vo2-mott-spiking-neuron-hardware

name: vo2-mott-spiking-neuron-hardware description: "VO2 Mott oscillator-based spiking neuron hardware for neuromorphic computing. Monolithic CMOS-BEOL integration of energy-efficient spiking neurons using vanadium dioxide phase-transition materials. Activation: vo2, mott, spiking neuron, neuromorphic hardware, phase-transition, BEOL integration."

VO2 Mott Spiking Neuron Hardware

Monolithic back-end-of-line (BEOL) integration of VO2-based spiking neurons on CMOS-compatible platforms for energy-efficient neuromorphic computing.

Metadata

• Source: arXiv:2604.21487v1
• Authors: Fabio Bersano, Cyrille Masserey, Vanessa Conti, Andrea Iaconeta, et al.
• Published: 2026-04-23
• Institution: EPFL, Switzerland

Core Methodology

Key Innovation

First demonstration of monolithic BEOL integration of one-transistor-one-VO2-memristor (1T-1MR) spiking neurons on CMOS-compatible platforms, achieving sub-20 pJ energy consumption per spike with scalable manufacturing.

Technical Framework

Device Architecture

• Configuration: 1T-1MR (one-transistor-one-memristor) compact architecture
• Substrate: Dielectrically isolated silicon-on-insulator (SOI) p-type junctionless field-effect transistors (JLFETs)
• VO2 Fabrication: Pulsed-laser deposition below 430°C
• Device Dimensions: 60 nm-thick VO2 with 6 μm² active area

Performance Characteristics

Key Phenomena

1. Gate-Tunable Oscillations: Frequency control via gate voltage
2. Non-Monotonic Frequency Dependence: Oscillation frequency depends non-monotonically on current and temperature
3. Bias-Dependent Stochastic Firing: Rich dynamical behavior for probabilistic computing
4. Voltage-Controlled Oscillator: Demonstrated functionality with active tunable resistive coupling

Implementation Guide

Prerequisites

• Cleanroom fabrication facilities
• Pulsed-laser deposition system
• SOI wafer with junctionless FETs
• Characterization equipment (oscilloscope, probe station)

Fabrication Steps

1. Substrate Preparation: SOI p-type JLFET fabrication
2. VO2 Deposition: Pulsed-laser deposition at <430°C
3. Device Patterning: Nanosheet device definition
4. BEOL Integration: Back-end-of-line metal routing
5. Characterization: Electrical and thermal testing

Circuit Configuration

RL Circuit Integration:
- Two-terminal VO2 device
- Series inductor L
- DC bias voltage
- Temperature control

Applications

• Neuromorphic Edge Computing: Ultra-low power AI at the edge
• Probabilistic Computing: Stochastic firing for Bayesian inference
• Oscillatory Neural Networks: Coupled oscillator computing
• Brain-Inspired Sensors: Event-driven sensory processing

Pitfalls

• Thermal Management: VO2 transition near room temperature requires precise thermal control
• Process Compatibility: BEOL temperature budget constraints (<430°C)
• Variability: Stochastic firing may require calibration for deterministic applications
• Scaling Challenges: Active area reduction while maintaining performance

Related Skills

• inhibitory-neuristor-mit: Complementary inhibitory neuron implementation
• neuromorphic-parametric-oscillators-v2: Alternative oscillatory neuromorphic approach
• neuromorphic-photonic-neuronsel: Photonic spiking neurons
• cmosx-mtj-neuron-nonlinear-classification: CMOS+X neuron approaches

References

• Bersano, F. et al. "Monolithically Integrated VO2 Mott Oscillators for Energy-Efficient Spiking Neurons." arXiv:2604.21487 (2026).