Related Experiment Video
Updated: Jun 2, 2026
09:48
Construction and Operation of a Light-driven Gold Nanorod Rotary Motor System
Published on: June 30, 2018
Investigating the Optical Properties of Gold Nanorods Using Forward and Inverse Design
Sabrina Islam1, Abu S M Mohsin1, Mohammed Belal Hossain Bhuian1
1Nanotechnology, AI, IoT, and Applied Machine Learning Research Group, Department of Electrical and Electronic Engineering, BRAC University, Kha 224 Bir Uttam Rafiqul Islam Avenue, Merul Badda, Dhaka 1212, Bangladesh.
ACS Omega
|June 1, 2026
Summary
Machine learning accurately predicts gold nanorod optical properties and designs nanoparticles for specific applications. This approach overcomes limitations of traditional methods for complex nanoparticle shapes.
Area of Science:
- * Nanophotonics and computational materials science.
- * Application of machine learning in optical property prediction and inverse design.
Background:
- * Nanoparticles exhibit unique optical properties influenced by size, shape, and composition, crucial for photonics and sensing.
- * Traditional theories (Mie, Mie-Gans) are limited to simple shapes, while advanced methods (FDTD) are computationally intensive.
- * Complex nanoparticle shapes require novel simulation and design approaches.
Purpose of the Study:
- * To develop and validate a machine and deep learning-based methodology for forward and inverse design of gold nanorod optical properties.
- * To overcome the limitations of classical theories and complex simulations for predicting optical behavior of nanoparticles.
- * To enable systematic design of nanoparticles with tailored optical responses for various applications.
Main Methods:
- * Evaluated five machine learning models for forward design, identifying XGB Regressor as optimal.
- * Utilized a tandem model for inverse design to determine gold nanorod dimensions (length, width, aspect ratio) for desired optical responses.
- * Simulated optical properties (absorption and scattering cross sections) across a 400-1500 nm wavelength range.
Main Results:
- * XGB Regressor achieved high accuracy in predicting gold nanorod optical properties (MSE: 0.0041, MAE: 0.0190, R²: 0.994).
- * Tandem model successfully identified nanorod dimensions for targeted optical responses (MSE: 0.000116, MAE: 0.004476, R²: 0.9521).
- * Demonstrated a robust integrated forward-inverse methodology for nanoparticle design.
Conclusions:
- * Machine learning offers an efficient and accurate approach for exploring nanoparticle optical properties, surpassing traditional methods.
- * The developed methodology facilitates the design of nanoparticles with specific optical characteristics for advanced applications.
- * Findings pave the way for utilizing tailored nanoparticles in fields like photothermal therapy, bioimaging, and solar cells.
