入門引路需口授,功夫無息法自修。

推薦一套科學家訪談錄:

我要說的,都在書裏了。理論研究的特點就是要大量高強度的閱讀,並且歸納總結出自己的風格。

還有美國各學校的暑研計劃、國內各高校的夏令營,比如南京大學理論化學暑期學校:

https://itcc.nju.edu.cn/summer/?

itcc.nju.edu.cn

網上的每年一度理論化學冬季學校:

winterschool.cc?

winterschool.cc

下邊這本書可以簡單科普一下。但要注意,書都有滯後性,以後做理論化學,QFT、ML、tensor network裡邊,不會個一兩招,根本寸步難行了。


本問題下一位匿名前輩另一個角度的觀點也很有見地,像小弟這樣半路出家的民科確實有些過於「眼中只有數值,缺乏對物理/化學的內在本質的理解和追求」(雖然我對他關於究竟什麼是有價值的「物理/化學內在本質」的看法不一定完全認同)。無論理論化學使用多少數學、物理和程序的手段,主要目的還是解決化學問題,如果缺少對化學的理解,確實落了下乘,也是另一種「一葉障目不見泰山」!

Reference:

想要成為一名理論化學家,需要做哪些前期準備??

www.zhihu.com圖標

利益相關:理論化學(電子結構理論)愛好者,目前在做excited-state polarizable embedding QM/MM方法開發以及寫code,離「理論化學家」的水平差了無數個數量級,惶恐謝邀。一家之言,井蛙之見,拋磚引玉,兼收並蓄,希望與各位志同道合的朋友們共勉。

本回答之中的英文部分引自Joe Subotnik課題組主頁上他給出建議的雄文Things to know: Tips for a Career as a Theoretical Chemist包括了對接觸理論化學各個階段的朋友的建議,很切題。這篇文章對小弟幫助很大,尤其是兩三年前濕實驗不成在深夜的淺色牀單下痛哭時,這篇文章讓剛剛按圖索驥地嘗試應用型量子化學計算(需要注意這種通過商業程序進行計算的過程一般來說並不包括在狹義「理論化學」的範圍內,更常用的指代是「計算化學」),對理論化學的世界觀還所知甚少的我接觸到了一些方法論。半個月前跟Joe面試聊天的時候問了他一些非常trivial的技術問題,之後我也提到他的這篇文章給了我很大的幫助,他聽了很高興,然而小弟目前還是沒有收到UPenn的offer(

回到這個問題本身,看到某些其他答主的回答,私以為更有價值的建議也應該更側重具體方法論而非只是general地強調應該多學數學物理計算機(「學而不思則罔,思而不學則殆」)。將理論化學分成三個主要分支(電子結構、動力學、統計力學)是較為傳統的做法(參考JCTC的模塊),三個分支相互之間也有非常多的交叉領域亟待研究/正在研究,但小弟對自己接觸甚少的後兩者還是不予置喙了(另外如Triborg老師所說,QFT、ML、tensor network已成為顯學,傳統功夫雖然源遠流長,但某些方面或許確實滯後。不過無論前沿進展如何,總要紮實地把基礎打好才能看到更深入的東西)。依小弟愚見,對電子結構理論研究而言,完全理想的起步過程應該從掌握線性代數、偏微分方程開始,系統學習量子力學、數學物理方法等數學物理專業核心課程(有前輩提到也應該學習理論力學、統計力學和凝聚態場論,做理論方向的科研不宜偏廢),同時可以閱讀Modern Quantum ChemistryMolecular Electronic-Structure Theory等聖經讀物,通過開源程序如PySCF或PSI4或是完全自己寫各種電子積分來實現量子化學的「Hello world」版SCF,加深對各種量子化學方法的理解。相對而言更為現實的起步過程,則可以根據實際研究需要,參考各種paper,review和聖經教材「哪裡不會點哪裡」。當然,在國內本科化學專業的培養模式下,或許甚至後者都顯得有些可望而不可即,也同樣需要曠日持久的努力,小弟目前也僅是半路出家的民科水平,正在努力實踐後一種起步過程。

不可否認的是,現階段理論化學在國內的發展水平不如國外,方法層面的前沿進展主要發表在JCTC、JCP等期刊上,JPC系列也時不時有一些方法文章,因此除了理論化學相關的技術性問題本身,學習英語和專業英語也尤其重要。在此不給出翻譯一是避免僭越地帶入主觀而引喻失義,二是對英文的掌握也是理論化學研究的一項基本要求。當然,國內外的科研環境有很大區別(無論是否侷限在理論化學領域),Joe也是年輕一代中非常強的rising star,他的經歷或許也並非像小弟這種能力平庸的理論化學愛好者可以企及,就像文中他自己所說,"Note: I make no guarantees. This advice reflects my experiences and what I think would work for my students and postdocs in general: it may not (and likely will not) work in all cases.",還希望各位見仁見智

Things to know: Tips for a Career as a Theoretical Chemist?

subotnikgroup.chem.upenn.edu

Things to know: Tips for a Career as a Theoretical Chemist


Year after year, I have given advice to undergraduates/graduates/postdocs looking to become professors of theoretical chemistry. The advice is not all mine; the majority of the advice below was imparted to me by my advisors (Martin Head-Gordon, Mark Ratner and Abe Nitzan) years ago. Rather than keep this information confidential, for the good of the community, I have decided to share this information broadly on-line. Have a good look: I hope this information will be helpful to those of you lining up to jump into this business. Note: I make no guarantees. This advice reflects my experiences and what I think would work for my students and postdocs in general: it may not (and likely will not) work in all cases.

Undergraduate Studies

If youre interested in theoretical chemistry, one of most important questions you face is, "What classes should I take in college?" Should I take chemistry classes? Physics classes? Math classes? Biology classes? Answer in short: Focus on physics classes. Physics classes are usually taught with more mathematical rigor than chemistry courses, and you will be expected to have a strong mathematical background in graduate school. To that end, as far as mathematics, make sure you take the most rigorous mathematical class you can in linear algebra. Linear algebra is at the heart of all theoretical chemistry and computational engineering, and top graduate programs will expect you to know this basic material. All international students entering graduate school in the USA will certainly have mastered this material.

Applying for Graduate School

When you apply for graduate school, its important to consider a schools reputation--both because that reputation will help you when you graduate, and also because you are likely to learn more from your peers at well-ranked schools. That being said, rankings are not everything and you are also likely to get less attention at a big school and potentially burn out. In the end, the most important indicator of your academic career will be your academic advisor. To that end, make sure you ask your undergraduate advisor: "Where should I apply for graduate schools?" You might be very surprised to learn that schools you were aspiring for are not necessarily the best schools to attend in practice. One small item: when you apply to graduate school, consider very carefully whom you say you want to work with. Those mentioned faculty members will be contacted and decide whether to admit you or not. So dont be lazy and forget to look through every faculty profile!

Accepting a Graduate School

One of the earliest mistakes you can make as a student is delaying your decision to accept one graduate school. If youre a good domestic student, you will likely be accepted to a handful (at least) of graduate schools. Thereafter, the temptation is to delay the painful decision of where to go, perhaps even until the absolute April 15 deadline.

This delay is a mistake. From the perspective of the graduate school, the PIs have a certain budget for the upcoming year and they need a certain number of incoming graduate students to matriculate in order for their research programs to continue. If one department does not believe you are coming, they will adjust their approach and they will begin to work down their waitlist and accept more students. In so doing, there may well be more incoming graduate students with whom you will compete for your top choice of advisor; this may hurt you in the end. It is much better for you to make up your mind early and accept.

Finally, if you have to delay your choice because of a lack of information, be sure to contact the schools under your consideration and tell them you are struggling with a decision. And if they write to you, be sure your write back. No faculty will ever be irreparably angry at you if you turn them down. The only way to make people angry is not to be honest about your intentions.

Picking an Advisor

There are a lot of different considerations when picking a PhD advisor. Do you want to work in a big group, a small group? Do you want to experience a lot of independence or very little independence? There are many advantages to working with a well-established group (assuming that you can survive with very little guidance). There may be networking advantages post graduation, infrastructure advantages, or even just peer advantages (i.e. your labmates will be very well accomplished likely). There is also a lot to be said for working with a younger professor who can give you a lot of attention and point you in the right direction. There is no magic bullet to picking an advisor, and there are many ways to succeed. In picking an advisor, just remember these two facts:

  1. Grad school is a long period of your life and its easy to burn out. Make sure you pick a lab that makes you want to get up in the morning. If you pick the right group, grad school can be the best time of your life: its an academic adventure without any testing. But if you pick the wrong group, grad school can be hell (plain and simple).
  2. Make sure you learn something in graduate school. Its a unique opportunity to spend five years thinking deeply about one particular topic, and it would be a shame to shy away from a difficult research project because you are afraid.

Writing a Paper

Making the Manuscript: The place to begin a paper is almost always with the figures. Decide which figures you want to include in your draft, make the figures, and then write the results section where you simply describe the figures. This is the easiest section to write and should be written first. Thereafter, you can usually finish up the discussion section and conclusions fairly easily. More often than not, the last section you will write will be the first section, the introduction.

In truth, the introduction is a crazy piece of literature because it aims to solve two contradictory goals. On the one hand, you want to introduce the reader to the big questions you are interested in, so you want the writing to be easy to follow for a novice. On the other hand, you want to convince the potential reviewer that you know the relevant literature, making sure you cite everyone in the field, which necessarily requires tailoring your words to fit the historical record. A good introduction can achieve both of these goals at the same time, but this is an art form that can take time to master; learning this skill is not essential for publishing a paper but it will help you in the long run.

Submission Process: Once youve written the manuscript, you can submit the article online very easily. For the first draft, no matter what the journal recommends, most journals will almost always accept a PDF (without either a doc file or a .tex folder).

Revision Process: In an ideal situation, you will get reviews of your manuscript within 2 months of journal submission. These reviews will usually be several paragraphs or text containing itemized lists of comments/questions. When replying to these reviews, just remember to address each of the questions/comments individually. Dont write paragraphs: reviewers dont want to read essays. Instead, just cut and paste the comments, and then respond to each comment individually and succinctly. At this point, you will also need to submit the full .doc or .tex file, with all graphics for publication.

Applying for a Postdoc

The choice of a postdoc advisor is perhaps the most important decision you will make in your academic career. Graduate students who are finishing up their PhDs can be at very different stages of development depending on their PhD labs: some have already learned to pick a research problem, some have not (depending on how much help and interaction there has been with the PhD advisor). Universities are aware of the widely different experiences, and you are expected to complete a postdoctoral fellowship in order to convince your future employer that you can be independent; your successful PhD program was not just dumb luck. If you can do well as a postdoc, you will prove yourself again... and now everyone will know you are talented: no one gets lucky twice.

When choosing a postdoc advisor, you need to look out for two things:

  1. Is the research different enough from your PhD research such that you can make a convincing case (when applying for jobs) that you have two success stories in your pocket? Moreover, will you be able to convince a future university that you are an expert in two different areas whom they should hire?
  2. In considering the group of your future advisor, ask yourself: how much guidance do you want from your future advisor versus how independent do you want to be? And how good are the other grad students/postdocs and will they help you? Note that if you are thinking about a career in teaching, you can take a postdoc at a small school and get some teaching experience, but whether this will help you or not will depend on the particular school you are aiming for.

Applying for Academic Jobs: Assembling your package

Applying for full time jobs is the bottleneck (i.e. choke point) of the academic process. For a job at a research heavy school, you are asking a university to invest at least $300,000 (and often much much more) into your research program. Some postdocs will get jobs, some will not. Luckily, in chemistry, the time frame for figuring out your trajectory should usually be between 2-4 years--unlike the biomedical sciences where postdocs can 8 (EIGHT!) full years. When applying for jobs, you will need to prepare four documents:

  1. A cover letter. For big research schools, this cover letter means just about nothing. It reads something like "Enclosed please find my application for the position of Assistant Professor at your institution. I include a teaching statement, CV, and research statement. Thanks for your consideration. Sincerely, Me." However, for a PUI, you need to put a lot more time into this cover letter. You need to state:
    1. Why you are applying for a job at that particular PUI?
    2. Why you think you would fit into that particular department?
    3. How you think you can help the department to improve? PUIs will look very carefully at this cover letter. By the way, most non-science departments will care about the cover letter as well, even at big universities; but chemistry departments will not care.
  2. A CV. This is straightforward.
  3. A teaching statement. Of course, you want to write something that highlights your writing skills, but the real point of this document is just to alert the hiring committee as to what courses you can teach.
  4. A research statement. Theoretical chemistry is a great field to study in part because the overhead for research is so low. You can change fields very quickly and so, for this reason, you do not need to stress about this research statement as much as you would as an experimentalist. When an experimentalist applies for a job, they will need to justify what equipment they need to buy. For this reason, the department must believe that the experiment is both important and doable, because once the money has been invested, no once can get the money back. By contrast, I think its very rare for a good theorist to actually do what they say they will do. A good theorist will look at the field of theoretical chemistry and pick the best projects possible on the fly, and pick the lowest lying fruit.That being said, the point of the research statement is for the hiring committee to see that you can think big, that you can identify big problems. You dont have to prove that you can solve all of these problems -- that will be investigated during the interview. Instead, for the research statement, just prove you can identify interesting problems. Ideally, in a research statement, you will state how your experiences in grad school and as a postdoc were complementary, giving you a unique background that will allow you to perform research beyond what either of your advisors can achieve. But this is just an ideal situation.

Take these documents and send them out everywhere you think you might want to go. Dont worry about applying too widely: you can always decline an invitation to interview. You never know what youll get. Plus, waiting for interviews is just about the most painful experience known to humankind in academia.

By the way, you might wonder: In this entire submission package, where did I actually describe my past research accomplishments so that I can present myself in the best light possible? Answer, you dont . Thats the job of your letter writers. In picking letter writers, you need three of them, including usually a.) your PhD advisor and b.) your postdoc advisor. Your third letter writer can be very helpful if it is someone well known who can comment on your accomplishments without the appearance of any ulterior motives.

Applying for Academic Jobs: Interviewing

Assuming you are lucky enough to get a job, you will be asked to come for a 1-2 day interview on campus. Now, although you will necessarily be nervous (especially for your first interview), this is actually a terrific opportunity to impress the faculty. Dont blow it. During the course of the interview, you need to give two talks:

  1. The public talk to the whole department, including students. This should be an easy talk to give because you are an expert in your published research, and should know much much more about this research area than any one on the faculty. The point of this talk is really to highlight what your expertise is, and to show that you can effectively communicate and be a good teacher.
  2. The chalk talk (or private talk) is given just to the faculty. This is the hardest talk you will ever give in your life because you need to propose a new line of research for your program, usually following the lines of your research statement. This is the point in the process where you will need to demonstrate that you can actually make progress on the research topics youve chosen and you should be prepared for some difficult questions during the chalk talk.

In general, your goals for this interview are twofold:

  1. To prove to the department that you wrote your research statement. You want to prove that you understand your research, that you were the one who made the major contributions (and that you are not merely your advisors pigeon). Moreover, you want to show your potential colleagues both that you can can get results and that you know a lot theory in general (so you are broad and deep).
  2. To prove to your colleagues that you can help contribute to their research portfolios. Actually, as a funny anecdote, I remember a friend of mine who gave a talk at a university and, at the end of her talk, the first question had nothing to do with her research at all! Instead, for the first question, a faculty member stood up and said "I have accumulated such and such data. How does your data help me explain my research results?" In other words, the focus was not the job candidate but rather the existing faculty member. This is life: its always me first. Your job is to get 7-8 faculty members on your side; if these candidates vote for you to get the job at the faculty meeting, youll get it.

Applying for Academic Jobs: Bargaining

If you are lucky enough to be chosen for a job, you will now need to begin the bargaining process. Of course, you should do the obvious things like ask to see friends offers, speak with your postdoc advisor, etc., to make sure youre getting a reasonable offer. If my experience is universal, though, I think very often that new faculty are afraid to bargain very hard. After all, youre amazed at that point in time that you have a job at all. What you might not realize, though, is that when you are bargaining with the department chair, the chair is usually on your side. At a university, the deans have all of the money usually and so the chair is really there to help you bargain with the dean: the chair (and the department) really wants you to succeed and theyre going to do what they can to help you. Plus, after a long review process, if all of the faculty has actually agreed on a single candidate, the last thing that anyone wants is to look for another candidate. So, in short, dont worry that bargaining is going to make anyone angry or cost you a job. You really need to give it a shot. The job is yours. Its hard to screw it up at this point by asking for too much (unless you are really obscene).

Taking an Extra Year:

After you have secured an academic position, my advice to all incoming faculty is to take a year off before beginning your professorship. Succeeding in academia is getting harder and harder each year, as the problems get more and more difficult to solve. Do yourself a favor and take a year to think about big things in a new place with a different research focus from anything you have seen before. Trust me, it can change your life.


利益相關方,從事這行20多年了。年輕時寫過強相關程序,後來做過計算生物、計算材料,現在主要幹點理論化學和計算物理。平時幫學生對付點小文章畢業,有想法了就讓學生在JACS/Angew和PRL上灌灌水。

下面說點我本人對理論化學現狀的一點看法。理論化學從早期由Langmuir和Lewis起步,再由Pauling把完整的量子理念帶入,最後到Hoffmann的前線軌道,從「道」的角度來說,理論化學到1970-80年代就已經發展完了,現在不管是理論化學,還是計算化學,做的都是「術」的範疇。所以後面發展的DFT/DMRG/QMC,乃至現在的ML/tensor network,解決的都是數值問題,而非化學問題。由此帶來的傾向是,新一代的研究者眼中越來越只有數值,而嚴重缺乏對物理/化學的內在本質的理解和追求,這一點和經過80-90年代學術訓練的上(上)一代研究者完全不同。同時,新一代研究者面臨的新情況是現在大量全新的數學工具被快速app化,使得這些工具被運用到其它領域的時間間隔大大壓縮。譬如對於一個自學能力較強的研究者來說,從ML小白變成專家只需要3-6個月時間,從而大大降低了專家門檻。這樣實際上使得理論化學處於一個非常尷尬的境地:一個新的博士畢業生也許會用各種ML演算法/軟體給出一大堆數據,也會用這些數據分析出一大堆結果(藉助各種軟體工具),並且寫出漂亮的一區論文,但實際上他/她很可能既不理解數學,也不理解化學。所以該領域目前人員嚴重過剩,且很難鑒別優劣,學術圈嚴重「內卷」,急需向產業界轉移。

對於有志於投身理論化學的新入行者,我的建議是盡量學習各種最新方法,但不用死摳每一個細節。在基本理論框架不變的情況下,現在的新技術很可能在5-10年就變成老技術。所以我們需要考慮的是能否建立新的理論框架;如果時機未到,那就用現有框架儘可能考慮如何解決一些業界覺得很難解決的實際問題,從而為新理論框架做準備。


去學物理和數學!!!


前提是你能遇到一個好的導師,否則都是白扯。

我們耳熟能詳的大牛,基本也都是大牛的學生。

比如,卡普拉斯(2013年諾貝爾獎)是鮑林(1954,1962年諾貝爾獎)的學生。卡普拉斯得學生又有klaus schulten等等

國內的唐敖慶(中科院院士,IAQMS院士)手下八大金剛中有5個院士,2個校長。

日本的福井謙一(1981年諾貝爾獎)的學生,諸熊奎治等人。

諸如此類數不勝數,但凡有成就的理論化學家都一定有個好的導師。但也要注意有的導師雖然自己很厲害,但是不一定有時間和精力管你,帶你。導致他的學生也並不能達到特別高的成就。


推薦閱讀:
相關文章