Perovskite LED Efficiency Reaches 37% Through Liquid-Phase Chemical Tuning

Researchers led by Emory University have boosted the quantum efficiency of light-emitting diodes made with metal-halide perovskite to a peak of 37%, advancing efforts to commercialize low-cost, flexible optoelectronics. The findings, published in the journal Nature and reported by TechXplore (https://techxplore.com/news/2026-10-chemical-fine-tuning-perovskite-efficiency.html), detail how fine-tuning chemical reactions during the liquid precursor stage overcomes previous efficiency ceilings.

Most previous research into perovskite LEDs reported quantum efficiency—the rate at which a device converts electrical current into light—in the 20% to 30% range. While inorganic LEDs made from semiconductors like gallium nitride are highly durable and efficient, their rigid structure makes them difficult to manufacture as flexible sheets or dense micro-pixel arrays. Conversely, organic LEDs provide flexibility and vibrant colors but remain expensive and less durable. Perovskite hybrids combine organic and inorganic materials, offering an alternative for low-cost, flexible devices operating in the near-infrared spectrum.

The multi-institutional team discovered that precursor solution processing dramatically affected the final crystalline material. During fabrication inside nitrogen-filled glove boxes, the researchers observed unexplained efficiency fluctuations in test batches. Subsequent testing showed that heating the precursor solution for exactly 20 minutes produced an optimal reaction pathway.

This heating duration promoted the synthesis of a linear molecular compound that enhances light conversion efficiency while suppressing a ring-shaped compound that degrades it. By purifying the linear compound and controlling the reaction pathway, the team achieved the 37% efficiency mark. Senior author Letian Dou, a chemistry professor at Emory who holds 20 patents including five related to LEDs, noted that the discovery moves the technology closer to commercial viability.

Co-authors on the paper include first author Sung-Doo Baek, co-first author Yu-Ting Yang, Xiyu Luo, and Hyuntae Choi of Emory; Libai Huang, Pengfei Wu, Hanjun Yang, Wenhao Shao, and Jeong Hui Kim of Purdue University; Mohammad Ali and Jay Guo of the University of Michigan, Ann Arbor; and Hui Taou Kok and Barry Rand of Princeton University.

Near-infrared LEDs are utilized in non-visible light applications, including smartphone facial recognition sensors, night-vision cameras, and noninvasive biomedical tissue imaging. The researchers plan to focus subsequent work on improving device operating lifetimes and testing applications with the Emory School of Medicine.