1. Introduction
Energy issues have always been important issues to be solved on Earth. The development of clean and renewable energy can promote the sustainable development of China and the world [1–3]. How to use large reserves, renewable and clean solar energy has become one of the directions explored by the scientific community. At present, there are many ways to convert and utilize solar radiation. For example, detection and sensing [4, 5], photovoltaic cells [6], desalination seawater treatment [7, 8], heat conversion utilization [9], photoelectric detection [10–13], etc. Solar absorbers play an important role in the above fields. Therefore, devices that can efficiently absorb solar energy are extremely important. From the early study of single-wave absorbers to the present study of ultra-wideband perfect absorbers, it is a long-term pursuit of researchers to achieve perfect capture in the full spectral range [14–17].
The absorber composed of metal-dielectric layer-metal has been widely studied [18–22]. In order to obtain better absorption intensity and wider absorption bandwidth, precious metal materials such as Au and Ag are often used in the production of absorbers, because precious metal plasma materials can produce strong plasma resonance and optical coupling [23–25]. However, its disadvantages are also obvious, such as high price, low temperature resistance and easy deformation. This refractory metal has become our new solution. Titanium is abundant on Earth, and Ti and its nitrides are widely used [26–28]. TiN has many characteristics such as a high melting point (2950 °C), high hardness (Mohs hardness 8–9), good thermal shock resistance and good corrosion resistance [29]. Therefore, we choose TiN plasma material for the study of ultra-wideband perfect absorbers and thermal emitters. Apart from that, we also used W with a melting point of 3427 °C and si3n4 with a melting point of 1900 °C, which are fire-resistant and heat-resistant, respectively. These materials are capable of efficient resonance absorption in the near-ultraviolet to near-infrared spectral range.
Thermal radiation is a way of conducting heat generated by electromagnetic waves radiated by objects due to temperature. The four laws of Kirchhoff's law of radiation, Planck's law of radiation distribution, Stephen–Boltzmann's law and Wien's law of displacement describe thermal radiation in detail [30]. Thermal radiation does not depend on any medium and can be carried out in vacuum. Solar air collectors, thermal imagers and spectrometers based on thermal radiation are used in various fields [31, 32].
According to the above ideas, we introduce an ultra-wideband perfect solar absorber and thermal emitter, which is composed of TiN and Si3N4 periodic nanodisk arrays, Ti and W double-layer substrates. Under the structure of the nanodisk, our absorber undergoes strong surface plasmon resonance and provides near-field coupling between the disks, which achieves efficient absorption performance for us. After simulation calculation, the average absorption rate obtained in the wavelength range of 200 nm–3000 nm reaches 95.93%, and the average absorption rate is 97.17% in the specific band of 2533 nm (A > 90%). After calculation, we are surprised to find that even at a temperature of 1500 K, the thermal emission efficiency reached 95.37%. Under the condition of atmospheric mass AM1.5, there is still a weighted average absorption rate of 98%, and the absorption loss was less than 2%. The near-perfect blackbody radiator is realized at a higher temperature. It provides ideas for future exploration of thermal emitters. In addition, the proposed structure is also insensitive to angle and has polarization independence [33, 34], which will be discussed in the future.
2. Structure design
This paper proposes a six-level ultra-wideband perfect absorber design. Figure 1, the absorber has a multi-layer nanodisk array, which is composed of W and Ti as the substrate and a double-layer TiN-SiO2 array film. After the finite difference time domain (FDTD) calculation, we evaluated the absorption performance of the model [35–37]. This method can also easily calculate the electromagnetic field of any material [38]. We select the perfect matching layer on the Z axis and use periodic boundaries on the X and Y axes. The mesh accuracy is set to 4.5, and taking 0.25 nm as the calculation step size. In figure 1(b), the heights of different nanodisks are H1 = 650 nm(TiN), H2 = 85 nm(Si3N4), H3 = 250 nm(TiN), H4 = 40 nm(Si3N4), Substrate film thickness are H5 = 100 nm(Ti), H6 = 150 nm(W). In figure 1(c), the period of the structure is P = 500 nm, the radius of the disk is R1 = 160 nm, and R2 = 80 nm. The absorption efficiency can be obtained by the calculation formula A = 1-T-R [39]. When the film thickness is thick enough, the reflectivity R can be ignored, so the absorption rate A is 1-R [40–42].
Figure 1. (a) Three-dimensional space diagram of nanodisk structure. (b) The material height distribution in the XOZ two-dimensional plane.(c) The XOY two-dimensional distribution map. (d) Solar absorber absorption efficiency diagram. |
3. Results and discussion
3.1.1. Absorption and bandwidth
We select the plane light containing ultraviolet light to infrared light (200 nm–3000 nm) as the incident wavelength. The results in figure 1(d) are obtained by FDTD simulation. The average absorption performance of the full-band (200 nm–3000 nm) is 95.93%. Over the wavelength span of 200 nm to 2733 nm (with a bandwidth of 2533 nm), it was observed that the solar absorber achieved an absorptivity of over 90%, with an average absorptivity of 97.17% in this specific range of wavelengths. We can also clearly see the strong absorption in visible light (380 nm–760 nm), and the average absorption is calculated to be 98.12%. A sufficiently thick base makes the overall transmittance basically 0. The main loss of absorption performance is concentrated in the far-infrared band after 2733 nm. In the absorption spectrum, we selected five wavelengths with high absorption rate, near-ultraviolet wavelength λ1 = 355.5 nm, visible wavelength λ2 = 504 nm and λ3 = 633 nm, near-infrared wavelength λ4 = 1048 nm and λ5 = 1324 nm, and the corresponding absorption rates were 99.62%, 99.84%, 99.97%, 99.86%, and 99.96%, respectively. The subsequent performance measurement of the above wavelengths can help us analyze how the absorber of this structure achieves ultra-wideband perfect absorption.
Table 1. Absorption performance of different absorbers. |
| Quote | Structure | Material | Absorption bandwidth (nm) | Full-band absorption efficiency |
|---|---|---|---|---|
| [43] | Biperiodic lattice array | Ti | 400–2000 | 91.4% |
| [44] | Cube periodic array | Ti | 354–1066 | 97.0% |
| [45] | Cone | TiN | 400–1500 | 99.6% |
| [46] | TiN disk square ring array | TiO2, TiN | 250–3000 | 94.0% |
| propose | Nanodisk periodic array | TiN | 200–3000 | 95.93% |
3.2.1. Field intensity distribution
Measuring the field strength on the absorber can help us understand how the absorber polarizes to achieve ultra-wideband absorption and obtain high thermal radiation intensity [47, 48]. Under different bands, we selected five wavelengths that perform better in absorption for field strength analysis. Figures 2(a) and (b) are the electric field intensity distribution maps of XOY plane and XOZ plane with near-ultraviolet wavelength λ1 = 355.5 nm. It can be clearly seen from the observation results that the outer surface of TiN nanoplates and Si3N4 thin films exhibit strong surface plasmon resonance. Figure 2(b) shows that the electromagnetic field is also observed in the outer region of the absorber, indicating that the plasma array excites the lattice resonance in a period of 500 nm. In figures 2(c) and (d), under the incident of visible light λ2 = 504 nm, the outer surface of the first layer of TiN nanodisks accumulates a high-intensity electric field, and there is also obvious strong coupling between adjacent disks. This is because the gap between the disks can be regarded as a cavity [49, 50]. After the incident light enters, the cavity membrane resonance is generated between the cavities. The generation of cavity mode resonance makes the energy bound between the disk gaps and the interaction between the cavities is strengthened. Figures 2(c) and (f) corresponds to the case of visible light wavelength λ3 = 633 nm, the strong effect occurs on the corner surface of TiN disk. Figure 2(g)–(j) are near-infrared light at 1048 nm and 1324 nm, respectively. There is a strong electric field on the corner surface of the third TiN nanodisk. Beyond the infrared wavelength range, the elongation of optical wavelengths results in an expansion of the resonant area for the third layer of TiN nanodisks and an augmentation of resonance between neighboring disks. Through the observation of these wavelengths, we can know that the reason why the absorber achieves ultra-wideband perfect absorption is that the structure has a strong effect on the near-field coupling of the structure [51].
Figure 2. (a), (c), (e), (g) and (i) are the electric field strength calculated on the XOY plane. (wavelength λ1–λ5). (b), (d), (f), (h) and (j) are the electric field strength calculated on the XOZ plane (wavelength λ1–λ5). |
3.3.1. Absorption energy analysis and heat dissipation efficiency analysis
By calculating the spectral absorptivity of the solar absorber can help to evaluate its heat collecting performance [52, 53]. At atmospheric mass AM1.5, the absorption efficiency (ŋA) of the solar absorber can be described as formula 1 [54]
$\begin{eqnarray}{{\rm{\eta }}}_{{\rm{A}}}=\frac{{\int }_{{\lambda }{\rm{M}}{\rm{i}}{\rm{n}}}^{{\lambda }{\rm{Max}}}{\rm{A}}\left({\rm{\omega }}\right){{\rm{I}}}_{{\rm{AM}}1.5}\left({\rm{\omega }}\right){\rm{d}}{\rm{\omega }}}{{\int }_{{\lambda }{\rm{M}}{\rm{i}}{\rm{n}}}^{{\lambda }{\rm{Max}}}{{\rm{I}}}_{{\rm{AM}}1.5}\left({\rm{\omega }}\right){\rm{d}}{\rm{\omega }}}.\end{eqnarray}$
In figure 3(a), the red shows the energy spectrum of atmospheric mass 1.5, and the grey shows the absorption spectrum. Figure 3(b) shows the energy loss (red) and absorption (black) of the absorber in the air. Through calculation, we found that the incident light is at 280 nm–3000 nm, the weighted average absorption rate reached more than 98%, and the absorption loss was less than 2%. This data further proves the excellent absorption effect of the absorber.
Figure 3. (a) and (b) are the distribution map of absorption energy and loss energy in the light range of 280 nm–3000 nm under air quality (AM) 1.5. |
Thermal divergence efficiency (ŋE) is an important parameter of the thermal emitter [55]. It can be used to evaluate the heat emission of the absorber system. According to Planck's blackbody radiation law in formula 2 , the calculation formula of thermal emission efficiency (ŋE) is shown. In formula 2 , the ideal blackbody spectral intensity is at frequency ω and temperature T [56]
$\begin{eqnarray}{\eta }_{E}=\frac{{\int }_{{\lambda }_{{m}_{i}n}}^{{\lambda }_{\max }}\varepsilon \left(\omega \right)\cdot {I}_{{BE}}\left(\omega ,T\right){\rm{d}}\omega }{{\int }_{{\lambda }_{{m}_{i}n}}^{{\lambda }_{\max }}{I}_{{BE}}\left(\omega ,T\right){\rm{d}}\omega }.\end{eqnarray}$
In figure 4(a), the heat emission characteristics of the heat emitter at 1500 K were analyzed. At wavelengths below 2000 nm, the thermal emitter has almost perfect thermal emission performance. The calculated thermal emission efficiency of the absorber at 1500 K temperature reaches 95.37%. It can be concluded that blackbody radiators and light sources based on refractory metals are possible, and further research directions are provided for the exploration of thermal emitters in the future.
Figure 4. (a) The thermal emission images of the absorber at 1500 k. (b), (c) and (d) correspond to the thermal emission images of case3-case5 at 1500 k, respectively. |
3.4.1. The performance difference of absorber under different structures
This paper also studies the absorption performance of the absorber under different structures in order to prove that our structure is the best. In case1 and case2 of figure 5(a), we removed the top two layers of nanodisks with a radius of 80 nm and the middle R2 = 160 nm, respectively, and in case3 and case4 we unified the radius of the nanodisks. The substrate structure parameters of the above five cases are the same as those of case6. Figure 5(b) is the absorption effect diagram corresponding to the six cases. In the case of six different colors, the structural absorption bandwidth and absorption efficiency of case6 is significantly better than those of the other cases. The absorption of the other cases in the near-infrared band is significantly lower. We know that electric field distribution analysis can provide a detailed explanation of the relevant physical mechanisms [57, 58]. Here, we also selected five absorption peaks of case1 λ = 1332 nm, case2 λ = 600 nm, case3 λ = 614 nm, case4 λ = 1610 nm and case5 λ = 518 nm to analyze their electric field distribution. As shown in figure 6 (k)–(t), the strong electric fields of case1-case4 are all on the surface of the nano-disc corner, and case5 has a strong electric field on the outer surface of TiN and the interface of Si3N4 film. In addition to the absorption rate, we also calculated the thermal emission performance of case3-case5. The absorption performance of case1 and case2 is obviously too poor to be calculated. As shown in figure 4, the thermal emission efficiency of (b)–(c) case3, case4 and case5 is 59.20%, 80.04% and 74.37%, respectively. The thermal emission performance of the three cases is weaker than that of the proposed structure. It basically proves the excellent absorption performance and thermal emission performance of this structure design.
Figure 5. (a) Two-dimensional images of absorbers with different structures. (b) Absorption curves corresponding to the six structures (200 mn–3000 nm). |
Figure 6. (k)–(t) are the electric field intensity images corresponding to different absorption peaks of case1-case5 (XOY plane and XOZ plane). |
3.5.1. Performance differences caused by changes in nanodisk height and radius parameters
By changing the thickness and radius of the nanodisk of the absorber, the influence of different thickness and radius on the absorption performance is calculated. In figures 7(a) and (b), the thickness of the top TiN layer (H1) and the thickness of the third TiN nanodisk (H3) are changed, respectively. Figure 7(a) shows the case when we change the first layer thickness to make H1 change from 600 nm to 700 nm, the absorption performance is slightly affected in the visible band, and there is a significant change in the near-infrared band, but the overall absorption performance remains above 90%. Compared with H1 = 650 nm, whether H1 increases or decreases, the absorption performance decreases significantly after the wavelength increases to 1000 nm. Figure 7(b) shows the increase of the third layer of nanodisk from 200 nm to 300 nm. With the increase of H3, the absorption of the infrared band decreased significantly, and the absorption effect was the worst when H3 reached 300 nm. After the wavelength reaches 1000 nm, with the increase of H1 and H3, the effect of light absorption is gradually improved, and the maximum bandwidth is also changed. Considering the overall light absorption and maximum bandwidth, we chose H1 = 650 nm and H3 = 250 nm.
Figure 7. (a) The absorption transformation caused by the increase of nanodisk height H1 from 600 nm. (b) The absorption transformation caused by the increase of nanodisk height H3 from 200 nm. (c) The absorption transformation caused by the increase of nanodisk radius R1 from 70 nm. (d) The absorption transformation caused by the increase of nanodisk radius R2 from 150 nm. |
We also examined the effects of changes in R1 and R2 at different levels of the disk. In figures 7(c) and (d), R1 varies from 70 nm to 90 nm, R2 varies from 150 nm to 170 nm, and the radius variation interval is 5 nm. In figure 7(c), when R1 is greater than 80 nm or less than 80 nm, the absorption efficiency has obvious attenuation, which is caused by the weakening of the near-field coupling effect. By observing figure 7, we can find that with the increase of R2, the absorption bandwidth also shows an increasing trend. This is because as the radius increases, the structure spacing of the bottom layer becomes closer, which in turn strengthens the plasma effect [59–61]. Finally, considering the bandwidth and absorption factors, we determined R1 = 80 nm and R2 = 160 nm.
3.6.1. Angle sensitivity analysis
For metasurface absorbers, the sensitivity of the incident angle determines many of the device's performance. Generally speaking, the less sensitive the incident angle, the better its performance [62–65]. To evaluate the performance of the absorption material, we proposed in real-world applications in figure 8(a), we performed absorption performance calculations (TE, TM) in different modes. Due to the symmetry of the structure, the absorptions of the two curves (TM and TE) are completely coincident. In addition, in figure 8(b), we also studied the absorption spectrum at the polarization angle of 0°–90°. The results show that the absorption intensity remains consistent with the change of polarization, which further verifies the insensitivity of the structure to polarization [66, 67]. In figure 8 (c) and (d), we also calculate the absorption efficiency image when the incident angle increases to 60° in TE and TM modes. Due to the high symmetry of our solar absorber structure, the absorption efficiency images in the two modes are consistent. Although the absorption rate changes with the increase of the incident angle, the average absorption rate of the perfect absorber remains above 85%, even if it increases to 60° in the whole band, indicating that the structure has a certain angle-insensitive characteristic.
Figure 8. (a) The absorption result of the absorber in two modes (TE and TM). (b) TE polarization to TM polarization (0° to 90°). (c) and (d) are the absorption effect of the incident light angle of 0°–60° in TE mode and TM mode. |
4. Conclusion
In summary, we propose a TiN-Si3N4 double-layer nanodisk solar absorber and thermal emitter, which can perfectly absorb broadband ultraviolet, visible and infrared light. By using the finite difference time domain (FDTD) method, the following results are obtained: in the incident wavelength range of 200 nm–3000 nm, the average absorption rate of the solar absorber reaches 95.93%. Especially the wavelength in the 400–700 nm has an average absorption of 98.12%. We use refractory metal materials Ti, TiN, W and high temperature resistant dielectric layer Si3N4 to achieve stable operation of the structure when the temperature reaches 1500 K. According to the Planck blackbody radiation law and the atmospheric absorption spectrum formula, the calculation results show that the thermal emission efficiency reaches 95.37% under high temperature conditions. In the case of atmospheric mass of AM1.5, the weighted average absorption rate exceeds 98%, and the absorption loss is less than 2%. The absorber also shows excellent polarization and angle-insensitive characteristics. It maintains a highly consistent absorption efficiency at 0°–90° polarization. In the TE and TM modes, the incident light angle reaches 60° and maintains an absorption efficiency of more than 85%, once again proving that our proposed solar absorbers and thermal emitters have good physical properties.