A recent study in Scientific Reports presented a graphene-based metamaterial as a solar absorber. The structure consisted of three layers: aluminum (Al) as the resonator, titanium nitride (TiN) as the middle substrate, and iron (Fe) as the ground layer. Machine learning (ML) techniques were used to optimize these materials for efficient solar thermal energy conversion.
Image Credit: PixieMe/Shutterstock.com
Background
To address the greenhouse effect of traditional power sources, various renewable energy technologies have been developed. Among these, solar power systems are widely used.
Solar absorbers made of Ti and Al exhibit conversion efficiencies of up to 95.2 % for solar thermal energy. In contrast, selective solar absorbers optimized using multi-island genetic algorithms achieve emission efficiencies of up to 96 %. The performance of solar devices can be further improved by incorporating Fe into a two-dimensional absorber structure. The properties of Fe, including its malleability, high strength, and hardness, make it suitable for solar absorber applications.
The proposed absorber design included Al, TiN, and Fe, with refractive indices of 3.212, 1.2887, and 2.865, respectively. Al was selected for its corrosion resistance, which protects the device from environmental factors such as humidity and temperature variations. TiN was used to enhance the retention of solar thermal energy absorbed by the Al resonator.
Methods
The Finite Element Method (FEM) was applied using COMSOL software to design the solar absorber. Initially, a 100 nm Fe ground layer was designed, which showed absorption rates of 78.47 % in the ultraviolet (UV), 82.83 % in the visible (Vis), and 84.01 % in the near-infrared (NIR) spectrum, with a maximum absorption of 92.1 %. The next 700 nm TiN substrate layer exhibited absorption rates of 69.98 % in UV, 72.32 % in Vis, and 81.48 % in NIR, with a maximum absorption of 96.36 %.
The Al resonator, without cylinders (500 nm), showed absorption rates of 88.86 % in UV, 90.36 % in Vis, and 91.97 % in NIR, with a maximum of 99.91 %. A complete Al resonator (without graphene) achieved absorption rates of 89.1 % in UV, 91.45 % in Vis, and 95.41 % in NIR, with a maximum of 99.95 %. The complete graphene absorber (0.34 nm graphene) was designed to show absorption rates of 90.1 % in UV, 92.08 % in Vis, and 95.6 % in NIR, with a maximum of 99.95 %. A lithographic process was used for the sequential deposition of Fe, TiN, graphene, and Al.
Absorbance values based on structural thickness variations were predicted using a local regression model (KNN). The model's performance was evaluated using the coefficient of determination (R2) and mean square error (MSE).
Results and Discussion
Varying the thickness of the Al resonator and Fe layer, from 100 to 500 nm, led to a decrease in absorbance from 92.94 % to 88.85 % and from 92.94 % to 85.71 %, respectively, at 2 µm. Increasing the TiN substrate thickness from 500 to 700 nm caused a decrease in absorbance from 97.14 % to 85.15 % at 2 µm.
Altering the absorber width from 360 to 410 nm decreased absorbance from 92.94 % to 88.85 % at 2 µm. Conversely, changing the radius of the resonator cylinders from 10 to 50 nm increased absorbance from 59.57 % to 62.62 % at 2 µm. The graphene layer also increased the absorption of the solar structure’s dielectric layer.
The effect of incident angle was assessed by varying it from 0 to 70 degrees to examine polarization effects. Both transverse electric and magnetic modes exhibited similar behavior, indicating insensitivity to polarization. The symmetrical structure of the absorber confined electromagnetic light to the dielectric part, achieving near-perfect absorption in the visible spectrum. Notably, electric and magnetic field radiation was more prominent in the upper part of the resonator compared to the substrate and ground layers.
ML optimization significantly reduced process time and simulation requirements. ML-based optimization of the multi-layer solar absorber structure took one-eighth of the time needed for traditional methods.
Conclusion
the study demonstrated the successful development of a solar absorber with a three-layer composition: an Al resonator, TiN middle substrate, and Fe ground layer. The absorber exhibited an absorbance of over 97 % for the 1.5–2.5 µm wavelength range and over 95 % for the 0.5–2.5 µm wavelength range.
The performance of the absorber was validated at wavelengths of 0.29, 0.58, 1, and 2 µm, with corresponding radiation output analyzed across the UV, Vis, and NIR regions. The proposed ML-based approach also reduced the optimization time by up to one-eighth compared to traditional methods. This methodology can assist in the design of solar absorbers for a range of applications, including water heating, lighting, ventilation, and electric vehicles.
Journal Reference
Aliqab, K., Han, B. B., Kumar, O. P., Alsharari, M., Armghan, A., Patel, S. K. (2024). Graphene metamaterial solar absorber using Al-TiN-Fe for efficient solar thermal energy conversion and optimization using machine learning. Scientific Reports. DOI: 10.1038/s41598-024-80485-0, https://www.nature.com/articles/s41598-024-80485-0
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.