25‐4: Methodology and Correlation of AI‐Based Design for OLED Materials

Abstract

AI based design for OLED materials are being tried in a variety of ways. An exemplary system is being developed to predict optical characteristics through machine learning (ML) with existing data. Once the performance descriptor is well defined and the quantum chemical calculation method is established, AI‐reverse design is expected to be possible. However, not all OLED emitting materials are equally capable of it. Different approaches are needed because the luminescence mechanism and its complexity of calculation are different depending on the material types. For pure fluorescence or even high efficiency phosphorescence, their luminescence mechanisms are relatively well defined and nearly irreversible and so the correlation between the calculation and performance could be better. If so, the reverse design is becoming possible and already it has begun to be tried a lot. However, in the case of TADF, the radiation and non‐radiation paths vary, ISC‐RISC is more reversible, and the controversy over luminescence mechanism remains. As a result, the calculating method of luminous efficiency has not yet been fully established. In this study, we want to report the consistency level of predicting characteristics of OLED materials using AI, and also discuss the difference between each emitting material types for reverse design. In particular, we also want to share the issues of calculating methods for TADF performance.

85‐3: Machine Learning Strategy Towards Inverse Design of Blue TADF Emitter: Training Excited State Properties Based on Density Functional Theory Calculations

“…GMPool's adaptive clustering capability improves the representation of hierarchical structures in bipolar single molecules that constitutes with donors, acceptors, and linkers. This enriched representation from DMPNN, facilitated by GMPool, serves as a robust input for the Variational Autoencoder (VAE), where the actual regression tasks are performed to identify valid molecules based on their OLED luminescent properties, such as k r , Δ𝐸 𝑆𝑇 , 𝜆, 𝑘 𝑅𝐼𝑆𝐶 , 𝑘 𝐼𝑆𝐶 , and PLQY [5]. Though our initial predictive accuracy was suboptimal, we managed to enhance it in each cycle through the implementation of active learning.…”

“…Here, we utilized graph neural network (GNN) to predict the characteristic intrinsic materials properties of TADF such as excited state energy levels and their transition properties, and charge transfer (CT) characters that is responsible for the narrow emission spectra. For the training of the GNN model DFT calculation results obtained from the TADF molecular assemblies out of basic fragments are used [5]. The validity of constructed GNN model is evaluated with experimental observations and improved iteratively.…”

“…Starting from conceptual development over the past decades it has continued to prosper in practical terms [3]. Along the way, computational workflows based on machine learning models rooted from deep neural networks for the inverse design of TADF materials is being attempted in various ways [4,5]. Traditionally, fragment-based molecular generation and highthroughput screening strategy has been normally utilized at a firstprinciples electronic structure level.…”

P‐104: Graph‐Based AI Workflow for OLED Materials Discovery

“…AI algorithms, in particular, Neural Networks (NNs) have a kind of universality for approximating a wide variety of interesting functions [8], and the inference of NNs is often extraordinarily fast. Recent attempts on applying AI to accelerate OLED materials screening have shown the great potential of AI in materials research [9,10,11,12].…”