April 5, 2023
Waseda University
Keio University
[Key Points]
World's first success in predicting molecular adsorption on solid surfaces using "Fujitsu Quantum-inspired Computing Digital Annealer" (hereinafter "Digital Annealer"), a quantum-inspired technology that is a form of next-generation computing.
Developed a new method to rapidly search for molecular adsorption configurations without causing a combinatorial explosion.
Enables accurate and rapid prediction of the optimal arrangement, especially for composite materials and other systems with a large number of possible molecular arrangements.
A research group from Waseda University, including Hiroshi Sampei and Koki Saegusa, first-year doctoral students at the Graduate School of Advanced Science and Engineering, and Professor Yasushi Sekine of the Faculty of Science and Engineering, in collaboration with Associate Professor Shu Tanaka of the Faculty of Science and Technology, Keio University, and ENEOS Corporation, has successfully predicted molecular adsorption on solid surfaces using "Digital Annealer," a quantum-inspired technology from Fujitsu Limited that is based on the annealing method, a form of next-generation computing. This is the world's first initiative to apply quantum-inspired technology to the prediction of molecular adsorption on solid surfaces.
This time, using "Digital Annealer," the team developed a new method to rapidly search for adsorption configurations without causing a combinatorial explosion. This method was found to be able to discover stable molecular arrangements much faster than conventional methods, especially in regions where many molecules are adsorbed. For example, a prediction for the adsorption of 16 molecules, which conventionally took 38,601 seconds, was completed with only 2,154 seconds of preparation time (required only once) and 132 seconds of computation by "Digital Annealer," demonstrating its overwhelming speed.
Furthermore, analysis of the results using Matlantis, a universal atomistic simulator provided by Preferred Computational Chemistry, Inc., a joint venture established by Preferred Networks, Inc. and ENEOS Corporation, confirmed that the predictions made by this method are correct. This method enables accurate and rapid prediction of optimal arrangements, particularly in composite materials and large-scale modeling where there are many combinations of molecular arrangements.
The results of this research were published online in the American Chemical Society's "JACS Au" on March 27, 2023 (local time).
For the full press release, please see below.