Related Experiment Video
Updated: Aug 19, 2025
07:11
ARL Spectral Fitting as an Application to Augment Spectral Data via Franck-Condon Lineshape Analysis and Color Analysis
Published on: August 19, 2021
2.6K
Neural network and decision tree-based machine learning tools to analyse the anion-responsive behaviours of emissive
Anik Sahoo1, Sohini Bhattacharya1, Subhamoy Jana2
1Department of Chemistry, Inorganic Chemistry Section, Jadavpur University, Kolkata 700032, India. sbaitalik@hotmail.com.
Dalton Transactions (Cambridge, England : 2003)
|December 5, 2022
Summary
Machine learning models, including decision trees, effectively analyzed anion-responsive ruthenium(II) complexes. These models accurately predict experimental data, reducing the need for extensive sensing experiments.
Area of Science:
- Coordination Chemistry
- Supramolecular Chemistry
- Machine Learning
Background:
- Ruthenium(II) complexes with H2pbbzim and substituted terpyridine ligands exhibit anion-responsive properties.
- The secondary coordination sphere contains imidazole NH moieties, enhancing anion sensing capabilities.
Purpose of the Study:
- To analyze the anion-responsive behaviors of two heteroleptic Ru(II) complexes using machine learning.
- To mimic logic gate operations (YES-NOT, set-reset flip-flop) using the complexes' responses.
- To develop predictive models for anion sensing experiments.
Main Methods:
- Synthesis and characterization of heteroleptic Ru(II) complexes.
- Spectroscopic (absorption, emission), electrochemical, and spectroelectrochemical measurements.
- Implementation of machine learning algorithms: Artificial Neural Networks (ANNs), Adaptive Neuro-Fuzzy Inference System (ANFIS), and Decision Tree (DT) regression.
Main Results:
- Complexes showed significant changes in spectral and electrochemical properties upon anion binding.
- Reversible responses were observed with acid treatment.
- Machine learning models successfully analyzed and predicted experimental data.
- Decision Tree regression demonstrated excellent accuracy with minimal error.
Conclusions:
- The studied Ru(II) complexes can function as components in molecular logic gates.
- Machine learning tools, particularly DT regression, provide a robust and efficient method for analyzing anion-responsive complex data.
- ML-based models offer a prospective approach to reduce experimental time and cost in sensing studies.
