High-Throughput Materials Discovery for Metal Halide Perovskite Light Emitters

In the past decade alone, thin-film metal-halide perovskite (MHP) solar cell power conversion efficiency (PCE) has increased from 3.8% to 25.5%. MHPs have small exciton binding energy, balanced electron, and hole mobilities and long exciton diffusion lengths. The high PCE achieved in solar cells has encouraged research into MHPs for use in LEDs and they have demonstrated high photoluminescence quantum yields (PLQY) and narrow peaks. This project will use an HTE approach and ML models to understand the origins of previous low emission efficiencies and large emission linewidths to inform research towards bright narrowband emitters. Single crystal fabrication benchtop methos require ~5 days to grow a single crystal and another day to characterize them. This means that a full-time experimental (FTE) researcher can only explore a few materials and conduct 5-10 experiments on each material over the course of a week. We have been able to accelerate this to up to 5000 experiments/week across 20 materials. An autonomous PL measuring setup was built to gather spectral emission in high-throughput to screen MHP consisting of two perpendicular motors, an excitation source (375 nm diode laser) and a detector (Ocean Optics 2000+ Spectrometer) in the configuration. The motors are programmed to move the substrate in a snake-like pattern, allowing the laser to excite each drop while the detector measures the spectral PL response of the crystal in the drop. The resulting spectra has been compiled into a dataset with the experimental parameters used to grow each crystal. This data was compiled into a dataset and used to train ML models to identify groups of materials that may share desired properties such as brightness, peak position and linewidth. 17 new materials were discovered using this method and published for further exploration by the perovskite community (ACS Cent. Sci. 2022, 8, 5, 571–580).

Machine-Learning Assisted Hole Transport Material Discovery for Quantum Dot LEDs

Thin-film LEDs, such as quantum dot LEDs (QD LEDs), require additional transport layers to aid in charge injection and extraction from the electrodes into the active material. The typical device architecture for a thin-film LED involves a pair of electrodes, a hole and electron transfer layer (HTLs and ETLs) and a light-emitting active layer sandwiched between the two. The transport layers inject charge carriers from the electrodes into the active material where they recombine radiatively and emit light. These transport layers are typically organic transport layers tend to be chemically unstable, along with being sensitive to moisture, temperature, and other atmospheric onditions. Inorganic alternatives to HTL materials are needed to produce the next generation of stable and bright QD LEDs with high emission efficiency.

This project uses an existing database of inorganic materials, synthesized, and characterized by the National Renewable Energy Laboratory (NREL) along with supplementary data from other databases to train ML models on characteristics integral to HTLs, such as conductivity, hole mobility and bandgaps, to predict high-performing, stable candidates.

\[\\[1in]\] \[\\[1in]\]

High-Throughput Discovery for Stable Blue Organic Emitters

OLEDs were first developed in the late 1980s by Tang et al and enjoyed a period of popularity in the decades following. In recent years, OLEDs have since taken a back seat to emerging technologies such as PLEDs and QDLEDs but they remain the dominant display technology for hand-held and wearable electronics. The field of OLEDs still requires research into more affordable fabrication and stable blue light emitters. This project uses a high-throughput screening approach to synthesize and characterize new organic light emitters that emit stable, blue light.