Chapter 22 Rapid progress
Bawendi is an expert in the field of perovskites, but he won the Nobel Prize in 23 not for his achievements in perovskites, but for his work on the chemical preparation method of quantum dots of innovation.
Quantum dots are tiny semiconductor particles with unique optical and electronic properties that can be used in LED, infrared detection, solar cells and other fields.
When Bawendi heard the other party choose perovskite, he smiled: "Wright, I'm not sure how good you are at doing experiments, because your paper didn't show much of your talent in this area.< br>
If you want to publish a paper in the field of perovskite, the requirements for experiments are very high. You need to be able to produce results that others cannot, or to observe phenomena that others cannot observe, and to be able to carry them out. Theorizing.”
Bawendi continued: "Of course if you want to publish a paper in a top journal, I know this is very important for Chinese students, because you need these supporting materials to prove your abilities so that you can strive for better results when you return to China.
From this point of view, perovskite is indeed a very good direction. There are no less than ten papers in the field of perovskite every year in Nature alone. With Science, this is indeed an easy topic. The direction of the results.
It’s just that I’m not sure how talented you are in doing experiments.”
Chen Yuanguang’s self-confidence is nothing compared to the topic given by Pangu: "I think I can."
Bawendi stared at his eyes for a while, then nodded: "ok, you think you can, then young man, let's start working!"
After Bawendi sent him a list of papers and related textbooks, he asked him to study on his own. This is the MIT style. The instructor assumes that you are a genius.
"Has Chen been to the laboratory?" Bawendi asked.
An Indian student said: "I have never seen him."
"Okay." Bavandi was a little confused.
When Chen Yuanguang sent him an email again, hoping that he would recommend some new papers, Ba Wendi replied to the email and asked: "Do you have any ideas about the papers?"
"Absolutely, I plan to study how to use machine learning methods to predict the high-throughput effects of antisolvents on perovskite stability." Chen Yuanguang replied.
After reading it, Ba Wendi realized that Chen Yuanguang had gone on the path he was familiar with, doing computational chemistry.
Machine learning is so good. In 2016, AlphaGo emerged and defeated Lee Sedol at the ninth level of Go. It was hailed as the first year of machine learning.
Two years have passed now, and everyone feels that artificial intelligence should be combined with this major, just like Internet+ was everywhere twenty years ago.
Now is artificial intelligence +, but artificial intelligence talents are hard to find. The prices offered by major companies such as Google, Amazon, and FB for talents in the field of artificial intelligence are sky-high, and some entrepreneurial companies offer even higher prices. One is higher than the other.
It is too difficult to recruit someone who understands both artificial intelligence and chemistry, let alone someone who can combine the two.
Even Ba Wendi couldn't find talents in this field. The appearance of Chen Yuanguang gave Ba Wendi some ideas, but he didn't expect that the other party would adapt so quickly and found him in just one month. Junction point.
"Wright, come to my office tomorrow and let's talk about specific research directions." After seeing Chen Yuanguang's reply, Bawendi stopped sending the email and called directly.
The next day, "Wright, please tell me roughly what you think." Bawendi handed the coffee to Chen Yuanguang.
“What I think is that we can use a combination of automatic characterization, chemical robotic synthesis technology and machine learning to explore how the choice of antisolvent affects the intrinsic stability of perovskite. For example, we choose different The ends are combined, such as MAPbI3, CsPbI3 and CsPbBr3, etc., used to synthesize some combination libraries, each library will have its own unique combination
I expected that we would be able to synthesize over a thousand ingredients in total, and then make each library twice using two different antisolvents: toluene and chloroform.
After synthesis, photoluminescence spectrum analysis is automatically performed every 5 minutes for a period of time. The specific length is appropriate. This requires experiments to determine.
Finally, nonnegative matrix factorization is used to map time- and composition-dependent optoelectronic properties.
By using this workflow for each library, we can find the impact of antisolvent choices on the intrinsic stability of the perovskite.
In fact, it may be a dynamic process. "Chen Yuanguang briefly talked about his idea.
This idea is very similar to the previous directed induced evolution of proteins, but the difference is that experiments are done first, a fixed process is set up through chemical robots, and then machine learning is used to Do analysis.
Bawendi nodded repeatedly after hearing this, thinking that it was indeed a good idea. No wonder Levitt was resentful when he called him: "Great idea, Wright, I support you.
It depends on the final result. If the result is good, it can be classified as nature or science.
If the result is not satisfactory, you can also publish it to a sub-journal of JACS or Nature. For scholars who are lagging behind the times in the field of chemistry, they will still give face to the most popular artificial intelligence algorithms. ”
Chen Yuanguang thought, the main reason is that the Internet is so good. Those who can combine perovskites and machine learning have either not yet grown up. After all, artificial intelligence has only been popular for two years, or they have never come to study for a Ph.D. and have long gone to the industry. Make a lot of money.
Just because this big money is nothing to him, he has the opportunity to study here.
"Professor, in addition, our laboratory does not have a graphics card. I need to buy NVIDIA's latest graphics card to do calculations." Chen Yuanguang mentioned.
Bawendi said helplessly: "It seems that I am also behind the times. To buy, you write an application form to me. After I sign it, you give it to Susan and let her go shopping."
"It's okay. The result is neither good nor bad. It may be a bit reluctant to publish in Nature, but it is more than enough to publish in JACS." Bawendi stared at the results for a long time and then said.
Chen Yuanguang said: "JACS is OK. When I was working on this project, I also worked on another project. This project can definitely be published in Nature."
Bawendi was shocked: "Tell me about it."< br>
Chen Yuanguang: “I built a cross-attribute deep learning framework on GitHub, which can be mainly used for predictive analysis of small material data.
Mainly using some ready-made data from the college for training This model first builds a large data set, and then builds small data sets with different attributes based on this large data set, and their models.
Through the framework extracted from these models, we can go. Directly input the physical properties as its calculation and experimental data sets, and finally get the results of its other properties. ”
To put it bluntly, it is about the topic of perovskite prediction, and I dug a layer deeper and came up with a more general result.
(End of this chapter)