EurekAlert! - Metasurfaces, ultra-compact optical devices capable of "precisely manipulating light," have shown great potential in augmented reality (AR) glasses, holographic projection, biosensing, and other fields. However, traditional manufacturing technologies face a dilemma: either they are costly and inefficient, or they lack sufficient performance, making large-scale mass production difficult. Recently, a team led by Yujin Park, Donghoe Kim, and Junsuk Rho from Pohang University of Science and Technology (POSTECH), South Korea, published a review paper in Optics and Photonics Research (Opt. Photonics Res.), systematically proposing two innovative strategies based on nanoimprint lithography (NIL). These strategies successfully address the "low refractive index" limitation of traditional processes, bringing the transition of metasurfaces from laboratory research to industrial production within reach.

image: Schematic overview of metasurface fabrication approaches, comparing E-beam lithography with mass-production strategies based on nanoimprint lithography, including particle-embedded resins and hybrid materials.

Credit: Jesus Yujin Park and Donghoe Kim and Junsuk Rho/Pohang University of Science and Technology (POSTECH)

Metasurfaces are composed of nanoscale "micro-optical units" (meta-atoms). Similar to traditional lenses and filters, they can control the direction, color, and phase of light, but they are hundreds of times thinner and lighter than conventional optical devices. To achieve such performance, two key conditions must be met during manufacturing: first, high-precision nanoscale patterning, and second, the use of high-refractive-index materials (which can "bend" light more strongly to enhance manipulation efficiency).

For a long time, electron beam lithography (EBL) has been the "gold standard" for metasurface manufacturing. It can create patterns with a precision of up to 80 nm, and when combined with deposition technologies like plasma-enhanced chemical vapor deposition (PECVD), it can produce metasurfaces using high-refractive-index materials, achieving an optical efficiency of up to 89%. However, EBL has obvious shortcomings: it draws patterns point by point like "writing by hand with a pen," and can only process small areas at a time. Not only is it extremely costly (the cost of a single batch production is 5-10 times that of NIL), but its throughput is also surprisingly low—it takes several days to process a single 12-inch wafer, making it completely unable to meet the needs of industrial mass production.

To solve the scalability issue, researchers once tried nanoimprint lithography (NIL). This technology replicates nanoscale patterns in batches using a prefabricated mold, similar to "stamping." Its throughput is more than 100 times that of EBL, while the cost is only 1/20 of EBL. However, a new problem emerged: the refractive index of the resin used in traditional NIL is only about 1.5 (close to ordinary glass). Metasurfaces made directly from this resin have an optical efficiency of less than 10%, so they can only be used as "temporary templates" and require additional processing to improve performance, which instead increases manufacturing complexity.

To address the refractive index problem of NIL, the team focused on proposing two solutions in the review. These solutions not only retain the advantages of NIL—high speed and low cost—but also enable metasurfaces to achieve performance comparable to that of EBL processes.

The first strategy is the "hybrid material method": first, nanoscale patterns are imprinted on low-refractive-index resin using NIL, and then a high-refractive-index thin film is coated on the surface of the patterns using "atomic layer deposition (ALD)" technology. For example, titanium dioxide (with a refractive index of 2.3-2.5) is used for visible light scenarios, and zirconium oxide (with a refractive index of over 2.2) is used for ultraviolet (UV) light scenarios. This thin film acts like a "high-refractive-index coat," significantly improving the optical efficiency of the resin patterns. The team's experimental data shows that visible-light metalenses manufactured using this method achieve a maximum focusing efficiency of 89.6% at wavelengths of 450-635 nm, which is almost the same as that of the EBL process. More importantly, this method can already be used for batch production on 12-inch wafers, reducing the manufacturing time per wafer to less than 2 hours.

The second strategy is the "particle-embedded resin (PER) method": high-refractive-index nanoparticles (such as titanium dioxide and silicon particles) are directly mixed into NIL resin, similar to "adding crystal powder to glue," turning the resin itself into a high-refractive-index material. This method does not require subsequent coating, allowing metasurfaces to be manufactured in one step, further simplifying the process. The team also optimized the PER method to address its shortcomings: to solve the problem of "structural damage during demolding," they developed a water-soluble polyvinyl alcohol (PVA) mold—during demolding, the mold only needs to be dissolved in water to obtain undamaged nanostructures; to solve the problem of "residual layers affecting light transmittance," they used a specially designed tape to precisely peel off the residual layers, reducing light scattering of the metasurfaces by more than 30%. Currently, the PER method can manufacture high-precision structures with an aspect ratio of 6, and infrared metalenses manufactured using this method achieve a focusing efficiency of 47% at a wavelength of 940 nm, which can be used for human blood vessel imaging.

The breakthroughs of the two strategies have also enabled metasurface applications to break free from the limitations of "flat surfaces and single wavelengths."

In terms of wavelength coverage, the zirconium oxide coating used in the hybrid material method is suitable for UV scenarios (such as optical components for deep UV lithography), while the silicon particles in the PER method are suitable for infrared scenarios (such as thermal imaging and LiDAR sensors). Titanium dioxide, whether used as a coating or particles, can work efficiently in the visible light range and can be applied to optical lenses for AR glasses and high-definition holographic displays.

In terms of substrate adaptability, the team also extended the PER method to biodegradable materials and curved surfaces: they mixed high-refractive-index particles into hydroxypropyl cellulose (HPC) resin to create metasurface labels that can be directly attached to the surface of fruits like apples. These labels not only have anti-counterfeiting functions but also can be dissolved in water, avoiding packaging pollution. At the same time, the PER method can also imprint patterns on curved glass surfaces, providing a low-cost manufacturing solution for the "curved optical windows" of automotive LiDAR.

Despite achieving significant progress, the team also objectively pointed out in the paper that there are still two major challenges to be solved in the large-scale manufacturing of metasurfaces: first, the peeling of residual layers in the PER method currently relies on manual adjustment of tape adhesion, and automated equipment needs to be developed in the future to ensure consistency in mass production; second, the ALD coating technology used in the hybrid material method currently requires multiple "precursor injection-cleaning" cycles to coat each thin film, resulting in slow speed, and the process needs to be optimized to improve throughput.

"Metasurfaces are expected to transform optical devices from being 'bulky' to 'thin and lightweight'—for example, reducing the thickness of camera lenses from several centimeters to the micrometer level," said Professor Junsuk Rho, the leader of the team. "Next, we will focus on promoting the combination of these two strategies with roll-to-roll (R2R) manufacturing technology. Our goal is to achieve 'hundred-meter-level' continuous production of metasurfaces, further reducing costs and enabling more consumer electronics and medical devices to use this new type of optical device."

Park Y, Kim D, Rho J. Nanoimprint lithography for scalable manufacturing of optical metasurfaces. Opt. Photonics Res. 2025(1):0001, https://doi.org/10.55092/opr20250001

Source from [https://www.eurekalert.org/news-releases/1107536].

Researchers have developed a high-fidelity 13-degree-of-freedom nonlinear model and an intelligent algorithm for wind turbine dynamic analysis. This framework accurately captures complex tower-blade interactions, including often-neglected torsional effects, achieving a remarkable agreement with high-fidelity benchmarks. Published in Smart Construction, this work provides a powerful and efficient tool for structural assessment and future optimization of large-scale wind energy systems.

image: The proposed intelligent analysis method bridges high-fidelity modeling and computational efficiency. It uses an iterative algorithm to identify optimal mode shapes, achieving a key response error of less than 3.5% against the high-fidelity benchmark OpenFAST, enabling faster and reliable wind turbine dynamic simulation.

Credit: Xuhong Zhou/Chongqing University, Jiepeng Liu/Chongqing University, Guoqing Huang/Chongqing University, Liang Cao/Hunan University, Maolin Dai/Chongqing University

The global push for sustainable energy has cemented wind power's role in the renewable transition. However, designing safe and cost-effective onshore wind turbines requires a deep understanding of their dynamic behavior under complex environmental loads. Traditional modeling approaches often struggle to balance computational efficiency with simulation accuracy, particularly in capturing the full coupled dynamics of the entire system.

Addressing this challenge, a research team led by Professor Xuhong Zhou from Chongqing University has developed an innovative nonlinear dynamic modeling and intelligent analysis framework for onshore wind turbines. Their study introduces a comprehensive 13-degree-of-freedom (13-DOF) multibody model derived using Euler-Lagrange formalism.

"This model provides a holistic view of wind turbine dynamics," explains Professor Guoqing Huang. "A key advancement is the explicit incorporation of the tower's torsional degree of freedom, an aspect often simplified in conventional models but critical for accurate load assessment in the upper tower sections."

The tower and blades are modeled as Euler-Bernoulli beams capable of capturing both bending and torsional deformations, with aerodynamic loads computed via an enhanced Blade Element Momentum theory. To tackle the critical challenge of selecting optimal vibration mode functions—which significantly impact computational cost and result accuracy—the team proposed an intelligent mode selection algorithm. This algorithm automatically identifies the most suitable mode shapes based on structural response convergence.

"A major hurdle in efficient simulation is choosing the right modal representations without sacrificing physical accuracy," says Professor Jiepeng Liu. "Our intelligent algorithm systematically optimizes this selection, striking a balance that avoids the prohibitive computational cost of high-fidelity commercial tools while maintaining high accuracy."

The numerical simulations, implemented symbolically in MATLAB?, were rigorously validated against OpenFAST, a widely recognized high-fidelity simulation tool from the National Renewable Energy Laboratory (NREL), using the NREL 5-MW reference turbine as a benchmark. The results demonstrated that the proposed model effectively captures nonlinear and coupled dynamic behavior.

"The validation showed a close agreement with OpenFAST outputs, with relative errors in key response metrics, such as tower-top and blade-tip displacements, maintained within 3.5%," notes Doctor Maolin Dai from Chongqing University. "This level of accuracy, achieved at a fraction of the computational expense, is highly promising for engineering applications."

This modeling framework offers a reliable tool for the structural dynamic assessment of existing turbines and establishes a solid foundation for future applications in optimization and control of large-scale wind energy systems. By enabling more accurate and efficient simulations, it can contribute to the design of lighter, safer, and more economically competitive wind turbine towers, which account for a significant portion of project costs.

"The framework is particularly suitable for preliminary design, parameter sensitivity studies, and dynamic response analysis," concludes Associate Professor Liang Cao from Hunan University. "It charts a clear path for developing next-generation, performance-driven design tools for the wind energy industry."

The team acknowledges future directions, including further theoretical refinement to capture more complex dynamic couplings and expansion of the model's validation under non-steady-state conditions like turbulent inflow.

This paper "Research on intelligent analysis method for dynamic response of onshore wind turbines" was published in Smart Construction (ISSN: 2960-2033), a peer-reviewed open access journal dedicated to original research articles, communications, reviews, perspectives, reports, and commentaries across all areas of intelligent construction, operation, and maintenance, covering both fundamental research and engineering applications. The journal is now indexed in Scopus, and article submission is completely free of charge until 2026.

Citation:

Dai M, Cao L, Huang G, Zhou X, Liu J. Research on intelligent analysis method for dynamic response of onshore wind turbines. Smart Constr. 2025; 20250028. https://doi.org/10.55092/sc20250028

Source from [https://www.eurekalert.org/news-releases/1106426].

Nashville, TN & Williamsburg, VA – 24 Nov 2025 – A new study published in Artif. Intell. Auton. Syst. delivers the first systematic cross-model analysis of prompt engineering for structured data generation, offering actionable guidance for developers, data scientists, and organizations leveraging large language models (LLMs) in healthcare, e-commerce, and beyond. Led by Ashraf Elnashar from Vanderbilt University, alongside co-authors Jules White (Vanderbilt University) and Douglas C. Schmidt (William & Mary), the research benchmarks six prompt styles across three leading LLMs to solve a critical challenge: balancing accuracy, speed, and cost in structured data workflows.

image: Evaluates six prompt styles (JSON, CSV, Prefix, YAML, Function, Hybrid) across three LLMs (ChatGPT-4o, Claude, Gemini) on datasets (Stories, Medical, Receipts), measuring Accuracy, Token Cost, and Time to reveal model trade-offs and optimal prompt choices.

Credit: Ashraf Elnashar, Jules White/Vanderbilt University, Douglas C. Schmidt/William & Mary

Structured data—from medical records and receipts to business analytics—powers essential AI-driven tasks, but its quality and efficiency depend heavily on how prompts are designed. “Prior research only scratched the surface, testing a limited set of prompts on single models,” said Elnashar, the study’s corresponding author and a researcher in Vanderbilt’s Department of Computer Science. “Our work expands the horizon by evaluating six widely used prompt formats across ChatGPT-4o, Claude, and Gemini, revealing clear trade-offs that let practitioners tailor their approach to real-world needs.”

Key Findings: Accuracy vs. Efficiency—A Clear Choice for Every Use Case

The team’s rigorous experiment, conducted across three datasets (personal stories, medical records, and receipts), measured accuracy, token cost (a key driver of API expenses), and generation time for each prompt style-LLM combination. The results uncovered distinct strengths in each model:

Claude emerged as the accuracy leader (85% overall), excelling with hierarchical prompt formats like JSON and YAML—ideal for complex, high-stakes tasks such as medical record generation where data integrity is non-negotiable.

ChatGPT-4o stood out for efficiency, delivering the lowest token usage (under 100 tokens for lightweight formats) and fastest processing times (4–6 seconds on average), making it perfect for cost-sensitive or real-time applications like e-commerce receipt processing.

Gemini offered a balanced middle ground, with solid performance across all metrics—though it showed variability with mixed-format prompts like Hybrid CSV/Prefix.

“Hierarchical formats like JSON and YAML boost accuracy but come with higher token costs, while lightweight options like CSV and simple prefixes cut latency without sacrificing much precision,” Elnashar explained. “For example, a healthcare provider handling patient data might prioritize Claude + JSON for accuracy, while an e-commerce platform could opt for ChatGPT-4o + CSV to process thousands of receipts efficiently.”

The study also highlighted a universal challenge: all LLMs struggled with narrative-style unstructured data (e.g., personal stories), with accuracy dropping to ~40% across prompt styles—underscoring the need for tailored approaches for different data types.

Practical Tools for Developers: Reusable Resources to Accelerate AI Workflows

Beyond insights, the research provides tangible value for the AI community. The team has made datasets, prompt templates, validation scripts, and design guidelines publicly available on GitHub (https://github.com/elnashara/EfficientStructuringMethods/tree/main), enabling reproducibility and immediate adoption.

“We wanted to move beyond theory—these resources let developers skip the trial-and-error and directly apply our findings to their pipelines,” said Jules White, co-author and professor at Vanderbilt’s Department of Computer Science. “Whether you’re building a medical data system or an e-commerce analytics tool, our work gives you a roadmap to choose the right prompt style and LLM.”

Looking Ahead: Expanding the Boundaries of Prompt Engineering

The study builds on the authors’ prior work focused on GPT-4o, now generalized to multiple models and prompt formats. Future research will explore LLMs’ robustness to noisy instructions, missing fields, and unseen schemas—critical considerations for real-world deployments. “As AI becomes more integrated into critical systems, we need to understand how these models perform when faced with the messiness of real data,” noted Schmidt, a professor in William & Mary’s Department of Computer Science.

This research was conducted without specific grant funding. The authors acknowledge the support of LLMs ChatGPT-4o, Claude, and Gemini for code generation, data visualization, and comparative evaluation.

About the Authors

Ashraf Elnashar: Department of Computer Science, Vanderbilt University (ashraf.elnashar@vanderbilt.edu)

Jules White: Department of Computer Science, Vanderbilt University

Douglas C. Schmidt: Department of Computer Science, William & Mary

About the Publication

Title: Prompt engineering for structured data: a comparative evaluation of styles and LLM performanceJournal: Artif. Intell. Auton. Syst.DOI: 10.55092/aias2025009

License: Creative Commons Attribution 4.0 International License

Source from [https://www.eurekalert.org/news-releases/1107970].

Researchers from the University of West Attica have introduced a comprehensive framework that redefines how products are designed for additive manufacturing (3D printing). Published in Advanced Manufacturing, the study emphasizes a system-level, process-aware approach that integrates design intent, material behavior, and sustainability—moving beyond geometry-focused rules to unlock the full potential of additive manufacturing.

image: The graphical abstract illustrates the transition from traditional CAD to process-aware design for additive manufacturing, highlighting key DfAM factors such as build orientation, anisotropy, tolerancing, and sustainability in a unified workflow.

Credit: Antreas Kantaros / University of West Attica

Additive Manufacturing (AM), commonly known as 3D printing, has matured from a rapid prototyping tool into a fully capable production technology used in aerospace, healthcare, and automotive industries. Yet, many design practices still follow principles developed for traditional, subtractive manufacturing methods.

Addressing this gap, the research team at the University of West Attica present a new perspective titled “Toward a Holistic Approach for Design for Additive Manufacturing: A Perspective on Challenges, Practical Insights, and Research Needs,” published in Advanced Manufacturing. The paper proposes a system-level and process-aware design mindset—one that aligns digital design tools with the physical realities of AM processes.

“Design for Additive Manufacturing should not merely ensure printability,” explains Dr. Kantaros. “It must connect material-process interactions, build orientation, tolerancing, and sustainability considerations to create designs that are innovative, reliable, and efficient.”

The study critiques existing DfAM methods that focus mainly on geometry optimization, arguing that such heuristics overlook critical factors such as anisotropy, thermal distortions, and lifecycle sustainability. Instead, the authors advocate for an integrated workflow that combines simulation, optimization, and AI-assisted manufacturability feedback within digital design environments.

Strategic applications such as part consolidation, mass customization, and functionally graded materials demonstrate how holistic DfAM thinking can improve product performance while reducing environmental impact and production costs.

“True innovation in additive manufacturing begins when design and manufacturing are no longer treated as separate stages,” adds co-author Professor Theodore Ganetsos. “By merging these perspectives, we can achieve sustainable, high-performance engineering solutions.”

This publication contributes to the growing movement to make DfAM education and practice more interdisciplinary—linking engineering, materials science, and industrial design—and positions additive manufacturing as a catalyst for sustainable, intelligent production systems.

Paper information: Kantaros, A.; Ganetsos, T.; Pallis, E.; Papoutsidakis, M. “Toward a Holistic Approach for Design for Additive Manufacturing: A Perspective on Challenges, Practical Insights, and Research Needs.” Advanced Manufacturing, 2025. DOI: https://doi.org/10.55092/am20250011

Source from [https://www.eurekalert.org/news-releases/1107818].

2025年青年科学基金项目A类（原国家杰青）名单，更新至333人，如下图所示（人工不完全统计整理，仅供参考）：

青年科学基金项目（A类）简介

青年科学基金项目（A类）（原国家杰出青年科学基金项目）是国家自然科学基金委员会设立的人才项目，1994年由国务院批准设立，旨在支持在基础研究方面已取得突出成绩的青年学者自主选择研究方向开展创新研究，促进青年科学技术人才成长，吸引海外人才，培养优秀学术带头人。该项目已成为我国高端人才资助计划的引领者，是国内仅次于两院院士的第二层次高端人才标尺，被视为科学家科研事业成功的重要源头和标志。

2020年12月起，包括国家杰出青年科学基金在内的特定项目已不再通过官方渠道向社会公开完整的资助名单。

2025年起，国家杰出青年科学基金项目更名为青年科学基金项目（A类），以深化人才项目改革，避免"帽子化"现象。

来源：“科学家之家”公众号，仅用于学术分享，如有侵权请联系删除。