Why mid-infrared light is worth controlling
The mid-infrared band of the electromagnetic spectrum occupies a privileged position in modern sensing. Molecules absorb and interact with light at these wavelengths in highly specific ways, producing spectral fingerprints that reveal what a substance actually is. That property underpins mid-infrared spectroscopy, a technique researchers rely on to identify biological materials, detect drugs and flag pollutants in the environment.
Sensing is not the only use. Mid-IR light can also carry information through free space, allowing data to travel through the air without cables or optical fibers. In both roles the practical payoff is the same: the tighter the control researchers have over mid-IR light, the more sensitive their detectors become and the faster their communication links can run. Improving that control has therefore become a central engineering goal.
The problem with most metasurfaces: they are frozen at fabrication
Metasurfaces have emerged as one of the most promising answers to that control problem. These are ultrathin structures patterned at the nanoscale, engineered so that their geometry shapes and steers light waves rather than relying on bulky conventional optics.
But there is a catch that has kept many of them out of real-world photonic systems. Most metasurfaces demonstrated so far are static: once the nanostructure is fabricated, its optical properties are locked in place. The device does what it was built to do and nothing else. A tunable optical component, one whose response can be changed on demand without rebuilding the chip, has been the missing piece.
Suspended crystalline silicon changes the equation
A research group led by Hatice Altug at the Bionanophotonic Systems Laboratory (BIOS) in EPFL's School of Engineering has now addressed that bottleneck. The team built metasurfaces on suspended membranes of crystalline silicon and tunes them by inducing mobile electrical charges inside the silicon itself. The result is a device whose response to light can be altered in real time without any modification to its physical structure.
The choice of material and architecture sits at the center of the advance. "By using crystalline silicon with the nanostructured layer suspended in air, we were able to greatly reduce optical losses that typically limit conventional mid-infrared platforms," Altug said.
Suspending the nanostructured layer instead of leaving it on a solid substrate avoids absorption and leakage pathways that drain performance in comparable mid-IR designs. The work has been published in Nature Communications.
Two routes to tuning, without touching the nanostructure
As BIOS senior scientist Ivan Sinev explained, the team demonstrated two distinct ways of tuning the metasurfaces while leaving their nanostructure untouched. One approach uses integrated electrodes as microheaters, applying localized thermal energy to shift how the structure behaves optically. The other relies on the charge-based principle the group emphasizes in describing the work: moving electrical carriers within the silicon to modify the material's optical response.
That distinction matters because it separates heat and charge as independent levers from the geometry of the device itself. Because the nanoscale pattern keeps doing its job either way, researchers are not forced to redesign and refabricate a new structure every time they need a different optical behavior.
From milliseconds to billionths of a second
The speed gain is one of the headline results. The EPFL devices tune mid-IR light on timescales measured in billionths of a second, a dramatic shift from the millisecond-scale switching typical of earlier tunable approaches.
For a photonic system, that difference is not a marginal refinement. Reconfiguring an optical element in milliseconds effectively rules it out of any application demanding rapid modulation, such as high-speed data encoding or measurement of fast-moving and transient samples.
Switching in billionths of a second opens both doors. A detector front-end could be reconfigured between measurements at a pace that keeps up with the signal it is analyzing. A free-space optical link could imprint information by modulating mid-IR light quickly enough to support far higher data rates.
Record optical performance and wafer-scale fabrication
The devices also post what the researchers describe as record optical performance, achieving more than an order of magnitude improvement in key metrics compared with earlier mid-IR platforms built from similar materials. That combination of tunability and sharply reduced loss is what makes the platform competitive rather than merely interesting.
Fabrication appears to scale as well. The team produced wafer-scale membranes consisting of 1 cm² chips, each carrying a 7×7 array of metasurfaces. Packing many structures onto a single chip points toward systems in which different regions of one device handle different optical tasks, or in which arrays are tuned in parallel.
What comes next
The near-term applications the researchers point to run along two tracks. On the sensing side, better mid-IR control translates into more sensitive detectors and more capable spectroscopy, useful for identifying biological samples, screening for drugs and monitoring pollutants. On the communications side, it supports faster free-space optical links that transmit data through the air. Both stand to benefit from a component that is ultrathin, low-loss and reconfigurable on command.
The larger significance may be architectural. For years the fixed nature of metasurfaces has confined them largely to static optical functions, while tunable alternatives depended on materials and mechanisms that sacrificed performance. By pairing suspended crystalline silicon with charge-based tuning, the EPFL group has shown that reconfigurability need not come at the cost of optical quality. If the approach scales as the wafer-level demonstration suggests, tunable mid-IR metasurfaces could move from laboratory curiosity to a standard building block in photonic systems designed for sensing and communication alike.
This article is based on reporting by Phys.org. Read the original article.
Originally published on phys.org
