Electrical engineers at Duke University report they have built a record-setting pyroelectric photodetector that uses trapped light in a plasmonic nanogap to capture a signal in just 125 picoseconds while operating at room temperature and drawing no external power. The achievement, described in Advanced Functional Materials and highlighted in Duke and Phys.org summaries, represents a leap of several orders of magnitude over previously reported thermal detector speeds and raises a pointed question: whether lower-cost, heat-based sensors can now compete with semiconductor detectors in high-speed applications. According to a summary of the work, the detector converts incoming light to heat and then to voltage so quickly that its response rivals the gigahertz bandwidths used in modern communications links.
From Microseconds to Picoseconds in One Design Shift
Pyroelectric detectors belong to a class of thermal photodetectors that generate an electrical signal when incoming light heats a thermally responsive material, triggering polarization changes inside the film. The appeal is simplicity: no cooling system, no bias voltage, and sensitivity across a wide spectrum. The tradeoff has always been speed. Conventional pyroelectric cameras needed bulky absorber layers to collect enough heat, and those thick films took microseconds to respond. A prior benchmark for on-chip-filter thermal cameras stood at 337 microseconds, a pace far too slow for telecommunications or real-time spectroscopy, and well outside the bandwidth needed to track ultrafast laser pulses or high-speed data streams.
The Duke team’s strategy was to make the absorber so thin that heat has almost nowhere to linger. By shrinking the active pyroelectric layer and pairing it with a plasmonic metasurface that concentrates incoming photons into an extremely small volume, the researchers eliminated the thermal lag that had hobbled every previous design. The result is a detector that operates at 2.8 GHz, equivalent to a response time of 125 picoseconds. That is roughly 2,700 times faster than the team’s own earlier experimental signals of approximately 700 picoseconds reported in a 2019 Nature Materials paper, and millions of times faster than the microsecond-class devices that defined the field just a few years ago. A separate Duke engineering brief notes that this speed record pushes pyroelectric technology into a performance regime previously thought to be the exclusive domain of semiconductor photodiodes and avalanche detectors.
Silver Nanocubes and the Physics of Trapped Light
The device architecture relies on silver nanocubes suspended approximately 10 nanometers above a gold film, forming what physicists call a nanogap cavity. When light enters the gap, it excites surface plasmons, collective oscillations of electrons at the metal interface, that trap and amplify the electromagnetic field in a space far smaller than the wavelength of the light itself. This plasmonic light trapping was first explored by the same group using film-coupled colloidal nanoantennas over a decade ago and has since been refined into a reliable fabrication method. The nanogap geometry forces nearly all incoming photon energy to convert to heat inside the ultrathin pyroelectric film sandwiched between the metals, rather than spreading across a large volume, which dramatically reduces the thermal mass the device must heat and cool during each detection event.
To verify the detector’s speed, researcher Eunso Shin devised a measurement setup using two distributed feedback lasers that could probe the device’s temporal response with high precision. The test confirmed a detection event in 125 picoseconds at room temperature with no external power supply, a combination the Duke team and accompanying coverage describe as unmatched for thermal photodetectors. Senior author Maiken Mikkelsen and Shin have pointed to the thinness of the pyroelectric layer as the core reason prior devices were slow: thicker materials stored heat longer, delaying the electrical response. By coupling the ultrathin film to a metasurface that delivers energy with extreme spatial precision, the team turned a traditionally sluggish sensor class into one that rivals the gigahertz speeds of cooled semiconductor detectors, while preserving the broadband and uncooled operation that make pyroelectric devices attractive in the first place.
Why Speed Alone Does Not Tell the Whole Story
Speed records attract attention, but the practical significance of this detector hinges on three additional traits. First, it can operate across a wide optical range rather than a narrow band, because the heat-based detection mechanism is inherently broadband and can, in principle, respond to wavelengths that deposit energy in the absorber. Second, it needs no cryogenic cooling, which means it could eventually be integrated into portable devices, wearables, or low-power sensor networks without the size and energy penalties of refrigerated infrared cameras. Third, the zero-power requirement means the detector generates its own signal voltage from absorbed heat, a feature that could simplify circuit design in battery-constrained systems and enable passive sensing nodes that wake only when illuminated.
These qualities together suggest a path toward affordable, compact thermal imaging at speeds that were previously the exclusive territory of expensive photonic detectors. Consider medical diagnostics: current wearable health monitors track heart rate and blood oxygen with optical LEDs, but many wearable thermal-sensing approaches are limited in how quickly they can track rapid temperature fluctuations. A pyroelectric detector operating at gigahertz frequencies could, in principle, help researchers study very fast thermal transients, though translating a nanoscale lab prototype into physiological measurements would require substantial additional validation and system-level engineering. Outside medicine, the same architecture could support ultrafast mid-infrared spectroscopy for chemical sensing, free-space optical communications that encode data in thermal signatures, or security cameras capable of tracking rapid motion in low-light conditions. That said, the Duke device is still a laboratory demonstration. The gap between a record-setting nanoscale prototype and a mass-produced chip is wide, and the team has not yet published data on fabrication yield, long-term stability, or integration with standard CMOS electronics, all of which will determine whether this concept can leave the lab.
A Decade of Incremental Gains, Then a Jump
The March 2026 result did not appear from nowhere. It sits at the end of a research arc that stretches back more than ten years, during which Mikkelsen’s group has repeatedly demonstrated how carefully engineered metal–dielectric–metal stacks can manipulate light and heat at the nanoscale. Early efforts focused on tuning the color and direction of reflected light using arrays of metallic nanostructures, work that laid the groundwork for later metasurfaces that could concentrate optical fields into subwavelength volumes. Those studies showed that by adjusting the size, spacing, and environment of nanoparticles, it was possible to sculpt the local density of optical states and dramatically increase absorption in targeted regions of a device.
Building on that foundation, the group began to explore how the same plasmonic principles could enhance photodetection, first by boosting the efficiency of conventional semiconductor photodiodes and then by coupling ultrathin thermal materials to resonant metallic cavities. The 2019 demonstration of on-chip multispectral imaging with light-trapping nanocubes provided a clear proof of concept that nanogap cavities could harvest and localize optical energy with high efficiency, albeit with microsecond response times. The new detector effectively takes that architecture and strips away the excess thermal mass, transforming a clever light concentrator into a platform for extreme-speed thermal sensing. As the latest measurements confirm, this combination of ultrathin pyroelectric films and nanogap plasmonics does more than set a record: it reframes what thermal detectors can be used for.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
