【Ten Years of Rapid Development in Chemical Equipment】1748-2024 – Amazing! Robot chemists make chemical experiments safer! !

1. Timing: An inevitable trend in scientific research development. At the end of 2011, Jiang Jun decided to return to China and join the University of Science and Technology of China, to work again with his supervisor Luo Yi, whom he had studied under while pursuing his PhD at the KTH Royal Institute of Technology in Sweden. Both are engaged in research in theoretical and computational chemistry, but with different focuses. Luo Yi focuses on the analysis of high-resolution single-molecule chemical imaging, while Jiang Jun is more interested in material design and the study of microscopic mechanisms. One day at the end of 2013, Luo Yi was on a business trip with Jiang Jun. On the way, Luo Yi raised a question: Theoretical calculations based on independent cases, although in line with traditional research paradigms, are too inefficient. Is it possible to improve efficiency through big data technology? After conducting research, Jiang Jun found that our country suffers from a severe lack of accumulation of scientific data, placing it in a very passive position academically. “We conducted a great deal of research, but the final results were published in foreign-language journals. To cite the data from these papers, one has to pay, and only access rights are provided; large-scale downloading is not allowed. ”Jiang Jun believes that to make good use of big data technology, it is necessary to establish a data system first. Since 2014, Jiang Jun has led students in collecting the chemical experience and knowledge \"accumulated\" in textbooks, papers, and patents, and over three years has built a comprehensive database of chemical materials specific to the Chinese people. Once there is enough data, the next question is how to use it. Technically, artificial intelligence is the best tool for handling big data. However, this is a completely new tool for Luo Yi and Jiang Jun. If you don’t understand, learn; if it’s difficult, face it. By leveraging their expertise in spectroscopy and catalysis, and applying artificial intelligence methods to their research, they have achieved good results. It is worth mentioning that several powerful descriptors discovered by Jiang Jun were featured in two special articles published by the journal Science. Although artificial intelligence technology is integrated, using big data to solve real-world chemical problems remains a challenge. “Because the data itself lacks logical connections based on knowledge. ”Luo Yi said. One day in 2017, Luo Yi came up with a good idea: could he take advantage of the extensive expertise of two professors at his university in the fields of luminescent materials and catalysts, and integrate their knowledge with large databases? Excitedly, he called Jiang Jun to his office and said, “Read all the papers published by Zhang Guoqing and Xiong Yujie; in this way, a ‘Zhang Guoqing brain’ and a ‘Xiong Yujie brain’ can be created.” When studying luminescent materials, the ‘Zhang Guoqing Brain’ is utilized’ ； When designing catalysts, the ‘Xiong Yujie Brain’ is brought into use. In this way, those of us with limited knowledge of chemistry can also make good use of data. ” In this way, Jiang Jun and his classmates manually labeled massive amounts of text data, gradually enabling computers to read and analyze documents on their own; it took only two years to develop a \"chemical brain\". “With the “chemical brain” in place, the next step is to find a way to equip it with practical “hands” to operate. In 2020, Andrew Cooper’s team at the University of Liverpool in the UK created the world’s first mobile robot lab assistant. After studying the relevant articles they published, Jiang Jun said to Luo Yi with confidence, \"We can not only use robots to conduct experiments, but we can also equip them with a ‘chemical brain’.\" ” That same year, at an interdisciplinary forum held at the University of Science and Technology of China, Jiang Jun met Shang Weiwei, an associate professor in the School of Information Science and Technology who was working on robotics research; the two of them immediately agreed to collaborate on developing robotic chemists. Thus, under Jiang Jun’s leadership, they quickly assembled a youth team that worked day and night to tackle the problem together. In December 2021, the world’s first data-intelligence-driven machine chemist was finally created in China. It integrates a \"chemical brain,\" combining two mobile robots, 19 intelligent chemical workstations, and high-throughput computing systems to cover the entire process of scientific research methodologies. In September 2022, the relevant research findings were published in **Science Reviews**. The review expert commented that the chemist’s “robotic systems, workstations, and intelligent chemical brain” in this machine are “the most advanced,” and that it “will have a tremendous impact on chemical science.” “Machine chemists cannot be created out of thin air one day. ”Luo Yi said, “This is a process of keeping up with the times, absorbing new knowledge, and using new technologies to solve problems; it is also an inevitable outcome of ongoing scientific research efforts.” ”

2 Geographical advantage: The strength of interdisciplinary collaboration. Creating a robotic chemist who can read, think, act, and ultimately solve problems is no easy task. Jiang Jun believes this is mainly due to the Chinese Academy of Sciences’ tradition of carrying out integrated research efforts, as well as the relaxed research environment at the University of Science and Technology of China and its advantages in interdisciplinary collaboration. He recalled that after a big data model was developed, it needed further validation through experiments, which meant greater investment. At the beginning of 2021, the Chinese Academy of Sciences released a list of major scientific issues in basic research, frontier fields, and key core technologies for that year, and “Machine scientists for data-driven chemistry, materials science, and biological sciences” was one of the 46 projects on the list. Jiang Jun mustered up the courage and decided to take on the challenge. Soon, his project application was approved. From 2021 to 2025, the Chinese Academy of Sciences will allocate 20 million yuan in funding to provide continuous support to Jiang Jun’s team over 5 years. Also in 2021, Jiang Jun submitted a \"Proposal Form for Key Projects in ‘Double First-Class’ Disciplines\" to the University of Science and Technology of China. After research, the school provided 10 million yuan in financial support. At the same time, the Hefei Center for Microscale Materials Science at the University of Science and Technology of China, along with the School of Chemistry and Materials Science, stepped in to provide the research team with a workplace and research equipment. “This response mechanism of rapid decision-making, quick support, and swift funding has been a great source of motivation for me, and it also provides timely and adequate support for research. In the field of artificial intelligence, technological advancement happens at an extremely fast pace; falling behind once means falling further behind thereafter. ”Jiang Jun said. USTC also has a natural advantage: it brings together researchers from various disciplinary backgrounds, who can complement each other’s strengths and work together to turn ideas into reality more quickly. For example, robots are an important part of the entire platform. Initially, after discussing it with Jiang Jun, Shang Weiwei was willing to change the team’s research focus and dedicate themselves to developing experimental robots. When Andrew Cooper visited the Machine Chemist Laboratory, what he envied the most was the fact that Jiang Jun and his team had their own robotics R&D team. Jiang Jun said with pride, \"Every line of code in the chemical robot’s instruction set was written by our team. If we want to create a new action, we can adjust the code at any time.\" Andrew Cooper, on the other hand, had to ask the company to change the code, because his robot was purchased. ”

3 People: The support of a young team. Just as young as the machine chemists is the R&D team, whose members are mainly energetic individuals born in the 1990s and 1995s. “The things these young people have created are far better than I imagined, and the results of each stage of research have exceeded expectations. ”Luo Yi said. For example, in the early stages of data collection, Xiao Hengyu, Huo Yaoyuan, and Feng Chao, who were undergraduates at the Youth Class College of the University of Science and Technology of China, demonstrated exceptional research and development capabilities. At first, Luo Yi and Jiang Jun held regular discussion sessions to provide guidance to the students. But at the seminar, the two of them found that they often couldn’t get a word in at all. What often happened afterward was that the students engaged in lively discussions, while the teacher listened quietly and watched with satisfaction. “It must be admitted that in the era of information science, young people are more sensitive and have a stronger understanding than we do; their way of thinking has long been ‘digitalized’. ”Thus, Luo Yi and Jiang Jun reached an agreement: to assign the most important and difficult tasks to the younger people. It turned out that their decision was correct. In the development of Machine Chemist, in addition to building a database, Xiao Hengyu also taught himself computer software and developed version 1.0 of an operating system. This operating system not only conforms to the operational conventions of chemical experiments, but also leverages the automation advantages of robots, as well as allowing the integration of intelligent models. Recently, Xiao Hengyu upgraded the operating system to version 2.0, enabling hardware plug-and-play, universal model migration, and cloud-based operation sharing. Professional engineers were amazed upon seeing this architecture: “Similar frameworks have only been seen in Siemens’ smart laboratories; they can only be developed by experienced engineering teams at great cost.” ” There is a saying in the field of artificial intelligence: there is as much human effort behind the intelligence as there is of intelligence itself. The most important thing in developing a “chemical brain” is to endow computers with the ability to understand substances. At that time, Zhang Baicheng, who was a master’s student, spent weeks working day and night on annotating the materials; he was overwhelmed by the amount of literature to read, yet the amount of work completed was minimal. Later, he changed his approach and teamed up with several classmates who had professional backgrounds to work together. Together, they built a chemical dictionary and organized reaction data, quickly designing the first generation of literature reading system and obtaining the first batch of high-quality annotated data. “Zhao Luyuan, born in the 1990s, leveraged his undergraduate knowledge of electronic engineering to focus on the renovation of chemical workstations. She remembered that when working with her junior colleague Cao Jiaqi to develop drivers for the solid sample injector, they were unable to accurately understand and utilize the instrument’s interfaces due to insufficient information. They consulted foreign manufacturers and agents on multiple occasions, only to be told “we can’t disclose that” or “it’s a trade secret”. “Later, we spent 3 weeks repeating 3,000 experiments of sending signals, waiting for feedback, and recording responses, gradually working backwards to figure out the underlying control logic, and successfully achieved remote control of the robot. ”Zhao Luyuan said that foreign manufacturers took the initiative to contact them, wanting to purchase the drivers they had developed. Things like this are not uncommon in the efforts of the Machine Chemist project. To Jiang Jun’s relief, these young people never limit themselves by their field of expertise; instead, they take the initiative to learn, face challenges head-on, and find ways to solve problems. At present, these students have become the backbone of the laboratory, each with their own special skills. Among them, Xiao Hengyu is working on designing a centralized, large-scale machine scientist platform; Zhang Baicheng is endowing robots with spectroscopic intelligence; and Zhao Luyuan is exploring chemical workstations suitable for industrial applications.

4 Performance: Exceptional research capabilities. Over thousands of years, humanity has continued to uncover the mysteries of the material world in the field of chemistry, creating many amazing items and materials. However, chemistry is also a subject that requires a bit of “luck”. There are many substances that can react, and different amounts will lead to different results. More importantly, the reaction vessel is entirely a “black box”; what happens inside and how the process unfolds cannot be calculated or measured in many cases. Therefore, in the field of chemistry, there are always some outcomes that arise unintentionally, as well as some situations of futile effort. It may take a long time to achieve the desired result. The emergence of machine chemists holds the potential to change this situation. Jiang Jun explained that, thanks to the precise automated operations of robots, machine chemists can carry out tasks \"without fatigue\". Meanwhile, the “chemical brain” carries out quantum chemical simulations in parallel, and by combining theoretical big data with experimental small data, it creates an artificial intelligence model with predictive capabilities; ultimately, this leads to the identification of experimental formulas that are predicted by the algorithm to be optimal overall. It is this integrated theory-practice research paradigm that enables the rapid identification of the best combinations from millions of possible formulations, thereby greatly accelerating the discovery process of new substances. Taking the development of a \"high-entropy non-precious metal oxygen-evolving catalyst\" as an example, Jiang Jun demonstrated the extraordinary capabilities of machine chemists. High-entropy materials are characterized by high disorder, high randomness, and high complexity; they can enhance the stability of energy batteries and are very important for the development of new energy sources. Using the traditional approach, researchers would have to select 5 elements from 29 non-precious metal elements to create over 550,000 different combination ratios, and the trial-and-error process would likely take 1,400 years. By reading 16,000 catalysis papers, the machine chemist independently selected five non-precious metal elements, and by integrating 25,000 sets of theoretical calculation data with 207 sets of full-process machine experiment data, it established and optimized a predictive model that reduced the development cycle to 5 weeks. Another example of the powerful research capabilities of machine chemists is a study on catalysts based on Martian meteorites. For a long time, migrating to Mars has been humanity’s dream. To realize this dream, the first challenge to overcome is the lack of oxygen in Mars’ atmosphere. Not long ago, machine chemists successfully created practical oxygen-producing electrocatalysts using Martian meteorites. After its publication in Nature Synthesis, this achievement generated a great deal of attention. Articles on the Nature website argue that this research “provides new technical methods for Mars exploration and the search for extraterrestrial civilizations”. Jiang Jun explained that if experiments are conducted manually, using 5 different types of Martian ores as raw materials, there are over 3.76 million possible combinations of formulas. Calculated based on at least 5 hours per experiment, it might take 2000 years to find the optimal formula. The machine chemist completed the experiment in just 5 weeks. The optimized catalyst formula provided can supply sufficient oxygen. Jiang Jun and his colleagues affectionately named the machine chemist “Xiao Lai,” signifying the beginning of a new era in the exploration of material science. Its combination with human wisdom will create entirely new possibilities in physical science, guiding humanity to continuously expand the boundaries of knowledge across the stars and oceans.

5 Goal: A new paradigm for chemistry research driven by data and intelligence. Globally, in the field of machine chemists, only the University of Glasgow in the UK and the University of Science and Technology of China have achieved automated literature reading capabilities, while the University of Liverpool in the UK and the University of Science and Technology of China have achieved robotic mobility. Most importantly, similar products developed by foreign counterparts have not covered the entire process including literature review, synthesis, characterization, testing, and data analysis; the types of research that can be conducted are relatively limited, and there is still a gap before such products can achieve intelligence. “This shows that China’s machine chemists are on par with their international counterparts in terms of research and development, and hold a leading edge in certain areas, boasting great potential for further development. ”Jiang Jun said. Drawing on the successful experiences of machine chemists, Luo Yi believes that in the future, science can be engineered; outstanding researchers can be brought together, with machines serving as the platform and data as the key element, to carry out organized scientific research. Luo Yi said that their goal is to rapidly expand the scale of the Machine Chemist technology platform, enhance its capabilities, and build the best “chemical hospital” in the world. “We aim to expand the scale of this platform and enable collaboration across multiple locations, in order to develop intelligent robots of various types – such as those that can fly in the air, move on the ground, or perform tasks at a desktop level. These robots will be capable of carrying out tens of thousands of chemical experiments per day, thus meeting the research and development needs in fields such as aerospace, healthcare, materials science, chemical manufacturing, and biomedicine. ” Currently, the R&D team is equipping robots with the ability to perceive and predict chemical processes. “For example, endowing robots with spectroscopic intelligence. Previously, the focus was on training robots to conduct experiments; now, it is hoped that robots can use spectroscopic intelligence to directly observe changes in the microscopic world, thereby enabling faster prediction of chemical changes and going beyond human understanding of experiments. ”Luo Yi said. Luo Yi hopes that more researchers will bring important scientific problems to the Machine Chemist platform in order to improve research efficiency and produce more new results. “We live in an era of rapid technological advancement; we must actively embrace new technologies and approach the changes in the world with a positive and open mindset. ”Luo Yi described the team’s ultimate goal as creating creative intelligent chemists, in order to establish a new paradigm for chemical research driven by data and intelligence at an early stage.

Thank you for sharing the information; the forum is even better with you!

Thank you for organizing and sharing the materials! Give a thumbs up and encouragement to the achievements of China’s chemical technology and equipment industry!