🔬 optics

Spatiotemporal Metasurface for Ultrafast All-Optical Wavefront Shaping

This paper introduces and experimentally demonstrates a fully all-optical, ultrafast, and reversible strategy for dynamic wavefront shaping in a semiconductor nanowire metasurface, where a spatially inhomogeneous femtosecond pump pulse induces transient permittivity modulation to reshape a probe pulse's wavefront with sub-picosecond switching speeds.

Original authors: Mert Akturk, Giulia Crotti, Andrea Schirato, Katsuya Tanaka, Isabelle Staude, Thomas Pertsch, Giulio Cerullo, Giuseppe Della Valle, Margherita Maiuri

Published 2026-08-04

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Mert Akturk, Giulia Crotti, Andrea Schirato, Katsuya Tanaka, Isabelle Staude, Thomas Pertsch, Giulio Cerullo, Giuseppe Della Valle, Margherita Maiuri

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine light not just as a beam that turns things on and off, but as a sculptor's clay. For centuries, scientists have been trying to mold this invisible clay into perfect shapes—focusing it into a sharp point, bending it around corners, or splitting it into rainbows. To do this, they built "metasurfaces," which are like ultra-thin, high-tech carpets covered in microscopic pillars. These pillars are so small that they can trick light into behaving in wild ways. Usually, however, these carpets are static; once you weave the pattern, the magic is locked in forever. If you want to change the trick, you have to throw the whole carpet away and weave a new one.

But what if you could change the magic trick while the show is still running? What if you could use a second beam of light to temporarily rewrite the rules of the first beam, turning a flat piece of glass into a lens, a mirror, or a prism in the blink of an eye? This is the dream of "dynamic" optics. It's the difference between a printed map that never changes and a GPS screen that reroutes you instantly when traffic appears. The challenge has been doing this fast enough to keep up with the speed of light itself, and doing it without needing bulky wires or slow-moving motors.

In this paper, a team of researchers has taken a giant leap toward that dream. They didn't just tweak an existing lens; they created a system where light writes its own lens on a surface made of tiny silicon nanowires. Using a clever "pump-and-probe" technique, they fired a shaped laser pulse (the pump) to temporarily change the properties of the silicon, and then sent a second pulse (the probe) through the modified area. The result? The second pulse was instantly defocused, as if it had passed through a diverging lens that appeared out of nowhere and vanished in less than a trillionth of a second.

Here is how they pulled off this optical magic trick.

The Setup: A Carpet of Tiny Wires
The researchers built a surface covered in a one-dimensional array of hydrogenated amorphous silicon (a-Si:H) nanowires. Think of these wires as a row of tiny, identical tuning forks. Under normal conditions, they are all the same, so when light hits them, it passes through evenly without changing its shape. It's like a crowd of people standing in a straight line; if everyone moves at the same speed, the line stays straight.

The Trick: Writing a Lens with Light
To make the surface do something new, the team used a "pump" laser pulse that was shaped like a long, thin oval (strongly astigmatic). This pulse was focused so that it was very bright in the middle and faded out toward the edges, like a spotlight on a stage. When this bright pulse hit the nanowires, it excited the electrons inside them, creating a temporary "hot" state.

Because the pump light was brighter in the center, the electrons in the middle of the array got hotter and more excited than those on the edges. This created a temporary, invisible landscape across the surface: the middle of the array became optically "thinner" (light traveled through it faster) than the edges. In the world of optics, this difference in speed is exactly what creates a lens. Just as a glass lens is thicker in the middle to slow down light and bend it inward, this silicon surface became "thinner" in the middle to speed up light and bend it outward.

The Result: A Lens That Appears and Disappears
The team then fired a second "probe" pulse, tuned to a specific color (691 nanometers), through this temporary landscape. In the simulations and experiments, they saw the probe beam spread out, or "defocus," exactly as if it had passed through a diverging lens.

The most impressive part is the speed. This lens didn't just appear; it appeared and disappeared in less than one picosecond (that's 0.000000000001 seconds). The effect was so fast that the beam didn't just change shape; it was reshaped and then the surface returned to normal before the beam had even finished passing through.

Why It's Not Just a Simple Dimmer
You might think, "Well, if the light gets absorbed in the middle, the beam will just get dimmer in the center, right?" The researchers were very careful to prove this wasn't just a dimming effect. They chose a specific color for the probe light where the silicon nanowires are very sensitive to changes in phase (the timing of the light wave) but not very sensitive to changes in brightness (amplitude).

To confirm this, they ran a control experiment using a plain, flat sheet of silicon with no nanowires. When they hit the flat sheet with the same shaped laser, the light just got dimmer in the middle. It didn't spread out or change shape. This proved that the magic wasn't just the laser heating the silicon; it was the specific arrangement of the nanowires that turned that heat into a shape-shifting lens. The nanowires acted as the translator, converting a simple change in material properties into a complex change in the light's path.

The Future: On-Demand Optics
The paper suggests that this isn't just a one-way street. By tweaking the color of the probe light, they could theoretically flip the effect. Instead of making the middle "thinner" (defocusing), they could make it "thicker" (focusing), creating a converging lens in the same amount of time.

This work establishes a new way to think about optical devices. Instead of building a library of different lenses and swapping them out, we might one day have a single, flat surface that can become any lens we need, instantly, just by shining a specific pattern of light on it. It's a step toward "reconfigurable" flat optics, where the functionality of a device is written by light itself, erased, and rewritten a trillion times a second. While this specific experiment used a simple one-dimensional array to demonstrate defocusing, the principles shown here suggest that complex, two-dimensional lenses and other optical tricks could be controlled with similar speed and precision in the future.

Technical Summary: Spatiotemporal Metasurface for Ultrafast All-Optical Wavefront Shaping

Problem Statement
Optical metasurfaces have established themselves as powerful platforms for controlling light at subwavelength scales by engineering the geometry and spatial arrangement of constituent meta-atoms to impart specific phase gradients. However, a fundamental limitation of conventional metasurfaces is that their optical response is permanently fixed once fabricated. While dynamic metasurfaces have been developed using thermal, mechanical, electrical, or phase-change mechanisms, these approaches often suffer from slow modulation speeds or rely on tuning pre-existing functionalities (e.g., shifting focal length) rather than generating new spatial phase gradients on demand. All-optical control offers the potential for ultrafast operation (femtosecond to picosecond timescales) via photoinduced nonlinearities, but realizing a strategy where a uniform metasurface can transiently acquire a complex wavefront-shaping capability—such as a lens—remains a challenge.

Methodology
The authors propose and demonstrate a strategy for dynamic, all-optical wavefront shaping using a spatiotemporally modulated metasurface. The core concept involves using a spatially structured pump beam to transiently "write" a phase gradient onto a uniform metasurface, rather than relying on a permanently encoded geometry.

System Design: The experimental platform consists of a one-dimensional periodic array of hydrogenated amorphous silicon (a-Si:H) nanowires on a silica substrate. The unit cell parameters are $P = 400$ nm, $w = 180$ nm, and $h = 220$ nm.
Excitation Scheme: A strongly astigmatic femtosecond pump pulse ( $\lambda_{pump} = 400$ nm, $\sim 100$ fs duration) is used to photoexcite the metasurface non-uniformly along the periodic $x$ -direction. The pump has a Gaussian intensity profile with a full width at half maximum (FWHM) of $I_x = 45$ $\mu$ m, while remaining uniform along the longitudinal $y$ -direction ( $I_y = 4$ mm).
Mechanism: The non-uniform pump absorption generates a spatially inhomogeneous distribution of hot electron-hole pairs. This induces a transient, spatially varying permittivity modulation $\Delta\varepsilon(x, t)$ across the metasurface. The authors identify a specific probe wavelength ( $\lambda_1 = 691$ nm) located on the red wing of the metasurface's resonance where the differential transmission $\Delta T \approx 0$ (isosbestic point) while the differential phase $\Delta \phi$ is maximized. In this regime, the photoexcitation predominantly modulates the phase of the transmitted light with minimal amplitude loss.
Simulation and Modeling: The study employs a multi-step numerical approach:
1. Periodic Model: Full-wave simulations of an infinite array to map the relationship between material perturbation ( $\Delta\varepsilon$ ) and the resulting differential transmission ( $\Delta T$ ) and phase ( $\Delta \phi$ ).
2. Non-Equilibrium Dynamics: A Two-Temperature Model (TTM) is solved locally for different positions along the metasurface to calculate the spatiotemporal evolution of carrier density $N(t)$ and lattice temperature $\Theta(t)$ . This accounts for one- and two-photon absorption, Drude-like carrier modulation, and thermo-optic effects.
3. Finite-Aperture Simulation: The calculated spatiotemporal permittivity map $\Delta\varepsilon(x, t)$ is applied to a finite-sized model (300 $\mu$ m wide, $\sim 730$ meta-atoms) to simulate the propagation of a delayed probe beam and the resulting wavefront reshaping.
Experimental Setup: Ultrafast wide-field pump-probe microscopy is used. A delayed probe pulse ( $\lambda_{probe} \approx 690$ nm) interrogates the excited region, and the transmitted beam is imaged onto a CMOS camera. The system is slightly defocused to convert transient phase modulations into measurable spatial intensity redistributions.

Key Results

Transient Defocusing: Experimental measurements and numerical simulations demonstrate that the photoexcited metasurface acts as a transient diverging cylindrical lens. At a pump-probe delay of $t = 0.1$ ps, the transmitted probe beam broadens significantly along the $x$ -direction, with the FWHM increasing from $0.28$ mm (unperturbed) to $0.44$ mm.
Differential Signal: The differential intensity signal ( $\Delta I/I_0$ ) exhibits a characteristic double-peaked profile, indicative of intensity redistribution from the center to the lateral regions of the beam. This signature appears within the temporal overlap of the pump and probe and decays on a sub-picosecond timescale (disappearing within $\sim 600$ fs).
Mechanism Validation: The quantitative agreement between the experimental data and the full-wave simulations confirms that the observed effect arises from the resonant conversion of the photoinduced spatial permittivity profile into a transient phase profile. The negative phase shift (optically thinner medium) at the center of the pump spot creates the defocusing effect.
Control Experiment: Measurements on an unpatterned a-Si:H film of comparable thickness show only homogeneous attenuation (negative signal) without the double-peaked structure. This confirms that the resonant nanostructuring is essential for converting the permittivity landscape into a phase-driven wavefront shaping element.
Switching Speed: The switching time for the wavefront modulation is below one picosecond, limited by the carrier relaxation dynamics.

Significance and Claims
The paper establishes a general strategy for ultrafast, all-optically reconfigurable wavefront shaping based on photoinduced-gradient metasurfaces. The authors claim that this work bridges ultrafast all-optical modulation and flat-optical wavefront engineering by demonstrating that optical functionalities can be "imprinted on demand" via the spatiotemporal profile of excitation light, rather than being permanently defined during fabrication.

The significance lies in:

Dynamic Analogue: Providing a fully all-optical and ultrafast route to the dynamic analogue of passive wavefront shaping found in static gradient metasurfaces.
Speed: Achieving reconfiguration rates (switching times $< 1$ ps) that far exceed thermal, mechanical, or electrical tuning mechanisms.
Versatility: The approach suggests that by selecting different probe wavelengths (e.g., operating on the blue wing of the resonance to induce a positive phase shift), the same platform could support transient converging lenses (focusing) in addition to the demonstrated defocusing.
Future Outlook: The work paves the way for high-speed, light-driven flat optical components, such as all-optically reconfigurable metalenses, where the functionality is controlled by light on ultrafast timescales.

More like this