Material design is shifting toward data-driven methods, where discovery increasingly relies on computational insights. Developing soft materials remains especially challenging, requiring careful selection of molecular building blocks and precise optimization of their arrangement in molecular space.

Researchers at Hokkaido University have developed an ultra-sticky hydrogel for wet environments by combining artificial intelligence (AI) with data mining. Since the softness of hydrogels typically limits adhesion, designing such materials is challenging and often relies on costly trial-and-error experiments. The new AI-driven approach overcomes these hurdles, as detailed in their Nature publication.

Figure 1. Ultra-Adhesive Underwater Hydrogels Created with Data Mining and AI

Designing Adhesive Hydrogels: Key Challenges

Hydrogels are water-filled polymer networks with tunable properties controlled by their chemical composition and structure. Computationally designing them for specific functions is challenging due to the vast diversity of polymer building blocks and the complexity of molecular arrangements and interactions, such as hydrogen bonding and van der Waals forces. Designing adhesive hydrogels for wet environments is even harder, as they swell in water, adding another layer of complexity to material optimisation. Figure 1 shows Ultra-Adhesive Underwater Hydrogels Created with Data Mining and AI.

Data-Driven Approaches Enable Breakthroughs

To create a hydrogel with strong, durable underwater adhesion, researchers mined the NCBI Protein database, which catalogs amino acid sequences responsible for adhesion in organisms such as bacteria, viruses, archaea, and eukaryotes. They synthetically adapted these sequences for hydrogel polymers, allowing systematic exploration of adhesive design. As co-lead author Hailong Fan explained, the goal was not just to mimic natural proteins but to use AI and data mining to uncover new design rules and push the limits of underwater adhesion.

Using the database insights, the team first designed and synthesized 180 bioinspired hydrogels, all showing superior adhesion compared to conventional versions. They then applied machine learning to optimize the designs, producing hydrogels with record-breaking underwater adhesion—instant and repeatable strengths above 1 MPa, about ten times greater than previous materials. These AI-designed hydrogels also worked across diverse surfaces in both fresh and saline water. As Fan noted, the breakthrough lies not only in the adhesive performance but in establishing a data-driven, AI-guided approach to material design beyond traditional biomimicry.

A Universal Adhesive

The team tested the top three hydrogels across diverse wet environments, showing long-lasting adhesion. One secured a rubber duck to a seaside rock, withstanding repeated wave impacts over multiple tide cycles. Another sealed a 20 mm hole in a pressurized water pipe, stopping the leak instantly and holding firm for five months. A third was implanted under the skin of mice, confirming its biocompatibility.

The hydrogels’ exceptional underwater adhesion opens opportunities in fields ranging from biomedical engineering—such as prosthetic coatings and wearable biosensors—to deep-sea exploration and marine farming. The researchers also emphasize that their data-driven strategy could be extended to designing other advanced soft materials [1]. Looking ahead, Fan notes the team plans to investigate the molecular mechanisms behind the adhesives and apply the approach to self-healing and biomedical hydrogels.

References:

1. https://physicsworld.com/a/super-sticky-underwater-hydrogels-designed-using-data-mining-and-ai/

Cite this article:

Janani R (2025), AI-Designed Super-Adhesive Underwater Hydrogels, AnaTechMaz, pp. 274