Schrödinger's Red Beyond 65,000 Pixel‐Per‐Inch by Multipolar Interaction in Freeform Meta‐Atom through Efficient Neural Optimizer
Abstract Freeform nanostructures have the potential to support complex resonances and their interactions, which are crucial for achieving desired spectral responses. However, the design optimization of such structures is nontrivial and computationally intensive. Furthermore, the current “black box”...
| Główni autorzy: | , , , , , , |
|---|---|
| Format: | Artykuł |
| Język: | English |
| Wydane: |
Wiley
2024-04-01
|
| Seria: | Advanced Science |
| Hasła przedmiotowe: | |
| Dostęp online: | https://doi.org/10.1002/advs.202303929 |
| _version_ | 1827296618014048256 |
|---|---|
| author | Ronghui Lin Vytautas Valuckas Thi Thu Ha Do Arash Nemati Arseniy I. Kuznetsov Jinghua Teng Son Tung Ha |
| author_facet | Ronghui Lin Vytautas Valuckas Thi Thu Ha Do Arash Nemati Arseniy I. Kuznetsov Jinghua Teng Son Tung Ha |
| author_sort | Ronghui Lin |
| collection | DOAJ |
| description | Abstract Freeform nanostructures have the potential to support complex resonances and their interactions, which are crucial for achieving desired spectral responses. However, the design optimization of such structures is nontrivial and computationally intensive. Furthermore, the current “black box” design approaches for freeform nanostructures often neglect the underlying physics. Here, a hybrid data‐efficient neural optimizer for resonant nanostructures by combining a reinforcement learning algorithm and Powell's local optimization technique is presented. As a case study, silicon nanostructures with a highly‐saturated red color are designed and experimentally demonstrated. Specifically, color coordinates of (0.677, 0.304) in the International Commission on Illumination (CIE) chromaticity diagram – close to the ideal Schrödinger's red, with polarization independence, high reflectance (>85%), and a large viewing angle (i.e., up to ± 25°) is achieved. The remarkable performance is attributed to underlying generalized multipolar interferences within each nanostructure rather than the collective array effects. Based on that, pixel size down to ≈400 nm, corresponding to a printing resolution of 65000 pixels per inch is demonstrated. Moreover, the proposed design model requires only ≈300 iterations to effectively search a thirteen‐dimensional (13D) design space – an order of magnitude more efficient than the previously reported approaches. The work significantly extends the free‐form optical design toolbox for high‐performance flat‐optical components and metadevices. |
| first_indexed | 2024-04-24T14:41:40Z |
| format | Article |
| id | doaj.art-65715dba58174dc8b92cf4a2e52fbb2d |
| institution | Directory Open Access Journal |
| issn | 2198-3844 |
| language | English |
| last_indexed | 2024-04-24T14:41:40Z |
| publishDate | 2024-04-01 |
| publisher | Wiley |
| record_format | Article |
| series | Advanced Science |
| spelling | doaj.art-65715dba58174dc8b92cf4a2e52fbb2d2024-04-02T20:51:56ZengWileyAdvanced Science2198-38442024-04-011113n/an/a10.1002/advs.202303929Schrödinger's Red Beyond 65,000 Pixel‐Per‐Inch by Multipolar Interaction in Freeform Meta‐Atom through Efficient Neural OptimizerRonghui Lin0Vytautas Valuckas1Thi Thu Ha Do2Arash Nemati3Arseniy I. Kuznetsov4Jinghua Teng5Son Tung Ha6Agency for Science, Technology and Research (A*STAR) Institute of Materials Research and Engineering (IMRE) 2 Fusionopolis Way, Innovis #08‐03 Singapore 138634 Republic of SingaporeAgency for Science, Technology and Research (A*STAR) Institute of Materials Research and Engineering (IMRE) 2 Fusionopolis Way, Innovis #08‐03 Singapore 138634 Republic of SingaporeAgency for Science, Technology and Research (A*STAR) Institute of Materials Research and Engineering (IMRE) 2 Fusionopolis Way, Innovis #08‐03 Singapore 138634 Republic of SingaporeAgency for Science, Technology and Research (A*STAR) Institute of Materials Research and Engineering (IMRE) 2 Fusionopolis Way, Innovis #08‐03 Singapore 138634 Republic of SingaporeAgency for Science, Technology and Research (A*STAR) Institute of Materials Research and Engineering (IMRE) 2 Fusionopolis Way, Innovis #08‐03 Singapore 138634 Republic of SingaporeAgency for Science, Technology and Research (A*STAR) Institute of Materials Research and Engineering (IMRE) 2 Fusionopolis Way, Innovis #08‐03 Singapore 138634 Republic of SingaporeAgency for Science, Technology and Research (A*STAR) Institute of Materials Research and Engineering (IMRE) 2 Fusionopolis Way, Innovis #08‐03 Singapore 138634 Republic of SingaporeAbstract Freeform nanostructures have the potential to support complex resonances and their interactions, which are crucial for achieving desired spectral responses. However, the design optimization of such structures is nontrivial and computationally intensive. Furthermore, the current “black box” design approaches for freeform nanostructures often neglect the underlying physics. Here, a hybrid data‐efficient neural optimizer for resonant nanostructures by combining a reinforcement learning algorithm and Powell's local optimization technique is presented. As a case study, silicon nanostructures with a highly‐saturated red color are designed and experimentally demonstrated. Specifically, color coordinates of (0.677, 0.304) in the International Commission on Illumination (CIE) chromaticity diagram – close to the ideal Schrödinger's red, with polarization independence, high reflectance (>85%), and a large viewing angle (i.e., up to ± 25°) is achieved. The remarkable performance is attributed to underlying generalized multipolar interferences within each nanostructure rather than the collective array effects. Based on that, pixel size down to ≈400 nm, corresponding to a printing resolution of 65000 pixels per inch is demonstrated. Moreover, the proposed design model requires only ≈300 iterations to effectively search a thirteen‐dimensional (13D) design space – an order of magnitude more efficient than the previously reported approaches. The work significantly extends the free‐form optical design toolbox for high‐performance flat‐optical components and metadevices.https://doi.org/10.1002/advs.202303929machine learningmetasurfacesmonte Carlo tree searchmultipole interferencestructural colors |
| spellingShingle | Ronghui Lin Vytautas Valuckas Thi Thu Ha Do Arash Nemati Arseniy I. Kuznetsov Jinghua Teng Son Tung Ha Schrödinger's Red Beyond 65,000 Pixel‐Per‐Inch by Multipolar Interaction in Freeform Meta‐Atom through Efficient Neural Optimizer Advanced Science machine learning metasurfaces monte Carlo tree search multipole interference structural colors |
| title | Schrödinger's Red Beyond 65,000 Pixel‐Per‐Inch by Multipolar Interaction in Freeform Meta‐Atom through Efficient Neural Optimizer |
| title_full | Schrödinger's Red Beyond 65,000 Pixel‐Per‐Inch by Multipolar Interaction in Freeform Meta‐Atom through Efficient Neural Optimizer |
| title_fullStr | Schrödinger's Red Beyond 65,000 Pixel‐Per‐Inch by Multipolar Interaction in Freeform Meta‐Atom through Efficient Neural Optimizer |
| title_full_unstemmed | Schrödinger's Red Beyond 65,000 Pixel‐Per‐Inch by Multipolar Interaction in Freeform Meta‐Atom through Efficient Neural Optimizer |
| title_short | Schrödinger's Red Beyond 65,000 Pixel‐Per‐Inch by Multipolar Interaction in Freeform Meta‐Atom through Efficient Neural Optimizer |
| title_sort | schrodinger s red beyond 65 000 pixel per inch by multipolar interaction in freeform meta atom through efficient neural optimizer |
| topic | machine learning metasurfaces monte Carlo tree search multipole interference structural colors |
| url | https://doi.org/10.1002/advs.202303929 |
| work_keys_str_mv | AT ronghuilin schrodingersredbeyond65000pixelperinchbymultipolarinteractioninfreeformmetaatomthroughefficientneuraloptimizer AT vytautasvaluckas schrodingersredbeyond65000pixelperinchbymultipolarinteractioninfreeformmetaatomthroughefficientneuraloptimizer AT thithuhado schrodingersredbeyond65000pixelperinchbymultipolarinteractioninfreeformmetaatomthroughefficientneuraloptimizer AT arashnemati schrodingersredbeyond65000pixelperinchbymultipolarinteractioninfreeformmetaatomthroughefficientneuraloptimizer AT arseniyikuznetsov schrodingersredbeyond65000pixelperinchbymultipolarinteractioninfreeformmetaatomthroughefficientneuraloptimizer AT jinghuateng schrodingersredbeyond65000pixelperinchbymultipolarinteractioninfreeformmetaatomthroughefficientneuraloptimizer AT sontungha schrodingersredbeyond65000pixelperinchbymultipolarinteractioninfreeformmetaatomthroughefficientneuraloptimizer |