Machine learning analysis of a Fano resonance based plasmonic refractive index sensor using U shaped resonators

Area of Science:

• Plasmonics
• Nanophotonics
• Optical Sensing

Background:

• Plasmonic sensors offer compact footprints and high sensitivity for refractive index (RI) measurement.
• Fano resonance (FR) in plasmonic nanostructures provides a basis for sensitive RI detection.

Purpose of the Study:

• To design and analyze a novel Fano resonance-based refractive index sensor.
• To investigate the sensor's performance using plasmonic nanostructures.
• To integrate machine learning for predictive analysis and resource optimization.

Main Methods:

• Utilized a metal-insulator-metal waveguide with U-shaped and inverted U-shaped resonators.
• Employed the finite-difference time-domain (FDTD) method for transmission spectrum analysis.
• Applied the Extreme Randomized Tree regression model for transmittance prediction.

Main Results:

• Achieved a high sensitivity of 571.4 nm/RIU and a figure of merit of 14,987 RIU⁻¹ for the first FR.
• The Extreme Randomized Tree model demonstrated high precision in predicting transmittance values (Adj-R²S close to 1).
• Reduced simulation time and resources by 90% through machine learning integration.

Conclusions:

• The designed plasmonic sensor exhibits excellent performance for RI sensing.
• Machine learning integration enhances predictive capabilities and computational efficiency.
• The sensor holds potential for advanced RI sensing applications.