1. Introduction

Electromagnetic metasurfaces are artificially engineered sheet materials with electromagnetic characteristics tailored over a broad range of wavelengths [1]. Wavelength-selective infrared (IR) thermal emissivity/absorptivity control is a key technology with applications in high-efficiency radiative cooling, bio sensing, fabrication of microbolometers and drying furnace. Nanophotonic daytime radiative coolers with different types of electromagnetic metasurfaces, including multilayer [2], photonic crystal [3], and metal-insulator-metal (MIM) structures have been proposed recently [4,5]. MIM structured metasurfaces have attracted much attention because of their high optical performance [6–10], and successful control of their spectral emissivity/absorptivity has been achieved with geometrically controlled electromagnetic resonances. MIM metasurfaces have been employed in infrared bio sensing applications [11,12] because biomolecules have infrared vibrational fingerprints. MIM configurations with broad absorption peaks have been used for surface-enhanced Raman scattering applications [13]. Uncooled microbolometers based on MIM metasurface absorbers have been proposed for thermal imaging applications [14]. Drying furnace employing MIM metasurface to dry a material with a flammable solvent at a low temperature has been designed [15].

The effective area of a metasurface is typically only in the order of a few millimeters [16,17]. The development of a metasurface usually requires high-cost fabrication technologies such as electron beam lithography [18,19], deep-UV lithography, and reactive ion etching [20], and scaling up is difficult. The nanoimprinting method [21], template stripping method [22], self-assembly technique [23], and colloidal mask etching method [24] are promising technologies for the fabrication of large-area metasurfaces. However, fabrication of meter-scale metasurfaces remains a significant challenge. The final goal for such meter-scale MIM metasurfaces is to be applied in nanophotonic radiative coolers, wavelength-selective building materials, and in the thermal radiation control of drying furnaces.

To overcome the difficulty and complexity of the fabrication process of large-area MIM metasurfaces, we employed a metallic patterning process employing photomask and wet etching technologies, which enabled the mass-production of large-area metasurfaces in a short time scale compared to the ordinary etching processes that require the vacuum evacuation. This study aims to experimentally and computationally demonstrate the enhanced selective thermal emissivity/absorptivity in the IR range of a meter-scale MIM metasurface. In the first section, the fabrication techniques of the MIM metasurface are introduced. Experimental observations, including emissivity measurements using Fourier-transform infrared (FTIR) spectroscopy, are then described. Finally, we present a physical picture of the underpinning electromagnetic resonances of the proposed MIM metasurfaces using computational and theoretical models.

2. Fabrication and measurement

Figure 1(a) shows a schematic of the proposed metasurface, i.e., a periodic metal disk pattern over a dielectric film, which is in turn placed atop a metal film. $\theta$ represents the incident angle of light. In this study, ceria (CeO₂) was chosen for the dielectric part of the proposed metasurface, whereas Aluminum (Al) was selected for the metal part. The latter was chosen because of its abundance in nature, ease of procurement, and low cost. The period of the unit cell is Λ = 4 μm and the diameter of the Al disks is w = 2.5 μm. The height of the metal disks and the thickness of the dielectric spacer are fixed at h = 60 nm and d = 270 nm, respectively.

 

Fig. 1 (a) Schematic of the proposed metasurface. Periodic Al disk patterns over a CeO₂ film that is placed atop of an Al film. (b) Picture of the fabricated metasurface with the patterned area on a 1 m × 1 m substrate. (c) SEM image of the fabricated metasurface from the top.

Download Full Size | PDF

Figure 1(b) is a photograph of the fabricated MIM metasurface, containing a patterned area on a 1 m × 1 m substrate. This metasurface was created by arranging 16 cm × 16 cm patterned samples side by side. The silver part of the substrate is the fabricated 16 cm × 16 cm patterned samples. The black part of the substrate is the coating (emissivity 0.94) used to measure the sample surface temperature, the results of which will be presented as part of a future study. Figure 1(c) displays the scanning electron microscopy (SEM) images of the top of the fabricated metasurface taken using a JEOL JCM-5100 scanning electron microscope. It can be seen that the patterns show excellent symmetry and structural parameters as designed. Once we provide the photomask, it becomes far easier to fabricate the Al disk patterns with a wet etching process. Therefore, the metasurface on a large-area such as that shown in Fig. 1(b) can be mass-produced.

Figure 2 shows the fabrication process of the MIM metasurface, which is composed of six main steps: (a) Sputtering of thin films onto the base glass; (b) Coating the top Al film with a photoresist; (c) Pattern transfer to the photoresist with UV light exposure through a photomask; (d) Developing the photoresist layer to form a resist pattern; (e) Removal of the exposed Al by wet etching; (f) Removal of the remaining photoresist. The chosen substrate was 0.7 mm thick glass (Corning EAGLE XG). The base Al film (100 nm), CeO₂ film (270 nm), and top Al film (60 nm) were sputtered onto the substrate using a sputtering apparatus (Shibaura CFS-8EP). Next, a positive type photoresist (Tokyo Ohka Kogyo THMR-iP3100 MM) was spin-coated onto the top Al film and prebaked for 3 minutes before being exposed to UV light (central wavelength of 365 nm) to form the resist pattern. The exposed Al was subsequently dissolved by wet etching. A mixed aqueous acid (55 wt% phosphoric acid, 10 wt% nitric acid, 15 wt% acetic acid) was applied as the etchant. It took only approximately 2 minutes to dissolve the unnecessary Al on the 16 cm × 16 cm sample. This is a quite shorter time compared to the conventional dry etching with vacuum evacuation which takes relatively long time for each etching process. Finally, the remaining photoresist was removed by acetone. It was found that the chemical damage caused by the wet etching process was not serious enough to impair the optical characteristics of the proposed metasurface.

 

Fig. 2 Schematic of the fabrication process of the proposed metasurface. (a) Thin film sputtering (b) Photoresist coating (c) UV light exposure through a photomask (d) Development of the photoresist layer to form a photoresist pattern (e) Removal of the exposed Al by wet etching (f) Removal of the remaining photoresist.

Download Full Size | PDF

3. Simulation and LC circuit model

In order to evaluate the fabricated metasurface, its spectral directional emissivity was measured and compared to numerical simulations based on the finite-difference time-domain (FDTD) method. Figure 3 shows the spectral directional emissivity obtained from the Lumerical FDTD software together with the observed spectral directional emissivity of the fabricated metasurface. For the FDTD simulation, the dielectric function of CeO₂ was obtained from Reference [25], and that of Al was obtained from the CRC Handbook [26]. The periodic boundary conditions were set for the x and y directions. Electromagnetic simulations were performed in three-dimensional computational domains using a non-uniform structured mesh. The sizes of the minimum meshes were set at 15 nm in the x and y directions and 5 nm in the z direction. A broadband linearly polarized plane wave incident perpendicularly from the plane wave source that was placed above the structure. A frequency domain power monitor was placed above the plane wave source in order to observe the waves reflected from the structure.

 

Fig. 3 Spectral directional emissivity obtained from the simulation of the proposed metasurface (red line), and the measured spectral directional emissivity of the fabricated metasurface (blue points).

Download Full Size | PDF

The base Al film was considered opaque. Therefore, the spectral directional emissivity could be obtained by applying Kirchhoff’s law, i.e., ${\epsilon }_{\lambda }=1-R_{\lambda },$ where $R_{\lambda }$is the reflectance of the proposed metasurface obtained from the FDTD simulation. FTIR (JASCO FT/IR-6000) was used to measure the spectral directional-hemispherical reflectance of the fabricated metasurface. Here, since the Al substrate of the fabricated metasurface is opaque, its spectral directional emissivity can also be calculated from the Kirchhoff’s law. For both the simulation and the measurement, the metasurface shows one strong emission peak around 9.8 μm, which is larger than the period length, thus nearly perfect emissivity/absorptivity could be achieved. Emissivity peak at 9.8 μm is in the range of the atmospheric window wavelength. This metasurface therefore can be applied to the radiative cooling device or wavelength-selective building materials which require high emissivity in the atmospheric window.

Several resonance mechanisms of wavelength-selective devices have been reported [27]. This emissivity peak is considered as the result of the excitation of magnetic polaritons (MP) [28]. MPs arise due to the strong coupling of the magnetic resonance inside the MIM structure with the external electromagnetic wave. Antiparallel oscillating currents occur in the top metallic structure and bottom metallic film under an incident time-varying magnetic field. The equivalent LC circuit model has been established as a useful tool for the design of MIM metasurfaces [29–34]. Figure 4(a) is a schematic of the equivalent LC circuit model applied to the proposed metasurface. $L_m=0.5{\mu }_{0}d$ corresponds to the magnetic or mutual inductance, where ${\mu }_0$ is the permeability of vacuum. $C_{\text{m}}=c_1{\epsilon }_{{\text{CeO}}_{\text{2}}}{\epsilon }_0{w}^2/d$ represents the parallel-plate capacitance where ${\epsilon }_{{\text{CeO}}_2}$ and ${\epsilon }_0$ are the dielectric constants of CeO₂ and vacuum, respectively, and $c_1=0.18$ is a numerical factor that takes into account the fringe effect or non-uniform charge distribution along the surface of the capacitor. $C_{\text{g}}={\epsilon }_{0}hw/\left(\Lambda -w\right)$ is used to approximate the gap capacitance between the Al disks. $L_{\text{e}}=-w/\left({\epsilon }_0{\omega }^{2}w\delta \right)\cdot {{\epsilon }^{\prime }}_{\text{Al}}/\left({{\epsilon }^{\prime }}_{\text{Al}}^2+{{\epsilon }^{″}}_{\text{Al}}^2\right)$ is employed to obtain the kinetic inductance of Al, where $\omega$ is the angular frequency. ${{\epsilon }^{\prime }}_{\text{Al}}$ and ${{\epsilon }^{″}}_{\text{Al}}$ represent the real and imaginary parts of the dielectric function of Al. $\delta$ is the effective penetration depth of Al and it is determined from $\delta =\lambda /\left(2\pi \kappa \right)$, where $\kappa$ is the extinction coefficient. The total impedance of this LC circuit model can be expressed as

 

Fig. 4 (a) Equivalent LC circuit model between the Al disk and the base Al plate separated by the CeO₂ spacer. (b) Electromagnetic field profiles at 9.8 µm calculated by FDTD simulation. The color counter shows the normalized magnitude of the square of the y-component magnetic field and the vectors show the direction and magnitude of the electric field.

Download Full Size | PDF

Figure 4(b) represents the electromagnetic field profiles of the proposed metasurface at 9.8 μm along the x-z plane at y = 0 nm. The color contour shows the normalized magnitude of the square of the y-component magnetic field, and the vectors show the direction and magnitude of the electric field. A highly localized y-component magnetic field enhancement can be observed in the CeO₂ film between the top and bottom Al. The electric field creates a closed current loop that creates an enhanced magnetic field and thus forms the MP.

Figure 5 shows the simulated spectral directional emissivity shift as a function of the disk diameter, w. The green triangles indicate the wavelength at peak emissivity calculated by the LC circuit model at each disk diameter. The shift of the emissivity peak matches the LC circuit model well. These results suggest that the electromagnetic resonance originated from the excitation of MP.

 

Fig. 5 Contour plot of the simulated spectral directional emissivity as a function of the wavelength and disk diameter. The green triangles indicate the peak emissivity wavelength calculated by the LC circuit model.

Download Full Size | PDF

4. Conclusion

A MIM structured Al-CeO₂ based metasurface that shows perfect emission/absorption in the IR region was designed and fabricated in this study. Large-area fabrication of the proposed metasurface was practically achieved by applying the wet etching method. The fabricated metasurface showed a high emissivity peak in the IR region (around 9.8 μm), which agreed well with the FDTD simulation results. Calculated electromagnetic field distribution of the proposed metasurface showed a highly localized y-component magnetic field enhancement in the CeO₂ film between the top and bottom Al. In addition, we calculated the spectral directional emissivity shifts as a function of the metal disk diameter, and we found the peak shift matched the LC circuit model. Therefore, it can be concluded that the emissivity peak is caused by MP.

This work will prove to be fundamental in the fabrication of the large-scale metasurface. Future investigationshould focus on the design and experiment of the large-scale metasurface under the high temperature condition to expand its applications.