Mobile phone lenses are becoming increasingly thin, and autonomous driving... “ eyes ” (The LiDAR) no longer spins around, AI Computing becomes both faster and more energy-efficient. —— These seemingly unrelated advances all conceal a revolution in photonics. “ Micro Revolution ” One of the main characters is a metasurface, which resembles a nanoscale… “ Dimming Magician ” , which can break the size and functional limitations of traditional silicon photonics chips; the other key player is nanoimprint lithography, which is equivalent to “ Magic Mass-Production Machine ” , bringing this cutting-edge magic from the laboratory into our everyday lives.

I. Silicon Photonics: The “Old Infrastructure” of the Optoelectronics Field

Let me first clarify the core concept: silicon photonics ( Silicon Photonics ), simply put, it is “ Using light instead of electricity ” technology of —— Using silicon as “ Stage ” (Core platform): Research is focused on enabling photons to perform generation, transmission, manipulation, and detection—just as electrons do within computer chips—ultimately compressing what were once bulky optical devices into a tiny chip. “ Optical chip ” , and can even be used to manufacture smartphone chips and computers CPU Similarly, mass production at low cost.

The smartphone chips and computers we use every day CPU , all use CMOS In manufacturing, silicon photonics leverages this mature technology to capitalize on a key advantage of silicon. —— High refractive index (simply put, light propagates more tightly within silicon and is less likely to scatter—roughly comparable to that of ordinary glass). 2 more than double), plus Ultra-precise A specialized structure gives rise to the core components. It is widely employed in fields such as optical communications, sensors, photonic computing, and LiDAR. 。

Integrated photonics It was once highly anticipated as a solution to the chip shortage. Size and Features the bottleneck , but The balance between size and functionality has always been that elusive one. “The elephant in the room” 。 Traditional silicon-based photonic devices are approaching their physical size limits, Just as even the most effective legacy infrastructure can no longer keep pace with emerging demands, traditional silicon photonics chips… “ Trouble ” Increasingly obvious:

Size Impasse : Due to the inherent limitations of silicon materials and conventional optical design, grating couplers Wait The dimensions of core components are mostly on the micrometer scale; just as a regular road cannot be shrunk into a narrow alleyway, they cannot be integrated into smaller, more portable devices—such as miniature sensors, AR glasses);

Single regulation Traditional silicon photonics devices can only guide light propagation in a rudimentary way—much like being able to control the direction of cars on a highway, but unable to regulate their speed, color, or make. —— Without the ability to precisely control key optical properties—such as phase, amplitude, and polarization—at the nanoscale, it becomes inadequate when confronted with complex light-field manipulation requirements. “ Unable to meet the demands. ”。

And “metasurface ( Metasurfaces ’)’s intervention is akin to the “spatial folding” magic conjured on a two-dimensional plane.

II. Metasurfaces: Nanoscale “Light‑Tuning Magicians”

The emergence of metasurfaces fundamentally breaks the limitations imposed by the intrinsic properties of conventional materials. Simply put, a metasurface is an ultra-thin “nanoscale patterned film” whose thickness is exceedingly small. It is densely populated with countless miniature structures—what we call “meta-atoms.” These are not actual atoms but rather carefully engineered micro- and nano‑scale building blocks that, when arranged in an orderly fashion, can transcend the inherent optical characteristics of natural materials, enabling precise control over the direction, intensity, and even the waveform of each incoming light beam.

There is also the concept of “subwavelength”: simply put, the dimensions of metamaterial atoms are smaller than the wavelength of light—too small for the human eye to perceive, typically only tens to hundreds of nanometers. It is precisely this “miniature size” that enables metasurfaces to overcome the diffraction limits of light propagation, condensing all the functions of conventional three-dimensional optical systems onto a single two-dimensional thin film—akin to transforming a bulky DSLR lens into a mere thin sticker. This is the most fundamental distinction between metasurfaces and traditional optical components.

The integration of metasurfaces and silicon photonics chips is far from a simple “stick‑on‑a‑sticker” approach. Recent research has evolved four sophisticated integration schemes, delivering a synergistic effect that surpasses the sum of its parts and catering to diverse application scenarios:

(Image source: the internet; please contact us for removal)

Embedded (In-type) : Performing surgery directly within the waveguide core, enabling mode conversion to occur in situ during light propagation;

On-type “Apply a protective coating”: Without disrupting the existing “highway,” a metasurface thin film is affixed to the surface of the silicon photonic waveguide. By leveraging the evanescent field generated during light propagation and its interaction with the metasurface, complex functionalities can be realized.

Hybrid-type “Strong‑strong synergy” combines the advantages of both embedded and overlay architectures, integrates multiple dimming mechanisms, and strives for ultimate light‑coupling efficiency and control precision, making it well suited to demanding applications such as high‑precision sensors and advanced optical computing.

Off-type “The external auxiliary device” is akin to equipping a silicon photonics chip with an “external light‑shaping lens,” tasked with addressing the “last mile” challenge of managing the light emitted by the chip.

If traditional silicon photonics is like building a “conventional highway” on a flat surface—allowing light to propagate in only one direction—then the introduction of metasurfaces is akin to constructing an “optical interchange” on that same plane, enabling light to steer, accelerate or decelerate, and even change its “form,” with control that is more efficient, more flexible, and more compact.

III. Three Major Application Scenarios: Metasurfaces Take Center Stage

Metasurfaces combined with silicon photonics chips—this “miniaturized, multifunctional” pairing is no mere laboratory fantasy; it has already begun to deliver tangible results, driving transformative change in three key areas and turning what was once deemed “impossible” into “possible.”

1. The “ultimate form” of all-solid-state LiDAR

Lidar serves as the “eyes” of autonomous driving and robotic navigation, enabling the detection of the surrounding environment and the identification of obstacles. Traditional silicon‑based lidar relies on mechanical rotating components to scan its beam—much like a person constantly turning their head to look around—which is not only laborious but also limits the field of view. In contrast, metasurfaces, with their nanoscale light‑modulation capabilities, can achieve nearly 180° of “mechanism‑free scanning”: without any physical rotation, they allow the beam to cover a wide area, while precise control over the beam’s direction and spatial distribution can be attained by tuning the arrangement of metasurface atoms.

Not only has it shrunk the LiDAR’s footprint to the millimeter scale, but it has also boosted scanning speed and stability while reducing power consumption—earning it the title of the “ultimate form” of all‑solid‑state LiDAR. This makes the “eyes” of autonomous driving both more compact and more reliable, while paving the way for low‑cost deployment.

2. Consumer-grade spectral analysis

Spectrometers are “specialized tools” for analyzing the composition of materials, indispensable in applications such as food safety, water‑borne pollutant detection, and microplastic analysis. However, conventional spectrometers remain bulky and costly, typically confined to laboratories or industrial settings. In contrast, metasurfaces exhibit extreme sensitivity to light wavelengths—acting like a “high‑precision optical sieve.” By engineering their surface structures, they can selectively filter and separate specific wavelengths, enabling precise identification of material constituents.

Leveraging this advantage, researchers can condense sophisticated spectral analysis capabilities onto silicon photonic chips just a few millimeters in size—making spectrometers as thin and portable as smartphone lenses. In the future, by integrating these miniature spectrometers into smartphones and smartwatches, we will be able to monitor food safety and environmental quality anytime, bringing professional-grade testing within reach of every household.

3. AI + Optics-Based Computing Brain

As AI technologies advance, traditional electronic computing is increasingly struggling. Optical computing, by contrast, can be seen as equipping AI with a “light‑based brain,” leveraging its advantages in parallel processing and high‑speed transmission to emerge as the next frontier of computing. At the heart of optical computing lies the optical neural network, which must efficiently handle complex spatial light fields. Metasurfaces, meanwhile, can simultaneously process multiple optical signals, performing intricate linear computations with extremely low energy consumption.

For the industry, the biggest concern is: Can this thing be mass-produced?

IV. Tackling Mass Production: Nanoimprinting Enables the “Magic” of Batch Replication

No matter how remarkable the capabilities of metasurfaces may be, if they cannot be manufactured in large quantities, they will remain confined to the laboratory and fail to enter our everyday lives. The core challenge lies in the fact that the micro‑ and nanostructures of metasurfaces demand extremely high fabrication precision, yet conventional manufacturing techniques either lack sufficient accuracy or are too slow to simultaneously achieve both “precision” and “mass production.”

At present, mainstream metasurface fabrication techniques all have notable limitations:

Electron-beam lithography (EBL): “Highly precise but too slow” It boasts extremely high machining precision, capable of precisely patterning nanoscale meta‑atoms, and is also compatible with silicon‑based materials. However, it employs a “point‑by‑point direct‑write” process—resulting in exceedingly slow speeds and prohibitively high costs; even in a single day, only a handful of devices can be fabricated, making large‑scale production entirely unfeasible.

Deep Ultraviolet Lithography (DUV): “Fast in volume but not precise” It is a “mass-production workhorse” for chip fabrication—fast and low-cost—but its resolution falls short of the required limit. Like using a thick paintbrush, it cannot render features as fine as those at the atomic scale, thus failing to meet the processing demands of meta‑surfaces.

Just then, Nanoimprint lithography (NIL) “Standing out” has become the “optimal solution” for the large-scale fabrication of metasurfaces. The underlying principle is straightforward: much like stamping a pattern with an ink pad, we first fabricate a “template stamp” bearing the desired metasurface‑atom array, then press this stamp onto a photoresist layer. After curing and demolding, the metasurface can be replicated in bulk. This approach is not only cost‑effective and rapid but also delivers high precision, seamlessly resolving the longstanding trade‑off between “precision” and “mass production.”

(Image source: the internet; please contact us for removal) Schematic diagram of the nanoimprint lithography process

However, nanoimprinting also has a “small drawback”: the refractive index of commercially available imprint resins is too low—around 1.5, comparable to that of ordinary glass—which makes it difficult to meet the stringent optical path‑length requirements of metasurfaces and thereby limits their optical performance. To address this issue, researchers at Pohang University of Science and Technology in South Korea published a review article titled “Nanoimprint lithography for scalable manufacturing of optical metasurfaces” in the journal Optics and Photonics Research, proposing two ingenious approaches that both preserve the mass‑production advantages of nanoimprinting and enhance its optical performance:

Hybrid material strategy: “Apply a layer of high-refractive-index thin film” First, an embossing process is used to pattern the metasurface on the substrate. Then, atomic layer deposition (ALD) is employed to deposit a thin film onto the patterned surface—this film has a high refractive index and, like a magnifying glass, confines light within the high‑index region, thereby enhancing the metasurface’s light‑modulation capabilities. This approach is well suited for applications with stringent optical performance requirements.

(Image source: the internet; please contact us for removal) Hybrid Fabrication of Metasurfaces Using ALD Technology

Nanoparticle-Doped Resin Strategy: “Direct Upgrading of the Resin” Simpler and more direct: by directly incorporating high‑refractive‑index nanoparticles into imprintable resin, ordinary resin is transformed into “high‑refractive‑index resin,” which can then be used in a single step to fabricate metasurfaces via nanoimprinting. This approach eliminates the need for additional thin‑film deposition, streamlining the process, and allows flexible tuning of the metasurface’s operating wavelength by simply switching the type of nanoparticles. Even more impressively, it enables the fabrication of metasurfaces on curved substrates—such as AR‑glasses lenses—a capability that conventional lithography cannot achieve, thereby making it compatible with a wider range of portable devices.

(Image source: the internet; please contact us for removal) Fabrication of Metasurfaces via a Nanoparticle-Doped Resin Strategy

An optimized version of metasurfaces combined with nanoimprinting ushers in a “dimension‑upgrading era” for silicon photonics.

In short, metasurfaces are not merely a “small patch” on conventional silicon photonics; they represent a “dimension‑raising revolution.” By breaking the size and functional constraints of traditional optical components, they transform silicon photonic chips from “single‑function highways” into “multifunctional optical interchanges.” Meanwhile, the optically optimized nanoimprint lithography technology serves as the “mass‑production engine” of this revolution, enabling metasurfaces to move beyond the laboratory and become mass‑producible products, thus bridging the gap between research and industry.

From the cutting edge of scientific research to industrial applications, the convergence of metasurfaces and silicon photonics is reshaping the landscape of the optoelectronics industry. In the future, as these technologies continue to advance, thinner smartphone lenses, smarter LiDAR systems for autonomous driving, more efficient AI‑powered optical computers, and more portable diagnostic devices will all emerge from the synergistic integration of metasurfaces and nanoimprint lithography—quietly transforming our daily lives and ushering the entire optoelectronics sector into a new era.