1. Introduction

The realization of perfect absorption [1,2] over a broad spectral region [3–5] has long been pursued and is desirable for numerous optoelectronic and energy applications, such as photodetection [6–8], photothermal conversion [9,10], thermal emission [11–13], solar steam generation [14–16], and solar thermophotovoltaics [17–19], to name a few. To date, broadband perfect absorption is usually realized in a lithographically patterned metasurface perfect absorber structure, by superimposing multiple resonant modes of resonators with different shapes and sizes [18,20,21], or of a single-design resonator with gradient geometrical parameters [22,23]. This approach, though successfully implemented over a wide range of spectral regions from visible [18,19,22,24,25], infrared (IR) [11,26,27], to terahertz (THz) [28–31], requires not only a series of complicated resonator designs in a large unit cell, but also advanced lithography as well as thin film deposition and etching techniques. The absorption bandwidth in the similar metasurface absorber structures could also be expanded either by reducing the quality factor of the otherwise narrowband plasmonic resonances, for example, replacing the low-loss noble metals (i.e., gold or silver) [32] with lossy materials such as tungsten [33–35] and metal nitrides [36–38], and loading lumped elements (i.e., resistor, inductor, capacitor, and diode) in the metallic resonators [39,40], or by the hybridization of propagating surface plasmon resonant modes with localized ones [41] or Fabry-Pérot resonances [42]. Planar thin-film stacks have also been employed for broadening absorption bandwidths, by utilizing the slow-wave resonances in hyperbolic metamaterials [43–47], the low quality-factor resonances in asymmetric Fabry-Pérot cavities [48], and the lattice scattering in one-dimensional photonic crystals backed by metal plane [49]. The fabrication of these multilayer structures can easily be scaled up for large-area applications as being lithography free, but the thicknesses of their constituent films need to be precisely controlled. In addition, the difficulty of making high-quality, thick dielectric and metal films makes it challenging for this broadband absorption scheme to be realized in the long wavelength regimes such as mid-IR and THz.

Apart from the periodic structures, on the other hand, it has been shown that disordered micro- and nanostructure arrays such as randomly distributed arrays of nanocones and nanowires [50–52], are of use to obtain broadband perfect absorption. These disordered structures achieve non-resonant absorption over a wide range of frequencies by minimizing reflection based on their graded refractive-index profile, and simultaneously by enhancing light scattering within the random structures. In comparison with the aforementioned periodic absorber structures, their formation does not rely on precise structural patterning nor repetitive thin-film deposition, thereby ensuring easy scalability (i.e., large-area fabrication) and manufacturability (i.e., low-cost production). However, these disordered surface structure arrays can be formed only on a few specific material platforms (e.g., silicon), which largely limits the scope of their practical applications. Ensemble of plasmonic nanoparticles with different geometries is an alternative disordered material system for achieving broadband perfect absorption. They can be randomly dispersed in a dielectric matrix to form nanocomposite films on metal coated substrate [53–55], realizing broadband perfect absorption by concatenating multiple resonant modes of nanoparticles, which essentially is the same strategy as that for the lithographically patterned broadband absorbers. A novel approach to achieving broadband perfect absorption using plasmonic nanoparticles is to deposit them onto the top surface and the inner walls of the nanopores of a three-dimensional AAO nanotemplate [15,56,57]. This type of absorbers has successfully demonstrated perfect absorption from visible all the way to mid-IR, attributed primarily to the overlapping and hybridization of localized plasmon resonant modes of the random-sized nanoparticles inside the nanopores, and has been utilized in applications including solar steam generation and solar desalination of water. The fabrication of this seemingly simple absorber structure, however, hinges on the non-trivial control of the gas pressure during the metal deposition, as the metal ions could just as easily land near the nanopore openings to form a perforated metal film on top of the AAO nanotemplate [58–60], instead of the desirable random-sized nanoparticles inside the nanopores. Moreover, it has been shown that the melting point of the metal nanoparticles is significantly lower than that of corresponding bulk and thin films [61]. Hence, metal nanoparticles could easily undergo a morphological transformation induced by the heating of their environments, greatly limiting the photo-to-thermal or photo-to-electric conversion efficiency of the nanoparticle-based broadband perfect absorbers, and preventing them from various high-temperature applications [62].

Here, we experimentally demonstrate TiN broadband metasurface perfect absorbers by conformally coating plasmonic TiN thin films onto disordered AAO nanotemplates. The disordered metasurface absorbers exhibit polarization-insensitive and weak angle-dependent perfect absorption over the entire visible and near-IR spectral regions (300 < λ < 2500 nm). We show that, by comparing the absorptance spectra of the absorbers with and without Al ground plane, the light scattering induced by the strong disorder of the nanopores is of critical importance for achieving broadband perfect absorption, whereas the ground plane plays a little role in enhancing the overall absorptance. We further show from numerical simulations that, the absorptance of the TiN-coated AAO nanotemplates highly depends on both the disorder and the porosity of the AAO nanotemplates, and more importantly, the strong absorption of the TiN deposited on the sidewall of the nanopores is the key to achieving broadband perfect absorption. Lastly, we demonstrate in a proof-of-concept experiment that, the TiN-coated AAO nanotemplates produce a greater temperature rise as compared to the uncoated ones under the same simulated solar illumination, illustrating their great potential for practical applications in photothermal conversion and solar energy harvesting. The TiN-coated AAO perfect absorbers possess many favorable attributes including lithography-free fabrication, low cost material, high thermal stability, and are more suitable for practical applications than those based on metal nanoparticles. Moreover, the demonstration of broadband perfect absorption employing plasmonic metal-nitride film coated disordered dielectric media provides a simple method for enhancing the efficiency of solar energy harvesting and the performance of hot-carrier based optoelectronic devices such as photodetectors and photovoltaic cells [63,64].

2. Results and discussion

Figure 1 illustrates the broadband disordered metasurface perfect absorber structures based on TiN thin-film coated AAO nanotemplates. The AAO nanotemplates are fabricated on Al thin-film coated glass substrate by a one-step anodization process. The Al thin film after the AAO self-assembly process either remains optically thick to serve as a back reflector (see Fig. 1(a)), or is completely anodized for absorbers without back metal plane. The pore size of the AAO nanotemplates is controlled by timing the wet etching in the pore widening step. Representative top-view and cross-sectional scanning electron micrographs (SEMs) of the fabricated AAO nanotemplates are shown in Fig. 1(b). The nanopores exhibit long-range disorder across the entire surface area (a few square centimeters), which is expected to play an important role in realizing broadband absorption of incident light. It can also be seen that the nanopores possess a clear distribution both in shapes and sizes, and their impacts on the absorber performance will be discussed later. The AAO nanotemplates are then conformally coated with a 25-nm thick TiN layer by atomic layer deposition (ALD) at elevated temperatures (Fig. 1(c)) [65]. The deposition condition of the ALD TiN films has been optimized so that they possess desirable strong plasmonic characters, evidenced by a large negative value of the real part (ɛ₁) of the dielectric function beyond the zero-crossing wavelength at λ ∼ 600 nm (Fig. 1(d)). Note that TiN has shown many favorable properties for plasmonics as compared to conventional noble metals, including low cost, broadband plasmonic response, complementary metal-oxide-semiconductor (CMOS) process compatibility, and high-temperature stability [66–69]. The TiN thin-film coated AAO absorbers should therefore be more suitable for practical applications especially in various harsh environments than the metal nanoparticle-based plasmonic absorbers. Also note that ALD is employed in this work for TiN deposition, as it allows for conformally coating a metal or dielectric layer over AAO and various three-dimensional hollow nanostructures even with very small openings [70,71], thereby overcoming the difficulty of coating the interior surfaces of deep vias or trenches that would have been encountered if using other deposition methods such as reactive sputtering.

 

Fig. 1. Broadband TiN disordered metasurface perfect absorbers. (a) Schematic of AAO nanotemplate with Al back reflector on glass substrate. (b) Top-view SEM of fabricated AAO nanotemplate. Numerical simulations in this work are constructed based on the geometry of nanopore highlighted by white solid curve. Inset: corresponding false-colored cross-sectional SEM. (c) Same as (a), but coated conformally with TiN. (d) Real (ɛ₁) and imaginary (ɛ₂) parts of extracted complex dielectric function of TiN films.

Download Full Size | PDF

The unpolarized reflectance of the Al backed AAO nanotemplates with and without the TiN conformal coating is experimentally characterized using a variable angle spectroscopic ellipsometer, by averaging the measured p- and s-polarized reflectance at 20° angle of incidence with reflection from a gold mirror as reference. The absorptance is then obtained by using absorptance = 1 – reflectance, as the transmission is completely blocked by the Al ground plane. Figure 2 plots the absorptance spectra of the AAO nanotemplates (thickness = 500 nm) of three different porosity values (P = 9%, 39%, and 57%) with (cyan) and without (black) the TiN coating. Their top-view SEMs before coating are also inserted in the spectra. Without coating TiN, the absorptance of the AAO nanotemplates regardless of their porosity drops gradually from the visible region onward and becomes basically wavelength independent in the entire near-IR region. This trend is easily understood as the influence of the nanopores is most significant when their sizes are comparable with the incident wavelengths, and toward longer wavelengths the nanotemplates can be essentially regarded as an effective composite layer of aluminum oxide (Al₂O₃) and air. The absorptance oscillations in the short wavelength region, most notably for P = 9%, result from interference of multiple reflections and have been observed elsewhere [58]. On the long wavelength side, the absorptance increases from ∼10% to ∼30% (or equivalently, reflectance decreases from ∼90% to ∼70%) when increasing P from 9% to 39%, due to the reduced refractive index of the AAO nanotemplates. When further increasing P to 59% the absorptance instead decreases, as the reflection from the back Al becomes significant. With the conformal TiN coating, the absorptance of all three AAO nanotemplates is considerably enhanced over the entire spectral region of interest (300 < λ < 2500 nm), and the absorptance is clearly higher for the one with larger P. For P = 57%, the TiN-coated AAO nanotemplates exhibit desirable broadband perfect absorption characteristics: the absorptance reaches ∼99% over the entire visible and part of the near-IR region up to λ ∼ 1200 nm, and maintains at least 92% for the longer wavelengths. The demonstrated perfect absorption bandwidth is much greater than that of many other metasurface perfect absorbers, and could potentially be expanded to cover the mid-IR wavelengths upon optimizing the absorber structure. For P = 39% and 9%, the absorptance is higher than 90% over the entire visible region, and drops to ∼80% and ∼60% in the long wavelength region, respectively. Clearly, the absorptance of the TiN-coated AAO absorbers depends strongly on the porosity of the AAO nanotemplates. It is worth noting that their absorptance should also be greatly influenced by the disorder of the AAO nanopores, but here the disorder of the three AAO nanotemplates is basically the same, evidenced by the similar diffusivity of the Fourier-transform images (see blue insets to Fig. 2) of their top-view SEMs. The impact of the disorder on the absorption of our TiN-coated AAO nanotemplates will be discussed later in detail.

 

Fig. 2. Absorptance spectra of AAO nanotemplates with (cyan) and without (black) TiN coating at different porosity values: (a) P = 9%, (b) P = 39%, and (c) P = 57%. Insets: top-view SEMs and corresponding Fourier transform images. All scale bars are 500 nm.

Download Full Size | PDF

The demonstrated perfect absorption of the TiN-coated AAO nanotemplates results from both the strong absorption of the highly plasmonic TiN films and the efficient light management of the disordered AAO nanotemplates. The latter is particularly critical, as from our numerical simulations (not shown) the planar TiN films absorb only ∼30% of the incident light over the entire spectral region of interest. In the TiN-coated AAO absorbers, we expect that the incident light before getting absorbed could partly undergo strong scattering in the AAO layer, and partly experience multiple reflections between the top TiN and the back Al mirror. To estimate quantitatively their separate contributions to the observed perfect absorption, we characterize both the reflectance and the transmittance of the TiN-coated AAO absorbers of the same porosity (i.e., P = 57%) but without the back Al mirror, so that the aforementioned multiple reflections can be minimized. Here the unpolarized transmittance is obtained by averaging the measured p- and s-polarized transmittance under normal incidence with transmission through bare glass substrate as reference. As shown in Fig. 3, by removing the back Al, the uncoated AAO nanotemplates transmit a significant portion of the incident light (> 50%, black solid curve in Fig. 3(b)), while reflecting only a few percent (< 10%, red solid curve in Fig. 3(a)) over nearly the entire spectral region of interest (500 < λ < 2500 nm). The scattering caused by the disordered AAO layer (cyan solid curve in Fig. 3(c)) is therefore ∼30 − 40%, obtained by using scattering = 1 – reflectance – transmittance. These experimental results are validated by full-wave numerical simulations using commercial software COMSOL Multiphysics, where the AAO layer is replaced by a planar Al₂O₃ layer of the same thickness (i.e., 500 nm) with a refractive index of 1.67: though no obvious change is observed in reflectance (red dashed curve in Fig. 3(a)), the transmittance of the Al₂O₃ layer (black dashed curve in Fig. 3(b)) increases by ∼30 − 40%, with the scattering (cyan dashed curve in Fig. 3(c)) decreasing from ∼30 − 40% to zero. Upon coating TiN, the reflectance of the AAO nanotemplates (red solid curve in Fig. 3(d)) slightly increases as compared to that of the uncoated ones (red solid curve in Fig. 3(a)), and the transmittance (black solid curve in Fig. 3(e)) drops to zero except for 400 < λ < 800 nm where a fairly small transmittance peak (see inset to Fig. 3(e)) is observed. It can also be seen that both spectra exhibit essentially the same spectral characteristics as the simulated reflectance (red dashed curve in Fig. 3(d)) and transmittance (black dashed curve in Fig. 3(e)) of planar TiN/Al₂O₃ thin-film stacks on glass substrate (in the simulations the complex dielectric function plotted in Fig. 1(d) and a refractive index of 1.5 are used for TiN and glass, respectively), as well as the experimental data of the TiN thin films sputtered on quartz substrate reported elsewhere [72]. The absorptance (cyan solid curve in Fig. 3(f)) shows a broadband feature similar to that observed for the TiN-coated AAO with back Al (see cyan solid curve in Fig. 2(c)): the absorptance is greater than 90% in the short wavelength region (300 < λ < 750 nm), and decreases to ∼70% when moving toward longer wavelengths. By comparing the absorptance spectra of the TiN-coated AAO with and without the Al back reflector (see Fig. S1 in Supplement 1), it is clear that with the Al back reflector, the absorptance is enhanced by ≤ 10% for λ < 750 nm and up to ∼20% at longer wavelengths, due to the increased optical path length of the incident light within the AAO nanotemplates. By further comparing the absorptance of the TiN-coated AAO without the back reflector to the simulated absorptance of the TiN/Al₂O₃ thin-film stack on glass (cyan dashed curve in Fig. 3(f)), it can be inferred that ∼40% of the incident light is scattered due to the strong disorder of the AAO nanotemplates, which agrees well with the data presented in Fig. 3(c). Therefore, light scattering due to the disorder of the AAO nanotemplates plays the predominant role in realizing the broadband perfect absorption of the TiN-coated AAO absorbers shown in Fig. 2(c).

 

Fig. 3. (a-c) Measured reflectance (red solid), transmittance (black solid), and scattering (cyan solid) of uncoated AAO nanotemplate without back Al on glass substrate, respectively. Simulated spectra of planar Al₂O₃ film on glass substrate are also plotted as dashed curves for comparison. Inset to (a): false-colored cross-sectional SEM. Scale bar is 200 nm. (d-f) Same as (a-c), but for TiN-coated AAO nanotemplate. Dashed lines are simulated spectra of TiN/Al₂O₃ thin-film stack on glass. Inset to (e): blow up of the measured transmittance spectra in short wavelength region.

Download Full Size | PDF

 

Fig. 4. (a-c) Simulated absorptance (cyan) and reflectance (black) spectra of AAO nanotemplate backed by Al when coated with TiN on (from top to bottom) all surfaces, pore sidewall, pore bottom, pore sidewall and bottom, and top surface for P = 57%, 39%, and 9%, respectively. The schematics on the right illustrate where TiN (yellow) is coated on AAO (cyan) in each case. (d) Simulated current density distribution in x-z plane indicated by dashed lines in (e) when all surfaces of AAO are coated with TiN for P = 57% at incident λ = 500, 800, and 1800nm, respectively. (e) Same as (d), but in x-y plane 10 nm below the top TiN surface.

Download Full Size | PDF

The three-dimensional nature of the AAO nanotemplates, apart from porosity and disorder, also has a strong impact on the absorption characteristics of the TiN-coated AAO absorbers. The TiN deposited on the top surface of the AAO nanotemplates and the bottom of the nanopores forms respectively plasmonic arrays of nanoholes and nanodisks, which could introduce propagating and localized plasmonic resonances. The TiN wrapping around the nanopore sidewall acting essentially as plasmonic arrays of nanoshells could also lead to specific resonant modes. We thus perform full-wave numerical simulations using COMSOL Multiphysics to elucidate the absorption characteristics of the AAO nanotemplates when coated with TiN films (thickness = 25 nm) on (Fig. 4(a)-(c) from top to bottom) all surfaces, pore sidewall, pore bottom, pore sidewall and bottom, and top surface for P = 57%, 39%, and 9%, respectively. In the simulations, a single nanopore in a unit cell with periodic boundary condition and extra fine mesh setting in lieu of random nanopore arrays is employed due to our limited computational resources. The shape of the nanopore without loss of generality is delineated based on the one enclosed by the white solid curve in the top-view SEM in Fig. 1(b), and its size is properly scaled in a fixed unit cell (size = 300 nm × 300 nm) for different values of P. It can be seen later that the shape of the nanopores does not significantly alter the overall absorption characteristics of the TiN-coated AAO nanotemplates. The thicknesses of the Al ground plane and the AAO barrier layer are 100 nm, and the height of the nanopore is 400 nm. For P = 57%, the simulated absorptance spectrum of the AAO nanotemplates with TiN coated only on the pore sidewall (second panel from the top in Fig. 4(a) well reproduces the broadband absorption feature of the experimental data shown in Fig. 2. Further coating TiN on the pore bottom (second panel from the bottom in Fig. 4(a)) only makes a slight difference in the absorptance spectrum, thus implying that TiN on the pore bottom is not critical for realizing broadband perfect absorption. However, the absorptance apparently decreases (especially at long wavelengths) when the top surface of the AAO is also coated (that is, all surfaces of the AAO are coated with TiN; see top panel in Fig. 4(a)) due to the increased reflectance (note that the reflectance spectrum in this case resembles that for a planar TiN film). On the other hand, when only the pore bottom is coated with TiN (middle panel in Fig. 4(a)), although two pronounced absorptance peaks are observed in the near-IR region, the absorptance in most of the visible region is < 50%. When only the top surface is coated (bottom panel in Fig. 4(a)), strong oscillations in both the absorptance and the reflectance spectra are clearly observed, resulting from the interference of multiple reflections between the top (TiN) and the bottom (back Al) mirrors. Note that in this case the absorptance and reflectance spectra are essentially the same as those for TiN/Al₂O₃/Al trilayer thin-film stacks on glass (see Fig. S2 in Supplement 1). Therefore, it is evident that the TiN on the pore sidewall plays the most important role in realizing broadband perfect absorption. This is further validated by simulated current density distribution in both x-z (Fig. 4(d)) and x-y (Fig. 4(e)) planes when all surfaces of the AAO (P = 57%) are coated, where most of the induced currents are found in the TiN on the pore sidewall, at incident λ = 500, 800, and 1800 nm, respectively. Our simulation results also confirm that at these wavelengths the electric field distribution of the AAO nanotemplates when coated with TiN on all surfaces, pore sidewall, and pore sidewall and bottom are all alike (see Fig. S3 in Supplement 1). These findings highlight the advantage of using ALD over sputtering for the fabrication of broadband perfect absorbers based on three-dimensional AAO nanotemplates [73]. When reducing the porosity to P = 39% and 9% (Fig. 4(b) and (c)), not only does the absorptance decrease but the absorption bandwidth becomes much narrower, which agrees with our experimental results shown previously.

Despite the overall good agreement between experimental results and numerical simulations, we find that the simulated absorptance of the TiN-coated AAO nanotemplates at near-IR wavelengths (see top panel in Fig. 4(a)) is apparently lower than that in the experiments (see Fig. 2(c)). As the maximum length of the irregular nanopore opening in the simulations is only ∼280 nm for P = 57%, this discrepancy should not result from the specific shape of the nanopore nor the incident polarization direction, evidenced by basically the same absorptance spectra when the geometry of the nanopore is changed to circle or the incident polarization is rotated by 90° (Supplement 1, Fig. S4). Moreover, seemingly from Fig. 2 one could easily relate the variation of the long-wavelength absorptance to porosity, yet it should be pointed out that the porosity of the nanotemplates does not cause the observed difference either. As can be seen in Supplement 1, Fig. S5(c), when there is absolutely no disorder (that is, AAO nanopores are perfectly circular and periodically arranged into a square array), the long-wavelength absorptance of the TiN-coated AAO nanotemplates remains lower than the measured values even with porosity as high as 68%. For smaller porosity (P = 7% and 25%, see Supplement 1, Figs. S5(a) and (b), respectively), the absorptance spectra become oscillatory and are similar to that for TiN/Al₂O₃/Al trilayer thin-film stacks (Supplement 1, Fig. S2). It is worth noting that the situation in Supplement 1, Fig. S5 is different than the experimental results shown in Fig. 2, which exhibits the enhancement of the absorptance of the TiN-coated AAO nanotemplates by increasing the porosity, when strong disorder of the nanopores is present. We believe that it is lack of long-range disorder comparable to the near-IR wavelengths in the simulations that leads to the observed difference in the absorptance spectra. It has been shown previously that both the absorptance level and the operation bandwidth of a metasurface absorber can be enhanced by significantly increasing the position disorder of its constituent plasmonic resonators [74]. We expect that the long-wavelength absorptance could be higher in the simulations by further introducing the disorder in a supercell containing multiple randomly distributed nanopores. We can further infer from our simulations that based on plasmonic metal-nitride film coated AAO nanotemplates, perfect absorption in the visible region could be readily realized by simply using AAO with periodically arranged nanopores (or by using lithographically patterned periodic hole arrays in a photoresist layer) as long as the porosity is optimized. However, strong disorder of the nanopore arrangement is unquestionably required to further expand the high absorption bandwidth to near-IR wavelengths.

For electromagnetic wave absorbers weak angle dependence on the absorption characteristics is highly desirable for practical applications. We characterize the absorptance of the TiN-coated AAO nanotemplates with back Al at incident angles from 20° to 70°. The results plotted in Fig. 5 demonstrate that our TiN-coated AAO nanotemplates possess excellent wide-angle broadband absorption performance. In the visible region, the absorptance is greater than 95% at incident angles up to 60°, and slightly drops to at least 86% at 70° incidence. The absorptance at large incident angles remains high when moving toward the near-IR region from the visible. At λ = 1200 nm, the absorptance is 88% and 76% at 60° and 70°, respectively. Moving further toward longer wavelengths, the absorptance decreases a bit further, to 70% at 60° and 60% at 70° for λ = 2500 nm. The observed decrease in absorptance when increasing incident angle results from higher reflectance due to larger impedance mismatch. As compared to other broadband metasurface perfect absorbers based on hybridization of multiple resonant modes or multilayer thin-film stacks, our TiN-coated AAO absorbers should exhibit more robust angular insensitivity due to the non-resonant absorption nature. Note that their broadband absorption characteristics are also independent of incident polarization states, evidenced by essentially the same absorptance for p- and s-polarization in the experiments (not shown) and the simulations (see Supplement 1, Figs. S4(a) and (b)). The demonstrated wide angle performance and polarization insensitivity, along with other favorable features such as easy and scalable fabrication, low cost material, and high thermal and chemical stability, make the TiN-coated AAO nanotemplates superior to many other types of broadband perfect absorbers previously reported, and suitable for a wide variety of applications, particularly for solar energy harvesting in which high temperature (> 1000°C) is required in order to achieve sufficiently high energy conversion efficiencies [19].

 

Fig. 5. Measured absorptance of TiN-coated AAO nanotemplate backed by Al with P = 57% as a function of angle of incidence and wavelength.

Download Full Size | PDF

Lastly, to demonstrate the TiN-coated AAO nanotemplates can in practice be employed as highly efficient broadband solar absorbers, in a proof-of-concept experiment we place a TiN-coated AAO nanotemplate sample along with an uncoated one (both without Al back reflector) under a Xenon arc lamp based solar simulator (Newport Inc.), and closely monitor their temperature change under illumination by using a hand-held IR imaging inspection camera (FLIR Inc.). As can be seen in Fig. 6, before the solar simulator is turned on, the temperatures of both samples are quite similar. After 3 minutes of illumination, it is clear that the TiN-coated AAO nanotemplate exhibits a much greater temperature rise as compared to that for the uncoated sample. This demonstration further verifies the spectroscopic data comparing the absorptance of the AAO nanotemplates with and without the TiN coating shown in Fig. 2 and 3. The observed solar-to-thermal conversion using TiN-coated AAO nanotemplates can be of use to greatly increase the evaporation rate of water in the desalination or purification process [56,57]. Moreover, the TiN-coated AAO nanotemplates could be integrated in a photoelectrochemical cell to enhance the hydrogen production from water splitting for solar energy storage [75]. The TiN-coated AAO nanotemplate backed by Al on its own right could also be employed as a high-efficiency broadband photodetector in the visible and near-IR regions, by utilizing the enhanced hot carrier excitation in the plasmonic TiN film.

 

Fig. 6. IR images of AAO nanotemplate with (larger sample on the left) and without (smaller sample on the right) TiN coating (a) before and (b) after 3 minutes of simulated solar illumination, showing much greater temperature rise for the TiN-coated AAO nanotemplate.

Download Full Size | PDF

3. Conclusion

In conclusion, a broadband TiN disordered metasurface perfect absorber is demonstrated in a TiN thin-film coated AAO nanotemplate structure, based on the strong non-resonant absorption of highly plasmonic TiN films and the efficient light management of disordered AAO nanotemplates. The key to realizing broadband perfect absorption in this structure is fully elucidated by both experimental results and numerical simulations. The demonstrated TiN-coated AAO absorbers are superior to many other types of broadband perfect absorbers previously reported, and could potentially be employed for various applications including solar thermophotovoltaics, water purification, solar energy storage, and photodetection.