Researchers Conduct Automated Semiconductor Study with Machine Learning.

Automated semiconductor research with machine learning

The emergence of new semiconductor thin materials requires quantitative analysis of the total number of RHEED records, which is time-consuming and requires expertise. To address this problem, researchers at the Tokyo University of Science identified a machine learning approach that can help automate RHEED data analysis.

Researchers conduct automated semiconductor research using machine learning.
Reflection High Energy Electron Diffraction (RHEED) is an imaging technique widely used to analyze the surface structures of materials grown via physical vapor deposition. However, RHEED produces huge amounts of data and is a skill-intensive tool to use. To address this problem, scientists at TUS and NIMS are using machine learning techniques to automate some of the more difficult parts of the analysis. Image Credit: Naoka Nagamura of the National Institute for Materials Science and Tokyo University of Science.

Their discoveries have the potential to significantly accelerate semiconductor research and lay the foundation for faster, more energy-efficient electronic devices.

Since its beginnings in the mid-twentieth century, the semiconductor industry has grown significantly, paving the way for the rapid digitization of society through the rapid communication and information technology that made it possible. Currently, as global energy demand tightens, there is a huge market for faster, more uniform and energy-efficient semiconductor devices.

In addition, advanced semiconductor methods have already reached the nanometer scale, and the structural analysis of semiconductor nanofilms is now used in the design of new high-performance materials.

RHEED (reflection high-energy electron diffraction) is a popular analytical technique for this purpose. RHEED can be used to evaluate the atomic structures that develop on the surface of thin films and could even receive structural adjustments as the thin film is synthesized.

Unfortunately, despite its many benefits, RHEED is sometimes compounded by the fact that production trends are difficult to define. In almost all cases, a highly skilled experimenter is needed to understand the vast amounts of data produced by RHEED in the type of diffraction patterns. So what if researchers could let machine learning do most of the work when it comes to RHEED data analysis?

dr. Naoka Nagamura, visiting professor at Tokyo University of Science (TUS) and senior researcher at Japan’s National Institute for Materials Science (NIMS), has led a group of researchers in this effort.

The group studied the potential of using machine learning to automatically analyze RHEED data in their most recent study, published online Sept. 9.e June 2022, in the international magazine Science and technology of advanced materials: methods

TUS and NIMS, Japan, collaborated on this project, which was funded by JST-PRESTO and JST-CREST. Ms. Asako Yoshinari, Prof. Masato Kotsugi from TUS and Dr. Yuma Iwasaki from NIMS contributed to the article.

The scientists focused on the surface superstructures that develop on the first atomic layers of pure monocrystalline silicon (one of the most versatile semiconductor materials), which depends on the number of indium atoms adsorbed and minute temperature fluctuations.

Atoms stabilize in periodic patterns other than those found in the crystal body in surface superstructures, which are atomic arrangements specific to crystal surfaces that depend on variations in the environment. Surface superstructures are of great interest for materials research because they often exhibit distinctive physical features.

To group samples into different clusters based on different metrics of similarity, the team’s first step was to use different hierarchical clustering techniques. This method is used to count the number of different superstructures of the surface. After experimenting with different methods, the scientists found that Ward’s method could most effectively track real-life phase transitions in surface superstructures.

The next step was to identify the ideal process parameters for creating each of the surface superstructures found. The main focus was on the indium deposition time for which each superstructure was most thoroughly made.

Other commonly used dimensionality reduction techniques, such as principal component analysis, did not work effectively. Remarkably, the best deposition timings for each superstructure can be accurately and autonomously determined using non-negative matrix factorization, a separate clustering and dimensionality reduction method.

Our efforts will help automate the work that typically requires time-consuming manual analysis by specialists. We believe our study has the potential to change the way materials research is done and allow scientists to spend more time on creative pursuits

dr. Naoka Nagamura, visiting professor at Tokyo University of Science

Overall, the results of this work should lead to innovative and practical applications of machine learning for materials science, an important area of ​​materials informatics. As a result, as current products and technology have improved with better materials, it would have an impact on how people live their daily lives.

Our approach can be used to analyze the superstructures grown not only on thin-film silicon single crystal surfaces, but also on metal crystal surfaces, sapphire, silicon carbide, gallium nitride and several other important substrates. Therefore, we expect our work to accelerate the research and development of the next generation of semiconductors and high-speed communication equipmentconcludes Dr. Nagamura.

Upcoming discoveries that help automate difficult data processing and reduce effort for scientists are definitely something researchers expect to see more of.

Magazine reference:

Yoshinari, A., et al† (2022) Skill agnostic analysis of high energy electron diffraction reflection patterns for Si(111) surface superstructures using machine learning. Science and technology of advanced materials: methodsdoi.org/10.1080/27660400.2022.2079942

Source: https://www.tus.ac.jp/en/

Leave a Comment

Your email address will not be published.