Abstract:Accurate knowledge of electronic molecular properties of excited states is fundamental for understanding the behavior of functional materials for organic electronics and sensors. In this work, we focus on determining the properties of the most intense peak in the electronic absorption spectra of organic molecules. For this purpose, we employed the quantum chemistry QM-symex dataset, which has approximately 173,000 organic molecules and time-dependent DFT (TD-DFT) data of the first ten electronic absorption transitions. Each one is identified by its Cartesian coordinates. From data in the original QM-symex, we built a new dataset named QM-symex-modif that contains molecules in Simplified Molecular Input Line Entry System (SMILES) format and properties related to the main electronic transition. We then employed twenty machine learning (ML) algorithms to investigate oscillator strengths, excitation energies, transition orbitals, and the highest occupied molecular orbitals (HOMOs). As inputs for the ML algorithms, we used several chemical descriptors for each molecule generated in the RDKit tool employing the corresponding SMILES format. The generated input descriptors significantly improved the accuracy of the ML predictions for these key photophysical properties. Very good mean absolute errors (MAEs) were obtained for the test set composed of 45,056 molecules, namely, an MAE of 0.035 for oscillator strengths, 0.09 eV for excitation energies, 1.24 and 0.62 for the initial and final transition molecular orbital (MO) numbers (i.e., for each molecule, their position in the MO listing) respectively, and 0.014 for HOMO numbers, with R² values consistently exceeding 0.94, thus demonstrating the accuracy of the models. Additionally, a Shapley additive explanation (SHAP) analysis was carried out to evaluate the importance of the input parameters for the investigated ML models. We found several interesting relationships involving the input parameters. In particular, molecular weight holds significant importance in our ML models for determining the target HOMO numbers and the transition orbitals.

Can Deep Learning Search for Exceptional Chiroptical Properties? The Halogenated [6]Helicene Case

Can Deep Learning Search for Exceptional Chiroptical Properties? The Halogenated [6]Helicene Case

Deep Learning for Optoelectronic Properties of Organic Semiconductors

Fingerprint-based Deep Neural Networks Can Model Thermodynamic and Optical Properties of Eumelanin DHI Dimers

Self-Learning Perfect Optical Chirality Via a Deep Neural Network

Deep peak property learning for efficient chiral molecules ECD spectra prediction

An artificial intelligence approach for tackling conformational energy uncertainties in chiroptical spectroscopies

Unraveling the amplified chiroptical responses of rylene-bladed quintuple [6]helicenes

Molecule Identification with Rotational Spectroscopy and Probabilistic Deep Learning

A General Method to Predict Optical Rotations of Chiral Molecules from Their Structures

Deep Learning Optical Spectroscopy Based on Experimental Database: Potential Applications to Molecular Design

Deep learning for circular dichroism of nanohole arrays

Deciphering Nonlinear Optical Properties in Functionalized Hexaphyrins via Explainable Machine Learning

Prediction of Human Liver Microsome Clearance with Chirality-Focused Graph Neural Networks

Flexible design of chiroptical response of planar chiral metamaterials using deep learning

Machine Learning-based Analysis of Electronic Properties as Predictors of Anticholinesterase Activity in Chalcone Derivatives

High-Throughput Virtual Screening of Organic Second-Order Nonlinear Optical Chromophores within the Donor-π-bridge-Acceptor Framework

Machine learning prediction of the most intense peak of the absorption spectra of organic molecules

Designing of halogenated and traditional donor–acceptors composites for high-performance nonlinear optical applications

Molecular fingerprint and machine learning to accelerate design of high‐performance homochiral metal–organic frameworks

Accelerated Discovery of Two-Dimensional Optoelectronic Octahedral Oxyhalides via High-Throughput Ab Initio Calculations and Machine Learning