Skip to main content



Archived content

NOTE: this is an archived page and the content is likely to be out of date.

Fujitsu and RIKEN Demonstrate AI's Utility in Material Design

Improving the development of solid electrolytes with high ionic conductivity for use in all-solid-state lithium-ion rechargeable batteries

Fujitsu Limited,RIKEN

Tokyo and Wako, Japan, March 16, 2018

Fujitsu Limited and RIKEN today announced that the RIKEN Center for Advanced Intelligence Project (AIP) and the RIKEN AIP-Fujitsu Collaboration Center(1) have conducted a trial using first principles calculations(2) and artificial intelligence (AI) technology in materials design. Conducted together with the AIP Center's Molecular Informatics Team, the trial predicted the composition of a solid-state electrolyte for use in all-solid-state lithium-ion rechargeable batteries that would provide high ionic conductivity, and then synthesized the material for testing and evaluation.

With this method, even if only a comparatively small amount of data could be obtained from the first principles calculation, which carries a significant computational load, by combining this data with AI methods, the researchers can efficiently find an optimal material composition and accelerate materials development. Going forward, there are high expectations for the utilization of materials informatics(3) technology, including the technology used in this trial, in the materials development field for components such as batteries, semiconductors, and magnetic materials.


Success in materials development has until now relied on the years of experience and keen insights of researchers and technicians. Also unavoidable was a lengthy list of failures.

On the other hand, first principles calculation can predict the characteristics of a specified material based on quantum mechanics, so it can predict the optimal composition of a new high-functionality material in advance of any experiments, which is useful in reducing the number of failed experiments. Because the computational load of such calculation is extremely large, however, when numerous computations for a variety of compositions were conducted at once, it could take up an enormous amount of time just to complete the computations.


The RIKEN AIP-Fujitsu Collaboration Center has been conducting research and development centered on the theme of "AI that predicts the unpredictable." As part of this theme, the group aims to resolve issues in materials development through a close collaboration between materials simulation, experiments and AI, achieving such results as shortening development times to a fraction of what they have been. Furthermore, the group is making efforts to discover new high-functionality materials with compositions or crystal structures that are difficult to think up.

Researchers are now able to limit the number of first principles calculations to a fraction of what would have previously been needed by combining it with Bayesian inference method(4), an AI technique. This technique was applied to a composite material composed of three types of lithium-containing oxoacid salts, a candidate material for a solid-state electrolyte aimed at fully solid lithium-ion rechargeable batteries(5), where Fujitsu Laboratories Ltd. has a proven track record. The technique succeeded in predicting the optimal composition to achieve a high lithium ion conductivity in a realistic time-frame, a first for this material (the right side of the figure shows the prediction results). Moreover, when the composite was actually synthesized and analyzed, the researchers confirmed that the material achieved a higher lithium ion transport rate than other compositions, with a composition close to the predicted one. This has verified the accuracy of the prediction and has brought the prospect of the development of new high functionality materials closer than ever.

Figure: Estimation of lithium ionic conductivityFigure: Estimation of lithium ionic conductivity


Lithium ionic conductivity is an important characteristic of solid electrolyte materials, as it is the factor that controls the speed at which a lithium battery charges and discharges. The results of this trial show that material informatics technology utilizing both materials simulation and AI methods is an effective tool in the efficient development of lithium-ion batteries with excellent charging and discharging characteristics, without the risk of leakage or ignition.

Future Plans

Fujitsu and RIKEN will continue to promote the advanced use of AI in materials development, working to further establish materials informatics technology that can be applied to a variety of materials. In addition, through the application of these technologies, the organizations will contribute to the efficient development of new materials.

  • [1] RIKEN AIP-Fujitsu Collaboration Center

    A collaboration center formed in April 2017 between the RIKEN Center for Advanced Intelligence Project (AIP) and Fujitsu Limited. The center undertakes joint research centered on the theme of creating AI that looks beyond the normal scope, which will support people in making good decisions based on accurate predictions of the future, even in the face of uncertain changes in the environment.

  • [2] First principles calculation

    A materials simulation method. Using electronic structure theory, grounded in quantum mechanics, it is capable of calculating various characteristics of materials and substances just from the number and types of atoms in them, without using empirical parameters. However, it has an extremely significant computational load compared with other methods.

  • [3] Materials informatics

    An effort to accelerate the search for new materials by combining and fusing technologies in the materials field, such as materials synthesis and analysis technologies and materials simulation, with such technologies as data science and AI. It is expected to significantly reduce the time and cost required for materials development.

  • [4] Bayesian inference method

    A method for probabilistically inferring the causes of a phenomenon based on observed realities, which is based on the way of thinking underlying Bayesian probability.

  • [5] All-solid-state batteries

    Batteries using a solid electrolyte in place of a liquid electrolyte. Ions move within the solid material. Because all materials in these batteries are solid, there is no danger of leaks or ignition. These batteries can be used in environments where existing batteries cannot be used, such as high temperature environments, and active development is ongoing for these batteries as next-generation batteries which can easily support higher voltages and capacities.


RIKEN is Japan's largest research institute for basic and applied research. Over 2500 papers by RIKEN researchers are published every year in leading scientific and technology journals covering a broad spectrum of disciplines including physics, chemistry, biology, engineering, and medical science. RIKEN's research environment and strong emphasis on interdisciplinary collaboration and globalization has earned a worldwide reputation for scientific excellence.
Twitter: @riken_en

About Fujitsu

Fujitsu is the leading Japanese information and communication technology (ICT) company offering a full range of technology products, solutions and services. Approximately 155,000 Fujitsu people support customers in more than 100 countries. We use our experience and the power of ICT to shape the future of society with our customers. Fujitsu Limited (TSE: 6702) reported consolidated revenues of 4.5 trillion yen (US$40 billion) for the fiscal year ended March 31, 2017. For more information, please see

Press Contacts

Public and Investor Relations Division

Company:Fujitsu Limited

All company or product names mentioned herein are trademarks or registered trademarks of their respective owners. Information provided in this press release is accurate at time of publication and is subject to change without advance notice.

Date: 16 March, 2018
City: Tokyo and Wako, Japan
Company: Fujitsu Limited ,RIKEN