[Explanation] In the mechanical chemist laboratory of the University of Science and Technology of China, researchers only need to input instructions, and a flexible robot will shuttle between the consoles, extending the robotic arm to prepare reagents.

This is the result of the team of Professor Jiang Jun, Department of Chemical Physics, University of Science and Technology of China. The team has developed the world's first machine chemist platform that integrates reading literature, designing experiments independently, and covering the entire process of material development. From millions of possible combinations Find optimal solutions to accelerate material development.

　　[Explanation] Recently, the research results have been published in the latest issue of the academic journal "National Science Review".

Industry experts believe that this achievement leads chemical research towards the future trend of digitization of knowledge understanding, instructional operation, and template creation, and will have a huge impact on chemical science.

　　[Concurrent] Jiang Jun, Professor of the Department of Chemical Physics, University of Science and Technology of China

　　The advantage of the machine chemist platform is to generate massive data through high-standard and repeatable machine experimental operations, and then extract digital knowledge graphs and artificial intelligence models from them, guide chemical robots to automatically optimize, and produce better and more efficient chemistry. products or new materials.

　　[Commentary] Jiang Jun introduced that the machine chemist platform consists of three parts: "chemical brain", robot experimenter and intelligent chemical workstation.

The core "chemical brain" builds a knowledge map by analyzing a large number of chemical experiments and theoretical data, and realizes the ability to read and understand literature, design chemical experiments, and optimize programs independently.

Data exchange and interaction can be carried out between the robot experimenter and the 15 chemical workstations, and the chemical experiments can be carried out with precise coordination.

　　[Concurrent] Jiang Jun, Professor of the Department of Chemical Physics, University of Science and Technology of China

　　Without a computing brain, its robot's thinking ability is very weak. It can only perform primary optimization based on the experimental data at that time. Based on the experimental data, it is easy to fall into this empirical trap. It can only find partial The optimal solution, then the computing brain we loaded is driven by a physical model, and the physical model can help us dig out a better formula from the data, so we can find the global optimal solution.

　　[Explanation] Since 2014, the team has started related scientific research work, from mobile robots to chemical workstations, from intelligent operating systems to scientific databases.

After 8 years, the team members overcame a series of difficulties and finally built the mechanochemist platform.

　　[Concurrent] Zhu Qing, associate researcher at the National Research Center for Microscale, University of Science and Technology of China

　　I think the biggest difficulty may be in two aspects. One is that we do not have any reference from our predecessors, because this is indeed the first project in the world to do such an intelligent scientist.

There is also a cross-integration involving multiple disciplines, because we have chemistry, machinery, computers, software, and robotics. This involves a lot of research. Subject.

　　[Commentary] There are nearly 30 members in Jiang Jun's team, and Zhao Luyuan, a 25-year-old doctoral student, is one of them.

Previously, when she and her team members were writing the driver program of the solid-state sampler, they could not accurately understand and call the instrument interface due to insufficient information, so it took 3 weeks to repeat the signal sending test 3000 times, and reversely tried out the underlying control logic , and finally successfully wrote a script to achieve remote control.

　　[Concurrent] Team member Zhao Luyuan

　　There are certainly difficulties, and the main thing to overcome is not to give up.

It feels like what Teacher Jiang said just now, some difficulties may really need to rely on themselves.

Then there is the need to continue to study here, and then include various trial and error. I think in the end, perseverance, talent is victory.

　　[Explanation] At present, the team has established a material gene innovation research platform and built the largest material database platform in Asia.

The machine chemist platform has stronger chemical intelligence and extensive new material development capabilities, covering photocatalytic and electrocatalytic materials, luminescent molecules, optical thin film materials and other fields. The scope of application will continue to expand as the platform is upgraded and expanded.

　　Reported by Zhang Jun, Chu Weiwei, Hefei, Anhui

Responsible editor: [Liu Xian]