Nanophotonics
Artificial Intelligence
5 Feb 2021

Inverse design of nanostructured optical solar reflectors

Optical solar reflectors (OSRs) are structures placed on the external surface of spacecraft that simultaneously reflect incoming sunlight and emit infrared radiation (heat). These structures serve two thermal management needs at the same time: dissipation of heat caused by illumination from the Sun and the excessive internal heat generated by onboard components.

Figure 1: Conventional optical solar reflector based on metallised quartz.
Figure 1: Conventional optical solar reflector based on metallised quartz.

Since the 1960s, metallised quartz tiles have been used for OSRs, where the quartz emits infrared radiation and lets the visible light through to the metal back reflector (Fig. 1) [1]. However, the weight, cost, brittleness, and rigidity of conventional OSRs leave room for improvement [2,3]. Flexible silvered polymers are one alternative to quartz tiles, but they are inferior to quartz OSRs in dissipative performance and quickly degrade [3]. Recently, a few groups have started investigating other novel materials and designs for OSRs. Such innovations include those based on metasurfaces [2-4], thermochromic phase-change materials with temperature-dependent emission [5-7], or both [8].

The dissipative performance of optical devices with metasurfaces or nanostructures can be determined using different simulation methods like, for instance, rigorous coupled-wave analysis (RCWA). Usually, the performance is found through a forward design process (Fig. 2a). To acheive better performance, the scientist must experiment with small changes in the device geometry to incrementally improve the reflection and emission spectra throughout the repeated simulation procedure. However, recent advances in machine learning techniques enable inverse design procedures that promise more efficient device development than the time-consuming forward processes (Fig. 2b) [9-14]. Thus, high-performance OSRs that expand the design options for future spacecraft may be constructed.

Figure 2: Overview of design processes. (a) Conventional forward design process. (b) Conventional inverse design process. Results from (a) are fed into the neural network that subsequently finds the geometry for the desired spectrum. (c) Inverse design process with backpropagation. The process is based on an iterative trial-and-error approach where the neural network inputs some initial nanostructure (NS) geometry to the simulation and gets feedback on how far from the desired spectrum it was.
Figure 2: Overview of design processes. (a) Conventional forward design process. (b) Conventional inverse design process. Results from (a) are fed into the neural network that subsequently finds the geometry for the desired spectrum. (c) Inverse design process with backpropagation. The process is based on an iterative trial-and-error approach where the neural network inputs some initial nanostructure (NS) geometry to the simulation and gets feedback on how far from the desired spectrum it was.

Project overview

With this project, we wish to build on prior inverse design research [14] and improve on the design process for nanostructure-based OSRs. Inverse design here refers to the process of creating a model, such as a neural network, to answer which material parameters provide the desired functionality. Using recent neural network architectures, such as Neural Radiance Fields (NeRF) [15] or sinusoidal representation networks [16], we aim to solve this inverse problem with high accuracy for a wide spectrum of wavelengths, ranging from near-ultraviolet to mid-infrared, with a superior inverse design process (Fig. 2c).

The structures suggested by the neural network may not necessarily be obtainable with current manufacturing techniques and might have other shortcomings aggravated by the harsh space environment. Thus, this problem will require an interdisciplinary approach that incorporates experimental validation at a later stage of the project.

In the spirit of the ACT, all research and code will be openly accessible. This study will, therefore, contribute to the democratisation of metamaterials research.

Currently, this project is in progress, with preliminary results being collected. Furthermore, we are actively investigating areas where this framework might be immediately profitable and other ways to expand on our work. One idea we are looking into is the use of this technology for next-generation solar sails [17].

References:

  1. K. N. Marshall and R. A. Breuch, "Optical solar reflector - A highly stable, low alpha sub S/epsilon spacecraft thermal control surface", Journal of Spacecraft and Rockets, 1968, Volume 5 (9), Pages 1051-1056, DOI: 10.2514/3.29420

  2. E. V. Shirshneva-Vaschenko, P. S. Shirshnev, Zh. G. Snezhnaia, L. A. Sokura, V. E. Bougrov, and A. E. Romanov, "Zinc oxide aluminum doped slabs for heat-eliminating coatings of spacecrafts", Acta Astronautica, 2019, Volume 163, Part A, Pages 107-111, ISSN 0094-5765, DOI: 10.1016/j.actaastro.2019.07.005

  3. K. Sun, C. A. Riedel, Y. Wang, A. Urbani, M. Simeoni, S. Mengali, M. Zalkovskij, B. Bilenberg, C. H. de Groot, and O. L. Muskens, "Metasurface optical solar reflectors using AZO transparent conducting oxides for radiative cooling of spacecraft", ACS Photonics, 2018, Volume 5 (2), Pages 495-501, DOI: 10.1021/acsphotonics.7b00991

  4. D. U. Yildirim, A. Ghobadi, M. C. Soydan, O. Atesal, A. Toprak, M. D. Caliskan, and E. Ozbay, "Disordered and Densely Packed ITO Nanorods as an Excellent Lithography-Free Optical Solar Reflector Metasurface", ACS Photonics, 2019, Volume 6 (7), Pages 1812-1822, DOI: 10.1021/acsphotonics.9b00636

  5. R. Beaini, B. Baloukas, S. Loquai, J. E. Klemberg-Sapieha, and L. Martinu, "Thermochromic VO2-based smart radiator devices with ultralow refractive index cavities for increased performance", Solar Energy Materials and Solar Cells, 2020, Volume 205, Article number 110260, ISSN 0927-0248, DOI: 10.1016/j.solmat.2019.110260

  6. H. Kim, K. Cheung, R. C. Y. Auyeung, D. E. Wilson, K. M. Charipar, A. Piqué, and N. A. Charipar, "VO2-based switchable radiator for spacecraft thermal control", Scientific Reports, 2019, Volume 9, Article number 11329, DOI: 10.1038/s41598-019-47572-z

  7. A. M. Morsy, M. T. Barako, V. Jankovic, V. D. Wheeler, M. W. Knight, G. T. Papadakis, L. A. Sweatlock, P. W. C. Hon, and M. L. Povinelli, "Experimental demonstration of dynamic thermal regulation using vanadium dioxide thin films", Scientific Reports, 2020, Volume 10, Article number 13964, DOI: 10.1038/s41598-020-70931-0

  8. K. Sun, C. A. Riedel, A. Urbani, M. Simeoni, S. Mengali, M. Zalkovskij, B. Bilenberg, C. H. de Groot, and O. L. Muskens, "VO2 Thermochromic Metamaterial-Based Smart Optical Solar Reflector", ACS Photonics, 2018, Volume 5 (6), Pages 2280-2286, DOI: 10.1021/acsphotonics.8b00119

  9. Z. Liu, D. Zhu, S. P. Rodrigues, K.-T. Lee, and W. Cai, "Generative Model for the Inverse Design of Metasurfaces", Nano Letters, 2018, Volume 18 (10), Pages 6570-6576, DOI: 10.1021/acs.nanolett.8b03171

  10. E. S. Harper, E. J. Coyle, J. P. Vernon, and M. S. Mills, "Inverse design of broadband highly reflective metasurfaces using neural networks", Physical Review B, 2020, Volume 101 (19), Article number 195104, DOI: 10.1103/PhysRevB.101.195104

  11. C. C. Nadell, B. Huang, J. M. Malof, and W. J. Padilla, "Deep learning for accelerated all-dielectric metasurface design", Optics Express, 2019, Volume 27 (20), Pages 27523-27535, DOI: 10.1364/OE.27.027523

  12. A. Mall, A. Patil, D. Tamboli, A. Sethi, and A. Kumar, "Fast design of plasmonic metasurfaces enabled by deep learning", Journal of Physics D: Applied Physics, 2020, Volume 53 (49), Pages 49LT01 DOI: 10.1088/1361-6463/abb33c

  13. X. Han, Z. Fan, Z. Liu, C. Li, and L. J. Guo, "Inverse design of metasurface optical filters using deep neural network with high degrees of freedom", InfoMat, 2020, Pages 1-11, DOI: 10.1002/inf2.12116

  14. Weiliang Jin, Wei Li, Meir Orenstein, and Shanhui Fan, "Inverse Design of Lightweight Broadband Reflector for Relativistic Lightsail Propulsion", ACS Photonics, 2020, Volume 7 (9), Pages 2350-2355, DOI: 10.1021/acsphotonics.0c00768

  15. B. Mildenhall, P. P. Srinivasan, M. Tancik, J. Barron, R. Ramamoorthi, and R. Ng, "Nerf: Representing scenes as neural radiance fields for view synthesis", European Conference on Computer Vision, 2020, Pages 405-421, DOI: 10.1007/978-3-030-58452-8_24

  16. V. Sitzmann, J. N. P. Martel, A. W. Bergman, D. B. Lindell, and G. Wetzstein, "Implicit neural representations with periodic activation functions", arXiv preprint, 2020, arXiv:2006.09661

  17. A. R. Davoyan, J. N. Munday, N. Tabiryan, G. A. Swartzlander, and L. Johnson, "Photonic materials for interstellar solar sailing", Optica, 2021, Volume 8 (5), Pages 722-734, DOI: 10.1364/OPTICA.417007

Hamburger icon
Menu
Advanced Concepts Team