Printed Volatile Memristors for Neuromorphic Computing Hardware

NU 2025-007

INVENTORS

• Mark Hersam*
• Shreyash Hadke
• Vinod Sangwan

SHORT DESCRIPTION 
This technology involves the development of printed volatile memristors designed for neuromorphic computing hardware. These memristors can be printed on flexible substrates, and mimic the complex spiking behavior of biological neurons.

BACKGROUND
Neuromorphic computing aims to emulate the neural architecture and operation of the human brain to overcome the limitations of conventional computing paradigms. Traditional approaches using silicon-based circuits fail to capture the adaptive complexity of biological neurons, resulting in inefficient and less scalable systems. The integration of solution-processed semiconductors on flexible substrates offers a promising alternative to existing silicon-based methods, facilitating the development of scalable and biocompatible artificial neurons.

ABSTRACT
This invention introduces a new class of printed nonlinear dynamical systems that replicate neurobiological computation. By overcoming the challenges associated with solution-based fabrication and integrating these devices into spiking circuits with bio-realistic characteristics, the technology enables scalable printing of volatile threshold switching memristors. These memristors offer an innovative approach to creating neuromorphic systems, enhancing their utility in biohybrid systems, brain-machine interfaces, and intelligent sensing applications. The ability to fabricate these devices on flexible substrates aligns them more closely with biological tissues, offering an advantage over traditional rigid silicon-based systems.

APPLICATIONS

• Neuromorphic computing hardware
  • Enhances computing systems by closely mimicking biological neural networks.
• Spiking neural networks
  • Supports networks that mimic the spiking behavior of neurons.
• Biohybrid systems
  • Integrates biological and artificial systems for advanced computing solutions.
• Brain-machine interface
  • Facilitates communication between neural networks and electronic devices.
• Neurostimulation
  • Offers potential for stimulating neural activity in medical applications.
• Intelligent sensing
  • Employs advanced sensing capabilities inspired by biological systems.
• Flexible electronics
  • Provides flexibility for electronic devices used in various applications.
• Printed electronics
  • Enables the creation of electronic components through additive manufacturing

ADVANTAGES

• Scalability and low-cost fabrication
  • Supports large-scale production of neuromorphic devices at reduced costs.
• Energy efficiency and biocompatibility
  • Enhances compatibility with biological systems while improving energy use.
• Complex spiking pattern replication
  • Mimics intricate neural spiking patterns more effectively than silicon transistors.
• Flexible and adaptive systems
  • Aligns closely with the dynamic nature of biological tissues.

PUBLICATIONS

Hersam, Mark and Hadke, Shreyash and Sangwan, Vinod "Journal Article" submitted to Nature, Availability TBD

IP STATUS
US Provisional Patent Application Filed.