TOKYO — A team from Hokkaido University has developed a gel with extremely strong adhesive power that can maintain its bond for long periods even when wet, following a search for materials using artificial intelligence (AI).
The adhesive is so strong that, when bonding two materials over a 1-square-meter area, it could support the weight of more than a dozen elephants without coming apart. The team envisages applications in repairing water and sewage pipes, constructing underwater structures such as offshore wind turbines and even fabricating three-dimensional organs.
Gel adhesives, which resemble jelly, are widely used in fields from construction to medicine because they can conform to uneven surfaces. However, it was difficult to develop a molecular structure that would provide flexibility, and maintaining adhesion on wet surfaces or underwater was a challenge.
This photo provided by Hokkaido University shows a rubber duck toy placed on a gel material developed by its research team. Even in rough seas, the toy sticks to a rock and does not come off.
The team focused on the proteins produced by bacteria and mollusks that exhibit excellent adhesive properties. Using AI and other tools, they analyzed 24,707 types of adhesive proteins registered with the U.S. National Institutes of Health (NIH) to design molecular structures with powerful adhesive strength. They succeeded in developing an adhesive that can withstand more than 100 metric tons of force per square meter underwater, which is apparently more than five times stronger than the best-performing gel materials to date.
According to the team, the adhesive was used to seal a 2-centimeter-diameter hole in a 3-meter-tall plastic pipe filled with water, and it has been preventing water leakage for over six months. The adhesive can also be reused. In another test, it was used to attach a rubber duck to a rock underwater, and the bond withstood both wave impact and tidal changes.
Jian Ping Gong, a professor at Hokkaido University specializing in polymer materials, commented, “We expect applications in a wide range of fields, including medicine and construction sites, including underwater sites. By using AI, we want to pursue even higher performance.”
The results were published in the Aug. 6 issue of the British scientific journal Nature at https://doi.org/10.1038/s41586-025-09269-4.
(Japanese original by Yasuyoshi Tanaka, Lifestyle, Science & Environment News Department; Video provided by Hokkaido University)
AloJapan.com