdNeuRAM) operating at low voltages.
- Identified a Pd-O-Hf interfacial configuration facilitating forming-free switching.
- Achieved reduced programming (43%) and read (38%) energy in spiking neural network tasks.
Conclusions:
- Interfacial engineering in Pd/HfO2 devices eliminates the need for electroforming.
- These devices offer multi-bit functionality, reduced variability, and lower energy consumption.
- Demonstrated potential for energy-efficient neuromorphic computing applications.
