By utilizing deep learning*1 and image analysis— technologies which have been making remarkable progress in recent years—for material inspection, Resonac Corporation (President: Hidehito Takahashi) has succeeded in automating product development and inspection, and in significantly reducing the time required for inspection.  For example, the inspection time for solder particles used in anisotropic conductive film can be reduced to 1/140th of the time it previously took, while the inspection time for graphite fiber used as a conductivity aid for cathodes and anodes of lithium-ion secondary batteries can be reduced to 1/3rd of the conventional time.

The inspection processes for solder particles and graphite fibers typically consisted of visual inspections.  Issues included the fact that it was time-consuming, involved many processes, and there were variations in the amount of time it took for different inspectors to perform the inspections.  Meanwhile, automating the inspection processes using deep learning required having a vast number of images that would make up the basis for learning, engineers specializing in deep learning, and computers with powerful data processing capabilities.

The team of Research Center for Computational Science and Informatics which is dedicated to development of the image analysis system, succeeded in automating material inspections by leveraging the capabilities of a team specialized in deep-learning-based image analysis and introducing unique technologies.  In the development process, the team 1) introduced advanced image processing technology that enables the system to learn and perform analysis using just a small number of images, 2) narrowed down the deep learning areas to be used for inspection, and 3) introduced a web application.  The team also communicated closely with those who would actually perform the work on the manufacturing floor and prepared high-quality images for deep learning.  This automated material inspection, and simultaneously eliminated the variations in the amount of time it took for different inspectors to perform inspections.

In order to obtain highly accurate results, we have developed a number of unique innovations, such as combining conventional image processing techniques with deep learning based on its operating principles and combining multiple deep learning models.  
Going forward, the Research Center for Computational Science and Informatics plans to increase the number of engineers capable of performing image analysis by developing in-house expertise in image analysis, and develop intra-company web applications that utilize image analysis and large language models.

(1)This image processing technology enables learning and analysis using just a small number of images.
While deep learning typically requires thousands of images as a basis for learning, we can now use deep learning with as little as a few dozen images.  This is possible due to a technique that allows us to obtain multiple images from a single image*2 and then transfer learning to a model that has already learned from other data is tuned using the target data.

(2)Narrowing down the deep learning areas to be used for inspections
 While it typically takes time to train engineers specializing in deep learning, the range of operations used in materials inspection is limited, so we were able to train engineers quickly by focusing on technologies for analyzing three essential areas (calculation, measurement, and classification).

(3)Web application deployment
 We have created a setup where deep learning calculations are performed on a high-performance server in the cloud, and provided users with a web application connected to the cloud server through a network, thereby enabling users to conduct image analysis on a business-use mobile PC.
 We also facilitated maintenance and management work on the part of software developers with the use of cloud.

Photographed image

Analysis example

Solder particles: The photographed image is analyzed, and the quantity, dimension, and roundness of the particles on the screen are measured automatically.

Photographed image

Analysis example

Graphite fibers: The photographed image is analyzed, and each fiber is identified separately.
(Even where fibers overlap, they are not misidentified as the same fiber.)

*1. Deep learning: A method used to teach artificial intelligence (AI) how to process information.
*2. Multiple images are obtained by making adjustments to a single image, such as left-right inversion, saturation, etc.

— For Contribution to the Electrification of Vehicles and Others through Co-Creation with External Parties —

Resonac Corporation (TOKYO: 4004) (President: Hidehito Takahashi) is set to launch full-scale operation of Power Module Integration Center (“PMiC”), which is located at Oyama Plant in Tochigi Prefecture, to enhance development activity of materials for power semiconductor and its package, power module, which is indispensable component for vehicle electrification. So far, Resonac’s evaluation and simulation functions were for internal development activities. Hereafter, however, these functions will be used for co-creation activities with customers such as automakers and semiconductor manufacturers. Along with expansion of the EV market, thermal management techniques are becoming the more important. Resonac intends to set PMiC as a base to propose new materials for power modules.

PMiC is a research center which has equipment for prototype assembly, evaluation and simulation. Power modules and power semiconductor devices are essential components for vehicle electrification. However, they cause large amount of electrical loss and generate heat during operation inside power electronics systems. To extend the range of EV by increasing efficiency or not to cause overheating, it is important to reduce electrical loss and dissipate heat by improving performance of materials inside power module, which is difficult for automakers and semiconductor manufacturers to realize by themselves.

To propose appropriate materials to customers in a speedy manner, Resonac launched PMiC in July 2021, which can build prototype power modules by adopting various materials of Resonac, evaluate them under conditions similar to those set by our customers, give feedback to product development groups of each material. From this year, Resonac will start utilizing PMiC’s functions of prototype assembly, evaluation and testing for customers to co-create technological innovations tracing back to material development stage. Resonac will help customers to shorten their power module development periods with proposals derived from evaluation of appropriate materials combination based on performance required from customers. In 2022, Resonac helped a customer to cut the number of prototype evaluation by half. Now Resonac aims to help the customer make their development lead time even shorter by 2025.

With the global expansion of xEV sales, by 2030, the size of the power module market is expected to grow to be 3.9 times of that of 2021*1. In particular, the demand for power modules composed of SiC will expand. Resonac has 25%*2 share of global market for SiC epitaxial wafers and supplies a range of materials related to power modules, such as Cu sintering paste, thermal interface materials, molding compound, cooling devices, resin water jackets and heat-resistant coating materials. In addition, Resonac has accumulated expertise through many years of power module evaluation, and engineers who have experience of power module development in automotive device manufacturers have joined Resonac. Taking advantage of these features, Resonac will achieve further growth in expanding power module market.

 
Power module packaged in-house for the purpose of evaluation

*1　Source: “Automotive Power Module Market 2021” published on September 30, 2021 by Yano Research Institute Ltd.
*2　Estimation by Resonac.

— With Computational Science and Materials Analysis Professionals Also Stationed at the Same Resonac R&D Center —

In January 2023, Resonac Corporation (TOKYO: 4004) (President and CEO: Hidehito Takahashi) is launching a project to develop new semiconductor materials for 6G, the next-generation telecommunication system standard following 5G. Resonac will implement this development project, which starts with the molecular design stage, in cooperation with ventures and universities at its “Stage for Co-creation” R&D center newly opened in Yokohama City, where teams of experts, including those in computational science and materials analysis, have been stationed since the end of December. Companies around the world are competing for the development of 6G systems, which are expected to be put into practical use in around 2030. Resonac aims to develop materials that will support high-frequency data transmission in the terahertz band.

With 6G set to be 100 times faster than 5G, new semiconductor materials are needed to substantially reduce the transmission loss associated with the increased communication speed. To meet this requirement, Resonac will develop materials and technologies for composite materials from scratch, starting with the synthesis of materials to create resins and ceramics for filler materials, as well as developing interface control technologies. In order to identify the optimal combination of materials to provide the required characteristics, Resonac will perform simulations and use AI systems in the molecular design stage to swiftly derive the required chemical structural formulas, a feat that would be impossible using conventional methods. Also, while it would take three months to manually test one combination of materials, the use of AI and other systems will make it possible to test one combination per day, or as many as 90 different combinations in three months.

Such development activities will be supported by a group of professionals with expertise in fundamental technologies for R&D. As a base for open innovation and Resonac’s R&D center, the “Stage for Co-creation” concentrates experts in computational science, materials analysis, manufacturing process technology & equipment management for mass production, and the management & evaluation of chemical safety. For example, the “Research Center for Computational Science and Informatics” has 70 specialists in simulation, AI and MI. These specialists have engaged in the process of developing catalysts from the molecular design stage in the petrochemicals and basic chemicals fields. Many of them are shifting their focus to the development of semiconductor materials to support the development project. Moreover, experts on mass production technology will also make contributions in the production process examination stage.

At the “Stage for Co-creation”, the solution of social issues on a medium-to long-term basis will be pursued as the driving R&D theme. The first initiatives to be implemented under this theme include plastic chemical recycling as well as the development of semiconductor materials for 6G. Toward the achievement of carbon neutrality by 2050, research is being carried out in order to identify a method to recover raw materials for plastics such as ethylene directly from waste plastics. For this, research personnel need not only to overcome technological challenges but also to listen to advice and opinions from those who contribute to the sorted collection of waste plastics and use recycled products. In this and other projects, personnel at the “Stage for Co-creation” will work toward the solution of social issues by promoting dialogue and co-creation with a wide range of stakeholders, including local governments and consumers. 

 
Stage for Co-creation (Yokohama City)