journal article·2026-04-15Open access

MxDiffusion: A Physics-Aware Maxwell’s Law-Guided Diffusion Model Strategy for Inverse Photonic Metasurface Design

Sujoy Mondal, Taehyuk Park, Sudipta Biswas, Alan X. Wang, Wenshan Cai

Nano Letters· Volume 26· Issue 14

Abstract

We introduce MxDiffusion, a hybrid physics- and data-driven diffusion-based framework that enables the efficient and highly accurate generation of photonic structures from target optical properties. The improved accuracy is achieved through a two-stage generation strategy, in which the first diffusion model is explicitly trained with Maxwell's equation-based loss to embed physical insight directly into the inverse design process, while the second model maps the physically consistent intermediate representation to the final structural geometry with significantly higher fidelity than solely data-driven approaches. The performance of MxDiffusion is validated on two representative applications: gold pattern optimization for random spectral responses and a tunable bandpass filter design based on a phase change material. In both cases, the proposed framework consistently outperforms a conventional data-driven diffusion model benchmark, particularly for out-of-training distribution design targets and highly constrained resonance conditions. These results demonstrate the efficacy and superiority of MxDiffusion as a general physics-guided inverse design paradigm.
Source: openalex_inverted_index · Quality: baseline

Research context

OpenAlex topics: Metamaterials and Metasurfaces Applications · Photonic Crystals and Applications · Optical Coatings and Gratings

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

Affiliations

School of Electrical and Computer Engineering

Georgia Institute of Technology

Department of Electrical and Computer Engineering

Baylor University

School of Materials Science and Engineering

Funding

National Aeronautics and Space Administration · Division of Materials Research

References (64)

Show source-provided references
1. 10.1021/acsnano.0c0942410.1021/acsnano.0c09424 ↗
2. 10.1038/s41566-020-0685-y10.1038/s41566-020-0685-y ↗
3. 10.1515/nanoph-2019-047410.1515/nanoph-2019-0474 ↗
4. 10.1515/nanoph-2025-004910.1515/nanoph-2025-0049 ↗
5. 10.1002/advs.20200292310.1002/advs.202002923 ↗
6. 10.1038/s41578-020-00260-110.1038/s41578-020-00260-1 ↗
7. 10.1002/lpor.20240152010.1002/lpor.202401520 ↗
8. 10.1002/adom.20250206210.1002/adom.202502062 ↗
9. 10.1021/acsphotonics.1c0185010.1021/acsphotonics.1c01850 ↗
10. 10.1515/nanoph-2023-052710.1515/nanoph-2023-0527 ↗
11. 10.1515/nanoph-2024-053610.1515/nanoph-2024-0536 ↗
12. 10.1038/s41377-018-0060-710.1038/s41377-018-0060-7 ↗
13. 10.1021/acsphotonics.7b0137710.1021/acsphotonics.7b01377 ↗
14. 10.1038/s41566-018-0246-910.1038/s41566-018-0246-9 ↗
15. 10.1364/OE.40196010.1364/OE.401960 ↗
16. 10.1039/D5NR02043C10.1039/D5NR02043C ↗
17. 10.1038/s41467-024-54983-810.1038/s41467-024-54983-8 ↗
18. 10.1038/s41586-025-09102-y10.1038/s41586-025-09102-y ↗
19. 10.1002/adma.20211002210.1002/adma.202110022 ↗
20. 10.1016/j.pquantelec.2023.10046910.1016/j.pquantelec.2023.100469 ↗
21. 10.1021/acsnano.1c0929810.1021/acsnano.1c09298 ↗
22. 10.1038/s41598-024-56522-310.1038/s41598-024-56522-3 ↗
23. 10.1038/s41377-023-01218-y10.1038/s41377-023-01218-y ↗
24. 10.1021/acsphotonics.0c0096010.1021/acsphotonics.0c00960 ↗
25. 10.1515/nanoph-2022-074610.1515/nanoph-2022-0746 ↗
26. 10.1364/OE.52798610.1364/OE.527986 ↗
27. 10.1038/s41467-019-09103-210.1038/s41467-019-09103-2 ↗
28. 10.1021/acs.nanolett.4c0306910.1021/acs.nanolett.4c03069 ↗
29. 10.1515/nanoph-2020-019710.1515/nanoph-2020-0197 ↗
30. 10.1021/acsphotonics.0c0146810.1021/acsphotonics.0c01468 ↗
31. 10.1038/s41566-020-0604-210.1038/s41566-020-0604-2 ↗
32. 10.1021/acs.nanolett.9b0397110.1021/acs.nanolett.9b03971 ↗
33. 10.1002/adma.20190111110.1002/adma.201901111 ↗
34. 10.1515/nanoph-2020-037610.1515/nanoph-2020-0376 ↗
35. 10.1021/acs.nanolett.8b0317110.1021/acs.nanolett.8b03171 ↗
36. 10.1002/adom.20217001910.1002/adom.202170019 ↗
37. 10.1016/j.engappai.2022.10525910.1016/j.engappai.2022.105259 ↗
38. 10.1002/adom.20230133710.1002/adom.202301337 ↗
39. 10.1021/acs.nanolett.9b0185710.1021/acs.nanolett.9b01857 ↗
40. 10.1021/acsphotonics.2c0087610.1021/acsphotonics.2c00876 ↗
41. 10.1021/acsphotonics.5c0055210.1021/acsphotonics.5c00552 ↗
42. 10.1002/adpr.20230015810.1002/adpr.202300158 ↗
43. 10.1515/nanoph-2024-050410.1515/nanoph-2024-0504 ↗
44. Liu, R.; Lehman, J.; Molino, P.; Petroski Such, F.; Frank, E.; Sergeev, A.; Yosinski, J. An intriguing failing of convolutional neural networks and the coordconv solution, in Proceedings of the 32nd Conference on Neural Information Processing Systems (NIPS) , Montréal, QC, Canada, 3–8 December 2018; pp 9628–9639.
45. 10.1515/nanoph-2023-029210.1515/nanoph-2023-0292 ↗
46. 10.1021/acsphotonics.5c0138410.1021/acsphotonics.5c01384 ↗
47. 10.1002/lpor.20230099810.1002/lpor.202300998 ↗
48. 10.1109/OJAP.2023.329214910.1109/OJAP.2023.3292149 ↗
49. 10.1364/OE.53738910.1364/OE.537389 ↗
50. 10.1145/361839810.1145/3618398 ↗
51. 10.1109/JLT.2024.339192410.1109/JLT.2024.3391924 ↗
52. 10.1088/2632-2153/ad743f10.1088/2632-2153/ad743f ↗
53. Chen, Y.; Bezick, M.; Wilson, B.; Kildishev, A.; Shalaev, V.; Boltasseva, A. Physics-informed latent diffusion model for high-efficiency metasurface optimization. Proc. SPIE, Photonic Computing: From Materials and Devices to Systems and Applications II . San Diego, California, USA, 18 September 2025, PC135810S.10.1117/12.3065521 ↗
54. 10.1021/acsphotonics.5c0099310.1021/acsphotonics.5c00993 ↗
55. Ho J. · Advances in Neural Information Processing Systems · 2020 · 33 · 6840
56. Song, Y.; Sohl-Dickstein, J.; Kingma, D. P.; Kumar, A.; Ermon, S.; Poole, B. Score-based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.134562020, 10.48550/arXiv.2011.13456 (accessed 2026–03–18).10.48550/arXiv.2011.13456 ↗
57. Sohl-Dickstein, J.; Weiss, E.; Maheswaranathan, N.; Ganguli, S. Deep unsupervised learning using nonequilibrium thermodynamics. In Proceedings of the 32nd International Conference on Machine Learning , PMLR, July 6–11, 2015, Lille, France, 37, pp 2256–2265.
58. Nichol, A. Q.; Dhariwal, P. Improved denoising diffusion probabilistic models. Proceedings of the 38th International Conference on Machine Learning , Virtual Event, 18–24 July, 2021, PMLR, 139, pp 8162–8171.
59. Ronneberger, O.; Fischer, P.; Brox, T. U-net: Convolutional networks for biomedical image segmentation. In Proceedings of the 18th International Conference on Medical Image Computing and Computer-Assisted Intervention , Springer, October 5–9, 2015, Munich, Germany: pp 234–241.10.1007/978-3-319-24574-4_28 ↗
60. Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A. N.; Kaiser, Ł.; Polosukhin, I. Attention is all you need. In Proceedings of the 31st International Conference on Neural Information Processing Systems (NIPS) , MIT Press, December 4–9, 2017, Long Beach, CA, USA; pp 5998–6008.
61. Ho, J.; Salimans, T. Classifier-free diffusion guidance . arXiv preprint arXiv:2207.12598 2022 10.48550/arXiv.2207.12598 (accessed 2026–03–18).10.48550/arXiv.2207.12598 ↗
62. Perez, E.; Strub, F.; De Vries, H.; Dumoulin, V.; Courville, A. Film: Visual reasoning with a general conditioning layer. In Proceedings of the 32nd AAAI conference on artificial intelligence ; February 2–7, New Orleans, Louisiana, USA, 2018, 32, pp 3942–3951.10.1609/aaai.v32i1.11671 ↗
63. 10.1016/j.materresbull.2021.11167910.1016/j.materresbull.2021.111679 ↗
64. Miscuglio, M.; Meng, J.; Yesiliurt, O.; Zhang, Y.; Prokopeva, L. J.; Mehrabian, A.; Hu, J.; Kildishev, A. V.; Sorger, V. J. Artificial synapse with mnemonic functionality using GSST-based photonic integrated memory. In 2020 International Applied Computational Electromagnetics Society Symposium (ACES) , IEEE, July 27–31, 2020, Monterey, CA, USA; pp 1–3.10.23919/ACES49320.2020.9196183 ↗