Frontier Research Institute for Interdisciplinary Sciences
Tohoku University

Actinomycetes are an important group of microorganisms in drug discovery and industry, known for their ability to produce beneficial compounds, including antibiotics. The production of these compounds is regulated by signaling molecules (self-regulatory molecules). These signaling molecules play a crucial role in the exceptional biosynthetic capabilities of actinomycetes; however, only a few have been clearly identified. One reason for this is that signaling molecules are produced at extremely low concentrations, making their analysis highly labor-intensive. In this study, FRIS Associate Professor Yuta Kudo and his group have developed two original methods for the rapid identification of signaling molecules: By co-culturing actinomycetes with a resin (adsorbent) that enhances the production of signaling molecules, they successfully isolated and determined the structure of these molecules at a scale 1/100 of conventional methods. They established a chemo-enzymatic synthesis method by combining organic synthesis with enzymatic reactions utilizing biosynthetic enzymes from actinomycetes. This method enabled the efficient synthesis of signaling molecules and their rapid identification from various actinomycetes employing 12 synthetic standards. Furthermore, this research marks the first discovery of optical isomers of signaling molecules from actinomycetes. By elucidating the structural diversity and distribution of these molecules, the study expands our understanding of the regulatory mechanisms governing secondary metabolism in actinomycetes. Moving forward, these findings are expected to contribute to the increased production of valuable compounds and the discovery of novel bioactive substances. The results of this study have been published in RSC Chemical Biology, a journal of the Royal Society of Chemistry (RSC), and were made available online as an early release on February 25, 2025. The paper is open access. Publication Details: •Title: Establishment and Demonstration of a Rapid Identification Method for Microbial Signaling Molecules Regulating Compound Production •Author: Yuta Kudo*, Keiichi Konoki and Mari Yotsu-Yamashita (*corresponding author) •Journal: RSC Chemical Biology •DOI: 10.1039/d5cb00007f •URL: https://pubs.rsc.org/en/content/articlelanding/2025/cb/d5cb00007f

Researchers at Tohoku University have developed a groundbreaking titanium-aluminum (Ti-Al)-based superelastic alloy. This new material is not only lightweight but also strong, offering the unique superelastic capability to function across a broad temperature range—from as low as –269°C, the temperature of liquid helium, to +127°C, which is above the boiling point of water. This discovery holds significant potential for a variety of applications, including those in space exploration and medical technology.   Figure 1: Lightweight flexible alloy developed in this study. Sheng Xu, an Assistant Professor at Tohoku University’s Frontier Research Institute for Interdisciplinary Sciences, emphasized the importance of the alloy’s wide operational temperature range. "This alloy is the first of its kind to maintain superelasticity at such an extreme range of temperatures while remaining lightweight and strong, which opens up a variety of practical applications that were not possible before. The alloy's properties make it ideal for future space missions, such as creating superelastic tires for lunar rovers to navigate the extreme temperature fluctuations on the Moon’s surface." The alloy’s flexibility at extremely low temperatures makes it a promising material useful at liquid hydrogen environment, holding promise for applications in the forthcoming Hydrogen Society and various other industries. Of course, the alloy can be used in everyday applications requiring flexibility, such as medical devices like stents.   Figure 2: Stress-strain curves at various temperatures for the Ti-Al-Cr superelastic alloy. The surface temperature ranges of Earth, Mars and Moon are also shown. Currently, most shape-memory alloys—materials capable of regaining their original shape after force is removed—are limited to specific temperature ranges. The new Ti-Al-based alloy overcomes this limitation, offering wide applicability in fields that require materials with exceptional strength and flexibility, from space exploration to everyday medical tools. The research team employed advanced techniques such as rational alloy design and precise microstructure control. By using phase diagrams, the researchers were able to select alloy components and their proportions. Additionally, they optimized processing and heat treatment methods to achieve the desired material properties. The implications of this study extend beyond immediate practical applications. "This discovery not only sets a new standard for superelastic materials but also introduces new principles for material design, which will undoubtedly inspire further breakthroughs in materials science," Xu added.   Details of the breakthrough were published in the journal Nature on February 26, 2025.   Figure 3: A comparison between Ti-Al-Cr alloy and other superelastic alloys in terms of lightness and operational temperature range.   Publication Details: Title: A lightweight shape-memory alloy with superior temperature-fluctuation resistance Authors: Yuxin Song#, Sheng Xu#*, Shunsuke Sato, Inho Lee, Xiao Xu, Toshihiro Omori*, Makoto Nagasako, Takuro Kawasaki, Ryoji Kiyanagi, Stefanus Harjo, Wu Gong, Tomáš Grabec, Pavla Stoklasová, Ryosuke Kainuma* Journal: Nature DOI: 10.1038/s41586-024-08583-7   Press Release: Tohoku University https://www.tohoku.ac.jp/en/press/a_lightweight_flexible_alloy_for_extreme_temperatures.html School of Engineering, Tohoku University https://www.eng.tohoku.ac.jp/english/news/detail-,-id,3130.html J-PARC Center https://www.j-parc.jp/c/en/press-release/2025/02/27001478.html

As the world moves toward sustainable energy, hydrogen will likely play an invaluable role as a clean and versatile fuel. Yet, adoption of hydrogen technologies hinges on overcoming key challenges in electrocatalysis, where costly and scarce platinum-group metals have long been the industry standard. Taking one step to rectify this, a research team including Assistant Professor Linda Zhang from FRIS has now developed a new strategy that fine-tunes electronic interactions at the atomic level. The study introduces an innovative electronic fine-tuning (EFT) approach to enhance the interactions between zinc (Zn) and ruthenium (Ru) species, resulting in a highly active and stable catalyst for both the oxygen reduction reaction (ORR) and hydrogen evolution reaction (HER). By anchoring Ru clusters onto hierarchically layered Zn-N-C nanosheets (denoted as Ru@Zn-SAs/N-C), the team has designed a material that outperforms commercial platinum-based catalysts. "Our work demonstrates how precise control over electronic structures can fundamentally reshape catalytic performance," says Hao Li, associate professor at Tohoku University's Advanced Institute for Materials Research (WPI-AIMR) and corresponding author of the paper. "By leveraging the synergy between Zn and Ru, we have developed a cost-effective alternative to conventional platinum catalysts, offering new possibilities for sustainable hydrogen production." Key to this breakthrough is the strong electronic metal-support interaction (EMSI) between Zn and Ru, which optimizes the adsorption energy of critical reaction intermediates. X-ray absorption spectroscopy and computational modeling confirm that this synergy shifts *OOH and *OH adsorption energies to an optimal balance, enhancing ORR efficiency. Simultaneously, Ru sites achieve near-ideal hydrogen binding free energy, placing the catalyst at the peak of theoretical HER activity. "This research is not just about replacing platinum," Li explains. "It's about understanding how electronic properties at the atomic level dictate catalytic efficiency. That knowledge allows us to design better, more accessible materials for real-world applications." (a) Schematic illustration of synthesis procedure for Ru@Zn-SAs/N-C catalysts, where the electrostatic potential diagram of the iso-surface value is 0.03 e Å-3; (b-d) SEM images with different magnification; (e) TEM image; (f-g) HRTEM images, (f) the inset is the corresponding particle-size distribution of Ru clusters and (g) the inset shows the Moiré images extracted from the FFT; (h) AC HADDF-STEM image and integrated pixel intensities; (i) AFM image and corresponding height profiles of Ru@Zn-SAs/N-C ©Hao Li et al. These findings have significant implications for the affordability and scalability of hydrogen energy. By reducing dependence on expensive platinum while improving performance, this research contributes to the development of cost-effective hydrogen fuel cells, water electrolysis systems, and sustainable industrial processes. X-ray absorption spectroscopic characterization of Ru@Zn-SAs/N-C. ©Hao Li et al. Looking ahead, the team plans to further refine the EFT strategy, improve catalyst stability under real-world conditions, and develop scalable production methods. Applications in zinc-air batteries, fuel cells, and carbon and nitrogen reduction reactions are also under investigation. The research has been made available through the Digital Catalysis Platform (DigCat), the largest experimental catalysis database to date, developed by the Hao Li Lab. Details of its findings were published in the journal Advanced Functional Materials. The article processing charge (APC) was supported by the Tohoku University Support Program. Theoretical calculation analysis. ©Hao Li et al. Publication Details: Title: Synergistic Effects of Ruthenium and Zinc Active Sites Fine Tune the Electronic Structures of Augmented Electrocatalysis Authors: Tingyu Lu, Jing Li, Jingwen Ying, Ningkang Peng, Linda Zhang, Yizhou Zhang, Di Zhang, Songbo Ye, Lin Xu, Dongmei Sun, Hao Li, Yanhui Gu, Yawen Tang Journal: Advanced Functional Materials DOI: 10.1002/adfm.202422594 Press Release: Tohoku University https://www.tohoku.ac.jp/en/press/electronic_finetuning_unlocks_superior_performance.html Advanced Institute for Materials Research (WPI-AIMR) https://www.wpi-aimr.tohoku.ac.jp/en/achievements/press/2025/20250219_001933.html

Researchers at Tohoku University have achieved a significant advancement in opto-magnetic technology, observing an opto-magnetic torque approximately five times more efficient than in conventional magnets. This breakthrough, led by Mr. Koki Nukui, Assistant Professor Satoshi Iihama, and Professor Shigemi Mizukami, has far-reaching implications for the development of light-based spin memory and storage technologies. Opto-magnetic torque is a method which can generate force on magnets. This can be used to change the direction of magnets by light more efficiently. By creating alloy nanofilms with up to 70% platinum dissolved in cobalt, the team discovered that the unique relativistic quantum mechanical effects of platinum significantly boost the magnetic torque. The study revealed that the enhancement of opto-magnetic torque was attributed to the electron orbital angular momentum generated by circularly polarized light and relativistic quantum mechanical effects. When circularly polarized light is incident perpendicular to the surface of a nano-thin film of cobalt-platinum alloy, which consists of cobalt and platinum, an opto-magnetic torque is generated (red and blue vectors) that changes the magnetization direction (black vector). The opto-magnetic torque consists of components out-of-plane (red vector) and in-plane (blue vector). ©Nukui et al. This achievement allows for the same opto-magnetic effect to be produced with only one-fifth of the previous light intensity, paving the way for more energy-efficient opto-magnetic devices. The findings not only provide new insights into the physics of electron orbital angular momentum in metallic magnetic materials but also contribute to the development of high-efficiency spin memory and storage technologies that use light to write information. "These improvements could result in faster and more energy-efficient devices in the future," explains Mizukami. The research aligns with the growing interest in opto-electronic fusion technologies, combining electronic and optical technologies for next-generation applications. This discovery marks a significant step forward in controlling nanomagnetic materials using light and magnetism. These findings were published in Physical Review Letters on January 2, 2025. Examples of experimental data on magnetization oscillation driven by opto- magnetic torque measured by the pump-probe time-resolved magneto-optical Kerr effect: (a) Cobalt nano-thin film; (b) Cobalt-Platinum nano-thin film (Platinum concentration is 65% atomic ratio); (c) Platinum concentration dependence of the magnitude of opto-magnetic torques evaluated from the measured magnetization oscillations. Both the in-plane and out-of-plane torques increase with the platinum concentration. ©Nukui et al. Publication Details: Title: Light-Induced Torque in Ferromagnetic Metals via Orbital Angular Momentum Generated by Photon Helicity Authors: Koki Nukui, Satoshi Iihama, Kazuaki Ishibashi, Shogo Yamashita, Akimasa Sakuma, Philippe Scheid, Grégory Malinowski, Michel Hehn, Stéphane Mangin, Shigemi Mizukami Journal: PHYSICAL REVIEW LETTERS DOI: 10.1103/PhysRevLett.134.016701 Press Release: Tohoku University https://www.tohoku.ac.jp/en/press/breakthrough_in_optomagnetic_technology_5x_torque.html Advanced Institute for Materials Research (WPI-AIMR) https://www.wpi-aimr.tohoku.ac.jp/en/achievements/press/2025/20250107_001905.html School of Engineering, Tohoku University https://www.eng.tohoku.ac.jp/english/news/detail-,-id,3107.html

The global climate crisis, driven by the depletion of fossil fuels and rising atmospheric CO2 levels, has intensified the need for sustainable energy solutions. Among these, the electrochemical CO2 reduction reaction (CO2RR), particularly when integrated with renewable energy sources, has emerged as a promising approach. This process not only mitigates CO2 emissions but also addresses energy storage challenges by converting CO2 into high-value, carbon-neutral fuels. One of the standout products of CO2RR is formic acid (HCOOH), valued for its versatility in industries such as tanning, textiles, and pharmaceuticals, as well as its role as a high-energy-density liquid hydrogen storage medium.   "Formic acid is an indispensable chemical in various industries, and its potential as a hydrogen carrier makes it a critical component for a sustainable energy future," said Xue Jia, an assistant professor at Tohoku University's Advanced Institute for Materials Research (WPI-AIMR). Recent techno-economic analyses have also highlighted the practicality and economic feasibility of synthesizing formic acid through CO2RR, emphasizing its adaptability for future industrial applications.   To advance the development of efficient CO2RR catalysts, Jia and her colleagues conducted a comprehensive study, analyzing over 2,300 experimental reports from the past decade. Their findings underscored the superior activity and selectivity of tin-based catalysts, such as Sn−N4−C single-atom catalysts (SAC) and polyatomic Sn, for HCOOH production. These catalysts consistently outperformed others, including metal-nitrogen-carbon (M−N−C) catalysts and various metals, in terms of formic acid Faradaic efficiency (FE). Figure 1: Summary of the experimental CO2RR performance of 2,348 reported catalysts via a large-scale data mining. ©Hao Li et al. A significant aspect of the study was the influence of pH on catalyst performance. The team's analysis revealed that the selectivity and activity of HCOOH production increase with pH levels, as demonstrated in catalysts like SnO2 and Bi0.1Sn. However, conventional theoretical models that treat pH-dependent energetic corrections as constants failed to accurately predict activity at the reversible hydrogen electrode (RHE) scale.   "By incorporating electric field effects and pH-dependent free energy formulations, we were able to analyze the selectivity and activity of catalysts under actual working conditions, which is a significant step forward," explained Hao Li, associate professor at WPI-AIMR. This advanced modeling approach provided critical insights into the reaction mechanism, enabling a deeper understanding of the pH-dependent behavior of Sn-based catalysts.   The study also addressed a longstanding challenge: understanding how the structural differences between single-atom and polyatomic Sn catalysts impact their performance. The team discovered that Sn−N4−C SAC exhibits a monodentate adsorption mode, while polyatomic Sn adopts a bidentate mode. These distinct adsorption modes result in opposite dipole moments for the intermediate OCHO, significantly influencing the catalysts' activity and selectivity for CO2RR.   "This structural sensitivity, combined with pH-dependent modeling, has provided a comprehensive understanding of Sn-based catalysts and aligned our predictions with experimental observations," said Linda Zhang, Assistant Professor at Tohoku University's Frontier Research Institute for Interdisciplinary Sciences (FRIS). The research highlights the importance of considering structural and kinetic factors, beyond conventional thermodynamic models, for precise catalyst design. Figure 2: Surface reconstruction analyses. ©Hao Li et al. The implications of this work extend beyond CO2RR. By employing advanced computational techniques, such as density functional theory (DFT) and machine learning force fields (MLFF), the researchers demonstrated the potential of tailoring catalysts for specific reaction conditions. This approach is expected to drive the development of high-performance systems for a range of electrocatalytic processes.   "Precise modeling and advanced computational techniques are enabling us to design catalysts tailored for specific reaction conditions, paving the way for more efficient CO2 reduction technologies," adds Li. The study's integration of experimental and theoretical perspectives marks a significant step toward addressing climate challenges through innovative catalyst design. The findings were published in the journal Angewandte Chemie International Edition, with the authors expressing their gratitude to the Tohoku University Support Program for covering the article processing charge. Figure 3: pH-dependent modelling and benchmarking between theory and experiments. ©Hao Li et al. Publication Details: Title: Divergent Activity Shifts of Tin-Based Catalysts for Electrochemical CO2 Reduction: pH-Dependent Behavior of Single-Atom versus Polyatomic Structures Authors: Yuhang Wang, Di Zhang, Bin Sun, Xue Jia, Linda Zhang,, Hefeng Cheng, Jun Fan, and Hao Li Journal: Angewandte Chemie International Edition DOI: 10.1002/anie.202418228 Press Release: Tohoku University https://www.tohoku.ac.jp/en/press/researchers_unlock_new_insights_into_tin_based_catalysts_for_electrochemical_co_reduction.html Advanced Institute for Materials Research (WPI-AIMR) https://www.wpi-aimr.tohoku.ac.jp/en/achievements/press/2025/20250115_001913.html  

Artificial intelligence (AI) and AI-enabled robots are becoming a bigger part of our daily lives. Real-time, flexible interactions between humans and robots are no longer just science fiction. As robots become smarter and more human-like in both behavior and appearance, they are transforming from mere tools to potential partners and social entities.   This rapid evolution presents significant challenges to our legal and ethical frameworks, including concerns about privacy, safety, and regulation in the context of AI and robots. The Cambridge Handbook of the Law, Policy, and Regulation for Human-Robot Interaction, published by Cambridge University Press on November 21, 2024, explores and addresses these emerging issues. It is now available online as of December 2024.   Edited by Woodrow Barfield, Yueh-Hsuan Weng, and Ugo Pagallo, three experts in AI-related legal issues, the handbook gathers insights from social sciences, computer science, and engineering. It is the first book to specifically address issues of law, policy, and regulation focusing on human-robot interaction.       “Humanities are crucial to AI development,” says Yueh-Hsuan Weng, Associate Professor at the Institute for Advanced Study (IAS), Kyushu University, and the Frontier Research Institute for Interdisciplinary Sciences (FRIS), Tohoku University (Cross-appointment). He is also a co-editor of the book. “Tech professionals can create cutting-edge systems, but without input from legal and humanities perspectives, these systems may struggle to coexist with humans. We hope this book serves as a compass for developers, ensuring AI systems better benefit our society.”   Comprising 46 chapters, the handbook is organized into four parts. The opening section introduces the legal and ethical challenges arising from human-robot interaction, addressing issues such as trust for robots and anthropomorphism—where non-human entities are given human-like emotions or intentions. The second section explores the societal impacts of human-robot interaction, discussing questions about whether AI entities should be granted legal personhood and what steps are needed for the growing integration of robots into human life.   The third section looks deeper into ethical, cultural, and value-based issues in human-robot interaction. A key aspect of AI governance is aligning AI’s value judgments with human values, which can vary across regions, contexts, and cultural value systems. Through a range of scenarios, including the role of robots in long-term assistance, their potential function in religious settings, and intercultural challenges, this chapter reveals the complexities of value alignment. The book concludes by discussing the legal challenges posed by AI’s integration into society, offering insights into how consumer law, criminal law, and constitutional law may need to evolve to accommodate intelligent systems. This handbook brings together authors from various countries and presents case studies from across the globe. By offering diverse perspectives, it provides valuable insights into the ethical dilemmas emerging from our personal interactions with robots, sparking a global dialogue on these issues. “A major issue I addressed in the book is the AI pacing problem,” says Weng. This refers to the gap between rapid AI advancements and the slower pace of legislation. While many countries and organizations are working on regulations for AI-enabled robots, creating comprehensive laws often struggles to keep up with AI’s progress. “Governance mechanisms have been proposed, ranging from ‘hard’ legislation to ‘soft’ ethical guidelines. What’s needed now are solutions that balance enforceability and flexibility.”   One solution Weng proposed in his chapter is global AI ethics standards developed by the Institute of Electrical and Electronics Engineers (IEEE), the world’s largest technical professional organization. Currently, Weng chairs a working group at the IEEE and is compiling a database of AI-related ethical cases from various countries, modularizing core issues and region-specific concerns, aiming to help developers navigate and apply them effectively.   The handbook also addresses critical topics like anthropomorphism, robots in healthcare, and privacy protection, all requiring continued focus and collaboration. As algorithms enable robots to perform human-like actions, such as robot dogs dancing jazz, these behaviors challenge traditional ethical expectations and may reshape how future generations perceive concepts like “dogs.” Meanwhile, when people, especially older adults, are unfamiliar with robots, they may view robotic caregivers as true companions, leading to emotional challenges. Ethical guidelines are needed to ensure responsible use in these sensitive contexts. Additionally, balancing high-quality services with data security remains an urgent task that demands innovative regulatory solutions.   Reflecting on these topics, Weng emphasizes, “As human-AI interactions become more common, I hope designers, manufacturers, and users of robots will engage with our book. Responsible research and innovation are crucial for the development of AI and robots, and this requires input from people across various societal sectors. We warmly invite everyone to explore this book and join us in creating IEEE’s global standards for AI ethics.”     Publication Details: Book Publised: The Cambridge Handbook on the Law, Policy and Regulation for Human-Robot Interaction Woodrow Barfield, Yueh-Hsuan Weng and Ugo Pagallo (Eds), Cambridge University Press Title: Ethical Design and Standardization for Robot Governance Author: Yueh-Hsuan Weng DOI: 10.1017/9781009386708 ISBN: 9781009386708 URL: https://www.cambridge.org/core/books/cambridge-handbook-of-the-law-policy-and-regulation-for-humanrobot-interaction/5740D8AEA42968E6A195BEDF5CBD0E5C Press Release: Kyushu University https://www.kyushu-u.ac.jp/en/researches/view/318    

FRIS has established the Interdisciplinary Platform for Advanced Health Sensing (Endowed Research Division) through a contribution from Milbon Co., Ltd. This division, led by Associate Professor Masaki Okumura from the Creative Interdisciplinary Research Division, brings together researchers from six different fields: protein science, physical organic chemistry, bio-measurement, mass spectrometry, structural biology, and cosmetic science. The division aims to conduct interdisciplinary research that contributes to the development of cosmetics and quasi-drugs. By integrating knowledge from each field, the division expects to develop cutting-edge measurement techniques for biological systems and gain new insights. Specifically, the research will encompass areas such as proteomics and drug molecular design, contributing to advances in the medical and drug discovery fields, with the goal of exploring new approaches beyond traditional cosmetic research.   With the establishment of this division, a new interdisciplinary platform will be built to accelerate the development of new cosmetics and aim for social implementation, from basic research to product development. Tohoku University, aiming to become a leading research university in Japan and on par with global standards, and Milbon, a pioneer in innovative hair and skin research, will collaborate to drive new innovations.   Press Release: Milbon Co., Ltd. https://prtimes.jp/main/html/rd/p/000000095.000028306.html (in Japanese)  

Researchers from Tohoku University and the Massachusetts Institute of Technology (MIT) have unveiled a new AI tool for high-quality optical spectra with the same accuracy as quantum simulations, but working a million times faster, potentially accelerating the development of photovoltaic and quantum materials.   Understanding the optical properties of materials is essential for developing optoelectronic devices, such as LEDs, solar cells, photodetectors, and photonic integrated circuits. These devices are pivotal in the semiconductor industry's current resurgence.   Traditional means of calculation using the basic laws of physics involve complex mathematical calculations and immense computational power, rendering it difficult to quickly test a large number of materials. Overcoming this challenge could lead to the discovery of new photovoltaic materials for energy conversion and a deeper understanding of the fundamental physics of materials through their optical spectra.   A team led by Nguyen Tuan Hung, an assistant professor at the Frontier Institute for Interdisciplinary Science (FRIS), Tohoku University, and Mingda Li, an associate professor at MIT’s Department of Nuclear Science and Engineering (NSE), did just that, introducing a new AI model that predicts optical properties across a wide range of light frequency, using only a material’s crystal structure as an input. Lead author Nguyen and his colleagues recently published their findings in an open-access paper in “Advanced Materials.”   “Optics is a fascinating aspect of condensed matter physics, governed by the causal relationship known as the Kramers-Krönig (KK) relation”, says Nguyen. “Once one optical property is known, all other optical properties can be derived using the KK relation. It is intriguing to observe how AI models can grasp physics concepts through this relation.”   Obtaining optical spectra with complete frequency coverage in experiments is challenging due to the limitations of laser wavelengths. Simulations are also complex, requiring high convergence criteria and incurring significant computational costs. As a result, the scientific community has long been searching for more efficient methods to predict the optical spectra of various materials.   “Machine-learning models utilized for optical prediction are called graph neural networks (GNNs),” points out Ryotaro Okabe, a chemistry graduate student at MIT. “GNNs provide a natural representation of molecules and materials by representing atoms as graph nodes and interatomic bonds as graph edges.”   Yet, while GNNs have shown promise for predicting material properties, they lack universality, especially in representations of crystal structures. To work around this conundrum, Nguyen and others devised a universal ensemble embedding, whereby multiple models or algorithms are created to unify the data representation.   "This ensemble embedding goes beyond human intuition but is broadly applicable to improve prediction accuracy without affecting neural network structures," explains Abhijatmedhi Chotrattanapituk, an electrical engineering and computer science graduate student at MIT.   The ensemble embedding method is a universal layer that can be seamlessly applied to any neural network model without modifying the neural network structures. “This implies that universal embedding can readily be integrated into any machine learning architecture, potentially making a profound impact on data science,” says Mingda Li.   This method enables highly precise optical prediction based solely on crystal structures, making it suitable for a wide variety of applications, such as screening materials for high-performance solar cells and detecting quantum materials.   Looking ahead, the researchers aim to develop new databases for various material properties, such as mechanical and magnetic characteristics, to enhance the AI model’s capability to predict material properties based solely on crystal structures.   Figure: An AI tool called GNNOpt can accurately predict optical spectra based solely on crystal structures and speed up the development of photovoltaic and quantum materials. Publication Details: •Title: Universal Ensemble-Embedding Graph Neural Network for Direct Prediction of Optical Spectra from Crystal Structures •Author: Nguyen Tuan Hung, Ryotaro Okabe, Abhijatmedhi Chotrattanapituk, Mingda Li •Journal: Advanced Materials •DOI: 10.1002/adma.202409175 •URL: https://onlinelibrary.wiley.com/doi/10.1002/adma.202409175 Press Release: Tohoku University https://www.tohoku.ac.jp/en/press/ai_speeds_up_discovery_of_energy_and_quantum_materials.html  

Quantum squeezing is a concept in quantum physics where the uncertainty in one aspect of a system is reduced while the uncertainty in another related aspect is increased. Imagine squeezing a round balloon filled with air. In its normal state, the balloon is perfectly spherical. When you squeeze one side, it gets flattened and stretched out in the other direction. This represents what is happening in a squeezed quantum state: you are reducing the uncertainty (or noise) in one quantity, like position, but in doing so, you increase the uncertainty in another quantity, like momentum. However, the total uncertainty remains the same, since you are just redistributing it between the two. Even though the overall uncertainty remains the same, this ‘squeezing’ allows you to measure one of those variables with much greater precision than before.   This technique has already been used to improve the accuracy of measurements in situations where only one variable needs to be precisely measured, such as in improving the precision of atomic clocks. However, using squeezing in cases where multiple factors need to be measured simultaneously, such as an object's position and momentum, is much more challenging.   In a research paper published in Physical Review Research, Tohoku University’s Dr. Le Bin Ho explores the effectiveness of the squeezing technique in enhancing the precision of measurements in quantum systems with multiple factors. The analysis provides theoretical and numerical insights, aiding in the identification of mechanisms for achieving maximum precision in these intricate measurements. "The research aims to better understand how quantum squeezing can be used in more complicated measurement situations involving the estimation of multiple phases," said Le. "By figuring out how to achieve the highest level of precision, we can pave the way for new technological breakthroughs in quantum sensing and imaging."   The study looked at a situation where a three-dimensional magnetic field interacts with an ensemble of identical two-level quantum systems. In ideal cases, the precision of the measurements can be as accurate as theoretically possible. However, earlier research has struggled to explain how this works, especially in real-world situations where only one direction achieves full quantum entanglement. This research will have broad implications. By making quantum measurements more precise for multiple phases, it could significantly advance various technologies. For example, quantum imaging could produce sharper images, quantum radar could detect objects more accurately, and atomic clocks could become even more precise, improving GPS and other time-sensitive technologies. In biophysics, it could lead to advancements in techniques like MRI and enhance the accuracy of molecular and cellular measurements, improving the sensitivity of biosensors used in detecting diseases early. "Our findings contribute to a deeper understanding of the mechanisms behind the improvement of measurement precision in quantum sensing," adds Le. "This research not only pushes the boundaries of quantum science, but also lays the groundwork for the next generation of quantum technologies." Looking ahead, Le hopes to explore how this mechanism changes with different types of noise and explore ways to reduce it.     Figure: A visual comparison between the familiar act of squeezing a lemon and the concept of quantum squeezing in a sensor. (License: CC BY-NC-SA) Publication Details: •Title: Squeezing-induced quantum-enhanced multiphase estimation •Author: Le Bin Ho •Journal: Physical Review Research •DOI: 10.1103/PhysRevResearch.6.033292 •URL: https://journals.aps.org/prresearch/pdf/10.1103/PhysRevResearch.6.033292   Press Release: Tohoku University https://www.tohoku.ac.jp/en/press/squeezing_increased_accuracy_out_of_quantum_measurements.html  

Researchers at Tohoku University and Utsunomiya University have made a breakthrough in understanding the complex nature of turbulence in structures called “accretion disks” surrounding black holes, using state-of-the-art supercomputers to conduct the highest-resolution simulations to date. An accretion disk, as the name implies, is a disk-shaped gas that spirals inwards towards a central black hole.   There is a great interest in studying the unique and extreme properties of black holes. However, black holes do not allow light to escape, and therefore cannot be directly perceived by telescopes. In order to probe black holes and study them, we look at how they affect their surroundings instead. Accretion disks are one such way to indirectly observe the effects of black holes, as they emit electromagnetic radiation that can be seen by telescopes.   “Accurately simulating the behaviour of accretion disks significantly advances our understanding of physical phenomena around black holes,” explains Yohei Kawazura, “It provides crucial insights for interpreting observational data from the Event Horizon Telescope.”   The researchers utilized supercomputers such as RIKEN's "Fugaku" (the fastest computer in the world up until 2022) and NAOJ's "ATERUI II" to perform unprecedentedly high-resolution simulations. Although there have been previous numerical simulations of accretion disks, none have observed the inertial range because of the lack of computational resources. This study was the first to successfully reproduce the "inertial range" connecting large and small eddies in accretion disk turbulence.   It was also discovered that "slow magnetosonic waves" dominate this range. This finding explains why ions are selectively heated in accretion disks. The turbulent electromagnetic fields in accretion disks interact with charged particles, potentially accelerating some to extremely high energies. Figure: Artistic image of accretion disk turbulence. The inset is the magnetic field fluctuations computed by the simulation of this study. ©Yohei Kawazura In magnetohydronamics, magnetosonic waves (slow and fast) and Alfvén waves make up the basic types of waves. Slow magnetosonic waves were found to dominate the inertial range, carrying about twice the energy of Alfvén waves. The research also highlights a fundamental difference between accretion disk turbulence and solar wind turbulence, where Alfvén waves dominate.   This advancement is expected to improve the physical interpretation of observational data from radio telescopes focused on regions near black holes. The study was published in Science Advances on August 28, 2024. Publication Details: Title: Inertial range of magnetorotational turbulence Authors: Yohei Kawazura and Shigeo S. Kimura Journal: Science Advances DOI: 10.1126/sciadv.adp4965 URL: https://www.science.org/doi/10.1126/sciadv.adp4965 Press Release: Tohoku University https://www.tohoku.ac.jp/en/press/supercomputer_simulations_reveal_the_nature_of_turbulence_in_black_hole_accretion_disks.html