See discussions, stats, and author profiles for this publication at: [Link]

net/publication/359053779

Tutorial on metalenses for advanced flat optics: Design, fabrication, and

critical considerations

Article in Journal of Applied Physics · March 2022

DOI: 10.1063/5.0078804

CITATIONS READS

14 1,217

5 authors, including:

Chihun Lee Junsuk Rho

Pohang University of Science and Technology Pohang University of Science and Technology
13 PUBLICATIONS 110 CITATIONS 272 PUBLICATIONS 9,151 CITATIONS

SEE PROFILE SEE PROFILE

Younghwan Yang
Pohang University of Science and Technology
64 PUBLICATIONS 1,275 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Metasurfaces View project

All content following this page was uploaded by Chihun Lee on 07 March 2022.

The user has requested enhancement of the downloaded file.

Tutorial on metalenses for advanced flat
optics: Design, fabrication, and critical
considerations
Cite as: J. Appl. Phys. 131, 091101 (2022); [Link]
Submitted: 15 November 2021 • Accepted: 28 January 2022 • Published Online: 04 March 2022

Seong-Won Moon, Chihun Lee, Younghwan Yang, et al.

COLLECTIONS

Paper published as part of the special topic on Metasurfaces for Photonic Devices

This paper was selected as an Editor’s Pick

ARTICLES YOU MAY BE INTERESTED IN

The crane operator's trick and other shenanigans with a pendulum

American Journal of Physics 90, 169 (2022); [Link]

Condensation of 2D exciton-polaritons in an open-access microcavity

Journal of Applied Physics 131, 093101 (2022); [Link]

Hopf bifurcation in a nonlocal nonlinear transport equation stemming from stochastic neural
dynamics
Chaos: An Interdisciplinary Journal of Nonlinear Science 27, 021101 (2017); https://
[Link]/10.1063/1.4976510

J. Appl. Phys. 131, 091101 (2022); [Link] 131, 091101

© 2022 Author(s).
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

Tutorial on metalenses for advanced flat optics:

Design, fabrication, and critical considerations
Cite as: J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804
Submitted: 15 November 2021 · Accepted: 28 January 2022 · View Online Export Citation CrossMark
Published Online: 4 March 2022

Seong-Won Moon,1 Chihun Lee,1 Younghwan Yang,1 Joohoon Kim,1 Trevon Badloe,1 Chunghwan Jung,2
Gwanho Yoon, and Junsuk Rho1,2,4,a)
3

AFFILIATIONS
1
Department of Mechanical Engineering, Pohang University of Science and Technology (POSTECH), Pohang 37673,
Republic of Korea
2
Department of Chemical Engineering, Pohang University of Science and Technology (POSTECH), Pohang 37673,
Republic of Korea
3
Department of Manufacturing Systems and Design Engineering, Seoul National University of Science and Technology,
Seoul 01811, Republic of Korea
4
POSCO-POSTECH-RIST Convergence Research Center for Flat Optics and Metaphotonics, Pohang 37673, Republic of Korea

Note: This paper is part of the Special Topic on Metasurfaces for Photonic Devices.
a)
Author to whom correspondence should be addressed: jsrho@[Link]

ABSTRACT
Metalenses comprised of artificial subwavelength structures known as meta-atoms have demonstrated abilities beyond conventional bulky
optical components by modulating the phase, amplitude, and polarization of light in an ultrathin planar form factor. In this Tutorial, we
present the fundamental principles and practical design procedures to exploit the abilities of metalenses, including achromaticity, high
numerical aperture, and tunability. The fundamental principles include both plasmonic and dielectric meta-atoms, which require different
physics to describe their light–matter interactions. In the phase modulation section, we compare the methods of physically implementing
phase via meta-atoms including both the propagation and geometric phase methods. Next, we cover the recent progress of nanofabrication
procedures from the perspective of the metalenses using materials such as titanium dioxide, gallium nitride, and hydrogenated amorphous
silicon. We further compare the various fabrication methods with regard to the resolution, size, cost, and optical properties of fabricated
metalenses. Then, we describe the critical considerations of metalenses including aberration-correction, numerical aperture, and tunability
for advanced flat optics. Herein, we provide a practical guide for the design, fabrication, and critical considerations of metalenses with exam-
ples of research from early works to more recent developments.

Published under an exclusive license by AIP Publishing. [Link]

I. INTRODUCTION surface machining. Flat lenses such as holographic and diffractive

Optical lenses are one of the fundamental components used to lenses are thinner and lighter than those refractive lenses, and they
focus and manipulate light. Refractive lenses, typically used in com- generally work through the diffractive properties of light to create
mercial optical systems, have extensively contributed to the devel- constructive interference at the desired focal spot.1,2 However, they
opment of science and fields including imaging, display, biology, suffer from low focusing efficiencies due to the shadowing effect.
laser processing, quantum optics, and optical communication. The low resolution of fabrication methods limits the achievable
They are designed through the local thickness and curvature of numerical aperture (NA), which hinders diffraction-based flat
glass to focus the light depending on its position. However, the lenses from versatile applications.
thickness of a few centimeter scale and weight of the glass results Metalenses are flat lenses that have explosively developed over
in a heavy and bulky formfactor. Additionally, aspherical refractive the last decade. As a specific application of the more general field
lenses capable of correcting spherical aberration require complex of metasurfaces, metalenses are able to manipulate the properties of

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-1

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

light such as amplitude, phase, polarization, absorption, reflectance, 1. Plasmonic resonator

and orbital angular momentum via arrays of subwavelength struc- Plasmonics refers to the coherent oscillation of the electrons at
tures, known as meta-atoms.3–23 Compared to the typical the interface between a metal and a dielectric. When light interacts
state-of-the-art refractive lenses, metalenses not only focus the light with the metal particles, free surface electrons align with the electric
more closely to the ideal Airy disk, but also give rise to high NA,
field of the incident light, and surface electric dipoles are induced.
varifocal, and multi-functional lenses that are revolutionarily light
This electric dipole produces the resonance phenomenon, which is
and thin.24–33 These characteristics have provided scientists and determined by the material properties, geometric shape, and
engineers with new ways to create versatile optical systems includ- natural harmonic vibration frequency. By controlling the resonance
ing cameras, bioimaging, augmented/virtual reality, sensors, three- at the interfaces, the wavelength size optical resonators such as
dimensional particle trapping, optical fibers, quantum sources, and
metal nanoparticles and nanostructures can modulate the phase,
other novel devices.34–41
amplitude, polarization, and dispersion of visible light.
Here, we introduce the design and fabrication processes of the V-shaped nanoantenna, consisting of two equal-length nano-
metalenses in detail and then discuss their critical considerations rods as shown in Fig. 2(a), demonstrated the possibility of a plas-
compared with conventional lenses. The design principles of the monic antenna for the metalenses.50 Two different resonant modes
constituent subwavelength structures and the mechanisms of how
(symmetric and antisymmetric) are excited by changing the direc-
metalenses are designed and operated are described first. Section II
tion and polarization state of the incident light (parallel or perpen-
introduces the general fabrication techniques for metalenses, and dicular). If the incident light is at an angle between parallel and
furthermore, for the realization of large-scale, cost-effective metal- perpendicular, both modes are simultaneously excited. The combi-
enses. Then, we move on to discuss the critical considerations of nation of the two modes and polarization states can cover the full
the metalenses, including aberration-correction, high NA, and tun-
0–2π phase space. However, due to the metal–dielectric interaction,
ability separated as external stimuli. Finally, we summarize the con-
the metal antenna must be thin to excite the electric dipole. As a
tents of this Tutorial. result, such electric dipole-based plasmonic metalenses achieve a
fairly low 25% theoretical limited efficiency on the cross-polarized
II. DESIGN PROCESS OF METALENSES component of the reflected light, even when ignoring absorption.
Designing metalenses for applicable optical devices such as To increase the efficiency of the reflective plasmonic metasur-
planar lenses,42,43 vortex converters,44 and polarimetry45 requires face, gap plasmon metasurfaces using metal–insulator–metal build-
several steps including calculating the required phase map and ing blocks were introduced.51 The thin dielectric layer induces a
designing the meta-atoms needed to physically implement it. A strong near-field coupling effect. The gap-surface plasmon, which
brief schematic of the metalens design process is shown in Fig. 1. denotes the strong magnetic field that comes from the insulator, can
The procedure consists of three steps: meta-atom design, phase achieve a high reflective efficiency. By changing the design parame-
profile calculation, and meta-atom to phase matching. The ters of the unit structure, different radiation phases can be achieved.
meta-atom design describes the building of a library of information Meanwhile, the amplitudes from similar structures are not much
about the optical properties depending on the materials and geome- different. Therefore, after determining the unit structure, only phase
tries of meta-atoms, which can cover full phase modulation with a modulation is required. As the metallic ground plane hinders the
suitable efficiency. Meanwhile, the required phase profiles are calcu- transmission mode, an H-shaped meta-atom was used to demon-
lated depending on applications such as beam focusing and steering strate 80% anomalous reflection efficiency.52 Unfortunately, metals
based on lens equations. Using both the meta-atom library and have inherent Ohmic losses in the optical frequency range, limiting
phase profile, metalenses are designed by implementing a form of their use for visible metalenses. Therefore, the majority of studies
phase manipulation. The most widely used implementing methods have utilized a high refractive index and low loss dielectric materials,
are the propagation and geometric phases. The detailed mechanisms as discussed in the next section.
of these methods will be discussed in the following chapters.
2. Dielectric resonator
A. Design of meta-atoms: Materials perspective In contrast to metallic modes determined by the plasmonic
Metalenses utilize light–matter interactions that are determined resonances, dielectric particles are able to strongly confine the inci-
by the structure (material properties and geometric shape) and inci- dent electromagnetic waves and excite both electric and magnetic
dent light (phase, wavelength, and polarization). Therefore, theoreti- resonant modes that can be decomposed as multipole partial
cally describing these interactions is essential for the intuitive design waves. The excited modes depend on the shape and material of the
of the meta-atoms. Depending on the type of material, meta-atoms meta-atom, as shown in Fig. 2(b).53 For dielectric meta-atoms to
can be broadly divided into either metallic structures, that utilize cover the full 2π phase modulation, multiple values of thickness
plasmonic resonances,46 or high-index dielectric structures that fun- and refractive index are required.
damentally rely on confining and scattering light through subwave- The mechanism behind the perfect reflection or transmission
length structures.47,48 The majority of studies about high-efficiency in dielectric materials can be described by the Kerker effect.54 In
metalenses use dielectric-based lenses, but studies on reflective or 1983, Kerker et al. demonstrated that back or forward scattering is
high-efficiency plasmonic metalesnses49 are still in progress. In the zero when the effective permittivity ε and permeability μ are com-
following section, we focus on the physical phenomena that occur parable. When the amplitude of the electric and magnetic multi-
when the incident light passes through the metalens. poles are equal and in phase, constructive interference occurs in

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-2

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

FIG. 1. The general procedure for designing metalenses. Metallic meta-atoms induce plasmonic responses, while high-index dielectric materials strongly confine the light
to provide both electric and magnetic multipole modes. By changing the materials and geometry, we can build a meta-atom library that can cover the full-phase modulation
(0–2π) with suitable efficiency. Phase profiles, that describe the required phase at each spatial location, are calculated from the lens equation for beam focusing. After gath-
ering a library of meta-atoms and their optical properties, combined with the required phase profile, a metalens can be achieved. Propagation and geometric phase are
common ways of realizing the desired phase profiles. The propagation phase relies on the fact that the geometric shape of meta-atoms interacts with light to provide differ-
ent phase retardations. The geometric phase, which is commonly known as the Pancharatnam-Berry phase (PB phase), uses a single meta-atom design with the phase
defined by the in-plane rotation angle. The figure is reprinted with permission from Genevet et al., Optica 4, 139 (2017). Copyright 2017 Optica Publishing Group; Aieta
et al., Nano Lett. 12, 4932 (2012). Copyright 2012 American Chemical Society.

the forward direction, and in the backward direction when the properties are interpreted as subwavelength waveguides for guiding
multipoles are completely out of phase.55 With these induced mag- the light in modes. Waveguide type meta-atoms have various
netic resonances, dielectric metalenses can achieve optical scatter- shapes such as rectangular, cylinder, coaxial cylinder, and so on,
ing without magnetic materials.56 and the effective index of the waveguide is related to shapes.57–60
The arrival time of light passing through a meta-atom as mode
delays along the propagation direction than other meta-atoms,
3. Dielectric waveguide
which leads to a local phase shift. Therefore, the effective index of
Resonators described in Secs. II A 1 and II A 2 control optical mode in meta-atoms and height of the meta-atoms is proportional
properties via resonance in meta-atoms. To fully control amplitude to phase shift. Phase shift Δw is expressed by61
and phase shift, strong resonances are required. However, such res-
onances generally occur in narrow bandwidths, leading to low 2π
transmittance and difficulty of phase modulation over a broadband Δw ¼ Δneff h, (1)
λ
range. Therefore, meta-atoms without resonance working in the
broadband range are required for advanced applications such as where λ is the wavelength of incident light and Δneff is the effective
full-color imaging; non-resonance meta-atoms acting as waveguide index difference between two meta-atoms; h is the height of
elements not only operate in the broadband regime but also show meta-atoms. The refractive index of background material (usually
high transmittance and full phase modulation. Their optical air) and material that make up the meta-atoms determine the

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-3

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

FIG. 2. Design of plasmonic and dielectric meta-atoms. (a) (i) The V-shaped antenna can handle two resonant modes (symmetric mode and antisymmetric) depending on
the angle of rotation, the angle between the two rods, and the angle of the incident light. (ii) Finite-difference time-domain (FDTD) results of the phase produced by each
antenna showing the full 2π phase shift. (iii) SEM image of the fabricated gold antenna on a silicon wafer. Reprinted with permission from Yu et al., Science 334, 333
(2011). Copyright 2011 The American Association for the Advancement of Science. (b) (i) Schematic of a silicon nanobeam array. (ii) The phase retardation of transverse
magnetic (TM) and transverse electric (TE) field of incident light. Each line denotes different beam widths; 100 nm (blue), 120 nm (red), 140 nm (green). Red squares indi-
cate experimental measurements of the 120 nm nanobeam, while the inset shows an SEM image. (iii) The phase wavefronts of the 120 nm Si nanobeams under 550 nm
illumination. (iv) Absorption spectra of the 120 width nanobeams under TE (green) and TM (blue) field incidence. Inset shows the magnetic and electric field profiles for
TE and TM incidence at a wavelength of 600 nm. Both electric and magnetic resonance were involved. Reprinted with permission from Lin et al., Science 335, 298 (2014).
Copyright 2014 The American Association for the Advancement of Science.

range of Δneff. To obtain phase shift Δw ¼ 2π, the height of B. Design of meta-atoms: Computational methods
meta-atoms is expressed by The shape of individual meta-atoms must be determined con-
sidering phase modulation or transmission efficiency. Analytical
λ solutions on the scattered light from arbitrary structure arrays do
h¼ : (2)
Δneff not exist except for the case of spherical particles; so numerical
simulations are required to calculate the reflection and transmission
To modulate the full phase of light, the decision to adjust coefficients for meta-atoms. In the case of propagation phase-based
the height and meta-atoms design should be considered with metasurfaces, we have to optimize the meta-atoms to achieve the
operating wavelength. For example, in cylinder-shaped wave- desired phase modulation while simultaneously maintaining high
guide meta-atoms composed of silicon, the phase shift occurs transmittance. On the other hand, PB phase-based metasurfaces
depending on the height and radius of the cylinder that deter- require the design of a single meta-atom with high polarization-
mines the effective index. The effective index can vary from conversion efficiency because the full-phase delay is achieved
neff  nair ¼ 1 to neff  nSi ¼ 3:57 at λ = 1 μm, so the maximum through the rotation of meta-atoms. A set of meta-atom designs is
of Δneff is 2.57. To obtain phase shift Δw ¼ 2π in this case, the necessary for the propagation phase, whereas just one meta-atom
height of meta-atoms should be over 389 nm given by Eq. (2). Note design is enough for the PB phase.
that since the effective index of waveguide type meta-atoms is gen- Rigorous-coupled wave analysis (RCWA) is a simulation tool
erally smaller than the materials of which it is made, the corre- widely used for designing meta-atoms. RCWA is more computa-
sponding height is the lower limit and should be designed higher. tionally efficient to obtain the optical properties of meta-atoms

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-4

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

compared to full-wave simulation methods including the finite- meta-atoms. In addition to the beam focusing using Eq. (3), various
difference time-domain (FDTD) and finite element methods studies have been conducted to design phase maps that can produce
(FEM) because it quickly provides semi-analytical solutions about beam steering,66,67 and aberration-corrected spectrometers36 by adjust-
electromagnetic fields propagating through periodic structures ing the lens equation.
through the sums of spatial harmonics in Fourier-space, based on
Floquet’s theorem.62 It is well known that the effectiveness of the D. Principles of phase modulation
RCWA is usually limited to the calculation of periodic structures;
After obtaining the proper phase map for the specific purpose, a
however, the periodic approximation is pretty valid in metalens
method of implementing the phase through meta-atoms is required.
because the target phase profile usually varies slowly compared to the
Typically, the phase distribution maps require full phase control from 0
dimension between any two neighboring structures; therefore, a peri-
to 2π and are discretized into a periodic array as the phase at each
odic condition is available to calculate the electromagnetic response of
spatial location must be fulfilled with a physical meta-atom. Design
meta-atoms. In special cases such as waveguide type meta-atoms,
mechanisms depend on several factors, including the material type
since meta-atoms composed of high refractive index materials have
(metal or dielectric), material properties (refractive index), and available
weak interactions with neighboring structures, they behave almost
fabrication methods. In this Tutorial, we mainly handle the periodic
independently, and thus, contributions due to periodic conditions are
structure-based metalenses that discretize the required phase map.
negligible.61,63 Therefore, the solutions of RCWA simulations can be
Although this is the representative way, we have to assume that each
used to quickly and accurately evaluate the performance of the
meta-atom has negligible interaction even if we design with periodic
meta-atoms. This assumption is undertaken not only in the RCWA
boundary conditions. The propagation phase and PB phase are typical
calculations but also in other methods such as the FDTD and FEM.
methods used to modulate the desired phase shift. Hybrid methods
Recently, an open-source RCWA software called MAXIM has
(combining both propagation and PB phase) and non-meta-atom-
been released to facilitate the easy computation of the optical prop-
based methods such as inverse design are also described.
erties of meta-atoms.64 MAXIM is incorporated with an intuitive
graphical user interface and does not require any complicated
installation process. The program allows the simulation of various 1. Propagation phase
kinds of structure arrays including multilayered thin films, gratings, The degree of freedom of meta-atoms depends on their basic
rectangular pillars, cylindrical pillars, circular cones, rectangular structure. For example, the width, length, and height of the rectan-
horns, and combinations of them all. These structures are widely gular meta-atoms can be adjusted, whereas the long and short axis
used in designing metalenses and general metasurfaces. The calcu- diameters can be changed for ellipsoid meta-atoms. By changing
lation accuracy of MAXIM is also validated by comparing simula- the design parameters, a library of phase delay and transmission
tion results with the commercial FDTD software. intensity can be calculated and stored. If the database can cover the
full phase shift (0–2π) using transverse electric and transverse mag-
C. Calculation of the phase profile netic modes, it is possible to design by selecting proper meta-atoms
Metalenses induce spatially discontinuous phase changes at according to the pre-calculated phase map [Fig. 3(a)].
the interface of two media. By applying the generalized Snell’s law, Arbabi et al. demonstrated that propagation phase methods
reflected and refracted light can be redirected arbitrarily by adjust- with ellipsoid meta-atoms can achieve complete control of the 2π
ing the structure and position of meta-atoms designed with specific phase with high efficiency for arbitrarily polarized incident light.68
optical properties. Based on this principle, to design practically For this, the general relation between input and output electric
applicable metalenses, obtaining the required phase delay map that fields can be expressed using the symmetric and unitary Jones
quantitatively describes the phase at each spatial location is matrix69 as shown in Eq. (4). It means that the metalens can cover
required. In general, the methods of obtaining the phase map the complete phase and polarization state if each pixel can make a
depend on the application of the lens such as for planar lenses and unitary and symmetric Jones matrix,
wave plates. For example, the lens equation, Eq. (3), is mainly used
for beam focusing or steering for planar lenses,65 Et ¼ TEi ,
 
ω pffiffiffiffiffiffiffiffiffiffiffiffiffiffi  Txx Txy
T¼ ,
w(r, ω) ¼  r2 þ f 2  f , (3) Tyx Tyy (4)
c
" # " #
Ext Exi
where w is the phase profile of the metalens that can ensure the focus- Et ¼ Ei ¼
t
ing within a diffraction-limited spot at the focal length f, r is the radial Ey Eyi
coordinate, and c is the light speed. Note that this equation is only
valid for normally incident spherical light of angular frequency ω. The where T is the transmission matrix, Ei is the incident light, Et is the
phase map obtained through the lens equation can be wrapped transmitted light; Txx, Txy, Tyx, and Tyy are complex transmittance
between 0 and 2π, and implemented as actual metalenses through the coefficients; Ext and Eyt are x- and y-polarized components of trans-
meta-atom library (Secs. II A and II B) and the phase modulation mitted lights, respectively; Exi and Eyi are x- and y-polarized compo-
method (Sec. II D). The phase is generally discretized (with more nents of the incident light, respectively. Coefficients are changed
steps producing less noise) to satisfy the conditions to reconstruct the depending on the type of meta-atoms and the detailed description
desired optical wave to obtain the final target phase map with distinct can be found in Ref. 43.

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-5

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

FIG. 3. The mechanism of propagation phase and PB phase methods. (a) Schematic of the propagation phase. Each geometry of the meta-atom has a different amplitude
(transmission or reflection) and phase retardation. As an example, cylindrical atoms with different heights and diameters are presented. To satisfy the desired phase map,
the meta-atoms need to cover the full phase range with high amplitudes to maximize efficiency. (b) Schematic of the PB phase. A single anisotropic structure with high
amplitudes is chosen, then by rotating the in-plane angle from 0 to π, the full phase coverage can be achieved for circularly polarized light. The amplitudes and phases
were calculated by rigorous coupled-wave analysis simulations.

Similarly, various metalenses have been designed based on the another technique, known as PB phase (or geometric phase), per-
methods of finding proper meta-atoms by resonance tuning. High forms full phase control by changing the in-plane orientation angle
NA and high-efficiency metalens has been designed with high- of identical meta-atoms. Consider the meta-atom under normal
contrast transmit arrays.70 There is a trade-off in that the focusing monochromatic plane wave incidence. As in Eq. (4), the relation-
efficiency is lowered as NA increases, but with a high refractive ship between the incident and transmitted E field can be expressed
index, materials reduce the coupling effect between meta-atoms, so in a Jones matrix form. If the meta-atom is rotated by an angle θ, it
the entire metalens can operate similarly to the designed phase. also can be described with Jones calculus72 as shown in Eq. (5).
Chen et al. developed broadband achromatic metalens that can Using the advanced transformation matrix,73 we can change the
operate at visible frequencies.71 The previous propagation metasur- linear polarized matrix to the circularly polarized transmission
face tried to find the proper meta-atoms from the pre-calculated matrix [Eq. (6)],
database to satisfy one desired phase map, whereas to design a
broadband achromatic metalens, the meta-atoms must simultane- T(θ) ¼ S1 (θ)TS(θ)
ously satisfy multiple phase maps considering both group delay   
and group delay dispersion over the desired bandwidth. Since the cos(θ) sin(θ) Txx Txy
¼
existing simple single structure meta-atom does not provide suffi- sin(θ) cos(θ) Tyx Tyy
 
cient room for design freedom, interleaved meta-atoms structures, cos(θ) sin(θ)
which contain multiple meta-atoms in each period, were utilized.  , (5)
sin(θ) cos(θ)

2. Pancharatnam–Berry phase  
While the propagation phase controls the phase modulation ^ ¼ p1ffiffiffi 1
Λ
1
; (6)
by adjusting the fill factor of meta-atoms (geometric shape), 2 j j

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-6

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

where Λ^ is the change of basis matrix from Cartesian to polar coor- polarization, they limit the design of more general
dinates. The circularly polarized transmission matrix is then given polarization-independent, broadband, and achromatic lenses. To
by Eq. (7), generalize the polarization of incident light, Mueller et al. com-
  bined the propagation and geometric phases by imparting a fully
^ 1 ^ Tþþ Tþ independent phase on each of any two orthogonal polarizations.74
T ¼ Λ T(θ)Λ ¼
circ
, (7)
Tþ T The meta-atom’s shape and orientation are independently deter-
1 mined through combining the propagation phase and PB phase
Tþþ (θ) ¼ ½Txx þ Tyy þ j(Txy  Tyx ), (8) methods as shown in Fig. 4(a). They define the required phase as
2
two terms; one is the angle-independent phase which can be imple-
1 mented through the propagation phase only, and the other is the
Tþ (θ) ¼ ½(Txx  Tyy )cos(2θ)  (Txy þ Tyx )sin(2θ) angle-dependent phase which dictates the PB phase contribution.
2
This combined method allows every meta-atom in the arbitrary
 j[(Txx  Tyy )sin(2θ) þ (Txy þ Tyx )cos(2θ)], (9)
spatial location to contribute to the beam manipulation.
So far, the majority of metalenses have been designed based
1 on matching each unit cell to the phase map from the lens equa-
Tþ (θ) ¼ ½(Txx  Tyy )cos(2θ)  (Txy þ Tyx )sin(2θ)
2 tion. However, this unit cell approach cannot achieve high NA and
þ j[(Txx  Tyy )sin(2θ) þ (Txy þ Tyx )cos(2θ)], (10) high efficiency at the same time.75 Large area broadband metal-
enses require complex and diverse meta-atoms that can cover the
1 group velocity and group velocity dispersion. By increasing the
T (θ) ¼ ½Txx þ Tyy  j(Txy  Tyx ), (11) periodicity of each atom, we can get a higher degree of design
2
freedom, but it necessarily involves higher-order modes and dif-
where the + and − signs denote the right- and left-circularly polar- fraction, which reduce the overall efficiency. Conversely, if the peri-
ized states, respectively. odicity is small, the design freedom is reduced. Therefore, recent
As shown in Eq. (7), the co-polarized elements [Eqs. (8) studies have been conducted to inverse design the total metalens as
and (11)] of the transmission matrix do not contain the angle a whole, rather than the meta-atoms,76,77 as shown in Fig. 4(b).
terms. This means that if we use meta-atoms with zero co-polarized This approach typically uses the geometric parameters at the input
transmission, we can adjust the total transmission matrix by only and focusing efficiency as the output to be optimized. The adjoint
manipulating θ. In other words, when the co-polarized terms method and statistical learning optimization have been applied as
become Txx = −Tyy, and Txy = Tyx we can obtain the new transmis- optimizers for the inverse design. As we have seen so far, there are
sion matrix [Eq. (12)], and transmitted E fields for each circular various design methods from the propagation phase, PB phase,
polarization direction [Eqs. (13) and (14)], hybrid methods to inverse design to create the various functionali-
  ties of metalenses such as achromatic, broadband, polarization-
0 ej2θ
Tcirc (θ) ¼ jTxy j2θ , (12) insensitive, scalable, and high efficiency. Still, there are limitations
e 0 of each design method, and there are many potentially exciting
avenues to be pursued.
Etþ ¼ jTxy Ei ej2θ , (13)
III. FABRICATION OF METALENSES
Et ¼ jTxy Eiþ e j2θ
: (14) A. General fabrication process
In this section, we will discuss fabrication methods of the met-
According to Eqs. (13) and (14), it is obvious that if we rotate
alenses from typical materials such as titanium dioxide (TiO2),
the meta-atom by θ, the resultant field changes by 2θ (−2θ).
gallium nitride (GaN), and hydrogenated amorphous silicon (a-Si:
Therefore, we can tune the full phase modulation from 0 to 2π by
H) to recently proposed ones including low-loss a-Si:H and nano-
rotating the meta-atom 0 to π [Fig. 3(b)]. When we design a meta-
particle embedded resin (nano-PER). As discussed, the meta-atom
surface based on the PB phase method, we need only one aniso-
of the metalenses should be subwavelength in scale and composed
tropic meta-atom with high transmission efficiency for circularly
of low-optical loss dielectrics with a high index of refraction.
polarized light. Therefore, PB phase-based metalenses have the
Typical complementary metal-oxide-semiconductor (CMOS) man-
advantage of a simple meta-atom design and fabrication. The ideal
ufacturing has been widely exploited to achieve sub-100 nm resolu-
case, when jTxy j ¼ jTyx j ¼ 1, 100% efficiency is possible by induc-
tion with low-loss dielectrics at a target wavelength. Atomic layer
ing the complete conversion of circularly polarized light. However,
deposition (ALD), metal organic chemical vapor deposition
due to Ohmic losses and coupling effects between the nanostruc-
(MOCVD), and plasma-enhanced chemical vapor deposition
tures, practical efficiencies are often less than 100%.
(PECVD) have been widely used for the fabrication of amorphous
titanium dioxide (a-TiO2), GaN, and a-Si:H thin films, respectively.
3. Combination of the propagation phase and However, conventional fabrication methods have some limitations
Pancharatnam–Berry phase such as small-area, low-throughput, and high-cost. To overcome
Since the propagation phase operates only for orthogonal and these limitations, various fabrication methods such as nanoimprint-
linear polarization, and the PB phase operates only for circular ing and deep ultraviolet photolithography are being developed to

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-7

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

FIG. 4. Schematics of other design methods. (a) Conceptual description of propagation phase, geometric phase, and hybrid (combined) methods. (i) linearly polarized
light fx, fy passes into meta-atoms with different dimensions and fixed orientation angle. When orthogonal linear inputs (red) are incident, the output states have the
same polarization state (green). (ii) All meta-atoms have the same dimensions but different orientation angles. The input circular polarized light (red) is converted to equal
and opposite magnitude circular polarized light (green). (iii) Meta-atoms have different shapes and orientation angles, i.e., combining the PB phase and propagation
phase. Arbitrary orthogonal states of polarization can be used. Reprinted with permission from Mueller et al., Phys. Rev. Lett. 118, 113901 (2017). Copyright 2017
American Physical Society. (b) Schematic and summary of an inverse design-based metalens. (i) The dashed line indicates a figure of merit for optical focusing. (ii)
Comparing the unit cell approach and inverse design method. The former uses incomplete, periodic-boundary meta-atoms; however, the latter uses non-periodic/unit-cell
approximation which can increase the design freedom. (iii) The result of the inverse design that depends on the NA using freeform (red) and const-z (blue). Constant z
means that the structural conditions limit the state along the z axis for a fabrication feasible design. Reprinted with permission from Chung et al., Opt. Express 28, 6945
(2020). Copyright 2020 Optica Publishing Group.

realize large-area metalens with high throughput and low cost. We The minimum deposition thickness of a-TiO2 is half of the
will specifically introduce the fabrication steps depending on the maximum width w among the designed meta-atoms. To prevent
composition of the metalenses and briefly introduce some typical the collapse of the exposed photoresist patterns, the chamber tem-
experimental results. perature should be lower than the glass transition temperature of
In the visible regime, a-TiO2 metalenses have been widely the photoresist. Then, reactive ion etching processes are conducted
demonstrated with ALD [Fig. 5(a), i].78 Although, the terms to etch a-TiO2 film on the photoresists. Finally, a-TiO2 metalenses
“a-TiO2” and “TiO2” are interchangeably used in meta-optics are fabricated after removing the unexposed photoresist with a suit-
research, we use the term “a-TiO2” to be more precise. First, the able remover. This process can be applied to other materials that
photoresist is spin-coated on a thoroughly cleaned glass substrate. can be deposited using ALD as demonstrated with ultraviolet-
The substrate is generally cleaned with acetone to remove organic metasurfaces with hafnium-dioxide (HfO2).80
traces and isopropyl alcohol.60 The options of coated photoresist Demonstrations of early a-TiO2 metalenses were optimized for
are undiluted positive-tone electron beam resist (ZEP)78 and poly single frequencies in the visible regime [Fig. 5(a), ii].81 Focusing effi-
(methyl methacrylate) (PMMA),60 with the former being a better ciencies reaching 66%, 50%, and 86% at the wavelengths of 660, 532,
option for providing a higher resolution.79 Electron-beam lithogra- and 405 nm were shown [Fig. 5(a), ii]. The working wavelength of
phy is conducted to pattern the coated resist, and then it is devel- a-TiO2 metalenses has been continuously getting broader. An achro-
oped. The a-TiO2 deposition is conducted with an ALD process. matic metalens over a bandwidth of 60 nm has been demonstrated

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-8

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

FIG. 5. Fabrication of all-dielectric metalenses and their SEM images. Fabrication of (a) titanium dioxide (TiO2), (b) gallium nitride (GaN), and (c) hydrogenated amorphous
silicon (a-Si:H) metalens fabrication. The SEM images shows fabricated metalenses with (a, ii)-(a, iv) TiO2, (b, ii)-(b, iv) GaN, and (c, ii)-(c, v) a-Si:H. Reprinted with per-
mission (a, i) from Devlin et al., Proc. Natl. Acad. Sci. U.S.A. 113, 10473–10478 (2016). Copyright 2016 Proceedings of the National Academy of Sciences of the United
States of America (a, ii) from Khorasaninejad et al., Science 352, 1190 (2016). Copyright 2016. Copyright 2019 The American Association for the Advancement of
Science, (a, iii) from Khorasaninejad et al., Nano Lett. 17, 1819 (2017). Copyright 2017 American Chemical Society, (a, iv) from Chen et al., Nat. Nanotechnol. 13, 220
(2018), (a, v) from Li et al, Sci. Adv. 7, eabe4458 (2021) Copyright 2021 The American Association for the Advancement of Science, (a, vi) from Hu et al., Nano Lett. 21,
4554 (2021). Copyright 2021 American Chemical Society, (b, i-iv) from Lin et al. Nat. Nanotechnol. 14, 227 (2019). Copyright 2019 Nature Publishing Group, (b, v) from
Chen et al., Nano Lett. 17, 6345 (2017). Copyright 2017 American Chemical Society, (b, vi) from Chen et al., Sci. Rep. 11, 6500 (2021). Copyright 2019 Nature Publishing
Group, (c, ii) from Pahlevaninezhad et al., Nat. Photonics. 12, 540 (2018), Copyright 2018, Nature Publishing Group, (c, iii) from Chantakit et al., Photonics Res. 8, 1435
(2020). Copyright 2020 Chinese Laser Press, (c, iv) from Shrestha et al., Light Sci. Appl. 7, 85 (2018).

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-9

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

with an aluminum substrate [Fig. 5(a), iii],82 and an achromatic met- Various functional infrared applications have been
alens has also been demonstrated with complex geometric a-TiO2 demonstrated with a-Si:H. Nano-optic endoscopes have been
nanostructures working over the full-visible spectrum [Fig. 5(a), iv] demonstrated with metalenses working at the NIR wavelength
and red, green, and blue (RGB) wavelengths [Fig. 5(a), v].71,83 (1240–1350 nm) [Fig. 5(c), ii].35 The 750 nm-thick a-Si:H layer is
Furthermore, this fabrication method can be applied to varifocal used, and in the fabrication process, it improves the resolution of
metalens by integrating with liquid crystals [Fig. 5(a), vi].84 the endoscope. Optical tweezers with an a-Si:H metalens have been
GaN is another option to demonstrate metalenses working at proven at the wavelength of 800 nm, with efficiencies that reach
visible frequencies [Fig. 5(b)].85–87 They are fabricated with 82.1% and 83.7%, for left-circularly polarized light (LCP) and
MOCVD and two-step dry etching. First, a double-polished right-circularly polarized light (RCP), respectively, as shown in
c-plane sapphire substrate is prepared. Thermal baking the sapphire Fig. 5[(c), iii].90 A polarization-independent achromatic metalens
substrate at 1100 °C can remove surface contamination and remove has been developed using 800 nm-thick a-Si:H meta-atoms, with its
physisorption of the native oxide on the substrate. GaN is deposited operating wavelength from 1200 to 1650 nm and efficiencies up to
with MOCVD using trimethylgallium (TMGa) and NH3 as precur- 50% [Fig. 5(c), iv].57
sors of Ga and N, respectively. A SiO2 hard mask is deposited on
the GaN film with PECVD. In the previous report, 400 nm-thick
SiO2 films were used for a hard mask of 800 nm-thick GaN films. B. Large-scale fabrication process
A photoresist is spin-coated on the GaN-SiO2 film. ZEP is typically Nano-PER enables one-step fabrication of the metalenses over
used for the photoresist. The desired patterns are then exposed a large area, with low cost and high throughput.92–94 Nanoimprint
using EBL and Cr-masks are deposited with an electron-beam lithography (NIL) makes meta-atoms by transferring from the pre-
evaporator followed by lift-off of the unexposed photoresist. The patterned master stamps.95 The procedure of NIL includes coating
Cr-mask patterns are transferred to the SiO2 using reactive ion- a resist on a substrate, heating or illuminating with UV light while
etching (RIE). The GaN patterns are formed after inductively pressing the resin with the patterned stamp, and demolding the
coupled-plasma reactive-ion etching (ICP-RIE) using BCL3/Cl2 stamp. Subsequently, the meta-atoms are transferred through the
chemistry. The GaN metalenses are finally fabricated after immer- resin, requiring no secondary operations such as evaporation or
sion in a buffered oxide etching (BOE) solution. etching. Conventional resins, which are polymers, have low refrac-
Achromatic,85,86 pixel-level routing,87 and highly efficient88 tive indices which are unable to confine light strongly, making it
metalenses have been demonstrated with GaN [Fig. 5(b), ii–vi]. difficult to use them as meta-atoms.96 By mixing nanoparticles into
The operating wavelength of the reported achromatic metalens is the typical imprint resin, nano-PER has been developed as a way to
from 400 to 660 nm, and its efficiency reaches 49% at the central increase the refractive index of the resin and fabricate metalenses
wavelength (530 nm) [Fig. 5(b), ii–iv]. The average efficiency is with a single step. The optical properties of the resin can be tuned
40% and its highest value at a single wavelength is 67%. A pixel- according to the type, density, and size of the nanoparticles. For
level routing metalens has been demonstrated with 600 nm high example, TiO2 and Si nano-PER have been used to demonstrate
GaN meta-atoms [Fig. 5(b), v]. It consists of three types of metalenses in the visible93 and NIR regime,94 respectively.
meta-atom, whose efficiencies experimentally reach 50.6%, 91.6%, Moreover, TiO2 nano-PER has been used to achieve over 90% con-
and 87% at wavelengths of 633, 532, and 430 nm, respectively. In version efficiency.97
another report,88 efficiencies of 89%, 84%, and 79% at wavelengths The fabrication of the nano-PER is similar to the NIL process
of 633, 532, and 405 nm were demonstrated [Fig. 5(b), vi]. [Fig. 6(a), i]. First, the master mold is fabricated using EBL
a-Si:H is the dominant material for metalenses working in the [Fig. 6(a), ii]. The master mold is then coated with hard polydime-
near-infrared (NIR) region. In meta-optic research fields, the term thylsiloxane (h-PDMS) and polydimethylsiloxane (PDMS) to make
“a-Si:H” is often used as “a-Si”; however, in this chapter, we used the a soft mold. The h-PDMS can replicate meta-atoms at a higher res-
more specific term “a-Si:H.” The fabrication of a-Si:H is conducted olution than typical PDMS. After hardening the soft mold, it is
with a cost-effective PECVD process [Fig. 5(c), i]. First, the glass sub- demolded from the hard mold [Fig. 6(a), iii]. After coating the soft
strate is prepared, and then a-Si:H is deposited using PEVCD with mold with the nano-PER, it is placed on a substrate and pressed
silane (SiH4) as the precursor. This is generally deposited at 200– while applying UV light or heat to harden. The final nano-PER
300 °C. The photoresist is spin-coated on the a-Si:H films.89 The structure is created on the substrate by demolding the soft mold
spin-coated photoresist is then exposed with EBL and developed [Fig. 6(a), iv]. This single-step fabrication has the advantages of
with a proper developer. A conductive polymer or aluminum layer low cost and high throughput. Given that the master mold is large
can be coated on the photoresist to prevent the accumulation of elec- enough, large-scale manufacturing is also possible. For example, a
trons in the following process.90 A Cr-mask is deposited using an 4 mm-diameter metalens has been demonstrated with Si
electron-beam evaporator. RIE is conducted to transfer the Cr-mask nano-PER.94
pattern to the a-Si:H structures. Finally, a-Si:H metalenses are fabri- A 1 cm diameter metalens at visible wavelengths has been fab-
cated after Cr wet etching.91 This fabrication process can be seen in a ricated with high throughput using deep-ultraviolet projection
video form in Ref. 54. Considering that visibly transparent a-Si:H89 lithography.98 Photolithography has been actively used as a fabrica-
has been recently developed with the low-temperature deposition tion method for micro/nano patterning. Original photolithography
PECVD methods, a-Si:H is an option for metalenses working at used ultraviolet i-line (365 nm) light. However, the resolution of
visible frequencies and can significantly reduce production costs of the i-line stepper is too low to produce visible metalenses. The dif-
the conventional a-TiO2 and GaN films. fraction limit of photolithography can be reduced by using a

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-10

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

FIG. 6. Fabrication method for high throughput and large-scale metalenses. (a) Single-step fabrication method using a nano-PER (i) Schematic of the single-step fabrica-
tion processes. (ii) SEM image of the master mold fabricated by EBL. (iii) SEM image of the soft mold composed of h-polydimethylsiloxane (PDMS) and PDMS. (iv) SEM
image of the final nano-PER structure (inset) optical microscopy images. Reprinted with permission from Yoon et al., Nat. Commun. 11, 2268 (2020). Copyright 2020
Springer Nature. (b) (i) Schematic of the metasurface fabrication process using deep ultraviolet (DUV) lithography. (ii) Top-view SEM image near the center, and the inset
shows the image of the edge of the lens. (iii) Zoom-in and tilted SEM image of the metalens. (iv) Image of 45 fabricated 1 cm metalens on a 100 mm diameter SiO2
wafer. Reprinted with permission from Park et al., Nano Lett. 19, 8673 (2019). Copyright 2019 American Chemical Society.

shorter wavelength of light. Therefore, deep ultraviolet (DUV) KrF method shows the size-scalability and possibility for
(248 nm) laser has been used for fabricating smaller feature sizes. mass-production of the metalenses.
Using KrF stepper lithography, 45 metalenses of 1 cm diameter The fabrication of a large metalens has also been conducted
were fabricated on a 4-in. fused silica wafer. This fabrication with KrF stepper [Fig. 6(b), i]. First, a 100 nm-thick chrome layer is

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-11

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

deposited on a 500 μm-thick and 100 mm diameter fused silica imperfectly focuses light due to monochromatic aberrations such as
wafer. Subsequently, a DUV photoresist layer and an antireflective spherical aberration, coma aberration, astigmatism, and distortion
coating (ARC) layer is coated on the Cr-coated substrate. The ARC for the monochromatic incident light. Among them, distortion and
layer suppresses the backreflection while exposing the DUV light, astigmatism are not the main interest because distortion can be
thus enhancing the accuracy of the lithography process. The corrected through postprocessing software astigmatism–caused by
pattern of the mask for the metalens design is transferred to the the difference between tangential and sagittal focal planes for the
photoresist layer with DUV projection stepper lithography. After off-axis incidence–which is more important when considering low
the exposure, the sample goes through a postexposure bake, water NA optical systems (NA <0.2). Therefore, here, we deal with the
rinse, and spin-dry process. Using reactive ion etching (RIE) and correction of spherical and coma aberrations of the metalenses.
inductive-coupled plasma RIE (ICP-RIE), the photoresist pattern is Most real lenses are spherical, meaning that one or both sides
then transferred to the ARC and Cr layers while acting as the etch are spherical, which is designed considering the paraxial approxi-
mask. The ARC and the photoresist are removed using O2 plasma. mation. In fact, the rays of light passing through the different loca-
Using the Cr pattern as the etch mask, the final SiO2 pattern is fab- tions of a spherical lens are focused on the different planes, causing
ricated by etching the SiO2 substrate using ICP-RIE with trifluoro- the focal spot to spread away from the ideal Airy disk. Metalenses
methane/argon (CHF3/Ar) plasma. Finally, the Cr mask is removed usually follow the hyperbolic lens equation [Eq. (3)], so they focus
using ICP-RIE with ICP-RIE with Cl2 plasma. light without spherical aberration. Since metalenses are able to
overcome spherical aberration with a subwavelength thickness, they
IV. CRITICAL CONSIDERATIONS FOR METALENSES have attracted attention for use as two-photon microscopy, endos-
copy, nonlinear imaging, spectrometers, and light-field
A. Correction of aberrations imaging.34,36,85,90,99–101 In 2016, sub-diffraction-limited imaging
In optical systems, aberration including monochromatic or using a 2 mm diameter metalens with an NA = 0.8 was demon-
chromatic aberration occurs because the light emitted from an strated.81 The metalens is composed of TiO2 meta-atoms which are
object is inaccurately focused. Geometrical problems such as spher- transparent in the visible range, as shown in Fig. 7(a). To modulate
ical or asymmetric curvature of lenses cause monochromatic aber- the wavefront of the incident light, the meta-atoms are appropri-
ration whereas the dispersive property of media or structures ately oriented according to the PB phase. The focal spot of the met-
causes chromatic aberration, which arises from the incorrect or alens is closer to diffraction-limit focusing compared to a
low-resolved data. Especially, the reproduced image is blurry, commercial state-of-the-art objective lens (100× Nikon CFI 60,
curved, or separated in color during the imaging process. Hence, NA = 0.8) at the wavelengths λ = 660, 532, and 405 nm. Under
reducing both aberrations in optical systems is critical to allow 532 nm incidence, the full width at half maximum (FWHM) of the
high-resolution and accurate imaging. focal spot of the metalens is 375 nm, whereas that of the commer-
cial objective lens is 600 nm, with the diffraction-limited FWHM
being 332 nm. Using a metalens instead of the commercial objec-
1. Monochromatic aberration tive lens, subwavelength nanoholes have been resolved. Versatile
An ideal lens focuses light in the shape of the Airy disk for diffraction-limited imaging systems using metalenses have been
the incident beam at an arbitrary angle, which is related to the size introduced. Non-destructive bioimaging enables the imaging of the
of the aperture of a conventional lens. However, a real lens inside of living creatures without cutting or breaking the sample,

FIG. 7. Monochromatic aberration-corrected metalenses. (a) Spherical aberration-corrected metalens composed of TiO2 nanopillars: (i) Optical image and SEM image of
the fabricated metalens, (ii) measured focal spot profile of the metalens (NA = 0.8) designed at λ = 660 nm, (iii) measured focal spot profile of a commercial objective lens
(100× Nikon CFI 60, NA = 0.8). Reprinted with permission from Khorasaninejad et al., Science 352, 1190 (2016). Copyright 2016 The American Association for the
Advancement of Science. (b) Schematic of the doublet metalens corrected for both spherical and coma aberration. Reprinted with permission from Arbabi et al., Nat.
Commun. 7, 13682 (2016). Copyright 2016 Springer Nature.

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-12

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

which has contributed to the development of medical applications. metalenses that well operate separately for red, green, and blue fre-
Endoscopy is a typical noninvasive bioimaging device based on quencies with low crosstalk were demonstrated. Each metalens acts
optical coherence tomography for obtaining in vivo cross-sectional as a binary Fresnel zone plate and is composed of plasmonic
images of the living samples, including the retina, by inserting an meta-atoms based on different materials for employing resonances
optical fiber into the body. However, astigmatism and spherical in a narrow bandwidth, as shown in Fig. 8(a). The focusing effi-
aberration of the catheter at the end of the optical fiber hinders ciency was measured to be less than 10% overall at the red, green,
high-resolution imaging. Pahlevaninezhad et al. implemented a and blue wavelengths, which is too low to use as a practical lens.
high-performing compact metalens catheter for use in endoscopy.90 Dielectric multilayer metalenses with a measured focusing effi-
High-resolution in vivo optical coherence tomography images were ciency of 52% were demonstrated.107 The three-wavelength metal-
obtained, compared to conventional catheters such as ball lenses ens triplet is composed of amorphous silicon resonators built into
and gradient index lenses. the PDMS. The silicon resonators are optimized independently to
On the other hand, coma aberration depends on the incident operate at 1180, 1400, and 1680 nm, enabling them to be integrated
angle of the light. When the incident light is off-axis, it does not into noninteracting metalenses for achromatic imaging. Multilayer
focus on the image plane, which makes tailored focal spots like a metalenses for focusing different wavelengths of light at different
coma. This prevents the lenses from acquiring wide field-of-view positions in the same focusing plane were also designed.
(FoV) images and confines the imageable area to small spots at the Another strategy to correct chromatic aberration in metalenses
center of the image plane. Conventional wide FoV lenses, known as is dispersion engineering.108 Conventional metalenses are generally
fisheye lenses, are bulky and complex because they are imple- designed according to the hyperbolic lens equation [i.e., Eq. (3)].
mented by compounding multiple complex lenses to reduce mono- The encoded phase profile in metalenses considers focusing at a
chromatic aberrations. Arbabi et al. introduced a miniature camera single wavelength. However, when the polychromatic light passes
that corrected both spherical and coma aberrations using a metal- through the metalens, the focal spot shifts along the optical axis as
ens doublet, as shown in Fig. 5(b).102 Doublet techniques adding a function of frequency, which gives rise to different focal lengths
optical component on the front focal length of a lens is widely used for different wavelengths of light. Taylor expansion helps to com-
to correct coma and spherical aberrations.103 The first metalens prehend the cause of chromatic aberration. The Taylor expansion
corrects the coma aberration and the second acts as a spherical of equation Eq. (3) with regard to the angular frequency can be
aberration-corrected lens. Both metalenses are composed of amor- expressed as
phous silicon meta-atoms arranged in a SU-8 polymer layer. The
monochromatic aberration-corrected metalens doublet shows high- 
performance with 70% focusing efficiency and wide FoV imaging @ w(r, ω)
w(r, ω) ¼ w(r, ωd ) þ (ω  ωd )
for the incident angles up to 30° at the NIR wavelength of 850 nm, @ω ω¼ωd

as shown in Fig. 7(b). In the visible region, Groever et al. demon- @ 2 w(r, ω)
þ (ω  ωd )2 þ    , (15)
2@ 2 ω ω¼ωd
strated spherical and coma aberration-free imaging using metalens
doublet.104 Metalens doublet composed of TiO2 nanopillars shows
performance with 50% focusing efficiency and 50° field of view at a
center wavelength of 532 nm. Recently, extremely wide FoV metal- where w is the target phase profile, r is the radial coordinate on the
enses have been demonstrated.105 The rectangular and H-shaped metalens, and ωd is the angular frequency of the designed fre-
meta-atoms based on lead telluride operate as resonators. Wide quency. The first term on the right-hand side indicates the lens
FoV panoramic imaging is shown by an aperture-assisted metalens profile and manipulates the transmitted wavefront for conversing
at the mid-infrared (MIR) wavelength of 5.2 μm. light. The second term is called the group delay, which describes
the time delay at which the polychromatic pulses reach the focal
2. Chromatic aberration point. Group delay dispersion, the third term on the right-hand
side of Eq. (15), leads to the spread out polychromatic pulse.
Chromatic aberration describes the effect when polychromatic Especially, group delay and group delay dispersion result in chro-
light spreads out at the image plane, which hiders the imaging matic aberration. To realize achromatic metalenses over a broad
system from full-color imaging. Generally, variations in the refrac- bandwidth, the group delay and group delay dispersion should be
tive index of the material at different frequencies, called dispersion, appropriately compensated for, as shown in Fig. 6(b).71 In the
is the main cause of chromatic aberration but structural dispersion visible regime, the required group delay and group delay dispersion
also exists. To realize achromatic metalens, two strategies have been are in the order of fs and fs.2 It is difficult to implement group
mainly reported: achromatic metalenses at several distinct frequen- delay and group delay dispersion on the order of femtoseconds in a
cies and broadband achromatic metalenses (BAMLs) over a certain nanostructure with a subwavelength scale due to relatively low
bandwidth. The former independently deals with the discrete fre- refractive index and low aspect ratio.
quencies, while the latter controls the structural dispersion of the Recently, however, BAMLs that take care of the dispersion
meta-atoms. have been reported. The phase profile for the proposed BAML can
Intuitively, if meta-atoms can control the properties of light be expressed as
without coupling of each frequency, metalenses can be designed to
focus light of different frequencies at the same location. In 2017, a
multispectral achromatic metalens was reported.106 Three cascaded wLens (r, λ) ¼ w(r, λmax ) þ Δw(r, λ), (16)

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-13

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

FIG. 8. Broadband achromatic metalenses: (a) three-layer achromatic metalens at three wavelengths in the visible region. Reprinted with permission from Avayu et al.,
Nat. Commun. 8, 14992 (2017). Copyright 2017 Springer Nature. (b) Schematic of the principle of broadband achromatic metalenses by group delay control. Reprinted
with permission from Chen et al., Nat. Nanotechnol. 13, 220 (2018). Copyright 2018 Springer Nature. (c) Intensity profile of the focal spot of achromatic metalens com-
posed of GaN nanostructures. (d) Full-color images captured by the broadband achromatic metalens. Reprinted with permission from Wang et al., Nat. Nanotechnol. 13,
227 (2018). Copyright 2018 Springer Nature for (c) and (d).

h pffiffiffiffiffiffiffiffiffiffiffiffiffiffi i1 1
Figures 8(c) and 8(d) show the focal spot for several wavelengths
Δw(r, λ) ¼  2π r2 þ f 2  f  , (17) and full-color imaging using a BAML composed of GaN under cir-
λ λmax cularly polarized incident light.
Chen et al. demonstrated polarization-insensitive metalens
where w is the required phase, r is the radial coordinate, f is a focal composed of coupled anisotropic meta-atoms.110 When the orien-
length, λmax is the max boundary of the interested wavelength tation angle of the meta-atoms is θ = 0° or 90°, the value of Eqs. (8)
band, and λ is the working wavelength. The phase profile is repre- and (11) are equal, both RCP and LCP experience the same phase
sented by two terms: the lens equation-based phase profile about shift when incident light passes the meta-atom, which allows the
the maximum target wavelength [the first term of Eq. (16)], and metalens to operate under any arbitrary polarization, as all polari-
the phase difference depending on the working wavelength zation states can be broken down into the RCP and LCP compo-
[Eq. (17)]. Wang et al. successfully designed a BAML by compen- nents. The BAMLs introduced here that work in the visible range
sating the dispersion phase profile using specially designed have diameters that are under 30 μm, which is too small for use in
meta-atoms.109 Each meta-atom has a geometric shape that satisfies practical optical systems. However, increasing the size of BAMLs is
the specific plasmonic resonance. As a result, the designed metalens far from trivial, and requires an exponentially large group delay,
has the same focal length for any arbitrary wavelength between which is proportional to the quality factor of the meta-atoms down
1200 and 1680 nm. As a follow-up study, BAML at visible wave- to the subwavelength scale, which is a very difficult challenge.
lengths was also designed by applying GaN meta-atoms with the To increase the size of single layer achromatic metalenses, Li
PB phase.85,86 GaN has a relatively high refractive index and low et al. designed an RGB achromatic metalens by combining the
loss in the visible regime, and can also be fabricated with a high propagation phase method and Fresnel lens with multiple zones.83
aspect ratio, allowing the excitation of higher-order cavity modes. Fresnel lenses originally have large chromatic aberration due to dis-
Therefore, GaN enables the compensation of a large group delay. persion induced from the phase discontinuities at the zone

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-14

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

TABLE I. Summary of achromatic metalenses. aberration in conventional refractive optics. The diffractive lens is
also limited by several challenges such as the shadowing effect, the
Eff. low refractive index of the material, and the low resolution of fabri-
Work Material Wavelength NA Diameter (%) Pol. cation. Although these challenges have been improved via the
Reference 70 TiO2 470–670 0.2 25 μm 20 CP immersion technique, which assists to raise the NA of the lens by
Reference 82 TiO2 470, 548, 0.3 1 cm 16 NP inserting a liquid between the target-object and lens, there are still
647 limitations when exploiting a singlet lens.
Reference 84 GaN 400–660 0.215 21.6 μm 40 CP Metalenses have successfully achieved a high NA. The key to
Reference 85 GaN 400–660 0.106 25 μm 40 CP the realization of high NA is the power of a metalens to manipulate
Reference 105 Au, Ag, 450, 550, 0.1 212 μm 8.7 NP light at the subwavelength scale. To reconstruct the original signal
Al 650 while ensuring resolution of the Abbe diffraction limit, the period-
Reference 106 Si 1180, 1400, 0.29 120 μm 22 NP icity of the meta-atoms must satisfy the Nyquist sampling criterion.
1680 Therefore, the periodicity of the meta-atoms must be smaller than
Reference 108 Au 1200–1680 0.268 55.5 μm 12 CP d. If the NA of metalens increases, the diffraction-limited resolu-
Reference 109 TiO2 460–700 0.2 26.4 μm 35 NP tion decreases and it is difficult to satisfy the Nyquist sampling cri-
terion. Since a TiO2 metalens with NA = 0.8 that overcomes the
limit of conventional lenses was demonstrated, follow-up studies
have provided demonstrations of metalens-imaging by the immer-
boundaries. The authors design the profile in the individual zone to sion technique in the visible range, as shown in Fig. 9(a).81,112 The
control phase and dispersion independently by meta-atoms. NA of the reported oil immersion metalens is up to 1.1 and is able
Furthermore, using the gradient-based local optimization method, to tightly focus the light with an average FWHM = 240 nm and a
zone transition locations of the metalens are optimized for construc- Strehl ratio of 0.98 at a wavelength of 532 nm. Commercial confocal
tive interference of the light from an individual zone of the metalens microscopy integrated with an oil-immersion metalens enables
at a focal spot for the red, green, and blue wavelengths. Introducing near-diffraction-limited imaging with a spatial resolution of
the multizone dispersion-engineered metalens allowed the realiza- 200 nm. These results prove the ability of metalenses to achieve
tion of large-scale RGB achromatic metalenses (1 cm diameter). higher NA than refractive and diffractive lenses. Liang et al. pro-
We summarize the reported achromatic metalenses with their posed ultrahigh-NA, high transmission metalenses in the visible
performance including work, material, working wavelength, NA, regime using crystalline silicon, which has not only a high refractive
diameter, focusing efficiency (Eff.), and working polarization (Pol.) index but also low loss at visible wavelengths.113 The silicon
in Table I. CP and NP indicate circularly and non-polarized light, meta-atoms are arranged with a periodicity of just 220 nm, which
respectively. allows for metalenses that achieve an ultrahigh NA of 0.98 in air
and 1.48 with oil immersion experimentally and 1.73 theoretically
[Fig. 9(b)]. These results are comparable to the state-of-the-art
B. High numerical aperture commercial objective lenses. A near-unity NA (>0.99) metalens at
The NA of a lens is a characteristic related to the entrance and 715 nm wavelength has been also reported [Fig. 9(c)].114 A pair of
exit pupil of an optical system, and it is defined as cylindrical meta-atoms assist to scatter most of the light at the
edges of the metalens in the large-angle direction toward the focal
NA ¼ nsinθ, (18) spot. Note that vector diffraction theory is required to accurately
analyze the properties of high NA metalenses. The longitudinal
where n indicates the refractive index of medium and θ means components of the electric field play a critical role in the focal spot
maximum half-angle that can pass through the lens. When the NA in high NA metalenses. To exploit vector theory, numerous
of the lens increases, it can collect a larger angle of incident light or ultrahigh-effective NA metalenses, called superoscillatory metal-
conversely, focus the light more tightly. This means that the enses, have been studied, but are not covered in this Tutorial.115,116
imaging system can receive high spatial frequency components of
light diffracted from an object and high-resolution images can be C. Tunability and reconfigurability
reconstructed. However, observing objects with a subwavelength
resolution is limited by the Abbe diffraction limit.111 According to Tunability or reconfigurability is a critical consideration of the
the Abbe diffraction limit, the minimum distance between two metalenses that needs to be developed for advanced optics such as
objects that can be resolved is expressed by augmented reality, smartphone cameras, and light detecting and
ranging (LiDAR).39,117–119 For example, augmented reality suffers
λ from the vergence-accommodation conflict, narrow field of view,
d¼ : (19) which provides awkward movements of display and eye fatigue to
2NA
users. To overcome these problems, augmented reality systems use
For refractive lenses, large curvatures of the interface are heavy and several compound lenses that can change focal length.
required to increase the NA, which leads to an increase in weight This often results in users feeling weary. The use of tunable metal-
and causes serious aberrations. Therefore, bulky and heavy com- enses can be an alternative to conventional compound lenses.
pound lenses are used to increase the NA while correcting for Tunable metalens can vary the focal length while maintaining their

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-15

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

FIG. 9. High NA metalenses: (a) Part of the measurement setup for water immersion metalens based on TiO2 meta-atoms (inset shows SEM image). Reprinted with per-
mission from Chen et al., Nano Lett. 17, 3188 (2017). Copyright 2017 American Chemical Society. (b) SEM image of super-high NA metalens (i) and target sample of
nanometric scale (ii) to estimate characteristics of metalens. (iii) Measured high-resolution optical images of the target sample. Reprinted with permission from Liang et al.,
Nano Lett. 18, 4460 (2018). Copyright 2018 American Chemical Society. (c) Schematic of a metalens with near-unity NA (i), and SEM images of the fabricated metalens
(ii and iii). Reprinted with permission from Paniagua-Domínguez et al., Nano Lett. 18, 2124 (2018). Copyright 2018 American Chemical Society.

ultrathin thickness and planar form factor. Therefore, tunable metal- acts as an electrically controllable wave-retarder in the metalens, it
enses have attracted attention in recent years, and various work has can convert polarization states of incident light for the induced
been reported.120–126 Various phase modulation mechanisms such as voltage. Under the LCP incidence and inducing voltage of 1.1 V, the
converting polarization, phase change of materials, and change incident light does not change polarization states, and metalens
period of meta-atoms can vary the focal length of tunable metalenses. focuses the light at the focal plane f1. The polarization state of incident
In this section, we introduce three types of tunable metalenses classi- light changes to RCP and the metalens focuses the light at the focal
fied as external stimuli: electrical, thermal, and mechanical stimuli. plane f2 when the voltage of 1.3 V is induced. Figure 10(a) shows the
Electrically tunable metalens can vary their focal length inducing schematic of the proposed metalens and the imaging results at the dif-
electrical voltages. From the point of view that almost all systems in ferent focal planes for the 1.1 V and 1.3 V. Bosch et al. demonstrated
the real world are controlled by electrical signals, tuning focal length continuous and reversible varifocal lens with an LC-encapsulated
by electric stimuli is very important. A typical approach for electrically dielectric metasurface.127 The electrostatic actuation of the LC can
tunable metalens is exploiting liquid crystals for converting polariza- tune the resonance frequency of meta-atom, resulting in the continu-
tion states of the incident light. Liquid crystal (LC) is a typical bire- ous tuning of focal length from f = 12 mm to f = 15 mm.
fringence material that has a different refractive index in the Electrically tunable metalenses can be designed with a non-
orthogonal direction. The optics axis of LC is rotated when the elec- patterned or patterned one. For an example of non-patterned
tric field induces. Because of these characteristics, the LC has tremen- graphene, metalenses composed of a gold array on single-layered
dously contributed to science and industry including display, LiDAR, graphene have been reported.130 The operating wavelength is
and sensing. The LC also provides electrical tunability to the metal- 30 μm and its focal length shift approaches 90.4 μm when the
enses.84,124,127,128 Recently, Badloe et al. designed the tunable bifocal Fermi level of graphene changes from 0.1 to 0.9 eV with a
metalens with liquid crystals in the visible region.129 Because the LC maximum focusing efficiency of 61.6%. Similarly, a C-shaped

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-16

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

FIG. 10. Tunable metalenses using various mechanisms: (a) electrically tunable metalens by using liquid crystals. Reprinted with permission from Badloe et al., Adv. Sci.
8, 2102646 (2021). Copyright 2021 Wiley. (b) Thermally tunable metalens based on phase change material. Reprinted with permission from Shalaginov et al., Nat.
Commun. 12, 1225 (2021). Copyright 2021 Nature Publishing Group. (c) Mechanically tunable metalens integrated with micro-electromechanical systems mirror. Reprinted
with permission from Sci. Adv. 7, eabg5639 (2021). Copyright 2021 The American Association for the Advancement of Science. (d) Mechanically tunable metalens oper-
ated by the strain of graphene. Reprinted with permission from Wei et al., ACS Nano 15, 4769 (2021). Copyright 2021 American Chemical Society.

nanostructure on a graphene layer has been investigated with a 60% at the wavelength of 405 nm with a thickness of metalens of
tunable focal length from 7.13 to 8.25 mm.131 For an example of <13 nm, and their focal length can change from 190 to 236.9 μm
patterned graphene, a ring-patterned multilayer metalens has when voltage is applied to graphene.
been proposed.132 Compared to the single-layered non-patterned Phase shift in a thermally tunable metalens has been obtained
graphene metalens, the patterned graphene metalens operates at by change of the complex refractive index of materials as a response
the visible wavelength. The highest focusing efficiency reaches to induced temperature.125,133,134 Phase change materials (PCMs)

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-17

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

are used as the main approach for thermally tunable reconfiguration methods have an intuitive and inexpensive opera-
metalens.135–137 The phase of PCMs changes from crystalline to tion but have problems with modulation speed and stability.
amorphous or vice versa when they get enough energy, which Additionally, the above mechanisms, which are manipulated by
results in the variation of the refractive index. Shalaginov et al. thermal and mechanical signals, have the disadvantage of being dif-
demonstrated a reconfigurable varifocal metalens working at a MIR ficult to combine with integrated circuits, but the development of
wavelength.138 They use a chalcogenide material, especially micro-heaters145,146 and MEMS22,122,141,142/NEMS(nanoelectrome-
Ge2Sb2Se4Te1 (GSST), as a non-volatile and low loss PCM in the chanical systems mirror)147 platforms alleviates the problem. The
MIR range. The proposed metalens is composed of meta-atoms LCs are the most traditional and widely known electrical optical
with various regular geometries, as shown in Fig. 10(b). Through tuning materials. They are usually treated as separate optical com-
hot-plate annealing, the GSST changes phase from the amorphous ponents such as linear polarizers, half-, and quarter-waveplates;14,21
to crystalline state, moving the focal plane from f1 = 1.5 mm to f2- however, it has been found that they can be incorporated into met-
= 2.0 mm, as shown in Fig. 10(b). Thermally tunable metalens alens’ background materials to modulate the refractive index.127,148
using vanadium dioxide that is a volatile PCM has been also Their versatile applications and high efficiency in broadband visible
reported, but it has high loss in the NIR and MIR regions.136 light make the LCs attractive, but their modulation speed is difficult
In mechanically tunable metalenses, mechanical stimuli such to reach in the kHz range. This property prevents the LC-based
as electrostatic actuating and stretching change the position, size, systems from being exploited in high-speed sensing such as LiDAR.
and shape of the metalens.22,121,122,126,139,140 Using microelectrome- In addition to the prementioned methods, various tunable mecha-
chanical systems mirror (MEMS) to control accurate distance of nisms, e.g., carrier density tuning,10 electrochemical reactions,149 are
two different metalenses, the continuously variable focal length of a being actively studied, but they are not covered in this paper. Since
metalens doublet has been reported.141 The focal length decreases the ideal tunable metalens with high-efficiency and fast modulation
the distance between the metalens. Metalens doublet enables a speeds in broadband wavelength does not yet exist, it is recom-
large change in the focal length of 35 μm with a 1 μm distance mended to choose a tunable mechanism suitable for their purpose.
change between the metalens. Recently, electrically driven dynamic
metalens combined with piezoelectric MEMS mirrors has been V. CONCLUSION
reported.142 The incident light excites gap-surface-plasmon and
In this Tutorial, we introduced the design, fabrication, and
Fabry–Pérot resonance in the metallic metalens and MEMS
critical considerations of the metalenses. In the design section, we
mirrors, where phase shifts occur depending on the size of
introduced the basic physical phenomena that underpin the light–
meta-atoms. The piezoelectric MEMS-based metalens operates at
matter interactions required for the metalenses, and the design
high efficiency and fast speed over a broadband range, which sug-
methods for plasmonic and dielectric meta-atoms of various
gests a possibility of more complicated dynamic functional metal-
shapes. Then, we described how to calculate the required phase
enses. Figure 10(c) shows the schematic and photo of the
profiles for the metalenses that can be implemented through a
piezoelectric MEMS-based metalens. Stretching the metalens can
variety of techniques. We presented the typical phase modulation
also change the focal length continuously. When the substrate is
methods such as PB and propagation phase, along with advanced
stretched, a different phase shift in each meta-atom occurs due to
optimization methods such as inverse design. Next, three fabrica-
the increase of the meta-atom periodicity. Wei et al. demonstrated
tion processes were explained in detail, one for each of the most
broadband zoom imaging in the visible regime using a varifocal
commonly used dielectric materials, TiO2, GaN, and a-Si:H.
stretchable metalens.143 They focus on the outstanding optical
Compared to early metasurface fabrication methods that exploit
properties of graphene that is dispersionless with the absence of a
conventional materials, new materials have been investigated to
bandgap ranging from the ultraviolet to the terahertz region. The
fabricate metalenses with low-production costs over large-scale
stretchable metalens is composed of graphene on PDMS substrate
areas.150–154 As an example of one way to achieve this, nano-PER-
and zoom imaging was demonstrated by stretching along its radial
based one-step printing fabrication processes have been introduced.
direction. Moreover, dielectric elastomer actuators (DEAs), also
The nano-PER fabrication process has been actively studied, enabling
known as artificial muscles in soft robotics, can be used to demon-
cost-effective, large-area, and high-throughput fabrication. Finally,
strate electrically tunable metalens. The DEA is an electroactive
tunable metalenses for dynamic imaging and sensing were discussed,
polymer that stretches when a voltage or electric field is applied.
as well as aberration-free and high NA metalenses for full-color
She et al. demonstrated electrically tunable metalens capable of
high-resolution imaging. Although not covered in this Tutorial, non-
simultaneous focal lens tuning, on-the-fly astigmatism, and image
local metalenses155,156 and deep learning-based design methods157,158
shift corrections using the DEA.121 By patterning the DEA, metal-
have also been actively introduced in recent years. We believe that
ens with asymmetric stretch as well as uniform and isotropic
this Tutorial could act as a simple and in-depth guideline for begin-
stretch can be achieved. Uniform and isotropic stretch enables
ners studying the fundamentals of the metalenses and can be a step-
defocus, asymmetric stretch enables astigmatism, and lateral dis-
ping stone into the vast sea of possibilities for metalenses that could
placement enables image shift corrections.
be opened up in the near future.
Although various mechanisms have been reported for tunable
metalens, the advantages and limitations of each mechanism are
relatively clear.23,144 Thermally responsive PCMs can make non- ACKNOWLEDGMENTS
volatile changes with high operating speed; however, precise control This work was financially supported by the
is difficult and wavelength range are limited. Mechanical POSCO-POSTECH-RIST Convergence Research Center program

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-18

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

18
funded by POSCO, an industry-university strategic grant funded by A. McClung, S. Samudrala, M. Torfeh, M. Mansouree, and A. Arbabi, Sci.
LG Innotek, and the National Research Foundation (NRF) grants Adv. 6, eabc7646 (2020).
19
(Nos. NRF-2019R1A2C3003129, CAMM-2019M3A6B3030637, A. M. Shaltout, K. G. Lagoudakis, J. van de Groep, S. J. Kim, J. Vučković,
V. M. Shalaev, and M. L. Brongersma, Science 365, 374 (2019).
and NRF-2019R1A5A8080290) funded by the Ministry of Science 20
C. Chen, W. Song, J.-W. Chen, J.-H. Wang, Y. H. Chen, B. Xu, M.-K. Chen,
and ICT of the Korean government. Y.Y. acknowledges the H. Li, B. Fang, J. Chen, H. Y. Kuo, S. Wang, D. P. Tsai, S. Zhu, and T. Li, Light
Hyundai Motor Chung Mong-Koo fellowship and the NRF fellow- Sci. Appl. 8, 99 (2019).
ship (No. NRF-2021R1A6A3A13038935) funded by the Ministry of 21
I. Kim, M. A. Ansari, M. Q. Mehmood, W.-S. Kim, J. Jang, M. Zubair,
Education of the Korean government. J.K. acknowledges the Y.-K. Kim, and J. Rho, Adv. Mater. 32, 2004664 (2020).
POSTECH Alchemist fellowship. 22
A. L. Holsteen, A. F. Cihan, and M. L. Brongersma, Science 365, 257 (2019).
23
C. Jung, G. Kim, M. Jeong, J. Jang, Z. Dong, T. Badloe, J. K. W. Yang, and
AUTHOR DECLARATIONS J. Rho, Chem. Rev. 121, 13013 (2021).
24
F. Aieta, P. Genevet, M. A. Kats, N. Yu, R. Blanchard, Z. Gaburro, and
Conflict of Interest F. Capasso, Nano Lett. 12, 4932 (2012).
25
The authors have no conflicts to disclose. M. D. Aiello, A. S. Backer, A. J. Sapon, J. Smits, J. D. Perreault, P. Llull, and
V. M. Acosta, ACS Photonics 6, 2432 (2019).
26
J. Sisler, W. T. Chen, A. Y. Zhu, and F. Capasso, APL Photonics 5, 056107 (2020).
Author Contributions 27
W. Liu, Z. Li, H. Cheng, C. Tang, J. Li, S. Zhang, S. Chen, and J. Tian, Adv.
S.-W. M., C.L., Y.Y., and J.K. contributed equally to this work. Mater. 30, 1706368 (2018).
28
H. Lv, X. Lu, Y. Han, Z. Mou, and S. Teng, Opt. Lett. 44, 2518 (2019).
29
S. Vo, D. Fattal, W. V. Sorin, Z. Peng, T. Tran, M. Fiorentino, and
DATA AVAILABILITY
R. G. Beausoleil, IEEE Photonics Technol. Lett. 26, 1375 (2014).
30
The data that support the findings of this study are available Z. Shi, M. Khorasaninejad, Y.-W. Huang, C. Roques-Carmes, A. Y. Zhu,
from the corresponding author upon reasonable request. W. T. Chen, V. Sanjeev, Z.-W. Ding, M. Tamagnone, K. Chaudhary,
R. C. Devlin, C.-W. Qiu, and F. Capasso, Nano Lett. 18, 2420 (2018).
31
J. Engelberg and U. Levy, Nat. Commun. 11, 1991 (2020).
REFERENCES 32
N. Mahmood, H. Jeong, I. Kim, M. Q. Mehmood, M. Zubair, A. Akbar,
1
K. Huang, F. Qin, H. Liu, H. Ye, C.-W. Qiu, M. Hong, B. Luk’yanchuk, and M. Saleem, M. S. Anwar, F. A. Tahir, and J. Rho, Nanoscale 11, 20571 (2019).
J. Teng, Adv. Mater. 30, 1704556 (2018). 33
S. Colburn, A. Zhan, and A. Majumdar, Optica 5, 825 (2018).
34
2
P. Lalanne and P. Chavel, Laser Photonics Rev. 11, 1600295 (2017). A. L. Holsteen, D. Lin, I. Kauvar, G. Wetzstein, and M. L. Brongersma, Nano
3
H. Li, B. Fang, C. Chen, S. Zhu, and T. Li, Sci. Rep. 10, 417 (2020). Lett. 19, 2267 (2019).
4 35
H. Liang, A. Martins, B.-H. V. Borges, J. Zhou, E. R. Martins, J. Li, and T. Chantakit, T. Chantakit, T. Chantakit, C. Schlickriede, B. Sain, F. Meyer,
T. F. Krauss, Optica 6, 1461 (2019). T. Weiss, N. Chattham, and T. Zentgraf, Photonics Res. 8, 1435 (2020).
5 36
M. Khorasaninejad, A. Y. Zhu, C. Roques-Carmes, W. T. Chen, J. Oh, M. Faraji-Dana, E. Arbabi, A. Arbabi, S. M. Kamali, H. Kwon, and A. Faraon,
I. Mishra, R. C. Devlin, and F. Capasso, Nano Lett. 16, 7229 (2016). Nat. Commun. 9, 4196 (2018).
6
X. Ni, S. Ishii, A. V. Kildishev, and V. M. Shalaev, Light Sci. Appl. 2, e72 37
M. Miyata, M. Nakajima, and T. Hashimoto, Opt. Express 28, 9996 (2020).
38
(2013). L. Li, Z. Liu, X. Ren, S. Wang, V.-C. Su, M.-K. Chen, C. H. Chu, H. Y. Kuo,
7
I. Kim, J. Jang, G. Kim, J. Lee, T. Badloe, J. Mun, and J. Rho, Nat. Commun. B. Liu, W. Zang, G. Guo, L. Zhang, Z. Wang, S. Zhu, and D. P. Tsai, Science
12, 3614 (2021). 368, 1487 (2020).
8 39
T. Stolt, J. Kim, S. Héron, A. Vesala, Y. Yang, J. Mun, M. Kim, M. J. Huttunen, G.-Y. Lee, J.-Y. Hong, S. Hwang, S. Moon, H. Kang, S. Jeon, H. Kim,
R. Czaplicki, M. Kauranen, J. Rho, and P. Genevet, Phys. Rev. Lett. 126, 033901 J.-H. Jeong, and B. Lee, Nat. Commun. 9, 4562 (2018).
40
(2021). Y.-Y. Xie, P.-N. Ni, Q.-H. Wang, Q. Kan, G. Briere, P.-P. Chen, Z.-Z. Zhao,
9
G. Yoon, D. Lee, K. T. Nam, and J. Rho, ACS Nano 12, 6421 (2018). A. Delga, H.-R. Ren, H.-D. Chen, C. Xu, and P. Genevet, Nat. Nanotechnol. 15,
10
J. Park, B. G. Jeong, S. Il. Kim, D. Lee, J. Kim, C. Shin, C. B. Lee, T. Otsuka, 125 (2020).
41
J. Kyoung, S. Kim, K.-Y. Yang, Y.-Y. Park, J. Lee, I. Hwang, J. Jang, S. H. Song, M. Plidschun, H. Ren, J. Kim, R. Förster, S. A. Maier, and M. A. Schmidt,
M. L. Brongersma, K. Ha, S.-W. Hwang, H. Choo, and B. L. Choi, Nat. Light Sci. Appl. 10, 57 (2021).
Nanotechnol. 16, 69 (2021). 42
J. Rho, MRS Bull. 45, 180 (2020).
11
X. Ni, Z. J. Wong, M. Mrejen, Y. Wang, and X. Zhang, Science 349, 1310 43
S.-W. Moon, Y. Kim, G. Yoon, and J. Rho, IScience 23, 101877 (2020).
44
(2015). E. Maguid, I. Yulevich, D. Veksler, V. Kleiner, M. L. Brongersma, and
12
H. Ren, X. Fang, J. Jang, J. Bürger, J. Rho, and S. A. Maier, Nat. Nanotechnol. E. Hasman, Science 352, 1202 (2016).
45
15, 948 (2020). D. Lee, S. Y. Han, Y. Jeong, D. M. Nguyen, G. Yoon, J. Mun, J. Chae, J. H. Lee,
13
P. C. Wu, R. A. Pala, G. Kafaie Shirmanesh, W.-H. Cheng, R. Sokhoyan, J. G. Ok, G. Y. Jung, H. J. Park, K. Kim, and J. Rho, Sci. Rep. 8, 12393 (2018).
M. Grajower, M. Z. Alam, D. Lee, and H. A. Atwater, Nat. Commun. 10, 3654 46
A. Pors and S. I. Bozhevolnyi, Opt. Express 21, 27438 (2013).
(2019). 47
H.-H. Hsiao, C. H. Chu, and D. P. Tsai, Small Methods 1, 1600064 (2017).
14
I. Kim, W.-S. Kim, K. Kim, M. A. Ansari, M. Q. Mehmood, T. Badloe, Y. Kim, 48
P. Genevet, F. Capasso, F. Aieta, M. Khorasaninejad, and R. Devlin, Optica 4,
J. Gwak, H. Lee, Y.-K. Kim, and J. Rho, Sci. Adv. 7, eabe9943 (2021). 139 (2017).
15 49
N. A. Rubin, G. D’Aversa, P. Chevalier, Z. Shi, W. T. Chen, and F. Capasso, J. Zhang, M. ElKabbash, R. Wei, S. C. Singh, B. Lam, and C. Guo, Light Sci.
Science 365, eaax1839 (2019). Appl. 8, 53 (2019).
16 50
A. H. Dorrah, N. A. Rubin, A. Zaidi, M. Tamagnone, and F. Capasso, Nat. N. Yu, P. Genevet, M. A. Kats, F. Aieta, J.-P. Tetienne, F. Capasso, and
Photonics 15, 287 (2021). Z. Gaburro, Science 334, 333 (2011).
17
W.-J. Joo, J. Kyoung, M. Esfandyarpour, S.-H. Lee, H. Koo, S. Song, 51
A. Pors, M. G. Nielsen, R. L. Eriksen, and S. I. Bozhevolnyi, Nano Lett. 13,
Y.-N. Kwon, S. H. Song, J. C. Bae, A. Jo, M.-J. Kwon, S. H. Han, S.-H. Kim, 829 (2013).
S. Hwang, and M. L. Brongersma, Science 370, 459 (2020). 52
S. Sun, Q. He, S. Xiao, Q. Xu, X. Li, and L. Zhou, Nat. Mater. 11, 426 (2012).

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-19

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

53
D. Lin, P. Fan, E. Hasman, and M. L. Brongersma, Science 345, 298 (2014). 88
M.-H. Chen, W.-N. Chou, V.-C. Su, C.-H. Kuan, and H. Y. Lin, Sci. Rep. 11,
54
A. L. Aden and M. Kerker, J. Appl. Phys. 22, 1242 (1951). 6500 (2021).
55 89
J. Jang, H. Jeong, G. Hu, C.-W. Qiu, K. T. Nam, and J. Rho, Adv. Opt. Mater. Y. Yang, G. Yoon, S. Park, S. D. Namgung, T. Badloe, K. T. Nam, and J. Rho,
7, 1970016 (2019). Adv. Mater. 33, 2005893 (2021).
56
W. Liu and Y. S. Kivshar, Opt. Express 26, 13085 (2018). 90
H. Pahlevaninezhad, M. Khorasaninejad, Y.-W. Huang, Z. Shi, L. P. Hariri,
57
S. Shrestha, A. C. Overvig, M. Lu, A. Stein, and N. Yu, Light Sci. Appl. 7, 85 D. C. Adams, V. Ding, A. Zhu, C.-W. Qiu, F. Capasso, and M. J. Suter, Nat.
(2018). Photonics 12, 540 (2018).
58
Y. Wang, Q. Chen, W. Yang, Z. Ji, L. Jin, X. Ma, Q. Song, A. Boltasseva, 91
I. Kim, D. Lee, and J. Rho, J. Visualized Exp. 163, e61334 (2020).
J. Han, V. M. Shalaev, and S. Xiao, Nat. Commun. 12, 5560 (2021). 92
K. Kim, G. Yoon, S. Baek, J. Rho, and H. Lee, ACS Appl. Mater. Interfaces 11,
59
M. A. Ansari, I. Kim, D. Lee, M. H. Waseem, M. Zubair, N. Mahmood, 26109 (2019).
T. Badloe, S. Yerci, T. Tauqeer, M. Q. Mehmood, and J. Rho, Laser Photonics 93
G. Yoon, K. Kim, D. Huh, H. Lee, and J. Rho, Nat. Commun. 11, 2268 (2020).
Rev. 13, 1900065 (2019). 94
G. Yoon, K. Kim, S.-U. Kim, S. Han, H. Lee, and J. Rho, ACS Nano 15, 698
60
A. Ndao, L. Hsu, J. Ha, J.-H. Park, C. Chang-Hasnain, and B. Kanté, Nat. (2021).
Commun. 11, 3205 (2020). 95
D. K. Oh, T. Lee, B. Ko, T. Badloe, J. G. Ok, and J. Rho, Front. Optoelectron.
61
M. Khorasaninejad and F. Capasso, Nano Lett. 15, 6709 (2015). 14, 229 (2021).
62
M. G. Moharam and T. K. Gaylord, J. Opt. Soc. Am. 71, 811 (1981). 96
T. Higashihara and M. Ueda, Macromolecules 48, 1915 (2015).
63 97
C. Gigli, Q. Li, P. Chavel, G. Leo, M. L. Brongersma, and P. Lalanne, Laser J. Kim, D. K. Oh, H. Kim, G. Yoon, C. Jung, J. K. Kim, T. Badloe, S. Kim,
Photonics Rev. 15, 2000448 (2021). Y. Yang, J. Lee, B. Ko, J. G. Ok, and J. Rho, arXiv:2109.01141 (2021).
64
G. Yoon and J. Rho, Comput. Phys. Commun. 264, 107846 (2021). 98
J.-S. Park, S. Zhang, A. She, W. T. Chen, P. Lin, K. M. A. Yousef, J.-X. Cheng,
65
A. V. Kildishev, A. Boltasseva, and V. M. Shalaev, Science 339, 1232009 and F. Capasso, Nano Lett. 19, 8673 (2019).
99
(2013). E. Arbabi, J. Li, R. J. Hutchins, S. M. Kamali, A. Arbabi, Yu. Horie, P. Van
66
R. A. Aoni, M. Rahmani, L. Xu, K. Zangeneh Kamali, A. Komar, J. Yan, Dorpe, V. Gradinaru, D. A. Wagenaar, and A. Faraon, Nano Lett. 18, 4943
D. Neshev, and A. E. Miroshnichenko, Sci. Rep. 9, 6510 (2019). (2018).
67
J. Yi, S. N. Burokur, and A. de Lustrac, Opt. Express 23, 12942 (2015). 100
C. Schlickriede, N. Waterman, B. Reineke, P. Georgi, G. Li, S. Zhang, and
68
A. Arbabi, Yu. Horie, M. Bagheri, and A. Faraon, Nat. Nanotechnol. 10, 937 T. Zentgraf, Adv. Mater. 30, 1703843 (2018).
101
(2015). C. Schlickriede, S. S. Kruk, L. Wang, B. Sain, Y. Kivshar, and T. Zentgraf,
69
J. Kim, Y. Yang, T. Badloe, I. Kim, G. Yoon, and J. Rho, InfoMat 3, 739 Nano Lett. 20, 4370 (2020).
102
(2021). A. Arbabi, E. Arbabi, S. M. Kamali, Y. Horie, S. Han, and A. Faraon, Nat.
70
A. Arbabi, Y. Horie, A. J. Ball, M. Bagheri, and A. Faraon, Nat. Commun. 6, Commun. 7, 13682 (2016).
7069 (2015). 103
D. A. Buralli and G. M. Morris, Appl. Opt. 28, 3950 (1989).
71
W. T. Chen, A. Y. Zhu, V. Sanjeev, M. Khorasaninejad, Z. Shi, E. Lee, and 104
B. Groever, W. T. Chen, and F. Capasso, Nano Lett. 17, 4902 (2017).
F. Capasso, Nat. Nanotechnol. 13, 220 (2018). 105
M. Y. Shalaginov, S. An, F. Yang, P. Su, D. Lyzwa, A. M. Agarwal, H. Zhang,
72
R. C. Jones, J. Opt. Soc. Am. 31, 488 (1941). J. Hu, and T. Gu, Nano Lett. 20, 7429 (2020).
73
C. Menzel, C. Rockstuhl, and F. Lederer, Phys. Rev. A 82, 053811 (2010). 106
O. Avayu, E. Almeida, Y. Prior, and T. Ellenbogen, Nat. Commun. 8, 14992
74
J. P. Balthasar Mueller, N. A. Rubin, R. C. Devlin, B. Groever, and F. Capasso, (2017).
Phys. Rev. Lett. 118, 113901 (2017). 107
Y. Zhou, I. I. Kravchenko, H. Wang, J. R. Nolen, G. Gu, and J. Valentine,
75
H. Chung, and O. D. Miller, Opt. Express 28, 6945 (2020). Nano Lett. 18, 7529 (2018).
76
M. M. R. Elsawy, A. Gourdin, M. Binois, R. Duvigneau, D. Felbacq, S. Khadir, 108
W. T. Chen, A. Y. Zhu, and F. Capasso, Nat. Rev. Mater. 5, 604 (2020).
P. Genevet, and S. Lanteri, ACS Photonics 8, 2498 (2021). 109
S. Wang, P. C. Wu, V.-C. Su, Y.-C. Lai, C. Hung Chu, J.-W. Chen, S.-H. Lu,
77
M. Zhou, D. Liu, S. W. Belling, H. Cheng, M. A. Kats, S. Fan, M. L. Povinelli, J. Chen, B. Xu, C.-H. Kuan, T. Li, S. Zhu, and D. P. Tsai, Nat. Commun. 8, 187
and Z. Yu, ACS Photonics 8, 2265 (2021). (2017).
78 110
R. C. Devlin, M. Khorasaninejad, W. T. Chen, J. Oh, and F. Capasso, Proc. W. T. Chen, A. Y. Zhu, J. Sisler, Z. Bharwani, and F. Capasso, Nat. Commun.
Natl. Acad. Sci. U.S.A. 113, 10473 (2016). 10, 355 (2019).
79
K. Koshelev, M. Ali Mohammad, T. Fito, K. L. Westra, S. K. Dew, and 111
E. Abbe, Arch. Mikrosk. Anat. 9, 413 (1873).
M. Stepanova, J. Vac. Sci. Technol. B 29, 06F306 (2011). 112
W. T. Chen, A. Y. Zhu, M. Khorasaninejad, Z. Shi, V. Sanjeev, and
80
C. Zhang, S. Divitt, Q. Fan, W. Zhu, A. Agrawal, Y. Lu, T. Xu, and H. J. Lezec, F. Capasso, Nano Lett. 17, 3188 (2017).
Light Sci. Appl. 9, 55 (2020). 113
H. Liang, Q. Lin, X. Xie, Q. Sun, Y. Wang, L. Zhou, L. Liu, X. Yu, J. Zhou,
81
M. Khorasaninejad, W. T. Chen, R. C. Devlin, J. Oh, A. Y. Zhu, and T. F. Krauss, and J. Li, Nano Lett. 18, 4460 (2018).
F. Capasso, Science 352, 1190 (2016). 114
R. Paniagua-Domínguez, Y. F. Yu, E. Khaidarov, S. Choi, V. Leong,
82
M. Khorasaninejad, Z. Shi, A. Y. Zhu, W. T. Chen, V. Sanjeev, A. Zaidi, and R. M. Bakker, X. Liang, Y. H. Fu, V. Valuckas, L. A. Krivitsky, and
F. Capasso, Nano Lett. 17, 1819 (2017). A. I. Kuznetsov, Nano Lett. 18, 2124 (2018).
83 115
Z. Li, P. Lin, Y.-W. Huang, J.-S. Park, W. T. Chen, Z. Shi, C.-W. Qiu, G. Yuan, K. S. Rogers, E. T. F. Rogers, and N. I. Zheludev, Phys. Rev. Appl.
J.-X. Cheng, and F. Capasso, Sci. Adv. 7, eabe4458 (2021). 11, 064016 (2019).
84 116
Y. Hu, X. Ou, T. Zeng, J. Lai, J. Zhang, X. Li, X. Luo, L. Li, F. Fan, and D. Tang, C. Wang, Z. Zhao, Y. Wang, M. Pu, X. Li, P. Gao, and X. Luo, Laser
H. Duan, Nano Lett. 21, 4554 (2021). Photonics Rev. 9, 713 (2015).
85
R. J. Lin, V.-C. Su, S. Wang, M. K. Chen, T. L. Chung, Yu. H. Chen, 117
T. Badloe, J. Lee, J. Seong, and J. Rho, Adv. Photonics Res. 2, 2000205
H. Yu. Kuo, J.-W. Chen, J. Chen, Yi.-T. Huang, J.-H. Wang, C. H. Chu, P. C. Wu, (2021).
T. Li, Z. Wang, S. Zhu, and D. P. Tsai, Nat. Nanotechnol. 14, 227 (2019). 118
C.-W. Lee, H. J. Choi, and H. Jeong, Nano Convergence 7, 3 (2020).
86 119
S. Wang, P. C. Wu, V.-C. Su, Y.-C. Lai, M.-K. Chen, H. Yu. Kuo, B. H. Chen, I. Kim, R. J. Martins, J. Jang, T. Badloe, S. Khadir, H.-Y. Jung, H. Kim,
Yu. H. Chen, T.-T. Huang, J.-H. Wang, R.-M. Lin, C.-H. Kuan, T. Li, Z. Wang, J. Kim, P. Genevet, and J. Rho, Nat. Nanotechnol. 16, 508 (2021).
S. Zhu, and D. P. Tsai, Nat. Nanotechnol. 13, 227 (2018). 120
K. Chen, Y. Feng, F. Monticone, J. Zhao, B. Zhu, T. Jiang, L. Zhang,
87
B. H. Chen, P. C. Wu, V.-C. Su, Yi.-C. Lai, C. H. Chu, I. C. Lee, J.-W. Chen, Y. Kim, X. Ding, S. Zhang, A. Alù, and C.-W. Qiu, Adv. Mater. 29, 1606422
Y. H. Chen, Y.-C. Lan, C.-H. Kuan, and D. P. Tsai, Nano Lett. 17, 6345 (2017). (2017).

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-20

Published under an exclusive license by AIP Publishing
 Journal of TUTORIAL [Link]/journal/jap
Applied Physics

121
A. She, S. Zhang, S. Shian, D. R. Clarke, and F. Capasso, Sci. Adv. 4, K. A. Richardson, C. Rivero-Baleine, H. Zhang, J. Hu, and T. Gu, Nat. Commun.
eaap9957 (2018). 12, 1225 (2021).
122
T. Roy, S. Zhang, Il. W. Jung, M. Troccoli, F. Capasso, and D. Lopez, APL 139
H.-S. Ee and R. Agarwal, Nano Lett. 16, 2818 (2016).
Photonics 3, 021302 (2018). 140
Y. Luo, C. H. Chu, S. Vyas, H. Y. Kuo, Y. H. Chia, M. K. Chen, X. Shi,
123
G. K. Shirmanesh, R. Sokhoyan, P. C. Wu, and H. A. Atwater, ACS Nano 14, T. Tanaka, H. Misawa, Y.-Y. Huang, and D. P. Tsai, Nano Lett. 21, 5133 (2021).
141
6912 (2020). E. Arbabi, A. Arbabi, S. M. Kamali, Yu. Horie, M. Faraji-Dana, and
124
C.-Y. Fan, C.-Y. Fan, T.-J. Chuang, K.-H. Wu, G.-D. J. Su, and G.-D. J. Su, A. Faraon, Nat. Commun. 9, 812 (2018).
Opt. Express 28, 10609 (2020). 142
C. Meng, P. C. V. Thrane, F. Ding, J. Gjessing, M. Thomaschewski, C. Wu,
125
A. Afridi, J. Canet-Ferrer, L. Philippet, J. Osmond, P. Berto, and R. Quidant, C. Dirdal, and S. I. Bozhevolnyi, Sci. Adv. 7, eabg5639 (2021).
ACS Photonics 5, 4497 (2018). 143
S. Wei, G. Cao, H. Lin, X. Yuan, M. Somekh, and B. Jia, ACS Nano 15, 4769
126
Y. Wei, Y. Wang, X. Feng, S. Xiao, Z. Wang, T. Hu, M. Hu, J. Song, (2021).
M. Wegener, M. Zhao, J. Xia, and Z. Yang, Adv. Opt. Mater. 8, 2000142 (2020). 144
Y. Che, X. Wang, Q. Song, Y. Zhu, and S. Xiao, Nanophotonics 9, 4407
127
M. Bosch, M. R. Shcherbakov, K. Won, H.-S. Lee, Y. Kim, and G. Shvets, (2020).
Nano Lett. 21, 3849 (2021). 145
Y. Wang, P. Landreman, D. Schoen, K. Okabe, A. Marshall, U. Celano,
128
S. Zhou, Z. Shen, Z. Shen, X. Li, S. Ge, Y. Lu, and W. Hu, Opt. Lett. 45, 4324 H.-S. P. Wong, J. Park, and M. L. Brongersma, Nat. Nanotechnol. 16, 667 (2021).
146
(2020). Y. Zhang, C. Fowler, J. Liang, B. Azhar, M. Y. Shalaginov, S. Deckoff-Jones,
129
T. Badloe, I. Kim, Y. Kim, J. Kim, and J. Rho, Adv. Sci. 8, 2102646 S. An, J. B. Chou, C. M. Roberts, V. Liberman, M. Kang, C. Ríos,
(2021). K. A. Richardson, C. Rivero-Baleine, T. Gu, H. Zhang, and J. Hu, Nat.
130
Z. Zhang, Z. Zhang, X. Qi, X. Qi, J. Zhang, C. Guo, and Z. Zhu, Opt. Express Nanotechnol. 16, 661 (2021).
28, 28101 (2020). 147
S. Chen, Z. Liu, H. Du, C. Tang, C.-Y. Ji, B. Quan, R. Pan, L. Yang, X. Li, C. Gu,
131
N. Ullah, W. Liu, G. Wang, G. Wang, Z. Wang, A. U. R. Khalid, B. Hu, X. Zhang, Y. Yao, J. Li, N. X. Fang, and J. Li, Nat. Commun. 12, 1299 (2021).
J. Liu, and Y. Zhang, Opt. Express 28, 2789 (2020). 148
S.-Q. Li, X. Xu, R. M. Veetil, V. Valuckas, R. Paniagua-Domínguez, and
132
S. Park, G. Lee, B. Park, Y. Seo, C. bin Park, Y. T. Chun, C. Joo, J. Rho, A. I. Kuznetsov, Science 364, 1087 (2019).
J. M. Kim, J. Hone, and S. C. Jun, Light Sci. Appl. 9, 98 (2020). 149
Y. Jin, J. Liang, S. Wu, Y. Zhang, L. Zhou, Q. Wang, H. Liu, and J. Zhu, Adv.
133
P. P. Iyer, R. A. DeCrescent, T. Lewi, N. Antonellis, and J. A. Schuller, Phys. Mater. 32, 2000058 (2020).
Rev. Appl. 10, 044029 (2018). 150
V.-C. Su, C. H. Chu, G. Sun, and D. P. Tsai, Opt. Express 26, 13148 (2018).
134
K. Chaudhary, M. Tamagnone, X. Yin, C. M. Spägele, S. L. Oscurato, J. Li, 151
G. Yoon, T. Tanaka, T. Zentgraf, and J. Rho, J. Phys. D: Appl. Phys. 54,
C. Persch, R. Li, N. A. Rubin, L. A. Jauregui, K. Watanabe, T. Taniguchi, P. Kim, 383002 (2021).
M. Wuttig, J. H. Edgar, A. Ambrosio, and F. Capasso, Nat. Commun. 10, 4487 152
D. K. Oh, H. Jeong, J. Kim, Y. Kim, I. Kim, J. G. Ok, and J. Rho, J. Mech. Sci.
(2019). Technol. 35, 837 (2021).
135
W. Bai, P. Yang, J. Huang, D. Chen, J. Zhang, Z. Zhang, J. Yang, and B. Xu, 153
G. Yoon, I. Kim, S. So, J. Mun, M. Kim, and J. Rho, Sci. Rep. 7, 6668 (2017).
Sci. Rep. 9, 5368 (2019). 154
G. Yoon, I. Kim, and J. Rho, Microelectron. Eng. 163, 7 (2016).
136 155
Y. Kim, P. C. Wu, R. Sokhoyan, K. Mauser, R. Glaudell, G. Kafaie H. Kwon, D. Sounas, A. Cordaro, A. Polman, and A. Alù, Phys. Rev. Lett.
Shirmanesh, and H. A. Atwater, Nano Lett. 19, 3961 (2019). 121, 173004 (2018).
137
S. Zhou, Y. Wu, S. Chen, S. Liao, H. Zhang, C. Xie, and M. Chan, J. Phys. D: 156
A. C. Overvig, S. C. Malek, and N. Yu, Phys. Rev. Lett. 125, 017402 (2020).
Appl. Phys. 53, 204001 (2020). 157
S. So, T. Badloe, J. Noh, J. Bravo-Abad, and J. Rho, Nanophotonics 9, 1041
138
M. Y. Shalaginov, S. An, Y. Zhang, F. Yang, P. Su, V. Liberman, J. B. Chou, (2020).
C. M. Roberts, M. Kang, C. Rios, Q. Du, C. Fowler, A. Agarwal, 158
S. So, D. Lee, T. Badloe, and J. Rho, Opt. Mater. Express 11, 1863 (2021).

J. Appl. Phys. 131, 091101 (2022); doi: 10.1063/5.0078804 131, 091101-21

Published under an exclusive license by AIP Publishing

View publication stats