Related Experiment Video
Updated: Jan 11, 2026
15:06
Measurement of Scattering Nonlinearities from a Single Plasmonic Nanoparticle
Published on: January 3, 2016
13.3K
Digital and plasmonic artificial neural networks-Improved nonlinear signal processing at high speed and low
Tobias Blatter1, Amane Zürrer1, Yannik Horst1
1ETH Zurich, Institute of Electromagnetic Fields (IEF), Gloriastrasse 35, 8092 Zürich, Switzerland.
Science Advances
|November 14, 2025
Summary
A new plasmonic artificial neural network (ANN) uses photonic and plasmonic modulators to reduce signal distortions. This compact, high-speed ANN offers a power-efficient alternative to traditional digital signal processing for faster data transmission.
Area of Science:
- Photonics
- Artificial Intelligence
- Signal Processing
Background:
- Increasing data rates necessitate advanced digital signal processing, leading to higher power consumption and costs.
- Nonlinear signal distortions pose a significant challenge in high-speed optical communication systems.
Purpose of the Study:
- To introduce a novel photonic/plasmonic artificial neural network (ANN) for direct mitigation of nonlinear signal distortions.
- To demonstrate the potential of plasmonic modulators in creating compact and efficient signal processing solutions.
Main Methods:
- Development of a plasmonic ANN utilizing plasmonic modulators.
- Comparison of the plasmonic ANN's performance against digital feed-forward equalizers, Volterra series, and digital ANNs.
- Experimental execution of the ANN on a plasmonic chip.
Main Results:
- The plasmonic ANN achieved superior signal-to-noise ratio (SNR) compared to classical equalizers with reduced computational effort.
- The ultracompact plasmonic ANN demonstrated high-speed operation and significantly lowered the demand for electronic processing.
- The chip-based plasmonic ANN showed remarkable equalization performance with minimal components.
Conclusions:
- The developed plasmonic ANN presents a viable path toward ultracompact, high-speed, and power-efficient signal processing.
- This technology offers a low-latency alternative to conventional electronic signal processing methods.
- Plasmonic ANNs hold promise for future optical communication systems demanding greater efficiency and performance.
