Machine learning and droplet-based printing accelerating the achievement of flexible terahertz metamaterial absorber

• Published in: Chemical engineering journal (Lausanne, Switzerland : 1996) Vol. 519; p. 165123
• Main Authors: Niu, Jianing, Chao, Xujiang, Lian, Hongcheng, Luo, Jun, Qi, Lehua
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.09.2025
• Subjects: Droplet-based printing; Environmental adaptability; Graphene micropattern; Machine learning; THz metamaterial absorber; Droplet-based printing; THz metamaterial absorber; Environmental adaptability; Graphene micropattern; Machine learning
• ISSN: 1385-8947
• DOI: 10.1016/j.cej.2025.165123

High-performance terahertz (THz) absorbers are crucial for enhancing stealth capabilities. However, traditional porous materials face significant challenges in balancing thinness with high absorptivity. Although flexible metamaterial absorbers can address this balance, their narrow absorption bandwidth and complex, substrate-limited manufacturing processes impede widespread application. Since machine learning-assisted design and additive manufacturing technology have created infinite possibilities for high-performance absorbers, in this work, we innovatively propose and fabricate a thin, broadband graphene-based THz metamaterial absorber (GMA-THz). The metamaterial micropattern is designed using neural network-based machine learning to ensure droplet-printability and broadband absorption performance. Graphene ink dispersed with cellulose nanofiber (CNF) has been formulated and studied for its dispersion stability and droplet-printing suitability. The developed strategy involves depositing CNF-dispersed graphene ink onto a CNF substrate to create graphene micropatterns with high conductivity of 10,600 S/m and strong adhesion enduring 5000 bending cycles. The resulting GMA-THz achieves >90% average absorptivity within the 1.16–1.63 THz range, with a thickness of only about 45 μm, demonstrating outstanding comprehensive performance. This GMA-THz also exhibits excellent flexibility with bend and fold deformation, and features waterproof capabilities, highlighting strong environmental adaptability. This study provides guidance for fabricating graphene metamaterial absorber through structural design, material tailoring, and printing process. [Display omitted] •GMA-THz is designed by machine learning and built via droplet printing technology.•CNF-dispersed graphene ink is deposited on CNF to improve conductivity and adhesion.•GMA-THz achieves 90% absorptivity in 1.16–1.63 THz with only 45 μm thickness.•GMA-THz is flexible, foldable, and waterproof, showing excellent overall performance.