Century-old semiconductors get a second life as the basis for cheaper, smaller infrared devices

Advanced Optical Materials and Nano Letters, 2026

Lead selenide and lead tin selenide are among the oldest semiconductors ever recorded, dating back more than a hundred years in the scientific literature. They have been well studied, well characterized, and largely set aside as modern semiconductor development moved in other directions. Now a team at Stanford University has brought them back.

In two papers, materials engineers led by assistant professor Kunal Mukherjee demonstrate how to integrate these vintage IV-VI semiconductor materials with modern chip infrastructure to produce a new type of infrared light-emitting diode and a method to control infrared light through temperature-induced structural changes. The work could lead to smaller, cheaper, and more practical infrared technologies for environmental monitoring, medical sensing, and industrial applications.

Bright despite billions of flaws

The first study, published in Advanced Optical Materials and co-led by former graduate student Jarod Meyer and former postdoctoral researcher Leland Nordin, describes the integration technique. The team used molecular beam epitaxy, building the crystal structures atom by atom, layer by layer, to marry lead selenide with gallium arsenide, a mainstream semiconductor material.

The resulting diodes emit infrared light at wavelengths between 4,000 and 5,000 nanometers, a range useful for sensing gases in the atmosphere, including greenhouse gases, and in medical settings, such as carbon dioxide monitoring.

The unexpected benefit was defect tolerance. The integrated materials contained billions of dislocations, structural defects, per square centimeter. Modern semiconductors typically cannot tolerate that level of imperfection. But these devices were surprisingly bright emitters despite the defects. This tolerance could significantly reduce manufacturing costs, because building defect-free nanoscale crystals is extremely difficult and expensive.

Controlling light by shifting crystal structure

The second paper, published in Nano Letters and led by graduate student Pooja Reddy, describes a method to modulate and control infrared light. By making small, precise temperature adjustments, the researchers induced the material to transition between two well-ordered crystal structures.

This structural shift changes how light travels through the crystal. The material transitions from transparent to opaque, enabling the light to be turned on and off or its intensity controlled. The shift also affects the light's phase and polarization, properties important for advanced sensing applications.

Mukherjee highlighted the scientific significance: most research in structural phase-change materials creates shifts between disordered and ordered states. Transitioning between two ordered states while maintaining the bond with gallium arsenide is the technically challenging and commercially relevant achievement.

Why infrared has lagged behind

Most LED development has focused on visible light. Infrared applications, invisible to the human eye, developed more slowly. As a result, current infrared technologies tend to be bulky, expensive, and inelegant compared to their visible-light counterparts.

The ability to integrate these updated materials with mainstream semiconductors and to induce controllable structural changes in the crystals opens a path toward modern, cost-competitive infrared devices operating at wavelengths out to nearly 10,000 nanometers. Potential applications include environmental monitoring to detect gas leaks, industrial and medical processes requiring precise infrared sensors, and non-invasive temperature measurement.

Because the underlying semiconductor materials are well studied, they could potentially be manufactured on existing chip-making infrastructure without expensive retooling of fabrication facilities.

What remains to be demonstrated

The papers represent five years of painstaking research. Building the crystal layers atom by atom using molecular beam epitaxy is a slow, technically demanding process that included late-night emergency runs to the lab during power outages. Scaling this technique to commercial production will require demonstrating that it can be done reliably and economically.

The defect tolerance is promising but needs to be validated over the full operating lifetime of real devices. Long-term reliability data does not yet exist. And while the wavelength range of the new LEDs is well suited to gas sensing, competing technologies in that spectral range already exist. The new devices will need to demonstrate clear advantages in cost, size, or performance to gain traction.

This work was funded by the National Science Foundation and the Army Research Office.