Inverse design of photonic surfaces via multi fidelity ensemble framework and femtosecond laser processing

npj Computational Materials, Год журнала: 2025, Номер 11(1)

Опубликована: Фев. 15, 2025

We demonstrate a multi-fidelity (MF) machine learning ensemble framework for the inverse design of photonic surfaces, trained on dataset 11,759 samples that we fabricate using high throughput femtosecond laser processing. The MF combines an initial low fidelity model generating solutions, with refines these solutions through local optimization. combined can generate multiple disparate sets laser-processing parameters each produce same target input spectral emissivity accuracy (root mean squared errors < 2%). SHapley Additive exPlanations analysis shows transparent interpretability complex relationship between and emissivity. Finally, is experimentally validated by fabricating evaluating surface designs it generates improved efficiency energy harvesting devices. Our approach provides powerful tool advancing surfaces in applications.

Язык: Английский

---

Inverse Design of Photonic Surfaces via High throughput Femtosecond Laser Processing and Tandem Neural Networks

Advanced Science, Год журнала: 2024, Номер 11(26)

Опубликована: Апрель 29, 2024

Abstract This work demonstrates a method to design photonic surfaces by combining femtosecond laser processing with the inverse capabilities of tandem neural networks that directly link fabrication parameters their resulting textured substrate optical properties. High throughput and characterization platforms are developed generate dataset comprising 35280 unique microtextured on stainless steel corresponding measured spectral emissivities. The trained model utilizes nonlinear one‐to‐many mapping between emissivity parameters. Consequently, it generates predominantly novel designs, which reproduce full range emissivities (average root‐mean‐squared‐error < 2.5%) using only compact region parameter space 25 times smaller than what is represented in training data. Finally, experimentally validated thermophotovoltaic emitter application. By synergizing laser‐matter interactions network capabilities, approach offers insights into accelerating discovery surfaces, advancing energy harvesting technologies.

Язык: Английский

---

Artificial intelligence driven laser parameter search: Inverse design of photonic surfaces using greedy surrogate-based optimization

Engineering Applications of Artificial Intelligence, Год журнала: 2025, Номер 143, С. 109971 - 109971

Опубликована: Янв. 13, 2025

Язык: Английский

---

Enhancing High-Degree-of-Freedom Meta-Atom Design Precision and Speed with a Tandem Generative Network

ACS Photonics, Год журнала: 2025, Номер unknown

Опубликована: Янв. 31, 2025

Язык: Английский

---

Generalizable Metamaterials Design Techniques Inspire Efficient Mycelial Materials Inverse Design

ACS Biomaterials Science & Engineering, Год журнала: 2025, Номер unknown

Опубликована: Фев. 3, 2025

Fungal mycelial materials can mimic numerous nonrenewable materials; they are even capable of outperforming certain at their own applications. Fungi's versatility makes mock leather, bricks, wood, foam, meats, and many other products possible. That said, there is currently a critical need to develop efficient design techniques. In materials, the wider field biomaterials, primarily limited costly forward New could be developed faster cheaper with robust inverse techniques, which not used within field. However, computational techniques will tractable unless clear concrete parameters defined for fungi, derived from genotype bulk phenotype characteristics. Through case studies comprehensive review metamaterials we identify three needs that must addressed implement in materials. These following: 1) heuristic search/optimization algorithms, 2) mathematical modeling, 3) dimensionality reduction Metamaterials researchers already use these adapted design. Then, suggest mycelium-specific as well how measure them. Ultimately, based on research current state design, synthesize generalizable paradigm applied or related fields.

Язык: Английский

---

Pixel-level metal blackbody microcavities via hierarchical laser writing

Science Advances, Год журнала: 2025, Номер 11(9)

Опубликована: Фев. 28, 2025

Conventional blackbody cavities, known for their near-unity broadband omnidirectional emissivity (absorptivity), are however constrained by large volume (e.g., >10 4 cm 3 ), imposing crucial restrictions on integration with existing devices. Here, we introduce the concept of metal microcavities, comprising thousands microscale periodic pores created metals, demonstrating excellent across visible and infrared (IR) ranges (exceeding 0.94 average from 0.25 to 20 μm). In long-wavelength IR (8 14 μm) region, was successfully achieved 100-μm-deep microcavities ultralow structural aspect ratios, facilitated laser-textured multiscale surface morphologies that substantially enhance light-trapping capabilities. Our findings demonstrate microcavity-based patterns can produce local emissivity, tunable radiative intensity gradients, wide-angle feasibility, high-temperature resistance, thereby enabling diverse applications in thermal displays such as illusion, encryption, grayscale imaging. Notably, these applicable various presenting considerable potential use extreme environments.

Язык: Английский

---

Developing Neural Networks for Inverse Design of Chiral Metamaterials

Laser & Photonics Review, Год журнала: 2025, Номер unknown

Опубликована: Апрель 8, 2025

Abstract Chiral metamaterials, renowned for their unique optical properties such as circular dichroism, are pivotal in applications like spectroscopy, sensing, and imaging. However, inherent asymmetry complex light‐matter interactions present substantial design challenges. This study harnesses the power of deep neural networks (DNNs) inverse chiral nanohole arrays (CNAs). A bidirectional network (Bi‐DNN) is developed to address one‐to‐many mapping issue, achieving high prediction accuracy (0.98). Various input‐output configurations examined, including combining inputs leveraging different models (e.g., using left‐handed circularly polarized light spectra predict right‐handed spectra), enhancing precision while reducing experimental workload. Additionally, potential CNAs high‐performance surface‐enhanced Raman spectroscopy substrates detection demonstrated. The Bi‐DNN enabled rapid accurate solutions, showing strong agreement with validations. These findings emphasize transformative role DNNs advancing metamaterial design, unlocking efficient customizable materials next‐generation sensing imaging technologies.

Язык: Английский

---

Topology optimization empowered dual-band second-order photonic topological insulators

Physical Review Applied, Год журнала: 2025, Номер 23(4)

Опубликована: Апрель 22, 2025

Язык: Английский

---