Logistic regression exploration of host-dopant compatibility in MR-TADF-based OLEDs

Area of Science:

• Materials Science
• Organic Electronics
• Photophysics

Background:

• Selecting optimal host materials for organic light-emitting diodes (OLEDs) is crucial for device performance.
• Multiple resonance thermally activated delayed fluorescence (MR-TADF) materials offer unique emissive properties.
• Current selection processes are time-consuming and resource-intensive.

Purpose of the Study:

• To accelerate the selection of optimal host materials for MR-TADF OLEDs.
• To establish predictive models for OLED performance based on host-dopant interactions.
• To identify key correlations between material properties and device metrics.

Main Methods:

• Fabrication of OLED devices using a specific MR-TADF dopant (DtCzB-mDS) and 14 common host materials.
• Application of machine learning, specifically logistic regression, to analyze experimental data.
• Derivation of empirical formulas linking host-dopant systems to device performance metrics.

Main Results:

• Empirical formulas were derived to predict OLED performance.
• A strong negative correlation was found between the highest occupied molecular orbital (HOMO) difference and maximum external quantum efficiency (EQEmax) (R=-0.67).
• A significant negative correlation was observed between HOMO difference and luminance (R=-0.54).

Conclusions:

• The study provides valuable insights into optimizing host material selection for MR-TADF OLEDs.
• Machine learning models can effectively predict and guide the selection of high-performance host materials.
• Understanding HOMO energy level alignment is critical for enhancing OLED efficiency and luminance.