Key Points:
- Metasurface definitions vary, but typically they are defined as a diffractive element comprising individual scattering elements that are spaced apart at subwavelength scale
- Metasurfaces are not new devices, they’ve existed for decades. In recent years, however, there has been a revitalization of the field, primarily due to advances in compute and nanofabrication paired with public and private investments that have followed several key results in the early 2000’s and 2010’s
- From a physical perspective, you can consider metasurfaces as arrays of coupled antennas, an engineered surface current distribution, or a distribution of isolated phase shifters. How you define a metasurface will largely hinge on your choice of model and approximations
Introduction
What is a metasurface? Depending on who you ask, you may get different answers. For the most part, answers will entail a few key descriptions such as “subwavelength”, “diffractive”, “induces abrupt change in properties of light”, or offers “arbitrary manipulation of incident radiation”. It’s important to provide some nuance to these descriptions so that one can understand what these elements truly are, and when and how they are actually useful for research and engineering problems.
A Very Brief History
Within the photonics community, the consensus appears to be that metasurfaces are thin (i.e., on the order of a wavelength thick, if using a dielectric material) optical elements that comprise arrays of patterned nanostructures that are spaced apart in a subwavelength manner. The subwavelength spacing of the nanostructures is by far the single most distinguishing and critical property of a metasurface. Metasurfaces themselves are not “new” by any reasonable definition. In the optical domain, these devices go back several decades, they were just never commonly referred to as metasurfaces until the 2010’s. At lower frequencies of the electromagnetic spectrum, they stretch back even further.
Over the last decade or so, there has been a significant revitalization in optical metasurface research, building off seminal works in the 1990’s and early 2000’s on subwavelength gratings by various research groups. Furthermore, this research on subwavelength blazed gratings built upon decades of earlier research on diffractive optical elements and holography, as ultimately almost all research does.
One may debate the reasons for the revitalization in the 2010’s, but one key difference between the 1990’s and the 2010’s is that both compute and nanofabrication became substantially cheaper and more accessible to academic facilities than at any time in history. This has fueled substantial public and private investment into metasurface R&D that funneled into academia as well as industry, facilitating advancements in design and nanofabrication enabling commercialization of metasurface technologies in various markets.
What is a metasurface?
Back to the question though, what is a metasurface actually? A common definition is that a metasurface is a type of diffractive optic that has the requirement that its constituent elements are spaced apart on a subwavelength grid. These constituent elements may be described as “scatterers”, “nanoantennas”, “meta-atoms”, or something else, but these are all just variants on the same idea, which is essentially a subwavelength chunk of material into which incident light couples, its properties are modified (i.e., polarization, amplitude, phase, and/or wavevector), and the light then outcouples into free space.
Through material selection and subsequent engineering of the geometry and orientation of the constituent elements, the effect(s) of the individual elements can be altered (i.e., shape, and material properties, determine function). The underlying physical mechanisms that give rise to these effects can differ, and there are various formalisms (e.g., coupled mode theory, Mie theory, and others) for describing these mechanisms, but the general idea and design principles still hold; select your material and determine your geometry in order to induce the desired changes you want on an incident wavefront. In this manner, metasurfaces are not really any different from any other type of optic, refractive, diffractive, or otherwise.
This glosses over a lot of detail of course, and there are a number of complications (and potential benefits) associated with metasurfaces that are worth mentioning. For one, while metasurface elements are sometimes designed as periodic arrays to simplify modeling, practical implementations are rarely periodic. More often, the designs are aperiodic or gradient in nature, with spatially varying elements across the surface to achieve the desired optical function. Modeling these structures is not always straightforward; full-wave simulations can sometimes be used, but they often provide little intuition. Analytical or semi-analytical models exist, but their accuracy and applicability vary depending on the operating regime and the nature of the structure.
Model Selection and Practical Relevance
This raises the question: do you really need to understand how metasurfaces work in detail? Does having intuition actually matter? In many routine engineering contexts, the answer is no, not really. If the design process is largely simulation-driven, and if the simulations are accurate and the fabrication is repeatable, then the details of the underlying physics might not need to be deeply understood for successful deployment. But in more advanced or unconventional applications, or when pushing the limits of performance, then yes, having physical intuition becomes absolutely critical. Understanding the coupling mechanisms, interference effects, and field distributions can reveal why a given design is underperforming, how to improve it, or how to extend the concept to new domains.
At one level of abstraction, metasurfaces can be thought of as arrays of isolated antennas that impart abrupt, localized transformations on an incident optical field that modify its amplitude, phase, or polarization in a spatially controlled way. At another level, the surface as a whole can be treated as a continuous distribution of induced surface currents, which are governed by the incident field and the geometry/materials of the structure. This formulation is often more general and captures effects such as mutual coupling between elements and complex field interactions, but it also introduces significant modeling complexity.
Summary
At the end of the day, what you have physically is a patterned slab of material. The oscillating fields of incoming radiation induce polarization in the structure, the orientation of which and to what extent depends on the material’s electromagnetic properties and the geometry of the nanostructures. Describing the surface as a collection of discrete “posts” or “scatterers” is a useful abstraction, but it is still just that, an abstraction. The elements are typically not truly independent, and their behavior can depend on the fields from neighboring elements. Like many models in physics, metasurface models are inherently phenomenological, as are Maxwell’s equations themselves. But that doesn’t mean they aren’t useful. In fact, their power comes precisely from how well they allow us to predict and engineer useful behavior.
So, in summary, a metasurface is a thin, subwavelength-patterned structure that enables precise control over light through engineered scattering at the nanoscale. The requirement of subwavelength spacing is what sets metasurfaces apart from more conventional diffractive optics, and it is this property that enables their compactness, versatility, and multifunctionality. While the field has a long and rich history, the revitalization of metasurfaces over the past couple decades has been driven by advances in design tools, nanofabrication technologies, and application demand. Ultimately, how you define a metasurface will largely depend on your choice of model, whether it’s isolated scatterers, surface currents, or something in between.
Additional References
Lalanne, Philippe, et al. “Blazed binary subwavelength gratings with efficiencies larger than those of conventional échelette gratings.” Optics letters 23.14 (1998): 1081-1083.
Lin, Dianmin, et al. “Dielectric gradient metasurface optical elements.” science 345.6194 (2014): 298-302.
Bomzon, Ze’ev, Vladimir Kleiner, and Erez Hasman. “Pancharatnam–Berry phase in space-variant polarization-state manipulations with subwavelength gratings.” Optics letters 26.18 (2001): 1424-1426.
Chang-Hasnain, Connie J., and Weijian Yang. “High-contrast gratings for integrated optoelectronics.” Advances in Optics and Photonics 4.3 (2012): 379-440.
Arbabi, Amir, et al. “Subwavelength-thick lenses with high numerical apertures and large efficiency based on high-contrast transmitarrays.” Nature communications 6.1 (2015): 7069.
Smith, David R., John B. Pendry, and Mike CK Wiltshire. “Metamaterials and negative refractive index.” science 305.5685 (2004): 788-792.
Yu, Nanfang, and Federico Capasso. “Flat optics with designer metasurfaces.” Nature materials 13.2 (2014): 139-150.omething in between.