After decades of intense research, surprises in the realm of semiconductors – materials used in microchips to control electrical currents – are few and far between. But with a pair of published papers, materials engineers at Stanford University debut a promising approach to using a well-studied semiconductor to improve infrared light-emitting diodes and sensors. They say the approach could lead to smaller, sleeker, and less expensive infrared technologies for environmental, medical, and industrial uses.
“We taught an old dog new tricks,” said senior author Kunal Mukherjee, an assistant professor of materials science and engineering at the Stanford School of Engineering , putting the work’s importance in perspective. “The so-called IV-VI materials we’re working with – lead selenide and lead tin selenide – are more than a hundred years old. They are among the oldest semiconductors historically recorded. We found a way to integrate them with modern technology to produce a new type of infrared diode and to control the infrared light in important ways.”
The new diode emits infrared light in a desirable range of longer wavelengths (4000-5000 nanometers) good for sensing gas in the air (think greenhouse gases in the sky) or in medical settings (think carbon dioxide meters).
The unexpected benefit of the new approach is that the resulting integrated devices are “defect tolerant.” That is, these materials work even if not built with absolute precision, which is nearly impossible to achieve in these nanoscale crystals. Such freedom could bring down the cost of new devices significantly. Likewise, because these semiconductors are well studied, they can potentially be manufactured on some existing chip-making infrastructure and wouldn’t require expensive retooling of fabrication facilities to produce at commercial scale.
The challenge, engineering-wise, is to combine the resurrected materials with modern technology. All semiconductors are formed of crystal lattices. The layers of these various materials must bond to one another electronically with no adhesives allowed. This is no easy task.
Together, the two papers represent five years of painstaking research using a technique – known as molecular beam epitaxy – in which the complex crystals are built atom-by-atom, layer-by-layer.
“It took all those years to figure out how to grow these materials properly, one layer of atoms at a time,” co-author Jarod Meyer, a former graduate student in Mukherjee’s lab, explained. “And that meant keeping a special piece of equipment like the molecular beam epitaxy running the whole time – including some 2 a.m. sprints to the lab because of power outages.”
The first study , published in the journal Advanced Optical Materials and co-led by Meyer and former postdoctoral researcher Leland Nordin, describes the integration technique and fabrication flow used to marry these old semiconductor materials with other mainstream crystals like gallium arsenide to create a simple and efficient infrared platform. They showed the diodes were surprisingly bright emitters despite having billions of defects, called dislocations, per square centimeter. Modern semiconductors can rarely tolerate these defect levels.
The second paper, appearing in the journal Nano Letters and led by graduate student Pooja Reddy, describes a clever method to manipulate the crystal structure, enabling modulation and control of infrared light through small-but-precise temperature adjustments that cause the material to change between two well-ordered crystal structures.
“Most research in these structural change-type materials creates shifts between disordered and ordered states of the crystal,” Mukherjee said, highlighting the paper’s primary scientific accomplishment. “To go between two ordered states while staying mated to gallium arsenide is actually the hard part and the selling point of the research.”
The shift in structure alters how the light travels through the crystal, transitioning from transparent to opaque to turn the light on and off or to control its intensity, or to induce changes to its phase and polarization.
Most technological development in LEDs has focused on visible light. By contrast, infrared (IR) applications – invisible to the human eye – were slower to develop and, consequently, today’s infrared technologies tend to be bulky, expensive, and inelegantly designed. Mukherjee hopes to change all that and foresees a new generation of IR devices that are modern, cost-competitive, and easily manufacturable.
The ability to integrate these updated materials with other semiconductors and the ability to induce physical changes in the crystals can now be harnessed by engineers to create novel infrared devices in a range of wavelengths out to nearly 10,000 nm. Mukherjee projected that these could be used for environmental monitoring to detect gas leaks, in industrial and medical processes requiring precise IR sensors, or in new devices for non-invasive temperature measurement, among other potential avenues.
“Infrared has been sort of overlooked in the LED space historically because we can’t see the light, but its contributions could be important,” Mukherjee said. “We hope these developments might lead to a new era of infrared advances.”
Contributing authors on the Advanced Optical Materials are from the University of Central Florida. Contributing authors to the Nano Letters paper include Arturas Vailionis, senior research scientist at Stanford Nano Shared Facilities , and researchers from the University of Washington.
This work was funded by the National Science Foundation and the Army Research Office.
Media contact
Jill Wu, School of Engineering: jillwu@stanford.edu
Advanced Optical Materials
Mid-Infrared LEDs Based on Lattice-Mismatched Hybrid IV–VI/III–V Heterojunctions
16-Feb-2026