Publication:
Learned Manifold Reparameterization for DRC-Compatible Inverse Design in Silicon Photonics
| cris.virtual.department | #PLACEHOLDER_PARENT_METADATA_VALUE# | |
| cris.virtual.orcid | #PLACEHOLDER_PARENT_METADATA_VALUE# | |
| cris.virtualsource.department | d1dac5b2-325d-4956-a26b-9583d552bac4 | |
| cris.virtualsource.orcid | d1dac5b2-325d-4956-a26b-9583d552bac4 | |
| dc.contributor.author | Danis, Bahrem Serhat | |
| dc.contributor.author | Vit, Aycan Deniz | |
| dc.contributor.author | Dasdemir, Ahmet Onur | |
| dc.contributor.author | Desdemir, Demet Baldan | |
| dc.contributor.author | Akcakoca, Enes | |
| dc.contributor.author | Yanmaz, Zeynep Ipek | |
| dc.contributor.author | Polat, Gulzade | |
| dc.contributor.author | Aydogan, Aytug | |
| dc.contributor.author | Magden, Abdullah | |
| dc.contributor.author | Magden, Emir Salih | |
| dc.date.accessioned | 2026-11-03T12:29:03Z | |
| dc.date.available | 2026-11-03T12:29:03Z | |
| dc.date.createdwos | 2026 | |
| dc.date.issued | 2026 | |
| dc.description.abstract | We recast fabrication-compatible photonic inverse design as an unconstrained optimization problem on a learned low-dimensional manifold, intrinsically eliminating design-rule violations. A deep generative network is used to simultaneously enforce minimum feature size, binary material composition in an SOI platform, and CMOS design-rule checks. Optimizations performed in this resulting latent space are mapped onto the full space of manufacturable device geometries. This procedure converts length-scale-constrained topology optimization into a standard stochastic gradient problem, removing the need for post-processing or device correction steps. We demonstrate separate generator models for electron-beam lithography (60 nm minimum feature) and photolithography (150 nm minimum feature), and show that the same framework scales across fabrication processes by adjusting kernel configurations and upsampling ratios. We validate this approach by designing three compact device classes using FDFD-based optimizations: broadband power splitters with 50/50, 30/70, and 10/90 splitting ratios (losses between 0.05–0.12 dB; 4.2 × 4.2 μm2), a wavelength demultiplexer (0.25 dB loss, 16 dB extinction; 7×7 μm2), and a TE0-to-TE1 mode converter (0.03 dB loss; 2.1×2.1 μm2). This approach establishes a scalable, process-agnostic backbone for fabrication-aware integrated photonic design. | |
| dc.description.wosFundingText | This work is supported by the Scientific and Technological Research Council of Turkey (TUBITAK) under grant number 122E214. | |
| dc.identifier.doi | 10.1117/12.3099627 | |
| dc.identifier.isbn | 979-8-9023-2167-5 | |
| dc.identifier.issn | 0277-786X | |
| dc.identifier.uri | https://imec-publications.be/handle/20.500.12860/60714 | |
| dc.language.iso | eng | |
| dc.provenance.editstepuser | greet.vanhoof@imec.be | |
| dc.publisher | SPIE-INT SOC OPTICAL ENGINEERING | |
| dc.source.beginpage | 1410407 | |
| dc.source.conference | MACHINE LEARNING IN PHOTONICS II | |
| dc.source.conferencedate | 2026-04-12 | |
| dc.source.conferencelocation | Strasbourg | |
| dc.source.journal | MACHINE LEARNING IN PHOTONICS II | |
| dc.source.numberofpages | 7 | |
| dc.title | Learned Manifold Reparameterization for DRC-Compatible Inverse Design in Silicon Photonics | |
| dc.type | Proceedings paper | |
| dspace.entity.type | Publication | |
| imec.internal.crawledAt | 2026-09-07 | |
| imec.internal.source | crawler | |
| imec.internal.wosCreatedAt | 2026-09-07 | |
| Files | ||
| Publication available in collections: |