2026-02-12 ヒューストン大学（UH）
On the left is the physical tree-like geometry of the thin films and on the right is the temperature map.
＜関連情報＞
- https://www.uh.edu/news-events/stories/2026/february/02122026-ghasemi-ai-data-centers-heat.php
- https://www.sciencedirect.com/science/article/abs/pii/S0017931025013626
- https://www.sciencedirect.com/science/article/abs/pii/S0017931025012372
極薄膜蒸着のためのPINN誘導多孔質階層構造 PINN-guided porous hierarchical structure for extreme thin film evaporation
Saber Badkoobeh Hezaveh, Amirmohammad Jahanbakhsh, Rojan Firuznia, Mohammadreza Borzooei, Long Chang, Hadi Ghasemi
International Journal of Heat and Mass Transfer Available online: 27 October 2025
DOI:https://doi.org/10.1016/j.ijheatmasstransfer.2025.128027
Highlights
- PINN-based design for porous hierarchical structures with enhanced critical heat flux
- Fabricated and experimentally tested, showing a CHF of 2785 W/cm at 93 K superheat
- Evaporation kinetics governed by interfacial liquid temperature, not the structure’s geometry
- Higher non-dimensional vapor pressure directly leads to higher achievable heat flux
Abstract
Thin-film evaporation has emerged as a highly efficient thermal management solution capable of sustaining extreme heat fluxes while minimizing thermal resistance. In this study, a physics-informed neural network model (PINN) was used not as a direct validation tool but as a design guide to find porous hierarchical structures with high critical heat flux (CHF). PINN insights were used to find optimized hierarchical structures to enhance capillary-driven liquid replenishment and interfacial heat flux. Through PINN guided design, we developed porous hierarchical structures with CHF of 1638 W/cm2 at 78 K superheat. On reduced vapor pressure environment of 3 kPa, these structures show CHF of 2785 W/cm2 at 93 K superheat. This superior performance is attributed to high contact line density of 15.25 µm-1 and 127 µm-1 in the structures fabricated with 500 nm and 60 nm Copper nanoparticles, respectively, obtained from geometric considerations of the porous structures, and then used as input parameters in the PINN analysis. Furthermore, we have determined the liquid film thickness and interfacial temperature in these extreme heat flux conditions. Most importantly, we found that interfacial evaporation is governed primarily by the temperature of the interfacial liquid, which is independent of the geometry of the hierarchical structure. Also, heat flux is directly correlated to the non-dimensional vapor pressure and higher heat fluxes are expected at higher normalized vapor pressures. Along with new fundamental insights, this work underscores the potential of PINN-guided design approaches in shaping of next-generation thermal management solutions.
階層構造における薄膜蒸着のための物理学に基づくニューラルネットワークベースのトポロジー最適化 Physics-informed neural network based topology optimization for thin-film evaporation in hierarchical structures
Amirmohammad Jahanbakhsh, Rojan Firuznia, Saber Badkoobeh Hezaveh, Mohammadreza Borzooei, Hadi Ghasemi
International Journal of Heat and Mass Transfer Available online: 3 October 2025
DOI:https://doi.org/10.1016/j.ijheatmasstransfer.2025.127902
Highlights
- Coupled topology optimization and PINN for thin-film evaporation design.
- Optimized structures reach CHF at lower superheat than literature benchmarks.
- Optimized structures are branched with a solid density near 0.5.
- Scalable design approach for thermal structures across various solids and fluids.
Abstract
Thin film evaporation through hierarchical structures is a promising approach for thermal management in electronics and photonics. However, identifying the optimal hierarchical structure for efficient thermal management remains an ongoing challenge. This study presents a coupled framework that integrates classical SIMP-based thermal topology optimization with a pretrained physics-informed neural network (PINN) for data-driven verification to final optimal hierarchical structures. The objective is to minimize thermal compliance in evaporative structures while ensuring physical fidelity. The findings suggest that topologically optimal structures are mostly in the form of branched structures with solid density of ≈ 0.5. These structures could achieve high critical heat flux (CHF) at much lower superheats compared to traditionally studied structures. In addition, even for optimal structures, higher density of solid–liquid contact line directly correlates to higher CHF values. This hybrid approach not only enhances computational efficiency but also bridges the gap between simulation and real-world physical behavior, paving the way for validated thermal design in advanced cooling systems.
