Related Experiment Video
Updated: Jan 12, 2026
09:11
Synthesis of Near-Infrared Emitting Gold Nanoclusters for Biological Applications
Published on: March 22, 2020
8.4K
Predicting the Optical Properties of Gold Nanoclusters Using Machine Learning Approach
Geraldine Sánchez-Dueñez1, Wladimiro Diaz-Villanueva2, Jorge Escorihuela1,3
1Institut de Ciència Molecular (ICMol), Universitat de València, C/Catedrático José Beltrán 2, 46980 Paterna, Spain.
ACS Omega
|November 3, 2025
Summary
Machine learning accurately predicts gold nanocluster (AuNC) optical properties. A GXBoost model, trained on over 200 articles, minimizes prediction errors, aiding in designing functional nanomaterials.
Area of Science:
- Nanoscience and Nanotechnology
- Computational Chemistry
- Materials Science
Background:
- Gold nanoclusters (AuNCs) exhibit optical properties sensitive to synthesis conditions, ligand type, and solvent.
- Optimizing AuNC synthesis and functionality requires advanced predictive tools.
Purpose of the Study:
- To develop a machine learning model for predicting the maximum emission wavelength of AuNCs.
- To explore critical variables influencing AuNC optical properties.
- To demonstrate the utility of machine learning in nanoscience material design.
Main Methods:
- Utilized a GXBoost algorithm trained on data from over 200 scientific articles.
- Employed One-Hot Encoding for data preparation.
- Validated the model using prediction vs. experimental and training/validation data.
- Performed independent regression for thiolated ligands other than GSH.
Main Results:
- Achieved low percentage errors (1.7%, 1.6%, 4.9%) for overall model validation.
- Obtained minimal training (0.01%) and test (3%) errors for independent regression with thiolated ligands.
- Identified key variables influencing AuNC optical characteristics.
Conclusions:
- Machine learning, specifically the GXBoost algorithm, is a powerful tool for predicting AuNC optical properties.
- This approach can significantly contribute to the efficient design and optimization of functional nanomaterials.
- The study highlights the potential of AI in accelerating materials discovery in nanoscience.
