Metalenses: Advances and Applications

PROGRESS REPORT

Metalenses www.advopticalmat.de Metalenses: Advances and Applications

Ming Lun Tseng, Hui-Hsin Hsiao, Cheng Hung Chu, Mu Ku Chen, Greg Sun, Ai-Qun Liu, and Din Ping Tsai*

and optoelectronics. Metasurfaces as 2D Metasurfaces, the 2D counterpart of artificial metamaterials, have attracted artificial photonic structures are com- much attention because of their exceptional ability to manipulate the electro- posed of subwavelength resonators (often [1–8] magnetic wave such as amplitude, phase, polarization, propagation direc- referred to as meta-atoms). The optical responses/functionalities of metasurfaces tion, and so on. Different from conventional lenses, metalenses based on the can be tailored by choosing the proper metasurface optics are truly flat and compact and exhibit superior perfor- geometrical parameters of these meta- mance. In this report, recent progress in the development of metalenses is atoms.[9–14] They have been applied to explored. First, the working principle and characteristics of metalenses are realize ultrathin invisible cloaking,[15] [16–20] discussed. Then, it is described how the dispersion aberration in metalenses metahologram, surface plasmon [21–23] can be eliminated to make them suitable for being employed in a range of launchers, compact nonlinear devices,[24–28] nanolasers,[29–34] and many applications that are difficult or impossible for traditional lenses. In addition, novel compact photonic devices.[11,35–57] various metalens-based applications are introduced, including imaging, high Among all the promising applications spectral resolution spectroscopy, and multiplex color routing. Furthermore, of metasurfaces, planar metalenses with a survey of reconfigurable and tunable metalenses is conducted. Finally, the superior performance and functionalities report concludes by addressing future prospects of metalenses. over their conventional counterparts are definitely one of the most exciting and important research areas. This is because metasurface lenses (metalenses) can not 1. Introduction only show better optical functionalities but also allow for much more compact device design than the conventional high-end In recent years, optical metasurfaces have attracted much objective lenses. In this article, we will review the progress in attention due to their great potential to advance photonics the development of metalenses. This article is organized into three major parts. First, the working mechanisms of chromatic Dr. M. L. Tseng, Dr. C. H. Chu, M. K. Chen, Prof. D. P. Tsai and achromatic metalenses will be introduced. The design Research Center for Applied Sciences principle of integrated-resonant units for correcting chromatic Academia Sinica aberration of a metalens will be discussed. Second, novel appli- Taipei 11529, Taiwan E-mail: [email protected], [email protected] cations of metalenses mentioned in the first part for full color Dr. M. L. Tseng, Dr. C. H. Chu, M. K. Chen, Prof. D. P. Tsai imaging, sensing, and spectroscopy will be reviewed. After the Department of Physics discussion of the passive metalenses, the emerging tunable National Taiwan University metalenses and their applications will be surveyed, followed by Taipei 10617, Taiwan concluding remarks on future prospects. Prof. H.-H. Hsiao Institute of Electro-Optical Science and Technology National Taiwan Normal University Taipei 11677, Taiwan 2. The Origin of Chromatic Aberration Prof. H.-H. Hsiao Graduate Institute of Biomedical Optomechatronics Conventional refractive lenses rely on gradual phase accumu- Taipei Medical University lation along the optical path traveled by a light beam. This is Taipei 11031, Taiwan accomplished either by controlling the surface topography or Prof. G. Sun by varying the spatial profile of the refractive index. The chro- Department of Engineering University of Massachusetts at Boston matic aberration in refractive optical components arises from Boston, MA 02125, USA material dispersion. In a material with normal dispersion, the Prof. A.-Q. Liu refractive index decreases with increasing wavelength so that School of Electrical and Electronic Engineering a refractive lens made of the same material has larger focal Nanyang Technological University length in red than in blue. Diffractive optical devices, on the Singapore 639798, Singapore other hand, operate as the interference of light through an The ORCID identification number(s) for the author(s) of this article amplitude or phase mask. The deflection angle of a diffrac- can be found under https://doi.org/10.1002/adom.201800554. tive grating varies with wavelength, where the rate of change DOI: 10.1002/adom.201800554 depends on the pitch of the grating. Such dispersion is opposite

Adv. Optical Mater. 2018, 1800554 1800554 (1 of 16) © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.advancedsciencenews.com www.advopticalmat.de to the standard refractive optical elements (known as “negative dispersion”),[58] where longer focal length corresponds to a Greg Sun received his Ph.D. shorter wavelength. In order to eliminate chromatic aberration, in Electrical Engineering from early effort has endeavored to achieve refractive or diffractive Johns Hopkins University, achromatic lenses by integrating or cascading them in the form Baltimore, USA in 1993. He of doublets and triplets to obtain the same focal length for mul- is now professor in engi- tiple wavelengths of interest. However, the composite configu- neering and physics at the ration makes the optical system bulky, complicated, and costly. University of Massachusetts Metasurfaces are thin optical elements comprised of spatially Boston where he also serves varying optical resonators with subwavelength separations that as the founding chair of the provide an entirely new mechanism for light control. In order Engineering Department. His to convert incident planar wavefront into a spherical one, one research interests are in sem- should impose the following phase profile[59] iconductor optoelectronics, silicon photonics, and nanophotonics. He is an associate 2π 22 2 editor for the Journal of Lightwave Technology. ϕλ(,xy,)−=ϕλ(0xy,0==,) −+xy+−ff (1) λ () Ai-Qun Liu received his where x and y are the spatial coordinates with respect to Ph.D. degree in Applied the center of lens (x = y = 0), and f is the focal length of the Mechanics from National designed lens. Since the required phase profile has a depend- University of Singapore (NUS) ence on wavelength, a metalens must exhibit independent in 1994. Currently, he is a pro- phase control at each desired wavelength to achieve achromatic fessor at School of Electrical focusing. However, if the desired phase coverage of 0−2π for & Electronic Engineering, metasurface-based devices can only be obtained by either var- Nanyang Technological ying the geometric dimensions of one set of meta-atoms or University (NTU). He serves by rotating their orientations, the outcome is that this one set as an editor and editorial of meta-atoms can only satisfy the required phase for a given board member of several jour- wavelength but not at other wavelengths which leads to chro- nals. Prof. Liu is interested in matic aberration. This causes the demonstrated devices in early the research fields of MEMS, nanophotonics, optofluidics, times suffer from strong chromatic aberration even though nano/microfluidics, and sensors for water and environ- metasurface-based optical elements can be designed to operate mental monitoring. He is the author of over 160 publica- over a broadband wavelength range. The dependence of the tions including peer-reviewed journal papers and two books. deflection angle or focal length on wavelength significantly Also, he is OSA Fellow, RCS Fellow, and SPIE Fellow. compromises their performance in full-color optical applications such as imaging and displaying. Din Ping Tsai received his Ph.D. degree in Physics from University of Cincinnati, USA 3. Metalens: Control of Chromatic Dispersion in 1990. He is the Director and Distinguished Research 3.1. Multiwavelength Achromatic Metalenses Fellow of Research Center for Applied Sciences, Academia Up to now much effort has focused on multiwavelength metade- Sinica since 2012. He is also vices aimed at improving the lens behavior at a discrete set of the Distinguished Professor wavelengths.[60–63] Cappasso and co-workers utilized an aperi- of Department of Physics at odic arrangement of coupled rectangular dielectric resonators National Taiwan University. He to provide the phase coverage at three wavelengths (Figure 1a). is a fellow of AAAS, APS, IEEE, This design allows the incident near-infrared light to have the OSA, SPIE, and EMA. His current research interests are same focal length at λ = 1300, 1550, and 1800 nm with meas- nanophotonics, near-field optics, plasmonics, metamaterials, ured efficiencies of 15%, 10%, and 21%, respectively.[64] green photonics, quantum photonics, and their applications. Composite metalenses in cascading or spatially multiplexing configurations have also been demonstrated, comprising of several sublenses with each one designed to fulfill the required frequency band.[65] Arbabi et al. utilized spatially multiplexing phase profile at a specific working wavelength. Avayu et al. high contrast dielectric metaslenses to achieve light focusing used a vertical stacking of three layers to demonstrate a triply at wavelengths of 915 and 1550 nm to the same focal plane red (650 nm), green (550 nm), and blue (450 nm) achromatic (Figure 1c). Two metalenses comprised of amorphous silicon metalens (Figure 1b). Each layer consists of disc-shaped nano- (a-Si) nanoposts are designed for the operation at two different particles made of different materials (aluminum, silver, and wavelengths and are combined via either large scale segmenta- gold), and the parameters of the nanodiscs are optimized to tion or meta-atom interleaving arrangement.[66] Since the cou- support localized surface plasmon resonances for a specific pling among the nanoposts is relatively weak because each of

Figure 1. Multiwavelength achromatic metalenses. a) Metasurfaces consisting of coupled rectangular a-Si resonators focus three wavelengths of 1300, 1550, and 1800 nm to the same line. b) A vertical stacking metalens made of three different metallic nanodiscs shows the same focal distance at red (650 nm), green (550 nm), and blue (450 nm). c) Polarization insensitive metalenses comprised of a-Si nanoposts in large scale segmentation (upper panel) and interleaving arrangement (lower panel). d) A composite metalens comprised of three randomly segmented Si-based sublenses achieves light focusing at three primary red, green, and blue wavelengths to the same focal length. e) Double-wavelength metalens designed with birefringent elliptical meta-atoms.[69] f) Multispectral metasurfaces based on the spin–orbit interaction in elliptic nanoapertures. (a) Reproduced with permission.[64] Copyright 2015, American Chemical Society. (b) Reproduced with permission.[65] Copyright 2017, Nature Publishing Group. (c) Reproduced with permission.[66] Copyright 2016, Nature Publishing Group. (d) Reproduced with permission.[68] Copyright 2016, American Chemical Society. (e) Reproduced with permission.[69] Copyright 2016, Optical Society of America. (f) Reproduced with permission. Copyright 2015, Nature Publishing Group.[70]

them behaves like a multimode truncated waveguide, a unit Lin et al. used interleaved Si-based metasurfaces to eliminate cell consisting of multiple meta-atoms has more parameters the chromatic aberration at three primary colors: red, green, to independently control the phases at different wavelengths. and blue wavelengths (480, 550, and 620 nm) (Figure 1d). Thus, a unit cell with four different nanoposts was also utilized Three metalenses were designed individually to focus light to effectively sample the desired phase profiles simultaneously of these three colors at the same focal spot. They were then at two wavelengths, achieving focusing efficiencies of 22% and divided into a set of small segments randomly distributed 65% at 915 and 1550 nm, respectively, for NA = 0.46.[67] across the whole area of the composite metasurfaces.[68] In

addition, birefringent elliptical nanoposts were used to inde- dioxide (SiO2) spacer above an aluminum reflector were used to pendently control the phases of two orthogonal polarizations simultaneously control the imparted phase and its wavelength at two different wavelengths (Figure 1e). With this approach derivative (phase dispersion). The zero dispersion metalens the focal spots of the metalens at 780 and 915 nm under y- and exhibits a highly diminished chromatic dispersion in the x-polarized, respectively, appeared at the same position.[69] The wavelength range of 1450–1590 nm and maintains efficiency measured efficiencies reach above 65% for all demonstrations around 50% in the whole working spectra. On the other hand, with different NA values (up to 0.7) at both wavelengths. Zhao the hyperpositive and hypernegative dispersion metalenses et al. demonstrated multispectral metasurfaces based on the have a dispersion that is twice and 3 1/2 times, respectively, spin–orbit interaction in elliptic nanoapertures to encode the larger than a regular negative one. phase information of many different wavelengths into a single Recently, Tsai and co-workers proposed the use of integrated- metasurface (Figure 1f). The designed metalens exhibited resonant units (IRUs) in metasurfaces to realize smooth and unvaried focal length at wavelengths of 532, 632.8, and 785 nm, linear phase dispersion in combination with geometric phase respectively.[70] for the design of reflective achromatic metalenses capable The design principles of these multiwavelength metal- of operating in broad infrared bandwidth of 1200–1680 nm enses mostly rely on the integration of several metalenses (Figure 2c).[72] The design utilized a sandwich structure working at different targeted wavelengths into a single device. formed by a couple of specially arranged gold (Au) nanorods, With the proper selection of constituent materials and careful SiO2 spacer, and Au back reflector as the basic IRU building design of meta-atoms, multiwavelength achromatic metalenses block. With careful design of the multiple resonances of the can be realized in near-IR and visible regime. This approach nanorods, the phase dispersion of IRUs can be manipulated can be effective for applications that only need a few discrete to achieve different phase-changing slopes. Larger phase com- operating wavelengths, such as in florescence microscopy pensation can be directly realized by adding more resonators and optical communication. However, with the increase of into the unit cells. In addition, the phase value at the refer- number (or bandwidth) of the working wavelengths, the design enced lens center is defined asϕ shift(λ) = p/λ + q to adjust the complexity goes up dramatically. At the same time, the unin- required phase compensation profile so that the phase disper- tended coupling and interference between metalenses become sion of designed IRU building blocks can appropriately satisfy unavoidable. These undesirable effects will ultimately limit the required phase profile at each sampling positions, where their effectiveness in certain applications such as white-light λλ λ p = maxmin and q =−Ψ min , and Ψ denotes the largest imaging and spectroscopy. A novel approach for the develop- λλmaxm− in λλmaxm− in ment of broadband achromatic metalenses is critically needed. additional phase shift between λmin and λmax at the central position of the metalens. A similar strategy was applied to aluminum (Al) sandwich structures to construct achromatic met- 3.2. Continuous Broadband Achromatic Metalenses alenses capable of operating in visible (Figure 2d).[73] As Al has higher plasma frequency than Au, their reduced absorption in The recent development in dispersion-engineered metasurfaces visible leads to better performance in efficiency compared with has led to the achievement of achromatic focusing over a con- Au. The demonstrated Al metalenses show unvaried focal dis- tinuous broadband. Several works have been demonstrated in tance from 400 to 667 nm with efficiency above 20% over the reflective mode to take advantages of the large reflection ampli- working bandwidth. tude and overall increase of the phase coverage of metallic While these reflective metalenses are truly innovative in reflectors. Khorasaninejad et al. experimentally demonstrated demonstrating broadband achromaticity, transmissive optical an achromatic metalens made of titanium dioxide (TiO2) components are much more attractive and highly desirable for nanopillars and a metallic mirror separated by a dielectric spacer many practical applications. Tsai and co-workers have recently layer that exhibited a constant focal length over the contin- utilized GaN-based IRUs to achieve an achromatic metalens uous 60 nm bandwidth from λ = 490 to 550 nm (Figure 2a).[71] operating in the entire visible region in transmission mode According to the calculated phase response of the nanopillars (Figure 3a).[74] Compared to the metallic IRUs that utilize near- with different widths at a specific wavelength, they found that field coupling between the resonators to achieve phase disper- several choices of widths can achieve the same phase at the sion modulation, the optical coupling of dielectric nanopillars wavelength of interest and provide different dispersive values is relatively weak, but the high-index dielectric resonator can be simultaneously. In addition, according to Equation (1), since the functionalized as a truncated waveguide supporting multiple phase at the referenced lens center can have distinct values for Fabry–Pérot resonances. Tsai and co-workers therefore used different wavelengths, they used it as a free parameter in the high-aspect ratio GaN nanopillars and complementary struc- optimization of the metasurface that minimizes the difference tures made of nanohole arrays drilled in a GaN film to accom- between the designed and implemented phase simultaneously plish large phase compensation. The resulting metalenses for all selected wavelengths. Using the same approach, they also demonstrated unvaried focal length from 400 to 660 nm with designed a metalens with a reverse chromatic dispersion where 40% average efficiency. In addition, Chen et al. utilized two the focal length increases with increasing wavelength. TiO2 nanofins in close proximity as coupled waveguides with Arbabi et al. utilized a reflective metasurface composed Pancharatnam–Berry phase method to independently tailor the of dielectric nanoposts to demonstrate versatile focusing phase, group delay, and group delay dispersion of metalenses effect with positive, zero, and hypernegative dispersions (Figure 3b). They achieved achromatic imaging from 470 to (Figure 2b).[58] Amorphous-Si nanoposts fabricated on a silicon 670 nm with efficiency of about 20% at 500 nm.[75]

Figure 2. Reflective achromatic metalenses operating in continuous broadband. a) Building blocks made of TiO2 nanopillars, a dielectric spacer, and a metallic back reflector were applied to realize an achromatic metalens with over 60 nm continuous bandwidth achromaticity.[71] b) Dispersion-engineered metasurfaces with positive, zero, and hypernegative dispersions.[58] c) Au-IRUs metalens displaying unvaried focal distance in a broad infrared bandwidth of 1200–1680 nm.[72] d) Visible Al-IRUs metalens achieving achromatic focusing from 400 to 667 nm.[73] (a) Reproduced with permission.[71] Copyright 2017, American Chemical Society. (b) Reproduced with permission.[58] Copyright 2017, Optical Society of America. (c) Reproduced with permission.[72] Copyright 2017, Nature Publishing Group. (d) Reproduced with permission.[73] Copyright 2018, Wiley-VCH.

Three different meta-atoms have been proposed so far to the need for the back-reflection layer restrict this type of metal- achieve broadband achromatic metalenses, namely, metallic enses to applications. nanorods (Figure 4a), dielectric nanoposts (Figure 4b), and In the approach of coupled dielectric nanoposts, meta- complementary dielectric resonators (Figure 4c). Using atoms of high refractive-index and low-loss dielectric materials metallic meta-atoms, the resonance layer can be made very such as TiO2 and GaN can have strong optical resonances thin (typically less than 100 nm). Metalenses can work in vis- with near-zero optical loss. In addition, metalenses composed ible with the use of aluminum as the constituent material. of dielectric materials can work in the transmissive mode. However, the intrinsic loss of these metallic meta-atoms and However, the coupling between dielectric nanoresonators

Figure 3. Broadband achromatic metalenses in transmission mode. Metalenses a) made of GaN nanopillars and nanoholes to demonstrate a con- [74] stant focal length over from 400 to 660 nm, and b) made of coupled TiO2 nanofins to achieve achromatic imaging from 470 to 670 nm. Scale bars: 500 nm. (a) Reproduced with permission.[74] Copyright 2018, Nature Publishing Group. (b) Reproduced with permission.[75] Copyright 2018, Nature Publishing Group. is typically a weak process, making it difficult to effectively achromatic metalenses aresummarized in Figure 5. Achro- manipulate the phase dispersion and maintain high efficiency matic metalenses with continuous working bandwidth are simultaneously. The complementary dielectric resonant connected by solid lines indicating their working spectral units, on the other hand, can be employed to overcome such range, while achromatic metalenses with discrete working a difficulty. Their optical resonances are associated with the wavelengths are marked by various symbols at the achromatic high-order waveguide-like resonances inside the meta-atoms wavelengths connected by dashed line. The square and tri- and their phase dispersion can be tailored by modifying the angle symbols indicate the operation in either transmissive geometrical parameters of the meta-atoms, allowing for easier or reflective mode, respectively. In summary, two approaches attainment of the desired phase dispersion. As a result, one have been reported to realize achromatic metalenses. In the can have the freedom to choose an optimized set of meta- first approach, multiple metalenses are integrated together. atoms to realize a high-efficiency metalens. Moreover, besides Due to the compact nature of metalenses, the integrated chromatic aberration, other aberrations, such as monochro- devices are still much thinner than the conventional lenses. matic,[76] coma,[77] astigmatism,[78] can be resolved by using The second approach is to introduce various meta-atoms with metalenses as well. The state-of-the-art achievements on well-designed dispersion curves. As a result, the phase profiles

Figure 4. Approaches to design a broadband achromatic metalens. Three kinds of meta-atoms have been used to design an achromatic metalens. a) Coupled metallic nanorods.[73] The electric field distribution associated to the coupling of three Al nanorods can be observed. b) A schematic of the resonant mode between two coupled dielectric nanoposts.[79] c) Simulated magnetic energy distribution on complementary dielectric resonan- tors.[74] (b) Reproduced with permission.[79] Copyright 2018, Nature Publishing Group. (c) Reproduced with permission.[74] Copyright 2018, Nature Publishing Group.

Figure 5. State-of-the-art achromatic metalenses. The last name of the first author, adopted material for meta-atoms, and the publication year are listed.

of the metalens at different working wavelengths can be pre- Chen et al. demonstrated a high-aspect-ratio titanium dioxide [80] cisely tailored. Although the device design is more challenging (TiO2) metalens with NA of 0.8 capable of achieving diffrac- in the second approach, it has several advantages in terms of tion-limited focal spots at the design wavelengths of 405, 532, device size and fabrication cost. and 660 nm where the highest efficiency can be as high as 86% (Figure 6a,b). High resolution and imaging qualities were also achieved.[80] Unfortunately, this metalens showed chromatic 4. Applications of Metalens aberration that was resulted from the phase profile changes at different wavelengthes. Recently, Wang et al. developed an As a planar optical device, metalens has the advantage for achromatic metalens composed of GaN meta-atoms that works being ultrathin, lightweight, and ultracompact. A wide variety for the entire visible region in transmission mode.[81] As shown of functionalities that are impossible to realize in conventional in Figure 6c, the chromatic aberration correction was demon- devices are now made possible with optical metalenses. In strated in the wavelength region from 400 to 660 nm, exhib- the following, we shall discuss the applications of metalenses, iting about 49% bandwidth to the central working wavelength. including imaging, chiral sensing, hyper spectroscopy, as well Results clearly show that the achromatic metalens can resolve as color routing. Perspectives on future directions of the met- features separated by microscaled distances without chromatic alens technology will also be discussed here. aberration. Full-color imaging pictures (Figure 6d) were gener- ated from the broadband achromatic metalens—an important milestone in the development of metasurfaces for metalens- 4.1. Imaging based full-color imaging.

Human beings arguably obtain most information from the outside world through our eyes—an unartificial imaging/ 4.2. Spectroscopy optical system perfected through milions of years of evolu- tion to develop the ability to focus and project colorful image As described above in Section 1, while the wavelength dis- on retina. Metalens, a very recent technological advancement, persion of metalens often times has a negative impact to its is an artificial optical system that realizes the focusing ability performance in optical systems, the dispersion can in some by using various designs and materials. Constructing high- applications be employed to improve specific optical systems quality images with this system has proven to be a challenge when carefully designed. In a conventional spectrometer, in owing to its low working efficiency in the visible region because order to enhance spectral resolution, the distance between the of the intrinsic absorption in metals that are often used in the grating and the detector needs to be large, which makes high metalenses. Although reflection mode operation has achieved resolution spectrometers particularly bulky. Chen et al. pro- efficiency over 80%,[18] the architecture and operation of reflec- posed an off-axis metalens (Figure 6e,f) as a hybrid component tive components in general tend to increase the complexity of that combines the functions of focusing and dispersing at dif- the imaging system. Recently, this obstacle has been circum- ferent wavelengths with high spectral resolution in a compact vented by replacing metals with high index dielectric materials, configuration.[82] Silicon nanofins were used as building blocks [80] [81] such as titanium oxide (TiO2) and gallium nitride (GaN). and the phase requirement was satisfied by rotating the TiO2

Figure 6. Novel applications of metalenses. a) Image of resolution test chart captured by a TiO2-based metalens with 530 nm illumination. The scale bars: 40 µm.[84] b) Images of the highlighted region in (a) at wavelengths of 480, 530, 590, and 620 nm, respectively. The scale bars: 5 µm. c) Image of resolution test chart taken from the GaN-based achromatic metalens. The scale bars: 4 µm.[81] d) Full-colour Erithacus rubecula images captured by the GaN-based achromatic metalens.[81] e) Schematic diagram of the off-axis metalens.[82] f) High spectral resolution at the focusing angle of 80°. A wavelength difference as small as 200 pm is resolved.[82] g) Schematic of a full-color routing metalens with light convergence and color filtering functionalities.[83] (a,b) Reproduced with permission.[80] Copyright 2016, American Association for the Advancement of Science. (c,d) Reproduced with permission.[74] Copyright 2018, Nature Publishing Group. (e,f) Reproduced with permission.[82] Copyright 2016, American Chemical Society. (g) Reproduced with permission.[83] Copyright 2018, American Chemical Society. meta-atoms. Taking advantage of the dispersion behavior of the metalenses, the multiplex color router offers more degrees of metalens, the focal-line shifting with different wavelengths can freedom to focus different wavelengths to arbitrary positions be utilized to resolve the incident wavelengths with high pre- individually. Furthermore, the GaN-based metalens reveals high cision. The spectral resolution increases with focusing angle. operation efficiency in the visible region where the efficiencies When the focusing angle is set at 80°, a high dispersive char- can reach 87%, 91.6%, and 50.6% at 430, 532, and 633nm, acteristic of 0.27 nm mrad−1 can be obtained in the telecom respectively. This work combines the advantages of low-cost, region. The operating efficiency can be as high as 90% in the semiconductor fabrication compatibility, and high efficiency region of 1.1–1.6 µm. Furthermore, the working bandwidth can that allow for miniaturization and integration of various optical be extended by integrating several metalenses on a single chip. components and applications, such as imaging sensors, high- With the validation of high spectral resolution, the super-dis- resolution lithography, optical communications, and so on. persive off-axis metalenses exhibit good feasibility for compact and portable optical applications. 5. Tunable Metalenses 4.3. Full-Color Routing Metalenses discussed all have one feature in common, i.e., their properties are fixed once made. There is another impor- As a demosntration of its practicality in device miniaturiza- tant direction of research that is emerging, namely, the tun- tion, Chen et al. proposed a concept to integrate a metalens in able metalenses with functionalities that can be actively varied. a complementary metal-oxide-semiconductor (CMOS) sensor Such actively variable functionalities are desirable in optical (Figure 6g).[83] The reported GaN-based dielectric metalens applications where versatility is required. To date, there are two can function as a color router to guide three primary colors common approaches that have been used to realize tunable into different spatial positions. To seperate different colors in metalenses. The first kind is based on reconfigurable metasur- free space, the equal optical path principle is adopted to design faces.[85,86] The working principle of this category of metasur- the various phase profiles for different colors. As a result, the faces relies on the ability to reconFigure physical dimensions three primary colors at 430, 532, and 633 nm can be focused to of the metasurfaces including the interspace between the meta- arbitrary positions. Different from the super-dispersive off-axis atoms and reshaping the individual meta-atoms to actively

manipulate the output wavefront. With the change of the elec- electrode was patterned on the substrate prior to the met- tromagnetic coupling and the scattering phase difference due alens transfer. The stretchable material used in this work is a to the change of near-field interaction between meta-atoms, type of electroactive polymer, which is sometimes referred to the resonance wavelength as well as the output wavefront of as artificial muscle.[97] It can compress along the direction of metasurfaces can be modulated accordingly. Another approach external bias,[97] therefore, when a vertical bias is applied on is to integrate active materials into the metalenses. Optical the polymer, it will expand laterally due to volume conserva- properties of semiconductors such as GaAs,[87] indium tin oxide tion. The lattice constant of the metalens on the top can be (ITO),[88,89] phase-change materials,[49,90–92] and graphene[93–96] manipulated by locally applying a bias on specific electrodes, can be actively tuned by applying external excitation. Because resulting in local tuning of the phase profile accordingly. This the optical resonances of meta-atoms are highly sensitive to enables the proposed metalens to adjust the output wavefront their dielectric background, by placing a metasurface nearby an for various applications of adaptive optics. Figure 7g shows active material, its optical response can be actively controlled as the results of focal length tuning from 50 mm to nominally well. In this seciton, we will briefly review the progress made in 100 mm with a tuning ratio larger than 100%. The lateral shift the development of reconfigurable and tunable metalenes. of the focus spot as well as the active correction of the astigmatism are realized by applying a bias on specific electrodes, as shown in Figure 7h,i. Notably, because the meta-atoms in 5.1. Stretchable Metalens the device are circular nanoposts, this metalens is insensitive to the polarization state of the incident light. This makes it very One of the most efficient ways to realize a tunable metalens useful in many applications that do not necessarily need polar- is to integrate a metalens in a stretchable substrate. Stretch- ized light but require high efficiency, such as astronomical and able materials[97] such as polydimethylsiloxane (PDMS) have bioimaging systems. been widely used in reconfigurable photonic devices for color generation,[98,99] resonance switch,[100–108] and miniature spec- troscope[109] due to their high flexibility, reversibility, as well 5.2. Microelectromechanical Systems (MEMS)- as stability. They are also inexpensive and can be prepared Reconfigurable Metalens easily. The fabrication of stretchable metalenses is generally based on two steps: nanofabrication of a metalens on a tem- MEMS have been functionally applied as a feasible approach plate (Figure 7a) followed by the pattern transfer (Figure 7b). to achieve reconfigurable metalenses. With MEMS tech- For example, Ee and Agarwal used e-beam lithography to fab- nology, reconfigurable metasurfaces for intensity and anisot- ricate a metalens consisting of gold/poly(methyl methacrylate) ropy switching have been demonstrated in optical[85] and THz (PMMA) bilayered nanoparticles on a silicon substrate and regimes.[116–119] Recently, Faraon and co-workers reported a casted a PDMS layer over it. By taking advantage of the poor MEMS-based metalens doublet that its focal length can be adhesion between the PMMA and silicon, the metalens was continuously tuned along the optical axis.[120] Its fabrication successfully transferred to the stretchable substrate (Figure 7c). flowchart and working principle are shown in Figure 8a,b, The function of the stretchable substrate is to change the inter- respectively. A pair of metalenses consisting of a-Si nanoposts space of meta-atoms in the metalens to modulate the phase at are made on two substrates: one a freestanding silicon nitride the positions of individual meta-atoms. Faraon and co-workers substrate and the other a glass substrate. A spacer made of found that the change of the phase profile along the stretch- optical resist is inserted between the two metalenses to con- able metalens is proportional to 1/(1+ε)2, where ε is the strain trol the interspace between them around 10 µm. Both are sur- ratio of the stretchable substrate (Figure 7d).[110] As a result, the rounded by metallic rings to form a vertical capacitor (Figure 8c). focusing spot of the metalens can be continuously moved along By applying a bias on the metallic rings, the interspace between the axis by gradually changing the strain applied to the met- two metalenses can be actively controlled, resulting in the alens (Figure 7e). dynamic change of the effect focal length (Figure 8d). A result The stretchable metalens can be integrated with electric actu- of the imaging obtained using this MEMS-based metalens is ators to work as an adaptive planar lens. Adaptive lenses are also demonstrated. By carefully controlling the bias applied important in applications such as astronomical imaging, optical on the device, the metalens can generate an image that clearly communication, and optical scanning microscope. However, reproduces the resolution chart as shown in Figure 8e. current adaptive lenses are either too bulky[111] or highly sensitive to light polarization.[112,113] To address this issue, Capasso and co-workers reported an electrically stretchable metalens 5.3. Addressable GHz Metalens with Liquid Metal capable of not only shifting the focal spot in three dimensions but also dynamically correcting the astigmatism.[114] The As the wavefront tuning of MEMS-based metalenses is usu- proposed device is a combination of a metalens composed of ally obtained by varying the interspace of the meta-atoms or silicon nanoposts and dielectric elastomer actuators, as shown the metalens layers, this tuning is typically accompanied with in Figure 7f–i. To fabricate this device, the metalens was ini- some form of overall spatial variation of the devices which is tially made on a silicon wafer and subsequently transferred not always desirable for some highly compact photonic devices to an off-the-shelf acrylate elastomer (VHB Tape 4905, as the in which the tolerance of such a spatial change is difficult. If stretchable substrate). Between the metalens and the substrate, the resonances of the meta-atoms can be individually tuned a single-walled carbon nanotube layer as optically transparent via their physical shape changes without the overall physical

Figure 7. Stretchable metalens. a) Left: As-fabricated gold metalens before being transferred to the stretchable substrates. Right: The scanning electron microscope (SEM) image of the template after the metalens was transferred. Scale bar: 400 nm. b) An example of nanostructure transfer technique. The metalens made on a Si substrate template can be transferred to the PDMS by taking advantage of the different substrate adhesion. c) A photograph of metalens before and after being stretched. Scale bar: 10 mm. d) The working principle of the stretchable metalens for focal length tuning. e) Experimental optical intensity profiles of the strained metalens (made of Si nanoposts in a PDMS substrate) along the axial plane and the corresponding focal plane images. The focal length was tuned from 600 µm to around 1380 µm. Scale bar: 5 µm.[115] f) Photograph of the electrically stretchable metalens. Results of g) focal length tuning, h) 3D shifting the focal spot, as well as i) astigmatism correction. Scale bar: 20 µm. (a)–(c) Reproduced with permission.[115] Copyright 2016, American Chemical Society.[115] (d),(e) Reproduced with permission.[110] Copyright 2016, Wiley-VCH.[110] (f)–(i) Reproduced with permission.[78] Copyright 2018, American Association for the Advancement of Science.

Figure 8. MEMS-tunable metalens. a) Fabrication process of the MEMS-tunable metalens. b) Working principle of the device. By controlling the interspace between two metalenses, the effective focal length can be tuned accordingly. c) Photographs of the metalens pair. Gold rings as the vertical capacity can be seen in the images. Scale bars: 100 µm. d) Measured effective focal length versus the applied DC voltage of the metalens doublet. e) Applica- tion of the metalens doublet for imaging. The scale bars: 10 µm. (a)–(e) Reproduced with permission.[120] Copyright 2018, Nature Publishing Group.

variation of the metalens, the ultimate degree of freedom for between amorphous and crystalline states.[49,90–92,128–133] dynamical wavefront manipulation can be achieved. To realize The phase states of phase-change material thin films can this, Liu and co-workers and Zheludev et al. developed a recon- be reversibly switched by applying laser/electrical pulses on figurable metalens consisting of microfluidic channels and them.[49,90–92,128,130,133,134] Recently, phase-change materials have liquid metal mercury. Liquid metals such as mercury and gal- been used in nanophotonic devices for optical switching,[135,136] instan have melting points at room temperature, and they have active beam steering,[136] and color change displaying.[137] similar electrical conductivities as copper and aluminum. That Phase-change materials can also be used in tunable metalenses makes them suitable to be used as active materials in metasur- due to their low loss from optical to mid-IR regime and fast face devices.[121–123] In the proposed metalens, mercury is filled phase-transition time.[132] For example, Giessen et al. reported in ring shaped microcavities (Figure 9a). By controlling the air a mid-IR tunable metalens comprising of a plasmonic nanorod pressure in a specific cavity via a pneumatic valve, a gap can array and a phase-change thin film. The plasmonic layer is a be introduced in a meta-atom. The size and location of the air combination of two metalenses consisting of gold nanorods gap in the ring can be precisely controlled. As such, the ampli- with different lengths (Figure 10a,b). When the phase-change tude and phase modulation by the individual meta-atom can layer is in the amorphous state, the metalens consisting of be controlled (Figure 9b). The tuning of focus spot is shown longer gold nanorods is in resonnace at the wavelength of in Figure 9c, where the corresponding phase profiles of the 3.2 µm while the other set of nanorods with shorter length is metalens consisting of different set of meta-atoms are shown off-resonance. When the phase-changing layer is switched to in Figure 9d. As can be seen in Figure 9c, the focal length of crystalline state, the metalens consisting of shorter nanorods is this GHz planar metalens can be continuously tuned from 5λ in resonance while the longer ones are off-resonance. Since the to 15λ by modifying the spatial phase gradient along the met- two metalenses are designed to have different phase profiles alens. While the reported metalens worked in GHz regime, it along the surface (Figure 10c), the focal length of this device can be scaled down to work in the optical regime by adopting can be switched by phase state change, as shown in Figure 10d. nanofluidics.[124–126] This category of tunable metalenses can be used to obtain mul- tifocal lengths when more metalenses are integrated into the plasmonic layer. 5.4. Tunable Metalenses with Phase-Change Material Alloys Phase-change material can not only serve as the active layer to tune the metalenses but also be used as a “canvas” [91,138–140] Phase-change materials such as Ge2Sb2Te 5 and AgInSbTe have for direct writing of metasurface on it. Wang et al. been widely used in commercial rewritable and nonvolatile demonstrated a tunable all-dielectric metalens using a pulsed optical disks and as media for storage cells in electronic phase- femtosecond laser to record grayscale patterns consisting of change memories due to their large optical/electrical contrast crystalline nanomarks on an amorphous phase-change film

Figure 9. Liquid metal metalens. a) Photographs of the liquid metal metalens. The device is composed of ring-shaped microfluidic civilities filled with liquid metal mercury. The individual meta-atoms can be reshaped by controlling the pressure in the air gap. b) Phase and amplitude modulation of the mercury meta-atom in different shapes and orientations. By controlling the geometrical parameters of the meta-atom, 2π-phase modulation can be achieved. c) Experiment and simulation results of the focal spot shift of the proposed metalens device and d) the corresponding phase profiles of metalens with different sets of the meta-atoms. (a)–(d) Reproduced with permission.[127] Copyright 2015, Wiley-VCH.

(Figure 10e).[139] As shown in Figure 10f, two closely spaced on Pancharatnam–Berry phase method which only works for super-oscillation metalenses capable of producing two focal circularly polarized incident light. This property makes ach- spots in free space were recorded using the titghtly focused romatic metalenses unsuitable in many practical applicaitons 780 nm femtosecond laser pulses. By carefully controlling the because of the need for extra components to convert inci- illumination parameters of the fs-laser pulses, the recorded dent light into a cicularly polarized one. At the moment, the metaleneses can be erased as well as rewritten on the phase- largest numerical aperture and best efficiency of the reported change thin film (the middle panel of Figure 10f). This result achromatic metalenses based on dielectric meta-atoms are points to the potential of the phase-change metalenses for a 0.2 and 40%, respectively. Achromatic metalenses clearly need range of on-demand applications, such as real-time optical improvement to increase their performance, and to this end, scanning, imaging, and optical data processing. novel working principles and more sophisticated design of meta-atoms need to be explored. Metalenses can be integrated into a range of photonic 6. Prospects and Conclusion devices to improve their performance and expand functionalities. For example, metalenses can be used in applications In this article, we have surveyed the recent development of of vivo imaging for diagnosis due to their compact sizes and the metalenses. One of the most exciting achievements is good imaging capability. In light of their better imaging quali- the realization of broadband achromatic metalenses with the ties and much smaller footprint, metalenses can be used to use of integrated-resonant units. These metalenses can be replace microlens arrays in optical devices such as light field made flat, compact, and integrated in other optical systems camera and hemispherical electronic eye camera to improve with great potential to replace their conventional bulky coun- their performance and augment their functionalities.[142] The terparts. With the design flexibility offered by the degrees of great scalabilty of metalenses also makes them possible to be freedom to vary the geometrical parameters, interspace, and integrated into tiny unmanned flying vehicles such as machine distribution of meta-atoms, metalenses can be made to per- insects. Also, integrating metalenses with various functionali- form functionalities that are challenging to realize with con- ties on the ends of optical fibers can be very useful in optical ventional lenses, such as compact spectroscopy and full-color communication applications. Many conventional diffractive routing. However, there remaining significant challenges optical elements can be replaced by them. Some interesting as metalenses are being improved to fulfill these potentials. research directions of metalenses are also just emerging and For example, current broadband achromatic metalenses rely have not been fully studied yet. For instance, metalenses

Figure 10. Phase-change-material-based metalens. a–d) Plasmonic metalens on a phase-change thin film for dynamic focal length tuning.[141] a) SEM image of the metalens. Scale bar: 5 µm. b) Transmission spectra of rods on an amorphous (left panel) and crystalline (right panel) phase-change thin film. c) Phase profiles of the metalens in two phase states at the working wavelength. d) Dynamic focal length tuning with the tunable metalens. e) All-dielectric metalens recorded on a phase-change thin film. A metalens can be readily written on a phase-change thin film by using the femtosecond laser pulses.[139] f) Dynamically tuning the functionality of the phase-change metalens. A pair of metalenses was written on the phase-change thin film initially. Two super-resolution focusing spots can be produced by the metalens pair. After the right metalens was erased, only the left focusing spot can be observed. However, after the erased metalens was rewritten, two super-resolution spots can be observed again, indicating the reversibility of the metalens. Scale bar: 10 µm. (a)–(d) Reproduced with permission.[141] Copyright 2017, Nature Publishing Group. (e,f) Reproduced with permission.[139] Copyright 2016, Nature Publishing Group.

working in the visible for nonlinear generation and focusing actively controlled have begun to emerge, promising advan- are recently reported.[143,144] Although their efficiencies are tages such as being more compact with faster response time. At quite low at the moment, they can be improved and in due time the moment, tunable metalenses incorporating active materials serve as multifunctional devices for imaging and scanning if have only been demonstrated in the near-infrared regime. The novel nonlinear materials (e.g., multiquantum-well semicon- next frontier would be to expand their working range to the vis- ductor heterostructures)[145] of higher efficiency are integrated ible regime while achieving ultrafast tuning speed. in the devices. Metalenses for manipulating thermal radiation In conclusion, metalenses have already been ready to be are another interesting topic. A few reports studying the use used in a number of novel devices including portable mobile of plasmonic metasurfaces to control thermal emission spectra devices, virtual reality headsets, as well as drones. They are also have appeared recently,[146,147] pointing to their potential appli- expected to further penetrate and make impact to all kinds of cations in thermal imaging and focusing. optical systems. In light of their exotic optical functionalities Another emerging area is the tunable metalenses enabled enabled by their versatile wavefront tuning capability as well by the incorporation of structures or materials whose geom- as ultrasmall sizes, metalenses could potentially revolutionize etries or properties can be actively varied on demand. So far, photonics and optoelectronics in the forseeable future. majority of the metalenses in this area are demonstrated with reconfigurable structures where the interspace between meta- atoms and their shapes can be changed via various mechanical mechanisms enabled by means such as stretchable substrates, Acknowledgements MEMS, and air presure that create gaps in liquid metal. Tun- M.L.T., H.-H.H., and C.H.C. contributed equally to this article. The able metalenses integrated with active materials that can be authors acknowledge financial support from Ministry of Science and

Technology, Taiwan (grant no. MOST-106-2745-M-002-003-ASP) and [21] Y. Bao, S. Zu, W. Liu, L. Zhou, X. Zhu, Z. Fang, Phys. Rev. B 2017, Academia Sinica (grant no. AS-103-TP-A06). Authors are also grateful 95, 081406. to the National Center for Theoretical Sciences, NEMS Research Center [22] Q. Jiang, Y. Bao, F. Lin, X. Zhu, S. Zhang, Z. Fang, of National Taiwan University, National Center for High- Performance Adv. Funct. Mater. 2017, 28, 1705503. Computing, Taiwan, and Research Center for Applied Sciences, Academia [23] J. Lin, J. P. Mueller, Q. Wang, G. Yuan, N. Antoniou, X. C. Yuan, Sinica, Taiwan for their support. G. Sun acknowledges support from Air F. Capasso, Science 2013, 340, 331. Force Office of Scientific Research (grant nos. FA9550-17-1-0354 and [24] A. Krasnok, M. Tymchenko, A. Alù, Mater. Today 2018, 8, 21. FA2386-17-1-4100). The authors thank Prof. Shuming Wang and Dr. Pin [25] K. Thyagarajan, J. Butet, O. J. Martin, Nano Lett. 2013, 13, 1847. Chieh Wu for their useful discussions. [26] Y. Zhang, N. K. Grady, C. Ayala-Orozco, N. J. Halas, Nano Lett. 2011, 11, 5519. [27] G. Li, S. Zhang, T. Zentgraf, Nat. Rev. Mater. 2017, 2, 17010. Conflict of Interest [28] J. B. Khurgin, G. Sun, W. T. Chen, W.-Y. Tsai, D. P. Tsai, Sci. Rep. 2015, 5, 17899. The authors declare no conflict of interest. [29] K. C. Shen, C. T. Ku, C. Hsieh, H. C. Kuo, Y. J. Cheng, D. P. Tsai, Adv. Mater. 2018, 30, e1706918. [30] J. S. Totero Gongora, A. E. Miroshnichenko, Y. S. Kivshar, A. Fratalocchi, Nat. Commun. 2017, 8, 15535. Keywords [31] Q. Zhang, G. Li, X. Liu, F. Qian, Y. Li, T. C. Sum, C. M. Lieber, metadevices, metalenses, metasurfaces, nanophotonic devices, planar Q. Xiong, Nat. Commun. 2014, 5, 4953. optics [32] W. Zhou, M. Dridi, J. Y. Suh, C. H. Kim, D. T. Co, M. R. Wasielewski, G. C. Schatz, T. W. Odom, Nat. Nanotechnol. 2013, 8, 506. Received: April 28, 2018 [33] N. I. Zheludev, S. L. Prosvirnin, N. Papasimakis, V. A. Fedotov, Revised: June 14, 2018 Nat. Photonics 2008, 2, 351. Published online: [34] G. Shangjr, S. Chih-Kang, Rep. Prog. Phys. 2016, 79, 086501. [35] A. Tittl, M. G. Harats, R. Walter, X. Yin, M. Schäferling, N. Liu, R. Rapaport, H. Giessen, ACS Nano 2014, 8, 10885. [36] K. Aydin, V. E. Ferry, R. M. Briggs, H. A. Atwater, Nat. Commun. [1] S. Jahani, Z. Jacob, Nat. Nanotechnol. 2016, 11, 23. 2011, 2, 517. [2] N. Meinzer, W. L. Barnes, I. R. Hooper, Nat. Photonics 2014, 8, 889. [37] C.-C. Chen, A. Ishikawa, Y.-H. Tang, M.-H. Shiao, D. P. Tsai, [3] C. M. Soukoulis, M. Wegener, Nat. Photonics 2011, 5, 523. T. Tanaka, Adv. Opt. Mater. 2015, 3, 44. [4] I. Staude, J. Schilling, Nat. Photonics 2017, 11, 274. [38] P. C. Wu, C. Y. Liao, J.-W. Chen, D. P. Tsai, Adv. Opt. Mater. 2017, [5] A. I. Kuznetsov, A. E. Miroshnichenko, M. L. Brongersma, 5, 1600581. Y. S. Kivshar, B. Luk’yanchuk, Science 2016, 354, aag2472. [39] P. C. Wu, G. Sun, W. T. Chen, K.-Y. Yang, Y.-W. Huang, Y.-H. Chen, [6] W. T. Chen, C. J. Chen, P. C. Wu, S. Sun, L. Zhou, G.-Y. Guo, H. L. Huang, W.-L. Hsu, H. P. Chiang, D. P. Tsai, Appl. Phys. Lett. C. T. Hsiao, K.-Y. Yang, N. I. Zheludev, D. P. Tsai, Opt. Express 2014, 105, 033105. 2011, 19, 12837. [40] X. Fang, M. Lun Tseng, J.-Y. Ou, K. F. MacDonald, D. Ping Tsai, [7] N. J. Halas, S. Lal, W.-S. Chang, S. Link, P. Nordlander, Chem. Rev. N. I. Zheludev, Appl. Phys. Lett. 2014, 104, 141102. 2011, 111, 3913. [41] X. Fang, M. L. Tseng, D. P. Tsai, N. I. Zheludev, Phys. Rev. Appl. [8] R. Ameling, H. Giessen, Laser Photonics Rev. 2013, 7, 141. 2016, 5, 014010. [9] B. Luk’yanchuk, N. I. Zheludev, S. A. Maier, N. J. Halas, [42] M. L. Tseng, X. Fang, V. Savinov, P. C. Wu, J.-Y. Ou, N. I. Zheludev, P. Nordlander, H. Giessen, C. T. Chong, Nat. Mater. 2010, 9, 707. D. P. Tsai, Sci. Rep. 2017, 7, 44488. [10] C. Wu Pin, T. Chen Wei, K.-Y. Yang, T. Hsiao Chih, G. Sun, Q. Liu [43] K.-C. Shen, C. Hsieh, Y.-J. Cheng, D. P. Tsai, Nano Energy 2018, 45, Ai, I. Zheludev Nikolay, P. Tsai Din, Nanophotonics 2012, 1, 131. 353. [11] M. L. Tseng, P. C. Wu, S. Sun, C. M. Chang, W. T. Chen, C. H. Chu, [44] N. I. Zheludev, Y. S. Kivshar, Nat. Mater. 2012, 11, 917. P. L. Chen, L. Zhou, D. W. Huang, T. J. Yen, D. P. Tsai, Laser Photonics [45] W. Cai, A. P. Vasudev, M. L. Brongersma, Science 2011, 333, 1720. Rev. 2012, 6, 702. [46] M. Kauranen, A. V. Zayats, Nat. Photonics 2012, 6, 737. [12] N. Liu, H. Liu, S. Zhu, H. Giessen, Nat. Photonics 2009, 3, 157. [47] C. M. Chang, M. L. Tseng, B. H. Cheng, C. H. Chu, Y. Z. Ho, [13] N. Liu, M. Hentschel, T. Weiss, A. P. Alivisatos, H. Giessen, Sci- H. W. Huang, Y.-C. Lan, D.-W. Huang, A. Q. Liu, D. P. Tsai, ence 2011, 332, 1407. Adv. Mater. 2013, 25, 1118. [14] W.-Y. Tsai, C.-M. Wang, C.-F. Chen, P. C. Wu, Y.-H. Chen, T.-Y. Chen, [48] W. T. Chen, M. L. Tseng, C. Y. Liao, P. C. Wu, S. Sun, Y.-W. Huang, P. R. Wu, J.-W. Chen, D. P. Tsai, Sci. Rep. 2017, 7, 42076. C. M. Chang, C. H. Lu, L. Zhou, D.-W. Huang, A. Q. Liu, D. P. Tsai, [15] X. Ni, Z. J. Wong, M. Mrejen, Y. Wang, X. Zhang, Science 2015, Opt. Express 2013, 21, 618. 349, 1310. [49] C. H. Chu, M. L. Tseng, C. Da Shiue, S. W. Chen, H.-P. Chiang, [16] Y.-W. Huang, W. T. Chen, W.-Y. Tsai, P. C. Wu, C.-M. Wang, G. Sun, M. Mansuripur, D. P. Tsai, Opt. Express 2011, 19, 12652. D. P. Tsai, Nano Lett. 2015, 15, 3122. [50] S. Sun, Q. He, S. Xiao, Q. Xu, X. Li, L. Zhou, Nat. Mater. 2012, 11, [17] Z.-L. Deng, J. Deng, X. Zhuang, S. Wang, K. Li, Y. Wang, Y. Chi, 426. X. Ye, J. Xu, G. P. Wang, R. Zhao, X. Wang, Y. Cao, X. Cheng, G. Li, [51] N. Segal, S. Keren-Zur, N. Hendler, T. Ellenbogen, Nat. Photonics X. Li, Nano Lett. 2018, 18, 2885. 2015, 9, 180. [18] G. Zheng, H. Mühlenbernd, M. Kenney, G. Li, T. Zentgraf, [52] A. Arbabi, E. Arbabi, Y. Horie, S. M. Kamali, A. Faraon, Nat. Photonics S. Zhang, Nat. Nanotechnol. 2015, 10, 308. 2017, 11, 415. [19] H. C. Wang, C. H. Chu, P. C. Wu, H. H. Hsiao, H. J. Wu, J. W. Chen, [53] F. Qin, L. Ding, L. Zhang, F. Monticone, C. C. Chum, J. Deng, W. H. Lee, Y. C. Lai, Y. W. Huang, M. L. Tseng, S. W. Chang, S. Mei, Y. Li, J. Teng, M. Hong, S. Zhang, A. Alù, C.-W. Qiu, D. P. Tsai, Small 2018, 14, e1703920. Sci. Adv. 2016, 2, e1501168. [20] L. Wang, S. Kruk, H. Tang, T. Li, I. Kravchenko, D. N. Neshev, [54] P. C. Wu, W. Y. Tsai, W. T. Chen, Y. W. Huang, T. Y. Chen, J. W. Chen, Y. S. Kivshar, Optica 2016, 3, 1504. C. Y. Liao, C. H. Chu, G. Sun, D. P. Tsai, Nano Lett. 2017, 17, 445.

[55] W. T. Chen, P. Török, M. R. Foreman, C. Y. Liao, W. Y. Tsai, [86] N. I. Zheludev, E. Plum, Nat. Nanotechnol. 2016, 11, 16. P. R. Wu, D. P. Tsai, Nanotechnology 2016, 27, 224002. [87] H.-T. Chen, W. J. Padilla, J. M. O. Zide, A. C. Gossard, A. J. Taylor, [56] H. Ren, X. Li, Q. Zhang, M. Gu, Science 2016, 352, 805. R. D. Averitt, Nature 2006, 444, 597. [57] S. Sun, K.-Y. Yang, C.-M. Wang, T.-K. Juan, W. T. Chen, C. Y. Liao, [88] Y.-W. Huang, H. W. H. Lee, R. Sokhoyan, R. A. Pala, K. Thyagarajan, Q. He, S. Xiao, W.-T. Kung, G.-Y. Guo, L. Zhou, D. P. Tsai, Nano S. Han, D. P. Tsai, H. A. Atwater, Nano Lett. 2016, 16, 5319. Lett. 2012, 12, 6223. [89] J. Park, J.-H. Kang, S. J. Kim, X. Liu, M. L. Brongersma, Nano Lett. [58] E. Arbabi, A. Arbabi, S. M. Kamali, Y. Horie, A. Faraon, Optica 2017, 17, 407. 2017, 4, 625. [90] C. M. Chang, C. H. Chu, M. L. Tseng, H.-P. Chiang, M. Mansuripur, [59] F. Aieta, M. A. Kats, P. Genevet, F. Capasso, Science 2015, 347, D. P. Tsai, Opt. Express 2011, 19, 9492. 1342. [91] C. H. Chu, C. Da Shiue, H. W. Cheng, M. L. Tseng, H.-P. Chiang, [60] P. Wang, N. Mohammad, R. Menon, Sci. Rep. 2016, 6, 21545. M. Mansuripur, D. P. Tsai, Opt. Express 2010, 18, 18383. [61] O. Eisenbach, O. Avayu, R. Ditcovski, T. Ellenbogen, Opt. Express [92] M. L. Tseng, B. H. Chen, C. H. Chu, C. M. Chang, W. C. Lin, 2015, 23, 3928. N.-N. Chu, M. Mansuripur, A. Q. Liu, D. P. Tsai, Opt. Express 2011, [62] J. Hu, C.-H. Liu, X. Ren, L. J. Lauhon, T. W. Odom, ACS Nano 19, 16975. 2016, 10, 10275. [93] N. Dabidian, S. Dutta-Gupta, I. Kholmanov, K. Lai, F. Lu, J. Lee, [63] M. Pu, X. Li, X. Ma, Y. Wang, Z. Zhao, C. Wang, C. Hu, P. Gao, M. Jin, S. Trendafilov, A. Khanikaev, B. Fallahazad, E. Tutuc, C. Huang, H. Ren, X. Li, F. Qin, J. Yang, M. Gu, M. Hong, X. Luo, M. A. Belkin, G. Shvets, Nano Lett. 2016, 16, 3607. Sci. Adv. 2015, 2, e1601102. [94] S. H. Mousavi, I. Kholmanov, K. B. Alici, D. Purtseladze, [64] M. Khorasaninejad, F. Aieta, P. Kanhaiya, M. A. Kats, P. Genevet, N. Arju, K. Tatar, D. Y. Fozdar, J. W. Suk, Y. Hao, A. B. Khanikaev, D. Rousso, F. Capasso, Nano Lett. 2015, 15, 5358. R. S. Ruoff, G. Shvets, Nano Lett. 2013, 13, 1111. [65] O. Avayu, E. Almeida, Y. Prior, T. Ellenbogen, Nat. Commun. 2017, [95] S. Kim, M. S. Jang, V. W. Brar, Y. Tolstova, K. W. Mauser, 8, 14992. H. A. Atwater, Nat. Commun. 2016, 7, 12323. [66] E. Arbabi, A. Arbabi, S. M. Kamali, Y. Horie, A. Faraon, Sci. Rep. [96] P. C. Wu, N. Papasimakis, D. P. Tsai, Phys. Rev. Appl. 2016, 6, 2016, 6, 32803. 044019. [67] E. Arbabi, A. Arbabi, S. M. Kamali, Y. Horie, A. Faraon, Optica [97] A. O’Halloran, F. O’Malley, P. McHugh, J. Appl. Phys. 2008, 104, 2016, 3, 628. 071101. [68] D. Lin, A. L. Holsteen, E. Maguid, G. Wetzstein, P. G. Kik, [98] M. L. Tseng, J. Yang, M. Semmlinger, C. Zhang, P. Nordlander, E. Hasman, M. L. Brongersma, Nano Lett. 2016, 16, 7671. N. J. Halas, Nano Lett. 2017, 17, 6034. [69] E. Arbabi, A. Arbabi, S. M. Kamali, Y. Horie, A. Faraon, Opt. Express [99] L. Zhu, J. Kapraun, J. Ferrara, C. J. Chang-Hasnain, Optica 2015, 2, 2016, 24, 18468. 255. [70] Z. Zhao, M. Pu, H. Gao, J. Jin, X. Li, X. Ma, Y. Wang, P. Gao, [100] L. Gao, Y. Zhang, H. Zhang, S. Doshay, X. Xie, H. Luo, D. Shah, X. Luo, Sci. Rep. 2015, 5, 15781. Y. Shi, S. Xu, H. Fang, J. A. Fan, P. Nordlander, Y. Huang, [71] M. Khorasaninejad, Z. Shi, A. Y. Zhu, W. T. Chen, V. Sanjeev, J. A. Rogers, ACS Nano 2015, 9, 5968. A. Zaidi, F. Capasso, Nano Lett. 2017, 17, 1819. [101] P. Gutruf, C. Zou, W. Withayachumnankul, M. Bhaskaran, [72] S. Wang, P. C. Wu, V.-C. Su, Y.-C. Lai, C. Hung Chu, J.-W. Chen, S. Sriram, C. Fumeaux, ACS Nano 2016, 10, 133. S.-H. Lu, J. Chen, B. Xu, C.-H. Kuan, T. Li, S. Zhu, D. P. Tsai, [102] A. Yang, A. J. Hryn, M. R. Bourgeois, W. K. Lee, J. Hu, G. C. Schatz, Nat. Commun. 2017, 8, 187. T. W. Odom, Proc. Natl. Acad. Sci. USA 2016, 113, 14201. [73] H.-H. Hsiao, Y. H. Chen, R. J. Lin, P. C. Wu, S. Wang, B. H. Chen, [103] Y. Cui, J. Zhou, V. A. Tamma, W. Park, ACS Nano 2012, 6, 2385. D. P. Tsai, Adv. Opt. Mater. 2018, 6, 1800031. [104] D. Yoo, T. W. Johnson, S. Cherukulappurath, D. J. Norris, S.-H. Oh, [74] S. Wang, P. C. Wu, V.-C. Su, Y.-C. Lai, M.-K. Chen, H. Y. Kuo, ACS Nano 2015, 9, 10647. B. H. Chen, Y. H. Chen, T.-T. Huang, J.-H. Wang, R.-M. Lin, [105] S. Aksu, M. Huang, A. Artar, A. A. Yanik, S. Selvarasah, C.-H. Kuan, T. Li, Z. Wang, S. Zhu, D. P. Tsai, Nat. Nanotechnol. M. R. Dokmeci, H. Altug, Adv. Mater. 2011, 23, 4422. 2018, 13, 227. [106] F. Lütolf, D. Casari, B. Gallinet, Adv. Opt. Mater. 2016, 4, 715. [75] W. T. Chen, A. Y. Zhu, V. Sanjeev, M. Khorasaninejad, Z. Shi, [107] X. Liu, Z. Huang, C. Zhu, L. Wang, J. Zang, Nano Lett. 2018, 18, E. Lee, F. Capasso, Nat. Nanotechnol. 2018, 13, 220. 1435. [76] A. Arbabi, E. Arbabi, S. M. Kamali, Y. Horie, S. Han, A. Faraon, [108] I. M. Pryce, K. Aydin, Y. A. Kelaita, R. M. Briggs, H. A. Atwater, Nat. Commun. 2016, 7, 13682. Nano Lett. 2010, 10, 4222. [77] B. Groever, W. T. Chen, F. Capasso, Nano Lett. 2017, 17, 4902. [109] S. C. Truxal, Y.-C. Tung, K. Kurabayashi, Appl. Phys. Lett. 2008, 92, [78] A. She, S. Zhang, S. Shian, D. R. Clarke, F. Capasso, Sci. Adv. 2018, 051116. 4, eaap9957. [110] S. M. Kamali, E. Arbabi, A. Arbabi, Y. Horie, A. Faraon, Laser Photonics [79] T. Zentgraf, Nat. Nanotechnol. 2018, 13, 179. Rev. 2016, 10, 1002. [80] M. Khorasaninejad, W. T. Chen, R. C. Devlin, J. Oh, A. Y. Zhu, [111] W. W. Arrasmith, Systems Engineering and Analysis of Electro-Optical F. Capasso, Science 2016, 352, 1190. and Infrared Systems, CRC Press, Boca Raton 2015. [81] S. Wang, P. C. Wu, V. C. Su, Y. C. Lai, M. K. Chen, H. Y. Kuo, [112] H. Itoh, N. Matsumoto, T. Inoue, Opt. Express 2009, 17, 14367. B. H. Chen, Y. H. Chen, T. T. Huang, J. H. Wang, R. M. Lin, [113] H. Ren, Y.-H. Fan, S. Gauza, S.-T. Wu, Appl. Phys. Lett. 2004, 84, C. H. Kuan, T. Li, Z. Wang, S. Zhu, D. P. Tsai, Nat. Nanotechnol. 4789. 2018, 13, 227. [114] A. She, S. Zhang, S. Shian, D. R. Clarke, F. Capasso, Sci. Adv. 2018, [82] M. Khorasaninejad, W. T. Chen, J. Oh, F. Capasso, Nano Lett. 4, eaap9957. 2016, 16, 3732. [115] H. S. Ee, R. Agarwal, Nano Lett. 2016, 16, 2818. [83] B. H. Chen, P. C. Wu, V. C. Su, Y. C. Lai, C. H. Chu, I. C. Lee, [116] W. M. Zhu, A. Q. Liu, X. M. Zhang, D. P. Tsai, T. Bourouina, J. W. Chen, Y. H. Chen, Y. C. Lan, C. H. Kuan, D. P. Tsai, Nano Lett. J. H. Teng, X. H. Zhang, H. C. Guo, H. Tanoto, T. Mei, G. Q. Lo, 2017, 17, 6345. D. L. Kwong, Adv. Mater. 2011, 23, 1792. [84] M. Khorasaninejad, F. Capasso, Science 2017, 358, eaam8100. [117] Y. H. Fu, A. Q. Liu, W. M. Zhu, X. M. Zhang, D. P. Tsai, J. B. Zhang, [85] J.-Y. Ou, E. Plum, J. Zhang, N. I. Zheludev, Nat. Nanotechnol. 2013, T. Mei, J. F. Tao, H. C. Guo, X. H. Zhang, J. H. Teng, N. I. Zheludev, 8, 252. G. Q. Lo, D. L. Kwong, Adv. Funct. Mater. 2011, 21, 3589.

[118] W. M. Zhu, A. Q. Liu, T. Bourouina, D. P. Tsai, J. H. Teng, [131] Y. Yin, T. Noguchi, H. Ohno, S. Hosaka, Appl. Phys. Lett. 2009, 95, X. H. Zhang, G. Q. Lo, D. L. Kwong, N. I. Zheludev, Nat. Commun. 133503. 2012, 3, 1274. [132] D. Loke, T. H. Lee, W. J. Wang, L. P. Shi, R. Zhao, Y. C. Yeo, [119] A. D. Boardman, V. V. Grimalsky, Y. S. Kivshar, S. V. Koshevaya, T. C. Chong, S. R. Elliott, Science 2012, 336, 1566. M. Lapine, N. M. Litchinitser, V. N. Malnev, M. Noginov, [133] M. Wuttig, N. Yamada, Nat. Mater. 2007, 6, 824. Y. G. Rapoport, V. M. Shalaev, Laser Photonics Rev. 2011, 5, [134] S. Raoux, W. Welnic, D. Ielmini, Chem. Rev. 2010, 110, 240. 287. [135] B. Gholipour, J. Zhang, K. F. MacDonald, D. W. Hewak, [120] E. Arbabi, A. Arbabi, S. M. Kamali, Y. Horie, M. Faraji-Dana, N. I. Zheludev, Adv. Mater. 2013, 25, 3050. A. Faraon, Nat. Commun. 2018, 9, 812. [136] C. H. Chu, M. L. Tseng, J. Chen, P. C. Wu, Y.-H. Chen, H.-C. Wang, [121] Q. Song, W. Zhang, P. C. Wu, W. Zhu, Z. X. Shen, P. H. J. Chong, T.-Y. Chen, W. T. Hsieh, H. J. Wu, G. Sun, D. P. Tsai, Laser Photonics Q. X. Liang, Z. C. Yang, Y. L. Hao, H. Cai, H. F. Zhou, Y. Gu, Rev. 2016, 10, 1063. G. Q. Lo, D. P. Tsai, T. Bourouina, Y. Leprince-Wang, A. Q. Liu, [137] P. Hosseini, C. D. Wright, H. Bhaskaran, Nature 2014, 511, 206. Adv. Opt. Mater. 2017, 5, 1601103. [138] P. Li, X. Yang, T. W. Mass, J. Hanss, M. Lewin, A. K. Michel, [122] P. C. Wu, W. Zhu, Z. X. Shen, P. H. J. Chong, W. Ser, D. P. Tsai, M. Wuttig, T. Taubner, Nat. Mater. 2016, 15, 870. A. Q. Liu, Adv. Opt. Mater. 2017, 5, 1600938. [139] Q. Wang, E. T. F. Rogers, B. Gholipour, C.-M. Wang, G. Yuan, [123] Q. H. Song, W. M. Zhu, P. C. Wu, W. Zhang, Q. Y. S. Wu, J. Teng, N. I. Zheludev, Nat. Photonics 2016, 10, 60. J. H. Teng, Z. X. Shen, P. H. J. Chong, Q. X. Liang, Z. C. Yang, [140] K. Dong, S. Hong, Y. Deng, H. Ma, J. Li, X. Wang, J. Yeo, L. Wang, D. P. Tsai, T. Bourouina, Y. Leprince-Wang, A. Q. Liu, APL Mater. S. Lou, K. B. Tom, K. Liu, Z. You, Y. Wei, C. P. Grigoropoulos, 2017, 5, 066103. J. Yao, J. Wu, Adv. Mater. 2018, 30, 1703878. [124] T. H. H. Le, T. Tanaka, ACS Nano 2017, 11, 9780. [141] X. Yin, T. Steinle, L. Huang, T. Taubner, M. Wuttig, T. Zentgraf, [125] J. Fritzsche, D. Albinsson, M. Fritzsche, T. J. Antosiewicz, H. Giessen, Light: Sci. Appl. 2017, 6, e17016. F. Westerlund, C. Langhammer, Nano Lett. 2016, 16, 7857. [142] H. C. Ko, M. P. Stoykovich, J. Song, V. Malyarchuk, W. M. Choi, [126] M. Wang, C. Zhao, X. Miao, Y. Zhao, J. Rufo, Y. J. Liu, T. J. Huang, C.-J. Yu, J. B. Geddes Iii, J. Xiao, S. Wang, Y. Huang, J. A. Rogers, Y. Zheng, Small 2015, 11, 4423. Nature 2008, 454, 748. [127] W. Zhu, Q. Song, L. Yan, W. Zhang, P.-C. Wu, L. K. Chin, H. Cai, [143] C. Schlickriede, N. Waterman, B. Reineke, P. Georgi, G. Li, D. P. Tsai, Z. X. Shen, T. W. Deng, S. K. Ting, Y. Gu, G. Q. Lo, S. Zhang, T. Zentgraf, Adv. Mater. 2018, 30, 1703843. D. L. Kwong, Z. C. Yang, R. Huang, A.-Q. Liu, N. Zheludev, [144] E. Almeida, O. Bitton, Y. Prior, Nat. Commun. 2016, 7, 12533. Adv. Mater. 2015, 27, 4739. [145] J. Lee, M. Tymchenko, C. Argyropoulos, P. Y. Chen, F. Lu, [128] D. Ielmini, A. L. Lacaita, Mater. Today 2011, 14, 600. F. Demmerle, G. Boehm, M. C. Amann, A. Alu, M. A. Belkin, [129] H. S. P. Wong, S. Raoux, S. Kim, J. Liang, J. P. Reifenberg, Nature 2014, 511, 65. B. Rajendran, M. Asheghi, K. E. Goodson, Proc. IEEE 2010, 98, [146] M. Makhsiyan, P. Bouchon, J. Jaeck, J.-L. Pelouard, R. Haïdar, 2201. Appl. Phys. Lett. 2015, 107, 251103. [130] C. M. Chang, Y. J. Liu, M. L. Tseng, N.-N. Chu, D.-W. Huang, [147] Y. Qu, Q. Li, K. Du, L. Cai, J. Lu, M. Qiu, Laser Photonics Rev. 2017, M. Mansuripur, D. P. Tsai, Phys. Status Solidi B 2012, 249, 1945. 11, 1700091.